Drug-Induced Liver Disease, Second Edition

Drug-Induced Liver DiseasE

Kaplowitz_978-0849398964_TP.indd1 1

6/6/07 4:00:05 PM

Drug-Induced Liver DiseasE Second Edition

Edited by

Neil Kaplowitz Keck School of Medicine, University of Southern California Los Angeles, California, USA

Laurie D. DeLeve Keck School of Medicine, University of Southern California Los Angeles, California, USA

Kaplowitz_978-0849398964_TP.indd2 2

6/6/07 4:00:05 PM

Informa Healthcare USA, Inc. 52 Vanderbilt Avenue New York, NY 10017 q 2007 by Informa Healthcare USA, Inc. Informa Healthcare is an Informa business No claim to original U.S. Government works Printed in the United States of America on acid-free paper 10 9 8 7 6 5 4 3 2 1 International Standard Book Number-10: 0-8493-9896-7 (hb: alk. paper) International Standard Book Number-13: 978-0-8493-9896-4 (hb: alk. paper) This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials or for the consequence of their use. No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC) 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data Drug-induced liver disease / edited by Neil Kaplowitz, Laurie D. DeLeve. – 2nd ed. p. ; cm. Includes bibliographical references and index. ISBN-13: 978-0-8493-9896-4 (hb : alk. paper) ISBN-10: 0-8493-9896-7 (hb : alk. paper) 1. Hepatotoxicology. 2. Liver–Effect of drugs on. 3. Drugs–Side effects. I. Kaplowitz, Neil. II. DeLeve, Laurie D., 1955[DNLM: 1. Liver Diseases–chemically induced. 2. Liver–drug effects. 3. Preparations–adverse effects. WI 700 D7943 2007] RC848.H48D78 2007 616.3’6–dc22

Visit the Informa Web site at www.informa.com and the Informa Healthcare Web site at www.informahealthcare.com

Pharmaceutical

2007014965

To our loving families and to the memory of Hy Zimmerman, who inspired us with his intellect and dedication in pioneering this field. We are proud to follow in his footsteps.

Preface to the Second Edition

The high productivity of the pharmaceutical industry has provided exciting, efficacious new drugs. However, with efficacy always comes the potential for toxicity, and the growth in new pharmaceuticals has been accompanied by several new drugs linked to liver toxicity. At the same time, epidemiological studies have found that drugs are now the most common causes of liver failure. This has led to a resurgence of interest in drug-induced liver disease in general and has spurred an influx of clinical researchers into this area. Basic research in this area has also thrived over the last decade, due to innovations in biomedical research that have given us tools that have provided new insights into the mechanisms of drug toxicity. The time was therefore ripe to compile a volume with contributions from scientists around the world with expertise in pathogenesis and clinical presentation, as well as authorities on the various categories of drugs and toxins of importance to this field. We have been gratified by the outstanding reviews of the first edition. However, we believe that in this fast-moving field, a book would only remain of value if it can be revised in a timely enough fashion to keep abreast of recent developments. In the second edition, 16 out of the 36 chapters have new authors or cover new topics. New topics include pharmacogenomics and toxicogenomics, causality assessment, risk factors for drug-induced liver disease, management of drug-induced liver disease, and mushroom poisoning. Pharmacogenomics and toxicogenomics are new fields that provide the hope for rational strategies for the pharmaceutical industry in weeding out toxic drugs earlier in development, but may also devise novel approaches to prevent drug toxicity in the susceptible few without exclusion of new drugs efficacious for the many. The new chapters “Causality Assessment” and “Risk Factors for Drug-Induced Liver Disease” expand the Diagnosis and Management section and provide more background on the fundamentals of the field. We have also added a chapter on management to address issues as such liver transplantation and the use of steroids and ursodeoxycholic acid, and we have addressed therapy in individual chapters as well. The second edition is divided into four sections. Section I focuses on mechanisms of hepatotoxicity, often illustrated by examples of specific drugs. The newly expanded Section II reviews general principles of clinical presentation, histopathology, predisposition to toxicity, diagnosis, and management. Each chapter in Section III examines a class of drugs, toxins, or drugs used within a clinical specialty. This section provides a systematic review of the major xenobiotics associated with drug-induced liver disease and also serves as a reference for clinicians dealing with a possible case of drug-induced liver disease. Section IV contains a completely rewritten chapter on drug toxicity from a regulatory perspective. We are pleased with the second edition, which has allowed us to improve and expand coverage of the field and to update the rapidly advancing knowledge of pathogenesis. We believe this volume will be a great value to hepatologists, physicians in all fields of medicine, toxicologists and pharmacologists, and scientists working in preclinical and clinical drug development in both academia and industry. Neil Kaplowitz Laurie D. DeLeve

Preface to the First Edition

With the ever-increasing exposure to pharmaceuticals, more and more examples of druginduced liver disease have been identified in recent years. At the same time, the basic science of hepatic pharmacology, toxicology, and immunology have exploded in the past 5 to 10 years with exciting new developments and insights. We are now poised at the very end of the 20th century with the opportunity to re-evaluate this important topic as we look to the promise of understanding, predicting, preventing, and healing a common problem in clinical medicine that is of importance to all branches of medicine and to anyone who prescribes pharmaceutical or alternative medications. Therefore, the editors believe that an authoritative, up-to-date volume with contributions by experts in basic pathogenesis, clinical pathology, and use of various categories of agents will be of great interest to a broad spectrum of medicine. In this regard, we have drawn upon worldwide expertise with about one-third of the chapters written by authors outside the United States. Innovations in methodology have had a major impact on research in drug-induced liver injury, and this has led to a greater understanding of the mechanisms involved. A few examples should illustrate the progress that has been made and is described in this book. The explosion of information on apoptosis has provided insight into the subtleties of drug-induced cell death. The use of molecular biological techniques has permitted the cloning of numerous genes encoding for P450 isoenzymes. This has made possible the expression of recombinant P450 enzymes and specific P450 antibodies. The availability of recombinant enzymes and specific inhibiting antibodies has facilitated studies to determine the contribution of individual P450 isoenzymes to the metabolism of specific drugs. Until quite recently, cholestasis was thought to be due to either mechanical obstruction of bile flow or cell toxicity that impeded the handling of bile. Improved techniques for isolating membrane vesicles and the cloning and characterization of hepatocyte membrane transporters have allowed the elucidation of a novel mechanism of cholestasis: drug-induced impairment of bile acid transporters in otherwise intact hepatocytes. As more investigators have taken advantage of relatively new methods to isolate pure nonparenchymal cells, there has been a rapid rise in information on the contribution of Kupffer cells, sinusoidal endothelial cells, and stellate cells to a variety of liver diseases, including drug- and toxin-induced liver injury. The concept of the mitochondrion as a major target of drug-induced toxicity was only raised in the early 1980s. Since then, toxicity of an everincreasing number of drugs has been linked to selective toxicity to the mitochondrion. Although reference is made in these examples to chapters on mechanisms in Section I, Section III reiterates many of these processes in the context of individual drugs that have been linked to one of these modes of toxicity. This book has been divided into three major sections. Section I examines hepatotoxicity from the perspective of the mechanisms, across categories of drugs, so that the principles involved can be explored in depth. Examples of drugs to which these mechanisms apply is provided, but the main focus is on the mechanism. Because the authors are experts who are writing about the current state-of-the-art in their own field, this information is useful to both clinicians who want to gain understanding of the fundamental principles as we understand them today, as well as to knowledgeable clinicians and investigators who wish to read about the newest advances. Section II provides a general outline of the clinical presentation, histopathology, and management of drug-induced hepatotoxicity. Chapter 12 systematically reviews the clinical

viii

Preface to the First Edition

presentation and pathological picture of the types of liver injury that can be induced by drugs and toxins. Chapter 14 reviews the factors that predispose an individual to drug toxicity, suggests strategies for monitoring patients at risk for toxicity, and provides information on preventive measures. The information provided in this section provides a basic framework for any clinician who might be confronted with xenobiotic-induced hepatotoxicity. Section III systematically reviews specific toxins implicated in drug-induced hepatotoxicity. Each chapter examines the toxicity induced by drugs or toxins within a specific pharmacological class or by drugs used within a clinical specialty. The current understanding of the mechanism of toxicity, risk factors for developing toxicity, histological characteristics, clinical manifestations, and management are discussed for each category of drugs. This section is of value to gastroenterologists and hepatologists who want a systematic review of druginduced liver disease. It also serves as a reference for clinicians in a variety of specialties who are confronted with a patient with liver disease that might be attributable to drug therapy. Neil Kaplowitz Laurie D. DeLeve

Contents

Preface to the Second Edition Preface to the First Edition Contributors

v vii

xiii

PART I. MECHANISMS OF LIVER INJURY 1.

Drug-Induced Liver Disease Neil Kaplowitz

1

2.

Cytochrome P450 Activation of Toxins and Hepatotoxicity F. Peter Guengerich

3.

Antioxidant Defense in Liver Injury: Oxidant Stress, Antioxidant Defense, and Liver Injury 33 Hartmut Jaeschke

4.

Hepatotoxicity Due to Mitochondrial Injury 49 Dominique Pessayre, Bernard Fromenty, Abdellah Mansouri, and Alain Berson

5.

Mechanisms of Cell Death and Relevance to Drug Hepatotoxicity Neil Kaplowitz

6.

Significance of Hepatobiliary Transporters for Drug-Induced Liver Disease Peter J. Meier and Christiane Pauli-Magnus

7.

Immunological Mechanisms in Drug-Induced Liver Injury Dwain L. Thiele

8.

Mechanistic Role of Acyl Glucuronides 125 Hilde Spahn-Langguth, Chunze Li, and Leslie Z. Benet

9.

Nonparenchymal Cells, Inflammatory Macrophages, and Hepatotoxicity Debra L. Laskin and Carol R. Gardner

10. Role of Tissue Repair in Liver Injury Harihara M. Mehendale

13

207

PART II. DIAGNOSIS AND MANAGEMENT 12. Clinicopathological Patterns of Drug-Induced Liver Disease Willis C. Maddrey

97

115

185

11. Genetic Susceptibility to Drug-Induced Liver Disease Mark Russo and Paul B. Watkins

85

223

159

x

Contents

13. Histopathology of Drug-Induced Liver Disease Gary C. Kanel 14. Risk Factors for Drug-Induced Liver Disease Laurie D. DeLeve

237 291

15. Genomics, Proteomics, and Metabolomics in the Diagnosis and Mechanisms of Drug-Induced Liver Disease 307 Andrew A. Stolz 16. Causality Assessment 325 Ma Isabel Lucena, Rau´l J. Andrade, Raquel Camargo, and Miren Garcı´a-Corte´s 17. Management of the Patient with Drug-Induced Liver Disease Thomas D. Boyer

345

PART III. HEPATOTOXICITY OF SPECIFIC DRUGS 18. Mechanisms of Acetaminophen-Induced Liver Disease Sidney D. Nelson and Sam A. Bruschi

353

19. Acetaminophen: Pathology and Clinical Presentation of Hepatotoxicity William M. Lee and George Ostapowicz

389

20. Mechanisms Underlying the Hepatotoxicity of Nonsteroidal Anti-inflammatory Drugs 407 Urs A. Boelsterli 21. Nonsteroidal Anti-inflammatory Drugs and Leukotriene Receptor Antagonists: Pathology and Clinical Presentation of Hepatotoxicity 439 James H. Lewis 22. Mechanism, Pathology, and Clinical Presentation of Hepatotoxicity of Anesthetic Agents 465 J. Gerald Kenna 23. Anticonvulsant Agents 485 Munir Pirmohamed and Steven J. Leeder 24. Hepatotoxicity of Psychotropic Drugs and Drugs of Abuse Dominique Larrey 25. Antibacterial and Antifungal Agents Richard H. Moseley

527

26. Hepatotoxicity of Antituberculosis Drugs Sumita Verma and Neil Kaplowitz 27. Hepatic Toxicity of Antiviral Agents Ulrich Spengler

507

547

567

28. Hepatotoxicity of Cardiovascular and Antidiabetic Medications Sidharth S. Bhardwaj and Naga P. Chalasani 29. Cancer Chemotherapy Laurie D. DeLeve

631

593

xi

Contents

30. Hepatotoxicity of Immunomodulating Agents and the Transplant Situation Timothy J. Davern 31. Methotrexate Controversies Adrian Reuben

663

683

32. Adverse Effects of Hormones and Hormone Antagonists on the Liver Shivakumar Chitturi and Geoffrey C. Farrell

707

33. Mushroom Poisoning: A Clinical Model of Toxin-Induced Centrilobular Necrosis Franc¸ois Durand and Dominique Valla 34. Hepatotoxicity of Herbal Medicines, Vitamins, and Natural Hepatotoxins Lawrence U. Liu and Thomas D. Schiano 35. Occupational and Environmental Hepatotoxicity Keith G. Tolman and Anthony S. Dalpiaz PART IV. REGULATORY PERSPECTIVES 36. Regulatory Perspectives John R. Senior Index

789

771

755

733

723

Contributors

Rau´l J. Andrade

Liver Unit, Hospital Virgen de la Victoria, Ma´laga, Spain

Leslie Z. Benet Department of Biopharmaceutical Sciences, University of California-San Francisco School of Pharmacy, San Francisco, California, U.S.A. Alain Berson E´quipe Mitochondries, INSERM, U773, Centre de Recherche Biome´dicale Bichat Beaujon, Faculte´ de Me´decine Xavier Bichat, Universite´ Paris 7 Denis Diderot, Paris, France Sidharth S. Bhardwaj Division of Gastroenterology and Hepatology, Indiana University School of Medicine, Indianapolis, Indiana, U.S.A. Urs A. Boelsterli Department of Pharmaceutical Sciences, School of Pharmacy, University of Connecticut, Storrs, Connecticut, U.S.A. Thomas D. Boyer Liver Research Institute, University of Arizona, Tucson, Arizona, U.S.A. Sam A. Bruschi Department of Medicinal Chemistry, University of Washington School of Pharmacy, Seattle, Washington, U.S.A. Raquel Camargo

Liver Unit, Hospital Virgen de la Victoria, Ma´laga, Spain

Naga P. Chalasani Division of Gastroenterology and Hepatology, Indiana University School of Medicine, Indianapolis, Indiana, U.S.A. Shivakumar Chitturi Department of Gastroenterology and Hepatology, Australian National University Medical School at the Canberra Hospital, Australian Capital Territory, Australia Anthony S. Dalpiaz Division of Gastroenterology, University of Utah School of Medicine, Salt Lake City, Utah, U.S.A. Timothy J. Davern Gastroenterology Division and Liver Transplant Program, University of California, San Francisco, California, U.S.A. Laurie D. DeLeve Keck School of Medicine, University of Southern California, Los Angeles, California, U.S.A. Franc¸ois Durand Service d’He´patologie, Hospital Beaujon, Clichy, France Geoffrey C. Farrell Department of Gastroenterology and Hepatology, Australian National University Medical School at the Canberra Hospital, Australian Capital Territory, Australia Bernard Fromenty E´quipe Mitochondries, INSERM, U773, Centre de Recherche Biome´dicale Bichat Beaujon, Faculte´ de Me´decine Xavier Bichat, Universite´ Paris 7 Denis Diderot, Paris, France Miren Garcı´a-Corte´s Liver Unit, Hospital Virgen de la Victoria, Ma´laga, Spain

xiv

Contributors

Carol R. Gardner Department of Pharmacology and Toxicology, Rutgers University, Piscataway, New Jersey, U.S.A. F. Peter Guengerich Department of Biochemistry and Center in Molecular Toxicology, Vanderbilt University School of Medicine, Nashville, Tennessee, U.S.A. Hartmut Jaeschke Department of Pharmacology, Toxicology, and Therapeutics, University of Kansas Medical Center, Kansas City, Kansas, U.S.A. Gary C. Kanel Department of Pathology, Keck School of Medicine, University of Southern California, Los Angeles, California, U.S.A. Neil Kaplowitz Keck School of Medicine, University of Southern California, Los Angeles, California, U.S.A. J. Gerald Kenna

AstraZeneca Safety Assessment, R&D Alderley Park, Macclesfield, Cheshire, U.K.

Dominique Larrey Service d’He´pato-Gastroente´rologie et Transplantation, Hoˆpital Saint Eloi, Montpellier, France Debra L. Laskin Department of Pharmacology and Toxicology, Rutgers University, Piscataway, New Jersey, U.S.A. William M. Lee

University of Texas Southwestern Medical Center, Dallas, Texas, U.S.A.

Steven J. Leeder Division of Pediatric Pharmacology and Medical Toxicology, Children’s Mercy Hospital and Clinics, Kansas City, Missouri, U.S.A. James H. Lewis Chunze Li

Georgetown University Medical Center, Washington, D.C., U.S.A.

Merck & Co., Inc., West Point, Pennsylvania, U.S.A.

Lawrence U. Liu Division of Liver Diseases and Recanati/Miller Transplantation Institute, The Mount Sinai Medical Center, New York, New York, U.S.A. Ma Isabel Lucena Departmento de Farmacologia Clinica, Hospital Virgen de la Victoria, School of Medicine, Ma´laga, Spain Willis C. Maddrey Department of Internal Medicine, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas, U.S.A. Abdellah Mansouri E´quipe Mitochondries, INSERM, U773, Centre de Recherche Biome´dicale Bichat Beaujon, Faculte´ de Me´decine Xavier Bichat, Universite´ Paris 7 Denis Diderot, Paris, France Harihara M. Mehendale Department of Toxicology, College of Pharmacy, University of Louisiana at Monroe, Monroe, Louisiana, U.S.A. Peter J. Meier

University of Basel, Basel, Switzerland

Richard H. Moseley Ann Arbor Veterans Affairs Healthcare System and Department of Internal Medicine, University of Michigan Health System, Ann Arbor, Michigan, U.S.A. Sidney D. Nelson Department of Medicinal Chemistry, University of Washington School of Pharmacy, Seattle, Washington, U.S.A.

xv

Contributors

George Ostapowicz

University of Texas Southwestern Medical Center, Dallas, Texas, U.S.A.

Christiane Pauli-Magnus

University Hospital Basel, Basel, Switzerland

Dominique Pessayre E´quipe Mitochondries, INSERM, U773, Centre de Recherche Biome´dicale Bichat Beaujon, Faculte´ de Me´decine Xavier Bichat, Universite´ Paris 7 Denis Diderot, Paris, France Munir Pirmohamed Department of Pharmacology and Therapeutics, University of Liverpool, Liverpool, U.K. Adrian Reuben Division of Gastroenterology and Hepatology, Medical University of South Carolina, Charleston, South Carolina, U.S.A. Mark Russo Department of Medicine, University of North Carolina, Chapel Hill, North Carolina, U.S.A. Thomas D. Schiano Division of Liver Diseases and Recanati/Miller Transplantation Institute, The Mount Sinai Medical Center, New York, New York, U.S.A. John R. Senior Office of Surveillance and Epidemiology, Center for Drug Evaluation and Research, Food and Drug Administration, Silver Spring, Maryland, U.S.A. Hilde Spahn-Langguth

German University in Cairo, New Cairo City, Egypt

Ulrich Spengler Department of General Internal Medicine, Rheinische Friedrich Wilhelms Universita¨t Bonn, Bonn, Germany Andrew A. Stolz Division of Gastrointestinal and Liver Diseases, Department of Medicine, Keck School of Medicine, University of Southern California, Los Angeles, California, U.S.A. Dwain L. Thiele Division of Digestive and Liver Diseases, Department of Internal Medicine, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas, U.S.A. Keith G. Tolman Division of Gastroenterology, University of Utah School of Medicine, Salt Lake City, Utah, U.S.A. Dominique Valla

Service d’He´patologie, Hopital Beaujon, Clichy, France

Sumita Verma Keck School of Medicine, University of Southern California, Los Angeles, California, U.S.A. Paul B. Watkins Department of Medicine, University of North Carolina, Chapel Hill, North Carolina, U.S.A.

PART I

1

MECHANISMS OF LIVER INJURY

Drug-Induced Liver Disease Neil Kaplowitz

Keck School of Medicine, University of Southern California, Los Angeles, California, U.S.A.

INTRODUCTION The goal of this chapter is to provide a broad overview of the subject of this book and to introduce a number of concepts that will be expanded upon in subsequent chapters. Drug-induced liver disease represents an important problem for the following major reasons: (1) approximately 1000 drugs have been implicated in causing liver disease at least on rare occasions (1); (2) in the United States drug-induced liver disease is the most common cause of acute liver failure, accounting for one-third to one-half of cases (2,3); although acetaminophen accounts for the bulk of these, other drugs are still a more frequent cause of acute liver failure than viral hepatitis and other causes (4); (3) in addition, drug-induced liver disease represents an important diagnostic/therapeutic challenge for physicians caring for patients presenting with liver disorders, since it can mimic all forms of acute or chronic liver disease (5); the frequency and economic impact of this problem is a major challenge for the pharmaceutical industry and regulatory bodies, especially since the toxic potential of some drugs is not evident in preclinical and phase 1 to 3 clinical testing. The incidence of drug-induced liver injury is not well established in the general population. In a population-based cohort study in France the incidence was 14 cases per 100,000 (0.014%) inhabitants (6), whereas an inpatient study from Switzerland found a higher incidence (1.4%) (7). CLINICAL OVERVIEW Drug-induced liver diseases can mimic all forms of acute and chronic hepatobiliary diseases (Table 1) (5,8). However, the predominant clinical presentations resemble acute icteric hepatitis (hepatocellular jaundice) or cholestatic liver disease. The former is of grave significance as the mortality approximates 10% irrespective of the specific drug (1,5,9,10). This is referred to as Hy’s Law after the late Hy Zimmerman, who noted that mortality from drug-induced hepatocellular jaundice ranged from 10% to 50%. Hy Zimmerman also noted that in most cases at risk for fatal outcome, aside from jaundice alanine aminotransferase (ALT) and aspartate aminotransferase were between 8 and 100! upper limit of normal (ULN) and alkaline phosphatase (Alk. Ptase) !3!ULN. Over the years the validity of Hy’s Law has held up in specific examples (Table 2) and is further verified in recent registries (9–11). This type of reaction is accompanied by systemic symptoms, jaundice, markedly elevated serum transaminases, ALT!ULN/Alk. Ptase.!ULNR5, and in the more severe cases, coagulopathy and encephalopathy indicative of acute (fulminant) liver failure. It is noteworthy that the height of the transaminases does not reliably predict severity except perhaps in the case of acute intrinsic toxins, e.g., acetaminophen. Cholestatic disease, although not usually life threatening, presents with jaundice, disproportionate increased Alk. Ptase, ALT!ULN/Alk. Ptase.!ULN%2, and pruritus; cholestatic reactions tend to resolve very slowly (i.e., months vs. weeks for hepatitis) The author of this chapter has relationships with the following corporations: Abbott, Adams Respiratory Therapy, Allergan, Amgen, Astra Zeneca, Avera, BG Medicine, Biogen, Boehringer/Ingelhem, Cadence, Daiichi Sankyo, DOV, Elan, Enanta, Encysive, GSK, Gtx, Incyte, ISIS, Janssen, Johnson & Johnson, Maxygen, Merck, Millenium, Ono, Pfizer, Rigel, Roche, Sankyo, TAP, Threshold, Teva, and Wyeth.

2

Kaplowitz

TABLE 1 Spectrum of Hepatic Manifestations of Drug-Induced Liver Disease Acute hepatitis Chronic hepatitisa Acute cholestasis Mixed hepatitis/cholestasis or atypical hepatitis Chronic cholestasisa Non-alcoholic steatohepatitis Fibrosis/cirrhosis Microvesicular fatty liver Veno-occlusive disease Peliosis hepatitis Adenoma and hepatocellular carcinoma a

Acetaminophen, isoniazid, troglitazone, bromfenac Nitrofurantoin, methyldopa, diclofenac, minocycline, dantrolene Amoxicillin clavulanic acid, erythromycins, sulindac, chlorpromazine, angiotensin-converting enzyme inhibitors Phenytoin, sulfonamides Chlorpromazine, numerous others on rare occasion Amiodarone, tamoxifen Methotrexate Valproic acid, nucleoside reverse transcriptase inhibitors Busulfan, cyclophosphamide Azathioprine, hormones Hormones

Drugs that cause chronic disease more frequently cause acute disease.

and on rare occasion lead to vanishing bile duct disease and biliary cirrhosis (12,13). Mixed injury patterns with intermediate ALT/Alk. Ptase. can resemble atypical hepatitis or granulomatous hepatitis. Individual drugs tend to exhibit a consistent pattern or clinicopathological signature of the reaction (Table 1) with characteristic latency and clinical presentation. However, some drugs may show several patterns: e.g., nimesulide can cause a short-latency, hypersensitivity-mediated cholestatic injury and a delayed idiosyncratic acute hepatitis-like reaction (14). Thus, although one pattern may predominate, crossover to other patterns is not unusual. Drug-induced liver disease can be predictable (high incidence and dose-related) or unpredictable (low incidence and may or may not be dose-related). Unpredictable reactions, also referred to as idiosyncratic, can be viewed as either immune-mediated hypersensitivity or nonimmune reactions. Most potent predictable hepatotoxins are recognized in the animal testing or clinical phase of drug development. Those that slip through are almost always unpredictable. Latency between the initiation of therapy and the onset of liver disease is a component of the signature of reactions to specific drugs and provides some clues as to the pathogenesis. Early onset within a few days (particularly if no previous exposure) is strong evidence for direct toxicity of the drug or its metabolite, which is characteristic of predictable reactions; acetaminophen overdose is an example (15). Unpredictable reactions manifested as overt or symptomatic disease usually occur with intermediate (one to eight weeks) or long latency (up to 12 months). Intermediate latency is characteristic of hypersensitivity reactions, but can be seen with nonimmune idiosyncrasy as well. The hypersensitivity reactions tend to be associated with fever, rash, and eosinophilia and a rapid positive rechallenge (5,8). Hepatotoxicity of sulindac (16), phenytoin (17), and amoxicillin–clavulanic acid (18) are typical examples. Most cases of cholestatic liver injury and chronic hepatitis caused by drugs are of the hypersensitivity type. It is important to recognize that these reactions may occur up to three to four weeks after a one to two week course of medication (e.g., amoxicillin–clavulanic acid). In contrast, the long-latency type of idiosyncratic reaction is characteristically not associated with features of hypersensitivity and the response to rechallenge is variable and delayed. Thus, one assumes that these events reflect TABLE 2 Drugs that Cause Hepatocellular Jaundice and Confirm Hy’s Law 1978 Iproniazid Isoniazid Phenytoin Halothane Cinchophen Dantrolene Nitrofurantoin

Later Phenylbutazone Ketoconazole Ticrynafen Valproic acid Enflurane Pemoline Labetalol Diclofenac Sulindac

1978 refers to first edition of Zimmerman’s textbook.

Probable Bromfenac Troglitazone Trovafloxacin Nefazodone

3

Drug-Induced Liver Disease

TABLE 3 Drugs Associated with Idiosyncratic Hepatitis Nonallergic Acarbose Benoxaprofenb Bosentan Bromfenacb Dantrolene Diclofenac Disulfiram Felbamate Flutamideb Isoniazid Isotretinoin Ketoconazole Labetalol a b

Allergic Leflunomide Nefazodoneb Nevirapine Pemoline Pyrazinamide Terbinafinea Tolcapone Troglitazoneb Trovafloxacinb Valproic acid Zafirlukast Zileuton

Allopurinol Diclofenac Dihydralazine Halothane Methyldopa Minocycline Nevirapine Nitrofurantoin Phenytoin Propylthiouracil ACE inhibitorsa Augmentinwa Phenothiazinesa

Erythromycinsa Sulfonamidesa Sulindaca Tricyclicsa

Cholestatic/mixed. Withdrawn from the market.

some type of late-onset change in the metabolism of the drug or the response to injury (repair or regeneration). Drugs associated with variable, long latency include isoniazid (19) and troglitazone (20). This type of idiosyncratic reaction is extremely challenging with respect to understanding the pathogenesis and predicting the problem in individual cases. Table 3 provides a list of drugs that are associated with idiosyncratic allergic and nonallergic reactions. A few can cause either allergic or nonallergic reactions, e.g., diclofenac and nevirapine. Low-frequency unpredictable reactions, either immune-mediated or not, often occur on a background, higher rate of mild, asymptomatic, and usually transient liver injury, which is detected as abnormal biochemical tests, particularly serum ALT. Generally, the biochemical abnormality defined as ALTO3!ULN may occur 10 to 20 times more frequently than overt disease. In almost all instances, the ALT returns to normal despite continued drug use. Thus, in the majority of patients with increased ALT some type of adaptation or “tolerance” occurs and in the minority progression to overt, severe injury occurs, which may reflect a failure to adapt. This issue is further complicated by the uncertain explanation for the very long latency in some of the idiosyncratic reactions. It should be emphasized that acute or chronic hepatitis induced by drugs subsides upon discontinuation of the drug without long-term sequelae with rare exception. A few reported cases of autoimmune hepatitis triggered by hypersensitivity drug reactions have continued on without the drug, but it is questionable as to whether this was drug-induced liver disease or underlying autoimmune chronic hepatitis. Scarring may persist after severe subacute or chronic injury but is of little consequence after removal of the drug cholestalic reactions are cholestatic reactions are not infrequently associated with loss of interlobular bile ducts. However, the development of cirrhosis or effects on longevity are exceedingly rare. PATHOGENESIS Hepatotoxicity of drugs can be principally metabolism-dependent, parent drug-dependent, or a combination of both (Fig. 1). Metabolism takes place largely in the liver, which accounts for its susceptibility to drug-induced injury (8). The metabolites may be electrophilic chemicals or free radicals that deplete glutathione (GSH), covalently bind to proteins, lipids, or nucleic acids, or induce lipid peroxidation. The consequences include hepatocellular necrosis, apoptosis, or sensitization to cytokines or inflammatory mediators produced by nonparenchymal cells. Alternatively, the reactive metabolites may covalently bind to or alter liver proteins such as cytochrome P450s (CYPs) leading to sensitization and immune-mediated injury. The immune phenomena nevertheless are metabolism dependent. Thus, the rare occurrence of immunemediated liver disease is often superimposed on a higher frequency of mild injury (abnormal ALT) suggesting that the drug has a mild toxic potential (e.g., phenytoin or halothane) but in rare individuals this toxic potential leads to more severe toxicity initiated by metabolism steps,

4

Kaplowitz

Haptenization Immune response Metabolite ” Covalent binding anger “D GSH depletion Reactive O2

Drug Mitochondria DNA Toxicity (mild)

Adaptation

Inflammatory/ toxic mediators

Recovery

Repair

Overt liver injury

FIGURE 1 Pathogenesis of drug-induced liver diseases. Upstream events in the hepatocytes affect viability of individual cells but sensitize to downstream processes leading to clinically overt organ damage. The latter involves a balance of effects of cytokines, chemokines, and inflammatory mediators, mainly produced by nonparenchymal cells and the effects on repair processes a such as regeneration.

but also heavily influenced by genetic and/or environmental factors that determine either an immune response or idiosyncratic reaction. Genetic polymorphisms of enzymes involving drug activation or detoxification have been implicated in the susceptibility to hypersensitivity reactions to sulfonamides (21,22), anticonvulsants (17,23), and tacrine (24). Presumably genetic polymorphisms of either major histocompatibility complex (MHC) I-dependent antigen presentation in hepatocytes or MHC II-dependent antigen presentation in macrophages, which have scavenged necrotic or apoptotic hepatocytes directly killed by the drug, may further contribute to determine the rare occurrence of these hypersensitivity reactions (25) which most often have an incidence of 1:1000 or less. Parent drug-dependent toxicity occurs as a result of the properties of the parent drug (or metabolite) to accumulate in organelles [weak bases such as amiodarone accumulate in mitochondria (26)], undergo nonspecific redox cycling (quinones cycle electrons from NADPH to O2 generating (OK 2 )), or specifically inhibit enzymes or transporters (nucleoside reverse transcriptase inhibitors block mitochondrial DNA polymerase (27)) or cyclosporin A inhibits canalicular transporters (28). In these cases, if the parent drug’s chemical properties account for direct toxicity, factors that enhance its availability (decreased metabolism or export) may increase susceptibility. Regardless of whether toxicity within a target liver cell (e.g., hepatocyte, sinusoidal endothelial cell, or bile duct cell) is parent drug- or metabolite-dependent, the ultimate severity of the liver disease in vivo may depend greatly on the subsequent downstream participation of toxic mediators released from various cell types and the recruitment of inflammatory cells as well as intracellular and tissue repair and regenerative responses. The toxic mediators include chemicals, such as NO and reactive oxygen metabolites, and the balance of cytokines that promote injury [e.g., tumor necrosis factor-a (TNFa), Interleukin-1 (IL-1), Interferon-g (IFNg), Interleukin-12 (IL-12), Interleukin-18 (IL-18)), or prevent injury (IL-4, IL-10, IL-13, monocyte chemotactic protein-1 (MCP-1)]. Thus, toxin may somewhat injure hepatocytes but then sensitize to the effects of an imbalance in injurious versus protective cytokines (Fig. 2). For example, the toxicity of carbon tetrachloride (CCl4) is abrogated in vivo by neutralizing TNF (29); the toxicity of acetaminophen is markedly enhanced in MCP-1 chemokine receptor Protection IL-4 IL-10 IL-13 MCP-1 Toxin

Injury IL-12 TNFα IFNγ IL-12

TNFα Sensitize hepatocytes IFNγ T(–) MCP-1 IL-10, 13

Lethal

FIGURE 2 Role of cytokine balance in determining susceptibility to toxins. Drugs or metabolites may directly injure hepatocytes to a minor extent, but may markedly sensitize to the lethal effects of TNFa and IFNg. The latter are modulated by cytokines that promote or inhibit their production or actions. Abbreviations: TNFa, tumor necrosis factor a; IFNg, interferon g.

5

Drug-Induced Liver Disease

Hepatocyte injury

Toxin HGF

Antiapoptosis

Regeneration

Acute liver Pro- or anti- failure apoptosis

IL-6

FIGURE 3 Role of cytokines and regeneration in toxin-induced liver disease. TNF activates NF-kB in nonparenchymal cells, leading to IL-6 production, and in hepatocytes, or with the regenerative response, worsening of the overt liver injury.

TNFα

LPS

Kupffer cell activation

knockouts associated with an enhanced TNFa response (30) and may be abrogated by inactivating Kupffer cells (31), although this is controversial. Similarly IL-10 null mice are sensitized to acetaminophen (32), whereas natural killer/natural killer cell with T cell receptor cell depletion, knockout of IFNg, and Fas/Fas ligand deficiency protect against acetaminophen (33). Thus, the direct and indirect influence of toxins on the production and balance of mediators, and genetic polymorphisms in these responses may play a major role in unmasking the overt toxic potential of a drug culminating in overt idiosyncratic toxicity. Another important factor that contributes to the extent of liver injury is the capacity of the liver to regenerate (Fig. 3). Thus, for example, TNFa will promote regeneration by acting on Kupffer cells (autocrine) to release IL-6, which will trigger, along with hepatocyte growth factor, regeneration. Interference with IL-6 (knockout) worsens CCl4 injury, and conversely, exogenous IL-6 treatment diminishes liver injury in wild-type mice (34). TNF also acts on hepatocytes through NF-kB signaling to promote survival gene transcription. If the toxin interferes with the latter pathway, TNFa-induced apoptosis may occur.

RISK FACTORS Regardless of whether hepatotoxicity is predictable (frequent) or unpredictable (rare), hypersensitivity-mediated or idiosyncratic, metabolism dependent or parent drug-dependent, the interplay of genetic and environmental risk factors influences susceptibility (Fig. 4) (35). Age, gender, concomitant drugs, and underlying diseases (e.g., hepatitis C virus, hepatitis

Toxic potential of drug - Reactive metabolite - Acylglucuronide - Mitochondrial effects

Genetic factors - Drug metabolism - Detoxification - Transport - Others

Environmental factors - Other drugs - Ethanol - Age - Underlying disease

FIGURE 4 Risk of hepatotoxicity. The ultimate development of hepatotoxicity is determined by the interplay of the toxic potential of the drug or its metabolites and the susceptibility of the host as determined by genetic and environmental factors, both of which influence gene expression.

6

Kaplowitz

TABLE 4 Risk Factors for Drug-Induced Hepatotoxicity Drug Methotrexate Isoniazid Acetaminophen Valproate Diclofenac Anticonvulsants Sulfonamides

Factors Chronic alcohol, obesity, diabetes, chronic hepatitis, psoriasis HBV, HCV, HIV, alcohol, older age, female, slow acetylator, rifampin, pyrazinamide Chronic alcohol, fasting, isoniazid Young age, anticonvulsants, genetic defects of mitochondrial b-oxidation and respiratory chain enzymes Female, osteoarthritis Genetic defect in detoxification HIV, slow acetylator, genetic defect in defense

Abbreviations: HBV, hepatitis B virus; HCV, hepatitis C virus.

B virus, HIV) have been most frequently identified. Table 4 lists examples of drugs and associated risk factors. With the advent of new technologies in genomics and proteomics, one can anticipate that new insights into the mechanisms of susceptibility and liver injury from drugs will be forthcoming (36). Some of the genetic factors to consider are listed in Table 5. Polymorphisms (CYPs, cytokines, MHC, etc.) and rare heterozygous mutations (b-oxidation, bile salt export pump) will need to be assessed.

DIAGNOSIS Establishing a diagnosis of drug-induced liver disease in an individual case is mainly based upon circumstantial evidence aided by the signature type of reactions (if known) with respect to latency and clinical characteristics as well as exclusion of other more plausible alternative causes. Additional information can be gained from the response to removal of the drug—rapid improvement in cytotoxic reactions and slow improvement in cholestatic reactions. A rechallenge with recrudescence of liver abnormalities is the most definitive evidence, but hardly ever justified and not always positive in idiosyncratic cases. A practical approach is to consider the diagnosis probable/possible if the signature latency and pattern of disease fit and other causes are excluded (viral hepatitis, ischemic hepatitis, biliary disease). The remaining cases are unlikely or unrelated depending on the completeness of the workup and the strength of the evidence in favor of an alternative diagnosis. This ad hoc approach is equivalent to diagnosing as yes, no, or maybe. The presence of autoantibodies to specific forms of CYP has been associated with hypersensitivity reactions to certain drugs (25,37,38). Although uncertain but intriguing significance with respect to pathophysiology, their presence may be helpful in the diagnosis of drug-induced liver disease in these special cases (Table 6). However, testing for these autoantibodies is mainly a research tool at present. Furthermore, sensitivity and specificity of the presence of these autoantibodies is uncertain and autoantibodies to certain CYPs have been TABLE 5 Possible Genetic Determinants of Risk 1. Drug metabolism (e.g., CYP polymorphisms) 2. Detoxification (e.g., GSH-related, epoxide hydrase) 3. Apoptosis and survival genes 4. Signal transduction (kinases and phosphatases) 5. MHC I and II 6. Cytokines/chemokines and receptors 7. Inflammatory mediators (Cox, NOS, etc.) 8. Regeneration/repair 9. Transporters (BSEP, MRP2, etc.) 10. Mitochondrial b-oxidation and respiratory chain 11. Structural integrity (e.g., cytokeratin 8/18) Abbreviations: Cox, cyclooxygenase; NOS, nitric oxide synthase; BSEP, bile salt export pump; MRP2, multidrug resistance associated protein family; MHC, major histocompatibility complex; CYP, cytochrome P450s; GSH, glutathione.

7

Drug-Induced Liver Disease

TABLE 6 Autoantibodies in Drug-Induced Liver Disease Autoantibody target a

CYP 2C9 CYP 1A2b CYP 3Ab CYP 2E1 mEH

Drug Tienilic acid Dihydralazine Anticonvulsants Halothane Germander

a

Also referred to as anli-LKM2 autoantibody. Also referred to as anti-LM autoantibody. Abbreviations: CYP, cytochrome P450; mEH, microsomal epoxide hydrolase.

b

observed in patients taking drugs without evidence of hepatotoxicity, limiting their diagnostic value. Several groups have attempted to generate quantitative systems designed to generate a numerical score that reflects the probability of a drug as the cause for liver disease (39–42). The Roussel-Uclaf casuality assessment method scoring system appears to be the most accurate (43,44) and puts numerical weight on the factors discussed above (Table 7) to generate a composite score that reflects the probability that liver injury is drug-induced. The advantage is that this system is less subjective than the ad hoc approach. This type of scoring system performs well when validated against well-documented cases of drug-induced liver disease. Specialists, the pharmaceutical industry, and regulatory bodies should be encouraged to use this scale. It also would be reasonable to apply the scoring system to individual case reports submitted to medical journals. Although it is not perfect and may not discriminate between multiple concurrently used candidate toxins, it does provide consistency and focuses the attention of the evaluator on most of the critical parameters that need to be considered in estimating the probability of causality. DRUG DEVELOPMENT Drug development involves a preclinical and clinical phase. Preclinical assessment is centered on animal testing using very high doses. Although animal testing is probably very reliable in screening out potent, predictable toxins, it is far less reliable in identifying a propensity for unpredictable toxicity. Experience has suggested that there are numerous false positives and false negatives. A better understanding of the mechanisms of unpredictable hepatotoxicity may eventually lead to the establishment of appropriate animal models that recapitulate the factors that determine susceptibility in humans. This will most likely be achieved with the use of transgenic and knockout mice to set up conditions that mimic human susceptibility. At present, some suggestion of hepatotoxicity at high doses in animals may at least warrant more careful or extensive assessment in the clinical phases of drug development. During phases I to III of drug testing the likelihood of encountering overt hepatotoxicity (i.e., jaundice and high transaminases) depends on the frequency of the reaction. Most idiosyncratic reactions occur in 1 in 1000 or more individuals and acute liver failure in 1 in 10,000 or more. Typically in clinical development, drugs are tested in 1500 to 2500 individuals. Exclusion of an overt reaction with 95% confidence, if the true incidence is 1:1000, would require 3000 treated patients, assuming all were exposed for the appropriate duration (e.g., six to nine months). This is usually not attained in clinical trials. Therefore, if one is not fortunate TABLE 7 Causality Assessment 1. 2. 3. 4. 5. 6. 7.

Latency Rate of resolution (dechallenge) Risk factors (age, alcohol, pregnancy) Exclusion of other causes (viral hepatitis, ischemia, biliary tract disease, alcohol) Concomitant drugs Track record (PDR, case reports) Rechallenge

Abbreviation: PDR, Physicians Desk Reference.

8

Kaplowitz

enough to identify overt hepatitis in one or two cases in the study population, the appearance of lesser signals needs to be the focus for scrutiny. It is very rare indeed to identify acute liver failure from idiosyncratic hepatotoxins in drug development. The rule of threes indicates that to identify (not miss) acute liver failure with 95% confidence that has a true incidence of 1 in 10,000 would require 30,000 study patients. Since the probability of identifying overt or life-threatening liver injury in clinical trials is so low, one must focus on the incidence of asymptomatic ALT and bilirubin elevations (45). The most sensitive parameter appears to be the incidence of ALTO3!ULN in drug-treated versus placebo control-treated patients. Depending on the study population, the incidence of 3!ALT in controls may vary from 0.1% to nearly 1.0%. Thus, an incidence of 2% to 3% or greater in the drug-treated patients would be unequivocal cause for closer scrutiny. Although this is a sensitive indicator, it is not entirely specific since there are drugs, e.g., statins, tacrine, aspirin, etc., that are associated with an increased incidence of 3!ALT, but have proved relatively safe in postmarketing experience (?false-positive signal). More specificity is gained by examining the height of the transaminases. An ALT increase of eightfold or greater is a more specific signal since this rarely occurs in controls. Even more specific is conjugated hyperbilirubinemia (R1.5-fold) associated with elevated ALT. The experience with troglitazone exemplifies the issue of identification of a signal. In a cohort of 2500 study patients, ALTO3!ULN occurred in nearly 2% (vs. !1% in controls); ALTO8!ULN occurred in 0.6% (vs. none of controls); and two cases with overt jaundice were observed (46). Thus, all the criteria for a hepatic signal were present at the time of approval of the drug. A similar premarketing experience was observed with bromfenac (47), which was also withdrawn postmarketing. A critical issue is what is the appropriate regulatory response to the occurrence of a signal? In some cases, particularly when the drug is not crucial, approval is denied. If the drug is critical, warnings and education of physicians and patients are very important and there may be justification for recommending monitoring ALT (see below) and/or restrictions on the use of the drug. Two recent examples are bosentan and ximelagatran, which had unequivocal signals for idiosyncratic hepatotoxicity, perhaps stronger than troglitazone (Table 8). Bosentan was approved with monitoring because the disease being treated is fatal, i.e., pulmonary hypertension, and no other efficacious treatment is available, and therefore a favorable risk-benefit analysis. Ximelagatran, a thrombin inhibitor, was not approved because several cases of very severe hepatotoxicity were seen and alternative anticoagulants are available. A number of antituberculous, HIV and cancer drugs have remained on the market because benefit outweighs risks. POSTMARKETING MONITORING The background incidence of drug-induced mild, reversible liver injury provides the rationale for monitoring or surveillance. From this background of mild injury a minority of individuals will emerge with overt disease. Thus, by stopping the medications at the first sign of mild injury one should prevent serious consequences. Although this seems a logical approach, a number of problems must be considered. First, the approach applies only to delayed reactions.

TABLE 8 Comparison of Idiosyncratic Hepatotoxins in Clinical Trials Troglitazone nZ ALTO3!ULN ALTO10!ULN ALTO3! and bili O2! ALTO8! and bili O3! Fatal acute liver failure

Bosentan

2500 1.8% 0.6%

658 12 14% 2 7%

0.08% 0

0.3% 0

a 3 cases (one confounded by HBV). Abbreviations: ALT, alanine aminotransferase; ULN, upper limit of normal; HBV, hepatitis B virus.

Ximelagatran 6948 7.8% 1.9% 0.5% 0.1% 0.01 0.04%a

Drug-Induced Liver Disease

9

Hypersensitivity reactions occur relatively early and evolve rapidly so educating patients about symptoms is crucial in early cessation of the offending drug. Second, one is sacrificing potentially very important therapy to a much larger number of patients than would actually develop overt disease; third, compliance with such approaches is known to be very poor; fourth, the rate of development of overt disease from the first appearance of elevated ALT needs to be gradual for monthly monitoring to be efficacious in preventing life-threatening disease. Testing more frequently than monthly is not practical, although the future development of a fingerstick ALT test that could be applied in a fashion similar to monitoring glucose might change this by improving compliance and allowing more frequent monitoring. In any case, monthly monitoring for delayed idiosyncratic reactions is the best approach available, but the efficacy of the approach is assumed and not proven. Furthermore, this should not substitute for the need to educate patients about symptoms of hepatotoxicity, such as fever, rash, malaise, fatigue, anorexia, gastrointestinal complaints, abdominal pain, dark urine; jaundice, pruritus, etc., and the need to report them to the physicians to ensure expeditious cessation of offending agents. Despite monthly monitoring, some adverse events may appear rapidly in the few weeks after a normal test. Ultimately, the most difficult challenge to the application of monitoring is cost-effectiveness—monthly monitoring is expensive and one needs to weigh this quantitatively against the morbidity and mortality of adverse liver events. There is no clear answer to the question of what incidence of serious adverse idiosyncratic hepatitis warrants monitoring, how frequently it should be performed, and for how long. Furthermore, the postmarketing occurrence of adverse events must be weighed against the benefit of the drug. Risk/benefit assessment is ill defined but ultimately becomes the crucial factor in recommending monitoring ALT versus removal of the drug from the market. In the case of the nonsteroidal anti-inflammatory drug, bromfenac, continued use of the drug in the face of infrequent, delayed idiosyncratic severe hepatotoxicity (47) could not be justified since many alternative treatments were available. In the case of troglitazone, the decision to withdraw was delayed and more complicated owing to the important and unique therapeutic properties of the drug in managing a serious medical condition, albeit with benefits that would not be evident for many years (i.e., the effect of long-term control of blood sugar on complications of diabetes). It was decided that the implementation of monthly monitoring would likely protect the users and the drug was continued. Although this strategy may have worked to some extent, the issue of compliance with monitoring and the possibility of occasional “rapid risers” meant that the population could not be completely protected. At the same time, several new drugs in this class were approved and after a year of postmarketing experience with the alternative new agents, it was concluded that these agents were much less likely to induce severe hepatotoxicity, leading to the withdrawal of troglitazone. A major issue in postmarketing surveillance is the frequency of serious adverse events compared to the background in the general population. Reliable data on the occurrence of hospitalization for idiopathic hepatitis and acute liver failure are very limited. However, several databases (Medicaid, HMO) suggest that hospitalization for cryptic acute hepatitis occurs in about 1:50,000 to 1:100,000 adult individuals in the general population each year (48–50). Acute liver failure is estimated to occur in up to 3000 individuals annually in the United States. About 10% to 20% of these cases are idiopathic with a resulting annual incidence of one to two cases per million individuals in the population. Thus, when a new drug is marketed and more than a few cases of unexplained acute liver failure are reported, concern should be raised. In the example of troglitazone, at least 30 to 40 cases of acute liver failure were reported to the FDA in the first year on the market from a population of about one million taking the drug. This was a much higher rate than predicted. Such a postmarketing signal may be less apparent when far fewer individuals are exposed. However, the current MedWatch system for reporting adverse events, with all its attendant problems regarding poor compliance and accurate causality assessment, has been reasonably successful in rapidly identifying problems with numerous drugs leading to withdrawal or severe restrictions of use.

10

Kaplowitz

CONCLUSIONS The liver is a particular target for drugs because of its role in clearing and metabolizing chemicals. The parent drug or more frequently the metabolites may either affect critical functions, or sensitize to the effects of cytokines or inflammatory cells, or elicit an immune response. This often occurs in an unpredictable fashion, implying that environmental and genetic factors alter the susceptibility to these adverse events. A wide range of liver diseases can occur as adverse events but the individual drug tends to induce a characteristic signature reaction with respect to latency and clinicopathological manifestations. Hepatotoxicity from drugs poses a major challenge in drug development and postmarketing surveillance. The future identification of the pathogenesis of idiosyncratic reactions represents the major challenge in this field and will likely advance rapidly with the application of methods of toxicogenomics and pharmacogenomics in the preclinical and clinical arenas. REFERENCES 1. Zimmerman H. Drug Hepatotoxicity. 2nd ed. Philadelphia, PA: Lippincott, 1999. 2. Ostapowicz G, Fontana RB, Larson AM, et al. Etiology and outcome of acute liver failure in the USA: preliminary results of a prospective multi-center study (abstr). Hepatology 1999; 30(4):221A. 3. Shakil A, Kramer D, Mazariegos G, et al. Acute liver failure: clinical features, outcome analysis, and applicability of prognostic criteria. Liver Transplant 2000; 16:163–9. 4. Larsen A, Polson J, Fontana R, et al. Acetaminophen-induced acute liver failure: results of a U.S. Multicenter, prospective study. Hepatology 2005; 42:1364–72. 5. Zimmerman H. Drug-induced liver disease. In: Schiff E, Sorrell M, Maddrey W, eds. Schiff’s Diseases of the Liver. 8th ed. Philadelphia, PA: Lippincott-Raven, 1999:973–1064. 6. Sgro C, Clinard F, Ouazir K, et al. Incidence of drug-induced hepatic injuries: a French populationbased study. Hepatology 2002; 36:451–5. 7. Meier Y, Cavallaro M, Roos M, et al. Incidence of drug-induced liver injury in medical inpatients. Eur J Clin Pharmacol 2005; 61:135–43. 8. Kaplowitz N. Drug metabolism and hepatotoxicity. In: Kaplowitz N, ed. Liver and Biliary Diseases. 2nd ed. Baltimore, MD: Williams & Wilkins, 1996:103–20. 9. Bjornsson E, Olsson R. Outcome and prognostic markers in severe drug-induced liver disease. Hepatology 2005; 42:481–9. 10. Andrade R, Lucena M, Fernandez M, et al. Drug-induced liver injury: an analysis of 461 incidences submitted to the Spanish registry over a 10-year period. Gastroenterology 2005; 129:512–21. 11. Lewis JH. “Hy’s law”, the “Rezulin rule” and other predictors of severe drug-induced hepatotoxicity: putting risk-benefit perspective. Pharmacoepidemiol Drug Saf 2006; 15:221–9. 12. Desmet VJ. Vanishing bile duct syndrome in drug-induced liver disease. J Hepatol 1997; 26:31–5. 13. Degott C, Feldmann G, Larrey D, et al. Drug-induced prolonged cholestasis in adults: a histological semiquantitative study demonstrating progressive ductopenia. Hepatology 1992; 15:244–51. 14. Van Steenberen W, Peeters P, DeBondt J, et al. Nimesulide-induced acute hepatitis: evidence from six cases. J Hepatol 1998; 29:135–41. 15. Pham T-V, Lu S, Kaplowitz N. Acetaminophen hepatotoxicity. In: Taylor MB, ed. Gastrointestinal Emergencies. 2nd ed Baltimore, MD: Williams & Wilkins, 1997:371–88. 16. Tarazi E, Harter JG, Zimmerman HJ, et al. Sulindac-associated hepatic injury. Analysis of 91 cases reported to the food and drug administration. Gastroenterology 1993; 104:569–74. 17. Shear N, Spielberg S. Anticonvulsant hypersensitivity syndrome: in vitro assessment of risk. J Clin Invest 1988; 82:1826–32. 18. Larrey D, Vial T, Micaleff A, et al. Hepatitis associated with amoxycillin-clavulanic acid combination. Report of 15 cases. Gut 1992; 33:368–71. 19. Thompson N, Caplin M, Hamilton M, et al. Anti-tuberculosis medication and the liver: dangers and recommendations in management. Eur Respir J 1995; 8:1384–8. 20. Murphy E, Davern T, Shakil O, et al. Troglitazone-induced fulminant hepatic failure. Dig Dis Sci 2000; 45:549–53. 21. Rieder M, Uetrecht J, Shear N, et al. Diagnosis of sulfonamide hypersensitivity reactions by in vitro “rechallenge” with hydroxylamine metabolites. Ann Intern Med 1999; 110:286–9. 22. Rieder M, Shear N, Kanee A, et al. Prominence of slow acetylator phenotype among patients with sulfonamide hypersensitivity reactions. Clin Pharmacol Ther 1991; 49:13–7. 23. Gennis M, Vemusi R, Burns E, et al. Familial occurrence of hypersensitivity to phenytoin. Am J Med 1991; 91:631–4. 24. Becquemont L, Lecoeur S, Simon T, et al. Glutathione S-transferase q genetic polymorphism might influence tacrine in Alzheimer’s patients. Pharmacogenetics 1997; 7:251–3.

Drug-Induced Liver Disease

11

25. Robin M, LeRoy M, Descatoire V, Pessayre D. Plasma membrane cytochromes P450 as neoantigens and autoimmune targets in drug-induced hepatitis. J Hepatol 1997; 26(1):23–30. 26. Berson A, DeBeco V, Letteron P, et al. Steatohepatitis-inducing drugs cause mitochondrial dysfunction and lipid peroxidation in rat hepatocytes. Gastroenterology 1998; 114:764–74. 27. Brinkman K, Hofstede H, Burger D, et al. Adverse effects of reverse transcriptase inhibitors: mitochondrial toxicity as common pathway. AIDS 1998; 12:1735–44. 28. Kowdley K, Keefe E. Hepatotoxicity of transplant immunosuppressive agents. Gastroenterol Clin North Am 1995; 24:991–1001. 29. Czaja M, Xu J, Alt E. Prevention of carbon tetrachloride-induced rat liver injury by soluble tumor necrosis factor receptor. Gastroenterology 1995; 108:1849–54. 30. Hogaboam C, Bone-Larson C, Steinhauser M, et al. Exaggerated hepatic injury due to acetaminophen challenge in mice lacking C–C chemokine receptor 2. Am J Pathol 2000; 156:1245–52. 31. Laskin D, Gardner C, Price V, Jollow D. Modulation of macrophage functioning abrogates the acute hepatotoxicity of acetaminophen. Hepatology 1995; 21:1045–50. 32. Bourdi M, Masubuchi Y, Reilly TP, et al. Protection against acetaminophen induced liver injury and lethality by interleukin 10: role of inducible nitric oxide synthase. Hepatology 2002; 35:289–98. 33. Liu Z-X, Govindarajan S, Kaplowitz N. Innate immune system plays a critical role in determining the progression and severity of acetaminophen hepatotoxicity. Gastroenterology 2004; 127:1760–74. 34. Cressman P, Greenbaum L, DeAngelis R, et al. Liver failure and defective hepatocyte regeneration in interleukin-6-deficient mice. Science 1996; 274:1379–83. 35. DeLeve L, Kaplowitz N. Prevention and therapy of drug-induced hepatic injury. In: Wolfe M, ed. Therapy of Digestive Disorders Philadelphia, PA: WB Saunders/Harcourt Brace & Company, 2000:334–48. 36. Nuwayser E, Bittner M, Trent J, et al. Microarrays and toxicology: the advent of toxicogenomics. Mol Carcinog 1999; 24:153–9. 37. Beaune P, Lecoeur S. Immunotoxicology of the liver: adverse reactions to drugs. J Hepatol 1997; 26(Suppl. 2):37–42. 38. Neuberger J. Immune mechanisms in drag hepatotoxicity. Clin Liver Dis 1998; 2:471–82. 39. Maria V, Victorino R. Development and validation of a clinical scale for the diagnosis of drug-induced hepatitis. Hepatology 1997; 26:664–9. 40. Benichou C. Criteria of drug induced liver disorders: report of an international consensus meeting. J Hepatol 1990; 11:272–6. 41. Danan G, Benichou C. Causality assessment of adverse reactions to drugs I. A novel method based on the conclusions of international consensus meetings: application to drug-induced liver injuries. J Clin Epidemiol 1993; 46:1323–30. 42. Benichou C, Danan G, Flahault A. Causality assessment of adverse reactions to drugs II. An original model or validation of drug causality assessment methods: case reports with positive rechallenge. J Clin Epidemiol 1993; 46:1331–6. 43. Lucena M, Camargo R, Andrade R, et al. Comparison of two clinical scales for causality assessment in hepatotoxicity. Hepatology 2001; 33:123–30. 44. Kaplowitz N. Causality assessment versus guilt by association in drug hepatotoxicity. Hepatology 2001; 33:308–10. 45. Kaplowitz N. Idiosyncratic drug hepatotoxicity. Nat Rev Drug Discov 2005; 4:489–99. 46. Watkins P, Whitcomb R. Hepatic dysfunction associated with troglitazone. N Engl J Med 1998; 338:916–7. 47. Moses P, Schroeder B, Alkhatib O, et al. Severe hepatotoxicity associated with bromfenac sodium. Am J Gastroenterol 1999; 94:1393–6. 48. Walker A, Cavanaugh R. The occurrence of new hepatic disorders in a defined population. Post Mark Surveill 1992; 6:107–17. 49. Carson J, Strom B, Duff A, et al. Safety of nonsteroidal anti-inflammatory drugs with respect to acute liver disease. Arch Intern Med 1993; 153:1331–6. 50. Dun M-S, Walker A, Kronlund K. Descriptive epidemiology of acute liver enzyme abnormalities in the general population of central Massachusetts. Pharmacoepidemiol Drug Saf 1999; 8:275–83.

2

Cytochrome P450 Activation of Toxins and Hepatotoxicity F. Peter Guengerich

Department of Biochemistry and Center in Molecular Toxicology, Vanderbilt University School of Medicine, Nashville, Tennessee, U.S.A.

INTRODUCTION The history of bioactivation of chemicals predates cytochrome P450 (P450) research, beginning with structure-activity studies of carcinogens by Fieser (1) and biochemical experiments by Millers (2,3). The enzymatic activation of inert chemicals by oxidation or other processes can generate electrophilic products that react with nucleophilic sites on DNA and proteins. The concept was extended to drugs and other chemicals and can provide an explanation for toxicity, following the classic studies of the Gillette and Brodie laboratories with acetaminophen and bromobenzene in the 1970s (4,5). In the case of carcinogenesis, causality can be demonstrated at least to the extent that mutagenicity can be demonstrated with defined DNA lesions (6). As discussed later, the evidence linking toxicities with covalent binding is considerable but must still be considered correlative (7,8). P450 research began with these and other studies in steroid metabolism (9,10) and microbial biochemistry (11). Much of the field is still centered in these areas. Drug metabolism is one area in which basic knowledge about P450 has had tremendous practical impact. Some of the key findings in the development of the field were the demonstration that the hemoprotein P450 is the component of the system that functions as the terminal oxidase [i.e., actually reacts O2 with the substrate (12)], the separation and reconstitution of the components of the microsomal system (13), the discovery that P450 is inducible (14,15), the demonstration of multiple P450s (16), and the purification (17), sequence analysis (18), and crystallization (19) of mammalian P450s. P450 ENZYMES The human genome contains 57 P450s, and these are named systematically based on sequence similarity (20,21). The sequence of a P450 does not provide direct insight into catalytic function, however. One approach to categorizing the human P450s is shown in Table 1. Some assignments are equivocal (e.g., 1B1 and 27A1), but the P450s involved in the metabolism of steroids, eicosanoids, and fat-soluble vitamins are clearly important in development and normal physiology. The P450s in the “xenobiotics” column are not considered critical but these are the enzymes involved in drug metabolism (22). In considering metabolism of all commercial drugs, w75% is attributed to P450s (Fig. 1A). Within this group, w95% is catalyzed by five P450s (Fig. 1B). Thus, the determination of which P450s are involved in the metabolism of a drug can usually be done quickly, and this information is expected for new drug candidates by the U.S. Food and Drug Administration (FDA). Determining which human P450s are involved in bioactivation and detoxication reactions can also be readily established in in vitro systems (25). Methods also exist to put The author of this chapter has relationships with the following corporations. Stock holdings: Abbot Laboratories, Glaxo SmithKline, and Orchid Biosciences. Consulting: Merck Research Laboratories, Schering-Plough Research Institute, Targacept, Biogen-Idec ProCor, Otsuka Pharmaceuticals, BioPure, and ThermoFisher Scientific.

14

Guengerich

TABLE 1 Classification of Human P450s Based on Major Substrate Class Sterols

Xenobiotics

Fatty acids

1A1 1A2 2A6 2A13 2B6 2C8 2C9 2C18 2C19 2D6 2E1 2F1 3A4 3A5 3A7

2J2 4A11 4B1 4F12

1B1 7A1 7B1 8B1 11A1 11B1 11B2 17A1 19A1 21A2 27A1 39A1 46A1 51A1

Eicosanoids 4F2 4F3 4F8 5A1 8A1

Vitamins 2R1 24A1 26A1 26B1 26C1 27B1

Unknown 2A7 2S1 2U1 2W1 3A43 4A22 4F11 4F22 4V2 4X1 4Z1 20A1 27C1

Source: From Refs. 22, 23.

in vitro findings into the context of human in vivo situations, although this is obviously more difficult in the case of bioactivation. Humans vary considerably in terms of amounts of each P450 and catalytic activity (Fig. 2), and some of the reasons will be discussed in the next section. This variability can be the basis of unexpected pharmacokinetics (Fig. 3). Ideally, the same pharmacokinetic pattern (Fig. 3) would be observed in all individuals, with the increase and decrease in plasma levels (and presumably levels of drug and metabolites in each tissue) being the same. However, an individual with a deficit of the P450 needed to metabolize the drug would show a response such as that shown in Figure 3 after receiving the same doses. Alternatively, unexpectedly rapid metabolism and low levels of the drug could result in therapeutic effectiveness. Further, if a P450 converts the drug to a toxic metabolite, the patient could be at higher risk. At this point, it is useful to consider differences between experimental animal models and humans. As discussed later, pharmacodynamic differences are often seen and these can influence drug efficacy and also toxicology. The choice of an appropriate test species is also important in considerations of metabolism and pharmacokinetics. The P450s in humans and laboratory mammals have similarity but are not identical, even when they have the same P450 name. As an example, the case of P450 1A2 is presented, where the rat and human enzymes have very similar activities toward the model substrate 7-methoxyresorufin, but human P450 1A2 has 10-fold higher catalytic efficiency than rat P450 1A2 in the activation of some FMO

UGT

NAT

1A1

MAO

1A2

Esterases

2B6

2C9

3A4(+5) P450 2D6

(A)

(B)

2C19

2E1

FIGURE 1 (A) Contribution of individual enzymes in (human) drug metabolism. (B) Contribution of individual P450 enzymes in drug metabolism. Abbreviations: UGT, UDP-glucuronosyl transferase; FMO, flavin-containing monooxygenase; NAT, N-acetyltransferase; MAO, monoamine oxidase. Source: Adapted from Ref. 24.

15

Cytochrome P450 Activation of Toxins

(A)

Nifedipine oxidation

0.2

10

5

8 6 4 2

0

(B)

0

39 98 100 107 108 109 110 112 115 116 118 123 126 129 130 132 133 134

0.4

P450 3A4

P450 2E1

39 98 100 107 108 109 110 112 115 116 118 123 126 129 130 132 133 134

chlorzoxazone 6-hydroxylation

P450 1A2

0

10

15

39 98 100 107 108 109 110 112 115 116 118 123 126 129 130 132 133 134

7-Ethoxyresorutin O-deethylation

0.6

(C)

FIGURE 2 Variation of activities of three P450s in human liver samples. A set of 18 organ donor samples was used to prepare liver microsomes and the indicated catalytic activities were measured (units of nmol product/formed/nmol total P450) (numbers indicate sample numbers). (A) P450 1A2, (B) P450 2E1, and (C) P450 3A4. Source: Adapted from Ref. 22.

heterocyclic amines to genotoxic products (26). One of the first exercises usually done with a new drug candidate is in vitro comparisons of metabolism to determine which animal species most resembles humans. The issue of metabolites that are unique to humans and not seen in toxicology test species is a matter of current concern with pharmaceutical companies and the FDA (27). Another point is that the FDA expects to have toxicology test animals exposed to multiples of both the parent drug and major human metabolites, and the choice of a species to use for testing can be a problem if metabolism is very different. Regulation of P450s Variation in P450 activities can be due to several factors. One issue is genetic polymorphism that control expression and/or activity, which are treated in chapter 11. The other issues are enzyme induction and inhibition, which underlie many of the drug–drug interactions attributed to pharmacokinetics. P450 Inhibition Many examples of P450 inhibition are known and contribute to drug interaction problems. Because of this, pharmaceutical companies screen new drug candidates for inhibition in vitro. At an early stage, the throughput can be at a high volume. The results are used (1) to select

Extensive Metabolizer (EM, normal)

C

Plasma level of drug

p,max

AUC

Poor Metabolizer (PM)

Time (arrows show repeated doses)

FIGURE 3 Effect of variability of a P450 (or other enzyme) on pharmacokinetics of a drug. Both the EM and PM patients received the same dose of the drug at each time interval (indicated with the arrow). The drug accumulates in the PM patient. Abbreviations: EM, extensive metabolizer; PM, poor metabolizer. Source: Adapted from Ref. 22.

16

Guengerich

among a series of candidates for those that should be less likely to cause problems and (2) to rationally decide which in vivo interaction studies will be most important. On the basis of the mechanism, P450 inhibition may be divided into two categories: (1) The first is competitive inhibition, the competition of two chemicals for a site on the same P450. Although several complex models are possible (28), the analysis is relatively straightforward and the information can be applied directly to pharmacokinetic equations. The inhibition is reversible. (2) The second type is mechanism-based, or “suicide,” inhibition, which is irreversible. In the course of catalytic action on the inhibitor (which is actually a substrate), at least a fraction of the reaction cycles results in inactivation of the enzyme (29). This process has the effect of lowering the concentration of the P450. Examples include components of oral contraceptives (30,31) and components of grapefruit juice (32). P450 Induction New drug candidates are often screened with in vitro induction assays in many pharmaceutical companies. Two major issues are of concern. The first is that a rough correlation among compounds exists between the inducibility of some of the P450s and liver tumors found in rodent cancer bioassays. This pattern is seen with P450s in the 1A, 2B, and 4A subfamilies. However, many exceptions to the trend exist, and the relevance of the rodent tumors to human cancer is questionable, in light of the successful use of omeprazole (P450 1A inducer) (33), the lack of tumors in individuals on long-term barbiturate therapy (P450 2B inducer) (34), and the limited responses to peroxisome proliferators (P450 4A inducer) in humans (35). The point should also be made that the P450 induction is not causal (of tumors) in the rodent models. The other reason to study induction is the potential for drug–drug interactions. Induction of a P450 by a drug can lead to changes in pharmacokinetics as a function of time, if the drug is metabolized by the induced P450. In addition, induction of a P450 by one drug can lead to increased metabolism of another. Classic examples are the early work of Remmer (14) and the loss of efficacy of oral contraceptives after use of barbiturates, rifampicin, or St. John’s wort (30,36). Although several mechanisms of P450 induction are possible and have been characterized in experimental systems, the most common mode of induction involves transcriptional regulation. A generalized scheme is shown in Figure 4. In general, a ligand is bound to a cytosolic receptor, which then forms a heterodimer with another protein in the nucleus and

R' L +

L R

R

L R

R'

(movement to nucleus?)

Coactiv L R DNA

R'

P450 gene

L R

R'

P450 gene

DNA

RNA pol

(increased access to promoter, start site)

Increased transcription

FIGURE 4 General scheme of P450 induction. A ligand L binds to a receptor R. The LR complex then interacts with another protein, R 0 , to yield a heterodimer. In the nucleus, this complex binds to a specific enhancer sequence, possibly with the assistance of a co-activator, and changes the chromatin structure to facilitate transcription by RNA polymerase. Source: Adapted from Ref. 22.

Cytochrome P450 Activation of Toxins

17

binds to a specific DNA sequence upstream of the P450-coding region. The DNA region is what is called an “enhancer” element. In some cases, a co-activator protein is also bound to the complex. The net effect of the binding of these proteins is to rearrange the structure of the gene, in the chromatin, to open the DNA promoter region, increasing the access for RNA polymerase and the rate of transcription of the P450 gene. Brief descriptions of the major characterized systems are provided here. The point should be made that some genes respond to multiple regulatory systems. The family 1 P450s are regulated by the aryl hydrocarbon receptor (AhR). A variety of polycyclic hydrocarbons and related materials are ligands, including some compounds in cigarette smoke. Binding of a ligand to the AhR involves the release of AhR from heat shock protein 90 and dimer formation with the AhR nuclear transferase protein (37). This heterodimer binds to xenobiotic regulatory element sites. This response explains the interactions of certain drugs with cigarette smoking, ingestion of charbroiled food, and use of certain drugs (e.g., omeprazole). The P450 2B subfamily genes are inducible by barbiturates, related drugs, and a variety of certain other chemicals (e.g., the pesticide DDTand the herbal medicine St. John’s wort). Induction involves the dimerization of the constitutive androgen receptor (CAR) with one of the retinoid X receptor (RXR) forms and interaction with an element far upstream of the start site (38). This system is still enigmatic in the initial response. Some artificial ligands do bind to CAR [e.g., “TCBOPOP” (39)] but apparently barbiturates and many other typical inducers do not (40). P450 3A subfamily enzymes respond to a variety of compounds, including steroids, macrolide antibiotics, and a variety of drugs. These inducers are ligands for the pregnane X receptor (PXR) which, after ligand binding, dimerizes with an RXR protein and then binds to a defined element on the P450 gene and, with the assistance of hepatic nuclear factor (HNF)-4a and possibly other co-activators (41,42), enhances transcription. The last case presented is that of the P450 4A subfamily enzymes [in humans, this consists of only P450 4A11 (and possibly 4A22), but experimental animals have multiple 4A genes]. The ligands, which can be drugs, fatty acids, or other chemicals (35), bind to the peroxisome proliferator–activated receptor a (PPARa). Liganded PPARa dimerizes with an RXR and the complex interacts with an element on the DNA. This dimer also involves the transcription of other genes and leads to the proliferation of peroxisomes, up to 25% of the cell volume, in rodents. The point should be made that CAR, PXR, PPAR, and RXR are all members of the so-called steroid receptor superfamily, which also includes the steroid hormone receptors, glucocorticoid receptor, vitamin D receptor, and retinoic acid receptor (43). The dimerization with RXR is important in that the interaction and the response are functions of the particular form of RXR involved and also its (retinoid) ligand. As mentioned earlier, several of these systems involve co-activators, e.g., HNF-4a in the case of P450 3A4, so use of cell lines with the appropriate factors is critical in mimicking the in vivo response (and also explaining the basis of tissue specificity). The point should also be made that other receptors in this family, e.g., liver X recptor and farnesoid X receptor, regulate the steroidogenic P450s (44), and some of these systems may be involved in “cross talk” with the regulatory components of the xenobioticmetabolizing P450s. Cell-based models can be used to test compounds for the ability to induce P450s, either by measuring P450 induction directly or using a reporter-based system, provided that the cells have the appropriate factors. Another consideration in using experimental animals to predict induction responses is that some species differences exist among the receptors. For instance, the response of PXR in rats, rabbits, and humans varies considerably with the ligands rifampicin and pregnenolone 16a-carbonitrile (41), and the results can be understood in the context of a few amino acid substitutions (45).

CONTEXTS OF TOXICITY One way to consider drug toxicity is classification with regard to general mechanism. Thus, it is possible to divide toxicities into five categories [Table 2; (46); see also chap. 5].

18

Guengerich

TABLE 2 Contexts of Drug Toxicity On-target Hypersensitivity and immune reactions Off-target Bioactivation to reactive intermediates Idiosyncratic Source: From Ref. 46.

The first of the five contexts, on-target or mechanism-based toxicity, is relatively straightforward to deal with. An example is the statins. Most of the toxicities can be explained by the intended action of the parent drug on 3-hydroxymethylglutaryl-CoA reductase, but in a different cell (e.g., muscle). The toxicity can be reduced by lowering the dose or, in this case, by administration of mevalonate, the product of the targeted enzyme. The second context is due to immune reactions resulting from modification of a protein by the parent drug. An example is penicillins and related b-lactams. This process is part of a “hapten” hypothesis (47). To expand this, there is a “danger” hypothesis (48), basically that the drug–macromolecule complex generates a “danger” signal that ultimately results in antibody/ T-cell responses (47). A further twist on this is the “p-i concept” of “direct pharmacological interaction of drugs with immune receptors” (47,49) (see chap. 7). The third case is off-target pharmacology. In this case, the parent drug interacts with additional targets that were not anticipated and yields unexpected pharmacology, which can result in toxicity. A relevant example is terfenadine, an antihistamine that was recalled in 1995 (Fig. 5). In this case, P450 inhibition is also an issue. Normally, terfenadine is rapidly oxidized by P450 3A4 to fexofenadine (50). Both terfenadine and fexofenadine bind to the H1 receptor and have antihistaminic activity. In most individuals, the oxidation is rapid and no terfenadine is found in plasma. However, individuals who use P450 3A4 inhibitors concurrently (e.g., erythromycin, ketoconazole) have lowered P450 3A4 activity and terfenadine accumulates.

Erythromycin Ketoconazole H+ N

CH3

CH3 Terfenadine

CH3

X P450 3A4

CH2OH

+ H+ N

Dehydrogenases or P450

CH3

H+ N

CH3 CH3

H+ N

CH3 –

CO 2 CH3

Fexofenadine

CH3 CH3

Arrhythmia, QT intervals Antagonists of H1 receptor (antihistamines)

Zwitterionic; does not cross blood/brain barrier

FIGURE 5 Terfenadine as an example of off-target toxicity and a role for metabolism. Terfenadine is extensively oxidized in three steps to a carboxylic acid (fexofenadine). Both the terfenadine and the product fexofenadine affect the target, the H1 receptor, and have antihistaminic activity. Terfenadine is normally not found in plasma following oral ingestion because of rapid metabolism. However, inhibitors of P450 3A4 block oxidation and cause terfenadine to accumulate. Terfenadine has off-target pharmacology and interacts with hERG receptors to produce cardiac problems; in some cases, abnormal QT intervals and fatal arrhythmias have been attributed to the increased terfenadine load. Abbreviation: hERG, human ether-ago-go related gene. Source: From Ref. 50.

19

Cytochrome P450 Activation of Toxins

O

O

CH3

HN

HN

CH3

O HN

O – SO3

OH P450s (2E1/1A2/3A4)

O HN

CH3

O gluc

Acetaminophen

CH3

HN

+ O

OH CH3

O

OH

N

CH3

SG GSH

OH

Protein O Iminoquinone

O HN

CH3

Cys-protein OH

FIGURE 6 Metabolism of acetaminophen and relevance to toxicity. At low doses, the metabolism of acetaminophen is largely through conjugation. Some oxidation does occur but GSH conjugation renders this inactive. At higher doses, the three conjugation pathways are not effective and larger amounts of the protein conjugates are formed and apparently contribute to toxicity. See chapter 18 for more details.

Terfenadine (but not fexofenadine) also binds to the human ether-ago-go related gene (hERG) receptor and inhibits ion channels, resulting in abnormal QT intervals and arrhythmias. Terfenadine, the first non-sedating antihistamine, was recalled and replaced with fexofenadine. The fourth context of toxicity is bioactivation, which will be discussed later. An example shown is that of acetaminophen (Fig. 6), where some oxidation to a reactive iminoquinone is always detected but can be tolerated, until the dose of the drug exceeds the ability of the conjugation systems to clear the compound. The concept of bioactivation, already presented, is generally associated with the concept of reaction of the products with proteins or other cellular macromolecules to cause damage (51), except in cases in which the products are themselves dangerous due to altered pharmacology (vide infra) or cause oxidative stress. Thus, this context has some similarity to the second one (Table 2), except that an immune system component is usually not invoked. Historically, the concept was what can be called the “critical protein hypothesis” (7), where ablation of the function of an individual protein was responsible for cell toxicity; i.e., that protein acts as a “master switch” for the cell. However, a more accepted current view is that cells have very complex networks and that disruption of those signaling pathways (46) or any of several of the systems involved in energy production (7) may lead to toxicity, with effects being cell and tissue specific (see also chaps. 4 and 5). The final context in this discussion is idiosyncratic, which is rare (!1/104 patients), a serious problem, and not understood (by definition). Dihydralazine is usually regarded as an example of an idiosyncratic drug issue, although closer analysis reveals aspects of metabolic activation, covalent binding, and possibly immunological components involved in the response (Fig. 7). BIOACTIVATION BY P450s AND OTHER ENZYMES Basic Aspects As indicated earlier in this chapter, bioactivation of drugs is one context of toxicity of drugs and other chemicals. As indicated, the phenomenon has been known since the 1930s and

20

Guengerich

HN

NH2 N N

HN

HN P450

HN

HN

Dihydralazine

NH2

N=O N N

N N

1A2

NH2

NHOH

HN

P450 1A2 adducts

NH2 Anti-LKM antibodies (recognize unmodified P450 1A2) Cause/ effect? Hepatitis

FIGURE 7 Possible mechanisms for dihydralazine-induced hepatitis. Dihydralazine is oxidized by P450 1A2 and the hydroxylamine and/or nitroso product form protein adducts. LKM antibodies are found in patients. Whether these antibodies are causal in hepatitis is not established. See also chapter 7. Abbreviation: LKM, liver kidney microsomal. Source: From Ref. 52.

extensively applied to drugs since the early 1970s. Although bioactivation may seem complex, the fundamentals are quite basic, even if the interpretation may not be straightforward. Seven principles are used to summarize this area (8): 1. The reactions are described by two basic types of chemistry. (a) The first is the reaction of an electrophile (E; formed from the drug) and a nucleophile (NucZprotein, DNA, or other component in the cell) EdC CNucdK/E

2.

3.

4. 5. 6.

Nuc

or (b) Free radical propagation, usually resulting from an odd-electron enzymatic process or lipid peroxidation. Radical processes and oxidative stress from reactive oxygen (or nitrogen) species have not been addressed extensively here, but in some cases are probably very important (53). How much reactive oxygen species contribute to the overall hepatoxicity of drugs is unclear (see chap. 3). The first product or the most obvious reactant may not be the one that reacts with the nucleophile. For instance, in some cases, an epoxide may be generated by a P450 but it will undergo a rearrangement to an acyl halide or other product that participates in the reaction (54–56). The stability (and also the reactivity) of reactive products may vary considerably (57). A short-lived product, e.g., aflatoxin B1 (AFB1) exo-8,9-epoxide (t1/2Zone second at neutral pH) (58) can be formed in the endoplasmic reticulum but still migrates to the cell nucleus to react with DNA very efficiently (Fig. 8). Other reactive products may have half-lives on the order of minutes to hours. In contrast to the shorter t1/2 products, these can often migrate from one cell to another and even one tissue to another. In vitro systems are good models for elucidating details but ultimately all covalent binding results must be considered in the context of in vivo situations. In in vitro settings, some factors may be overemphasized or biased. The dose is an issue or, in many cases, the issue, going back 500 years to an axiom of toxicology from Paracelsus (60). Covalent binding can be an index of toxicity and often correlates (4,61). Some pharmaceutical companies are using data from in vitro and in vivo covalent binding studies to at least stratify candidates for development (62). However, many exceptions are known, i.e., nontoxic compounds show covalent binding and toxic compounds do not bind. Thus, caution is advised in interpretation.

O

OCH3

O O H

DNA

GS

OCH3

O

O O

OCH3

1.5 ×

106

H 2O (H+)

M–1 min–1

HO

H

O

OCH3

O

AFB1 GSH conjugate

O H O

OH

O

42 min–1 + 1.3 × 105 exo–8.9– epoxide AFB1 M–1 min–1 + GST M1 1.2 × 10−5 M–1 min–1 (H component)

O 1.8 × 105 M–1 min–1 H O

P450 3A4

O

O

AFB1 Guanine–N 7 adduct

O H O

H

O

O

O O

O OCH3 AFB1 dihydrodiol

O

OH

(H–) 0.12 min–1 (at pH 7)

OH–

HO

O

O

HO

O

O

OCH3

O

O

O

OCH3

O

OCH3

O

1.8 x 106 M–1 min–1

O

+

O

Protein

2.4 × 102 M–1 min–1 (BSA) to 2.4 × 103 M–1 min–1 (N 2–AcLys)

AFB1 monoaldehydes

HO

OH

HO

OH

O

OCH3

AFB1 dialdehyde

HO

OH

AFAR AKR 7A2,7A3

O

1.2 × 105 M–1 min–1 O

HO

O

O

AFB1 Lys adduct

Lys N

O

FIGURE 8 Metabolism of AFB1 and relevance to toxicity. The major human enzymes involved in each reaction are included and rates have been determined. First-order rates are listed in units of reciprocal time. Units of minK1 MK1 indicate either second-order rate constants (for chemical reactions) or catalytic efficiency (kcat/Km) for enzymatic reactions measured using the most relevant purified recombinant human enzymes. The exo-8,9-epoxide plays a central role in this process and appears to be the only product that reacts with DNA. The dialdehyde reacts with proteins (BSA). The dehydrogenase AFAR is now classified as an AKR. Several other oxidation products (AFQ, AFM, endo epoxide) are not shown, which are all less toxic than AFB1. Abbreviations: AFAR, aflatoxin aldehyde reductase; AFB1, aflatoxin B1; AFQ, aflatoxin Q1; AFM, aflatoxin M1; AKR, aldo-keto reductase; BSA, bovine serum albumin. Source: From Ref. 59.

AFB1

O H O

H

O

Gua–N 7

OH

Cytochrome P450 Activation of Toxins

21

22

Guengerich

7. Numerous other factors influence the relevance of covalent binding to toxicity. Among these are receptor-medicated events (i.e., with the parent drug itself), the disruption of particular signaling events by the reactive products, the ability of the cells to repair damage, the intervention of immunological processes, and cell proliferation (see also chaps. 4, 5, and 11). Examples of P450-Dependent Bioactivation Acetaminophen This is a classic example of P450 bioactivation (4) (see also chap. 18). It is also a very practical issue because, as pointed out elsewhere in this book, this is one of the most commonly used drugs and overdoses are a leading cause of liver failure today (see also chap. 19). At low doses, the conjugating enzymes allow only a small amount of acetaminophen to be oxidized (P450 2E1, 3A4, 1A2) to the reactive iminoquinone, a reactive “Michael acceptor” (Fig. 6). With higher doses and in situations in which the cofactors needed for the various conjugations are depleted (e.g., GSH, UDP-glucuronic acid, 3 0 -phosphoadenosine 5 0 -phosphosulfate, as in nutritionally challenged states), more acetaminophen is activated and reacts with proteins. Which proteins are the most critical targets has been considered extensively (7,63) but is still not clear (see chap. 18). The process is somewhat more complex in terms of other processes not shown in Figure 6. The iminoquinone is reduced back to acetaminophen by NADPH-P450 reductase, hydrolysis occurs to yield benzoquinone, and peroxidases can also be involved in the activation (64,65). Aflatoxin B1 AFB1 is not a drug but is a mycotoxin of considerable significance regarding human exposure, in that it is one of the most potent (and prevalent) hepatocarcinogens known (66). P450 oxidations of AFB1 produce several products, all of which are less dangerous except one, the exo-8,9-epoxide (Fig. 8). All of the genotoxicity can be explained in terms of this epoxide, specifically the exo isomer (67). Even though the t1/2 of the epoxide is one second under physiological conditions (58), the compound can migrate from the endoplasmic reticulum to the cell nucleus to bind efficiently to DNA (68). Alternatively, the epoxide can be enzymatically conjugated with GSH (69) or (nonenzymatically) hydrolyze to a dihydrodiol. The dihydrodiol is unstable and undergoes reversible ring opening to a dialdehyde species (58). The dialdehyde reacts with protein lysines (but not with DNA) (59), but enzymatic reduction of the dialdehyde occurs and prevents the covalent binding (70). Measurements of AFB1–lysine adducts in human serum albumin have been used to assess AFB1 exposure and biologically relevant exposure (71). Trichloroethylene Although trichloroethylene (TCE) was used formerly as an anesthetic, the major concerns are from industrial and environmental exposure to this solvent. Oxidation of TCE by P450s, especially P450 2E1 (72), yields two primary products (Fig. 9). One is chloral (hydrate), formed via rearrangement of an enzyme–substrate intermediate (Fig. 9) (74). (Chloral is a known sedative, although this is probably not an issue in the pharmacology of TCE.) The other major oxidation product is TCE oxide, which has a t1/2 of 12 seconds under physiological conditions (55,74). Labeling experiments with 18O and 2H (55) support a mechanism for reaction with lysines that does not involve a direct reaction by the epoxide but instead by an acyl halide rearrangement product (Fig. 10). TCE oxide [and its rearrangement product(s)] does not react well with DNA (75). The reactions with protein lysine groups yield stable amides, both N6-formyl lysine and N6dichloroacetyl lysine (55). Modern electrospray mass spectrometry methods can be used to quantify the number and amount of the two adducts formed on an oligopeptide or small protein in vitro (55). In the course of such analysis, it was found that some of the adducts are unstable at basic or neutral pH (75). These appear to be adducts formed with serine, cysteine, and tyrosine (75). The t1/2 for loss under physiological conditions is wone hour. When model

23

Cytochrome P450 Activation of Toxins

Cl FeO3+

+

Cl

Cl

H

Cl

FeO2+

Cl Cl

H

O H

+ Cl

FeIII

Cl Cl Cl

H

FeIV

H

Cl

HO Cl

O

Cl Cl

Cl Cl H

O

Cl O

Cl

Cl

Cl

H Cl N N N N Heme (porphyrin) adduct

FIGURE 9 Oxidation of TCE by P450. The activated form of P450 is indicated as the peroxyferryl complex, FeO3C. Reaction with TCE yields a one-electron oxidation product and then a carbocation intermediate, which can rearrange to chloral or react with heme. The epoxide is unstable (t1/2Z12 seconds) and undergoes rapid reaction (Fig.10) but does not form chloral. Abbreviation: TCE, trichloroethylene. Source: From Refs. 73, 74.

enzymes with these three residues critical for catalytic activity were partially inactivated with TCE oxide, the catalytic activity slowly returned with a t1/2 of wone hour, suggesting the loss of the acylating entity and “recovery” of the enzymes (75). Such unstable protein adducts would not have been detected using more traditional, slower analytical methods, and the existence of such a phenomenon raises several questions. On one hand, the transient nature of the adducts might argue that a given amount of protein damage is not as serious as first thought from consideration of the level of binding. Alternatively, one can take the other view that the extra burden on an enzyme involved in cell regulation may be serious, if the regulatory system is distorted for even a short period of time. Troglitazone This compound was the first in the thiazolidinedione group of PPAR activators marketed to treat diabetes (see chap. 28). However, the drug was withdrawn in 2000 due to hepatotoxicity, with the incidence of liver transplant or fatality due to liver failure being 1/104 (47). The FDA withdrew the drug following 560 reports of hepatotoxicity. X O

X

H

X

a O

O

H b aX NH2-Lys b

O

O

H

NH-Lys

− O

+ Cl H Cl

X

O

NH-Lys H

O X NH2-Lys

X

O

X

NH-Lys

H

H

O NH-Lys

FIGURE 10 Reactions of TCE oxide (Fig. 9) with lysine residues in proteins. Evidence for the proposed pathways was obtained using 18O and 2H labeling studies. The acyl halide intermediates are the species that react with the protein, not the epoxide itself. Abbreviation: TCE, trichloroethylene. Source: From Ref. 55.

24

Guengerich

Although the exact cause of the hepatotoxicity has not been firmly established, the covalent binding of troglitazone is considerable in model systems. P450 3A4 is involved in the pathways that can lead to electrophilic products (Fig. 11). One pathway (Fig. 11A) involves the formation of a quinone, a potent Michael acceptor capable of reacting with nucleophilic groups in proteins. The other pathway involves oxidation of the thiazolidinone ring (Fig. 11B). Although troglitazone was withdrawn from the market, several other “glitazones” with the thiazolidinone structure are now marketed. If the oxidation of that ring is the cause of covalent binding (and covalent binding is related to toxicity) then one might expect these to also be toxic. They appear not to be, but part of the reason may be that the doses used (!10 mg QD) are more than an order of magnitude less than used with troglitazone.

RELATIVE CONTRIBUTION OF METABOLISM TO DRUG TOXICITY The contexts of drug toxicity have been discussed earlier. How does metabolism fit into this picture, i.e., in what fraction of the cases are metabolites really involved in toxicity of drugs? Two aspects can be considered. One is metabolism of drugs to stable products. The process of metabolism almost always occurs to some extent, and the products generally have attenuated activity toward the same target (as the parent drug). In some cases, the product may show enhanced biological activity at the same site, and this is really a “pro-drug” phenomenon. [An example is the oxidation of the antihypertensive losartan to a carboxylic acid by P450 2C9 (79)]. Recently, the FDA has expressed concern about safety assessment of drug metabolites that are unique to humans (compared to experimental animals) [FDA/CDER Draft: Guidance for Safety Testing of Drug Metabolites, 2005]. However, few drugs appear to be transformed to products that have pharmacology totally different from the parent drug (46), even if the possibility cannot be completely excluded. Examples from the non-pharmaceutical world include the oxidation of the anesthetic/solvent TCE to the sedative chloral (vide supra) and the oxidation of the pesticide methoxychlor to an estrogen (80). The other issue is metabolism of drugs to reactive products. Indeed, all of the four examples cited by FDA/CDER in the 2005 Draft on Guidance for Safety Testing of Drug Metabolites (vide supra) involve reactive products, which cannot usually be tested in the usual manner because they are unlikely to reach the target site. Reactive products are an interesting topic and have a tendency to reach the open literature because of their interesting chemistry. However, what fraction of drug problems really involves bioactivation? One way of looking at the issue involves consideration of acetaminophen itself, probably the most widely used drug today in the United States. (see chap. 19). Nearly, one-half of druginduced liver injury can be related to drugs, and about one-half of these cases are due to overdoses of acetaminophen. The statement can be accurately made that all acetaminopheninduced injury is the result of bioactivation, even if unknown other events are also important in determining the course of injury (7,81,82). Another way of analyzing the roles of metabolism and covalent binding is the analysis of Walgren et al. (83). Of six drugs that have been withdrawn from the market (in a given recent time period), five are known to have reactive metabolites (benzoxaprofen, iproniazid, nefazodone, tienilic acid, and troglitazone). Of 15 drugs with “Black Box” warnings, eight have reactive metabolites (dacarbazine, dantrolene, felbamate, flutamide, isoniazid, ketoconazole, tolcapone, and valproic acid). In considering the above two sets of compounds, 62% have reactive metabolites. Further, at least 14 drugs known to produce reactive metabolites have warnings about hepatotoxicity (acetaminophen, carbamazepine, clozapine, diclofenac, disulfiram, halothane, leflunomide, methyldopa, rifampicin, tacrin, tamoxifen, terbinafine, ticlopidine, and zileuton). Another 14 drugs with reactive metabolites were never approved in the United States (alpidem, amineptine, amodiaquine, cinchophen, glufenine, ibufenac, isoxanine, niperotiden, perhexiline, pirprofen, and tibroquinol) (83). Looking at the amassed literature, one might conclude that bioactivation is a prominent aspect of drug toxicity. Indeed, several pharmaceutical companies now conduct early in vitro screens to stratify compounds to avoid those most likely to show reactive products (62). However, the above discussion may present a biased view of the reasons why new drug

Troglitazone

O

O

(B)

N H

O

O

N H

O– S+

O

O

O

· O

S

S

O

NH2

SG

GSH, H2O –CO2

O

N=C=O

OH

O

O

GSH

OH O

CH2

O

S O

OH

O

GSH

SG

O

O

HO

O

–O

O

O

O

O

–H2O

O

OH

OH

S+

S

OH

GSH

O

SG

O

SG

SG

S

NH2 SG

O

FIGURE 11 Activation of troglitazone by P450 3A4. Pathway (A) results in the formation of a benzoquinone. Pathway (B) involves oxidation of the thiazolidinedione ring. Source: From Refs. 47, 76 78.

HO

O

S

(A)

·

O

Cytochrome P450 Activation of Toxins

25

26

Guengerich

candidates fail. First, when reactive metabolites are found, there is almost universally no demonstrated mechanism of causality with disease, as of yet. Another issue is that the true cause of toxicity of many failed drug candidates is never investigated, and there is a strong possibility that many show off-target pharmacology as the context of toxicity (vide supra). Unfortunately, a complete analysis of the causes of candidate failure may never be available because of the expense of analyzing mechanisms for large numbers of compounds that are not viable as leads. One final important point should be made. Although many of the toxicities, particularly those considered idiosyncratic, are not considered to be dose-dependent, there is a clear inverse correlation between the dose and the proclivity for toxicity. Most of the drugs that have been withdrawn from the market were used at high doses (O100 mg/day) (83) and few problems have been seen at doses on the order of 10 mg/day (47). This information might be considered as evidence for roles of bioactivation and either protein modification or oxidative stress, in that there is a hint of a threshold. The relationship is not absolute and one can consider the case of ceruvistatin, which was used at sub-milligram per day doses and withdrawn due to muscle and kidney problems. However, this is probably a case of on-target toxicity, in that the effect is due to the intended action of the drug (albeit in a different tissue) (84). Thus, the conclusion is that drug toxicity is the result of several factors, including the dose, the usage (chronic use is more of a problem than temporary use), the activation to reactive metabolites and existence of detoxication pathways, and the inherent on- and off-target pharmacology. More problems arise when a single pathway is dominant in the metabolism steps and large variations in the population can have major impacts (as opposed to situations where several enzymes contribute to detoxication and the result is a dampening of the effects of changes in any single system) (7,85).

APPROACHES TO SAFETY PREDICTION OF NEW DRUGS Toxicity, particularly hepatotoxicity, has become one of the most serious concerns in the development of new drugs. A major problem is the development of toxicity problems late in the process of clinical trials, after considerable investment in a drug candidate. One issue is the prediction of chemical groups that make drugs more likely to be activated to reactive products (86). This issue has been considered by medicinal chemists, and a number of entities are now either avoided completely or, if any assays suggest toxicity in early screening, changed to reduce toxicity while maintaining efficacy (47,86). Although some chemical entities are clearly more prone to cause problems than others (e.g., thiophenes), this approach cannot completely avoid problems in that even a simple phenyl ring is only from one to three steps away from a reactive product. One view of the drug development process is shown in Figure 12. A generation ago the loop cycled only between chemistry (drug design and synthesis) and biology (efficacy and selectivity testing). At that time, most of the screening was with animals (in vivo). Today, the field involves more in vitro approaches, particularly with systems of human origin. A separate assay loop is involved, with “developability screening” and predictions. Many of the desirable properties of a drug candidate are related to P450s, transporters, and detoxication processes. The overall goal is to select the most desirable candidates before proceeding to the more complex (and expensive) safety studies and clinical trials. In many respects, the prediction of pharmacokinetics has been considerably improved with better understanding of the P450s, conjugating enzymes, and transporters. However, early prediction of toxicity is more difficult (7,46). The basic problem is that the key events in most types of toxicity remain poorly understood (see chap. 5). Genotoxicity is an exception, to a large extent, in that simple mutations can be scored quickly. Some of the issues in the biology of cell toxicity have been introduced elsewhere (see chap. 5). Several newer approaches are being examined in both industrial and academic settings (46) (see chap. 15).

27

Cytochrome P450 Activation of Toxins

Target Lead compounds

Design & synthesis

Efficacy & selectivity testing

Candidate

"Developability screening" & predictions

Toxicology pharmaceutics ADME: Pharmacokinetics Pharmacodynamics Cytochrome P450 inhibition Cytochrome P450 induction Permeability Transporter interactions Instrinsic clearance Reaction phenotyping Reactive metabolites

Detailed physicochemical, ADME, & safety workup

Clinical trials

FIGURE 12 A typical strategy used in drug development. A target receptor or enzyme is established on the basis of knowledge about the biology for the disease of interest. Lead compounds are developed, usually by screening procedures (sometimes by rational design). These candidates enter a series of reiterative development loops. Efficacy and selectivity testing involve pharmacological assays based on the target; developing screening and predictions involve the aspects shown at the right; design and synthesis involve more chemistry. The goal of the process is to develop candidates that are less likely to fail in the detailed later studies and the clinical trials. Thanks are extended to W. G. Humphreys (Bristol-Myers Squibb) for the concept. Abbreviation: ADME, absorption, distribution, metabolism, and excretion. Source: Adapted from Ref. 87.

1. One approach is in silico prediction. The difficulty is that few targets directly related to toxicity have been identified and structures of even fewer are available, so modeling of receptor–ligand interactions is possible in only a few cases (e.g., possibly Ah receptor). Thus, most approaches are based on quantitation–structure relationships, mostly involving structural entities and crude toxicity parameters. 2. The second approach is transcriptonomics, or mRNA profiling. A number of experiments have been done in this arena (88,89), with the general goal of finding early changes in patterns that will be predictive of toxicity. Because the extent of changes in gene expression is extensive (one can easily find changes in expression of 10% of the genes due to treatment in many experiments), the analysis is not trivial and enlightened new approaches to the informatics are necessary. Another complication is that in vivo changes are probably much more valuable than work done in cell cultures; extrapolation to humans is still an issue. 3. The third approach is proteomics. Briefly, four aspects can be considered. The first is changes in the levels of expression of individual genes, similar to the approach used in transcriptome profiling. The same problems arise as in the consideration of mRNA profiles. Another aspect is the analysis of changes in posttranslational events, such as phosphorylation (redox changes are generally even more difficult). The third approach is to map drug modification of individual proteins (90). To date the modification of a single protein has not clearly defined a role in toxicity, but as increased information becomes available more insight may be possible. The fourth approach, like the second and third of these, uses mass spectrometry approaches and involves the use of the techniques to profile patterns of localization of proteins within tissues (91,92).

28

Guengerich

4. Metabonomics (or sometime “metabolomics”) is the analysis of changes in the profiles of low Mr compounds, usually using blood or urine samples. The methods are usually NMR spectroscopy and mass spectrometry (93). The point is not to look at the drug or its metabolites but instead at changes in the profiles of endogenous compounds. These are not necessarily identified, but the goal is to find changes in patterns that can be related to toxicity and serve as early predictors, e.g., with pattern sorting using principal components analysis and related approaches. 5. Another approach is the use of covalent binding data. The point is to stratify lead candidates to find those that produce less binding. Two major approaches are used. One is in vitro (cell-free) systems searching for conjugates of the drug with GSH or a surrogate thiol (62,94). This aspect can be done without radiolabeled material. Quantitation of in vivo (animal) results is done subsequently, with radiolabeled drug. A rough decision cutoff value of 50 pmol adduct/mg protein has been proposed (62), but obviously this value would be a function of dose, time, etc. The overall plan in these strategies is to find better and earlier biomarkers that can be used effectively in the development process and even in clinical trials and marketed drugs. A critical issue with candidates found in animals is extrapolation to humans. Another lingering problem is the idiosyncratic toxicities. With time, more of these will be classified into other “contexts of toxicity” (vide supra). However, the problem of eliminating all possible adverse reactions in large clinical populations without a mechanism or an animal model is still a formidable task. CONCLUSIONS Metabolism is an important issue in drug efficacy and toxicity. The P450s are the most important enzymes in these processes, but other enzymes have roles, including transporters. The regulation of P450s is understood largely in terms of transcription and classic receptor– ligand systems. Metabolism by P450s and other systems clearly has a role in the toxicity of many drugs, generally either deactivating the drug or activating it to a reactive form. There are several contexts of drug toxicity, and exactly what the relative contributions are to overall injury is not clear. However, a sizeable function of toxicity is somewhat related to reactive products, although even when these compounds are produced they appear to require other factors for toxicity. A number of newer approaches to toxicity prediction are being considered, but the validation of biomarkers of toxicity is still an important issue. REFERENCES 1. Fieser LF. Carcinogenic activity, structure and chemical reactivity of polynuclear hydrocarbons. Am J Cancer 1938; 34:37–124. 2. Miller EC, Miller JA. The presence and significance of bound amino azodyes in the livers of rats fed p-dimethylaminoazobenzene. Cancer Res 1947; 7:468–80. 3. Mueller GC, Miller JA. The metabolism of 4-dimethylaminoazobenzene by rat liver homogenates. J Biol Chem 1948; 176(11):535–44. 4. Jollow DJ, Mitchell JR, Potter WZ, et al. Acetaminophen-induced hepatic necrosis. II. Role of covalent binding in vivo. J Pharmacol Exp Ther 1973; 187(1):195–202. 5. Zampaglione N, Jollow DJ, Mitchell JR, et al. Role of detoxifying enzymes in bromobenzene-induced liver necrosis. J Pharmacol Exp Ther 1973; 187(1):218–27. 6. Basu AK, Essigmann JM. Site-specifically modified oligodeoxynucleotides as probes for the structural and biological effects of DNA-damaging agents. Chem Res Toxicol 1988; 1(1):1–18. 7. Park BK, Kitteringham NR, Maggs JL, et al. The role of metabolic activation in drug-induced hepatotoxicity. Annu Rev Pharmacol Toxicol 2005; 45(1):177–202. 8. Guengerich FP. Principles of covalent binding of reactive metabolites and examples of activation of bis-electrophiles by conjugation. Arch Biochem Biophys 2005; 433(2):369–78. 9. Dorfman R, Cook JW, Hamilton JB. Conversion by the human of the testis hormone, testosterone, into the urinary androgen, androsterone. J Biol Chem 1939; 130:285–95. 10. Ryan KJ. Conversion of androstenedione to estrone by placental microsomes. Biochim Biophys Acta 1958; 27(3):658–9.

Cytochrome P450 Activation of Toxins

29

11. Bradshaw WH, Conrad HE, Corey EJ, et al. Degradation of (C)-camphor. J Am Chem Soc 1959; 81(20):5007. 12. Cooper DY, Levine S, Narasimhulu S, et al. Photochemical action spectrum of the terminal oxidase of mixed function oxidase systems. Science 1965; 147:400–2. 13. Lu AYH, Coon MJ. Role of hemoprotein P-450 in fatty acid w-hydroxylation in a soluble enzyme system from liver microsomes. J Biol Chem 1968; 243(6):1331–2. 14. Remmer H. The acceleration of evipan oxidation and the demethylation of methylaminopyrine by barbiturates. Naunyn–Schmiedeberg’s Arch Exp Pathol Pharmakol 1959; 237:296–307. 15. Conney AH, Miller EC, Miller JA. The metabolism of methylated aminoazo dyes. V. Evidence for induction of enzyme synthesis in the rat by 3-methylcholanthrene. Cancer Res 1956; 16(15):450–9. 16. Sladek NE, Mannering GJ. Induction of drug metabolism. II. Qualitative differences in the microsomal N-demethylating systems stimulated by polycyclic hydrocarbons and by phenobarbital. Mol Pharmacol 1969; 5(2):186–99. 17. Haugen DA, van der Hoeven TA, Coon MJ. Purified liver microsomal cytochrome P-450: separation and characterization of multiple forms. J Biol Chem 1975; 250(9):3567–70. 18. Fujii-Kuriyama Y, Mizukami Y, Kawajiri K, et al. Primary structure of a cytochrome P-450: coding nucleotide sequence of phenobarbital-inducible cytochrome P-450 cDNA from rat liver. Proc Natl Acad Sci USA 1982; 79(9):2793–7. 19. Williams PA, Cosme J, Sridhar V, et al. Mammalian microsomal cytochrome P450 monooxygenase: strutural adaptations for membrane binding and functional diversity. Mol Cell 2000; 5(1):121–31. 20. Nebert DW, Adesnik M, Coon MJ, et al. The P450 gene superfamily: recommended nomenclature. DNA 1987; 6(1):1–11. 21. http://drnelson.utmem.edu/CytochromeP450.html, accessed on April 25, 2007. 22. Guengerich FP. Human cytochrome P450 enzymes. In: Ortiz de Montellano PR, ed. Cytochrome P450: Structure, Mechanism, and Biochemistry. 3rd ed. New York: Kluwer Academic/Plenum Press, 2005:377–531. 23. Guengerich FP, Wu Z-L, Bartleson CJ. Function of human cytochrome P450s: characterization of the remaining orphans. Biochem Biophys Res Commun 2005; 338(1):465–9. 24. Williams JA, Hyland R, Jones BC, et al. Drug-drug interactions for UDP-glucuronosyltransferase substrates: a pharmacokinetic explanation for typically observed low exposure (AUCi/AUC) ratios. Drug Metab Dispos 2004; 32(11):1201–8. 25. Guengerich FP, Shimada T. Oxidation of toxic and carcinogenic chemicals by human cytochrome P-450 enzymes. Chem Res Toxicol 1991; 4(4):391–407. 26. Turesky RJ, Constable A, Richoz J, et al. Differences in activation of heterocyclic aromatic amines by rat and human liver microsomes and by rat and human cytochromes P450 1A2. Chem Res Toxicol 1998; 11(8):925–36. 27. http://www.fda.gov/cder/guidance/6366dft.htm, accessed on April 25, 2007. 28. Segel IH. Simple inhibition systems. Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and Steady-State Enzyme Systems. New York: Wiley, 1975:109. 29. Guengerich FP. Inhibition of drug metabolizing enzymes: molecular and biochemical aspects. In: Woolf TF, ed. Handbook of Drug Metabolism New York: Marcel Dekker, 1999:203–27. 30. Guengerich FP. Oxidation of 17a-ethynylestradiol by human liver cytochrome P-450. Mol Pharmacol 1988; 33(5):500–8. 31. Guengerich FP. Mechanism-based inactivation of human liver cytochrome P-450 IIIA4 by gestodene. Chem Res Toxicol 1990; 3(4):363–71. 32. Schmiedlin-Ren P, Edwards DJ, Fitzsimmons ME, et al. Mechanisms of enhanced oral availability of CYP3A4 substrates by grapefruit constituents. Drug Metab Dispos 1997; 25(11):1228–33. 33. Watkins PB. Omeprazole induction of cytochrome P-450IA2: the importance of selecting the appropriate human model. Hepatology 1993; 17(4):748–50. 34. Olsen JH, Boice JD, Jr., Jensen JPA, et al. Cancer among epileptic patients exposed to anticonvulsant drugs. J Natl Cancer Inst 1989; 81(10):803–8. 35. Capen CC, Dayan AD, Green S. Receptor-mediated mechanisms in carcinogenesis: an overview. Mut Res 1995; 333(1–2):215–24. 36. Bolt HM, Kappus H, Bolt M. Effect of rifampicin treatment on the metabolism of oestradiol and 17a-ethinyloestradiol by human liver microsomes. Eur J Clin Pharmacol 1975; 8(39):301–7. 37. Williams SN, Dunham E, Bradfield CA. Induction of P450 enzymes: receptors. In: Ortiz de Montellano PR, ed. Cytochrome P450: Structure, Mechanism, and Biochemistry. 3rd ed. New York: Kluwer Academic/Plenum Press, 2005:323–46. 38. Park Y, Li H, Kemper B. Phenobarbital induction mediated by a distal CYP2B2 sequence in rat liver transiently transfected in situ. J Biol Chem 1996; 271(39):23725–8. 39. Honkakoski P, Moore R, Gynther J, et al. Characterization of phenobarbital-inducible mouse Cyp2b10 gene transcription in primary hepatocytes. J Biol Chem 1996; 271(16):9746–53. 40. Honkakoski P, Negishi M. Regulation of cytochrome P450 (CYP) genes by nuclear receptors. Biochem J 2000; 347(2):321–37.

30

Guengerich

41. Goodwin B, Redinbo MR, Kliewer SA. Regulation of CYP3A gene transcription by the pregnane X receptor. Annu Rev Pharmacol Toxicol 2002; 42(1):1–23. 42. Tirona RG, Lee W, Leake BF, et al. The orphan nuclear receptor HNF4a determines PXR- and CARmediated xenobiotic induction of CYP3A4. Nat Med 2003; 9(2):220–4. 43. Mangelsdorf DJ, Evans RM. The RXR heterodimers and orphan receptors. Cell 1995; 83(6):841–50. 44. Lehmann JM, Kliewer SA, Moore LB, et al. Activation of the nuclear receptor LXR by oxysterols defines a new hormone response pathway. J Biol Chem 1997; 272(6):3137–40. 45. Tirona RG, Leake BF, Podust LM, et al. Identification of amino acids in rat pregnane X receptor that determine species-specific activation. Mol Pharmacol 2004; 65(1):36–44. 46. Liebler DC, Guengerich FP. Elucidating mechanisms of drug-induced toxicity. Nat Rev Drug Discov 2005; 4(5):410–20. 47. Kalgutkar AS, Gardner I, Obach RS, et al. A comprehensive listing of bioactivation pathways of organic functional groups. Curr Drug Metab 2005; 6(3):161–225. 48. Uetrecht JP. New concepts in immunology relevant to idiosyncratic drug reactions: the “danger hypothesis” and innate immune system. Chem Res Toxicol 1999; 12(5):387–95. 49. Gerber BO, Pichler WJ. Cellular mechanisms of T cell mediated drug hypersensitivity. Curr Opin Immunol 2004; 16(6):732–7. 50. Yun C-H, Okerholm RA, Guengerich FP. Oxidation of the antihistaminic drug terfenadine in human liver microsomes: role of cytochrome P450 3A(4) in N-dealkylation and C-hydroxylation. Drug Metab Dispos 1993; 21(3):403–9. 51. Miller EC, Miller JA, Sapp RW, et al. Studies on the protein-bound aminoazo dyes formed in vivo from 4-dimethyl-aminoazobenzene and its C-monomethyl derivatives. Cancer Res 1949; 9(6):336–43. 52. Bourdi M, Larrey D, Nataf J, et al. A new anti-liver endoplasmic reticulum antibody directed against human cytochrome P-450 IA2: a specific marker of dihydralazine-induced hepatitis. J Clin Invest 1990; 85(6):1967–73. 53. Berlett BS, Stadtman ER. Protein oxidation in aging, disease, and oxidative stress. J Biol Chem 1997; 272(33):20313–6. 54. Liebler DC, Guengerich FP. Olefin oxidation by cytochrome P-450: evidence for group migration in catalytic intermediates formed with vinylidene chloride and trans-1-phenyl-1-butene. Biochemistry 1983; 22(24):5482–9. 55. Cai H, Guengerich FP. Mechanism of aqueous decomposition of trichloroethylene oxide. J Am Chem Soc 1999; 121(50):11656–63. 56. Yoshioka T, Krauser JA, Guengerich FP. Tetrachloroethylene oxide: hydrolytic products and reactions with phosphate and lysine. Chem Res Toxicol 2002; 15(8):1096–105. 57. Guengerich FP, Liebler DC. Enzymatic activation of chemicals to toxic metabolites. CRC Crit Rev Toxicol 1985; 14(3):259–307. 58. Johnson WW, Harris TM, Guengerich FP. Kinetics and mechanism of hydrolysis of aflatoxin B1 exo-8,9oxide and rearrangement of the dihydrodiol. J Am Chem Soc 1996; 118(35):8213–20. 59. Guengerich FP, Arneson KO, Williams KM, et al. Reaction of aflatoxin B1 oxidation products with lysine. Chem Res Toxicol 2002; 15(6):780–92. 60. Borzelleca JF. Profiles in toxicology—paracelsus: herald of modern toxicology. Toxicol Sci 2000; 53(1):2–4. 61. Boyd MR, Burka LT. In vivo studies on the relationship between target organ alkylation and the pulmonary toxicity of a chemically reactive metabolite of 4-ipomeanol. J Pharmacol Exp Ther 1978; 207(3):687–97. 62. Evans DC, Watt AP, Nicoll-Griffith DA, et al. Drug-protein adducts: an industry perspective on minimizing the potential for drug bioactivation in drug discovery and development. Chem Res Toxicol 2004; 17(1):3–16. 63. Qiu Y, Benet LZ, Burlingame AL. Identification of the hepatic protein targets of reactive metabolites of acetaminophenin in vivo in mice using two-dimensional gel electrophoresis and mass spectrometry. J Biol Chem 1998; 273(28):17940–53. 64. Nelson SD, Dahlin DC, Rauckman EJ, et al. Peroxidase-mediated formation of reactive metabolites of acetaminophen. Mol Pharmacol 1981; 20(1):195–9. 65. Potter DW, Hinson JA. The 1- and 2-electron oxidation of acetaminophen catalyzed by prostaglandin H synthase. J Biol Chem 1987; 262(3):974–80. 66. Busby WF, Wogan GN. Aflatoxins. In: Searle CE, ed. Chemical Carcinogens. 2nd ed. Washington, D.C.: American Chemical Society, 1984:945–1136. 67. Iyer R, Coles B, Raney KD, et al. DNA adduction by the potent carcinogen aflatoxin B1: mechanistic studies. J Am Chem Soc 1994; 116(5):1603–9. 68. Johnson WW, Guengerich FP. Reaction of aflatoxin B1 exo-8,9-epoxide with DNA: kinetic analysis of covalent binding and DNA-induced hydrolysis. Proc Natl Acad Sci USA 1997; 94(12):6121–5. 69. Johnson WW, Ueng Y-F, Mannervik B, et al. Conjugation of highly reactive aflatoxin B1 8,9-exoepoxide catalyzed by rat and human glutathione transferases: estimation of kinetic parameters. Biochemistry 1997; 36(11):3056–60.

Cytochrome P450 Activation of Toxins

31

70. Guengerich FP, Cai H, McMahon M, et al. Reduction of aflatoxin B1 dialdehyde by rat and human aldo-keto reductases. Chem Res Toxicol 2001; 14(6):727–37. 71. Groopman JD, Kensler TW, Links JM. Molecular epidemiology and human risk monitoring. Toxicol Lett 1995; 82/83:763–9. 72. Guengerich FP, Kim D-H, Iwasaki M. Role of human cytochrome P-450 IIE1 in the oxidation of many low molecular weight cancer suspects. Chem Res Toxicol 1991; 4(2):168–79. 73. Guengerich FP, Macdonald TL. Chemical mechanisms of catalysis by cytochromes P-450: a unified view. Acc Chem Res 1984; 17(1):9–16. 74. Miller RE, Guengerich FP. Oxidation of trichloroethylene by liver microsomal cytochrome P-450: evidence for chlorine migration in a transition state not involving trichloroethylene oxide. Biochemistry 1982; 21(5):1090–7. 75. Cai H, Guengerich FP. Reaction of trichloroethylene oxide with proteins and DNA: instability of adducts and modulation of functions. Chem Res Toxicol 2001; 14(1):54–61. 76. Kassahun K, Pearson PG, Tang W, et al. Studies on the metabolism of troglitazone to reactive intermediates in vitro and in vivo. Evidence for novel biotransformation pathways involving quinone methide formation and thiazolidinedione ring scission. Chem Res Toxicol 2001; 14(1):62–70. 77. Reddy VB, Karanam BV, Gruber WL, et al. Mechanistic studies on the metabolic scission of thiazolidinedione derivatives to acyclic thiols. Chem Res Toxicol 2005; 18(5):880–8. 78. Yamazaki H, Shibata A, Suzuki M, et al. Oxidation of troglitazone to a quinone-type metabolite catalyzed by cytochrome P450 2C8 and 3A4 in human liver microsomes. Drug Metab Dispos 1999; 27(11):1260–6. 79. Stearns RA, Chakravarty PK, Chen R, et al. Biotransformation of losartan to its active carboxylic acid metabolite in human liver microsomes: role of cytochrome P4502C and 3A subfamily members. Drug Metab Dispos 1995; 23(2):207–15. 80. Stresser DM, Kupfer D. Human cytochrome P450-catalyzed conversion of the proestrogenic pesticide methoxychlor into an estrogen. Drug Metab Dispos 1998; 26(9):868–74. 81. Dahlin DC, Miwa GT, Lu AYH, et al. N-acetyl-p-benzoquinone imine: a cytochrome P-450-mediated oxidation product of acetaminophen. Proc Natl Acad Sci USA 1984; 81(5):1327–31. 82. Lee SST, Buters JTM, Pineau T, et al. Role of CYP2E1 in the hepatotoxicity of acetaminophen. J Biol Chem 1996; 271(20):12063–7. 83. Walgren JL, Mitchell MD, Thompson DC. Role of metabolism in drug-induced idiosyncratic hepatotoxicity. Crit Rev Toxicol 2005; 35(4):325–61. 84. Johnson TE, Zhang X, Bleicher KB, et al. Statins induce apoptosis in rat and human myotube cultures by inhibiting protein geranylgeranylation but not ubiquinone. Toxicol Appl Pharmacol 2004; 200(3):237–50. 85. Dorne JL, Walton K, Renwick AG. Human variability in CYP3A4 metabolism and CYP3A4-related uncertainty factors for risk assessment. Food Chem Toxicol 2003; 41(2):201–24. 86. Kalgutkar AS, Soglia JR. Minimising the potential for metabolic activation in drug discovery. Expert Opin Drug Metab Toxicol 2005; 1(1):91–141. 87. Guengerich FP. Life and times in biochemical toxicology. Int J Toxicol 2005; 24(1):5–21. 88. Thomas RS, Rank DR, Penn SG, et al. Identification of toxicologically predictive gene sets using cDNA microarrays. Mol Pharmacol 2001; 60(6):1189–94. 89. Hayes KR, Bradfield CA. Advances in toxicogenomics. Chem Res Toxicol 2005; 18(3):403–14. 90. Hansen BT, Davey SW, Ham AJL, et al. P-Mod: an algorithm and software to map modifications to peptide sequences from tandem MS data. J Proteome Res 2005; 4(20):358–68. 91. Witzmann FA, Grant RA. Pharmacoproteomics in drug development. Pharmacogenomics J 2003; 3(2):69–76. 92. Chaurand P, Schwartz SA, Caprioli RM. Assessing protein patterns in disease using imaging mass spectrometry. J Proteome Res 2004; 3(2):245–52. 93. Lindon JC, Nicholson JK, Holmes E, et al. Contemporary issues in toxicology the role of metabonomics in toxicology and its evaluation by the COMET project. Toxicol Appl Pharmacol 2003; 187(3):137–46. 94. Gan J, Harper TW, Hsueh MM, et al. Dansyl glutathione as a trapping agent for the quantitative estimation and identification of reactive metabolites. Chem Res Toxicol 2005; 18(5):896–903.

3

Antioxidant Defense in Liver Injury: Oxidant Stress, Antioxidant Defense, and Liver Injury Hartmut Jaeschke

Department of Pharmacology, Toxicology, and Therapeutics, University of Kansas Medical Center, Kansas City, Kansas, U.S.A.

INTRODUCTION The liver, the body’s largest solid organ, performs a substantial number of vital metabolic functions and is the main organ for drug and xenobiotic metabolism. These functions require an extensive aerobic metabolism to generate sufficient quantities of ATP in mitochondria. However, this metabolic activity causes a continuous formation of reactive oxygen species (ROS). In addition, drug metabolism and cell injury with inflammation can dramatically increase the oxidant stress burden for each individual cell and the organ. This review will focus on the general discussion of reactive oxygen and peroxynitrite formation, description of antioxidant systems in different cellular and vascular compartments, and analysis of potential adverse consequences of excessive oxidant stress in the liver. REACTIVE OXYGEN AND NITROGEN INTERMEDIATES Molecular oxygen (3O2) can be reduced by one-electron steps to superoxide ðOK 2 Þ, hydrogen peroxide (H2O2), the hydroxyl radical (OH%), and then to water (Fig. 1). Superoxide is not very stable and dismutates rapidly to form hydrogen peroxide and singlet oxygen (1O2), another ROS. However, in the presence of nitric oxide (%NO), superoxide reacts preferably with NO to generate peroxynitrite (ONOOK). The rate of peroxynitrite formation depends on the concentrations of both NO and superoxide (first-order kinetics) and this reaction is near diffusion controlled (1,2). With the ubiquitous presence of carbon dioxide (CO2)/bicarbonate in vivo, peroxynitrite reacts rapidly with CO2 to form reactive intermediates, which are highly effective oxidizing and nitrating species (1). In addition, peroxynitrite can be protonated to form peroxynitrous acid (ONOOH), which is a powerful oxidant. Hydrogen peroxide can be reductively cleaved to the extremely reactive hydroxyl radical in the presence of transition metals (Fenton reaction). However, if phagocytes release myeloperoxidase (MPO), then hypochlorous acid (HOCl), another potent oxidant, is generated. In addition to the described primary reactive intermediates (Fig. 1), a number of secondary radicals can also be formed, e.g., alkyl (R%), peroxy (ROO%), and alkoxy (RO%) radicals. In general, the secondary radicals are less reactive and more selective in their target. Formation and steady-state concentrations of any of these described reactive oxygen and nitrogen species in vivo are dependent on a number of factors including the formation rates of the precursors, detoxification reactions, pH, and availability of transition metals. INTRACELLULAR AND VASCULAR SOURCES OF OXIDANTS Mitochondria Superoxide and hydrogen peroxide are the main initial ROS generated in all liver cell types and the vascular space. A major continuous intracellular source of superoxide formation is the electron transport chain of mitochondria (3). Approximately 2% of total oxygen utilized in a cell is reduced to superoxide (3). NADH dehydrogenase (complex I) and ubiquinone–cytochrome b

34

Jaeschke

OCI– H2O

MPO 3

O2

e

•

O2

–

e • – O2

– CI

H2O2 1

Fe

O2

e 2+

OH Fe

•

e

H2O

3+

•

NO

ONOO – ONOOH

CO2

•

NO2CO3

•

–

– NO2 + CO3

•

•

FIGURE 1 Reactive oxygen species generated by oneelectron reduction steps of molecular oxygen (3O2). During the spontaneous dismutation of superoxide ðOK 2 Þ, singlet oxygen (1O2) is formed. If nitric oxide (†NO) is present, superoxide reacts with NO and forms peroxynitrite (ONOOK) and peroxynitrous acid (ONOOH). Peroxynitrite reacts with carbon dioxide to generate nitrating species such as the (†NO2) radical. Neutrophils release the enzyme myeloperoxidase (MPO) to form hypochlorite (OClK).

complex (complex III) release superoxide even under physiological conditions. The highest mitochondrial superoxide formation is observed during slow resting state 4 respiration, i.e., when the components of the respiratory chain are mainly in the reduced form (3,4). Mitochondrial superoxide can increase substantially when mitochondria are damaged. When superoxide is released from the electron transport chain, it can combine with NO to form peroxynitrite. It was postulated that mitochondria contain a nitric oxide synthase (NOS) (5). However, the existence of a mitochondrial NOS was questioned (6); nevertheless, peroxynitrite formation in mitochondria can occur when NO from an extramitochondrial source diffuses into mitochondria (6). In addition to the electron transport chain of the inner mitochondrial membrane, monoamine oxidases, which are located in the outer membrane, generate substantial amounts of hydrogen peroxide during the oxidative deamination of biogenic amines (7). Because of the localization, monoamine oxidases contribute to the oxidant stress in mitochondria and in the cytosol. A mitochondrial oxidant stress has been demonstrated in connection with mitochondrial dysfunction during hypoxia–reoxygenation (8), chemical hypoxia (9), extracellular oxidant stress (10), and the toxicity of ethanol (11) and bile acids (12). In the case of acetaminophen hepatotoxicity, reactive oxygen (13) and peroxynitrite formation (14) has been localized to the mitochondria. Microsomes During phase I metabolism of xenobiotics, the microsomal P450 enzyme system can release activated oxygen intermediates. The formation of hydrogen peroxide and superoxide has been documented in isolated microsomes (15). However, during in vivo drug metabolism there is little evidence for increased oxidant stress (16,17), suggesting less leakage of ROS from cytochrome P450 enzymes in the intact cell than in isolated microsomes. On the other hand, cytochrome P450 2E1 is considered a major source of ROS during metabolism of ethanol in hepatocytes (18). In addition, drug metabolism can lead to secondary oxidant stress in the liver, e.g., injury to mitochondria (13) and mobilization of transition metals (19). A severe oxidant stress can be generated by metabolism of redox-cycling agents such as diquat (20), paraquat (21), and menadione (22). These compounds are reduced by P450 reductase to a radical species, which can reduce oxygen to superoxide thereby regenerating the parent compound. Redoxcycling agents can undergo numerous cycles before they are excreted and create an enormous oxidant stress and cause severe liver damage (20). Peroxisomes These cell organelles contain a number of oxidases, e.g., fatty acyl-CoA oxidase, amino acid oxidase, and urate oxidase, which generate hydrogen peroxide as a regular product. Owing to the very high levels of catalase in peroxisomes, the adverse effects are limited under physiological conditions. However, high fat diet and drugs that are peroxisome proliferators cause an increase in fatty acyl-CoA oxidase and potentiate the oxidant stress in this cell organelle (23,24).

Antioxidant Defense in Liver Injury

35

Cytosol Xanthine dehydrogenase is a major enzyme in the cytosol of all liver cells. Although the specific activity of this enzyme is the same in all three cell types, hepatocytes contain more than 85% of the total enzyme activity in the liver (25). Prolonged periods of ischemia (26) and certain drug toxicities (13) can cause a proteolytic cleavage of the enzyme, resulting in the loss of the capability to bind the cofactor NADC. Instead, the enzyme acts as an oxidase by using molecular oxygen as electron acceptor, which leads to formation of superoxide and hydrogen peroxide in the cytosol. Some time ago, xanthine oxidase (XO) was considered the main intracellular source of reactive oxygen formation during ischemia–reperfusion. However, it is questionable if XO can actually generate a quantitatively relevant oxidant stress in hepatocytes (27). The restricted availability of the substrate, hypoxanthine or xanthine, may be the limiting factor for the duration and extent of XO-mediated reactive oxygen formation (8). It was suggested that XO might be a relevant source of reactive oxygen in Kupffer cells (25) and, after release by hepatocytes and binding to endothelial cells, a major source of oxidant stress in vascular lining cells (28). Vascular Oxidant Stress Kupffer cell activation and hepatic neutrophil recruitment can occur during drug metabolism (29–32), ischemia–reperfusion (33–35), endotoxemia (36,37), alcoholic hepatitis (38), and obstructive cholestasis (39,40), and may contribute to liver injury. A number of inflammatory mediators activate and prime Kupffer cells and neutrophils for enhanced superoxide formation including activated complement factors (41), TNF-a (42), and platelet-activating factor (43). Kupffer cells are in a fixed position within the sinusoidal lining. Superoxide, generated by NADPH oxidase in Kupffer cells, is released into the sinusoidal lumen and the space of Disse. Because of the close proximity to other cells, Kupffer cell-derived reactive oxygen can directly cause cell injury, which can be inhibited by vascular antioxidants (44). In contrast to Kupffer cells, neutrophils adherent to vascular endothelial cells release cytotoxic mediators only when excessively stimulated. However, this is rarely the case under realistic pathophysiological conditions in vivo. Injury occurs mainly after chemotactic stimulation, transmigration, and adherence of the neutrophil to hepatocytes (45,46). These processes require a number of adhesion molecules including b2 integrins and intercellular adhesion molecule-1 (ICAM-1) (47). Upregulation of the b2 integrin Mac-1 (CD11b/CD18) and the adhesion through this receptor (48) is critical for neutrophil-induced reactive oxygen formation. In support of this hypothesis, enhanced Mac-1 expression was shown in every model where neutrophils contribute to liver injury (39,49–53). In addition, antibodies against Mac-1 attenuated the postischemic oxidant stress by neutrophils and protected against neutrophil-induced liver injury (36,50). Furthermore, the neutrophil-induced oxidant stress and injury were attenuated in CD18 gene knockout mice after obstructive cholestasis (39) or a-naphthyl-isothionate treatment (54). The role of reactive oxygen in the mechanism of neutrophil-mediated cell killing remains controversial. Coculture systems showed that activated neutrophils damage hepatocytes by protease release and not oxidant stress over a time frame of 15 hour in vitro (55,56). However, recent findings suggest that reactive oxygen generated by transmigrated and adherent neutrophils causes not only an oxidant stress in the vasculature but also intracellularly (37,39,57). The appearance of chlorotyrosine protein adducts (37,39,40) and HOCl-modified protein epitopes (58) during neutrophil cytotoxicity clearly documents the presence of neutrophil-derived oxidants (HOCl) in the target cell. In addition, the higher susceptibility of glutathione peroxidase (GPx) gene knockout mice to neutrophil cytotoxicity indicates that neutrophils can kill hepatocytes by reactive oxygen, a process that requires not more than one hour after neutrophil attack (57). How can we explain the drastic differences between results of in vivo experiments and the coculture system in vitro? The most likely explanation for the opposite results is in the role of hepatocytes. In vivo, hepatocytes are exposed to the same inflammatory mediators as neutrophils resulting in the upregulation of adhesion molecules such as ICAM-1 (40,49,59) as well as the formation and release of CXC chemokines (60,61).

36

Jaeschke

Chemokines and ICAM-1 are important for neutrophil chemotaxis and adherence to hepatocytes (60), which is the final activating step for neutrophil degranulation and long-lasting adherence-dependent reactive oxygen formation. In contrast, all coculture experiments were done with control hepatocytes. Since neutrophils do not attack healthy cells (37,46), they do not firmly adhere in vitro and the cytotoxicity is dependent on the excessive stimulation with inflammatory mediators. Although reactive oxygen may be generated, it may not be long enough to cause toxicity and not in close proximity to hepatocytes. Therefore, the cytotoxicity in vitro is mainly caused by slow proteolytic digestion and not by the rapid targeted killing by reactive oxygen (45,46).

PATHOPHYSIOLOGICAL CONSEQUENCES OF OXIDANT STRESS Lipid Peroxidation Oxidant stress in the liver can cause lipid peroxidation (LPO), which is still a frequently hypothesized mechanism of cell injury. However, on a quantitative basis, the magnitude of LPO in vivo is mostly insufficient to directly cause cell death (62). Excessive intracellular superoxide formation alone does not kill hepatocytes by LPO even after depletion of glutathione (GSH) (63). Extensive hepatic LPO in vivo was only observed when, in addition to reactive oxygen formation, the cellular antioxidants, e.g., vitamin E and GSH, were depleted (64,65), high levels of polyunsaturated fatty acids were present in membranes (65), and iron was mobilized from intracellular stores (19). If some of these factors come together, massive LPO with severe cell injury ensues. However, under most realistic pathophysiological conditions, LPO is minimal and is not likely to be directly responsible for cell damage. Does this mean LPO is not important? LPO products are potent chemotactic factors for neutrophils and can modulate superoxide formation (66). In addition, LPO products may enhance chemokine formation (67). These observations may explain the role of LPO products in maintaining an inflammatory response beyond the initial mediator formation (68). Furthermore, LPO products were shown to promote induction of collagen gene expression in activated stellate cells and contribute to fibrosis (69). Thus, LPO products can be important signaling molecules under certain pathophysiological conditions. Peroxynitrite Formation The reaction of peroxynitrite with carbon dioxide yields nitrating species which react preferentially with tyrosine (70). During acetaminophen hepatotoxicity, a mitochondrial oxidant stress (13) and NO derived at least in part from the inducible NOS (71) cause peroxynitrite formation in mitochondria (14). Aided by the massive depletion of cytosolic and mitochondrial GSH, peroxynitrite proved to be a critical mediator in the mechanism of cell injury (71,72) and trapping of it did not only strongly attenuate cell death but also promoted cell cycle activation and regeneration (73). Although the detailed mechanisms are not completely understood, peroxynitrite appears to be involved in the mitochondrial membrane permeability transition (MPT) pore opening (74). The MPT leads to breakdown of the mitochondrial membrane potential and ATP depletion (75) and the release of intermembrane proteins such as endonuclease G and apoptosis-inducing factor, which translocate to the nucleus and cause DNA fragmentation (14,76). These events are essential for the development of acetaminopheninduced necrosis (77). Nitrotyrosine residues were also detected in the liver during hepatic ischemia–reperfusion injury (78,79) and during alcohol-induced liver injury (80), but the pathophysiological relevance is less clear (81). Since ischemic stress promotes vasoconstriction in the hepatic vasculature (82), NO formation in the liver is also critical for maintenance of liver blood flow (83,84). In many situations, peroxynitrite formation, especially in the presence of scavengers such as GSH and NADH, is less damaging than prolonged vasoconstriction and ischemia (84,85).

Antioxidant Defense in Liver Injury

37

Reactive Oxygen and Cell Death Reactive oxygen can cause hepatocellular necrosis without gross cell damage by LPO. The mechanism of this necrotic cell death is linked to the opening of the MPT pore, which causes mitochondrial uncoupling and loss of the membrane potential (86). A significant oxidant stress causes oxidation of mitochondrial pyridine nucleotides and the formation of ROS in mitochondria, both of which increase mitochondrial free Ca2C(10). The MPT can be induced by an increase in mitochondrial Ca2C directly (87) or through the activation of mitochondrial serine proteases (calpains) (88). Cytosolic calpains can induce membrane blebbing by degrading cytoskeletal proteins (89). These events lead to rapid necrotic cell death of hepatocytes within one hour (10,86). Reactive oxygen can also induce cell death through apoptosis (90–94). Generation of superoxide with menadione triggers caspase-mediated apoptotic cell death dependent on activation of the MAPkinase JNK in both hepatocyte cell lines (92) and primary cultured hepatocytes (22). Induction of ERK1/2 attenuated superoxide-induced apoptosis (22,92). In contrast to superoxide, higher levels of H2O2 induce necrotic cell death in hepatocytes (22). The reason for the lack of apoptosis after H2O2 exposure may be the oxidative inhibition of caspases by high concentrations of H2O2 (95). This mechanism may be responsible for the delayed apoptosis of activated neutrophils at an inflammatory site (96). In addition, since only H2O2 and HOCl generated by neutrophils can actually cross cell membranes and induce an intracellular oxidant stress in target cells, this could explain why neutrophils cause necrotic cell death and not apoptosis (37,97). Reactive Oxygen and Gene Transcription The activation of many transcription factors including nuclear factor-kappa B (NF-kB) and activating protein-1 (AP-1) can be induced or modulated by ROS (98,99). Despite extensive experimental data, the molecular mechanism of the redox sensitivity of these transcription factors is still not completely clear (100). However, a number of these transcription factors have thiols, which are critical for their activity, and are known to be regulated at least in part by thioredoxin (Trx) (98,99,101). Pro-inflammatory cytokines (e.g., TNF-a, interleukin (IL)-1), chemokines (e.g., IL-8), adhesion molecules (e.g., ICAM-1, VCAM-1 and E-selectin), and stress genes (e.g., heme oxygenase-1 (HO-1)) are regulated by these redox-sensitive transcription factors. Therefore, reactive oxygen can significantly enhance an inflammatory response thereby indirectly contributing to cell damage. In the liver, TNF-a formation can be modulated by oxidant stress. Antioxidants inhibited endotoxin-induced NF-kB activation and the formation of TNF mRNA and protein in isolated Kupffer cells (102). In support of these results, TNF-a formation in vivo could be prevented by the radical scavenger dimethyl sulfoxide (103) and endotoxin-induced TNF-a generation was three times higher in GPx knockout mice (57). Furthermore, increased nonheme iron concentrations and increased reactive oxygen formation in hepatic macrophages isolated from alcohol-treated animals responded to endotoxin exposure with higher NF-kB activation and elevated transcription of cytokines and chemokines (104). Together these data clearly indicate that gene transcription in Kupffer cells can be modulated by reactive oxygen. However, not only pro-inflammatory genes, but also genes which protect against oxidant stress can be activated by ROS (101). In general, these genes have an antioxidant response element (ARE) within their promoter. AREs are regulated by the transcription factor NF-E2-related factor 2 (Nrf2), which translocates from the cytosol into the nucleus. Nrf2 is normally associated with the cytoplasmic inhibitor, Kelchlike ECH-associated protein 1 (Keap1) in the cytosol. When thiols of Keap1 are oxidized, Nrf2 is released and can enter the nucleus (105). Nrf2 activation occurs after acetaminophen treatment and limits liver cell injury possibly through upregulation of HO-1 and other protective genes (106,107). Induction of HO-1 in hepatocytes during hemorrhagic shock and resuscitation are dependent on the activation of AP-1 (108). Both the activation of AP-1 and the induction of HO-1 could be inhibited by antioxidants (109). HO-1 generates the antioxidant biliverdin and the vasodilator carbon monoxide, both of which may contribute to the hepatoprotective effect of HO-1 induction (109,110).

38

Jaeschke

Stellate cells regulate sinusoidal blood flow and are, after transformation, the major source of extracellular matrix protein formation leading to fibrosis. There is considerable evidence that reactive oxygen and LPO can stimulate fibrogenesis (69,111). ROS induce or modulate transforming growth factor b1-induced collagen a1 (I) gene expression in vivo (112,113). Furthermore, reactive oxygen can stimulate chemokine transcription in stellate cells (114) thereby enhancing the inflammatory response. Although there is extensive evidence that ROS can directly cause cell death or contribute to modulation of transcription factor activation and gene expression in liver cells, it should be kept in mind that much of the data on ROS toxicity and redox regulation generated in cultured cells have to be interpreted with caution. Gene expression profiles of primary cultured hepatocytes change rapidly after isolation and are considerably different from the intact liver in vivo (115). Since liver cells in culture are in a state of oxidant stress due to hyperoxia (21% oxygen), adaptive changes in response to these artificial culture conditions may modulate the susceptibility to oxidant stress and trigger redox regulation of signaling pathways, which may have limited relevance in vivo (116). ANTIOXIDANT DEFENSE SYSTEMS The continuous formation of reactive oxygen and reactive nitrogen species during physiological functions of liver cells and the potential for a substantially increased oxidant stress under many pathophysiological conditions (Fig. 1) require an effective defense system against these reactive intermediates. Because of the variety of oxygen and nitrogen metabolites formed and their different localization and reactivity, a sophisticated, multilevel network of antioxidant enzymes and small molecules is operative in every liver cell (Fig. 2.). Enzymatic Defense Mechanisms Superoxide is removed by superoxide dismutases (SODs) in the major cellular compartments (117). Cu2C/Zn2C-SOD (SOD1) is mainly located in the cytosol and nuclear matrix whereas Mn3C-SOD (SOD2) is present in the mitochondria (118). Superoxide first reduces the redoxactive metal (Cu2C or Mn3C) which yields molecular oxygen; a second superoxide molecule is then reduced to hydrogen peroxide by the metal ion. The reaction of superoxide with SOD is diffusion limited. The high intracellular SOD levels (approx. 10 mM) keep the steady-state levels of superoxide in the range of 1 to 10 pM (2,4). Since superoxide is not a very toxic molecule by itself and the spontaneous dismutation has the same reaction products, then why is it beneficial to have these high levels of SOD? In addition to the fact that SOD catalyzed dismutation avoids the formation of singlet oxygen (Fig. 1), the main reason for the importance of SOD might be to limit peroxynitrite generation (1,2). Using the rate constants for the reaction of superoxide with SOD (2.4!109/M/sec) and NO (2!1010/M/sec), the rate of disappearance for superoxide is

2H2O Catalase

·– O2

Superoxide dismutases

·

O2, A H2O2, AH2

H2O2

Ferritin Fe2+

NO 2 GSH

GSSG

ONOO – H+

ONOO H

Glutathione peroxidase 2 GSH

GSSG

·

OH

Vitamin E

Fe3+ Glutathione peroxidase

2H2O

NO2– +H2O

LPO

FIGURE 2 Cellular antioxidant defense mechanisms include enzymes for the rapid metabolism of reactive oxygen and nitrogen species, binding proteins for transition metals (e.g., ferritin), and chain-breaking antioxidants (e.g., vitamin E). Abbreviations: GSH, glutathione; GSSG, glutathione disulfide; LPO, lipid peroxidation.

Antioxidant Defense in Liver Injury

39

20,000/sec with SOD (10 mM) and 200/sec with NO (estimated physiological concentrations: 10 nM) (1). Thus, under physiological conditions, SOD prevents peroxynitrite formation. However, if the NO concentrations are increased (10 mM), e.g., during an inflammatory response, the reaction with NO increases to 40,000/sec (1). This indicates that SOD cannot prevent intracellular peroxynitrite formation under inflammatory conditions. This situation is even more critical in the extracellular space where SOD levels are much lower. Hydrogen peroxide (H2O2) formed by SOD is degraded by catalase, GPx, or peroxiredoxin (Prx). Most of the catalase enzyme activity is located in peroxisomes. Mammalian catalase is a hemeprotein which reduces hydrogen peroxide to water by utilizing electrons from either hydrogen peroxide (catalase reaction) or from other small molecules such as ethanol or methanol (peroxidase reaction) (119). Catalase is inducible in the liver by caloric restriction, phenobarbital, and hypolipidemic drugs, e.g., clofibrate (119). The main function of catalase is to metabolize hydrogen peroxide generated by oxidases in peroxisomes. Only under extreme conditions will any relevant amount of hydrogen peroxide escape peroxisomes or will significant amounts of cytosolic hydrogen peroxide be detoxified by catalase (120). Cellular glutathione peroxidase (cGPx-1), one of the four known selenium-dependent GPxs, is ubiquitously expressed (121). The other three forms include GPx-2 (gastrointestinal GPx), GPx-3 (plasma GPx), and GPx-4 (phospholipid hydroperoxide GPx) (121). cGPx-1 is located in the cytosol (75%) and mitochondria (25%). The enzyme contains selenium in the form of selenocysteine, which is critical for the catalytic function (121). cGPx-1 can reduce peroxides, including hydrogen peroxide and organic peroxides (121). Although there is evidence in vitro that cGPx-1 can also reduce peroxynitrite (122), the peroxynitrite-dependent injury after acetaminophen overdose was not enhanced in GPx-1 gene knockout mice (72). In contrast to its low specificity for peroxide substrates, the enzyme requires GSH as cofactor. Glutathione disulfide (GSSG) is rapidly reduced back to GSH by glutathione reductase and NADPH. Since the reductase is the rate-limiting step of this cycle, GSSG accumulates to some degree and can be excreted into bile and plasma (16). The removal of GSSG either by the reductase (O95%) or export from hepatocytes (!5%) protects protein sulfhydryl groups from oxidation by high levels of GSSG. Mitochondria take up and release GSH but are not able to export GSSG (123). Consequently, reduction of GSSG within the mitochondria is the only option to prevent GSSG accumulation. As in the case of acetaminophen toxicity, this cannot always be avoided (13). A member of the glutathione-S-transferase family (GST-B) was identified as the enzyme responsible for the GPx activity in selenium-deficient mice (124). GST-B, which is inducible by selenium deficiency (125), can only metabolize organic hydroperoxide but not hydrogen peroxide (126). However, despite this adaptation, selenium-deficient animals are more susceptible to reactive oxygen-induced liver injury (127). Another enzyme, GPx-4, which selectively uses lipid hydroperoxides as substrates, was identified (128). This enzyme is located in the mitochondria, nuclei, and microsomes and is involved in the metabolism of peroxidized lipids and therefore inhibits the propagation of LPO (129). Trxs and Prxs In mammalian cells, thioredoxin-1 (Trx-1) is located intracellularly in the cytosol, but is also found in the nucleus, and extracellularly in plasma (101). Trx-2 is detected primarily in mitochondria. Trx contains redox-active dithiols which can be oxidized to the corresponding disulfide (Fig. 3). The oxidized form of Trx-1 and -2 is reduced by the respective thioredoxin reductase-1 (TR1) and -2 (TR2) (101). Multiple functions of these enzymes have been reported, which are mainly complementary to the GSH/GPx system (101). One important function of Trx is to reverse oxidative changes in proteins including reduction of protein disulfides (Fig. 3). This function does not only restore protein function, but can also regulate the activity of certain transcription factors. Reduction of critical thiol groups by Trx increases the DNA binding of NFkB and AP-1 after nuclear translocation (130). Trx can function as antioxidant by reducing hydrogen peroxide and peroxynitrite (101,131). However, this requires the combined activity of Trx and peroxiredoxins (Prxs) (Fig. 4) (131). Mammalian cells express at least six Prx isoforms (Prx I–VI) (131). Prx I, II, and VI are located in the cytosol, Prx III and V reside in the mitochondria, and Prx IV is present in the extracellular space. Common to all Prx family

40

Jaeschke

Trx

Pr-SH

S

Pr-S-S-Pr

S

TR

TR

SeH SH

FAD

TR

TR

Trx SH SH

Se

S

NADPH

NADP

FADH2

+

FIGURE 3 Reduction of protein disulfides by thioredoxin (Trx). Trx is regenerated by thioredoxin reductase, (TR) which uses electrons derived from NADPH. Source: Adapted from Ref. 101.

members is a conserved cysteine residue near the N-terminal, which is the primary cysteine oxidized by hydrogen peroxide. In this reaction, the sulfhydryl group (cys-SH) is oxidized to the sulfenic acid (cys-SOH) as intermediate, which is then reduced by a second cysteine residue to the disulfide (Prx-S–S-Prx; Fig. 4). This disulfide can be reduced by Trx. Due to the high concentration of Prx in the cytosol (0.1–0.8% of soluble proteins), the Prx/Trx system is even more effective than GPx/GSH system in removing low levels of hydrogen peroxide (131,132). However, under certain conditions, Prx-SOH can be further oxidized to the sulfinic acid form (Prx-SO2H), which can no longer be reduced by Trx (131). Prx can be recovered through reduction by another enzyme, sulfiredoxin (Fig. 4) (133). However, this reaction is slow. Thus, hyperoxidation results in a temporary inactivation of Prx, which may allow accumulation of hydrogen peroxide for signal transduction purposes (132). Low Molecular Weight Antioxidants Ascorbate (vitamin C), a-tocopherol (vitamin E), and GSH are examples of low molecular weight antioxidants. a-Tocopherol is the most effective chain-breaking compound in biological membranes (134) However, as for most antioxidants, the intracellular concentrations of a-tocopherol are not high enough to be a relevant hydroxyl radical scavenger. However, it effectively reduces peroxyl radicals (ROO%), one of the less reactive secondary radicals, to the lipid hydroperoxide which can then be metabolized by GPx-4 (134). Thus, a-tocopherol prevents the propagation of the radical chain by avoiding the formation of new alkyl radicals. The a-tocopherol radical can be reduced by ascorbate and thiols such as GSH (135). Ascorbate is regenerated in the aqueous phase by a GSH-dependent dehydroascorbate reductase or a ATP R-SH ADP R-S-S-R

Prx

Sulfiredoxin

SO2H Sulfinic acid

H2O2 H2O2

Prx SH

Prx

Prx-SH

Prx

SOH

S

Sulfenic acid

S Prx

Trx S

S Trx SH SH

FIGURE 4 Detoxification of hydrogen peroxide (H2O2) by peroxiredoxin (Prx). During the catalytic cycle, the sulfenic acid is reduced by Prx-SH and thioredoxin (Trx). However, excess H2O2 may further oxidize Prx to the sulfinic acid, which is inactive. Sulfiredoxin can regenerate the active enzyme in a multistep process. Source: Adapted from Ref. 131.

Antioxidant Defense in Liver Injury

41

NADH-dependent semidehydroascorbate reductase (134). Thus, low molecular weight antioxidants act together to interrupt radical chain reactions and to divert radicals away from sensitive areas, e.g., hydrophobic membranes, to the aqueous phase (136). The critical importance of these compounds as defense systems to protect membranes has been shown in acetaminophen hepatotoxicity. Despite the substantial mitochondrial oxidant stress and peroxynitrite formation (13,14,72), LPO is not a relevant mechanism of cell injury for this drug in vivo (137). However, in animals fed with vitamin E-deficient diet, LPO becomes the predominant injury mechanism with complete destruction of the liver within 1 to 4 hour after administration of acetaminophen (64). GSH is the most important water-soluble antioxidant (138). It is used as a cofactor for GPxs and GSTs and is important for the maintenance of protein sulfhydryl groups. GSH is also an effective scavenger of peroxynitrite and HOCl. The key functional group of GSH is the cysteine sulfhydryl moiety, which is less susceptible to autoxidation than the same moiety in the isolated amino acid. GSH is synthesized intracellularly by two ATP-dependent enzymes: g-glutamylcysteine synthetase and glutathione synthetase. Due to its protease resistant g-glutamyl bond, GSH cannot be degraded by intracellular proteases. Thus, for the cellular turnover, GSH has to be exported from the cell and is degraded by g-glutamyltranspeptidase, which is present on the surface of epithelial cells in the kidney, lung, and intestine and also in the biliary tract. Approximately 90% of the GSH in plasma is supplied by the sinusoidal GSH release of hepatocytes (139). Cellular export proteins for GSH include members of the multidrug resistance-associated proteins (Mrps), which are members of the ATP-binding cassette superfamily, and organic anion-transporting polypeptides (Oatps) (140). Mrps use ATP and Oatps use an electrochemical gradient as the driving force (140). Mrp2 appears to be the main carrier for GSH in the canalicular membrane. Oatp1 and 2 contribute to the export across the sinusoidal membrane (140). Mrp4, located at the basolateral membrane, is responsible for cotransport of GSH with bile acids (141). In addition to the biliary export of GSH, Mrp2 is also the main carrier for GSSG and GSH conjugates (142,143). Much less is known about a sinusoidal GSSG transport. However, functional studies showed a release of GSSG into the sinusoids in the isolated perfused liver (63,144) as well as in vivo (145). Quantitative estimations of GSSG formation and release even during severe oxidant stress indicated that 1% to 5% of all GSSG formed is exported and 95% to 99% is reduced by glutathione reductase (63). Approximately 80% of the exported GSSG is released into bile and 20% into the sinusoid (144). Thus, the biliary GSSG efflux is the most sensitive marker of intracellular oxidant stress. GSH is not only present in the cytosol, but also in other cellular compartments including mitochondria (15% of total cellular GSH) (146). However, GSH is not synthesized in mitochondria but has to be taken up from the cytosol. The carriers involved include the dicarboxylate and 2-oxoglutarate carriers in the inner mitochondrial membrane (147). The transport of GSH into the mitochondrial matrix, which occurs against an electrochemical gradient, is facilitated by the release of phosphate (dicarboxylate carrier) or oxoglutarate (oxoglutarate carrier) (147). A depletion of mitochondrial GSH impairs the detoxification mechanisms in this cell organelle and can lead to increased oxidative injury, loss of mitochondrial function, and cell death (146–149). Reactive oxygen escaping from mitochondria can induce NF-kB activation and promote gene transcription (149). On the other hand, any GSSG formed in the mitochondria through the activity of GPx cannot be exported into the cytosol (122); GSSG has to be reduced or it will accumulate as has been shown during acetaminophen toxicity (13). Metal-Binding Proteins Free radical processes such as LPO are dependent on the availability of redox-active transition metals. Therefore, another defense strategy is to keep metal ions such as Fe2C/Fe3C or CuC tightly bound to transport or storage proteins. Metal-binding proteins include ferritin, transferrin, and lactoferrin for iron, ceruloplasmin for copper, and metallothionein for other metals (150). Because of its large number of cysteines, metallothionein may also act directly or indirectly as antioxidant (151).

42

Jaeschke

ANTIOXIDANT DEFENSE IN NONPARENCHYMAL CELLS The previous sections described various antioxidant strategies in the liver, i.e., hepatocytes. However, a limited number of studies suggest that similar systems are operative in nonparenchymal cells (152,153). In general, activities of SODs and selenium-dependent GPx are similar in nonparenchymal cells and in hepatocytes (152). Likewise, the GSH contents on a nmol/mg cellular protein basis are very similar in all cell types. However, due to the much smaller cell size of nonparenchymal cells compared to hepatocytes, the total GSH content in nonparenchymal cells represents less than 5% of the liver GSH content (152). Thus, the detoxification capacity for ROS is only a fraction of hepatocytes. Interestingly, Kupffer and endothelial cells adapt differently to an inflammatory stimulus. During endotoxemia, Kupffer cells modulate pro-oxidant pathways resulting in increased superoxide formation (154). In contrast, endothelial cells upregulate SOD and GPx activities, the glucose transporter GLUT1, and the key enzymes of the carbohydrate metabolism (153,154). This response supports the detoxification potential for reactive oxygen in endothelial cells and helps to maintain the integrity of the vascular lining cells. ANTIOXIDANT DEFENSE IN THE VASCULAR SPACE Inflammatory cells can release reactive oxygen and nitrogen species into the vascular space and generate a substantial oxidant stress. Plasma antioxidants include albumin, transferrin, lactoferrin, ceruloplasmin, haptoglobin, urate, ascorbate, vitamin E, bilirubin, and extracellular SOD and GPx activities (reviewed in Ref. 155). However, the main problems of plasma antioxidant systems are the low concentrations of the scavengers and the low activities of the enzymes, which makes them much less effective than their intracellular counterparts. The exceptions are the metal transport proteins which bind metals with high affinity and, therefore, virtually eliminate free iron from plasma (155). Extracellular Cu2C/Zn2C-SOD (EC-SOD, SOD3) (156) and extracellular selenium-dependent GPx (eGPx) (121,157) are proteins distinct from the cellular enzymes. However, their biological relevance is not clear. EC-SOD can bind to surface proteoglycans on the endothelial cell surface (158). It can be speculated that this SOD may protect endothelial cells from a vascular oxidant stress induced by phagocytes and may limit peroxynitrite formation. Plasma eGPx is dependent on the cofactor GSH. In contrast to the mM Km value of the enzyme (157), plasma GSH concentrations are generally in the 5 to 200 mM range (139,159). Thus, it would be expected that eGPx is not very effective in removing peroxides from plasma. A more liver-specific antioxidant defense system has recently been recognized (160,161). GSH can be oxidized in the vasculature during ischemia–reperfusion (34,160) and endotoxemia (159), reflecting a Kupffer cell-induced oxidant stress. In this situation, the plasma GSH levels are substantially increased (100–200 mM in the hepatic vein) due to enhanced release of GSH from hepatocytes (159). These elevated levels, which may even be higher in the space of Disse, are sufficient to make GSH a fairly effective trapping agent for ROS. In support of this conclusion, depletion of plasma GSH concentrations aggravated inflammatory liver injury (160) and an increase above baseline levels protected against injury (161,162). The oxidation of GSH occurred in the vascular space (34) and was not enzymatically catalyzed (160). In vitro studies showed that only H2O2 reacts with GSH and forms GSSG (161). Other relevant oxidants, i.e., hypochlorite and peroxynitrite, are trapped by GSH but generate mainly higher oxidation states such as sulfenic or sulfinic acid and very little GSSG (72). These data are supported by in vivo experiments which showed that NOS inhibitors increased plasma GSH and GSSG levels during hepatic inflammation (85,86). In contrast, NO donors decreased GSH and GSSG concentrations in plasma (86). This suggests that when NO and superoxide are generated in the vascular space, at least some of these molecules react to form peroxynitrite which can be trapped by GSH. Thus, GSH appears to be an important scavenger of reactive oxygen and nitrogen species in the vascular space under inflammatory conditions. Another important antioxidant system in plasma is selenoprotein P (163). The plasma concentrations of this protein are in the range of 25 to 30 mg/mL. The protein contains 7 to 10 selenocysteine and 17 cysteine residues in each molecule (163). Thus, it should be able to

Antioxidant Defense in Liver Injury

43

function as a scavenger for reactive oxygen and peroxynitrite. Since many different organs can produce and release this protein, it was postulated that selenoprotein P acts as an antioxidant in the interstitial space (163). The importance of this selenoprotein P was demonstrated in experiments where the redox-cycling agent diquat or GSH depletion with phorone induced LPO and liver injury in selenium-deficient animals (164,165). Treatment of these animals with a dose of selenium, which was sufficient to restore selenoprotein P levels in plasma but did not affect the low levels of GPx in plasma or liver tissue, prevented LPO and liver injury (164,165). These observations suggest a significant role of selenoprotein P as antioxidant in plasma. REFERENCES 1. Squadrito GL, Pryor WA. Oxidative chemistry of nitric oxide: the role of superoxide, peroxynitrite and carbon dioxide. Free Radic Biol Med 1998; 25:392–403. 2. Koppenol WH. The basic chemistry of nitrogen monoxide and peroxynitrite. Free Radic Biol Med 1998; 25:385–91. 3. Turrens JF. Mitochondrial formation of reactive oxygen species. J Physiol 2003; 552:335–44. 4. Cadenas E, Davies KJA. Mitochondrial free radical generation, oxidative stress and aging. Free Radic Biol Med 2000; 29:222–30. 5. Tatoyan A, Giulivi C. Purification and characterization of a nitric-oxide synthase from rat liver mitochondria. J Biol Chem 1998; 273:11044–8. 6. Tay YM, Lim KS, Sheu FS, et al. Do mitochondria make nitric oxide? no? Free Radic Res 2004; 38(6):591–9. 7. Hauptmann N, Grimsby J, Shih JC, et al. The metabolism of tyramine by monoamine oxidase A/B causes oxidative damage to mitochondrial DNA. Arch Biochem Biophys 1996; 335:295–304. 8. Jaeschke H, Mitchell JR. Mitochondria and xanthine oxidase both generate reactive oxygen species after hypoxic damage in isolated perfused rat liver. Biochem Biophys Res Commun 1989; 160:140–7. 9. Gores GJ, Flarsheim CE, Dawson TL, et al. Swelling, reductive stress, and cell death during chemical hypoxia in hepatocytes. Am J Physiol 1989; 257:C347–54. 10. Nieminen AL, Byrne AM, Herman B, et al. Mitochondrial permeability transition in hepatocytes induced by t-BuOOH—NAD(P)H and reactive oxygen species. Am J Physiol 1997; 272:C1286–94. 11. Kukielka E, Dicker E, Cederbaum AI. Increased production of reactive oxygen species by rat liver mitochondria after chronic ethanol treatment. Arch Biochem Biophys 1994; 309:377–86. 12. Sokol RJ, Winklhofer-Roob BM, Devereaux MW, et al. Generation of hydroperoxides in isolated rat hepatocytes and hepatic mitochondria exposed to hydrophobic bile acids. Gastroenterology 1995; 109:1249–56. 13. Jaeschke H. Glutathione disulfide formation and oxidant stress during acetaminophen-induced hepatotoxicity in mice in vivo: the protective effect of allopurinol. J Pharmacol Exp Ther 1990; 255:935–41. 14. Cover C, Mansouri A, Knight TR, et al. Peroxynitrite-induced mitochondrial and endonucleasemediated nuclear DNA damage in acetaminophen hepatotoxicity. J Pharmacol Exp Ther 2005; 315(2):879–87. 15. Kuthan H, Ullrich V. Oxidase and oxygenase function of the microsomal cytochrome P-450 monooxygenase system. Eur J Biochem 1982; 126:583–8. 16. Lauterburg BH, Smith CV, Hughes H, et al. Biliary excretion of glutathione and glutathione disulfide in the rat. J Clin Invest 1984; 73:124–33. 17. Bajt ML, Knight TR, Lemasters JJ, et al. Acetaminophen-induced oxidant stress and cell injury in cultured mouse hepatocytes: protection by N-acetyl cysteine. Toxicol Sci 2004; 80(2):343–9. 18. Wu D, Cederbaum AI. Oxidative stress mediated toxicity exerted by ethanol-inducible CYP2E1. Toxicol Appl Pharmacol 2005; 207(Suppl. 2):70–6. 19. Jaeschke H, Kleinwaechter C, Wendel A. NADH-dependent reductive stress and ferritin-bound iron in allyl alcohol-induced lipid peroxidation in vivo: the protective effect of vitamin E. Chem Biol Interact 1992; 81:57–68. 20. Smith CV, Hughes H, Lauterburg BH, et al. Oxidant stress and hepatic necrosis in rats treated with diquat. J Pharmacol Exp Ther 1985; 235:172–7. 21. Brigelius R, Anwer MS. Increased biliary GSSG-secretion and loss of hepatic glutathione in isolated perfused rat liver after paraquat treatment. Res Commun Chem Pathol Pharmacol 1981; 31:493–502. 22. Conde de la Rosa L, Schoemaker MH, Vrenken TE, et al. Superoxide anions and hydrogen peroxide induce hepatocyte death by different mechanisms: involvement of JNK and ERK MAP kinases. J Hepatol 2006; 44(5):918–29. 23. Lazarow PB, De Duve C. A fatty acyl-CoA oxidizing system in rat liver peroxisomes: enhancement by clofibrate, a hypolipidemic drug. Proc Natl Acad Sci USA 1976; 73:2043–6.

44

Jaeschke

24. Conway JG, Neptun DA, Garvey LK, et al. Role of fatty acyl coenzyme A oxidase in the efflux of oxidized glutathione from perfused livers of rats treated with the peroxisome proliferator nafenopin. Cancer Res 1987; 47:4795–800. 25. Wiezorek JS, Brown DH, Kupperman DE, et al. Rapid conversion to high xanthine oxidase activity in viable Kupffer cells during hypoxia. J Clin Invest 1994; 94:2224–30. 26. Engerson TD, McKelvey TG, Rhyne DB, et al. Conversion of xanthine dehydrogenase to oxidase in ischemic rat tissues. J Clin Invest 1987; 79:1564–70. 27. Jaeschke H. Reactive oxygen and ischemia/reperfusion injury of the liver. Chem Biol Interact 1991; 79:115–36. 28. Houston M, Estevez A, Chumley P, et al. Binding of xanthine oxidase to vascular endothelium. J Biol Chem 1999; 274:4985–94. 29. Laskin DL, Gardner CR, Price VF, et al. Modulation of macrophage functioning abrogates the acute hepatotoxicity of acetaminophen. Hepatology 1995; 21:1045–50. 30. Lawson JA, Farhood A, Hopper RD, et al. The hepatic inflammatory response after acetaminophen overdose: the role of neutrophils. Toxicol Sci 2000; 54:509–16. 31. Liu ZX, Han D, Gunawan B, et al. Neutrophil depletion protects against murine acetaminophen hepatotoxicity. Hepatology 2006; 43(6):1220–30. 32. Cover C, Liu J, Farhood A, et al. Pathophysiological role of the acute inflammatory response during acetaminophen hepatotoxicity. Toxicol Appl Pharmacol 2006; 216:98–107. 33. Jaeschke H, Farhood A, Smith CW. Neutrophils contribute to ischemia/reperfusion injury in rat liver in vivo. FASEB J 1990; 4:3355–9. 34. Jaeschke H, Farhood A. Neutrophil and Kupffer cell—induced oxidant stress and ischemia– reperfusion injury in rat liver in vivo. Am J Physiol 1991; 260:G355–62. 35. Jaeschke H, Bautista AP, Spolarics Z, et al. Superoxide generation by Kupffer cells and priming of neutrophils during reperfusion after hepatic ischemia. Free Radic Res Commun 1991; 15:277–84. 36. Jaeschke H, Farhood A, Smith CW. Neutrophil-induced liver cell injury in endotoxin shock is a CD11b/CD18—dependent mechanism. Am J Physiol 1991; 261:G1051–6. 37. Gujral JS, Hinson JA, Farhood A, et al. NADPH oxidase-derived oxidant stress is critical for neutrophil cytotoxicity during endotoxemia. Am J Physiol Gastrointest Liver Physiol 2004; 287(1):G243–52. 38. Bautista AP. Chronic alcohol intoxication induces hepatic injury through enhanced macrophage inflammatory protein-2 production and intercellular adhesion molecule-1 expression in the liver. Hepatology 1997; 25:335–42. 39. Gujral JS, Farhood A, Bajt ML, et al. Neutrophils aggravate acute liver injury during obstructive cholestasis in bile duct-ligated mice. Hepatology 2003; 38(2):355–63. 40. Gujral JS, Liu J, Farhood A, et al. Functional importance of ICAM-1 in the mechanism of neutrophilinduced liver injury in bile duct-ligated mice. Am J Physiol Gastrointest Liver Physiol 2004; 286(3):G499–507. 41. Jaeschke H, Farhood A, Bautista AP, et al. Complement activates Kupffer cells and neutrophils during reperfusion after hepatic ischemia. Am J Physiol 1993; 264:G801–9. 42. Bautista AP, Schuler A, Spolarics Z, et al. Tumor necrosis factor-alpha stimulates superoxide anion generation by perfused rat liver and Kupffer cells. Am J Physiol 1991; 261:G891–5. 43. Bautista AP, Spitzer JJ. Platelet activating factor stimulates and primes the liver, Kupffer cells and neutrophils to release superoxide anions. Free Radic Res Commun 1992; 7:195–209. 44. Bilzer M, Jaeschke H, Vollmar AM, et al. Prevention of Kupffer cell-induced oxidant injury in rat liver by atrial natriuretic peptide. Am J Physiol 1999; 276:G1137–44. 45. Jaeschke H, Smith CW. Mechanisms of neutrophil-induced parenchymal cell injury. J Leukoc Biol 1997; 61:647–53. 46. Jaeschke H. Mechanisms of Liver Injury. II. Mechanisms of neutrophil-induced liver cell injury during hepatic ischemia–reperfusion and other acute inflammatory conditions. Am J Physiol Gastrointest Liver Physiol 2006; 290(6):G1083–8. 47. Jaeschke H. Cellular adhesion molecules: regulation and role in the pathogenesis of liver disease. Am J Physiol 1997; 273:G602–11. 48. Shappell SB, Toman C, Anderson DC, et al. Mac-1 (CD11b/CD18) mediates adherence-dependent hydrogen peroxide production by human and canine neutrophils. J Immunol 1990; 144:2702–11. 49. Essani NA, Fisher MA, Farhood A, et al. Cytokine-induced hepatic intercellular adhesion molecule-1 (ICAM-1) mRNA expression and its role in the pathophysiology of murine endotoxin shock and acute liver failure. Hepatology 1995; 21:1632–9. 50. Jaeschke H, Farhood A, Bautista AP, et al. Functional inactivation of neutrophils with a Mac-1 (CD11b/CD18) monoclonal antibody protects against ischemia–reperfusion injury in rat liver. Hepatology 1993; 17:915–23. 51. Witthaut R, Farhood A, Smith CW, et al. Complement and tumor necrosis factor-a contribute to Mac-1 (CD11b/CD18) upregulation and systemic neutrophil activation during endotoxemia in vivo. J Leukoc Biol 1994; 55:105–11.

Antioxidant Defense in Liver Injury

45

52. Lawson JA, Burns AR, Farhood A, et al. Functional importance of E- and L-selectin for neutrophilinduced liver injury during endotoxemia in mice. Hepatology 2000; 32:990–8. 53. Bajt ML, Farhood A, Jaeschke H. Effects of CXC chemokines on neutrophil activation and sequestration in hepatic vasculature. Am J Physiol Gastrointest Liver Physiol 2001; 281(5):G1188–95. 54. Kodali P, Wu P, Lahiji PA, et al. ANIT toxicity toward mouse hepatocytes in vivo is mediated primarily by neutrophils via CD18. Am J Physiol Gastrointest Liver Physiol 2006; 291:G355–63. 55. Mavier P, Preaux A-M, Guigui B, et al. In vitro toxicity of polymorphonuclear neutrophils to rat hepatocytes: evidence for a proteinase-mediated mechanism. Hepatology 1988; 8:254–8. 56. Harbrecht BG, Biliar TR, Curran RD, et al. Hepatocyte injury by activated neutrophils in vitro is mediated by proteases. Ann Surg 1993; 218:120–8. 57. Jaeschke H, Ho Y-S, Fisher MA, et al. Glutathione peroxidase-deficient mice are more susceptible to neutrophil-mediated hepatic parenchymal cell injury during endotoxemia: importance of an intracellular oxidant stress. Hepatology 1999; 29:443–50. 58. Hasegawa T, Malle E, Farhood A, et al. Generation of hypochlorite-modified proteins by neutrophils during ischemia–reperfusion injury in rat liver: attenuation by ischemic preconditioning. Am J Physiol Gastrointest Liver Physiol 2005; 289(4):G760–7. 59. Farhood A, McGuire GM, Manning AM, et al. Intercellular adhesion molecule-1 (ICAM-1) gene expression and its role in neutrophil-induced ischemia–reperfusion injury in the liver. J Leukoc Biol 1995; 57:368–74. 60. Nagendra AR, Mickelson JK, Smith CW. CD18 integrin and CD54-dependent neutrophil adhesion to cytokine-stimulated human hepatocytes. Am J Physiol 1997; 272:G408–16. 61. Dorman RB, Gujral JS, Bajt ML, et al. Generation and functional significance of CXC chemokines for neutrophil-induced liver injury during endotoxemia. Am J Physiol Gastrointest Liver Physiol 2005; 288(5):G880–6. 62. Mathews WR, Guido DM, Fisher MA, et al. Lipid peroxidation as molecular mechanism of liver cell injury during reperfusion after ischemia. Free Radic Biol Med 1994; 16:763–70. 63. Jaeschke H, Benzick EA. Pathophysiological consequences of enhanced intracellular superoxide formation in isolated perfused rat liver. Chem Biol Interact 1992; 84:55–68. 64. Wendel A, Feuerstein S, Konz KH. Acute paracetamol intoxication of starved mice leads to lipid peroxidation in vivo. Biochem Pharmacol 1979; 28(13):2051–5. 65. Jaeschke H, Kleinwaechter C, Wendel A. The role of acrolein in allyl alcohol-induced lipid peroxidation and liver cell damage in mice. Biochem Pharmacol 1987; 36:51–7. 66. Curzio M, Esterbauer H, Di Mauro C, et al. Chemotactic activity of the lipid peroxidation product 4-hydroxynonenal and homologous hydroxyalkenals. Biol Chem Hoppe Seyler 1986; 367:321–9. 67. Jayatilleke A, Shaw S. Stimulation of monocyte interleukin-8 by lipid peroxidation products: a mechanism for alcohol-induced liver injury. Alcohol 1998; 16:119–23. 68. Liu P, Vonderfecht SL, McGuire GM, et al. The 21-aminosteroid tirilazad mesylate protects in a model of endotoxin shock and acute liver failure in rats. J Pharmacol Exp Ther 1994; 271:438–45. 69. Poli G, Parola M. Oxidative damage and fibrogenesis. Free Radic Biol Med 1997; 22:287–305. 70. Beckman JS. Oxidative damage and tyrosine nitration from peroxynitrite. Chem Res Toxicol 1996; 9:836–44. 71. Gardner CR, Laskin JD, Dambach DM, et al. Reduced hepatotoxicity of acetaminophen in mice lacking inducible nitric oxide synthase: potential role of tumor necrosis factor-alpha and interleukin10. Toxicol Appl Pharmacol 2002; 184(1):27–36. 72. Knight TR, Ho YS, Farhood A, et al. Peroxynitrite is a critical mediator of acetaminophen hepatotoxicity in murine livers: protection by glutathione. J Pharmacol Exp Ther 2002; 303(2):468–75. 73. Bajt ML, Knight TR, Farhood A, et al. Scavenging peroxynitrite with glutathione promotes regeneration and enhances survival during acetaminophen-induced liver injury in mice. J Pharmacol Exp Ther 2003; 307(1):67–73. 74. Kon K, Kim JS, Jaeschke H. Increase of cytosolic ferrous iron induces the mitochondrial permeability transition in acetaminophen-induced toxicity to mouse hepatocytes (abstract). Hepatology 2004; 40:647A. 75. Kon K, Kim JS, Jaeschke H, et al. Mitochondrial permeability transition in acetaminophen-induced necrosis and apoptosis of cultured mouse hepatocytes. Hepatology 2004; 40(5):170–9. 76. Bajt ML, Cover C, Lemasters JJ, et al. Nuclear translocation of endonuclease G and apoptosisinducing factor during acetaminophen-induced liver injury. Toxicol Sci 2006; 94:217–25. 77. Jaeschke H, Bajt ML. Intracellular signaling mechanisms of acetaminophen-induced liver cell death. Toxicol Sci 2006; 89(1):31–41. 78. Liu P, Yin K, Nagele R, et al. Inhibition of nitric oxide synthase attenuates peroxynitrite generation, but augments neutrophil accumulation in hepatic ischemia–reperfusion in rats. J Pharmacol Exp Ther 1998; 284:1139–46. 79. Skinner KA, Crow JP, Skinner HB, et al. Free and protein-associated nitrotyrosine formation following rat liver preservation and transplantation. Arch Biochem Biophys 1997; 342:282–8.

46

Jaeschke

80. McKim SE, Gabele E, Isayama F, et al. Inducible nitric oxide synthase is required in alcohol-induced liver injury: studies with knockout mice. Gastroenterology 2003; 125(6):1834–44. 81. Jaeschke H. Molecular mechanisms of hepatic ischemia–reperfusion injury and preconditioning. Am J Physiol Gastrointest Liver Physiol 2003; 284(1):G15–26. 82. Clemens MG, Bauer M, Pannen BH, et al. Remodeling of hepatic microvascular responsiveness after ischemia/reperfusion. Shock 1997; 8(2):80–5. 83. Nishida J, McCuskey RS, McDonnell D, et al. Protective role of NO in hepatic microcirculatory dysfunction during endotoxemia. Am J Physiol 1994; 267:G1135–41. 84. Wang Y, Mathews WR, Guido DM, et al. Inhibition of nitric oxide synthesis aggravates reperfusion injury after hepatic ischemia and endotoxemia. Shock 1995; 4:282–8. 85. Wang Y, Lawson JA, Jaeschke H. Differential effect of 2-aminoethyl-isothiourea, an inhibitor of the inducible nitric oxide synthase, on microvascular blood flow and organ injury in models of hepatic ischemia–reperfusion injury and endotoxemia. Shock 1998; 10:20–5. 86. Nieminen AL, Saylor AK, Tesfai SA, et al. Contribution of the membrane permeability transition to lethal injury after exposure of hepatocytes to t-butylhydroperoxide. Biochem J 1995; 307:99–106. 87. Byrne AM, Lemasters JJ, Nieminen AL. Contribution of increased mitochondrial free Ca2C to the mitochondrial permeability transition induced by tert-butylhydroperoxide in rat hepatocytes. Hepatology 1999; 29:1523–31. 88. Aguilar HI, Botla R, Arora AS, et al. Induction of the mitochondrial permeability transition by protease activity in rats: a mechanism of hepatocyte necrosis. Gastroenterology 1996; 110:558–66. 89. Miyoshi H, Umeshita K, Sakon M, et al. Calpain activation in plasma membrane bleb formation during tert-butyl hydroperoxide-induced rat hepatocyte injury. Gastroenterology 1996; 110:1897–904. 90. Czaja MJ. Induction and regulation of hepatocyte apoptosis by oxidative stress. Antioxid Redox Signal 2002; 4(5):759–67. 91. Kurose I, Higuchi H, Miura S, et al. Oxidative stress-mediated apoptosis of hepatocytes exposed to acute ethanol intoxication. Hepatology 1997; 25:368–78. 92. Czaja MJ, Liu H, Wang Y. Oxidant-induced hepatocyte injury from menadione is regulated by ERK and AP-1 signaling. Hepatology 2003; 37(6):1405–13. 93. Rauen U, Polzar B, Stephan H, et al. Cold-induced apoptosis in cultured hepatocytes and liver endothelial cells: mediation by reactive oxygen species. FASEB J 1999; 13:155–68. 94. Motoyama S, Minamiya Y, Saito S, et al. Hydrogen peroxide derived from hepatocytes induces sinusoidal endothelial cell apoptosis in perfused hypoxic rat liver. Gastroenterology 1998; 114:153–63. 95. Hampton MB, Orrenius S. Dual regulation of caspase activity by hydrogen peroxide: implications for apoptosis. FEBS Lett 1997; 414:552–6. 96. Fadeel B, Ahlin A, Henter JI, et al. Involvement of caspases in neutrophil apoptosis: regulation by reactive oxygen species. Blood 1998; 92:4808–18. 97. Jaeschke H, Fisher MA, Lawson JA, et al. Activation of caspase 3 (CPP32)-like proteases is essential for TNF-alpha-induced hepatic parenchymal cell apoptosis and neutrophil-mediated necrosis in a murine endotoxin shock model. J Immunol 1998; 160(7):3480–6. 98. Gabbita SP, Robinson KA, Stewart CA, et al. Redox regulatory mechanisms of cellular signal transduction. Arch Biochem Biophys 2000; 376(1):1–13. 99. Haddad JJ. Antioxidant and prooxidant mechanisms in the regulation of redox(y)-sensitive transcription factors. Cell Signal 2002; 14(11):879–97. 100. Li X, Karin M. Is NF-kB the sensor of oxidative stress? FASEB J 1999; 13:1137–43. 101. Watson WH, Yang X, Choi YE, et al. Thioredoxin and its role in toxicology. Toxicol Sci 2004; 78(1):3–14. 102. Bellezzo JM, Leingang KA, Bulla GA, et al. Modulation of lipopolysaccharide-mediated activation of rat Kupffer cells by antioxidants. J Lab Clin Med 1998; 131:36–44. 103. Essani NA, Fisher MA, Jaeschke H. Inhibition of NF-kB activation by dimethyl sulfoxide correlates with suppression of TNF-a formation, reduced ICAM-1 gene transcription and protection against endotoxin-induced liver injury. Shock 1997; 7:90–6. 104. Tsukamoto H, Lin M, Ohata M, et al. Iron primes hepatic macrophages for NF-kappaB activation in alcoholic liver injury. Am J Physiol 1999; 277:G1240–50. 105. Dinkova-Kostova AT, Holtzclaw WD, Cole RN, et al. Direct evidence that sulfhydryl groups of Keap1 are the sensors regulating induction of phase 2 enzymes that protect against carcinogens and oxidants. Proc Natl Acad Sci USA 2002; 99(18):11908–13. 106. Goldring CE, Kitteringham NR, Elsby R, et al. Activation of hepatic Nrf2 in vivo by acetaminophen in CD-1 mice. Hepatology 2004; 39(5):267–76. 107. Okawa H, Motohashi H, Kobayashi A, et al. Hepatocyte-specific deletion of the keap1 gene activates Nrf2 and confers potent resistance against acute drug toxicity. Biochem Biophys Res Commun 2006; 339(1):79–88.

Antioxidant Defense in Liver Injury

47

108. Rensing H, Jaeschke H, Bauer I, et al. Differential activation pattern of redox-sensitive transcription factors and stress-inducible dilator systems heme oxygenase-1 and inducible nitric oxide synthase in hemorrhagic and endotoxic shock. Crit Care Med 2001; 29(10):1962–71. 109. Rensing H, Bauer I, Peters I, et al. Role of reactive oxygen species for hepatocellular injury and heme oxygenase-1 gene expression following hemorrhage and resuscitation. Shock 1999; 12:300–8. 110. Bauer M, Bauer I. Heme oxygenase-1: redox regulation and role in the hepatic response to oxidative stress. Antioxid Redox Signal 2002; 4(5):749–58. 111. Kaplowitz N, Tsukamoto H. Oxidative stress and liver disease. In: Boyer JL, Ockner RK, eds. Progress in Liver Diseases, Vol. 14. Philadelphia: W.B. Saunders, 1996:131–59. 112. Chojkier M, Houglum K, Lee KS, et al. Long- and short-term D-alpha-tocopherol supplementation inhibits liver collagen alpha1(I) gene expression. Am J Physiol 1998; 275:G1480–5. 113. Garcia-Trevijano ER, Iraburu MJ, Fontana L, et al. Transforming growth factor beta1 induces the expression of alpha1(I) procollagen mRNA by a hydrogen peroxide-C/EBPbeta-dependent mechanism in rat hepatic stellate cells. Hepatology 1999; 29:960–70. 114. Xu Y, Rojkind M, Czaja MJ. Regulation of monocyte chemoattractant protein 1 by cytokines and oxygen free radicals in rat hepatic fat-storing cells. Gastroenterology 1996; 110:1870–7. 115. Boess F, Kamber M, Romer S, et al. Gene expression in two hepatic cell lines, cultured primary hepatocytes, and liver slices compared to the in vivo liver gene expression in rats: possible implications for toxicogenomics use of in vitro systems. Toxicol Sci 2003; 73(2):386–402. 116. Halliwell B. Oxidative stress in cell culture: an under-appreciated problem? FEBS Lett 2003; 540(1–3):3–6. 117. Fridovich I. Superoxide radical and superoxide dismutases. Annu Rev Biochem 1995; 64:97–112. 118. Slot JW, Geuze HJ, Freeman BA, et al. Intracellular localisation of the copper–zinc and manganese superoxide dismutases in rat liver parenchymal cells. Lab Invest 1986; 55:363–71. 119. Calabrese EJ, Canada AT. Catalase: its role in xenobiotic detoxification. Pharmac Ther 1989; 44:297–307. 120. Oshino N, Chance B. Properties of glutathione release observed during reduction of organic hydroperoxide, demethylation of aminopyrine and oxidation of some substances in perfused rat liver, and their implications for the physiological function of catalase. Biochem J 1977; 162:509–25. 121. Brigelius-Flohe R. Tissue-specific functions of individual glutathione peroxidases. Free Radic Biol Med 1999; 27(9–10):951–65. 122. Sies H, Sharov VS, Klotz LO, et al. Glutathione peroxidase protects against peroxynitrite-mediated oxidations. J Biol Chem 1997; 272:27812–7. 123. Olafsdottir K, Reed DJ. Retention of oxidized glutathione by isolated perfused rat liver mitochondria during hydroperoxide treatment. Biochim Biophys Acta 1988; 964:377–82. 124. Lawrence RA, Burk RF. Glutathione peroxidase activity in selenium-deficient rat liver. Biochem Biophys Res Commun 1976; 71:952–8. 125. Hill KE, Burk RF, Lane JM. Effect of selenium depletion and repletion on plasma glutathione and glutathione-dependent enzymes in the rat. J Nutr 1987; 117:99–104. 126. Burk RF, Nishiki K, Lawrence RA, et al. Peroxide removal by selenium-dependent and seleniumindependent glutathione peroxidases in hemoglobin-free perfused rat liver. J Biol Chem 1978; 253:43–6. 127. Burk RF, Lawrence RA, Lane JM. Liver necrosis and lipid peroxidation in the rat as the result of paraquat and diquat administration. J Clin Invest 1980; 65:1024–31. 128. Ursini F, Maiorino M, Gregolin C. The selenoenzyme phospholipid hydroperoxide glutathione peroxidase. Biochim Biophys Acta 1985; 839:62–70. 129. Ursini F, Bindoli A. The role of selenium peroxidases in the protection against oxidative damage of membranes. Chem Phys Lipids 1987; 44:255–76. 130. Nishiyama A, Masutani H, Nakamura H, et al. Redox regulation by thioredoxin and thioredoxinbinding proteins. IUBMB Life 2001; 52(1–2):29–33. 131. Rhee SG, Chae HZ, Kim K. Peroxiredoxins: a historical overview and speculative preview of novel mechanisms and emerging concepts in cell signaling. Free Radic Biol Med 2005; 38(12):1543–52. 132. Wood ZA, Poole LB, Karplus PA. Peroxiredoxin evolution and the regulation of hydrogen peroxide signaling. Science 2003; 300(5619):650–3. 133. Jeong W, Park SJ, Chang TS, et al. Molecular mechanism of the reduction of cysteine sulfinic acid of peroxiredoxin to cysteine by mammalian sulfiredoxin. J Biol Chem 2006; 281(20):14400–7. 134. Chow CK. Vitamin E and oxidative stress. Free Radic Biol Med 1991; 11:215–32. 135. Wefers H, Sies H. The protection by ascorbate and glutathione against microsomal lipid peroxidation is dependent on vitamin E. Eur J Biochem 1988; 174:353–7. 136. Sies H. Strategies of antioxidant defense. Eur J Biochem 1993; 215:213–9. 137. Knight TR, Fariss MW, Farhood A, et al. Role of lipid peroxidation as a mechanism of liver injury after acetaminophen overdose in mice. Toxicol Sci 2003; 76(1):229–36. 138. Han D, Hanawa N, Saberi B, et al. Mechanisms of Liver Injury. III. Role of glutathione redox status in liver injury. Am J Physiol Gastrointest Liver Physiol 2006; 291(1):G1–7.

48

Jaeschke

139. Lauterburg BH, Adams JD, Mitchell JR. Hepatic glutathione homeostasis in the rat: efflux accounts for glutathione turnover. Hepatology 1984; 4:586–90. 140. Ballatori N, Hammond CL, Cunningham JB, et al. Molecular mechanisms of reduced glutathione transport: role of the MRP/CFTR/ABCC and OATP/SLC21A families of membrane proteins. Toxicol Appl Pharmacol 2005; 204(3):238–55. 141. Rius M, Hummel-Eisenbeiss J, Hofmann AF, et al. Substrate specificity of human ABCC4 (MRP4)mediated cotransport of bile acids and reduced glutathione. Am J Physiol Gastrointest Liver Physiol 2006; 290(4):G640–9. 142. Chen C, Hennig GE, Manautou JE. Hepatobiliary excretion of acetaminophen glutathione conjugate and its derivatives in transport-deficient (TR-) hyperbilirubinemic rats. Drug Metab Dispos 2003; 31(6):798–804. 143. Muller M, Jansen PL. Molecular aspects of hepatobiliary transport. Am J Physiol 1997; 272(6 Pt 1):G1285–303. 144. Jaeschke H. Glutathione disulfide as index of oxidant stress in rat liver during hypoxia. Am J Physiol 1990; 258:G499–505. 145. Adams JD, Lauterburg BH, Mitchell JR. Plasma glutathione and glutathione disulfide in the rat: regulation and response to oxidative stress. J Pharmacol Exp Ther 1983; 227:749–54. 146. Fernandez-Checa JC, Kaplowitz N. Hepatic mitochondrial glutathione: transport and role in disease and toxicity. Toxicol Appl Pharmacol 2005; 204(3):263–73. 147. Lash LH. Mitochondrial glutathione transport: physiological, pathological and toxicological implications. Chem Biol Interact 2006; 163:54–7. 148. Colell A, Garcia-Ruiz C, Miranda M, et al. Selective glutathione depletion of mitochondria by ethanol sensitizes hepatocytes to tumor necrosis factor. Gastroenterology 1998; 115:1541–51. 149. Garcia-Ruiz C, Colell A, Morales A, et al. Role of oxidative stress generated from the mitochondrial electron transport chain and mitochondrial glutathione status in loss of mitochondrial function and activation of transcription factor nuclear factor-kappa B: studies with isolated mitochondria and rat hepatocytes. Mol Pharmacol 1995; 48:825–34. 150. Jaeschke H. In: Sipes IG, McQueen CA, Gandolfi AJ, eds. Antioxidant Defense Mechanisms. Comprehensive Toxicology, Vol. IX Oxford: Elsevier, 1997:181–97. 151. Hidalgo J, Campmany L, Borras M, et al. Metallothionein response to stress in rats: role of free radical scavenging. Am J Physiol 1988; 255:E518–24. 152. DeLeve LD. Glutathione defense in non-parenchymal cells. Semin Liver Dis 1998; 18:403–13. 153. Spolarics Z. Endotoxemia, pentose cycle, and the oxidant/antioxidant balance in the hepatic sinusoid. J Leukoc Biol 1998; 63:534–41. 154. Spolarics Z, Wu JX. Role of glutathione and catalase in H2O2 detoxification in LPS-activated hepatic endothelial and Kupffer cells. Am J Physiol 1997; 273:G1304–11. 155. Halliwell B, Gutteridge JMC. The antioxidants of human extracellular fluids. Arch Biochem Biophys 1990; 280:1–8. 156. Zelko IN, Mariani TJ, Folz RJ. Superoxide dismutase multigene family: a comparison of the CuZnSOD (SOD1), Mn-SOD (SOD2), and EC-SOD (SOD3) gene structures, evolution, and expression. Free Radic Biol Med 2002; 33(3):337–49. 157. Takahashi K, Avissar N, Whitin J, et al. Purification and characterization of human plasma glutathione peroxidase: a selenoglycoprotein distinct from the known cellular enzyme. Arch Biochem Biophys 1987; 256:677–86. 158. Karlsson K, Lindahl U, Marklund SL. Binding of human extracellular superoxide dismutase C to sulphated glycosaminoglycans. Biochem J 1988; 256:29–33. 159. Jaeschke H. Enhanced sinusoidal glutathione efflux during endotoxin-induced oxidant stress in vivo. Am J Physiol 1992; 263:G60–8. 160. Jaeschke H. Vascular oxidant stress and hepatic ischemia/reperfusion injury. Free Radic Res Commun 1991; 12–13:737–43. 161. Liu P, Fisher MA, Farhood A, et al. Beneficial effect of extracellular glutathione against reactive oxygen-mediated reperfusion injury in the liver. Circ Shock 1994; 43:64–70. 162. Bilzer M, Baron A, Schauer R, et al. Glutathione treatment protects the rat liver against injury after warm ischemia and Kupffer cell activation. Digestion 2002; 66(1):49–57. 163. Burk RF, Hill KE. Selenoprotein P: an extracellular protein with unique physical characteristics and a role in selenium homeostasis. Annu Rev Nutr 2005; 25:215–35. 164. Burk RF, Hill KE, Awad JA, et al. Pathogenesis of diquat-induced liver necrosis in selenium-deficient rats: assessment of the roles of lipid peroxidation and selenoprotein P. Hepatology 1995; 21:561–9. 165. Burk RF, Hill KE, Awad JA, et al. Liver and kidney necrosis in selenium-deficient rat depleted of glutathione. Lab Invest 1995; 72:723–30.

4

Hepatotoxicity Due to Mitochondrial Injury Dominique Pessayre, Bernard Fromenty, Abdellah Mansouri, and Alain Berson quipe Mitochondries, INSERM, U773, Centre de Recherche BiomØdicale Bichat Beaujon, FacultØ de MØdecine Xavier Bichat, UniversitØ Paris 7 Denis Diderot, Paris, France

INTRODUCTION There are three main mechanisms for drug-induced liver injury. The most frequent mechanism is the formation of reactive drug metabolites that can be directly toxic or can cause immune reactions (1,2). The other is drug-induced impairment of mitochondrial function, which may decrease fat oxidation (thus causing steatosis) and/or energy production (thus causing cell dysfunction or cell death) (3–7). A third mechanism involves the opening of the mitochondrial permeability transition (MPT) pore, causing either necrosis or apoptosis. Although a few parent drugs can trigger this transition (8,9), it is more commonly caused by reactive metabolites, which can lead to MPT either through direct toxicity or through immune reactions (8,9). Thus, even when liver injury is initially due to the formation of reactive metabolites, mitochondrial damage plays a major role in the final mechanisms of cell death (8,9). Therefore, most forms of drug-induced liver lesions initially or secondarily involve mitochondrial injury (9). ORIGIN AND STRUCTURE OF MITOCHONDRIA Some 1.5 to 2 billion years ago, a precursor of present day eukaryotes engaged in a parasitic/symbiotic partnership with a wild bacterium (10–12). This precursor allowed the bacterium to reside and divide within its cytoplasm. In exchange, the bacterium used the emerging oxygen atmosphere to completely degrade fuels into CO2 and water, thus providing the host with considerable energy (11,12). Like their bacterial ancestors, mitochondria have two membranes. A circular outer membrane surrounds the intermembrane space, while an inner membrane with inner folds (the mitochondrial cristae) limits the mitochondrial matrix (12). Although most of the initial bacterial genes have been transferred to the nucleus of the host, mitochondria have retained a small genome, located in the matrix (12). Each cell contains many copies of this circular, double-stranded genome, as there are several copies of mitochondrial DNA (mtDNA) in a single mitochondrion and many mitochondria per cell (10). This mtDNA encodes 13 polypeptides of the respiratory chain, while nuclear DNA encodes the remaining respiratory polypeptides and all other mitochondrial proteins. These proteins are synthesized in the cytoplasm (usually with a mitochondrial targeting presequence) and are then imported into the mitochondrial membranes or the matrix, where the presequence is cleaved (10). The authors of this chapter have relationships with the following corporations: Dominique Pessayre has been a consultant for Abbott, Actelion, Astellas, Astra Zeneca, Aventis, Bayer, Ciba-Geigy, DuPont, EliLilly, Fournier Laboratories, Glaxo SmithKline, Helsinn Healthcare, Hoechst-Marion-Roussel, HoffmanLaRoche, Novartis, Parke-Davis, Pfizer, Rhone Poulenc-Rorer, Roussel-UCLAF, Sanofi-Synthelabo, Servier Laboratories, Solvay, TROPHOS and Wyeth. Bernard Fromenty has been a consultant for Aventis and Bristol Myers Squibb. The laboratory of Dominique Pessayre and Bernard Fromenty at INSERM has benefited from research grants from Astra Zeneca, Beaufour, Bristol-Myers-Squibb, Cassenne Laboratories, Ciba Geigy, Helsinn Healthcare, Hoffmann-LaRoche, Knoll, Roger Bellon, Sanofi-Synthelabo, Schering-Plough, Servier Laboratories, and Sigma-Tau.

50

Pessayre et al.

Respiratory chain polypeptides are located in the inner membrane, except for cytochrome c (cyt. c), which is targeted to the intermembrane space, but is mostly associated with cardiolipin on the inner membrane (13). Although the initial shortening of very-long and long-chain fatty acids is mediated by enzymes located in the inner membrane, most of the other enzymes involved in fatty acid b-oxidation and the tricarboxylic acid cycle are located in the mitochondrial matrix (10). ROLES OF MITOCHONDRIA Fat Oxidation and Energy Production Mitochondria play a major role in fat oxidation and energy production. The entry of long-chain fatty acids into mitochondria is modulated by carnitine palmitoyltransferase I (CPT I, Fig. 1). This outer membrane enzyme is inhibited by malonyl-CoA, which is the first step in the synthesis of fatty acids (14). This inhibition normally ensures a reciprocal relationship between fatty acid synthesis and the mitochondrial uptake, and oxidation of fatty acids (14). Once fatty acids are taken up by mitochondria, they are targeted towards oxidation (14). Fatty acyl-CoAs (FA-CoAs) are split by b-oxidation cycles into acetyl-CoA subunits, which can either form ketone bodies or can, like other fuels, be completely degraded to CO2 by the tricarboxylic acid cycle (Fig. 1). The NADH and FADH2 that are generated by both b-oxidation and the tricarboxylic acid cycle are then re-oxidized by the mitochondrial respiratory chain (12). This regenerates the NADC and FAD necessary for other cycles of fuel oxidation (Fig. 1), and also initiates the process of energy production (3). NADH and FADH2 transfer their electrons to the first complexes of the respiratory chain (12). Although a fraction of these electrons react with oxygen to form reactive oxygen species

Uptake CPT I

Synthesis FFA FA-CoA

CYTOSOL TG

VLDL

Secretion

MITOCHONDRIA O2

H2O

e–

Cyt. c oxidase

NAD+, FAD

β-Oxidation

MATRIX

Tricarboxylic acid cycle

e–

NADH, FADH2 ADP CO2

INTERMEMBRANE SPACE

H+

ATP ANT

O2 O2– H+ ATP synthase

FIGURE 1 Fat metabolism and energy production in hepatocytes and mitochondria. FFA are either synthesized in the liver or transferred from the intestine or adipose tissue. Long-chain FFA form FA-CoA thioesters, which either enter mitochondria, a step regulated by CPT I, or are esterified into triglycerides that are stored in the cytoplasm or secreted as VLDLs. Inside mitochondria, the FA-CoA thioesters are cut and oxidized by the b-oxidation and tricarboxylic acid cycles, generating NADH and FADH2 that transfer electrons to the respiratory chain. Although part of these electrons react with oxygen to form the superoxide anion radical and other ROS, most electrons migrate up to cyt. c oxidase, where they safely combine with oxygen and protons to form water. The transfer of electrons along the respiratory chain is associated with the extrusion of protons from the mitochondrial matrix into the intermembrane space. The reentry of protons in the matrix through ATP synthase transforms ADP into ATP, which is then extruded by the ANT, in exchange for cytosolic ADP. Abbreviations: FA-CoA, fatty acyl-CoA; CPT I, carnitine palmitoyltransferase I; VLDLs, very lowdensity lipoproteins; cyt. c, cytochrome c; ANT, adenine nucleotide translocator; TG, triglycerides; FFA, free fatty acids.

Hepatotoxicity Due to Mitochondrial Injury

51

(ROS) as discussed further on, most electrons migrate all the way along the respiratory chain, up to cyt. c oxidase, where four electrons react in quick succession with oxygen and protons to safely form water, without any release of ROS (Fig. 1). During this transfer of electrons along the respiratory chain, protons are extruded from the mitochondrial matrix into the intermembrane space (Fig. 1) (12). This creates a large electrochemical potential across the inner membrane whose potential energy is then used to generate ATP. When ADP is high, protons reenter the matrix through the F0 portion of ATP synthase, causing the rotation of a molecular rotor within the F 1 portion of ATP synthase and the conversion of ADP into ATP (Fig. 1). The adenine nucleotide translocator (ANT) then extrudes the formed ATP, in exchange for cytosolic ADP (12). Thus, mitochondria turn fat and other fuels into CO2 and water, providing most of the cell’s ATP. However, the price of this oxidative phosphorylation is a high formation rate of ROS (12). Mitochondrial ROS Formation A small fraction of the electrons transferred to complexes I and III of the respiratory chain directly react with oxygen to form the superoxide anion radical (Fig. 1) and other ROS, including peroxynitrite (12). Mitochondria are the main source of ROS in the cell (15). Although mitochondria actively repair ROS-mediated mtDNA lesions (16), mtDNA is very sensitive to ROS-induced damage, due to its proximity to the inner membrane (the main cellular source of ROS) and the absence of protective histones. ROS oxidize mtDNA bases, and these modifications can cause errors during either mtDNA replication (17) or repair (18), leading to point mutations. ROS also cause mtDNA strand breaks, which can cause mtDNA deletions (19). The mitochondrial theory of aging suggests that the accumulation of these mtDNA lesions eventually decreases mtDNA-encoded polypeptide synthesis and the transfer of electrons along the respiratory chain (15). Whenever the flow of electrons is blocked at some point along the respiratory chain, then the respiratory chain components, which are located upstream to the block, become overly reduced and directly transfer their electrons to O2, thus increasing the basal formation of ROS (20). This increased ROS formation further increases ROS-induced damage to mtDNA, causing more impairment of respiration and more ROS formation (10). This vicious circle could explain why mtDNA deletions (21), and some point mutations (22), which are uncommon before the age of 40 years, exponentially accumulate during old age. The high mitochondrial formation rate of ROS, and the oxidative damage it causes to mtDNA and nuclear DNA, is probably one important mechanism of aging (10,23). Another role of mitochondria is to modulate cell death. Cell Survival and Cell Death The partnership with parasitic bacteria may have been initially rather dangerous, because wild bacteria tend to proliferate when they are well fed. The problem was solved when the transcription and replication of the bacterial/mitochondrial genome were placed under the control of nuclear genes, thus transforming wild bacteria into tame mitochondria (24). Before this could occur, however, the host may have found ways to partially invalidate the proliferating bacteria, while bacteria may have disabled any mutant host that took advantage of this control mechanism. The present day sequel to this double warfare is a shared decisionmaking process, in which both symbiotic partners play a role in regulating cell death (25). Mitochondria are involved in this decision through the regulation of the permeability of their outer membrane and also through either closure or opening of a pore in the inner membrane, called the MPT pore (Fig. 2) (9,25). Pore closure allows the cell to survive, while pore opening causes cell death (Fig. 2) (9). Pore opening allows massive reentry of protons through the inner membrane, causing collapse of the mitochondrial membrane potential and interrupting mitochondrial ATP synthesis. If the pore opens quickly in all mitochondria, severe ATP depletion prevents apoptosis (an energy-requiring process) and causes necrosis (Fig. 2) (26). Pore opening also causes matrix expansion, rupture of the spherical outer membrane,

52

Pessayre et al.

Normal mitochondria

MPT in all mitochondria

MPT in some mitochondria

MITO

ATP

MPT H+

No ATP

MITO

ATP

MITO

ATP

MPT H+

No ATP

MPT H+

No ATP

Cytochrome c Caspases

LIFE

NECROSIS

APOPTOSIS

FIGURE 2 Regulation of life and death by the MPT pore. Pore closure allows cell survival, whereas pore opening causes cell death. MPT causes the reentry of protons into the mitochondrial matrix, thus bypassing ATP synthase and preventing ATP formation. Pore opening also causes matrix expansion, outer membrane rupture and release of mitochondrial cyt. c from the intermembrane space into the cytosol. If the pore opens in all mitochondria, decreased ATP synthesis causes ATP depletion and necrosis. If the pore only opens in some mitochondria, unaffected mitochondria keep synthesizing ATP (thus avoiding necrosis), while permeabilized mitochondria release cytochrome c, which activates caspases, causing apoptosis. Abbreviations: MPT, mitochondrial permeability transition; cyt. c, cytochrome c; MITO, mitochondria.

herniation of the inner membrane and matrix through the outer membrane gap, and cyt. c release from the intermembrane space into the cytosol (27,28). If the pore only opens in some mitochondria, unaffected mitochondria keep synthesizing ATP, while permeabilized mitochondria release cyt. c, which activates caspases in the cytosol, causing apoptosis (Fig. 2) (9,27,28). CONSEQUENCES OF IMPAIRED MITOCHONDRIAL FUNCTION Primary Impairment of Fatty Acid Oxidation Primary impairment of fatty acid oxidation causes hepatic steatosis (Fig. 3) (3). Let us recall that two main pathways can remove fat from the liver. First, the mitochondrial b-oxidation and tricarboxylic acid cycles can degrade fatty acids into CO2 and water. Second, triglycerides (TG) can be secreted as very low-density lipoprotein (VLDL). In the lumen of the endoplasmic reticulum (ER), microsomal triglyceride transfer protein (MTP) lipidates apolipoprotein B (Apo B) into triglyceride-rich VLDL particles, which then follow vesicular flow to the plasma membrane to be secreted. Drugs can cause steatosis by impairing mitochondrial b-oxidation (3), by inhibiting MTP activity and thus hepatic VLDL secretion (29), or, quite frequently, by concomitantly inhibiting both MTP activity and mitochondrial b-oxidation (30). Mild inhibition of mitochondrial b-oxidation is not enough; severe impairment is required to trigger steatosis (3). In this case, the free fatty acids (FFA) which are taken up by the liver or synthesized within the liver are insufficiently oxidized by the deficient mitochondria and are instead esterified into TG, which accumulate within the cytoplasm of hepatocytes, thus causing hepatic steatosis (Fig. 3) (3). Acute impairments of fatty acid b-oxidation typically cause microvesicular steatosis (3). In this peculiar form of steatosis, numerous tiny lipid vesicles leave the nucleus in the center of the cell and give the hepatocyte a “foamy”, “spongiocytic” appearance. However, when b-oxidation is more chronically impaired, mixed forms of steatosis can also occur.

53

Hepatotoxicity Due to Mitochondrial Injury

Microvesicular steatosis

CYTOSOL Lipid peroxidation Fat vesicle

FFA

TG

FA-CoA

Time Macrovacuolar steatosis

Toxic to mitochondria

MITOCHONDRIA Fatty acid β-oxidation CO2

NAD+ FAD

O2 e–

NADH FADH2

+

H

MATRIX

ADP

H2O Respiration

e– H+

INTERMEMBRANE SPACE

Steatosis

Gluconeogenesis

ATP

Glucose

Cell dysfunction

FIGURE 3 Primary impairment of fatty acid oxidation and its consequences. The impairment of FA-CoA b-oxidation in mitochondria increases FFA, and TG, which accumulate as small lipid vesicles that can progressively coalesce into larger vacuoles. The impairment of fat oxidation deprives the cell of an important source of energy during fasting. It also inhibits gluconeogenesis so that hypoglycemia may occur. In extrahepatic organs, this limits pyruvate oxidation as an alternative source of energy. Finally, the toxic effects of lipid peroxidation products, FFA and FA-CoA on mitochondrial further decrease energy formation, thus causing cell dysfunction. Abbreviations: FFA, free fatty acids; FA-CoA, fatty acyl-CoA; TG, triglycerides.

Some hepatocytes are then filled with tiny lipid vesicles, while others exhibit large fat vacuoles or both small vesicles and larger vacuoles. These associations and transitions suggest that tiny lipid vesicles can progressively coalesce into larger vacuoles (Fig. 3). Indeed, prolonged causes of steatosis rather tend to cause macrovacuolar steatosis (3). In this other form of steatosis, hepatocytes are distended by a single, large vacuole of fat, displacing the nucleus to the periphery of the cell. Primary impairment of fatty acid oxidation also secondarily impairs mitochondrial energy production (3). Fatty acid oxidation represents the main cellular source of energy between meals, and subjects whose mitochondrial b-oxidation is impaired, do not tolerate fasting (31). The inhibition of b-oxidation secondarily inhibits hepatic gluconeogenesis (3). Fasting may trigger hypoglycemia in these patients (31), thus hampering energy production from this still oxidizable fuel. Furthermore, fasting also causes massive adipose tissue lipolysis, thus flooding the liver with FFA, which are not oxidized by the deficient mitochondria and therefore accumulate in hepatocytes (3). FFA and their dicarboxylic acid derivatives inhibit and uncouple mitochondrial respiration, further decreasing energy production (3). Furthermore, steatosis leads to lipid peroxidation whose reactive products damage the respiratory chain and mtDNA (10). All these effects may cause an energy deficit in cells (Fig. 3), which may cause cell dysfunction in different organs. These patients may develop mild liver failure (sometimes with renal failure and pancreatitis) and severe brain dysfunction, causing coma and death (31). Primary Impairment of Mitochondrial Respiration Impairment of mitochondrial respiration decreases energy formation and, depending on the severity of the deficit, can cause either cell dysfunction or cell death (Fig. 4) (3). Moderate impairment of respiration only causes cell dysfunction. Severe impairment of respiration can cause liver cell death, cholestasis, and fibrosis (32–34).

54

Pessayre et al.

CYTOSOL FA-CoA

TG

β-oxidation

NAD+ FAD

Pyruvate oxidation

NADH FADH2 Acetyl-CoA CO2

MATRIX

MITOCHONDRIA H2O

O2 e− e− H+

Respiration e−− e−

O2–

e

H+

ADP

O2

INTERMEMBRANE SPACE ATP

Steatosis

Lactic acidosis

Cell dysfunction Cell death

Oxidative stress mtDNA aging

FIGURE 4 Primary impairment of respiration and its consequences. The block in the flow of electrons in the respiratory chain decreases ATP formation, thus causing cell dysfunction or cell death. It also causes the accumulation of electrons in upstream respiratory chain complexes, thus increasing ROS formation, which can cause oxidative stress and aging of mtDNA. Finally, the block in electron flow also decreases the re-oxidation of NADH into NADC. A first consequence of the lack of NADC is to decrease the b-oxidation of fatty acids, thus causing steatosis. A second consequence is to decrease the oxidation of pyruvate by the pyruvate dehydrogenase complex. Pyruvate is not degraded and, due to the excess in NADH, it is increasingly reduced into lactate, which can trigger lactic acidosis. Abbreviation: mtDNA, mitochondrial DNA.

Impairment of respiration also blocks the transfer of electrons along the respiratory chain, causing over-reduction of the respiratory chain components located upstream, which react with oxygen to form the superoxide anion radical (Fig. 4) (20). The increased ROS formation could damage mtDNA and further impair respiration. ROS could also contribute to the appearance of necroinflammation and fibrosis, as discussed further on. Finally, severe impairment of respiration secondarily impairs mitochondrial b-oxidation (35). Normally, the NADH, which is formed by b-oxidation, is then re-oxidized by the mitochondrial respiratory chain, thus regenerating the NADC required for fatty acid b-oxidation. When respiration is severely impaired, NADC regeneration is insufficient to sustain b-oxidation (35). This secondarily impairs b-oxidation and causes microvesicular steatosis (Fig. 4) (3). The lack of NADC also inhibits the oxidation of pyruvate by the pyruvate dehydrogenase complex. Instead, due to high NADH levels, pyruvate is reduced into lactate, whose accumulation can trigger lactic acidosis (Fig. 4). Common Features Thus, primary impairment of b-oxidation causes both microvesicular steatosis and secondary energy deficiency (causing cell dysfunction), while primary impairment of respiration causes both energy deficiency (leading to cell dysfunction) and secondary impairment of b-oxidation (causing microvesicular steatosis). Thus, irrespective of the initial defect, drug-induced mitochondrial dysfunction may associate features of both steatosis and cell dysfunction. However, each of these features can predominate according to the initial mechanism.

Hepatotoxicity Due to Mitochondrial Injury

55

DIVERSE MECHANISMS IMPAIRING MITOCHONDRIAL FUNCTION Drugs may impair mitochondrial function by several mechanisms (3–7). They can degrade mtDNA, inhibit or terminate mtDNA replication, inhibit mtDNA transcription, impair the stability of mtDNA transcripts, inhibit the translation of mtDNA transcripts, inhibit or uncouple mitochondrial respiration, sequester CoA (which is required for fatty acid activation before b-oxidation), and/or directly inhibit b-oxidation enzymes, ATP synthase, or the ANT. Degradation of mtDNA Alcohol Ethanol abuse causes oxidative stress in the liver (36). The metabolism of ethanol first into acetaldehyde and then into acetate by alcohol dehydrogenase and aldehyde dehydrogenase is associated with the conversion of NADC into NADH (37). This conversion increases the NADH/NADC ratio and thus the NADPH/NADPC ratio, which can reduce ferric iron into ferrous iron, a potent generator of the hydroxyl radical (37). Other ROS sources are increased levels of the ROS-generating cytochrome P450 (CYP) 2E1 (38) and increased mitochondrial ROS formation in hepatocytes (39), as well as the activation of the NADPH oxidase of Kupffer cells (10). ROS cause oxidative damage to mitochondrial lipids (40), proteins (41), and DNA (41,42) in intoxicated animals. A single, high dose of ethanol causes extensive degradation and depletion of hepatic mtDNA within two hours in mice (42,43). Depletion of mtDNA also occurs in skeletal muscle, heart, and brain (43). Depletion of mtDNA can be prevented by 4-methylpyrazole, which blocks ethanol metabolism, or by melatonin, vitamin E, or ubiquinone, three antioxidant drugs (42,43). After a single alcoholic binge, damaged mtDNA molecules are quickly repaired or resynthesized de novo, and mtDNA depletion is only transient (42,43). In contrast, after four days of daily alcohol administration, the accumulation of non-repaired (bulky) DNA lesions on mtDNA inhibits the resynthesis of mtDNA, due to the limited number of intact mtDNA templates. As a consequence, mtDNA depletion lasts for several days after interruption of alcohol intoxication (44). The repetition of mtDNA strand breaks during chronic alcoholism can also cause mtDNA deletions. The prevalence of hepatic mtDNA deletions is increased in alcoholics with microvesicular steatosis, but not in patients with alcoholic hepatitis or cirrhosis (45,46). The latter conditions increase liver cell turnover (47,48), which could eliminate mutated mtDNA genomes if cells with a high proportion of mutated genomes fail to replicate and/or are progressively eliminated through apoptosis. When ethanol ingestion is stopped, alcoholinduced mtDNA deletions disappear quickly in white blood cells (49), which have a quick cell turnover. Alcohol abuse can cause diverse liver lesions. Macrovacuolar steatosis seems to be due to a combination of different mechanisms (50). These include an increased hepatic expression of sterol regulatory element-binding protein-1 causing increased hepatic fatty acid synthesis (51), decreased hepatic expression of MTP, which may sometimes limit hepatic lipoprotein secretion (52), and finally, impaired hepatic fatty acid oxidation (53). The latter effect is due to the excessive reduction of NADC into NADH during the metabolism of ethanol. The lack of NADC then slightly impairs mitochondrial b-oxidation and markedly inhibits the tricarboxylic acid cycle (53). Microvesicular steatosis may be due to a combination of this mild inhibition of b-oxidation, together with severe ROS-dependent damage to mitochondrial lipids, proteins, and DNA, thus further impairing mitochondrial function (46). Necroinflammation and fibrogenesis seem to be mediated by increased intestinal absorption of endotoxin, activation of toll-like receptor 4 on macrophages (54), induction of inducible nitric oxide synthase (55), increased formation of tumor necrosis factor-a (TNF-a), and finally increased apoptosis, which triggers transforming growth factor-b synthesis and collagen synthesis by stellate cells (56). Acetaminophen (Paracetamol) The inadvertent or deliberate intake of large doses of acetaminophen leads to an extensive formation of N-acetyl-p-benzoquinone-imine. This electrophilic metabolite depletes hepatic glutathione (GSH) and protein thiols, increases cell calcium, damages mitochondria, increases

56

Pessayre et al.

the formation of ROS including peroxynitrite, activates c-Jun N-terminal kinase (JNK), and causes MPT and liver cell necrosis, as discussed later on (see section MPT). Hepatic mtDNA was rapidly depleted after a paracetamol overdose in mice, whereas nuclear DNA, albeit partially fragmented, was not significantly depleted (57). The rapid depletion of mtDNA may be due to the DNA damage caused by peroxynitrite and other ROS (57). The absence of a rapid rebound in mtDNA levels may suggest impaired mtDNA resynthesis. Possible reasons could be the development of cell necrosis and the poisoning of topoisomerase II by N-acetyl-pbenzoquinone-imine (58). Impairment of mtDNA Replication by Drugs that Inhibit Topoisomerases and/or Bind to DNA The replication of mtDNA can be inhibited by drugs, which either intercalate between DNA bases or bind strongly to DNA grooves (59). Furthermore, intercalating drugs can inhibit and/or poison DNA topoisomerases, and these effects can also impair mtDNA replication (59). Topoisomerases play an important role in DNA replication and transcription (60). While type I topoisomerases act as monomers and only cut one strand of DNA, type II topoisomerases act as dimers or multimers and cut both strands of DNA. These enzymes cut the phosphodiester DNA backbone by forming a covalent bond between the liberated phosphorus of DNA and a tyrosine of the enzyme. Normally, this reaction is quickly reversible: topoisomerases transiently cut DNA, thus allowing DNA strand(s) to cross the gap, and then promptly reseal the DNA gap. Several antibacterial drugs (4-quinolones, novobiocine) or anticancer drugs (amsacrine, etoposide, anthracyclines, ellipticines, and actinomycins) are topoisomerase inhibitors (60). Although a few inhibitors inhibit the initial cutting of DNA by topoisomerases, most inhibitors instead inhibit mostly the resealing of DNA. These drugs, which are said to “poison” topoisomerases, increase the number of enzyme-bound DNA complexes (60). These complexes are called “cleavable complexes,” because detergents can destabilize the complexes so that the DNA ends are no longer held together, and the protein-linked DNA breaks are revealed. Both the inhibition and the “poisoning” of topoisomerases are deleterious to cells. The collision of a transcription complex or a replication fork against the accumulated topoisomerase-associated DNA breaks interrupts RNA or DNA synthesis, and can lead to real (non-topoisomerase– bound) double-strand breaks and to gene translocations, which can trigger apoptosis and/or cancer (61). Because mitochondria contain both a type I topoisomerase (62) and a bacterial-like type II topoisomerase (63), topoisomerase inhibitors or poisons can affect the replication of mtDNA. Quinolone Antibiotics The 4-quinolone antibiotics inhibit gyrase (a bacterial type II topoisomerase) and also the mitochondrial type II topoisomerase (64). Ciprofloxacin blocks the resealing of mtDNA breaks, causing accumulation of protein-linked double-strand mtDNA breaks (65). Ciprofloxacin and nalidixic acid progressively decrease mtDNA in cultured mammalian cells, impairing mitochondrial respiration and cell growth (64). 4-Quinolone antibiotics can cause cholestasis, steatosis, and necrosis in treated patients (66,67), and both trovafloxacin and alatrofloxacin were taken off the market because of an unacceptable risk of fulminant liver failure. However, it remains unknown whether mtDNA depletion actually occurs in humans or experimental animals treated with quinolone antibiotics. Alternative mechanisms for quinolone-induced hepatitis could include lysosomal membrane permeabilization and MPT (68), altered expression of mitochondrial proteins (69), and the occurrence of immune reactions in some patients (70). Tacrine and Tamoxifen The reversible, cholinesterase inhibitor, tacrine, has been used for the symptomatic treatment of Alzheimer’s disease. Monitoring of serum alanine aminotransferase (ALT) activity and tolerance-dependent, stepwise escalation of the doses were recommended, because the drug increased ALT activity in 50% of recipients, usually after about six weeks of treatment (71). The weak base tacrine is taken up by mitochondria, where it may cycle back and forth across the

Hepatotoxicity Due to Mitochondrial Injury

57

mitochondrial inner membrane, uncoupling respiration, and wasting energy without ATP production (72). First-pass metabolism in the liver ensures low exposure in other organs, explaining why the liver is selectively injured (72). The electrophoretic accumulation of tacrine within mitochondria also explains why this organelle is a selective target (72,73). Tacrine intercalates between mtDNA bases, poisons topoisomerases, and decreases the synthesis of mtDNA in mice (73). This leads to a progressive depletion of mtDNA in mice, eventually followed by the death of a few hepatocytes by necrosis or apoptosis (73). Tamoxifen is an antiestrogen, which is used in the treatment of advanced breast cancer, but can cause steatohepatitis, particularly in overweight women (74). This cationic amphiphilic drug accumulates electrophoretically within mitochondria, where it directly inhibits mitochondrial respiration and mitochondrial b-oxidation, thus causing steatosis (75). In addition, tamoxifen intercalates between DNA bases, and inhibits topoisomerases and mtDNA synthesis (75). Tamoxifen progressively depletes hepatic mtDNA in mice (75). Pentamidine Pentamidine is used in the prevention and treatment of Pneumocystis carinii infections. Although pentamidine inhibits the topoisomerases of P. carinii and trypanosomes, it has little effect on mammalian topoisomerases (59). However, pentamidine binds to the minor groove of duplex DNA and can deplete mtDNA in mammalian cells in vitro (59). Pentamide may also inhibit mitochondrial translation (76). Ethidium Bromide, Ditercalinium, Dequalinium, and 1-Methyl-4-Phenylpyridinium Ion Ethidium bromide, ditercalinium, and dequalinium are cationic drugs, which are concentrated electrophoretically in the mitochondrial matrix, where they intercalate between mtDNA bases, and inhibit mtDNA synthesis, to cause progressive depletion of mtDNA (77–79). 1-Methyl-4-phenylpyridinium ion (MPPC) is an oxidative metabolite of a mitochondrial toxin, which causes Parkinson’s disease in humans. This positively charged metabolite accumulates into mitochondria where it inhibits mtDNA synthesis and depletes mtDNA (80). Polyamine Analogs and Methylglyoxal bis(Guanine Hydrazone) The polycationic nature of polyamines leads to their strong interaction with the successive phosphate groups of the DNA backbone, on the major groove of DNA. This strong binding causes conformational changes in DNA (59). The polycationic nature of these compounds also causes their electrophoretic accumulation within the mitochondrial matrix (59). Methylglyoxal bis(guanine hydrazone) and several polyamine analogs, including N1,N12-bis(ethyl)spermine and N1,N8-bis(ethyl)spermidine, have been shown to progressively decrease mtDNA levels in cultured cell lines (81,82). Impairment of mtDNA Replication by Nucleoside Analogs Incorporated into mtDNA 2 0 ,3 0 -Dideoxynucleosides and Abacavir Several 2 0 ,3 0 -dideoxynucleosides are used in patients with the human immunodeficiency virus. These include 3 0 -azido-2 0 ,3 0 -dideoxythymidine (zidovudine, AZT), 2 0 ,3 0 -dideoxycytidine (zalcitabine, ddC), 2 0 ,3 0 -dideoxyinosine (didanosine, ddI), 2 0 ,3 0 -didehydro-3 0 -deoxythymidine (stavudine, d4T), and (K)-2 0 -deoxy-3 0 -thiacytidine (lamivudine, 3TC). A related molecule is abacavir, although this analog contains a cyclopentene–methanol moiety instead of a dideoxyribose moeity. The normal 5 0 -hydroxyl group found in deoxyribose is present in the sugar analog of these diverse molecules, allowing the formation of the triphosphate derivative and the incorporation of the analog into a nascent chain of DNA (Fig. 5). However, the normal 3 0 -hydroxyl group of deoxyribose is absent in these analogs. Once a single molecule of the analog has been incorporated, the DNA molecule lacks a 3 0 -hydroxyl group. No other nucleotide can be incorporated, interrupting DNA replication (Fig. 5) (83,84). The effects of these dideoxynucleosides depend on the ability of various polymerases to incorporate them into DNA. The human immunodeficiency virus reverse transcriptase can perform this incorporation, impairing reverse transcription of the viral RNA (83). In contrast, the

58

Pessayre et al.

Normal deoxyribose (5'-OH and 3'-OH)

5'OH

P

Dideoxyribose (5'-OH but no 3'-OH)

P

(No 3'OH-end)

mtDNA Interruption of mtDNA replication

mtDNA depletion Acquired mitochondrial cytopathy

FIGURE 5 Interruption of mtDNA replication by dideoxynucleosides. Once DNA polymerase g has incorporated a single molecule of dideoxynucleoside into a growing chain of mtDNA, the DNA now lacks a 3 0 hydroxyl group, and no other nucleotide can be incorporated, unless the analogue can be removed by the proofreading activity of polymerase g. Interruption of mtDNA replication can cause mtDNA depletion and an acquired mitochondrial cytopathy. Abbreviation: mtDNA, mitochondrial DNA.

DNA polymerases, which act in the nucleus, barely perform this incorporation, thus allowing the therapeutic use of these analogs (84). However, DNA polymerase g, which acts in mitochondria, also incorporates the dideoxynucleoside analogs into growing chains of mtDNA. This incorporation terminates mtDNA replication (Fig. 5) (85,86), unless the dideoxynucleotide analog is eventually removed by the proofreading, 3 0 -5 0 -exonuclease activity of polymerase g (87) However, polymerase g has a poor proofreading efficiency with some dideoxynucleotides, which are correctly base-paired, although they have an abnormal phospho-sugar analog backbone. Thus, the rate for the incorporation of these analogs into mtDNA is much faster than the rate for their removal (87). Furthermore, after one analog molecule has been eventually removed, and elongation has transiently resumed, another molecule of the analog is likely to be incorporated again further on. This new blocking molecule will have to be removed again, and so on. The end result is to considerably slow down the rate of efficient, complete mtDNA replication (87). Even in post-mitotic tissues, there is a constant turnover of mitochondria, requiring basal replication of mtDNA. When mtDNA replication is markedly slowed down, mtDNA levels may progressively decrease (Fig. 5). For reasons, which are not fully understood, different dideoxynucleosides tend to have differential effects on mtDNA in diverse organs. Although AZT can occasionally cause mtDNA depletion in the liver (88), it has been suggested that the so-called “D-drugs,” namely ddC, ddI, and d4T are more likely to cause hepatic mtDNA depletion than AZT, 3TC, or abacavir (89). In a patient with lactic acidosis after treatment with both ddI and d4T, all mitochondrial complexes were markedly decreased, except for complex II, which is only encoded by nuclear DNA (90). The impaired synthesis of mtDNA-encoded respiratory chain polypeptides can partially block the flow of electrons in the respiratory chain (91). Whenever the flow of electrons in the respiratory chain is partially blocked, the accumulation of electrons in complex I and complex III increases their reaction with oxygen, and the formation of the superoxide anion. In the case of AZT, this effect could be further aggravated by the inhibitory effect of this analog on the ANT (92). The lack of the ANT in knock out mice blocks the exchange of mitochondrial ATP for cytosolic ADP (20). This prevents the reentry of protons through ATP synthase and causes a high mitochondrial potential, which blocks the flow of electrons in the respiratory chain, and increases mitochondrial ROS formation, thus causing precocious appearance of mtDNA deletions (20). In mice, AZT increased peroxide formation by hepatic mitochondria and caused extensive oxidation of guanosine into 8-hydroxydeoxyguanosine in mouse liver mtDNA (93). Like AZT-treated mice, asymptomatic HIV-infected patients treated with AZT had higher urinary excretion of 8-hydroxydeoxyguanonise than untreated patients (94).

Hepatotoxicity Due to Mitochondrial Injury

59

As previously discussed, oxidative damage to mtDNA can cause both point mutations and deletions. In patients treated with diverse nucleoside reverse transcriptase inhibitors, heteroplasmic point mutations were shown to accumulate in the mtDNA of peripheral blood cells (95). Likewise, mtDNA deletions were more prevalent in the sperm of patients treated with various nucleoside analogs for 12 months or more than in those treated for lesser lengths of time (96). In one of these patients, multiple deletions were present after six months of treatment, whereas none had been found before treatment (96). In another patient treated with 3TC, d4T, and saquinavir, most of the hepatic mtDNA existed as six distinct deleted forms (97). This patient developed lactic acidosis, although hepatic mtDNA was not depleted (97). Thus, once mtDNA deletions and point mutations have developed, severe mitochondrial dysfunction may occur, even though tissue mtDNA level (which is the sum of normal and mutated mtDNA) may be normal. Tissues, whose mtDNA is decreased or abnormal, attempt to increase both mtDNA replication and transcription by inducing compensatory increases in mitochondrial biogenesis. In human volunteers treated for two weeks with either d4T/3TC or AZT/3TC, a decrease in mtDNA-encoded messenger RNAs was associated with an increased expression of peroxisome proliferator receptor g coactivator 1 (PGC1), nuclear respiratory factor 1, and mitochondrial transcription factor A in adipose tissue (98). All three factors are master regulators of mitochondrial biogenesis (99) and may help attenuate the adverse effects of nucleoside reverse transcriptase inhibitors. Indeed, an increase in the number of muscle mitochondria is observed in patients with AZT-induced myopathy (86). Mitochondrial proliferation can also occur in the liver (100). In one patient with d4T-induced lactic acidosis, extensive hepatic mitochondrial proliferation led to tightly packed mitochondria on electron microscopy, and to an aspect of oncocytic hepatocytes (cells with a pink granular cytoplasm) on light microscopy (100). Fialuridine and Ganciclovir Clinical trials testing the efficacy of fialuridine in patients with chronic hepatitis B were interrupted after several patients developed microvesicular steatosis and unmanageable lactic acidosis, sometimes with pancreatitis, neuropathy, or myopathy (101). This complication was unexpected because fialuridine possesses both a 5 0 -hydroxyl group and a 3 0 -hydroxyl group, so that the incorporation of a single molecule of fialuridine into DNA should not immediately terminate mtDNA replication. However, when several adjacent molecules of fialuridine are successively incorporated, further activity of DNA polymerase g is inhibited, decreasing mtDNA replication and mtDNA levels (102). Ganciclovir is mainly used in the treatment of cytomegalovirus infections (103). This nucleoside analog is converted into ganciclovir-monophosphate by viral kinases encoded by cytomegalovirus, varicella/zoster, and herpes simplex viruses (103). Cellular kinases then form ganciclovir-triphosphate, which can lead to the incorporation of ganciclovir nucleotides into growing chains of DNA (103). The acyclic, pseudo-sugar analog present in ganciclovir has two hydroxyl groups so that its incorporation does not terminate DNA replication. However, incorporated ganciclovir molecules may cause DNA helix perturbations, which may block the next DNA replication cycle, when the ganciclovir-modified DNA chain has to serve as a replication template (103). Ganciclovir is also incorporated in mtDNA and triggers mtDNA depletion, ultrastructural mitochondrial lesions and steatosis, together with apoptosis (104). Decreased Synthesis and Stability of mtDNA Transcripts in Cells Treated with Interferon-a Interferon-a is used in patients with chronic viral hepatitis B or C. Interferons induce 2 0 ,5 0 -oligoadenylate synthases, which synthesize 2 0 ,5 0 -oligoadenylates in the presence of double-stranded RNAs (105). The 2 0 ,5 0 -oligoadenylates activate RNase L, which, furthermore, is induced by interferons-a and -b (105). RNase L cleaves RNAs (105). This may cleave the nuclear DNA-encoded mRNA of mitochondrial transcription factor A (106), thus decreasing mtDNA transcription and the synthesis of mitochondrial mRNAs (107). Furthermore, RNase L is also present in mitochondria, where it degrades mitochondrial mRNAs (108). Thus, interferon-a decreases both the synthesis and also the stability of mitochondrial transcripts

60

Pessayre et al.

(107,108), which can eventually decrease mtDNA-encoded respiratory chain polypeptides and mitochondrial respiration in cultured cells (109). Although it is not known whether these in vitro effects also occur in interferon-treated patients, it is noteworthy that some adverse effects of interferon-a, such as minor blood dyscrasias, myalgias, paresthesias, convulsions or depression (110), or hepatic steatosis (111), resemble those observed in mild forms of inborn mitochondrial cytopathies. Decreased Translation of mtDNA Transcripts Erythromycins Erythromycins are amphiphilic cationic drugs, which accumulate in acidic compartments including lysosomes, where they can inhibit phospholipases to cause phospholipidosis (112). These antibiotics bind to the 50S ribosomal subunit of bacteria to inhibit the transfer of amino acids from the aminoacyl-tRNA to the peptide chain (113). Erythromycins can also inhibit mitochondrial protein synthesis, to some extent (114). They can cause megamitochondria (115) and can trigger sensorineural hearing loss (116). Finally, erythromycins are transformed into reactive metabolites, which may covalently bind to proteins, thus forming neoantigens, which may trigger immunization and hepatitis (117). It is unknown whether mitochondrial effects can contribute to erythromycin-induced cholestatic or mixed hepatitis (possibly by causing bile duct lesions and the release of immunizing neoantigens). Chloramphenicol and Thiamphenicol Chloramphenicol and thiamphenicol also bind to the 50S ribosomal subunit to inhibit protein synthesis in both bacteria and mitochondria (118). Mitochondrial dysfunction is probably involved in the reversible bone marrow suppression induced by chloramphenicol (119). However, it remains unknown whether other effects of chloramphenicol, including aplastic anemia and cholestatic hepatitis, are due to mitochondrial dysfunction and/or reactive metabolite formation (119). Sequestration of CoA or Carnitine and/or Direct Inhibition of b-Oxidation Aspirin Aspirin is quickly hydrolyzed into salicylic acid, which is activated into salicylyl-CoA on the outer mitochondrial membrane (120). Extensive salicylyl-CoA formation sequesters extramitochondrial CoA, leaving insufficient CoA to activate long-chain fatty acids and preventing their entry into mitochondria and b-oxidation (121). Even though lethal overdoses of aspirin frequently cause microvesicular steatosis (122), therapeutic doses do not, although they can trigger Reye’s syndrome in children with viral infections. Interferon-a, TNF-a, and nitric oxide are released during viral infections, and all impair mitochondrial function. As discussed above, interferon-a decreases the synthesis and stability of mitochondrial transcripts (107,108). Nitric oxide reversibly inhibits mitochondrial respiration (123) and may open the MPT pore (124). TNF-a also inhibits respiration and opens the MPT pore (8). Nevertheless, viral infections rarely cause Reye’s syndrome, suggesting that these endogenous substances do not impair mitochondrial function enough to trigger the disease. However, if children take aspirin during a viral illness, the added effects of salicylate on mitochondrial function may sufficiently impair mitochondrial function to trigger the syndrome in some children. The potentiating effect of aspirin is based on the following evidence. In the past, 93% of children with Reye’s syndrome had received aspirin during an acute viral illness (125), and children with Reye’s syndrome had received aspirin more frequently than those with similar viral diseases not followed by Reye’s syndrome (126). When aspirin use was advised against in feverish children, there was a corresponding decline in the use of aspirin and the incidence of Reye’s syndrome in the United States (127). The few residual cases of Reye’s syndrome now mainly occur in children with another potentiating factor, namely a latent genetic defect in mitochondrial b-oxidation enzymes (128).

61

Hepatotoxicity Due to Mitochondrial Injury

∆4-Ene-valproate

CYP

Mito

Valproate

Uncoupling Ca2+ Other stimuli

Mito

∆2, ∆4-Diene-valproyl-CoA

Valproyl-CoA

MPT

Inactivation of β-oxidation enzymes?

Sequestration of CoA

Cell death

Fatty acid β-oxidation

Pyruvate oxidation

FIGURE 6 Mitochondrial effects of valproate. Valproate is extensively transformed into valproyl-CoA in mitochondria, thus sequestering intramitochondrial CoA and impairing mitochondrial fatty acid b-oxidation and pyruvate oxidation. Valproate is also desaturated by CYP into D4-ene-valproate, which forms D4-ene-valproyl-CoA in mitochondria, and then D2,D4-diene-valproyl-CoA, an electrophilic metabolite, which may inactivate b-oxidation enzymes. Finally, valproic acid uncouples mitochondrial respiration, thus favoring MPT and cell death. Abbreviations: CYP, cytochrome P450; MPT, mitochondrial permeability transition; Mito, mitochondria.

Yet another effect of salicylate is to slightly uncouple mitochondrial respiration (121) and open the MPT pore (129), as discussed further on. The latter effect could contribute to Reye’s syndrome and could also be involved in the spotty liver cell death observed in patients receiving high therapeutic doses of aspirin (130). Valproic Acid Valproic acid is a branched-chain fatty acid used in several forms of seizures. Like natural short-chain fatty acids, valproic acid can enter mitochondria without undergoing previous activation. Inside the mitochondria, valproic acid is extensively transformed into valproyl-CoA (131). The sequestration of intramitochondrial CoA inhibits the b-oxidation of long-, medium-, and short-chain fatty acids (Fig. 6) (131,132). The lack of CoA may also inhibit pyruvate dehydrogenase which requires CoA as a necessary cofactor. This may explain why valproate markedly decreases mitochondrial respiration from pyruvate, although it has little effect on the respiration supported by malate and glutamate (133). Several other effects contribute to the mitochondrial toxicity of valproate (Fig. 6). CYPs 2C9 and 2A6 desaturate the outer carbons of valproate, forming 4-ene-valproate (134). This metabolite is activated into 4-ene-valproyl-CoA inside mitochondria (135,136). The first dehydrogenation step of the b-oxidation cycle then forms 2,4-diene-valproyl-CoA, which is a chemically reactive metabolite that may inactivate b-oxidation enzymes (136,137). This CYPinvolving pathway could explain why the hepatotoxicity of valproate is enhanced by the concomitant administration of the CYP-inducing antiepileptic drugs, phenytoin and carbamazepine, which increase the formation of 4-ene-valproate (138). Finally, valproic acid has an uncoupling effect, which favors MPT pore opening (Fig. 6), as discussed further on. An asymptomatic increase in serum aminotransferase activity, which normalizes with either dose reduction or drug discontinuation, is frequent during administration of this antiepileptic agent (139). A much less frequent side effect is a Reye’s-like syndrome (140), which occurs mainly (but not exclusively) in very young children and between the first and fourth month of treatment. Histologically, centri- and midzonal microvesicular steatosis is associated with centrizonal necrosis, and sometimes cirrhosis (140). This combination of microvesicular fat and liver cell death may be related to the dual effect of valproic acid, which both inhibits mitochondrial b-oxidation and opens the MPT pore (7), as discussed further on.

62

Pessayre et al.

The inhibition by valproate of both fatty acid b-oxidation and pyruvate-supported respiration may explain why valproate administration can aggravate both inborn b-oxidation defects (141,142) and mitochondrial cytopathies (143–145). Panadiplon Development of the anxiolytic drug panadiplon was interrupted due to increased serum transaminases activity during clinical trials (146). Panadiplon was converted into cyclopropane carboxylic acid, which sequesters coenzyme A and carnitine, and inhibits the mitochondrial b-oxidation of fatty acids (146). Pivampicillin Administration of pivampicillin results in the extensive formation and urinary excretion of pivaloylcarnitine, thus depleting free carnitine and inhibiting fatty acid oxidation (147). Tetracyclines Tetracycline itself and the various tetracycline derivatives produce extensive microvesicular steatosis of the liver in experimental animals (148,149). This is due to the dual effect of these antibiotics, which inhibit both the mitochondrial b-oxidation of fatty acids (148,149) and the hepatic secretion of VLDLs (Fig. 7) (149,150). The latter effect occurs at doses, which do not inhibit protein synthesis (150), and may be due to the inhibition of MTP activity by tetracycline (30). At presently administered oral doses, tetracycline may produce minor degrees of hepatic steatosis of no clinical severity in humans. However, severe microvesicular steatosis has occurred in the past during the intravenous administration of high doses of tetracycline (151). Predisposing factors included impaired renal function (which may decrease tetracycline elimination) and pregnancy (which may impair mitochondrial function, as discussed later on) (151). The syndrome usually appeared after four to ten days of tetracycline infusion. Microvesicular steatosis has also been observed after intravenous administration of several other tetracycline derivatives (3,151). 2-Arylpropionate Nonsteroidal Anti-inflammatory Drugs Several nonsteroidal anti-inflammatory drugs are 2-arylpropionate derivatives. Hepatic injury due to these drugs consists of hepatitis and/or microvesicular steatosis of the liver. The latter condition has been observed with pirprofen, naproxen, ibuprofen, and ketoprofen (152–155).

Fat Apo B

Fat

ER

MTP

Microsomal triglyceride transfer protein VLDL

VLDL

VLDL Fat

Fat

VLDL secretion Tetracycline, Amiodarone, Amineptine, Tianeptine, Pirprofen

MITO

CO2

FIGURE 7 Dual effects of tetracycline and some other steatogenic drugs on fat oxidation and hepatic lipid secretion. Several drugs inhibit the b-oxidation of fatty acids in MITO, and also inhibit MTP activity. The latter effect decreases the lipidation of Apo B into VLDL in the lumen of the ER, thus decreasing hepatic VLDL secretion. Abbreviations: MITO, mitochondria; MTP, microsomal triglyceride transfer protein; Apo B, apolipoprotein B; VLDL, very low-density lipoprotein; ER, endoplasmic reticulum.

Hepatotoxicity Due to Mitochondrial Injury

63

2-Arylpropionates have an asymmetric carbon and exist as either S(C)- or R(K)enantiomers. Only the S(C)-enantiomer inhibits prostaglandin synthesis, whereas only the R(K)-enantiomer is converted into the acyl-CoA derivative. Nevertheless, both the S(C)enantiomer and the R(K)-enantiomer of ibuprofen inhibit the b-oxidation of medium- and short-chain fatty acids (156). Inhibition of b-oxidation has also been observed with pirprofen, tiaprofenic acid, and flurbiprofen (157). Amineptine and Tianeptine These antidepressant drugs have a tricyclic moiety and a heptanoic side chain. They rarely cause immunoallergic hepatitis due to the formation of reactive metabolites by CYP (158,159). Exceptionally, they can also cause microvesicular steatosis, due to impaired b-oxidation (3). Both amineptine and tianeptine are metabolized by the b-oxidation of their heptanoic side chain, forming the 5-carbon and 3-carbon derivatives (160,161). In the presence of these drugs, mitochondria are thus exposed to C7, C5, and C3 analogs of natural fatty acids. These analogs reversibly inhibit the b-oxidation of medium- and short-chain fatty acids (162,163). Female Sex Hormones About 1 in 13,000 pregnant women develop microvesicular steatosis during the last trimester of pregnancy (164). Untreated, the disease progresses to coma, kidney failure, and hemorrhage, and leads to the death of the mother and child in 75% to 85% of cases. In contrast, rapid termination of pregnancy results in the delivery of a healthy child and rapid resolution of the mother’s disease in most cases (165). Both pregnancy and the administration of estradiol and progesterone alter mitochondrial ultrastructure and function in mice (166,167). However, these effects are mild; b-oxidation is only slightly impaired and microvesicular steatosis does not develop in these mice (166,167). Similarly, most human pregnancies do not cause acute fatty liver. Therefore, additional factors are probably required to trigger this syndrome. Partial deficiency of long-chain 3-hydroxyacyl-CoA dehydrogenase (LCHAD), which is part of the trifunctional membrane-bound b-oxidation enzyme, has been reported in some women with acute fatty liver of pregnancy (168). Mothers with a single defective LCHAD allele, who are unlucky enough to marry a heterozygous carrier, and then to conceive a fetus with two defective alleles develop the disease (168), whereas those who bear an unaffected child usually have uncomplicated pregnancies. Although the fetus itself may not use fatty acids for energy production, the human placenta metabolizes fatty acids and may therefore release toxic 3-hydroxy, and dicarboxylic fatty acids in the feto-maternal circulation during LCHADnegative conceptions (169). However, the high frequency of LCHAD deficiency in some publications of acute fatty liver of pregnancy does not reflect the real prevalence of this association in an unselected series of acute fatty liver of pregnancy. We did not detect the most prevalent LCHAD mutation in any of 14 consecutive women with histologically confirmed acute fatty liver of pregnancy (170). These findings suggest that the LCHAD deficiency is a possible, albeit uncommon, cause of acute fatty liver of pregnancy in unselected cases. Furthermore, although one in 70 French persons are heterozygous for the A985G mediumchain acyl-CoA dehydrogenase (MCAD) mutation (171), which accounts for 89% of all deficient MCAD alleles, none of the 14 women with acute fatty liver of pregnancy carried the A985G MCAD mutation (170). These observations should not prevent screening for various defects in mitochondrial b-oxidation in women with acute fatty liver of pregnancy (particularly those with recurrent disease or with a still-born or unhealthy child). However, together with epidemiological data, they do suggest that such defects are rarely involved. Indeed, with few exceptions, when pregnancy is terminated early, the child is healthy and the acute fatty liver of pregnancy does not recur in subsequent pregnancies. Drugs, food, stress, infections, an autoimmune reaction triggered by the foreign child, or placental ischemia associated with preeclampsia may perhaps trigger the syndrome in different women.

64

Pessayre et al.

Glucocorticoids Glucocorticoids inhibit acyl-coenzyme A dehydrogenases and produce microvesicular steatosis of the liver in mice (172). Glucocorticoids can cause steatosis (151) and steatohepatitis (173) in humans. Calcium Hopantenate Calcium hopantenate has caused several cases of Reye’s-like syndrome in Japan (174). Pantothenic acid is a constituent of CoA, and calcium hopantenate may decrease CoA and inhibit mitochondrial b-oxidation (174). Inhibition of b-Oxidation and/or Respiration: Role in Steatohepatitis or Hepatitis Amiodarone, 4,4 0 -Diethyaminoethoxyhexestrol, and Perhexiline Amiodarone, 4,4 0 -diethylaminoethoxyhexestrol, and perhexiline are cationic amphiphilic compounds. They have a lipophilic moiety and an amine function, which can become protonated (and thus positively charged). This structure is responsible for the two liver lesions that occur with these drugs, namely lysosomal phospholipidosis and steatohepatitis (7). The uncharged, lipophilic form of these drugs crosses the lysosomal membrane (175). In the acidic lysosomal milieu, the unprotonated drug molecule is protonated and trapped, since it cannot cross back through the lysosomal membrane. It reaches high intralysosomal concentrations and forms non-covalent but tight complexes with phospholipids, thus hampering the action of intralysosomal phospholipases (175). Phospholipids are not degraded, and the phospholipids–drug complexes progressively accumulate as myelin-like figures in enlarged lysosomes (175). Although phospholipidosis is frequent and may be constant in patients receiving these drugs, it appears to have no clinical consequence, since it often occurs without clinical symptoms or biochemical disturbances (176). However, the cationic amphiphilic structure of these drugs also causes impaired mitochondrial function (Fig. 8) (177–181). The unprotonated, lipophilic form easily crosses the mitochondrial outer membrane and is protonated in the acidic intermembranous space (177–181). This positively charged, protonated form is electrophoretically “pushed” inside mitochondria by the high electrochemical potential existing across the inner mitochondrial membrane, and thus reaches high intramitochondrial concentrations (177–181). It remains unknown whether this transport occurs through inner membrane transporters or directly through the lipid bilayer, thanks to charge delocalization. The high intramitochondrial concentrations of these drugs inhibit b-oxidation (causing steatosis) and also block the transfer of electrons along the respiratory chain (177–181). Respiratory chain components become overly reduced and increasingly transfer their electrons to oxygen to form the superoxide anion radical and other ROS (181). This increased ROS formation causes lipid peroxidation (181), and, like ethanol (10), could increase cytokine production (10). Both lipid peroxidation and cytokines can cause steatohepatitis lesions (10). Prolonged administration of these three drugs can cause typical steatohepatitis lesions, with steatosis, necrosis, Mallory bodies, a mixed inflammatory cell infiltrate (containing neutrophils), fibrosis, and even cirrhosis (182–187). Tamoxifen As already mentioned, the antiestrogen, tamoxifen is a cationic amphiphilic drug, which is electrophoretically transported into the mitochondrial matrix, where it achieves high concentrations that directly inhibit both mitochondrial b-oxidation and respiration (75). In addition, tamoxifen intercalates between DNA bases, and inhibits mtDNA synthesis, which depletes mtDNA in mice (75). Tamoxifen can cause steatohepatitis (188,189) particularly in overweight women (74). Although tamoxifen has been shown to impair lysosomal acidification (190) and cause intralysosomal storage of polar lipids after administration of high doses to animals (191), phospholipidosis does not seem to have been reported in human livers.

65

Hepatotoxicity Due to Mitochondrial Injury

A (Amiodarone, perhexiline, DEAEH)

Respiratory chain

A

H+

AH+ + −

CYTOSOL

MITOCHONDRIA INTERMEMBRANE SPACE MATRIX

High concentrations Respiratory chain β-oxidation

Respiration e−

Steatosis

e−− e e−

O2

O2:

ROS

FIGURE 8 Effects of amphiphilic cationic drugs on mitochondrial function. After crossing the outer membrane, the uncharged tertiary amine (A) of amiodarone, perhexiline or DEAEH is protonated in the acidic intermembrane space. The positively charged molecule (AHC) is then electrophoretically pushed by the mitochondrial membrane potential into the matrix. High intramitochondrial concentrations inhibit both b-oxidation (causing steatosis) and oxidative phosphorylation (increasing the formation of ROS). The latter may oxidize fat deposits, causing lipid peroxidation, which, together with ROS-induced cytokine production, could cause steatohepatitis. Abbreviations: DEAEH, diethylaminoethoxyhexestrol; ROS, reactive oxygen species.

Buprenorphine This morphine analog is used as a substitution drug in heroin addicts. The sublingual route is used, to partly prevent extensive first pass metabolism in the liver. At high concentrations, buprenorphine inhibits both mitochondrial b-oxidation and respiration in rat hepatocyte mitochondria (192). Much lower concentrations are observed in humans, and the drug is usually well tolerated. However, cytolytic hepatitis and steatosis have been observed in a few patients (193). Predisposing factors could include intravenous buprenorphine misuse (resulting in higher concentrations) and concomitant exposure to viruses, other drugs or ethanol, all of which impair mitochondrial function (193). Benzarone and Benzbromarone Benzarone and benzbromarone, despite their structural analogy with amiodarone, are not cationic drugs but phenolic compounds. Benzarone and benzbromarone both uncouple and inhibit respiration at low concentrations (194). Although these two drugs also impair mitochondrial b-oxidation, this effect requires higher concentrations (194). In humans, benzarone and benzbromarone can cause hepatocellular liver injury (195,196). Antimalarial Drugs Chloroquine and most other antimalarial drugs are cationic compounds which accumulate in the acidic vacuole of the malaria parasite to disrupt its function due to alkalinization (197). Chloroquine also accumulates at high concentrations in the lysosomes of the host and can cause phospholipidosis (198). Chloroquine, primaquine, and quinine have been shown to impair respiration in rat liver mitochondria (199). In addition, some antimalarial drugs, including primaquine (200) and amodiaquine (201) form reactive metabolites. Metabolic activation may play an important role in amodiaquine-induced agranulocytosis and hepatitis (202).

66

Pessayre et al.

Inhibition of ATP Synthase or the ANT When cells lack energy, cytosolic ADP is high, and ADP enters the mitochondrial matrix through the ANT. The intramitochondrial ADP stimulates the reentry of protons through ATP synthase, and the energy, which is liberated by this reentry, is harnessed in the synthesis of ATP from ADP. The ANT then extrudes the formed ATP, in exchange for cytosolic ADP. Concomitantly, the decrease in the membrane potential due to the reentry of protons allows more electrons to flow within the respiratory chain, thus increasing mitochondrial respiration. Like drugs blocking electron flow within the respiratory chain, the inhibitors of the ANT or of ATP synthase also inhibit the mitochondrial respiration, which is triggered by ADP. However, unlike respiratory chain inhibitors, the inhibitors of the ANT or of ATP synthase do not inhibit the respiration triggered by uncouplers. The latter compounds stimulate respiration by causing the reentry of protons directly through the inner membrane, thus bypassing the need for either the ANT or ATP synthase. Inhibitors of ATP Synthase ATP synthase is a wonderful biological engine (203). Its transmembrane, F0 portion contains a nonrotating moiety, and a rotor made of c subunits (c-ring rotor). The reentry of protons through a channel located at the junction of the non-rotating moiety and the c-ring, triggers the rotation of the c-ring and the attached g stalk. The g stalk rotates within a cap made of three pairs of alternating a and b subunits (a3b3 cap), which is kept immobile by the a/b2/d stator (203). The stalk is asymmetrical. When it rotates within the a3b3 cap, it causes rhythmic deformations of the catalytic b subunits, which may alternatively tighten around ADP and Pi, close to form ATP and fully open to release ATP. Importantly, ATP synthase is a reversible machine. When the mitochondrial membrane potential is low, for example during ischemia, ATP synthase can function in the reverse mode. Its ATPase activity then consumes the meager amounts of ATP formed by glycolysis to pump protons into the intermembrane space, in an attempt to partially restore the mitochondrial membrane potential. This ATP-depleting effect can be attenuated by the binding of a matrix protein, which is termed the inhibitor of F1ATPase (IF1) (204). During ischemia, anaerobic glycolysis causes lactate formation and matrix acidification, which triggers the binding of IF1 to the F1 moiety of ATP synthase thus inhibiting its ATPase activity. This attenuates the selfish use of cell ATP by de-energized mitochondria. Organotin compounds (205) and several natural toxins, such as apoptolidin (206), aurovertin (207), citreoviridin (208), efrapeptins (209), oligomycin, and venturicidin (205), are potent inhibitors of ATP synthase. These toxins block aerobic ATP formation by mitochondria, and can damage cells, which cannot synthesize enough ATP through glycolysis. However, these toxins also block anaerobic ATP consumption by hypoxic mitochondria. Paradoxically, they can transiently protect hypoxic cells from necrosis by preventing ATP synthase from further depleting ATP in hypoxic cells (210). ATP synthase activity is also inhibited by high, supraphysiological concentrations of estrogens (211), and by several phenolic phytochemicals present in human diet, such as resveratrol, curcumin, genistein, or quercetin (212). Inhibitors of the ANT The ANT exchanges ATP for ADP across the inner membrane, and also modulates MPT (213). The ANT achieves two different conformational states (214). Atractyloside (or carboxyatractyloside) binds to, and stabilizes the cytosolic c-state of the transporter (214) and can trigger MPT (11). Bongkrekic acid binds to, and stabilizes the matrix m-state of the transporter (214), and inhibits MPT (11). Both inhibitors block the transport of adenine nucleotide and can decrease cell ATP. The impairment of the ANT also increases the membrane potential, blocks electron flow, causes the over-reduction of respiratory chain complexes, and increases mitochondrial ROS formation, thus causing mtDNA lesions (20).

67

Hepatotoxicity Due to Mitochondrial Injury

Like bongkrekate, the HIV protease inhibitor, nelfinavir binds the ANT on its matrix side, and inhibits MPT and apoptosis (215). In contrast, long-chain acyl-CoA esters bind and inhibit ANT from both sides of the inner membrane (216), and can trigger MPT (217). Uncoupling of Mitochondrial Respiration by Protonophoric Drugs Several drugs can translocate protons from the intermembrane space into the mitochondrial matrix (Fig. 9). This occurs with either cationic or acidic compounds, the latter including both phenolic and carboxylic compounds. Cationic drugs, such as amiodarone, 4,4 0 -diethylaminoethoxyhexestrol, perhexiline, tacrine, and buprenorphine, have an amine function (A) that is protonated (AHC) in the acidic intermembrane space. This positively charged molecule is electrophoretically pushed across the inner membrane by the membrane potential (72,181). It releases its proton in the alkaline matrix, thus reforming the uncharged molecule (A), which may cross back through the inner membrane lipid bilayer, to be protonated again in the intermembrane space, ready for another cycle of proton translocation (72). Carboxylic and other acidic compounds, including natural FFA and several nonsteroidal anti-inflammatory drugs (218,219), cross the inner membrane lipid bilayer in the uncharged form (R–COOH) and release HC in the alkaline matrix. Then the anionic form (R–COOK) is pulled into the intermembrane space by the mitochondrial membrane potential. This second crossing of the inner membrane may occur through anion transporters (220,221). Inside the acidic intermembrane space, the uncharged acid (R–COOH) is formed again, ready for another cycle of proton translocation. The reentry of protons into the intermembrane space increases basal respiration (Fig. 9). As explained above, the flow of electrons along the respiratory chain is coupled with the extrusion of protons from the mitochondrial matrix into the intermembranous state (Fig. 1).

NORMAL

UNCOUPLING

MATRIX

Consequences

DRUG

MATRIX

Heat generation

H+ O2 e−

O2

H2O Respiratory chain

e–

ATP

Respiration Electron flow

H+

H+ ADP

H2O

ATP synthase

ADP ATP H+

H+

ATP Cell dysfunction

IM

IM

Cell death

FIGURE 9 Opposite effects of uncouplers on mitochondrial respiration and ATP formation. Uncouplers translocate protons across the IM. The reentry of protons into the mitochondrial matrix decreases the membrane potential, thus unleashing the flow of electrons in the respiratory chain and increasing mitochondrial respiration. However, ATP synthase is bypassed, and this increased respiration produces heat instead of ATP. Energy deprivation can cause cell dysfunction or cell death. Abbreviation: IM, inner membrane.

68

Pessayre et al.

Once a high membrane potential is achieved, this high potential blocks the flow of electrons in the respiratory chain. Uncouplers cause the reentry of protons into the mitochondrial matrix and decrease the mitochondrial membrane potential (72,181), unleashing the flow of electrons in the respiratory chain and increasing basal oxygen consumption (Fig. 9). However, ATP synthase is bypassed, and this increased respiration occurs in vain, to produce heat instead of ATP. Severe uncoupling decreases cell ATP and can cause cell dysfunction or cell death (72). Unlike respiratory chain inhibitors, which impair the re-oxidation of NADH into NADC and may secondarily inhibit mitochondrial b-oxidation (which requires NADC), uncouplers increase respiration, the regeneration of NADC, and tend to increase mitochondrial b-oxidation (179). Steatosis does not occur, unless the drug has other effects on mitochondrial function. Induction of Uncoupling Protein 2 Another mechanism that could uncouple respiration is induction of the uncoupling protein 2 (UCP2). When UCP2 is highly expressed in yeast cells, this inner mitochondrial membrane protein has an uncoupling effect (222). UCP2 could sustain the cycling of FFA across the inner membrane, by permitting the transport of the charged anionic form (RCOOK) from the mitochondrial matrix into the intermembrane space, where RCOOK is protonated to RCOOH, which can cross the lipid bilayer to translocate one proton into the matrix (223). The UCP2 mRNA is expressed in hepatocytes and increased by FFA, ROS, lipopolysaccharide, TNF-a, interleukin-1, and peroxisome proliferator–activated receptor agonists in rodents (224,225). It is still unclear, however, whether the UCP2 protein itself or some other related protein is expressed in hepatocytes (226). Any uncoupling due to UCP2 is probably too mild to cause deleterious pathological changes, although it might accelerate hepatic fatty acid b-oxidation, and might also aggravate the toxicity of ATP-depleting substances (224). MITOCHONDRIAL PERMEABILITY TRANSITION Some drugs directly cause MPT, while other drugs first form reactive metabolites which then trigger permeability transition in the setting of direct toxicity or immune reactions. Parent Drugs Betulinic Acid and Lonidamide Betulinic acid is a pentacyclic triterpene proposed as an anticancer drug. Betulinic acid triggers MPT in isolated mitochondria, even without added calcium, and causes apoptosis in treated cells (227). Betulinic acid also inhibits topoisomerases I and II, which may also contribute to its proapoptotic effects (228). Lonidamide is another investigational antineoplastic agent. Lonidamide targets the ANT and triggers MPT (in the absence of added calcium) and apoptosis (229,230). Hydrochloroquine Mitochondrial effects can also occur as a consequence of lysosomal effects. Indeed, the lysosomotropic antimalarial drug, hydrochloroquine, releases cathepsin B from lysosomes, and causes the translocation of Bcl-2-associated protein x (Bax) from the cytosol to mitochondria, where Bax associates with its analog, Bak to trigger MPT and apoptosis (231). Peripheral Benzodiazepine Receptor Ligands The peripheral benzodiazepine receptor (PBR) is located on the outer mitochondrial membrane and interacts with the MPT pore. In different experimental conditions, PBR ligands can either inhibit or augment apoptosis by modulating MPT. 4 0 -Chlorodiazepam and another selective PBR ligands decreased the mitochondrial membrane potential and triggered apoptosis in hepatic stellate cells (232). In fibroblasts, PBR ligands, albeit not toxic by themselves, increased the MPT and cell death caused by proapoptotic substances, such as TNF-a (233). Likewise, a low concentration of the PBR ligand, alpidem, was not toxic to hepatocytes but increased TNFa-mediated toxicity (234).

Hepatotoxicity Due to Mitochondrial Injury

69

Salicylic Acid, Valproic Acid, Diclofenac, Nimesulide, and Other Acidic Drugs Carboxylic and other acidic drugs can transport protons across the inner mitochondrial membrane. The reentry of protons into the matrix decreases the mitochondrial membrane potential, unleashes the flow of electrons in the respiratory chain and increases mitochondrial respiration, which increases mitochondrial NADH consumption. The resulting NADH depletion can impair ROS detoxification. Let us recall that mitochondrial NADP transhydrogenase uses NADH and the mitochondrial membrane potential to regenerate NADPH from NADPC(235). In turn, NADPH is needed by GSH reductase to reduce GSH disulfide back to GSH (235). Finally, reduced GSH is needed by mitochondrial GSH peroxidase to inactivate the hydrogen peroxide formed by manganese superoxide dismutase (235). Therefore, an insufficient availability of NADH and/or a marked decrease in the mitochondrial membrane potential can impair the detoxification of hydrogen peroxide, and can help trigger MPT (235–237). The opening of the pore in a few mitochondria then allows the egress of NADH, NADC, NADPH, and NADPC, thus further impairing ROS inactivation (237). Pore opening also causes matrix swelling and outer membrane rupture, which releases cyt. c from the damaged mitochondria, thus increasing succinate-supported ROS formation by the affected mitochondria (238). Finally, MPT pore opening also releases calcium from mitochondria. Both the increased ROS formation and calcium release can then trigger MPT in other mitochondria, resulting in a self-amplifying, propagating wave. Thus, several anionic uncouplers, including salicylic acid and valproic acid, have been shown to facilitate calcium-triggered MPT in isolated mitochondria (239). In isolated rat hepatocytes incubated with very low concentrations of albumin, the acidic, nonsteroidal anti-inflammatory drug, nimesulide, uncouples mitochondrial respiration, depletes NAD(P)H, decreases the mitochondrial membrane potential, depletes ATP, and causes cell death (240). Similar effects are observed with the uncoupling drug, diclofenac, whose toxicity towards rat hepatocytes is attenuated by cyclosporin A (241). In addition to the increased NADH consumption due to uncoupling and increased respiration, yet another mechanism for diclofenac-induced NADPH depletion may be redox cycling between 5-hydroxydiclofenac and N,5-dihydroxydiclofenac (242). This metabolic cycle may consume NADPH both during the CYP-mediated oxidation of 5-hydroxydiclofenac into N,5-dihydroxydiclofenac and then during its subsequent reduction into 5-hydroxydiclofenac (242). Like diclofenac, several other acidic nonsteroidal anti-inflammatory drugs caused severe ATP depletion and cell death in isolated rat hepatocytes, including diflunisal, flufenamic acid, mefenamic acid, tolfenamic acid, indomethacin, fenoprofen, and flurbiprofen (243). Troglitazone Troglitazone, a peroxisome proliferator–associated receptor-g agonist, was removed from the market after more than 40 cases of acute liver failure had been reported to the Food and Drug Administration, and once safer analogs had become available (244). Although troglitazone caused mixed hepatitis in some patients, it mostly triggered hepatocellular, sometimes severe, liver injury (245). Serum aminotransferases could continue to rise for many days or weeks after stopping treatment, and the final resolution of liver injury could take months (244). One effect of troglitazone is to inhibit the canalicular bile salt export pump (246). Although this inhibition may contribute to cholestasis in patients with mixed hepatitis, it cannot account for the severe, life-threatening, hepatocellular injury observed in other patients. A possible mechanism for hepatocellular injury is CYP-mediated metabolic activation of troglitazone on both its a-tocopherol moiety and its thiazolidinedione moiety (247). Another possible mechanism is that troglitazone triggers c-Jun N-terminal protein kinase activation, Bid truncation, MPT, mitochondrial membrane potential collapse, mitochondrial cyt. c release, ROS formation, and apoptosis in hepatic cell lines (248,249). These proapoptotic mitochondrial effects occur even in hepatic human cell lines with very low-CYP expression, thus excluding a role of metabolic activation in these effects (248). Which of these diverse effects actually cause hepatitis in humans is unknown. Conceivably, hepatitis could be due to different mechanisms in different subjects. Alternatively, several effects of troglitazone may act together to trigger hepatitis in one patient. For example,

70

Pessayre et al.

increased bile acid levels, mitochondrial effects and/or the direct toxicity of reactive metabolites may kill a few hepatocytes, thus permitting the immunization of some patients against hepatic proteins modified by the covalent binding of reactive metabolites. A positive lymphocyte stimulation test or the presence of hepatic eosinophils or granulomas, have been observed in some cases, suggesting immunoallergic or autoimmune mechanisms in at least some patients (250). Direct Toxicity of Reactive Metabolites The most frequent mechanism of drug-induced hepatitis is the formation of chemically reactive metabolites (1). Free radicals cause lipid peroxidation, while electrophilic metabolites react with GSH or covalently bind to hepatic macromolecules. When only small amounts of electrophilic metabolites are formed, they are detoxified by GSH, and direct toxicity does not occur. When large amounts of electrophilic metabolites are formed, direct toxicity can occur (1). The initial cellular mechanisms causing metabolite-mediated hepatocyte apoptosis (Fig. 10) were studied with germander (251). This medicinal plant was marketed for use in weight control diets (252). This popular indication and the natural medicine fad led to largescale utilization and an epidemic of hepatitis in France (252). Germander contains furano neoclerodane diterpenoids, which are responsible for the in vivo hepatotoxicity of germander in mice (253). In vitro, these furano diterpenoids are activated by CYP3A into electrophilic metabolites (254). Extensive formation of GSH conjugates exceeds the capacity of hepatocytes to resynthesize GSH (254). The resulting GSH depletion causes oxidation of protein thiols, Drug Reactive metabolite

DNA damage

GSH depletion

Covalent binding

MITO

ATP

p53 Bax

Protein thiols

Ca2+

Inactivation of Ca2+-ATPases

Transglutaminase

MPT Ca2+

Arachidonic acid

Cytochrome c

Phospholipase A2

Caspases

Endonucleases CAD/ICAD Cross-linked protein scaffold

DNA fragmentation

FIGURE 10 Involvement of mitochondria in reactive metabolite-mediated direct toxicity. The extensive formation of reactive metabolites may cause GSH depletion, covalent binding to protein thiols and also DNA damage, leading to p53 and Bcl-2-associated x protein (Bax) overexpression. GSH depletion and covalent binding decrease protein thiols and inactivate Ca2C-ATPases. The increase in cell Ca2C activates Ca2C-dependent tissue transglutaminase (forming a cross-linked protein scaffold), endonucleases (contributing to DNA fragmentation) and phospholipase A2 (releasing arachidonic acid). The overexpression of Bax, the oxidation of protein thiols causing disulfide bond formation in the protein structure of the MPT pore, as well as the increases in arachidonic acid and intramitochondrial Ca2C open the MPT pore in some mitochondria. Unaffected mitochondria keep synthesizing ATP, while permeabilized mitochondria release cytochrome c, which activates caspases. Caspase-3 cuts the ICAD, allowing this nuclease (CAD) to enter the nucleus and fragment DNA. Abbreviations: GSH, glutathione; MPT, mitochondrial permeability transition; ICAD, inhibitor of caspase-activated deoxyribonuclease; CAD, caspase-activated deoxyribonuclease.

Hepatotoxicity Due to Mitochondrial Injury

71

which are further decreased by the covalent binding of the metabolite (254). Protein thiol oxidation decreases the activity of plasma membrane calcium ATPases whose role is the constant extrusion of calcium from hepatocytes (255), thus increasing cytosolic Ca2C (Fig. 10) (254). Increased cell Ca2C activates Ca2C-dependent tissue transglutaminase (forming a crosslinked protein scaffold), endonucleases (causing internucleosomal DNA fragmentation), and phospholipase A2 (releasing arachidonic acid) (254). The final cellular events in metabolite-mediated apoptosis were studied with skullcap (28). This medicinal plant also contains diterpenoids transformed into reactive metabolites by CYP3A (28). In addition to the effects mentioned above, effects on MPT were also studied. MPT pore opening was observed, probably due to several mechanisms (Fig. 10). The oxidation of protein thiols can form disulfide bonds within the MPT pore structure, causing pore opening. The increase in cytosolic Ca2C increases intramitochondrial Ca2C, a potent stimulus for pore opening. Increased cell calcium levels may also act indirectly by increasing phospholipase A2 activity and by releasing arachidonic acid, which triggers MPT (256). Finally, reactive metabolites may also damage DNA. This DNA damage increases p53, a transcriptional activator of Bax, which translocates to mitochondria and opens the MPT pore (28). MPT pore opening in some mitochondria causes matrix expansion, outer mitochondrial membrane rupture, and release of mitochondrial cyt. c from the intermembrane space into the cytosol (Fig. 10) (28). Cytosolic cyt. c has been shown to be associated with apoptosis protein-activating factor-1 (apaf-1), causing activation of procaspase-9 (11). The latter activates effector procaspases, including procaspase-3 (11). Caspases cut cytosolic, cytoskeletal, and nuclear proteins, contributing to the ultrastructural lesions of apoptosis (257). Although it was not documented in the skullcap study (28), outer membrane rupture also releases apoptosis inducing factor, triggering large-sized DNA fragmentation (258), while caspase-3 cuts the inhibitor of caspase-activated deoxyribonuclease (ICAD), allowing caspase-activated deoxyribonuclease (CAD) to enter the nucleus and cause further DNA fragmentation (Fig. 10) (259). The internucleosomal DNA fragmentation and apoptotic cell death caused by skullcap diterpenoids were decreased by inhibitors acting on either one of the successive steps shown in Figure 10 (28). Indeed, apoptotic cell death was prevented when metabolic activation was inhibited by a CYP3A inhibitor, when the depletion of cellular thiols was attenuated by GSH precursors, or when the activation of Ca2C-dependent enzymes and Ca2C-induced MPT were inhibited by a calcium/calmodulin inhibitor (28). Cyclosporin A, an MPT inhibitor, prevented cyt. c release, caspase-3 activation, and cell death (28). Finally, aurintricarboxylic acid, an endonuclease and caspase inhibitor and Ac-DEVD-CHO, a caspase-3 inhibitor, also prevented apoptosis (28). Likewise, the hepatotoxicity of acetaminophen towards isolated mouse hepatocytes involves GSH depletion, oxidative stress, JNK activation, and MPT (260,261). A JNK inhibitor attenuates acetaminophen-induced toxicity in vitro (261) and in vivo (262). Immune Reactions Drugs that form small amounts of reactive metabolites do not cause direct toxicity (1,2). These drugs can nevertheless cause hepatitis in some patients, due to immune reactions (1,2). Mechanism of Immunization The immune system recognizes proteins that differ from those of the individual (9). Peptides derived from both self and foreign proteins are transported to the cell surface, where major histocompatibility complex (MHC) class II molecules present them to helper T lymphocytes (9). Somatic clonal mutations of the T cell receptor (TCR) provide a vast array of different lymphocytes. The helper T lymphocytes, which recognize normal peptides are deleted or rendered anergic. Only the helper T lymphocytes, whose receptor recognizes something else, remain active. Some of these lymphocytes can recognize a viral peptide and start the immunization process. The drawback to this system is that a self-protein modified by the covalent binding of a reactive metabolite will also differ from the normal self, which, in some subjects, can trigger immune reactions directed against the modified protein (1,2,9).

72

Pessayre et al.

Mechanisms of Cell Death Hepatocytes express MHC class I molecules that present peptides for possible recognition by cytotoxic T lymphocytes (Fig. 11). When normal peptides are presented, no cytotoxic T lymphocytes recognize these peptides (since autoreactive T lymphocytes are normally deleted or inactivated). In viral hepatitis, viral peptides are presented and recognized by the TCR of cytotoxic T lymphocytes (263,264). Similar mechanisms are probably involved in druginduced immunoallergic hepatitis (9). Due to the alkylation of proteins by reactive metabolites, hepatocytes may express alkylated peptides on their MHC class I molecules (Fig. 11) (9). These peptides differ from normal peptides and may be recognized by the TCR of cytotoxic T lymphocytes (Fig. 11) (9). Cytotoxic lymphocytes perform euthanasia: they help their diseased targets commit suicide (Fig. 11) (9). They express Fas ligand (Fas L) on their surface and express or release TNFa at contact sites with their cellular targets, and they also release granzyme B (263,264). The binding of Fas L to Fas causes the assembly and oligomerization of a signaling complex, including Fas, Fas-associated protein with death domain (FADD) and procaspase-8, which selfactivates (265). The binding of TNF-a to its receptor (TNFR1) has similar effects. TNFR1 associates with TNFR1-associated death domain, which recruits FADD, which recruits and activates pro-caspase-8 (265). After Fas stimulation, the receptor-mediated activation of caspase-8 is weak in hepatocytes, but is amplified by a mitochondrial loop (266). Caspase-8 cuts BH3 interacting domain death agonist (Bid) into truncated Bid (tBid) (267), which causes a conformational change in Bax (268). Bax translocates to mitochondria (269), and associates with Bak (270) to permeabilize the outer mitochondrial membrane (271). MPT also occurs in some mitochondria, causing matrix expansion and outer membrane rupture (27). Mitochondrial cyt. c translocates to the cytosol and activates caspase-9 (272), which activates caspase-3, which then reinitiates the loop

CYTOTOTOXIC T LYMPHOCYTE TCR MHC class I

TNF-α

FAS L Modified peptide

Covalently bound reactive metabolite

HEPATOCYTE tBid Bax

Modified protein

Granzyme B

OM rupture Bax Bak

MPT

Permeabilized OM

Cyt. c Caspases APOPTOSIS

FIGURE 11 Involvement of mitochondria in reactive metabolite-mediated immunoallergic hepatitis. Chemically reactive metabolites covalently bind to hepatic proteins. After protein processing into small peptides, this may lead to the presentation of modified peptides on MHC class I molecules. Some cytotoxic T lymphocytes express a TCR able to recognize these modified structures. Cytotoxic T lymphocytes express Fas L on their surface. They express or release TNF-a, and they release granzyme B. All three substances cut Bid into (tBid), which causes a conformational change in Bax, which then migrates to the mitochondria and associates with Bak to permeabilize the OM. Opening of the MPT pore also occurs in some mitochondria, causing matrix expansion and OM rupture. cyt. c released from permeabilized and ruptured mitochondria activates caspases, which trigger apoptosis. Abbreviations: MHC, major histocompatibility complex; TCR, T cell receptor; Fas L, Fas ligand; Bid, BH3-interacting domain death agonist; tBid, truncated Bid; Bax, Bcl-2-associated x protein; OM, outer membrane; MPT, mitochondrial permeability transition; cyt. c, cytochrome c.

73

Hepatotoxicity Due to Mitochondrial Injury

by activating caspase-8 (273,274) and by directly cutting Bid (275). After several turns of the loop (266), the extensive activation of effector caspases triggers apoptosis. Similar effects occur in TNF-a-induced cell death, with MPT again playing a major role (276,277). Thus, immunoallergic drug-induced hepatitis, like viral hepatitis, involves mitochondrial effects in the final mechanism of cell death (Fig. 11) (9).

IDIOSYNCRASY When investigational drug molecules cause frequent adverse effects in humans, they are not released on the market. At therapeutic doses, drug-induced toxicity only occurs in a few recipients. Reasons behind the susceptibility of these few subjects are incompletely understood. However, both metabolic factors and comorbidity diseases may modulate the hepatotoxicity of drugs impairing mitochondrial function (Fig. 12). Metabolic Factors When the parent drug itself, rather than a metabolite, impairs mitochondrial function, toxicity can be enhanced by any factor decreasing drug elimination. For example, renal insufficiency, which decreases tetracycline elimination, was a risk factor for severe microvesicular steatosis following high intravenous doses of tetracycline (151). Although chloramphenicol inhibits mitochondrial protein synthesis, it is detoxified by glucuronide formation. Its mitochondrial toxicity at high doses was increased in premature or newborn babies, whose glucuroconjugating activity is immature (278). Perhexiline maleate inhibits mitochondrial fat oxidation and energy production, but is detoxified through the formation of hydrosoluble metabolites by CYP2D6 (279). Patients genetically deficient in CYP2D6 were at increased risk of developing perhexiline-induced liver injury (279). Comorbidity Factors Several different causes can add their effects to additively impair mitochondrial function and trigger liver injury (Fig. 12). Thus, Reye’s syndrome can be triggered by the combination of a viral infection and aspirin use, or the combination of a previously latent genetic defect in b-oxidation enzymes and a viral infection (127,128). The prevalence of microvesicular steatosis after large intravenous doses of tetracyclines seemed to be increased by pregnancy (151), which impairs fatty acid oxidation in mice (166). Valproate can inhibit both fatty acid oxidation and pyruvate oxidation, and can aggravate both inborn b-oxidation defects and inborn mitochondrial cytopathies (141–145).

DRUG ± Alcohol

± Obesity diabetes

± Inborn mitochondrial cytopathy

Detoxification ± Viruses

Severe mitochondrial dysfunction

Adverse effects

± Pregnancy

± NO Cytokines ± Inborn β-oxidation defect

FIGURE 12 Metabolic and comorbidity factors involved in idiosyncrasy. The susceptibility of a few subjects to drug-induced mitochondrial dysfunction may be due to impaired detoxification and/or to one or several comorbidity factors, which, like the drug, also impair mitochondrial function.

74

Pessayre et al.

Obesity can cause insulin resistance and hepatic steatosis, which may trigger mitochondrial dysfunction and nonalcoholic steatohepatitis (99). Both obesity and tamoxifen additively increase the risk of steatohepatitis (74).

CONCLUSIONS Drugs or their reactive metabolites can trigger MPT and can cause either necrosis (due to ATP depletion) or apoptosis (due to caspase activation). Necrosis can also be due to drug-induced uncoupling or inhibition of mitochondrial respiration (causing ATP depletion). Drugs may also inhibit b-oxidation and cause microvesicular steatosis, through various mechanisms. Drugs can sequester coenzyme A (aspirin, valproic acid), inhibit mitochondrial b-oxidation enzymes (tetracyclines, 2-arylpropionate anti-inflammatory drugs, amineptine and tianeptine, glucocorticoids, amiodarone, perhexiline, and tamoxifen), impair mitochondrial structure and function (female sex hormones), inhibit mtDNA replication (dideoxynucleosides, fialuridine, tacrine, tamoxifen), decrease the synthesis and stability of mtDNA transcripts (interferon-a), or cause structural damage to mtDNA and mtDNA depletion (alcohol) (7). Severe impairment of oxidative phosphorylation can secondarily inhibit b-oxidation, and can thus cause steatosis. A single drug may have several different effects on mitochondrial function (e.g., valproic acid), and several causes may add their deleterious effects on mitochondrial function (e.g., aspirin and viral infections; valproate and inborn mitochondrial cytopathies or inborn b-oxidation defects). The requirement for these comorbidity factors may play an important role in the apparent idiosyncratic occurrence of these adverse effects. When b-oxidation is impaired, fatty acids are poorly oxidized by mitochondria and are instead esterified into TG which initially accumulate as small lipid vesicles that can progressively coalesce with time into larger vacuoles. Impaired energy production due to the inability to oxidize fatty acids, as well as the mitochondrial toxicity of FFA, dicarboxylic acids and lipid peroxidation products, may explain the severity of microvesicular steatosis, which can cause liver failure, coma, and death (7). Mitochondrial mechanisms for drug-induced toxicity have only been described recently and are rarely investigated during the preclinical development of new drug molecules. However, cases of microvesicular steatosis have led to the recall of diethylaminoethoxyhexestrol (DEAEH), the discontinuation of clinical trials with fialuridine, a limited use of perhexiline or tacrine, as well as early therapeutic misadventures with tetracyclines and valproic acid. We suggest that new drug molecules should be screened for possible mitochondrial effects before they are released on the market. REFERENCES 1. Pessayre D. Role of reactive metabolites in drug-induced hepatitis. J Hepatol 1995; 23(Suppl. 1):16–24. 2. Robin MA, Le Roy M, Descatoire V, Pessayre D. Plasma membrane cytochromes P450 as neoantigens and autoimmune targets in drug-induced hepatitis. J Hepatol 1997; 26(Suppl. 1):23–30. 3. Fromenty B, Pessayre D. Inhibition of mitochondrial beta-oxidation as a mechanism of hepatotoxicity. Pharmacol Ther 1995; 67:101–54. 4. Fromenty B, Pessayre D. Impaired mitochondrial function in microvesicular steatosis. Effects of drugs, ethanol, hormones and cytokines. J Hepatol 1997; 26(Suppl. 2):43–53. 5. Fromenty B, Berson A, Pessayre D. Microvesicular steatosis and steatohepatitis: role of mitochondrial dysfunction and lipid peroxidation. J Hepatol 1997; 26(Suppl. 1):13–22. 6. Pessayre D, Mansouri A, Haouzi D, Fromenty B. Hepatotoxicity due to mitochondrial dysfunction. Cell Biol Toxicol 1999; 15:367–73. 7. Pessayre D, Fromenty B, Mansouri A. Drug-induced steatosis and steatohepatitis. In: Lemasters JJ, Niemenen AL, eds. Mitochondria in Pathogenesis. New York: Plenum Press, 2001:489–517. 8. Pessayre D, Feldmann G, Haouzi D, Fau D, Moreau A, Neuman M. Hepatocyte apoptosis triggered by natural substances (cytokines, other endogenous substances and foreign toxins). Cameron RG, Feuer G, eds. Apoptosis and Its Modulation by Drugs. Handbook of Experimental Pharmacology, Vol. 142. Heidelberg: Springer, 2000:59–108.

Hepatotoxicity Due to Mitochondrial Injury

75

9. Pessayre D, Haouzi D, Fau D, Robin MA, Mansouri A, Berson A. Withdrawal of life support, altruistic suicide, fratricidal killing and euthanasia by lymphocytes: different forms of drug-induced hepatic apoptosis. J Hepatol 1999; 31:760–70. 10. Pessayre D, Berson A, Fromenty B, Mansouri A. Mitochondria in steatohepatitis. Semin Liver Dis 2001; 22:57–69. 11. Green DR, Reed JC. Mitochondria and apoptosis. Science 1998; 281:1309–12. 12. Wallace DC. Mitochondrial disease in man and mouse. Science 1999; 283:1482–8. 13. Saraste M. Oxidative phosphorylation at the fin de sie`cle. Science 1999; 283:1488–93. 14. McGarry JD, Foster DW. Regulation of hepatic fatty acid oxidation and ketone body production. Annu Rev Biochem 1980; 49:395–420. 15. Shigenaga MK, Hagen TM, Ames BN. Oxidative damage and mitochondrial decay in aging. Proc Natl Acad Sci USA 1994; 91:10771–8. 16. LeDoux SP, Driggers WJ, Hollensworth BS, Wilson GL. Repair of alkylation and oxidative damage in mitochondrial DNA. Mutat Res 1999; 434:149–59. 17. Kuchino Y, Mori F, Kasai H, et al. Misreading of DNA templates containing 8-hydroxydeoxyguanosine at the modified base and adjacent residues. Nature 1987; 327:77–9. 18. Pinz KG, Shibutani S, Bogenhagen DF. Action of mitochondrial DNA polymerase g at sites of base loss or oxidative damage. J Biol Chem 1995; 270:9202–6. 19. Berneburg M, Grether-Beck S, Ku¨rten V, Ruzicka T, Briviba K, Sies H, Krutmann J. Singlet oxygen mediates the UVA-induced generation of the photoaging-associated mitochondrial common deletion. J Biol Chem 1999; 274:15345–9. 20. Esposito LA, Melov S, Panov A, Cottrell BA, Wallace DC. Mitochondrial disease in mouse results in increased oxidative stress. Proc Natl Acad Sci USA 1999; 96:4820–5. 21. Cortopassi GA, Shibata D, Soong NW, Arnheim N. A pattern of accumulation of a somatic deletion of mitochondrial DNA in aging human tissues. Proc Natl Acad Sci USA 1992; 89:7370–4. 22. Michikawa Y, Mazzucchelli F, Bresolin N, Scarlato G, Attardi G. Aging-dependent large accumulation of point mutations in the human mtDNA control region for replication. Science 1999; 286:774–9. 23. Sohal RS, Weindruch R. Oxidative stress, caloric restriction, and aging. Science 1996; 273:59–63. 24. Wiesner RJ. Adaptations of mitochondrial gene expression to changing cellular energy demands. News Physiol Sci 1997; 12:178–83. 25. Kroemer G, Reed JC. Mitochondrial control of cell death. Nat Med 2000; 6:513–9. 26. Leist M, Single B, Castoldi AF, Ku¨hnle S, Nicotera P. Intracellular adenosine triphosphate (ATP) concentration: a switch in the decision between apoptosis and necrosis. J Exp Med 1997; 185:1481–6. 27. Feldmann G, Haouzi D, Moreau A, et al. Opening of the mitochondrial permeability transition pore causes matrix expansion and outer membrane rupture in Fas-mediated hepatic apoptosis in mice. Hepatology 2000; 31:674–83. 28. Haouzi D, Lekehal M, Moreau A, et al. Cytochrome P450-generated reactive metabolites cause mitochondrial permeability transition, caspase activation and apoptosis in rat hepatocytes. Hepatology 2000; 32:303–11. 29. Wettereau JR, Gregg RE, Harrity TW, et al. An MTP inhibitor that normalizes atherogenic lipoprotein levels in WHHL rabbits. Science 1998; 282:751–4. 30. Lette´ron P, Sutton A, Mansouri A, Fromenty B, Pessayre D. Inhibition of microsomal triglyceride transfer protein: another mechanism for drug-induced steatosis in mice. Hepatology 2003; 38:133–40. 31. Saudubray JM, Martin D, de Lonlay P, et al. Recognition and management of fatty acid oxidation defects: a series of 107 patients. J Inherit Metab Dis 1999; 22:488–502. 32. Morris AAM. Mitochondrial respiratory chain disorders and the liver. Liver 1999; 19:357–68. 33. Morris AAM, Taanman JW, Blake J, et al. Liver failure associated with mitochondrial DNA depletion. J Hepatol 1998; 28:556–63. 34. Bioulac-Sage P, Parrot-Roulaud F, Mazat JP, et al. Fatal neonatal liver failure and mitochondrial cytopathy (oxidative phosphorylation deficiency): a light and electron microscopic study of the liver. Hepatology 1993; 18:839–46. 35. Watmough NJ, Bindoff LA, Birch-Machin MA, et al. Impaired mitochondrial b-oxidation in a patient with an abnormality of the respiratory chain. Studies in skeletal muscle mitochondria. J Clin Invest 1990; 85:177–84. 36. Lette´ron P, Duchatelle V, Berson A, et al. Increased ethane exhalation, an in vivo index of lipid peroxidation, in alcohol abusers. Gut 1993; 34:409–14. 37. Kukielka E, Cederbaum AI. NADH-dependent microsomal interaction with ferric complexes and production of reactive oxygen species. Arch Biochem Biophys 1989; 275:540–50. 38. Dai Y, Rashba-Step J, Cederbaum AI. Stable expression of human cytochrome P450 2E1 in HepG2 cells: characterization of catalytic activities and production of reactive oxygen intermediates. Biochemistry 1993; 32:6928–37.

76

Pessayre et al.

39. Kukielka E, Dicker E, Cederbaum AI. Increased production of reactive oxygen species by rat liver mitochondria after chronic ethanol treatment. Arch Biochem Biophys 1994; 309:377–86. 40. Nordmann R, Ribie`re C, Rouach H. Implication of free radical mechanisms in ethanol-induced cellular injury. Free Radic Biol Med 1992; 12:219–40. 41. Wieland P, Lauterburg BH. Oxidation of mitochondrial proteins and DNA following administration of ethanol. Biochem Biophys Res Commun 1995; 213:815–9. 42. Mansouri A, Gaou I, De Kerguenec C, et al. An alcoholic binge causes massive degradation of hepatic mitochondrial DNA in mice. Gastroenterology 1999; 117:181–90. 43. Mansouri A, Demeilliers C, Amsellem S, Pessayre D, Fromenty B. Acute ethanol administration oxidatively damages and depletes mitochondrial DNA in mouse liver, brain, heart and skeletal muscles. Protective effects of antioxidants. J Pharmacol Exp Ther 2001; 298:737–43. 44. Demeilliers C, Maisonneuve C, Grodet A, et al. Impaired adaptive resynthesis and prolonged depletion of hepatic mitochondrial DNA after repeated alcohol binges in mice. Gastroenterology 2002; 123:1278–90. 45. Fromenty B, Grimbert S, Mansouri A, Beaugrand M, Erlinger S, Ro¨tig A, Pessayre D. Hepatic mitochondrial DNA deletion in alcoholics: association with microvesicular steatosis. Gastroenterology 1995; 108:193–200. 46. Mansouri A, Fromenty B, Berson A, et al. Multiple hepatic mitochondrial DNA deletions suggest premature oxidative aging in alcoholic patients. J Hepatol 1997; 27:96–102. 47. Hillan KJ, Logan MC, Ferrier RK, Bird GLA, Bennett GL, McKay IC, MacSween RNM. Hepatocyte proliferation and serum hepatocyte growth factor levels in patients with alcoholic hepatitis. J Hepatol 1996; 24:385–90. 48. Ballardini G, Groff P, Zoli M, et al. Increased risk of hepatocellular carcinoma development in patients with cirrhosis and with high hepatocellular proliferation. J Hepatol 1994; 20:218–22. 49. Tsuchishina M, Tsustumi M, Shiroeda H, Yano H, Ueshima Y, Shimanaka K, Takase S. Study of mitochondrial DNA deletion in alcoholics. Alcohol Clin Exp Res 2000; 24:12S–15. 50. Lieber CS. Hepatic, metabolic and toxic effects of ethanol: 1991 update. Alcohol Clin Exp Res 1991; 15:573–92. 51. You M, Fischer M, Deeg MA, Crabb DW. Ethanol induces fatty acid synthesis pathways by activation of sterol regulatory element-binding protein (SREBP). J Biol Chem 2002; 277:29342–7. 52. Lin MCM, Li JJ, Wang EJ, Princler GL, Kauffman FC, Kung HF. Ethanol down-regulates the transcription of microsomal triglyceride transfer protein gene. FASEB J 1997; 11:1145–52. 53. Grunnet N, Kondrup J. The effect of ethanol on the b-oxidation of fatty acids. Alcohol Clin Exp Res 1986; 10:64S–8. 54. Uesugi T, Froh M, Arteel GE, Bradford BU, Thurman RG. Toll-like receptor 4 is involved in the mechanism of early alcohol-induced liver injury in mice. Hepatology 2001; 34:101–8. 55. McKim SE, Ga¨bele E, Isayama F, et al. Inducible nitric oxide synthase is required in alcohol-induced liver injury: studies with knockout mice. Gastroenterology 2003; 125:1834–44. 56. Canbay A, Taimr P, Torok N, Higuchi H, Friedman S, Gores GJ. Apoptotic body engulfment by a human stellate cell line is profibrogenic. Lab Invest 2003; 83:655–63. 57. Cover C, Mansouri A, Knight TR, Bajt ML, Lemasters JJ, Pessayre D, Jaeschke H. Peroxynitriteinduced mitochondrial and endonuclease-mediated nuclear DNA damage in acetaminophen hepatotoxicity. J Pharmacol Exp Ther 2005; 315:879–87. 58. Bender RP, Lindsey RH, Burden DA, Osheroff N. N-acetyl-benzoquinone imine, the toxic metabolite of acetaminophen, is a topoisomerase II poison. Biochemistry 2004; 43:3731–9. 59. Rowe TC, Weissig V, Lawrence JW. Mitochondrial DNA metabolism targeting drugs. Adv Drug Deliv Rev 2001; 49:175–87. 60. Schneider E, Hsiang YH, Liu LF. DNA topoisomerases as anticancer drug targets. Adv Pharmacol 1990; 21:149–83. 61. Li TK, Liu LF. Tumor cell death induced by topoisomerase-targeting drugs. Annu Rev Pharmacol Toxicol 2001; 41:53–77. 62. Zhang H, Barcelo´ JM, Lee B, Kohlhagen G, Zimonjic DB, Popescu NC, Pommier Y. Human mitochondrial topoisomerase I. Proc Natl Acad Sci USA 2001; 98:10608–13. 63. Lin JH, Castora FJ. DNA topoisomerase II from mammalian mitochondria is inhibited by the antitumor drugs, m-ASA and VM-26. Biochem Biophys Res Commun 1991; 176:690–7. 64. Lawrence JW, Darkin-Rattray S, Xie F, Neims AH, Rowe TC. 4-Quinolones cause a selective loss of mitochondrial DNA from mouse L1210 Leukemia cells. J Cell Biochem 1993; 51:165–74. 65. Lawrence JW, Claire DC, Weissig V, Rowe TC. Delayed cytotoxicity and cleavage of mitochondrial DNA in ciprofloxacin-treated mammalian cells. Mol Pharmacol 1996; 50:1178–88. 66. Labowitz JK, Silverman WB. Cholestatic jaundice induced by ciprofloxacin. Dig Dis Sci 1997; 42:192–4. 67. Lopez-Navidad A, Domingo P, Cadafalch J, Farrerons J. Norfloxacin-induced hepatotoxicity. J Hepatol 1990; 11:277–8.

Hepatotoxicity Due to Mitochondrial Injury

77

68. Boya P, Andreau K, Poncet D, et al. Lysosomal membrane permeabilization induces cell death in a mitochondria-dependent fashion. J Exp Med 2003; 197:1323–34. 69. Liguori MJ, Anderson MG, Bukofzer S, et al. Microarray analysis in human hepatocytes suggests a mechanism for hepatotoxicity induced by trovafloxacin. Hepatology 2005; 41:177–86. 70. Lazarczyck DA, Goldstein NS, Gordon SC. Trovafloxacin hepatotoxicity. Dig Dis Sci 2001; 46:925–6. 71. Watkins PB, Zimmermann HJ, Knapp MJ, Gracon SI, Lewis KW. Hepatotoxic effects of tacrine administration in patients with Alzheimer’s disease. JAMA 1994; 271:992–8. 72. Berson A, Renault S, Lette´ron P, et al. Uncoupling of rat and human mitochondria: a possible explanation for tacrine-induced liver dysfunction. Gastroenterology 1996; 110:1878–90. 73. Mansouri A, Haouzi D, Descatoire V, et al. Tacrine inhibits topoisomerases and DNA synthesis to cause mitochondrial DNA depletion and apoptosis in mouse liver. Hepatology 2003; 38:715–25. 74. Bruno S, Maisonneuve P, Castellana P, et al. Incidence and risk factors for non-alcoholic steatohepatitis: prospective study of 5408 women enrolled in Italian tamoxifen chemoprevention trial. Br Med J 2005; 330:932. 75. Larosche I, Lette´ron P, Mansouri A, Vadrot N, Feldmann G, Fromenty B, Pessayre D. Tamoxifen inhibits mitochondrial function, topoisomerases and mitochondrial DNA synthesis and causes progressive depletion of mitochondrial DNA and steatosis in mouse liver. J Hepatol 2005; 42:23–4. 76. Zhang Y, Bell A, Perlman PS, Leibowitz MJ. Pentamidine inhibits mitochondrial intron splicing and translation in Saccharomyces cerevisiae. RNA 2000; 6:937–51. 77. Wiseman A, Attardi G. Reversible tenfold reduction in mitochondria DNA content of human cells treated with ethidium bromide. Mol Gen Genet 1978; 167:51–63. 78. Esnault C, Brown SC, Segal-Bendirdjian E, Coulaud D, Mishal Z, Roques BP, Le Pecq JB. Selective alteration of mitochondrial function by dirtercalinium (NSC 335153), a DNA bisintercalating agent. Biochem Pharmacol 1990; 39:109–22. 79. Schneider Berlin KR, Ammini CV, Rowe TC. Dequalinium induces a selective depletion of mitochondrial DNA from HeLa human cervical carcinoma cells. Exp Cell Res 1998; 245:137–45. 80. Miyako K, Kai Y, Irie T, Takeshige K, Kang D. The content of intracellular mitochondrial DNA is decreased by 1-methyl-4-phenylpyridium ion (MPPC). J Biol Chem 1997; 272:9605–8. 81. Nass MM. Analysis of methylglyoxal bis(guanylhydrazone)-induced alterations of hamster tumor mitochondria by correlated studies of selective rhodamine binding, ultrastructural damage, DNA replication and reversibility. Cancer Res 1984; 44:2677–88. 82. Bergeron RJ, Neims AH, McManis JS, Hawthorne TR, Vinson JR, Bortell R, Ingeno MJ. Synthetic polyamine as antineoplastics. J Med Chem 1988; 31:1183–90. 83. Mitsuya H, Broder S. Inhibition of the in vitro infectivity and cytopathic effect of human T-lymphotrophic virus type III/lymphadenopathy-associated virus (HTLV-III/LAV) by 2 0 ,3 0 dideoxynucleosides. Proc Natl Acad Sci USA 1986; 83:1911–5. 84. Yarchoan R, Mitsuya H, Myers CE, Broder S. Clinical pharmacology of 3 0 -azido-2 0 ,3 0 -dideoxythymidine (zidovudine) and related dideoxynucleosides. N Engl J Med 1989; 321:726–38. 85. Chen CH, Cheng YC. Delayed cytotoxicity and selective loss of mitochondrial DNA in cells treated with the anti-human immunodeficiency virus compound 2 0 ,3 0 -dideoxycytidine. J Biol Chem 1989; 264:11934–7. 86. Lewis W, Dalakas MC. Mitochondrial toxicity of antiviral drugs. Nat Med 1995; 1:417–22. 87. Johnson AA, Ray AS, Hanes J, Suo Z, Colacino JM, Anderson KS, Johnson KE. Toxicity of antiviral nucleoside analogs and the human mitochondrial DNA polymerase. J Biol Chem 2001; 276:40847–57. 88. Chariot P, Drogou I, de Lacroix-Szmania I, et al. Zidovudine-induced mitochondrial disorder with massive liver steatosis, myopathy, lactic acidosis and mitochondrial DNA depletion. J Hepatol 1999; 30:156–60. 89. Walker UA, Ba¨uerle J, Laguno M, et al. Depletion of mitochondrial DNA in liver under antiretroviral therapy with didanosine, stavudine or zalcitabine. Hepatology 2004; 39:311–7. 90. Brivet FG, Nion I, Me´garbane B, Slama A, Brivet P, Rustin P, Munnich A. Fatal lactic acidosis and liver steatosis associated with didanosine and stavudine treatment: a respiratory chain dysfunction? J Hepatol 1999; 30:364–5. 91. Lewis W, Copeland WC, Day BJ. Mitochondrial DNA depletion, oxidative stress and mutation: mechanisms of dysfunction from nucleoside reverse transcriptase inhibitors. Lab Invest 2001; 81:777–90. 92. Barile M, Valenti D, Passarella S, Quagliariello E. 3 0 -Azido-3 0 -deoxythymidine uptake into isolated rat liver mitochondria and impairment of ADP/ATP translocator. Biochem Pharmacol 1997; 53:913–20. 93. de la Asuncion JG, del Olmo ML, Sastre J, Pallardo FV, Vina J. Zidovudine (AZT) causes an oxidation of mitochondrial DNA in mouse liver. Hepatology 1999; 29:985–7. 94. de la Asuncion JG, del Olmo ML, Sastre J, Millan A, Pellin A, Pallardo FV. AZT treatment induces molecular and ultrastructural oxidative damage to muscle mitochondria. Prevention by antioxidant vitamins. J Clin Invest 1998; 102:4–9.

78

Pessayre et al.

95. Martin AM, Hammond E, Nolan D, et al. Accumulations of mitochondrial DNA mutations in human immunodeficiency virus-infected patients treated with nucleoside-analogue reverse transcriptase inhibitors. Am J Hum Genet 2003; 72:549–60. 96. White DJ, Mital D, Taylor S, St John KJC. Sperm mitochondrial DNA deletions as a consequence of long-term highly active antiretroviral therapy. AIDS 2001; 15:1061–2. 97. Bartley PB, Westacott L, Boots RJ, Lawson M, Potter JM, Hyland VJ, Woods ML, II. Large hepatic mitochondrial DNA deletions associated with L-lactic acidosis and highly active antiretroviral therapy. AIDS 2001; 15:419–20. 98. Mallon PWG, Unemori P, Sedwell R, et al. In vivo, nucleoside reverse transcriptase inhibitors alter expression of both mitochondrial and lipid metabolism genes in the absence of depletion of mitochondrial DNA. J Infect Dis 2005; 191:1686–96. 99. Pessayre D, Fromenty B. NASH: a mitochondrial disease. J Hepatol 2005; 42:928–40. 100. Kamal MA, French SW. Drug-induced mitochondrial biogenesis in a liver biopsy. Exp Mol Pathol 2004; 77:201–4. 101. McKenzie R, Fried MW, Sallie R, et al. Hepatic failure and lactic acidosis due to fialuridine (FIAU), an investigational nucleoside analogue for chronic hepatitis B. N Engl J Med 1995; 333:1099–105. 102. Lewis W, Levine ES, Griniuviene B, et al. Fialuridine and its metabolites inhibit DNA polymerase g at sites of multiple adjacent analog incorporation, decrease mtDNA abundance, and cause mitochondrial structural defects in cultured hepatoblasts. Proc Natl Acad Sci USA 1996; 93:3592–7. 103. Thust R, Tomicic M, Klo¨cking R, Voutilainen N, Wutzler P, Kaina B. Comparison of the genotoxic and apoptosis-inducing properties of ganciclovir and penciclovir in Chinese hamster ovary cells transfected with the thymidine kinase gene of herpes simplex virus-1: implications for gene therapeutic approaches. Cancer Gene Ther 2000; 7:107–17. 104. Herraiz M, Beraza N, Solano A, et al. Liver failure caused by herpes simplex virus thymidine kinase plus ganciclovir therapy is associated with mitochondrial dysfunction and mitochondrial DNA depletion. Human Gene Ther 2003; 14:463–72. 105. Samuel CE. Antiviral actions of interferons. Clin Microbiol Rev 2001; 14:778–809. 106. Inagaki H, Matsushima Y, Ohshima M, Kitagawa Y. Interferons suppress mitochondrial gene transcription by depleting mitochondrial transcription factor A (mtTFA). J Interferon Cytokine Res 1997; 17:263–9. 107. Shan B, Vazquez E, Lewis JA. Interferon selectively inhibits the expression of mitochondrial genes: a novel pathway for interferon-mediated responses. EMBO J 1990; 9:4307–14. 108. Le Roy F, Bisbal C, Silhol M, Martinand C, Lebleu B, Salehzada T. The 2-5A/RNase L/RNase inhibitor (RNI) pathway regulates mitochondrial mRNAs stability in interferon-a-treated H9 cells. J Biol Chem 2001; 276:48473–82. 109. Lewis JA, Huq A, Najarro P. Inhibition of mitochondrial function by interferon. J Biol Chem 1996; 22:13184–90. 110. Okanoue T, Sakamoto S, Itoh Y, et al. Side effects of high-dose interferon therapy for chronic hepatitis C. J Hepatol 1996; 25:283–91. 111. Caste´ra L, Kalinsky E, Bedossa P, Tertian G, Buffet C. Macrovacuolar steatosis induced by interferon alfa therapy for chronic myelogenous leukemia. Liver 1999; 19:259–60. 112. Montenez JP, Van Bambeke F, Piret J, Brasseur R, Tulkens PM, Mingeot-Leclercq MP. Interactions of macrolide antibiotics (erythromycin A, roxithromycin, erythromycylamine [Dirithromycin] and azithromycin) with phospholipids: computer-aided conformation analysis and studies on acellular and cell culture models. Toxicol Appl Pharmacol 1999; 156:129–40. 113. Oleinick NL, Corcoran JW. Two types of binding of erythromycin to ribosomes from antibioticsensitive and resistant Bacillus subtilis 168. J Biol Chem 1969; 244:727–35. 114. Anandatheerthavarada HK, Vijayasarathy C, Bhagwat SV, Biswas G, Mullick J, Avadhani NG. Physiological role of the N-terminal processed P4501A1 targeted to mitochondria in erythromycin metabolism and reversal of erythromycin-mediated inhibition of mitochondrial protein synthesis. J Biol Chem 1999; 274:6617–25. 115. Karbowski M, Kurono C, Wozniak M, et al. Free radical-induced megamitochondria formation and apoptosis. Free Radic Biol Med 1999; 26:396–409. 116. Ress BD, Gross EM. Irreversible sensorineural hearing loss as a result of azithromycin ototoxicity. Ann Otol Rhinol Laryngol 2000; 109:435–7. 117. Pessayre D, Larrey D, Funck-Brentano C, Benhamou JP. Drug interactions and hepatitis produced by some macrolide antibiotics. J Antimicrob Chemother 1985; 16:181–94. 118. Kroon AM, de Vries H. The effects of chloramphenicol on the biogenesis of mitochondria of rat liver in vivo. FEBS Lett 1969; 3:208–10. 119. Yunis AA. Chloramphenicol toxicity: 25 years of research. Am J Med 1989; 87:44N–8. 120. Killenberg PG, Davidson ED, Webster LT. Evidence for a medium-chain fatty acid coenzyme A ligase (adenosine monophosphate) that activates salicylate. Mol Pharmacol 1971; 7:260–8.

Hepatotoxicity Due to Mitochondrial Injury

79

121. Deschamps D, Fisch C, Fromenty B, Berson A, Degott C, Pessayre D. Inhibition by salicylic acid of the activation and thus oxidation of long-chain fatty acids. Possible role in the development of Reye’s syndrome. J Pharmacol Exp Ther 1991; 259:894–904. 122. Partin JS, Daugherty CC, McAdams AJ, Partin JC, Schubert WK. A comparison of liver ultrastructure in salicylate intoxication and Reye’s syndrome. Hepatology 1984; 4:687–90. 123. Borutaite´ V, Brown GC. Rapid reduction of nitric oxide by mitochondria, and reversible inhibition of mitochondrial respiration by nitric oxide. Biochem J 1996; 315:295–9. 124. Susin SA, Zamzami N, Kroemer G. Mitochondria as regulators of apoptosis: doubt no more. Biochim Biophys Acta 1998; 1366:151–65. 125. Hurwitz ES, Barrett MJ, Bregman D, Gunn WJ, Schonberger LB, Fairweather WR, Drage JS, Lamontagne JR, Kaslow RA, Burlington DB, et al. Public health service study on Reye’s syndrome and medications. Report of the pilot phase. N Engl J Med 1985; 313:849–57. 126. Forsyth BW, Horwitz RI, Acampora D, et al. New epidemiologic evidence confirming that bias does not explain the aspirin/Reye’s syndrome association. JAMA 1989; 261:2517–24. 127. Remington PL, Rowley D, McGee H, Hall WN, Monto AS. Decreasing trends in Reye syndrome and aspirin use in Michigan, 1979 to 1984. Pediatrics 1986; 77:93–8. 128. Rowe PC, Valle D, Brusilow SW. Inborn errors of metabolism in children referred with Reye’s syndrome: a changing pattern. JAMA 1988; 260:3167–70. 129. Oh KW, Qian T, Brenner DA, Lemasters JJ. Salicylate enhances necrosis and apoptosis mediated by the mitochondrial permeability transition. Toxicol Sci 2003; 73:44–52. 130. Zimmerman HJ. Effects of aspirin and acetaminophen on the liver. Arch Intern Med 1981; 141:333–42. 131. Ponchaut S, Van Hoof F, Veitch K. In vitro effects of valproate and valproate metabolites on mitochondrial oxidations. Relevance of CoA sequestration to the observed inhibitions. Biochem Pharmacol 1992; 43:2435–42. 132. Turnbull DM, Bone AJ, Bartlett K, Koundakjian PP, Sherratt HSA. The effects of valproate on intermediary metabolism in isolated rat hepatocytes and intact rats. Biochem Pharmacol 1983; 32:1887–92. 133. Silva MFB, Ruiter JPN, Ijlst L, Jakobs C, Duran M, de Almeida IT, Wanders RJA. Valproate inhibits the mitochondrial pyruvate-driven oxidative phosphorylation. J Inherit Metab Dis 1997; 20:397–400. 134. Sadeque AJM, Fisher MB, Korzekwa KR, Gonzalez FJ, Rettie AE. Human CYP2C9 and CYP2A6 mediate formation of the hepatotoxin 4-ene-valproic acid. J Pharmacol Exp Ther 1997; 283:698–703. 135. Bjorge SM, Baillie TA. Inhibition of medium-chain fatty acid b-oxidation in vitro by valproic acid and its unsaturated metabolite, 2-n-propyl-4-pentenoic acid. Biochem Biophys Res Commun 1985; 132:245–52. 136. Kassahun K, Farrell K, Abbott F. Identification and characterization of the glutathione and N-acetylcysteine conjugates of (E)-2-propyl-2,4-pentadienoic acid, a toxic metabolite of valproic acid, in rats and humans. Drug Metab Dispos 1991; 19:525–35. 137. Kassahun K, Hu P, Grillo MP, Davis MR, Jin L, Baillie TA. Metabolic activation of unsaturated derivatives of valproic acid. Identification of novel glutathione adducts formed through coenzyme A-dependent and -independent mechanisms. Chem Biol Interact 1994; 90:253–75. 138. Levy RH, Rettenmeier AW, Anderson GD, et al. Effects of polytherapy with phenytoin, carbamazepine and stiripentol on formation of 4-ene-valproate, a hepatotoxic metabolite of valproic acid. Clin Pharmacol Ther 1990; 48:225–35. 139. Farrell G. Drug-Induced Liver Disease. London: Churchill-Livingstone, 1994. 140. Zimmerman HJ, Ishak KG. Valproate-induced hepatic injury: analyses of 23 fatal cases. Hepatology 1982; 2:591–7. 141. Njølstad PR, Skjeldal OH, Agsteribbe E, Huckriede A, Wannag E, Søvik O, Waaler PE. Medium chain acyl-CoA dehydrogenase deficiency and fatal valproate toxicity. Pediatr Neurol 1997; 16:160–2. 142. Kottlors M, Jaksch M, Ketelsen UP, Weiner S, Glocker FX, Lu¨cking CH. Valproic acid triggers acute rhabdomyolysis in a patient with carnitine palmitoyltransferase type II deficiency. Neuromuscul Disord 2001; 11:757–9. 143. Chabrol B, Mancini J, Chretien D, Rustin P, Munnich A, Pinsard N. Valproate-induced hepatic failure in a case of cytochrome c oxidase deficiency. Eur J Pediatr 1994; 153:133–5. 144. Lam CW, Lau CH, Williams JC, Cahn YW, Wong LJC. Mitochondrial myopathy, encephalopathy, lactic acidosis and stroke-like episodes (MELAS) triggered by valproate therapy. Eur J Pediatr 1997; 156:562–4. 145. Kra¨henbu¨hl S, Brandner S, Kleinle S, Liechti S, Straussmann D. Mitochondrial diseases represent a risk factor for valproate-induced fulminant liver failure. Liver 2000; 20:346–8. 146. Ulrich RG, Bacon JA, Brass EP, Cramer CT, Petrella DK, Sun EL. Metabolic, idiosyncratic toxicity of drugs: overview of the hepatic toxicity induced by the anxiolytic, panadiplon. Chem Biol Interact 2001; 134:251–70.

80

Pessayre et al.

147. Melegh B, Pap M, Molna´r D, Masszi G, Kopcsa´nyi G. Carnitine administration ameliorates the changes in energy metabolism caused by short-term pivampicillin medication. Eur J Pediatr 1997; 156:795–9. 148. Fre´neaux E, Labbe G, Lette´ron P, et al. Inhibition of the mitochondrial oxidation of fatty acids by tetracycline in mice and in man: possible role in microvesicular steatosis induced by this antibiotic. Hepatology 1988; 8:1056–62. 149. Labbe G, Fromenty B, Fre´neaux E, Morzelle V, Lette´ron P, Berson A, Pessayre D. Effects of various tetracycline derivatives on in vitro and in vivo b-oxidation of fatty acids, egress of triglycerides from the liver, accumulation of hepatic triglycerides, and mortality in mice. Biochem Pharmacol 1991; 41:638–41. 150. Deboyser D, Goethals F, Krack G, Roberfroid M. Investigation into the mechanism of tetracyclineinduced steatosis: study in isolated hepatocytes. Toxicol Appl Pharmacol 1989; 97:473–9. 151. Zimmerman HJ. Hepatotoxicity: The Adverse Effects of Drugs and Other Chemicals on the Liver. New York: Appleton-Century-Crofts, 1978. 152. Bravo JF, Jacobson MP, Mertens BF. Fatty liver and pleural effusion with ibuprofen therapy. Ann Intern Med 1997; 87:200–1. 153. Victorino RM, Silveira JC, Baptista A, de Moura MC. Jaundice associated with naproxen. Postgrad Med J 1980; 56:368–70. 154. Danan G, Trunet P, Bernuau J, et al. Pirprofen-induced fulminant hepatitis. Gastroenterology 1985; 89:210–3. 155. Dutertre JP, Bastides F, Jonville AP, De Muret A, Sonneville A, Larrey D, Autret E. Microvesicular steatosis after ketoprofen administration. Eur J Gastroenterol Hepatol 1991; 3:953–4. 156. Fre´neaux E, Fromenty B, Berson A, et al. Stereoselective and nonstereoselective effects of ibuprofen enantiomers on mitochondrial b-oxidation of fatty acids. J Pharmacol Exp Ther 1990; 255:529–35. 157. Gene`ve J, Hayat-Bonan B, Labbe G, et al. Inhibition of mitochondrial b-oxidation of fatty acids by pirprofen. Role in microvesicular steatosis due to this nonsteroidal anti-inflammatory drug. J Pharmacol Exp Ther 1987; 242:1133–7. 158. Gene`ve J, Larrey D, Amouyal G, Belghiti J, Pessayre D. Metabolic activation of the tricyclic antidepressant amineptine by human liver cytochrome P-450. Biochem Pharmacol 1987; 36:2421–4. 159. Larrey D, Tinel M, Lette´ron P, Maurel P, Loeper J, Belghiti J, Pessayre D. Metabolic activation of the new tricyclic antidepressant tianeptine by human liver cytochrome P-450. Biochem Pharmacol 1990; 40:545–50. 160. Sbarra C, Castelli MG, Noseda A, Fanelli R. Pharmacokinetics of amineptine in man. Eur J Drug Metab Pharmacokin 1981; 6:123–6. 161. Grislain L, Gele´ P, Bertrand M, Luijten W, Bromet N, Salvadori C, Kamoun A. The metabolic pathways of tianeptine, a new antidepressant, in healthy volunteers. Drug Metab Dispos 1990; 18:804–8. 162. Le Dinh T, Fre´neaux E, Labbe G, et al. Amineptine, a tricyclic antidepressant, inhibits the mitochondrial oxidation of fatty acids and produces microvesicular steatosis of the liver in mice. J Pharmacol Exp Ther 1988; 247:745–50. 163. Fromenty B, Fre´neaux E, Labbe G, Deschamps D, Larrey D, Lette´ron D, Pessayre D. Tianeptine, a new tricyclic antidepressant metabolized by b-oxidation of its heptanoic side chain, inhibits the mitochondrial oxidation of medium and short chain fatty acids in mice. Biochem Pharmacol 1989; 38:3743–51. 164. Kaplan MM. Acute fatty liver of pregnancy. N Engl J Med 1985; 313:367–70. 165. Ebert EC, Sun EA, Wright SH, Decker JP, Librizzi RJ, Bolognese RJ, Lipshutz WH. Does early diagnosis and delivery in acute fatty liver of pregnancy lead to improvement in maternal and infant survival? Dig Dis Sci 1984; 29:453–5. 166. Grimbert S, Fromenty B, Fisch C, et al. Decreased mitochondrial oxidation of fatty acids in pregnant mice: possible relevance to development of acute fatty liver of pregnancy. Hepatology 1993; 17:628–37. 167. Grimbert S, Fisch C, Deschamps D, Berson A, Fromenty B, Feldmann G, Pessayre D. Effects of female sex hormones on liver mitochondria in non-pregnant female mice: possible role in acute fatty liver of pregnancy. Am J Physiol 1995; 268:G107–15 (Gastrointest Liver Physiol 31). 168. Ibdah JA, Bennet MJ, Rinaldo P, Zhao Y, Gibson B, Sims HF, Strauss AW. A fetal fatty-acid oxidation disorder as a cause of liver disease in pregnant women. N Engl J Med 1999; 340:1723–31. 169. Shekhawat P, Bennett MJ, Sadovsky Y, Nelson DM, Rakheja D, Strauss AW. Human placenta metabolizes fatty acids: implications for fetal fatty acid oxidation disorders and maternal liver diseases. Am J Physiol Endocrinol Metab 2003; 284:E1098–105. 170. Mansouri A, Fromenty B, Durand F, Degott C, Bernuau J, Pessayre D. Assessment of the prevalence of genetic metabolic defects in acute fatty liver of pregnancy. J Hepatol 1996; 25:781. 171. Fromenty B, Mansouri A, Bonnefont JP, Courtois F, Munnich A, Rabier D, Pessayre D. Most cases of medium-chain acyl-CoA dehydrogenase deficiency escape detection in France. Hum Genet 1996; 97:367–8.

Hepatotoxicity Due to Mitochondrial Injury

81

172. Lette´ron P, Brahimi-Bourouina N, Robin MA, Moreau A, Feldmann G, Pessayre D. Glucocorticoids inhibit mitochondrial matrix acyl-CoA dehydrogenases and fatty acid b-oxidation. Am J Physiol 1997; 272:G1141–50 (Gastrointest Liver Physiol 35). 173. Itoh S, Igarashi M, Tsukada Y, Ichinoe A. Non-alcoholic fatty liver with alcoholic hyalin after longterm glucocorticoid therapy. Acta Hepato-Gastroenterol 1977; 24:415–8. 174. Noda S, Umezaki H, Yamamoto K, Araki T, Murakami T, Ishii N. Reye-like syndrome following treatment with the pantothenic acid antagonist, calcium hopantenate. J Neurol Neurosurg Psychiatry 1988; 51:582–5. 175. Kodavanti UP, Mehendale HM. Cationic amphiphilic drugs and phospholipid storage disorder. Pharmacol Rev 1990; 42:327–54. 176. Guigui B, Perrot S, Berry JP, et al. Amiodarone-induced hepatic phospholipidosis: a morphological alteration independent of pseudoalcoholic liver disease. Hepatology 1988; 8:1063–8. 177. Fromenty B, Fisch C, Labbe G, et al. Amiodarone inhibits the mitochondrial b-oxidation of fatty acids and produces microvesicular steatosis of the liver in mice. J Pharmacol Exp Ther 1990; 255:1371–6. 178. Fromenty B, Fisch C, Berson A, Lette´ron P, Larrey D, Pessayre D. Dual effect of amiodarone on mitochondrial respiration. Initial protonophoric uncoupling effect followed by inhibition of the respiratory chain at the levels of complex I and complex II. J Pharmacol Exp Ther 1990; 255:1377–84. 179. Fromenty B, Lette´ron P, Fisch C, Berson A, Deschamps D, Pessayre D. Evaluation of human blood lymphocytes as a model to study the effects of drugs on human mitochondria. Effects of low concentrations of amiodarone on fatty acid oxidation, ATP levels and cell survival. Biochem Pharmacol 1993; 46:421–32. 180. Deschamps D, De Beco V, Fisch C, Fromenty B, Guillouzo A, Pessayre D. Inhibition by perhexiline of oxidative phosphorylation and the b-oxidation of fatty acids: possible role in pseudoalcoholic liver lesions. Hepatology 1994; 19:948–61. 181. Berson A, De Beco V, Lette´ron P, et al. Steatohepatitis-inducing drugs cause mitochondrial dysfunction and lipid peroxidation in rat hepatocytes. Gastroenterology 1998; 114:764–74. 182. Simon JB, Manley PN, Brien JF, Armstrong PW. Amiodarone hepatotoxicity simulating alcoholic liver disease. N Engl J Med 1984; 311:167–72. 183. Poucell S, Ireton J, Valencia-Mayoral P, et al. Amiodarone-associated phospholipidosis and fibrosis of the liver. Light, immunohistochemical, and electron microscopic studies. Gastroenterology 1984; 86:926–36. 184. Lewis JH, Ranard RC, Caruso A, et al. Amiodarone hepatotoxicity: prevalence and clinicopathologic correlations among 104 patients. Hepatology 1989; 9:679–85. 185. De La Inglesia FA, Feuer G, Takada A, Matsuda Y. Morphologic studies on secondary phospholipidosis in humans. Lab Invest 1974; 4:539–49. 186. Shikata T, Kanetaka T, Endo Y, Nagashima K. Drug-induced generalized phospholipidosis. Acta Pathol Jpn 1972; 22:517–31. 187. Pessayre D, Bichara M, Feldmann G, Degott C, Potet F, Benhamou JP. Perhexiline maleate-induced cirrhosis. Gastroenterology 1979; 76:170–7. 188. Cortez-Pinto H, Baptista A, Camilo MA, de Costa EB, Valente A, de Moura MC. Tamoxifenassociated hepatitis—report of three cases. J Hepatol 1995; 23:95–7. 189. Oien KA, Moffat D, Curry GW, Dickson J, Habeshaw T, Mills PR, MacSween RNM. Cirrhosis with steatohepatitis after adjuvant tamoxifen. Lancet 1999; 353:36–7. 190. Altan N, Chen Y, Schindler M, Simon SM. Tamoxifen inhibits acidification in cells independent of the estrogen receptor. Proc Natl Acad Sci USA 1999; 96:4432–7. 191. Ioannidis N. Lipidosis induced in rat uteri by high doses of tamoxifen. Ann Anat 1998; 180:315–9. 192. Berson A, Fau D, Fornacciari R, et al. Mechanism for experimental buprenorphine hepatotoxicity: major role of mitochondrial dysfunction versus metabolic activation. J Hepatol 2001; 34:261–9. 193. Berson A, Gervais A, Cazals D, et al. Hepatitis after buprenorphine misuse in heroin addicts. J Hepatol 2001; 34:346–50. 194. Kaufmann P, Torok M, Hanni A, Roberts P, Gasser R, Kra¨henbu¨hl S. Mechanisms of benzarone and benzbromarone-induced hepatic toxicity. Hepatology 2005; 41:925–35. 195. Babany G, Larrey D, Pessayre D, Degott C, Rueff B, Benhamou JP. Chronic active hepatitis caused by benzarone. J Hepatol 1987; 5:332–5. 196. van der Klauw MM, Houtman PM, Stricker BH, Spoelstra P. Hepatic injury caused by benzbromarone. J Hepatol 1994; 20:376–9. 197. Ginsburg H, Geary TG. Current concepts and new ideas on the mechanism of action of quinolonecontaining antimalarials. Biochem Pharmacol 1987; 36:1567–76. 198. Hostetler KY, Reasor M, Yazaki PJ. Chloroquine-induced phospholipid fatty liver. Measurement of drug and lipid concentrations in rat liver lysosomes. J Biol Chem 1985; 260:215–9. 199. Katewa SD, Katyare SS. Treatment with antimalarials adversely affects the oxidative energy metabolism in rat liver mitochondria. Drug Chem Toxicol 2004; 27:41–53. 200. Va´squez-Vivar J, Augusto O. Hydroxylated metabolites of the antimalarial drug primaquine. Oxidation and redox cycling. J Biol Chem 1992; 267:6848–54.

82

Pessayre et al.

201. Maggs JL, Kitteringham NR, Breckenridge AM, Park BK. Autoxidative formation of a chemically reactive intermediate from amodiaquine, a myelotoxin and hepatotoxin in man. Biochem Pharmacol 1987; 36:2061–2. 202. Larrey D, Castot A, Pessayre D, et al. Amodiaquine-induced hepatitis. A report of seven cases. Ann Intern Med 1986; 104:801–3. 203. Boyer PD. What makes ATP synthase spin? Nature 1999; 402:247–9. 204. Cabezo´n E, Montgomery MG, Leslie AGW, Walker JE. The structure of bovine F1-ATPase in complex with its regulatory protein IF1. Nat Struct Biol 2003; 10:744–50. 205. Matsuno-Yagi A, Hatefi Y. Studies on the mechanism of oxidative phosphorylation. Different effects of F0 inhibitors on unisite and multisite ATP hydrolysis by bovine submitochondrial particles. J Biol Chem 1993; 268:1539–45. 206. Salomon AR, Voehringer DW, Herzenberg LA, Khosla C. Apoptolidin, a selective cytotoxic agent, is an inhibitor of F0F1-ATPase. Chem Biol 2001; 8:71–80. 207. Ebel RE, Lardy HA. Influence of aurovertin on mitochondrial ATPase activity. J Biol Chem 1975; 250:4992–5. 208. Linnett PE, Mitchell AD, Osselton MD, Muhleirn LJ, Beechey RB. Citreoviridin, a specific inhibitor of the mitochondrial adenosine triphosphatase. Biochem J 1978; 170:503–10. 209. Cross RL, Kohlbrenner WE. Efrapeptin prevents modification by phenylglyoxal of an essential arginyl residue in mitochondrial adenosine triphosphatase. J Biol Chem 1978; 253:4865–73. 210. Qian T, Herman B, Lemasters J. The mitochondrial permeability transition mediates both necrotic and apoptotic death of hepatocytes exposed to Br-A23187. Toxicol Appl Pharmacol 1999; 154:117–25. 211. Zheng J, Ramirez VD. Rapid inhibition of rat brain mitochondrial proton F0F1-ATPase activity by estrogens: comparison with NaC, KC-ATPase of porcine cortex. Eur J Pharmacol 1999; 368:95–102. 212. Zheng J, Ramirez VD. Inhibition of mitochondrial proton F0F1-ATPase/ATP synthase by polyphenolic phytochemicals. Br J Pharmacol 2000; 130:1115–23. 213. Kokoszka JE, Waymire KG, Levy SE, et al. The ADP/ATP translocator is not essential for the mitochondrial permeability transition pore. Nature 2004; 427:461–5. 214. Buchanan BB, Eiermann W, Riccio P, Aquila H, Klingenberg M. Antibody evidence for different conformational states of ADP/ATP translocator protein isolated from mitochondria. Proc Natl Acad Sci USA 1976; 73:2280–4. 215. Weaver JGR, Tarze A, Moffat TC, et al. Inhibition of adenine nucleotide translocator pore function and protection against apoptosis in vivo by an HIV protease inhibitor. J Clin Invest 2005; 115:1828–38. 216. Chua BH, Shrago E. Reversible inhibition of adenine nucleotide translocation by long chain acylCoA esters in bovine heart mitochondria and inverted submitochondrial particles. J Biol Chem 1977; 252:6711–4. 217. Furuno T, Kanno T, Arita K, et al. Roles of long chain fatty acids and carnitine in mitochondrial membrane permeability transition. Biochem Pharmacol 2001; 62:1037–46. 218. Petrescu I, Tarba C. Uncoupling effects of diclofenac and aspirin in the perfused liver and isolated hepatic mitochondria of rat. Biochem Biophys Acta 1997; 1318:385–94. 219. Moreno-Sa´nchez R, Bravo C, Va´squez C, Ayala G, Silveira LH, Ma´rtinez-Lavı´n M. Inhibition and uncoupling of oxidative phosphorylation by nonsteroidal anti-inflammatory drugs: study in mitochondria, submitochondrial particles, cells and whole heart. Biochem Pharmacol 1999; 57:743–52. 220. Brustrovetski N, Klingenberg M. The reconstituted ADP/ATP carrier can mediate HC transport by free fatty acids, which is further stimulated by Mersalyl. J Biol Chem 1994; 269:27329–36. 221. Wieckowski MR, Wojtczak L. Involvement of the dicarboxylate carrier in the protonophoric action of long-chain fatty acids in mitochondria. Biochem Biophys Res Commun 1997; 232:414–7. 222. Fleury C, Nererova M, Collins S, et al. Uncoupling protein-2: a novel gene linked to obesity and hyperinsulinemia. Nat Genet 1997; 15:269–72. 223. Jaburek M, Varecha M, Gimeno RE, et al. Transport function and regulation of mitochondrial uncoupling proteins 2 and 3. J Biol Chem 1999; 274:26003–7. 224. Cortez-Pinto H, Lin HZ, Yang SQ, Da Costa SOD, Diehl AM. Lipids up-regulate uncoupling protein 2 expression in rat hepatocytes. Gastroenterology 1999; 116:1184–93. 225. Fleury C, Sanchis D. The mitochondrial uncoupling protein-2: current status. Int J Biochem Biophys 1999; 31:1261–78. 226. Pecqueur C, Alves-Guerra MC, Gelly C, et al. Uncoupling protein 2, in vivo distribution, induction upon oxidative stress, and evidence for translational regulation. J Biol Chem 2001; 276:8705–12. 227. Fulda S, Scaffidi C, Susin SA, Krammer PH, Kroemer G, Peter ME, Debatin KM. Activation of mitochondria and release of mitochondrial apoptogenic factors by betulinic acid. J Biol Chem 1998; 273:33942–8. 228. Chowdhury AR, Mandal S, Goswami A, et al. Dehydrobetulinic acid induces apoptosis in Leishmania donovani by targeting DNA topoisomerases I and II: implications is antileishmanial therapy. Mol Med 2003; 9:26–36.

Hepatotoxicity Due to Mitochondrial Injury

83

229. Ravagnan L, Marzo I, Costantini P, et al. Lonidamide triggers apoptosis via a direct, Bcl-2-inhibited effect on the mitochondrial permeability transition pore. Oncogene 1999; 18:2537–46. 230. Belzacq AS, El Hamel C, Vieira HL, et al. Adenine nucleotide translocator mediates the mitochondrial membrane permeabilization induced by lonidamide, arsenic acid and CD437. Oncogene 2001; 15:7579–87. 231. Boya P, Gonzalez-Polo RA, Poncet D, et al. Mitochondrial membrane permeabilization is a critical step of lysosome-initiated apoptosis induced by hydroxychloroquine. Oncogene 2003; 22:3927–36. 232. Fisher R, Schmitt M, Bode JG, Ha¨ussinger D. Expression of the peripheral-type benzodiazepine receptor and apoptosis induction in hepatic stellate cells. Gastroenterology 2001; 120:1212–26. 233. Pastorino JG, Simbula G, Yamamoto K, Glascott PA, Jr, Rothman RJ, Farber JL. The cytotoxicity of tumor necrosis factor depends on induction of the mitochondrial permeability transition. J Biol Chem 1996; 271:29792–8. 234. Berson A, Descatoire V, Sutton A, et al. Toxicity of alpidem, a peripheral benzodiazepine receptor ligand, but not zolpidem in rat hepatocytes. Role of mitochondrial permeability transition and metabolic activation. J Pharmacol Exp Ther 2001; 299:793–800. 235. Kowaltowski AJ, Castilho RF, Vercesi AE. Mitochondrial permeability transition and oxidative stress. FEBS Lett 2001; 495:12–15. 236. Catisti R, Vercesi AE. The participation of pyridine nucleotides redox state and reactive oxygen species in the fatty acid-induced permeability transition in rat liver mitochondria. FEBS Lett 1999; 464:97–101. 237. Brunner M, Moeslinger T, Spieckermann PG. Regulation of cyclosporin A sensitive mitochondrial permeability transition by the redox state of pyridine nucleotides. Comp Biochem Physiol B 2001; 128:31–41. 238. Cai J, Jones DP. Superoxide in apoptosis. Mitochondrial generation triggered by cytochrome c loss. J Biol Chem 1998; 273:11401–4. 239. Trost LC, Lemasters JJ. The mitochondrial permeability transition: a new pathophysiological mechanism for Reye’s syndrome and toxic liver injury. J Pharmacol Exp Ther 1996; 278:1000–5. 240. Mingatto FE, Rodrigues T, Pigoso AA, Uyemura SA, Curti C, Santos AC. The critical role of mitochondrial energetic impairment in the toxicity of nimesulide to hepatocytes. J Pharmacol Exp Ther 2002; 303:601–7. 241. Masubuchi Y, Nakayama S, Horie T. Role of mitochondrial permeability transition in diclofenacinduced hepatocyte injury in rats. Hepatology 2002; 35:544–51. 242. Bort R, Ponsoda X, Jover R, Go´mez-Lecho´n J, Castell JV. Diclofenac toxicity to hepatocytes: a role for drug metabolism in cell toxicity. J Pharmacol Exp Ther 1999; 288:65–72. 243. Masubuchi Y, Saito H, Horie T. Structural requirements for the hepatotoxicity of nonsteroidal antiinflammatory drugs in isolated rat hepatocytes. J Pharmacol Exp Ther 1998; 287:208–13. 244. Watkins PB. Insight into hepatotoxicity: the troglitazone experience. Hepatology 2005; 41:229–30. 245. Murphy EJ, Davern TJ, Shakil AO, et al. Troglitazone-induced fulminant hepatic failure. Acute Liver Failure Study Group. Dig Dis Sci 2000; 45:549–53. 246. Funk C, Ponelle C, Scheuermann G, Pantze M. Cholestatic potential of troglitazone as a possible factor contributing to troglitazone-induced hepatotoxicity: in vivo and in vitro interaction at the canalicular bile salt export pump (Bsep) in the rat. Mol Pharmacol 2001; 59:627–35. 247. Kassahun K, Pearson PG, Tang W, et al. Studies on the metabolism of troglitazone to reactive intermediates in vitro and in vivo. Evidence for novel pathways involving quinone methide formation and thiazolidinedione ring scission. Chem Res Toxicol 2001; 14:62–70. 248. Bae MA, Song BY. Critical role of c-Jun N-terminal protein kinase activation in troglitazone-induced apoptosis of human HepG2 hepatoma cells. Mol Pharmacol 2003; 63:401–8. 249. Shishido S, Koga H, Harada M, et al. Hydrogen peroxide overproduction in megamitochondria of troglitazone-treated human hepatocytes. Hepatology 2003; 37:136–47. 250. Kohlroser J, Mathai J, Reichheld J, Banner BF, Bonkovsky HL. Hepatotoxicity due to troglitazone: report of two cases and review of adverse events reported to the United States Food and Drug Administration. Am J Gastroenterol 2000; 95:272–6. 251. Fau D, Lekehal M, Farrell G, et al. Diterpenoids from germander, a herbal medicine, induce apoptosis in isolated rat hepatocytes. Gastroenterology 1997; 113:1334–46. 252. Larrey D, Vial T, Pauwels A, Castot A, Biour M, David M, Michel H. Hepatitis after germander (Teucrium chamaedrys) administration: another instance of herbal medicine hepatotoxicity. Ann Intern Med 1992; 117:129–32. 253. Loeper J, Descatoire V, Lette´ron P, et al. Hepatotoxicity of germander in mice. Gastroenterology 1994; 106:464–72. 254. Lekehal M, Pessayre D, Lereau JM, Moulis C, Fouraste´ I, Fau D. Hepatotoxicity of the herbal medicine, germander. Metabolic activation of its furano diterpenoids by cytochrome P450 3A depletes cytoskeleton-associated protein thiols and forms plasma membrane blebs in rat hepatocytes. Hepatology 1996; 24:212–8.

84

Pessayre et al.

255. Bellomo G, Orrenius S. Altered thiol and calcium homeostasis in oxidative hepatocellular injury. Hepatology 1985; 5:876–82. 256. Penzo D, Petronilli V, Angelin A, et al. Arachidonic acid released by phospholipase A2 activation triggers Ca2C-dependent apoptosis through mitochondrial pathway. J Biol Chem 2004; 279:25219–25. 257. Evan G, Littlewood T. A matter of life and death. Science 1998; 281:1317–22. 258. Susin SA, Lorenzo HK, Zamzami N, Marzo I, Snomw BE, Brothers GM, Mangion J, Jacotot E, Costantini P, Loeffler M, et al. Molecular characterization of mitochondrial apoptosis-inducing factor. Nature 1999; 397:441–6. 259. Sakahira H, Enari M, Nagata S. Cleavage of CAD inhibitor in CAD activation and DNA degradation during apoptosis. Nature 1998; 391:96–9. 260. Reid AB, Kurten RC, McCullough SS, Brock RW, Hinson JA. Mechanism of acetaminophen-induced hepatotoxicity: role of oxidative stress and mitochondrial permeability transition in freshly isolated mouse hepatocytes. J Pharmacol Exp Ther 2005; 312:508–16. 261. Matsumaru K, Ji C, Kaplowitz N. Mechanism for sensitization to TNF-induced apoptosis by acute glutathione depletion in murine hepatocytes. Hepatology 2003; 37:1425–34. 262. Gunawan B, Liu ZX, Kaplowitz N. JNK plays a major role in acetaminophen hepatotoxicity. Hepatology 2004; 40:276A–7. 263. Kondo T, Suda T, Fukuyama H, Adachi M, Nagata S. Esssential roles of the Fas ligand in the development of hepatitis. Nat Med 1997; 3:409–13. 264. Ando K, Hiroishi K, Kaneko T, et al. Perforin, Fas/Fas ligand, and TNF-a pathways as specific and bystander killing mechanisms of hepatitis C virus-specific human CTL. J Immunol 1997; 158:5283–91. 265. Nagata S. Apoptosis by death factor. Cell 1997; 88:355–65. 266. Bajt ML, Lawson JA, Vonderfecht SL, Gujral JS, Jaeshcke H. Protection against Fas receptor-mediated apoptosis in hepatocytes and non-parenchymal cells by a caspase-8 inhibitor in vivo: evidence for a post-mitochondrial processing of caspase-8. Toxicol Sci 2000; 58:109–17. 267. Li H, Zhu H, Xu C, Yuan J. Cleavage of BID by caspase 8 mediates the mitochondrial damage in the Fas pathway of apoptosis. Cell 1998; 94:491–501. 268. Desagher S, Osen-Sand A, Nichols A, et al. Bid-induced conformational change of Bax is responsible for mitochondrial cytochrome c release during apoptosis. J Cell Biol 1999; 144:891–901. 269. Wolter KG, Hsu YT, Smith CL, Nechushtan A, Xi XG, Youle RJ. Movement of Bax from the cytosol to mitochondria during apoptosis. J Cell Biol 1997; 139:1281–92. 270. Nechushtan A, Smith CL, Lamensdorf I, Yoon SH, Youle RJ. Bax and Bak coalesce into mitochondriaassociated clusters during apoptosis. J Cell Biol 2001; 153:1265–76. 271. Wei MC, Zong WX, Cheng EHY, et al. Proapoptotic Bax and Bak: a requisite gateway to mitochondrial dysfunction and death. Science 2001; 292:727–30. 272. Saleh A, Srinivasula SM, Acharya S, Fishel R, Alnemri ES. Cytochrome c and dATP-mediated oligomerization of Apaf-1 is a prerequisite for procaspase-9 activation. J Biol Chem 1999; 274:17941–5. 273. Van de Craen M, Declercq W, Van den Brande I, Fiers W, Vandenabeele P. The proteolytic procaspase activation network: an in vitro analysis. Cell Death Differ 1999; 6:1117–24. 274. Slee EA, Harte MT, Kluck RM, et al. Ordering the cytochrome c-initiated caspase cascade: hierarchical activation of caspases-2, -3, -6, -7, -8, and -10 in a caspase-9-dependent manner. J Cell Biol 1999; 144:281–92. 275. Bossy-Wetzel E, Green DR. Caspases induce cytochrome c release from mitochondria by activating cytosolic factors. J Biol Chem 1999; 274:17484–90. 276. Angermu¨ller S, Kunstle G, Tiegs G. Pre-apoptotic alterations in hepatocytes of TNFa-treated galactosamine-sensitized mice. J Histochem Cytochem 1998; 46:1175–83. 277. Tafani M, Schneider TG, Pastorino JG, Farber JL. Cytochrome c-dependent activation of caspase-3 by tumor necrosis factor requires induction of the mitochondrial permeability transition. Am J Pathol 2000; 156:2111–21. 278. Weiss CF, Glazko AJ, Weston JK. Chloramphenicol in the newborn infant. A physiologic explanation for its toxicity when given in excessive doses. N Engl J Med 1960; 262:787–94. 279. Morgan MY, Reshef R, Shah RR, Oates NS, Smith RL, Sherlock S. Impaired oxidation of debrisoquine in patients with perhexiline liver injury. Gut 1984; 10:1057–64.

5

Mechanisms of Cell Death and Relevance to Drug Hepatotoxicity Neil Kaplowitz

Keck School of Medicine, University of Southern California, Los Angeles, California, U.S.A.

INTRODUCTION The purpose of this chapter will be to provide a brief overview of the subject of cell death and then to focus on what is known about the role of apoptosis and necrosis in drug hepatotoxicity. OVERVIEW OF CELL DEATH It is currently recognized that the demise of cells reflects the triggering of the activation of a death program either by death receptor signaling or from a source of intracellular stress leading to apoptosis versus the massive loss of cell integrity from overwhelming stress leading to necrosis (1,2). The former involves shrinkage and nuclear disassembly (apoptosis) and the latter involves swelling and lysis (necrosis). The type of cell death and the susceptibility to death-inducing stimuli vary greatly from cell type to cell type and from transformed cells to normal cells. In the liver, death of hepatocytes is the major event leading to organ failure but in special circumstances the sinusoidal endothelial cells (e.g., veno-occlusive disease) (3) or bile duct epithelium (vanishing duct syndrome) (4) may be a key target. Hepatocyte death accounts for the key findings in drug-induced hepatitis, namely elevated serum aspartate aminotransferase, alanine aminotransferase, and functional disorders due to parenchymal extinction, such as jaundice and coagulopathy. Apoptotic Cell Death Apoptosis is a form of cell death that involves the shrinkage and disassembly of the nucleus and cytoskeleton so that the cell is broken down into small fragments (Councilman bodies seen in histology of liver), which undergo rapid clearance by phagocytosis by surrounding cells or professional phagocytes; apoptotic cell death (unless massive) tends not to elicit inflammation and is therefore a mechanism for “quiet” removal. The entire “machinery” of apoptosis is listed in Table 1. The process of dismantling involves the participation of proteolytic enzymes, caspases, which are present in zymogen forms and are activated in a cascade from initiator to executioner members of this class (5). The trigger for the activation process to begin occurs either at the cell surface, where a death receptor binds a ligand, or from an intracellular stress that initiates the process independent of death receptors (Fig. 1). Death receptors of major significance in liver include tumor necrosis factor (TNFa) receptor 1 (TNFR1) and Fas, which bind TNFa (soluble or membrane bound) and FasL (on T cells), respectively. When ligand binds, it causes aggregation of receptors leading to conformational changes on the cytoplasmic side so that adaptor or scaffolding proteins associate with the receptor, i.e., TNFR-associated death domain and Fas-associated death domain (TRADD and FADD). These then bind procaspase 8 causing it to self-activate by cleavage to release caspase 8. In some cell types, The author of this chapter has relationships with the following corporations: Abbott, Adams Respiratory Therapy, Allergan, Amgen, Astra Zeneca, Avera, BG Medicine, Biogen, Boehringer/Ingelhem, Cadence, Daiichi Sankyo, DOV, Elan, Enanta, Encysive, GSK, Gtx, Incyte, ISIS, Janssen, Johnson & Johnson, Maxygen, Merck, Millenium, Ono, Pfizer, Rigel, Roche, Sankyo, TAP, Threshold, Teva, and Wyeth.

86

Kaplowitz

TABLE 1 The Machinery of Apoptosis Key participants Death receptors (Fas, TNF-R1) Caspases (Initiators caspase 8,9,10) (Executioners caspase 3,6,7) Bid, Bax, Bak, Bim, Puma, Noxa, Bad Mitochondrial proteins (cytochrome c, AIF, Smac/Diablo), ?MPT Other factors p53 Ceramide Oxidative stress ER stress, (GADD 153/CHOP) Cathepsins Granzyme JNK, p38 kinases NOC O2_ Survival factors NF-kB (lAPs etc.) PI-3-Kinase/Akt IAPs (XIAP, IAP-1, IAP-2) FLIP HSP NOC O2_ Bcl2, Bcl-XL, Mcl-1 Abbreviations: IAP, apoptosis inhibitors; NF-kB, nuclear factor-kappa B; TNF-R1, tumor necrosis factor R1; HSP, heat shock protein; MPT, mitochondrial permeability transition; AIF, apoptosis inducing factor; JNK, c-jun-N-terminal kinase; FLIP, inhibitory protein; ER stress, endoplasmic reticulum stress.

sufficient initiator caspase 8 is released to activate procaspase 3 to produce sufficient caspase 3 (executioner) to carry out the actual apoptosis. However, in hepatocytes the death receptorinduced formation of caspase 8 is insufficient to activate caspase 3 directly and an amplification mechanism is required, which involves the participation of mitochondria with the release of intermembrane proteins such as cytochrome c leading to the assembly on a cytoplasmic scaffold (apaf-l) for cytochrome c, procaspase 9, and ATP (apoptosome) (6). Self-cleavage releases caspase 9, an initiator caspase, which then cleaves procaspase 3 to release its active form. Caspase 3 then cleaves a number of specific proteins as well as procaspase 6, 7, and 2, which may have their respective specific targets (7). When Fas is ligated the death-inducing signaling Fasl TNFα

Fas TNFR-1

FADD TRADD (DISC)

Caspase 8 Intracellular stress

tBid

Oxidative stress ER stress DNA damage

Bax, Bak Apoptosis

Caspase 3 Caspase 9 Apaf-1 ATP Procaspase 9 (apoptosome)

Mitochondria Cytochrome c

?MPT

AIF Smac/Diablo Procaspases 2,9 (?3)

FIGURE 1 Apoptosis cascade: death receptor and intracellular stress pathways emphasizing the central role of mitochondria.

Cell Death and Drug Hepatotoxicity

87

complex (DISC) assembles at the plasma membrane, whereas with TNFR1 the complex forms in two stages: complex 1 forms at the plasma membrane when TRADD, receptor interacting protein (RIP), and TNF receptor associated factor 2 (TRAF2) associate with the cytoplasmic side of the receptor; this leads to activation of the IKB kinase or inhibitor of KB kinase–nuclear factor kB (IKK-NF-kB) pathway and mitogen activated protein kinases, complex II then forms after intracellular translocation of TRADD, RIP, TRAF2 without TNFR1 leading to association of FADD and procaspase 8, and propagation of the death pathways. However, the death pathway will not progress in the face of NF-kB survival gene expression which inhibits multiple steps in the pathway. Mitochondria participate in apoptosis in many cell types. With death receptor signaling, the generation of activated caspase 8 cleaves a Bcl2 family member, Bid, to form tBid. tBid causes Bax to insert in mitochondria and Bak to self-aggregate (6,8). Either Bax or Bak or both cause the outer mitochondrial membrane to become permeable, leading to the release of cytochrome c. The intermembrane space contains other proteins that may participate, including some procaspases, AIF, and Smac/Diablo. The latter binds apoptosis inhibitors, such as IAP1, IAP2, and XIAP. The IAPs are cytoplasmic proteins that inhibit caspases so the release of Smac/ Diablo incapacitates these caspase inhibitors allowing apoptosis to proceed (9,10). Considerable controversy and uncertainty centers around the precise mechanism of release of intermembrane proteins with regard to the role of the mitochondrial permeability transition (MPT). This is a pore, composed of inner- and outer-membrane proteins, which is normally closed but when opened causes depolarization and release of proteins. The pore consists of the outer-membrane voltage-dependent anion channel, and peripheral benzodiazepine receptor, inner membrane adenine nucleotide translocase (ANT), matrix cyclophilin D (which binds cyclosporin A), and cytosol hexokinase and creatine kinase. The pore is opened by oxidative stress and excess CaC2. The pore contains functional vicinal thiols so that it is responsive to changes in the thiol-disulfide status of the milieu; i.e., disulfide formation opens the pore (11). Opening is inhibited by cyclosporin A and promoted by atractyloside and brongkrekic acid binding to ANT. MPT plays a key role in ischemic cell death but current evidence suggests a much less important role in extrinsic and intrinsic pathway mediated apoptosis. It is not clear if MPT is important in drug hepatotoxicity but, since cyclosporine A protects against acetaminophen, a role for MPT in necrosis is suggested. Recently, selective mitochondrial outermembrane permeabilization (MOMP) has been defined as the key mechanism for allowing release of intermembrane mitochondrial proteins, such as cytochrome c. Proapoptotic Bcl2 members such as Bax promote MOMP and antiapoptotic members such as Bcl2 and Bcl-XL, inhibit opening. Thus, Bcl2 family govern MOMP. Permeabilization is mediated by Bax and Bak, which are directly activated by tBid or Bim, whereas Bcl2 and Bcl-XL bind the latter, preventing their activation of Bax and Bak. However, other proapoptotic Bcl2 members, referred to as “derepressors,” do not directly activate Bax or Bak but compete for binding to Bcl2 and Bcl-XL, releasing tBid or Bim. Members of the family which act in this way include Puma, Noxa, and Bad. Another protein, p53, which is not a Bcl2 member can also directly activate Bax. Activation of Bax and Bak involves oligomerization of both and membrane insertion of Bax, which creates pores in the outer mitochondrial membrane. Bax and Bak appear to have redundant function [see (12) and (13) for review]. Apoptosis may also be triggered by events within the cell that occur downstream of death receptors, caspase 8, and/or tBid. This usually involves some type of stress—oxidative stress, DNA damage, endoplasmic reticulum stress, etc. Drug toxicity might cause any of these phenomena to occur. The resultant stress usually leads to participation of mitochondria but the precise mechanisms leading to their participation are not well established but could include alterations in the balance of Bcl2 members, the participation of p53, or direct effects on the MPT pore. A number of factors serve to inhibit apoptosis. Some are under the control of the transcription factor, NF-kB (14), and include the IAPs (9) and inducible nitric oxide synthase (15) (NO inhibits caspases). Stress kinases, such as c-jun-N-terminal kinase (JNK) and p38, have been associated with pro and anti-apoptotic effects (16). Heat shock protein (HSP) can inhibit caspases (17). Hepatocytes are extremely resistant to lethal actions of TNFa because

88

Kaplowitz

TNFa not only activates the apoptosis cascade, but also activates NF-kB leading to upregulation of survival genes. The resistance to TNFa can be overcome by inhibition of transcription (18) or depletion of glutathione (GSH) (17–22). The former blocks the production of survival gene products, whereas the mechanism for the latter is less certain, but may include increased susceptibility of mitochondria to oxidative stress or alterations in redox control of kinases and transcription factors (tipping the balance towards proapoptotic events). Necrotic Cell Death Necrosis is a lytic cell death that usually involves cell swelling and rupture due to loss of the ability to maintain ion gradients and active transport as a consequence of profound ATP depletion. Thus, loss of mitochondrial function plays a key role. The determination of whether cells will die by apoptosis or necrosis appears to depend on how severely impaired mitochondria become. Some critical level of ATP is required for the function of the apoptosome (23). Furthermore, if extensive reactive oxygen species (ROS) or NO production occurs, mitochondria may release cytochrome c but the ROS and NO may inactivate caspases leaving the cell to progressively swell and lyse owing to loss of mitochondrial electron transport. Thus, depending on the triggering phenomenon, cells may be committed to die as a result of effects on mitochondria and the mode of death depends on the status of ATP and the extent of inhibition of caspases. In other circumstances, the effects on mitochondria may be only sufficient to cause apoptosis or may be so profound as to lead to rapid necrosis. This phenomenon has been referred to as programed necrosis (or abortive apoptosis). Either death receptor ligation (extrinsic mechanism) or intracellular stress (intrinsic mechanism) may activate the pathways (caspase 8, proapoptotic Bcl-2 family, JNK) of programed cell death upstream of mitochondria but executioner caspases may be inhibited while mitochondrial functional deterioration proceeds to ATP depletion and oncotic necrosis. Other unusual phenomena that overlap apoptosis and necrosis have been described in special circumstances and have been described as aponecrosis, paraptosis, and caspase-independent apoptosis (24–28). These are not well understood but underscore the concept that cell death may not simply occur as apoptosis or necrosis, but may represent a continuum or spectrum.

CELL DEATH IN DRUG HEPATOTOXICITY Surprisingly, little is known about the role of apoptosis versus necrosis in drug hepatotoxicity. It seems reasonable to assume that immune hypersensitivity reactions directed at the liver involve apoptosis. The histological picture of mononuclear cell infiltration and spotty “necrosis” of individual hepatocytes (Councilman bodies), resembling the histological picture of viral hepatitis, supports this view; viral hepatitis mainly induces apoptotic death of liver cells. The immune-mediated killing is directed at hepatocytes through antigen recognition major histocompatibility complex 1 (MHC1) and the likely participation of either FasL binding or the porin-dependent delivery of granzyme to the cytoplasm. Granzymes can directly cleave procaspase 8 upstream of mitochondria or perhaps procaspase 3 downstream (29). The potential for the participation of the Fas pathway is evident from the fact that agonistic monoclonal antibody to Fas can induce fulminant hepatic failure in mice as a consequence of massive hepatocellular apoptosis (18). The role of apoptosis in the action of direct hepatotoxins such as CC14 or acetaminophen has been investigated, whereas the role in idiosyncratic delayed reactions is not known. As noted in the chapter by Laskin in this volume, activation of Kupffer cells plays an important role in direct toxicity of CC14 and acetaminophen. In the case of CC14, the role of TNFa produced by Kupffer cells has been well established in that toxicity is abrogated by immunoneutralization of TNFa (30) or use of TNFR1 knockouts (31). In the case of acetaminophen, some controversy exists. Although macrophage inhibitors protect against acetaminophen (32), TNFa knockout mice are not protected (33), however, the susceptibility to acetaminophen is enhanced in C–C chemokine receptor 2 (CCR2) knockouts, an effect that

89

Cell Death and Drug Hepatotoxicity

can be attenuated by immunoneutralization of TNFa or Interferon g (IFNg) (34); CCR2 expression thus is protective through regulation of cytokine generation by MCP-1 (Fig. 2). Furthermore, others have shown that neutralization of TNFa seems to slow the development of acetaminophen-induced injury and speed its resolution (35,36). Interleukin-10 knockout has been shown to sensitize to acetaminophen (37). Depletion of natural killer cells and natural killer cells with T-cell receptor cells, IFNg knockout and deficiency of Fas or FasL all have been shown to diminish acetaminophen hepatotoxicity without affecting its metabolism (38). Limited data are available with other hepatotoxicants. Lipopolysccharide (LPS) enhances allyl alcohol toxicity by a mechanism independent of TNFa but abrogated by pentoxifylline (39). As discussed below, hepatotoxicants may be directly lethal to hepatocytes or may sensitize hepatocytes to the lethal actions of TNFa or other components of the innate immune system. In either case, there is considerable uncertainty as to the mode of cell death. CCl4 When rats were administered a modest dose of carbon tetrachloride (CCl4), a mixture of apoptosis and necrosis was observed. Although TUNEL staining may not be completely reliable in distinguishing apoptosis from necrosis in situ, the TUNEL staining was correlated with characteristic cellular morphological changes of apoptosis as well as DNA laddering (40). It is difficult to quantitate the extent of apoptosis since these cells are rapidly phagocytozed. However, apoptosis was appreciable. It is speculated that the dose of CC14 may be a factor with massive exposure leading to overwhelming necrosis and lesser exposure leading to a mixture of necrosis and apoptosis. It is uncertain but certainly plausible that the apoptosis, and perhaps the necrosis, are triggered by TNFa in the CCL4-sensitized hepatocytes. Acetaminophen Considerable controversy has existed concerning the mode of liver cell death induced by acetaminophen and its mechanism. The classic view is that covalent binding of N-acetyl-pbenzoquinone imine (NAPQI) to critical proteins mediates the lethal toxicity but it should be recognized that covalent binding does not occur until all the cell GSH has been depleted, including mitochondrial GSH for covalent binding in this organelle (which correlates with Wt

CCR2–/–

Antisera NPC

NPC TNF α IFN γ

CCR2 (–)

APAP

(–

IFN γ APAP MCP1

MCP1 HC

)

TNF α

HC

Lethal

FIGURE 2 Mechanism for increased acetaminophen hepatotoxicity in chemokine CCR2 null mice compared to Wt. APAP acts on HC to sensitize to the lethal actions of TNFa and IFNg while upregulating MCP-1. MCP-1 acting via CCR2 on NPC, particularly Kupffer cells, downregulates the cytokines. In null mice, MCP-1 does not have a receptor leading to augmented TNFa and IFNg production and lethal effects on the sensitized hepatocytes. Abbreviations: CCR2, c receptor 2; Wt, wild-type; APAP, acetaminophen; HC, hepatocytes; TNFa, tumor necrosis factor a; IFNg, interferon g; MCP-1, monocyte chemoattractant protein-1; NPC, nonparenchymal cells.

90

Kaplowitz

killing). Thus, profound GSH depletion itself may exert a lethal action owing to the loss of the defense against endogenous ROS normally produced in mitochondria (and possibly enhanced by the action of TNF) leading to lethal oxidative stress. In cultured mouse hepatocytes, the suspectibility to acetaminophen killing is potentiated by carmustine (BCNU) inhibition of glutathione disulfide (GSSG) reductase and abrogated by iron chelators (41,42). Liposome-encapsulated superoxide dismutases (SOD) protects rats against acetaminopheninduced liver injury without altering covalent binding or GSH depletion (43). Similar results have been observed with an SOD mimic (44) and intravenous SOD itself (45). Although oxidative stress appears to play a major role in acetaminophen-induced liver injury, the contribution of apoptosis versus necrosis is less certain. Earlier histological evaluation suggested the appearance of apoptotic hepatocytes at the proximal edge of the necrotic zone (46). Subsequent studies in the rat have suggested that nearly half the dead cells are apoptotic by morphological criteria with confirmation by the appearance of electrophoretic DNA laddering (47). Others have confirmed these findings more recently using TUNEL staining and increased caspase 9 and 3 activities (44). In contrast, Jaeschke and coworkers have observed that caspases are inhibited after acetaminophen treatment (? direct arylation or inhibitory effects of reactive oxygen or nitrogen species) (48) so appearance of TUNEL positive cells and DNA fragmentation in the liver after acetaminophen administration suggests a role for Ca2C-activated endonuclease and a necrotic cell death (49,50). Indeed, prevention of acetaminophen-induced increased cell calcium protects against DNA fragmentation (51). Furthermore, DNase 1 null mice exhibit significant protection against acetaminophen induced NADC and ATP depletion and centrilobular necrosis. This fits with the concept that acetaminophen disrupts CaC2 homeostasis and releases Dnase 1 from intracellular stores leading to DNA damage which then activates PARP which consumes NADC and ATP. DNA damage contributes to toxicity by further depleting ATP (52). Thus, the bulk of evidence favors the view that acetaminophen-induced profound GSH depletion leads to lethal oxidative stress with the mode of cell death being caspase-independent and therefore presumably necrotic. The appearance of TUNEL-positive cells and DNA fragmentation in this case are features of necrotic cell death, underscoring the lack of specificity of these phenomena. However, it remains possible that a significant contribution of cytokine-induced apoptosis in acetaminophen-sensitized hepatocytes occurs, particularly at lower doses of acetaminophen or in midzonal cells not overwhelmed by the production of NAPQI or that the mode of cell death involves an overlap or caspase-independent apoptosis. Although oxidative stress appears to be an important contributor to acetaminophen toxicity and necrosis is the major mode of cell death, recent evidence has challenged the view that the necrotic demise is simply a consequence of the nonspecific direct effects of overwhelming oxidative stress leading to “accidental” necrosis. It has been recognized in a number of contexts that necrosis can result from the engagement of pro-death programs employed by the apoptotic pathway with an abortive attempt to induce apoptosis leading to a necrotic demise, referred to as programmed necrosis. In the case of acetaminophen, oxidative stress leads to sustained activation of JNK. A chemical inhibitor of JNK as well as silencing JNK expression is markedly protective against acetaminophen toxicity in cultured hepatocytes and in vivo (53). This protection is seen despite identical GSH depletion and covalent binding. Two forms of JNK are expressed in the liver. Silencing JNK2 rather than JNK1 was partially protective. This was confirmed in knockout mice in which JNK2 deletion offered partial protection (53). The targets of JNK2 in this programmed necrosis remain to be fully explored. Bax translocation to mitochondria was seen early after acetaminophen exposure and was blocked by the JNK inhibitor (53). Also, JNK itself translocated to mitochondria suggesting that proteins in the mitochondria may be important targets. In this model, the mode of cell death is necrosis probably because profound ATP depletion from mitochondrial dysfunction caused by direct and indirect effects of acetaminophen as well as oxidative stress/GSH depletion prevent the participation of executioner caspase activity. Other work has demonstrated that sustaining some critical level of ATP through anaerobic glycolysis will favor apoptosis in response to acetaminophen (21,54). Similarly, antioxidants can cause a switch from necrosis to apoptosis in hepatocytes dying from acetaminophen poisoning (21,22). It is uncertain whether the mode of cell death is important in the ultimate outcome of

Cell Death and Drug Hepatotoxicity

91

acetaminophen toxicity but it is intriguing to speculate that the severity of liver injury and survival may be influenced in vivo as the diversion to apoptosis theoretically could limit collateral damage and inflammation in response to the free release of necrotic cellular content. Another interesting hypothesis is that dying cells release contents which can injure surrounding cells. Data support the view that Dnase1 (52) and calpains (55,56) are released into the surrounding milieu and induce collateral damage. Role of GSH in Susceptibility to FasL and TNFa Fas-mediated apoptotic killing of hepatocytes in vivo was prevented by phorone-induced profound GSH depletion, an effect associated with inhibition of caspase 3 activation (57). Antioxidants could not replace GSH, suggesting a direct effect of GSH on caspase activity, i.e., redox maintenance of protein thiols critical for enzymatic activity. On the other hand, chronic GSH depletion sensitized the liver to Fas-mediated apoptosis (58). This effect correlated with GSH depletion-induced upregulation of p53 (presumably via oxidative stress) and consequent upregulation of Bax. However, it is unclear whether the increase in Bax or other effects of decreased GSH mediate the increased susceptibility. The effect of GSH depletion in sensitizing to the lethal actions of TNFa has received considerable attention. Intracellular signaling following the binding of TNFa to TNFR1 leads to increased reactive oxygen metabolite production in mitochondria (19). It has been proposed that this involves the activation of acidic sphingomyelinase and the release of ceramide and its products, which then act on mitochondria to block electron transport at the complex IIIubiquinone cycle leading to auto-oxidation of O2 from the buildup of electrons (59). Goosens et al. were the first to demonstrate this in a murine fibrosarcoma cell; they showed that depletion of mitochondrial GSH markedly enhanced the TNFa-induced oxidative stress and killing, whereas lesser depletion of mainly cytosol GSH was ineffective (60). A critical point in their studies was that GSH depletion had to be delayed for several hours after administration of TNFa to avoid the suppression of TNFa signaling, hinting at the critical nature of the timing. Fernandez-Checa, Kaplowitz, and coworkers subsequently observed that the selective depletion of mitochondrial GSH by chronic ethanol feeding rendered hepatocytes susceptible to TNFa killing, but the killing was mainly in the form of necrotic cell death (19). Subsequently, it was found that profound depletion of GSH by diethylmaleate sensitized hepatocytes to TNFa-induced apoptotic cell death (21). Thus, it appears that depletion of mitochondrial GSH sensitizes to TNFa-induced oxidative stress and lethal actions but the mode of cell death depends on the condition: chronic ethanol appears to interfere with the apoptotic machinery in some fashion. Conflicting results have been published on the effect of GSH depletion on the hepatotoxicity of LPS (or TNFa)Cgalactosamine. Aside from the issue of acute versus chronic GSH depletion mentioned above, gradual GSH depletion with buthionine sulfoximine pretreatment sensitized the liver (58) and hepatocytes (20), or protected in another study (61), whereas acute depletion sensitized in one study (62) or protected in several studies (58,61,63). However, in our own work, we have observed sensitization of mouse hepatocytes to TNFa-induced apoptosis by acute GSH depletion and a variety of redox modulators (21,22,64). Similar results with GSH depletion have been reported by Fausto and coworkers using a well-differentiated, nontransformed mouse hepatocyte cell line (20). Role of NF-kB Galactosamine, actinomycin D, and a-amanitin induce a transcriptional arrest, which markedly sensitized to TNFa either directly administered, stimulated by exogenous LPS, or endogenously produced (65). Transcriptional arrest in hepatocytes interferes with TNFa-induced NF-kB dependent survival gene expression while leaving unopposed TNFa-induced apoptotic signaling via DISC, caspase 8, and tBid or ceramide leading to effects on mitochondria (Fig. 3). The role of stress kinases is less certain, i.e., pro-versus anti-apoptotic effects. However, NF-kB responsive genes inhibit JNK (66,67).

92

Kaplowitz

TNFα TNF-R1 Stress kinases Caspase 8

NF-KB Actinomycin D Galactosamine α–Amanitin Oxidative stress GSH depletion

tBid Ceramide Mitochondria Survival Apoptosis (necrosis)

FIGURE 3 Opposing pathways of death and survival induced by TNFa. Abbreviation: TNFa, tumor necrosis factor a.

Since GSH depletion also sensitizes to TNFa at least under some experimental conditions, it is of interest to understand the effect of GSH depletion on NF-kB activation and transactivation. In Molt-4 cells, GSH depletion inhibited NF-kB activation and GSSG inhibited DNA binding (68). GSH depletion has also been shown to inhibit NF-kB-dependent transcription in response to oxidative stress in Jurkat cells (69). In contrast, GSH depletion stimulated JNK activation, which then phosphorylates c-jun leading to increased AP-1 transactivation of antioxidative stress genes, including GSH-S-transferases (GST) and g-glutamylcysteine synthetase. Thiol-disulfide control of the stress kinase pathway has been shown to be exerted at the level of apoptosis signal-inducing kinase (a kinase upstream of JNK and p38) by redoxresponsive thioredoxin (70) and at the level of JNK by GST-Pi monomer (71). Thus, at present the mechanism by which GSH depletion sensitizes to TNFa-induced apoptosis/necrosis is not entirely clarified and may include decreased defense against mitochondrial oxidative stress, inhibition of NF-kB activation or transactivation, or increased activation of stress kinases. NF-kB activation limits oxidative stress by promoting upregulation of several antioxidant genes including mitochondrial MnSOD and ferritin-heavy chain. Thus inhibition of NF-kB may promote oxidative stress which contributes to TNF killing by various mechanisms including sustained JNK activation and direct effects on mitochondria. In view of the importance of NF-kB in survival it is of considerable interest that oxidative stress inhibits this pathway. Indeed, in the context of toxin-induced liver damage as well as most forms of liver disease this mechanism could be far more relevant than the specific maneuvers designed to inhibit transcription (galactosamine, actinomycin D) and translation (cycloheximide). Of considerable interest is the fact that the level of oxidative stress or redox modulation (e.g., extent of GSH depletion) determines the step affected. Marked GSH depletion, exogenous H2O2 or intracellular oxidative stress (antimycin A, BCNU), sensitize hepatocytes to TNF induced apoptosis by inhibiting IKK activation (64). The precise details of what aspect of this activation process is impaired is currently not known but preliminary evidence suggests that the activation or turnover of RIP in the TNFR1/TRADD/TRAF2 signaling complex may be involved (Lou H and Kaplowitz N unpublished observations). TRAF2 normally ubiquinates RIP which leads to binding of IKK. Interestingly, modest GSH depletion which does affect mitochondrial GSH pool or promote ROS release from mitochondria inhibits the transactivation of NF-kB without affecting IKK, IkBa phosphorylation and degradation or NF-kB translocation to the nucleus; preliminary evidence suggests that DNA binding of NF-kB is not altered under these conditions, although the precise mechanism of inhibition of transcriptional activity is not clear (Lou H and Kaplowitz N, unpublished observations). CONCLUSIONS Hepatotoxicity of drugs and chemicals involves lethal effects on hepatocytes or other cell types in the liver. Hepatotoxicants may illicit an immune response leading to apoptosis or affect liver cells in one of two ways: direct killing or sensitization to cytokines. The mode of cell death in

Cell Death and Drug Hepatotoxicity

93

these circumstances may involve apoptosis, necrosis, or a mixture of the two. In the case of acetaminophen, downstream of covalent binding and GSH depletion, a type of programmed necrosis is involved with a critical role of JNK. REFERENCES 1. Kaplowitz N. Cell death at the millennium: implications for liver diseases. Clin Liver Dis 2000; 4:1–23. 2. Kaplowitz N. Mechanisms of liver cell injury. J Hepatol 2000; 32:39–47. 3. DeLeve L, Wang X, Kuhlenkamp J, Kaplowitz N. Toxicity of azathioprine and monocrotaline in murine sinusoidal endothelial cells and hepatocytes: the role of glutathione and relevance to hepatic venoocclusivc disease. Hepatology 1996; 23:589–99. 4. Odin JA, Huebert RC, Casciola-Rosen L, LaRusso N, Rosen N. Bcl-2-dependent oxidation of pyruvate dehydrogenase-E2, a primary biliary cirrhosis autoantigen, during apoptosis. J Clin Invest 2001; 108:223–32. 5. Thornberry NA, Lazebnik Y. Caspase: enemies within. Science 1998; 281:1312–6. 6. Zhao Y, Li S, Childs E, et al. Activation of pro-death Bcl-2 family proteins and mitochondria apoptosis pathway in TNFa-induced liver injury. J Biol Chem 2001; 276:27432–40. 7. Slee EA, Adrain C, Martin SJ. Executioner caspase-3, -6, and -7 perform distinct, nonredundant roles during the demolition phase of apoptosis. J Biol Chem 2001; 276:7320–6. 8. Wei MC, Zong WX, Chang E, et al. Pro-apoptotic BAX and BAK: a requisite gateway to mitochondrial dysfunction and death. Science 2001; 292:727–30. 9. Goyal L. Cell death inhibition: keeping caspases in check. Cell 2001; 104:805–8. 10. Chauhan D, Hideshima T, Rosen S, Reed JC, Kharbanda S, Anderson KC. Apaf-1/cytochrome c-independent and Smac-dependent induction of apoptosis in multiple myeloma cells. J Biol Chem 2001; 276:24453–6. 11. Zamzani N, Marzo I, Susin S, et al. Thiol crosslinking agent diamide overcomes the apoptosisinhibitory effect of Bcl-2 by enforcing mitochondrial permeability transition. Oncogene 1998; 16:1055–63. 12. Bouchier-Hayes L, Lartigue L, Newmeyer D. Mitochondria: pharmacological manipulation of cell death. J Clin Invest 2005; 115:2640–7. 13. Letai A. Pharmacological manipulation of Bcl-2 family members to control cell death. J Clin Invest 2005; 115:2644–8. 14. Aggarwal B. Apoptosis and nuclear factor-kB: a tale of association and dissociation. Biochem Pharmacol 2000; 60:1033–9. 15. Kim Y-M, Talanian RV, Billiar TR. Nitric oxide inhibits apoptosis by preventing increases in caspase-3like activity via two distinct mechanisms. J Biol Chem 1997; 272:31138–48. 16. Kyriakis J, Banerjee P, Nilolakaki E, et al. The stress-activated protein kinase subfamily of c-Jun kinases. Nature 1994; 369:156–60. 17. Masser DD, Antoine C, Bourget L, et al. Role of the human heat shock protein hsp 70 in protection against stress-induced apoptosis. Mol Cell Biol 1997; 17:5317–27. 18. Feng G, Kaplowitz N. Colchicine protects mice from the lethal effect of agonistic anti-Fas antibody. J Clin Invest 2000; 105:329–39. 19. Colell A, Garcia-Ruiz C, Miranda M, et al. Selective depletion of mitochondrial GSH by ethanol sensitizes to TNFa. Gastroenterology 1998; 115:1541–51. 20. Pierce RH, Campbell JS, Stephenson AB, et al. Disruption of redox homeostasis in tumor necrosis factor-induced apoptosis in a murine hepatocyte cell line. Am J Pathol 2000; 157:221–36. 21. Nagai H, Matusmaru K, Feng G, Kaplowitz N. GSH depletion causes oxidative stress-dependent necrosis and sensitization to TNFa-induced, oxidative stress-independent apoptosis in cultured mouse hepatocytes. Hepatology 2002; 36:55–64. 22. Matsumaru K, Ji C, Kaplowitz N. Mechanisms for sensitization to TNF-induced apoptosis by acute glutathione depletion in murine hepatocytes. Hepatology 2003; 37:1425–34. 23. Leist M, Single A, Castoldi A, et al. Intracellular adenosine triphosphate (ATP) concentration, a switch in the decision between apoptosis and necrosis. J Exp Med 1997; 185:1481–6. 24. De`as O, Dumont C, MacFarlane M, et al. Caspase-independent cell death induced by anti-CD2 or staurosporine in activated human peripheral T lymphocytes. J Immunol 1998; 161:3375–83. 25. Forrnigli L, Papucci L, Tani A, et al. Aponecrosis: morphological and biochemical exploration of a syncretic process of cell death sharing apoptosis and necrosis. J Cell Physiol 2000; 182:41–9. 26. Sperandio S, deBelle I, Bredesen DE. An alternative, nonapoptotic form of programmed cell death. Proc Natl Acad Sci USA 2000; 97:14376–81. 27. Ruefli AA, Smyth MJ, Johnstone RW. HMBA induces activation of a caspase-independent cell death pathway to overcome P-glycoprotein-mediated multidrug resistance. Blood 2000; 95:2378–85. 28. Hamatake M, Iguchi K, Hirano K, Ishida R. Zinc induces mixed types of cell death, necrosis, and apoptosis, in Molt-4-cells. J Biochem 2000; 128:933–9.

94

Kaplowitz

29. Pinkoski MJ, Waterhouse NJ, Heibein JA, et al. Granzyme B-mediated apoptosis proceeds predominantly through a Bcl-2-inhibitable mitochondrial pathway. J Biol Chem 2001; 276:12060–7. 30. Czaja M, Xu J, Alt E. Prevention of carbon tetrachloride-induced rat liver injury by soluble tumor necrosis factor receptor. Gastroenterology 1995; 108:1849–54. 31. Morio LA, Chiu H, Sprowles KA, et al. Distinct roles of tumor necrosis-a and nitric oxide in acute liver injury induced by carbon tetrachloride in mice. Toxicol Appl Pharmacol 2001; 172:44–51. 32. Laskin DL, Gardner CR, Price VF, Jollow DJ. Modulation of macrophage functioning abrogates the acute hepatotoxicity of acetaminophen. Hepatology 1995; 21:1045–50. 33. Boess F, Bopst M, Althaus R, et al. Acetaminophen hepatotoxicity in tumor necrosis factor/lymphotoxin-a gene knockout mice. Hepatology 1998; 27:1021–9. 34. Hogaboam CM, Bone-Lasron CL, Steinhauser ML, et al. Exaggerated hepatic injury due to acetaminophen challenge in mice lacking c–c chemokine receptor 2. Am J Pathol 2000; 156:1245–52. 35. Blazka ME, Wilmer JL, Holladay SD, et al. Role of proinflammatory cytokines in acetaminophen hepatotoxicity. Toxicol Appl Pharmacol 1995; 133:43–52. 36. Blazka ME, Elwell MR, Holladay SD, et al. Histopathology of acetaminophen-induced liver changes. Role of interleukin lot and tumor necrosis factor a. Toxicol Pathol 1996; 24:181–9. 37. Bourdi M, Masubuchi Y, Reilly TP, et al. Protection against acetaminophen-induced liver injury and lethality by interleukin-10: role of inducible nitric oxide synthase. Hepatology 2002; 35:289–98. 38. Liu Z-X, Govindarajan S, Kaplowitz N. Innate immune system plays a critical role in determining the progression and severity of acetaminophen hepatotoxicity. Gastroenterology 2004; 127:1760–74. 39. Sneed RA, Buchweitz JP, Jean PA, Ganey PE. Pentoxifylline attenuates bacterial lipopolysecharideinduced enhancement of allyl alcohol hepatotoxicity. Toxicol Sci 2000; 56:203–10. 40. Shi J, Aisaki K, Ikawa Y, Wake K. Evidence of hepatocyte apoptosis in rat liver after the administration of carbon tetrachloride. Am J Pathol 1998; 153:515–25. 41. Adamson GM, Harman AW. Oxidative stress in cultured hepatocytes exposed to acetaminophen. Biochem Pharmacol 1993; 145:2289–94. 42. Arnaiz SL, Lesuy S, Cutrin JC, Boveris A. Oxidative stress by acute acetaminophen administration in mouse liver. Free Radic Biol Med 1995; 19:303–10. 43. Nakae D, Yamamoto K, Yoshiji H, et al. Liposome-encapsulated superoxide dismutase prevents liver necrosis induced by acetaminophen. Am J Pathol 1990; 136:787–95. 44. Ferret P-J, Hammoud R, Tulliez M, et al. Detoxification of reactive oxygen species by a nonpeptidyl mimic of superoxide dismutase cures acetaminophen-induced acute liver failure in the mouse. Hepatology 2001; 33:1173–80. 45. Mirochnitchenko O, Weisbrot-Lefkowitz M, Reuhl K, et al. Acetaminophen toxicity: opposite effects of two forms of glutathione peroxidase. J Biol Chem 1999; 274:10349–55. 46. Dixon MF, Dixon B, Aparicio SR, Loney DP. Experimental paracetamol-induced hepatic necrosis. A light and electron microscope, and histochemical study. J Pathol 1975; 116:17–29. 47. Ray SD, Mumaw VR, Raje RR, Fariss MW. Protector of acetaminophen-induced hepatocellu-lar apoptosis and necrosis by cholesteryl hemisuccinate pretreatment. J Pharmacol Exp Ther 1996; 279:1470–83. 48. Lawson JA, Fisher MA, Simmons CA, et al. Inhibition of a Fas receptor (CD95)-induced hepatic caspase activation and apoptosis by acetaminophen in mice. Toxicol Appl Pharmacol 1999; 156:179–86. 49. Ray SD, Sorge CL, Raucy JL, Corcoran GB. Early loss of large genomic DNA in vivo with accumulation of Ca2C in the nucleus during acetaminophen-induced liver injury. Toxicol Appl Pharmacol 1990; 106:346–51. 50. Shen W, Kamendulis LM, Ray SD, Corcoran GB. Acetaminophen-induced cytotoxicity in cultured mouse hepatocytes: correlation of nuclear Ca2C accumulation and early DNA fragmentation with cell death. Toxicol Appl Pharmacol 1991; 111:242–54. 51. Ray SD, Kamendulis LM, Gurule MW, et al. Ca2C antagonists inhibit DNA fragmentation and toxic cell death induced by acetaminophen. FASEB J 1993; 7:453–63. 52. Napirei M, Basnakian A, Apostolov E, Mannerz H. Deoxyribonuclease 1 aggravates acetaminopheninduced liver necrosis in male CD-1 mice. Hepatology 2006; 43:297–305. 53. Gunawan B, Liu Z-X, Han D, Hanawa N, Gaarde W, Kaplowitz N. c-Jun-N-Terminal kinase plays a major role in murine acetaminophen hepatotoxicity. Gastroent 2006; 131:165–78. 54. Kon K, Kim J, Jaeschke H, Lemasters JJ. Mitochondrial permeability transition in acetaminopheninduced necrosis and apoptosis of cultured mouse hepatocytes. Hepatology 2004; 40(5):1170–9. 55. Limaye P, Apte U, Shankar K, Bucci T, et al. Capain release from dying hepatocytes mediates progression of acute liver injury induced by model hepatotoxicants. Toxicol Appl Pharmacol 2003; 191:211–26. 56. Limaye PB, Bhave VS, Palkar PS, et al. Upregulation of calpastatin in regenerating and developing rat liver: role in resistance against hepatotoxicity. Hepatology 2006; 44(2):379–88. 57. Hentze H, Kunstle G, Volbracht C, et al. CD95-mediated murine hepatic apoptosis requires an intact glutathione status. Hepatology 1999; 30:177–85.

Cell Death and Drug Hepatotoxicity

95

58. Haouzi D, Lekehal M, Tinel M, et al. Prolonged, but not acute, glutathione depletion promotes Fasmediated mitochondrial permeability transition and apoptosis in mice. Hepatology 2001; 33:1181–8. 59. Fernandez-Checa J, Kaplowitz N, Garcia-Ruiz C, et al. GSH transport in mitochondria: defense against TNFa-induced oxidative stress and alcohol-induced defect. Am J Physiol 1997; 273:G7–17. 60. Goosens V, Grooten J, DeVos K, Fiers W. Direct evidence for tumor necrosis factor-induced mitochondrial reactive oxygen intermediates and their involvement in cytotoxicity. Proc Natl Acad Sci USA 1995; 92:8115–9. 61. Jones JJ, Fan J, Nathens AB, et al. Redox manipulation using the thiol-oxidizing agent diethyl-maleate prevents hepatocellular necrosis and apoptosis in a rodent endotoxemia model. Hepatology 1999; 30:714–24. 62. Xu Y, Jones BE, Neufeld DS, Czaja MJ. Glutathione modulates rat and mouse hepatocyte sensitivity to tumor necrosis factor a toxicity. Gastroenterology 1998; 115:1229–37. 63. Hentze H, Gantner F, Kolb SA, Wendel A. Depletion of hepatic glutalhione prevents death receptordependent apoptotic and necrotic liver injury in mice. Am J Pathol 2000; 156:2045–56. 64. Han D, Hanawa N, Saberi B, Kaplowitz N. Hydrogen peroxide and redox modulation sensitize primary mouse hepatocytes to TNF-induced apoptosis. Free Radic Biol Med 2006; 41(4):627–39. 65. Leist M, Gantner F, Naumann H, et al. Tumor necrosis factor-induced apoptosis during the poisoning of mice with hepatotoxins. Gastroenterology 1997; 112:923–34. 66. De Smaele E, Zazzeroni F, Papa S, et al. Induction of gadd45(beta) by NF-(kappa)B down-regulates pro-apoptotic JNK signalling. Nature 2001; 414:308–12. 67. Tang G, Minemoto Y, Dibling B, et al. Inhibition of JNK activation through NF-(kappa)B target genes. Nature 2001; 414:313–7. 68. Mihm S, Galter D, Droge W. Modulation of transcription factor NF-kB activity by intracellular glutathione levels and by variations of the extracellular cysteine supply. FASEB J 1995; 9:246–52. 69. Ginne-Pease ME, Whisler RL. Optimal NF-kB mediated transcriptional responses in Jurkat T cells exposed to oxidative stress are dependent on intracellular glutathione and costimulatory signals. Biochem Biophys Res Commun 1996; 226:695–702. 70. Sitoh M, Nishitoh H, Fuju M, et al. Mammalian thioredoxin is a direct inhibitor of apoptosis signalregulating kinase (ASK)-l. EMBO J 1998; 17(9):2596–606. 71. Adler V, Yin Z, Fuchs S, et al. Regulation of JNK by GSTp. EMBO J 1999; 18:1321–34.

6

Significance of Hepatobiliary Transporters for Drug-Induced Liver Disease Peter J. Meier

University of Basel, Basel, Switzerland

Christiane Pauli-Magnus

University Hospital Basel, Basel, Switzerland

INTRODUCTION Despite rigorous safety requirements in drug development and a better understanding of mechanisms of drug action and toxicity, drug-induced liver injury remains one of the main consequences of drug-related toxicity. Drug-induced liver injury accounts for approximately 2% to 5% of hospitalizations for jaundice, 10% of cases of hepatitis in all adults and more than 40% of hepatitis cases in adults older than 50 (1). According to the consensus conference of the Council for International Organization of Medical Sciences, druginduced liver injury can be classified into hepatocellular, cholestatic and mixed types of liver damage depending on serum biochemistry markers (2). While hepatocellular injury is mainly characterized by the elevation of serum transaminases, cholestatic liver damage is reflected by increased levels of alkaline phosphatase, gamma-glutamyl transpeptidase, and conjugated bilirubin in serum. For most hepatocellular forms of drug-induced liver injury, the underlying pathophysiological mechanism is poorly understood. Only a minority of cases can be related to intrinsic hepatotoxins. Most cases are still attributed to idiosyncratic reactions, where immunoallergic mechanisms resulting from hypersensitivity and aberrant metabolism of the suspected drug are thought to represent the predominant pathophysiological pathways. In addition, it has become evident that expression and function of hepatobiliary transport systems localized at the two polar surface membranes of hepatocytes may play an important role in certain forms of drug-induced liver disease. For instance, basolateral (sinusoidal) uptake transporters play a key role in controlling hepatic drug exposure, thereby determining the amount of potential toxins being taken up into liver cells. On the other hand, function and expression of apical (canalicular) transport systems determine the amount of hepatic drug clearance from hepatocytes into the bile canaliculus. Furthermore, apical transport systems are critical for the efflux of bile salts and other bile components from hepatocytes into the bile canaliculus. Any functional disturbance of these apical efflux systems for bile salts might therefore predispose to the development of cholestatic liver cell damage through intracellular accumulation of toxic bile constituents. Such functional disturbances may arise from a drug-mediated inhibition of apical transport systems involved in bile formation. Furthermore, function and expression of hepatocellular transporters might also be altered by environmental and genetic factors contributing to the development of drug-induced liver disease in susceptible individuals. Furthermore, bile ductular reabsorption of bile salts and drugs and cholehepatic shunting might contribute to changes in bile composition and to hepatic accumulation of drugs and toxins. This chapter summarizes current knowledge about the function of hepatocellular uptake and efflux systems involved in bile formation and discusses environmental and genetic risk factors that might affect the function of these systems.

98

Meier and Pauli-Magnus

PHYSIOLOGY OF HEPATOBILIARY TRANSPORT AND BILE FORMATION Bile formation is maintained by distinct transport systems expressed at the two polar surface domains of liver cells. While basolateral transporters are responsible for hepatic uptake of bile salts from sinusoidal blood, canalicular transporters maintain the excretion of cholephilic compounds into the bile canaliculus. These systems are subject to extensive transcriptional and posttranscriptional regulation, allowing adaptational changes in response to the intracellular accumulation of bile salts [reviewed in (3,4)]. After canalicular secretion, bile composition undergoes further modification in the bile canaliculi, involving reabsorption and secretion processes maintained by apical and basolateral transport systems in cholangiocytes. Figure 1 shows a schematic of hepatocellular and bile ductular transport proteins involved in the uptake and efflux of endogenous and exogenous (xenobiotic) cholephilic compounds. Hepatic Transport Systems Properties, Function, and Regulation of Basolateral (Sinusoidal) Transporters So far, sodium-dependent and -independent transport pathways have been identified to play a key role in hepatic uptake of endogenous and exogenous substances from sinusoidal blood

MRP4

GR

NTCP

MDR1

ABCG5/G8

PXR

MDR3

OATP1A2 HNF1α

BCRP

OATP1B1

BSEP

FIC 1

FXR

MRP2

SHP-1 OATP1B3

MRP3

Hepatocyte Cholehepatic shunt pathway Cholangiocyte

OATP1A2

MRP3 Basolateral

Ostαostβ

ASBT MRP2

Apical

FIGURE 1 Bile salt transporters in human liver and cholangiocytes. Abbreviations: Efflux transporters (black symbols): BSEP, bile salt export pump; MDR, multidrug resistance protein; MRP, multidrug resistance associated protein; ABCG5/G8, ATP-binding cassette G5 and G8 transporters; BCRP, breast cancer resistance protein; Ostalpha/Ostbeta, organic solute transporter alpha/beta; FIC1, familial cholestasis type 1. Uptake transporters (grey symbols): ASBT, apical sodium-dependent bile salt transporter; NTCP, sodium-taurocholate cotransporting polypeptide; OATP, organic anion transporting polypeptide; OCT, organic cation transporter; OAT, organic anion transporter. HNF1a, hepatocyte nuclear factor 1a; SHP-1, small heterodimeric partner 1; PXR, pregnane X receptor; FXR, farnesoid X receptor; GR, glucocorticoid receptor. Further details are given in the text.

Hepatocellular Drug-Induced Liver Injury

99

plasma (Fig. 1). Major substrates of these uptake transporters are summarized in Table 1. The sodium-dependent pathway is represented by the sodium-taurocholate–cotransporting polypeptide (NTCP) [SLC10A1; reviewed in (5)], the substrate specificity of which is essentially limited to conjugated bile salts and certain sulfated steroids. NTCP accounts for more than 80% of conjugated (i.e., taurocholate and glycocholate) but for less than 50% of unconjugated (i.e., cholate) bile salt uptake (5). In response to cholestasis, NTCP is suppressed through farnesoid X receptor (FXR)-mediated induction of the small heterodimeric partner 1 (SHP-1), thereby preventing the hepatocyte from further accumulating toxic bile salts (13,14). In contrast, the sodium-independent pathway is represented by different members of the organic anion–transporting polypeptide (OATP) superfamily of solute transporters (SLO family, former SLC21) [reviewed in (6)]. In the human liver, highest expressions are found for OATP1B1 (SLCO1B1) and its 80% sequence homolog OATP1B3 (SLCO1B3), which are both predominantly, if not exclusively, expressed in the liver. With the exception of OATP2B1 (SLCO2B1), the substrate specificity of which seems to be limited to bromosulfophthalein (BSP) and steroid sulfates, OATP1A2 (SLCO1A2), OATP1B1, and OATP1B3 exhibit overlapping transport activities for conjugated and unconjugated bile salts, BSP, neutral steroids, steroid sulfates and glucuronides, and selected organic cations (6). Furthermore, numerous drugs are the substrates of OATPs, including the antihistamine fexofenadine, opioid peptides, digoxin, the hydroxymethylglutaryl CoA-reductase inhibitor pravastatin, the angiotensin-converting enzyme inhibitor enalapril, or the antimetabolite methotrexate (6). In addition, OATP1B1 and OATP1B3 mediate the uptake of the hepatotoxins phalloidin and microcystin into human liver (15–17), while hepatic uptake of amanitin, the most dangerous natural toxins causing hepatic failure, seems to be exclusively mediated by OATP1B3 (18). Similar to NTCP, the expression of OATP1B1 is downregulated during cholestasis through bile acid-mediated activation of SHP-1, which leads to a repression of hepatocyte nuclear factor 1a, the major transcriptional activator of OATP1B1 (19,20). In contrast, cholestasis leads to an FXR-mediated activation of hepatic OATP1B3 (21), which might constitute an escape mechanism promoting the hepatocellular clearance of xenobiotics during cholestasis. On the posttranscriptional level, sodium-dependent and -independent hepatocellular uptake systems are mainly regulated by cAMP-mediated dephosphorylation processes, which is controlled by phosphoinositide-3-kinase (PI3K)/ protein kinase B (22–24). Furthermore, PDZK1 was demonstrated to be a critical determinant for the proper subcellular localization and function of rat Oatp1a1 (25). In addition, the sodium-independent uptake systems involve the organic anion and organic cation transporter family of solute carriers (SLC22), the major substrates of which are summarized in Table 1. The SLC22 family belongs to a gene family separate from OATPs. OAT2 (SLC22A7) is the only transporter of the organic anion transporter/organic cation transporter (OAT/OCT) family expressed in human liver and is believed to be liver specific [reviewed in (8)]. OCT1 (SCL22A1) is expressed in human liver, as well as in kidneys, small intestine, and colon (8). After disruption of OCT1 in mice, the biliary excretion of some but not all substrates of OCT1 was largely reduced, indicating that impaired OCT1 function may reduce the biliary excretion of specific drugs and affect their efficacy and/or toxicity (26). Furthermore, there is speculation that hepatotoxicity of anionic drugs or xenobiotics may be reduced by comedication with substrates or inhibitors of OAT2 (8). However, the exact role of OAT2 and OCT1 for hepatic uptake of drugs and bile constituents remains to be established. Besides these uptake systems, the basolateral hepatocyte membrane also localizes several ATP-dependent efflux pumps. These transporters belong to the family of multidrug resistance– associated proteins (MRPs ABCC), which are multispecific transporters for different organic anions that are summarized in Table 1 [reviewed in (27)]. Among the MRP family of ATPbinding cassette (ABC) transporters, the five members MRP1 (ABCC1), MRP3 (ABCC3), MRP4 (ABCC4), MRP5 (ABCC5), and MRP6 (ABCC6) have been implicated in the cellular efflux of drug-glutathione, -glucuronide, and -sulfate conjugates (MRP1), the efflux of bile salts (MRP3), the transport of nucleoside analog drugs such as zidovudine, lamivudine, and stavudine (MRP4), and of the cyclic nucleosides cAMP and cGMP, as well as methotrexate and the purine analogs 6-mercaptopurine and 6-thioguanine (MRP4 and MRP5) (11), and glutathione-Sconjugates (MRP6). The transcriptional regulation of basolaterally expressed MRPs is not

Bile salts and conjugates bilirubin glucuronide, Acetaminophene-glucuronide etoposide-glucuronide, conjugated steroids, glucuronide conjugates methothrexate, morphine-3-glucuronide Bile salts, conjugated steroids, cyclic nucleotides Leucovorin, 6-mercaptopurine, methotrexate, (cAMP and cGMP), glutathione, folate, nucleoside analogs, thioguanine prostaglandins E1 and E2, thromboxane, urate Cyclic nucleotides (cAMP and cGMP), Folate Methotrexate, 6-mercaptopurine, nucleoside analogs, thioguanine Glutathione S-conjugates

ABCC3 ABCC4 ABCC5 ABCC6

MRP4

MRP5

MRP6

Pseudoxantoma elasticum

Inherited disease

Details summarized in this table refer to findings in different species (for details see the text). For simplification, only the human transporter symbol is listed in the table. Abbreviations: NTCP, sodium-taurocholate cotransporting polypeptide; OAT, organic anion transporter; OCT, organic cation transporter; MRP, multidrug resistance associated protein; BSP, bromosulfophthalein; TEA, tetraethylammonium; NSAID, nonsteroidal anti-inflammatory drug; DPDPE, [d-penicillamine(2,5)]enkephalin.

ABCC1

Multidrug resistance associated MRP1 proteins (ABCC family) (11,12) MRP3

Antibiotics, antiviral drugs, diuretics, methotrexate, NSAID, ochratoxin A, uricosurics Vinca alkaloids, anthracyclines

Cyclic nucleotides, prostaglandins E2 and F2a Leukotriene C4

OAT2

OATP2B1 SLCO2B1 Steroid conjugates OCT1 SLC22A1 Prostaglandins E2 and F2a serotonin

Ajmalinium, BSP, DPDPE and deltorphin II, fexofenadine, microcystin, N-methyl-quinine and quinidine, ouabain, opioid receptor agonists rocuronium BSP Acyclovir, desipramine, ganciclovir, metformin, Nmethylquinine, 1-methyl-4-phenylpyridinium, TEA

Amanitin, BSP, digoxin, methothrexate, microcystin, ouabain, phalloidin, rifampicin

Benzylpenicillin, BSP, irinotecan, methotrexate, microcystin, phalloidin, pravastatin, rifampicin

Major drug and toxin substrates

SLC22A7

Organic cation and anion transporters (SLC22 family) (8 10)

SLC10A1

Major endogenous substrates

Conjugated and unconjugated bile salts, steroid conjugates OATP1B1 SLCO1B1 Conjugated and unconjugated bile salts, conjugated and cyclic peptides, eicosanoids, steroid conjugates, thyroid hormones, unconjugated bilirubin OATP1B3 SLCO1B3 Conjugated and unconjugated bile salts, leukotriene C4, linear and cyclic peptides, monoglucuronosyl bilirubin, steroid conjugates, thyroid hormones OATP1A2 SLCO1A2 Conjugated and unconjugated bile salts, prostaglandin E2, steroid conjugates, thyroid hormones

NTCP

Sodium bile salt cotransporters (SLC10 family) (5) Organic anion transporting polypeptides (SLCO family) (6,7)

Gene

Characteristics of Basolateral (Sinusoidal) Hepatic Transporters

Member

TABLE 1

Family

100 Meier and Pauli-Magnus

Hepatocellular Drug-Induced Liver Injury

101

fully elucidated. Studies in mice support the notion that Mrp3 and Mrp4 are induced through a pregnane X receptor (PXR)-mediated pathway (28). Properties, Function, and Regulation of Apical (Canalicular) Transporters The secretion of bile salts and xenobiotics across the canalicular membrane of hepatocytes is mediated by various ABC transporters (Fig. 1). With the exception of familial intrahepatic cholestasis type 1 (FIC1) (ATP8B1), which is thought to play a role in the regulation of the enterohepatic bile acid pool and in the elimination of hydrophobic substances from the enterohepatic circulation (29), canalicular transporters involved in bile formation and hepatic drug clearance belong to different members of the superfamily of ABC transporters. These include members of the family of multidrug resistance (MDR) P-glycoproteins, such as MDR1 (ABCB1), MDR3 (ABCB4), and the bile salt export pump (BSEP) (ABCB11). In addition, the canalicular membrane localizes MRP2 (ABCC2), the ABC half transporters breast cancer resistance protein (BCRP, ABCG2), ABCG5 and ABCG8 (ABCG5 and ABCG8). Major substrates of these efflux systems expressed at the canalicular membrane of hepatocytes are summarized in Table 2. Within the family of MRPs, BSEP and MDR3 are the two highly conserved members, which are involved in the secretion of cholephilic compounds from the liver cell into the bile canaliculus [reviewed in (30,31)]. BSEP constitutes the predominant bile salt efflux system of hepatocytes and mediates the cellular excretion of numerous conjugated bile salts, such as taurine- or glycine-conjugated cholate, chenodeoxycholate, and deoxycholate (50,51). In addition, BSEP has recently been shown to actively transport the cholesterol-lowering drug pravastatin (32). MDR3 was shown to function as an ATP-dependent phospholipid flippase, translocating phosphatidylcholine from the inner to the outer leaflet of the canalicular membrane [reviewed in (33)]. Canalicular phospholipids are then solubilized by canalicular bile salts to form mixed micelles, thereby protecting cholangiocytes from the detergent properties of bile salts. In addition to these processes, MRP2, the only canalicular member of the MRP family, mediates the canalicular transport of glucuronidated and sulfated bile salts. MRP2 is the main driving force for bile salt independent bile flow through canalicular excretion of reduced glutathione. Furthermore, MRP2 transports a wide spectrum of organic anions, including bilirubin diglucuronide, glutathione conjugates, leukotriene C4, and divalent bile salt conjugates as well as drug substrates, such as cancer chemotherapeutic agents, uricosurics, and antibiotics [reviewed in (35)]. Transcriptional regulation of BSEP, MDR3, and MRP2 is mediated by the FXR (52–55). FXR-mediated activation of BSEP and MDR3 leads to increased bile salt efflux and the formation of mixed micelles in the biliary tree during cholestatic episodes, thereby preventing toxic effects of bile salts on hepatocytes and cholangiocytes. Furthermore, FXR-mediated induction of MRP2 expression in human hepatocytes might constitute another compensatory mechanism during cholestasis (55). On the posttranscriptional level, targeting of Bsep, Mdr2 (the rat homolog of human MDR3), and Mrp2 to the canalicular membrane (56–58) is mediated by PI3K and protein kinase C isoforms. The exact contribution of MDR1 to hepatic bile formation remains to be established, but it is thought to contribute to the canalicular excretion of drugs and other xenobiotics into the bile. Its broad substrate specificity and physiological expression in various tissues with excretory and protective functions make MDR1 one of the major determinants of drug disposition and toxicity. Substrates are many neutral and positively charged organic compounds, including various chemotherapeutic and immunosuppressant agents, antiarrhythmic drugs, HIV protease inhibitors, and antifungals (Table 2) (59,60). In contrast to BSEP, MDR3, and MRP2, MDR1 is upregulated via the PXR on the transcriptional level, which in addition to endogenous ligands is activated by different xenobiotics such as rifampicin or the herbal remedy St John’s Wort (61–63). This pathway is thought to be part of a general cellular detoxification mechanism, as MDR1 is the key transporter protein involved in the cellular efflux of numerous drugs and xenobiotics. The ABC half transporter BCRP (ABCG2) shows the highest expression levels in mammary epithelium and placenta, where it plays an important role in the efflux of a variety of xenobiotics (37–44). Recently, BCRP has been shown to in vitro transport sulfated bile salt conjugates, such as taurolithocholate sulfate (45). In addition to conferring a MDR phenotype against a variety of chemotherapeutic agents, BCRP is also involved in the biliary

ABCG2

ABCG5/G8

BCRP

ABCG5/G8

ABCB4 ABCB11 ABCC2

MDR3 BSEP MRP2

Cholesterol

Phosphatidylcholine Conjugated bile salts Bile salt conjugates, bilirubin diglucuronide, estrogen conjugates, estrone-3-sulfate, glutathione conjugates, leukotrienes C4, D4, and E4 Bile salt conjugates, glucuronide conjugates, sulfate conjugates

?

Major endogenous substrates

Aflatoxin, albendazole, 9-aminocamptothecin, antifolates, bisantrene, cimetidine, daunorubicine, doxirubicine, fenbendazole, flavopiridol, fluoroquinolone antibiotics, mitoxantrone, nitrofurantoin, pantoprazole, pitavastatin, sulfasalazin, topotecan Plant sterols

Pravastatin Antibiotics, BSP, methotrexate, ochratoxin A, pravastatin, uricosurics

? Antibiotics, anticancer drugs, antidiarrheal agents, antiemetics, beta-blockers, calcium channel blockers, cardiac glycosides, debrisoquine, H1 and H2 antihistamines, HIV protease inhibitors, immunosuppressants, lipid-lowering drugs, losartane, morphine, phenytoin, quinidine, rifampicin

Major drug and toxin substrates

Sitosterolemia

PFIC3 PFIC2 Dubin Johnson syndrome

PFIC1

Inherited disease

Details summarized in this table refer to findings in different species (for details see the text). For simplification, only the human transporter symbol is listed in the table. Abbreviations: ABC, ATP-binding cassette; FIC1, familial cholestasis type 1; MDR, multidrug resistance protein; BSEP, bile salt export pump; MRP, multidrug resistance associated protein; BCRP, breast cancer resistance protein; ABCG5/G8, ATP-binding cassette G5 and G8 transporters; PFIC, progressive familial intrahepatic cholestasis; BSP, bromosulfophthalein.

ABC half transporters (37 49)

Multidrug resistance associated proteins (ABCC family) (35,36)

ATP8B1 ABCB1

Gene

FIC1 MDR1

Member

Characteristics of Apical (Canalicular) Hepatic Transporters

P-type ATPase Multidrug resistance proteins (ABCB family) (30 34)

Family

TABLE 2

102 Meier and Pauli-Magnus

Hepatocellular Drug-Induced Liver Injury

103

excretion of drugs, such as fluoroquinolone antibiotics, cimetidine, or pitavastatin (39,43,46). It might therefore be speculated that BCRP contributes to the hepatocellular excretion of bile salts and xenobiotics. Furthermore, the heterodimeric transporters ABCG5 and ABCG8 (ABCG5 and ABCG8) have been identified as the apical transport systems involved in the hepatobiliary excretion of plant sterols and cholesterol [reviewed in (47,48)]. Overexpression of ABCG5/ ABCG8 in transgenic mice led to an increase in biliary cholesterol secretion and a reduced intestinal absorption of dietary cholesterol, providing strong evidence for ABCG5/ABCG8 being involved in hepatocellular secretion and intestinal efflux of cholesterol (47). However, the possible role of these ABC half transporters for hepatic drug clearance and the development of drug-induced cholestasis remain to be determined. Transcriptional regulation of BCRP most likely involves the aryl hydrocarbon receptor, which leads to an increase in BCRP expression (64). Furthermore, BCRP is induced through hypoxia while the role of estradiol on BCRP expression is controversial (65,66). In contrast, ABCG5 and ABCG8 are direct targets of the oxysterol-dependent liver X receptor alpha and -beta (67). Bile Ductular Transport Systems Properties and Function of Basolateral and Apical Cholangiocyte Transporters Uptake of bile salts from canalicular bile into cholangiocytes is mediated by the apical sodiumdependent bile salt transporter (ASBT; SCL10A2) (Fig. 1). ASBT belongs to the superfamily of solute carriers and is identical with the gene product expressed in the terminal ileum of the small intestine (5). Furthermore, the uptake of bile salts involves the OATP1A2, which belongs to the OATP superfamily of sodium-independent solute transporters. After their uptake into cholangiocytes, bile salts are effluxed at the basolateral cholangiocyte membrane into the peribiliary plexus via an anion exchange mechanism (68). From here, bile salts reach the portal circulation and undergo the cholehepatic shunt pathway. MRP3, a basolaterally expressed member of the family of MRPs, contributes to the efflux of bile salts from cholangiocytes (69). Moreover, MRP2 was recently localized in gallbladder-derived biliary epithelial cells, where it might contribute to taurocholate homeostasis (69). In addition, a splicing variant of rat ASBT could be localized to the basolateral membrane of cholangiocytes, where it is proposed to function as a bile salt efflux protein. However, the contribution of this truncated protein to bile salt efflux in human cholangiocytes has not been established (70). Furthermore, the heterodimeric organic solute transporter (OST) (OSTalpha–OSTbeta) was recently found to be expressed in the basolateral membrane of cholangiocytes, where it is thought to contribute to bile acid and sterol transport into the peribiliary plexus (71). ROLE OF HEPATOCELLULAR TRANSPORT SYSTEMS IN DRUG-INDUCED LIVER INJURY Animal models and in vitro studies of drug-induced liver injury reveal different carrier-related mechanisms that might be relevant to the development of hepatic damage (Fig. 2). The amount of drug and its metabolites reaching the hepatocyte and hence the canalicular membrane is determined by the function and expression of basolateral transport processes. Accordingly, increased hepatic uptake of xenobiotics might be associated with hepatic damage. On the apical hepatocyte membrane, decreased canalicular transporter function could lead to intracellular accumulation of bile constituents and xenobiotics, with consecutive toxic liver cell damage (72). Hepatic injury involving a dysfunction of canalicular transporter proteins is predominantly cholestatic, as it is associated with impaired bile secretion. Furthermore, there is speculation whether bile ductular reabsorption of bile salts and drugs and cholehepatic shunting could result in changes in bile composition and promote hepatic drug accumulation. It is, however, likely that transporter-mediated mechanisms of drug-induced hepatic damage coexist with other pathophysiological pathways of liver injury such as immunoallergic reaction. For instance, in ductular forms of drug-induced liver injury such as the vanishing bile duct syndrome, a drug or its metabolite is thought to trigger an immune response against biliary epithelium (73). It might also be speculated that immunoallergic reactions might indirectly alter bile formation through impairment of hepatocellular transporter function. In

104

Meier and Pauli-Magnus

Increased hepatic uptake

Decreased hepatic uptake

Basolateral

Hepatocellular transport Increased hepatic secretion

Increased biliary reabsorption

Decreased hepatic secretion

Canalicular

Ductular transport

FIGURE 2 Transporter-related mechanisms in the development of liver injury. Potentially harmful (shaded boxes) or protective (white boxes) transporter-related mechanisms. Further details are given in the text.

the latter case, the functional state of hepatocellular transporter proteins might be essential to determine the susceptibility to such superimposed damage. Environmental Risk Factors Drug–Transporter Interactions The possible impact of a drug-mediated inhibition of basolateral transport processes is only beginning to emerge. A study investigating the effect of rifampin on basolateral OATP function could demonstrate a significant inhibition of transport activity only for OATP1B3, while OATP1A2, OATP1B1, and OATP1B2 functions was unaffected (74). Furthermore, recent data indicate that NTCP-mediated uptake of taurocholate is inhibited by rifampicin, rifamycin SV, glibenclamide, and cyclosporine A (75), as well as by the protease inhibitors ritonavir, saquinavir, and efavirenz (76). While the significance of these observations for the development of drug-induced liver injury is not clear, there is speculation that inhibition of OATP-mediated uptake of hepatotoxins such as amanitin, microcystin, and phalloidin could prevent toxic liver injury. For instance, OATP1B3-mediated uptake of amanitin was inhibited in transfected MDCKII cells by OATP1B3 inhibitors such as cyclosporine A and rifampicin, as well as by some antidotes used in the past for the treatment of human amatoxin poisoning (18). Inhibition of hepatocellular uptake transport might therefore constitute a promising approach for the treatment of human poisoning with certain hepatotoxins. On the canalicular side, various drugs are thought to cause cholestasis through inhibition of BSEP function. In vitro studies with rat Bsep revealed that cyclosporine A, rifampin, bosentan, troglitazone, and glibenclamide inhibit ATP-dependent taurocholate transport (77–79). While most of these drugs directly cis-inhibit BSEP function in a competitive manner, estradiol 17b-glucuronide and progesterone metabolites indirectly trans-inhibit BSEP after their secretion into the bile canaliculus by Mrp2 (Fig. 3) (77). Recently, these data could be confirmed for human BSEP, where vectorial transport of fluorescent bile acids was significantly inhibited by rifampicin, rifamycin SV, glibenclamide, and cyclosporine A in LLCPK1 cells stably expressing NTCP and BSEP at the basolateral and apical sides, respectively (75). Furthermore, in vitro testing of over 40 structurally different drugs and natural compounds in BSEP-expressing Sf9 cells revealed that in addition to the above-mentioned compounds, the calcium channel inhibitors nifedipine, fendiline, and nicardipine strongly inhibited BSEPmediated taurocholate transport, while bepridil and the anticancer drugs vinblastine and actinomycin were moderate inhibitors (80). In addition, ritonavir, saquinavir, and efavirenz

105

Hepatocellular Drug-Induced Liver Injury

Cis-inhibitors Cyclosporine A Rifampin Bosentan Troglitazone Glibenclamide Nifedipine Fendiline Nicardipine Bepridil Vinblastin Actinomycin Ritonavir Saquinavir Efavirenz

BSEP Bile acids

Cis-inhibition

Trans-Inhibitors Estradiol 17β− glucuronide progesterone metabolites

Trans-inhibition

Drug MRP2

FIGURE 3 Mechanisms of BSEP inhibition. Abbreviations: BSEP, bile salt export pump; MRP2, multidrug resistance associated protein 2. Further details are given in the text.

were recently identified to inhibit biliary excretion of taurocholate in BSEP-transfected SF9 cells and cultured human rat hepatocytes, offering a possible explanation for the hepatotoxic potential of antiretroviral therapy (76). On the other hand, verapamil, cyclosporine, and vinblastine showed in vitro transport by MDR3, which could potentially lead to concentration-dependent inhibition of phospholipid flippase activity (81). An additional mechanism of cholestatic cell damage was recently proposed by Fouassier et al., who investigated the effect of bosentan on Mrp2-mediated canalicular bile formation (82). It could be shown that bosentan stimulates and significantly increases Mrp2-dependent bilirubin excretion and bile salt-independent bile flow, while biliary lipid secretion was profoundly inhibited and uncoupled from bile salt secretion. Inhibition of biliary lipid secretion was not seen in Mrp2 transport-deficient rats, which suggests that translocation of organic anions across the canalicular membrane is a prerequisite for the occurrence of the uncoupling effect. Thus, Mrp2-induced choleresis might dilute bile salts in the bile canaliculi below the concentration required for solubilization of phosphatidylcholine and cholesterol. Consequently, decreased biliary phospholipid secretion does not necessarily mean a defect in the canalicular phospholipid flippase MDR3 but could also be explained by a physicochemical disequilibrium in bile composition. In contrast, data supporting a contribution of MDR1 to hepatic drug clearance are not available so far. However, such a contribution could be suspected based upon the important role of MDR1 as a source of many clinically relevant drug interactions in other tissues with excretory function, such as small intestinal enterocytes or proximal tubulus cells of kidneys (83). As MDR1 transports numerous drugs with hepatotoxic potential such as amiodarone, HIV protease inhibitors, or chemotherapeutics (60), it can be speculated that altered hepatic MDR1 function at the canalicular membrane might contribute to decreased biliary elimination of xenobiotics, thereby promoting hepatotoxicity. In addition to these hepatic sources of drug toxicity, it is conceivable that drug reabsorption by cholangiocytes and cholehepatic shunting contributes to the development of hepatic damage. For instance, the nonsteroidal anti-inflammatory drug sulindac was shown to induce hypercholeresis in rats, a phenomenon that is explained by cholehepatic shunting (84). The resulting hepatic accumulation of sulindac and the inhibition of canalicular bile salt transport might contribute to the potential of this drug to produce cholestasis (85,86). Age and Gender It is well established that age over 50 years and female gender are associated with an increased risk to develop drug-induced hepatic damage, and it was therefore speculated whether these factors affect the expression of hepatocellular transporters (Table 3). On the basolateral side, Simon and coworkers could demonstrate decreased sodium-dependent bile salt uptake in female hepatocytes, which was partly due to a selective decrease in Ntcp expression (87). Furthermore, a recent study revealed gender-specific expression of hepatic Oatp1a1 and

?

? ? ? ? Anticancer drugs, calcium channel blockers, HIV protease inhibitors, immunosuppressants quinidine Verapamil, cyclosporine, vinblastine

OATP1B1 OATP1B3 OATP1A2 OATP2B1 OCT1 OAT2 MRP1

MRP3

MRP4 MRP5 MRP6 FIC1 MDR1

Fumitremorgin, imanitib, novobiocin ?

MRP2

BCRP ABCG5/G8

Liver disease

4 in cholestatic alcoholic hepatitis, Y in PBC, [ in CCL4 toxicity, [ in acetaminphen toxicity, 4in PBC, HCV, and submassive necrosis 4 in CCL4 and acetaminophen toxicity ?

4 in cholestatic alcoholic hepatitis, 4in PBC, HCV, and submassive necrosis Y in cholestatic alcoholic hepatitis, 4in PBC, HCV, and submassive necrosis

? ? [ in CCL4 toxicity, [ in PBC, HCV, and submassive necrosis 4 in cholestatic alcoholic hepatitis, Y in CCL4 toxicity, [ in acetaminphen toxicity, [ in PBC, HCV, and submassive necrosis [ in CCL4 toxicity, [ in acetaminphen toxicity ? ? ? 4 in cholestatic alcoholic hepatitis, [ in PBC, HCV, and submassive necrosis

Y in cholestatic hepatitis, Y in CCL4 toxicity, 4 in acetaminophen toxicity Y in PBC

Gender

? ?

MaleOfemale ?

4

4

4

4

4

4

FemaleOmale ? ? ? 4

FemaleOmale

? ? ? ? ? 4

? ? ? ? ? FemaleZmale

Female!male

? ? ? ? ? ? ?

?

Age

References

(96 99)

(88,90,94)

(92,94)

(92,94)

(92,94,95)

(88,93)

(88,92,93)

(8) (91) (88,92)

(90) (18,74)

(75,76,87 89)

Details summarized in this table refer to findings in different species (for details see the text). For simplification, only the human transporter symbol is listed in the table. Abbreviations: NTCP, sodium-taurocholate cotransporting polypeptide; OATP, organic anion transporting polypeptide; OCT, organic cation transporter; MRP, multidrug resistance associated protein; FIC1, familial cholestasis type 1; MDR, multidrug resistance protein; BSEP, bile salt export pump; BCRP, breast cancer resistance protein; ABCG5/G8, ATP-binding cassette G5 and G8 transporters; CCL4, carbon tetrachloride; PBC, primary biliary cirrhosis; HCV, hepatitis C virus.

Cyclosporine A, rifampin, bosentan, troglitazone, glibenclamide, estradiol 17bglucuronide, progesterone metabolites, nifedipine, fendiline, nicardipine, bepridil, vinblastine, actinomycin, itonavir, saquinavir, efavirenz ?

BSEP

MDR3

Rifampicin, rifamycin SV, glibenclamide, cyclosporin A, ritonavir, saquinavir, efavirenz ? Rifampin, cyclosporine A ? ? cGMP ? ?

NTCP

Inhibiting substances

Non-Genetic Factors Associated with Impaired Hepatocellular Transporter Expression or Function

Transporter

TABLE 3

106 Meier and Pauli-Magnus

Hepatocellular Drug-Induced Liver Injury

107

Oatp1a4 in mice, with Oatp1a1 being male predominant, whereas Oatp1a4 is female predominant (100). In contrast, hepatic Mrp3 and Mrp4 expressions are female predominant in mice (93). However, as no human data are so far available on gender-specific regulation of basolateral transporter proteins, the impact of such findings for the development of liver toxicity is unclear. A recent study investigating the extent of interindividual variability in the canalicular expression of BSEP, MDR3, MRP2, and MDR1 in over hundred samples of human liver tissue could demonstrate significant interindividual variability of canalicular transporter expression, with 15% to 20% of individuals being classified as low or very low expressers of at least one of the investigated proteins (94). However, this variability could not be related to demographic data such as age or gender. Differences in the susceptibility to develop drug-induced cholestasis could therefore not be related to age- and gender differences in baseline expression levels of these canalicular transporter proteins. Preexisting Liver Disease It is well known that preexisting liver disease is associated with a worse outcome of druginduced hepatotoxicity. However, it is still controversial whether preexisting liver disease is also a susceptibility factor for the development of hepatotoxic drug side effects. There is indication that underlying liver pathology might be associated with altered expression and function of canalicular transporter proteins. It could recently been shown that murine hepatic Ntcp and Mrp expressions changed under the administration of the hepatotoxicants acetaminophen and carbon tetrachloride (CCl4). Specifically, CCl4 led to a significant decrease in basolateral Ntcp and Mrp1 expressions, while Mrp4 levels were increased. Conversely, acetaminophen had no influence on Ntcp expression but increased hepatic Mrp3 and Mrp4 contents (88). On the canalicular side, both toxins induced Mrp2 expression, while Bcrp levels remained unchanged (88). In humans, cholestatic alcoholic hepatitis leads to reduced hepatic mRNA and protein expression levels of NTCP and BSEP, whereas mRNA levels of MRP2, MRP3, MDR1, and MDR3 remain unchanged (Table 3). In contrast, early stages of primary biliary cirrhosis (stages I and II) and mild cholestasis were not associated with changes in hepatocellular transporter expression (94,101,102). The impact of such disease-associated adaptive changes in hepatocellular transporter expression on cholestatic drug side effects has not been investigated so far; however, it can be speculated that downregulation of BSEP in certain pathological conditions constitutes an additional risk factor for the development of drug-induced cholestatic liver injury. Genetic Risk Factors In addition to these environmental risk factors, genetics are a major determinant of hepatocellular transporter function. Only limited information is so far available on the functional consequences of genetic variation in basolateral transporter systems. Tirona et al. identified a total of 14 non-synonymous SLC1B1 SNPs encoding OATP1B1 in a population of African- and European-Americans (103), six of which exhibited reduced in vitro uptake of the OATP1B1 substrates estrone-3-sulfate and estradiol-17b-glucuronide. OATP1B1 genetic variants have also been associated with interindividual differences in hepatic disposition of pravastatin and irinotecan, respectively (104–107). While the impact of these observations for the development of cholestasis remains to be studied, these data indicate that polymorphic OATP1B1 function and expression are a determinant of hepatic exposure to OATP1B1 substrates. High OATP expression phenotypes might therefore constitute a risk factor to develop hepatotoxicity due to high hepatocellular uptake of drugs and hepatotoxins, while low expression phenotypes should be protective. On the canalicular side, mutations in the ABCB11 and ABCB4 genes encoding BSEP and MDR3 are a well established cause of inherited cholestatic syndromes, such as progressive and benign forms of familial cholestasis (108–110). A variety of different mutations and polymorphisms has so far been described for inherited forms of cholestasis, leading to absent protein expression or expression of a non-functional transporter (72). Furthermore, mutations and polymorphisms in these two genes have been associated with acquired forms of cholestasis,

108

Meier and Pauli-Magnus

TABLE 4 Mutations of BSEP (ABCB11) and MDR3 (ABCB4) Associated with Drug-Induced Liver Injury Gene ABCB11

ABCB4

Amino acid change

Type of liver injury

V444A

Cholestatic and hepatocellular

D676Y G855R

Cholestatic Cholestatic

E297GC R432T

Mixed

I764L L1082Q

Cholestatic Hepatocellular

IVS5C2TO Cholestatic CCA934T

Implicated drug b-Lactam antibiotics, hormonal contraceptives, proton pump inhibitors, neuroleptics, statins Fluvastatin Ethinyl estradiol/ gestagen NSAID Risperidone Amoxicilline/ clavulanic acid Ethinyl estradiol/ levonorgestrel

Protein expression

Protein function

Y

4

? ?

4 Y

4

Y

? ?

? ?

?

?

Associated condition

References (115)

ICP

(115) (115)

BRIC2

(108) (115) (115)

PFIC3

(116)

Abbreviations: ICP, intrahepatic cholestasis of pregnancy; BRIC2, benign recurrent intrahepatic cholestasis; PFIC, progressive familial intrahepatic cholestasis; NSAID, nonsteroidal anti-inflammatory drug; BSEP, bile salt export pump.

such as intrahepatic cholestasis of pregnancy (ICP), primary biliary cirrhosis, and primary sclerosing cholangitis, and very recently with the susceptibility to develop drug-induced cholestasis (Table 4) (111–115). Patients with ICP or benign forms of familial cholestasis were occasionally reported to exhibit increased susceptibility to certain drugs. For instance, increased susceptibility to oral contraceptives or postmenopausal hormone replacement therapy is a frequent phenomenon in patients with ICP (112,117), while different anti-inflammatory drugs were suspected to induce cholestatic episodes in a patient with benign recurrent intrahepatic cholestasis (108). These observations favor the concept that a genetically determined canalicular transporter deficiency is the common pathophysiological denominator for the development of cholestasis under different extrinsic and intrinsic challenges in affected patients. In addition to such disease-causing mutations, a frequent polymorphism in ABCB11 has recently been associated with a threefold increased risk to develop cholestatic side effect under treatment with different drugs, such as b-lactam antibiotics, oral contraceptives, psychotropic drugs, and proton pump inhibitors (115). Very interestingly, this polymorphism, which leads to a valine to alanine exchange at the highly conserved position 444 of the BSEP protein, has also been associated with decreased hepatic BSEP content in human liver tissue samples, offering a mechanistic explanation for this observation (94). It can therefore be speculated that ABCB11 and ABCB4 genetic variants found to be associated with different cholestatic conditions also predispose to the occurrence of cholestasis under treatment with certain drugs. Mutations in ABCC2 encoding MRP2 results in the Dubin–Johnson syndrome (DJS), a disease characterized by conjugated hyperbilirubinemia. Overt jaundice was observed in patients with subclinical DJS under treatment with oral contraceptives (118), probably due to a metabolite-induced competitive inhibition of bilirubin diglucuronide excretion. As for BSEP and MDR3, a genetically determined MRP2 deficiency syndrome therefore also predisposes to hepatic drug side effects. No association was found between the presence of two frequent ABCC2 polymorphisms (V1188E and C1515Y) and the susceptibility for cholestatic drug side effects (unpublished data). Such an association was suspected based upon the observation that hepatic MRP2 expression was significantly influenced by these ABCC2 polymorphisms in healthy liver tissue (94). Functionally relevant genetic polymorphisms affecting drug disposition have also been described for ABCB1 encoding MDR1. For instance, decreased MDR1 function has been found

Hepatocellular Drug-Induced Liver Injury

109

with a synonymous polymorphism in exon 26 (C3435T). Homozygous carriers of this polymorphism, which is linked to a non-synonymous polymorphism in exon 21 (G2677T), were initially shown to exhibit increased plasma levels of the MDR1 substrate digoxin (119,120). Furthermore, the C3435T polymorphism could be associated with differences in treatment outcome in HIV infection and acute lymphatic leukemia (121–123). However, study results remain controversial, as some authors found increased MDR1 expression and function in carriers of the variant alleles in positions 2677 and 3435 (124), while others failed to show functional differences between the two groups (125,126). The impact of these polymorphisms on the canalicular excretion of drugs and its metabolites therefore remains to be delineated. Recent studies have also identified frequently occurring non-synonymous polymorphisms in the ABC half transporter genes ABCG5/ABCG8 and ABCG2. Although some of these variants result in significantly altered transport capacity and substrate handling of the encoded proteins, a possible role of these polymorphisms for hepatic drug handling and the development of liver injury has not been studied (127,128).

CONCLUSIONS From the examples delineated in this chapter, it is apparent that hepatocellular and bile ductular transporters play an important role in the physiological circulation of bile constituents and in controlling hepatic exposure to potential hepatotoxins. Numerous examples underscore that an acquired dysfunction of bile salt transporters due to drug interactions or environmental factors is one important cause of drug-induced cholestatic side effects. In addition, it has become increasingly evident that genetics are a major determinant of hepatocellular transporter function and expression, thereby determining an individual’s susceptibility to develop cholestasis. The challenge in the future will consist in integrating these environmental and genetic determinants of drug toxicity in a comprehensive system, allowing the cautious use of problematic drugs in susceptible patients. REFERENCES 1. Ishak KG, Zimmerman HJ. Morphologic spectrum of drug-induced hepatic disease. Gastroenterol Clin North Am 1995; 24(4):759–86. 2. Benichou C. Criteria of drug-induced liver disorders. Report of an international consensus meeting. J Hepatol 1990; 11(2):272–6. 3. Eloranta JJ, Kullak-Ublick GA. Coordinate transcriptional regulation of bile acid homeostasis and drug metabolism. Arch Biochem Biophys 2005; 433(2):397–412. 4. Trauner M, Boyer JL. Bile salt transporters: molecular characterization, function, and regulation. Physiol Rev 2003; 83(2):633–71. 5. Hagenbuch B, Dawson P. The sodium bile salt cotransport family SLC10. Pflugers Archiv 2004; 447(5):566–70. 6. Hagenbuch B, Meier PJ. Organic anion transporting polypeptides of the OATP/SLC21 family: phylogenetic classification as OATP/SLCO superfamily, new nomenclature and molecular/functional properties. Pflugers Archiv 2004; 447(5):653–65. 7. Hagenbuch B, Meier PJ. The superfamily of organic anion transporting polypeptides. Biochim Biophys Acta 2003; 1609(1):1–18. 8. Koepsell H, Endou H. The SLC22 drug transporter family. Pflugers Archiv 2004; 447(5):666–76. 9. Koepsell H, Schmitt BM, Gorboulev V. Organic cation transporters. Rev Physiol Biochem Pharmacol 2003; 150:36–90. 10. Burckhardt BC, Burckhardt G. Transport of organic anions across the basolateral membrane of proximal tubule cells. Rev Physiol Biochem Pharmacol 2003; 146:95–158. 11. Borst P, Evers R, Kool M, et al. A family of drug transporters: the multidrug resistance–associated proteins. J Natl Cancer Inst 2000; 92(16):1295–302. 12. Borst P, de Wolf C, van de Wetering K. Multidrug resistance–associated proteins 3, 4, and 5. Pflugers Arch-Eur J Physiol 2007; 453:661–73. 13. Denson LA, Sturm E, Echevarria W, et al. The orphan nuclear receptor, shp, mediates bile acidinduced inhibition of the rat bile acid transporter, ntcp. Gastroenterology 2001; 121(1):140–7. 14. Neimark E, Chen F, Li X, et al. Bile acid-induced negative feedback regulation of the human ileal bile acid transporter. Hepatology 2004; 40(1):149–56.

110

Meier and Pauli-Magnus

15. Fischer WJ, Altheimer S, Cattori V, et al. Organic anion transporting polypeptides expressed in liver and brain mediate uptake of microcystin. Toxicol Appl Pharmacol 2005; 203(3):257–63. 16. Meier-Abt F, Faulstich H, Hagenbuch B. Identification of phalloidin uptake systems of rat and human liver. Biochim Biophys Acta 2004; 1664(1):64–9. 17. Fehrenbach T, Cui Y, Faulstich H, et al. Characterization of the transport of the bicyclic peptide phalloidin by human hepatic transport proteins. Naunyn Schmiedebergs Arch Pharmacol 2003; 368(5):415–20. 18. Letschert K, Faulstich H, Keller D, et al. Molecular characterization and inhibition of amanitin uptake into human hepatocytes. Toxicol Sci 2006; 91(1):140–9. 19. Jung D, Kullak-Ublick GA. Hepatocyte nuclear factor 1 alpha: a key mediator of the effect of bile acids on gene expression. Hepatology 2003; 37(3):622–31. 20. Jung D, Hagenbuch B, Gresh L, et al. Characterization of the human OATP-C (SLC21A6) gene promoter and regulation of liver-specific OATP genes by hepatocyte nuclear factor 1 alpha. J Biol Chem 2001; 276(40):37206–14. 21. Jung D, Podvinec M, Meyer UA, et al. Human organic anion transporting polypeptide 8 promoter is transactivated by the farnesoid X receptor/bile acid receptor. Gastroenterology 2002; 122(7):1954–66. 22. Mukhopadhyay S, Ananthanarayanan M, Stieger B, et al. Sodium taurocholate cotransporting polypeptide is a serine, threonine phosphoprotein and is dephosphorylated by cyclic adenosine monophosphate. Hepatology 1998; 28(6):1629–36. 23. Webster CR, Anwer MS. Role of the PI3K/PKB signaling pathway in cAMP-mediated translocation of rat liver Ntcp. Am J Physiol 1999; 277(6 Pt 1):G1165–72. 24. Webster CR, Blanch C, Anwer MS. Role of PP2B in cAMP-induced dephosphorylation and translocation of NTCP. Am J Physiol 2002; 283(1):G44–50. 25. Wang P, Wang JJ, Xiao Y, et al. Interaction with PDZK1 is required for expression of organic anion transporting protein 1A1 on the hepatocyte surface. J Biol Chem 2005; 280(34):30143–9. 26. Jonker JW, Wagenaar E, Mol CA. Reduced hepatic uptake and intestinal excretion of organic cations in mice with a targeted disruption of the organic cation transporter 1 (Oct1 [Slc22a1]) gene. Mol Cell Biol 2001; 21(16):5471–7. 27. Homolya L, Varadi A, Sarkadi B. Multidrug resistance–associated proteins: export pumps for conjugates with glutathione, glucuronate or sulfate. Biofactors 2003; 17(1–4):103–14. 28. Schuetz EG, Strom S, Yasuda K, et al. Disrupted bile acid homeostasis reveals an unexpected interaction among nuclear hormone receptors, transporters, and cytochrome P450. J Biol Chem 2001; 276(42):39411–8. 29. Stieger B. FIC1: another bile salt carrier within the enterohepatic circulation? J Hepatol 2001; 35(4):522–4. 30. Kullak-Ublick GA, Stieger B, Meier PJ. Enterohepatic bile salt transporters in normal physiology and liver disease. Gastroenterology 2004; 126(1):322–42. 31. Meier PJ, Stieger B. Bile salt transporters. Annu Rev Physiol 2002; 64:635–61. 32. Hirano M, Maeda K, Hayashi H, et al. Bile salt export pump (BSEP/ABCB11) can transport a nonbile acid substrate, pravastatin. J Pharmacol Exp Ther 2005; 314(2):876–82. 33. van Erpecum KJ. Biliary lipids, water and cholesterol gallstones. Biol Cell 2005; 97(11):815–22. 34. Fromm MF. The influence of MDR1 polymorphisms on P-glycoprotein expression and function in humans. Adv Drug Deliv Rev 2002; 54(10):1295–310. 35. Borst P, Zelcer N, van de Wetering K. MRP2 and 3 in health and disease. Cancer Lett 2006; 234(1):51–61. 36. Nies AT, Keppler D. The apical conjugate efflux pump ABCC2 (MRP2). Pflugers Arch-Eur J Physiol 2006; 453:643–59. 37. Adachi Y, Suzuki H, Schinkel AH, et al. Role of breast cancer resistance protein (Bcrp1/Abcg2) in the extrusion of glucuronide and sulfate conjugates from enterocytes to intestinal lumen. Mol Pharmacol 2005; 67(3):923–8. 38. Breedveld P, Zelcer N, Pluim D, et al. Mechanism of the pharmacokinetic interaction between methotrexate and benzimidazoles: potential role for breast cancer resistance protein in clinical drugdrug interactions. Cancer Res 2004; 64(16):5804–11. 39. Merino G, Alvarez AI, Pulido MM, et al. Breast cancer resistance protein (BCRP/ABCG2) transports fluoroquinolone antibiotics and affects their oral availability, pharmacokinetics, and milk secretion. Drug Metab Dispos 2006; 34(4):690–5. 40. van Herwaarden AE, Wagenaar E, Karnekamp B, et al. Breast cancer resistance protein (Bcrp1/Abcg2) reduces systemic exposure of the dietary carcinogens aflatoxin B1, IQ and Trp-P-1 but also mediates their secretion into breast milk. Carcinogenesis 2006; 27(1):123–30. 41. Merino G, Jonker JW, Wagenaar E, et al. The breast cancer resistance protein (BCRP/ABCG2) affects pharmacokinetics, hepatobiliary excretion, and milk secretion of the antibiotic nitrofurantoin. Mol Pharmacol 2005; 67(5):1758–64. 42. Merino G, Jonker JW, Wagenaar E, et al. Transport of anthelmintic benzimidazole drugs by breast cancer resistance protein (BCRP/ABCG2). Drug Metab Dispos 2005; 33(5):614–8.

Hepatocellular Drug-Induced Liver Injury

111

43. Pavek P, Merino G, Wagenaar E, et al. Human breast cancer resistance protein: interactions with steroid drugs, hormones, the dietary carcinogen 2-amino-1-methyl-6-phenylimidazo(4,5-b)pyridine, and transport of cimetidine. J Pharmacol Exp Ther 2005; 312(1):144–52. 44. Zaher H, Khan AA, Palandra J, et al. Breast cancer resistance protein (Bcrp/abcg2) is a major determinant of sulfasalazine absorption and elimination in the mouse. Mol Pharm 2006; 3(1):55–61. 45. Suzuki M, Suzuki H, Sugimoto Y, et al. ABCG2 transports sulfated conjugates of steroids and xenobiotics. J Biol Chem 2003; 278(25):22644–9. 46. Hirano M, Maeda K, Matsushima S, et al. Involvement of BCRP (ABCG2) in the biliary excretion of pitavastatin. Mol Pharmacol 2005; 68(3):800–7. 47. Klett EL, Patel S. Genetic defenses against noncholesterol sterols. Curr Opin Lipidol 2003; 14(4):341–5. 48. Kosters A, Kunne C, Looije N, et al. The mechanism of Abcg5/Abcg8 in biliary cholesterol secretion in mice. J Lipid Res 2006; 47:1959–66. 49. Jonker JW, Merino G, Musters S, et al. The breast cancer resistance protein BCRP (ABCG2) concentrates drugs and carcinogenic xenotoxins into milk. Nat Med 2005; 11(2):127–9. 50. Byrne JA, Strautnieks SS, Mieli-Vergani G, et al. The human bile salt export pump: characterization of substrate specificity and identification of inhibitors. Gastroenterology 2002; 123(5):1649–58. 51. Noe J, Stieger B, Meier PJ. Functional expression of the canalicular bile salt export pump of human liver. Gastroenterology 2002; 123(5):1659–66. 52. Ananthanarayanan M, Balasubramanian N, Makishima M, et al. Human bile salt export pump promoter is transactivated by the farnesoid X receptor/bile acid receptor. J Biol Chem 2001; 276(31):28857–65. 53. Huang L, Zhao A, Lew JL, et al. Farnesoid X receptor activates transcription of the phospholipid pump MDR3. J Biol Chem 2003; 278(51):51085–90. 54. Plass JR, Mol O, Heegsma J, et al. Farnesoid X receptor and bile salts are involved in transcriptional regulation of the gene encoding the human bile salt export pump. Hepatology 2002; 35(3):589–96. 55. Kast HR, Goodwin B, Tarr PT, et al. Regulation of multidrug resistance–associated protein 2 (ABCC2) by the nuclear receptors pregnane X receptor, farnesoid X-activated receptor, and constitutive androstane receptor. J Biol Chem 2002; 277(4):2908–15. 56. Kurz AK, Graf D, Schmitt M, et al. Tauroursodesoxycholate-induced choleresis involves p38(MAPK) activation and translocation of the bile salt export pump in rats. Gastroenterology 2001; 121(2):407–19. 57. Misra S, Ujhazy P, Gatmaitan Z, et al. The role of phosphoinositide 3-kinase in taurocholate-induced trafficking of ATP-dependent canalicular transporters in rat liver. J Biol Chem 1998; 273(41):26638–44. 58. Misra S, Ujhazy P, Varticovski L, et al. Phosphoinositide 3-kinase lipid products regulate ATPdependent transport by sister of P-glycoprotein and multidrug resistance associated protein 2 in bile canalicular membrane vesicles. Proc Natl Acad Sci USA 1999; 96(10):5814–9. 59. Kim RB, Fromm MF, Wandel C, et al. The drug transporter P-glycoprotein limits oral absorption and brain entry of HIV-1 protease inhibitors. J Clin Invest 1998; 101(2):289–94. 60. Kim RB, Wandel C, Leake B, et al. Interrelationship between substrates and inhibitors of human CYP3A and P-glycoprotein. Pharm Res 1999; 16(3):408–14. 61. Greiner B, Eichelbaum M, Fritz P, et al. The role of intestinal P-glycoprotein in the interaction of digoxin and rifampin. J Clin Invest 1999; 104(2):147–53. 62. Durr D, Stieger B, Kullak-Ublick GA, et al. St John’s Wort induces intestinal P-glycoprotein/MDR1 and intestinal and hepatic CYP3A4. Clin Pharmacol Ther 2000; 68(6):598–604. 63. Geick A, Eichelbaum M, Burk O. Nuclear receptor response elements mediate induction of intestinal MDR1 by rifampin. J Biol Chem 2001; 276(18):14581–7. 64. Ebert B, Seidel A, Lampen A. Identification of BCRP as transporter of benzo[a]pyrene conjugates metabolically formed in Caco-2 cells and its induction by Ah-receptor agonists. Carcinogenesis 2005; 26(10):1754–63. 65. Krishnamurthy P, Ross DD, Nakanishi T, et al. The stem cell marker Bcrp/ABCG2 enhances hypoxic cell survival through interactions with heme. J Biol Chem 2004; 279(23):24218–25. 66. Imai Y, Ishikawa E, Asada S, et al. Estrogen-mediated post transcriptional down-regulation of breast cancer resistance protein/ABCG2. Cancer Res 2005; 65(2):596–604. 67. Yu L, York J, von Bergmann K, et al. Stimulation of cholesterol excretion by the liver X receptor agonist requires ATP-binding cassette transporters G5 and G8. J Biol Chem 2003; 278(18):15565–70. 68. Benedetti A, Di Sario A, Marucci L, et al. Carrier-mediated transport of conjugated bile acids across the basolateral membrane of biliary epithelial cells. Am J Physiol 1997; 272(6 Pt 1):G1416–24. 69. Rost D, Konig J, Weiss G, et al. Expression and localization of the multidrug resistance proteins MRP2 and MRP3 in human gallbladder epithelia. Gastroenterology 2001; 121(5):1203–8. 70. Lazaridis KN, Tietz P, Wu T, et al. Alternative splicing of the rat sodium/bile acid transporter changes its cellular localization and transport properties. Proc Natl Acad Sci USA 2000; 97(20):11092–7.

112

Meier and Pauli-Magnus

71. Ballatori N, Christian WV, Lee JY, et al. OSTalpha-OSTbeta: a major basolateral bile acid and steroid transporter in human intestinal, renal, and biliary epithelia. Hepatology 2005; 42(6):1270–9. 72. Pauli-Magnus C, Stieger B, Meier Y, et al. Enterohepatic transport of bile salts and genetics of cholestasis. J Hepatol 2005; 43(2):342–57. 73. Desmet VJ. Vanishing bile duct syndrome in drug-induced liver disease. J Hepatol 1997; 26(Suppl. 1):31–5. 74. Vavricka SR, Van Montfoort J, Ha HR, et al. Interactions of rifamycin SV and rifampicin with organic anion uptake systems of human liver. Hepatology 2002; 36(1):164–72. 75. Mita S, Suzuki H, Akita H, et al. Inhibition of bile acid transport across NTCP and BSEP co-expressing LLC-PK1 cells by cholestasis-inducing drugs. Drug Metab Dispos 2006; 34:1575–81. 76. McRae MP, Lowe CM, Tian X, et al. Ritonavir, Saquinavir and Efavirenz, but not Nevirapine, inhibit bile acid transport in human and rat hepatocytes. J Pharmacol Exp Ther 2006; 318:1068–75. 77. Stieger B, Fattinger K, Madon J, et al. Drug- and estrogen-induced cholestasis through inhibition of the hepatocellular bile salt export pump (Bsep) of rat liver. Gastroenterology 2000; 118(2):422–30. 78. Fattinger K, Funk C, Pantze M, et al. The endothelin antagonist bosentan inhibits the canalicular bile salt export pump: a potential mechanism for hepatic adverse reactions. Clin Pharmacol Ther 2001; 69(4):223–31. 79. Funk C, Ponelle C, Scheuermann G, et al. Cholestatic potential of troglitazone as a possible factor contributing to troglitazone-induced hepatotoxicity: in vivo and in vitro interaction at the canalicular bile salt export pump (Bsep) in the rat. Mol Pharmacol 2001; 59(3):627–35. 80. Hirano H, Kurata A, Onishi Y, et al. High-speed screening and QSAR analysis of human ATPbinding cassette transporter ABCB11 (bile salt export pump) to predict drug-induced intrahepatic cholestasis. Mol Pharm 2006; 3(3):252–65. 81. Smith AJ, van Helvoort A, van Meer G, et al. MDR3 P-glycoprotein, a phosphatidylcholine translocase, transports several cytotoxic drugs and directly interacts with drugs as judged by interference with nucleotide trapping. J Biol Chem 2000; 275(31):23530–9. 82. Fouassier L, Kinnman N, Lefevre G, et al. Contribution of mrp2 in alterations of canalicular bile formation by the endothelin antagonist bosentan. J Hepatol 2002; 37(2):184–91. 83. Fromm MF. Importance of P-glycoprotein at blood-tissue barriers. Trends Pharmacol Sci 2004; 25(8):423–9. 84. Bolder U, Trang NV, Hagey LR, et al. Sulindac is excreted into bile by a canalicular bile salt pump and undergoes a cholehepatic circulation in rats. Gastroenterology 1999; 117(4):962–71. 85. Giroux Y, Moreau M, Kass TG. Cholestatic jaundice caused by sulindac. Can J Surg 1982; 25(3):334–5. 86. McIndoe GA, Menzies KW, Reddy J. Sulindac (Clinoril) and cholestatic jaundice. NZ Med J 1981; 94(697):430–1. 87. Simon FR, Fortune J, Iwahashi M, et al. Characterization of the mechanisms involved in the gender differences in hepatic taurocholate uptake. Am J Physiol 1999; 276(2 Pt 1):G556–65. 88. Aleksunes LM, Scheffer GL, Jakowski AB, et al. Coordinated expression of multidrug resistance– associated proteins (Mrps) in mouse liver during toxicant-induced injury. Toxicol Sci 2006; 89(2):370–9. 89. Rost D, Herrmann T, Sauer P, et al. Regulation of rat organic anion transporters in bile salt-induced cholestatic hepatitis: effect of ursodeoxycholate. Hepatology 2003; 38(1):187–95. 90. Oswald M, Kullak-Ublick GA, Paumgartner G, et al. Expression of hepatic transporters OATP-C and MRP2 in primary sclerosing cholangitis. Liver 2001; 21(4):247–53. 91. Buist SC, Klaassen CD. Rat and mouse differences in gender-predominant expression of organic anion transporter (Oat1-3; Slc22a6-8) mRNA levels. Drug Metab Dispos 2004; 32(6):620–5. 92. Ros JE, Libbrecht L, Geuken M, et al. High expression of MDR1, MRP1, and MRP3 in the hepatic progenitor cell compartment and hepatocytes in severe human liver disease. J Pathol 2003; 200(5):553–60. 93. Maher JM, Cheng X, Tanaka Y, et al. Hormonal regulation of renal multidrug resistance–associated proteins 3 and 4 (Mrp3 and Mrp4) in mice. Biochem Pharmacol 2006; 71(10):1470–8. 94. Meier Y, Pauli-Magnus C, Zanger UM, et al. Interindividual variability of canalicular ATP-bindingcassette (ABC)-transporter expression in human liver. Hepatology 2006; 44(1):62–74. 95. Fromm MF. P-glycoprotein: a defense mechanism limiting oral bioavailability and CNS accumulation of drugs. Int J Clin Pharmacol Ther 2000; 38(2):69–74. 96. Allen JD, van Loevezijn A, Lakhai JM, et al. Potent and specific inhibition of the breast cancer resistance protein multidrug transporter in vitro and in mouse intestine by a novel analogue of fumitremorgin C. Mol Cancer Ther 2002; 1(6):417–25. 97. Ozvegy-Laczka C, Hegedus T, Varady G, et al. High-affinity interaction of tyrosine kinase inhibitors with the ABCG2 multidrug transporter. Mol Pharmacol 2004; 65(6):1485–95. 98. Shiozawa K, Oka M, Soda H, et al. Reversal of breast cancer resistance protein (BCRP/ABCG2)mediated drug resistance by novobiocin, a coumermycin antibiotic. Int J Cancer 2004; 108(1):146–51.

Hepatocellular Drug-Induced Liver Injury

113

99. Merino G, van Herwaarden AE, Wagenaar E, et al. Sex-dependent expression and activity of the ATP-binding cassette transporter breast cancer resistance protein (BCRP/ABCG2) in liver. Mol Pharmacol 2005; 67(5):1765–71. 100. Cheng X, Maher JM, Lu H, et al. Endocrine regulation of gender-divergent mouse organic anion transporting polypeptide (Oatp) expression. Mol Pharmacol 2006; 70:1291–7. 101. Zollner G, Fickert P, Silbert D, et al. Adaptive changes in hepatobiliary transporter expression in primary biliary cirrhosis. J Hepatol 2003; 38(6):717–27. 102. Zollner G, Fickert P, Zenz R, et al. Hepatobiliary transporter expression in percutaneous liver biopsies of patients with cholestatic liver diseases. Hepatology 2001; 33(3):633–46. 103. Tirona RG, Leake BF, Merino G, et al. Polymorphisms in OATP-C: identification of multiple allelic variants associated with altered transport activity among European- and African-Americans. J Biol Chem 2001; 276:35669–75. 104. Nishizato Y, Ieiri I, Suzuki H, et al. Polymorphisms of OATP-C (SLC21A6) and OAT3 (SLC22A8) genes: consequences for pravastatin pharmacokinetics. Clin Pharmacol Ther 2003; 73(6):554–65. 105. Mwinyi J, Johne A, Bauer S, et al. Evidence for inverse effects of OATP-C (SLC21A6) 5 and 1b haplotypes on pravastatin kinetics. Clin Pharmacol Ther 2004; 75(5):415–21. 106. Niemi M, Schaeffeler E, Lang T, et al. High plasma pravastatin concentrations are associated with single nucleotide polymorphisms and haplotypes of organic anion transporting polypeptide-C (OATP-C, SLCO1B1). Pharmacogenetics 2004; 14(7):429–40. 107. Nozawa T, Minami H, Sugiura S, et al. Role of organic anion transporter OATP1B1 (OATP-C) in hepatic uptake of irinotecan and its active metabolite, 7-ethyl-10-hydroxycamptothecin: in vitro evidence and effect of single nucleotide polymorphisms. Drug Met Disp 2005; 33(3):434–9. 108. Noe J, Kullak-Ublick GA, Jochum W, et al. Impaired expression and function of the bile salt export pump due to three novel ABCB11 mutations in intrahepatic cholestasis. J Hepatol 2005; 43(3):536–43. 109. Strautnieks SS, Bull LN, Knisely AS, et al. A gene encoding a liver-specific ABC transporter is mutated in progressive familial intrahepatic cholestasis. Nat Genet 1998; 20(3):233–8. 110. van Mil SW, van der Woerd WL, van der Brugge G, et al. Benign recurrent intrahepatic cholestasis type 2 is caused by mutations in ABCB11. Gastroenterology 2004; 127(2):379–84. 111. Jacquemin E, Cresteil D, Manouvrier S, et al. Heterozygous non-sense mutation of the MDR3 gene in familial intrahepatic cholestasis of pregnancy. Lancet 1999; 353(9148):210–1. 112. Jacquemin E, De Vree JM, Cresteil D, et al. The wide spectrum of multidrug resistance 3 deficiency: from neonatal cholestasis to cirrhosis of adulthood. Gastroenterology 2001; 120(6):1448–58. 113. Pauli Magnus C, Lang T, Meier Y, et al. Sequence analysis of bile salt export pump (ABCB11) and multidrug resistance p-glycoprotein 3 (ABCB4, MDR3) in patients with intrahepatic cholestasis of pregnancy. Pharmacogenetics 2004; 14(2):91–102. 114. Pauli-Magnus C, Kerb R, Fattinger K, et al. BSEP and MDR3 haplotype structure in healthy Caucasians, primary biliary cirrhosis and primary sclerosing cholangitis. Hepatology 2004; 39(3):779–91. 115. Lang C, Meier Y, Stieger B, et al. Mutations and polymorphisms in the bile salt export pump and the multidrug ressitance protein 3 associated with drug-induced liver injury. Pharmacogenet Genomics 2007; 17:47–60. 116. Ganne-Carrie N, Baussan C, Grando V, et al. Progressive familial intrahepatic cholestasis type 3 revealed by oral contraceptive pills. J hepatol 2003; 38(5):693–4. 117. Leevy CB, Koneru B, Klein KM. Recurrent familial prolonged intrahepatic cholestasis of pregnancy associated with chronic liver disease. Gastroenterology 1997; 113(3):966–72. 118. Lindberg MC. Hepatobiliary complications of oral contraceptives. J Gen Intern Med 1992; 7(2):199–209. 119. Hoffmeyer S, Burk O, von Richter O, et al. Functional polymorphisms of the human multidrugresistance gene: multiple sequence variations and correlation of one allele with P-glycoprotein expression and activity in vivo. Proc Natl Acad Sci USA 2000; 97(7):3473–8. 120. Hitzl M, Drescher S, van der Kuip H, et al. The C3435T mutation in the human MDR1 gene is associated with altered efflux of the P-glycoprotein substrate rhodamine 123 from CD56C natural killer cells. Pharmacogenetics 2001; 11(4):293–8. 121. Fellay J, Marzolini C, Meaden ER, et al. Response to antiretroviral treatment in HIV-1-infected individuals with allelic variants of the multidrug resistance transporter 1: a pharmacogenetics study. Lancet 2002; 359(9300):30–6. 122. Jamroziak K, Mlynarski W, Balcerczak E, et al. Functional C3435T polymorphism of MDR1 gene: an impact on genetic susceptibility and clinical outcome of childhood acute lymphoblastic leukemia. Eur J Haematol 2004; 72(5):314–21. 123. Plasschaert SL, Groninger E, Boezen M, et al. Influence of functional polymorphisms of the MDR1 gene on vincristine pharmacokinetics in childhood acute lymphoblastic leukemia. Clin Pharmacol Ther 2004; 76(3):220–9. 124. Kim RB, Leake BF, Choo EF, et al. Identification of functionally variant MDR1 alleles among European Americans and African Americans. Clin Pharmacol Ther 2001; 70(2):189–99.

114

Meier and Pauli-Magnus

125. Drescher S, Schaeffeler E, Hitzl M, et al. MDR1 gene polymorphisms and disposition of the P-glycoprotein substrate fexofenadine. Br J Clin Pharmacol 2002; 53(5):526–34. 126. Pauli-Magnus C, de Young J, Herskowitz I, et al. Ethnic variation in MDR1 haplotype structure. Clin Pharm Ther 2002; 71(2):35–115. 127. Kajinami K, Brousseau ME, Nartsupha C, et al. ATP binding cassette transporter G5 and G8 genotypes and plasma lipoprotein levels before and after treatment with atorvastatin. J Lip Res 2004; 45(4):653–6. 128. Morisaki K, Robey RW, Ozvegy-Laczka C, et al. Single nucleotide polymorphisms modify the transporter activity of ABCG2. Cancer Chemother Pharmacol 2005; 56(2):161–72.

7

Immunological Mechanisms in Drug-Induced Liver Injury Dwain L. Thiele

Division of Digestive and Liver Diseases, Department of Internal Medicine, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas, U.S.A.

INTRODUCTION Immune-mediated injury plays a primary or secondary role in many forms of drug-induced liver injury (DILI) (1,2). This mechanism of DILI is suggested clinically by the presence of features of systemic immune activation such as concomitant fever, rash, atypical lymphocytes, and/or eosinophilia. When drug rechallenge occurs, immune-mediated mechanisms of DILI are confirmed by the development of an accelerated onset of equally or more severe liver injury indicating the presence of immunological memory, the signature of an adaptive immune response (1,3). In many such cases, hepatic drug metabolism has been found to produce reactive metabolites that haptenate or alkylate self proteins which in turn elicit adaptive immune responses and associated liver injury in genetically susceptible individuals (2,4–6). More recently, it has been recognized that drugs that directly injure hepatocytes may also trigger innate immune responses apparently directed at damaged hepatocytes (2,7,8). In many, but not all, forms of putative immune-mediated drug injury, antibody responses directed at altered self or autoantigens can be detected. However, while antibody responses may be directly implicated in extrahepatic immune complex-mediated diseases or hemolytic syndromes associated with some forms of immune-mediated liver injury, there is a poor correlation between such antibody responses and either the presence or severity of DILI or other forms of immune-mediated liver injury (9–11). Thus, while antibodies can theoretically induce cellular injury either through fixation of complement or triggering of cell-mediated antibody dependent cell-mediated cytotoxicity, there is little evidence that such mechanisms alone commonly trigger liver injury. Cytokine responses also play a prominent role in both the innate and adaptive immune response to viral, auto, or altered self antigens. Proinflammatory cytokines such as tumor necrosis factor (TNF), interleukin-1, and interleukin-6 that are elicited during the course of innate and adaptive immune responses are now well recognized causes of intrahepatic cholestasis mediated by the role that these cytokines play in the transcriptional regulation of hepatobiliary transport systems (12,13). Accumulation of bile salts during cholestasis can in turn induce hepatocyte or biliary epithelial cell apoptosis and liver enzyme abnormalities (14,15). Fortunately however, the cholestatic effects of intrahepatic or systemic proinflammatory cytokine responses rarely are associated with severe hepatocellular injury or acute liver failure. Rather, the component of the immune response most clearly implicated as a cause of severe acute hepatocellular injury during the course of viral, autoimmune or drug-induced hepatitis is the cell-mediated cytotoxicity mediated by cytotoxic T lymphocytes (CTL), natural killer cells (NK), and/or natural killer cell with T cell receptor (NKT) cells (7,16–18).

The author of this chapter has a relationship with the following corporation: membership in the Speaker’s Bureau for Schering.

116

Thiele

STIMULI THAT TRIGGER IMMUNE-MEDIATED LIVER INJURY Development of T cell cytokine responses or CTL-mediated immune injury in the liver is dependent upon activation and differentiation of antigen-specific effector cells that are then triggered to kill target cells. Each step in this process is dependent upon processing and presentation of peptide antigens within the binding groove of major histocompatibility complex (MHC) molecules by liver cells, or specialized antigen-presenting cells to T cells expressing receptors that recognize the MHC—peptide complex (19). Initial steps in this activation cascade also are dependent upon costimulatory signals delivered by signaling molecules on antigen- presenting cells, by inflammatory cytokines, and by T helper cytokines that promote activation, proliferation, and differentiation of antigen-specific T cells (20). Indeed, a common requirement for the induction of adaptive immune responses is not only the presence of an appropriate antigenic stimulus for T and B lymphocytes but also the other “danger signals” commonly provided by bacterial and viral infections to trigger innate immune responses that serve the role of adjuvants which potentiate adaptive immunity (21). However, once activated, memory T helper cell or CTL recognition of antigen bearing target cells, and triggering of cytolytic effector function is largely dependent only upon engagement of T cell antigen receptor complexes by target cell MHC-peptide complexes, (Table 1) (22). Antigenic stimuli responsible for production of altered self antigens responsible for eliciting adaptive immune responses to drugs such as tienilic acid and dihydralazine have been well defined (6,23). In each case, a Cytochrome P450 (CYP) isoenzyme converts the parent drug into a reactive metabolite that in turn binds the CYP isoenzyme to form a neoantigen that is identified by specific antibodies present in the sera of patients who have developed DILI during the course of ingestion of the parent drug. Presumably, this occurs because the modified CYP enzyme now contains drug-haptenated peptide sequences that serve as both the T helper and B cell epitopes required to elicit a classic T cell-dependent antibody response (24). Indeed, the presence of peripheral blood mononuclear lymphocytes that proliferate in response to drugs or drug metabolite altered proteins supports the presence of such T cell responses during the course of such forms of immune-mediated liver injury (25–28). The peptide epitopes recognized by CD4 and CD8 T cells in association with Class II and Class I MHC molecules, respectively usually represent different epitopes of the eliciting protein antigen as MHC Class II molecules typically bind 12 amino acid peptides processed in the endolysosomal compartment of antigen-presenting cells whereas Class I MHC molecules

TABLE 1 Examples of Receptors that Regulate Cytotoxic Lymphocyte Effector Function Effector cell CTL T helper NKT NK (C/K NKT)

Receptor CD3/TCR/CD8 CD3/TCR/CD4 CD3/TCR/CD8 Killer Ig-like receptors KIR2DS1 KIR2DS2 KIR2DL1 KIR2DL2/3 KIR3DS1 KIR3DL2 Lectin-like receptors CD94/NKG2A CD94/NKG2C Natural cytotoxicity receptors NKG2D NKp44/46 CD16 (Fc Receptor)

Ligand

Effect on cytotoxicity

MHC Class ICpeptide MHC Class IICpeptide CD1dCglycolipid

Activates Potentiates Activates

HLA Cw2/4/5/6 HLA Cw1/3/7/8 HLA Cw2/4/5/6 HLA Cw1/3/7/8 HLA Bw4 HLA Aw3/11Cpeptide

Activates Activates Inhibits Inhibits Activates Inhibits

HLA-E HLA-E

Inhibits Activates

MICA/B, ULPB Viral hemagglutinins IgG

Activates Activates Activates

Abbreviations: MIC, MHC class I chain related; ULPB, UL-16 binding proteins; CD, cluster of differentiation; HLA, human leucocyte antigen; IgG, immunoglobulin G; CTL, cytotoxic T lymphocytes; MHC, major histocompatibility complex; TCR, T cell receptor.

Immunology of Drug-Induced Liver Injury

117

bind eight to nine amino acid peptides produced by protein degradation by the proteasome (29,30). B cell epitopes, in turn, also typically differ from T cell epitopes as Immunoglobulin (Ig) molecules can bind not only linear or conformational protein epitopes but also a variety of other nonprotein epitopes. Viruses or other pathogens commonly present a multitude of such antigenic epitopes to the host immune system and usually elicit all components of a host adaptive immune response, likely because of the large numbers of nonself (CD, cluster of differentiation) CD4 T cell, CD8 T cell, and B cell epitopes that are presented. However, the range of nonself epitopes produced by small molecule metabolites of drugs binding to a single CYP metabolizing isoenzyme, likely constitutes a much more limited repertoire of potential T and B cell epitopes. This may account for the low frequency of patients that develop the full spectrum of immunoallergic responses to drugs that appear to uniformly produce reactive, protein binding metabolites in all recipients (2,31). However, there are examples of other drugs such as sulfamethoxazole that appear to produce reactive metabolites that either modify multiple MHC binding peptides or, like superantigens, directly bind to or modify MHC molecules and thereby activate polyclonal T cell responses (32). In addition, when peripheral blood mononuclear cells (PBMC) from patients with DILI are cultured with the implicated drug in in vitro cultures along with prostaglandin inhibitors to block suppressor cell function, lymphocyte proliferative responses manifested by 3H-thymidine incorporation are detected in greater than 50% of DILI patients but not in lymphocytes from controls (25). Thus, in a significant fraction of patients with immune-mediated DILI, specific lymphocyte responses appear to be directed at the drug itself or drug metabolites produced by PBMC. An alternate explanation for the low frequency of clinically significant immune responses to drugs that produce chemically reactive metabolites in all recipients, may be the lack of the types of innate immune response stimulation during drug sensitization events that occur during infectious processes when pathogen-associated molecular patterns (PAMP) such as endotoxin, present during bacterial infections, or double stranded or single stranded RNA molecules present during viral infections stimulate toll-like receptors (TLRs) on NK, macrophage, or dendritic cells (2). In contrast to T cells that require prior antigenspecific activation to develop a full repertoire of cytokine and cytotoxic effector mechanisms, populations of circulating and resident hepatic NK cells expressing fully differentiated cytolytic effector mechanisms are present in the absence of any prior sensitization. These cells are also major sources of proinflammatory cytokines, such as TNF and interferon early in the course of immune responses (33,34). Of note, NK, and/or NKT cells also constitute a major fraction of the resident lymphocytes present in the liver (35,36) and thus are among the first lymphocytes exposed to drug-induced alterations in the liver. These effector cells require inhibitory signals from normal autologous cells to prevent unwanted cell death or cytokine responses (Table 1) (33,36). The best characterized of these inhibitory signals are those mediated by interactions of HLA Class I molecules with killer immunoglobulin-like receptors (KIR) family receptors, or with the CD94/NKG2A lectin like receptor (34). Other KIR family receptors, the CD94/NKG2C lectin like receptor, the Ig binding receptor CD16, and a heterogeneous group of natural cytotoxicity receptors can also activate NK cells and trigger target cell killing (33,34). Of note these natural cytotoxicity receptors that are capable of triggering NK cell cytotoxicity, not only include the TLRs capable of response to PAMPs, such as endotoxin or double stranded RNA, but also NKG2D, a receptor for MHC Class I chain (MIC) related stress induced proteins MICA and MICB (34) likely to be expressed in hepatocytes injured by directly hepatotoxic drug metabolites (2). Indeed, valproate has been found to be a potent inducer of MICA and MICB protein expression in hepatocellular carcinoma lines that in turn exhibit increased susceptibility to lysis by NK cells (8). Moreover, in animal models of acetaminophen induced hepatotoxicity, NK, and NKT cells also have been found to play a role in amplifying the degree of hepatotoxicity produced by sublethal doses of this direct hepatotoxin (7). As many drugs capable of inducing DILI such as antibiotics or antipyretics are commonly administered to patients with bacterial or viral infections, it is plausible that additive NK activation signals provided by TLR signaling triggered by PAMPs expressed by these infectious agents, NKG2D responses to stress proteins expressed by hepatocytes mildly injured by drug metabolites and CD16 activation by

118

Thiele

antibodies to drug-haptenated hepatocyte proteins could in concert trigger either exacerbation of mild hepatotoxicity initiated by a directly hepatotoxic drug or serve as the initial stimulus to development of a classic immunoallergic drug response that may continue to evolve even after resolution of the initial illness (2). MECHANISMS OF LYMPHOCYTE-MEDIATED CYTOTOXICITY Cytotoxic lymphocytes kill target cells by two major pathways (Fig. 1). The first is a rapid kinetics pathway dependent upon exocytosis from preformed secretory granules of granule effector molecules including the membrane-disrupting protein, perforin, and the apoptosisinducing granzymes (37). The second, delayed kinetics pathway is dependent upon engagement of target cell TNF receptor (TNFR) family death receptors (e.g. Fas, TNFR1, DR4, DR5) by Fas ligand (FasL), TNF-related apoptosis-inducing ligand (TRAIL), membranebound TNF or secreted TNF (38–40). Perforin was initially identified by its ability to form pores in target cell membranes (41) but subsequently was found to induce target cell apoptosis and efficiently kill nucleated target cells only when functioning in conjunction with granzymes A and/or B (42). These observations initially led to the suggestion that perforin generated pores in target cell membranes allowed the entry of proapoptotic granzymes into target cells. However, subsequent experimental data reveal that granzymes enter target cells independently of perforin by receptor-mediated endocytosis after binding to the mannose 6-phosphate/ insulin-like growth factor II receptor (43) or by other pathways (44) and that perforin facilitates target cell killing by releasing granzyme molecules from endolysosomal compartments into the cytosol or nucleus where they are capable of initiating an apoptotic cascade (45).

TNF, LT-α

Target Cell Fas ligand, TNF, TRAIL Fas, TNFRI, DR4, DR5

TCR

MHC Class I Caspase activation, +Antigen mitochondrial dysfunction

CD8

Intracellular signaling

Granzymes P P Apoptosis

Perforin (p) Granzymes

IFN-Y

Granule exocytosis

Cytotoxic T Lymphocyte

FIGURE 1 Effector mechanisms employed by cytotoxic T lymphocytes. Following activation via the T cell receptor (TCR), cytotoxic lymphocytes are triggered to exocytose contents of cytolytic granules and to mobilize death receptor ligands such as TNF, Fas ligand and TRAIL that are capable of inducing apoptosis in susceptible target cells.

119

Immunology of Drug-Induced Liver Injury

Granzymes A and B, as well as various other granzymes expressed by mouse (granzymes C, D, E, F, G, and K) and human (granzymes H, K, and M) cytotoxic lymphocytes (40,46) are all members of a family of neutral serine proteases that also includes cathepsin G, neutrophil elastase, and proteinase 3 (expressed by neutrophils) and various mast cell chymases and tryptases. Granzyme B is an aspase (47) that cleaves peptide bonds after aspartic acid residues that are part of an extended (I/V) EPD motif in protein substrates such as procaspase 3 (48,49), other procaspases (50–53), the inhibitor of caspase-activated DNAse (ICAD) (54), and Bid (55,56). Granzyme B cleavage of these protein substrates results in caspase and ICAD activation as well as tBid translocation to mitochondria and thus initiates or facilitates multiple apoptotic mechanisms. Granzyme B-deficient CTL are defective in rapid killing of target cells (57) but in prolonged assays mediate delayed target cell killing. Granzyme A is a tryptase that induces target cell apoptosis by pathways independent of those involved in granzyme B induced apoptosis (58). Granzyme A-induced apoptosis proceeds with slower kinetics than granzyme B induced apoptosis and appears to be mediated by activation of an endoplasmic reticulumassociated caspase-independent nuclease that induces singlestranded DNA nicks (59). The role of other granzymes expressed in mouse (granzymes C, D, E, F, G, and K) and human (granzymes H, K, and M) cytotoxic lymphocytes in cell-mediated cytotoxicity has been less extensively studied, but at least some of these molecules also have been shown to induce target cell apoptosis when present in combination with perforin (60,61). The death receptor pathway of cell-mediated cytotoxicty was initially viewed as a secondary or complementary pathway of target cell killing by CTL (38–40) that seemed to play even less of a role in NK cell killing (33,36). However, initial lack of appreciation for the often major role for death receptor-mediated killing by immune effector cells was, likely, based in part on the fact that killing by this mechanism is entirely dependent upon expression by the target cell of appropriate cell surface death receptors such as Fas, TNFR1, DR4 or DR5, and requires the use of longer assays that detect the more delayed, slower kinetics induction of target cell apoptosis by this pathway. When activated CTL (38–40) and, under at least some circumstances, activated NK cells (62), contact target cells expressing appropriate death receptors of the TNFR family, the effector cells translocate FasL, TNF, or TRAIL to the cell surface where these ligands are then able to engage their receptors on target cells (32). This interaction leads to assembly at the cell surface membrane of a multiprotein signaling complex. Following association of FADD (Fas-associated protein with death domain) and other adapter proteins with the cytoplasmic domain of this signaling complex, caspase 8 is recruited and activated and in turn initiates a cascade of caspase and mitochondrial apoptogenic protein activation that leads to target cell apoptosis (63,64). THE ROLE OF THE HEPATOCYTE IN IMMUNE-MEDIATED LIVER INJURY One of the unique aspect of apoptotic death mechanisms such as those induced by cytotoxic lymphocytes is that the target cell plays a role in determining the mechanism and timing of its demise. Thus, susceptibility to FasL, TNF or TRAIL induced death is not only determined by the level of cell surface expression of death receptors on target cells but also by the level of target cell expression of proapoptogenic proteins such as BH3 interacting domain death protein (BID) and of intracellular inhibitors of the caspase and BID family proteins that execute apoptosis (65,66). Expression of serine protease inhibitors (serpins) that selectively inhibit granzyme B or other granzymes within the cytosol of target cells also regulates induction of apoptosis by the granule exoctyosis pathway of CTL and NK effector function (67–69). As detailed in Table 2, the sensitivity of hepatocytes to various cytotoxic lymphocyte effector TABLE 2 Sensitivity of Hepatocytes to Lymphocyte Cytotoxic Effector Mechanisms Effector mechanism Fas ligand/Fas TRAIL/death receptor 5 PerforinCgranzymes

Normal hepatocytes

Virally infected hepatocytes

Sensitive Resistant Sensitive

Sensitive Sensitive Resistant

120

Thiele

mechanisms is related in part to the presence or absence of viral infection and the cytokine responses that such infection generates (62,66–68). While resting hepatocytes are resistant to TRAIL induced apoptosis, viral infected hepatocytes become sensitive to this death receptor-mediated mechanism of immune-mediated cell injury (65,70,71). In contrast, while HBV transgenic murine hepatocytes appear sensitive to ex vivo activated HBV specific CTL killing mediated by the perforin and granzyme dependent granule exocytosis pathway (72), hepatocytes infected by viruses that elicit strong innate immune responses are induced to express high levels of the granzyme B inhibitors, protease inhibitor 9 (PI-9)/serine protease inhibitor 6 (SPI-6) (69) and become highly resistant to perforin and granzyme dependent cytotoxicity mechanisms (73,74). Both normal and virally infected hepatocytes exhibit sensitivity to TNF induced apoptosis (74–77). However, the principal mechanism of immune-mediated hepatocellular injury in most murine experimental models is engagement of hepatocyte Fas receptors by FasL expressed by CTL or NK cells (7,74,76,78–81). When normal mice are infused with a proapoptogenic anti-Fas monoclonal antibody, extensive hepatocellular apoptosis and acute liver failure evolves under conditions in which minimal Fas-mediated injury is apparent in other organs (80). Virally infected hepatocyte also remain more sensitive to this cytotoxic lymphocyte effector pathway than to any of the other well characterized CTL or NK effector mechanisms (74). Indeed in animal models of acetaminophen induced hepatotoxicity in which NK and NKT cells mediated increased levels of liver injury, Fas deficient mice exhibit protection from liver injury (7) thereby implying a major role for FasL/Fas effector mechanisms in NK and NKT cell-mediated DILI.

SUMMARY Drug-induced immune-mediated hepatic injury is a fortunately rare but potentially severe form of liver injury that arises in an idiosyncratic manner. Reactive metabolites derived during intrahepatic metabolism appear to play a central role in eliciting some forms of immunemediated DILI. In addition, an evolving body of literature has also implicated triggering of innate NK and NKT cell responses as another independent pathway towards induction of immune injury that amplifies directly hepatotoxic drug reactions. It remains to be determined whether such NK and NKT cell activation also serve as an important component of the innate immune response required to initiate adaptive immunity and the full spectrum of immunoallergic drug reactions observed in the small subset of patients exhibiting the most severe forms of immune-mediated DILI. REFERENCES 1. Liu ZX, Kaplowitz N. Immune-mediated drug-induced liver disease. Clin Liver Dis 2002; 6:755–74. 2. Seguin B, Uetrecht J. The danger hypothesis applied to idiosyncratic drug reactions. Curr Opin Allergy Clin Immunol 2003; 3:235–42. 3. Bissell DM, Gores GJ, Laskin DL, Hoofnagle JH. Drug-induced liver injury: mechanisms and test systems. Hepatology 2001; 33:1009–13. 4. Mieli-Vergani G, Vergani D, Tredger JM, Eddleston AL, Davis M, Williams R. Lymphocyte cytotoxicity to halothane altered hepatocytes in patients with severe hepatic necrosis following halothane anaesthesia. J Clin Lab Immunol 1980; 4:49–51. 5. Vergani D, Mieli-Vergani G, Alberti A, Neuberger J, Eddleston AL, Davis M, Williams R. Antibodies to the surface of halothane-altered rabbit hepatocytes in patients with severe halothane-associated hepatitis. N Engl J Med 1980; 303:66–71. 6. Bonierbale E, Valadon P, Pons C, Desfosses B, Dansette PM, Mansuy D. Opposite behaviors of reactive metabolites of tienilic acid and its isomer toward liver proteins: use of specific anti-tienilic acid– protein adduct antibodies and the possible relationship with different hepatotoxic effects of the two compounds. Chem Res Toxicol 1999; 12:286–96. 7. Liu ZX, Govindarajan S, Kaplowitz N. Innate immune system plays a critical role in determining the progression and severity of acetaminophen hepatotoxicity. Gastroenterology 2004; 127:1760–74. 8. Armeanu S, Bitzer M, Lauer UM, et al. Natural killer cell-mediated lysis of hepatoma cells via specific induction of NKG2D ligands by the histone deacetylase inhibitor sodium valproate. Cancer Res 2005; 65:6321–9.

Immunology of Drug-Induced Liver Injury

121

9. Njoku DB, Greenberg RS, Bourdi M, Borkowf CB, Dake EM, Martin JL, Pohl LR. Autoantibodies associated with volatile anesthetic hepatitis found in the sera of a large cohort of pediatric anesthesiologists. Anesth Analg 2002; 94:243–9. 10. Rehermann B, Nascimbeni M. Immunology of hepatitis B virus and hepatitis C virus infection. Nat Rev Immunol 2005; 5:215–29. 11. Thiele DL. Autoimmune hepatitis. Clin Liver Dis 2005; 9:635–46. 12. Wagner M, Trauner M. Transcriptional regulation of hepatobiliary transport systems in health and disease: implications for a rationale approach to the treatment of intrahepatic cholestasis. Ann Hepatol 2005; 4:77–99. 13. Geier A, Dietrich CG, Voigt S, et al. Cytokine-dependent regulation of hepatic organic anion transporter gene transactivators in mouse liver. Am J Physiol Gastrointest Liver Physiol 2005; 289:6831–41. 14. Higuchi H, Bronk SF, Takikawa Y, Werneburg N, Takimoto R, El-Deiry W, Gores GJ. The bile acid glycochenodeoxycholate induces trail-receptor 2/DR5 expression and apoptosis. J Biol Chem 2001; 276:38610–8. 15. Faubion WA, Guicciardi ME, Miyoshi H, et al. Toxic bile salts induce rodent hepatocyte apoptosis via direct activation of Fas. J Clin Invest 1999; 103:137–45. 16. Thimme R, Wieland S, Steiger C, Ghrayeb J, Reimann KA, Purcell RH, Chisari FV. CD8(C) T cells mediate viral clearance and disease pathogenesis during acute hepatitis B virus infection. J Virol 2003; 77:68–76. 17. Chisari FV, Ferrari C. Hepatitis B virus immunopathogenesis. Annu Rev Immunol 1995; 13:29–60. 18. Ando K, Moriyama T, Guidotti LG, et al. Mechanisms of class I restricted immunopathology. A transgenic mouse model of fulminant hepatitis. J Exp Med 1993; 178:1541–54. 19. Kilgore NE, Ford ML, Margot CD, Jones DS, Reichardt P, Evavold BD. Defining the parameters necessary for T-cell recognition of ligands that vary in potency. Immunol Res 2004; 29:29–40. 20. Wang S, Chen L. T lymphocyte co-signaling pathways of the B7-CD28 family. Cell Mol Immunol 2004; 1:37–42. 21. Matzinger P. An innate sense of danger. Semin Immunol 1998; 10:399–415. 22. Chandok MR, Farber DL. Signaling control of memory T cell generation and function. Semin Immunol 2004; 16:285–93. 23. Bourdi M, Tinel M, Beaune PH, Pessayre D. Interactions of dihydralazine with cytochromes P4501A: a possible explanation for the appearance of anti-cytochrome P4501A2 autoantibodies. Mol Pharmacol 1994; 45:1287–95. 24. Manns MP, Obermayer-Straub P. Cytochromes P450 and uridine triphosphate-glucuronosyltransferases: model autoantigens to study drug-induced, virus-induced, and autoimmune liver disease. Hepatology 1997; 26:1054–66. 25. Maria VA, Victorino RM. Diagnostic value of specific T cell reactivity to drugs in 95 cases of drug induced liver injury. Gut 1997; 41:534–40. 26. Hofer T, Becker EW, Weigand K, Berg PA. Demonstration of sensitized lymphocytes to trimethoprim/sulfamethoxazole and ofloxacin in a patient with cholestatic hepatitis. J Hepatol 1992; 15:262–3. 27. Tsutsui H, Terano Y, Sakagami C, Hasegawa I, Mizoguchi Y, Morisawa S. Drug-specific T cells derived from patients with drug-induced allergic hepatitis. J Immunol 1992; 149:706–16. 28. Robin MA, Le Roy M, Descatoire V, Pessayre D. Plasma membrane cytochromes P450 as neoantigens and autoimmune targets in drug-induced hepatitis. J Hepatol 1997; 1:23–30. 29. Groothuis T, Neefjes J. The ins and outs of intracellular peptides and antigen presentation by MHC class I molecules. Curr Top Microbiol Immunol 2005; 300:127–48. 30. Eisenlohr LC, Rothstein JL. Antigen processing and presentation. Cancer Treat Res 2005; 123:3–36. 31. Pichler WJ. Modes of presentation of chemical neoantigens to the immune system. Toxicology 2002; 182:49–54. 32. Burkhart C, Britschgi M, Strasser I, Depta JP, von Greyerz S, Barnaba V, Pichler WJ. Non-covalent presentation of sulfamethoxazole to human CD4CT cells is independent of distinct human leucocyte antigen-bound peptides. Clin Exp Allergy 2002; 32:1635–43. 33. Orange JS, Ballas ZK. Natural killer cells in human health and disease. Clin Immunol 2006; 118:1–10. 34. O’Connor GM, Hart OM, Gardiner CM. Putting the natural killer cell in its place. Immunology 2006; 117:1–10. 35. Mercer JC, Ragin MJ, August A. Natural killer T cells: rapid responders controlling immunity and disease. Int J Biochem Cell Biol 2005; 37:1337–43. 36. Vermijlen D, Luo D, Froelich CJ, et al. Hepatic natural killer cells exclusively kill splenic/blood natural killer-resistant tumor cells by the perforin/granzyme pathway. J Leukoc Biol 2002; 72:668–76. 37. Ashton-Rickardt PG. The granule pathway of programmed cell death. Crit Rev Immunol 2005; 25:161–82. 38. Kagi D, Vignaux F, Ledermann B, et al. Fas and perforin pathways as major mechanisms of T cellmediated cytotoxicity. Science 1994; 265:528–30.

122

Thiele

39. Braun MY, Lowin B, French L, Acha-Orbea H, Tschopp J. Cytotoxic T cells deficient in both functional fas ligand and perforin show residual cytolytic activity yet lose their capacity to induce lethal acute graft-versus-host disease. J Exp Med 1996; 183:657–61. 40. Russell JH, Ley TJ. Lymphocyte-mediated cytotoxicity [Review]. Annu Rev Immunol 2002; 20:323–70. 41. Podack ER, Young JD, Cohn ZA. Isolation and biochemical and functional characterization of perforin 1 from cytolytic T-cell granules. Proc Natl Acad Sci USA 1985; 82:8629–33. 42. Nakajima H, Park HL, Henkart PA. Synergistic roles of granzymes A and B in mediating target cell death by rat basophilic leukemia mast cell tumors also expressing cytolysin/perforin. J Exp Med 1995; 181:1037–46. 43. Motyka B, Korbutt G, Pinkoski MJ, et al. Mannose 6-phosphate/insulin-like growth factor II receptor is a death receptor for granzyme B during cytotoxic T cell-induced apoptosis. Cell 2000; 103:491–500. 44. Trapani JA, Sutton VR, Thia KY, et al. A clathrin/dynamin- and mannose-6-phosphate receptorindependent pathway for granzyme B-induced cell death. J Cell Biol 2003; 160:223–33. 45. Metkar SS, Wang B, Aguilar-Santelises M, et al. Cytotoxic cell granule-mediated apoptosis: perforin delivers granzyme B-serglycin complexes into target cells without plasma membrane pore formation. Immunity 2002; 16:417–28. 46. Kam CM, Hudig D, Powers JC. Granzymes (lymphocyte serine proteases): characterization with natural and synthetic substrates and inhibitors. Biochim Biophys Acta 2000; 1477:307–23. 47. Thornberry NA, Rano TA, Peterson EP, et al. A combinatorial approach defines specificities of members of the caspase family and granzyme B. Functional relationships established for key mediators of apoptosis. J Biol Chem 1997; 272:17907–11. 48. Darmon AJ, Nicholson DW, Bleackley RC. Activation of the apoptotic protease CPP32 by cytotoxic T-cell-derived granzyme B. Nature 1995; 377:446–8. 49. Martin SJ, Amarante-Mendes GP, Shi L, et al. The cytotoxic cell protease granzyme B initiates apoptosis in a cell-free system by proteolytic processing and activation of the ICE/CED-3 family protease, CPP32, via a novel two-step mechanism. EMBO J 1996; 15:2407–16. 50. Yang X, Stennicke HR, Wang B, et al. Granzyme B mimics apical caspases. Description of a unified pathway for trans-activation of executioner caspase-3 and -7. J Biol Chem 1998; 273:34278–83. 51. Talanian RV, Yang X, Turbov J, et al. Granule-mediated killing: pathways for granzyme B-initiated apoptosis. J Exp Med 1997; 186:1323–31. 52. Shi L, Chen G, MacDonald G, et al. Activation of an interleukin 1 converting enzyme-dependent apoptosis pathway by granzyme B. Proc Natl Acad Sci USA 1996; 93:11002–7. 53. Duan H, Orth K, Chinnaiyan AM, Poirier GG, Froelich CJ, He WW, Dixit VM. ICE-LAP6, a novel member of the ICE/Ced-3 gene family, is activated by the cytotoxic T cell protease granzyme B. J Biol Chem 1996; 271:16720–4. 54. Thomas DA, Du C, Xu M, Wang X, Ley TJ. DFF45/ICAD can be directly processed by granzyme B during the induction of apoptosis. Immunity 2000; 12:621–32. 55. Barry M, Heibein JA, Pinkoski MJ, Lee SF, Moyer RW, Green DR, Bleackley RC. Granzyme B shortcircuits the need for caspase 8 activity during granule-mediated cytotoxic T-lymphocyte killing by directly cleaving Bid. Mol Cell Biol 2000; 20:3781–94. 56. Sutton VR, Davis JE, Cancilla M, et al. Initiation of apoptosis by granzyme B requires direct cleavage of bid, but not direct granzyme B-mediated caspase activation. J Exp Med 2000; 192:1403–14. 57. Shresta S, MacIvor DM, Heusel JW, Russell JH, Ley TJ. Natural killer and lymphokine-activated killer cells require granzyme B for the rapid induction of apoptosis in susceptible target cells. Proc Natl Acad Sci USA 1995; 92:5679–83. 58. Shresta S, Graubert TA, Thomas DA, Raptis SZ, Ley TJ. Granzyme A initiates an alternative pathway for granule-mediated apoptosis. Immunity 1999; 10:595–605. 59. Beresford PJ, Zhang D, Oh DY, et al. Granzyme A activates an endoplasmic reticulum-associated caspase-independent nuclease to induce single-stranded DNA nicks. J Biol Chem 2001; 276:43285–93. 60. Johnson H, Scorrano L, Korsmeyer SJ, Ley TJ. Cell death induced by granzyme C. Blood 2003; 101:3093–101. 61. Kelly JM, Waterhouse NJ, Cretney E, Browne KA, Ellis S, Trapani JA, Smyth MJ. Granzyme M mediates a novel form of perforin-dependent cell death. J Biol Chem 2004; 279:22236–42. 62. Screpanti V, Wallin RP, Grandien A, Ljunggren HG. Impact of FASL-induced apoptosis in the elimination of tumor cells by NK cells. Mol Immunol 2005; 42:495–9. 63. Locksley RM, Killeen N, Lenardo MJ. The TNF and TNF receptor superfamilies: integrating mammalian biology. Cell 2001; 104:487–501. 64. Chen G, Goeddel DV. TNF-R1 signaling: a beautiful pathway. Science 2002; 296:1634–5. 65. Zender L, Hutker S, Mundt B, et al. NFkappaB-mediated upregulation of bcl-xl restrains TRAILmediated apoptosis in murine viral hepatitis. Hepatology 2005; 41:280–8. 66. Li S, Zhao Y, He X, et al. Relief of extrinsic pathway inhibition by the Bid-dependent mitochondrial release of Smac in Fas-mediated hepatocyte apoptosis. J Biol Chem 2002; 277:26912–20. 67. Bots M, Kolfschoten IG, Bres SA, et al. SPI-CI and SPI-6 cooperate in the protection from effector cellmediated cytotoxicity. Blood 2005; 105:1153–61.

Immunology of Drug-Induced Liver Injury

123

68. Bird PI. Regulation of pro-apoptotic leucocyte granule serine proteinases by intracellular serpins. Immunol Cell Biol 1999; 77:47–57. 69. Barrie MB, Stout HW, Abougergi MS, Miller BC, Thiele DL. Antiviral cytokines induce hepatic expression of the granzyme B inhibitors, proteinase inhibitor 9 and serine proteinase inhibitor 6. J Immunol 2004; 172:6453–9. 70. Zheng SJ, Wang P, Tsabary G, Chen YH. Critical roles of TRAIL in hepatic cell death and hepatic inflammation. J Clin Invest 2004; 113:58–64. 71. Zhang HG, Xie J, Xu L, et al. Hepatic DR5 induces apoptosis and limits adenovirus gene therapy product expression in the liver. J Virol 2002; 76:5692–700. 72. Nakamoto Y, Guidotti LG, Pasquetto V, Schreiber RD, Chisari FV. Differential target cell sensitivity to CTL-activated death pathways in hepatitis B virus transgenic mice. J Immunol 1997; 158:5692–7 (erratum appears in J Immunol 1999 July 15; 163(2):1092). 73. Tay CH, Welsh RM. Distinct organ-dependent mechanisms for the control of murine cytomegalovirus infection by natural killer cells. J Virol 1997; 71:267–75. 74. Kafrouni MI, Brown GR, Thiele DL. Virally infected hepatocytes are resistant to perforin-dependent CTL effector mechanisms. J Immunol 2001; 167:1566–74. 75. Guicciardi ME, Deussing J, Miyoshi H, et al. Cathepsin B contributes to TNF-alpha-mediated hepatocyte apoptosis by promoting mitochondrial release of cytochrome c. J Clin Invest 2000; 106:1127–37. 76. Liu ZX, Govindarajan S, Okamoto S, Dennert G. Fas- and tumor necrosis factor receptor 1-dependent but not perforin-dependent pathways cause injury in livers infected with an adenovirus construct in mice. Hepatology 2000; 31:665–73. 77. Elkon KB, Liu CC, Gall JG, et al. Tumor necrosis factor alpha plays a central role in immune-mediated clearance of adenoviral vectors. Proc Natl Acad Sci USA 1997; 94:9814–9. 78. Abougergi MS, Gidner SJ, Spady DK, Miller BC, Thiele DL. Fas and TNFR1, but not cytolytic granuledependent mechanisms, mediate clearance of murine liver adenoviral infection. Hepatology 2005; 41:97–105. 79. Chirmule N, Moscioni AD, Qian Y, Qian R, Chen Y, Wilson JM. Fas–Fas ligand interactions play a major role in effector functions of cytotoxic T lymphocytes after adenovirus vector-mediated gene transfer. Hum Gene Ther 1999; 10:259–69. 80. Ogasawara J, Watanabe-Fukunaga R, Adachi M, et al. Lethal effect of the anti-Fas antibody in mice. Nature 1993; 364:806–9 (erratum appears in Nature 1993 October 7; 365(6446):568).

8

Mechanistic Role of Acyl Glucuronides Hilde Spahn-Langguth

German University in Cairo, New Cairo City, Egypt

Chunze Li

Merck & Co., Inc., West Point, Pennsylvania, U.S.A.

Leslie Z. Benet

Department of Biopharmaceutical Sciences, University of California-San Francisco School of Pharmacy, San Francisco, California, U.S.A.

INTRODUCTION Only a few years ago, it was generally recognized by pharmaceutical scientists that phase-II metabolites of drugs, such as acyl glucuronide conjugates, are readily excreted following their formation in the body and that these metabolites are neither active nor reactive. We and others have shown that this is not generally true (1,2). Acyl glucuronide conjugates—as opposed to ether and other glucuronides—are, in fact, reactive metabolites, capable of undergoing hydrolysis, intramolecular acyl migration, and covalent binding to proteins, both in vitro and in vivo. This newly recognized reactivity has an important, but still poorly defined, bearing on biological distribution and metabolism of a widely prescribed class of drugs. It may also be directly associated with the perplexing toxicity of many carboxylic acid–containing drugs (3,4). It is striking that out of 47 drugs withdrawn from U.S., British, and Spanish markets from 1964 through 1993 owing to severe toxicity (3,4), 10 are carboxylic acids, all of which are metabolized by humans to acyl glucuronides. Conjugation with glucuronic acid is the major route for the elimination of xenobiotic and endogenous compounds with a carboxylic acid function (1). These acyl-linked glucuronides are chemically reactive electrophiles (2) and have been shown to be susceptible to hydrolysis, transacylation by methanol (5), ammonia (6), ethanethiol (7), and glutathione (8,9), and reaction with chemical nucleophiles such as 4-(p-nitrobenzyl)pyridine (10). Acyl-linked glucuronides have been observed to undergo intramolecular nucleophilic substitution reactions with the hydroxyl groups on the glucuronic acid moiety, resulting in the intramolecular migration of the xenobiotic moiety from the 1-b-O-position to the 2-, 3-, and 4-position of the glucuronic acid ring. Such intramolecular acyl migration and hydrolysis may occur during biological sample handling and, of particular relevance, also under the pH and temperature conditions found in vivo. Earlier studies not employing correct sample stabilization procedures yielded inaccurate measures of the pharmacokinetics of carboxylic acid–containing drugs as well as their glucuronides. In addition to hydrolysis and intramolecular acyl migration, acyl glucuronides also readily react with the nucleophiles on proteins, both in vivo and in vitro. Covalent modification The author of this chapter has relationships with the following corporations: Founder & Board Member: AvMAX, INC. Corporate and Scientific Advisory Boards: CNSBio Pty., Ltd. Comentis (Formerly Athenagen, Inc.). Eastman Drug Delivery Sciences. Gelmed Sciences. Hurel Corp. Impax Laboratories Inc. Institute for One World Health. Josman Laboratories. Life Cycle Pharma A/S. Limerick Neurosciences. Lipocine, Inc. Millennium Drug Safety & Disposition. Panacea Biotec. Savient Pharamaceuticals. Silico Insights. TRF Pharma. UMD Inc. Active Consultancies and Confidentiality Agreements: Aditech Pharma AB. Advanced Cardiovascular Systems. Allergan. Amgen. Boehringer Ingelheim. Canada Customs and Revenue Agency. Exelixis. Forsight Labs. Genentech. Incyte. Johnson & Johnson PR&D. McNeil Consumer & Speciality Pharma. Merck Research Laboratories. Mutual Pharmaceuticals. Nordic Biotech. Osteologix. Rigel, Inc. Sanofi aventis. Schering-Plough. Virochem Pharma.

126

Spahn-Langguth et al.

of cellular proteins by acyl glucuronides has been suggested to mediate the rare, but potentially fatal, idiosyncratic hypersensitivity associated with carboxylic acids (1). The mechanisms responsible for the initiation of such immune-type toxic side effects, including anaphylaxis and drug-induced liver injury (11), remain poorly understood. A current explanation for the different types of hypersensitivity reactions caused by drugs or other small molecules is the “hapten hypothesis” (12). Small foreign molecules such as drugs are not immunogenic by themselves, but may become so after covalent attachment to endogenous carrier proteins (such as albumin) that facilitate recognition by the immune system. However, more recent investigations have revealed covalent binding to cellular targets as well. As a result of exposure to high acyl glucuronide concentrations, the liver appears to be the major target of adduct formation. Intrahepatic targets are, e.g., uridine 5 0 -diphosphate-glucuronosyltransferases (UGTs) and tubulin. Altered protein activity and, potentially, a contribution to hepatotoxicity may result. Any change in acyl glucuronide disposition, such as the rate of glucuronide formation or membrane transport, may affect the extent of exposure and adduct formation. In general, the extent of the exposure of the organism or an organ to the potential immunogen or cellular toxicant is one possible determinant for the occurrence of adverse reactions, as depicted schematically in Figure 1. This chapter will focus on the chemical reactivity of acyl glucuronides. We will summarize the general properties and newer aspects of formation and degradation of acyl glucuronides, and their reversible and irreversible binding to plasma and tissue proteins in vitro and in vivo. The selective modification of tissue proteins by acyl glucuronides and their potential drug-induced organ toxicity (especially liver) will also be discussed. OVERVIEW ON MAJOR TYPES OF CHEMICAL REACTIVITY OF ACYL GLUCURONIDES Because of the susceptibility of the acyl group, which links the aglycone and the glucuronic acid moieties to nucleophilic substitution reactions, the biosynthetically formed b-1-O-acyl glucuronides General scheme for reactive metabolites Drug

Excretion

Reactive metabolite(s)

Protein adduct

Immunogen

Antibody

Stable metabolites

Direct toxicity

For acyl glucuronides

Aglycone Glucuronide and isomer formation

Other metabolites Excretion

Glucuronide and isomer cleavage

Acyl glucuronide and isomers Adduct formation via Adduct cleavage acylation or glycation Protein adduct

Antigen Immunemediated response

Antibody

FIGURE 1 General interrelationship between toxicity (including immune response; hapten hypothesis ) and disposition of a drug with reactive metabolites (left ). Scheme focusing on major processes relevant for the disposition of reactive acyl glucuronides (right ), for which an antigen character of adducts was shown and antibodies have been detected, also in vivo, in some cases. Acyl glucuronides are generating increasing interest as potential mediators of hypersensitivity reactions and cellular toxicity. Source: From Ref. 2.

127

Mechanistic Role of Acyl Glucuronides HOOC HO HO

is

lys

dro

HOOC HO HO

Hy O

R

O 1-β-O -acyl glucuronide

HOOC HO HO

HO HO

R

HO HO R

O 1-α-O -acyl glucuronide

OH

HOOC

O O

R

R

O aglycone

OH

O 2-β-O -acyl glucuronide

HOOC

O OH O

OH glucuronic acid HOOC

O OH

HO

O

HO O R

O

HOOC HO OH

2-α-O -acyl glucuronide

OH

OH

O 3-β-O -acyl glucuronide

O O

R O

R

O

O

HOOC O HO

O

OH OH

3-α-O -acyl glucuronide

OH

OH

4-β-O -acyl glucuronide

R O

O

O

HOOC O HO

O OH

OH

4-α-O -acyl glucuronide

FIGURE 2 Hydrolysis and isomer formation through acyl migration and anomerization. Acyl glucuronides are hydrolyzed (with water being the attacking nucleophile) under back-formation to aglycone and glucuronic acid. When the hydroxyl group on glucuronic acid (e.g., C-2) is the attacking nucleophile, the aglycone starts to migrate from, e.g., C-1 to C-2 in the glucuronic acid (acyl migration). Migration may occur from C-1 to C-2, C-2 to C-3, etc. The reaction, which is catalyzed by a hydroxide ion, is reversible. Furthermore, positional isomers of glucuronides can ring-open and mutarotate yielding a- in addition to b-anomers, although it was believed not to occur, also back-formation of C-1 isomers appears possible, at least to some extent. All anomers may undergo hydrolysis yielding glucuronic acid and aglycone.

are subject to different types of nucleophilic attacks on the electron-deficient carbonyl carbon atom (Fig. 2). The outcome of these reactions will depend on the nature of the attacking nucleophile. Transacylation and Hydrolysis 1. If the nucleophile is water, the reaction is actually a hydrolysis reaction, yielding the free aglycone and glucuronic acid (Fig. 2). Hydrolysis can be both spontaneous (buffer solution) and enzymatic (biofluid) hydrolysis. 2. When the neighboring hydroxyl group on glucuronic acid (on C-2) is the attacking nucleophile, the result is that the aglycone actually migrates from the anomeric carbon C-1 to C-2 and the process is known as an (intramolecular) acyl migration, a rearrangement reaction. The reaction is catalyzed by hydroxide ion. Naturally, the migration reaction can proceed from C-1 to C-2, C-3, and C-4, and in both directions (reversible). Even the backmigration from a-2 to a-1 and b-2 to b-1, which has been previously believed to be impossible, has indeed been found to occur to a minor extent. Unlike the biosynthetic 1-Oacyl glucuronide itself, the glucuronide isomers readily ring-open and mutarotate giving a- and b-anomers, which can then undergo further acyl-migration reactions (Fig. 2). 3. Intermolecular transacylation with, e.g., proteins. This is the intermolecular analogue of reaction (2). Here the attacking group is some nucleophilic function on the protein, e.g., –OH, –NH2, –SH.

Glycation 1. Formation of protein adducts from the rearrangement isomers via a Schiff’s base reaction between the open-chain aldose and an amino group. Here again, the attacking group is an amino group on, e.g., a protein, but the attack is on the open-chain aldose of a rearranged glucuronide, and not on the pyranose form. The final product of this reaction, following an Amadori rearrangement, is a ketoamine, which is chemically stable toward hydrolysis. A complicating factor in analytical and kinetic studies of such acyl glucuronides is the fact that all of these reactions are in reality coupled together; so for example, the rearranged

128

Spahn-Langguth et al.

glucuronides are potential substrates for a Schiff’s base reaction. Also, any of these rearranged glucuronides (reaction 2) can be attacked by water and hydrolyzed (reaction 1); an intermolecular transacylation (reaction 3) can take place with any of the intramolecular rearrangement products of reaction 2 and so on. Although acyl glucuronides are chemically reactive electrophilic metabolites, they are relatively stable when compared with other classes of reactive metabolites that do not survive the environment where they are formed, as opposed to acyl glucuronides that reach the systemic circulation and are excreted into urine and bile. BIOCHEMICAL ASPECTS OF ACYL GLUCURONIDATION Conjugation with glucuronic acid is the major route for the biotransformation and elimination of carboxylic acid–containing drugs (1,2). Under normal conditions, acyl glucuronides are formed primarily in the liver and excreted predominantly through the urine in humans (Fig. 3). The formation of acyl glucuronides is catalyzed by a membrane-bound enzyme, UGT (EC 2.4.1.17), which transfers the glucuronic acid from uridine 5 0 -diphosphate (UDP)-glucuronic acid (UDPGA) to the carboxyl group of the aglycone, resulting in ester-linked glucuronides. The mechanism of the reaction catalyzed by UGT is an SN2-type reaction. The anomeric center undergoes inversion during the enzymatic transfer of a-D-glucuronic acid in UDPGA to the acceptor substrate, resulting in the formation of the b-configuration (Fig. 3). UGT is a family of closely related isoenzymes mainly located in the endoplasmic reticulum and exhibiting different, but overlapping, substrate specificities (13). Studies with Gunn rats, which are genetically deficient in bilirubin glucuronidation, revealed that the isoform(s) involved in glucuronidation of carboxylic acid–containing drugs was different from those responsible for

1

COOH O 5 4 OH 2 3 OH OOH O

O P O O O P O O

Liver

NH O

N

O

OHOH UDP-glucuronic acid (UDPGA, alpha) O C UGT OH R UDP

R

O C

OH

Beta-glucuronidase

COOH O 5 4 OH R O 1 C 2 3 OH HO O

Bile

Intestines

1-O-acyl glucuronide (beta) Feces Plasma

Urine

FIGURE 3 The formation and elimination of b-1-O-acyl glucuronide. Abbreviations: UDP, uridine 5 0 -diphosphate; UDPGA, uridine 5 0 -diphosphate-glucuronic acid; UGT, uridine 5 0 -diphosphate-glucuronosyltransferase.

Mechanistic Role of Acyl Glucuronides

129

bilirubin acyl conjugation, at least for the arylpropionic acids (14). Recently, several human liver UGTs have been cloned and the cDNAs expressed in heterologous cell lines. This technological advance has allowed the assessment of the functional specificity of these UGTs. Of these, UGT1A3, UGT1A9, and UGT2B7 appear to be key isoforms in the glucuronidation of a wide range of xenobiotic carboxylic acids (13). In expressed enzyme systems, major substrates for UGT1A3 were clofibric acid, fenoprofen, ibuprofen (R/S ratio in activityZ1.6), ketoprofen, naproxen, valproic acid, and ciprofibric acid. Diflunisal, fenoprofen, ibuprofen, ketoprofen, naproxen, 4-aminosalicylic acid, bumetanide, and furosemide were glucuronidated by UGT1A9, while zomepirac, benoxaprofen, clofibric acid, tiaprofenic acid, fenoprofen, ibuprofen, ketoprofen, valproic acid, and naproxen were glucuronidated by UGT2B7 (15). Bilirubin glucuronidation occurs mainly via UGT1A1. The major organ for glucuronidation is the liver, and UGT1A3 and UGT1A9 are highly expressed in the liver. However, in the liver, deconjugating enzymes are present as well, such as b-glucuronidase (a lysosomal enzyme) and carboxylesterases, which codetermine the net hepatic glucuronidation rate since they hydrolyze acyl glucuronides. SYNTHESIS, ISOLATION, AND CHARACTERIZATION OF ACYL GLUCURONIDES Biosynthesis of Acyl Glucuronides Owing to the difficulty and expense of synthesizing the labile acyl glucuronides by chemical methods (1,16), alternative biosynthetic methods are preferable for the preparation of acyl glucuronides. Acyl glucuronide metabolites may be synthesized in vitro using crude enzyme mixtures derived from animal tissue (e.g., liver microsomes) or using in vitro purified enzyme systems (e.g., immobilized enzymes). Many investigators have successfully utilized these two methods to obtain small quantities of acyl glucuronides. Ruelius et al. (17) synthesized oxaprozin glucuronide by combining labeled oxaprozin aglycone and UDPGA using a crude enzyme preparation from a rhesus monkey liver homogenate, which resulted in 7.5% conversion of the parent aglycone to 1-O-acyl glucuronide. A similar incubation with sheep liver microsomes by the addition of inhibitors of hydrolytic enzymes (e.g., esterases and b-glucuronidase) was utilized to prepare zomepirac- [14C]glucuronide, with an improved yield of 10.1% (18). Grubb et al. (19) also successfully biosynthesized fenofibric and clofibric [14C]glucuronides by incubation of aglycone and radiolabeled UDPGA with rabbit and human hepatic microsomes, leading to 14% and 36% conversion of UDP[14C]GA to the glucuronides, respectively. High yield of tolmetin glucuronide (O60% of aglycone) with freshly prepared sheep liver microsomes (20 mg protein/mL) has been achieved in our laboratory (Ojingwa and Benet, unpublished results) with the following concentrations: 1 mM aglycone, 10 mM UDPGA, 10 mM MgCl2 in 100 mM Tris–HCl buffer (pH 6.9) containing the enzyme inhibitor phenylmethylsulfonyl fluoride (2 mM) and 1,4-saccharic acid lactone (20 mM), and the detergent Triton X-100 (0.2%). Immobilized UGT, covalently bound to cyanogen bromide-activated agarose (20) or Sepharose (21), has been used for the preparation of a variety of glucuronide conjugates. Using partially purified immobilized liver UGT, van Breemen and Fenselau (22) were able to synthesize a series of aglycone-labeled 1-O-acyl glucuronides. The yields obtained were benoxaprofen glucuronide, 2%; clofibric glucuronide, 3%; flufenamic glucuronide, 7%; and indomethacin glucuronide, 28%. Using similar immobilized UGT preparations, Bradow et al. (23) also synthesized small quantities of 1-O-acyl glucuronides of salicylic acid, S-benoxaprofen, and D9-11-carboxy-tetrahydrocannabinol. The above two in vitro enzymatic biosynthetic methods are fairly useful for the preparation of small quantities of acyl glucuronides, particularly for the glucuronides radiolabeled in either the drug or glucuronic acid moieties (1). However, for preparation of larger quantities, in vivo biosynthesis has significant advantages. Mostly, this has been accomplished by harvesting glucuronides from the urine of drug-dosed humans or animals. In 1981, the isolation of probenecid acyl glucuronide from urine by ethyl acetate extraction and HPLC purification was described by Eggers and Doust (24). Such extraction and purification methods from human urine in subjects dosed with the parent drugs have been successfully

130

Spahn-Langguth et al.

used to prepare the acyl glucuronides of a number of carboxylic acids, including zomepirac (25), tolmetin (26), diflunisal (27), beclobric acid (28), carprofen (29), etodolac (30), suprofen (31), ibuprofen (32), furosemide (33), and mefenamic acid (34). Relatively large-scale preparations of clofibric glucuronide and fenofibric glucuronides (19) were also achieved from the urine of rabbits dosed with the corresponding carboxylic acids, and of salicylic acid (34), valproic acid (35), and zomepirac (36) from rat urine and bile. A shortcoming of acyl glucuronides isolated from human or animal excretory fluids may be the fact that they may represent mixtures of different acyl glucuronides when more than one acyl glucuronide is present due to phase-I metabolism of the aglycone. Examples of carboxylic acid drugs undergoing sequential phase-I–phase-II metabolism are diclofenac (4 0 -hydroxy diclofenac acyl glucuronide), etodolac (6-hydroxy etodolac acyl glucuronide), fenoprofen (4-hydroxy fenoprofen acyl glucuronide), naproxen (desmethyl naproxen acyl glucuronide), and gemfibrozil (M1–M4 acyl glucuronides) (2,37). Separation and Quantification of Acyl Glucuronides Two classes of analytical methods, indirect and direct, are available to quantify acyl glucuronides. The indirect methods, which were used almost exclusively before the early 1980s, involve enzymatic and alkaline hydrolysis of the ester bond, leading to release of the parent aglycone. Since only 1-b-O-acyl glucuronide can be hydrolyzed by b-glucuronidase and both 1-O-acyl glucuronides and its b-glucuronidase-resistant isomers are labile in alkaline solution, a differentiation between 1-b-O-acyl glucuronide and its positional isomers is possible by fractional hydrolysis. With the improvement of HPLC techniques, the development of direct HPLC methods for glucuronides allowed investigators to chromatographically separate the different components in the glucuronide fraction and then to study the chemical properties (such as stability) of acyl glucuronides under different conditions. Compared to indirect methods, the direct procedures are more convenient and sensitive. Sinclair and Caldwell (38) reported one of the first HPLC separations of different glucuronide isomers of clofibric acid. Zomepirac, its b-1-O-acyl glucuronide, and four isomers were assayed on a reversed-phase (C18) HPLC column utilizing a mixture of methanol and sodium acetate buffer as the mobile phase (25). A similar HPLC condition, including an ion-pairing reagent, tetrabutylammonium, was successfully applied to tolmetin, allowing simultaneous quantification of all the glucuronide conjugates (39). Because of the various forms of isomerism, the analytical problems may, depending on the nature of the glucuronide conjugate, become rather complex. Isomerization is possible not only via intramolecular rearrangement by acyl migration (2-, 3-, and 4-O-isomers), but also via isomerization of the sugar group, yielding furanose as opposed to pyranose structures. Except for the C-l position (b-1-O-acyl glucuronides), a- and b-anomeric forms may occur in addition to open-chain forms and lactones (40). As a consequence, numerous isomers of the enzymatically formed b-1-O-acyl glucuronide may be found. Several authors have reported the presence of more than three isomers in addition to the b-1-O-acyl glucuronides. Hansen-Moller et al. (41) isolated and identified the a- and b-anomers of three positional isomers of diflunisal glucuronide. Dickinson et al. (42) separated six structural isomers of salicylic acid glucuronide and speculated on the presence of a- and b-anomers of the three 2-, 3-, and 4-O-isomers. Eight isomeric peaks of furosemide glucuronide other than 1-O-acyl glucuronide and the free acid were observed utilizing gradient HPLC by our group (43). Resolution of the diastereomeric (R)- and (S)-glucuronides of 2-arylpropionic acids could also be achieved by HPLC on octadecylsilane (ODS) stationary phases. We described the resolution of diastereomeric naproxen glucuronides as well as the glucuronides of various other 2-arylpropionic acids [e.g., flunoxaprofen (44,45), benoxaprofen (45,46), carprofen (29), and fenoprofen (47)] on Ultrasphere ODS using a gradient of acetonitrile (ACN) in 9 or 10 mM tetrabutylammonium (TBA) buffer (pH 2.5) with elution order (S)- before (R)-glucuronide. Fournel-Gigleux et al. (48) used a Lichrosorb Hibar RT column and ACN/trifluoroacetic acid (TFA)/water (19:0.04:81) to resolve the diastereomeric conjugates of 2-phenylpropionic acid (elution order (R)- before (S)-glucuronide). el Mouelhi et al. (49) employed different ACN/ ammonium acetate or phosphate buffer systems to resolve the stereoisomeric conjugates of

131

Mechanistic Role of Acyl Glucuronides

TABLE 1 HPLC Conditions for Selected Acyl Glucuronides of Xenobiotic Carboxylic Acids Compound

Column

Running buffer

References

254 254

ODS ODS

23 49

Diclofenac (C)

313/365 290/365 245 226 226 280

ODS ODS ODS ODS ODS ODS

Diflunisal (C)

226

ODS

Etodolac Fenofibric acid

280 290

ODS Octyl

Fenoprofen (R/S) Flufenamic acid (C) Flunoxaprofen (R/S) Furosemide (C)

272 254 313/365 233/289

ODS ODS ODS ODS

Gemfibrozil (C) Ibufenac

284/316 214

Cyano ODS

Ibuprofen

214

ODS

Isoxepac Ketoprofen (R/S)

254 254

ODS ODS

Mefenamic acid

280

Octyl

Naproxen (R/S)

275/355

ODS

Oxaprozin (C)

280

Octyl

2-Phenylpropionic acid (R/S) Probenecid

254

ODS

ACN/0.05 M KH2PO4, pH 4.5 ACN/0.01 M phosphate buffer, pH 6.5 ACN/0.01 M TBA, pH 2.5 ACN/9 mM TBA ACN/water/TFA (40:60:0.04) MeOH/water/TFA (40:60:0.1) ACN/5 mM TBA MeOH/0.05 M ammonium acetate, pH 4.5 (50:50) MeOH/0.01 M Na2HPO4, pH 2.7, 4% (w/v) Na2SO4 (54:46) MeOH/0.01 M TFA (47:53 v/v) ACN/10 mM phosphate buffer, 5 mM TEA, pH 7.5 (45:55) ACN/10 mM TBA, pH 2.5 ACN/0.05 M KH2PO4, pH 4.5 ACN/0.01 M TBA, pH 2.5 ACN/0.08 M phosphoric acid (30:70) ACN/10 mM TBA, pH 3.5 MeOH/0.01 M TFA, pH 2.2 (55:45) MeOH/0.01 M TFA, pH 2.2 (57:43) ACN/phosphoric acid (0.2%) ACN/10 mM TBA, 1 mM potassium phosphate, pH 4.3 ACN/0.05 M ammonium acetate, pH 4.5 (30:70) ACN/66 mM ammonium acetate, pH 6 (25:75) ACN/0.05 M phosphate buffer (26:74) ACN/TFA/water (19:0.04:81)

254

Octyl

Salicylic acid (C)

240

ODS

Suprofen

295

ODS

Tolmetin (C)

313

Octyl

Zomepirac (C)

313

ODS

Benoxaprofen (C) Benoxaprofen (R/S) Carprofen (R/S) Clofibric acid (C)

Detection (nm)

MeOH/water/acetic acid (50:50:1), 40 mM TBA MeOH/0.1 M sodium phosphate buffer, pH 2.7 MeOH/0.01 M sodium acetate, pH 5.1 (37.5:62.5) MeOH/0.01 M TBA, 0.05 M sodium acetate, pH 4.5 MeOH/0.01 M sodium acetate, pH 5.1

46 53 50 38 19 54 55 30 19 47 23 44 33 56 32 32 57 52 34 58 17 48 59 42 31 39 25

Note: C denotes the simultaneous assay of isomeric conjugates resulting from acyl migration; R/S denotes the separation of diastereomeric glucuronides. Abbreviations: ACN, acetonitrile; MeOH, methanol; ODS, octadecylsilane; TBA, tetrabutylammonium; TEA, tetraethylammonium; TFA, trifluoroacetic acid. Source: Partially from Ref. 2.

naproxen, ibuprofen, and benoxaprofen on an Ultrasphere ODS column with elution order (S)- before (R)-conjugate for all compounds. HPLC separation of (R)- and (S)-carprofen glucuronides, using Lichrosorb RP18 column and ACN/water/TFA (40:60:0.04) system, were also reported by Georges et al. (50), with the elution order (R) before (S). Additional analytical studies with naproxen glucuronide were published by Buszewski et al. (51) and with ketoprofen glucuronides by Chakir et al. (52). The HPLC conditions for the separation of acyl glucuronides of various carboxylic acids are summarized in Table 1.

132

Spahn-Langguth et al.

Structural Characterization of Acyl Glucuronides General procedures for the structural elucidation of glucuronides were summarized by Heirwegh and Compernolle (60). Different analytical methods have been used to identify the structures of acyl glucuronides and their isomers, including mass spectrometry (MS) and nuclear magnetic resonance (NMR) spectrometry. Compernolle et al. (61) utilized gas chromatography/MS in their determination of the structures of bilirubin glucuronide isomers. The structures of furosemide glucuronide and its isomerization products were confirmed by negative-ion thermospray liquid chromatography/MS by Rachmel et al. (33). These scientists detected the abundant (MK1)2 ion at mass 505, the aglycone fragment at m/z 329, and the characteristic sugar fragment ion of mass 175 in the spectra of the b-1-O-acyl glucuronide and the isomers, whereas an ion at m/z 221 was noted only in the case of the b-1 conjugate. Liquid chromatography–electrospray ionization MS was used to differentiate glucuronides of the anti-HIV drug candidate PA-457 (62). This methodology has also been applied for other well-known drugs. The complex phase-I/phase-II metabolic scheme of rhein in different species was elucidated and among many different mono- and bis-conjugates, an acyl glucuronide was detected in rabbit and man, and mixed (ether-ester–bis) glucuronides in rabbits (63). For gemfibrozil and its oxidative metabolites (hydroxylated derivatives of the carboxylic acid as well as a bicarboxylic acid) Hermening confirmed that all conjugates are acyl glucuronides, including a bis-acyl glucuronide (64). Eggers and Doust (24) have used 13C-NMR studies to confirm the isomerization of probenecid glucuronide. Smith and Benet (65), using 1H-NMR, confirmed that the four fractions that could be separated by HPLC were positional isomers of zomepirac glucuronide. The structural assignments for flufenamic acid and (S)-benoxaprofen (23) were also confirmed by 1H-NMR. In 1988, Hansen-Moller et al. (41) described a simultaneous separation of the aand b-anomers of three positional isomers of diflunisal glucuronide, in addition to the b-1-Oacyl glucuronide. Using two-dimensional NMR spectrometry, they were able to identify these six different a- and b-anomers. Similarly, the a- and b-isomers of flufenamic acid glucuronide were also found by 1H-NMR and their structures were confirmed by a series of successive decoupling experiments (23). The available data suggest that anomerization is a general phenomenon for C-2–C-4 isomers of all acyl glucuronides. More recently, using stopped-flow HPLC–1H-NMR, Mortensen et al. (66) found evidence for the regeneration of the b-1-O-acyl glucuronide and the a-1-O-acyl glucuronide from C-2-isomers. STABILITY OF ACYL GLUCURONIDES Acyl glucuronides are generally less stable than other glucuronides (2). Hydrolysis and intramolecular acyl migration are two major reactions contributing to this instability. Despite the common elements of their chemical structure, acyl glucuronide reactivity can differ enormously from compound to compound [e.g., (62,67,68)]. Electronic, steric, and other factors were hypothesized to contribute to these differences. Hydrolysis of Acyl Glucuronides—The Problem of Sample Treatment Hydrolysis of an acyl glucuronide leads to regeneration of the pharmacologically active parent drug. Potential catalysts include hydroxide ion, b-glucuronidases, serum albumin, and esterases. Rates of hydrolysis are dependent on pH and temperature, with more rapid degradation of the enzymatically formed b-1-O-acyl glucuronide at higher pH, also at physiological pH, than at a more acidic level. Hydrolysis of an acyl glucuronide conjugate occurs readily in biological samples, for example, in urine and plasma, in vitro under laboratory conditions and during storage. The rate of chemical hydrolysis decreases significantly in cold and acidic conditions (pH 3–4), but hydrolysis may still occur slowly during freezing and especially during thawing (69). This may result in substantial increase in the concentration of the parent compound and may be responsible for some of the variation in the apparent extent of the amount excreted unchanged in the urine of some drugs as reported by

Mechanistic Role of Acyl Glucuronides

133

different investigators. Acyl glucuronides can undergo substantial hydrolysis to the parent aglycone in vivo and this may be due to enzymatic cleavage by b-glucuronidase or nonspecific esterases under physiological conditions. Degradation of the conjugates in bile and intestines will contribute to the enterohepatic recirculation of the parent compound (Fig. 3). The same catalysts that can hydrolyze the acyl glucuronide can also hydrolyze all the migration isomers with the one notable exception: b-glucuronidase hydrolyzes the b-1-O-acyl glucuronide, yet none of the isomers. Predictability of Hydrolysis Rates When the calculated overall hydrolysis rates were compared with the formation rates of aglycones, a linear trend was observed. Thus, the most reactive compounds are generally also the compounds with the highest rates of hydrolysis and vice versa. However, until recently, a formal systematic structure–reactivity relationship has not been constructed. In earlier studies with acyl glucuronides, a general pattern was observed in the relation between the structure of the aglycone and the rates of hydrolysis (for defined experimental conditions); as we reported, the bulkier the substituents on the a-carbon, the slower the hydrolysis (70). More quantitatively, the rates of hydrolysis can be linearly related to the STERIMOL parameters of the a-substituents in the case of aliphatic aglycones or ortho-substituents in the case of aromatic aglycones (71,72). Moreover, a group of aromatic aglycones, using the STERIMOL parameter B1 (one of the parameters for steric influence of substituents), together with the lowest unoccupied molecular orbital (LUMO) energies gave a linear regression model with an adjusted r2 value of 0.94 (71). Overall, the major determinant for acyl glucuronide reactivity in the different data sets appeared to be the bulkiness of the substituents in proximity of the electron-deficient carbonyl carbon atom (Elgabarty et al., in prep.). Intramolecular Acyl Migration of Acyl Glucuronides Intramolecular acyl group rearrangement is a well-established reaction in carbohydrate chemistry (73) and is mechanistically related to alkaline hydrolysis. Migration of the acyl moiety occurs from the 1-carbon hydroxyl group to the neighboring 2-, 3-, and 4-hydroxyl groups of the glucuronyl moiety (Fig. 2). This results in the formation of b-glucuronidase– resistant glucuronic acid esters that exhibit chromatographic properties different from the b-1O-acyl glucuronide. Intramolecular acyl migration was first demonstrated for bilirubin glucuronide. Studies with endogenous bilirubin-IXa-glucuronides collected from bile demonstrated a sequential migration of the original biosynthetic 1-O-acyl glucuronide to 2-, 3-, and 4O-isomers (40,61). Subsequently, studies of acyl glucuronides of various xenobiotic carboxylic acids have shown intramolecular acyl migration to be a general phenomenon for acyl glucuronides (1). However, the velocities of the different processes deviate according to the respective structural characteristics of the compounds (71). The mechanism of acyl migration is well established and proceeds via nucleophilic attack by the neighboring hydroxyl group and formation of an ortho-ester intermediate (23,73). In situ mechanistic studies with 1H-NMR spectroscopy of HPLC-purified isomers have determined the order of migration to be from the biosynthetic glucuronide to the 2-O-isomer followed by formation of the 3- and 4-O-isomers. The studies of Bradow et al. (23) indicated that there is no evidence for rearrangements beyond nearest-neighbor hydroxyl groups. Migrations between the positional isomers are readily reversible, except that reformation of the biosynthetic, higher energy b-1-O-acyl glucuronide is highly unfavorable. Reformation of the parent b-1-O-acyl glucuronide was believed to be very unlikely, yet recent work has documented that reversible 2-O- to 1-O-acyl migrations occur in both the b- and the a-configurations (66,74,75). As early as 1988, Hansen-Moller et al. (41) presented evidence for minor regeneration of diflunisal b-1-O-acyl glucuronide from the corresponding b-4-O-acyl isomer at pH 8 or a mixture of the acyl-migrated isomers at pH 8.5. However, other studies on rearrangement of acyl glucuronides concluded that this particular migration seems not to occur. Recently, a deviation from first-order kinetics at late time points was observed in S-naproxen b-1-O-acyl glucuronide (74). The authors concluded that the a-1-O-acyl isomer was formed in all reaction

134

Spahn-Langguth et al.

media and its formation must be considered a general mechanism in the rearrangement scheme of S-naproxen glucuronide regardless of the incubation conditions. They then included an equilibrium between the b-1-O-acyl glucuronide and its b-2-O-acyl isomer, yet favoring the b-2O-acyl isomer. The back-formation of b-1-O-acyl glucuronide is believed to be so minor that it has not been observed in the majority of studies of acyl glucuronide rearrangement kinetics. Akira et al. (75) demonstrated for probenecid that the b-1-O-acyl isomer exists in an equilibrium favoring the corresponding a/b-2-O-acyl isomer by showing that a low level of the b-1-O-acyl isomer was rapidly formed from the b-3-O-acyl isomer via the b-2-O-acyl isomer. Factors that Influence the Degradation of Acyl Glucuronides The loss of 1-O-acyl glucuronides (including hydrolysis and acyl migration) follows apparent first-order kinetics over the measurable concentration range. Subsequent loss of the rearranged isomers is generally much slower than that of 1-O-acyl glucuronide. The disappearance of zomepirac 1-O-acyl glucuronide and the formation of isomers and parent zomepirac in vitro at pH 7.4 and 378C is depicted in Figure 4, which demonstrates that intramolecular acyl migration under physiological conditions is the predominating reaction in the early stages of the in vitro incubations, whereas hydrolysis of 1-O-acyl glucuronide and its isomers becomes the more important reaction at later times or under alkali conditions (25). The rate of acyl migration and hydrolysis varies among different compounds and is influenced by many factors. Both hydrolysis and rearrangement are accelerated at alkaline pH and with increasing temperature. Hasegawa et al. (25) described an apparent first-order pH-dependent degradation (including acyl migration and hydrolysis) of zomepirac glucuronide with minimal isomerization occurring at pH 5, similar to what was reported for isoxepac (57), valproic acid (76), and furosemide (33,43). Degradation half-lives for the b-1-O-acyl glucuronides for various compounds in the physiological pH range are summarized in Table 2. Large variation has been noted in the degradation of acyl glucuronides, from highly labile glucuronides like that of diclofenac (54) and tolmetin (26) (no substitution at the

Peak height ratio

0.9

0 0

1

2

3

4

5 6 Hours

7

8

22

45

69

141

FIGURE 4 Time-dependent degradation of Zgl to its isomeric conjugates (2-, 3-, and 4-O-iso) and hydrolysis Z in 0.1 M phosphate buffer, pH 7.4, 378C (2-, 3-, and 4-iso represent the a/b-2-O-, a/b-3-O-, and a/b-4-O-acyl isomers, respectively). Solid circles, Zgl; solid squares, Z; open diamonds, 2-O-iso; open squares, 3-O-iso; open triangles, 4-Oiso; open stars, an unidentified product. Abbreviations: Zgl, zomepirac glucuronide; Z, zomepirac. Source: From Ref. 25.

135

Mechanistic Role of Acyl Glucuronides

TABLE 2 Apparent First-Order Degradation Half-Lives of b-1-O-Acyl Glucuronides at pH 7 to 7.5 at 378C Compound

pH

Buffer type/medium

t1/2 (hours)

Beclobric acid

7.4

25.7 (C), 22.7 (K)

28

Benoxaprofen Carprofen

7.4 7.4

Phosphate buffer (0.15 M) Tris HCl buffer (0.05 M) Phosphate buffer (0.067 M) Krebs Ringer phosphate buffer Phosphate buffer Tris maleate (0.1 M) Phosphate buffer (0.05 M) Phosphate buffer Phosphate buffer (0.1 M) Phosphate buffer (0.15 M) Sodium phosphate Phosphate buffer Tris HCl buffer (0.05 M) Tris maleate (0.05 M) Phosphate buffer (0.1 M) Phosphate buffer (0.15 M) Phosphate buffer (0.15 M) Phosphate buffer Urine Phosphate buffer Phosphate buffer Sodium phosphate (0.1 M) Phosphate buffer (0.1 M) Phosphate buffer (0.2 M) Phosphate buffer Phosphate buffer (0.15 M) Phosphate buffer (0.1 M) Phosphate buffer Phosphate buffer (0.1 M) Phosphate buffer (0.1 M) Phosphate buffer (0.1 M) Phosphate buffer (0.1 M) Phosphate buffer (0.1 M)

2 (R), 4.1 (S) 2.5 (R), 3.5 (S)

46 50

1.7 (R), 2.9 (S)

77

7.3 7a 3a

78 19 8

7.4 Clofibric acid

7.4 7 7.5

Diclofenac Diflunisal Etodolac

7.4 7.4 7.4

Fenoprofen Flufenamic acid Flunoxaprofen Furosemide Gemfibrozil Ibufenac

7.4 7.4 7.4 7.4 7.4 7.4

Ibuprofen

7.4

Indomethacin Isoxepac Ketoprofen Mefenamic acid Naproxen

7.4 7 7.4 7.4 7.4

Oxaprozin Probenecid Salicylic acid Suprofen

7.4 7.4 7.4 7.4

Telmisartan Tolmetin Valproic acid Wy-18251 Wy-41770 Zenarestat Zomepirac

7.4 7.4 7.4 7.4 7.4 7.4 7.4

0.47 and 0.51 0.60 and 0.67 20

References

54 79,80 30

0.99 (R), 1.95 (S) 7 4.5 (R), 8 (S) 5.289 44 1.1

81 78 44 33 82 32

3.3

32

1.4 0.29a 0.66 (R), 1.26 (S) 16.5 0.92 (R), 1.75 (S)

78 57 83 34 58

1.3 0.4 and 0.27 1.55 1.4 26 0.26 60a 0.38 14 0.42 0.45

17,84 59,75 42 31 67 26 36 84 84 85 25

a

Half-life was calculated from the data given in the reference. Source: References prior to 1992 from Ref. 2.

a-carbon) to most stable species like gemfibrozil (82) (disubstitution at the a-carbon) and valproic acid (35) (highly steric-hindered dipropyl substitution at the a-carbon). More interestingly, a single replacement of a chloro group (a better electron-withdrawing group than fluoro) on benoxaprofen by a fluoro group, which becomes flunoxaprofen, leads to an increase in the stability of the acyl glucuronides. The half-lives of (R)- and (S)-benoxaprofen glucuronides are two and four hours, respectively (46), whereas the half-lives of acyl glucuronides of (R)- and (S)-flunoxaprofen are 4.5 and 8.8 hours, respectively (44). We initially hypothesized that the degradation rate of acyl glucuronide at physiological conditions is predictable based on the chemical structure of the acid and depends on: (i) the degree of substitution at the a-carbon to the carboxylic acid (69), and (ii) electron-withdrawing or -donating groups at the a-carbon or on the conjugated aromatic ring. And in fact, considering both steric hindrance and electronic effects on the chemical reactivity of acyl glucuronides appears to give a better understanding of the marked differences in degradation rates of

136

Spahn-Langguth et al.

structurally diverse carboxylic acid–containing drugs. The systematic correlation of structural parameters with the degradation rates for the b-1-O-acyl glucuronides, however, gives evidence that steric parameters are by far more relevant. Degradation rates of acyl glucuronides depend not only on the pH and temperature, but also on the nature of the solution (e.g., buffer, organic solvent, plasma, blood, urine, bile). Furosemide glucuronide (86) was shown to degrade much faster in bile (t1/2Z19.5 minutes) and a supernatant solution of a duodenal homogenate from rabbits (t1/2Z1.2 minutes) than in buffer (pH 7.4, t1/2Z5.3 hours). Degradation of zomepirac glucuronide in blood and plasma was found to be faster than in buffer (87). Ruelius et al. (17) also found accelerated degradation of oxaprozin glucuronide in human serum albumin (HSA) and plasma. Indeed, they showed that albumin was catalytic for all three reactions (intramolecular acyl migration, hydrolysis, and covalent binding). Reports in the literature suggest that the effects of albumin or plasma on the stability of acyl glucuronide conjugates and their isomers vary with the drugs studied. HSA has been shown to enhance the degradation rates of acyl glucuronides of many carboxylic acid–containing drugs, including zomepirac (87), oxaprozin (17,88), fenoprofen (81), etodolac (30), ketoprofen (89), naproxen (58), clofibric acid (19), gemfibrozil (82), and diclofenac (54). An opposite (stabilizing) effect of HSA was observed for tolmetin glucuronide, but bovine serum albumin (BSA) causes an increase in the rate of hydrolysis (26). In the presence of HSA, degradation of diflunisal (79), salicylic acid (42), mefenamic acid (34), and furosemide (90) glucuronides in albumin solution was retarded in comparison with that found in buffer alone, while no significant change in the degradation rate of ibufenac glucuronide was observed with or without HSA (32). These data suggest that the effect of HSA toward acyl glucuronides is strongly dependent on the chemical structure of the aglycone moiety. To explain the accelerated degradation of oxaprozin glucuronide in the presence of HSA, Ruelius et al. (17) hypothesized that the degradation reaction of oxaprozin glucuronide in HSA proceeds through the formation of a reversible complex of oxaprozin glucuronide with HSA at the benzodiazepine-binding site [site II, as classified by Sudlow et al. (91)]. Located within this site is a reactive tyrosine, which appears to be the nucleophile responsible for mediating all the reactions, including hydrolysis, acyl migration, and covalent binding. Support for this hypothesis was obtained when other agents, like naproxen, decanoic acid, and oxaprozin itself, that strongly bind to the benzodiazepine site inhibited both hydrolysis and acyl migration (17,88). Watt and Dickinson (79) proposed a similar mechanism to explain the protective effect of albumin on the degradation of acyl glucuronide, by introducing two binding sites: one a reversible binding site and the other the primary site catalyzing the degradation of acyl glucuronide (catalytic site). If the reversible binding site happens to be the catalytic site, albumin may accelerate the degradation of acyl glucuronides, like oxaprozin. Otherwise, it may retard the degradation rate of acyl glucuronides, such as diflunisal and salicylic acid. The correctness of such speculation clearly requires further investigation.

REVERSIBLE BINDING OF ACYL GLUCURONIDES TO PROTEINS As mentioned above, by introducing the reversible binding of oxaprozin glucuronide to HSA, Ruelius et al. (17) could well explain the catalytic effect of HSA on the hydrolysis and acyl migration of acyl glucuronides. They also hypothesized that a correlation exists between reversible binding and irreversible (covalent) binding to plasma proteins, with reversible binding acting as a preliminary or an intermediate step (17,88). Measurements of reversible plasma protein binding of glucuronide conjugates, however, are rare. With respect to acyl glucuronides, the lack of data mainly results from experimental difficulties since the studies need to be carried out under physiological conditions (378C, pH 7.4), i.e., the conditions at which acyl glucuronides are not stable. By rapid ultrafiltration, reversible binding of acyl glucuronides to HSA has been studied for carprofen (50,53), zomepirac (92), tolmetin (92), ketoprofen (89), fenoprofen (93), naproxen (58), and furosemide (90). Interestingly, significant binding to HSA was found for the b-1-O-acyl glucuronides as well as for their positional isomers (92). Stereoselective reversible binding of (R)- and (S)-glucuronides of carprofen, naproxen, fenoprofen, and ketoprofen to HSA was also observed. (S)-glucuronides of carprofen

Mechanistic Role of Acyl Glucuronides

137

(50,53) and fenoprofen (93) had a higher affinity to HSA than the (R)-glucuronides. An opposite result, (R)-glucuronide having a higher affinity than the (S)-diastereomer, was observed for naproxen glucuronide (58). Presumably, reversible binding of acyl glucuronides to proteins acts as a preliminary step for irreversible (covalent) binding. Studies of reversible binding might help us better understand the mechanism of covalent binding and the HSA effect on the stability of acyl glucuronides. COVALENT BINDING OF ACYL GLUCURONIDES TO PROTEINS Hydrolysis and intramolecular acyl migration of acyl glucuronide conjugates of carboxylic acid–containing compounds have been extensively documented (1,2). A third reaction manifesting the inherent chemical electrophilicity of acyl glucuronides involves their capacity to act as substrates for the covalent binding of the aglycone to tissue proteins, notably albumin. Covalent binding, first described for bilirubin in 1966 (94), was demonstrated to be dependent on the presence of bilirubin acyl glucuronides in vitro and in vivo. van Breemen and Fenselau (22) reported covalent binding of flufenamic acid, indomethacin, clofibric acid, and benoxaprofen to BSA, when acyl glucuronides were incubated with the protein in vitro, and suggested that the mechanism involved transacylation with the free sulfhydryl group of cysteine residues. Ruelius and coworkers (17,88) documented the in vitro covalent binding of oxaprozin to HSA via acyl glucuronide and concluded, on the basis of extensive inhibition studies, that the site of covalent binding to HSA was a tyrosine residue located within the benzodiazepine-binding site (transacylation with the hydroxyl group of tyrosine). We have demonstrated that such covalent binding occurs in vivo in humans, as well as in vitro, for zomepirac (95), tolmetin (26,96,97), carprofen (77), fenoprofen (81), beclobric acid (28), naproxen (58), and diclofenac (54), and also for a number of carboxylic acid metabolites that are formed in vivo following gemfibrozil dosage (37), while McKinnon and Dickinson have shown such binding for diflunisal and probenecid (98), William et al. for valproic acid (35), Sallustio et al. for clofibric acid (99) and gemfibrozil (82), Shipkova et al. for mycophenolic acid (100), and Ebner et al. (67) for telmisartan. Procedures to Assay Covalent Binding Generally, the extent of covalent binding is quantified as the amount of aglycone that remains bound to proteins after an exhaustive washing procedure, which is then liberated after treatment with strong base (101). Proteins are usually precipitated by the addition of, for example, ice-cold isopropanol and acidic ACN or an ACN/ethanol mixture. The protein pellet obtained after centrifugation is washed several times (at least 5–10 times) with methanol/diethylether (3:1; vortexing and sonication, followed by centrifugation) to remove the reversibly bound aglycone and conjugates from the proteins. Aglycone–protein adducts are then dissolved in sodium hydroxide solution at 708C to 808C overnight to release the bound aglycone. The liberated aglycone can be either quantified by scintillation counting if the aglycone is radiolabeled or quantified by HPLC for nonradiolabeled aglycone (95). Similar procedures were used to study the binding of acyl glucuronides to DNA in vitro. The extent of covalent binding, determined by such an indirect assay, is usually expressed as picomols or nanomols of covalently bound aglycone per milligram of protein. This indirect procedure could be applied for both in vitro and in vivo covalent binding studies and the bound aglycone (drug) can be assayed specifically, even stereospecifically, by HPLC. However, these methods are labor-intensive and have provided very limited sequence information on the target proteins adducted, and thus the identities of the target proteins are usually unknown. Conventionally, following separation, MS, NMR, immunological and radioisotope methods are used to detect and identify protein targets for reactive drug metabolites (102). The aglycone-modified protein targets were identified by SDS-PAGE (or two-dimensional SDS-PAGE) and direct autoradiography or fluorography if radiolabeled aglycone is available or the aglycone is highly fluorescent. Direct quantification of aglycone fluorescence or radioactivity related to certain macromolecule fractions may also be performed following blotting

138

Spahn-Langguth et al.

(45). Because of the limited availability of radiolabeled compounds, more recently, immunochemical methods have been developed and have become the preferred methods to detect and identify xenobiotics covalently bound to proteins. The immunochemical methods involve the production of highly selective polyclonal antibodies by immunization of animals (e.g., rabbits) with aglycone-linked immunogenic protein. Via these methods, the protein covalent binding can be determined both quantitatively and qualitatively. Western blots permit detection of individual proteins targeted by the reactive metabolite and ELISA techniques permit quantification of protein adducts in tissues and subcellular fractions (103). In addition, immunohistochemical analysis permits assessment of distribution of covalent adducts among tissues and localization within individual cell types (104). The immunochemical methods have been successfully utilized to investigate the mechanisms of tissue toxicity of halothane, acetaminophen, diclofenac (105), and a variety of other chemicals. The results produced by such methods, however, are highly variable between laboratories: Sometimes, a completely different pattern of protein adducts is detected, mainly due to the different specificity of polyclonal antibodies produced among different laboratories. Antibody-based methods, in general, are frequently limited by the availability, quality, and specificity of antibodies to protein adducts (102). In Vitro Covalent Binding of Acyl Glucuronides to Plasma and Tissue Proteins As a consequence of the chemical reactivity of acyl glucuronides, a large number of carboxylic acid–containing drugs have been demonstrated to covalently bind to plasma proteins, especially albumin in vitro, including nonsteroidal anti-inflammatory drugs [benoxaprofen (23), indomethacin (23), flufenamic acid (23), oxaprozin (17), zomepirac (95), tolmetin (26), carprofen (77), fenoprofen (81), naproxen (58), diclofenac (54), diflunisal (79), salicylic acid (42), etodolac (30), suprofen (31), ibuprofen (32), ibufenac (32), ketoprofen (89), and mefenamic acid (34)], the uricosuric drug probenecid (59), the antihyperlipoproteinemic agents [clofibric acid (19,22), fenofibric acid (19,106), gemfibrozil (82), and beclobric acid (28)], the diuretic agent furosemide (90), and the antiepileptic drug valproic acid (35), as well as the acyl glucuronides of the oxidative metabolites of, e.g., gemfibrozil (64). From these in vitro studies, the extent of covalent binding was found to be clearly dependent on time (22,88), pH (26,31,95), glucuronide concentration (107), and origin of albumin (26,79). For oxaprozin glucuronide (17,88), the highest yield of protein adduct was obtained after the glucuronide and HSA were incubated at pH 7 for approximately one hour at 378C. Similarly, maximum covalent binding to HSA for zomepirac glucuronide occurred after one-hour incubation at pH 9, although the level of protein adducts decreased rapidly after this time owing to the instability of the adducts at this pH. High concentrations of adduct were also observed after six-hour incubation of zomepirac glucuronide and HSA at pH 7 and 8 at 378C (95). The in vitro covalent binding of suprofen glucuronide to HSA was shown to increase with increasing pH at 378C and to be time dependent (31). The extent of covalent binding of ketoprofen glucuronide (107) to albumin was proportional to acyl glucuronide concentration over the range studied (from 11.62 to 69.72 mM). Watt and Dickinson (79) showed that covalent binding of diflunisal glucuronide was greater with fatty-acid–free HSA than with rat plasma albumin and human and rat plasma proteins, and suggested that the different animal origins and the state of purity of albumin might be important for the stability and covalent binding of acyl glucuronides. Similar findings were also reported for tolmetin glucuronide (26). The extent of covalent binding of tolmetin glucuronide with BSA was much less than, but the rate of adduct formation was the same as, that with HSA. In addition to 1-O-acyl glucuronide, the isomeric conjugates could also form covalent protein adducts. Isomeric conjugates of zomepirac glucuronide (18,95) were found to covalently bind to HSA, at somewhat decreased extents as compared with the b-1-O-acyl glucuronide itself (% bound: C1OC2OC4OC3). Reports in the literature suggest that certain isomeric conjugates were even more reactive toward proteins than the b-1-O-acyl glucuronide. Isomers of suprofen glucuronide exhibited time-dependent covalent binding and this binding was 38% higher than that of the b-1-O-acyl glucuronide (31). Similarly, protein adduct formation of valproic acid (35), salicylic acid (42), etodolac (30), and diflunisal (108) was

139

Mechanistic Role of Acyl Glucuronides

shown to be much more rapid and extensive from isomeric glucuronide conjugates than from the b-1-O-acyl glucuronides. However, not all of the isomeric conjugates are important for the covalent binding. Ruelius et al. (17) reported that only the b-1-O-acyl glucuronide of oxaprozin, not the isomers, led to significant irreversible binding. Studies performed by Dubois et al. (107), as well by our laboratory (109), suggest that HSA is the major binding protein with respect to covalent binding to plasma proteins; for example, no covalent binding was detected with fibrinogen and g-globulins, and only 0.14% of ketoprofen was bound to a- and b-globulins after a three hour incubation. However, covalent binding is not restricted to albumin. Bailey et al. (110) have shown that zomepirac glucuronide and its isomers covalently modified microtubular protein in a dose-dependent manner and suggested that perturbation of the tubulin/microtubulin dynamics might contribute to the hepatotoxicity of certain acidic drugs. In vitro studies of covalent binding of tolmetin glucuronide to tissue homogenates from rat and sheep indicated that the extent of tissue covalent binding was comparable to that detected with albumin and plasma proteins (109). Similarly, incubation of rat liver microsomes with [14C]diclofenac showed that diclofenac covalently bound to hepatic microsome proteins varied as a function of exposure time and the concentration of the cofactor, UDPGA (111). Hepatic microsomes incubated with [I4C]UDPGA and nonradiolabeled diclofenac resulted in similar covalent binding of the radiolabeled compound to microsomal protein, which was significantly decreased in the presence of 7,7, 7-triphenylheptyl-UDP, a specific inhibitor of UGT.

HOOC

HOOC

O

HO HO

OH

O

O

HO HO

O

OH

OH

R Protein

X O Protein

HX

R

X=NH S O

Acylated protein

(A)

HOOC HO O

O

HOOC O

HO

OH

OH

O

O

R

O

HO

H

OH

H N

H H Stable ketoamine O

Protein

Protein

H2N

HOOC

HOOC O

OH

R Open-ring isomer

R Glucuronide isomer

HO

O

Amadori

RearrangementO

HO

O

R

HO

OH

N

Protein

H

Schiff's base

(B)

FIGURE 5 Proposed mechanisms for covalent binding of carboxylic acids to proteins via their acyl glucuronides. (A) Nucleophilic displacement (acylation). This represents the direct acyl group transfer that occurs under loss of glucuronic acid. (B) Via Schiff base intermediate (glycation). Following acyl migration, ring-opening occurs in the isomer and the aldehyde structure of the acylated glucuronic acid may react with a primary amino group in proteins forming a Schiff’s base. While this step is reversible, stabilization of the adduct may occur via Amadori rearrangement.

140

Spahn-Langguth et al.

Mechanism of Covalent Binding of Acyl Glucuronides to Proteins The mechanism of the irreversible binding of acyl glucuronide to proteins has been investigated extensively. Basically, two pathways have been proposed, each resulting in a different type of adduct (Fig. 5). The first is a nucleophilic displacement reaction whereby protein nucleophiles, including sulfhydryl, hydroxyl, and amine groups, react with the facile carbonyl carbon of the acyl glucuronide. This mechanism leads to the acylation of proteins giving rise to thioester-, oxygen ester-, and amide-linked conjugates. The consequence of this reaction would be the direct covalent linkage of the drug to the target protein without the glucuronic acid moiety. Evidence of the involvement of nucleophilic groups such as –SH of cysteine residues (22), –OH of tyrosine residues (88), and –NH2 of lysine residues (112,113) in the formation of covalent adducts between protein and various acyl glucuronides has been documented. The second mechanism of covalent binding of acyl glucuronides to proteins is analogous to the nonenzymatic glycosylation of albumin (114) and requires prior acyl migration of the drug moiety away from the biosynthetic b-1-O-acyl glucuronide to permit ring-opening of the carbohydrate. The reactive aldehyde group so exposed can then reversibly form an imine (Schiff’s base) with an amine group on protein. Subsequent Amadori rearrangement could then yield a stable ketoamine derivative. Thus, in contrast to the transacylation mechanism, both drug and glucuronic acid moieties (still linked together by an ester group) become bonded to proteins. This mechanism was first proposed for the covalent binding of zomepirac to plasma protein (95). In covalent binding studies with zomepirac and tolmetin glucuronides, iminetrapping agents (cyanide or cyanoborohydride) significantly increased the extent of covalent binding of the drug, supporting this mechanism. Furthermore, isomeric glucuronide conjugates released after extensive washing and subsequent acid treatment (18,95) gave evidence that the glucuronic acid moiety is part of the adduct. In vitro studies with clofibryl and fenofibryl glucuronides (19) showed that covalent binding of 14C to proteins was markedly higher after incubation of HSA with clofibryl or fenofibryl glucuronide labeled with 14C on the glucuronyl moiety, compared with the label on the aglycone. Binding of 14C to HSA was 14.9-fold (24-hour incubation) higher for clofibryl glucuronide and 5.9-fold (24-hour incubation) higher for fenofibryl glucuronide when labeled on the glucuronyl moiety than on the aglycone (19). This is consistent with the proposed Schiff’s base mechanism, in which the glucuronyl moiety becomes covalently bound to proteins. The two mechanisms proposed for adduct formation contrast sharply. The simpler transacylation mechanism involves nucleophilic attack at the ester group by –SH, –OH, or –NH2 groups on protein. The drug moiety itself thus becomes directly linked to the protein via a thioester, ester, or amide bond, and glucuronic acid is lost. Under physiological pH conditions, relative facile transacylation reactions might be expected of the 1-O-acyl glucuronide itself, but not of its 2-, 3-, or 4-isomers, since only in the b-1-O-acyl glucuronide is the carboxyl group of the drug linked to the glucuronic acid moiety via an acetal. Conversely, the Schiff’s base mechanism for adduct formation requires prior migration of the drug moiety away from the 1-position of the glucuronic acid ring and thus is operative for the isomers but not the 1-O-acyl glucuronide itself. According to this mechanism, the glucuronic acid moiety, still bearing the ester-linked drug, becomes bound to an amine group on protein via an imine formation (Schiff’s base). Because the reversibility of acyl migration only included the reformation of the parent acyl glucuronide to a minor extent (Fig. 2), the two mechanisms of adduct formation are theoretically distinguishable on the basis of whether the glucuronide or its isomers are the better substrate. The isomeric conjugates of some xenobiotic carboxylic acids, such as diflunisal (108), valproic acid (35), and salicylic acid (42), have been shown to be more reactive toward protein than the corresponding b-1-O-acyl glucuronide, supporting the Schiff’s base (glycosylation) mechanism. On the other hand, Ruelius et al. (17) presented strong evidence favoring a transacylation mechanism for covalent binding of oxaprozin to HSA. After incubation of radiolabeled oxaprozin glucuronide with HSA at pH 7 for one hour, 22% of the radioactivity became attached to HSA, but only 0.6% when the label was in the glucuronic acid moiety. Furthermore, only 2.1% attachment of label to HSA occurred after incubation with the 2-isomer

Mechanistic Role of Acyl Glucuronides

141

of [14C]oxaprozin glucuronide. Smith et al. (18,95) showed that zomepirac–HSA adduct formation from zomepirac acyl glucuronide was roughly comparable to that from its purified 2-isomer and greater than that from its purified 4- and 3-isomers, over a 45-minute incubation with HSA at pH 7.4 and 378C. Munafo et al. (26) found that the rate of covalent binding of tolmetin to HSA was 10 times greater for tolmetin glucuronide than for a mixture of its isomers (predominantly the 3-isomers, generated in situ from tolmetin glucuronide by preincubation in albumin-free buffer). All of these researchers suggested that more than one mechanism was operative. MS analysis of tryptic digests of albumin adducts (115,116) provided direct evidence that the in vitro binding of tolmetin glucuronide to HSA occurs via both mechanisms. Similar findings were reported for benoxaprofen glucuronide (117). Results obtained with naproxen also suggest that covalent binding to HSA via its acyl glucuronides proceeds through both transacylation and glycation mechanisms. Naproxen glucuronide rapidly forms an adduct that may be unstable, and the protein adduct from the 2-O-acyl glucuronide appears as important for the covalent binding as that from the 1-O-acyl glucuronide (118). It is possible and probable that both mechanisms occur concurrently in vivo. In Vivo Covalent Binding of Acyl Glucuronides to Proteins In Vivo Plasma Protein Binding The in vivo formation of covalently bound plasma protein adducts by acyl glucuronides has now been demonstrated in humans for a large number of compounds, including beclobric acid (28), clofibric acid (99), carprofen (77), diclofenac (54), diflunisal (98), fenoprofen (81), gemfibrozil (119), ketoprofen (107), probenecid (98), salicylic acid (42), tolmetin (96,97), valproic acid (35), and zomepirac (95), and also to some extent for the acyl glucuronides of flunoxaprofen and benoxaprofen (45) and those of oxidative metabolites of gemfibrozil (64), where covalent binding was significant for the dicarboxylic acid metabolite (M3) only (120). Studies with repetitive gemfibrozil dosage in rats showed higher binding for the major gemfibrozil metabolite M3 (dicarboxylic acid) than for parent drug (64). From these in vivo studies, the extent of protein binding of acyl glucuronides in vivo was found to correlate well with the extent of the exposure of acyl glucuronides, which is measured as the area-under-the-curve (AUC). Increase in plasma glucuronide concentrations leads to higher covalent binding. Increased adduct formation can thus be expected during chronic dosage or with decreased renal clearance of the glucuronide as in renal failure or resulting from a drug–drug interaction. Indeed, adduct concentrations in elderly patients treated with tolmetin were significantly higher than those in the control group of elderly patients given a single dose (96). Significant accumulation of protein adducts of tolmetin has also been observed in healthy human volunteers after a 10-day multiple dosing regimen of tolmetin. The bound levels after administration of multiple doses were approximately 10 times higher than those after a single dose was given to the same subjects (97). Valproic acid adducts were measurable in the plasma of epileptic patients on chronic drug therapy (35). Coadministration of probenecid and zomepirac resulted in an increase in the amount of covalent binding and an increase in exposure to zomepirac glucuronide plasma concentrations (121). Covalent binding of diflunisal and probenecid has been investigated after administration of multiple doses of each drug. After a six-day regimen in healthy human volunteers of oral diflunisal with concomitant administration of oral probenecid during the last two days, measurable covalent binding of both drugs via their acyl glucuronide metabolite has been observed (98). In Vivo Tissue Protein Binding and Processes Affecting Tissue Binding Unlike many reactive metabolites that may never leave the place of formation, acyl glucuronides may not only be directly excreted into bile, but are stable enough to reach the circulation and subsequently be excreted into the urine. Hence, covalent adducts are not only to be expected at the organ(s) of biosynthesis. The in vivo formation of covalent adducts with tissue proteins was demonstrated for diflunisal in liver, kidney, skeletal muscle, and small and large intestine of rats given the drug (122,123), as well as urinary bladder tissue proteins (124). Following daily diflunisal dosing, the adduct concentration increased in all tissues over time and declined slowly after cessation of

[1]

[1]

[1]

[1] D or M

D or M

D or M

D or M

D = Drug M = Metabolite

[50]

[50]

[5000]

[5000]

[1]

[1]

[1]

D or M

D or M

D or M

D or M

(A)

Concentration in: Blood [1] Hepatocyte [50] Bile [5000]

[50]

[50]

[1]

Blood flow OCT1

MRP3

OATP1B3 OATP2B1

OCT3

MRP5

OATP1B1

MRP4

Hepatocyte Tight junction

MRP2 MDR1 MDR3 BSEP BCRP

ABCG5 ABCG8

Canalicular membrane

MRP1 MRP6

(B)

Sinusoidal membrane

NTCP

Na+

OAT2

Blood flow

Plasma proteins Microsomal proteins Endoplasmatic reticulum Canalicular membrane proteins

Tubulin Cytoskeleton

(C)

FIGURE 6 (Caption on facing page)

Chromatin Nucleus

Plasma proteins

Mechanistic Role of Acyl Glucuronides

143

drug administration with a half-life of approximately 20 hours (124). Following repetitive gemfibrozil dosage in rats, significant covalent binding of gemfibrozil was detected in liver and intestinal tissues (64). Similarly, chronic dosing of rats with clofibric acid over a 21-day period resulted in higher concentrations of clofibric acid covalently bound to liver proteins. Concentration of tissue protein adducts seemed to increase linearly with time, with no indication of steady state having been achieved by 21 days (99). In vivo covalent binding of carboxylic acids to tissue proteins has also been documented for diclofenac, sulindac, and ibuprofen in mice liver (125), and for zomepirac and valproic acid in rat liver (126). The primary site of glucuronidation is the liver and, therefore, this organ is exposed to high concentrations of acyl glucuronides. Since the extent of covalent binding correlates with the exposure, high adduct levels are expected in the liver. But in contrast to the welldocumented adduct formation of xenobiotic carboxylic acid–containing drugs to plasma proteins, few studies have investigated the intracellular adduct formation in organs exposed to the drugs or their glucuronides. At the major site of glucuronide formation, the endoplasmatic reticulum, there appears to be some probability for covalent binding, and also at the canalicular membrane, where, by far, the highest concentrations of conjugates result. The high exposure of cellular structures in the liver to the reactive glucuronides is not only due to metabolism, but also due to the presence of transport systems. In the interplay with intracellular UGTs, this may lead to a considerable concentration gradient between the blood, the intrahepatic space, and the bile (Fig. 6A). When comparing the concentrations between perfusate, intracellular space, and bile, a ratio of up to 1:50:5000 has been found, e.g., in liver perfusions (127). Efficient transport processes appear to lead to the differences between these compartments: In general, the aglycones may have different acidities according to their structure and may move across membranes via passive diffusion or be substrates of transport systems. Since glucuronides are generally acidic (pKa 3–4 irrespective of the pKa of the aglycone), mostly more hydrophilic than their aglycones, and mainly ionized and highly polar at physiological pH, carrier systems are of particular importance for their movement across membranes, from the site of formation to extrahepatic compartments and to the bile (Fig. 6B). Seitz et al. (128) demonstrated that the reactive diclofenac glucuronide was selectively transported into rat bile via the canalicular conjugate export pump multidrug resistance protein 2 (mrp2) and that hepatobiliary transport is critical for diclofenac covalent binding to proteins in the biliary tree. By comparing the covalent binding pattern in normal Wistar rats with that

FIGURE 6 (Facing page) (A) Potential concentration gradient for drug conjugates in different spaces related to hepatocytes. Due to their polarity, membrane transport of acyl glucuronides is carrier mediated, resulting in a significant concentration gradient between sinusoidal space (blood), hepatocyte, and bile. A ratio of 1:50:5000 may be resulting. Intracellular macromolecules and particularly those facing the biliary tree are exposed to high concentrations of reactive conjugates. (B) Relevant transport systems in hepatocytes. Basolateral uptake and efflux as well as canalicular (apical) efflux are the most relevant processes (in addition to UGT-catalyzed acyl glucuronide formation)that define the exposure of intrahepatic macromolecules to reactive acyl glucuronides. For basolateral uptake, members of the oatp/OATP family appear most relevant, which were suggested to act as bidirectional transporters, carrying out hepatic uptake of anions under normal conditions, but acting as efflux systems under pathological conditions such as cholestasis. In canalicular efflux, several ATP-dependent transporters may be involved, the most important of which appears to be the multidrug resistance protein 2 (mrp2/MRP2). For basolateral efflux, e.g., MRP3 has been identified as an inducible efflux transporter for amphiphilic-conjugated organic anions. Changes in membrane transporter expression and activities can significantly affect acyl glucuronide disposition. This will affect aglycone levels (because of back-formation) as well as the exposure to reactive acyl glucuronides and may, hence, affect clinical toxicity and clinical efficacy of acyl glucuronide-forming drugs. The black semicircular arrows denote ATP-consuming processes. (C) Sites of adduct formation in hepatocytes (in addition to plasma protein adducts). Upon exposure to high acyl glucuronide concentrations, e.g., dipeptidyl peptidase IV (canalicular membrane protein), UGTs and tubulin have been identified as intrahepatic targets of adduct formation. Moreover, acyl glucuronides were found to interact with DNA, but the relevance of this interaction has not yet been clearly defined. In general, adduct formation results in altered activity of the macromolecule (e.g., protein) and may contribute to hepatotoxicity. Abbreviations: BCRP, breast cancer resistance protein (ABCG2); BSEP, bile salt export pump (ABCB11); D, drug; M, metabolite; MDR1, multidrug resistance protein 1(P-glycoprotein, ABCB1); MDR3, multidrug resistance protein 3 (ABCB4); MRP1 MRP 6, multidrug resistance associated protein (ABCC1 ABCC6); NCTP, NaC taurocholate co-transporting polypeptide; OAT; multispecific organic anion transporter (SLC22 family); OATP, organic anion transporting polypeptide (SLCO family); OCT, organic cation transporter (SLC22 family); UGT, uridine 5 0 -diphosphate glucuronosyltransferase; sterol transporters (ABCG5, ABCG8).

144

Spahn-Langguth et al.

in mutant mrp2 transport-deficient (TR) rats, the authors found that a major protein adduct of an apparent molecular mass of 118 kDa was selectively detected by immunoblotting in isolated canalicular, but not basolateral, membrane subfractions of wild-type rats, whereas no adducts could be identified in livers of TR rats (128). ATP-dependent mrp2/MRP2 (Fig. 6B), a canalicular efflux pump, appears to be the major transport system involved in the biliary excretion of conjugates/organic anions. Results available regarding its relevance are mainly from rat studies, but it must be concluded that similar relevance exists for humans. However, it is known that the molecular weight threshold for biliary excretion is greater in humans (approx. 450–500) than in rats (approx. 300–400). Other important processes, which codetermine the amount of reactive metabolites in the hepatocyte are, for basolateral uptake, members of the organic anion transporting polypeptide (oatp/OATP) family, which were suggested to act as bidirectional transporters, carrying out hepatic uptake of anions under normal conditions, but act as efflux systems under pathological conditions such as cholestasis, and for basolateral efflux (e.g., MRP3), identified as an inducible efflux transporter for amphiphilic, conjugated organic anions (15). MRP3 expression is elevated in livers of patients with Dubin–Johnson syndrome (with lack of MRP2) and in patients with cirrhosis. Hence, it is anticipated that changes in membrane transporter expression may occur in diseased states (129) and activities can significantly affect acyl glucuronide disposition. Changes in the disposition of the acyl glucuronide will lead to a change in aglycone levels and, presumably, also a change in the cellular exposure of macromolecules to the reactive metabolite and, consequently, affect the adduct formation depicted in Figure 6C. At the major site of glucuronide formation, the endoplasmatic reticulum, there appears to be high probability for covalent binding, as well as at the canalicular membrane, where by far, the highest conjugate levels result. One complicating factor in the interpretation of intrahepatic covalent binding is the fact that carboxylate drugs can also be substrates for acyl-CoA ligases and form reactive acyl-CoA thioesters, capable of forming covalent protein adducts in vivo. Possibly, both acyl glucuronides and acyl-CoAs contribute to the overall protein adduct formation, at least in the liver. Plasma adduct levels are somehow predictable, since the extent of adduct formation is usually related to the extent of exposure (70). Because of the different mechanisms for adduct formation in tissue, either via acyl glucuronide or via CoA thioester (130), it has to be assumed that the concentrations of acyl glucuronides in plasma are poor predictors of intrahepatic adduct formation (131). Stability of Protein Adducts At present, very little is known about the pharmacokinetics of the covalently bound protein adducts formed by carboxylic acid drugs in plasma and tissues, although the in vivo stability (half-life) of the formed adduct may be important for the potential immunogenic effects of a hapten (11). From the currently available data, it is evident that the plasma protein adducts are long-lived, with half-lives much greater than those of their parent carboxylic acids and acyl glucuronide conjugates. Tolmetin–protein adducts persisted in plasma beyond the period when concentrations of tolmetin and its glucuronide were measurable (97). Specifically, tolmetin–plasma protein adducts exhibited an average half-life of approximately 4.8 days, whereas tolmetin and its glucuronide had a half-life of five hours (96). McKinnon and Dickinson (98) have reported terminal half-lives for the plasma protein adducts of diflunisal and probenecid in humans of 10 and 13.5 days, respectively. The half-lives of (K)- and (C)beclobric acid plasma protein adducts in humans were reported to be 1.75 and 2.9 days, respectively (28). Adduct half-lives determined for gemfibrozil upon repetitive dosage in young and elderly patients were in the range of 2 to 2.5 days (64). For the structurally and closely related 2-arylpropionic acids, flunoxaprofen and benoxaprofen, significantly different adduct half-lives were found after single doses. For flunoxaprofen, the adduct t1/2 was 2 to 2.5 days, while that of benoxaprofen adducts was in the range of four days, with no difference between the two enantiomers (132). These values are significantly shorter than the half-life of albumin in humans (17–23 days) and may represent clearance of adducts formed with plasma proteins other than albumin, or may be caused by the breakdown or partial breakdown of relatively unstable adducts, e.g., via a loss of the acyl residue from the ketoamine that was generated via the glycation mechanism (Schiff’s base), independent of the longer turnover rates

145

Mechanistic Role of Acyl Glucuronides

for the protein itself. In vitro studies of diflunisal glucuronide (108) with HSA revealed a biphasic decline with an apparent terminal half-life of about 28 days. Kitteringham et al. (133) have also demonstrated that the clearance of dinitrobenzene–albumin adducts was dependent on the degree of substitution of the albumin, with clearance increasing as epitope density was increased, which may be another factor contributing to the elimination of plasma adducts formed from acyl glucuronides. Owing to the long-lived protein adduct, steady–state adduct concentrations may not be achieved until months after the commencement of chronic dosing, with significant accumulation at steady state. These long-lived adducts may lead to enhanced uptake by antigen-presenting cells (e.g., macrophages), resulting in greater possibilities to be processed and presented to the immune system. Selectivity of Covalent Binding of Acyl Glucuronides to Particular Tissue Proteins Evidence has accumulated to show that the protein covalent binding via acyl glucuronides is not random, but rather selective with respect to the proteins targeted (Fig. 7) and may include other types of macromolecules as well (132–134). Therefore, selective binding to specific cellular target proteins may correlate better with toxicity than total protein covalent binding. Identification of protein targets in the liver includes differentiation on the basis of molecular weight with SDS-PAGE analysis, autoradiography, and fluorography of immunoblots. Major protein bands, at which binding was detected for various compounds, are the 70-, 110-, and 140kDa bands [e.g., diclofenac (135,110,70); zomepirac (135,110,70); diflunisal (135,110,70); sulindac (135,110)] (15). Using a fluorescence detection technique, covalently bound flunoxaprofen and benoxaprofen (45) were associated with a 39- and 62-kDa protein in a rat hepatic microsome system in the presence of UDPGA. Proteins of 110 and 70 kDa were the major liver protein targets 7 0.7 0.6

6

0.4

Adduct density [mole/mole protein × 10–3]

Tolmetin

Furosemide

Zomepirac

0.5

Gemfibrozil Gemfibrozil M2

0.3

5

0.2

Gemfibrozil M4

Gemfibrozil M1 Beclobric acid

0.1

4

Gemfibrozil M3 Telmisartan

0.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.24

3 R-Fenoprofen 2 R-Carprofen 1

S-Fenoprofen

S-Carprofen Furosemide

0 0.0 r 2 =0.9939;

0.2

0.4

r =0.9969; p =0.9969;

0.6

0.8

y =0.1316 + 3.5457x

1.0

1.2

1.4

1.6

1.8

2.0

K [h–1]

FIGURE 7 Predictability of covalent binding in vitro. Plot of maximum epitope densities (mols of drug bound per mol of protein!103) versus degradation rate constant k (1/hr) for the in vitro incubation of various acyl glucuronides (1 mm) in the presence of human serum albumin (0.5 mM). Degradation rates reflect both acyl migration and hydrolysis. Results were obtained from seven different studies performed in our laboratories with purified b-1-O-acyl glucuronides and from the literature for telmisartan. (The data points for (C)- and (K)-beclobric acid cannot be distinguished on the scales used.) Source: From Refs. 64, 67, 70.

146

Spahn-Langguth et al.

modified by covalent attachment of these drugs in in vivo studies in rats (68). Similarly, immunochemical detection of diclofenac adducts in mouse liver homogenates, after oral treatment of mice with diclofenac, revealed a dose-dependent formation of four major protein adducts with apparent molecular masses of 50, 70, 110, and 140 kDa (105). Dose- and time-dependent covalent modifications of hepatic proteins by diclofenac were also detected in rats given diclofenac (136). Subcellular fractionation of rat liver homogenate from diclofenactreated rats showed that a 50-kDa microsomal protein and 110-, 140-, and 200-kDa plasma membrane proteins were preferentially modified by diclofenac. Hargus et al. (136) presented evidence implicating UGT-dependent glucuronidation in the formation of the 110-, 140-, and 200-kDa diclofenac protein adducts in vitro in rat liver homogenate, while the adduct formation of the 50-kDa microsome protein was shown to be cytochrome P450 dependent. Using immunofluorescence and immunohistochemistry, the majority of the diclofenac adducts were detected on the plasma membrane and localized within the bile canalicular membrane (136). The 110-kDa band from diclofenac-dosed rats (a band also modified by zomepirac, diflunisal, sulindac, and ibuprofen) was identified to be at least in part made up by the membrane enzyme dipeptidyl peptidase (CD26), which is localized to the apical (bile canalicular) membrane in hepatocytes and could well become a target during biliary excretion of acyl glucuronides through this membrane (131). On the other hand, studies undertaken by Kretz-Rommel and Boelsterli (137) have reported immunochemical identification of 50-, 60-, 80-, and 126-kDa adducts, which were expressed in cultured rat hepatocytes exposed to diclofenac in vitro, and of 60- and 80-kDa adducts, expressed in vivo in livers of rats given diclofenac. The 60-kDa protein was also detected in a UGT-dependent microsomal incubation of diclofenac and UDPGA, with radiolabel on either diclofenac or UDPGA (111). Furthermore, Gil and coworkers (138) have reported the detection of a major 60-kDa adduct generated in vitro when rat and human hepatocytes were cultured with diclofenac. The reasons for the different patterns found in the different laboratories are unclear at present, but contributing factors could include differences in the model system (in vivo/in vitro rat, mouse, and cultured hepatocytes), samples (liver homogenate and subcellular fractions), and specificity of antisera. A similar pattern of covalent binding was found with other carboxylic acid–containing drugs using drug-specific antibodies. Diflunisal and zomepirac (126) were shown to produce major 110-, 140-, and 200-kDa hepatic protein adducts in vivo, similar to the results found with diclofenac. A different pattern of protein modification was detected in the livers of clofibric acid– and valproic acid–treated rats (126). A 70-kDa protein adduct was detected in clofibric acid–treated rats, while a 140-kDa protein and several other proteins with smaller molecular weight (e.g., 40, 43, and 55 kDa) were detected in livers of valproic acid–treated rats. The major protein adduct observed with sulindac (125) was the 110-kDa protein, with low levels found for the 140- and 200-kDa proteins. All the sulindac-modified protein adducts were shown to be concentrated in a subfraction derived from the bile canalicular region of the hepatocyte plasma membrane. Ibuprofen was the least toxic of carboxylic acids tested and predominantly bound covalently to a 60-kDa protein with only relatively low levels of a 110-kDa adduct (125). Such selective modification of plasma membrane proteins by carboxylic acids, possibly containing new antigenic determinants, could become particularly important if the immune system were involved in the pathogenesis of carboxylic acid–induced liver injury (104).

STEREOCHEMICAL ASPECTS OF ACYL GLUCURONIDES Many of the carboxylic acid drugs belong to the class of 2-arylpropionic acids (profens), which have a chiral center at the carbon 2 of the propionic acid side chain. Only (S)-enantiomers have significant anti-inflammatory activity (139). Nevertheless, the clinically used profens are marketed as racemates with the notable exception of naproxen. A unique feature of the metabolism of this class of compounds is the inversion at the chiral center (carbon 2), generally referred to as chiral inversion, which is unidirectional in mammalian organisms. The pharmacologically inactive (R)-enantiomer is usually transformed to the active (S)-antipode, whereas the reverse reaction does not occur.

147

Mechanistic Role of Acyl Glucuronides

In general, the principal urinary metabolites of the profen drugs are their acyl glucuronides. Acyl glucuronidation of chiral carboxylic acids was reported to be enantioselective (140). Investigations of substrate enantioselectivity in the formation of acyl glucuronides have been performed in vitro with microsomes, solubilized microsomal protein, and immobilized protein obtained from animal and human liver as sources of UGTs. Species-dependent enantioselective formation of conjugates of naproxen, ibuprofen, and benoxaprofen were described by el Mouelhi et al. (49). Similar results for benoxaprofen glucuronidation have been reported by Spahn et al. (46). Preferential glucuronidation of the (R)-enantiomer with rat liver microsomes was observed with various 2-arylpropionic acids, including 2-phenylpropionic acid (48), flunoxaprofen, flurbiprofen, indoprofen, pirprofen, benoxaprofen, and carprofen, with the exception of naproxen (135,141,142). In vitro glucuronidation studies of (R)- and (S)-ketoprofen in liver microsomes from a number of animal species demonstrated that the rate of glucuronidation of (S)-aglycone was 4.5-fold faster than that of (R)-enantiomer in dog liver microsomes, whereas no significant stereoselectivity was found in human, rat, or rabbit liver microsomes (52). With sheep liver microsome preparations, glucuronide yields were higher for (R)-flunoxaprofen (44,143) and (R)-fenoprofen (47) than for their respective (S)glucuronides. Glucuronidation studies with enzyme-induced liver microsomes performed by Fournel-Gigleux et al. (48) with 2-phenylpropionic acid as the substrate clearly demonstrated that acyl glucuronide formation is significantly induced by phenobarbital, whereas other inducers (dexamethasone and 3-methylcholanthrene) lead to a minor increase of glucuronidation. The S/R ratio of acyl glucuronidation was not affected by any of the inducers (Spahn-Langguth, unpublished results). In each of these studies, potential interference from stereoselective degradation of acyl glucuronides was minimized by rapid sample quenching, lowering the incubation pH to 5.5, or by the addition of specific esterase and b-glucuronidase inhibitors to prevent the enzymatic hydrolysis of acyl glucuronides (140,142). In addition to stereoselective acyl glucuronidation, degradation of the diastereomeric acyl glucuronides (including hydrolysis and acyl migration) of various chiral carboxylic acids has also been shown to be stereoselective (140). In studies with benoxaprofen (46), flunoxaprofen (44), carprofen (77), naproxen (58), ketoprofen (52), and fenoprofen (81), the apparent first-order degradation half-lives of the (S)-acyl glucuronides were approximately twofold longer than those of their corresponding (R)-acyl glucuronides in a protein-free buffer system at pH 7.4, 378C (Table 2). Stereoselective degradation of carprofen glucuronides under different conditions and the influence of HSA were characterized by our group (70). When (R)- and (S)-carprofen glucuronides were incubated at pH 7, 7.4, and 8 at 378C in phosphate buffer, degradation was highly stereoselective at pH 7. Stereoselectivity decreased while degradation velocity increased with higher pH, as summarized in Table 3. At all pH values, the (R)-

TABLE 3 Degradation Half-Lives of Carprofen b-1-O-Acyl Glucuronides: Influence of pH, Temperature, and Addition of Albumin on the Velocities of Degradation and Their Enantioselectivities Half-life (hours)

1 pH effect at 378C pH 7 pH 7.4 pH 8 2 Temperature dependence at pH 7.4 48C 258C 378C 3 Effect of HSA, pH 7.4, 378C Without HSA With 30 mM HSA (fatty-acid free) With 30 mM HSA (fraction V) Abbreviation: HSA, human serum albumin. Source: From Ref. 140.

(S)-glucuronide

(R)-glucuronide

S/R ratio

6.42 2.90 0.85

2.60 1.72 0.60

2.43 1.69 1.41

O100 11.8 2.90

O100 7.80 1.72

1 1.51 1.69

2.90 1.55 1.82

1.72 2.80 1.78

1.69 0.55 1.02

148

Spahn-Langguth et al.

glucuronide conjugate of carprofen degraded more rapidly than the (S)-glucuronide. When HSA was added to the incubation medium, the stability of (S)-glucuronide was decreased, whereas the apparent half-life of (R)-glucuronide increased. Georges et al. (50) published comparable data and explained the effect of HSA on the selective stabilization and protection of (R)-carprofen, yet a catalyst for hydrolysis of both enantiomers. Interestingly, the effect of fattyacid–free HSA was much greater than that of fraction V HSA (110). This finding suggests that seemingly trivial differences in HSA purity could be important for the stability and chemical reactivity of acyl glucuronides (129). A similar stereoselective effect of HSA on the stability of diastereomeric glucuronides has been observed for naproxen glucuronides. The addition of HSA to the incubation medium not only increased the degradation rate of naproxen glucuronide, but also caused a change in the stereoselective stability, where the (R)-naproxen glucuronide became more stable than the (S)-glucuronide (58). In summary, for those compounds with a chiral a-carbon, the R-enantiomer is always less stable; this is observed uniformly in beclobric acid, flunoxaprofen, benoxaprofen, carprofen, fenoprofen, naproxen, and ketoprofen. Given the reported prevalence of the steric factor, the reason behind this stereospecific behavior is probably steric as well. In addition—under the experimental conditions employed in general—the larger contribution to the instability of acyl glucuronides out of the processes described is due to acyl migration; hence, this stereospecific behavior should be related to a steric effect acting specifically on this migration step. It is apparent that—for the acyl-migration reaction in the S-enantiomer—the bulkier group is always oriented in a fashion that hinders the attack by the hydroxyl group on C-2 in glucuronic acid, i.e., the R-enantiomer offers a comparatively less crowded environment for the C-2 hydroxyl group on glucuronic acid to attack the carbonyl group (71). In contrast to the well-documented stereoselective degradation studies of acyl glucuronides, relatively few studies have examined the potentially stereoselective nature of covalent binding of chiral carboxylic acids to protein via their acyl glucuronide conjugates. The covalent binding of carprofen to HSA after one-hour incubation was higher with (S)-carprofen glucuronide than (R)-glucuronide, whereas after 24 hours, covalent binding was significantly higher for (R)-carprofen glucuronide incubations (77). In vitro covalent binding was also found to be higher for (R)-naproxen (57) than for (S)-naproxen when a 50-mM concentration of each epimeric glucuronide was incubated with HSA under physiological conditions (pH 7.4, 378C). This stereoselective difference was observed with an HSA-containing medium as well as in rat and human plasma. No significant diastereoselective difference between the glucuronides of the two beclobric acid enantiomers was detected with respect to the extent of in vitro covalent binding to HSA (28). Volland et al. (81) described significantly more covalent binding for the (R)-enantiomer of fenoprofen to human plasma protein in vitro; however, in vivo, this stereoselectivity was reversed. This provides a clear example of competing enantioselective metabolism since (R)-fenoprofen is subjected to significant chiral inversion in humans, which will increase the exposure to (S)-fenoprofen and its glucuronide relative to its optical enantiomer in vivo. Since acyl glucuronidation, stability under physiological conditions, and extent of covalent binding of diastereoisomeric acyl glucuronides to plasma proteins are often stereoselective (140), one might consider toxicity of a racemic compound to be more extensive due to one enantiomer than to its antipode. However, at this stage, prediction of toxicity of carboxylic acids, including considerations concerning stereoselectivity, resulting from unstable acyl glucuronides is only speculative. Whether only (S)-enantiomers of the profen drugs should be marketed is still debatable. PREDICTABILITY OF THE COVALENT BINDING OF ACIDIC DRUGS The accumulated data from a number of studies suggest that the extent of covalent binding for carboxylic acid drugs in vitro may be predicted on the basis of the degradation rate constant (including hydrolysis and acyl migration) of the glucuronide conjugate. A synthesis of published data (70) on the covalent binding of several acyl glucuronides indicates that there is a good linear correlation between the apparent first-order disappearance rate constant for an b-1-O-acyl glucuronide in buffer, which is a measure of its chemical reactivity, and the

149

Mechanistic Role of Acyl Glucuronides

maximum covalent binding observed when the glucuronide is incubated with HSA in vitro (Fig. 7). Acyl glucuronides of arylacetic acid (a-unsubstituted) such as tolmetin and zomepirac exhibit the highest covalent binding and lowest stability (highest degradation rate). The intermediate stable glucuronides of 2-arylpropionic acids (mono a-substituted), such as carprofen and fenoprofen, have lower covalent binding. Lowest covalent binding is observed for the most stable fully substituted carboxylic acids, such as beclobric acid. Figure 7 summarizes data from our laboratories (obtained over a period of several years) with respect to in vitro degradation rates of the b-1-O-acyl glucuronides and the in vitro covalent binding for different drug molecules (including new data on gemfibrozil metabolites and telmisartan data from the literature), suggesting that the extent of in vitro covalent binding to albumin is predictable based on the chemical structure of the acid and depends on the degree of substitution at the a-carbon to the carboxylic acid. It can be argued that the global degradation rate constant of b-1-O-acyl glucuronides (Zsummation of hydrolysis and acyl migration) cannot be used as a predictor of potential covalent binding to proteins. Covalent binding occurs also—or even preferentially—via the acyl-migrated isomers, and hence a distinction has to be made between the rate of hydrolysis and the rate of acyl migration. Pursuing this idea, Bolze et al. (144) attempted to reproduce the previously published correlation (70) using their microsome-based screening model, in which small-scale biosynthetic generation of the acyl glucuronides and HSA adduct formation were performed with eight drugs (diclofenac, fenoprofen, furosemide, ibuprofen, ketoprofen, suprofen, tolmetin, and zomepirac). When the global hydrolysis rate (of all the isomers) weighted by the percentage of isomerization was used, a good correlation with covalent binding was found (r2Z0.94). The authors concluded that the extent of covalent binding could be predicted on the basis of acyl glucuronide hydrolysis rate combined with acyl-migration propensity. Since there was no discrimination between the behavior of the conjugates of individual enantiomers of fenoprofen, ketoprofen, suprofen, or ibuprofen, the measured values are actually averaged values for two different molecules. The relationship for in vivo covalent binding would be more complex than that of in vitro binding. The degree of covalent binding to plasma proteins should depend, at least, on the plasma concentrations of the acyl glucuronides and the degradation rate of each conjugate. The plasma concentrations of acyl glucuronides vary with the drug studied, and are dependent on the rate of formation, its degradation, and elimination, as well as the administered dose. Acyl glucuronides of some carboxylic acids may reach significant concentrations in plasma of humans, as shown for zomepirac (121), tolmetin (145), diflunisal (98), beclobric acid (28), and etodolac (30), while no oxaprozin (17) and fenofibric acid (106) glucuronides have been detected in human plasma. In vivo studies (Table 4) with five carboxylic acid drugs, at their usual therapeutic doses, in five different sets of healthy volunteers, showed a 30-fold variation TABLE 4 In Vivo Bound Drug, Area Under the Plasma Drug Glucuronide Concentration Time Curve, and In Vitro Acyl Glucuronide Degradation Rates Parent compound

Bound drug (mole/ mole protein)!104

AUC glucuronide (mole!hr/L)!106

Bound/ AUC (10L2)

k (1/hr)

Tolmetin Zomepirac (R)-Fenoprofen (S)-Fenoprofen Racemic carprofen (C)-Beclobric acid (K)-Beclobric acid

2.77G1.54 2.33G0.45 1.02G0.32 3.23G0.85 1.92G1.28 0.12G0.03 0.20G0.11

3.72G0.95 6.41G2.14 6.31G5.65 60.4G24.7 40.9G7.3 8.16G1.34 8.31G1.63

0.75 0.36 0.16 0.054 0.047 0.015 0.024

1.78 1.54 0.71 0.36 0.32 0.031 0.027

Measurement of maximum amount of drug covalently bound to human serum albumin and area under the plasma concentration time curve (AUC) for the glucuronide conjugates measured in five different groups of healthy volunteers following oral dosing of either 400 mg of tolmetin (145), 100 mg of zomepirac (95), 600 mg of racemic fenoprofen (81), 50 mg of racemic carprofen (77), or 100 mg of racemic beclobric acid (28). When covalently bound drug is normalized to AUC for the respective glucuronide conjugates, an excellent correlation with the in vitro degradation rate constant (k) is obtained with an r2 value of 0.873. Abbreviation: AUC, area-under-the-curve. Source: From Ref. 70.

150

Spahn-Langguth et al.

in AUCs for acyl glucuronides, whereas the maximum plasma protein binding showed a 25-fold variation. Since for each drug there is a direct relationship between the amount of covalent binding and the extent of exposure of acyl glucuronide (AUC), we normalize bound drug to glucuronide AUC for comparison with in vitro glucuronide degradation rates, yielding a highly significant linear correlation (r2Z0.873). The findings presented in Table 4 suggest that the in vivo covalent binding of acidic drugs to albumin in humans is also predictable on the basis of the degradation rate constant of the glucuronide conjugate when the extent of covalent binding is corrected for the levels of the glucuronide present in plasma.

POTENTIAL TOXICOLOGICAL SIGNIFICANCE OF THE REACTIVE ACYL GLUCURONIDES It has been hypothesized that acyl glucuronides, owing to their reactive nature, may have a role in the observed toxicities associated with administration of a number of acidic compounds. It is striking that out of 47 drugs withdrawn from U.S., British, and Spanish markets from 1964 through 1993 owing to severe toxicity (5,6), 10 are carboxylic acids. These drugs—alclofenac, bendazac, benoxaprofen, fenclofenac, ibufenac, indoprofen, pirprofen, suprofen, ticrynafen, and zomepirac—are primarily metabolized by humans to acyl glucuronides. For all these discontinued carboxylic acids, the most frequent types of adverse reaction leading to the decision to discontinue the products were idiosyncratic toxicities such as liver damage, serious skin reactions, and renal toxicity, sometimes associated with fever, rash, and eosinophilia. Covalent binding of carboxylic acids to proteins via their common reactive intermediates, acyl glucuronides, has been proposed to mediate such idiosyncratic toxicities associated with carboxylic acid–containing drugs (11,146). Both direct toxic effects and immune-mediated toxicity (hypersensitivity reactions) have been suggested as a possible mechanism of idiosyncratic liver injury (147). With direct toxicity, covalent protein binding via acyl glucuronides may disrupt the normal physiological function of a “critical” protein or some critical regulatory pathway, leading to cellular necrosis. Alternatively, the chemically reactive acyl glucuronides of carboxylic acids can act as a hapten and initiate an immune reaction that may be mediated via a specific humoral (antibody) response, a cellular response (T lymphocytes), or a combination of both (12,148). In most cases, the differentiation of these two forms of idiosyncratic toxicity is largely empirical as it is based on the clinical symptoms; for example, manifestations such as rash, fever, lymphadenopathy, and eosinophilia all suggest drug hypersensitivity (immunemediated toxicity). The lack of the clinical hallmarks of immunoallergic reactions, combined with the nature of the histological changes, may suggest a direct toxic reaction. At present, the exact mechanisms responsible for the initiation and perpetuation of carboxylic acid–associated idiosyncratic organ (especially liver) toxicity and anaphylaxis remain poorly understood. Although it has not been ultimately proven that immune reactions are causally involved in such toxicities, a number of reports from the literature have provided evidence that immune-mediated toxicity plays an important role (1,149). Drug-specific antibodies have been detected in aspirin-hypersensitive patients (150) and in patients receiving valproic acid therapy (35). Immunization with the mouse albumin conjugate of tolmetin glucuronide has been demonstrated to stimulate an antibody response in mice (151). Antiadduct antibodies formed in mice following the administration of tolmetin–albumin adducts appeared to be specific for the aglycone and some cross-reactivity was observed for the structurally related carboxylic acids and their glucuronides. Kretz-Rommel and Boelsterli have characterized the selective covalent binding of diclofenac to rat and mice liver proteins in vivo, in cultured hepatocytes (137), and in subcellular incubations (111). Since these selective protein adducts were shown not to exhibit a direct cytotoxic effect in short-term cultures of hepatocytes, the authors proposed that such selective covalent binding may be involved in the development of an immunogenic reaction in vivo (152). To confirm such a hypothesis, a murine ex vivo/in vitro mixed lymphocyte hepatocyte culture model was developed (152). Cultured hepatocytes from C57BL mice, preexposed to nontoxic concentrations of diclofenac, were cocultured with splenocytes derived from mice immunized with a synthetic diclofenac–protein adduct, i.e., diclofenac covalently linked to the carrier protein, keyhole limpet hemocyanin

Mechanistic Role of Acyl Glucuronides

151

(KLH). Splenocyte-mediated cytotoxicity was demonstrated by massively increased alanine aminotransferase release, an indicator of hepatocyte injury, apparent at 48 hours in coculture and present only in those cultures pretreated with diclofenac, not in the untreated controls or in the cocultures treated with KLH alone. These experiments imply a role for T cells in diclofenacdependent cell killing and further support the possibility of immune-mediated toxicities (152). In addition to drug hypersensitivity, direct disruption of function of critical proteins or important regulatory pathways by reactive acyl glucuronides may be involved in the idiosyncratic toxicities of carboxylic acids. For certain carboxylic acids, both mechanisms may be operative simultaneously. The rare, but potentially lethal, idiosyncratic adverse reactions of carboxylic acids are highly host dependent. The risk of the unpredictable hepatocytic injuries posed by carboxylic acids is small (153); nevertheless, fulminant hepatitis may develop in susceptible patients, who may have abnormal metabolism or excretion of carboxylic acids, resulting in the overproduction and accumulation of reactive acyl glucuronides. Genetic and environmental variations in acyl glucuronidation, canalicular or sinusoidal secretion, and renal clearance of acyl glucuronides may all contribute to enhanced susceptibility, but these pathophysiological abnormalities have been poorly investigated. Specifically, interindividual variations in acyl glucuronidation in humans should be better characterized. Furthermore, the interindividual variations in canalicular excretion of acyl glucuronides may also need further characterization. Results from different laboratories indicate the important role of transporters of acyl glucuronides in the selective covalent binding of carboxylic acids to proteins. Identification of potential genetic and environmental factors in the susceptible individuals will not only allow us to identify the vulnerable individuals, but also help us better understand the mechanism of toxicity. ACKNOWLEDGMENTS Preparation of the manuscript and the studies in the authors’ laboratory were supported in part by NIH grant GM36633. The authors appreciate the assistance of Hossam Eldin Elgabarty (GUC) in the preparation of part of the figures for this manuscript. REFERENCES 1. Faed EM. Properties of acyl glucuronides: implications for studies of the pharmacokinetics and metabolism of acidic drugs. Drug Metab Rev 1984; 15:1213–49. 2. Spahn-Langguth H, Benet LZ. Acyl glucuronides revisited: is the glucuronidation process a toxification as well as detoxification mechanism? Drug Metab Rev 1992; 24:5–47. 3. Zimmerman HJ. Update of the hepatotoxicity due to classes of drugs in common clinical use: nonsteroidal drugs, anti-inflammatory drugs, antibiotics, antihypertensives, and cardiac and psychotropic agents. Semin Liver Dis 1990; 10:322–38. 4. Zimmerman HJ. Hepatic injury associated with nonsteroidal anti-inflammatory drugs. In: Lewis AJ, Furst DE, eds. Nonsteroidal Antiinflammatory Drugs: Mechanisms and Clinical Use. New York: Marcel Dekker, 1994:171–94. 5. Bakke OM, Wardell WM, Lasagna L. Drug discontinuations in the United Kingdom and the United States, 1964 to 1983: issues of safety. Clin Pharmacol Ther 1984; 35:559–67. 6. Bakke OM, Manocchia M, de Abajo F, Kaitin KI, Lasagna L. Drug safety discontinuations in the United Kingdom, the United States and Spain from 1974 to 1993, a regulatory perspective. Clin Pharmacol Ther 1995; 58:108–17. 7. Stogniew M, Fenselau C. Electrophilic reactions of acyl-linked glucuronides, formation of clofibrate mercapturate in human. Drug Metab Dispos 1982; 10:609–13. 8. Shore LJ, Fenselau C, King AR, Dickinson RG. Characterization and formation of the glutathione conjugate of clofibric acid. Drug Metab Dispos 1995; 23:119–23. 9. Grillo MP, Benet LZ. In vitro studies of tolmetin metabolism in fresh isolated rat hepatocytes. Identification of a tolmetin–glycine amino acid conjugate. ISSX Proc 1995; 8:228. 10. van Breemen RB, Fenselau C. Reaction of 1-O-acyl glucuronides with 4-(p-nitrobenzyl)pyridine. Drug Metab Dispos 1986; 14:197–201. 11. Boelsterli UA, Zimmerman HJ, Kretz-Rommel A. Idiosyncratic liver toxicity of nonsteroidal antiinflammatory drugs: molecular mechanisms and pathology. Crit Rev Toxicol 1995; 25:207–35. 12. Park BK, Coleman JW, Kitteringham NR. Drug disposition and drug hypersensitivity. Biochem Pharmacol 1987; 36:581–90.

152

Spahn-Langguth et al.

13. Burchell B, Brierley CH, Rance D. Specificity of human UDP-glucuronosyl transferases and xenobiotic glucuronidation. Life Sci 1995; 57:1819–31. 14. Magdalou J, Chajes V, Lafaurie C, Siest G. Glucuronidation of 2-arylpropionic acids pirprofen, flurbiprofen, and ibuprofen by liver microsomes. Drug Metab Dispos 1990; 18:692–7. 15. Sallustio BC, Sabordo L, Evans AM, Nation RL. Hepatic disposition of electrophilic acyl glucuronide conjugates. Curr Drug Metab 2000; 1:163–80. 16. Panfil I, Lehman PA, Zimniak P, Ernst B, Franz T, Lester R, Radominska A. Biosynthesis and chemical synthesis of carboxyl-linked glucuronide of lithocholic acid. Biochim Biophys Acta 1992; 1126:221–8. 17. Ruelius RW, Kirkham SK, Young EM, Jassen FW. Reaction of oxaprozin-1-O-acyl glucuronide in solutions of human plasma and albumin. Adv Exp Med Biol 1986; 197:431–41. 18. Smith PC, Benet LZ, McDonagh AF. Covalent binding of zomepirac glucuronide to proteins: evidence for a Schiff base mechanism. Drug Metab Dispos 1990; 18:639–44. 19. Grubb N, Weil A, Caldwell J. Studies of the in vitro reactivity of clofibryl and fenofibryl glucuronides, evidence for protein binding via a Schiff base mechanism. Biochem Pharmacol 1993; 46:357–64. 20. Parikh I, MacGlashan DW, Fenselau C. Immobilized glucuronosyltransferase for the synthesis of conjugates. J Med Chem 1976; 19:296–9. 21. Fenselau C, Pallante S, Parikh I. Solid-phase synthesis of drug glucuronides by immobilized glucuronosyltransferase. J Med Chem 1976; 19:679–83. 22. van Breemen RB, Fenselau C. Acylation of albumin by 1-O-acyl glucuronides. Drug Metab Dispos 1985; 13:316–20. 23. Bradow G, Kan L, Fenselau C. Studies of intramolecular rearrangements of acyl-linked glucuronides using salicylic acid, flufenamic acid, and (S)- and (R)-benoxaprofen and confirmation of isomerization in acyl-linked D9-11-carboxyltetrahydrocannabinol glucuronide. Chem Res Toxicol 1989; 2:316–24. 24. Eggers NJ, Doust K. Isolation and identification of probenecid acyl glucuronide. J Pharm Pharmacol 1981; 33:123–4. 25. Hasegawa J, Smith PC, Benet LZ. Apparent intramolecular acyl migration of zomepirac glucuronide. Drug Metab Dispos 1982; 10:469–73. 26. Munafo A, McDonagh AF, Smith PC, Benet LZ. Irreversible binding of tolmetin glucuronic acid ester to albumin in vitro. Pharm Res 1990; 7:21–7. 27. Watt JA, King AR, Dickinson RG. Contrasting systemic stabilities of the acyl and phenolic glucuronides of diflunisal in the rat. Xenobiotica 1991; 21:403–15. 28. Mayer S, Mutschler E, Benet LZ, Spahn-Langguth H. In vitro and in vivo irreversible plasma protein binding of beclobric acid enantiomers. Chirality 1993; 5:120–5. 29. Iwakawa S, Suganumna T, Lee S, Spahn H, Benet LZ, Lin ET. Direct determination of diastereomeric carprofen glucuronides in human plasma and urine and preliminary measurements of stereoselective metabolic and renal elimination after oral administration of carprofen in man. Drug Metab Dispos 1989; 17:414–9. 30. Smith PC, Song WQ, Rodriguez RJ. Covalent binding of etodolac acyl glucuronide to albumin in vitro. Drug Metab Dispos 1992; 20:962–5. 31. Smith PC, Liu JH. Covalent binding of suprofen acyl glucuronide to albumin in vitro. Xenobiotica 1993; 23:337–48. 32. Castillo M, Smith PC. Disposition and reactivity of ibuprofen and ibufenac acyl glucuronides in vivo in the Rhesus monkey and in vitro with human serum albumin. Drug Metab Dispos 1995; 23:566–72. 33. Rachmel A, Hazelton GA, Yergey AL, Liberalo DJ. Furosemide 1-O-acyl glucuronide, in vitro biosynthesis and pH-dependent isomerization to b-glucuronidase-resistant forms. Drug Metab Dispos 1985; 13:705–10. 34. McGurk KA, Remmel R, Hosagrahara VP, Tosh D, Burchell B. Reactivity of mefenamic acid 1-O-acyl glucuronide with proteins in vitro and ex vivo. Drug Metab Dispos 1996; 24:842–9. 35. William AM, Worrall S, Jersey JD, Dickinson RG. Studies on the reactivity of acyl glucuronides. III. Glucuronide-derived adducts of valproic acid and plasma protein and anti-adduct antibody in humans. Biochem Pharmacol 1992; 43:745–55. 36. King AR, Dickinson RG. The utility of the bile-exteriorized rat as a source of reactive acyl glucuronides: studies with zomepirac. J Pharmacol Toxicol Methods 1996; 36:131–6. 37. Hermening A, Gra¨fe AK, Baktir G, Mutschler E, Spahn-Langguth H. Gemfibrozil and its metabolites—Quantification of aglycones, acyl glucuronides, and covalent adducts in samples from preclinical and clinical kinetic studies. J Chromatogr B, Biomed Appl 2000; 741:129–44. 38. Sinclair KA, Caldwell J. The formation of b-glucuronidase resistant glucuronides by intramolecular rearrangement of glucuronic acid conjugates at mild alkaline pH. Biochem Pharmacol 1982; 31:953–7.

Mechanistic Role of Acyl Glucuronides

153

39. Hyneck ML, Smith PC, Unseld E, Benet LZ. High-performance liquid chromatographic determination of tolmetin glucuronide and its isomeric conjugates in plasma and urine. J Chromatogr 1987; 420:349–56. 40. Blanckaert N, Compernolle F, Leroy P, Van Hautte R, Fevery J, Heirwegh KPM. The fate of bilirubinIX a glucuronide in cholestasis and during storage in vitro. Biochem J 1978; 171:203–14. 41. Hansen-Moller J, Cornett C, Dalgaard L, Hansen SH. Isolation and identification of the rearrangement products of diflunisal 1-O-acyl glucuronide. J Pharm Biomed Anal 1988; 6:229–40. 42. Dickinson RG, Baker PV, King AR. Studies of the reactivity of acyl glucuronides. VII. Salicyl acyl glucuronide reactivity in vitro and covalent binding of salicylic acid to plasma protein of human taking aspirin. Biochem Pharmacol 1994; 47:469–76. 43. Sekikawa H, Yagi N, Lin ET, Benet LZ. Apparent intramolecular acyl migration and hydrolysis of furosemide glucuronide in aqueous solution. Biol Pharm Bull 1995; 18:134–9. 44. Spahn H, Iwakawa S, Benet LZ. Stereoselective formation and degradation of flunoxaprofen glucuronides in microsomal incubations. Pharm Res 1988; 5(Suppl.):S200. 45. Dahms M, Spahn-Langguth H. Covalent binding of acidic drugs via reactive intermediates: detection of benoxaprofen and flunoxaprofen adducts in biological material. Die Pharmazie 1996; 51:874–81. 46. Spahn H, Iwakawa S, Lin ET, Benet LZ. Procedures to characterize in vitro and in vivo enantioselective glucuronidation properly: studies with benoxaprofen glucuronides. Pharm Res 1989; 6:125–32. 47. Volland C, Benet LZ. In vitro enantioselective glucuronidation of fenoprofen. Pharmacology 1991; 43:53–60. 48. Fournel-Gigleux S, Hammar-Hansen C, Motassim N, et al. Substrate specific and enantioselectivity of arylcarboxylic acid glucuronidation. Drug Metab Dispos 1988; 16:627–34. 49. el Mouelhi M, Ruelius HW, Fenselau C, Dulik DM. Species-dependent enantioselective glucuronidation of three 2-arylpropionic acids, naproxen, ibuprofen, and benoxaprofen. Drug Metab Dispos 1987; 15:767–72. 50. Georges H, Persle N, Buronfosse T, Fournel-Gigleux S, Netter P, Magdalou J, Lapicque F. In vitro stereoselective degradation of carprofen glucuronide by human serum albumin: characterization of sites and reactive amino acids. Chirality 2000; 12:53–62. 51. Buszewski B, el Mouelhi M, Albert K, Bayer E. Influence of the structure of chemically bonded C18 phase on HPLC separation of naproxen glucuronide diastereomers. J Liq Chromatogr 1990; 13:505–24. 52. Chakir S, Maurice M, Magdalou J, et al. High-performance liquid chromatographic enantioselective assay for the measurement of ketoprofen glucuronidation by liver microsomes. J Chromatogr B 1994; 654:61–8. 53. Iwakawa S, Spahn H, Benet LZ, Lin ET. Stereoselective binding of the glucuronide conjugates of carprofen enantiomers to human serum albumin. Biochem Pharmacol 1990; 39:949–53. 54. Yang DY, Benet LZ. Stability of diclofenac acyl glucuronide and irreversible binding to plasma proteins in vitro. Pharm Res 1995; 12:S–377. 55. Dickinson RG, King AR. Reactivity considerations in the analysis of glucuronide and sulfate conjugates of diflunisal. Ther Drug Monit 1989; 11:712–20. 56. Sallustio BC, Fairchild BA. Biosynthesis, characterization and direct high-performance liquid chromatographic analysis of gemfibrozil 1-O-b-acylglucuronide. J Chromatogr B 1995; 665:345–53. 57. Illing HP, Wilson ID. pH-Dependent formation of b-glucuronidase resistant conjugates from the biosynthetic ester glucuronide of isoxepac. Biochem Pharmacol 1981; 30:3381–4. 58. Bischer A, Zia-Amirhosseini P, Iwaki M, McDonagh AF, Benet LZ. Stereoselective binding properties of naproxen glucuronide diastereomers to proteins. J Pharmacokinet Biopharm 1995; 23:379–95. 59. Hansen-Moller J, Schmit U. Rapid high-performance liquid chromatographic assay for the simultaneous determination of probenecid and its glucuronide in urine. Irreversible binding of probenecid to serum albumin. J Pharm Biomed Anal 1991; 9:65–73. 60. Heirwegh KP, Compernolle F. Micro-analytic detection and structure elucidation of esterglycoside. Biochem Pharmacol 1979; 28:2109–14. 61. Compernolle F, Blanckaert N, Heirweg KPM. The fate of bilirubin-IX a glucuronides in cholestatic bile: sequential migration of the 1-acylaglycone to 2-, 3-, and 4-positions of glucuronic acid. Biochem Soc Trans 1977; 5:317–9. 62. Wen Z, Stern ST, Martin DE, Lee KH, Smith PC. Structural characterization of anti-HIV drug candidate PA-457 [3-O-(3 0 ,3 0 -dimethylsuccinyl)-betulinic acid] and its acyl glucuronides in rat bile and evaluation of in vitro stability in human and animal liver microsomes and plasma. Drug Metab Dispos 2006; 34:1436–42. 63. Dahms M, Lotz R, Lang W, Renner U, Bayer E, Spahn-Langguth H. Elucidation of phase-I- and phase-II metabolic pathways of rhein: species-differences and their potential relevance in toxicology. Drug Metab Dispos 1997; 25:442–52.

154

Spahn-Langguth et al.

64. Hermening A. Reactive metabolites of lipid-lowering agents—kinetics of acyl glucuronides of gemfibrozil and its phase-I metabolites. PhD thesis, Johann Wolfgang Goethe-Universita¨t Frankfurt/Main, 1998. 65. Smith PC, Benet LZ. Characterization of the isomeric esters of zomepirac glucuronide by proton NMR. Drug Metab Dispos 1986; 14:503–5. 66. Mortensen RW, Corcoran O, Cornett C, Sidelmann UG, Lindon JC, Nicholson JK, Hansen SH. SNaproxen-b-1-O-acyl glucuronide degradation kinetic studies by stopped-flow high-performance liquid chromatography–1H NMR and high-performance liquid chromatography–UV. Drug Metab Dispos 2001; 29:375–80. 67. Ebner T, Heinzel G, Prox A, Beschke K, Wachsmuth H. Disposition and chemical stability of telmisartan 1-O-acyl glucuronide. Drug Metab Dispos 1999; 27:1143–9. 68. Dong JQ, Liu J, Smith PC. Role of benoxaprofen and flunoxaprofen acyl glucuronides in covalent binding to rat plasma and liver proteins in vivo. Biochem Pharmacol 2005; 70:937–48. 69. Upton RA, Williams RL, Buskin JN, Jones RM. Effects of probenecid on ketoprofen kinetics. Clin Pharmacol Ther 1982; 31:705–12. 70. Benet LZ, Spahn H, Iwakawa H, et al. Prediction of covalent binding of acidic drugs in man. Life Sci 1993; 53:PL141–146. 71. Elgabarty H, Spahn-Langguth H. Different hydrolytic stabilities of acyl glucuronides: a study on structural parameters. BioVision—The World Life Sciences Forum 2006. Alexandria (Abstract). 72. Verloop A. The STERIMOL Approach to Drug Design. New York: Marcel Dekker, 1987. 73. Haine AH. Relative reactivities of hydroxyl groups in carbohydrates. Adv Carbohydr Chem Biochem 1976; 33:101–9. 74. Corcoran O, Mortensen RW, Hansen SH, Troke J, Nicholson JK. HPLC/1H NMR spectroscopic studies of the reactive a-1-O-acyl isomer formed during acyl migration of S-naproxen b-1-O-acyl glucuronide. Chem Res Toxicol 2001; 14:1363–70. 75. Akira K, Uchijima T, Hashimoto T. Rapid internal acyl migration and protein binding of synthetic probenecid glucuronides. Chem Res Toxicol 2002; 15:765–72. 76. Dickinson RG, Hooper WD, Eadie MJ. pH-Dependent rearrangement of the biosynthetic ester glucuronide of valproic acid to b-glucuronidase-resistant forms. Drug Metab Dispos 1984; 12:247–52. 77. Iwakawa S, Spahn H, Benet LZ, Lin ET. Carprofen glucuronide: stereoselective degradation and interaction with human serum albumin. Pharm Res 1988; 5(Suppl.):S–214. 78. van Breemen RB, Fenselau CC, Dulik DM. Activated phase II metabolites: comparison of alkylation by 1-O-acyl glucuronides and acyl sulfates. Adv Exp Med Biol 1986; 197:423–9. 79. Watt J, Dickinson RG. Reactivity of diflunisal acyl glucuronide in human and rat plasma and albumin solutions. Biochem Pharmacol 1990; 39:1067–75. 80. Williams AM, Dickinson RG. Studies on the reactivity of acyl glucuronides. Modulation of reversible and covalent interaction of diflunisal acyl glucuronide and its isomers with human plasma protein in vitro. Biochem Pharmacol 1994; 47:457–67. 81. Volland C, Sun H, Dammeyer J, Benet LZ. Stereoselective degradation of the fenoprofen acyl glucuronide enantiomers and irreversible binding to plasma protein. Drug Metab Dispos 1991; 19:1080–6. 82. Sallustio BC, Fairchild BA, Panall PR. Interaction of human serum albumin with the electrophilic metabolite 1-O-gemfibrozil-beta-D-glucuronide. Drug Metab Dispos 1997; 25:55–60. 83. Akira K, Taira T, Hasegawa H, Sakuma C, Shinohara Y. Studies on the stereoselective internal acyl migration of ketoprofen glucuronide using 13C labeling and nuclear magnetic resonance spectroscopy. Drug Metab Dispos 1998; 26:457–64. 84. Ruelius RW, Young EM, Kirkman SK, Schilings RT, Sisenwine SF, Jassen FW. Biological fate of acyl glucuronides in the rat: the role of rearrangement, intestinal enzyme and reabsorption. Biochem Pharmacol 1985; 34:451–2. 85. Tanaka Y, Suzuki A. Enzymatic hydrolysis of zenarestat 1-O-acyl glucuronide. J Pharm Pharmacol 1994; 46:235–9. 86. Sekikawa H, Yagi N, Oda K, et al. Biliary excretion of furosemide glucuronide in rabbits. Biol Pharm Bull 1995; 18:447–53. 87. Smith PC, Hasegawa J, Langendijk PJ, Benet LZ. Stability of acyl glucuronides in blood, plasma, and urine: studies with zomepirac. Drug Metab Dispos 1985; 13:110–2. 88. Wells DS, Janssen FW. Interactions between oxaprozin glucuronide and human serum albumin. Xenobiotica 1987; 17:1437–49. 89. Hayball PJ, Nation RL, Bochner F. Stereoselective interactions of ketoprofen glucuronides with human plasma and serum albumin. Biochem Pharmacol 1992; 44:291–9. 90. Mizuma T, Benet LZ, Lin ET. Interaction of human serum albumin with furosemide glucuronide: a role of albumin in isomerization, hydrolysis, reversible binding and irreversible binding of a 1-Oacyl glucuronide metabolite. Biopharm Drug Dispos 1999; 20:131–6. 91. Sudlow G, Birkett DJ, Wade DN. Further characterization of specific drug binding sites on human serum albumin. Mol Pharmacol 1976; 12:1052–61.

Mechanistic Role of Acyl Glucuronides

155

92. Ojingwa JC, Spahn-Langguth H, Benet LZ. Reversible binding of tolmetin, zomepirac, and their glucuronide conjugates to human serum albumin and plasma. J Pharmacokinet Biopharm 1994; 22:19–40. 93. Bischer A, Iwaki M, Zia-Amirhosseini P, Benet LZ. Stereoselective reversible binding properties of the glucuronide conjugates of fenoprofen enantiomers to human serum albumin. Drug Metab Dispos 1995; 23:900–3. 94. Kuenzle CC, Maier C, Ruttner JR. The nature of four bilirubin fractions from serum and of three bilirubin fractions from bile. J Lab Clin Med 1966; 67:294–306. 95. Smith PC, McDonagh AF, Benet LZ. Irreversible binding of zomepirac to plasma in vitro and in vivo. J Clin Invest 1986; 77:934–9. 96. Munafo A, Hyneck ML, Benet LZ. Pharmacokinetics and irreversible binding of tolmetin and its glucuronic acid esters in the elderly. Pharmacology 1993; 47:309–17. 97. Zia-Amirhosseini P, Ojingwa J, Spahn-Langguth H, McDonagh AF, Benet LZ. Enhanced covalent binding of tolmetin to proteins in humans after multiple dosing. Clin Pharmacol Ther 1994; 55:21–7. 98. McKinnon GE, Dickinson RG. Covalent binding of diflunisal and probenecid to plasma protein in humans: persistence of the adducts in the circulation. Res Commun Chem Pathol Pharmacol 1989; 3:1033–9. 99. Sallustio BC, Knights KM, Roberts BJ, Zacest R. In vivo covalent binding of clofibric acid to human plasma proteins and rat liver proteins. Biochem Pharmacol 1991; 42:1421–5. 100. Shipkova M, Beck H, Voland A, Armstrong VW, Grone HJ, Oellerich M, Wieland E. Identification of protein targets for mycophenolic acid acyl glucuronide in rat liver and colon tissue. Proteomics 2004; 4:2728–38. 101. Pohl LR, Branchflower RV. Covalent binding of electrophilic metabolites to macromolecules. Methods Enzymol 1981; 77:43–50. 102. Zhou S. Separation and detection methods for covalent drug protein adducts. J Chromatogr B Analyt Technol Biomed Life Sci 2003; 797:63–90. 103. Pumford NR, Halmes NC, Hinson JA. Covalent binding of xenobiotics to specific proteins in the liver. Drug Metab Rev 1997; 29:39–57. 104. Cohen SD, Pumford NR, Khairallah EA, Boekelheide K, Pohl LR, Amouzadeh HR, Hinson JA. Selective protein covalent binding and target organ toxicity. Toxicol Appl Pharmacol 1997; 143:1–12. 105. Pumford NR, Myers TG, Davila JC, Highet RJ, Pohl LR. Immunochemical detection of liver protein adducts of the nonsteroidal antiinflammatory drug diclofenac. Chem Res Toxicol 1993; 6:147–50. 106. Wel A, Guichard JP, Caldwell J. In: Siest G, Magdalou J, Burchell B, eds. Interactions between fenofibryl glucuronides and human serum albumin or human plasma. Cellular and Molecular Aspect of Glucuronidation. Vol. 173. Montrouge, France: Colloque INSERM/John Libbey Eurotext Ltd, 1988:233–6. 107. Dubois N, Lapicque F, Maurice M, et al. In vitro irreversible binding of ketoprofen glucuronide to plasma proteins. Drug Metab Dispos 1993; 21:617–23. 108. Dickinson RG, King AR. Studies on the reactivity of acyl glucuronides. II. Interaction of diflunisal acyl glucuronide and its isomers with human serum albumin in vitro. Biochem Pharmacol 1991; 42:2301–6. 109. Ojingwa JC, Spahn-Langguth H, Benet LZ. Irreversible binding of tolmetin to macromolecules via its glucuronide: binding to blood constituents, tissue homogenates and subcellular fractions in vitro. Xenobiotica 1994; 24:495–506. 110. Bailey MJ, Worrall S, de Jersey J, Dickinson RG. Zomepirac acyl glucuronide covalently modifies tubulin in vitro and in vivo and inhibits its assembly in an in vitro system. Chem Biol Interact 1998; 115:153–66. 111. Kretz-Rommel A, Boelsterli UA. Mechanism of covalent adduct formation of diclofenac to rat hepatic microsome proteins, retention of the glucuronic acid moiety in the adduct. Drug Metab Dispos 1994; 22:956–61. 112. McDonagh AF, Palma LA, Lauff JJ, Wu T-W. Origin of mammalian biliprotein and rearrangement of bilirubin glucuronide in vivo in the rat. J Clin Invest 1984; 74:763–70. 113. van Breemen RB, Fenselau C, Mogilevsky W, Odell GB. Reaction of bilirubin glucuronides with serum albumin. J Chromatogr 1986; 383:287–392. 114. Garlick RL, Mazer JS. The principal site of non-enzymatic glycosylation of human serum albumin in vivo. J Biol Chem 1983; 258:6142–6. 115. Ding A, Ojingwa JC, McDonagh AF, Burlingame AL, Benet LZ. Evidence for covalent binding of acyl glucuronides to serum albumin via an imine mechanism as revealed by tandem mass spectrometry. Proc Natl Acad Sci USA 1993; 90:3797–801. 116. Ding A, Zia-Amirhosseini P, McDonagh AF, Burlingame AL, Benet LZ. Reactivity of tolmetin glucuronide with human serum albumin, identification of binding sites and mechanisms of reaction by tandem mass spectrometry. Drug Metab Dispos 1995; 23:369–75. 117. Qiu Y, Burlingame A, Benet LZ. Mechanisms for covalent binding of benoxaprofen glucuronide to human serum albumin, studies by tandem mass spectrometry. Drug Metab Dispos 1998; 26:246–56.

156

Spahn-Langguth et al.

118. Iwaki M, Ogiso T, Inagawa S, Kakehi K. In vitro regioselective stability of beta-1-O- and 2-Oglucuronides of naproxen and their covalent binding to human serum albumin. J Pharm Sci 1999; 88:52–7. 119. Sallustio BC, Foster DJ. Reactivity of gemfibrozil 1-O-b-acyl glucuronide, pharmacokinetics of covalently bound gemfibrozil–protein adducts in rats. Drug Metab Dispos 1995; 23:892–9. 120. Spahn-Langguth H, Gra¨fe AK, Mayer S, Bu¨schges R, Benet LZ. Fibrates: the occurrence of active and reactive metabolites. In: Reid E, Hill HM, Wilson ID, eds. Biofluid and Tissue Analysis of Drugs Including Hyperlipidemics. Methodological Surveys in Bioanalysis of Drugs. Vol. 23. Cambridge, Lodon: The Royal Society of Chemistry, 1994:87–102. 121. Smith PC, Langendijk PNJ, Bosso JA, Benet LZ. Effect of probenecid on the formation and elimination of acyl glucuronides: studies with zomepirac. Clin Pharmacol Ther 1985; 38:121–7. 122. King AR, Dickinson RG. Studies on the reactivity of acyl glucuronides. IV. Covalent binding of diflunisal to tissues of rat. Biochem Pharmacol 1993; 45:1043–7. 123. Williams AM, Worrall S, Jersey JD, Dickinson RG. Studies on the reactivity of acyl glucuronides. VIII. Generation of an antiserum for the detection of diflunisal-modified proteins in diflunisal-dosed rats. Biochem Pharmacol 1995; 49:209–17. 124. Dickinson RG, King AR. Studies on the reactivity of acyl glucuronides. V. Glucuronide-derived covalent binding of diflunisal to bladder tissue of rats and its modulation by urine pH and b-glucuronidase. Biochem Pharmacol 1993; 46:1175–82. 125. Wade LT, Kenna JG, Caldwell J. Immunochemical identification of mouse hepatic protein adducts derived from the nonsteroidal anti-inflammatory drugs diclofenac, sulindac, and ibuprofen. Chem Res Toxicol 1997; 10:546–55. 126. Bailey MJ, Dickinson RG. Chemical and immunochemical comparison of protein adduct formation of four carboxylate drugs in rat liver and plasma. Chem Res Toxicol 1996; 9:659–66. 127. Sabordo L, Sallustio BC, Evans AM, Nation RL. Hepatic disposition of the acyl glucuronide 1-Ogemfibrozil-beta-D-glucuronide: effects of dibromosulfophthalein on membrane transport and aglycone formation. J Pharmacol Exp Ther 1999; 288:414–20. 128. Seitz S, Kretz-Rommel A, Elferink RPJO, Boelsterli UA. Selective protein adduct formation of diclofenac glucuronide is critically dependent on the rat canalicular conjugate export pump (Mrp2). Chem Res Toxicol 1998; 11:513–9. 129. Laouari D, Yang R, Veau C, Blanke I, Friedlander G. Two apical multidrug transporters, P-gp and MRP2, are differently altered in chronic renal failure. Am J Physiol Renal Physiol 2001; 280:F636–45. 130. Grillo MP, Chiellini G, Tonelli M, Benet LZ. Effect of alpha-fluorination of valproic acid on valproylS-acyl-CoA formation in vivo in rats. Drug Metab Dispos 2001; 29:1210–5. 131. Bailey MJ, Dickinson R. Acyl glucuronide reactivity in perspective: biological consequences. Chem Biol Interact 2003; 145:117–37. 132. Spahn-Langguth H, Dahms M, Hermening A. Acyl glucuronides: covalent binding and its potential relevance. In: Snyder R, Greim H, eds. Proceedings of the 5th International Meeting in Biological Reactive Intermediates (BRI-V) 1995, Mu¨nchen. New York: Plenum Press, 1996. 133. Kitteringham NR, Maggs JL, Newby S, Park BK. Drug-protein conjugates. VIII. The metabolic fate of the dinitrophenyl hapten conjugated to albumin. Biochem Pharmacol 1995; 34:1763–71. 134. Sallustio BC, DeGraaf YC, Weekley JS, Burcham PC. Bioactivation of carboxylic acid compounds by UDP-glucuronosyltransferases to DNA-damaging intermediates: role of glycoxidation and oxidative stress in genotoxicity. Chem Res Toxicol 2006; 19:683–91. 135. Spahn H. Assay method for product formation in in vitro enzyme kinetic studies of uridine disphosphate glucuronyltransferases: 2-arylpropionic acid enantiomers. J Chromatogr 1988; 430:368–75. 136. Hargus SJ, Amouzedeh HR, Pumford NR, Myers TG, McCoy SC, Pohl LR. Metabolic activation and immunochemical localization of liver protein adducts of the nonsteroidal anti-inflammatory drug diclofenac. Chem Res Toxicol 1994; 7:575–82. 137. Kretz-Rommel A, Boelsterli UA. Selective protein adducts to membrane proteins in cultured rat hepatocytes exposed to diclofenac: radiochemical and immunochemical analysis. Mol Pharmacol 1994; 45:237–44. 138. Gil ML, Ramirez MC, Terencio MC, Castell JV. Immunochemical detection of protein adducts in cultured human hepatocytes exposed to diclofenac. Biochim Biophys Acta 1995; 1272:140–6. 139. Shen TY. Nonsteroidal anti-inflammatory agents. Burger’s Medicinal Chemistry. In: Wolf ME, ed. 4th ed. Part III. New York: Wiley Interscience, 1981:1205–71. 140. Spahn-Langguth H, Benet LZ, Zia-Amirhosseini P, Iwakawa H, Langguth P. Kinetics of reactive phase II metabolites: stereochemical aspects of formation of epimeric acyl glucuronides and their reactivity. In: Aboul-Enein HY, Wainer IW, eds. The Impact of Stereochemistry on Drug Development and Use. Chemical Analysis Series, Vol. 142. New York: Wiley, 1997:125–70.

Mechanistic Role of Acyl Glucuronides

157

141. Spahn H, Benet LZ. Enantioselectivity of hepatic UDP-glucuronyltransferase in rat liver microsomes towards 2-arylpropionic acids: glucuronidation of naproxen enantiomers. In: Aiache JM, Hirtz J, eds. Third European Congress on Biopharmaceutics and Pharmacokinetics Proceedings, Vol. II Freiburg: Experimental Pharmacokinetics, 1987:261–8. 142. Hayball P. Formation and reactivity of acyl glucuronides: the influence of chirality. Chirality 1995; 7:1–9. 143. Spahn H, Iwakawa S, Ojingwa J, Benet LZ. Glucuronidation of flunoxaprofen enantiomers by UDPGTs from different sources. In: International Conference on Pharmaceutical Sciences and Clinical Pharmacology. Jerusalem, May/June 1988. 144. Bolze S, Bromet N, Gay-Feutry C, Massiere F, Boulieu R, Hulot T. Development of an in vitro screening model for the biosynthesis of acyl glucuronide metabolites and the assessment of their reactivity toward human serum albumin. Drug Metab Dispos 2002; 30:404–13. 145. Hyneck ML, Smith PC, Munafo A, McDonagh AF, Benet LZ. Disposition and irreversible plasma protein binding of tolmetin in humans. Clin Pharmacol Ther 1988; 44:107–14. 146. Pumford NR, Halmes NC. Protein targets of xenobiotic reactive intermediates. Annu Rev Pharmacol Toxicol 1997; 37:91–117. 147. Pirmohamed M, Madden S, Park BK. Idiosyncratic drug reactions metabolic bioactivation as a pathogenic mechanism. Clin Pharmacokinet 1996; 31:215–30. 148. Pohl LR, Satoh H, Christ DD, Kenna JG. The immunological and metabolic basis of drug hypersensitivities. Annu Rev Pharmacol 1988; 28:367–87. 149. Benet LZ, Spahn H. Acyl migration and covalent binding of drug glucuronides—potential toxicity mediators. In: Siest G, Magdalou J, Burchell B, eds. Cellular and Molecular Aspects of Glucuronidation. Vol. 173. Montrougc, France: Colloque INSERM/John Libbey Eurotext Ltd, 1988:261–9. 150. Amos HE, Wilson DV, Taussig MJ, Carlton SJ. Hypersensitivity reactions to acetylsalicylic acid. Clin Exp Immunol 1971; 8:563–72. 151. Zia-Amirhosseini P, Harris R, Brodsky FM, Benet LZ. Hypersensitivity to nonsteroidal antiinflammatory drugs. Nat Med 1995; 1:2–4. 152. Kretz-Rommel A, Boelsterli UA. Cytotoxic toxicity of T cells and non-T cells from diclofenacimmunized mice against cultured syngeneic hepatocytes exposed to diclofenac. Hepatology 1995; 22:213–22. 153. Garcia Rodriguez LA, Williams R, Derby LE, Dean AD, Jick H. Acute liver injury associated with nonsteroidal anti-inflammatory drugs and the role of risk factor. Arch Intern Med 1994; 154:311–6.

9

Nonparenchymal Cells, Inflammatory Macrophages, and Hepatotoxicity Debra L. Laskin and Carol R. Gardner

Department of Pharmacology and Toxicology, Rutgers University, Piscataway, New Jersey, U.S.A.

INTRODUCTION Considerable evidence has accumulated over the past several years demonstrating that hepatotoxicity induced by a diverse group of drugs and chemicals is due not only to a direct effect of these compounds on the liver, but also indirectly to the actions of inflammatory mediators released by nonparenchymal cells, in particular hepatic sinusoidal cells, including macrophages, endothelial cells, and stellate cells, as well as infiltrating leukocytes. Following exposure of experimental animals to hepatotoxicants, these cells become “activated.” This involves alterations in their functional and biochemical properties leading to the release of an array of proinflammatory and cytotoxic mediators that have the capacity to promote liver damage. These findings, together with the observation that hepatotoxicity can be modified by agents that modulate inflammatory cell and mediator activity, provide direct evidence that these cells contribute to tissue injury. The mediators involved in the cytotoxic process include reactive oxygen and reactive nitrogen intermediates, proinflammatory cytokines, proteolytic enzymes, eicosanoids, and/or bioactive lipids released at sites of injury. Whereas some of these mediators are directly cytotoxic (e.g., hydrogen peroxide, nitric oxide, peroxynitrite), others degrade the extracellular matrix (e.g., collagenase, elastase) and/or promote inflammatory cell adhesion and infiltration, and nonparenchymal cell proliferation and activation (e.g., interleukin-1 (IL-1), interleukin-6 (IL-6), tumor necrosis factor-a (TNFa), transforming growth factor-b, (TGFb), platelet-activating factor (PAF), chemokines, and colony-stimulating factors). There is also evidence that some of the mediators produced by activated nonparenchymal cells and inflammatory macrophages can modify hepatocyte protein and nucleic acid biosynthesis, as well as cytochrome P450–mediated xenobiotic metabolism, which may also contribute to hepatotoxicity. In this chapter, experimental evidence implicating nonparenchymal cells and inflammatory macrophages and mediators produced by these cells in hepatotoxicity is reviewed. HEPATIC NONPARENCHYMAL CELLS AND TISSUE INJURY The liver is comprised of two main cell populations: parenchymal cells or hepatocytes and nonparenchymal cells, the majority of which reside within the hepatic sinusoids, positioned between the arterial vasculature and the parenchyma. Sinusoidal cells consist mainly of Kupffer cells, endothelial cells, and stellate cells. Small populations of pit cells, the natural killer (NK) cells of the liver, natural killer T (NKT) cells, and dendritic cells are also found within the hepatic sinusoids. In addition to their normal physiological function in the liver, evidence suggests that each of these sinusoidal cell populations has the potential to contribute to xenobiotic-induced liver injury. Moreover, cross talk between these cells may augment their protoxicant activity. Kupffer Cells and Inflammatory Macrophages Kupffer cells constitute approximately 20% of the hepatic sinusoidal cells and represent the largest population (80–90%) of all the macrophages in the body. They are predominantly localized in periportal and central regions of the liver lobule and anchored to the lumen of the

160

Laskin and Gardner

endothelium by long cytoplasmic processes (1). Thus, they are well positioned to remove particulate and foreign materials from the portal circulation, primarily through the process of phagocytosis. Kupffer cells possess Fc, C3, and CRIg receptors, as well as scavenger receptors, carbohydrate receptors, pattern recognition receptors, such as toll-like receptor-4 (TLR-4), and cell adhesion molecules, which facilitate their ability to phagocytose opsonized and nonopsonized particles, apoptotic and damaged cells, neutrophils, and tumor cells (2–18). A major function of Kupffer cells is uptake and detoxification of gut-derived endotoxin (19,20). This is accomplished by binding of lipopolysaccharide (LPS), the toxic moiety in endotoxin, in association with LPS-binding protein, to CD14, TLR-4, and/or scavenger receptors such as macrosialin (CD68) (4,13,21–24). Kupffer cells are among the most active secretory cells in the body releasing hundreds of different products with inflammatory, growth promoting, and regulatory activity. These include superoxide anion, hydrogen peroxide, nitric oxide, peroxynitrite, proteolytic enzymes, and eicosanoids that aid in antigen destruction (7,12,25–32). They also release a number of different cytokines with immunoregulatory and proinflammatory actions, including TNFa, IL-1, IL-6, IL-8, IL-10, IL-18, PAF, TGFb, MIP-2, and interferon (12,32–37). Kupffer cells also elaborate growth factors involved in regulating the proliferation of hepatocytes, endothelial cells, stellate cells as well as inflammatory macrophages (30,32,38–41). The ability of liver macrophages to release mitogenic factors for hepatocytes suggests that these cells may also participate in the wound repair process (42–44). Because of their continuous exposure to endotoxin in the portal circulation, Kupffer cells are in a constant state of activation and are therefore primed to respond to tissue injury. Thus, after exposure to inflammatory stimuli, Kupffer cells exhibit markedly-increased chemotactic and phagocytic activity and display significantly greater oxidant-dependent and oxidantindependent cytotoxicity (6–9,25–29,45–49). Moreover, release of cytotoxic and proinflammatory mediators by these cells is greatly increased. These findings, together with the observation that hepatic macrophages increase in number in response to tissue injury, suggest that these cells have the capacity to modulate both normal and pathological processes in the liver. Recent studies have demonstrated that Kupffer cells, like resident macrophages present in other tissues, express major histocompatibility complex (MHC) class II antigens and act as antigenpresenting cells for the induction of specific T-lymphocyte responses (50–53). In addition, Kupffer cells have been reported to be required for the recruitment of dendritic cells into the liver resulting in increased number of locally available antigen-presenting cells (54). These findings demonstrate that Kupffer cells also contribute to specific immune responses of the liver to antigens. A growing body of literature has been generated over the past several years that has provided strong evidence implicating liver macrophages in hepatotoxicity induced by a diverse group of agents (Table 1). These include acetaminophen, endotoxin, carbon tetrachloride, galactosamine, 1,2-dichlorobenzene, allyl alcohol, cadmium, thioacetamide, and ethanol. With each of these compounds, hepatotoxicity is abrogated or prevented by pretreatment of experimental animals with agents such as gadolinium chloride, which block resident Kupffer cell activity. The fact that the mechanisms underlying tissue injury induced by these different toxicants are distinct suggests that an involvement of macrophages may be a critical step in the pathogenic process leading to hepatotoxicity. One of the first lines of evidence suggesting that macrophages contribute to hepatotoxicity is based on the observation that there are increased number of these cells in the liver following exposure of animals to hepatotoxicants (7,26,66,74,99,100,109,114). These cells are typically observed in the liver prior to histological evidence of frank necrosis. Moreover, their specific location within the liver lobule varies with the chemical agent and is directly correlated with areas of the tissue that subsequently exhibit signs of injury. For example, after administration of acetaminophen, carbon tetrachloride, or thioacetamide, agents that induce centrilobular hepatic necrosis, macrophages are observed in these regions of the liver (26,55,75,76,111). In contrast, macrophages that localize in the liver following endotoxin, phenobarbital, Corynebacterium parvum, or galactosamine treatment of rats are scattered in clusters throughout the liver lobule, which is consistent with patterns of injury observed after exposure to these toxins (7,66,101,114,115).

161

Nonparenchymal Cells and Inflammatory Macrophages

TABLE 1 Agents Whose Toxicity Is Associated with Macrophages and Inflammatory Mediators Toxicant Acetaminophen Endotoxin Carbon tetrachloride Ethanol 1,2-Dichlorobenzene Galactosamine Cadmium Allyl alcohol Thioacetamide

Response of liver macrophages

References

Increased number, chemotaxis, phagocytosis, cytotoxicity, ROI, RNI, IL-1, TNFa, HO-1, chemokines, eicosanoids Increased number, chemotaxis, phagocytosis, ROI, RNI, IL-1, TNFa, IL-6, IL-18, PAF, lipids, ICAM-1, COX-2 Increased number, ROI, RNI, IL-1, IL- 10, TNFa, IL-6, TGFb Increased ICAM-1, ROI, RNI, IL-1, TNFa, MIP-2, CINC, TGFb, IL-8, MCP-1, RANTES Increased ROI Increased number, increased ROI, TNFa Increased phagocytosis, IL-1, TNFa, CINC Increased ROI, TNFa Increased number, ROI, HSP70, TNFa, IL-6, TGFb, myeloperoxidase; Decreased IGFBP-3

(26,55 65) (6 9,25,28 30,33,35,46,49,66 73) (74 84) (85 94) (95 98) (47,99 105) (106 108) (109,110) (111 113)

Abbreviations: ROI, reactive oxygen intermediates; RNI, reactive nitrogen intermediates; IL, interleukin; TNFa, tumor necrosis factor-a; HO-1, heme oxygenase-1; PAF, platelet-activating factor; ICAM-1, intercellular adhesion molecule-1; COX-2, cyclooxygenase-2; TGFb, transforming growth factor-b; MIP-2, macrophage inflammatory protein-2; CINC, cytokine-induced neutrophil chemoattractant; MCP1, monocyte chemotactic protein-1; RANTES, regulated on activation, normal T cell expressed, and secreted; HSP70, heat shock protein-70; IGFBP-3, insulin-like growth factor binding protein-3.

Inflammatory cell accumulation in tissues is generally considered to be a relatively early marker of tissue injury. It is likely that the cells accumulating in the liver after hepatotoxicant exposure consist of both resident Kupffer cells and mononuclear phagocytes that have infiltrated into the tissue in response to damage. Both of these cell populations are highly sensitive to early-response cytokines such as TNFa and IL-1, rapidly generated at sites of tissue injury and become activated. Under homeostatic conditions, macrophage activation is carefully regulated. However, following exposure of experimental animals to hepatotoxicants, resident Kupffer cells and inflammatory macrophages may become “overactivated” or “hyperresponsive” and produce excessive quantities of cytotoxic mediators. In this regard, a number of studies have demonstrated that macrophages isolated from livers of hepatotoxicant-treated animals exhibit morphological and functional properties of activated mononuclear phagocytes (Fig. 1) (116). Thus, these cells appear larger and more stellate than cells from untreated rats, are highly vacuolated, and display an increased cytoplasmic:nuclear ratio (7,25,26,114). In addition, macrophages from animals treated with hepatotoxicants such as phenobarbital, acetaminophen, or endotoxin adhere to and spread on culture dishes more rapidly than resident Kupffer cells (7,26,114). These properties are characteristic of morphologically activated macrophages. Liver macrophages from animals treated with hepatotoxicants also exhibit varying degrees of functional activation including increased expression of cell adhesion molecules, and enhanced phagocytic, chemotactic, cytotoxic, and metabolic activity, as well as increased release of superoxide anion, hydrogen peroxide, nitric oxide, peroxynitrite, proteolytic enzymes,

(A)

(B)

FIGURE 1 Transmission electron micrographs of Kupffer cells from (A) control and (B) endotoxin-treated mice. Original magnification: 1500X.

162

Laskin and Gardner

eicosanoids, IL-1, IL-6, TNFa, and chemokines (7,9,25,26,33,45–49,77,78,85,114,115,117,118). Activated Kupffer cells and infiltrating macrophages are thought to promote hepatic damage through the release of these toxic secretory products (119). A second line of evidence supporting a role for macrophages in hepatotoxicity is derived from experiments in which animals are pretreated with agents that either inhibit or enhance macrophage activity and tissue injury is then assessed. Data from these studies clearly demonstrate that the degree of hepatic injury induced by a number of different chemicals is directly correlated with macrophage functioning. Thus, agents that depress macrophage functioning reduce toxicity, while compounds that augment macrophage activity enhance tissue injury. For example, drugs such as hydrocortisone, dextromethorphan, certain synthetic steroids, and natural substances that block inflammatory responses have been reported to protect against liver injury induced by carbon tetrachloride, endotoxin/galactosamine, and acetaminophen (120–122). Similarly, the accumulation of macrophages in the liver and subsequent toxicity of acetaminophen, carbon tetrachloride, or endotoxin is prevented by pretreatment of animals with gadolinium chloride, carbon particles, dextran sulfate, or liposome-encapsulated dichloromethylene diphosphonate, compounds known to depress macrophage activity (123–127). Hepatoprotective effects of gadolinium chloride against 1,2-dichlorobenzene, diethyldicarbamate, galactosamine, ethanol, endotoxin, allyl alcohol, cadmium, thioacetamide, and fumonisin B1 induced injury have also been described (87,95,106,107,109,128–132). Studies have also demonstrated that activation of hepatic macrophages augments liver injury induced by toxic xenobiotics. Thus, pretreatment of rats with macrophage activators such as endotoxin, glucan, vitamin A, or latex beads aggravates liver injury induced by carbon tetrachloride, galactosamine, allyl alcohol, and C. parvum (28,46,77,78,102,103,130,133–135). Taken together, these observations support the hypothesis that macrophages contribute to hepatotoxicity. The specific mediators released by these cells that are involved in the pathogenic process appears to depend on the nature of the hepatotoxicant as well as the levels of the mediator generated in the tissue, the timing of its release, and the extent to which other inflammatory signals are generated. A number of recent studies have suggested that some populations of macrophages accumulating in the liver in response to hepatotoxicant-induced injury play a role in the resolution of inflammation and in initiation of tissue repair. These infiltrating macrophage subpopulations are antigenically and phenotypically different from resident Kupffer cells (64) and release a distinct spectrum of anti-inflammatory and growth-promoting mediators (136–139). Moreover, blocking the recruitment of these macrophage subpopulations into the liver augments tissue injury induced by acetaminophen as well as carbon tetrachloride (62,64,136,140). These observations support a dual role of macrophages in hepatotoxicity, initially contributing to tissue injury and subsequently orchestrating tissue repair (43,136,140). Recent studies have focused on analyzing mechanisms regulating macrophage and endothelial cell activation following hepatotoxicant exposure. It has been suggested that this process involves inappropriate or prolonged activation of biochemical signaling pathways in the cells leading to increased gene expression and inflammatory mediator production. For example, following acetaminophen administration to animals, a rapid increase in nuclear binding activity of the transcription factor, NF-kB has been observed in the liver (141–143). Similar effects have been described after treatment of animals with carbon tetrachloride, ethanol, endotoxin, or galactosamine (118,144–150). NF-kB is a ubiquitous transcription factor known to regulate the activity of numerous genes involved in inflammatory responses including inducible nitric oxide synthase (NOSII), cyclooxygenase-2 (COX-2), TNFa, and intercellular adhesion molecule-1 (ICAM-1) (151). Enhanced NF-kB activity induced by toxicants presumably modulates liver injury through an effect on the synthesis of these mediators (152,153). This idea is supported by the finding that mice lacking the p50 subunit of NF-kB do not generate TNFa and are protected from carbon tetrachloride–induced toxicity (79). In contrast, the loss of NF-kB p50 had no effect on acetaminophen-induced hepatotoxicity suggesting that NF-kB p50dependent responses do not play a major role in the pathogenesis of toxicity in this model (143). It is interesting to note that activation of the NF-kB pathway is not required for hepatocyte proliferation (154). However, in a model of diethylnitrosamine-induced hepatocarcinogenesis, NF-kB was found to control the transcription of mediators in Kupffer cells that induce

Nonparenchymal Cells and Inflammatory Macrophages

163

hepatocyte proliferation (155). These data demonstrate that the specific role of transcription factors such as NF-kB in the response to tissue injury depends on the hepatotoxicants and the cell type activated. Increased nuclear binding activity of the transcription factor alkaline phosphatase-1 (AP-1) has also been described in the liver after treatment of animals with acetaminophen, carbon tetrachloride, endotoxin, or cadmium (76,149,156–158). The fact that this activity is prevented by pretreatment of animals with gadolinium chloride demonstrates that Kupffer cells are important in this response (76,141,156). Other signaling molecules implicated in macrophage responsiveness to hepatotoxicants, and in particular their role in wound repair includes the transcription factors c-jun and c-fos which regulate a battery of genes involved in cell proliferation, as well as signal transducers and activator of transcription 3 (STAT3) and early growth reponse-1 (Egr-1). These transcription factors have been reported to be activated in the liver following administration of carbon tetrachloride, acetaminophen, endotoxin, or ethanol (147,150,159–162) and their activity may be important in determining the outcome of the hepatotoxic response. Cell surface receptors including TLR-4 and a novel family of related molecules known as triggering receptor on myeloid cells (TREM), as well as receptor tyrosine kinase, Ron/STK are also upregulated in the liver following exposure to toxicants such as endotoxin or alcohol (16,17,163,164). This has been linked to the activation of various mitogen-activated protein kinases, including JNK, p38 MAP kinase, and PI-3 kinase/AKT (161,164–167). These findings suggest a potential mechanism regulating transcription factor activity and production of inflammatory mediators in the liver by macrophages and other nonparenchymal cells following exposure to hepatotoxicants. Endothelial Cells Endothelial cells form the walls of the hepatic sinusoids and represent the major fraction of hepatic sinusoidal cells (approximately 50%). Unlike endothelial cells in other vascular beds, hepatic sinusoidal endothelial cells are devoid of basement membrane (168). Moreover, they possess pores or fenestrae, which provide an opportunity for direct contact between plasma and hepatocytes. Thus, sinusoidal endothelial cells function as a selective barrier between the blood and the liver parenchyma. Endothelial cells also possess unique “bristle-coated” membrane invaginations and vesicles, and lysosome-like vacuoles, and are thought to play a role in the clearance of macromolecules from the circulation. Through Fc, carbohydrate, scavenger receptors, and TLR, endothelial cells endocytose a variety of particles in the portal circulation, including glycoproteins, lipoproteins, albumin, lactoferrin, hyaluronic acid, collagen, and aldehyde-modified proteins (2,4,5,10,168–180). It has been reported that endocytosis is upregulated in endothelial cells when Kupffer cell functioning is impaired (170,172,173,181–183). In response to inflammatory cytokines and bacterially derived LPS, hepatic endothelial cells, like Kupffer cells, are activated to release mediators that regulate the function of parenchymal and nonparenchymal liver cells. These include chemokines, IL-1, IL-6, PAF, fibroblast growth factor, interferons, endothelin, eicosanoids, proteolytic enzymes, reactive oxygen, and nitrogen intermediates (25,30,33,184–190). In normal liver, endothelial cells have been found to express various adhesion molecules, including ICAM-1, ICAM-2, LFA-3, VLA-5, and VAP-1 (191). Expression of these molecules, as well as P-selectin, E-selectin, and vascular cell adhesion molecule-1 (VCAM-1), which facilitate inflammatory cell emigration into the liver, is also upregulated in sinusoidal endothelial cells activated following hepatotoxicant exposure (6,191–193). These studies suggest that endothelial cells are important in inflammatory responses in the liver. The observation that endothelial cells express CD40, CD54, CD80, CD86, and MHC class I and II molecules, which are markers of antigen-presenting cells, indicate that they may also play a role in immune surveillance and potentially in the development of tolerance in the liver (168,194,195). A number of studies have demonstrated that endothelial cells increase in number and become activated following exposure of experimental animals to hepatotoxicants, such as acetaminophen, endotoxin, or ethanol (25,29,33,57,190,196). Like activated hepatic macrophages, these cells appear larger and more granular than cells from untreated rats, and

164

Laskin and Gardner

produce increased amounts of reactive oxygen and nitrogen intermediates, eicosanoids, endothelin, IL-1, IL-6, TGFb, fibroblast growth factor, interferon, and connective tissue growth factor (CTGF) (25,57,164,186,188,190,197). Moreover, expression of cell adhesion molecules, as well as receptors for TLR-4, TREM-1, TREM-3, Ron/STK, TNFa, and IL-6, are upregulated on these cells and their proliferative capacity increases (6,30,164,198–206). Recent studies have demonstrated that sinusoidal endothelial cells also produce fibronectin in response to injury (207–209). The ability of endothelial cells to produce and respond to these mediators may represent an important mechanism by which they participate in inflammatory, immune, and regeneration reactions associated with hepatotoxicity. As observed in hepatic macrophages, endothelial cell activation in the liver after hepatotoxicants exposure is associated with the upregulation of various biochemical signaling pathways including p38 MAP kinase, as well as the transcription factors NF-kB and AP-1 (90,164,210) and these may contribute to increased endothelial cell responsiveness. Stellate Cells Stellate cells, also referred to as Ito cells, fat-storing cells, perisinusoidal cells, and lipocytes, constitute approximately 20% of the hepatic sinusoidal cells. These cells normally reside in a quiescent, resting state within the space of Disse between endothelial cells and hepatocytes or between hepatocytes. Morphologically, stellate cells resemble fibroblasts in that they possess numerous extensions, as well as dilated rough endoplasmic reticulum. Stellate cells store vitamin A, which is localized in intracellular lipid droplets in the form of retinyl esters (211,212). Stellate cells also have the capacity to synthesize large quantities of extracellular matrix proteins, including types I, III, and IV collagen, as well as matrix metalloproteinases (MMP) and tissue inhibitors of metalloproteinase, and there is evidence that they play a major role in collagen synthesis in both normal and fibrotic liver (213–217). It has been suggested that stellate cells can contribute to inflammatory responses in the liver. Following exposure of animals to toxicants such as ethanol or carbon tetrachloride, stellate cells undergo a process of activation (218–220). This involves a loss of lipid droplets and vitamin A storage capacity, migration to sites of liver injury, and transformation into highly proliferative myofibroblast-like cells (219,221,222). Activated stellate cells also express increased quantities of the cell adhesion molecules, ICAM-1, and VCAM-1, as well as receptors for C5a, endothelin, eicosanoids, TNFa, IL-1, and platelet-derived growth factor (PDGF) (223–229). They are also primed to release cytotoxic and inflammatory mediators including IL-1, IL-6, IL-10, PAF, colony-stimulating factor-1, nitric oxide, hydrogen peroxide, superoxide anion, eicosanoids, gelatinase, fibronectin, TGFb, endothelin, macrophage chemotactic protein-1, and CINC (215,230–237). Activation and transformation of stellate cells during the pathogenesis of tissue injury and fibrosis, as well as collagen deposition, appear to be mediated by cytokines and growth factors elaborated by parenchymal and nonparenchymal liver cells. These are largely divided into mitogenic mediators (TGFa, PDGF, IL-1, TNFa, and insulin-like growth factor) that stimulate proliferation and transformation of stellate cells and fibrogenic mediators including TGFb and IL-6 that induce collagen gene expression (228). Hepatic fibrosis represents the liver’s wound healing response to injury and is characterized by excessive accumulation of interstitial matrix components within the tissue. A number of factors have been proposed to initiate and perpetuate the fibrogenic process in stellate cells including the accumulation of inflammatory cytokines and growth factors, alterations in the extracellular matrix, and oxidative stress (238). Xenobiotics such as alcohol or carbon tetrachloride can induce fibrogenesis by activating stellate cells. This can occur through the generation of lipid peroxides from damaged hepatocytes and/or oxidants and cytokines released from activated Kupffer cells and inflammatory macrophages (239,240). Recent studies have shown that stellate cells have the capacity to phagocytose apoptotic bodies derived from damaged hepatocytes, which markedly stimulates expression of TGFb1 mRNA and collagen and induces NADPH oxidase (241,242). During the pathogenesis of fibrosis, stellate cells exhibit increased sensitivity to inflammatory mediators such as TNFa (243). This can enhance the production of chemotactic and fibrogenic mediators by liver cells and may contribute to the maintenance of an inflammatory infiltrate dominated by macrophages (240,244).

Nonparenchymal Cells and Inflammatory Macrophages

165

Lymphocytes NK and NKT cells are found in great abundance in the liver and make up nearly 20% of liver leukocytes (14,245). NK cells are important in innate immune responsiveness as they produce cytokines, chemokines, and can lyse transformed cells (245,246). NKT cells consist of a subpopulation of lymphocytes that coexpress NK and T-cell receptors (245,247). Both cell populations can be activated to mediate cytotoxic activity by producing large quantities of IFN-g and have been implicated in endotoxin, concanavalin A, and acetaminophen-induced liver injury (248–250). INFLAMMATORY MEDIATORS IMPLICATED IN HEPATOTOXICITY Among the more prominent proinflammatory and cytotoxic mediators that have been implicated in hepatotoxicity are cytokines, reactive oxygen intermediates, reactive nitrogen intermediates, bioactive lipids, and hydrolytic enzymes. These mediators are likely to act in concert to promote hepatotoxicity. Cytokines Cytokines are cell-derived proteins that act in an autocrine and paracrine manner to regulate immune and inflammatory responses. Hepatic nonparenchymal cells and inflammatory macrophages are known to release a number of different cytokines that may play a role in the pathogenesis of tissue injury. Whereas some of these promote inflammatory responses (e.g., IL-1, IL-6, TNFa, interferon-g, TGFb, chemokines), others exert anti-inflammatory activity (e.g., IL-4, IL-10, IL-13). The overall outcome of the inflammatory response depends on the balance between levels of pro- and anti-inflammatory cytokines that are generated in the liver. Proinflammatory Cytokines IL-1 and TNFa are low-molecular weight, multifunctional proteins that induce a number of both distinct and overlapping functions (251–253). They are produced in large part by macrophages and are thought to play a prominent role in initiating the inflammatory response. Both IL-1 and TNFa stimulate the production of chemotactic factors and upregulate expression of cell adhesion molecules, thus promoting phagocyte margination and emigration to sites of injury. IL-1 also exerts mitogenic effects on macrophages and endothelial cells and induces the release of prostaglandins (PG), metalloproteinases, and colony-stimulating factor (85,133,251–254). In the liver, IL-1 and TNFa activate Kupffer cells and infiltrating macrophages for cytotoxicity and stimulate the release of cytotoxic mediators, including reactive nitrogen intermediates and reactive oxygen intermediates (32,251–255). They also induce the release of IL-1, IL-6, colonystimulating factor, PAF, and eicosanoids from parenchyma1 and nonparenchymal liver cells (32,33). In conjunction with IL-6, IL-1 and TNFa regulate hepatocyte acute-phase protein and gene expression and cytochrome P450 activity (251–254,256,257). TNFa is unique among inflammatory cytokines in that it has the capacity to induce cytotoxicity directly. In hepatocytes, TNFa stimulates nitric oxide production and induces both necrosis and apoptosis (258–261). Recent studies have demonstrated that TNFa is also a potent mitogen and that it plays a key role in regulating hepatocyte proliferation (262–265). This suggests that TNFa may also be important in the maintenance of hepatic homeostasis and in the repair of the liver following injury. This is supported by findings that TNFa upregulates antioxidant defense and stimulates both nonparenchymal and parenchymal liver cells to produce mediators involved in extracellular matrix remodeling, including TGFb, CTGF, and MMP-9, following acetaminophen-induced injury (142,161,266). Cytokines such as IL-1, IL-6, and TNFa, as well as interferon-g, which are known to activate macrophages, have been directly implicated in hepatotoxicity in a number of experimental models. Following exposure of animals to ethanol, endotoxin, turpentine, carbon tetrachloride, cadmium, zymosan, galactosamine, dimethylnitrosamine, or acetaminophen, expression of these cytokines increases in the liver (33,58,59,68,75,88,143,266–273). Moreover, many of the observed clinical features of liver disease and injury including fever, inflammation, cirrhosis, and

166

Laskin and Gardner

acute-phase protein production can be induced by administration of proinflammatory cytokines (251,274–276). Conversely, administration of neutralizing antibodies to IL-1, TNFa, IL-6, or interferon-g, soluble cytokine receptors, or cytokine receptor antagonists reduces inflammatory cell accumulation, acute-phase protein production, and tissue injury induced by toxicants such as carbon tetrachloride, acetaminophen, endotoxin, ethanol, allyl alcohol, zymosan, and cadmium (58,62,64,104,105,108,110,256,273,277–281). Protection against toxicants by blocking antibodies is paralleled in many models by results obtained using transgenic animals. For example, recent studies have demonstrated that mice lacking the gene for the p55 TNFa receptor-1 (TNFR1), or expressing only the membrane-bound form of the cytokine are protected from the toxic effects of carbon tetrachloride, ethanol, or the combination of endotoxin and galactosamine (79,89,105,282). Similarly, mice lacking the gene for IL-1 receptor or overexpressing IL-1 receptor antagonist exhibit an attenuated inflammatory response to turpentine (280), and IL-6 knockout mice are protected against the toxicity of zymosan (283). In contrast, mice deficient in IL-6 exhibit a greater hepatotoxic response to carbon tetrachloride (80,284), and mice lacking TNFR1 or the soluble form of TNFa are hypersensitive to the toxic effects of acetaminophen (142,161,266). These findings suggest that cytokines can exert both protective and proinflammatory/cytotoxic activity, which depend on the toxicant, levels of cytokine produced, its timing of appearance in the liver, and the extent to which other inflammatory mediators are generated. This is exemplified by results obtained with acetaminophen. In this model, TNFa appears to play a dual role in hepatotoxicity. Thus, when released in large quantities early in the pathogenic process by Kupffer cells, TNFa contributes to toxicity by causing damage to the endothelium, microthrombosis, and tissue injury (251–254). In contrast, TNFa released 24 to 48 hours later by infiltrating macrophages is protective, upregulating antioxidant defense, and inducing hepatocyte proliferation and extracellular matrix turnover (142,161,262–266). The fact that expression of chemokines and chemokine receptors including MCP-1 and CCR2, which regulate inflammatory macrophage infiltration into the liver, are significantly reduced in TNFaK/K and TNFR1K/K mice treated with acetaminophen, when compared to wild-type controls, demonstrate that TNFa signaling through TNFR1 is a critical mediator of the recruitment of inflammatory macrophages into the liver after acetaminopheninduced injury and that these inflammatory macrophages are important in the repair of damaged tissue. TGFb Studies with neutralizing antibodies and transgenic animal models have also provided evidence for a critical role of TGFb in sinusoidal cell activation and fibrosis. TGFb is produced by activated liver macrophages in response to injury and infection. In addition to its autocrine actions, TGFb acts on stellate cells to prolong their survival and induce collagen gene expression (240,243,285,286). In addition, TGFb promotes extracellular matrix production and mediates hepatocyte apoptosis (243). Following exposure of animals to fibrogenic doses of toxicants such as carbon tetrachloride, vitamin A, or alcohol, production of TGFb increases in the liver (81,236,287). These findings, together with the observation that carbon tetrachloride– induced increases in collagen deposition are reduced by 80% in transgenic mice with a targeted disruption of the TGFb gene, or in mice treated with antibodies to TGFb (288) and that fibrosis did not develop in mice with a targeted disruption of Smad3 (a transcriptional factor in TGFb receptor signal transduction) (289), provide strong support for an involvement of TGFb in tissue injury and fibrosis. Chemokines More recent studies have focused on another class of proinflammatory cytokines that exhibit chemotactic activity. These belong to a superfamily of low-molecular-weight proteins that play a key role in orchestrating the inflammatory response. Chemotactic cytokines or chemokines are divided into two subfamilies: C–X–C proteins (e.g., IL-8 or CINC), which are mainly neutrophil chemoattractants, and C–C chemokines (e.g., MIP-1, MIP-2, MCP-1, MCP-2, MCP-3, RANTES), which induce migration and activation of macrophages/monocytes and lymphocytes (290). Continuous local release of chemokines at sites of injury is thought to mediate the ongoing migration of effector cells into lesions during inflammatory responses. Chemokines

Nonparenchymal Cells and Inflammatory Macrophages

167

such as MIP-1a, MCP-1, RANTES, and CINC have been implicated in a variety of pathogenic processes in the liver including chemically-induced toxicity (64,91,291). These chemokines, which are produced in large part by Kupffer cells and endothelial cells (86,92,292), are upregulated in the liver after administration of endotoxin, ethanol, cadmium, or acetaminophen to animals (61,64,91,93,99,107,143,266,291,293). However, the precise role of chemokines in the pathogenesis of toxicity is controversial. Whereas some studies have indicated that they contribute to injury (250,294), others suggest that they may in fact act to reduce hepatotoxicity (62,64,143,295), which is most likely related to the production of anti-inflammatory mediators by newly infiltrated phagocytes (296). This is supported by findings that mice lacking the gene for CCR2, the receptor for MCP-1, are hypersensitive to the toxic effects of acetaminophen, a response that is correlated with increases in TNFa and interferon-g in the liver (62). Moreover, administration of MCP-1 protects mice from endotoxin toxicity and decreases hepatic TNFa levels (297). These data support the concept that inflammatory cytokines can both prevent and augment hepatotoxicity (298). Hepatocytes treated with toxicants like acetaminophen, galactosamine, or alcohol have also been reported to release phagocyte chemotactic and activating factors (99,299,300). Biochemical characterization studies have suggested that these factors are members of the chemokine family. Production of chemokines by hepatocytes is upregulated in response to Kupffer cell–derived TNFa and IL-1 (301,302). Thus, parenchymal cells apparently participate in inflammatory cell recruitment into the liver and activation during the pathogenesis of injury. Anti-inflammatory Cytokines Anti-inflammatory cytokines such as IL-4, IL-10, and IL-13 are also expressed in the liver following hepatotoxicant exposure (56,136,266,272,296,303,304). These cytokines facilitate the recovery of the liver from acute injury, inhibiting the production of proinflammatory cytokines (303,305–307) and stimulating the generation of extracellular matrix proteins (308). A recent study has demonstrated that IL-4 and IL-13 act on stellate cells to induce type 1 collagen formation and to suppress proliferation, suggesting that these cytokines may play a role in initiating repair processes (308). Anti-inflammatory cytokines also augment the production of IL-1 receptor antagonist (309). That these cytokines are important in toxicity is supported by the findings that administration of IL-13 protects mice from lethal endotoxemia and that anti-IL-13 antibodies significantly decrease survival rate (304). Similarly, hepatic fibrosis is increased after repeated carbon tetrachloride administration (303) and acetaminophen-induced hepatotoxicity is enhanced in IL-10 knockout mice (272). Reactive Oxygen Intermediates Reactive oxygen intermediates including superoxide anion, hydrogen peroxide, and hydroxyl radical are produced in significant quantities in cells by a variety of oxido-reductase reactions and during mitochondrial respiration. Although under physiological conditions, these mediators destroy invading pathogens and foreign materials, when generated in excessive amounts, they can cause oxidative injury. This includes cell membrane, protein, and DNA damage, lipid peroxidation, and cytotoxicity (310–313). Peroxidation of membrane lipids by reactive oxygen intermediates can also induce the formation and release of other inflammatory mediators, including PG, thromboxanes (Tx), and leukotrienes (LT) (see below). Macrophages, and in some models, endothelial cells and stellate cells isolated from the livers of hepatotoxicanttreated rats, have been reported to be activated to release increased quantities of reactive oxygen intermediates (7,25,26,28,82,87,95,114). Moreover, stimulation of hepatic macrophages to produce additional reactive oxygen intermediates by administration of agents such as retinol, glucan, or latex beads augments liver injury induced by agents such as C. parvum, carbon tetrachloride, 1,2-dichlorobenzene, and galactosamine. In contrast, administration of antioxidants like superoxide dismutase, catalase, allopurinol, N-acetylcysteine, methyl palmitate, endotoxin, or quinone derivatives is hepatoprotective (46,63,66,69,77,96,102,128,134,314–316). These studies support the hypothesis that oxygen-derived free radicals contribute to the pathogenesis of chemically-induced hepatotoxicity.

168

Laskin and Gardner

Reactive oxygen intermediates also appear to play an important role in fibrosis. Both expression and synthesis of TGFb are modulated via redox-sensitive reactions (237). Moreover, activation of stellate cells, as well as expression of metalloproteinases and their inhibitors, is dependent on reactive oxygen intermediates and lipid peroxidation products. The importance of oxidants in fibrosis is underscored by the finding that there is marked oxidative stress in the liver in most chronic disease processes affecting the tissue. It has been suggested that reactive oxidants contribute to both the onset and the progression of fibrosis induced by alcohol, carbon tetrachloride, viruses, iron, copper overload, cholestasis, and hepatic blood congestion (57,220,237). Reactive Nitrogen Intermediates Nitric oxide and its oxidation products have been implicated in altered hepatic functioning following xenobiotic exposure and in tissue injury (317–319). Nitric oxide is generated from 1-arginine by the NADPH-dependent enzyme, nitric oxide synthase. Three major isoforms of NOS have been identified: types I and III, which are produced in cells constitutively and are calcium and calmodulin dependent, and type II nitric oxide synthase (NOSII), which is induced after activation of cells by bacterially derived pathogens or cytokines (320). Whereas the type I form is largely localized in neuronal tissue, the type III form is found in vascular endothelium. In contrast, NOSII has been identified in both resident and inflammatory liver macrophages, as well as in hepatocytes, endothelial cells, stellate cells, smooth muscle cells, fibroblasts, and certain epithelial cells (29,30,55,56,230,231,258,298,317–322). Toxicity associated with excessive nitric oxide production is generally thought to be due to the actions of NOSII (298,323). Nitric oxide is a small, relatively stable free radical gas that readily diffuses into cells and cell membranes where it reacts with molecular targets such as heme- and thiol-containing proteins and amines (320,323). This can result in decreased cellular proliferation and nucleic acid biosynthesis, as well as altered enzyme activity, cytotoxicity, and apoptosis (317,321,323,324). Nitric oxide also binds to heme-containing proteins and this can result in either inhibition or activation of enzymes involved in hepatic drug metabolism. It has also been established that nitric oxide produced by macrophages is involved in the destruction of certain intracellular pathogens and tumor cells and in cytostasis (255,320,325,326). Nitric oxide is also known to react rapidly with superoxide anion generating peroxynitrite, a relatively long-lived cytotoxic oxidant that has been implicated in stroke, heart disease, and immune complex– mediated pulmonary edema (327–330). Peroxynitrite can also induce lipid peroxidation and can react directly with sulfhydryl groups in cell membranes leading to cytotoxicity and/or apoptosis (331–333). Peroxynitrite can also react with metals or metalloproteinases such as superoxide dismutase to form nitronium ion, a potent and toxic nitrosylating species (334). After treatment of animals with hepatotoxicants such as acetaminophen, carbon tetrachloride, ethanol, or endotoxin, Kupffer cells, as well as inflammatory macrophages, sinusoidal endothelial cells, stellate cells, and/or hepatocytes have been reported to express NOSII and to produce excessive quantities of nitric oxide (31,55,56,78,83,94,229,230,266,322,335,336). This has been correlated with nitrotyrosine staining of the liver (56,266,337). However, the role of nitric oxide and peroxynitrite in hepatotoxicity is controversial. Thus, while some studies have suggested that their actions are toxic, in other models, reactive nitrogen intermediates appear to play a protective role. For example, in animals pretreated with inhibitors of NOSII, such as aminoguanidine, or in transgenic mice with a targeted disruption of NOSII, hepatotoxicity induced by acetaminophen or endotoxin is significantly reduced (31,55,56,319,338,339). In contrast, hepatotoxicity is augmented in NOSII knockout mice treated with carbon tetrachloride (79). Similar increases in carbon tetrachloride- or endotoxin/C. parvum-induced hepatotoxicity have been described in animals pretreated with NOSII inhibitors (83,340–342), which is thought to be due to the ability of nitric oxide to reduce levels of cytotoxic oxidants (327–329). These data indicate that the relative pathological or protective roles of nitric oxide and peroxynitrite in toxicity depends on the nature of the toxicant and the extent to which tissue injury is mediated by reactive oxygen intermediates.

Nonparenchymal Cells and Inflammatory Macrophages

169

Bioactive Lipids Bioactive lipids constitute a broad range of mediators with both pro- and anti-inflammatory activity. The largest group is eicosanoids, which are derived from membrane-bound arachidonic acid. PG and Tx are generated from arachidonic acid via the enzyme COX. Two major isoforms of this enzyme have been identified: a constitutive form (COX-1), which is thought to provide cytoprotective function, and an inducible form (COX-2), which is involved in the generation of inflammatory PG. Metabolism of arachidonic acid via the enzyme lipoxygenase leads to the formation of LT. Although activated liver macrophages, as well as endothelial cells, stellate cells, and hepatocytes have been reported to synthesize a large number of different eicosanoids including LTB4, TxA2, PGE2, PGD2, PGF2g, and PGI2, their response to these mediators is distinct (32,70,71,184,231,343). This is most likely due to differential expression of eicosanoid receptors on these cells (227). The precise role of these eicosanoids in hepatotoxicity is unclear. PG such as PGE2 and PGD2 are known to play a key role in regulating inflammatory and immune reactions and also have the capacity to modify hepatocyte carbohydrate metabolism, calcium homeostasis, as well as protein synthesis and phosphorylation (72,343,344). Enhanced release of PG has been described following exposure of animals to toxins such as acetaminophen, ethanol, and endotoxin (267,343,345–347). Moreover, administration of COX inhibitors to animals prevents tissue injury induced by these toxicants (72,103,121,343,347–349). Similarly, inhibition of TxB2 synthase protects against endotoxic shock and liver injury (350,351). In contrast, recent studies have demonstrated that some PG may be hepatoprotective, presumably because of their ability to block inflammatory mediator production. Thus, PGE2 pretreatment prevents endotoxin-induced liver injury by downregulating TNFa and IL-12 and upregulating the anti-inflammatory cytokine IL-10 (352,353). Similarly, protection against galactosamine-induced hepatotoxicity by administration of PGE1 was correlated with decreased TNFa release (354). A number of LT also exhibit proinflammatory activity and are thought to play a role in chemically-induced tissue injury (100,343,353). For example, LTB4 is known to be a potent polymorphonuclear leukocyte chemoattractant and to induce monocyte IL-1, TNFa, and hydrogen peroxide production (71,355–357). Thus, release of LTB4 by macrophages in the liver following hepatotoxicant exposure may constitute a local control mechanism for the recruitment and activation of inflammatory cells. Kupffer cells and endothelial cells have been shown to express mRNA for 5-lipoxygenase, a major enzyme mediating the production of LT, while LTC4 synthase mRNA has been identified mainly in hepatocytes and endothelial cells (358). Endotoxin administration increases the expression of LTC4 synthase mRNA, and protein in hepatocytes, which may contribute to hepatocellular injury during inflammation (358). The finding that administration of lipoxygenase inhibitors or antagonists to mice protected against galactosamine-induced hepatitis suggests that LT have the capacity to contribute to inflammatory liver disease and injury (99,349). PAF is a phospholipid mediator that has also been implicated in tissue injury. It is released by a variety of cell types including macrophages, neutrophils, and endothelial cells and is thought to act in an autocrine and paracrine manner to amplify and propagate early stages of the inflammatory response. Thus, PAF released from inflammatory phagocytes stimulates macrophage and neutrophil chemotaxis and oxidative metabolism and nitric oxide generation (73,185,359–361). Following exposure of animals to endotoxin, liver macrophages and endothelial cells produce increased quantities of PAF (70,73,360). Interestingly, these cells also express increased numbers of functionally active receptors for PAF (362). Upregulation of PAF receptors may represent an important mechanism underlying macrophage and endothelial cell activation following hepatotoxicant exposure. In support of this possibility is the finding that administration of a PAF receptor antagonist reduces tissue injury induced by endotoxin (363). Hydrolytic Enzymes Macrophages and endothelial cells activated by inflammatory stimuli can also generate proteolytic and lysosomal enzymes. These include various proteases, lipases, MMP,

170

Laskin and Gardner

plasminogen activator, acid phosphatase, and cathepsin D (27,32,223,266,364–368). These can act directly on hepatocyte membranes to induce damage. Several of these enzymes have been shown to play a role in macrophage-mediated target cell destruction, as well as in altered hepatocyte functioning (27,32,223) and similar effects may occur in vivo after hepatotoxicant exposure. In contrast, the MMP (e.g., collagenase, gelatinase, and stromelysin) may contribute to recovery from liver fibrosis and play a role in fibrolysis during cirrhosis (368–370).

CONCLUSION Evidence has accumulated over the past several years demonstrating that chemically-induced toxicity is a multifactorial process involving direct tissue injury, as well as a cascade of protein and lipid mediators generated by cells in the liver. These include not only resident cells (hepatocytes, Kupffer cells, stellate cells, and endothelial cells), but also infiltrating leukocytes. Inflammatory mediators including cytokines, growth factors, and reactive mediators, including nitric oxide, peroxynitrite, superoxide anion, hydrogen peroxide, hydroxyl radicals, and eicosanoids, produced by activated nonparenchymal cells and/or infiltrating leukocytes may be cytotoxic, proinflammatory, and can compromise normal liver functioning or may help restore normal tissue structure and function (Fig. 2). Defining the precise role of each of these mediators in tissue injury and tissue repair is essential for our understanding of the mechanism of action of hepatotoxic chemicals and for devising steps to prevent or abrogate toxicity.

Toxicant

Hepatocytes

Necrosis

Repair Fibrosis

IM

IM Macrophages endothelial cells stellate cells

FIGURE 2 Model for the role of macrophages, endothelial cells, and stellate cells in hepatotoxicity. Toxicants such as acetaminophen and carbon tetrachloride cause injury to hepatocytes. This leads to the release of cytokines and/or growth factors that recruit and activate Kupffer cells, endothelial cells, stellate cells, and inflammatory macrophages to sites of injury. These cells release inflammatory mediators such as reactive oxygen and nitrogen intermediates, TNFa, IL-1, bioactive lipids, and/or hydrolytic enzymes that contribute to the development of hepatic necrosis and fibrosis. Infiltrating macrophages also release anti-inflammatory cytokines and growth factors that downregulate the inflammatory response and initiate tissue repair. The outcome of the hepatotoxic response depends on the balance between the pro- and anti-inflammatory mediators released. Abbreviation: IM, inflammatory mediators.

Nonparenchymal Cells and Inflammatory Macrophages

171

ACKNOWLEDGMENT

This work was supported by USPHS National Institutes of Health Grants GM034310, ES004738, AR055073, and ES005022.

REFERENCES 1. Bouwens L, Baekeland M, De Zanger R, et al. Quantitation, tissue distribution and proliferation kinetics of Kupffer cells in normal rat liver. Hepatology 1986; 6(4):718–22. 2. Caperna TJ, Garvey JS. Antigen handling in aging. II. The role of Kupffer and endothelial cells in antigen processing in Fischer 344 rats. Mech Ageing Dev 1982; 20(3):205–21. 3. Terpstra V, van Berkel TJ. Scavenger receptors on liver Kupffer cells mediate the in vivo uptake of oxidatively damaged red blood cells in mice. Blood 2000; 95(6):2157–63. 4. van Oosten M, van de Bilt E, van Berkel TJ, et al. New scavenger receptor-like receptors for the binding of lipopolysaccharide to liver endothelial and Kupffer cells. Infect Immun 1998; 66(11):5107–12. 5. Sano A, Taylor ME, Leaning MS, et al. Uptake and processing of glycoproteins by isolated rat hepatic endothelial and Kupffer cells. J Hepatol 1990; 10(2):211–6. 6. Ahmad N, Gardner CR, Yurkow EJ, et al. Inhibition of macrophages with gadolinium chloride alters intercellular adhesion molecule-1 expression in the liver during acute endotoxemia in rats. Hepatology 1999; 29(3):728–36. 7. Pilaro AM, Laskin DL. Accumulation of activated mononuclear phagocytes in the liver following lipopolysaccharide treatment of rats. J Leukoc Biol 1986; 40(1):29–41. 8. Gardner CR, Wasserman AJ, Laskin DL. Differential sensitivity of tumor targets to liver macrophage- mediated cytotoxicity. Cancer Res 1987; 47(24 Pt. 1):6686–91. 9. Gardner CR, Wasserman AJ, Laskin DL. Liver macrophage mediated cytotoxicity involves phagocytosis of tumor targets. Hepatology 1991; 14(2):318–24. 10. Løvdal T, Andersen E, Brech A, et al. Fc receptor mediated endocytosis of small soluble immunoglobulin G immune complexes in Kupffer cells and endothelial cells from rat liver. J Cell Sci 2000; 113(Pt. 18):3255–66. 11. Shi J, Fujieda H, Kokubo Y, et al. Apoptosis of neutrophils and their elimination by Kupffer cells in rat liver. Hepatology 1996; 24(5):1256–63. 12. Bykov I, Ylipaasto P, Eerola L, et al. Phagocytosis and LPS-stimulated production of cytokines and prostaglandin E2 is different in Kupffer cells isolated from the periportal or perivenous liver region. Scand J Gastroenterol 2003; 38(12):1256–61. 13. Naito M, Hasegawa G, Ebe Y, et al. Differentiation and function of Kupffer cells. Med Electron Microsc 2004; 37(1):16–28. 14. Parker GA, Picut CA. Liver immunobiology. Toxicol Pathol 2005; 33(1):52–62. 15. Helmy KY, Katschke KJ, Jr., Gorgani NN, et al. CRIg: a macrophage complement receptor required for phagocytosis of circulating pathogens. Cell 2006; 124(5):915–27. 16. Zuo G, Gong J, Liu C, et al. Synthesis of Toll-like receptor 4 in Kupffer cells and its role in alcoholinduced liver disease. Chin Med J (Engl) 2003; 116(2):297–300. 17. Gustot T, Lemmers A, Moreno C, et al. Differential liver sensitization to toll-like receptor pathways in mice with alcoholic fatty liver. Hepatology 2006; 43(5):989–1000. 18. Yamashina S, Takei Y, Ikejima K, et al. Ethanol-induced sensitization to endotoxin in Kupffer cells is dependent upon oxidative stress. Alcohol Clin Exp Res 2005; 29(Suppl. 12):246S–50. 19. Mathison JC, Ulevitch RJ. The clearance, tissue distribution, and cellular localization of intravenously injected lipopolysaccharide in rabbits. J Immunol 1979; 123(5):2133–43. 20. Nolan J. Endotoxin, reticuloendothelial function and liver injury. Hepatology 1981; 1(5):458–65. 21. Lukkari TA, Jarvelainen HA, Oinonen T, et al. Short-term ethanol exposure increases the expression of Kupffer cell CD14 receptor and lipopolysaccharide binding protein in rat liver. Alcohol 1999; 34(3):311–9. 22. Su GL, Klein RD, Aminlari A, et al. Kupffer cell activation by lipopolysaccaride in rats: role for lipopolysacchaide binding protein and toll-like receptor 4. Hepatology 2000; 31(4):932–6. 23. Su GL, Gong KQ, Fan MH, et al. Lipopolysaccharide-binding protein modulates acetaminopheninduced liver injury in mice. Hepatology 2005; 41(1):187–95. 24. Shnyra A, Lindberg AA. Scavenger receptor pathway for lipopolysaccharide binding to Kupffer and endothelial liver cells in vitro. Infect Immun 1995; 63(3):865–73. 25. McCloskey TW, Todaro JA, Laskin DL. Lipopolysaccharide treatment of rats alters antigen expression and oxidative metabolism in hepatic macrophages and endothelial cells. Hepatology 1992; 16(1):191–203. 26. Laskin DL, Pilaro AM. Potential role of activated macrophages in acetaminophen hepatotoxicity. I. Isolation and characterization of activated macrophages from rat liver. Toxicol Appl Pharmacol 1986; 86(2):204–15.

172

Laskin and Gardner

27. Tanner A, Keyhani A, Reiner R, et al. Proteolytic enzymes released by liver macrophages may promote injury in a rat model of hepatic damage. Gastroenterology 1981; 80(4):647–54. 28. Liu P, McGuire GM, Fisher MA, et al. Activation of Kupffer cells and neutrophils for reactive oxygen formation is responsible for endotoxin-enhanced liver injury after hepatic ischemia. Shock 1995; 3(1):56–62. 29. Laskin DL, Heck DE, Gardner CR, et al. Distinct patterns of nitric oxide production in hepatic macrophages and endothelial cells following acute exposure of rats to endotoxin. J Leukoc Biol 1994; 56(6):751–8. 30. Feder LS, Laskin DL. Regulation of hepatic endothelial cell and macrophage proliferation and nitric oxide production by GM-CSF, M-CSF and IL-1b following acute endotoxemia. J Leukoc Biol 1994; 55(4):507–13. 31. Laskin DL. Xenobiotic-induced inflammation and injury in the liver. In: McCuskey RS, Earnest DL, eds. Hepatic and Gastrointestinal Toxicology. New York: Pergamon Press, 1997:151–64. 32. Decker K. Biologically active products of stimulated liver macrophages (Kupffer cells). Eur J Biochem 1990; 192(2):245–61. 33. Feder LS, Todaro JA, Laskin DL. Characterization of interleukin-1 and interleukin-6 production by hepatic endothelial cells and macrophages. J Leukoc Biol 1993; 53(2):126–32. 34. Tsutsui H, Matsui K, Okamura H, et al. Pathophysiological roles of interleukin-18 in inflammatory liver diseases. Immunol Rev 2000; 174:192–209. 35. Tsutsui H, Adachi K, Seki E, et al. Cytokine-induced inflammatory liver injuries. Curr Mol Med 2003; 3(6):545–59. 36. Kojima Y, Suzuki S, Tsuchiya Y, et al. Regulation of pro-inflammatory and anti-inflammatory cytokine responses by Kupffer cells in endotoxin-enhanced reperfusion injury after total hepatic ischemia. Transpl Int 2003; 16(4):231–40. 37. Dahle MK, Overland G, Myhre AE, et al. The phosphatidylinositol 3-kinase/protein kinase B signaling pathway is activated by lipoteichoic acid and plays a role in Kupffer cell production of interleukin-6 (IL-6) and IL-10. Infect Immun 2004; 72(10):5704–11. 38. Noji S, Tashiro K, Koyama E, et al. Expression of hepatocyte growth factor gene in endothelial and Kupffer cells of damaged rat livers, as revealed by in situ hybridization. Biochem Biophys Res Commun 1990; 173(1):42–7. 39. Takeishi T, Hirano K, Kobayashi T, et al. The role of Kupffer cells in liver regeneration. Arch Histol Cytol 1999; 62(5):413–22. 40. Roth S, Gong W, Gressner AM. Expression of different isoforms of TGF-b and the latent TGF-b binding protein (LTBP) by rat Kupffer cells. J Hepatol 1998; 29(6):915–22. 41. Zhang X, Yu WP, Gao L, et al. Effects of lipopolysaccharides stimulated Kupffer cells on activation of rat hepatic stellate cells. World J Gastroenterol 2004; 10(4):610–3. 42. Seki E, Tsutsui H, Iimuro Y, et al. Contribution of Toll-like receptor/myeloid differentiation factor 88 signaling to murine liver regeneration. Hepatology 2005; 41(3):443–50. 43. Duffield JS. The inflammatory macrophage: a story of Jekyll and Hyde. Clin Sci (Lond) 2003; 104(1):27–38. 44. Granado M, Martin AI, Priego T, et al. Inactivation of Kupffer cells by gadolinium administration prevents lipopolysaccharide-induced decrease in liver insulin-like growth factor-I and IGF-binding protein-3 gene expression. J Endocrinol 2006; 188(3):503–11. 45. Abril ER, Simm WM, Earnest DL. Kupffer cell secretion of cytotoxin cytokines is enhanced by hypervitaminosis A. In: Wisse E, Knook DL, Decker K, eds. Cells of the Hepatic Sinusoid. Vol. 2 Rijswijk, The Netherlands: Kupffer Cell Foundation, 1989:73–5. 46. Arthur MJP, Bentley IS, Tanner A, et al. Oxygen-derived free radicals promote hepatic injury in the rat. Gastroenterology 1985; 89(5):1114–22. 47. Shiratori Y, Takikawa H, Kawase T, et al. Superoxide anion generating capacity and lysosomal enzyme activities of Kupffer cells in galactosamine-induced hepatitis. Gastroenterol Jpn 1986; 21(2):135–44. 48. Mochida S, Ogata I, Ohta Y, et al. In situ evaluation of the stimulatory state of hepatic macrophages based on their ability to produce superoxide anions in rats. J Pathol 1989; 158(1):67–71. 49. Bautista AP, Meszaros K, Bojta J, et al. Superoxide anion generation in the liver during the early stages of endotoxemia in rats. J Leukoc Biol 1990; 48(2):123–8. 50. Rogoff TM, Lipsky PE. Antigen presentation by isolated guinea pig Kupffer cells. J Immunol 1980; 124(4):1740–4. 51. Itoh Y, Okanoue T, Morimoto M, et al. Functional heterogeneity of rat liver macrophages: interleukin-1 secretion and Ia antigen expression in contrast with phagocytic activity. Liver 1992; 12(1):26–33. 52. Lohse AW, Knolle PA, Bilo K, et al. Antigen-presenting function and B7 expression of murine sinusoidal endothelial cells and Kupffer cells. Gastroenterology 1996; 110(4):1175–81. 53. Maemura K, Zheng Q, Wada T, et al. Reactive oxygen species are essential mediators in antigen presentation by Kupffer cells. Immunol Cell Biol 2005; 83(4):336–43.

Nonparenchymal Cells and Inflammatory Macrophages

173

54. Matsuno K, Nomiyama H, Yoneyama H, et al. Kupffer cell-mediated recruitment of dendritic cells to the liver crucial for a host defense. Dev Immunol 2002; 9(3):143–9. 55. Gardner CR, Heck DE, Yang CS, et al. Role of nitric oxide in acetaminophen-induced hepatotoxicity in the rat. Hepatology 1998; 27(4 Pt. 2):748–54. 56. Gardner CR, Laskin JD, Dambach DM, et al. Reduced hepatotoxicity of acetaminophen in mice lacking inducible nitric oxide synthase: potential role of tumor necrosis factor-a and interleukin-10. Toxicol Appl Pharmacol 2002; 184(1):27–36. 57. Laskin DL. Nonparenchymal cells and hepatotoxicity. Semin Liver Dis 1990; 10(4):293–304. 58. Blazka ME, Wilmer JL, Holladay SD, et al. Role of proinflammatory cytokines in acetaminophen hepatotoxicity. Toxicol Appl Pharmacol 1995; 133(1):43–52. 59. Blazka ME, Elwellk MR, Holladay SD, et al. Histopathology of acetaminophen-induced liver changes: role of interleukin 1a and tumor necrosis factor a. Toxicol Pathol 1996; 24(2):181–9. 60. Chiu H, Brittingham JA, Laskin DL. Differential induction of heme oxygenase-1 in macrophages and hepatocytes during acetaminophen-induced hepatotoxicity in the rat: effects of hemin and biliverdin. Toxicol Appl Pharmacol 2002; 181(2):106–15. 61. Lawson JA, Farhood A, Hopper RD, et al. The hepatic inflammatory response after acetaminophen overdose: role of neutrophils. Toxicol Sci 2000; 54(2):509–16. 62. Hogaboam CM, Bone-Larson CL, Steinhauser ML, et al. Exaggerated hepatic injury due to acetaminophen challenge in mice lacking C–C chemokine receptor 2. Am J Pathol 2000; 156(4):1245–52. 63. Nakae D, Yamamoto K, Yoshiji H, et al. Liposome-encapsulated superoxide dismutase prevents liver necrosis induced by acetaminophen. Am J Pathol 1990; 136(4):787–95. 64. Dambach DM, Watson LM, Gray KR, et al. Role of CCR2 in macrophage migration into the liver during acetaminophen-induced hepatotoxicity in the mouse. Hepatology 2002; 35(5):1093–103. 65. Michael SL, Pumford NR, Mayeux PR, et al. Pretreatment of mice with macrophage inactivators decreases acetaminophen hepatotoxicity and the formation of reactive oxygen and nitrogen species. Hepatology 1999; 30(1):186–95. 66. Shiratori Y, Tanaka M, Hai K, et al. Role of endotoxin-responsive macrophages in hepatic injury. Hepatology 1990; 11(2):183–92. 67. Ahmad N, Chen LC, Gordon MA, et al. Regulation of cyclooxygenase-2 by nitric oxide in activated hepatic macrophages during acute endotoxemia. J Leukoc Biol 2002; 71(6):1005–11. 68. Aono K, Isobe K, Kiuchi K, et al. In vitro and in vivo expression of inducible nitric oxide synthase during experimental endotoxemia: involvement of other cytokines. J Cell Biochem 1997; 65(3):349–58. 69. Ghezzi P, Saccardo B, Bianchi M. Role of reactive oxygen intermediates in the hepatotoxicity of endotoxin. Immunopharmacology 1986; 12(3):241–4. 70. Decker K. Eicosanoids, signal molecules of liver cells. Semin Liver Dis 1985; 5(2):175–90. 71. Keppler D, Hagmann W, Rapp S, et al. The relation of leukotrienes to liver injury. Hepatology 1985; 5(5):883–91. 72. Rodriguez de Turco EB, Spitzer JA. Eicosanoid production in nonparenchymal liver cells isolated from rats infused with E. coli endotoxin. J Leukoc Biol 1990; 48(6):488–94. 73. Mustafa SB, Flickinger BD, Olson MS. Suppression of lipopolysaccharide-induced nitric oxide synthase expression by platelet-activating factor receptor antagonists in the rat liver and cultured rat Kupffer cells. Hepatology 1999; 30(5):1206–14. 74. Thompson WD, Jack AS, Patrick RS. Possible role of macrophages in transient hepatic fibrogenesis induced by acute carbon tetrachloride injury. J Pathol 1980; 130(2):65–73. 75. Orfila C, Lepert JC, Alric L, et al. Expression of TNF-a and immunohistochemical distribution of hepatic macrophage surface markers in carbon tetrachloride-induced chronic liver injury in rats. Histochem J 1999; 31(10):677–85. 76. Camandola S, Aragno M, Cutrin JC, et al. Liver AP-1 activation due to carbon tetrachloride is potentiated by 1,2-dibromoethane but is inhibited by a-tocopherol or gadolinium chloride. Free Radic Biol Med 1999; 26(9/10):1108–16. 77. elSisi AE, Earnest DL, Sipes IG. Vitamin A potentiation of carbon tetrachloride hepatotoxicity: role of liver macrophages and active oxygen species. Toxicol Appl Pharmacol 1993; 119(2):295–301. 78. Chamulitrat W, Blazaka ME, Jordan SJ, et al. Tumor necrosis factor-a and nitric oxide production in endotoxin-primed rats administered carbon tetrachloride. Life Sci 1995; 57(24):2273–80. 79. Morio LA, Chiu H, Sprowles KA, et al. Distinct roles of tumor necrosis factor-a and nitric oxide in acute liver injury induced by carbon tetrachloride in mice. Toxicol Appl Pharmacol 2001; 172(1):44–51. 80. Katz A, Chebath J, Friedman J, et al. Increased sensitivity of IL-6-deficient mice to carbon tetrachloride hepatotoxicity and protection with an IL-6 receptor-IL-6 chimera. Cytokines Cell Mol Ther 1998; 4(4):221–7. 81. Armendariz-Borunda J, Katai H, Jones CM, et al. Transforming growth factor-b gene expression is transiently enhanced at a critical stage during liver regeneration after CCl4 treatment. Lab Invest 1993; 69(3):283–94.

174

Laskin and Gardner

82. Sipes IG, el Sisi AE, Sim WW, et al. Reactive oxygen species in the progression of CCl4-induced liver injury. Adv Exp Med Biol 1991; 283:489–97. 83. Zhu W, Fung PC. The roles played by crucial free radicals like lipid free radicals, nitric oxide, and enzymes NOS and NADPH in CCl4-induced acute liver injury of mice. Free Radic Biol Med 2000; 29(9):870–80. 84. Thompson K, Maltby J, Fallowfield J, et al. Interleukin-10 expression and function in experimental murine liver inflammation and fibrosis. Hepatology 1998; 28(6):1597–606. 85. Kamimura S, Tsukamoto H. Cytokine gene expression by Kupffer cells in experimental alcoholic liver disease. Hepatology 1995; 22(4 Pt. 1):1304–9. 86. Bukara M, Bautista AP. Acute alcohol intoxication and gadolinium chloride attenuate endotoxininduced release of CC chemokines in the rat. Alcohol 2000; 20(2):193–203. 87. Yokoyama H, Fukuda M, Okamura Y, et al. Superoxide anion release into the hepatic sinusoid after an acute ethanol challange and its attenuation by Kupffer cell depletion. Alcohol Clin Exp Res 1999; 23(Suppl. 4):71S–5S. 88. Fang C, Lindros KO, Badger TM, et al. Zonated expression of cytokines in rat liver: effect of chronic ethanol and the cytochrome P450 2E1 inhibitor, chlormethiazole. Hepatology 1998; 27(5):1304–10. 89. Yin M, Wheeler MD, Kono H, et al. Essential role of tumor necrosis factor alpha in alcohol-induced liver injury in mice. Gastroenterology 1999; 117(4):942–52. 90. Spitzer JA, Zheng M, Kolls JK, et al. Ethanol and LPS modulate NF-kappaB activation, inducible NO synthase and COX-2 gene expression in rat liver cells in vivo. Front Biosci 2002; 7:a99–108. 91. Maher JJ. Rat hepatocytes and Kupffer cells interact to produce interleukin-8 (CINC) in the setting of ethanol. Am J Physiol 1995; 269(4 Pt. 1):G518–23. 92. Bautista AP. Impact of alcohol on the ability of Kupffer cells to produce chemokines and its role in alcoholic liver disease. J Gastroenterol Hepatol 2000; 15(4):346–56. 93. Afford SC, Fisher NC, Neil DAH, et al. Distinct patterns of chemokine expression are associated with leukocyte recruitment in alcoholic hepatitis and alcoholic cirrhosis. J Pathol 1998; 186(1):82–9. 94. Matsumoto H, Nishitani Y, Minowa Y, et al. Role of Kupffer cells in the release of nitric oxide and change of portal pressure after ethanol perfusion in the rat liver. Alcohol Alcohol 2000; 35(1):31–4. 95. Hoglen NC, Younis HS, Hartley DP, et al. 1,2-Dichlorobenzene-induced lipid peroxidation in male Fischer 344 rats is Kupffer cell dependent. Toxicol Sci 1998; 46(2):376–85. 96. Guanawardhana L, Mobley SA, Sipes IG. Modulation of 1,2-dichlorobenzene hepatotoxicity in the Fischer- 344 rat by a scavanger of superoxide anions and an inhibitor of Kupffer cells. Toxicol Appl Pharmacol 1993; 119(2):205–13. 97. Younis HS, Hoglen NC, Kuester RK, et al. 1,2-Dichlorobenzene-mediated hepatocellular oxidative stress in Fischer-344 and Sprague–Dawley rats. Toxicol Appl Pharmacol 2000; 163(2):141–8. 98. Younis HS, Parrish AR, Sipes IG. The role of hepatocellular oxidative stress in Kupffer cell activation during 1,2-dichlorobenzene-induced hepatotoxicity. Toxicol Sci 2003; 76(1):201–11. 99. Shiratori Y, Hai K, Takada H, et al. Mechanism of accumulation of macrophages in galactosamineinduced liver injury: effect of lipoxygenase inhibitors on chemotaxis of spleen cells. Pathobiology 1992; 60(6):316–21. 100. Jonker AM, Kijkhuis FW, Kroese FG, et al. Immunopathology of acute galactosamine hepatitis in rats. Hepatology 1990; 11(4):622–7. 101. MacDonald JR, Beckstead JH, Smuckler EA. An ultrastructural and histochemical study of the prominent inflammatory response in D(C)-galactosamine hepatotoxicity. Br J Exp Pathol 1987; 68(2):189–99. 102. Shiratori Y, Kawase T, Shiina S, et al. Modulation of hepatotoxicity by macrophages in the liver. Hepatology 1988; 8(4):815–21. 103. Tiegs G, Wolter M, Wendel A. Tumor necrosis factor is a terminal mediator in galactosamine/endotoxin-induced hepatitis in mice. Biochem Pharmacol 1989; 38(4):627–31. 104. Hishinuma I, Nagakawa J, Hirota K, et al. Involvement of tumor necrosis factor-a in development of hepatic injury in galactosamine-sensitized mice. Hepatology 1990; 12(5):1187–91. 105. Nowak M, Gaines GG, Rosenberg J, et al. LPS-induced liver injury in D-galactosamine-sensitized mice requires secreted TNF-alpha and the TNF-p55 receptor. Am J Physiol Regul Integr Comp Physiol 2000; 278(5):R1202–9. 106. Rikans LE, Yamano T. Mechanisms of cadmium-mediated acute hepatotoxicity. J Biochem Mol Toxicol 2000; 14(2):110–7. 107. Yamano T, DeCicco LA, Rikans LE. Attenuation of cadmium-induced liver injury in senecent male Fischer 344 rats: role of Kupffer cells and inflammatory cytokines. Toxicol Appl Pharmacol 2000; 162(1):68–75. 108. Kayama F, Yoshida T, Elwell MR, et al. Role of tumor necrosis factor-alpha in cadmium-induced hepatotoxicity. Toxicol Appl Pharmacol 1995; 131(2):224–34. 109. Przybocki JM, Reuhl KR, Thurman RG, et al. Involvement of nonparenchymal cells in oxygendependent hepatic injury by allyl alcohol. Toxicol Appl Pharmacol 1992; 115(1):57–63.

Nonparenchymal Cells and Inflammatory Macrophages

175

110. Sneed RA, Buchweitz JP, Jean PA, et al. Pentoxifylline attenuates bacterial lipopolysaccharideinduced enhancement of allyl alcohol hepatotoxicity. Toxicol Sci 2000; 56(1):203–10. 111. Ide M, Yamate J, Machida Y, et al. Emergence of different macrophage populations in hepatic fibrosis following thioacetamide-induced acute hepatocyte injury in rats. J Comp Pathol 2003; 128(1):41–51. 112. Andre´s D, Sanchez-Reus I, Bautista M, et al. Depletion of Kupffer cell function by gadolinium chloride attenuates thioacetamide-induced hepatotoxicity. Expression of metallothionein and HSP70. Biochem Pharmacol 2003; 66(6):917–26. 113. Novosyadlyy R, Dargel R, Scharf JG. Expression of insulin-like growth factor-I and insulin-like growth factor binding proteins during thioacetamide-induced liver cirrhosis in rats. Growth Horm IGF Res 2005; 15(5):313–23. 114. Laskin DL, Robertson FM, Pilaro AM, et al. Activation of liver macrophages following phenobarbital treatment of rats. Hepatology 1988; 8(5):1051–5. 115. Arthur MJP, Kowalski-Sanders P, Wright R. Corynebacterium parvum-elicited hepatic macrophages demonstrate enhanced respiratory burst activity compared with resident Kupffer cells in the rat. Gastroenterology 1986; 91(1):174–81. 116. Laskin DL, Gardner CR. Nonparenchymal cells, inflammatory macrophages, and hepatotoxicity. In: Kaplowitz N, DeLeve LD, eds. Drug-Induced Liver Disease. New York: Marcel Dekker, 2003:183– 211. 117. Armendariz-Borunda J, Seyer JM, Postlethwaite AE, et al. Kupffer cells from carbon tetrachlorideinjured rat livers produce chemotactic factors for fibroblasts and monocytes: the role of tumor necrosis factor-alpha. Hepatology 1991; 14(5):895–900. 118. Nagy LE. Recent insights into the role of the innate immune system in the development of alcoholic liver disease. Exp Biol Med (Maywood) 2003; 228(8):882–90. 119. Laskin DL, Pendino KJ. Macrophages and inflammatory mediators in tissue injury. Annu Rev Pharmacol Toxicol 1995; 35:655–77. 120. Sudhir S, Budhiraja RD. Comparison of the protective effect of Withaferin-A and hydrocortisone against CCl4 induced hepatotoxicity in rats. Indian J Physiol Pharmacol 1992; 36(2):127–9. 121. Ben-Zvi Z, Weissman-Teitellman B, Katz S, et al. Acetaminophen hepatotoxicity: is there a role for prostaglandin synthesis? Arch Toxicol 1990; 64(4):299–304. 122. Li G, Liu Y, Tzeng NS, et al. Protective effect of dextromethorphan against endotoxic shock in mice. Biochem Pharmacol 2005; 69(2):233–40. 123. Edwards MJ, Keller BJ, Kauffman FC, et al. The involvement of Kupffer cells in carbon tetrachloride toxicity. Toxicol Appl Pharmacol 1993; 119(2):275–9. 124. Laskin DL, Gardner CR, Price VF, et al. Modulation of macrophage functioning abrogates the acute hepatotoxicity of acetaminophen. Hepatology 1995; 21(4):1045–50. 125. Stenger RJ, Petrelli M, McPath DCP, et al. Modification of carbon tetrachloride hepatoxicity by prior loading of the reticuloendothelial system with carbon particles. Am J Pathol 1969; 57(3):689–97. 126. Deaciuc IV, Bagby GJ, Niesman MR, et al. Modulation of hepatic sinusoidal endothelial cell function by Kupffer cells: an example of intercellular communication in the liver. Hepatology 1994; 19(2):464–70. 127. Goldin RD, Ratnayaka ID, Breach CS, et al. Role of macrophages in acetaminophen (paracetamol)induced hepatotoxicity. J Pathol 1996; 179(4):432–5. 128. Ishiyama H, Ogino K, Hobara T. Kupffer cells in rat liver injury induced by diethyldithiocarbamate. Eur J Pharmacol 1995; 292(2):135–41. 129. Iimuro Y, Yamamoto M, Kohno H, et al. Blockade of liver macrophages by gadolinium chloride reduces lethality in endotoxemic rats—analysis of mechanisms of lethality in endotoxemia. J Leukoc Biol 1994; 55(6):723–8. 130. Al-Tuwaijri A, Akdamar K, Di Luzio NR. Modification of galactosamine-induced liver injury in rats by reticuloendothelial system stimulation or depression. Hepatology 1981; 1(2):107–13. 131. Ide M, Kuwamura M, Kotani T, et al. Effects of gadolinium chloride (GdCl3) on the appearance of macrophage populations and fibrogenesis in thioacetamide-induced rat hepatic lesions. J Comp Pathol 2005; 133(2–3):92–102. 132. He Q, Kim J, Sharma RP. Fumonisin B1 hepatotoxicity in mice is attenuated by depletion of Kupffer cells by gadolinium chloride. Toxicology 2005; 207(1):137–47. 133. Galanos C, Freudenberg MA, Reuter W. Galactosamine-induced sensitization to the lethal effects of endotoxin. Proc Natl Acad Sci (USA) 1979; 76(11):5939–43. 134. Sauer JM, Hooser SB, Badger DA, et al. Alterations in chemically induced tissue injury related to alltrans-retinol pretreatment in rodents. Drug Metabol Rev 1995; 27(1–2):299–323. 135. Sauer JM, Sipes IG. Modulation of chemical-induced lung and liver toxicity by all-trans-retinol in the male Sprague–Dawley rat. Toxicology 1995; 105(2–3):237–49. 136. Ju C, Reilly TP, Bourdi M, et al. Protective role of Kupffer cells in acetaminophen-induced hepatic injury in mice. Chem Res Toxicol 2002; 15(12):1504–13. 137. Olthoff KM. Molecular pathways of regeneration and repair after liver transplantation. World J Surg 2002; 26(7):831–7.

176

Laskin and Gardner

138. Roggin KK, Papa EF, Kurkchubasche AG, et al. Kupffer cell inactivation delays repair in a rat model of reversible biliary obstruction. J Surg Res 2000; 90(2):166–73. 139. Taub R, Bansal M, Lei W. Roles of cytokines and growth factors in liver regeneration, repair, and fibrosis after liver injury. In: Kaplowitz N, DeLeve LD, eds. Drug-Induced Liver Disease. New York: Marcel Dekker, 2003:213–25. 140. Duffield JS, Forbes SJ, Constandinou CM, et al. Selective depletion of macrophages reveals distinct, opposing roles during liver injury and repair. J Clin Invest 2005; 115(1):56–65. 141. Blazka ME, Germolec DR, Simeonova P, et al. Acetaminophen-induced hepatotoxicity is associated with early changes in NF-kB and NF-IL6 DNA binding activity. J Inflamm 1995; 47(3):138–50. 142. Chiu H, Gardner CR, Dambach DM, et al. Role of p55 tumor necrosis factor receptor 1 in acetaminophen-induced antioxidant defense. Am J Physiol Gastrointest Liver Physiol 2003; 285(5):G959–66. 143. Dambach DM, Durham SK, Laskin JD, et al. Distinct roles of NF-kB p50 in the regulation of acetaminophen-induced inflammatory mediator production and hepatotoxicity. Toxicol Appl Pharmacol 2006; 211(2):157–65. 144. Ahmad N, Chen C, Martey CA, et al. Acute endotoxemia is associated with increased STAT1 and NF-kB nuclear binding activity in hepatic macrophages and endothelial cells. Toxicol Sci 2000; 54(1):140. 145. Fox ES, Cantrell CH, Leingang KA. Inhibition of the Kupffer cell inflammatory response by acute ethanol: NF-kB activation and subsequent cytokine production. Biochem Biophys Res Commun 1996; 225(1):134–40. 146. Essani NA, McGuire GM, Manning AM, et al. Endotoxin-induced activation of the nuclear transcription factor kB and expression of e-selectin messenger RNA in hepatocytes, Kupffer cells, and endothelial cells. J Immunol 1996; 156(8):2956–63. 147. Gruebele A, Zawaski K, Kaplan D, et al. Cytochrome P4502E1- and cytochrome P4502B1/2B2catalyzed carbon tetrachloride metabolism: effects on signal transduction as demonstrated by altered immediate-early (c-Fos and c-Jun) gene expression and nuclear AP-1 and NF-kB transcription factor levels. Drug Metabol Dispos 1996; 24(1):15–22. 148. Nanji AA, Jokelainen K, Rahemtulla A, et al. Activation of nuclear factor kappa B and cytokine imbalance in experimental alcoholic liver disease in the rat. Hepatology 1999; 30(4):934–43. 149. Taniguchi M, Takeuchi T, Nakatsuka R, et al. Molecular process in acute liver injury and regeneration induced by carbon tetrachloride. Life Sci 2004; 75(13):1539–49. 150. McMullen MR, Pritchard MT, Wang Q, et al. Early growth response-1 transcription factor is essential for ethanol-induced fatty liver injury in mice. Gastroenterology 2005; 128(7):2066–76. 151. Celec P. Nuclear factor kappa B–molecular biomedicine: the next generation. Biomed Pharmacother 2004; 58(6-7):365–71. 152. Liu SL, Degli Esposti S, Yao T, et al. Vitamin E therapy of acute CCl4-induced hepatic injury in mice is associated with inhibition of nuclear factor kappa B binding. Hepatology 1995; 22(5):1474–81. 153. Kim SH, Chu HJ, Kang DH, et al. NF-kappa B binding activity and cyclooxygenase-2 expression in persistent CCl4-treated rat liver injury. J Korean Med Sci 2002; 17(2):193–200. 154. Chaisson ML, Brooling JT, Ladiges W, et al. Hepatocyte-specific inhibition of NF-kappaB leads to apoptosis after TNF treatment, but not after partial hepatectomy. J Clin Invest 2002; 110(2):193–202. 155. Maeda S, Kamata H, Luo JL, et al. IKKbeta couples hepatocyte death to cytokine-driven compensatory proliferation that promotes chemical hepatocarcinogenesis. Cell 2005; 121(7):977–90. 156. Blazka ME, Bruccoleri A, Simeonova P, et al. Acetaminophen-induced hepatotoxicity is associated with early changes in AP-1 DNA binding activity. Res Commun Mol Pathol Pharmacol 1996; 92(3):259–73. 157. Tran-Thi TA, Decker K, Baeuerle P. Differential activation of transcription factors NF-kB and AP-1 in rat liver macrophages. Hepatology 1995; 22(2):613–9. 158. Liu J, Kadiiska MB, Corton JC, et al. Acute cadmium exposure induces stress-related gene expression in wild-type and metallothionein-I/II-null mice. Free Radic Biol Med 2002; 32(6):525–35. 159. Schiaffonati L, Tiberio L. Gene expression in liver after toxic injury: analysis of heat shock response and oxidative stree-inducible genes. Liver 1997; 17(4):183–91. 160. Wang XD, Liu C, Chung J, et al. Chronic alcohol intake reduces retinoic acid concentration and enhances AP-1 (c-Jun and c-Fos) expression in rat liver. Hepatology 1998; 28(3):744–50. 161. Chiu H, Gardner CR, Dambach DM, et al. Role of tumor necrosis factor receptor 1 (p55) in hepatocyte proliferation during acetaminophen-induced toxicity in mice. Toxicol Appl Pharmacol 2003; 193(2):218–27. 162. Gashler A, Sukhatme VP. Early growth response protein 1 (Egr-1): prototype of a zinc-finger family of transcription factors. Prog Nucleic Acid Res Mol Biol 1995; 50:191–224. 163. Feng JM, Shi JQ, Liu YS. The effect of lipopolysaccharides on the expression of CD14 and TLR4 in rat Kupffer cells. Hepatobiliary Pancreat Dis Int 2003; 2(2):265–9. 164. Chen LC, Laskin DL, Laskin JD. Regulation of inflammatory mediator production in the liver during acute endotoxemia. Toxicologist 2006; 90(1):349.

Nonparenchymal Cells and Inflammatory Macrophages

177

165. Gunawan BK, Liu ZX, Han D, et al. c-Jun-N-terminal kinase plays a major role in murine acetaminophen hepatotoxicity. Gastroenterology 2006; 131(1):165–78. 166. Klintman D, Li X, Santen S, et al. p38 mitogen-activated protein kinase-dependent chemokine production, leukocyte recruitment, and hepatocellular apoptosis in endotoxemic liver injury. Ann Surg 2005; 242(6):830–8 (discussion 838–839). 167. Cheng YJ, Liu MY. Modulation of tumor necrosis factor-alpha and oxidative stress through protein kinase C and P42/44 mitogen-activated protein kinase in lead increases lipopolysaccharide-induced liver damage in rats. Shock 2005; 24(2):188–93. 168. Enomoto K, Nishikawa Y, Omori Y, et al. Cell biology and pathology of liver sinusoidal endothelial cells. Med Electron Microsc 2004; 37(4):208–15. 169. van Oosten M, van Amersfoort ES, van Berkel TJ, et al. Scavenger receptor-like receptors for the binding of lipopolysaccharide and lipoteichoic acid to liver endothelial and Kupffer cells. J Endotoxin Res 2001; 7(5):381–4. 170. Praaning-van Dalen DP, De Leeuw AM, Brower A, et al. Rat liver endothelial cells have a greater capacity that Kupffer cells to endocytose N-acetylglucosamine- and mannose-terminated glycoproteins. Hepatology 1987; 7(4):672–9. 171. Dini L, Lentini A, Diez GD, et al. Phagocytosis of apoptotic bodies by liver endothelial cells. J Cell Sci 1995; 108(Pt 3):967–73. 172. Kosugi I, Muro H, Shirasawa H, et al. Endocytosis of soluble IgG immune complex and its transport to lysosomes in hepatic sinusoidal endothelial cells. J Hepatol 1992; 16(1–2):106–14. 173. Smedsrod B, Pertoft H, Gustanfson S, et al. Scavenger functions of the liver endothelial cell. Biochem J 1990; 266(2):313–27. 174. Oynebraten I, Hansen B, Smedsrod B, et al. Serglycin secreted by leukocytes is efficiently eliminated from the circulation by sinusoidal scavenger endothelial cells in the liver. J Leukoc Biol 2000; 67(2):183–8. 175. Zhou B, Oka JA, Singh A, et al. Purification and subunit characterization of the rat liver endocytic hyaluronan receptor. J Biol Chem 1999; 274(48):33831–4. 176. Yoshioka T, Yamamoto K, Kobashi H, et al. Receptor-mediated endocytosis of chemically modified albumins by sinusoidal endothelial cells and Kupffer cells in rat and human liver. Liver 1994; 14(3):129–37. 177. Esbach S, Stins MF, Brouwer A, et al. Morphological characterization of scavenger receptormediated processing of modified lipoproteins by rat liver endothelial cells. Exp Cell Res 1994; 210(1):62–70. 178. Duryee MJ, Freeman TL, Willis MS, et al. Scavenger receptors on sinusoidal liver endothelial cells (SECs) are involved in the uptake of Aldehyde-modified proteins. Mol Pharmacol 2005; 68(5):1423–30. 179. Nakajou K, Horiuchi S, Sakai M, et al. CD36 is not involved in scavenger receptor-mediated endocytic uptake of glycolaldehyde- and methylglyoxal-modified proteins by liver endothelial cells. J Biochem (Tokyo) 2005; 137(5):607–16. 180. Hansen B, Longati P, Elvevold K, et al. Stabilin-1 and stabilin-2 are both directed into the early endocytic pathway in hepatic sinusoidal endothelium via interactions with clathrin/AP-2, independent of ligand binding. Exp Cell Res 2005; 303(1):160–73. 181. Steffan A-M, Gendrault J-L, McCuskey RS, et al. Phagocytosis, an unrecognized property of murine endothelial liver cells. Hepatology 1986; 6(5):830–6. 182. Shiratori Y, Jin’nai H, Teraoka H, et al. Phagocytic properties of hepatic endothelial cells and splenic macrophages compensating for a decreased phagocytic function of Kupffer cells in the chronically ethanol-fed rats. Exp Cell Biol 1989; 57(6):300–9. 183. Bogers WMJM, Stad RK, Janssen DJ, et al. Kupffer cell depletion in vivo results in clearance of largesized IgA aggregates in rats by liver endothelial cells. Clin Exp Immunol 1991; 85(1):128–36. 184. Hashimoto N, Watanabe T, Shiratori Y, et al. Prostanoid secretion by rat hepatic sinusoidal endothelial cells and its regulation by exogenous adenosine triphosphate. Hepatology 1995; 21(6):1713–8. 185. Gardner CR, Laskin JD, Laskin DL. Platelet-activating factor-induced calcium mobilization and oxidative metabolism in hepatic macrophages and endothelial cells. J Leukoc Biol 1993; 53(2):190–6. 186. Spolarics Z. Endotoxin stimulates gene expression of ROS-eliminating pathways in rat hepatic endothelial and Kupffer cells. Am J Physiol 1996; 270(4 Pt. 1):G660–6. 187. Misquith S, Wattiaux-De Coninck S, Wattiaux R. Intracellular degradation by liver endothelial cells. Mol Cell Biochem 1989; 91(1–2):63–74. 188. Eakes AT, Olson MS. Regulation of endothelin synthesis in hepatic endothelial cells. Am J Physiol 1998; 274(6 Pt 1):G1068–76. 189. Mizoguchi Y, Ichikawa Y, Kioka K, et al. Effects of arachidonic acid metabolites and interleukin-1 on platelet activating factor production by hepatic sinusoidal endothelial cells from mice. J Gastroenterol Hepatol 1991; 6(3):283–8.

178

Laskin and Gardner

190. Arii S, Imamura M. Physiological role of sinusoidal endothelial cells and Kupffer cells and their implications in the pathogenesis of liver injury. J Hepatobiliary Pancreat Surg 2000; 7(1):40–8. 191. Ohira H, Abe K, Yokokawa J, et al. Adhesion molecules and CXC chemokines in endotoxin-induced liver injury. Fukushima J Med Sci 2003; 49(1):1–13. 192. Komatsu Y, Shiratori Y, Kawase T, et al. Role of polymorphonuclear leukocytes in galactosamine hepatitis: mechanism of adherence to hepatic endothelial cells. Hepatology 1994; 20(6):1548–56. 193. Scoazec JY, Feldmann G. The cell adhesion molecules of hepatic sinusoidal endothelial cells. J Hepatol 1994; 20(2):296–300. 194. Knolle PA, Gerken G. Local control of the immune response in the liver. Immunol Rev 2000; 174:21–34. 195. Knolle PA, Limmer A. Control of immune responses by savenger liver endothelial cells. Swiss Med Wkly 2003; 133(37–38):501–6. 196. Sarphie G, D’Souza NB, Van Thiel DH, et al. Dose- and time-dependent effects of ethanol on functional and structural aspects of the liver sinusoid in the mouse. Alcohol Clin Exp Res 1997; 21(6):1128–36. 197. Rachfal AW, Brigstock DR. Connective tissue growth factor (CTGF/CCN2) in hepatic fibrosis. Hepatol Res 2003; 26(1):1–9. 198. Deaciuc IV, Alappat JM, D’Souza NB. Effect of acute and chronic alcohol administration to rats on the expression of interleukin-6 cell-surface receptors of hepatic parenchymal and nonparenchymal cells. Alcohol Clin Exp Res 1994; 18(5):1207–14. 199. Deaciuc IV, D’Souza NB, Spitzer JJ. Tumor necrosis factor-alpha cell-surface receptors of liver parenchymal and nonparenchymal cells during acute and chronic alcohol administration to rats. Alcohol Clin Exp Res 1995; 19(2):332–8. 200. Nanji AA, Griniuviene B, Yacoub LK, et al. Intercellular adhesion molecule-1 expression in experimental alcoholic liver disease: relationship to endotoxemia and TNF alpha messenger RNA. Exp Mol Pathol 1995; 62(1):42–51. 201. Nanji AA, Tahan SR. Association between endothelial cell proliferation and pathologic changes in experimental alcoholic liver disease. Toxicol Appl Pharmacol 1996; 140(1):101–7. 202. Wisse E, Braet F, Luo D, et al. Structure and function of sinusoidal lining cells in the liver. Toxicol Pathol 1996; 24(1):100–11. 203. van Oosten M, van de Bilt E, de Vries HE, et al. Vascular adhesion molecule-1 and intercellular adhesion molecule-1 expression on rat liver cells after lipopolysaccharide administration in vivo. Hepatology 1995; 22(5):1538–46. 204. Mousa SA. Expression of adhesion molecules during cadmium hepatotoxicity. Life Sci 2004; 75(1):93–105. 205. Edwards S, Lalor PF, Nash GB, et al. Lymphocyte traffic through sinusoidal endothelial cells is regulated by hepatocytes. Hepatology 2005; 41(3):451–9. 206. Uhrig A, Banafsche R, Kremer M, et al. Development and functional consequences of LPS tolerance in sinusoidal endothelial cells of the liver. J Leukoc Biol 2005; 77(5):626–33. 207. Thiele GM, Duryee MJ, et al. Rat sinusoidal liver endothelial cells (SECs) produce pro-fibrotic factors in response to adducts formed from the metabolites of ethanol. Biochem Pharmacol 2005; 70(11):1593–600. 208. Jarnagin WR, Rockey DC, Koteliansky VE, et al. Expression of variant fibronectins in wound healing: cellular source and biological activity of the EIIIA segment in rat hepatic fibrogenesis. J Cell Biol 1994; 127(6 Pt 2):2037–48. 209. George J, Wang SS, Sevcsik AM, et al. Transforming growth factor-beta initiates wound repair in rat liver through induction of the EIIIA-fibronectin splice isoform. Am J Pathol 2000; 156(1):115–24. 210. Bautista AP. Chronic alcohol intoxication primes Kupffer cells and endothelial cells for enhanced CC-chemokine production and concomitantly suppresses phagocytosis and chemotaxis. Front Biosci 2002; 7:a117–25. 211. Blomhoff R, Wake K. Perisinusoidal stellate cells of the liver: important roles in retinol metabolism and fibrosis. FASEB J 1991; 5(3):271–7. 212. Hendriks HF, Verhoofstad WA, Brouwer A, et al. Perisinusoidal fat-storing cells are the main vitamin A storage cells in rat liver. Exp Cell Res 1985; 160(1):138–49. 213. Ogata I, Mochida S, Tomiya T, et al. Minor contribution of hepatocytes to collagen production in normal and early fibrotic rat livers. Hepatology 1991; 14(2):361–7. 214. Shiratori Y, Ichida T, Geerts A, et al. Modulation of collagen synthesis by fat-storing cells, isolated from CCl4- or vitamin A- treated rats. Dig Dis Sci 1987; 32(11):1281–9. 215. Arthur MJ, Stanley A, Iredale JP, et al. Secretion of 72 kDa type IV collagnease/gelatinase by cultured human lipocytes. Analysis of gene expression, protein synthesis and proteinase activity. Biochem J 1992; 287(Pt 3):701–7. 216. Friedman SL. Cellular sources of collagen and regulation of collagen production in liver. Semin Liver Dis 1990; 10(1):20–9.

Nonparenchymal Cells and Inflammatory Macrophages

179

217. Friedman SL, Roll FJ, Boyles JK, et al. Hepatic lipocytes: the principle collagen-producing cells of normal rat liver. Proc Natl Acad Sci (USA) 1985; 82(24):8681–5. 218. French SW, Takahashi H, Wong K, et al. Ito cell activation induced by chronic ethanol feeding in the presence of different dietary fats. Alcohol Alcohol Suppl 1991; 1:357–61. 219. Gressner AM. Transdifferentiation of hepatic stellate cells (Ito cells) to myofibroblasts: a key event in hepatic fibrogenesis. Kidney Int Suppl 1996; 54:S39–45. 220. Kim KY, Choi I, Kim SS. Progression of hepatic stellate cell activation is associated with the level of oxidative stress rather than cytokines during CCl4-induced fibrogenesis. Mol Cell 2000; 10(3):289–300. 221. Ikeda K, Wakahara T, Wang YQ, et al. In vitro migratory potential of rat quiescent hepatic stellate cells and its augmentation by cell activation. Hepatology 1999; 29(6):1760–7. 222. Friedman SL. Molecular mechanisms of hepatic fibrosis and principles of therapy. J Gastroenterol 1997; 32(3):424–30. 223. Knittel T, Dinter C, Kobold D, et al. Expression and regulation of cell adhesion molecules by hepatic stellate cells (HSC) of rat liver: involvement of HSC in recruitment of inflammatory cells during hepatic tissue repair. Am J Pathol 1999; 154(1):153–67. 224. Hellerbrand C, Wang SC, Tsukamoto H, et al. Expression of intracellular adhesion molecule 1 by activated hepatic stellate cells. Hepatology 1996; 24(3):670–6. 225. Schieferdecker HL, Schlaf G, Koleva M, et al. Induction of functional anaphylatoxin C5a receptors on hepatocytes by in vivo treatment of rats with IL-6. J Immunol 2000; 164(10):5453–8. 226. Gandhi CR, Uemura T, Kuddus RH. Endotoxin causes up-regulation of endothelin receptors in cultured hepatic stellate cells via nitric oxide-dependent and -independent mechanisms. Br J Pharmacol 2000; 131(2):319–27. 227. Fennekohl A, Schieferdecker HL, Jungermann K, et al. Differential expression of prostanoid receptors in hepatocytes. Kupffer cells, sinusoidal endothelial cells and stellate cells of rat liver. J Hepatol 1999; 30(1):38–47. 228. Tsukamoto H. Cytokine regulation of hepatic stellate cells in liver fibrosis. Alcohol Clin Exp Res 1999; 23(5):911–6. 229. Wong L, Yamasaki G, Johnson RJ, et al. Induction of beta-platelet-derived growth factor receptor in rat hepatic lipocytes during cellular activation in vivo and in culture. J Clin Invest 1994; 94(4):1563–9. 230. Helyar L, Bundschuh DS, Laskin JD, et al. Hepatic fat storing cells produce nitric oxide and hydrogen peroxide in response to bacterially-derived lipopolysaccharide. In: Knook DL, Decker K et al, eds. Cells of the Hepatic Sinusoid, Vol. 4. Amsterdam: Kupffer Cell Foundation, 1993:394–6. 231. Helyar L, Bundschuh DS, Laskin JD, et al. Induction of hepatic Ito cell nitric oxide production by acute endotoxemia. Hepatology 1994; 12(6):1509–15. 232. Athari A, Hanecke K, Jungermann K. Prostaglandin F2 alpha and D2 release from primary Ito cell cultures after stimulation with noradrenaline and ATP but not adenosine. Hepatology 1994; 20(1 Pt 1): 142–8. 233. Xu Y, Rojkind M, Czaja MJ. Regulation of monocyte chemoattractant protein 1 by cytokines and oxygen free radicals in rat hepatic fat-storing cells. Gastroenterology 1996; 110(6):1870–7. 234. Knittel T, Janneck T, Muller L, et al. Transforming growth factor beta 1-regulated gene expression of Ito cells. Hepatology 1996; 24(2):352–60. 235. Ramadori G, Knittel T, Odenthal M, et al. Synthesis of cellular fibronectin by rat liver fat-storing (Ito) cells: regulation by cytokines. Gastroenterology 1992; 103(4):1313–21. 236. Davis BH, Kramer RT, Davidson NO. Retinoic acid modulates rat Ito cell proliferation, collagen, and transforming growth factor beta production. J Clin Invest 1990; 86(6):2062–70. 237. Poli G. Pathogenesis of liver fibrosis: role of oxidative stress. Mol Aspects Med 2000; 21(3):49–98. 238. Britton RS, Bacon BR. Intracellular signaling pathways in stellate cell activation. Alcohol Clin Exp Res 1999; 23(5):922–5. 239. Friedman SL. Stellate cell activation in alcoholic fibrosis–an overview. Alcohol Clin Exp Res 1999; 23(5):904–10. 240. Reeves HL, Friedman SL. Activation of hepatic stellate cells–a key issue in liver fibrosis. Front Biosci 2002; 7:d808–26. 241. Higuchi H, Gores GJ. Mechanisms of liver injury: an overview. Curr Mol Med 2003; 3(6):483–90. 242. Zhan SS, Jiang JX, Wu J, et al. Phagocytosis of apoptotic bodies by hepatic stellate cells induces NADPH oxidase and is associated with liver fibrosis in vivo. Hepatology 2006; 43(3):435–43. 243. Prosser CC, Yen RD, Wu J. Molecular therapy for hepatic injury and fibrosis: where are we? World J Gastroenterol 2006; 12(4):509–15. 244. Sprenger H, Kaufmann A, Garn H, et al. Differential expression of monocyte chemotactic protein-1 (MCP-1) in transforming rat hepatic stellate cells. J Hepatol 1999; 30(1):88–94. 245. Wiltrout RH. Regulation and antimetastatic functions of liver-associated natural killer cells. Immunol Rev 2000; 174:63–76. 246. Morris MA, Ley K. Trafficking of natural killer cells. Curr Mol Med 2004; 4(4):431–8.

180

Laskin and Gardner

247. Kronenberg M, Gapin L. The unconventional lifestyle of NKT cells. Nat Rev Immunol 2002; 2(8):557–68. 248. Emoto M, Miyamoto M, Yoshizawa I, et al. Critical role of NK cells rather than V alpha 14(C)NKT cells in lipopolysaccharide-induced lethal shock in mice. J Immunol 2002; 169(3):1426–32. 249. Takeda K, Hayakawa Y, Van Kaer L, et al. Critical contribution of liver natural killer T cells to a murine model of hepatitis. Proc Natl Acad Sci USA 2000; 97(10):5498–503. 250. Liu ZX, Govindarajan S, Kaplowitz N. Innate immune system plays a critical role in determining the progression and severity of acetaminophen hepatotoxicity. Gastroenterology 2004; 127(6):1760–74. 251. Larrick JW, Kunkel SL. The role of tumor necrosis factor and interleukin 1 in the immunoinflammatory response. Pharm Res 1988; 5(3):129–39. 252. Whicher JT, Evans SW. Cytokines in disease. Clin Chem 1990; 36(7):1269–81. 253. Cerami A. Inflammatory cytokines. Clin Immunol Immunopathol 1992; 62(1 Pt 2):S3–10. 254. Punjabi CJ, Laskin JD, Hwang S-M, et al. Enhanced production of nitric oxide by bone marrow cells and increased sensitivity to macrophage colony-stimulating factor (CSF) and granulocyte-macrophage CSF after benzene treatment of mice. Blood 1994; 83(11):3255–62. 255. Lavnikova N, Drapier J-C, Laskin DL. A single exogenous stimulus activates rat macrophages for nitric oxide production and cytotoxicity. J Leukoc Biol 1993; 54(4):322–8. 256. Sharma RJ, Macallan DC, Sedgwick P, et al. Kinetics of endotoxin-induced acute-phase protein gene expression and its modulation by TNF-a monoclonal antibody. Am J Physiol 1992; 262(5 Pt 2):R786–93. 257. Sujita K, Okuno F, Tanaka Y, et al. Effect of interleukin 1 (IL-1) on the level of cytochrome P-450 involving IL-1 receptor on the isolated hepatocytes of rat. Biochem Biophys Res Commun 1990; 168(3):1217–22. 258. Curran RD, Billiar TR, Stuehr DJ, et al. Multiple cytokines are required to induce hepatocyte nitric oxide production and inhibit total protein synthesis. Ann Surg 1990; 212(4):462–9. 259. Wang JH, Redmond HP, Watson RW, et al. Role of lipopolysaccharide and tumor necrosis factor-a in induction of hepatocyte necrosis. Am J Physiol 1995; 269(2 Pt. 1):G297–304. 260. Leist M, Gantner F, Bohlinger I, et al. Murine hepatocyte apoptosis induced in vitro and in vivo by TNF-a requires transcriptional arrest. J Immunol 1994; 153(4):1778–88. 261. Schwabe RF, Brenner DA. Mechanisms of Liver Injury. I. TNF-alpha-induced liver injury: role of IKK, JNK, and ROS pathways. Am J Physiol Gastrointest Liver Physiol 2006; 290(4):G583–9. 262. Satoh M, Yamazaki M. Tumor necrosis factor stimulates DNA synthesis of mouse hepatocytes in primary culture and is suppressed by transforming growth factor beta and interleukin 6. J Cell Physiol 1992; 150(1):134–9. 263. Wajant H, Pfizenmaier K, Scheurich P. Tumor necrosis factor signaling. Cell Death Differ 2003; 10(1):45–65. 264. Beyer HS, Theologides A. Tumor necrosis factor-alpha is a direct hepatocyte mitogen in the rat. Biochem Mol Biol Int 1993; 29(1):1–4. 265. Yamada Y, Kirillova I, Peschon JJ, et al. Initiation of liver growth by tumor necrosis factor: deficient liver regeneration in mice lacking type I tumor necrosis factor receptor. Proc Natl Acad Sci USA 1997; 94(4):1441–6. 266. Gardner CR, Laskin JD, Dambach DM, et al. Exaggerated hepatotoxicity of acetaminophen in mice lacking tumor necrosis factor receptor-1. Potential role of inflammatory mediators. Toxicol Appl Pharmacol 2003; 192(2):119–30. 267. McClain CJ, Barve S, Deaciuc IV, et al. Cytokines in alcoholic liver disease. Semin Liver Dis 1999; 19(2):205–19. 268. Scotte M, Hiron M, Masson S, et al. Differential expression of cytokine genes in monocytes, peritoneal macrophages and liver following endotoxin- or turpentine-induced inflammation in rat. Cytokine 1996; 8(2):115–20. 269. Horn TL, O’Brien TD, Schook LB, et al. Acute hepatotoxicant exposure induces TNFR-mediated hepatic injury and cytokine/apoptotic gene expression. Toxicol Sci 2000; 54(1):262–73. 270. Wang J, Wendel A. Studies on the hepatotoxicity of galactosamine/endotoxin or galactosamine /TNF in the perfused mouse liver. Biochem Pharmacol 1990; 39(2):267–70. 271. Schook LB, Lockwood JF, Yang SD, et al. Dimethylnitrosamine (DMN)-induced IL-1 beta, TNFalpha, and IL-6 inflammatory cytokine expression. Toxicol Appl Pharmacol 1992; 116(1):110–6. 272. Bourdi M, Masubuchi Y, Reilly TP, et al. Protection against acetaminophen-induced liver injury and lethality by interleukin 10: role of inducible nitric oxide synthase. Hepatology 2002; 35(2):289–98. 273. Ishida Y, Kondo T, Ohshima T, et al. A pivotal involvement of IFN-g in the pathogenesis of acetaminophen-induced acute liver injury. FASEB J 2002; 16(10):1227–36. 274. Shedlofsky SI, McClain CJ. Hepatic dysfunction due to cytokines. In: Kimball ES, ed. Cytokines and Inflammation. Boca Raton, FL: CRC Press, 1991:235–73. 275. Dinarello CA. Interleukin-1 and its biologically related cytokines. Adv Immunol 1989; 44:153–205. 276. Tracey KJ, Beutler B, Lowry SF, et al. Shock and tissue injury induced by recombinant human cachectin. Science 1986; 234(4775):470–4.

Nonparenchymal Cells and Inflammatory Macrophages

181

277. Fiedler VB, Loof I, Sander E, et al. Monoclonal antibody to tumor necrosis factor-alpha prevents lethal endotoxin sepsis in adult rhesus monkeys. J Lab Clin Med 1992; 120(4):574–88. 278. Czaja MJ, Xu J, Alt E. Prevention of carbon tetrachloride-induced rat liver injury by soluble tumor necrosis factor receptor. Gastroenterology 1995; 108(6):1849–54. 279. Dinarello CA. Interleukin-1 and interleukin-1 antagonism. Blood 1991; 77(8):1627–52. 280. Josephs MD, Solorzano CC, Taylor M, et al. Modulation of the acute phase response by altered expression of the IL-1 type 1 receptor or IL-1ra. Am J Physiol Regul Integr Comp Physiol 2000; 278(4):R824–30. 281. Iimuro Y, Gallucci RM, Luster MI, et al. Antibodies to tumor necrosis factor alfa attenuate hepatic necrosis and inflammation caused by chronic exposure to ethanol in the rat. Hepatology 1997; 26(6):1530–7. 282. Yamada Y, Fausto N. Deficient liver regeneration after carbon tetrachloride injury in mice lacking type 1 but not type 2 tumor necrosis factor receptor. Am J Pathol 1998; 152(6):1577–89. 283. Cuzzocrea S, de Sarro G, Constantino G, et al. Role of interleukin-6 in a non-septic shock model induced by zymosan. Eur Cytokine Netw 1999; 10(2):191–203. 284. Kovalovich K, DeAngelis PA, Li W, et al. Increased toxin-induced liver injury and fibrosis in interleukin-6-deficient mice. Hepatology 2000; 31(1):149–59. 285. Saile B, Matthes N, Knittel T, et al. Transforming growth factor-b and tumor necrosis factor-a inhibit both apoptosis and proliferation of activated rat hepatic stellate cells. Hepatology 1999; 30(1):196–202. 286. Dooley S, Delvoux B, Lahme B, et al. Modulation of transforming growth factor beta response and signaling during transdifferentiation of rat hepatic stellate cells to myofibroblasts. Hepatology 2000; 31(5):1094–106. 287. Jarvelainen HA, Fang C, Ingelman-Sundberg M, et al. Effect of chronic coadministration of endotoxin and ethanol on rat liver pathology and proinflammatory and anti-inflammatory cytokines. Hepatology 1999; 29(5):1503–10. 288. Hellerbrand C, Stefanovic B, Giordano F, et al. The role of TGFb1 in initiating hepatic stellate cell activation in vivo. J Hepatol 1999; 30(1):77–87. 289. Schnabl B, Kweon YO, Frederick JP, et al. The role of Smad3 in mediating mouse hepatic stellate cell activation. Hepatology 2001; 34(1):89–100. 290. Feng L. Role of chemokines in inflammation and immunoregulation. Immunol Res 2000; 21(2-3): 203–10. 291. Czaja MJ, Geerts A, Xu J, et al. Monocyte chemoattractant protein 1 (MCP-1) expression occurs in toxic rat liver injury and human liver disease. J Leukoc Biol 1994; 55(1):120–6. 292. Ohkubo K, Masumoto T, Horiike N, et al. Induction of CINC (interleukin-8) production in rat liver by non-parenchymal cells. J Gastroenterol Hepatol 1998; 13(7):696–702. 293. Kopydlowski KM, Salkowski CA, Cody MJ, et al. Regulation of macrophage chemokine expression by lipopolysaccharide in vitro and in vivo. J Immunol 1999; 163(3):1537–44. 294. Zhang P, Xie M, Zagorski J, et al. Attenuation of hepatic neutrophil sequestration by anti-CINC antibody in endotoxic rats. Shock 1995; 4(4):262–8. 295. Koniaris LG, Zimmers-Koniaris T, Hsiao EC, et al. Cytokine-responsive gene-2/IFN-inducible protein-10 expression in multiple models of liver and bile duct injury suggests a role in tissue regeneration. J Immunol 2001; 167(1):399–406. 296. Rai RM, Loffreda S, Karp CL, et al. Kupffer cell depletion abolishes induction of interleukin-10 and permits sustained overexpression of tumor necrosis factor alpha messenger RNA in the regenerating rat liver. Hepatology 1997; 25(4):889–95. 297. Zisman DA, Kunkel SL, Strieter RM, et al. MCP-1 protects mice in lethal endotoxemia. J Clin Invest 1997; 99(12):2832–6. 298. Gardner CR, Laskin DL. Protective and pathologic roles of nitric oxide in tissue injury. In: Laskin JD, Laskin DL, eds. Cellular and Molecular Biology of Nitric Oxide. New York: Marcel Dekker, 1999:225–46. 299. Shiratori Y, Takada H, Hai K, et al. Generation of chemotactic factor by hepatocytes isolated from chronically ethanol-fed rats. Dig Dis Sci 1992; 37(5):650–8. 300. Laskin DL, Pilaro AM, Ji S. Potential role of macrophages in acetaminophen hepatotoxicity, II. Mechanisms of macrophage accumulation and activation. Toxicol Appl Pharmacol 1986; 86(2):216–26. 301. Mawet E, Shiratori Y, Hikiba Y, et al. Cytokine-induced neutrophil chemoattractant release from hepatocytes is modulated by Kupffer cells. Hepatology 1996; 23(2):353–8. 302. Thornton AJ, Ham J, Kunkel SL. Kupffer cell-derived cytokines induce the synthesis of a leukocyte chemotactic peptide, interleukin-8, in human hepatoma and primary hepatocyte cultures. Hepatology 1991; 14(6):1112–22. 303. Louis H, Van Laethem J-L, Wu W, et al. Interleukin-10 controls neutrophilic infiltration, hepatocyte proliferation, and liver fibrosis induced by carbon tetrachloride in mice. Hepatology 1998; 28(6):1607–15.

182

Laskin and Gardner

304. Matsukawa A, Hoagboam CM, Lukacs NW, et al. Expression and contribution of endogenous IL-13 in an experimental model of sepsis. J Immunol 2000; 164(5):2738–44. 305. Minty A, Chalon P, Derocq J-M, et al. Interleukin-13 is a new human lymphokine regulating inflammatory and immune responses. Nature 1993; 362(6417):248–50. 306. de Waal Malefyt R, Figdor CG, Huijbens RG, et al. Effects of IL-13 on phenotype, cytokine production, and cytotoxic function of human monocytes. Comparison with IL-4 and modulation by IFN-g or IL-10. J Immunol 1993; 151(11):6370–81. 307. Loyer P, Ilyin G, Abdel-Razzak Z, et al. Interleukin 4 inhibits the production of some acute-phase proteins by human hepatocytes in primary culture. FEBS Lett 1993; 336(2):215–20. 308. Sugimoto R, Enjoji M, Nakamuta M, et al. Effect of IL-4 and IL-13 on collagen production in cultured LI90 human hepatic stellate cells. Liver Int 2005; 25(2):420–8. 309. Gabay C, Porter B, Guenette D, et al. Interleukin-4 (IL-4) and IL-13 enhance the effect of IL-1 beta on production of IL-1 receptor antagonist by human primary hepatocytes and hepatoma HepG2 cells: differential effect on C-reactive protein production. Blood 1999; 93(4):1299–307. 310. Babior BM. Oxidants from phagocytes: agents of defense and destruction. Blood 1984; 64(5):959–66. 311. Halliwell B, Gutteridge JMC. Oxygen toxicity, oxygen radicals, transition metals and disease. Biochem J 1984; 219(1):1–14. 312. Rubin R, Farber JL. Mechanisms of the killing of cultured hepatocytes with hydrogen peroxide. Arch Biochem Biophys 1984; 228(2):450–9. 313. Sevanian A, Hochstein P. Mechanisms and consequences of lipid peroxidation in biological systems. Annu Rev Nutr 1985; 5:365–90. 314. Towner RA, Reinke LA, Janzen EG, et al. In vivo magnetic resonance imaging study of Kupffer cell involvement in CCl4-induced hepatotoxicity in rats. Can J Physiol Pharmacol 1994; 72(5):441–6. 315. Amimoto T, Matsura T, Koyama SY, et al. Acetaminophen-induced hepatic injury in mice: the role of lipid peroxidation and effects of pretreatment with coenzyme Q10 and alpha-tocopherol. Free Radic Biol Med 1995; 19(2):169–76. 316. Sugino K, Dohi K, Yamada K, et al. Changes in the levels of endogenous antioxidants in the liver of mice with experimental endotoxemia and the protective effects of antioxidants. Surgery 1989; 105(2 Pt 1):200–6. 317. Nussler AK, Beger H-G, Liu ZZ, et al. Nitric oxide, hepatocytes and inflammation. Res Immunol 1995; 146(9):671–7. 318. Milbourne EA, Bygrave FL. Does nitric oxide play a role in liver function? Cell Signal 1995; 7(4):313–8. 319. Losser M-R, Payen D. Mechanisms of liver damage. Sem Liver Dis 1996; 16(4):357–67. 320. Moncada S, Palmer RM, Higgs EA. Nitric oxide: physiology, pathophysiology, and pharmacology. Pharmacol Rev 1991; 43(2):109–42. 321. Lavnikova N, Lakhotia A, Patel N, et al. Cytostasis is required for IL-1 induced nitric oxide production in transformed hamster fibroblasts. J Cell Physiol 1996; 169(3):532–7. 322. Laskin DL, Rodriguez del Valle M, Heck DE, et al. Hepatic nitric oxide production following acute endotoxemia in rats is mediated by increased nitric oxide synthase gene expression. Hepatology 1995; 22(1):223–34. 323. Gross SS, Wolin MS. Nitric oxide: pathophysiological mechanisms. Annu Rev Physiol 1995; 57:737–69. 324. Hodgson PD, Renton KW. The role of nitric oxide generation in interferon-evoked cytochrome P450 down-regulation. Int J Immunopharmacol 1995; 17(12):995–1000. 325. MacMicking J, Xie QW, Nathan C. Nitric oxide and macrophage function. Annu Rev Immunol 1997; 15:323–50. 326. Lavnikova N, Burdelya L, Lakhotia A, et al. Macrophage and interleukin-1 induced nitric oxide production and cytostasis in hamster tumor cells varying in malignant potential. J Leukoc Biol 1997; 61(4):452–8. 327. Beckman J. The double-edged role of nitric oxide in brain function and superoxide-mediated injury. J Dev Physiol 1991; 15(1):53–9. 328. Beckman JS, Crow JP. Pathological implications of nitric, superoxide and peroxynitrite formation. Biochem Soc Trans 1993; 21(2):330–4. 329. Freeman B. Free radical chemistry of nitric oxide. Looking at the dark side. Chest 1994; 105(Suppl. 3): 79S–84. 330. Beckman JS, Koppenol WH. Nitric oxide, superoxide, and peroxynitrite: the good, the bad, and ugly. Am J Physiol 1996; 271(5 Pt. 1):C1424–37. 331. Radi R, Beckman JS, Bush KM, et al. Peroxynitrite oxidation of sulfhydrals. The cytotoxic potential of superoxide and nitric oxide. J Biol Chem 1991; 266(7):4244–50. 332. Radi R, Beckman JS, Bush KM, et al. Peroxynitrite-induced membrane lipid peroxidation: the cytotoxic potential of superoxide and nitric oxide. Arch Biochem Biophys 1991; 288(2):481–7. 333. Lin KT, Xue JY, Nomen M, et al. Peroxynitrite-induced apoptosis in HL-60 cells. J Biol Chem 1995; 270(28):16487–90.

Nonparenchymal Cells and Inflammatory Macrophages

183

334. Ischiropoulos H, Zhu L, Chen J, et al. Peroxynitrite-mediated tyrosine nitration catalyzed by superoxide dismutase. Arch Biochem Biophys 1992; 298(2):431–7. 335. Rockey DC, Chung JJ. Regulation of inducible nitric oxide synthase and nitric oxide during hepatic injury and fibrogenesis. Am J Physiol 1997; 273(1 Pt. 1):G124–30. 336. Spitzer JA, Spitzer JJ. Lipopolysaccharide tolerance and ethanol modulate hepatic nitric oxide production in a gender-dependent manner. Alcohol 2000; 21(1):27–35. 337. Hinson JA, Pike SL, Pumford NR, et al. Nitrotyrosine-protein adducts in hepatic centrilobular areas following toxic doses of acetaminophen in mice. Chem Res Toxicol 1998; 11(6):604–7. 338. Wright CE, Rees DD, Moncada S. Protective and pathological roles of nitric oxide in endotoxin shock. Cardiovasc Res 1992; 26(1):48–57. 339. Ialenti A, Ianaro A, Moncada S, et al. Modulation of acute inflammation by endogenous nitric oxide. Eur J Pharmacol 1992; 211(2):177–82. 340. Harbrecht BG, Billiar TR, Stadler J, et al. Inhibition of nitric oxide synthesis during endotoxemia promotes intrahepatic thrombosis and an oxygen radical-mediated hepatic injury. J Leukoc Biol 1992; 52(4):390–4. 341. Muriel P. Nitric oxide protection of rat liver from lipid peroxidation, collagen accumulation, and liver damage induced by carbon tetrachloride. Biochem Pharmacol 1998; 56(6):773–9. 342. Taylor BS, Alarcon LH, Billiar TR. Inducible nitric oxide synthase in the liver: regulation and function. Biochemistry (Mosc) 1998; 63(7):766–81. 343. vom Dahl S, Hallbrucker C, Lang F, et al. Role of eicosanoids, inositol phosphates and extracellular Ca2C in cell-volume regulation of rat liver. Eur J Biochem 1991; 198(1):73–83. 344. Ogle CK, Wu JZ, Alexander JW, et al. The effects of in vivo administration of endotoxin on the functions and interaction of hepatocytes and Kupffer cells. Prostaglandins 1991; 41(2):169–83. 345. Enomoto N, Ikejima K, Yamashina S, et al. Kupffer cell-derived prostaglandin E2 is involved in alcohol-induced fat accumulation in rat liver. Am J Physiol Gastrointest Liver Physiol 2000; 279(1):G100–6. 346. Horton AA, Wood JM. Effects of inhibitors of phospholipase A2, cyclooxygenase and thromboxane synthase on paracetamol hepatotoxicity in the rat. Eicosanoids 1989; 2(2):123–9. 347. Mancuso G, Cusumano V, Cook JA, et al. Efficacy of tumor necrosis factor alpha and eicosanoid inhibitors in experimental models of neonatal sepsis. FEMS Immunol Med Microbiol 1994; 9(1):49–54. 348. Shiratori Y, Tanaka M, Umihara J, et al. Leukotriene inhibitors modulate hepatic injury induced by lipopolysaccharide-activated macrophages. J Hepatol 1990; 10(1):51–61. 349. Tiegs G, Wendel A. Leukotriene-mediated liver injury. Biochem Pharmacol 1988; 37(13):2569–73. 350. Casey LC, Fletcher JR, Zmudka MI, et al. The role of thromboxane in primate endotoxin shock. J Surg Res 1985; 39(2):140–9. 351. Smith EF, III, Jugus M, Kinter LB. Effect of thromboxane receptor antagonist, BM 13.505, on the sequelae of endotoxemia in the conscious rat. Eicosanoids 1998; 1(1):27–33. 352. Takano M, Nishimura H, Kimura Y, et al. Prostaglandin E2 protects against liver injury after Escherichia coli infection but hampers the resolution of the infection in mice. J Immunol 1998; 161(6):3019–25. 353. Muntane´ J, Montero JL, Lozano JM, et al. TNF-a but not IL-1a is correlated with PGE1-dependent protection against acute D-galactosamine-induced liver injury. Can J Gastroenterol 2000; 14(3):175–80. 354. Muntane´ J, Rodriguez FJ, Segado O, et al. TNF-a dependent production of inducible nitric oxide is involved in PGE1 production against acute liver injury. Gut 2000; 47(4):553–62. 355. Hagmann W, Denzlinger C, Keppler D. Role of peptide leukotrienes and their hepatobilliary elimination in endotoxin action. Circ Shock 1984; 14(4):223–35. 356. Henderson WR. The role of leukotrienes in inflammation. Ann Int Med 1994; 121(9):684–97. 357. Goetzl EJ, Pickett WC. Novel structural determinants of the human neutrophil chemotactic activity of leukotriene B. J Exp Med 1981; 153(2):482–7. 358. Shimada K, Navarro J, Goeger DE, et al. Expression and regulation of leukotriene-synthesis enzymes in rat liver cells. Hepatology 1998; 28(5):1275–81. 359. Dieter P, Schulze-Specking A, Decker K. Differential inhibition of prostaglandin and superoxide production by dexamethasone in primary cultures of rat Kupffer cells. Eur J Biochem 1986; 159(3):451–7. 360. Sakaguchi T, Nakamura S, Suzuki S, et al. Participation of platelet-activating factor in the lipopolysaccharide-induced liver injury in partially hepatectomized rats. Hepatology 1999; 30(4):959–67. 361. Anderson BO, Bensard DD, Harken AH. The role of platelet activating factor and its antagonists in shock, sepsis and multiple organ failure. Surg Gynecol Obstet 1991; 172(5):415–24. 362. Gardner CR, Laskin JD, Laskin DL. Distinct biochemical responses of hepatic macrophages and endothelial cells to platelet-activating factor during endotoxemia. J Leukoc Biol 1995; 57(2):269–74.

184

Laskin and Gardner

363. Yue TL, Farhat M, Rabinovici R, et al. Protective effect of BN 50739, a new platelet-activating factor antagonist, in endotoxin-treated rabbits. J Pharmacol Exp Ther 1990; 254(3):976–81. 364. Tanner A, Keyhani A, Wright R. The influence of endotoxin in vitro on hepatic macrophage lysosomal enzyme release in different models of hepatic injury. Liver 1983; 3(3):151–60. 365. Magilavy DB, Zhan R, Black DD. Modulation of murine hepatic lipase activity by exogenous and endogenous Kupffer-cell activation. Biochem J 1993; 292(Pt 1):249–52. 366. Winwood PJ, Schuppan D, Iredale JP, et al. Kupffer cell-derived 95-kd type IV collagenase/gelatinase B: characterization and expression in cultured cells. Hepatology 1995; 22(1):304–15. 367. Peinado-Onsurbe J, Soler C, Galan X, et al. Involvement of catecholamines in the effect of fasting on hepatic endothelial lipase activity in the rat. Endocrinology 1991; 129(5):2599–606. 368. Hironaka K, Sakaida I, Matsumura Y, et al. Enhanced interstitial collagenase (matrix metalloproteinase-13) production of Kupffer cell by gadolinium chloride prevents pig serum-induced rat liver fibrosis. Biochem Biophys Res Commun 2000; 267(1):290–5. 369. Arthur MJ. Degradation of matrix proteins in liver fibrosis. Pathol Res Pract 1994; 190(9–10):825–33. 370. Ueno T, Sujaku K, Tamaki S, et al. OK-432 treatment increases matrix metalloproteinase-9 production and improves dimethylnitrosamine-induced liver cirrhosis in rats. Int J Mol Med 1999; 3(5):497–503.

10

Role of Tissue Repair in Liver Injury Harihara M. Mehendale

Department of Toxicology, College of Pharmacy, University of Louisiana at Monroe, Monroe, Louisiana, U.S.A.

INTRODUCTION Historically, the fate of a drug and its adverse effects in the body were estimated based on the established rules of toxicokinetics and toxicodynamics. It was observed that drugs and toxic chemicals follow the rules of absorption, distribution, metabolism, and excretion (1). Often, central to the pharmacological actions of many drugs was their phase I metabolism mediated by the drug-biotransforming enzymes, such as cytochrome P450 and others. After the drugs are absorbed and distributed in the body, they undergo metabolism via a combination of phase I and II biotransforming enzymes to generate water-soluble metabolites, which would be easily excreted from the body (2). However, it was observed that metabolism of drugs and toxicants may also lead to generation of highly reactive metabolites and free radicals which attack cellular macromolecules and inflict cell and tissue injury (2). While this is widely accepted as a generalized mechanism of initiation of liver injury, it was understood that continuation or progressive expansion of injury occurs through other mechanisms, these mechanisms have largely remained obscure until recently (3–6). Liver is the main site of drug and toxicant metabolism since the hepatocytes are a reservoir of microsomal and cytosolic phase I and II drug-metabolizing enzymes (2). This has made liver a prime target for drug- and toxicant-induced injury. The degree of liver injury was thought to be proportional to the generation of reactive metabolites of the drugs or toxicants via drug-metabolizing enzymemediated bioactivation. It is now known that such oversimplified concepts overlook the imposing effects of biological responses to toxic injury which control the final toxic outcomes. Very little was known about the opposing toxicodynamic response of tissue repair following chemical-induced liver injury (7–9). The extraordinary ability of liver to regenerate upon surgical resection or tissue injury has been known since prehistoric times (10). Liver regeneration has been studied in detail in a variety of animal models after two-thirds partial hepatectomy, rodents serving as the principal models (11,12). Studies have revealed many individual steps of the intricate signal transduction network consisting of chemokines, cytokines, growth factors, and hormones that governs liver regeneration following surgical removal of liver in partial hepatectomy (11). Investigations during last quarter century have revealed that a similar dynamic regeneration response or tissue repair occurs following cell death and tissue injury after exposure to toxic drugs and chemicals (13–18). Upon initiation of toxic injury, a cascade of distress signals is triggered (Fig. 1), which stimulates surrounding healthy cells to divide in order to replace the dead cells (13,20–24). However, such promitogenic signaling is inhibited after exposure to high doses resulting in inhibition of tissue repair (16,20,21,25). These findings have been critical in understanding the underlying mechanism of the dose–response for restorative tissue repair. Further investigations have revealed that tissue repair is affected by a variety of factors, including species (26,27) and strain (28), age (29–32), nutrition (33,34), caloric restriction (35), and disease (36–39). Although tissue repair has been studied in other tissues, such as blood (40), lung (41), and kidney (42), this chapter will focus primarily on liver tissue repair. These studies indicate that ability to mount an effective tissue repair following toxicant exposure can impact the final outcome, viz. regression of injury and recovery or escalation of injury and organ failure leading to survival or death following drug overdose or

186

Mehendale

Stage I

Stage II

Chemical toxicants

· Free radical formation · Reactive intermediate formation · Lipid peroxidation · Blockade of endogenous pathways

? −

Initiation of injury

?

Death

Distress signals

Progression of injury

Progression of injury

Tissue repair

+

?

Regression of injury

Survival

FIGURE 1 Mechanistic two-stage model of toxicity. During stage I, toxic chemicals initiate tissue injury via the well-established bioactivation-based events. Injury further progresses due to some unknown mechanisms. During stage II, infliction of injury stimulates restorative tissue repair response after treatment with low to moderate doses of the toxicant (C tissue repair) leading to prompt regression of injury permitting animal survival. In contrast, high dose of the toxicants inhibits tissue repair (Ktissue repair) further leading to unrestrained progression of injury, and animal death. Source: Adapted from Ref. 19.

other toxic exposures. Detailed studies of tissue repair following toxicant-induced injury have led to a two-stage model of toxicity (3). TWO-STAGE MODEL OF TOXICITY Numerous studies have established the determining effect of compensatory tissue repair in the final outcome of liver injury, i.e., progression or regression of injury (13,18,23,29,33–35,37,38, 43–48). These studies have made it possible to recognize the existence of two distinct stages of toxicity (Fig. 1). Stage I is the initiation stage in which drugs and toxicants initiate liver injury through well-established mechanisms (Fig. 1) tempered by the net effect of bioactivation of the drug, or other mechanisms of causation of injury and detoxification processes. Stage II is represented by progression or regression phase of injury corresponding with the absence or presence of restorative tissue repair, respectively. Repopulation of lost cells and tissue repair stimulated after exposure to low to moderate doses of toxicants restrain injury and restore liver structure and function resulting in recovery (16,25), while high doses of toxicants inhibit restorative tissue repair leading to unrestrained progression of liver injury and animal death (3,16,25,48). The two-stage model of toxicity emphasizes the critical role of opposing and dynamic interplay of progression and regression of acute toxic injury in determining the final outcome (5). TISSUE REPAIR FOLLOWS A DOSE–RESPONSE Intuitively, one would expect that restorative tissue repair response obeys the cardinal rule of dose–response, just as the pharmacologic or toxicologic actions of drugs and toxicants do. Indeed, time-course studies with increasing doses of liver toxicants have revealed that tissue repair follows the golden rule of toxicology, dose–response (16,43,44,48). Liver tissue repair increases in a dose-dependent fashion until a threshold dose is reached. Low to moderate doses stimulate tissue repair incrementally with the dose. However, with each incremental dose, there is a corresponding delay in the onset of tissue repair (16). Timely onset of tissue repair is important because during the time lost before the onset of tissue repair, tissue injury progresses (16,34,36,39). To an extent up to the threshold dose, higher incremental response in tissue repair offsets the delay due to increased toxicant dose as illustrated by thioacetamideinduced hepatic injury–tissue repair model (16). At doses beyond the threshold, tissue repair is inhibited and what little restorative tissue repair does occur, is much delayed and too little to arrest the accelerated progression of injury leading to liver failure and death (16). This concept works for all hepatotoxicants tested thus far (19). As noted, a classic example of the dose dependency of tissue repair is thioacetamideinduced liver injury and tissue repair (16). In contrast to many other hepatotoxicants, thioacetamide offers the advantage of large window of time (3.5–7 days) before liver failure

187

Role of Tissue Repair in Liver Injury

and death of animals. This is a distinct advantage over the other classic hepatotoxicants, such as CCl4, acetaminophen (APAP), CHCl3, etc. where animals die from lethal doses within 12 to 24 hours (25,37,44,49). With thioacetamide, the incline slope of injury after its initiation as well as the decline slope of injury that follows the onset of restorative tissue repair can be examined. Thioacetamide is known to be eliminated with a t1/2 of 2.5 hours, although it is known to be dose dependent because of bioactivation saturation kinetics (50,51). In a seminal study, male Sprague–Dawley rats were exposed to four increasing doses of thioacetamide, 50, 150, 300, and 600 mg/kg. Changes in liver injury and tissue repair were measured after the injection of thioacetamide over a time-course of 0 to 96 hours. Surprisingly, liver injury induced by the first three doses of thioacetamide did not yield a dose–response over a sixfold range. No deaths occurred with these three doses. After the administration of high dose (600 mg/kg), known to be lethal, initiation of liver injury was significantly lower during the early time-points. However, injury aggressively progressed only beyond 48 to 60 hours after the administration of thioacetamide, well after complete elimination of this toxicant (Fig. 2). As expected, with this high dose of thioacetamide, 90% mortality was observed. Tissue repair (3H-thymidine incorporation and PCNA analysis) was inhibited and much delayed after this high dose (Fig. 2). A negligible increase in tissue repair was observed as late as 72 hours following thioacetamide administration, which was too late and too little to rescue the rats from aggressive expansion of injury, liver failure, and animal death. These data elegantly illustrate the dose-dependent stimulation of tissue repair until a threshold dose (between 300 and 600 mg thioacetamide/kg) and accordingly, tissue repair is inhibited beyond the threshold dose (16). The dose-dependent increase in tissue repair has been established with a number of toxicants (Table 1), such as CCl4 (25), chloroform (CHCl3) (44), trichloroethylene (TCE), and allyl alcohol (AA) (6,56).

50 mg/kg 150 mg/kg 300 mg/kg 600 mg/kg

ALT x 102 UL

10 8 6 4 2 0

0

6

12

24

36

48

72

96

3H-T

Incorporation (DPM/μg DNA)

(A)

40 30 20 10 0

0

6

12

24

36

48

72

Hours after TA administration (B)

96

FIGURE 2 Dose response for liver injury and tissue repair after thioacetamide administration to rats. Male Sprague Dawley rats were divided into four groups. At time zero, rats received i.p. injection of 50, 150, 300, and 600 mg thioacetamide/kg. Controls received normal saline. (A) Plasma ALT measured as a marker of liver injury. (B) [3H]-Thymidine incorporation into hepatocellular nuclear DNA as a marker of hepatocellular regeneration. Abbreviation: ALT, alanine aminotransferase. Source: Adapted from Ref. 16.

188

Mehendale

TABLE 1 Examples of Drugs and Toxicants Known to Stimulate Tissue Repair in Liver Chemical Single chemicals Acetaminophen Allyl alcohol Carbon tetrachloride Chloroform 1,2-dichlorobenzene Thioacetamide Trichloroethylene Mixtures and/or combinations ChlordeconeCcarbon tetrachloride IsopropanolCcarbon tetrachloride ChloroformCtrichloroethylene ChloroformCtrichloroethyleneCallyl alcohol ThioacetamideCchloroformCtrichloroethyleneCallyl alcohol

Reference (37,52) (6,43) (36,53,54) (6,44) (28) (16,23,35,39,47,50,55) (6,56) (30,31,57) (6) (6) (6,43)

Dose–Response for Tissue Repair for Mixtures of Hepatotoxicants Several studies have established that mixtures of toxicants also stimulate a dose-dependent tissue repair (6,43). Studies with binary, ternary, and quaternary mixtures, such as TCEC CHCl3, TCECCHCl3CAA, and TCECCHCl3CAACTA suggest that tissue repair is stimulated with the lower doses of mixtures and is inhibited at higher doses (6), suggesting that dose–response relationships for restorative tissue repair are preserved for mixtures of toxicants just as for the individual compounds (6). Although a variety of toxicants individually and in mixtures (Table 1) stimulate tissue repair in a dose-dependent fashion, and the exact mechanisms regulating this response remain to be investigated (6,18,43,44). Detailed analysis of cell division cycle following toxicant exposure over a time-course indicates that high doses inhibit cell cycle progression, especially between G1- and S-phase of cell cycle (18,25,33,34,36,39,46,58–63). The cytokine/growth factormediated signaling and other genes such as cyclin D1 involved in the cell cycle and the effect of high-dose treatment on these factors are of continuing interest (23). Tissue Repair as a Determinant of the Final Outcome of Toxicity Although time-course studies of tissue repair as a function of various doses of toxicants indicate that tissue repair plays an important role in the final outcome, i.e., survival versus death in many studies (3,18,37,48), conclusive evidence supporting such a conclusion comes from a number of interventional studies primarily employing two strategies: (1) antimitosis studies where tissue repair was deliberately inhibited and (2) preplacement of tissue repair in autoprotection and heteroprotection models. Additional evidence also comes from studies in which the influence of existing disease on toxicodynamics of tissue repair has been investigated. One very successful strategy to demonstrate the critical importance of restorative tissue repair in driving the final outcome of liver injury is to intervene with cell division (Table 2) and tissue repair that oppose progression of injury. Colchicine is an antimitotic agent that inhibits cell division by two separate mechanisms, thus arresting cell division (66). First, DNA synthesis is inhibited so that cells cannot enter the S-phase of cell division cycle (67). Second, it also inhibits microtubular formation so that the cells that are in advanced stages of cell division cycle, cannot divide (66). In a classic colchicine antimitosis experiment, transient liver toxicity of a very low dose of CCl4 (100 ml/kg, i.p.), which is normally overcome by 24 hours, was enhanced by colchicine (1 mg/kg) administration two hours prior to CCl4 administration (53). In colchicine-treated rats, recovery was delayed by 48 hours. Enhancement and prolongation were neither associated with increased bioactivation of CCl4 nor any other toxicokinetic considerations. Colchicine administration at crucial time-points well after toxicant-initiated injury (150 and 300 mg thioacetamide/kg), but prior to or during tissue repair resulted in

189

Role of Tissue Repair in Liver Injury

TABLE 2 Colchicine Antimitosis After Completion of Drug Bioactivation Leads to Expansion of Liver Injury Chemicals Healthy animal models AcetaminophenCcolchicine Carbon tetrachlorideCcolchicine ChlordeconeCcarbon tetrachlorideCcolchicine 1,2-DichlorobenzeneCcolchicine ThioacetamideCcolchicine ThioacetamideCacetaminophenCcolchicine (heteroprotection) ThioacetamideCthioacetamide (high dose) (autoprotection) Physiologically diseased animal models Diabetic mice Type 1 diabetes Type 2 diabetes

Reference (23) (53) (30,57,64) (28) (65) (46) (59) (37) (52)

Colchicine is administered at a dose of 1 mg/kg, i.p., approximately six hours before the anticipated stimulation of S-phase DNA synthesis. Since colchicine is a complete antimitotic agent (inhibition of S-phase, inhibition of microtubule function), it has been very useful in antimitotic intervention.

complete inhibition of cell proliferation and tissue repair. In essence, this resulted in transformation of these normally nonlethal doses of thioacetamide (150 and 300 mg/kg) into 100% lethal doses (65). Analysis of tissue repair indicated that colchicine-inhibited cell proliferation and tissue repair allowing the liver injury to progress and expand leading to liver failure and animal death. Similar results were obtained in another model of toxicity of a combination of prior administration of a nontoxic dose of chlordecone and a single marginally toxic dose of CCl4 (30). Previous studies had indicated that 45-day-old male Sprague–Dawley rats exposed to nontoxic dose of chlordecone (10 ppm for 15 days in the diet)CCCl4 (100 ml/kg, i.p., single injection) exhibit 25% mortality. Treatment of chlordeconeCCCl4-exposed 45-day-old rats with colchicine (1 mg/kg, i.p.), resulted in increased mortality from 25% to 85%, which was associated with a significant decrease in tissue repair in the colchicine-treated rats (64). Similar increase in mortality was observed in Fischer 344 (F344) rats treated with 1,2-dichlorobenzene (o-DCB) upon antimitotic intervention with colchicine (Tables 1 and 2) (68). Diabetic mice are resilient to normally lethal doses of hepatotoxicants, such as APAP, carbon tetrachloride, bromobenzene (37,52), and thioacetamide (38). Neither toxicokinetics nor bioactivation of these toxicants was altered in diabetic mice indicating that resiliency could not be due to these factors (37,38,52). Investigations revealed that resiliency is due to the advancement of normally quiescent hepatocytes to S-phase of cell division cycle in the livers of diabetic mice. Inhibition of this cell enhancement by colchicine intervention in the diabetic mice before the administration of the toxicants led to loss of this resiliency (37,38,52). Collectively, these observations indicate that antimitotic intervention of tissue repair leads to progression of liver injury regardless of how low the initial injury is and highlight the importance of tissue repair in the final outcome of toxicity (Table 2). Another strategy to study the role of tissue repair in the final outcome of toxicity is by preplacement of tissue repair using autoprotection and heteroprotection models (13,16,17,19,60,69). By administration of low dose of drug “X” to initiate liver injury, tissue repair is stimulated that further protects against a subsequently administered lethal dose of the same drug “X” (autoprotection) or an entirely different drug “Y” (heteroprotection). The first small dose of the drug “X” initiates promitogenic cellular signals and essentially preplaces tissue repair, which serves to inhibit progression of injury initiated by the subsequently administered normally lethal dose of drug “X” or “Y” and prevents liver failure on the one hand and augments restorative tissue repair on the other, protecting the animals. Autoprotection has been studied using CCl4 (60), thioacetamide (59), and APAP (13,17), while heteroprotection has been investigated using thioacetamide and APAP combination (46). Preplacement of tissue repair can also be achieved by surgical two-thirds resection of liver by partial hepatectomy before toxicant treatment (45,70,71). Liver regeneration after 70% partial hepatectomy protects the animals from a lethal challenge of CCl4 or 67-fold higher lethal effects of chlordeconeCCCl4 combination due to attenuation of the progression phase of injury

190

Mehendale

(29,71–73). Recently, autoprotection and heteroprotection have also been demonstrated using thioacetamide in type 1 diabetic rats (74), which are sensitive to hepatotoxicants (62,75–77) suggesting that it may be possible to desensitize them by promitogenic intervention, a concept that might prove to be helpful in developing drug intervention strategies in the future. It should be noted that in these examples of autoprotection and heteroprotection, liver injury initiated by the high dose of toxicants is not decreased by the prior administration of the priming agents (46,59,60). Even though the same massive and normally lethal liver injury is reached, and is lethal in unprimed animals, the primed animals overcome this injury as a result of sustainable early onset of tissue repair stimulated by the priming dose. These observations suggest that it may be possible to replace the priming step of this model with appropriate molecular stimulus to activate promitogenic signaling even after normally life-threatening drug overdose. Essentially, in an acute toxicity paradigm, the absence or presence of tissue repair response leads to either progression or regression of injury, respectively. Injury regresses upon the onset of timely and robust tissue repair because the dividing/newly divided cells are resilient to progression of injury (29,30,71–73,78–80). Large doses of drugs and toxicants inhibit tissue repair, but priming with a low-dose or partial hepatectomy leads to sustainable promitogenic signaling and tissue repair such that injury of even a large dose can be overcome. These observations highlight the importance of tissue repair in biomedicine for potential therapeutic intervention.

RELATIONSHIP BETWEEN HIGHER INITIATION OF LIVER INJURY AND TISSUE REPAIR For a long time, the general assumption has been that the higher the dose of a drug or toxicant, the higher the formation of reactive metabolite. Also, it has been generally assumed that higher doses are more toxic because of higher initiation of toxic injury resulting from greater amounts of reactive metabolite. However, time-course studies in models of protection and enhancement of the final toxic outcomes have revealed that the ultimate toxic outcomes are completely dichotomous even though the initial liver injury is identical. Such studies have provided compelling evidence in support of a new paradigm that separates the extent of initial injury and the final toxic outcome (Fig. 3). Equitoxic studies have shown that the extent of initial injury does not determine the ultimate outcome of liver toxicity (35,39,47,62,63,81). Several studies have shown that moderate diet restriction (65% of ad libitum feeding) is known to have a plethora of beneficial effects and one of them is protection against a normally lethal dose of thioacetamide (35,47). Protection is due to stimulation of timely onset and robust liver tissue repair even though initial liver injury of thioacetamide is increased twofold because of fourfold induction of cytochrome P4502E1 (CYP2E1), which bioactivates this toxicant (35,47,81). The reason for this discrepancy between the induction of CYP2E1 and actual initiation of liver injury is because bioactivation of thioacetamide follows saturable zero-order kinetics (50). In equitoxicity studies, in which the dose of thioacetamide needed to initiate liver injury in dietrestricted rats equal to that observed in the ad libitum-fed rats (600 mg/kg, yielding 90% mortality) was found to be 50 mg/kg, upon administration of this dose, all of the diet-restricted rats survived, and 90% of the ad libitum-fed rats did not survive, even though the bioactivationmediated liver injury of thioacetamide was same indicating that tissue repair and survival do not depend on the extent of initial liver injury (47). Even though diet restriction increases thioacetamide-initiated liver injury because of fourfold induction of CYP2E1, which bioactivates thioacetamide, the rats (35,47,81) and mice (82) overcome liver injury due to stimulated tissue repair. Other examples of disconnect between the extent of initial liver injury and the ultimate toxic outcome are available (36,39,62,63,74–77). Type 1 and 2 diabetic rats are very sensitive to the hepatotoxic effects of thioacetamide (75–300 mg/kg) and several other toxicants due to inhibited hepatic cell division and tissue repair (36,39,62,63,74–77,83). Progression of liver injury occurs well after all of thioacetamide (t1/2Z!2 hours for 300 mg/kg dose) has been eliminated (50). Because hepatic microsomal CYP2E1 is induced in type 1 diabetic rats, the

191

Role of Tissue Repair in Liver Injury

Stage I Stage II Infliction of Progression injury of

Injury

Metabolism & Reactive metabolite generation

Tissue repair stimulated

Progression of injury

Regression of injury

Time

FIGURE 3 Initiation and progression/regression of liver injury. Injury is initiated by any one of a number of mechanisms depending on the dose of the drug or toxicant. With low to medium doses, soon after the initiation of tissue injury two separate events follow that determine the outcome of that injury. One event causes the injury to progress and expand. The other is a prompt stimulation of cell division and tissue repair. Although injury may enter the progressive phase, it will regress soon after stimulation of cell division and tissue repair. In the absence of timely cell division such as with very high doses as in drug overdose, in which case cell division is inhibited, injury progresses, and expands leading to liver failure. The mechanisms of injury progression are not well understood, but a number of interventional studies indicate that injury progression may be mediated by hydrolytic enzymes (death proteins) that leak out of necrosing cells and upon activation by extracellular high Ca2C attack their substrates in the plasma membranes of the neighboring healthy or partially affected cells destroying those cells. The hydrolytic enzymes of those cells will perpetuate the injury progression until inhibited by administered chemical inhibitor intervention or by endogenous inhibitors of the hydrolytic enzymes, such as the cysteine protease, calpain. This concept is further illustrated in Figure 7.

initial liver injury of thioacetamide is markedly increased (39). However, when the CYP2E1 in the diabetic rats was equalized to that in the ad libitum-fed rats by pretreatment with diallyl sulfide rendering the initiation of thioacetamide-induced bioactivation-mediated liver injury in diabetic rats equal to that in the nondiabetic rats, liver injury progressed only in the diabetic rats leading to liver failure and death, while all the nondiabetic rats survived illustrating that progression of injury is independent of initial liver injury (63). These observations provide the most compelling arguments that other factors such as the tissue’s response to injury play a critical role in determining the final outcome of initiated injury. Autoprotection and heteroprotection experiments also provide evidence that it is not the initial injury which determines the ultimate outcome of toxicity, but the status of tissue repair (46,59,60). FACTORS AFFECTING TISSUE REPAIR A number of physiological factors influence liver tissue repair. These include species, strain, age, nutrition, caloric restriction, fatty liver, and disease (Table 3). Therefore, in assessing druginduced liver disease, rather wide-ranging diversity of toxic responses may be expected (4,6,7,19). While many factors are considered as underlying causes of such unpredictable diversity in toxic responses, the role of tissue repair as a significant factor in overcoming liver injury has only sparsely been considered. The above factors that influence tissue repair response may help in explaining the wide-ranging interspecies, inter-strain, and interindividual differences in toxic responses to drugs and toxicants. Species and Strain Differences in Tissue Repair Studies with gerbils and Sprague–Dawley rats suggested that the LD50 of CCl4 was 35-fold lower (0.08 ml/kg in gerbils vs. 2.5 ml/kg in rats) in the gerbils (26). Higher CCl4-induced toxicity in gerbils compared to rats could be explained by extremely sluggish tissue repair in the gerbils compared to the rats (4,27). Gerbils are also remarkably resistant to chlordeconeamplified toxicity of CCl4 (9,10). Inhibition of the negligible restorative tissue repair by

192

Mehendale

TABLE 3 Factors Affecting Tissue Repair Induced by Low to Moderate Doses of Drugs and Toxicants Factor Species Strain Age Nutrition Glucose loading Fatty acid supplementation (palmitic acid) Diet restriction Diabetes Type 1 and type 2 rats Type 1 and type 2 mice Combination mixture Leading to inhibition ChlordeconeCCCl4 ChlordeconeCCHCl3 Leading to enhancement IsopropanolCCCl4 PhenobarbitalCCCl4

Effect on tissue repair

Reference

Higher tissue repair in rats as compared to Mongolian gerbils (26) Higher tissue repair in F344 rats as compared to (28) Sprague Dawley rats Higher tissue repair in neonatal (20-day old) and old (29,30,31,57) (14- and 24-month old) rats as compared to young adults (2- to 3-month old) Inhibited tissue repair Enhanced tissue repair

(34) (33)

Enhanced tissue repair

(20,21,81)

Inhibited tissue repair Enhanced tissue repair

(36,39,62,74 77,83) (23,37,38,52)

Inhibition in adult rats Inhibition in adult mice

(14,15,84)

Enhanced tissue repair protects from death in spite of potentiated liver injury Enhanced tissue repair protects against death in spite of markedly potentiated liver injury

(85) (86)

chlordeconeCCCl4 in gerbils is inconsequential because chlordecone-amplified CCl4 toxicity is known to be due to inhibited CCl4-induced increase in restorative tissue repair (4,87). Similar species difference was noticed between rats and mice under disease conditions (37–39). Type 1 and 2 diabetic rats are highly sensitive to thioacetamide-induced liver injury at even normally nonlethal dose of thioacetamide because of compromised tissue repair response (4,19,23,36–39,52,62,63,74–77,83). This sensitivity is not toxicant specific: diabetic rats are highly sensitive to AA, bromobenzene, CCl4, and thioacetamide. However, diabetic mice are completely refractory to liver injury induced by a lethal dose of thioacetamide and APAP due to their ability to mount effective tissue repair response (23,37,38,52). A classic example of strain difference in tissue repair can be seen in the difference between F344 and Sprague–Dawley rats when exposed to o-DCB (27). It was observed that F344 rats experience 10-fold higher liver injury after exposure to 0.2, 0.6, and 1.2 ml o-DCB/kg compared to the Sprague–Dawley rats treated with the same doses. However, the mortality induced by o-DCB in F344 rats is identical to that in the Sprague–Dawley rats, and this was attributed to much higher liver tissue repair in the F344 rats compared to Sprague–Dawley rats (28). Even though the liver injury induced by bioactivation of o-DCB was 10-fold higher in the F-344 rats compared to the Sprague–Dawley rats, that injury never progresses because of prompt and high tissue repair in the F344 rats compared to that in the Sprague–Dawley rats. The significantly higher tissue repair in F344 rats enables them to escape o-DCB-induced liver injury despite being 10-fold higher than the Sprague–Dawley rats (28,68). Age as a Determinant of Tissue Repair Age is an important factor that influences the extent of tissue repair in drug- and toxicantinduced liver injury (29–32,57,78). Newborn animals are capable of mounting faster and efficient tissue repair during early developing age compared to adults. This is known for CCl4, and chlordeconeCCCl4-amplified liver injury in 20-day-old neonatal and 2-month-old young adult Sprague–Dawley rats (29,30,57,87,88). The 20-day-old neonates are resistant to the CCl4 and chlordeconeCCCl4-induced liver injury compared to the young adult rats. Additional studies revealed that the mechanism underlying such resistance in the neonatal rats was the ongoing liver cell proliferation in these rats with growing livers. At 20 days,

Role of Tissue Repair in Liver Injury

193

following birth, the livers of the young rats are still under development and are able to mount an effective tissue repair. It also takes less time for cells to go through the cell division cycle to divide in the neonate and through the early developmental age. This ability is lost in the adults at two months of age when most of the hepatocytes are in quiescence and fail to divide and mount an effective tissue repair upon administration of chlordeconeCCCl4 combination. Furthermore, in the neonatal liver, the proto-oncogenes such as transforming growth factor-a (TGF-a), c-fos, H-ras, and K-ras are expressed at much higher levels and at much earlier timepoints following toxicant exposure (88). These data indicate that higher and timely expression of these proto-oncogenes stimulates timely tissue repair in the neonatal liver which plays a crucial role in the resistance exhibited by the neonates until 30 days of age. Recently, the reason for the resiliency of newly divided/dividing cells has been found to be due to overexpression of calpastatin (CAST) in dividing cells (52,89), an endogenous inhibitor of calpain, known to inhibit calpain-mediated progression of liver injury (54,90). Surprisingly, rats during advanced age also exhibit a prompt and timely tissue repair upon challenge with the combination of chlordeconeCCCl4 (31,90). F344 rats from three age groups, 3, 14, and 24 months, were exposed to the chlordeconeCCCl4 combination. The 14- and 24-month-old rats exhibited higher survival and liver tissue repair as compared to the 3-month-old rats. No difference in the bioactivation of CCl4 was observed between the 14and 24-month-old versus the 3-month-old rats. These data suggest that the tissue repair is not only intact in the old animals but also surprisingly enhanced after challenge with chlordeconeCCCl4 combination. The mechanisms and cellular signaling behind this enhanced tissue repair are yet to be investigated. Interestingly, such age difference was not evident when old rats were challenged with lethal dose of CCl4 alone. Taken together, these data indicate that ability to mount tissue repair following toxicant exposure varies among different age groups and may have a significant impact on the drug development and risk assessment. Effect of Nutrition on Tissue Repair Macronutrients such as carbohydrates, proteins, and lipids affect tissue repair responses. Modulation of the nutritional factors can directly affect the final outcome of toxicity by modifying the tissue repair response (19,33,34,58). Studies with “glucose loading” in rats indicated that 15% glucose supplementation in drinking water for eight days in addition to their normal food consumption inhibits liver tissue repair after exposure to thioacetamide, CHCl3, and CCl4. Glucose loading had no effect on the CYP450-mediated metabolism of thioacetamide but substantially decreased the restorative tissue repair response (34). Glucose loading did not affect insulin levels, nor was the normal blood glucose level affected, but hepatic content of glycogen was increased. Supplementation of diet with equicaloric levels (8% in the diet) of palmitic acid, a preferred source of energy for the periportal hepatocytes, along with its mitochondrial carrier, L-carnitine protected the rats from a lethal dose of thioacetamide (33). Therefore, higher caloric intake associated with glucose loading is not a factor in these studies. These studies suggest that excessive glucose in the body can adversely affect tissue repair, an observation further supported by findings that diabetic rats exhibit inhibited tissue repair following exposure to thioacetamide, CCl4, and other toxicants. It is possible that the inhibitory effect of higher glucose on restorative tissue repair might be mediated by advanced glycosylated end products, but this has not been systematically investigated. Caloric Restriction Diet restriction, known for its ability to slow aging, decrease cancer incidence, and to decrease other age-associated immunological disorders, is also known to protect from toxicity of various drugs and toxicants such as isoproterenol that induces cardiotoxicity, ozone-induced lung toxicity, and thioacetamide-induced liver injury (20,21,35,47,81). Male Sprague–Dawley rats and Swiss Webster mice subjected to 35% diet restriction for 21 days exhibit higher survival (70% in diet restriction vs. 10% in ad libitum) after administration of a normally lethal dose (600 mg/kg) of thioacetamide (35,47,82). Diet-restricted rats survive in spite of the higher liver

194

Mehendale

injury due to induction of thioacetamide-bioactivating enzyme CYP2E1 (81). The mechanism of increased survival in the diet-restricted rats is the earlier onset and robust liver tissue repair (35,47). After thioacetamide administration, promitogenic cellular signaling was promptly enhanced in the diet-restricted rats and mice (20,21,82). Cellular signaling pathways including interleukin-6-mediated JAK-STAT pathway, TGF-a and HGF-mediated MAPK pathway, and peroxisome proliferator receptor a (PPARa)-mediated signaling pathway were all upregulated in the diet-restricted rats resulting in a timely tissue repair response (20,21). It is known that most of this protection is PPARa mediated because protection was lost in diet-restricted PPARaK/K mice (82). Liver Tissue Repair in Diabetes A large prospective cohort study indicates that diabetic patients have higher risk of suffering from hepatic failure compared to nondiabetic population (91). Increased incidences of hepatotoxicity have also been observed in diabetic patients receiving drug therapies, including methotrexate, acarbose, metformin, and troglitazone. Neither the mechanisms nor the predisposing factors underlying liver injury in diabetics are clearly understood. Animal studies designed to examine the mechanisms of diabetes-modulated hepatotoxicity have traditionally focused only on bioactivation/detoxification of drugs and toxicants. It is becoming clear that bioactivation-based initiation of liver injury is only one of the mechanisms by which lifethreatening hepatotoxicity occurs. Once injury is initiated, additional events determine the final outcome of hepatotoxic injury. Foremost among them are two leading mechanisms: first, biochemical mechanisms that lead to progression or regression of injury; second, whether or not adequate liver tissue repair is stimulated on time to mitigate injury and to restore liver function. These observations and drug sensitivity to troglitazone and other members of this class of antidiabetic drugs have increased interest in understanding the basis for sensitivity of diabetes to drug-induced liver toxicity. Considerable animal data suggest that diabetes is a likely predisposing factor for hepatotoxic sensitivity. Streptozotocin-induced type 1 (insulin dependent) diabetic rats exhibit inhibited tissue repair following treatment with a nonlethal dose of thioacetamide (300 mg/kg), leading to 100% mortality (39,62,63). The diabetic rats experience increased liver injury, partly due to induction of thioacetamide-bioactivating liver microsomal enzyme CYP2E1 (63). However, inhibition of CYP2E1 in diabetic rats by a relatively specific inhibitor, diallyl sulfide, failed to protect diabetic rats from thioacetamide-induced liver failure and mortality. Although diallyl sulfide treatment decreased initial bioactivation-based injury (stage I of the two-stage model of toxicity) to the same level as seen in the nondiabetic rats, only diabetic rats failed to stimulate an effective tissue repair response, leading to progression of initial liver injury to culminate in liver failure, and death (63). Other studies with type 1 diabetic rat model have revealed that the mechanism behind inhibition of tissue repair following thioacetamide challenge is downregulation of MAPK pathway and nuclear factor kappa B (NF-kB)-mediated downstream signaling (Fig. 4) (75–77). To investigate the modulation of tissue repair in type 2 diabetes, which afflicts 90% of all diabetic patients, a high-fat diet plus streptozotocin-induced model of type 2 diabetes (noninsulin-dependent diabetes) was developed (36). Studies with these diabetic rats revealed inhibition of tissue repair following CCl4- and thioacetamide-induced hepatotoxicity (Fig. 5) (36,83). The mechanism behind inhibition of tissue repair in diabetes is downregulation of MAPK and NF-kB signaling pathways. In the type 1 rat model, tissue repair is inhibited in the absence of insulin, and in type 2 model even in the presence of insulin, tissue repair is inhibited suggesting that downregulation of cell division and tissue repair are unlikely to be insulin related (36,39). Taken together, these data provide substantial evidence that diabetic rats are sensitive to toxicant-induced liver injury due to inhibition of restorative tissue repair response. In stark contrast to rats, type 1 diabetic mice are not only not sensitive, but are also resilient to normally lethal dose of thioacetamide- and APAP-induced hepatotoxicity, due to increased tissue repair (23,37,38). Stimulation of tissue repair and protection of mice against APAP-induced hepatotoxicity are associated with increased expression of the nuclear receptor, PPARa in the diabetic liver. Protection against normally lethal dose of APAP, induced PPARa

195

Role of Tissue Repair in Liver Injury

EGFR

IL-6

3

1

STAT-3

MAPKs

Cdk4/6

P16/p21

Cyclin D1 2 pRb S

G1

Tissue repair

FIGURE 4 Molecular events regulating G1- to S-phase of the cell cycle. Previous work has shown that downregulated nuclear factor kappa B-regulated CD1 signaling explains impaired tissue repair in thioacetamide-treated diabetic rats. (1) In nondiabetic rats, EGFR MAPK signaling is stimulated upon administration of thioacetamide. However, in diabetic condition, this signaling is downregulated. The inhibitory effect of thioacetamide on EGFR MAPK signaling is increased in diabetic condition upon exposure to thioacetamide. (2) In nondiabetic rats, EGFR MAPKs signaling accelerates the expression of CD1, which complexes with its counterparts, cdk4 or cdk6 to inactivate tumor suppressor gene, Rb by phosphorylation. Phosphorylation of Rb allows the entry of hepatocytes from G1- to S-phase of the cell cycle. In contrast to this, administration of thioacetamide to diabetic rats results in impaired cdk expression and inhibited phosphorylation of Rb. (3) In the nondiabetic rats challenged with thioacetamide, cdk activity is negatively regulated by cdk inhibitors, p16 and p21, which are further controlled by IL-6-STAT-3 signaling pathway. Increased expression of cdk inhibitors via higher STAT-3 signaling results in decreased cdk activation and inhibited phosphorylation of Rb, explaining impaired G1- to S-phase progression of the cell division cycle in the thioacetamide-treated diabetic rats. Abbreviations: CD1, cyclin D1; Rb, retinoblastoma. Source: Adapted from Refs. 75,77.

expression, and stimulated liver tissue repair are all lost in diabetic PPARaK/K mice (23). The increased tissue repair in diabetic mice is partly explained by timely signaling via PPARa to stimulate cell cycle genes such as cyclin D1 (23). Gene expression studies in the liver of diabetic mice with and without APAP treatment revealed a 10-fold increase in expression of cyclin D1 in diabetic mice after drug treatment. Recently, type 2 diabetic mice were shown to be protected against a lethal dose of APAP (52). Toxicokinetics, bioactivation, metabolism of APAP to glucuronide, and hepatic glutathione were not altered in the diabetic liver suggesting that

ND ND + STZ HFD HFD + STZ (Diabetic) ∗ ∗

20 ∗∗ ∗ ∗ 10

∗

∗

∗

∗

∗ ∗

∗ ∗ ∗

! ∗

! ∗

3H-T

(dpm/μg DNA × 100)

30

∗ 0

0

12

24

36

48

72

Hours after CCI4 administration

96

FIGURE 5 3[H]-Thymidine incorporation into hepatonuclear DNA over a time-course after CCl4 administration to male Sprague Dawley rats. On day 24, DB and NDB rats received CCl4 (2 ml/kg, i.p., in 1:1 corn oil). Blood samples were collected under diethyl ether anesthesia at various timepoints (0 96 hours) after CCl4 administration (nZ 4/time-point/group except for diabetic group nZ12 at 36- and 48-hour time-point to allow survival of enough rats for time-course studies). *Significantly different from respective 0-hour time-point. !Significantly different from the NDB normal diet-fed control-injected citrate buffer (ND) at the same time-point. p%0.05. Abbreviations: DB, diabetic; NDB, nondiabetic. Source: Adapted from Ref. 36.

196

Mehendale

DB

ND (non-DB)

c

(A)

(B) N

c

c

(C)

(D)

N

N c c

(E)

(F)

N

N

c

c (G)

(H)

FIGURE 6 Effect of type 2 diabetes on APAP-induced hepatic necrosis in male Swiss Webster mice. Typical photomicrographs of liver sections from ND mice (right) and DB mice (HFDCSTZ) (left) at 0, 6, 12, and 24 hours after APAP administration. (A), (C), (E), and (G): ND mice 0, 6, 12, and 24-hours of post-treatment with APAP. (B), (D), (F), and (H): DB mice 0, 6, 12, and 24 hours of post-treatment with APAP. c, central vein; N, areas of necrosis. H&E original magnification, 200!. Abbreviations: APAP, acetaminophen; DB, diabetic; ND, non-DB. Source: Adapted from Ref. 52.

protection could not be explained by decreased bioactivation or by increased detoxification mechanisms. Protection was due to early and robust stimulation of liver tissue repair following the initiation of liver injury by bioactivation of APAP (Fig. 6) (52). Type 2 diabetes increased cells in S-phase by eightfold in normally quiescent livers of these mice. Immunohistochemical staining of liver sections from APAP-treated diabetic mice revealed overexpression of CAST, the endogenous inhibitor of the death protein, calpain (52,54,90). Antimitotic intervention of diabetes-associated cell division with colchicine before APAP administration resulted in 70% mortality in APAP-treated colchicine-intervened diabetic mice. These studies suggest that advancement of cells in the cell division cycle and higher tissue repair protect diabetic mice by preventing the progression of APAP-initiated liver injury that normally progresses leading to hepatic failure and death (52).

Role of Tissue Repair in Liver Injury

197

PROGRESSION OF LIVER INJURY Overdose of the commonly used analgesic drug APAP accounts for approximately 50% of the cases of acute liver failure and is the leading cause of liver transplantations in the United States. Because shortage of donors limits liver transplantation, stimulation of the native liver to repair and restore functional tissue via regeneration may offer an alternative method for treatment of patients with failing liver. These patients have little chance of survival. There is a need to understand and develop alternative strategies to turn these hopeless prognoses into hopeful prognoses. Efficient and timely liver regeneration can prevent drug-induced acute liver failure in experimental animal models as well as in humans. Monitoring the markers of liver regeneration in drug overdose victims has been proposed as a prognostic predictor (92–94). Under a variety of circumstances in which liver regeneration is stimulated experimentally either by chemical injury (e.g., low dose of thioacetamide, APAP) or by two-thirds partial hepatectomy (46,72,73), the liver is known to exhibit resistance toward the higher doses of the hepatotoxicants. Liver cell division during postnatal development in newborns is known to impart resistance against toxicity (29,30,95). Hepatocytes isolated from the regenerating liver exhibit resistance against hepatotoxicants in vitro (32,79). Although evidence suggests that the dividing cells in the regenerating liver are resistant to hepatotoxicity, the exact mechanism has remained elusive. Early literature indicates that resistance was due to lower bioactivation of drugs and toxicants but this mechanism is unlikely because bioactivation of model toxicants is not significantly decreased. Even after the offending hepatotoxic drug is eliminated from the body, injury continues until the liver heals itself via liver regeneration and tissue repair (3,80). A major hurdle in understanding the mechanism of resistance of newly divided cells is in part the incomplete understanding of how injury initiated by a toxicant expands even after the offending drug has been eliminated. On exposure to low doses of hepatotoxicants, the liver mounts proportional restorative regeneration to overcome toxicity (3,80). However, high doses of hepatotoxic drugs inhibit liver regeneration, which results in rapid and accelerated progression of injury and liver failure (3,80). Liver regeneration primarily curtails stage II injury (3). Several studies have shown that acute toxic injury develops in two distinct stages: stage I, initiation of injury and stage II, progression of injury as illustrated in Figure 1 (5,16,48,55,56,58, 59,81,84,96–98). Extensive evidence suggests that tissue repair is a dynamic opposing force that curtails stage II or progression of injury from developing into an organ failure (97). While much is known about the endogenous bioactivation mechanisms which lead to initiation of cellular and tissue injury, mechanisms of progression of injury have remained obscure. The increase in injury over time is thought to be due to slower production of reactive metabolites from the residual parent compound over a time-course. However, toxicokinetic studies do not support this notion. The toxicokinetics of the model hepatotoxicants such as thioacetamide, CCl4, and APAP indicate that most of the toxicants are excreted from the body within the first 24 hours by conjugation reaction mediated by phase II drug-metabolizing enzymes and other elimination processes (31,37,51). However, the time-course of injury suggests that the liver injury increases and progresses well beyond the 24 hours (16,25,37) suggesting that the progression of injury initiated by the toxicants, takes place in toxicant-independent fashion. Three mechanisms have been proposed to explain the progression of injury: (1) contribution of inflammatory cells to injury (99), (2) production of free radicals (98), and (3) leakage of degradative enzymes from the dying and injured cells (100). Activated resident Kupffer cells and the neutrophils recruited to the site of parenchymal liver injury have been considered as the primary culprits in destroying surrounding healthy cells as the result of nonspecific action (55). However, recent evidence suggests that the contribution of the inflammatory cells does not or is insufficient to mediate progression of injury (96,101). The second leading theory regarding progression of injury is production of free radicals and oxidative stress, and subsequent lipid peroxidation that propagates injury (102). Though antioxidants prevent/delay the tissue damage partially (99,103), progression of injury still occurs. Blocking lipid peroxidation fails to prevent progression of injury and subsequent lethality. These findings suggest that progression of injury is driven by other mechanisms and

198

Toxicant

Mehendale

1.3 mM Ca

2+

Calpain

2+

μM Ca

Phospholipase Dying cell

Nucleases (e.g., DNAse-1) Acid phosphatase

N AMP, CMP

Cathepsin-D Neighboring cell ECM

FIGURE 7 Role of death proteins in progression of injury. Schematic of the lytic action of death proteins released from dying hepatocytes in the extracellular space on the neighboring hepatocytes. Upon infliction of cellular injury by bioactivation-based events, the injured cells eventually are lysed releasing their contents into the extracellular space. Thus, the potent degradative enzymes ( death proteins ) enter the extracellular environment rich in Ca2C, are activated, and hydrolyze their substrates contained in the plasma membrane of the neighboring healthy cells or cells partly affected by toxicants, causing progression of injury until the augmented tissue repair is mounted or is severely inhibited. Source: Adapted from Ref. 54.

free radicals and oxidative stress may be a part of result of such mechanisms, it cannot be the primary driver of injury progression. The third relatively less-studied theory that may explain the progression of injury is the leakage of degradative enzymes or “death proteins” from the dying and injured cells, which may destroy neighboring cells causing progression of injury (100). However, this mechanism has not been explored in a systematic manner. Recent studies have considered whether such cellular degradative enzymes released upon initiation of injury from necrosed hepatocytes mediate progression of injury (Fig. 7). Whereas mechanisms of initiation of injury or stage I injury induced by hepatotoxic drugs are widely known to be formation of reactive metabolites, generation of free radicals, blockage of biochemical pathways, and so forth, the mechanism of progression of injury or stage II injury remains unknown. A recent study, has demonstrated that expansion of toxicant-initiated injury is mediated by the degradative enzymes released from dying hepatocytes, among which calpain appears to play a major role (Fig. 6) (19). A series of experiments affirmed that calpain spills out from the necrosed hepatocytes and is further activated in the extracellular high Ca2C environment. Activated calpain degrades the plasma membrane and cytoskeletal proteins from the surrounding hepatocytes, spreading the injury in a self-perpetuating manner (19). Numerous reports have linked activation of calpain in the pathogenesis of various disorders such as neurodegenerative disorders, including Alzheimer’s disease, epilepsy, cerebral ischemia/excitotoxicity, demyelination, cataracts, ischemic liver and ischemic renal injury, toxic renal injury, and muscular dystrophy. Studies also have shown that the overexpression of CAST, the endogenous inhibitor of calpain, is protective against these disorders where calpain activation is pathogenic (54,90). Such studies have provided substantial evidence that cysteine protease, calpain, plays a predominant role in progression of injury (54). Calpain is known to degrade several membrane and cytoskeletal proteins, including fodrin/spectrin, talin, filamin, and other macromolecules pivotal for cellular integrity (104–107), and thereby may cause progression of injury. In our studies, calpain inhibition using a specific calpain inhibitor, CBZ-VAL-PHEmethyl ester (CBZ) administered one hour after CCl4 in rats, led to 50% reduction in CCl4induced mortality. In order to establish whether this protection is due to inhibition of progressive phase of liver injury (stage II) a nonlethal dose of CCl4 (2 ml/kg, i.p.) was used. CBZ was administered to one group of rats one hour after the injection of CCl4. The other group received only the vehicle (DMSO, 0.2 ml/kg, i.p.), used for CBZ. Time-course measurements of liver injury assessed by plasma alanine aminotransferase elevation indicated that progression of liver injury initiated by CCl4 was substantially decreased by CBZ intervention. Histopathology of liver also confirmed protection against the progression of injury (Fig. 8) (54). Calpain inhibition also protected against APAP-induced progression of injury and subsequent mortality in mice (54). These findings indicate that calpain’s role in progression of injury is neither species specific nor toxicant specific. In both cases, calpain inhibitor had no effect on the major bioactivating enzyme CYP2E1. In vitro incubation studies

199

Role of Tissue Repair in Liver Injury

CCl4 + DMSO

CCl4 + CBZ

(A)

(B)

(C)

(D)

(E)

(F)

(G)

(H)

FIGURE 8 Histopathological analysis of H&E-stained liver sections over a time-course after CCl4 (2 ml/kg, i.p.) with and without calpain inhibitor, CBZ (60 mg/kg, i.p.) administered one hour after CCl4 to male Sprague Dawley rats. (A) CCl4C DMSO and (B) CCl4CCBZ at three hours; (C) CCl4CDMSO and (D) CCl4CCBZ at 24 hours; (E) CCl4CDMSO and (F) CCl4C CBZ at 48 hours; (G) CCl4CDMSO and (H) CCl4CCBZ at 72 hours. Source: Adapted from Ref. 54.

with microsomes also did not reveal any change in the catalytic activity of CYP2E1 enzyme even with 500-fold concentration range of CBZ. Covalent binding of 14CCl4-derived radiolabel in rat liver and 14C- APAP-derived radiolabel in mouse liver was unaltered regardless of the administration of CBZ (54). These observations strongly suggest that the calpain inhibitor CBZ does not alter the bioactivation of these toxicants. Observations such as increase in calpain leakage with increase in liver injury, decrease in calpain-mediated degradation of fodrin, a substrate of calpain, in CBZ-treated rats, and ability of calpain to induce cell death in isolated primary hepatocytes in vitro further support involvement of calpain in progression of injury. Calpain inhibition resulted in prevention of progression of injury, paving the way for tissue repair to take over and restore the hepatic architecture in place of the dead tissue mass. However, the mechanism for how the dividing cells escape calpain-induced cell death has been elusive. Recent study revealing overexpression of CAST, the endogenous inhibitor of calpain, may explain the mechanism of resiliency of new cells against progression of injury (89).

200

Mehendale

REGRESSION OF INJURY Substantial research has shown that timely liver regeneration can prevent progression of injury leading to positive prognosis (4,13,16,17,19,20,21,23,28–31,34,35,37,38,43,46,47,52,57–59,69,74, 78,89,92,93,108). However, the mechanism by which restorative regeneration prevents progression of injury has remained elusive for a long time. The hypothesis that the dividing and newly divided hepatocytes overexpress CAST and thereby prevent the calpain-mediated onslaught and progression of injury by inhibiting calpain was tested. CAST (109,110) is known to inhibit both Ca2C activated as well as a proenzyme form of calpain (111–114). CAST is mainly localized in the cytoplasm and plasma membrane (115), and it is also present in the nucleus during cell division (116). In recent studies, expression of CAST, an endogenous inhibitor of calpain, was examined (89) in three liver cell division models known to be resistant to hepatotoxicity, viz. liver regeneration stimulated by CCl4-induced liver injury, 70% partial hepatectomy, and postnatal liver development (29,30,45,73). The hypothesis that increased CAST in the dividing hepatocytes affords resistance against progression of injury was investigated (89). In all three models, CAST was upregulated in the dividing/newly divided hepatocytes and declined to normal levels with the cessation of cell division. To test whether CAST overexpression confers resistance against hepatotoxicity, CAST was overexpressed in the livers of normal adult Swiss Webster mice using adenoviral transfection before challenging them with APAP overdose. These mice exhibited markedly decreased progression of liver injury allowing a 57% survival (Fig. 9). Serum enzyme measurement and histopathology of the liver indicated that injury progression was markedly slower in the CAST-expressing mice treated with APAP (Fig. 10). In contrast, liver injury progressed in mice not expressing CAST and mice died of hepatic failure. Whereas APAP-bioactivating enzymes and covalent binding of the APAP-derived reactive metabolites remained unaffected, degradation of calpain-specific target substrates such as a-fodrin was significantly reduced in these mice. Whether CAST overexpression could be used as a therapeutic strategy to prevent progression of liver injury where liver regeneration is severely hampered remains uninvestigated. In other experimental animal models such as type 2 diabetic mice, overexpression of CAST has been observed in livers of mice known to be resistant to normally lethal doses of APAP and thioacetamide (52). Interestingly, because more than the normal number of cells advance to S-phase after induction of diabetes and overexpress CAST, calpain-mediated progression of liver injury initiated by APAP never takes hold. Consequently, liver structure and function are restored promptly allowing the mice to overcome injury initiated by the normally lethal dose of APAP (52). Type 1 diabetic mice are also resilient to normally lethal dose of APAP and thioacetamide because of augmented tissue repair due to the hepatic cells advancing to S-phase (23,37). Presumably, CAST expression in the liver cells that advance to S-phase in type 1 diabetic mice before the administration of the toxicants

AD/CAST+APAP AD/LacZ+APAP

Percent survival

100

PBS+APAP

80 60 40 20 0 0

10 20 30 40 50 Hours after APAP administration

60

FIGURE 9 Kaplan Meier survival plot after APAP administration. Male Swiss Webster mice (25 29 g) were divided into three groups. Group I received PBS (100 mL), group II received Ad/calpastatin (CAST, 0.5!1011 v.p. in 100 mL PBS), while group III received Ad/LacZ (0.5!1011 v.p. in 100 mL PBS) by tail vein injection. All three groups were administered a single dose of APAP (600 mg/kg, i.p., in 0.45% NaCl, pH 8) five hours before the end of the fourth day after the respective pretreatments. All mice were observed six times on the first day and twice daily thereafter for 14 days. All deaths occurred between 10 and 12 hours after APAP administration. Thereafter, all remaining mice survived. Abbreviation: APAP, acetaminophen. Source: Adapted from Ref. 89.

201

Role of Tissue Repair in Liver Injury 9000

Ad/LacZ+APAP

(a)

Ad/CAST+APAP

(b)

Plasma ALT (UL)

cv

cv

6000

∗

(c)

3000

(d)

cv ∗ 0

0

4

cv 5

Hours after APAP administration

FIGURE 10 Liver injury assessment after APAP administration with pretreatment with either Ad/CAST or Ad/LacZ. (Left) Plasma ALT measured as an index of liver injury after the administration of APAP (600 mg/kg, i.p.) with pretreatment with either Ad/CAST (adenovirally calpastatin expressing) or Ad/LacZ (control, not expressing calpastatin). Zero-hour control valuesZ42.5C12.8. *Values significantly lower than the group receiving Ad/LacZCAPAP alone at the same time-point (nZ5). p!05. (Right) Representative photomicrographs of the paraffin-embedded liver sections stained with H&E after the administration of APAP (600 mg/kg, i.p.) with pretreatment with either Ad/CAST [(b) four hours after APAP and (d) eight hours after APAP], or Ad/LacZ [(a) four hours after APAP and (c) eight hours after APAP]. MagnificationZ200!. Arrows indicate necrotic area, CV represents central vein. Abbreviations: APAP, acetaminophen; ALT, alanine aminotransferase. Source: Adapted from Ref. 89.

helps to keep the initiated liver injury from progressing. Once these higher number of preexisting S-phase cells divide, the expansion of injury by injury progression mechanisms is prevented, while other cells divide until dead cell mass and function are restored to achieve full recovery. SIGNIFICANCE OF TISSUE REPAIR Extensive evidence gathered during the last quarter century supports the role of tissue repair as an important determining factor affecting the final outcome of toxic injury. Tissue repair is a dose-dependent dynamic process, affected by several factors including species, strain, age, nutrition, caloric restriction, and disease. Various interventional strategies, detailed signal transduction studies, and genomic and proteomic studies have revealed that tissue repair plays a decisive role in determining survival or death of an animal exposed to a toxicant. Recent findings suggesting involvement of calpain and other death proteins in progression of injury aid in our understanding of a general paradigm of acute toxicity (Fig. 7). These data argue for consideration of tissue repair as a factor in risk assessment and drug development strategies. Consideration of endogenous restorative response to toxicity induced by a test chemical would be helpful in resolving imprecise risk assessment issues and may offer an explanation for interindividual variation in adverse drug/toxicant effects. Similarly, assessment of tissue repair stimulated by a test compound may provide additional mechanistic information extremely valuable for drug development. Taken together, these data indicate that assessment of tissue repair initiated by toxicants upon exposure can have enormous impact on public health. CONCLUSIONS Tissue repair is a dynamic restorative cell proliferation and tissue regeneration response stimulated in order to overcome acute toxicity and recover organ/tissue structure and function. Extensive evidence in rodent models using structurally and mechanistically diverse hepatotoxicants such as APAP, carbon tetrachloride, chloroform, DCB, thioacetamide, TCE, and AA has

202

Mehendale

demonstrated that tissue repair plays a critical role in determining the final outcome of toxicity, i.e., recovery from injury and survival or progression of injury leading to liver failure and death. Tissue repair is a complex process governed by intricate cellular signaling involving a number of chemokines, cytokines, growth factors, nuclear receptors leading to promitogenic gene expression and cell division. Tissue repair also encompasses regeneration of hepatic extracellular matrix and angiogenesis, the processes necessary to completely restore the structure and function of the liver tissue lost to toxicant-induced initiation of injury followed by progression of that injury mediated by the lytic enzymes. The lytic enzymes such as calpain, that leak out of the cells necrosing from the initiated injury, destroy the perinecrotic partially affected as well as the unaffected cells by degrading their plasma membrane upon activation by Ca2C in the Ca2C-rich extracellular milieu. New insights have emerged over the last quarter century indicating that tissue repair follows a dose–response. Tissue repair increases with dose until a threshold dose is reached, beyond which it is delayed and impaired due to inhibition of cellular signaling resulting in runaway secondary events causing tissue destruction, organ failure, and death. Prompt and adequately stimulated tissue repair response to toxic injury is critical for recovery from toxic injury. Tissue repair is modulated by a variety of factors, including species, strain, age, nutrition, and disease, causing marked changes in susceptibility and toxic outcome. This chapter has focused on the properties of tissue repair, different factors affecting tissue repair, and the mechanisms that govern tissue repair and progression of injury. It also highlights the significance of tissue repair as a target for drug development strategies and an important consideration in the assessment of risk from exposure to toxicants. REFERENCES 1. Rozzman K, Klaassen C. Absorption, distribution and excretion of toxicants. In: Klaassen C, ed. Caserett and Doull’s Toxicology: Basic Science of Poisons. New York: McGraw Hill, 2001:107–32. 2. diBethizy JD, Hayes J. Metabolism: a determinant of toxicity. In: Hayes AW, ed. Principles and Methods of Toxicology. Philadelphia, PA: Taylor and Francis, 2001:77–136. 3. Mehendale HM. Role of hepatocellular regeneration and hepatolobular healing in the final outcome of liver injury. A two-stage model of toxicity. Biochem Pharmacol 1991; 42(6):1155–62. 4. Mehendale HM. Amplified interactive toxicity of chemicals at nontoxic levels: mechanistic considerations and implications to public health. Environ Health Perspect 1994; 102(Suppl. 9):139–49. 5. Mehendale HM. Injury and repair as opposing forces in risk assessment. Toxicol Lett 1995; 82–83:891–9. 6. Soni MG, Mehendale HM. Role of tissue repair in toxicologic interactions among hepatotoxic organics. Environ Health Perspect 1998; 106(Suppl. 6):1307–17. 7. Mehendale HM, Roth RA, Gandolfi AJ, et al. Novel mechanisms in chemically induced hepatotoxicity. FASEB J 1994; 8(15):1285–95. 8. Plaa G. Detection and evaluation of chemically induced liver injury. In: Hayes A, ed. Principles and Methods of Toxicology. Philadelphia, PA: Taylor and Francis Publishers, 2001:1145–88. 9. Plaa GL. Chlorinated methanes and liver injury: highlights of the past 50 years. Annu Rev Pharmacol Toxicol 2000; 40:42–65. 10. Michalopoulos GK, DeFrances MC. Liver regeneration. Science 1997; 276(5309):60–6. 11. Fausto N, Laird AD, Webber EM. Liver regeneration. 2. Role of growth factors and cytokines in hepatic regeneration. FASEB J 1995; 9(15):1527–36. 12. Taub R. Liver regeneration 4: transcriptional control of liver regeneration. FASEB J 1996; 10(4):413–27. 13. Dalhoff K, Laursen H, Bangert K, et al. Autoprotection in acetaminophen intoxication in rats: the role of liver regeneration. Pharmacol Toxicol 2001; 88(3):135–41. 14. Lockard VG, Mehendale HM, O’Neal RM. Chlordecone-induced potentiation of carbon tetrachloride hepatotoxicity: a morphometric and biochemical study. Exp Mol Pathol 1983; 39(2):246–55. 15. Lockard VG, Mehendale HM, O’Neal RM. Chlordecone-induced potentiation of carbon tetrachloride hepatotoxicity: a light and electron microscopic study. Exp Mol Pathol 1983; 39(2):230–45. 16. Mangipudy RS, Chanda S, Mehendale HM. Tissue repair response as a function of dose in thioacetamide hepatotoxicity. Environ Health Perspect 1995; 103(3):260–7. 17. Shayiq RM, Roberts DW, Rothstein K, et al. Repeat exposure to incremental doses of acetaminophen provides protection against acetaminophen-induced lethality in mice: an explanation for high acetaminophen dosage in humans without hepatic injury. Hepatology 1999; 29(2):451–63.

Role of Tissue Repair in Liver Injury

203

18. Soni MG, Ramaiah SK, Mumtaz MM, et al. Toxicant-inflicted injury and stimulated tissue repair are opposing toxicodynamic forces in predictive toxicology. Regul Toxicol Pharmacol 1999; 29(2 Pt 1):165–74. 19. Mehendale HM. Tissue repair: an important determinant of final outcome of toxicant-induced injury. Toxicol Pathol 2005; 33(1):41–51. 20. Apte UM, Limaye PB, Desaiah D, et al. Mechanisms of increased liver tissue repair and survival in diet-restricted rats treated with equitoxic doses of thioacetamide. Toxicol Sci 2003; 72(2):272–82. 21. Apte UM, Limaye PB, Ramaiah SK, et al. Upregulated promitogenic signaling via cytokines and growth factors: potential mechanism of robust liver tissue repair in calorie-restricted rats upon toxic challenge. Toxicol Sci 2002; 69(2):448–59. 22. Gardner CR, Laskin JD, Dambach DM, et al. Exaggerated hepatotoxicity of acetaminophen in mice lacking tumor necrosis factor receptor-1. Potential role of inflammatory mediators. Toxicol Appl Pharmacol 2003; 192(2):119–30. 23. Shankar K, Vaidya VS, Corton JC, et al. Activation of PPAR-alpha in streptozotocin-induced diabetes is essential for resistance against acetaminophen toxicity. FASEB J 2003; 17(12):1748–50. 24. Tomiya T, Ogata I, Fujiwara K. Transforming growth factor alpha levels in liver and blood correlate better than hepatocyte growth factor with hepatocyte proliferation during liver regeneration. Am J Pathol 1998; 153(3):955–61. 25. Rao PS, Mangipudy RS, Mehendale HM. Tissue injury and repair as parallel and opposing responses to CCl4 hepatotoxicity: a novel dose–response. Toxicology 1997; 118(2–3):181–93. 26. Cai ZW, Mehendale HM. Lethal effects of CCl4 and its metabolism by Mongolian gerbils pretreated with chlordecone, phenobarbital, or mirex. Toxicol Appl Pharmacol 1990; 104(3):511–20. 27. Cai Z, Mehendale HM. Hepatotoxicity and lethality of halomethanes in Mongolian gerbils pretreated with chlordecone, phenobarbital or mirex. Arch Toxicol 1991; 65(3):204–12. 28. Kulkarni SG, Duong H, Gomila R, et al. Strain differences in tissue repair response to 1,2-dichlorobenzene. Arch Toxicol 1996; 70(11):714–23. 29. Cai Z, Mehendale HM. Resiliency to amplification of carbon tetrachloride hepatotoxicity by chlordecone during postnatal development in rats. Pediatr Res 1993; 33(3):225–32. 30. Dalu A, Mehendale HM. Efficient tissue repair underlies the resiliency of postnatally developing rats to chlordeconeCCCl4 hepatotoxicity. Toxicology 1996; 111(1–3):29–42. 31. Murali B, Korrapati MC, Warbritton A, et al. Tolerance of aged Fischer 344 rats against chlordeconeamplified carbon tetrachloride toxicity. Mech Ageing Dev 2004; 125(6):421–35. 32. Sanz N, Diez-Fernandez C, Fernandez-Simon L, et al. Necrogenic and regenerative responses of liver of newly weaned rats against a sublethal dose of thioacetamide. Biochim Biophys Acta 1998; 1384(1):66–78. 33. Chanda S, Mehendale M. Role of nutritional fatty acid and L-carnitine in the final outcome of thioacetamide hepatotoxicity. FASEB J 1994; 8(13):1061–8. 34. Chanda S, Mehendale HM. Nutritional impact on the final outcome of liver injury inflicted by model hepatotoxicants: effect of glucose loading. FASEB J 1995; 9(2):240–5. 35. Ramaiah SK, Bucci TJ, Warbritton A, et al. Temporal changes in tissue repair permit survival of dietrestricted rats from an acute lethal dose of thioacetamide. Toxicol Sci 1998; 45(2):233–41. 36. Sawant SP, Dnyanmote AV, Shankar K, et al. Potentiation of carbon tetrachloride hepatotoxicity and lethality in type 2 diabetic rats. J Pharmacol Exp Ther 2004; 308(2):694–704. 37. Shankar K, Vaidya VS, Apte UM, et al. Type 1 diabetic mice are protected from acetaminophen hepatotoxicity. Toxicol Sci 2003; 73(2):220–34. 38. Shankar K, Vaidya VS, Wang T, et al. Streptozotocin-induced diabetic mice are resistant to lethal effects of thioacetamide hepatotoxicity. Toxicol Appl Pharmacol 2003; 188(2):122–34. 39. Wang T, Fontenot RD, Soni MG, et al. Enhanced hepatotoxicity and toxic outcome of thioacetamide in streptozotocin-induced diabetic rats. Toxicol Appl Pharmacol 2000; 166(2):92–100. 40. Sivarao DV, Mehendale HM. 2-Butoxyethanol autoprotection is due to resiliance of newly formed erythrocytes to hemolysis. Arch Toxicol 1995; 69(8):526–32. 41. Barton CC, Bucci TJ, Lomax LG, et al. Stimulated pulmonary cell hyperplasia underlies resistance to alpha-naphthylthiourea. Toxicology 2000; 143(2):167–81. 42. Vaidya VS, Shankar K, Lock EA, et al. Renal injury and repair following S-1,2-dichlorovinyl-Lcysteine administration to mice. Toxicol Appl Pharmacol 2003; 188(2):110–21. 43. Anand SS, Murthy SN, Vaidya VS, et al. Tissue repair plays pivotal role in final outcome of liver injury following chloroform and allyl alcohol binary mixture. Food Chem Toxicol 2003; 41(8):1123–32. 44. Anand SS, Soni MG, Vaidya VS, et al. Extent and timeliness of tissue repair determines the doserelated hepatotoxicity of chloroform. Int J Toxicol 2003; 22(1):25–33. 45. Cai ZW, Mehendale HM. Protection from CCl4 toxicity by prestimulation of hepatocellular regeneration in partially hepatectomized gerbils. Biochem Pharmacol 1991; 42(3):633–44. 46. Chanda S, Mangipudy RS, Warbritton A, et al. Stimulated hepatic tissue repair underlies heteroprotection by thioacetamide against acetaminophen-induced lethality. Hepatology 1995; 21(2):477–86.

204

Mehendale

47. Ramaiah SK, Soni MG, Bucci TJ, et al. Diet restriction enhances compensatory liver tissue repair and survival following administration of lethal dose of thioacetamide. Toxicol Appl Pharmacol 1998; 150(1):12–21. 48. Calabrese EJ, Mehendale HM. A review of the role of tissue repair as an adaptive strategy: why low doses are often non-toxic and why high doses can be fatal. Food Chem Toxicol 1996; 34(3):301–11. 49. Mehendale HM, Klingensmith JS. In vivo metabolism of CCl4 by rats pretreated with chlordecone, mirex, or phenobarbital. Toxicol Appl Pharmacol 1988; 93(2):247–56. 50. Chilakapati J, Shankar K, Korrapati MC, et al. Saturation toxicokinetics of thioacetamide: role in initiation of liver injury. Drug Metab Dispos 2005; 33(12):1877–85. 51. Porter WR, Gudzinowicz MJ, Neal RA. Thioacetamide-induced hepatic necrosis. II. Pharmacokinetics of thioacetamide and thioacetamide-S-oxide in the rat. J Pharmacol Exp Ther 1979; 208(3):386–91. 52. Sawant SP, Dnyanmote AV, Mitra MS, et al. Protective effect of type 2 diabetes on acetaminopheninduced hepatotoxicity in male Swiss-Webster mice. J Pharmacol Exp Ther 2006; 316(2):507–19. 53. Rao VC, Mehendale HM. Effect of antimitotic agent colchicine on carbon tetrachloride toxicity. Arch Toxicol 1993; 67(6):392–400. 54. Limaye PB, Apte UM, Shankar K, et al. Calpain released from dying hepatocytes mediates progression of acute liver injury induced by model hepatotoxicants. Toxicol Appl Pharmacol 2003; 191(3):211–26. 55. Luster MI, Simeonova PP, Gallucci RM, et al. Role of inflammation in chemical-induced hepatotoxicity. Toxicol Lett 2001; 120(1–3):317–21. 56. Soni MG, Mangipudy RS, Mumtaz MM, et al. Tissue repair response as a function of dose during trichloroethylene hepatotoxicity. Toxicol Sci 1998; 42(2):158–65. 57. Dalu A, Warbritton A, Bucci TJ, et al. Age-related susceptibility to chlordecone-potentiated carbon tetrachloride hepatotoxicity and lethality is due to hepatic quiescence. Pediatr Res 1995; 38(2):140–8. 58. Chanda S, Mehendale HM. Hepatic cell division and tissue repair: a key to survival after liver injury. Mol Med Today 1996; 2(2):82–9. 59. Mangipudy RS, Chanda S, Mehendale HM. Hepatocellular regeneration: key to thioacetamide autoprotection. Pharmacol Toxicol 1995; 77(3):182–8. 60. Thakore KN, Mehendale HM. Role of hepatocellular regeneration in CCl4 autoprotection. Toxicol Pathol 1991; 19(1):47–58. 61. Thakore KN, Mehendale HM. Effect of phenobarbital and mirex pretreatments on CCl4 autoprotection. Toxicol Pathol 1994; 22(3):291–9. 62. Wang T, Shankar K, Ronis MJ, et al. Potentiation of thioacetamide liver injury in diabetic rats is due to induced CYP2E1. J Pharmacol Exp Ther 2000; 294(2):473–9. 63. Wang T, Shankar K, Bucci TJ, et al. Diallyl sulfide inhibition of CYP2E1 does not rescue diabetic rats from thioacetamide-induced mortality. Toxicol Appl Pharmacol 2001; 173(1):27–37. 64. Dalu A, Rao PS, Mehendale HM. Colchicine antimitosis abolishes resiliency of postnatally developing rats to chlordecone-amplified carbon tetrachloride hepatotoxicity and lethality. Environ Health Perspect 1998; 106(9):597–606. 65. Mangipudy RS, Rao PS, Mehendale HM. Effect of an antimitotic agent colchicine on thioacetamide hepatotoxicity. Environ Health Perspect 1996; 104(7):744–9. 66. Fitzgerald PH, Brehaut LA. Depression of DNA synthesis and mitotic index by colchicine in cultured human lymphocytes. Exp Cell Res 1970; 59(1):27–31. 67. Tsukamoto I, Kojo S. Effect of colchicine and vincristine on DNA synthesis in regenerating rat liver. Biochim Biophys Acta 1989; 1009(2):191–3. 68. Kulkarni SG, Warbritton A, Bucci TJ, et al. Antimitotic intervention with colchicine alters the outcome of o-DCB-induced hepatotoxicity in Fischer 344 rats. Toxicology 1997; 120(2):79–88. 69. Mehendale HM, Thakore KN, Rao CV. Autoprotection: stimulated tissue repair permits recovery from injury. J Biochem Toxicol 1994; 9(3):131–9. 70. Uryvaeva IV, Faktor VM. Resistance of regenerating liver to hepatotoxins. Biull Eksp Biol Med 1976; 81(3):283–5. 71. Kodavanti PR, Joshi UM, Mehendale HM, et al. Chlordecone (Kepone)-potentiated carbon tetrachloride hepatotoxicity in partially hepatectomized rats—a histomorphometric study. J Appl Toxicol 1989; 9(6):367–75. 72. Bell AN, Young RA, Lockard VG, et al. Protection of chlordecone-potentiated carbon tetrachloride hepatotoxicity and lethality by partial hepatectomy. Arch Toxicol 1988; 61(5):392–405. 73. Kodavanti PR, Joshi UM, Young RA, et al. Protection of hepatotoxic and lethal effects of CCl4 by partial hepatectomy. Toxicol Pathol 1989; 17(3):494–505. 74. Devi SS, Philip BK, Warbritton A, et al. Prior administration of a low dose of thioacetamide protects type 1 diabetic rats from subsequent administration of lethal dose of thioacetamide. Toxicology 2006; 226:107–17. 75. Devi SS, Mehendale HM. The role of NF-kappaB signaling in impaired liver tissue repair in thioacetamide-treated type 1 diabetic rats. Eur J Pharmacol 2005; 523(1–3):127–36.

Role of Tissue Repair in Liver Injury

205

76. Devi SS, Mehendale HM. Microarray analysis of thioacetamide-treated type 1 diabetic rats. Toxicol Appl Pharmacol 2006; 212(1):69–78. 77. Devi SS, Mehendale HM. Disrupted G1 to S phase clearance via cyclin signaling impairs liver tissue repair in thioacetamide-treated type 1 diabetic rats. Toxicol Appl Pharmacol 2005; 207(2):89–102. 78. Abdul-Hussain SK, Mehendale HM. Ongoing hepatocellular regeneration and resiliency toward galactosamine hepatotoxicity. Arch Toxicol 1992; 66(10):729–42. 79. Roberts E, Ahluwalia MB, Lee G, et al. Resistance to hepatotoxins acquired by hepatocytes during liver regeneration. Cancer Res 1983; 43(1):28–34. 80. Ruch RJ, Klaunig JE, Pereira MA. Selective resistance to cytotoxic agents in hepatocytes isolated from partially hepatectomized and neoplastic mouse liver. Cancer Lett 1985; 26(3):295–301. 81. Ramaiah SK, Apte U, Mehendale HM. Cytochrome P4502E1 induction increases thioacetamide liver injury in diet-restricted rats. Drug Metab Dispos 2001; 29(8):1088–95. 82. Corton JC, Apte U, Anderson SP, et al. Mimetics of caloric restriction include agonists of lipidactivated nuclear receptors. J Biol Chem 2004; 279(44):46204–12. 83. Sawant SP, Dnyanmote AV, Warbritton A, et al. Type 2 diabetic rats are sensitive to thioacetamide hepatotoxicity. Toxicol Appl Pharmacol 2006; 211(3):221–32. 84. Purushotham KR, Lockard VG, Mehendale HM. Amplification of chloroform hepatotoxicity and lethality by dietary chlordecone (kepone) in mice. Toxicol Pathol 1988; 16(1):27–34. 85. Rao PS, Dalu A, Kulkarni SG, et al. Stimulated tissue repair prevents lethality in isopropanolinduced potentiation of carbon tetrachloride hepatotoxicity. Toxicol Appl Pharmacol 1996; 140(2):235–44. 86. Kodavanti PR, Kodavanti UP, Faroon OM, et al. Pivotal role of hepatocellular regeneration in the ultimate hepatotoxicity of CCl4 in chlordecone-, mirex-, or phenobarbital-pretreated rats. Toxicol Pathol 1992; 20(4):556–69. 87. Mehendale H. Mechanism-based predictioxns of dose–response relationships: why low doses or CCl4 are nontoxic. Belle Newslett 1994; 2:1–7. 88. Dalu A, Cronin GM, Lyn-Cook BD, et al. Age-related differences in TGF-alpha and proto-oncogenes expression in rat liver after a low dose of carbon tetrachloride. J Biochem Toxicol 1995; 10(5):259–64. 89. Limaye PB, Bhave VS, Palkar PS, et al. Upregulation of calpastatin in regenerating and developing rat liver: role in resistance against hepatotoxicity. Hepatology 2006; 44(2):379–88. 90. Mehendale HM, Limaye PB. Calpain: a death protein that mediates progression of liver injury. Trends Pharmacol Sci 2005; 26(5):232–6. 91. El-Serag HB, Everhart JE. Diabetes increases the risk of acute hepatic failure. Gastroenterology 2002; 122(7):1822–8. 92. Schmidt LE, Dalhoff K. Alpha-fetoprotein is a predictor of outcome in acetaminophen-induced liver injury. Hepatology 2005; 41(1):26–31. 93. Schmidt LE, Dalhoff K. Serum phosphate is an early predictor of outcome in severe acetaminopheninduced hepatotoxicity. Hepatology 2002; 36(3):659–65. 94. Horn KD, Wax P, Schneider SM, et al. Biomarkers of liver regeneration allow early prediction of hepatic recovery after acute necrosis. Am J Clin Pathol 1999; 112(3):351–7. 95. Zhang BH, Gong DZ, Mei MH. Protection of regenerating liver after partial hepatectomy from carbon tetrachloride hepatotoxicity in rats: role of hepatic stimulator substance. J Gastroenterol Hepatol 1999; 14(10):1010–7. 96. Lawson JA, Farhood A, Hopper RD, et al. The hepatic inflammatory response after acetaminophen overdose: role of neutrophils. Toxicol Sci 2000; 54(2):509–16. 97. Mehendale HM. Toxicodynamics of low level toxicant interactions of biological significance: inhibition of tissue repair. Toxicology 1995; 105(2–3):251–66. 98. Slater TF. Free-radical mechanisms in tissue injury. Biochem J 1984; 222(1):1–15. 99. Czaja MJ, Xu J, Ju Y, et al. Lipopolysaccharide-neutralizing antibody reduces hepatocyte injury from acute hepatotoxin administration. Hepatology 1994; 19(5):1282–9. 100. Cotran RS, Kumar V, Collins T. Cell pathology I: cell injury and cell death. In: Robbins RC, Kumar V, Collins T, eds. Pathologic Basis of Disease. Philadelphia, PA: W.B. Saunders Company, 1999:1–30. 101. Ju C, Reilly TP, Bourdi M, et al. Protective role of Kupffer cells in acetaminophen-induced hepatic injury in mice. Chem Res Toxicol 2002; 15(12):1504–13. 102. Kellogg EW, III, Fridovich I. Superoxide, hydrogen peroxide, and singlet oxygen in lipid peroxidation by a xanthine oxidase system. J Biol Chem 1975; 250(22):8812–7. 103. Blazka ME, Wilmer JL, Holladay SD, et al. Role of proinflammatory cytokines in acetaminophen hepatotoxicity. Toxicol Appl Pharmacol 1995; 133(1):43–52. 104. Carragher NO, Frame MC. Calpain: a role in cell transformation and migration. Int J Biochem Cell Biol 2002; 34(12):1539–43. 105. Croall DE, DeMartino GN. Calcium-activated neutral protease(calpain) system: structure, function, and regulation. Physiol Rev 1991; 71(3):813–47.

206

Mehendale

106. Miyoshi H, Umeshita K, Sakon M, et al. Calpain activation in plasma membrane bleb formation during tert-butyl hydroperoxide-induced rat hepatocyte injury. Gastroenterology 1996; 110(6):1897–904. 107. Saido TC, Sorimachi H, Suzuki K. Calpain: new perspectives in molecular diversity and physiological–pathological involvement. FASEB J 1994; 8(11):814–22. 108. Vaidya VS, Shankar K, Lock EA, et al. Role of tissue repair in survival from s-(1,2-dichlorovinyl)-Lcysteine-induced acute renal tubular necrosis in the mouse. Toxicol Sci 2003; 74(1):215–27. 109. Nishiura I, Tanaka K, Yamato S, et al. The occurrence of an inhibitor of Ca2C-dependent neutral protease in rat liver. J Biochem (Tokyo) 1978; 84(6):1657–9. 110. Maekawa A, Lee JK, Nagaya T, et al. Overexpression of calpastatin by gene transfer prevents troponin I degradation and ameliorates contractile dysfunction in rat hearts subjected to ischemia/ reperfusion. J Mol Cell Cardiol 2003; 35(10):1277–84. 111. Waxman L, Krebs EG. Identification of two protease inhibitors from bovine cardiac muscle. J Biol Chem 1978; 253(17):5888–91. 112. Suzuki K, Imajoh S, Emori Y, et al. Calcium-activated neutral protease and its endogenous inhibitor. Activation at the cell membrane and biological function. FEBS Lett 1987; 220(2):271–7. 113. Inomata M, Kasai Y, Nakamura M, et al. Activation mechanism of calcium-activated neutral protease. Evidence for the existence of intramolecular and intermolecular autolyses. J Biol Chem 1988; 263(36):19783–7. 114. McClelland P, Hathaway DR. The calpain–calpastatin system in vascular smooth muscle. FEBS Lett 1991; 290(1–2):55–7. 115. Adachi Y, Ishida-Takahashi A, Takahashi C, et al. Phosphorylation and subcellular distribution of calpastatin in human hematopoietic system cells. J Biol Chem 1991; 266(6):3968–72. 116. Lane RD, Allan DM, Mellgren RL. A comparison of the intracellular distribution of mu-calpain, m-calpain, and calpastatin in proliferating human A431 cells. Exp Cell Res 1992; 203(1):5–16.

11

Genetic Susceptibility to Drug-Induced Liver Disease Mark Russo and Paul B. Watkins

Department of Medicine, University of North Carolina, Chapel Hill, North Carolina, U.S.A.

INTRODUCTION Most drugs capable of causing severe liver injury do so on an idiosyncratic basis. Characteristics of idiosyncratic drug-induced liver injury (DILI) include lack of a clear dose relationship and occurrence in a small subset of treated patients. A reasonable but as yet unproven hypothesis is that the main factors or “traits” that render this subset of patients susceptible are inherited (i.e., genetic). We will first discuss general approaches and challenges to DILI genetic association studies. We will then summarize the data obtained in studies to date. We will also discuss reasons why these investigations have met with limited success and why studies in the near future may be more fruitful. THE APPROPRIATE SOURCE FOR DILI CASES AND CONTROLS A standard approach to determine genetic contribution to disease susceptibility involves comparing the prevalence of specific genetic variations in patients with and without the disease of interest (i.e., cases vs. controls). Typically, the genetic variation assessed accounts for altered function of a protein suspected to be involved in the pathogenesis of the disease of interest. This genetic variation is termed a polymorphism if it affects greater than 1% of the population. The techniques involved in detecting genetic polymorphisms are now straightforward; the key to accurate association studies is therefore confident identification of a sufficient number of patients with and without the disease of interest. For genetic studies on DILI, prospective clinical trials are very useful. They represent the best opportunity to identify patients who are not susceptible to DILI from a given drug (controls), particularly, if liver chemistries are frequently monitored during the study. Prospective clinical trials are also the best means to identify patients susceptible to mild DILI as indicated by elevations in serum alanine aminotransferase (ALT). ALT elevations attributed to treatment can occur in a substantial proportion of the subjects in some clinical trials, enabling sufficient sample sizes for both cases and controls for some association studies. Clinical trials also have limitations in terms of genetic studies on DILI. This is because most drug trials mandate that therapy be discontinued when the serum ALT rises above three to five times the upper limits of normal. It cannot be assumed that all patients with ALT elevations would have developed serious DILI had they continued treatment. This is because ALT elevations often reverse without stopping therapy with the offending medication, a process termed “adaptation” (1–3). Adaptation is demonstrated by isoniazid (INH) where up to 15% of treated patients experience ALT elevations exceeding three times the upper limits of normal, yet the serum ALT returns to normal despite continued therapy in most of these patients (4). Less than 1% of INH-treated patients develop symptomatic hepatitis necessitating discontinuation of treatment. The adaptation phenomenon has been observed with many other drugs, including troglitazone (5), a drug that was withdrawn from the market due to multiple Paul B. Watkins has relationships with the following corporations: consultation agreements with Astra Zeneca, BG Medicine, Actelion, Critical Therapies Inc., Fleets, Inc., Glaxo SmithKline, Hoffman-LaRoche, Novartis, Mc Neil, Pfizer, and Xanodyne.

208

Russo and Watkins

cases of acute liver failure. It is assumed that patients who adapt are not at substantial risk for subsequent toxicity and can continue to be safely treated with the medication. Hence, the population of greatest interest for genetic studies are the subset of patients with ALT elevations who would not adapt had treatment been continued. Short of liver failure, there are no agreed upon criteria to identify patients susceptible to progressive liver injury from a drug. It seems reasonable to include patients who develop hepatocellular jaundice since these individuals appear to have a substantial mortality (6). Patients susceptible to progressive liver injury defined in this way are rarely observed in clinical trials in part because such patients may be too rare to be present even in a relatively large clinical trial and because therapy may be stopped before jaundice would occur in those individuals. For the above reasons, prospective clinical trials are of limited value in determining genetic factors that underlie susceptibility to progressive DILI. Such patients are generally best identified as they declare themselves among the many thousands of exposed patients outside of clinical trials. IDENTIFICATION OF DILI CASES A problem in studying DILI is the inherent difficulty in making a confident diagnosis; the diagnosis of DILI is generally one of exclusion and can rarely be made with absolute certainty. For example, no cause can be identified for acute liver failure in approximately 20% of patients presenting to major medical centers with this condition (7); if these patients were being treated with known hepatotoxic drugs, they probably would be erroneously considered DILI cases. Cases misclassified as DILI undoubtedly exist in published genetic studies, diluting the strength of true associations. Genetic association studies on DILI have also suffered because standardized criteria for defining a DILI case as hepatocellular or cholestatic (8–10) have not been uniformly applied. This is important since the mechanisms involved in hepatocellular and cholestatic liver injury differ. Even when the diagnosis of DILI seems assured for a given patient, it may be difficult to confidently identify the responsible drug due to polypharmacy or surreptitious use of overthe-counter or herbal preparations. Identifying the specific drug responsible for DILI is important because many genetic factors underlying susceptibility are likely to be drug specific. Erroneous identification of the implicated drug would further reduce the power of many association studies. EARLY STUDIES: THE LYMPHOCYTE TOXICITY ASSAY Some of the earliest indications of genetic susceptibility factors for DILI involved study of lymphocytes obtained from patients who had recovered from DILI (11,12). In these studies, lymphocytes freshly obtained from these patients are added to a mixture containing the implicated drug and liver microsomes (which contain most of the enzymes implicated in reactive metabolite generation). The rate of lymphocyte death is assessed and compared to the rate observed when the identical assay is performed with lymphocytes obtained from people who did not experience DILI (11). Lymphocytes contain much of the detoxification machinery present in the liver, most notably microsomal epoxide hydrolase and enzymes involved in glutathione synthesis and conjugation (13–15). The assumption is that if a patient experienced DILI and has increased lymphocyte toxicity from the implicated drug, then susceptibility is due to deficient detoxification machinery. The sensitivity detected by this assay is presumed to have a genetic basis. This is because relatives of patients with DILI sometimes demonstrate enhanced lymphocyte toxicity to the implicated drug, even though they have never received treatment with drug (12). Studies performed to date with the lymphocyte toxicity assay are listed in Table 1 (11,12,16,17). It can be seen that most of the studies are quite old. The reason why this assay is not in current use is not clear, but may relate to difficulty in obtaining robust and reproducible results. It may be important to reinitiate these studies because the drug-induced liver injury network (18) (DILIN; described later) will be collecting lymphocytes for immortalization from patients with DILI and controls. If lymphocytes indeed contain some of the machinery relevant

209

Genetic Susceptibility to Drug-Induced Liver Disease

TABLE 1 Studies on Drug Hepatotoxicity Using Lymphocyte Toxicity Assays Study

Drug

Results

Farrell 1985 (12)

Halothane

Larrey 1988 (16)

Amineptine

Increased cell death in cases compared with healthy controls Increased cell death in cases compared with drug-treated controls

Shear 1986 (17)

Sulfonamide

Spielberg 1981 (11)

Phenytoin

Increased cell death from sulfonamide metabolites in cases compared with healthy controls Increased cell death in cases compared with healthy controls

Proposed mechanism Deficiency in epoxide hydrolase Deficiency in enzymes that detoxify amineptine metabolites other than epoxide hydrolase or glutathione Deficiency in enzymes that detoxify sulfonamide metabolites Deficiency in epoxide dehyrolase

Abbreviation: HCO, healthy controls.

to DILI susceptibility, and if this machinery is not lost during the immortalization process, investigators would have an unlimited tissue for mechanistic studies. ASSOCIATION STUDIES TARGETING DRUG METABOLIZING ENZYMES Majority of DILI genetic studies published to date are based on the “reactive metabolite hypothesis.” This hypothesis, which has been widely held for more than four decades, states that most DILIs result from production of reactive metabolites from the parent drug in the liver (19–23). According to this hypothesis, the reactive metabolite must accumulate within the hepatocyte to levels that exceed a critical “threshold.” Once this threshold is crossed, a cascade of events is initiated that culminates in hepatocyte death (Fig. 1). The concept underlying most

Safe pathways (CYPs 2D6, 2C9, 2C19, UGT1A1)

Drug

Reactive Metabolite

Bioactivation (CYP1A2 CYP2E1 NAT–2)

Stress

Elimination

Detoxification (GSTM1 GSTT1 epoxide hydrolase NAT–2)

Innate immune response (IL–4, Il–10) Acquired immune response (HLA)

Recovery

Progressive injury

FIGURE 1 Mechanisms underlying DILI produced by reactive metabolites and genes investigated in association studies. Reactive metabolites can cause stress to hepatocytes. Deficiency in the activity of enzymes or transporters involved in the safe elimination of drugs should increase susceptibility to DILI because more drug is shunted to bioactivation pathways. Susceptibility to DILI should also be increased by deficiency in the activity of enzymes or transporters involved in safe removal of reactive metabolites (detoxification). Genetic factors that predispose to accumulation of a reactive metabolite could be quite drug specific. Once hepatocytes become stressed, factors including the innate and acquired immune response appear to determine whether liver injury subsides (recovery) or progresses to a serious medical problem. These factors, which may be less drug specific, may account for why ALT elevations due to a drug often resolve despite continuing treatment. The role of genes targeted in published DILI association studies (Tables 2 and 3) is noted. Abbreviations: CYP, cytochrome P450; UGT1A1, UDP-glucuronosyltransferase-1; NAT2, N-acetyltransferase 2; GSTM, glutathione-S-transferase mu; GSTT, glutathione-S-transferase theta; HLA, human leukocyte antigen; IL, interleukin; DILI, drug-induced liver injury; ALT, alanine aminotransferase.

210

Russo and Watkins

of the published association studies is that DILI can result from an inherited tendency to accumulate the relevant reactive metabolite. For a given drug, a tendency to accumulate its reactive metabolite could theoretically result from increased activity of the enzyme involved in the production of this metabolite. However, a genetic basis for significant increase in the activity of an enzyme is unusual; substantial increases in enzyme activity generally reflect environmental but not genetic factors (i.e., induction) (24). Investigators have therefore generally focused on the possibility that susceptibility reflects inheritance of dysfunctional enzymes involved in either detoxification or the safe elimination of the drug (which would result in shunting more parent through the reactive metabolite pathway) (1,23). As the specific enzymes involved in these processes were identified for a given drug, it was appreciated that some are polymorphic. The simplest approach to association studies is to determine whether “poor metabolizers” for a given enzyme were overrepresented among patients with DILI. Table 2 summarizes studies that have examined associations between susceptibility to DILI and polymorphisms in drug metabolizing enzymes. Only complete reports with more than five DILI cases are included. Single case reports and brief reports or letters are excluded (46–51). Studies are listed as “positive” if the investigator believed the results support the association sought. In many of these studies, genetic testing was not performed. Rather, the subjects were given drug “probes” that appear to selectively measure the activity of the target enzyme. In these tests, the probe is administered and the enzyme activity is inferred from the drug clearance or surrogates for clearance such as measurement of metabolites in urine (52). A potential issue with the probe approach is that the liver phenotype after recovery from DILI might not be identical to the pre-DILI phenotype (whereas genomic DNA does not change). Only one study administered a drug probe to patients prior to starting therapy with the implicated drug (43). An important point is that where associations were reported (Table 1), none appeared to be sufficiently robust to justify screening of patients prior to drug therapy and very few of the associations have been replicated by others. Indeed, the conclusions of some studies are contradictory. For example, several steps in the metabolism of INH involve N-acetyltransferase 2 (NAT2). The population can be divided into fast and slow acetylators based on a polymorphism in NAT2 (53). Rapid acetylators are reported to be at increased risk for INH hepatotoxicity in some studies (32,40), but more recent studies report higher risk in slow acetylators (30,33,35). Both findings have biological plausibility. Rapid acetylators could be at higher risk for DILI because acetyl-INH is hydrolyzed to acetylhydrazine, which is converted to toxic metabolites by cytochrome P450 2E1 (CYP2E1) (40). However, some toxic intermediates produced from INH are detoxified by NAT2, which would put slow acetylators at higher risk (30). Negative association studies published to date (Table 2) may not mean the associations sought are not present. As previously discussed, the ability to detect true associations would be diminished by inclusion of cases that did not have DILI and/or lumping of serious DILI and asymptomatic ALT elevations (which should not have identical susceptibility factors). In addition, most published studies to date have involved relatively small number of cases and controls. Relatively large sample sizes may be required in DILI association studies because the contribution of any single polymorphism to susceptibility is likely to be small. Indeed, it is not yet clear that susceptibility to DILI is mainly genetic in origin as non-genetic (i.e., environmental) factors are also likely to be important (54). Moreover, the genetic contribution to susceptibility will probably not reflect the contribution of a single polymorphism. Polymorphisms are by definition present in more than 1% of the population and the prevalence of progressive, idiosyncratic DILI is never this great. Hence, only a subset of patients with a given polymorphism could be susceptible to progressive DILI. One possibility is that genetic predisposition to DILI involves variation in a single gene which occurs in the population with much lower frequency than polymorphisms. A more likely scenario is that susceptibility to DILI involves polymorphisms in multiple genes: The most highly susceptible individual will therefore have functional variation in several different genes that each reflect polymorphisms (55). This idea is supported by a study involving tacrine, an anticholinesterase inhibitor used to treat Alzheimer’s disease. Glutathione transferases detoxify reactive metabolites generated by metabolism of tacrine by cytochrome P450 1A2 (CYP1A2) (56–58). Among 141 patients treated with tacrine, 72% of individuals null for both (Text continues on page 215)

Hepatotoxic drug(s)

Source of cases and controls

Due to only 3 atrium cases, HC/CH elevation in bilirubin in Higher mean mephenytoin correlations are not none of the 3 atrium cases hydroxylation index in the 3 significant. No correlations and 10 of the 21 other DILI atrium cases compared with evident between other DILI HCO and to other DILI cases. cases and controls 2 atrium cases were classified as CYP2D6 deficient

Few cases and only mild 7 cases of hepatotoxicity, 2 hepatotoxicity cases heterozygous for TPMT*3A mutation, pZ 0.005, TPMT activity did not correlate with hepatotoxicity

HC, AST or ALT O2! ULN

One UGT1A6 SNP associated with low enzyme activity

Comments

Strong association with known Cases were not true DILI but Gilbert’s disease selective failure to conjugate polymorphism pZ2!10K22 bilirubin

SNPs in UGT1A6 most associated with DILI

Results

Total bilirubin O2 mg/dL

HC ALT O1.5! ULN severe cases not mentioned

Type of liver injury

(Continued )

Study demonstrated that Significant association with Subjects recruited HC level of ALT elevation not strength of associations may stated, number jaundiced not GSTT1 in low-exposure VCM from PVC plants, vary depending on the level of stated genotype, high-exposure divided into lowexposure VCM group more likely to and highhave CYP2E1 c2/c2 exposure groups genotype, p!0.01

Subjects with neurological disorders, predominantly myasthenia gravis 1 hospital

135 cases/274 Recruited from controls receiving clinical trials drug without DILI 127 cases/909 drug- Subjects from treated controls clinical trial

Number of cases and controls

TPMT genotype 129 patients/115 and HCO phenotype

UGT1A1/ genotype

UGT1A1/ genotype

Gene/genotype or phenotype

Horsmans Phenobarbital 3 atrium cases/21 CYP2C19/phe 1994 (28) febarbamwith DILI from notype ate difebarother drugsa (mephenybamate toin hydroxycombination lation index) (Atriumw) and other drugsa CYP2D6 pheno 23 HCO type/(dextromethorphan metabolic ratio) Huang 1997 VCM GSTT1, GSTM1, 251 male workers (29) CYP2E1/ genotype

Danoff 2004 N-(3,4(26) dimethoxy cinnamoyl)anthranilic acid Heckmann Azathioprine (27)

Positive studies Acuna 2002 Tolcapone (25)

Source first author

TABLE 2 Results from Studies on Genetic Associations of Drug-Induced Liver Injury

Genetic Susceptibility to Drug-Induced Liver Disease

211

Hepatotoxic drug(s)

Sulfonamides

Shear 1986 (17)

Shimizu INH and 2005 (35) rifampin

Roy 2001 (34)

INH, rifampin, ethambutol, strepto mycin INH

Ohno 2000 (33)

Mitchell INH 1975 (32)

Huang 2003 INH (31)

Huang 2002 INH (30)

Source first author

Number of cases and controls

Source of cases and controls Type of liver injury

Results

Comments

NAT2/genotype

HC included all cases with ALT Higher rate of DILI in slow Most patients had mild DILI and 33 cases/191 treated Tapei Veterans General Hospital O2! ULN, number with acetylators versus controls could continue treatment. All patients who did bilirubin elevations not given (26.4% vs. 11.1%, pZ0.013) received pyrazinamide, not develop rifampin, and ethambutol for toxicity two months Controlled for NAT2 effect CYP2E1 (c1 and 49 cases/269 treated Taipei Veterans HC included all cases with ALT Higher prevalence of c1/c1 observed in many of these c2 alleles) without DILI General Hospital O2! ULN, number with CYP2E1 genotype in DILI same patients in earlier study and NAT2/ bilirubin elevations not given cases than controls (20% vs. (30). Most patients had mild genotype 9%, pZ0.009). Slow DILI and could continue acetylators with c1/c1 treatment. All received genotype were 7.3 times pyrazinamide, rifampin, and more likely to develop DILI ethambutol for two months compared with rapid acetylators with the c2/c2 genotype Sulfamethazine 26 cases Hospital HC, mild to severe hepatitis 86% of cases with probable INH phenotype hepatitis were rapid acetylators compared with expected frequency of 45% Slow acetylator genotype NAT2 77 cases/194 HCO TB clinic HC, cases have AST or ALT associated with increased O2! baseline or 1.5! risk of hepatotoxicity (RRZ ULN, number of cases with 28; 95% CI: 26 30) jaundice not reported Cases were on other HC, all cases had total bilirubin Higher prevalence of null GSTT1, GSTM1, 33 cases/33 treated Single hospital, antituberculosis drugs, India O3 mg/dL mutation for GSTM1 in patients who did NAT2/ including rifampin, but all patients with DILI (52% vs. not develop genotype cases had severe DILI 24% in controls, p!0.05) toxicity 6 cases/20 HCO Hospital HC/mixed 3 cases with All 6 cases were slow Small number of cases N-Acetylation jaundice acetylators, greater than phenotype/ expected, significantly caffeine metabolite ratio NAT2/genotype 10 cases/32 treated TB clinic HC, cases had ALT or AST O 80% slow acetylators and 9.1% No cases had jaundice. controls 2! ULN rapid acetylators developed Contradicts studies 45, 46 elevated AST or ALT, p!0.05

Gene/genotype or phenotype

TABLE 2 Results from Studies on Genetic Associations of Drug-Induced Liver Injury (Continued)

212 Russo and Watkins

Fontana Tacrine 1996 (43)

Negative studies Aithal 2000 Diclofenac (42)

Yamamoto INH, rifampin, 1986 (40) streptomycin Aithal 2004 Diclofenac (41)

Hospital and HC, 3 with liver failure, 6 had regional referral jaundice Single center

24 cases/100 HCO

37 patients tested before starting therapy, 12 had ALT O3! ULN

CYP2C9/ genotype

HC, only 2 cases with ALT O 10! ULN and no jaundice

HC, 9 patients with jaundice

3 hospitals

24 cases/48 control on diclofenac/4 to 321 HCO depending upon the polymorphism studied

Interleukin-10 and interleukin-4 receptor genotype

CYP1A2 phenotype/ (caffeine breath test)

HC elevation in AST or ALT

Hospital

(Continued )

Polymorphisms in CYP2C9 were Included in vitro studies not associated with DILI suggesting little CYP2C9 activity differences related to alleles studied No correlation between CYP1A2 Only prospective phenotyping activity and ALT study, CYP3A4 and CYP2D6 activities measured in a subset with no correlations evident

IL-10-627A and IL-4-590T more Included two control groups demonstrating results may common in cases compared vary depending on the choice with HCO. Possession of both of control group polymorphisms more common in cases than those treated without DILI

14 rapid acetylators in cases and 7 rapid acetylators in controls

HC ALT O3! ULN, number Combined GSTM1 and T1 null Tacrine very rarely causes with bilirubin elevations not genotypes associated with severe liver injury despite given 2.84 relative risk (pZ0.001) high incidence of ALT elevations HC, ALT O3! C677T mutation more common Mild cases of hepatotoxicity in patients discontinuing treatment for ALT elevation (RRZ2.38; 95% CI: 1.06 5.34) Mostly HC (not stated how Significant association with DILI Similar genetic association many jaundice cases) and GSTT1 GSTM1 null previously reported with DILI genotype (pZ0.008) from tacrine (36) All cases had jaundice All cases were poor sulfoxidizers Sulfoxidation enzyme unknown. CYP2D6 phenotype also compared with 22% normal measured controls, p!0.001 and 23.8% liver disease controls, p!0.001, cases had normal hydroxylation capacity

Hospital

12 subjects/HCO/ liver disease controls

Multicenter, Japanese

Multicenter study of rheumatoid arthritis

Recruited from multicenter study

104 cases/182 treated controls

Sulfoxidation, hydroxylation/phenotype (s-carboxymethylL-cysteine) NAT2/phenotype/caffeine

GSTT1, GSTM1/ 25 cases/85 treated genotype controls

Watanbe Troglitazone 2003 (38)

Watson Chlorpro1988 (39) mazine

MTHFR genotype

van Ede (37) Methotrexate

30 cases/147 folinic acid controls, 79 placebo controls

GSTT1, GSTM1, 52 cases/89 treated genotype controls

Simon 2000 Tacrine (36)

Genetic Susceptibility to Drug-Induced Liver Disease

213

Amineptine

Halothane

Larrey 1989 (16)

Ranek 1993 (45)

CYP2D6 phenotype/ (dextrome thorphan metabolic ratio) Phenotype CYP2D6/ dextromethor phan CYP2D6 phenotype/ debrisoquine hydroxylation

Gene/genotype or phenotype

Single center

7 cases/16 HCO

HC, 4 cases had jaundice, 2 had encephalopathy

Presenting to single Mostly mixed center

8 cases/17 HCO

HC, CH 28 had bilirubin elevations

Type of liver injury

Not stated

Source of cases and controls

51cases/103 HCO

Number of cases and controls

Comments

No differences between cases and controls in metabolism of probe drugs

Small number of subjects

Small number of subjects All cases were CYP2D6 extensive metabolizersprevalence not different from controls

CYP2D6 phenotype distribution Few cases with DILI due to not different between DILI same drug patients and controls

Results

b

Amineptine, amitrypyline, chlorpromazine, erythromycin, fipexide, clotiazepam, papaverine, pirprofen, diclofenac, glafenine, ketoconazole, amodiaquine, INH, nilutamide, trimetazidine, and dapsone. Amineptine, amodiaquine, diclofenac, erythromycin, amitryptyline, clometacin, dapsone, dextroproxyphene, INH, chlorpromazine, diltiazem, enalapril, gold salts, ibuprofen, indapamide, ketoconazole, levomepromazine, methyldopa, papverine, piroxicam, and tienilic. Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; CH, cholestatic liver injury; CI, confidence interval; GGT, g-glutamyl transpeptidase; GSTM, glutathione-S-transferase mu; GSTT, glutathione-S-transferase theta; HC, hepatocellular liver injury; HCO, healthy controls; INH, isoniazid; MTHFR, methylenetetrahydrofolate reductase; NAT2, N-acetyltransferase 2; P, phenotype with probe assay; SNP, single nucleotide polymorphism; TPMT, thiopurine S-methyltransferase; UGT1A, UDP-glucuronosyltransferase-1; ULN, upper limit of normal; VCM, vinyl chloride monomer.

a

Many drugs

b

Hepatotoxic drug(s)

Larrey 1989 (44)

Source first author

TABLE 2 Results from Studies on Genetic Associations of Drug-Induced Liver Injury (Continued)

214 Russo and Watkins

Genetic Susceptibility to Drug-Induced Liver Disease

215

glutathione transferases M1 and T1 developed ALT elevations three times the upper limit of normal (ULN) compared with 30% of patients not deficient in either glutathione transferase (36). This could explain why other smaller studies (48,49) that searched for associations between tacrine toxicity and either glutathione transferase M1 or T1 were unsuccessful. Another illustration of success when multiple polymorphisms were considered involves INH hepatotoxicity (31). The investigators reported that slow acetylators who also possess homozygous wild-type alleles of CYP2E1 (c1/c1) were 7.4 times more likely to develop INH hepatitis compared with rapid acetylators homozygous for the CYP2E1 c2/c2 alleles (31). It also seems likely that future association studies will be more successful if they incorporate polymorphisms in drug transport proteins (59,60). Drugs and their metabolites generally do not passively diffuse across the hepatocyte membranes. Rather, flux both into and back out of the hepatocyte is usually the result of specific drug transporters present on the apical or canalicular membrane. Transport at the basolateral membrane typically involves transporters in the organic anion transporter (OAT) and organic anion transporting polypeptide (OATP) families, whereas efflux into bile typically involves the canalicular transporters multidrug resistance protein2 (MRP2), breast cancer resistance protein, and multi-drug resistance1 (MDR-1). Several functional polymorphisms have been described in drug transporters (61). GENETIC VARIATION IN EVENTS NOT INVOLVING METABOLISM According to the reactive metabolite hypothesis, accumulation of a reactive metabolite above a critical threshold initiates a cascade of events that culminates in liver injury (Fig. 1). A relatively recent discovery is that in rodents, chemical or genetic modification of these “downstream” events can greatly influence the extent of liver injury (2). It logically follows that functional variation in the genes influencing these events could account for susceptibility to DILI. There is great interest in this possibility because the increased susceptibility related to these downstream events may be less drug specific than those involved in the accumulation of reactive metabolites. Hypotheses based on downstream events may therefore be easier to test since it might be possible to group cases of DILI due to multiple drugs, increasing sample size. Clues to specific genes involved in critical downstream events have emerged from studies of acetaminophen toxicity in rodents. These studies have implicated intrahepatocyte signaling pathways, certain cytokines, and the innate immune system (2,62,63). Functional polymorphisms have been identified in some of the implicated cytokines (64), but to date only a single study (41) has investigated the role of cytokine polymorphisms in susceptibility to DILI. These investigators reported a higher prevalence of both a polymorphism in interleukin-10 (IL-10627A) and a polymorphism in interleukin-4 (IL-4-590T) among 24 patients who developed DILI from diclofenac than among 48 patients treated with diclofenac who did not develop DILI. These associations have some biological plausibility because mice with homologous deletion (“knock out”) of the IL-10 or IL-6 genes demonstrate enhanced susceptibility to acetaminophen hepatotoxicity (65,66). The IL-10-627A polymorphism is associated with lower expression of IL-10; however, the IL-4-590T polymorphism is associated with higher expression of IL-4 (41). IL-4 can induce hepatotoxicity. ROLE OF THE ACQUIRED IMMUNE SYSTEM Downstream events can also include activation of the acquired immune system. This can occur when reactive metabolites produce new antigens by covalently binding to hepatocyte proteins. Some forms of DILI clearly appear to involve the acquired immune system, most notably those drugs, such as halothane and phenytoin, that typically produce fever, skin rash, and peripheral eosinophilia. Even in the absence to clinical signs of hypersensitivity, the acquired immune system may have a role. For example, lymphocytes obtained from patients who have recovered from a DILI episode sometimes proliferate in response to exposure to the implicated drug ex vivo, even when there was no clinical evidence of hypersensitivity (67,68). It should be noted that in addition to promoting hepatocyte injury, components of the acquired immune system may also be involved in reducing injury and may therefore play a role in the adaptation response (69).

216

Russo and Watkins

The ability to mount an immune response to new antigens is clearly influenced by genetic factors, and family predisposition to DILI from halothane and phenytoin represented some of the first evidence of inherited predisposition to DILI (11,12). The best studied genetic polymorphisms that influence immune response involve the human leukocyte antigen (HLA) genes, which are part of a much larger collection of genes termed the major histocompatability complex (70). Proteins coded by the HLA genes are expressed on cell membranes and present foreign antigens generated within the cells to circulating T cells. T cells will not react to antigens unless presented in this way; hence, the HLA molecules perform a critical role in many immune responses. The HLA genes are highly polymorphic and the HLA phenotype of an individual has traditionally been determined by serological cytotoxicity methods. Samples of lymphocytes are added to wells that contain different antibodies to specific HLA proteins. If the HLA antibody recognizes its antigen on the patient’s lymphocytes, these cells will lyse when complement is added. The pattern of wells showing cell loss reveals the combination of HLA antigens present on the patient’s lymphocytes. As the genetic basis for the HLA antigens has been characterized, more recent studies have inferred HLA phenotype from analysis of genomic DNA. Several studies have examined the associations between HLA polymorphisms and susceptibility to DILI (Table 2). Significant associations between DILI and HLA are difficult to establish because numerous polymorphisms are typically measured and there is generally no apriori hypothesis. Appropriate statistical methods must therefore be applied to correct for multiple comparisons (71–73). A reasonable conclusion from the studies summarized in Table 3 is that susceptibility to some forms of cholestatic DILI probably involves HLA polymorphisms. Studies have been underpowered to exclude true associations in other forms of liver injury.

DILI NOT RELATED TO REACTIVE METABOLITES Drugs can cause liver injury by mechanisms that do not involve reactive metabolites. For example, high levels of accumulation of the parent drug may cause liver toxicity in some instances. Accumulation of parent perhexiline may account for why three out of four patients who recovered from perhexiline hepatotoxicity were found to be CYP2D6 poor metabolizers, and the fourth patient had “substantial impairment” in CYP2D6 activity (51). Drugs that cause selective injury to mitochondria typically cause a Reye’s-like syndrome. Inherited deficiencies in mitochondrial enzymes predispose children to aspirin-induced Reye’s syndrome and appear to predispose children to DILI from other drugs, including valproic acid (81–83). Another mechanism that may account for DILI is drug or drug metabolite inhibition of the bile salt excretory polypeptide (BSEP), the major transport protein responsible for efflux of bile acids across the canalicular membrane. During clinical trials of the 5-lipoxygenase inhibitor, bosentan, ALT elevations exceeding 3! ULN were observed in more than 10% of treated patients and rare cases of symptomatic liver injury were observed (84). Bosentan has been shown to inhibit BSEP, resulting in the accumulation of toxic bile acids (85). A reasonable hypothesis is that patients susceptible to bosentan DILI may have functional polymorphisms in BSEP or other transport proteins capable of removing bile acids from the hepatocyte as they accumulate. As a final note, genes involved in bile acid homeostasis seem to play an adaptive role during DILI due to diverse drugs. During recovery from liver injury due to acetaminophen and/or carbon tetrachloride in rodents, there is upregulation of BSEP and downregulation of the Na(C)-dependent taurocholate-cotransporting polypeptide, the major protein that pumps bile acids into the hepatocyte from sinusoidal blood (86). Similar changes have been observed in human liver during the recovery phase of acute injury, including injury due to DILI (87). These changes should result in the reduction in intracellular concentrations of potentially toxic bile acids, thus protecting the stressed cell. Impaired ability to respond in this way, which could result from inherited defects in the regulation of transporters or in other genes involved in

HLA A, B, and C method not stated

HLA A, B, C, and D

Nitrofurantoin derivatives

38 cases/837 health controls 7/154 healthy controls 50 cases

No associations found

DR2, Aw24 Bw52 increased in jaundiced cases, Bw44 increased in cases without jaundice No associations found

No associations found

No association between HLA haplotypes and DILI

71/2163 healthy controls 17 cases

56/290

22/134

13 cases classified as chronic reactions, 16 cases had jaundice

All cases had jaundice

24 cases had jaundice

Few cases from a single drug. Types of liver injury included HC, CH, mixed, 50 cases had jaundice Number of cases with jaundice not stated

Defined DILI as AST or ALT O5! ULN or total bilirubin O1.5 mg/dL or any increase in AST or ALT with symptoms, unable to identify culprit drug in some cases. Number of cases with jaundice not given

Cholestatic injury and all cases had jaundice

Single drug, all cases had liver biopsy, all had jaundice, 8 patients had eosinophilia and 2 patients had a skin rash

Heterogeneous group of cases with 63 different drugs included, wide range of injury. 89 cases had jaundice

Comments

Cinitapride, ebrotidine, ebrotidine/piroxicam, famotidine, lanosprazole, omeprazole, ranitidine, repaglinide/metformin, stanozolol/tamoxifen, clopidogrel, gemfibrozil, lovastatin, simvastatin, ticlopidine, atorvastatin, captopril, diltiazem, enalapril, irbersartan, losartan/clarithromycin, propafenone, danazol, clomifene, ethinylestradiol, tibolone, carbimazole, amoxicillin, amoxicillin/clavulinic acid, ampicillin/sulbactam, azithromycin, erythromycin, ceftriaxone, INH, midecamycin/ciprofloxacin, norfloxacin, pyrazinamide, rifampin/INH/pyrazinamide, rifampin/INH, roxithromycin, anastrazole/acarbose, asparaginase, azathioprine, estramustine, fluorouracil, flutamide, diclofenac, celecoxib, glycosoaminoglycan, ibuprofen, indomethacin, leflunomide, naproxen/tetrazepam, nimesulide, rofecoxib, tetrabamate, acetylsalicylic acid, amitrypyline, bentazepam, carbamazepine, chlorpromazine, and citalopram. b Amineptine, chlorpromazine, diclofenac, amitryptyline, tetrahydroaminoacridine, allopurinol, amodiaquine, amoxicillin/clavulinic acid, clomipramine, dapsone, dextropropoxyphene, erythromycin, exophone, glafenene, ketoconazole, indomethacin, pirprofen, plethoryl, acenocoumarol, acetylsalicylic acid, atrium, captopril, clometacin, clotiazepam, cyamemazine, dihydrazine, ethionamide, fipexide, iproniazid, isotretonoin, levomepromazine, lisinopril, nilotimide, oxomemazine, papaverine, phenytoin, piroxicam, quinidine, ranitidine, and viloxazine. Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; CH, cholestatic liver injury; DILI, drug-induced liver injury; HC, hepatocellular liver injury; HLA, human leukocyte antigen; INH, isoniazid; NIH, National Institutes of Health; ULN, upper limit of normal.

a

Ranek 1993 (45) Stricker 1988 (80)

Halothane

DRB1*1501 DRB5* Amoxicillin/ 0101 DQA1*0102 DQB1*0602 clavulinic acid Sharma 2002 INH, rifampin, DRB1*03, DQA1*0102, DQB1*0201, HLA (77) ethambutol, typing by PCR pyrizinamide, strepto mycin Phenotype studies HLA-A, HLA-B, HLA-DR, HLA-DQ typing by Berson 1994 40 different NIH lymphocytotoxicity technique and (78) drugsb antibody-coated microspheres Eade 1981 (79) Halothane HLA-A and HLA-B locus antigens. Lymphocytotoxicity assay Otsuka 1985 Halothane HLA-DR2 typing by NIH lymphocytotoxicity (72) technique

O’Donohue 2000 (76)

Increased frequency of DRB1*15 and DQB1*06 and decreased frequency of DRB1*07 and DQB1*02 among cases with cholestatic/mixed DILI only 57% cases had DRB1*1501 DRB5*0101 DQB1*0602 haplotype compared with 11.7% of controls, p!0.000005 70% cases had DRB1*1501 genotype compared with 20% of controls, pZ2.5! 10-6 DQB1*0201 more frequent and DQA1*0102 less frequent in cases

Results

140/635

DRB1*1501 DRB5*0101 DQB1*0602, HLA- 35/300 D typing by PCR, HLA-A and HLA-B typing by alloantisera

Number of cases/controls

Hautekee 1999 Amoxicillin/ (75) clavulinic acid

HLA

DRB1*15, DQB1*06, DRB1*07, DQB1*02, HLA typing by PCR

Drug

Genotype studies Andrade 2004 58 different (74) drugsa

Source

TABLE 3 Studies on the Association Between HLA and DILI

Genetic Susceptibility to Drug-Induced Liver Disease

217

218

Russo and Watkins

maintaining bile acid homeostasis, could theoretically contribute to DILI susceptibility from multiple drugs. THE FUTURE The techniques involved in detecting genetic polymorphisms are now straightforward and the associated costs are dropping substantially. The major limiting factor for rigorous association studies will therefore be collection of genomic DNA from a sufficient number off people who are clearly susceptible to progressive DILI from a given drug, as well as collection suitable controls. For the reasons discussed earlier, it will be important to separately analyze patients with progressive symptomatic liver injury as opposed to ALT elevations alone. It will also be important to distinguish different clinical and histological presentations of liver disease. To address this need, the National Institute for Diabetes and Digestive and Kidney Diseases is sponsoring the Drug Induced Liver Injury Network (DILIN) that began studies in 2004 (18). Detailed clinical and laboratory data are obtained from each subject and undergo a rigorous causality assessment process. Each subject also provides genomic DNA, serum, and lymphocytes for immortalization. In addition, each subject is enrolled in a registry and agrees to be recontacted for up to 20 years. This should allow ancillary studies including pedigree analyses and determination of phenotypes associated with identified genotypes. In terms of genetic analyses, it will probably be most cost-effective to have a single laboratory performing all genotyping and ultimately direct sequencing of genes. This will facilitate quality control, provide internal consistency, and will address the growing desire for oversight in genetic studies where links remain between the DNA samples and their donors. Once the genetic analyses of the DILIN cases and controls are available, there will be no shortage of hypotheses to test because possible loci for susceptibility are rapidly emerging from many different areas of research. For instance, any time liver injury is observed in a genetically manipulated mouse, potential susceptibility genes may be suggested. An example of this is the presence of liver disease in mice engineered to overexpress the human steroid and xenobiotic receptor, steroid and xenobiotic receptor (SXR) (88) (also termed pregnane X receptor). Since many drugs that cause liver injury are ligands for SXR, susceptibility to DILI could involve genetic variation in SXR or many of the genes activated by SXR. Likewise, studies on genetic determinants for susceptibility to human liver disease will continue to produce new target genes. For example, a functional polymorphism in the lymphocyte receptor chemokine (C-C motif) receptor 5 (CCR5) has been linked to susceptibility to primary sclerosing cholangitis (89). Mice with homologous deletion of this receptor have greatly increased susceptibility to acute liver failure from concanavalin A (90). Candidate susceptibility genes will also evolve from the application of toxicogenomics to preclinical toxicity testing. There have been and will continue to be attempts to identify mRNA transcripts that can be used as biomarkers for liver toxicity preclinical models. Several teams of investigators have treated rats with various hepatotoxic drugs and used mRNA chip technology to search for early changes in many thousands of mRNAs (91–93). New genes involved in initiating, promoting, or responding to hepatotoxicity are being identified and provide new potential loci for susceptibility. Finally, it should be noted that it is possible to search for associations with polymorphisms across the entire genome without having a specific hypothesis (94). However, such “unbiased” analyses generally require large number of cases and controls and it is unclear how fruitful this approach will be for very rare events, such as progressive DILI. In summary, the formation of the DILIN and similar networks, the rapid advances in DNA analysis techniques, and the ongoing identification of potentially relevant genes should greatly speed identification of the genetic determinants of susceptibility to DILI. REFERENCES 1. Watkins PB. Idiosyncratic liver injury: challenges and approaches. Toxicol Pathol 2005; 33:1–5. 2. Kaplowitz N. Idiosyncratic drug hepatotoxicity. Nat Rev Drug Discov 2005; 4:489–99.

Genetic Susceptibility to Drug-Induced Liver Disease

219

3. O’Brien PJ, Slaughter MR, Swain A, et al. Repeated acetaminophen dosing in rats: adaptation of hepatic antioxidant system. Hum Exp Toxicol 2000; 19:277–83. 4. Fountain FF, Tolley E, Chrisman CR, Self TH. Isoniazid hepatotoxicity associated with treatment of latent tuberculosis infection: a 7-year evaluation from a public health tuberculosis clinic. Chest 2005; 128:116–23. 5. Watkins PB, Whitcomb RW. Hepatic dysfunction associated with troglitazone. N Engl J Med 1998; 338:916–7. 6. Reuben A. Hy’s law. Hepatology 2004; 39:574–8. 7. Ostapowicz G, Fontana RJ, Schiodt FV, et al. Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States. Ann Intern Med 2002; 137:947–54. 8. Benichou C. Criteria of drug-induced liver disorders. Report of an international consensus meeting. J Hepatol 1990; 11:272–6. 9. Benichou C, Danan G, Flahault A. Causality assessment of adverse reactions to drugs—II. An original model for validation of drug causality assessment methods: case reports with positive rechallenge. J Clin Epidemiol 1993; 46:1331–6. 10. Lucena MI, Camargo R, Andrade RJ, Perez-Sanchez CJ, Sanchez De La Cuesta F. Comparison of two clinical scales for causality assessment in hepatotoxicity. Hepatology 2001; 33:123–30. 11. Spielberg SP, Gordon GB, Blake DA, Goldstein DA, Herlong HF. Predisposition to phenytoin hepatotoxicity assessed in vitro. N Engl J Med 1981; 305:722–7. 12. Farrell G, Prendergast D, Murray M. Halothane hepatitis. Detection of a constitutional susceptibility factor. N Engl J Med 1985; 313:1310–4. 13. Spielberg SP. In vitro human pharmacogenetics of reactive drug metabolite detoxification. Prog Clin Biol Res 1983; 135:107–18. 14. Spielberg SP. In vitro assessment of pharmacogenetic susceptibility to toxic drug metabolites in humans. Fed Proc 1984; 43:2308–13. 15. Spielberg SP, Gordon GB. Glutathione synthetase-deficient lymphocytes and acetaminophen toxicity. Clin Pharmacol Ther 1981; 29:51–5. 16. Larrey D, Berson A, Habersetzer F, et al. Genetic predisposition to drug hepatotoxicity: role in hepatitis caused by amineptine, a tricyclic antidepressant. Hepatology 1989; 10:168–73. 17. Shear NH, Spielberg SP, Grant DM, Tang BK, Kalow W. Differences in metabolism of sulfonamides predisposing to idiosyncratic toxicity. Ann Intern Med 1986; 105:179–84. 18. Hoofnagle JH. Drug-induced liver injury network (DILIN). Hepatology 2004; 40:773 (Accessed May 2, 2007 at http://www.dilin.dcri.duke.edu). 19. Gillette JR. Keynote address: man, mice, microsomes, metabolites, and mathematics 40 years after the revolution. Drug Metab Rev 1995; 27:1–44. 20. Pessayre D. Role of reactive metabolites in drug-induced hepatitis. J Hepatol 1995; 23:16–24. 21. Williams DP, Kitteringham NR, Naisbitt DJ, Pirmohamed M, Smith DA, Park BK. Are chemically reactive metabolites responsible for adverse reactions to drugs? Curr Drug Metab 2002; 3:351–66. 22. Park BK, Kitteringham NR, Maggs JL, Pirmohamed M, Williams DP. The role of metabolic activation in drug-induced hepatotoxicity. Annu Rev Pharmacol Toxicol 2005; 45:177–202. 23. Walgren JL, Mitchell MD, Thompson DC. Role of metabolism in drug-induced idiosyncratic hepatotoxicity. Crit Rev Toxicol 2005; 35:325–61. 24. Xu C, Li CY, Kong AN. Induction of phase I, II and III drug metabolism/transport by xenobiotics. Arch Pharm Res 2005; 28:249–68. 25. Acuna G, Foernzler D, Leong D, et al. Pharmacogenetic analysis of adverse drug effect reveals genetic variant for susceptibility to liver toxicity. Pharmacogenomics J 2002; 2:327–34. 26. Danoff TM, Campbell DA, McCarthy LC, et al. A Gilbert’s syndrome UGT1A1 variant confers susceptibility to tranilast-induced hyperbilirubinemia. Pharmacogenomics J 2004; 4:49–53. 27. Heckmann JM, Lambson EM, Little F, Owen EP. Thiopurine methyltransferase (TPMT) heterozygosity and enzyme activity as predictive tests for the development of azathioprine-related adverse events. J Neurol Sci 2005; 231:71–80. 28. Horsmans Y, Lannes D, Pessayre D, Larrey D. Possible association between poor metabolism of mephenytoin and hepatotoxicity caused by Atrium, a fixed combination preparation containing phenobarbital, febarbamate and difebarbamate. J Hepatol 1994; 21:1075–9. 29. Huang CY, Huang KL, Cheng TJ, Wang JD, Hsieh LL. The GSTT1 and CYP2E1 genotypes are possible factors causing vinyl chloride induced abnormal liver function. Arch Toxicol 1997; 71:482–8. 30. Huang YS, Chern HD, Su WJ, et al. Polymorphism of the N-acetyltransferase 2 gene as a susceptibility risk factor for antituberculosis drug-induced hepatitis. Hepatology 2002; 35:883–9. 31. Huang YS, Chern HD, Su WJ, et al. Cytochrome P450 2E1 genotype and the susceptibility to antituberculosis drug-induced hepatitis. Hepatology 2003; 37:924–30. 32. Mitchell JR, Thorgeirsson UP, Black M, et al. Increased incidence of isoniazid hepatitis in rapid acetylators: possible relation to hydranize metabolites. Clin Pharmacol Ther 1975; 18:70–9. 33. Ohno M, Yamaguchi I, Yamamoto I, et al. Slow N-acetyltransferase 2 genotype affects the incidence of isoniazid and rifampicin-induced hepatotoxicity. Int J Tuberc Lung Dis 2000; 4:256–61.

220

Russo and Watkins

34. Roy B, Chowdhury A, Kundu S, et al. Increased risk of antituberculosis drug-induced hepatotoxicity in individuals with glutathione S-transferase M1 “null” mutation. J Gastroenterol Hepatol 2001; 16:1033–7. 35. Shimizu Y, Dobashi K, Mita Y, et al. DNA microarray genotyping of N-acetyltransferase 2 polymorphism using carbodiimide as the linker for assessment of isoniazid hepatotoxicity. Tuberculosis (Edinb) 2006; 86:374–81. 36. Simon T, Becquemont L, Mary-Krause M, et al. Combined glutathione-S-transferase M1 and T1 genetic polymorphism and tacrine hepatotoxicity. Clin Pharmacol Ther 2000; 67:432–7. 37. van Ede AE, Laan RF, Blom HJ, et al. The C677T mutation in the methylenetetrahydrofolate reductase gene: a genetic risk factor for methotrexate-related elevation of liver enzymes in rheumatoid arthritis patients. Arthritis Rheum 2001; 44:2525–30. 38. Watanabe I, Tomita A, Shimizu M, et al. A study to survey susceptible genetic factors responsible for troglitazone-associated hepatotoxicity in Japanese patients with type 2 diabetes mellitus. Clin Pharmacol Ther 2003; 73:435–55. 39. Watson RG, Olomu A, Clements D, Waring RH, Mitchell S, Elias E. A proposed mechanism for chlorpromazine jaundice—defective hepatic sulphoxidation combined with rapid hydroxylation. J Hepatol 1988; 7:72–8. 40. Yamamoto T, Suou T, Hirayama C. Elevated serum aminotransferase induced by isoniazid in relation to isoniazid acetylator phenotype. Hepatology 1986; 6:295–8. 41. Aithal GP, Ramsay L, Daly AK, et al. Hepatic adducts, circulating antibodies, and cytokine polymorphisms in patients with diclofenac hepatotoxicity. Hepatology 2004; 39:1430–40. 42. Aithal GP, Day CP, Leathart JB, Daly AK. Relationship of polymorphism in CYP2C9 to genetic susceptibility to diclofenac-induced hepatitis. Pharmacogenetics 2000; 10:511–8. 43. Fontana RJ, Turgeon DK, Woolf TF, Knapp MJ, Foster NL, Watkins PB. The caffeine breath test does not identify patients susceptible to tacrine hepatotoxicity. Hepatology 1996; 23:1429–35. 44. Larrey D, Tinel M, Amouyal G, et al. Genetically determined oxidation polymorphism and drug hepatotoxicity. Study of 51 patients. J Hepatol 1989; 8:158–64. 45. Ranek L, Dalhoff K, Poulsen HE, et al. Drug metabolism and genetic polymorphism in subjects with previous halothane hepatitis. Scand J Gastroenterol 1993; 28:677–80. 46. Larrey D, Henrion J, Heller F, Babany G, Degott C, Pessayre D, Benhamou JP. Metoprolol-induced hepatitis: rechallenge and drug oxidation phenotyping. Ann Intern Med 1988; 108:67–8. 47. Seybold U, Landauer N, Hillebrand S, Goebel FD. Senna-induced hepatitis in a poor metabolizer. Ann Intern Med 2004; 141:650–1. 48. Becquemont L, Lecoeur S, Simon T, Beaune P, Funck-Brentano C, Jaillon P. Glutathione S-transferase theta genetic polymorphism might influence tacrine hepatotoxicity in Alzheimer’s patients. Pharmacogenetics 1997; 7:251–3. 49. De Sousa M, Pirmohamed M, Kitteringham NR, Woolf T, Park BK. No association between tacrine transaminitis and the glutathione transferase theta genotype in patients with Alzheimer’s disease. Pharmacogenetics 1998; 8:353–5. 50. Zielinska E, Niewiarowski W, Bodalski J. The arylamine N-acetyltransferase (NAT2) polymorphism and the risk of adverse reactions to co-trimoxazole in children. Eur J Clin Pharmacol 1998; 54:779–85. 51. Morgan MY, Reshef R, Shah RR, Oates NS, Smith RL, Sherlock S. Impaired oxidation of debrisoquine in patients with perhexiline liver injury. Gut 1984; 25:1057–64. 52. Streetman DS, Bertino JS, Jr., Nafziger AN. Phenotyping of drug-metabolizing enzymes in adults: a review of in-vivo cytochrome P450 phenotyping probes. Pharmacogenetics 2000; 10:187–216. 53. Boukouvala S, Fakis G. Arylamine N-acetyltransferases: what we learn from genes and genomes. Drug Metab Rev 2005; 37:511–64. 54. Ganey PE, Luyendyk JP, Maddox JF, Roth RA. Adverse hepatic drug reactions: inflammatory episodes as consequence and contributor. Chem Biol Interact 2004; 150:35–51. 55. Watkins PB, Seeff LB. Drug-induced liver injury: summary of a clinical research single topic conference. Hepatology 2006; 43:618–31. 56. Spaldin V, Madden S, Pool WF, Woolf TF, Park BK. The effect of enzyme inhibition on the metabolism and activation of tacrine by human liver microsomes. Br J Clin Pharmacol 1994; 38:15–22. 57. Madden S, Woolf TF, Pool WF, Park BK. An investigation into the formation of stable, protein-reactive and cytotoxic metabolites from tacrine in vitro. Studies with human and rat liver microsomes. Biochem Pharmacol 1993; 46:13–20. 58. Tingle MD, Pirmohamed M, Templeton E, et al. An investigation of the formation of cytotoxic, genotoxic, protein-reactive and stable metabolites from naphthalene by human liver microsomes. Biochem Pharmacol 1993; 46:1529–38. 59. Chandra P, Brouwer KL. The complexities of hepatic drug transport: current knowledge and emerging concepts. Pharm Res 2004; 21:719–35. 60. Ho RH, Kim RB. Transporters and drug therapy: implications for drug disposition and disease. Clin Pharmacol Ther 2005; 78:260–77.

Genetic Susceptibility to Drug-Induced Liver Disease

221

61. Cervenak J, Andrikovics H, Ozvegy-Laczka C, et al. The role of the human ABCG2 multidrug transporter and its variants in cancer therapy and toxicology. Cancer Lett 2006; 234:62–72. 62. Prandota J. Important role of proinflammatory cytokines/other endogenous substances in druginduced hepatotoxicity: depression of drug metabolism during infections/inflammation states, and genetic polymorphisms of drug-metabolizing enzymes/cytokines may markedly contribute to this pathology. Am J Ther 2005; 12:254–61. 63. Mehendale HM, Limaye PB. Calpain: a death protein that mediates progression of liver injury. Trends Pharmacol Sci 2005; 26:232–6. 64. De Maio A, Torres MB, Reeves RH. Genetic determinants influencing the response to injury, inflammation, and sepsis. Shock 2005; 23:11–17. 65. Masubuchi Y, Bourdi M, Reilly TP, Graf ML, George JW, Pohl LR. Role of interleukin-6 in hepatic heat shock protein expression and protection against acetaminophen-induced liver disease. Biochem Biophys Res Commun 2003; 304:207–12. 66. Bourdi M, Masubuchi Y, Reilly TP, et al. Protection against acetaminophen-induced liver injury and lethality by interleukin 10: role of inducible nitric oxide synthase. Hepatology 2002; 35:289–98. 67. Maria VA, Victorino RM. Immunological investigation in hepatic drug reactions. Clin Exp Allergy 1998; 28(Suppl. 4):71–7. 68. Pichler WJ. Direct T-cell stimulations by drugs—bypassing the innate immune system. Toxicology 2005; 209:95–100. 69. Uetrecht J. Current trends in drug-induced autoimmunity. Autoimmun Rev 2005; 4:309–14. 70. Piertney SB, Oliver MK. The evolutionary ecology of the major histocompatibility complex. Heredity 2006; 96:7–21. 71. Follmann DA. Comparing HLA antigen frequencies between two groups of patients. Stat Med 2003; 22:1999–2013. 72. Otsuka S, Yamamoto M, Kasuya S, Ohtomo H, Yamamoto Y, Yoshida TO, Akaza T. HLA antigens in patients with unexplained hepatitis following halothane anesthesia. Acta Anaesthesiol Scand 1985; 29:497–501. 73. Bothamley GH. Treatment, tuberculosis, and human leukocyte antigen. Am J Respir Crit Care Med 2002; 166:907–8. 74. Andrade RJ, Lucena MI, Alonso A, Garcia-Cortes M, Garcia-Ruiz E, Benitez R, Fernandez MC, Pelaez G, Romero M, Corpas R, et al. HLA class II genotype influences the type of liver injury in druginduced idiosyncratic liver disease. Hepatology 2004; 39:1603–12. 75. Hautekeete ML, Horsmans Y, Van Waeyenberge C, et al. HLA association of amoxicillin-clavulanate—induced hepatitis. Gastroenterology 1999; 117:1181–6. 76. O’Donohue J, Oien KA, Donaldson P, et al. Co-amoxiclav jaundice: clinical and histological features and HLA class II association. Gut 2000; 47:717–20. 77. Sharma SK, Balamurugan A, Saha PK, Pandey RM, Mehra NK. Evaluation of clinical and immunogenetic risk factors for the development of hepatotoxicity during antituberculosis treatment. Am J Respir Crit Care Med 2002; 166:916–9. 78. Berson A, Freneaux E, Larrey D, et al. Possible role of HLA in hepatotoxicity. An exploratory study in 71 patients with drug-induced idiosyncratic hepatitis. J Hepatol 1994; 20:336–42. 79. Eade OE, Grice D, Krawitt EL, et al. HLA A and B locus antigens in patients with unexplained hepatitis following halothane anaesthesia. Tissue Antigens 1981; 17:428–32. 80. Stricker BH, Blok AP, Claas FH, Van Parys GE, Desmet VJ. Hepatic injury associated with the use of nitrofurans: a clinicopathological study of 52 reported cases. Hepatology 1988; 8:599–606. 81. Fromenty B, Pessayre D. Inhibition of mitochondrial beta-oxidation as a mechanism of hepatotoxicity. Pharmacol Ther 1995; 67:101–54. 82. Gerber N, Dickinson RG, Harland RC, et al. Reye-like syndrome associated with valproic acid therapy. J Pediatr 1979; 95:142–4. 83. Young RS, Bergman I, Gang DL, Richardson EP, Jr. Fatal Reye-like syndrome associated with valproic acid. Ann Neurol 1980; 7:389. 84. Kenyon KW, Nappi JM. Bosentan for the treatment of pulmonary arterial hypertension. Ann Pharmacother 2003; 37:1055–62. 85. Fattinger K, Funk C, Pantze M, et al. The endothelin antagonist bosentan inhibits the canalicular bile salt export pump: a potential mechanism for hepatic adverse reactions. Clin Pharmacol Ther 2001; 69:223–31. 86. Aleksunes LM, Scheffer GL, Jakowski AB, Pruimboom-Brees IM, Manautou JE. Coordinated expression of multidrug resistance-associated proteins (Mrps) in mouse liver during toxicantinduced injury. Toxicol Sci 2006; 89:370–9. 87. Ros JE, Libbrecht L, Geuken M, Jansen PL, Roskams TA. High expression of MDR1, MRP1, and MRP3 in the hepatic progenitor cell compartment and hepatocytes in severe human liver disease. J Pathol 2003; 200:553–60. 88. Xie W, Barwick JL, Downes M, et al. Humanized xenobiotic response in mice expressing nuclear receptor SXR. Nature 2000; 406:435–9.

222

Russo and Watkins

89. Eri R, Jonsson JR, Pandeya N, et al. CCR5-Delta32 mutation is strongly associated with primary sclerosing cholangitis. Genes Immun 2004; 5:444–50. 90. Moreno C, Gustot T, Nicaise C, et al. CCR5 deficiency exacerbates T-cell-mediated hepatitis in mice. Hepatology 2005; 42:854–62. 91. Heijne WH, Kienhuis AS, van Ommen B, Stierum RH, Groten JP. Systems toxicology: applications of toxicogenomics, transcriptomics, proteomics and metabolomics in toxicology. Expert Rev Proteomics 2005; 2:767–80. 92. Ruepp SU, Tonge RP, Shaw J, Wallis N, Pognan F. Genomics and proteomics analysis of acetaminophen toxicity in mouse liver. Toxicol Sci 2002; 65:135–50. 93. Ruepp S, Boess F, Suter L, et al. Assessment of hepatotoxic liabilities by transcript profiling. Toxicol Appl Pharmacol 2005; 207:161–70. 94. Lawrence RW, Evans DM, Cardon LR. Prospects and pitfalls in whole genome association studies. Philos Trans R Soc Lond B Biol Sci 2005; 360:1589–95.

PART II

12

DIAGNOSIS AND MANAGEMENT

Clinicopathological Patterns of Drug-Induced Liver Disease Willis C. Maddrey

Department of Internal Medicine, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas, U.S.A.

INTRODUCTION The liver ranks high on the list of targets affected by adverse reactions to therapeutic or environmental agents (1–4). Awareness of drug-induced reactions affecting the liver has become increasingly a matter of concern. The clinical manifestations of drug-induced liver injury cover the full spectrum of liver disorders. It is often difficult and occasionally impossible to separate clinical manifestation caused by a drug from those caused by the underlying illness. Of particular concern are those drug-induced injuries that lead to clinically severe liver disease and acute liver failure (5). Much attention has been focused on drugs that were evaluated and approved, which later were withdrawn because of hepatotoxicity. The challenge is to improve the methods by which potential hepatotoxicity is identified so as to assure that an accurate assessment of risk is made. Hepatotoxicity has been a major reason that therapeutic agents with established favorable efforts have failed during preapproval trials or have been withdrawn following release. In the absence of specific tests to establish a drug as the cause of a liver disease, it is often impossible to confidently establish a cause-effect relationship between the use of a drug and the appearance of an injury. Judgment, often based on circumstantial data, is often the basis for the diagnosis. There is hardly a drug in use that has not been proven, or at least suggested, to cause some type of adverse reaction affecting the liver. The frequency, type and possible extent of injury must be considered. It is well recognized that minimal biochemical abnormalities do not necessarily mean clinically important liver injury is occurring. Clinical acceptance and success of a drug in the market depend on perceptions of efficacy and recognition of risks. Concern regarding hepatotoxicity has limited the use of many drugs. There are many reasons that the hepatotoxic potential of a drug might not be recognized until after the drug has been approved and is in widespread use (6). The sample size of patients studied may be relatively small and surely is not as broad as the population that will be exposed once a drug is on the market. Even with several thousand patients receiving a drug preapproval, rare events may be missed.

The author of this chapter has relationships with the following corporations: consultation agreements with Merck, Pfizer, Bristol-Myers Squibb, Sanofi-Aventis, Glaxo SmithKline, Novartis, Schering-Plough, Johnson & Johnson, Alinea, Daiichi Sankyo, Roche, Valeant, Wyeth, Biogen-IDEC, Astrellas, Encysive, Actelion, Aegerion and Abbott.

224

Maddrey

SPECTRUM OF DRUG-INDUCED LIVER DISORDERS Hepatic manifestations of drug-induced liver injury can mimic the entire spectrum of liver diseases (Table 1) (1–3). Minimal biochemical abnormalities appearing shortly after initiation of a drug occur in patients in whom there is no clinical evidence of liver disease. Some patients manifest hepatotoxicity as an acute hepatitis, acute liver failure (fulminant hepatocellular failure), chronic hepatitis, or cirrhosis. Cholestatic disorders ranging from those that are so mild as to be found only on routine biochemical testing to symptomatic cholestatic syndromes TABLE 1 Spectrum of Drug-Induced Liver Disorders Type of injury Hepatocellular injury Elevated aminotransferase levels Acute hepatitis

Chronic hepatitis

Acute hepatic failure Cholestatic reactions Cholestasis Simulate primary biliary cirrhosis Simulate primary sclerosing cholangitis Granulomas Simulate alcoholic hepatitis Steatohepatitis

Features Often asymptomatic Mimics acute viral hepatitis

May closely resemble auto immune hepatitis

Overwhelming liver failure Often prolonged course Oral contraceptives may simulate bile duct obstruction Antimitochondrial antibody negative Wide spectrum of diseases with and without evidence of hypersensitivity reaction

Phospholipidosis Vascular lesions Perisinusoidal fibrosis Peliosis hepatis

Hepatomegaly

Hepatic vein obstruction Veno-occlusive disease Sinusoidal dilation

Congestive hepatopathy Congestive hepatopathy Hepatomegaly

Neoplasms Hepatic adenoma Cholangiocarcinoma Angiosarcoma Hepatocellular carcinoma a

Drugs withdrawn after marketing. Abbreviation: INH, isoniazid.

Selected examples Almost all drugs INH Ketoconazole Troglitazonea Bromfenaca Diclofenac Methyldopa Minocycline Nitrofurantoin Methyldopa Oxyphenisatin Halothane INH Chlorpromazine Benoxaprofena Chlorpromazine Floxuridine Phenylbutazone Carba-mazepine (Table 2) Amiodarone Amiodarone Tamoxifen Several drugs used to treat HIV infection Amiodarone Vitamin A Oral contraceptives Anabolic steroids Azathioprine Oral contraceptives Cytotoxics Cytotoxics Oral contraceptives Oral contraceptives Anabolic steroids Anabolic steroids Thorotrast Vinyl chloride Anabolic steroids Thorotrast Danazol

Clinicopathological Patterns of Drug-Induced Liver Disease

225

TABLE 2 Hepatic Granulomas from Therapeutic Drugs Allopurinol Phenylbutazone Sulfonamides Carbamazepine Quinidine Hydralazine Methyldopa Phenytoin Amoxicillin clavulanic acid Procainamide d-Penicillamine

closely resembling primary biliary cirrhosis and primary sclerosing cholangitis are established reactions to several drugs. The finding of fat in the liver, both microvesicular and macrovesicular, may result as an expected event because of the known mechanism of action of a drug (especially those that are likely to cause mitochondrial injury), or as a clinically important untoward event occurring in a few individuals (7). There has been increased attention to the effects of drugs on mitochondrial respiration, which may lead to microvesicular fat, fatty acid accumulation, and decreased adenosine triphosphate levels (8,9). A few drugs are established causes of hepatic granulomatous inflammation indistinguishable from those found in a variety of infections and sarcoidosis (Table 2) (10–12). Some drugs (e.g., amiodarone) lead to an acquired phospholipidosis. With some drugs, hepatotoxicity is manifested as hepatic vein obstruction (Budd-Chiari syndrome). Tumors induced or promoted by therapeutic drugs range from benign hepatic adenomas, long associated with long-term use of oral contraceptives, to angiosarcomas, cholangiocarcinomas, and hepatocellular carcinomas. IDENTIFICATION AND DIAGNOSIS OF DRUG-INDUCED LIVER INJURY: THE ISSUE OF ESTABLISHING CAUSALITY The difficulties in establishing a drug as the cause for a liver injury and in determining its importance reflect the protean manifestations of drug-induced hepatic injury and the absence of specific diagnostic features. Drug-induced liver disease is usually clinically indistinguishable from liver injury of other causes and may only be detected through awareness of the possibility, suspicion, careful history, and inquisitive persistence by the clinician which should include a search for exposure to herbal medicines as well as potential toxins in the workplace or the environment. Many drugs cause minimal elevations in biochemical tests of the liver that are not accompanied by any signs or symptoms suggesting liver disease. These patients are identified only through random or preplanned blood testing. These elevations (especially of aminotransferase levels) may represent transient hepatic injury to which the patient rapidly adapts following the introduction of the new chemical compound. The adaptational responses may include changes in the innate immune system or the development of alternative pathways of disposition which develop leading to resolution of the injury (5,13). The mechanisms which lead to hepatic adaptation to the introduction of a new agent are under careful study. The ability of the liver to adjust (adapt) pathways favoring disposition and defense systems, which minimize or combat the effects of a toxic intermediate may determine which patients have a transient increase in aminotransferase levels that return to normal despite continued drug administration and which patients develop progressive severe injury. The failure to adapt may affect the ability to open new pathways of disposition, handle transient reactive intermediates, or be related to the immune response (or lack thereof) to the new agent (5,13–15). Most drugs that are associated with severe liver injury are found to cause transient elevations in aminotransferases in w3% to 15% of patients. Such is the case with isoniazid (INH) which causes minor, transient aminotransferase increase in 10% to 20% of patients but causes clinically apparent liver injury in only a few. It may be that the failure to adapt, related to, metabolic idiosyncrasy conditioned by age, gender, or the effects of other drugs explain a part of what we

226

Maddrey

now define as an idiosyncratic response. Forthcoming studies evaluating gene array responses (especially in the broad area of innate responses) to new agents will be of great interest and potential clinical importance (16,17). Alternatively, finding elevated levels of biochemical tests soon after introduction of a drug, and during a time when there are no symptoms or signs of liver injury, may indicate that the liver injury may progress and lead to clinically apparent disease. The unresolved dilemmas are how to develop ways to identify individuals who are susceptible and to determine effective ways to detect an adverse reaction that is likely to progress before serious liver injury develops. The diagnosis of hepatic injury caused by a drug is usually shrouded in uncertainty and based on circumstantial evidence, depending largely on suspicion by the clinician who recognizes that the time of onset and type of liver injury may be compatible with that of an adverse reaction to a therapeutic or environmental agent. This is especially true with a drug that has no recognized history of causing liver injury. The problem is compounded by the status of the underlying disease, concomitant illnesses and polypharmacy. Ingenuity and persistence are required to determine whether a liver abnormality represents an adverse drug reaction and to establish whether a drug or environmental agent is actually the cause. For example, in a patient who develops an angiosarcoma of the liver, exposure to the inciting agent, vinyl chloride, may have occurred many years before. The development of reliable tests to detect hepatitis A–E has made the task of excluding viral hepatitis more straightforward and certain. The finding of a positive antimitochondrial antibody test in a patient who has jaundice and biochemical evidence of cholestasis may go far to resolve concern as to whether the patient has primary biliary cirrhosis or a drug-induced syndrome that resembles the disorder. However, the clinician must remain aware that patients with well-defined liver disorders of all types may develop a superimposed drug-induced injury. Since the clinical and laboratory abnormalities of drug-induced injuries are often indistinguishable from liver disorders from other causes, the strongest supportive evidence implicating a drug may be resolution of manifestations of liver injury (dechallenge) following withdrawal of the drug. The problem in a patient, who is receiving many drugs, is which one or more to stop where evidence of injury is found. Knowledge of the most likely culprits helps in the decision process. In patients who have drug-induced hepatocellular injury, there is usually a marked decrease in elevated aminotransferase levels within two weeks of removing the drug. However, with some drugs, continued evidence of liver injury may persist for weeks to months. The longer the abnormalities persist, the greater the anxiety that hepatic failure might develop or that the diagnosis of a drug-induced injury is not correct. In patients who have predominantly cholestatic injury, there may be a delay of weeks or months before elevated alkaline phosphatase and bilirubin levels return to or towards normal after drug withdrawal. Rechallenge with a suspected drug to establish a diagnosis is seldom necessary and in patients in whom manifestations of acute hepatitis have occurred (especially if associated with jaundice), may be contraindicated. Histological evaluation of the liver is rarely diagnostic, often allowing recognition of type and extent of injury present, rather than clearly incriminating a drug or environmental agent as the cause. The dissociation between the considerable extent of injury found in liver biopsy and clinical evidence of liver disease, which may be mild or absent, is often striking. There are additional difficulties in determining a drug-induced injury in a patient who has another known factor that could explain the liver injury. An example includes the heightened toxicity of acetaminophen in patients who are chronic users of alcohol. Acetaminophen Ingestion of excessive amounts of acetaminophen (O10–15 g), often in suicidal attempts, predictably leads to liver injury ranging from acute hepatitis to acute liver failure and in some patients, death (18–22). Acetaminophen hepatotoxicity is the most frequent cause of acute liver failure in the United States (22). In therapeutic doses (%4 g/day), acetaminophen is usually quite safe and well tolerated. However, the safety margin of patients who are regular users of alcohol appears to be diminished and these individuals are at increased risk of developing acetaminophen-induced liver injury (18,19,23). The effects of chronic use of alcohol

Clinicopathological Patterns of Drug-Induced Liver Disease

227

on acetaminophen metabolism and disposition offer an explanation for the heightened susceptibility. Hepatic injury from acetaminophen is caused by the damaging effects of a reactive metabolic product, N-acetyl-benzoquinone-imide (NAPQI). Acetaminophen is predominantly metabolized by conjugation reactions to form sulfate and glucuronide metabolites, which are excreted in the urine. A lesser amount is metabolized by cytochrome P450 2E1 (CYP 2E1) to form NAPQI, which is bound to intracellular glutathione and excreted in the urine as mercapturic acid. When large doses of acetaminophen are ingested, the ability of the liver to form sulfate and glucuronide metabolites is overwhelmed and metabolism by CYP 2E1 to NAPQI becomes of much greater importance. In these situations, the capacity of glutathione to serve as an effective hepatoprotectant is exceeded, and the hepatocyte is vulnerable to an attack by reactive damaging intermediates. The downstream events that lead to cellular injury and death involve a variety of cytokines and chemokines including nitric oxide. Depletion of GSH caused by prolonged fasting may add to the injury. Furthermore, the important role of the innate immune system in the causation and progression of acetaminophen-induced liver injury has been suggested again indicating the critical role of the ability to adapt to injury on outcome (15). Careful questioning to elicit factors that predispose patients to hepatic injury from acetaminophen in nonsuicidal situations is important. First and most important is the dose of acetaminophen. Patients may have underestimated or understated the amount ingested, especially since acetaminophen is present in many widely used combination products. The intracellular concentration of NAPQI and dose of acetaminophen are clearly associated. Second is the concomitant use of alcohol. CYP 2E1 is the P450 subspecies involved both in metabolism of ethanol and in the metabolism of acetaminophen. Prolonged regular use of ethanol induces P450 2E1 activity. In individuals who are regularly using alcohol, doses of acetaminophen near or even within the suggested therapeutic range may lead to liver injury, especially if there is a coexistent decrease in intracellular glutathione. CYP 2E1 is induced in patients regularly using alcohol, therefore, potentially leading to more acetaminophen being metabolized to yield NAPQI. In addition, the intracellular concentration of glutathione may be lowered in patients who regularly use alcohol. No clinical features specifically define these patients. Suspicion, careful history, and determination of blood acetaminophen levels are required. As discussed earlier the susceptibility of an individual to develop acetaminophen toxicity may in large measure relate to the ability to effectively adapt to the toxic intermediate and the decrease in GSH. CLINICAL FEATURES OF DRUG-INDUCED LIVER DISEASE A few generalizations may be drawn regarding clinical and laboratory manifestations of liver injury from therapeutic drugs and environmental agents, especially those which cause hepatocellular injury. There may be few, if any, clinical signs suggesting liver injury, even in a patient who has biochemical and histological evidence of considerable damage. Often, symptoms are attributed to the disorder for which the drug is being given. Early symptoms, sometimes associated with drug-induced injuries, are usually nonspecific and include loss of appetite, fatigue, lassitude, and occasionally a dull discomfort more prominent in the right upper quadrant of the abdomen. These are the same signs and symptoms found (or not found!) in patients who have chronic viral hepatitis or alcohol-induced liver disease. With a few drugs, there is the concomitant presence of fever, rash, or eosinophilia—the hallmarks of hypersensitivity reactions. With many drugs, the appearance of clinical jaundice in a patient with hepatic injury is an indication of an adverse prognosis, with a fatal outcome occurring in approximately 10% (3,24,25). Therefore, jaundice appearing in a patient who has or might have a drug-induced liver disease is a cause for concern. Isoniazid A remarkable range of manifestations of hepatocellular injury can be caused by INH (3,23,26,27). Approximately, 1% of patients receiving INH develop clinically evident hepatic injury and some develop acute and occasionally overwhelming hepatitis. 10% to 20% of patients receiving INH have some increase in aminotransferase levels with onset within the

228

Maddrey

first several days to weeks after beginning administration of the agent, and the vast majority of these patients are asymptomatic. In most, there is a return to or toward normal (adaptation), despite continued use of INH. Several important susceptibility factors affect the likelihood of developing severe hepatic injury. INH hepatitis is rare in patients below 20 years of age, whereas patients older than 35 years have an incidence of liver disease of at least 1.5% (3,23). Prodromal signs and symptoms are vague. If clinically apparent jaundice develops, there is an approximate 10% mortality. There is general agreement that hepatotoxicity from INH results from the effects of an intermediary metabolite. There is evidence that genetic polymorphisms of CYP 2E1 and N-acetyltransferase may significantly influence susceptibility to INH drug-induced hepatitis. Furthermore, there are reports that show an increased susceptibility to INH induced liver injury in patients who have tuberculosis and chronic hepatitis C (28). The specific toxin has not been definitely established but may well be a hydrazine derivative. Concomitant use of rifampicin appears to increase the likelihood of an adverse reaction. Continued use of INH after the appearance of even nonspecific symptoms is associated with a likelihood of developing severe liver injury (26). Heightened awareness of the risk of INH-induced liver injury and regular monitoring of aminotransferase levels in patients receiving INH have proven effective in identifying evidence of hepatic injury at the time when withdrawal of the drug will lead to resolution. Nonsteroidal Anti-inflammatory Drugs Clinically significant adverse reactions affecting the liver are fortunately rare with the Nonsteroidal Anti-Inflammatory Drugs (NSAIDs) presently in use (1–4,29). However, reactions of many types do occur and need to be recognized as drug-related. The spectrum of liver manifestations resulting from NSAIDs encompasses minimal abnormalities in biochemical tests in asymptomatic patients to acute hepatitis, cholestatic hepatitis, and, in rare instances, acute hepatic failure. Particular attention was directed to these drugs when benoxaprofen was removed from the market following recognition of a progressive downhill course and death from hepatic and renal failure in a number of patients (3,30). Elderly females were especially vulnerable to develop severe injury. Bromfenac Bromfenac, a NSAID approved for short term (10 days or less) use in the management of pain, was withdrawn from the marketplace in 1998 shortly after its release because of several instances of severe hepatocellular necrosis and acute liver failure requiring liver transplantation (31). Several deaths were attributed to the use of bromfenac (32–34). Most patients in whom severe hepatic toxicity developed had received the drug for longer than the approved 10-day course. Sulindac Occasionally, patients receiving sulindac present with evidence of acute liver injury (35). Liver injury from sulindac appears within a few days to six weeks after therapy is initiated. Fever, rash, eosinophilia, and edema are frequently found in association with evidence of liver injury. Many of the patients have a predominantly cholestatic injury which is relatively mild. There have been a few deaths. The mechanism of sulindac-related injury is uncertain but likely results from an immune reaction to a metabolic product. Diclofenac The NSAID that has received particular scrutiny as regards hepatotoxicity is diclofenac (36). Liver injury from this drug presents, predominantly as hepatocellular injury, with several instances of severe hepatocellular necrosis and death. Females appear to be at increased risk. Onset of liver abnormalities most often appears within three months of beginning of the therapy, although in a few patients, a much longer presymptomatic interval has been noted. The role of prospective monitoring of biochemical tests in identifying early injury, and thereby reducing the risk of developing severe injury, is uncertain.

Clinicopathological Patterns of Drug-Induced Liver Disease

229

SIGNALS OF HEPATOTOXICITY Therapeutic drugs that are likely to damage the liver in many recipients at doses needed to elicit a response are usually identified during preapproval evaluation and discarded. The process of determining safety of a new agent extends over several years and observations are required in several thousands of patients before approval is granted. However, the rarer the event, the more likely a signal will be missed (Table 3). During preapproval testing, clinical, and laboratory manifestations indicating actual or potential hepatotoxicity are recorded and evaluated. There are several levels of concern (Table 3). Signals indicating hepatotoxicity that may be seen in before approval include the appearance of any instances of overt hepatocellular failure leading to death or liver transplantation. Even one such patient brings the proposed drug under great scrutiny and consideration as to whether development should continue. One level less severe is the recognition of patients who have acute hepatitis with symptoms of malaise, anorexia, right upper quadrant abdominal pain, and especially jaundice. Most of these patients survive although, as noted previously, clinical jaundice carries an ominous prognosis. The most difficult signals to interpret are elevated aminotransferase or alkaline phosphatase levels in patients who are asymptomatic. As a general guideline, slight increases in alanine aminotransferase (ALT) (!3!upper limit of normal (ULN)) in asymptomatic patients who received a new agent, and in whom there were normal aminotransferase levels before beginning the drug, continue in the trials with follow-up evaluations to determine if the events are transient and adaptation occurs. Patients who have elevations to !3!ULN to ! 8!ULN (and with some agents O3!ULN to !5!ULN), even when asymptomatic, are evaluated more extensively including immediate redetermination to note whether further increases are occurring. Many drug evaluation protocols have mandatory drug withdrawal if the aminotransferase level is O8!ULN (and in some instances O5!ULN) even in asymptomatic patients. There are several general observations regarding drug-induced liver disease that serve as guides: 1. 2. 3.

4.

Clinical manifestations of drug-induced hepatotoxicity are in the main indistinguishable from those of liver disease caused by other etiologies. Therefore, the diagnosis is often (almost always) made after exclusion of other possible etiologies. The risk of death from drug-induced hepatocellular injury is greater than that of cholestatic injury at least in the short run. In patients who develop hepatocellular injury from a drug, the appearance of clinically apparent liver disease, especially when associated with clinical jaundice, has a much less favorable prognosis than in patients who have acute viral hepatitis with an apparently similar degree of initial injury including the presence of clinical jaundice. Identification of a patient who develops acute hepatic failure leading to death, or the requirement for liver transplantation, will lead to considerable scrutiny and possible withdrawal of a drug. In these situations, there is at least serious consideration of institution of a monitoring schedule in an effort to detect injury at a time that withdrawal is likely to be effective in avoiding severe liver disease. Unfortunately, the value of monitoring systems in identifying patients who will develop acute liver failure has not been established.

TABLE 3 Signals Regarding Hepatotoxicity Major

Intermediate Minor

Development of acute liver failure Development of symptoms Onset of clinically apparent jaundice Appearance of ascites, encephalopathy, coagulopathy ALTO8!ULN ALTO5!ULN ALTO3!ULN Any elevation of ALT (!3!ULN) in an asymptomatic patient

Abbreviations: ALT, alanine aminotransferase; ULN, upper limit of normal.

230

Maddrey

5.

Histological evidence of injury, especially in patients who have hepatocellular injury, is often more severe than is suggested by clinical signs or laboratory studies. 6. Even large and extensive testing programs in which several thousand patients are evaluated may not detect an idiosyncratic event that occurs in the range of 1 in 10,000 to 1 in 100,000 individuals. Therefore, compilation of data in the first one or two postrelease years, when many are exposed, may be necessary to identify toxicity. 7. A few drugs slip through the safety screens during preapproval evaluation and must be withdrawn based on unfavorable experience in the marketplace. 8. Hepatic injury from a drug may have a somewhat specific signature as regards time of onset, type of injury, and propensity to develop severe disease (e.g., hepatocellular or cholestatic manifestations). However, exceptions occur and illegible signatures are frequent. 9. In general, drugs that cause hepatocellular injury are more likely to produce serious, even life-threatening injury than are drugs that cause cholestatic injury. 10. Some drugs (e.g., phenylbutazone) lead to two patterns of injury (10). In those patients, in whom granulomatous inflammation is found, liver disease tends to be less severe than in those in whom hepatocellular injury in the absence of granulomas is found (1–4).

ASSESSMENT OF POSSIBLE DRUG-INDUCED HEPATOTOXICITY It may be difficult or impossible to assess a drug’s contribution to hepatic injury in a patient who has an underlying disease known also to produce liver injury (1–4,13). There may be masking of the effects of the drug on the liver because of abnormalities associated with the underlying disease. Examples would include overlooking drug-induced hepatic injury in patients who have chronic viral hepatitis or cirrhosis, or in those with HIV infection and acquired immune deficiency syndrome, settings in which many drugs are used thereby presenting special difficulties (37–41). There are particular difficulties in determining a druginduced cause of liver injury in patients (especially those who are older) who are receiving many agents. When herbal medications and vitamins are also considered, the numbers of medications taken by an individual can be remarkably large. Furthermore, the timing of stopping, restarting, and substitution of drugs adds to the difficulty. An example of the complexities encountered is found in an older female over 50 years of age who is receiving antihypertensives, drugs to control hyperlipidemia, drugs for diabetes, and medications for heart failure or asthma. In addition, decisions regarding attribution of an injury to a drug becomes difficult in patients who are receiving many drugs often prescribed by several physicians and herbals taken without the knowledge of the physicians (1–3). In these situations, the clinician often must make a judgment call and withdraw the drug (or drugs) suspected of causing an injury, and then observe whether the liver abnormalities resolve. The need for more specific markers such as a characteristic genomic signature to incriminate an agent is apparent. The dilemma is heightened when one or more of the suspected drugs is definitely required for a serious underlying illness. In some patients, considerable liver damage may occur and progress without any clinical signs or symptoms in the early stages. There are many examples of liver injury progressing subclinically until there has been damage that is irreversible. Examples include progressive fibrosis and cirrhosis induced in some by prolonged use of methotrexate, and the hepatic failure that may develop in patients who have received amiodarone over prolonged intervals (3). Amiodarone Amiodarone, a benzofuran derivative used in the treatment of ventricular and atrial tachyarrhythmias, is an established cause of hepatic injury and acquired phospholipidosis (3,42–44). There are multiple side effects from amiodarone including pulmonary, thyroid, corneal, renal, and neurological toxicities. Liver injury is well recognized. Amiodarone is a cationic amphiphilic compound that accumulates in lysosomes. The drug and its major

Clinicopathological Patterns of Drug-Induced Liver Disease

231

metabolite desethylamiodarone, are stored in lysosomes within hepatocytes and bile duct epithelium, leading to phospholipidosis. Evidence of hepatic toxicity may appear within the first several months of beginning therapy with amiodarone or may become apparent after more than a year of treatment. Manifestations of liver injury may be subtle and include anorexia and fatigue. Hepatomegaly is often present. Types of liver injury associated with amiodarone in addition to phospholipidosis include acute liver failure, cholestatic hepatitis, steatohepatitis, and cirrhosis (44). Elevations in aminotransferase levels are found in 15% to 50% of patients, usually in the range of 2 to 10 times, the ULN. Occasionally, severe cholestasis occurs. Amiodarone-induced liver injury may closely simulate hepatic injury caused by alcohol with fibrosis, Mallory bodies, and active cirrhosis. The relation of the phospholipidosis, which results from a drug-induced inhibition of lysosomal phospholipases to the hepatocellular injury is uncertain. An unfortunate feature of amiodarone-induced injury is that even upon recognition of the drug-induced liver injury and withdrawal of the drug, there may be continued damage for months caused by the release of accumulated active drug from lysosomal reservoirs (45). Some patients have died from decompensated liver disease. There are no reliable ways to predict when hepatic toxicity from amiodarone is near a dangerous level and no way to accelerate removal of the drug from the lysosomal stores. RISK-BENEFIT CONSIDERATIONS With some drugs, the decision is made to accept the risk of hepatotoxicity to favorably treat a serious problem, especially if there are few, if any, alternatives. Such was the case with the drug tacrine used in the treatment of Alzheimer’s disease (46). Even though half of all patients receiving tacrine exhibited increases in serum aminotransferases, the possible benefits of the drug led to the decision to approve it albeit with a stringent monitoring schedule. Tacrine Tacrine is a reversible cholinesterase inhibitor used in the treatment of Alzheimer’s disease (46). Approximately, 50% of about 2500 patients who received the drug during clinical trials had ALT elevations. Women were more likely to have elevations than were men. ALT levels greater than three times the ULN occurred in 25%, and greater than 20 times the upper level of normal in 2%. Ninety percent of initial ALT elevations occurred within 12 weeks of beginning therapy (46). Elevations were noted after 12 weeks of therapy in only 10% of patients. Eosinophilia appeared to be associated with increased ALT levels, although fatigue, malaise, nausea, and vomiting did not occur more frequently in patients with elevated ALT levels compared to these manifestations in patients in the trials in whom ALT elevations did not occur. Through use of a monitoring program, patients who have considerable elevations in aminotransferase levels are identified and the drug withdrawn. P450 1A2 has a major role in tacrine metabolism and may produce a reactive intermediate (47). Fortunately, in most patients, there is resolution of the abnormal elevations of aminotransferases within several weeks after drug withdrawal. However, at least one death has been suggested to have been the result of tacrine-induced liver injury (48). Troglitazone Drugs that have established benefit, yet show evidence of hepatic toxicity, may remain on the market until either safer drugs that achieve the same benefit are developed, or the accumulation of evidence of severe hepatotoxicity leads to a decision to withdraw the agent. An example is troglitazone, which has been withdrawn because of hepatotoxicity and has been replaced by pioglitazone and rosiglitazone. Troglitazone, a thiazolidinedione agent, which is a peroxisome proliferator-activated receptors (PPARg) agonist, was shown to decrease hepatic glucose output and to increase insulin-dependent glucose metabolism in skeletal muscle. However, the drug was withdrawn from clinical use because of hepatic toxicity (5,49). Several instances of acute liver failure

232

Maddrey

leading to death or the need for liver transplantation occurred after the drug was approved in 1997 and was subsequently widely used. In the prerelease clinical trials of troglitazone elevations in aminotransferase levels to O3!ULN was found in 1.9% of the patients as compared to an incidence of 0.6% in patients receiving placebo. However, it was noted that minimal elevations of ALT were more frequent in patients receiving placebo possibly the result of favorable effects of troglitazone on underlying nonalcoholic fatty liver diseases. Two treated patients became clinically jaundiced during the trial which in retrospect was the strongest signal of potential hepatic failure. In many patients, biochemical abnormalities seen during the trial returned to normal during continued drug treatment (adaptation) or following drug withdrawal. Troglitazone-induced hepatic injuries were predominantly hepatocellular. Several patients developed acute liver failure and died or required liver transplantation (50–53). No fully supportable mechanism to explain troglitazone-induced liver injury has been established. In 2000, it was decided to withdraw the drug from the market. Other drugs in the PPARg agonist family (pioglitazone and rosiglitazone) have been approved and have not shown hepatotoxicity similar to that found with troglitazone. Two patients who presumably developed hepatic failure from rosiglitazone have been reported (54,55). DRUG-INDUCED CHRONIC HEPATITIS Several drugs cause chronic hepatitis syndromes that may be indistinguishable from autoimmune hepatitis (56–59). It is most important to recognize the drug cause for the liver injury. Misdiagnosing those patients as having autoimmune hepatitis may lead to the institution of corticosteroid therapy and continuation of the drug. The corticosteroids may blunt the manifestations of the injury while continued drug use leads to further damage. Generally, drugs that cause chronic hepatitis are taken for prolonged intervals and the extent of the injury correlates to some extent with the duration of therapy. It is uncertain whether the chronic hepatitis results from continued ongoing acute injury from the drug administered over a prolonged interval or whether the drug unmasks an injury in a genetically susceptible patient. The liver injury produced by some drugs including minocycline, nitrofurantoin, and methyldopa clinically and serologically closely mimics autoimmune hepatitis type I. Most of the reported cases have been in women. Most have increased serum globulin levels, and display the presence of autoantibodies, especially increased titers of serum antinuclear antibodies (ANA). The clinical onset of illness may be as an acute hepatitis in a patient in whom liver biopsy changes suggesting long-standing disease are found (an acute on chronic pattern), or the illness may develop as insidious progressive hepatic failure in a patient who has hepatosplenomegaly and evidence of portal hypertension or ascites. On liver biopsy, chronic inflammation including plasma cells is often found. Minocycline Minocycline, a tetracycline derivative used in the long-term treatment of acne, has been reported to cause several types of liver damage including acute hepatitis, often with features of hypersensitivity and rarely acute liver failure. Occasionally, cholestatic features predominate (60–65). A chronic hepatitis syndrome with features simulating autoimmune hepatitis has been reported. Furthermore, minocycline has been implicated in causing a drug-induced lupus syndrome (60,64). Although most patients who have minocycline-induced liver disease have been women, both sexes have been affected. Some patients have reported joint and muscle aches and pains as well as muscle stiffness. Hyperglobulinemia and the presence of ANA and anti-DNA antibodies have been reported. Nitrofurantoin Several types of liver injury, including acute hepatitis, cholestatic hepatitis, and chronic hepatitis, have been attributed to adverse reactions to nitrofurantoin (1–3,59,66–68). Nitrofurantoin-induced chronic hepatitis has occurred almost exclusively in women who

Clinicopathological Patterns of Drug-Induced Liver Disease

233

are middle-aged or older, and the most usual presentation is the insidious development of liver damage. Many have received nitrofurantoin for urinary antisepsis for longer than six months. Ascites, hypoalbuminemia, and hyperglobulinemia have been prominent features. Liver biopsy shows chronic hepatitis with bridging necrosis and occasionally cirrhosis. The hepatic manifestations of nitrofurantoin-induced injury often improve following withdrawal of the drug. Some patients have died of progressive liver failure even following drug withdrawal. Oxyphenisatin The first agent recognized as causing drug-induced chronic hepatitis was oxyphenisatin, a former component of several laxatives, which led to chronic hepatitis, cirrhosis, and liver failure—especially in older women who had received prolonged exposure (69). The clinical resemblance of oxyphenisatin-induced liver injury to the progressive liver disease of autoimmune hepatitis was often so close that many of these patients were treated with corticosteroids while continuing the drug. Once the drug relationship was noted and the agent withdrawn, resolution of at least the acute ongoing component of the injury occurred, although some patients were left with considerable damage. A quite similar chronic hepatitis syndrome occurred in patients receiving long-term treatment with the once widely used antihypertensive medication methyldopa (70). Tienilic Acid and Dihydralazine Two drugs, dihydralazine and tienilic acid, have been implicated as the cause of chronic hepatitis resembling autoimmune hepatitis, in which there is evidence of formation of antibodies against components of the cytochrome P450 system. Tienilic acid was transiently on the market as a uricosuric diuretic and was withdrawn following recognition of hepatotoxicity (3,71). Laboratory and clinical manifestations of liver disease caused by this drug, as well as by dihydralazine, are quite similar to those found in autoimmune hepatitis type I. Of additional interest is the observation that patients who developed liver injury from tienilic acid often had extensive hepatic injury in a setting in which there was development of anti-LKM2 antibodies. These antibodies were targeted against the cytochrome P450 2C9, the enzyme that catalyzes the hydroxylation of tienilic acid, therefore establishing that a drug can induce production of an autoantibody. Several of these patients had histological findings compatible with those found in classic chronic hepatitis. In dihydralazine-induced hepatitis, the predominant autoantibody is directed against cytochrome P450 1A2, which is important in the metabolism of the drug (58). Several patients receiving dihydralazine manifested features of chronic hepatitis. MECHANISMS OF INJURY: EFFECTS ON CLINICAL AND PATHOLOGICAL MANIFESTATIONS For some drugs there is evidence that genetically controlled pathways of metabolism play important role in determining which individuals are likely to have an adverse reaction affecting the liver (1–3,13,16,17). Undoubtedly in the future, identification of genetic control of susceptibility factors will become even more important and useful. A well-studied example of genetic susceptibility to hepatic injury is in the oxidative polymorphism of debrisoquine4-hydroxylase (CYP2D6), an enzyme important in the metabolism of several drugs (1–3). Individuals who have genetically determined impairment of debrisoquine-4-hydroxylase (up to 10% of the population) are at increased risk of developing an adverse reaction due to increased blood levels if exposed to a group of drugs that are metabolized by the enzyme, such as propranolol, quinidine, and desipramine, and are at increased risk of hepatic injury from perhexiline maleate, due to increased accumulation of the parent drug. Diphenylhydantoin An additional interesting story with genetic implications is that of adverse hepatic reactions that occur in patients receiving diphenylhydantoin. The hepatic injury that occasionally

234

Maddrey

develops in these patients may be severe, with intense liver necrosis often occurring as part of a syndrome that includes fever, exfoliative dermatitis, and eosinophilia (Stevens-Johnson syndrome) (72,73). The onset of evidence of an adverse reaction is usually within four weeks of beginning the drug. Up to half of the affected patients, who develop the full syndrome, died. Many features of diphenylhydantoin injury suggest important roles for immunological (hypersensitivity) reactions. However, it has been established that many of the patients have a genetically determined defect in detoxification, the nature of which is not certain (74). REFERENCES 1. Farrell GC. Drug-Induced Liver Disease. Edinburgh: Churchill-Livingstone, 1994. 2. Larrey D. Drug-induced liver diseases. J Hepatol 2000; 32:77–88. 3. Zimmerman HJ. Hepatotoxicity: The Adverse Effects of Drugs and Other Chemicals on the Liver. 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins, 1999. 4. Maddrey WC. Drug-induced hepatotoxicity. J Clin Gastroenterol 2005; 39:S83–9. 5. Watkins PB. Idiosyncratic liver injury: challenges and approaches. Toxicol Pathol 2005; 33:1–5. 6. Lee WM, Senior JR. Recognizing drug-induced liver injury: current problems, possible solutions. Toxicol Pathol 2005; 33:155–64. 7. Bryant AE, III, Dreifuss FE. Valproic acid hepatic fatalities. III. U.S. experience since 1986. Neurology 1996; 46:465–9. 8. Berson A, DeBeco V, Lette´ron P, et al. Steatohepatitis-inducing drugs cause mitochondrial dysfunction and lipid peroxidation in rat hepatocytes. Gastroenterology 1998; 114:764–74. 9. Day CP, James OFW. Steatohepatitis: a tale of two “Hits”? Gastroenterology 1998; 114:842–4. 10. Ishak KG, Kirchner JP, Dhar JK. Granulomas and cholestic-hepatocellular injury associated with phenylbutazon: report of two cases. Am J Dig Dis 1977; 22:611–7. 11. Mitchell MD, Boitnott JK, Arregui A, Maddrey WC. Granulomatous hepatitis associated with carbamazepine therapy. Am J Med 1981; 71:733–5. 12. Maddrey WC. Granulomas of the liver. In: Schiff ER, Sorrell M, Maddrey WC, eds. Schiff’s Diseases of the Liver. 8th ed. Philadelphia, PA: Lippincott-Raven, 1999:1571–85. 13. Kaplowitz N. Idiosyncratic drug hepatotoxicity. Nat Rev Drug Discov 2005; 4:489–99. 14. Walgren JL, Mitchell MD, Thompson DC. Role of metabolism in drug-induced idiosyncratic hepatotoxicity. Crit Rev Toxicol 2005; 35:325–61. 15. Liu Z-X, Govindarajan S, Kaplowitz N. Innate immune system plays a critical role in determining the progression and severity of acetaminophen hepatotoxicity. Gastroenterology 2004; 127:1760–74. 16. Evans WE, McLeod HL. Pharmacogenomics—drug disposition, drug targets, and side effects. N Engl J Med 2003; 348:538–47. 17. Weinshilboum R. Inheritance and drug response. N Engl J Med 2003; 348:529–37. 18. Seeff LB, Cuccherini BS, Zimmerman HJ, Adler E, Benjamin SB. Acetaminophen hepatotoxicity in alcoholics. Ann Intern Med 1986; 104:399–404. 19. Schenker S, Maddrey WC. Subliminal drug–drug interactions: users and their physicians take notice. Hepatology 1991; 13:995–8. 20. Zimmerman HJ, Maddrey WC. Acetaminophen (paracetamol) hepatotoxicity with regular intake of alcohol: analysis of instances of therapeutic misadventure. Hepatology 1995; 22:767–73. 21. O’Grady JG. Broadening the view of acetaminophen hepatotoxicity. Hepatology 2005; 42:1252–4. 22. Larson AM, Polson J, Fontana RJ, et al. Acetaminophen-induced acute liver failure: results of a United States multicenter, prospective study. Hepatology 2005; 42:1364–72. 23. Black M, Mitchell JR, Zimmerman HJ, Ishak KG, Epler GR. Isoniazid-associated hepatitis in 114 patients. Gastroenterology 1975; 69:289–302. 24. Bjo¨rnsson E, Olsson R. Outcome and prognostic markers in severe drug-induced liver disease. Hepatology 2005; 42:481–9. 25. Andrade RJ, Lucena MI, Ferna´ndez MC, et al. Drug-induced liver injury: an analysis of 461 incidences submitted to the spanish registry over a 10-year period. Gastroenterology 2005; 129:512–21. 26. Maddrey WC, Boitnott JK. Isoniazid hepatitis. Ann Intern Med 1973; 79:1–12. 27. Maddrey WC. Isoniazid-induced liver disease. Semin Liver Dis 1981; 1:77–84. 28. Ungo JR, Jones D, Ashkin D, et al. Antituberculosis drug-induced hepatotoxicity: the role of hepatitis C virus and the human immunodeficiency virus. Am J Respir Crit Care Med 1998; 157:1871–6. 29. Tolman KG. Hepatotoxicity of non-narcotic analgesics. Am J Med 1998; 105:13S–9S. 30. Taggart HM, Alderdice JM. Fatal cholestatic jaundice in elderly patients taking benoxaprofen. BMJ 1982; 284:1372. 31. Hunter EB, Johnston PE, Tanner G, Pinson CW, Awad JA. Bromfenac(duract)-associated hepatic failure requiring liver transplantation. Am J Gastroenterol 1999; 1999:2299–301. 32. Fontana RJ, McCashland TM, Benner KG, et al. Acute liver failure associated with prolonged use of bromfenac leading to liver transplantation. Liver Transpl Surg 1999; 5:480–4.

Clinicopathological Patterns of Drug-Induced Liver Disease

235

33. Moses PL, Schroeder B, Alkhatib O, Ferrentino N, Suppan T, Lidofsky SD. Severe hepatotoxicity associated with bromfenac sodium. Am J Gastroenterol 1999; 94:1393–6. 34. Rabkin JM, Smith MJ, Orloff SL, Corless CL, Stenzel P, Olyaei AJ. Fatal fulminant hepatitis associated with bromfenac use. Ann Pharmacother 1999; 33:945–7. 35. Tarazi EM, Harter JG, Zimmerman HJ, Ishak KG, Eaton RA. Sulindac-associated hepatic injury: analysis of 91 cases reported to the food and drug administration. Gastroenterology 1993; 104:569–74. 36. Banks AT, Zimmerman HJ, Ishak KG, Harter JG. Diclofenac-associated hepatotoxicity: analysis of 180 cases reported to the food and drug administration as adverse reactions. Hepatology 1995; 22:820–7. 37. Fortgang IS, Belitsos PC, Chaisson RE, Moore RD. Hepatomegaly and steatosis in HIV-infected patients receiving nucleoside analog antiretroviral therapy. Am J Gastroenterol 1995; 90:1433–6. 38. Brau N, Leaf HL. Severe hepatitis in three AIDS patients treated with indinavir. Lancet 1997; 349:924–6. 39. Vergis E, Paterson DL, Singh N. Indinavir-associated hepatitis in patients with advanced HIV infection. Int J STD AIDS 1998; 9:53. 40. Sulkowski MS, Thomas DL, Chaisson RE, Moore RD. Hepatotoxicity associated with antiretroviral therapy in adults infected with human immunodeficiency virus and the role of hepatitis C or B virus infection. JAMA 2000; 283:74–80. 41. Sulkowski MS. Hepatotoxicity associated with antiretroviral therapy containing HIV-1 protease inhibitors. Semin Liver Dis 2003; 23:183–94. 42. Rigas B, Rosenfeld LE, Barwick KW, et al. Amiodarone hepatotoxicity: a clinicopathologic study of five patients. Ann Intern Med 1986; 104:348–51. 43. Rinder HM, Love JC, Wexler R. Amiodarone hepatotoxicity. N Engl J Med 1986; 314:318–9. 44. Snir Y, Pick N, Riesenberg K, Yanai-Inbar I, Zirkin H, Schlaeffer F. Fatal hepatic failure due to prolonged amiodarone treatment. J Clin Gastroenterol 1995; 20:265–6. 45. Chang C-C, Petrelli M, Tomashefski JF, McCullough AJ. Severe intrahepatic cholestasis caused by amiodarone toxicity after withdrawal of the drug: a case report and review of the literature. Arch Pathol Lab Med 1999; 123:251–6. 46. Watkins PB, Zimmerman HJ, Knapp MJ, Gracon SI, Lewis KW. Hepatotoxic effects of tacrine administration in patients with Alzheimer’s disease. JAMA 1994; 271:992–8. 47. Becquemont L, Ragueneau I, Le Bot MA, Riche C, Funck-Brentano C, Jaillon P. Influence of the CYP1A2 inhibitor fluvoxamine on tacrine pharmacokinetics in humans. Clin Pharmacol Ther 1997; 61:619–27. 48. Blackard WG, Jr., Sood GK, Crowe DR, Fallon MB. Tacrine: a cause of fatal hepatotoxicity? J Clin Gastroenterol 1998; 26:57–9. 49. Watkins PB, Whitcomb RW. Hepatic dysfunction associated with troglitazone. N Engl J Med 1998; 338:916–7. 50. Gitlin N, Julie NL, Spurr CL, Lim KN, Juarbe HM. Two cases of severe clinical and histologic hepatotoxicity associated with troglitazone. Ann Intern Med 1998; 129:36–8. 51. Fukano M, Amano S, Sato J, et al. Subacute hepatic failure associated with a new antidiabetic agent, troglitazone: a case report with autopsy examination. Hum Pathol 2000; 31:250–3. 52. Murphy EJ, Davern TJ, Shakil AO, et al. Troglitazone-induced fulminant hepatic failure. Dig Dis Sci 2000; 45:549–53. 53. Kohlroser J, Mathai J, Reichheld J, Banner BF, Bonkovsky HL. Hepatotoxicity due to troglitazone: report of two cases and review of adverse events reported to the United States food and drug administration. Am J Gastroenterol 2000; 95:272–6. 54. Bonkovsky HL, Azar R, Bird S, Szabo G, Banner BF. Severe cholestatic hepatitis caused by thiazolidinediones: risks associated with substituting rosiglitazone for troglitazone. Dig Dis Sci 2002; 47:1632–7. 55. Forman LM, Simmons DA, Diamond RH. Hepatic failure in a patient taking rosiglitazone. Ann Intern Med 2000; 132:118–21. 56. Maddrey WC, Boitnott JK. Drug-induced chronic liver disease. Gastroenterology 1977; 72:1348–53. 57. Farrell GC. Drug-induced hepatic injury. J Gastroenterol Hepatol 1997; 12:S242–50. 58. Bourdi M, Larrey D, Nataf J, et al. Anti-liver endoplasmic reticulum autoantibodies are directed against human cytochrome P-450IA2. J Clin Invest 1990; 85:1967–73. 59. Black M, Rabin L, Schatz N. Nitrofurantoin-induced chronic active hepatitis. Ann Intern Med 1980; 92:62–4. 60. Golstein PE, Deviere J, Cremer M. Acute hepatitis and drug-related lupus induced by minocycline treatment. Am J Gastroenterol 1997; 92:143–6. 61. Tamm M, Sieber C. Moxonidine-induced cholestatic hepatitis. Lancet 1997; 350:1822. 62. Bhat G, Jordan J, Jr., Sokalski S, Bajaj V, Marshall R, Berkelhammer C. Minocycline-induced hepatitis with autoimmune features and neutropenia. J Clin Gastroenterol 1998; 27:74–5. 63. Angulo JM, Sigal LH, Espinoza LR. Coexistent minocycline-induced systemic lupus erythematosus and autoimmune hepatitis. Semin Arthritis Rheum 1998; 28:187–92.

236

Maddrey

64. Schlienger RG, Bircher AJ, Meier CR. Minocycline-induced lupus: a systematic review. Dermatology 2000; 200:223–31. 65. Teitelbaum JE, Perez-Atayde AR, Cohen M, Bousvaros A, Jonas MM. Minocycline-related autoimmune hepatitis. Arch Pediatr Adolesc Med 1998; 152:1132–6. 66. Sharp JR, Ishak KG, Zimmerman HJ. Chronic active hepatitis and severe hepatic necrosis associated with nitrofurantoin. Ann Intern Med 1980; 92:14–19. 67. Stricker BHCh, lok APR, Claas FHJ, Van Parys GE, Desmet VJ. Hepatic injury associated with the use of nitrofurans: a clinicopathological study of 52 reported cases. Hepatology 1988; 8:599–606. 68. Schattner A, Von der Walde J, Kozak N, Sokolovskaya N, Knobler H. Nitrofurantoin-induced immune-mediated lung and liver disease. Am J Med Sci 1999; 317:336–40. 69. Reynolds TB, Peters RL, Yamada S. Chronic active and lupoid hepatitis caused by a laxative, oxyphenisatin. N Engl J Med 1971; 285:813–20. 70. Maddrey WC, Boitnott JK. Severe hepatitis from methyldopa. Gastroenterology 1975; 68:351–60. 71. Zimmerman HJ, Lewis JH, Ishak KG, Maddrey WC. Ticrynafen-associated hepatic injury: analysis of 340 cases. Hepatology 1984; 4:315–23. 72. Spielberg SP, Gordon GB, Blake DA, Goldstein DA, Herlong HF. Predisposition to phenytoin hepatotoxicity assessed in vitro. N Engl J Med 1981; 305:722–7. 73. Mullick FG, Ishak KG. Hepatic injury associated with diphenylhydantoin therapy: a clinicopathological study of 20 cases. Am J Clin Pathol 1980; 74:442–52. 74. Gennis MA, Vemuri R, Burns EA, Hill JV, Miller MA, Spielberg SP. Familial occurrence of hypersensitivity to phenytoin. Am J Med 1991; 91:631–4.

13

Histopathology of Drug-Induced Liver Disease Gary C. Kanel

Department of Pathology, Keck School of Medicine, University of Southern California, Los Angeles, California, U.S.A.

INTRODUCTION As the liver is the major site for drug metabolism, it is not surprising that drug toxicity and adverse drug reactions would incite variable functional, histological, and ultrastructural hepatic abnormalities (1–14). Up to 10% of patients with abnormal liver tests are found to have drug- or toxin-induced hepatic injury, with the incidence rising to over 40% in patients over the age of 50 (15). Drug-induced liver injury is estimated to occur in up to 1.4% of all hospitalized patients and in from 2% to 5% of hospitalized patients with jaundice, and is responsible for up to 15% to 20% of cases of intrahepatic cholestasis. In addition, up to one-half of all cases of acute liver failure and 20% to 50% of cases of nonviral chronic hepatitis are associated with drug-induced liver damage (11,12,16–22). The type of liver cell injury may be intrinsic and dose dependent, whereby the mechanism may relate either to the formation of free radicals or electrophilic intermediates, or to the production of active oxygen species, which, like free radicals, lead to lipid peroxidation and direct distortion and destruction of liver cell membranes (9,23–29). On the other hand, liver cell damage may be idiosyncratic and dose independent, i.e., dependent on host susceptibility, and may be either immunologically or metabolically mediated, with inflammatory mediators playing a role (12,16,28,30–33). A wide variety of hepatic histological changes have been documented as secondary to drugs and toxins (Table 1) (28,34,35); in addition, up to 1000 drugs and toxins have been implicated in causing these histological changes (36,37). Even herbal medicines are now increasingly recognized as a cause of liver injury (38). Although the morphological features are usually reversible with discontinuation of the medication and toxin exposure, unfortunately, in severe (fulminant) hepatitis and certain forms of chronic hepatitis, discontinuance of the drug does not alleviate the sometimes drastic outcomes. This chapter divides drugs and toxins into the various histological features seen on biopsy. HEPATOCELLULAR INJURY Lobular Necrosis with Minimal to Absent Inflammation This type of liver cell injury is usually related to direct effects of the drug or toxin itself or its metabolites (9,12,24,27–29,39). Unlike drug-induced hypersensitivity reactions, the type of liver cell necrosis can be predicted, and is most often zonal in distribution. Usually, the liver cell injury is coagulative in type, whereby the damaged liver cells become shrunken, with eosinophilic cytoplasm and hyperchromatic nuclei with eventual nuclear pyknosis and karyorrhexis. Although an inflammatory reaction is not characteristic of this type of liver cell injury, a histological response to the necrotic hepatocytes may secondarily occur, with this type of inflammatory reaction both neutrophilic and histiocytic. A zonal nature is often characteristic of specific drugs. Most frequently, the injury is perivenular (zone 3), but other zones may be specifically affected (Table 2; Figs. 1–7). In the more severe cases, bridging confluent necrosis may be seen involving two zones or the entire lobule, and is usually associated with high mortality. Often the borders of the areas of necrosis are sharply defined and distinct from the adjacent viable hepatocytes. The spared liver cells with time may show a hydropic change not representing liver cell injury but instead representing regenerative activity. Sometimes fatty

238

Kanel

TABLE 1 Drug and Toxin-Induced Liver Cell Injury Morphologic Variants Morphology Hepatocellular injury Lobular necrosis with minimal to absent inflammation Zonal: Perivenular (zone 3) Midzonal (zone 2) Periportal (zone 1) Diffuse (confluent) Lobular necrosis with inflammation Lobular confluent necrosis with inflammation Fatty change Granulomas Mallory bodies Cholestatic injury Cholestasis, simple Cholestasis with inflammation Bile duct injury Inflammation by neutrophils Inflammation by lymphocytes, ductopenia (duct loss) Periductal fibrosis Vascular injury Sinusoids Peliosis Dilatation Venoocclusive disease Thrombosis, fibrous obliteration Vasculitis Portal fibrosis Progression to cirrhosis Without progression to cirrhosis Neoplasia Benign Malignant Miscellaneous Inclusions Hepatocytes Reticuloendothelial Pigments Lipochrome Hemosiderin Radiopaque Anthracite Gold

Examples

Carbon tetrachloride, acetaminophen, mushrooms Beryllium, dioxane Phosphorus, ferrous sulfate Halogenated hydrocarbons, mushrooms Isoniazid, phenytoin Niacin, troglitazone Ethanol, methotrexate, corticosteroids Allopurinol, phenylbutazone Amiodarone, ethanol Oral contraceptives, cyclosporin A Isoniazid, carbamazepine Phenytoin Chlorpromazine Floxuridine Anabolic steroids Oral contraceptives Pyrrolizidine alkaloids, cyclophosphamide Ethanol, oral contraceptives Phenylbutazone Methotrexate, ethanol Arsenic, vinyl chloride Oral contraceptives, anabolic steroids Aflatoxins, thorotrast Procainamide Thorotrast Phenacetin Ethanol, iron (oral/parenteral) Thorotrast (Coal miners, city dwellers) Gold sodium thiomalate

Source: From Refs. 14, 28, 34.

change secondary to intrinsic damage may also occur. When there is impediment to bile flow, cholestasis may also be present. Lobular Necrosis with Inflammation As opposed to direct injury, drugs may induce a hypersensitivity reaction (12,28,31,32). Patients with this type of drug-induced injury may exhibit both clinical and histological features of acute hepatitis (Table 3; Figs. 8–17). The portal tracts show an inflammatory infiltrate that is most often lymphocytic, although coexisting eosinophils and sometimes neutrophils may also be seen. The parenchyma shows variable degrees of spotty necrosis either without a zonal distribution pattern, or slight accentuation in zone 3 (perivenular zone) in early-stage disease. The hepatocytes often show both hydropic ballooning change as well as formation of individual cell necrosis (“acidophil” bodies) with an associated usually mononuclear (lymphocytic) inflammatory infiltrate and Kupffer cell hyperplasia. Although cholestasis may also be seen, it usually is not pronounced except in cases of severe hepatitis when there is significant (Text continues on page 243.)

239

Histopathology of Drug-Induced Liver Disease

TABLE 2 Lobular Necrosis with Minimal to Absent Inflammation Perivenular (zone 3) or perivenular and midzonal (zones 2 and 3) Alpha-methyldopa Metoprolol Acetaminophen Mithramycin Aflatoxins Mushrooms Carbon tetrachloride Phalloidin Chloroform Propylthiouracil Copper Pyrrolizidine alkaloids Dimethylnitrosamine Tannic acid Ethionamide Tetrachloroethane Halogenated hydrocarbons Urethane Ketoconazole Midzonal (zone 2) Beryllium Dioxane Ngaione

Periportal (zone 1) Albitocin Allyl formate Endotoxin from Proteus vulgaris Ferrous sulfate Phosphorus

Perivenular (zone 3) or periportal (zone 1) Cocaine

Diffuse confluent Galactosamine Halogenated hydrocarbons Mushrooms 2-Nitropropane Phenelzine Trinitrotoluene

Drugs listed in bold italic font indicate well-documented examples of the toxicity.

FIGURE 1 Lobular necrosis with minimal to absent inflammation: acetaminophen. This low-power photomicrograph shows extensive coagulative necrosis of the perivenular and midzonal regions, with periportal sparing. The portal tract in the center of the field is unremarkable.

240

Kanel

FIGURE 2 Lobular necrosis with minimal to absent inflammation: acetaminophen. High power of a portal tract shows a normal hepatic venule, hepatic arteriole, and single interlobular bile duct. No atypia is seen to any of these components, and virtually no portal inflammatory infiltrate is present.

FIGURE 3 Lobular necrosis with minimal to absent inflammation: acetaminophen. The perivenular zone in the early stage of the disease shows liver cells undergoing coagulative-type necrosis, with shrinkage of the cytoplasm, and pyknosis of the nuclei. No lobular inflammation is seen.

Histopathology of Drug-Induced Liver Disease

241

FIGURE 4 Lobular necrosis with minimal to absent inflammation: acetaminophen. By 6 to 10 days after initial injury, the damaged hepatocytes within the perivenular zone are phagocytized by hypertrophic Kupffer cells and macrophages, as seen in the center of this field. The viable adjacent liver cells show a clearness of the cytoplasm without ballooning change or inflammation, and represent regenerating liver cells.

FIGURE 5 Lobular necrosis with minimal to absent inflammation: cocaine. There is extensive coagulative necrosis involving the periportal hepatocytes. Bile ducts and cholangioles are seen, with no inflammatory infiltrates within the portal tract. The perivenular hepatocytes were spared in this case (not shown in this field).

242

Kanel

FIGURE 6 Lobular necrosis with minimal to absent inflammation: cocaine. The portal tract in the center of the field shows only a minimal lymphocytic infiltrate. The periportal hepatocytes are unremarkable with no lobular inflammation; however, the midzonal hepatocytes at the edges of the field show extensive coagulative necrosis.

FIGURE 7 Lobular necrosis with minimal to absent inflammation: mushrooms. There is extensive confluent necrosis of all the hepatocytes, these cells having dropped out, with red blood cells taking their place in the hepatic cord remnants as well as within the sinusoids. Occasional Kupffer cells and circulating lymphocytes can be seen within the sinusoids, without any lobular inflammation. Some residual fat is still apparent.

Histopathology of Drug-Induced Liver Disease

243

TABLE 3 Lobular Necrosis with Inflammation Alpha-methyldopa Aspirin Benzarone Bupropion Carbutamide Chaparral Chlorpromazine Clarithromycin Clometacin Dantrolene Dapsone Diclofenac Dideoxyinosine Dihydralazine Disulfiram Ethanol Etretinate Fenofibrate

Germander Halogenated hydrocarbons Ibuprofen Indomethacin Isoniazid Jin Bu Huan Ketoconazole Lisinopril Methotrexate Minocycline Naproxen Niacin Nifedipine Nitrofurantoin Oxacillin Oxaprozin Oxyphenisatin Papaverine

Paraaminosalicylic acid Pemoline Perhexiline maleate Phenylbutazone Phenytoin Pirprofen Propylthiouracil Rifampin Sulfadoxine Sulfasalazine Sulfonamides Suloctidil Ticrynafen Toxic oil (rapeseed) Trazodone Troglitazone

Drugs listed in italic font indicate well-documented examples of the toxicity.

impediment to bile flow. Although the histology in many ways is similar to that seen in acute viral hepatitis, the degree of portal infiltrates in drug-induced injury is usually not as striking. In addition, a helpful clue to drug-induced injury is a prominent portal eosinophilic infiltrate, which unfortunately is only seen in a minority of cases of drug-induced liver cell injury. In instances of ongoing necroinflammatory change, a chronic hepatitis may also ensue, with persistently abnormal aminotransferase elevations. The histological features may show progression with time, with variable degrees of periportal inflammatory activity (periportal or (Text continues on page 248.)

FIGURE 8 Lobular necrosis with inflammation: bupropion. The parenchyma shows distortion of the cord sinusoid pattern by variable hydropic change of the hepatocytes and numerous foci of necroinflammatory change, the inflammatory cells chiefly lymphocytic.

244

Kanel

FIGURE 9 Lobular necrosis with inflammation: halogenated hydrocarbons. The portal tract shows a prominent lymphocytic infiltrate, with mild bile duct proliferation.

FIGURE 10

(See color insert) Lobular necrosis with inflammation: halogenated hydrocarbons.

Histopathology of Drug-Induced Liver Disease

245

FIGURE 11 Lobular necrosis with inflammation: nitrofurantoin. This low power image shows diffuse necroinflammatory change and Kupffer cell hyperplasia, the inflammatory cells chiefly lymphocytic.

FIGURE 12 Lobular necrosis with inflammation: nitrofurantoin. This high power image highlights the diffuse necroinflammatory change and Kupffer cell hyperplasia, with the inflammatory cells chiefly lymphocytic.

246

Kanel

FIGURE 13 Lobular necrosis with inflammation: phenytoin. A mild but diffuse necroinflammatory change is seen, with mild hydropic change of the liver cells. A mild increase in circulating lymphocytes within the sinusoids is also present.

FIGURE 14 Lobular necrosis with inflammation: rifampin. Mild lobular necroinflammatory change and Kupffer cell hyperplasia are seen.

Histopathology of Drug-Induced Liver Disease

247

FIGURE 15 Lobular necrosis with inflammation: sertraline. Prominent lymphocytic infiltrates are seen within the sinusoids as well as infiltrating into the hepatic cords. The liver cells show variation in nuclear size. Clusters of pigmentladen macrophages, the pigment representing lipochrome, are also evident within the sinusoids, representing phagocytized debris from damaged liver cells.

FIGURE 16 Lobular necrosis with inflammation: methotrexate. Mild lobular inflammation, mild macrovesicular fatty change, prominent variation in nuclear size, and occasional glycogenated nuclei of hepatocytes are present in this patient on long-term methotrexate therapy.

248

Kanel

FIGURE 17 Lobular necrosis with inflammation: methotrexate. This high power photomicrograph from a patient on long-term methotrexate therapy shows the parenchyma with a mild lymphocytic infiltrate and Kupffer cell hyperplasia. The individual hepatocytes are enlarged, with dysplastic binucleated and trinucleated hepatocytes maintaining a normal nuclear:cytoplasmic ratio. Considerable variation in nuclear size is also seen, this change characteristic of methotrexate-induced injury.

interface hepatitis, “piecemeal” necrosis), portal fibrosis, and bridging fibrosis if the drug is not discontinued (see “Portal Fibrosis” for drugs that may cause a chronic hepatitis with fibrosis). Although cirrhosis may eventually occur, this feature nowadays is uncommon. Lobular Confluent Necrosis with Inflammation The more severe forms of acute hepatitis are associated with significant liver cell necrosis with prominent liver cell dropout and associated collapse of the reticulin framework (confluent or submassive necrosis), and may clinically present as a fulminant hepatitis (Table 4; Figs. 18–26) (40–42). The necrosis usually involves an entire zonal population of cells, most TABLE 4 Lobular Confluent Necrosis with Inflammation Alpha-methyldopa Allopurinol Bromofenac Captopril Carbamazepine Chaparral Chinese herbal medicines Chlordiazepoxide Clarithromycin Cimetidine Dacarbazine Dideoxyinosine Erythromycin Ethacrynic Ethionamide

Germander Gold sodium thiomalate Halogenated hydrocarbons Hydralazine Indomethacin Iprocloziden Isoniazid Ketoconazole Mithramycin Mitomycin C Niacin Nitrofurantoin Pemoline Phenelzine Phenylbutazone

Drugs listed in italic font indicate well-documented examples of the toxicity.

Phenytoin Piroxicam Probenecid Prochlorperazine Propylthiouracil Sulfamethoxazole Sulfasalazine Sulfonamides Ticrynafen Troglitazone Valproic acid

Histopathology of Drug-Induced Liver Disease

249

FIGURE 18 Lobular confluent necrosis with inflammation: halogenated hydrocarbons. Extensive necrosis is seen involving all of the perivenular and midzonal regions on this low-power photomicrograph, with bridging (confluent) necrosis between the different lobules as well.

FIGURE 19 Lobular confluent necrosis with inflammation: halogenated hydrocarbons. Medium power shows a portal tract with a moderate lymphocytic infiltrate. The parenchyma shows striking perivenular and midzonal necrosis and dropout of hepatocytes with a prominent lymphocytic infiltrate; these inflammatory cells are also seen in and amongst the viable periportal hepatocytes.

250

Kanel

FIGURE 20 Lobular confluent necrosis with inflammation: halogenated hydrocarbons. High power of the perivenular and midzones shows total necrosis and dropout of hepatocytes with a prominent lymphocytic infiltrate, Kupffer cell hyperplasia, and histiocytic reaction to the necrotic hepatocytes.

FIGURE 21 Lobular confluent necrosis with inflammation: isoniazid. Low power shows the basic architecture to be intact. Three portal tracts with prominent lymphocytic infiltrates are seen at the edges of the field. The parenchyma, however, shows a lymphocytic infiltrate with total liver cell necrosis and dropout (massive hepatic necrosis). No viable hepatocytes are seen.

Histopathology of Drug-Induced Liver Disease

251

FIGURE 22 Lobular confluent necrosis with inflammation: isoniazid. The portal tract exhibits a moderate lymphocytic infiltrate with mild bile duct proliferation. The adjacent parenchyma shows necroinflammatory change with lymphocytic infiltrates and areas of liver cell dropout. There are still residual hepatocytes seen, enabling possible regeneration and resolution of the liver damage, although the mortality in these cases is still high.

FIGURE 23 Lobular confluent necrosis with inflammation: niacin. The portal tract shows a moderate mixed inflammatory infiltrate consisting of lymphocytes with occasional eosinophils and neutrophils. Mild bile duct proliferation is seen.

252

Kanel

FIGURE 24 Lobular confluent necrosis with inflammation: niacin. The perivenular and midzones show total loss of hepatocytes. A lymphocytic infiltrate can be seen, with occasional small clusters of plasma cells.

FIGURE 25 Lobular confluent necrosis with inflammation: sertraline. The perivenular zone shows total necrosis of hepatocytes, with numerous pigment-laden histiocytes and Kupffer cells present. Acute hemorrhage is also seen involving the sinusoids and the empty hepatic cords.

Histopathology of Drug-Induced Liver Disease

253

FIGURE 26 Lobular confluent necrosis with inflammation: troglitazone. The periportal zone shows total dropout of hepatocytes with collapse of the reticulin framework. The overlying expanded portal tract shows a lymphocytic infiltrate with bile ductular proliferation.

often the perivenular zone (zone 3), although at times the periportal zone (zone 1) may be affected; however, more than one zone is frequently involved. When all the three zones are affected, a panacinar (massive) necrosis is present, associated with an ominous prognosis. A portal and lobular inflammatory component is present and is predominantly lymphocytic with a prominent Kupffer cell reaction. The hepatocytes that are viable show variable and often prominent ballooning degeneration with an accompanying mononuclear inflammatory infiltrate. Regenerative activity may also be seen, although the incidence of recovery is oftentimes meager. In this type of hepatitis, cholestasis in the surviving lobules may be pronounced when associated with impaired regeneration. Fatty Change Fatty change may be due to a number of factors, including [1] inability of the liver cells to excrete synthesized fat due to defective or deficient assembly of the lipid transport moiety apoproteinVLDL (very-low-density-lipoprotein), [2] increased mobilization of lipids from peripheral stores, [3] increased synthesis but decreased oxidation of fatty acids, and [4] mitochondrial dysfunction (24,26,43–46). The type of fat may be either macrovesicular (equal to or larger than the liver cell nucleus) or microvesicular (smaller than the nucleus) (Table 5; Figs. 27–37). A “mixed” pattern may also be seen. In addition, sometimes the microvesicles may be extremely small (foamy change) and difficult to identify on routine histological sections unless the cut sections are thin (1–2 mm). Although sometimes a zonal distribution pattern is seen, often the feature is spotty or diffuse. Macrovesicular fatty change may be the only histological feature present, without accompanying liver cell necrosis, whereby the change is more often incidental without liver cell dysfunction. Pure microvesicular fat, on the other hand, is often associated with liver cell damage even without an inflammatory infiltrate. Liver cell necrosis may also be present with

254

Kanel

TABLE 5 Fatty Change Macrovesicular Acetaminophen Alpha-methyldopa Amanitin Amiodarone Aspirin Azidothymidine (AZT) Bleomycin Borates Cadmium Carbon tetrachloride Chloroform Chromate Cisplatin Clometacin Cocaine Corticosteroids Cyanamide Dantrolene Dichloroethylene Dimethylformamide Ethanol

Ethionine Ethyl bromide Ethyl chloride Etretinate Fialuridine Floxuridine Flurazepam Glucocorticoids Gold sodium thiomalate Halogenated hydrocarbons Hydrazine Ibuprofen Indomethacin Isoniazid L-asparaginase Methimazole Methotrexate Methyl bromide Methyl chloride Methyldichloride Microcycline

Minocycline Mitomycin C Mushrooms Nitrofurantoin Organic solvents Orotic acid Perhexiline maleate Phenylbutazone Phosphorus Rifampin Sulfasalazine Sulindac Tamoxifen Tannic acid Tetrachloroethane Ticrynafen Total parenteral nutrition Trichlorethylene Uranium compounds Warfarin

Microvesicular Aflatoxins Amineptine Amiodarone Antiemetics Aspirin Boric acid Camphor Chlortetracycline Cocaine Demeclocycline Desferrioxamine Didanosine Dideoxyinosine Dimethylformamide Ethanol

Ethionine Fialuridine Hopantenate Hypoglycin A Ibuprofen Ketoprofen Margosa oil Methyl salicylate Mushrooms Oxytetracycline Pennyroyal oil Pentenoic acid Phalloidin Phosphorus Piroxicam

Pirprofen Pyrrolizidine alkaloids Rolitetracycline Tetracycline Thallium compounds Tolmetin Valproic acid Vitamin A a Warfarin

Fatty changeCinflammation (steatohepatitis) Amiodarone Methotrexate Sulfasalazine Naproxen Spironolactone Nifedipine Perhexiline maleate Phospholipidosis Amiodarone Amitriptyline Chloramphenicol Chloroquine Chlorpheniramine

Chlorpromazine Coralgil Gentamycin Perhexiline maleate Promethazine

Sulfamethoxazole-trimethoprim Thioridazine

Drugs listed in italic font indicate well-documented exaples of the toxicity. Fat within sinusoidal stellate cells

a

fatty change, either without an accompanying inflammatory reaction (e.g., mushroom hepatotoxicity) or with a mononuclear and/or neutrophilic infiltrate (e.g., amiodarone hepatotoxicity) (45,47,48). The latter condition is also termed “steatohepatitis” and is sometimes associated with Mallory body deposition. The histological features may then mimic those seen in alcoholic hepatitis, necessitating careful history with appropriate laboratory interpretation. For example, in drug-induced steatohepatitis, the aspartate and alanine aminotransferase (AST (Text continues on page 260.)

Histopathology of Drug-Induced Liver Disease

FIGURE 27

255

Fatty change: ethanol. Diffuse macrovesicular fatty change is present.

FIGURE 28 Fatty change: mushrooms. This low-power photomicrograph shows that virtually all the cells contain macrovesicular fat. Close inspection would also demonstrate in this case severe liver cell dropout (refer to Fig. 7), a feature also seen with mushroom hepatotoxicity.

256

Kanel

FIGURE 29 Fatty change: methotrexate. Most of the hepatocytes contain macrovesicular fat. Microvesicular fat, defined as fat droplets smaller than the liver cell nucleus, is also occasionally present but is not prominent.

FIGURE 30

Fatty change: cocaine. The periportal hepatocytes exhibit predominantly microvesicular fat.

Histopathology of Drug-Induced Liver Disease

FIGURE 31

257

(See color insert) Fatty change: ethanol.

FIGURE 32 Fatty change: tetracycline. Virtually all the hepatocytes contain microvesicular fat. This type of liver cell injury associated with tetracycline is seen virtually only in cases of intravenous administration of the medication.

258

Kanel

FIGURE 33

(See color insert) Fatty change: Tetracycline (Oil Red O).

FIGURE 34

(See color insert) Fatty change: Tetracycline (autofluorescence).

Histopathology of Drug-Induced Liver Disease

259

FIGURE 35 Fatty change with inflammation (steatohepatitis): methotrexate. A moderate degree of predominantly macrovesicular fatty change is present, with some degree of microvesicular fat as well. In addition, there is mild spotty necroinflammatory change with lymphocytic infiltrates.

FIGURE 36 Fatty change with inflammation (steatohepatitis): sulfasalazine. Macrovesicular fatty change is present. Focal necroinflammatory change with a prominent lymphocytic infiltrate is also seen.

260

Kanel

FIGURE 37 Fatty change: vitamin A. The small droplet microvesicular fat is present within the perisinusoidal stellate (Ito) cells. In hypervitaminosis A, these cells contain abundant amounts of the vitamin, which also appears intensely green by autofluorescence of fresh frozen tissue.

and ALT) activities are usually equally elevated or the ALT is greater than the AST (43), in contrast to the increased AST:ALT ratio in alcoholic liver disease. A variant of fatty change involves foamy lipid deposition within hepatocytes (phospholipidosis) (24,48,49). These phospholipids accumulate within the lysosomes due to the inhibition of phospholipase A from lipid hydrolysis. Many of the drugs that are associated with this morphologic feature may also be responsible for Mallory body deposition.

Granulomas Hepatic granulomas are loosely defined as distinct clusters of inflammatory cells, and may be seen more frequently within the lobules in drug-induced injury, although portal tracts may also be involved (Table 6; Figs. 38–41). Granulomas are the result of a cellular immune reaction by the hepatic reticuloendothelial system to a drug or toxin (50). The granulomas may be small, poorly circumscribed, and may contain a mixed inflammatory infiltrate consisting of lymphocytes, histiocytes, neutrophils, and eosinophils (inflammatory type). These granulomas may infrequently contain multinucleated giant cells. Granulomas may also be sharply circumscribed, and composed chiefly of lymphocytes and activated macrophages having clear large irregular nuclei and abundant eosinophilic cytoplasm (epithelioid type), often with multinucleated giant cells. Central necrosis is seldom seen in drug-induced granulomatous necrosis, and coalescence of granulomas, a feature sometimes seen in sarcoidosis or tuberculosis, is uncommon in drug-induced injury; however, more often than not, the histological changes in drug-induced granulomatous hepatitis are indistinguishable from other causes of hepatic granulomas. The diagnosis then rests on exclusion (51).

Histopathology of Drug-Induced Liver Disease

261

TABLE 6 Granulomas Acitretin Allopurinol Alpha-methyldopa Amiodarone Amoxicillin clavulanate Aprindine Aspirin Azapropazone Barium Bacille Calmette-Guerin (BCG) therapy or vaccination Beryllium Carbamazepine Carbutamide Cephalexin Chlorpromazine Chlorpropamide Copper Dapsone Detajmium Diazepam Dideoxyinosine Diltiazem Dimethicone Disopyramide Feprazone Glibenclamide Gold sodium thiomalate

Halogenated hydrocarbons Hydralazine Interferon Isoniazid Mestranol Metahydrin Methimazole Methotrexate Metolazone Mineral oil Nitrofurantoin Nomifensine Norethindrone Norethynodrel Norgestrel Oral contraceptives Oxacillin Oxyphenbutazone Oxyphenisatin Papaverine Penicillin Phenazone Phenprocoumon Phenylbutazone Phenytoin Polyvinyl pyrrolidone Prajmalium

Procarbazine Pronestyl Quinidine Quinine Ranitidine Salicylazosulfapyridine Seatone (green-lipped mussel extract) Silica Succinylsulfathiazole Sulfamethoxazole trimethoprim Sulfasuxidine Sulfadiazine Sulfadimethoxine Sulfadoxine pyrimethamine Sulfanilamide Sulfasalazine Sulfathiazole Sulfonamide Sulfonylurea agents Tacrine Thorotrast Tocainide Tolbutamide Trichlormethiazide Verapamil

Drugs listed in italic font indicate well-documented examples of the toxicity.

FIGURE 38 Granuloma: ethanol. Up to 10% of biopsies in patients with alcoholic liver disease may demonstrate these clusters of histiocytes within the lobule, forming ill-defined granulomas that rarely contain multinucleated giant cells. Some of these histiocytes contain fine droplet fat. The term lipogranuloma is used when the fat within the histiocytes is more prominent.

262

Kanel

FIGURE 39 Granuloma: sulfasalazine. The portal tract shows a well-demarcated granuloma composed of lymphocytes and epithelioid histiocytes without multinucleated giant cells. A moderate lymphocytic infiltrate is also seen within the portal tract.

FIGURE 40 Granuloma: sulfasalazine. The parenchyma shows a granuloma composed of lymphocytes and histiocytes. A moderate degree of fatty change is also seen.

263

Histopathology of Drug-Induced Liver Disease

FIGURE 41 Granuloma: mineral oil. This medium-sized cluster of histiocytes contains an abundant amount of fat (lipogranuloma) and represents accumulation of mineral oil. These lipogranuloma are usually seen within the portal tracts, although occasionally they may also be apparent within the parenchyma.

Mallory Bodies Mallory bodies represent ropey eosinophilic cytoplasmic inclusions within hepatocytes. These are most characteristic of both acute and chronic alcoholic liver injury, although a variety of nonalcoholic liver diseases (such as nonalcoholic steatohepatitis, primary biliary cirrhosis, Wilson’s disease) as well as injury secondary to various drugs and toxins may also be associated with Mallory body deposition (Table 7; Figs. 42 and 43) (24,46,52–54). The Mallory bodies in part represent proliferation and derangement of intermediate filaments, which in part constitute the cytoskeleton of the hepatocyte (55). The Mallory bodies located within the liver cell cytoplasm may appear alone or be associated with an inflammatory component that is usually but not always neutrophilic. The cells containing the Mallory bodies have a tendency to be located within the perivenular zone (zone 3), especially in alcoholic liver disease in the noncirrhotic stage, although exceptions with other drugs do occur. When induced by alcohol, associated sinusoidal collagen deposition and fatty change are the characteristic features. TABLE 7 Mallory Bodies Amiodarone Collidine Coralgil Dideoxyinosine 4,4 0 -Diethylaminoethoxyhexestrol Diethylstilbestrol Diltiazem Estrogens Ethanol Glucocorticoids

Griseofulvin Methotrexate Nicardipine Nifedipine Perhexiline maleate Stilbestrol Tamoxifen Tetracycline Valproic acid Vitamin A

Drugs listed in italic font indicate well-documented examples of the toxicity.

264

Kanel

FIGURE 42 Mallory bodies: ethanol. Many of the liver cells contain a ropey eosinophilic intracytoplasmic material that represents Mallory bodies. The liver cells are also hydropic, and there is a moderate neutrophilic infiltrate with prominent intrasinusoidal collagen deposition; these features are commonly seen in alcoholic hepatitis.

FIGURE 43 Mallory bodies: amiodarone. Many of the hepatocytes contain a ropey eosinophilic intracytoplasmic material that represents Mallory bodies. This photomicrograph also demonstrates intrasinusoidal collagen deposition, mild fatty change, and a neutrophilic infiltrate, and in many ways morphologically resembles alcoholic hepatitis (refer to Fig. 38).

Histopathology of Drug-Induced Liver Disease

265

TABLE 8 Cholestasis, Simple Anabolic steroids Azathioprine Cyclosporin A Cytosine arabinoside Danazol Estrogens Ethchlorvynol Flucloxacillin Fluoxymesterone Fosinopril

Gold sodium thiomalate Infliximab Mestranol Methandrostenolone Methimazole Methyltestosterone Norethindrone Norethynodrel Norgestrel Oral contraceptives

Oxymetholone Piroxicam Prochlorperazine Sulindac Tamoxifen Total parenteral nutrition Warfarin

Drugs listed in italic font indicate well-documented examples of the toxicity.

Although some drugs such as amiodarone may also demonstrate histological features quite similar to alcoholic liver disease, subtle differences can occur; for instance, the Mallory bodies tend to have a periportal (zone 1) accentuation in amiodarone-induced hepatic injury.

CHOLESTATIC INJURY Cholestasis, Simple This form of cholestatic liver cell injury is limited to impaired transport and secretion of bile without an accompanying inflammatory infiltrate or injury to bile ducts (Table 8; Fig. 44) (16,17,56–58). The bile is most often conspicuous in the perivenular zone (zone 3), and is histologically manifested by both an intracytoplasmic and intracanalicular component. Less frequently, the midzone and periportal zone may also be involved. In the latter instance,

FIGURE 44 Cholestasis, simple: oral contraceptives. Bile can be seen within dilated canaliculi. No necrosis or lobular inflammation is present.

266

Kanel

proliferation of cholangioles containing bile concretions may be seen, often associated with a mild neutrophilic infiltrate in and amongt the cholangioles. The interlobular bile ducts are spared. The hepatocytes are histologically uninvolved. Portal inflammatory changes are minimal to absent. What is most important in histological diagnosis is the assessment of the interlobular bile ducts, as one of the most common causes of cholestasis is extrahepatic biliary tract obstruction. The interlobular bile ducts in obstruction are characteristically increased in number, often ectatic, and may show periductal edema, periductal fibrosis, and/or acute cholangitis. In drug-induced cholestatic liver cell injury, the interlobular bile ducts are usually normal, exceptions being the rare examples of direct bile duct damage caused by certain drugs (refer to “Bile Duct Injury”); however, in very early stages of bile duct obstruction, the portal duct changes may be subtle, and other parameters such as imaging studies may be most important in identifying the cause. Cholestasis with Inflammation Cholestatic drug-induced liver cell injury may also be associated with a lobular inflammatory infiltrate (Table 9; Figs. 45–50) (56). The inflammation is usually mild with the cholestatic component more striking. The inflammation is usually composed of mononuclear cells. Cholestasis with inflammation is often enhanced in the perivenular zone (zone 3), although TABLE 9 Cholestasis with Inflammation Alpha-methyldopa Acetaminophen Acetohexamide Allopurinol Amineptine Aminoglutethimide Aminosalicylic acid Amitriptyline Amoxicillin clavulanate Aprindine Atenolol Azathioprine Benoxaprofen Captopril Carbamazepine Carbarsone Carbimazole Carisoprodol Cefadroxil Cefazolin Chlorambucil Chlordiazepoxide Chlorothiazide Chlorpromazine Chlorpropamide Chlortetracycline Chlorthalidone Cimetidine Cisplatin Clarithromycin Clorazepate Cyclosporin A Dacarbazine Dantrolene Diazepam Diclofenac Disopyramide Enalapril

Erythromycin Ethchlorvynol Ethionamide Flucloxacillin Fluoxymesterone Fluphenazine Flurazepam Flutamide Glibenclamide Gold sodium thiomalate Griseofulvin Halogenated hydrocarbons Haloperidol Ibuprofen Imipramine Indomethacin Iodipamide meglumine Isocarboxazid Jin Bu Juan Isoniazid Ketoconazole Meprobamate 6-Mercaptopurine Naproxen Niacin Nifedipine Nitrofurantoin Nomifensine Oxacillin Oxaprozin Oxyphenisatin Papaverine Paraaminosalicylic acid Penicillamine Penicillin Perphenazine Phenobarbital Phenylbutazone

Drugs listed in italic font indicate well-documented examples of the toxicity.

Phenytoin Piperazine Piroxicam Pizotyline Polythiazide Prajmalium Procainamide Prochlorperazine Propoxyphene Propylthiouracil Quinethazone Ranitidine Rifampin Sulfamethoxazole trimethoprim Sulfasalazine Sulfonamides Sulindac Tamoxifen Thiabendazole 6-Thioguanine Thiopental Thioridazine Ticlopidine Tocainide Tolazamide Tolbutamide Total parenteral nutrition Toxic oil (rapeseed) Tranylcypromine Triazolam Trifluoperazine Trimethobenzamide Tripelennamine Troleandomycin Valproic acid Verapamil Zimelidine

Histopathology of Drug-Induced Liver Disease

267

FIGURE 45 Cholestasis with inflammation: allopurinol. The perivenular zone shows mild lobular necrosis and inflammation by lymphocytes. Intracellular bile is also present in some of the hepatocytes, as well as within the Kupffer cells that have phagocytized the damaged hepatocytes. Mild macrovesicular and microvesicular fatty change is also present.

FIGURE 46 Cholestasis with inflammation: chlorpromazine. A predominantly lymphocytic infiltrate can be seen in the perivenular zone, with occasional neutrophils also apparent. Bile is present, mostly intracellular, but can also be seen within occasional slightly dilated canaliculi.

268

Kanel

FIGURE 47 Cholestasis with inflammation: clarithromycin. Many of the cells in this field represent enlarged macrophages and Kupffer cells that contain bile pigment which was phagocytized from damaged liver cells that originally contained intracellular bile. The adjacent viable hepatocytes also contain irregular intracellular clumps of bile pigment. A moderate lymphocytic infiltrate can also be seen within the sinusoids and in the areas of necrosis.

in severe cases the features may be diffused. In contrast to simple cholestasis, cholestasis with inflammation is associated with a portal inflammatory infiltrate, which may be predominantly lymphocytic but also may include eosinophils and neutrophils. Interlobular bile ducts are generally unremarkable and do not show signs of obstruction (e.g., ectasia, periductal edema, or periductal fibrosis).

FIGURE 48 Cholestasis with inflammation: clarithromycin. Dilated cholangioles at the edge of a portal tract contain abundant bile. The adjacent portal tract shows a mild lymphocytic infiltrate.

Histopathology of Drug-Induced Liver Disease

269

FIGURE 49 Cholestasis with inflammation: ketoconazole. Bile can be seen within dilated canaliculi within the parenchyma, as well as within dilated cholangioles located at the border of the portal tract and the parenchyma. A mild lobular lymphocytic infiltrate with Kupffer cell hyperplasia is present, as is a mild lymphocytic infiltrate within the adjacent portal tract.

FIGURE 50 Cholestasis with inflammation: niacin. The perivenular zone shows a mild lymphocytic infiltrate with Kupffer cell hyperplasia. Occasional dilated canaliculi containing bile are also present.

270

Kanel

BILE DUCT INJURY Interlobular bile ducts may show histological damage, evident by variable hydropic change of the cytoplasm, nuclear irregularity with pyknosis, and individual cell necrosis (17,59,60). There is most often an accompanying inflammatory reaction oriented to the bile ducts (Table 10; Figs. 51–54), and this may be neutrophilic (acute cholangitis) or lymphocytic (nonsuppurative cholangitis). Persistence of duct inflammation and damage may eventually lead to duct loss (ductopenia) (17,59). When ductopenia occurs, other etiologies for duct loss, such as primary biliary cirrhosis, primary sclerosing cholangitis, and autoimmune hepatitis (autoimmune cholangitis) must be considered. An uncommon bile duct change induced by floxuridine in the treatment of hepatic tumors (61–63) is periductal fibrosis (biliary sclerosis), which histologically mimics bile duct obstruction and primary sclerosing cholangitis. VASCULAR INJURY A number of different vascular hepatic lesions may be seen in drug-induced injury (Table 11; Fig. 55) (64). Weakening and damage to the lobular reticulin network can cause microcystic lobular lesions that fill with red blood cells, these small cysts being usually but not always devoid of an endothelial lining (peliosis hepatis) (59,65–68). Sometimes these lesions may be large enough to be visualized on imaging; rarely they may rupture with intra-abdominal hemorrhage. Sinusoidal dilatation may be induced in livers that also demonstrate peliosis, but also may be seen in the periportal or midzones secondary to oral contraceptive usage (59,69–70). Venoocclusive disease (sinusoidal obstruction syndrome) is associated with endothelial damage of the terminal hepatic venules and perivenular sinusoids, with endothelial loss, intraluminal occlusion, and variable perivenular sinusoidal collagen deposition (71). It is seen in conjunction with exposure to pyrrolizidine alkaloids, but also is seen in patients after bone marrow transplantation on various conditioning regimens (71–74). Hepatic vein thrombosis has been reported in patients on oral contraceptives (75). Chronic alcoholic liver disease shows characteristic fibrous obliteration of the terminal hepatic venules and sublobular veins due to the activation of stellate cells (76–78). Arteritis seldom directly involves the arterioles, and usually the small arteries are also spared. Medium-sized vessels are most often affected; hence, TABLE 10 Bile Ducts: Inflammation and Injury Inflammation by neutrophils Allopurinol Hydralazine Chlorpromazine Sulindac Chlorpropamide Phenytoin Flucloxacillin Inflammation by lymphocytes, ductopenia Acetaminophen Cimetidine Ajmaline (alkaloid isolated from Clindamycin Rauwolfia serpentina) Cromolyn Allopurinol Cyproheptadine Amineptine Diazepam Amitriptyline Dicloxacillin Amoxicillin-clavulanate Erythromycin Ampicillin Flucloxacillin Arsenic Haloperidol Azathioprine Imipramine Carbamazepine Methylenediamine Carbutamide Methyltestosterone Chlorpromazine Paraquat Chlorpropamide Phenylbutazone Chlorthiazide Phenytoin Periductal fibrosis Floxuridine Drugs listed in italic font indicate well-documented examples of the toxicity.

Piroxicam Prochlorperazine Sporidesmin Sulfamethoxazole-trimethoprim Sulfonurea agents Tetracycline Thiabendazole Tiopronin Tolazemide Tolbutamide Toxic oil (rapeseed) Trifluoperazine Troleandomycin Xenelamine

Histopathology of Drug-Induced Liver Disease

271

FIGURE 51 Bile duct injury: chlorpromazine. The interlobular bile duct shows considerable cytologic atypia, with lymphocytes surrounding and infiltrating into the duct epithelium (nonsuppurative cholangitis).

FIGURE 52 Bile duct injury: chlorpropamide. The interlobular bile ducts show cytologic atypia and are surrounded by mononuclear cells. The duct at the left of the field also shows considerable nuclear irregularity. The portal inflammatory infiltrate consists of lymphocytes with scattered eosinophils.

272

FIGURE 53

Kanel

(See color insert) Bile duct injury: Chlorpropamide.

FIGURE 54 Bile duct injury: chlorpropamide. The portal tract shows bile duct loss (ductopenia) and a mixed inflammatory infiltrate with numerous eosinophils.

273

Histopathology of Drug-Induced Liver Disease

TABLE 11

Vascular

Sinusoids Peliotic lesions Anabolic steroids Arsenic Azathioprine Busulfan Danazol Diethylstilbesterol Estrone Fluoxymesterone Glucocorticoids Hydroxyprogesterone Hydroxyurea

Medroxyprogesterone 6-Mercaptopurine Methandrostenolone Methotrexate Methyltestosterone Norethandrolone Oxymetholone Phalloidin Steroids, endogenous production (adrenal tumor) Tamoxifen

Testosterone 6-Thioguanine Thorotrast Vinyl chloride Vitamin A

Dacarbazine Dactinomycin Danazol Daunorubicin Decarbazine Dimethylbusulfan Dimethylnitrosamine Doxorubicin Estramustine Floxuridine Germander Indicine Mate tea

Mechlorethamine 6-Mercaptopurine Mitomycin C Pyrrolizidine alkaloids Tamoxifen 6-Thioguanine Urethane Vinblastine Vincristine Vitamin A

Dilatation Oral contraceptives Metoclopramide Veno-occlusive disease Adriamycin Aflatoxins Anabolic steroids Arsenic Azathioprine Busulfan Carboplatin Carmustine (BCNU) Cisplatin Cyclophosphamide Cysteamine Cytarabine Cytosine arabinoside

Hepatic vein thrombosis/fibrous obliteration Dacarbazine Oral contraceptives Estrogens Total parenteral nutrition Ethanol Vasculitis Allopurinol Chlorothiazide Chlorpropamide Penicillin Phenylbutazone

Phenytoin Sulfonamides

Drugs listed in italic font indicate well-documented examples of the toxicity.

the morphological features can be missed on biopsy unless a larger portal tract is present for evaluation.

PORTAL FIBROSIS Both toxin exposure and long-term usage of certain drugs may cause a chronic liver disease with portal fibrosis that, in some instances, may progress with time to cirrhosis (Table 12; Figs. 56–58). Hepatic stellate cells play a central role in extracellular fibrogenesis with these drugs and toxins. Following injury, the stellate cells undergo morphological and phenotypic alterations resulting in their activation with eventual extracellular matrix deposition (79–81). Knowledge of these drugs with appropriate screening on biopsy has virtually eliminated development of advanced liver disease, an excellent example being methotrexate in the treatment of rheumatoid arthritis. When cirrhosis develops, the subtype may be macronodular

274

Kanel

FIGURE 55 (See color insert ) Vascular, veno-occlusive disease: Busulfan (conditioning regimen for bone marrow transplantion) (trichrome).

(regenerative nodules O3 mm in diameter) or micronodular (regenerative nodules %3 mm in diameter). A biliary pattern (“geographic” appearance of the nodules) has also been described in instances where the primary liver injury is directed towards the bile ducts, with eventual duct depletion (ductopenia). Sometimes portal fibrosis will occur with clinical manifestations of portal hypertension, without eventual progression of the liver disease to the cirrhotic stage (“noncirrhotic portal fibrosis”); when secondary to thorotrast, polyvinyl chloride, or arsenic exposure, this hepatic disorder is also associated with the development of malignant TABLE 12 Portal Fibrosis Cirrhosis, or portal fibrosis with progression to cirrhosis on serial biopsies Alpha-methyldopa Ferrous fumarate Acetaminophen Isoniazid Acetohexamide Lisinopril Amiodarone 6-Mercaptopurine Chlorpromazine Methotrexate Coralgil Methyltestosterone Dantrolene Nitrofurantoin Diclofenac Oxyphenisatin Ethanol Papaverine Etretinate Pemoline Fenofibrate Perhexiline maleate Portal fibrosis without progression to cirrhosis (non-cirrhotic portal fibrosis) Anabolic steroids Thorotrast Arsenic Toxic oil (rapeseed) Azathioprine Vinyl chloride Copper Vitamin A Drugs listed in italic font indicate well-documented examples of the toxicity. Cardiac cirrhosis

a

Phenylbutazone Propylthiouracil Pyrrolizidine alkaloidsa Tamoxifen Thiabendazole Total parenteral nutrition Tricrynafen Valproic acid

Histopathology of Drug-Induced Liver Disease

FIGURE 56

275

Fibrosis: methotrexate. Portal fibrosis with bridging without regenerative nodule formation is present.

FIGURE 57 Fibrosis: methotrexate. This low-power photomicrograph shows thin fibrous septa with small regenerative nodule formation (micronodular cirrhosis) in a patient who was on low dose but daily underwent methotrexate therapy for psoriasis for many years.

276

Kanel

FIGURE 58 Fibrosis: ethanol. This low-power photomicrograph from a liver explant shows a well-established to advanced cirrhosis in an alcoholic patient who had stopped drinking approximately one year before the transplant. No fatty change, sinusoidal collagen, or Mallory bodies are present.

primary tumors [most often angiosarcoma (see “Neoplasia”) and cholangiocarcinoma] (28,35,82,83). NEOPLASIA Certain drugs and toxins may with time induce the formation of both benign and malignant neoplasms (Table 13; Figs. 59–62) (16,28,35,84) which may be single or multiple within the liver. Oral contraceptive usage and liver cell adenomas is a well-known example (85), with case reports of focal nodular hyperplasia associated as well (86). These adenomas may disappear when the drug is discontinued, although more often the mass lesions persist. Liver cell adenomas have also been reported to progress to hepatocellular carcinoma (87), although the aggressiveness of this malignant lesion is questionable, with no known cases having metastasized. Hepatocellular carcinoma and cholangiocarcinoma are both associated with long-term drug or toxin exposure. Angiosarcoma, which is a very rare primary liver cell neoplasm, has a remarkably high incidence in patients with exposure to thorotrast (thorium dioxide, an alpha-, beta-, gamma- emitter used from 1930 to 1953 as an arteriographic agent), arsenic, copper, or polyvinyl chloride (82,88,89). MISCELLANEOUS Inclusions Inclusion bodies secondary to drugs and toxins can be seen within liver cell cytoplasm, liver cell nuclei, and Kupffer cells (Table 14; Figs. 63 and 64). The inclusions may be distinct and well circumscribed (e.g., procainamide usage); however, sometimes the liver cell cytoplasm has a diffuse “ground-glass” appearance, with or without distinct inclusions, as seen with phenobarbital usage and cyanamide exposure. The inclusions are in part secondary to hypertrophy of

277

Histopathology of Drug-Induced Liver Disease

TABLE 13

Neoplasia/Mass Lesions

Benign Liver cell adenoma

Focal nodular hyperplasia

Anabolic steroids Danazol Mestranol Oral contraceptives Toxic oil (rapeseed)

Estrogens

Nodular regenerative hyperplasia Anabolic steroids Azathioprine Busulfan Copper Corticosteroids Dactinomycin Ethanol Oral contraceptives 6-Thioguanine Thorotrast Toxic oil (rapeseed) Vinyl chloride

Malignant Hepatocellular carcinoma

Angiosarcoma

Aflatoxins Anabolic steroids Arsenic Danazol Ethanol Methotrexate Oral contraceptives Thorotrast Vinyl chloride

Anabolic steroids Arsenic Copper Diethylstilbestrol Oral contraceptives Phenelzine Thorotrast Vinyl chloride

Cholangiocarcinoma Alpha-methyldopa Anabolic steroids Isoniazid Oral contraceptives Thorotrast

Drugs listed in italic font indicate well-documented examples of the toxicity.

FIGURE 59 Neoplasm, liver cell adenoma: oral contraceptives. The tumor is composed of cytologically benign hepatocytes having a normal nuclear: cytoplasmic ratio and forming trabecular cords one to two cells thick.

278

Kanel

FIGURE 60 Neoplasm, atypical liver cell adenoma: oxymethalone. The hepatocytes show mild cytologic atypia and form pseudoacinar-type structures often containing bile. The hepatic cords are one to two cells thick.

FIGURE 61 Neoplasm, cholangiocarcinoma: thorotrast. This image is taken from the liver of a patient with thorotrast exposure who developed cholangiocarcinoma. The non-tumor liver seen in this field shows a portal tract with a number of histiocytes containing the finely to coarsely granular thorotrast pigment.

279

Histopathology of Drug-Induced Liver Disease

FIGURE 62 Neoplasm, cholangiocarcinoma: thorotrast. This image is from the same liver seen in Figure 55, and shows a well-differentiated cholangiocarcinoma. An adjacent well-differentiated cholangiocarcinoma is present.

the smooth endoplasmic reticulum (90). Particulate material showing positive birefringence under polarized light may be identified within portal tracts and occasionally within Kupffer cells in long-term intravenous drug users, and represents the injectant used. Pigments Various pigments secondary to drugs and toxins (Table 15; Figs. 65 and 66) may also be identified, and can be confirmed with special stains (Ziehl–Neelsen acid-fast stain to highlight some examples of lipochrome, and Perls’ iron stain to highlight hemosiderin). Although lipochrome is generally most prominent in the elderly population, in some instances this pigment deposition can be enhanced in individuals on certain medications such as phenacetin

TABLE 14

Inclusions

Hepatocytes Cytoplasmic Procainamide Cyanamide Nuclear Lead Ground glass-like hepatocytes

Azathioprine Chlorpromazine Glucocorticoids Phenobarbital Phenytoin

Kupffer Cells, Portal Macrophages Polyvinyl pyrrolidone Silicone rubber (damaged cardiac prosthetic valves) Talc, particulate material a Thorotrast Drugs listed in italic font indicate well-documented examples of the toxicity. a Also extracellular within portal tracts.

280

Kanel

FIGURE 63 Inclusions: procainamide. Almost all the hepatocytes contain well-demarcated round to oval intracytoplasmic eosinophilic inclusions that take up most of the liver cell cytoplasm. Although these inclusions in some way resemble the ground-glass cells seen in chronic hepatitis B virus infection, these inclusions stain negatively on immunoperoxidase stain for the hepatitis B surface antigen.

FIGURE 64

(See color insert) Inclusions: particulate injectant (IV drug users) (polarized light).

281

Histopathology of Drug-Induced Liver Disease

TABLE 15

Pigments

Lipochrome (Hepatocytes) Aminopyrine Aspirin Carbamazepine Cascara sagrada Chenodeoxycholic acid Chlordecone (kepone) Chlorinated biphenyls Chlorpromazine Halogenated hydrocarbons Insecticides Nitrofurantoin Phenacetin Phenothiazines Rifampin Black pigments (Kupffer cells, portal macrophages) Anthracite Gold sodium thiomalate Titanium

Hemosiderin (Hepatocytes, Kupffer cells, portal macrophages)

Radiopaque (Kupffer cells, portal macrophages)

Cimetidine Ethanol Hexachlorobenzene Iron (oral/parenteral)

Thorotrast

Drugs listed in italic font indicate well-documented examples of the toxicity.

(16). Certain pigments may have a tinctorial quality that is unique, such as the grey-green thorotrast pigment, or the dark black granular anthracotic pigment sometimes seen in city dwellers and coal miners; these pigments are most often seen in portal macrophages and Kupffer cells, although anthracotic pigment is more often extracellular. (Text continues on page 289.)

FIGURE 65 Pigment: anthracite. This coarsely granular black pigment is seen within histiocytes as well as free within the portal tracts. This type of pigment can be seen most commonly in long-term city dwellers.

282

FIGURE 66

Kanel

(See color insert) Pigment: Thorotrast.

TABLE 16 Drugs/Toxins Listed in Alphabetical Order with Associated Histologic Changes Table # 2 a

3 4 5

b c d e

a b c d

6 7 8 9 10 a b c 11 a b

Histological changes Lobular necrosis with minimal to absent inflammation Perivenular (zone 3) or perivenular and midzonal (zones 2 and 3) Midzonal (zone 2) Periportal (zone 1) Perivenular (zone 3) or periportal (zone 1) Diffuse confluent Lobular necrosis with inflammation Lobular confluent necrosis with inflammation Fatty change Macrovesicular Microvesicular Plus inflammation (steatohepatitis) Phospholipidosis Granulomas Mallory bodies Cholestasis, simple Cholestasis with inflammation Bile ducts: inflammation/injury Inflammation by neutrophils Inflammation by lymphocytes, ductopenia Periductal fibrosis Vascular Sinusoids, peliotic lesions Sinusoids, dilatation

Table # c d e 12 a b 13 a b c d e f 14 a b c d 15 a b c d

Histological changes Terminal hepatic venules, venoocclusive disease Hepatic vein, thrombosis/fibrous obliteration Arteries/arterioles, vasculitis Portal fibrosis With progression to cirrhosis Without progression to cirrhosis Neoplasia/mass lesions Liver cell adenoma Focal nodular hyperplasia Nodular regenerative hyperplasia Hepatocellular carcinoma Angiosarcoma Cholangiocarcinoma Inclusions Hepatocytes, cytoplasmic Hepatocytes, nuclear Hepatocytes, ground glass-like Kupffer cells/portal macrophages Pigments Lipochrome Hemosiderin Radiopaque Black (Continued)

283

Histopathology of Drug-Induced Liver Disease

TABLE 16

Drugs/Toxins Listed in Alphabetical Order with Associated Histologic Changes (Continued )

Acetaminophen Acetohexamide Acitretin Adriamycin Aflatoxins Ajmaline Allopurinol Allyl formate Alpha-methyldopa Albitocin Amanitin Amineptine Aminoglutethimide Aminopyrine Aminosalicylic acid Amiodarone Amitriptyline Amoxicillin clavulanate Ampicillin Anabolic steroids Anthracite Antiemetics Aprindine Arsenic Aspirin Atenolol Azapropazone Azathioprine Azidothymidine (AZT) Bacille CalmetteGuerin (BCG) Barium Benoxaprofen Benzarone Beryllium Bleomycin Borates Boric acid Bromofenac Bupropion Busulfan Cadmium Camphor Captopril Carbamazepine Carbarsone Carbimazole Carbon tetrachloride Carboplatin Carbutamide Carisoprodol Carmustine (BCNU) Cascara sagrada Cefadroxil Cefazolin Cephalexin Chaparral

2a,5,9,10b,12a 9,12a 6 11c 2a,5b,11c,13d 10b 4,6,9,10a,10b,11e 2c 2a,3,4,5a,6,9,12a,13f 2c 5 5b,9,10b 9 15a 9 5a,5b,5c,5d,6,7,12a 5d,9,10b 6,9,10b 10b 8,11a,11c,12b,13a,13c,13d,13e,13f 15d 5b 6,9 10b,11a,11c,12b,13d,13e 3,5a,5b,6,15a 9 6 8,9,10b,11a,11c,12b,13c,14c 5a 6 6 9 3 2b,6 5a 5a 5b 4 3 11a,11c,13c 5a 5b 4,9 4,6,9,10b,15a 9 9 2a,5a 11c 3,6,10b 9 11c 15a 9 9 6 3,4

Chenodeoxycholic acid Chlorambucil Chloramphenicol Chlordecone (kepone) Chlordiazepoxide Chlorinated biphenyls Chloroform Chloroquine Chlorothiazide Chlorpheniramine Chlorpromazine Chlorpropamide Chlortetracycline Chlorthalidone Chlorothiazide Chromate Cimetidine Cisplatin Clarithromycin Clindamycin Clometacin Clorazepate Cocaine Collidine Copper Coralgil Corticosteroids Cromolyn Cyanamide Cyclophosphamide Cyclosporin A Cyproheptadine Cysteamine Cytarabine Cytosine arabinoside Dacarbazine Dactinomycin Danazol Dantrolene Dapsone Daunorubicin Dacarbazine Demeclocycline Desferrioxamine Detajmium Diazepam Dichloroethylene Diclofenac Dicloxacillin Didanosine Dideoxyinosine 4,4 -Diethylaminoethoxyhexestrol Diethylstilbestrol Dihydralazine Diltiazem Dimethicone

15a 9 5d 15a 4,9 15a 2a,5a 5d 9,11e 5d 3,5d,6,9,10a,10b,12a,14c,15a 6,9,10a,10b,11e 5b,9 9 10b 5a 4,9,10b,15b 5a,9,11c 3,4,9 10b 3,5a 9 2d,5a,5b 7 2a,6,12b,13c,13e 5d,7,12a 5a,13c 10b 5a,14a 11c 8,9 10b 11c 11c 8,11c 4,9,11c,11d 11c,13c 8,11a,11c,13a,13d 3,5a,9,12a 3,6 11c 11c 5b 5b 6 6,9,10b 5a 3,9,12a 10b 5b 3,4,5b,6,7 7 7,11a,13e 3 6,7 6 (Continued)

284

Kanel

TABLE 16 Drugs/Toxins Listed in Alphabetical Order with Associated Histologic Changes (Continued ) Dimethylbusulfan Dimethylformamide Dimethylnitrosamine Dioxane Disopyramide Disulfiram Doxorubicin Enalapril Endotoxin from Proteus vulgaris Erythromycin Estramustine Estrogens Estrone Ethacrynic Ethanol Ethchlorvynol Ethionamide Ethionine Ethyl bromide Ethyl chloride Etretinate Fenofibrate Feprazone Ferrous fumarate Ferrous sulfate Fialuridine Floxuridine Flucloxacillin Fluoxymesterone Fluphenazine Flurazepam Flutamide Fosinopril Galactosamine Gentamicin Germander Glibenclamide Glucocorticoids Gold sodium thiomalate Griseofulvin Halogenated hydrocarbons Haloperidol Hexachlorobenzene Hopantenate Hydralazine Hydrazine Hydroxyprogesterone Hydroxyurea Hypoglycin A Ibuprofen Imipramine Indicine Indomethacin Infliximab

11c 5a,5b 2a,11c 2b 6,9 3 11c 9 2c 4,9,10b 11c 7,8,11d,13b 11a 4 3,5a,5b,7,11d,12a,13c,13d,15b 8,9 2a,4,9 5a,5b 5a 5a 3,5a,12a 3,12a 6 12a 2c 5a,5b 5a,10c,11c 8,9,10a,10b 8,9,11a 9 5a,9 9 8 2e 5d 3,4,11c 6,9 5a,7,11a,14c 4,5a,6,8,9,15d 7,9 2a,2e,3,4,5a,6,9,15a 9,10b 15b 5b 4,6,10a 5a 11a 11a 5b 3,5a,5b,9 9,10b 11c 3,4,5a,9 8

Insecticides Interferon Iodipamide meglumine Iprocloziden Iron (oral/ parenteral) Isocarboxazid Isoniazid Jin Bu Huan Ketoconazole Ketoprofen l-asparaginase Lead Lisinopril Margosa oil Mate tea Mechlorethamine Medroxyprogesterone Meprobamate 6-Mercaptopurine Mestranol Metahydrin Metolazone Methandrostenolone Methimazole Methotrexate Methyl salicylate Methyl bromide Methyl chloride Methyldichloride Methylenediame Methyltestosterone Metoclopramide Metolazone Metoprolol Microcycline Mineral oil Minocycline Mithramycin Mitomycin C Mushrooms Naproxen Ngaione Niacin Nicardipine Nifedipine Nitrofurantoin 2-Nitropropane Nomifensine Norethandrolone Norethindrone Norethynodrel Norgestrel Oral contraceptives Organic solvents Orotic acid Oxacillin

15a 6 9 4 15b 9 3,4,5a,6,9,12a,13f 3,9 2a,3,4,9 5b 5a 14b 3,12a 5b 11c 11c 11a 9 9,11a,11c,12a 6,8,13a 6 6 8,11a 5a,6,8 3,5a,5c,6,7,11a,12a,13d 5b 5a 5a 5a 10b 8,10b,11a,12a 11b 6 2a 5a 6 3,5a 2a,4 4,5a,11c 2a,2e,5a,5b 3,5c,9 2b 3,4,9 7 3,5c,7,9 3,4,5a,6,9,12a,15a 2e 6,9 11a 6,8 6,8 6,8 6,8,11b,11d,13a,13c,13d,13e,13f 5a 5a 3,6,9 (Continued)

285

Histopathology of Drug-Induced Liver Disease

TABLE 16

Drugs/Toxins Listed in Alphabetical Order with Associated Histologic Changes (Continued )

Oxaprozin Oxymetholone Oxyphenbutazone Oxyphenisatin Oxytetracycline Papaverine Paraaminosalicylic acid Paraquat Pemoline Penicillamine Penicillin Pennyroyal oil Pentenoic acid Perhexiline maleate Perphenazine Phalloidin Phenacetin Phenazone Phenelzine Phenobarbital Phenothiazines Phenprocoumon Phenylbutazone Phenytoin Phosphorus Piperazine Piroxicam Pirprofen Pizotyline Polythiazide Polyvinyl pyrrolidone Prajmalium Probenecid Procainamide Procarbazine Prochlorperazine Promethazine Pronestyl Propoxyphene Propylthiouracil Pyrrolizidine alkaloids Quinethazone Quinidine Quinine Ranitidine Rifampin Rolitetracycline Salicylazosulfapyridine Seatone Silica Silicone rubber Spironolactone Sporidesmin Steroids, endogenous production

3,9 8,11a 6 3,6,9,12a 5b 3,6,9,12a 3,9 10b 3,4,12a 9 6,9,11e 5b 5b 3,5a,5c,5d,7,12a 9 2a,5b,11a 15a 6 2e,4,13e 9,14c 15a 6 3,4,5a,6,9,10b,11e,12a 3,4,6,9,10a,10b,11e,14c 2c,5a,5b 9 4,5b,8,9,10b 3,5b 9 9 6,14d 6,9 4 6,9,14a 6 4,8,9,10b 5d 6 9 2a,3,4,9,12a 2a,5b,11c,12a 9 6 6 6,9 3,5a,9,15a 5b 6 6 6 14d 5c 10b 11a

Stilbestrol Succinylsulfathiazole Sulfadoxine Sulfamethoxazole Sulfamethoxazole trimethoprim Sulfasalazine Sulfasuxidine Sulfonamides Sulfonurea agents Sulindac Suloctidil Sulfadiazine Sulfadimethoxine Sulfadoxine pyrimethamine Sulfanilamide Sulfasalazine Sulfathiazole Sulfonylurea agents Tacrine Talc, particulate material Tamoxifen Tannic acid Testosterone Tetrachloroethane Tetracycline Thallium compounds Thiabendazole 6-Thioguanine Thiopental Thioridazine Thorotrast Ticlopidine Ticrynafen Tiopronin Titanium Tocainide Tolazamide Tolbutamide Tolmetin Total parenteral nutrition Toxic oil (rapeseed) Tranylcypromine Trazodone Triazolam Trichloroethylene Trichlormethiazide Trifluoperazine Trimethobenzamide Trinitrotoluene Tripelennamine Troglitazone Troleandomycin

7 6 3 4 5d,6,9,10 3,4,5a,5c,9 6 3,4,9,11e 10b 5a,8,9,10a 3 6 6 6 6 6 6 6 6 14d 5a,7,8,9,11a,11c,12a 2a,5a 11a 2a,5a 5b,7,10b 5b 9,10b,12a 9,11a,11c,13c 9 5d,9 6,11a,12b,13c,13d,13e,13f,14d,15c 9 3,4,5a,12a 10b 15d 6,9 9 6,9,10b 5b 5a,8,9,11d,12a 3,9,10b,12b,13a,13c 9 3 9 5a 6 9,10b 9 2e 9 3,4 9,10b (Continued)

286

Kanel

TABLE 16 Drugs/Toxins Listed in Alphabetical Order with Associated Histologic Changes (Continued ) Uranium compounds Urethane Valproic acid Verapamil Vinblastine

5a 2a,11c 4,5b,7,9,12a 6,9 11c

Vincristine Vinyl chloride Vitamin A Warfarin Xenylamine Zimelidine

11c 11a,12b,13c,13d,13e 5b,7,11a,11c,12b 5a,5b,8 10b 9

The table # refers to the previously listed tables in this chapter; in instances where more than one histological feature occurs for any drug or toxin, when appropriate the feature or features more frequently seen are listed in bold.

SUMMARY The histological changes seen in the liver in drug-induced and toxic hepatic injury are complex. A whole spectrum of morphological changes are observed, and, unfortunately, with rare exceptions (e.g., demonstration of thorotrast pigment), no histological features are diagnostic. In fact, in the case of hypersensitivity-induced injury, it is common that various subtypes of morphological injury may occur for the same drug in different patients. For instance, in isoniazid-induced injury, an acute hepatitis-like picture usually occurs, manifested by lobular necrosis and inflammation (Table 3); however, in some cases, cholestasis with inflammation (Table 9) may be seen, and in the severe cases, lobular confluent necrosis and inflammation (Table 4) may occur; therefore, isoniazid can be seen listed in these three separate tables. Table 16 is therefore an added table, listing in alphabetical order all the drugs and toxins in Tables 1–15, noting all the morphologic changes that can be seen associated with that particular drug or toxin, but also highlighting the feature or features most commonly seen. The distinction in the vast majority of cases rests upon eliminating other causes of liver disease, as no reliable approach outside of discontinuing the medication and observing improvement of liver tests is feasible. Usually, the degree of active liver disease manifested by monitoring of hepatic function will resolve within one to two weeks in hepatitic reactions but may take months in cholestatic reactions, although in some instances the abnormal liver tests may persist for considerable time periods with either. In instances of hypersensitivity reactions, rechallenging the patient will demonstrate conclusively the diagnosis, but should be approached cautiously to avoid serious liver cell injury. REFERENCES 1. Larrey D. Drug-induced liver diseases. J Hepatol 2000; 32(Suppl.):77–88. 2. Amacher DE. Serum transaminase elevations as indicators of hepatic injury following the administration of drugs. Regul Toxicol Pharmacol 1998; 27:119–30. 3. Lee WM. Drug-induced hepatotoxicity. N Engl J Med 1995; 333:1118–27. 4. Ishak KG, Zimmerman HJ. Morphologic spectrum of drug-induced liver disease. Gastroenterol Clin North Am 1995; 24:759–86. 5. Hall P de la M. Histopathology of drug-induced liver disease. In: Farrell GC, ed. Drug-Induced Liver Diseases. Edinburgh: Churchill Livingston, 1994:115–51. 6. Black M. Drug-induced liver disease. Clin Liver Dis 1998; 2:457–647. 7. Lee WM. Drug-induced hepatotoxicity. N Engl J Med 2003; 349:474–85. 8. Maddrey WC. Drug-induced hepatotoxicity: 2005. J Clin Gastroenterol 2005; 39(Suppl. 2):S83–9. 9. Park K, Williams DP, Naisbitt DJ, et al. Investigation of toxic metabolites during drug development. Toxicol Appl Pharmacol 2005; 207 (2 Suppl.): 425–34. 10. Novak D, Lewis JH. Drug-induced liver disease. Curr Opin Gastroenterol 2003; 19:203–15. 11. Meier Y, Cavallaro M, Roos M, et al. Incidence of drug-induced liver injury in medical inpatients. Eur J Clin Pharmacol 2005; 61:135–43. 12. Lee WM, Schiodt FV. Drug-induced liver disease. In: Yamada T, Alpers DH, Kaplowitz N, Laine L, Owyang C, Powell DW, eds. Textbook of Gastroenterology. 4th ed. Philadelphia, PA: Lippincott Williams & Wilkins, 2003:2352–65. 13. Jaeschke H, Gores GJ, Cederbaum AI, et al. Mechanisms of hepatotoxicity. Toxicol Sci 2002; 65:166–76. 14. Kaplowitz N. Drug metabolism and hepatotoxicity. In: Kaplowitz N, ed. Liver and Biliary Diseases. 2nd ed. Baltimore, MD: Williams & Wilkins, 1996:103–20. 15. Benhamos JP. Drug-induced hepatitis: clinical aspects. In: Fillastre JP, ed. Hepatotoxicity of Drugs. Rouen: Universite´ de Rouen, 1985:22–30.

Histopathology of Drug-Induced Liver Disease

287

16. Zimmerman HJ. Hepatotoxicity: The Adverse Effects of Drugs and Other Chemicals on the Liver. Philadelphia, PA: Lippincott Williams & Wilkins, 1999. 17. Zimmerman HJ, Lewis JH. Drug-induced cholestasis. Med Toxicol 1987; 2:112–60. 18. Matzen P, Malchow-Moller A, Hilden J, et al. Computer icterus group: differential diagnosis of jaundice: a pocket diagnostic chart. Liver 1984; 4:360–71. 19. Dossing M, Sonne J. Drug-induced hepatic disorders. Incidence, management and avoidance. Drug Saf 1993; 9:441–9. 20. Devictor D, Desplanques L, Debray D, et al. Emergency liver transplantation for fulminant liver failure in infants and children. Hepatology 1992; 16:1156–62. 21. Hoofnagle JH, Carithers RL, Jr., Shapiro C, et al. Fulminant hepatic failure: summary of a workshop. Hepatology 1995; 21:240–52. 22. Lewis JH, Zimmerman HJ. Drug-induced liver disease. Med Clin North Am 1989; 73:775–92. 23. Zimmerman HJ, Ishak KG. Hepatic injury due to drugs and toxins. In: MacSween RB, Anthony P, Scheuer P, Burt AD, Portmann BC, eds. Pathology of the Liver. 3rd ed. Edinburgh: Churchill Livingston, 1994:563–633. 24. Zimmerman HJ. Drug-induced liver disease. In: Schiff ER, Sorrell MF, Maddrey WC, eds. Schiff’s Diseases of the Liver. 8th ed. Philadelphia, PA: Lippincott-Raven, 1999:973–1064. 25. DeLeve LD, Kaplowitz N. Mechanism of drug-induced liver disease. Gastroenterol Clin North Am 1995; 24:787–810. 26. Pessayre D. Role of reactive metabolites in drug-induced hepatitis. J Hepatol 1995; 23(Suppl. 1):16–24. 27. Park BK, Kitteringham NR, Maggs JL, et al. The role of metabolic activation in drug-induced hepatotoxicity. Annu Rev Pharmacol Toxicol 2005; 45:177–202. 28. Zimmerman HJ, Ishak KG. Hepatic injury due to drugs and toxins. In: MacSween RNM, Burt AD, Portmann BC, Ishak KG, Scheuer PJ, Anthony PP, eds. Pathology of the Liver. 4th ed. London: Churchill Livingstone, 2002:621–709. 29. Lazerow SK, Abdi MS, Lewis JH. Drug-induced liver disease 2004. Curr Opin Gastroenterol 2005; 21:283–92. 30. Neuberger J. Immune mechanisms in drug hepatotoxicity. Clin Liver Dis 1998; 2:471–82. 31. Watkins PB. Idiosyncratic liver injury: challenges and approaches. Toxicol Pathol 2005; 33:1–5. 32. Ju C. Immunological mechanisms of drug-induced liver injury. Curr Opin Drug Discov Devel 2005; 8:38–43. 33. Ganey PE, Luyendyk JP, Maddox JF, et al. Adverse hepatic drug reactions: inflammatory episodes as consequence and contributor. Chem Biol Interact 2004; 150:35–51. 34. Kanel GC. Histopathology of drug-induced liver disease. In: Kaplowitz N, DeLeve LD, eds. DrugInduced Liver Disease. New York, NY: Marcel Dekker Inc., 2003:243–86. 35. Kanel GC, Korula J. Atlas of Liver Pathology. 2nd ed. Philadelphia, PA: Elsevier, 2005:78–115. 36. Farrell GC. Drug-induced chronic active hepatitis. In: Farrell GC, ed. Drug-Induced Liver Disease. Edinburgh: Churchill Livingston, 1994:413–30. 37. Zimmerman HJ. Update of hepatotoxicity due to classes of drugs in common clinical use: nonsteroidal drugs, anti-inflammatory drugs, antibiotics, antihypertensives, and cardiac and psychotropic agents. Semin Liver Dis 1990; 10:322–38. 38. Stedman C. Herbal hepatotoxicity. Semin Liver Dis 2002; 22:195–206. 39. Bromer MQ, Black M. Acetaminophen hepatotoxicity. Clin Liver Dis 2003; 7:351–67. 40. Bjornsson E, Olsson R. Suspected drug-induced liver fatalities reported to the WHO database. Dig Liver Dis 2006; 38:33–8. 41. Chojkier M. Troglitazone and liver injury: in search of answers. Hepatology 2005; 41:237–46. 42. Schiodt FV, Lee WM. Fulminant liver disease. Clin Liver Dis 2003; 7:331–49. 43. Dianzani MU. Toxic liver injury by protein synthesis inhibitors. Prog Liver Dis 1976; 5:232–45. 44. Pessayre D, Mansouri A, Haouzi D, et al. Hepatotoxicity due to mitochondrial dysfunction. Cell Biol Toxicol 1999; 15:367–73. 45. Stravitz RT, Sanyal AJ. Drug-induced steatohepatitis. Clin Liver Dis 2003; 7:435–51. 46. Lefkowitch JH. Morphology of alcoholic liver disease. Clin Liver Dis 2005; 9:37–53. 47. Lewis JH, Ranard RC, Caruso A, et al. Amiodarone hepatotoxicity: prevalence and clinicopathologic correlations among 104 patients. Hepatology 1989; 9:679–85. 48. Poucell S, Ireton J, Valencia-Mayoral P, et al. Amiodarone associated phospholipidosis and fibrosis of the liver: light, immunohistochemical and electron microscopic studies. Gastroenterology 1984; 86:926–36. 49. Lullman H, Lullman-Rauch R. Drug induced lysosomal storage disease of the liver. In: Fillastre W, ed. Hepatotoxicity of Drugs. Rouen: Universite´ de Rouen, 1986:127–37. 50. Kanel GC, Reynolds TB. Hepatic granulomas. In: Kaplowitz N, ed. Liver and Biliary Diseases. 2nd ed. Baltimore, MD: Williams & Wilkins, 1996:455–62. 51. Ishak KG, Zimmerman HJ. Drug-induced and toxic granulomatous hepatitis. Baillieres Clin Gastroenterol 1988; 2:463–80. 52. French SW. The Mallory body: structure, composition and pathogenesis. Hepatology 1981; 1:76–83.

288

Kanel

53. Kanel GC. Hepatic lesions resembling alcoholic liver disease. In: Ferrell LD, ed. Diagnostic Problems in Liver Pathology. Philadelphia, PA: Hanley & Belfus Inc., 1994:77–104. 54. Zetterman RK. Nonalcoholic steatohepatitis. In: Schiff ER, Sorrell MF, Maddrey WC, eds. Schiff’s Diseases of the Liver. 8th ed. Philadelphia, PA: Lippincott-Raven, 1999:1179–83. 55. Phillips MJ, Poucell S, Patterson J, Valencia P. The Liver: An Atlas and Text of Ultrastructural Pathology. New York, NY: Raven Press, 1987:393–446. 56. Simon FR. Drug-induced cholestasis: pathobiology and clinical features. Clin Liver Dis 1998; 2:483–99. 57. Levy C, Lindor KD. Drug-induced cholestasis. Clin Liver Dis 2003; 7:311–30. 58. Mohi-ud-din R, Lewis JH. Drug- and chemical-induced cholestasis. Clin Liver Dis 2004; 8:95–132. 59. Ishak KG. The liver. In: Riddell RH, ed. Pathology of Drug-Induced and Toxic Diseases. New York, NY: Churchill Livingston, 1982:457–513. 60. Guebel AP, Sempoux C, Rahier J. Bile duct disorders. Clin Liver Dis 2003; 7:295–309. 61. Ludwig J, Kim CH, Wiesner GH, et al. Floxuridine-induced sclerosing cholangitis: an ischemic cholangiopathy. Hepatology 1989; 9:215–8. 62. Kemeny MM, Battifora H, Blayney DW, et al. Sclerosing cholangitis after continuous hepatic artery infusion of FUDR. Ann Surg 1985; 202:176–81. 63. Aldrighetti L, Arru M, Ronzoni M, et al. Extrahepatic biliary stenosis after hepatic arterial infusion (HAI) of floxuridine (FudR) for liver metastases from colorectal cancer. Hepatogastroenterology 2001; 48:1302–7. 64. Gitlin N. Drug-induced hepatic vascular abnormalities. Clin Liver Dis 1998; 2:591–606. 65. Zafrani ES, Pinaudeau Y, Dhumeaux D. Drug-induced vascular lesions of the liver. Arch Intern Med 1983; 143:495–502. 66. Bagheri SA, Boyer JL. Peliosis hepatis associated with androgenic-anabolic steroid therapy: a severe form of hepatic injury. Ann Intern Med 1974; 81:610–8. 67. Yanoff M, Rawson AJ. Peliosis hepatis: an anatomic study with demonstration of two varieties. Arch Pathol 1964; 77:159–65. 68. Staub PG, Leibowitz CB. Peliosis hepatis associated with oral contraceptive use. Australas Radiol 1996; 40:172–4. 69. Altmann H-W. Drug-induced liver reactions: a morphological approach. Curr Top Pathol 1980; 69:69–142. 70. Winkler K, Poulsen H. Liver disease with periportal sinusoidal dilatation: a possible complication to contraceptive steroids. Scand J Gastroenterol 1975; 10:699–704. 71. DeLeve LD, McCuskey RS, Xiangdong W, et al. Characterization of a reproducible rat model of hepatic veno-occlusive disease. Hepatology 1999; 29:1779–91. 72. Shulman HM, Fisher LB, Schoch HG, et al. Venoocclusive disease of the liver after marrow transplantation: histological correlates of clinical signs and symptoms. Hepatology 1994; 19:1171–80. 73. Lee JL, Gooley T, Bensinger W, et al. Veno-occlusive disease of the liver after busulfan, melphalan, and thiotepa conditioning therapy: incidence, risk factors, and outcome. Biol Blood Marrow Transplant 1999; 5:306–15. 74. Schiano TD. Liver injury from herbs and other botanicals. Clin Liver Dis 1998; 2:607–30. 75. Maddrey WC. Hepatic vein thrombosis (Budd-Chiari syndrome): possible association with the use of oral contraceptives. Semin Liver Dis 1987; 7:32–9. 76. French SW, Nash J, Shitabata P, et al. Pathology of alcoholic liver disease. VA Cooperative Study Group 119. Semin Liver Dis 1993; 13:154–69. 77. Nakano M, Worner TM, Lieber CS. Perivenular fibrosis in alcoholic liver injury: ultrastructural and histologic progression. Gastroenterology 1982; 83:777–85. 78. Goodman ZD, Ishak KG. Occlusive venous lesions in alcoholic liver disease: a study of 200 cases. Gastroenterology 1982; 83:786–96. 79. Patel T. Apoptosis in hepatic pathophysiology. Clin Liver Dis 2000; 4:295–317. 80. Rockey DC. The cell and molecular biology of hepatic fibrogenesis: clinical and therapeutic implications. Clin Liver Dis 2000; 4:319–55. 81. Li D, Friedman SL. Liver fibrogenesis and the role of hepatic stellate cells: new insights and prospects for therapy. J Gastroenterol Hepatol 1999; 14:618–33. 82. Popper H, Thomas LB, Telles NC, et al. Development of hepatic angiosarcoma in man induced by vinyl chloride, thorotrast, and arsenic: comparison with cases of unknown etiology. Am J Pathol 1978; 92:349–76. 83. Almoudarres M, Vega KJ, Trotman BW. Noncirrhotic portal hypertension in the adult: case report and review of the literature. J Assoc Acad Minor Phys 1998; 9:53–5. 84. Bosch FX, Ribes J, Cleries R, et al. Epidemiology of hepatocellular carcinoma. Clin Liver Dis 2005; 9:191–211. 85. Edmondson HA, Henderson B, Benton B. Liver-cell adenomas associated with use of oral contraceptives. N Engl J Med 1976; 294:470–2.

Histopathology of Drug-Induced Liver Disease

289

86. Tajada M, Nerin J, Ruiz MM, et al. Liver adenoma and focal nodular hyperplasia associated with oral contraceptives. Eur J Contracept Reprod Health Care 2001; 6:227–30. 87. Korula J, Yellin A, Kanel G, et al. Hepatocellular carcinoma coexisting with hepatic adenoma: incidental discovery after long-term oral contraceptive use. West J Med 1991; 155:416–8. 88. Falk H. Vinyl chloride-induced hepatic angiosarcoma. Princess Takamatsu Symp 1987; 18:39–46. 89. Centeno JA, Mullick FG, Martinez L, et al. Pathology related to chronic arsenic exposure. Environ Health Perspect 2002; 110(Suppl. 5):883–6. 90. Pamperl H, Gradner W, Frittlich H, et al. Influence of long-term anti-convulsant treatment on liver ultrastructure in man. Liver 1984; 4:294–300.

14

Risk Factors for Drug-Induced Liver Disease Laurie D. DeLeve

Keck School of Medicine, University of Southern California, Los Angeles, California, U.S.A.

INTRODUCTION The number of individuals who develop liver toxicity is very small, even for drugs considered particularly hepatotoxic. The major reason for this is that intrinsic hepatotoxicity, toxicity that predictably occurs at a given dose, is most commonly weeded out during preclinical testing if the toxic dose is close to the therapeutic range (i.e., narrow therapeutic index). Drug overdose is the major setting for intrinsic hepatotoxicity. The most commonly seen form of drug-induced liver injury is idiosyncratic hepatotoxicity. By definition, idiosyncratic hepatotoxicity occurs within the therapeutic dose range, but only in susceptible individuals. This chapter examines some of the factors that lead to this susceptibility. Susceptibility to idiosyncratic hepatotoxicity will depend on genetic or environmental risk factors that determine drug disposition and metabolism, tissue susceptibility to toxicity, and adaptation. Most of the risk factors that have been identified are the ones which modify drug disposition or metabolism. Fewer risk factors have been identified that relate to tissue susceptibility to toxicity. The author is not aware of any risk factors that have been identified, which modify the ability of the hepatocyte to adapt. The best understood form of idiosyncratic toxicity occurs in the normal dose range due to elevated tissue concentrations, i.e., due to changes in drug disposition or metabolism. Elevated liver levels of a toxic drug may be due to decreased intestinal first-pass clearance (see example in section on “Nutrition”), decreased protein binding in the blood (see example in section on “Age”), decreased export from the hepatocyte (see below in paragraph on adaptation), decreased hepatic or renal clearance (see example in section on “Aging”), or changes in metabolism of the drug (see sections on “Genetic and Familial Predisposition,” “Drug Interactions,” and “Alcohol”). Tissue susceptibility to toxicity may be due to variety of mechanisms. Increased intracellular concentrations of a toxic parent compound may be due to changes in drug disposition or reduced drug metabolism. Increased concentrations of a toxic metabolite without increased concentrations of the parent metabolite can occur with increased formation or decreased detoxification of the toxic metabolite. Pre-existing injury to a target organelle due to damage from other drugs or liver disease could predispose a tissue to injury. Impairment of liver regeneration will lead to failure to compensate for cell death and thereby increases injury. Mehendale and colleagues have demonstrated the importance of liver regeneration as a determinant of extent of injury using acetaminophen and several experimental hepatotoxins in rodent models [see recent review (1)] and these studies will be discussed in chapter 10. Another novel example of the importance of liver regeneration has recently been described in abstract form for isoniazid (INH) and todralazine: these arylamine derivatives inhibit histone acetylation, which could inhibit liver regeneration (2). Adaptation is a process which has long been surmised to occur based on the observation that liver test abnormalities can spontaneously normalize in some patients. The normalization of liver tests suggests that the cell adapts to the drug by mechanisms which reduce cell death. Possible mechanisms might include downregulation of metabolic activation of the drug, upregulation of drug detoxification, increased cellular repair processes, decreased uptake, or enhanced export of drugs. There is experimental evidence for some of these mechanisms. Dr. DeLeve is a consultant for Johnson & Johnson, Ono USA, and Wyeth-Ayerst.

292

DeLeve

TABLE 1 Risk Factors for Drug-Induced Liver Disease Age Sex Genetic and familial predisposition Drug interactions Cross-reactivity Alcohol abuse Nutritional status Underlying liver disease Other diseases

Repeated exposure to increasing doses of acetaminophen protects mice from subsequent lethal dosing by decreased P450 activation and increased glutathione detoxification (3). Toxic doses of acetaminophen or carbon tetrachloride decrease expression of various basolateral uptake transporters and increase expression of genes involved in drug efflux into bile and blood in mice (4,5). It remains to be demonstrated that these changes in transporters do indeed confer protection. The observation that drug-induced liver test abnormalities can normalize in one individual but progress to liver injury in another indicates the interindividual variation in adaptation, but at present, we do not know what determines who will adapt and who will not. At this point in time, we have limited ability to predict who is at risk for liver injury. Conversely, no risk factor can be identified in many cases of drug-induced liver disease. If the currently known risk factors are so poorly predictive of toxicity, what is the value of knowing risk factors? Despite our limited ability to predict who will develop toxicity, physicians should be aware of risk factors and avoid prescribing drugs to at-risk patients. For the consultant who attempts to uncover the cause of liver test abnormalities or liver injury, the presence of risk factors for drug injury will increase the likelihood that a drug is the cause of the injury. The remainder of the chapter will review specific risk factors that have been identified (Table 1). AGE Around 20% of hepatitis is drug induced in the geriatric population, as compared with 2% to 5% for all age groups (6). There are several reasons why drug toxicity is more common with aging. The elderly use more medication than younger individuals. According to the National Health and Nutrition Examination Survey (NHANES) III data collected between 1988 and 1994, 39% of adults ages 19 to 64 use prescription drugs versus 74% of those 65 years and older (7). The elderly are also much more likely to take multiple medications and the incidence of adverse reactions rises exponentially with polypharmacy (8–11). According to the NHANES III data, 19% of individuals ages 19 to 64 use two or more medications. In individuals 65 to 74 years of age, 51% use two or more medications and 12% use five or more prescription drugs. The major concern about the use of multiple drugs is the increased risk of drug interactions. The increased number of medications may also reflect more severe illnesses, such as renal insufficiency, which might predispose to drug toxicity. With increasing age, the likelihood of previous exposure to a drug increases and repeated exposure predisposes to immunoallergic reactions. There is also an age-related impairment of drug disposition and metabolism, which will be reviewed in the following paragraphs. Some of the increased risks described in the elderly may be spurious: for individuals who are on multiple medications, there is an increased likelihood that disease symptoms will falsely be attributed to a drug. With aging, there is a gradual decline in renal function (12) and hepatic function that may contribute to drug accumulation. The age-dependent decline in hepatic blood flow has been estimated to be 20% to 50% (13–16) and the decrease in liver mass 20% to 30% (13). Serum albumin and a-1-acid glycoprotein, the two major blood carrier proteins of drugs, decline with age (17,18). There is an age-dependent increase in body fat that peaks between ages 40 and 60 and a decrease in muscle mass with age that continues to decline in the very aged (19–23). These changes in body composition will alter the volume of distribution of drugs depending on their tissue distribution.

Risk Factors

293

Pregnane X receptor (PXR), farnesoid X receptor (FXR), and constitutive androgen receptor (CAR) are nuclear hormone receptors that regulate the transcription of genes for many of the enzymes involved in drug metabolism. All three of these nuclear hormone receptors are obligate partners with retinoid X receptor (RXRa). RXRa protein levels decline in senescent rats (24). If RXRa declines in humans, this would have an impact on drug metabolism regulated by PXR, FXR, and CAR, notably phase I (e.g., CYP3A4, CYP2B, CYP1A1, CYP2C), phase II [UDP-glucose (UDPG)-transferase and sulfotransferase], and phase III (multidrug resistance 1 and multidrug resistance-associate protein 2) metabolism. Age-related changes in expression of P450 and various conjugating enzymes have been reported in experimental animals and in humans (25,26), but this has not been a consistent observation (27,28). An in-depth review of the literature (29) suggests that there is an age-related decrease in clearance of CYP3A substrates and some studies suggest that the age-related decrease is more likely to occur in men. Capillarization is the loss of fenestrae in sinusoidal endothelial cells with increased basement membrane formation in the space of Disse (30). Capillarization precedes fibrosis in humans and in experimental animal models. With aging, there is a loss of sinusoidal endothelial cell fenestration, thickening of sinusoidal endothelial cells, and some formation of basement membrane, as observed in rats, dogs, and humans. Because the formation of basement membrane is not as extensive as may occur in fibrosis-associated capillarization, this aging-related capillarization has been termed pseudocapillarization (31–33). The effect on drug metabolism of this incomplete degree of capillarization with aging has not been studied, but capillarization in cirrhosis impairs drug metabolism. Capillarization impedes oxygen diffusion and therefore reduces oxygen transfer to the hepatocyte (32,34,35). Decreased hepatocyte oxygenation reduces oxidative metabolism (e.g., P450 metabolism of propranolol), but does not alter conjugation reactions. Loss of fenestrae creates a physical barrier to drug clearance by the liver. Fenestration allows the passage of protein-bound drug into the space of Disse. This allows the liver to take up a larger fraction of drug than the original free (non-protein bound) fraction entering the afferent blood. When drug enters the space of Disse, there is uptake of the free fraction, re-equilibration of free and protein bound, and then uptake of more of the newly formed free fraction during a single pass through the liver. This allows more of the drug to be cleared than just the free fraction during a single pass through the liver. When the sinusoidal endothelial cells lose fenestrae, there is a barrier to the passage of protein-bound drug into the space of Disse and uptake is restricted to the free fraction in blood. In summary, aging-related early capillarization would be predicted to reduce oxidative drug metabolism and hepatic drug clearance as has been shown in fibrosis-related capillarization. Treatment for tuberculosis with INH, rifampin (Rif), or pyrazinamide (PZA) has a high incidence of elevated serum transaminases and of frank liver injury. Older patients have an increased incidence of elevated serum transaminases and hepatotoxicity from anti-tubercular regimens, but it is unclear which drug(s) is the source of the increased risk in the elderly (36–40). One study of 1000 patients treated with INH found that serum transaminases more than five times the upper limit of normal were more frequent with age (39), whereas another study of 519 patients found an association between age and PZA-induced but not with INHinduced elevations of serum transaminases (37). Benoxaprofen was withdrawn from the market because of cholestatic hepatitis and renal failure, particularly in the elderly, that lead to fatalities (41–46). Benoxaprofen is excreted in bile and urine as the glucuronide conjugate. The parenchymal dysfunction was generally not severe. Liver toxicity may therefore not be the ultimate cause of death, but cholestasis may have contributed to accumulation of drug. The age-related risk may be due to decline in renal function, as the half-life in a population with an average age of 82 was almost four times longer than in a population with an average age half of that (47). Impaired clearance in the elderly due to renal insufficiency and drug-induced cholestasis may have predisposed to accumulation of benoxaprofen and benoxaprofen glucuronide. Valproate (VPA)-induced microvesicular steatosis, often with necrosis, has a high case fatality rate (48). The major risk factors are age less than 2, as well as polypharmacy, certain metabolic defects, and psychomotor retardation. Most cases occur in adolescents and children over age 2, but this reflects a greater absolute number of patients treated with VPA in this age group (49). In the 1980s, the incidence of hepatic fatalities was reported to be w1/500 to 1/800

294

DeLeve

in the high-risk group of patients between age 0 and 2 on multiple anticonvulsants and with mental impairment as compared to an incidence of w1/45,000 to 1/49,000 in patients over age 2 on monotherapy (50,51). There was a nearly fivefold decrease in hepatic fatality rates subsequently, due to a more widespread awareness of risk factors that lead to a change in practice (51). SEX Women use drugs more often than men (7,52). In the age group of 19- to 64-year old, 14% of males and 24% of females use two or more prescriptions: a female-to-male ratio of 1.7 (7). In two large studies, 61% to 66% of drug-induced liver injury occurred in women (53,54), which is approximately a 1.5-fold increase of women over men. Thus, in adults, the sex distribution of prescription drug use and of drug-induced liver toxicity is comparable. According to the NHANES III data, the gap in prescription use by women and men begins to appear after puberty (7) and the sex difference in incidence of drug-induced liver injury seems to appear after age 20 (53). Taken together, these data suggest that increased exposure to prescription medication may account for much of the overall increased frequency of drug-induced liver injury in women. Although exposure to drugs may account for much of the elevated risk in females, there remains some sex-dependent increase that is not due to increased exposure. Some of the increased risk in females may be due to differences in volume of distribution and drug metabolizing enzymes, but only for the dose-dependent responses (55,56). There are some sex-dependent differences that are specific for certain drugs. Halothane hepatotoxicity is 1.5- to 2-fold more common in females than in males (57–61). Diclofenac is used more commonly by women, but even when that is factored in there is a twofold increased risk of hepatotoxicity in females that is not seen for nonsteroidal anti-inflammatory drugs (NSAIDs) as a class (62). On the other hand, amoxicillin–clavulanate hepatotoxicity is more common in males (63,64). One particular type of liver injury which is more common in females is drug-induced chronic hepatitis. This form of liver injury shares several features with autoimmune hepatitis, notably female preponderance, presence of autoantibodies (anti-nuclear antibody, anti-smooth muscle antibody) and hyperglobulinemia, and response to steroids. Chronic hepatitis can occur with minocycline, a-methyldopa, diclofenac, nitrofurantoin, and oxyphenisatin. GENETIC AND FAMILIAL PREDISPOSITION One of the greatest challenges that lies before us in the field of drug-induced liver injury is identifying who among the few are at risk. The ultimate goal will be to be able to use a drug in the vast majority of patients who are not at risk for hepatotoxicity with exclusion of the rare individual who is at risk. Currently, the best hope for this is the development of pharmacogenomic or metabolomic characterization of individuals that is affordable and clinically practical. Genomic, proteomic, and metabolomic approaches have been utilized using various drugs and experimental toxins in laboratory animals to identify patterns that will predict risk for hepatotoxicity (65–74). As discussed in chapter 15, “omic” approaches are being used for cancer chemotherapy, but additional development will be needed before they can be practically applied on a routine basis to predict an individual’s risk of drug-induced liver disease. The more immediate application of these “omic” approaches is likely to be in the drug development phase. There is a genetic predisposition to liver injury from phenytoin, carbamazepine, and phenobarbital (75,76). All three drugs are anti-seizure medications that are metabolized by P450 to arene oxides. In vitro studies using a murine metabolic activating system for the drugs combined with lymphocytes from patients who incur liver toxicity from one or more of these compounds suggest that these patients inadequately detoxify the arene oxide metabolites. Family members of the patients also demonstrate increased toxicity to their lymphocytes in the in vitro assay, suggesting that this may be a genetic defect. The genetic detoxification defect is not specific for any one of these three drugs. Of 10 patients, 7 demonstrated toxicity to all three

Risk Factors

295

medications and 40 of 50 patients demonstrated susceptibility of their lymphocytes in vitro to all three medications. At this point, no specific mutations have been identified that might lead to this detoxification defect. Using the same approach described in the preceding paragraph, these investigators also demonstrated a detoxification defect to sulfonamide in patients with liver injury and their family members (77,78). Arylamines such as sulfonamides are predominantly metabolized by N-acetylation and to a lesser degree metabolized by P450 to a hydroxylamine. In slow acetylators, metabolism shifts towards formation of the hydroxylamine, which is further metabolized to the more toxic nitrososulfamethoxazole metabolite. Ninety percent of patients who develop sulfonamide toxicity are slow metabolizers (78,79). Thus, slow acetylator status is likely necessary to develop sulfonamide toxicity. However, since half of all Caucasians and Blacks are slow acetylators but only 0.1% of the population who are exposed to sulfonamides develop toxicity, slow acetylator status is not sufficient to cause sulfonamide toxicity. Slow acetylator status may need to be combined with a defect in detoxification of the nitrososulfamethoxazole metabolite (80). Chlorpromazine can induce cholestatic hepatitis that usually resolves with prompt discontinuation of the drug, but that can progress to chronic cholestasis with the vanishing bile duct syndrome. Metabolism is through sulfoxidation and hydroxylation. Sulfoxide metabolites are not toxic, whereas hydroxylated and quinone-free radical metabolites are toxic metabolites. Chlorpromazine metabolism was examined in patients with a history of chlorpromazine-induced cholestasis (81). The 12 patients were all poor sulfoxidizers and extensive debrisoquine hydroxylators, which should shift metabolism towards the more toxic metabolites. However, 20% to 25% of the population are poor metabolizers and the majority of the population are extensive debrisoquine hydroxylators, so that poor sulfoxidation and extensive debrisoquine hydroxylation cannot account for the entire risk of developing chlorpromazine-induced cholestasis.

DRUG INTERACTIONS A drug may increase the risk of hepatotoxicity from another drug by inducing formation of a toxic metabolite, by inhibiting detoxification, or by competing for detoxification. Induction of the formation of a toxic metabolite can occur in patients on chronic treatment with INH, which will induce cytochrome P4502E1 (CYP2E1). CYP2E1 is the major isozyme responsible for the formation of N-acetyl-para-benzoquinoneimine (NAPQI) from acetaminophen. Co-administration of acetaminophen and INH will result in competition for the active site of CYP2E1 and reduced formation of NAPQI. However, when acetaminophen is given between doses of INH, the drug may interact with CYP2E1 free of INH. If INH has induced CYP2E1, increased amounts of NAPQI will be formed and the threshold for acetaminophen toxicity will be lowered (82). The combination of Rif with INH leads to earlier onset of toxicity and higher incidence of serum transaminase elevation and liver injury than for INH alone. The median time to onset of liver injury is one month for INH alone versus 15 days for INH plus Rif (83). The increased risk of liver injury may be additive rather than synergistic. In a meta-analysis of 34 clinical studies, the incidence of liver injury from INH alone was 1.6%, for Rif alone was 1.1%, and for INH and Rif was 2.6% (84). The postulated mechanism of the interaction is induction of INH metabolism (Fig. 1 in chap. 26). Reactive toxic metabolites are produced by oxidation of hydrazine by microsomal CYP enzymes. Rif is a potent agonist of the nuclear receptor PXR, which regulates multiple pathways of drug metabolism including microsomal CYP. However, pharmacokinetic studies on the effect of Rif on INH metabolism have not clearly established a mechanism that would support a synergistic toxicity. Studies in a slow acetylator and a fast acetylator could not detect an effect of Rif on metabolism of acetylisoniazid or monoacetylhydrazine (85), which are both metabolized by amidase. Other investigators have found that Rif induces amidase with increased direct formation of hydrazine from INH, particularly in slow acetylators (86). VPA toxicity is substantially higher when it is administered in conjunction with other anticonvulsants. There are two major activation pathways of VPA (see chapter 23). One

296

DeLeve

pathway is through P450 activation by CYP2C9 and to a lesser degree by CYP2A6 to 4-ene-VPA and the chemically reactive metabolites, which may inhibit the enzymes of b-oxidation (87,88). The other pathway follows conjugation with coenzyme A, which may lead to depletion of coenzyme A or to inhibition of mitochondrial oxidation by VPA-CoA (89–93). The interaction with other anticonvulsants may be related to their ability to induce P450 metabolism of VPA with increased production of 4-ene-VPA. CYP2C9 is transcriptionally regulated by two constitutive androstane receptor (CAR)-response elements, which bind CAR and PXR nuclear receptors, and a glucocorticoid-response element, which binds glucocorticoid receptors (94–97). Although phenobarbital is a known activator of CAR, its ability to induce CYP2C9 is mediated by PXR (97), which binds to the CAR-response element (94,95,97,98). Phenytoin induces CYP2B and murine CYP2C29 through binding to CAR (99,100) and this raises the possibility that phenytoin might induce CYP2C9 through CAR. However, the basis for the interaction of VPA with other anticonvulsants cannot be confirmed until the mechanism of VPA toxicity is more clearly established. The only consistent abnormality found in VPA toxicity is impairment of b-oxidation (49,101), but studies have not consistently found a correlation between liver dysfunction and formation of VPA metabolites (49,102,103).

CROSS-REACTIVITY A history of hepatotoxicity for a given drug should be considered a significant risk factor for structurally similar drugs. These reactions have been variously described as cross-reactivity, cross-sensitization, or cross-hepatotoxicity. Cross-reactivity may be caused by structural similarity that elicits an immunoallergic reaction or by shared genetic polymorphism in metabolism. Earlier in the chapter, the example was given of the genetic detoxification defect that predisposes to cross-reactivity to the arene oxide anticonvulsants, phenobarbital, phenytoin, and carbamazepine. Similarly, the genetic detoxification defect for sulfonamides would lead to cross-reactivity for various sulfa drugs. There are case reports of cross-reactivity between angiotensin-converting enzyme inhibitors (104), but this does not always occur (105). Lack of cross-reactivity in some patients may reflect the existence of two mechanisms of toxicity: an eicosanoid-mediated mechanism (105) and an immunoallergic form (106). Cross-reactivity has been described for erythromycin salts (107,108), ampicillin and cefuroxime (109), the furans nitrofurantoin and furazolidone (110), the haloalkane anesthetics (111–113) (see chapter 22 for an in-depth discussion), the proprionic NSAIDs naproxen and fenoprofen (114), the tricyclic antidepressants amineptine and clomipramine (115), and between the tricyclic antidepressants trimipramine and desipramine and the phenothiazine cyamemazine (116).

ALCOHOL ABUSE Heavy alcohol use can lower the threshold for acetaminophen toxicity through induction of CYP2E1 and by decreased detoxification by glutathione. CYP2E1 is the major isozyme responsible for formation of NAPQI, the toxic metabolite of acetaminophen. Metabolism of acetaminophen is largely through glucuronidation and sulfation, but an increase of the small fraction metabolized by CYP2E1 will increase the hepatotoxic potential. The interaction between alcohol and acetaminophen via CYP2E1 is a complex one. Heavy alcohol intake induces CYP2E1 leading to increased formation of NAPQI, but induction will disappear over time after discontinuing alcohol use. If alcohol is present at the time of acetaminophen ingestion then alcohol will also compete for CYP2E1. For alcohol to increase the risk of acetaminophen through its effect on CYP2E1, acetaminophen would need to be ingested after blood alcohol levels have declined but before CYP2E1 induction disappears. The second interaction is thought to be through reduction of mitochondrial glutathione. Ethanol prevents mitochondrial uptake of glutathione from the cytosol, leading to selective depletion of mitochondrial glutathione (117,118). Selective depletion of mitochondrial glutathione detoxification and subsequent toxicity to mitochondria is a key event in acetaminophen toxicity (119). There is

Risk Factors

297

only a narrow window of time when CYP2E1 is induced but not occupied by alcohol and mitochondrial glutathione is depleted and this is likely why the interaction of two such commonly ingested compounds does not frequently lead to liver injury (120,121). Nevertheless, there are over 100 case reports of this interaction (122,123) and prescribers should reduce the dose of acetaminophen in individuals with heavy alcohol intake. The literature sometimes states that alcoholism is a risk factor for anti-tuberculosis treatment. Large studies have not found a significant increase in liver injury in alcoholics treated with a regimen of INH and Rif (37,40,124). Serum transaminase elevations may be more common in alcoholics (124), but overt liver injury and the need for dosage adjustment may be less common in alcoholics (40). Heavy alcohol use is described as a risk factor for methotrexate. The other risk factors that are described are obesity and diabetes. One of the most common histological patterns ascribed to methotrexate toxicity shares the same features as (non) alcoholic steatohepatitis with or without fibrosis. This raises the question of whether (non) alcoholic steatohepatitis related to excess alcohol ingestion, diabetes, or obesity predisposes to methotrexate toxicity (125) or whether cases of (non) alcoholic steatohepatitis may be falsely ascribed to methotrexate. NUTRITIONAL STATUS CYP3A and p-glycoprotein work synergistically in the small intestine to metabolize certain drugs, in the so-called “drug efflux-metabolism alliance,” (126) which is a major determinant of intestinal first-pass metabolism. A CYP3A metabolized drug diffuses into the epithelium of the small intestine and a certain fraction will be metabolized by CYP3A. A fraction of the unmetabolized will be effluxed back into the lumen by p-glycoprotein and undergo another cycle of diffusion into the epithelium and possible metabolism by CYP3A. Grapefruit juice (127–132), pomegranate juice (133), star fruit juice (134), and pomelo juice (135) inhibit CYP3A. The consequence of CYP3A inhibition for drug metabolism has been best studied for grapefruit juice. Grapefruit juice increases oral availability of certain drugs by inhibiting CYP3A metabolism in the intestine (128,132). In addition, grapefruit juice inhibits p-glycoprotein-mediated drug transport of some drugs, which would synergize with the effect on CYP3A (136). A concrete example of decreased intestinal first-pass metabolism leading to liver disease would be a patient on cyclosporin A who develops cyclosporin A-induced cholestasis after ingesting sufficient grapefruit juice to increase cyclosporin A uptake. The inhibitory activity of grapefruit juice is due to multiple components present in grapefruit juice (136,137). Fasting reduces the threshold for acetaminophen toxicity in experimental animals (138) and in humans (122,139). There are three mechanisms that may contribute to the effect of fasting. Fasting depletes hepatic stores of glycogen, which in turn leads to loss of UDPG. Decreased UDPG decreases the ability of the liver to replenish UDP-glucuronic acid stores that have been consumed by conjugation with toxic concentrations of acetaminophen (140). Fasting also depletes hepatic glutathione (138). Thus, fasting reduces detoxification by two of the most critical pathways: glucuronidation, which is the high-capacity conjugating pathway, and glutathione, which is critical for detoxification of NAPQI. In addition, fasting induces CYP2E1 in rodents (141–143), although this may not be the case in humans. The widespread use of acetaminophen and the relatively small number of case reports suggest that fasting as a risk factor must work in conjunction with other factors. UNDERLYING LIVER DISEASE Any insult to the liver, including drug-induced liver injury, may precipitate liver failure in patients with chronic liver disease that borders on decompensation. However, there are few examples where underlying liver disease predisposes to drug-induced liver disease. There are several examples of drugs that reportedly have increased risk of liver test abnormalities in patients with hepatitis C. There are two problems interpreting these case reports or small case series. First, liver test abnormalities fluctuate in hepatitis C, so that the use of a drug at the peak

298

DeLeve

TABLE 2 Studies Evaluating Liver Disease as a Risk Factor for Anti-tuberculosis Drugs Risk factor Hepatitis

Drug(s)

Positive association with PZA, no association with INH or Rif Baseline aspartate aminotransferase INH/Rif/PZA and alanine aminotransferase elevations; hepatitis B virus and hepatitis C virus not significant risk factor HbeAgC INH HbsAgC/HbeAgK INH/Rif/EMBGPZA

Outcome Moderate hepatotoxicity

References (37)

Serum transaminase elevations and/or symptomatic hepatotoxicity

(147)

Symptomatic hepatotoxicity Liver test abnormalities

(148) (149)

Abbreviations: HbeAg, hepatitis B e antigen; HbsAg, hepatitis B surface antigen; INH, isoniazid; PZA, pyrazinamide; Rif, rifampin.

of virus-induced liver test abnormalities can falsely lead to the belief that the drug is the cause of the abnormalities. Second, if there is one case report in the literature suggesting such an association, it becomes a precedent for a second case report. A study by Chalasani and colleagues provides an excellent example of how to avoid such ascertainment bias (144). The purpose of the study was to examine whether patients with baseline liver test abnormalities have greater risk of statin-induced hepatotoxicity than patients with normal baseline liver test abnormalities. The inclusion of an untreated cohort demonstrated that the liver test abnormalities in the statin-treated group did not exceed that in the untreated cohort. In patients with nonalcoholic steatohepatitis, pre-existing liver test abnormalities did not predispose patients to liver injury from rosiglitazone (145,146). Viral hepatitis and baseline liver test abnormalities have been examined as risk factors for anti-tuberculosis treatment with varying findings. Studies such as those cited in Table 2 suggest that viral hepatitis and/or liver test abnormalities might be a risk, but this conclusion is undermined by the contrasting findings from these studies (Table 2). Some of the most clear-cut data on liver disease as a risk factor for drug-induced liver disease come from studies on highly active antiretroviral therapy. Chapter 27 in this book reviews viral hepatitis, fibrosis, and cirrhosis as risk factors for liver injury from HIV/AIDS medication. OTHER DISEASES HIV/AIDS predisposes to cotrimoxazole hepatotoxicity with an incidence of around 20% (150,151). Studies do not support the concept that acetylator status contributes to the risk in this population (152,153). One postulated mechanism relates to the glutathione deficiency in this population, as glutathione prevents oxidation of hydroxylamine to the more toxic nitrososulfamethoxazole metabolite (154–157). An alternate mechanism may be that the altered immune function predisposes to hypersensitivity (158). HIV/AIDS may also predispose to liver test abnormalities in patients treated with anti-tuberculosis agents (159). Juvenile rheumatoid arthritis (160–163) and systemic lupus erythematosus (164) are often cited as risk factors for salicylate-induced hepatotoxicity. However, high doses of salicylates have been used to treat these disorders and the hepatotoxicity is related to serum levels of salicylate (160–163) and duration of therapy (165). The incidence of hepatotoxicity was found to be higher in patients with active versus quiescent systemic lupus erythematosus (164), but salicylate levels were not reported for these two patient populations. Disease activity was not a significant factor for patients with juvenile rheumatoid arthritis (161). The relative risk for patients with rheumatoid arthritis was reported to be 11-fold higher than for patients with osteoarthritis when the number of prescriptions was taken into account, but the prescribed dose was not considered (62). It has been suggested that rheumatoid arthritis might be a risk factor for sulfasalazine, based on studies which found that three-quarters of the cases of sulfasalazine-induced hepatotoxicity occurred in patients with rheumatoid arthritis and only one-quarter in patients with inflammatory bowel disease (53,166). However, consumption data of sulfasalazine for the

Risk Factors

299

two patient populations were not reported, so it remains unclear whether this is due to increased susceptibility conferred by rheumatoid arthritis or to a higher number of patients at risk. The risk of diclofenac hepatotoxicity is twice as high in patients with osteoarthritis compared to patients with rheumatoid arthritis (167). At present, it is unclear why osteoarthritis might predispose to this. As described earlier in this chapter, diabetes has been described as a risk factor for methotrexate toxicity, but nonalcoholic steatohepatitis confounds this relationship (see section on Alcohol). REFERENCES 1. Mehendale HM. Tissue repair: an important determinant of final outcome of toxicant-induced injury. Toxicol Pathol 2005; 33:41–51. 2. Murata K, Sugimoto K, Hamada M. A novel mechanism for drug-induced liver failure: inhibition of histone acetylation by arylamine derivatives. Gastroenterology 2006; 130:A749. 3. Shayiq RM, Roberts DW, Rothstein K, et al. Repeat exposure to incremental doses of acetaminophen provides protection against acetaminophen-induced lethality in mice: an explantation for high acetaminophen dosage in humans without hepatic injury. Hepatology 1999; 29:451–63. 4. Aleksunes LM, Scheffer GL, Jakowski AB, Pruimboom-Brees IM, Manautou JE. Coordinated expression of multidrug resistance-associated proteins (Mrps) in mouse liver during toxicantinduced injury. Toxicol Sci 2006; 89:370–9. 5. Aleksunes LM, Slitt AM, Cherrington NJ, Thibodeau MS, Klaassen CD, Manautou JE. Differential expression of hepatic transporter genes in response to acetaminophen and carbon tetrachloride. Toxicol Sci 2005; 83:44–52. 6. Schenker S, Bay M. Drug disposition and hepatotoxicity in the elderly. J Clin Gastroenterol 1994; 18:232–7. 7. Anonymous. National health and nutrition examination survey (NHANES III): Patterns of prescription drug use in the United States, 1988-1994. In; 1998, http//www.cdc.gov/nchs/ data/nhanes/databriefs/preuse.pdf. 8. Kellaway GSM, McCrae E. Intensive monitoring for adverse drug effects in patients discharged from acute medical wards. NZ Med J 1973; 78:525–8. 9. May FE, Stewart RB, Cluff LE. Drug interactions and multiple drug administration. Clin Pharmacol Ther 1977; 22:322–8. 10. Smith JW, Seidl LG, Cluff L. Studies on the epidemiology of adverse drug reactions. Ann Intern Med 1966; 65:629–40. 11. Williamson J, Chopin JM. Adverse reactions to prescribed drugs in the elderly: a multicentre investigation. Age Ageing 1980; 9:73–80. 12. Fehrman-Ekholm I, Skeppholm L. Renal function in the elderly (O70 years old) measured by means of iohexol clearance, serum creatinine, serum urea, and estimated clearance. Scand J Urol Nephrol 2004; 38:73–7. 13. Herrlinger C, Klotz U. Drug metabolism and drug interactions in the elderly. Best Pract Res Clin Gastroenterol 2001; 15:897–918. 14. Zoli M, Magalotti D, Bianchi G, et al. Total and functional hepatic blood flow decrease in parallel with ageing. Age Ageing 1999; 28:29–33. 15. Le Couteur DG, McLean AJ. The aging liver. Drug clearance and an oxygen diffusion barrier hypothesis. Clin Pharmacokinet 1998; 34:359–73. 16. Wynne HA, Cope LH, Mutch E, Rawlins MD, Woodhouse KW, James OF. The effect of age upon liver volume and apparent liver blood flow in healthy man. Hepatology 1989; 9:297–301. 17. Veering BT, Burm AGL, Souverijn JHM, Serree JMP, Spierdijk J. The effect of age on serum concentrations of albumin and a1-acid glycoprotein. Br J Clin Pharmacol 1990; 29:201–6. 18. Lunde AV, Barrett-Connor E, Morton DJ. Serum albumin and bone mineral density in healthy older men and women: the Rancho Bernardo study. Osteoporos Int 1998; 8:547–51. 19. Kyle UG, Genton L, Hans D, et al. Total body mass, fat mass, fat-free mass, and skeletal muscle in older people: cross-sectional differences in 60-year-old persons. J Am Geriatr Soc 2001; 49:1633–40. 20. Vermeulen A, Goemaere S, Kaufman JM. Testosterone, body composition and aging. J Endocrinol Invest 1999; 22:110–6. 21. Borkan GA, Hults DE, Gerzof SG, Burrows BA, Robbins AH. Relationships between computed tomography tissue areas, thicknesses and total body composition. Ann Hum Biol 1983; 10:537–45. 22. Greenblatt DJ, Sellers EM, Shader RI. Drug therapy: drug disposition in old age. N Engl J Med 1982; 306:1081–8.

300

DeLeve

23. Zamboni M, Mazzali G, Zoico E, et al. Health consequences of obesity in the elderly: a review of four unresolved questions. Int J Obes 2005; 29:1011–29. 24. Chao C, Youssef J, Rezaiekhaleigh M, Birnbaum LS, Badr M. Senescence-associated decline in hepatic peroxisomal enzyme activities corresponds with diminished levels of retinoid X receptor alpha, but not peroxisome proliferator-activated receptor alpha. Mech Ageing Dev 2002; 123:1469–76. 25. George J, Byth K, Farrell GC. Age but not gender selectively affects expression of individual cytochrome P450 proteins in human liver. Biochem Pharmacol 1995; 50:727–30. 26. Thomas RP, Guigneaux M, Wood T, Evers BM. Age-associated changes in gene expression patterns in the liver. J Gastrointest Surg 2002; 6:445–53 (discussion 454). 27. Agrawal AK, Shapiro BH. Constitutive and inducible hepatic cytochrome P450 isoforms in senescent male and female rats and response to low-dose phenobarbital. Drug Metab Dispos 2003; 31:612–9. 28. Gorski JC, Vannaprasaht S, Hamman MA, Ambrosius WT, Bruce MA, Haehner-Daniels B, Hall SD. The effect of age, sex, and rifampin administration on intestinal and hepatic cytochrome P450 3A activity. Clin Pharmacol Ther 2003; 74:275–87. 29. Cotreau MM, von Moltke LL, Greenblatt DJ. The influence of age and sex on the clearance of cytochrome P450 3A substrates. Clin Pharmacokinet 2005; 44:33–60. 30. Schaffner F, Popper H. Capillarization of hepatic sinusoids in man. Gastroenterology 1963; 44:239–42. 31. Cogger VC, Mross PE, Hosie MJ, Ansselin AD, McLean AJ, Le Couteur DG. The effect of acute oxidative stress on the ultrastructure of the perfused rat liver. Pharmacol Toxicol 2001; 89:306–11. 32. Le Couteur DG, Cogger VC, Markus AM, et al. Pseudocapillarization and associated energy limitation in the aged rat liver. Hepatology 2001; 33:537–43. 33. Le Couteur DG, Fraser R, Cogger VC, McLean AJ. Hepatic pseudocapillarisation and atherosclerosis in ageing. Lancet 2002; 359:1612–15. 34. Le Couteur DG, Hickey H, Harvey PJ, Gready J, McLean AJ. Hepatic artery flow and propranolol metabolism in perfused cirrhotic rat liver. J Pharmacol Exp Ther 1999; 289:1553–8. 35. Harvey PJ, Gready JE, Yin Z, Le Couteur DG, McLean AJ. Acute oxygen supplementation restores markers of hepatocyte energy status and hypoxia in cirrhotic rats. J Pharmacol Exp Ther 2000; 293:641–5. 36. Gro¨nhagen-Riska C, Hellstrom P-E, Fro¨seth B. Predisposing factors in hepatitis induced by isoniazid treatment of tuberculosis. Am Rev Respir Dis 1978; 118:461–6. 37. Schaberg T, Rebhan K, Lode H. Risk factors for side-effects of isoniazid, rifampin and pyrazinamide in patients hospitalized for pulmonary tuberculosis. Eur Respir J 1996; 9:2026–30. 38. van den Brande P, van Steenbergen W, Vervoort G, Demedts M. Aging and hepatotoxicity of isoniazid and rifampin in pulmonary tuberculosis. Am J Respir Crit Care Med 1995; 152:1705–8. 39. Byrd RB, Horn BR, Solomon DA, Griggs GA. Toxic effects of isoniazid in tuberculosis prophylaxis. Role of biochemical monitoring in 1,000 patients. J Am Med Assoc 1979; 241:1239–41. 40. Dossing M, Wilcke JT, Askgaard DS, Nybo B. Liver injury during antituberculosis treatment: an 11-year study. Tuber Lung Dis 1996; 77:335–40. 41. Duthie A, Nicholls A, Freeth M, Moorhead P, Triger D. Fatal cholestatic jaundice in elderly patients taking benoxaprofen [letter]. Br Med J 1982; 285:62. 42. Firth H, Wilcock GK, Esiri M. Side effects of benoxaprofen. Br Med J 1982; 284:1784. 43. Fisher BM, McArthur JD. Side effects of benoxaprofen. Br Med J 1982; 284:1783. 44. Goudie BM, Birnie GF, Wathinson G, MacSween RN, Kissen LH, Cunningham NE. Jaundice associated with the use of benoxaprofen. Lancet 1982; 1:959–1982. 45. Prescott LF, Leslie PJ. Side effects of benoxaprofen. Br Med J 1982; 284:1783. 46. Taggart HM, Alderdice JM. Fatal cholestatic jaundice in elderly patients taking benoxaprofen. Br Med J 1982; 284:1372. 47. Hamdy RC, Murnang B, Perera N, Woodcock K, Koch IM. The pharmacokinetics of benoxaprofen in elderly subjects. Eur J Rheumatol Inflamm 1982; 5:69–75. 48. Zimmerman HJ, Ishak KG. Valproate-induced hepatic injury: analyses of 23 fatal cases. Hepatology 1982; 2:591–7. 49. Konig SA, Siemes H, Blaker F, et al. Severe hepatotoxicity during valproate therapy: an update. Epilepsia 1994; 35:1005–15. 50. Dreifuss FE, Santilli N, Langer DH, Sweeney KP, Moline KA, Menander KB. Valproic acid hepatic fatalities: a retrospective review. Neurology 1987; 37:379–85. 51. Dreifuss FE, Langer DH, Moline KA, Maxwell JE. Valproic acid hepatic fatalities. II. U.S. experience since 1984. Neurology 1989; 39:201–7. 52. Roe CM, McNamara AM, Motheral BR. Gender- and age-related prescription drug use patterns. Ann Pharmacother 2002; 36:30–9.

Risk Factors

301

53. Friis H, Andreasen PB. Drug-induced hepatic injury: an analysis of 1100 cases reported to the Danish Committee on Adverse Drug Reactions between 1978 and 1987. J Int Med 1992; 232:133–8. 54. Pillans PI. Drug associated hepatic reactions in New Zealand: 21 years experience. NZ Med J 1996; 109:315–9. 55. Fletcher CV, Acosta EP, Strykowski JM. Gender differences in human pharmacokinetics and pharmacodynamics. J Adolesc Healthc 1994; 15:619–29. 56. Sagraves R, Raebel MA, Jermain DM, O’Connell MB, Wermeling DP. ACCP White Paper: women as research subjects. Pharmacotherapy 1993; 13:534–42. 57. Carney FMT, Van Dyke RA. Halothane hepatitis: a critical review. Anesth Anal 1972; 51:135–66. 58. Jenner MA, Plummer JL, Cousins MJ. Halothane reductive metabolism in an adult surgical population. Anaesth Intensive Care 1990; 18:395–9. 59. Martinez-Roig A, Cami J, Llorens-Terol J, de la Torre R, Perich F. Acetylation phenotype and hepatotoxicity in the treatment of tuberculosis in children. Pediatrics 1986; 77:912–5. 60. Moult PJA, Sherlock S. Halothane-related hepatitis. A clinical study of twenty-six cases. QJMed 1975; 44:99–114. 61. Walton B, Simpson BR, Strunin L, Doniach D, Perrin J, Appleyard AJ. Unexplained hepatitis following halothane. Br Med J 1976; 1:1171–6. 62. Garcia Rodriguez LA, Williams R, Derby LE, Dean AD, Jick H. Acute liver injury associated with nonsteroidal anti-inflammatory drugs and the role of risk factors. Arch Intern Med 1994; 154:311–6. 63. Larrey D, Vial T, Micaleff A, Babany G, Morichau-Beauchant M, Michel H, Benhamou JP. Hepatitis associated with amoxycillin-clavulanic acid combination: report of 15 cases. Gut 1992; 33:368–71. 64. Gresser U. Amoxicillin–clavulanic acid therapy may be associated with severe side effects-review of the literature. Eur J Med Res 2001; 6:139–49. 65. Welch KD, Wen B, Goodlett DR, et al. Proteomic identification of potential susceptibility factors in drug-induced disease. Chem Res Toxicol 2005; 18:924–33. 66. Welch KD, Reilly TP, Bourdi M, et al. Genomic identification of potential risk factors during acetaminohen-induced liver disease in susceptible and resistant strains of mice. Chem Res Toxicol 2006; 19:223–33. 67. Coen M, Ruepp SU, Lindon JC, et al. Integrated application of transcriptomics and metabonomics yields new insight into the toxicity due to paracetamol in the mouse. J Pharm Biomed Anal 2004; 35:93–105. 68. Coen M, Lenz EM, Nicholson JK, Wilson ID, Pognan F, Lindon JC. An integrated metabonomic investigation of acetaminophen toxicity in the mouse using NMR spectroscopy. Chem Res Toxicol 2003; 16:295–303. 69. Clayton TA, Lindon JC, Cloarec O, et al. Pharmaco-metabonomic phenotyping and personalized drug treatment. Nature 2006; 440:1073–7. 70. Minami K, Saito T, Narahara M, et al. Relationship between hepatic gene expression profiles and hepatotoxicity in five typical hepatotoxicant-administered rats. Toxicol Sci 2005; 87:296–305. 71. Reilly TP, Bourdi M, Brady JN, et al. Expression profiing of acetaminophen liver toxicity in mice using microarray technology. Biochem Biophys Res Commun 2001; 282:321–8. 72. Williams DP, Garcia-Allan C, Hanton G, et al. Time course toxicogenomic profiles in CD-1 mice after nontoxic and nonlethal hepatotoxic paracetamol administration. Chem Res Toxicol 2004; 17:1551–61. 73. Yamamoto T, Kikkawa R, Yamada H, Horii I. Identification of oxidative stress-related proteins for predictive screening of hepatotoxicity using a proteomic approach. J Toxicol Sci 2005; 30:213–27. 74. Devi SS, Mehendale HM. Microarray analysis of thioacetamide-treated type 1 diabetic rats. Toxicol Appl Pharmacol 2006; 212:69–78. 75. Shear NH, Spielberg SP. Anticonvulsant hypersensitivity syndrome. In vitro assessment of risk. J Clin Invest 1988; 82:1826–32. 76. Spielberg SP, Gordon GB, Blake DA, Goldstein DA, Herlong HF. Predisposition to phenytoin hepatotoxicity assessed in vitro. N Engl J Med 1981; 305:722–7. 77. Rieder MJ, Uetrecht J, Shear NH, Cannon M, Miller M, Spielberg SP. Diagnosis of sulfonamide hypersensitivity reactions by in-vitro “rechallenge” with hydroxylamine metabolites. Ann Intern Med 1989; 110:286–9. 78. Shear NH, Spielberg SP, Grant DM, Tang BK, Kalow W. Differences in metabolism of sulfonamides predisposing to idiosyncratic toxicity. Ann Intern Med 1986; 105:179–84. 79. Rieder MJ, Shear NH, Kanee A, Tang BK, Spielberg SP. Prominence of slow acetylator phenotype among patients with sulfonamide hypersensitivity reactions. Clin Pharmacol Ther 1991; 49:13–17. 80. Cribb AE, Miller M, Leeder JS, Hill J, Spielberg SP. Reactions of the nitroso and hydroxylamine metabolites of sulfamethoxazole with reduced glutathione. Implications for idiosyncratic toxicity. Drug Metab Dispos 1991; 19:900–6. 81. Watson RGP, Olomu A, Clements D, Waring RH, Mitchell RH, Elias E. A proposed mechanism for chlorpromazine jaundice-defective hepatic sulphoxidation combined with rapid hydroxylation. J Hepatol 1988; 7:72–8.

302

DeLeve

82. Chien JY, Peter RM, Nolan CM, et al. Influence of polymorphic N-acetyltransferase phenotype on the inhibition and induction of acetaminophen bioactivation with long-term isoniazid. Clin Pharmacol Ther 1997; 61:24–34. 83. Durand F, Jebrak G, Pessayre D, Fournier M, Bernuau J. Hepatotoxicity of antitubercular treatments. Rationale for monitoring liver status. Drug Saf 1996; 15:394–405. 84. Steele MA, Burk RF, DesPrez RM. Toxic hepatitis with isoniazid and rifampin: a meta-analysis. Chest 1991; 99:465–71. 85. Jenner PJ, Ellard GA. Isoniazid-related hepatotoxicity: a study of the effect of rifampicin administration on the metabolism of acetylisoniazid in man. Tubercle 1989; 70:93–101. 86. Sarma GR, Immanuel C, Kailasam S, Narayana AS, Venkatesan P. Rifampin-induced release of hydrazine from isoniazid. A possible cause of hepatitis during treatment of tuberculosis with regimens containing isoniazide and rifampin. Am Rev Respir Dis 1986; 133:1072–5. 87. Rettenmeier AW, Prickett KS, Gordon WP, et al. Studies on the biotransformation in the perfused rat liver of 2-n-propyl-4-pentenoic acid, a metabolite of the antiepileptic drug valproic acid. Evidence for the formation of chemically reactive intermediates. Drug Metab Dispos 1985; 13:81–96. 88. Baillie TA. Metabolic activation of valproic acid and drug-mediated hepatotoxicity. Role of the terminal olefin, 2-n-propyl-4-pentenoic acid. Chem Res Toxicol 1988; 1:195–9. 89. Krahenbuhl S, Mang G, Kupferschmidt H, Meier PJ, Krause M. Plasma and hepatic carnitine and coenzyme A pools in a patient with fatal, valproate induced hepatotoxicity. Gut 1995; 37:140–3. 90. Ponchaut S, van Hoof F, Veitch K. In vitro effects of valproate and valproate metabolites on mitochondrial oxidations. Relevance of CoA sequestration to the observed inhibitions. Biochem Pharmacol 1992; 43:2435–42. 91. Li J, Norwood DL, Mao LF, Schulz H. Mitochondrial metabolism of valproic acid. Biochemistry 1991; 30:388–94. 92. Bjorge SM, Baillie TA. Inhibition of medium-chain fatty acid beta-oxidation in vitro by valproic acid and its unsaturated metabolite, 2-n-propyl-4-pentenoic acid. Biochem Biophys Res Commun 1985; 132:245–52. 93. Millington DS, Bohan TP, Roe CR, Yergey AL, Liberato DJ. Valproylcarnitine: a novel drug metabolite identified by fast atom bombardment and thermospray liquid chromatography–mass spectrometry. Clin Chim Acta 1985; 145:69–76. 94. Gerbal-Chaloin S, Daujat M, Pascussi J-M, Pichard-Garcia L, Vilarem M-J, Maurel P. Transcriptional regulation of CYP2C9 gene. Role of glucocorticoid receptor and constitutive androstane receptor. J Biol Chem 2002; 277:209–17. 95. Ferguson SS, LeCluyse EL, Negishi M, Goldstein JA. Regulation of human CYP2C9 by the constitutive androstane receptor: discovery of a new distal binding site. Mol Pharmacol 2002; 62:737–46. 96. Pascussi JM, Gerbal-Chaloin S, Drocourt L, Maurel P, Vilarem MJ. The expression of CYP2B6, CYP2C9 and CYP3A4 genes: a tangle of networks of nuclear and steroid receptors. Biochim Biophys Acta 2003; 1619:243–53. 97. Chen Y, Ferguson SS, Negishi M, Goldstein JA. Induction of human CYP2C9 by rifampicin, hyperforin, and phenobarbital is mediated by the pregnane X receptor. J Pharmacol Exp Ther 2004; 308:495–501. 98. Chen Y, Kissling G, Negishi M, Goldstein JA. The nuclear receptors constitutive androstane receptor and pregnane X receptor cross-talk with hepatic nuclear factor 4alpha to synergistically activate the human CYP2C9 promoter. J Pharmacol Exp Ther 2005; 314:1125–33. 99. Wang H, Faucette S, Moore R, Sueyoshi T, Negishi M, LeCluyse E. Human constitutive androstane receptor mediates induction of CYP2B6 gene expression by phenytoin. J Biol Chem 2004; 279:29295–301. 100. Jackson JP, Ferguson SS, Moore R, Negishi M, Goldstein JA. The constitutive active/androstane receptor regulates phenytoin induction of Cyp2c29. Mol Pharmacol 2004; 65:1397–404. 101. Siemes H, Nau H, Schultze K, et al. Valproate (VPA) metabolites in various clinical conditions of probable VPA-associated hepatotoxicity. Epilepsia 1993; 34:332–46. 102. Levy RH, Rettenmeier AW, Anderson GD, et al. Effects of polytherapy with phenytoin, carbamazepine, and stiripentol on formation of 4-ene-valproate, a hepatotoxic metabolite of valproic acid. Clin Pharmacol Ther 1990; 48:225–35. 103. Sugimoto T, Muro H, Woo M, Nishida N, Murakami K. Valproate metabolites in high-dose valproate plus phenytoin therapy. Epilepsia 1996; 37:1200–3. 104. Hagley MT, Benak RL, Hulisz DT. Suspected cross-reactivity of enalapril- and captopril-induced hepatotoxicity. Ann Pharmacother 1992; 26:780–1. 105. Zannad F, Marie PY, Aliot E, Trechot P, Gilgenkrantz JM. Cytolytic hepatitis with captopril but not with enalapril. Cardiovasc Drugs Ther 1988; 1:669–70. 106. Alderman CP. Adverse effects of the angiotensinogen-converting enzyme inhbitors. Ann Pharmacother 1996; 30:55–61.

Risk Factors

303

107. Keefe EB, Reis TC, Berland JE. Hepatotoxicity to both erythromycin estolate and erythromycin ethylsuccinate. Dig Dis Sci 1982; 27:701–4. 108. Tolman KG, Sannella JJ, Freston JW. Chemical structure of erythromycin and hepatotoxicity. Ann Intern Med 1974; 81:58–60. ´ , Arhan M, Altiparmak E. Cholestatic attack due to ampicilline and 109. Ko´klu´ S, Yu´ksel O, Faruk Yolcu O cross-reactivity to cefuroxime. Ann Pharmacother 2004; 38:1539–40. 110. Engel JJ, Vogt TR, Wilson DE. Cholestatic hepatitis after administration of furan derivatives. Arch Intern Med 1975; 135:733–5. 111. Joshi PH, Conn HO. The syndrome of methoxyflurane-associated hepatitis. Ann Intern Med 1974; 80:395–401. 112. Lewis JH, Zimmerman HJ, Ishak KG, Mullick FG. Enflurane hepatotoxicity. A clinicopathological study of 24 cases. Ann Intern Med 1983; 98:984–92. 113. Sigurdsen J, Hreidarsson AB, Thiodleifsson B. Enflurane hepatitis. A report of a case with a previous history of halothane hepatitis. Acta Anaesthesiol Scand 1985; 29:495–6. 114. Andrejak M, Davion T, Gineston JL, Capron JP. Cross hepatotoxicity between non-steroidal antiinflammatory drugs. Br Med J 1987; 295:180–1. 115. Larrey D, Rueff B, Pessayre D, et al. Cross hepatotoxicity between tricyclic antidepressants. Gut 1986; 27:726–7. 116. Remy AJ, Larrey D, Pageaux GP, Ribstein J, Ramos J, Michel H. Cross hepatotoxicity between tricyclic antidepressants and phenothiazines. Eur J Gastroenterol Hepatol 1995; 7:373–6. 117. Ferna´ndez-Checa JC, Garcı´a-Ruiz C, Ookhtens M, Kaplowitz N. Impaired uptake of glutathione by hepatic mitochondria from chronic ethanol-fed rats. Tracer kinetic studies in vitro and in vivo and susceptibility to oxidant stress. J Clin Invest 1991; 87:397–405. 118. Ferna´ndez-Checa JC, Hirano T, Tsukamoto H, Kaplowitz N. Mitochondrial glutathione depletion in alcoholic liver disease. Alcohol 1993; 10:469–75. 119. Tirmenstein MA, Nelson SD. Subcellular binding and effects on calcium homeostasis produced by acetaminophen and a nonhepatotoxic regioisomer, 3 0 -hydroxyacetanilide, in mouse liver. J Biol Chem 1989; 264:9814–9. 120. Zhao P, Kalhorn TF, Slattery JT. Selective mitochondrial glutathione depletion by ethanol enhances acetaminophen toxicity in rat liver. Hepatology 2002; 36:326–35. 121. Zhao P, Slattery JT. Effects of ethanol dose and ethanol withdrawal on rat liver mitochondrial glutathione: implication of potentiated acetaminophen toxicity in alcoholics. Drug Metab Dispos 2002; 30:1413–7. 122. Zimmerman HJ, Maddrey WC. Acetaminophen (paracetamol) hepatotoxicity with regular intake of alcohol: analysis of instances of therapeutic misadventure. Hepatology 1995; 22:767–73. 123. Seeff LB, Cuccherini BA, Zimmerman HJ, Adler E, Benjamin SB. Acetaminophen hepatotoxicity in alcoholic: a therapeutic misadventure. Ann Intern Med 1986; 104:399–404. 124. Cross FS, Long MW, Banner AS, Snider DE, Jr. Rifampin-isoniazid therapy of alcoholic and nonalcoholic tuberculous patients in a U.S. Public Health Service cooperative therapy trial. Am Rev Respir Dis 1980; 122:349–53. 125. Langman G, Hall P, Todd G. Role of non-alcoholic steatohepatitis in methotrexate-induced liver injury. J Gastroenterol Hepatol 2001; 16:1395–401. 126. Benet LZ, Cummins CL. The drug efflux-metabolism alliance: biochemical aspects. Adv Drug Deliv Rev 2001; 50:S3–11. 127. Bailey DG, Spence JD, Munoz C, Arnold JMO. Interaction of citrus juices with felodipine and nifedipine. Lancet 1991; 337:268–2691. 128. Ducharme MP, Warbasse LH, Edwards DJ. Disposition of intravenous and oral cyclosporine after administration with grapefruit juice. Clin Pharmacol Ther 1995; 57:485–91. 129. Kupferschmidt HHT, Riem Ha H, Ziegler WH, Meier PJ, Kra¨henbu¨hl S. Interaction between grapefruit juice and midazolam in humans. Clin Pharmacol Ther 1995; 58:20–8. 130. Edwards DJ, Bernier SM. Naringin and naringenin are not the primary CYP3A inhibitors in grapefruit juice. Life Sci 1996; 59:1025–30. 131. Soons PA, Vogels BAPM, Roosemalen MCM, et al. Grapefruit juice and cimetidine inhibit stereoselective metabolism of nitrendipine in humans. Clin Pharmacol Ther 1991; 50:394–403. 132. Lown KS, Bailey DG, Fontana RJ, et al. Grapefruit juice increases felodipine oral availability in humans by decreasing intestinal CYP3A protein expression. J Clin Invest 1997; 99:2545–53. 133. Hidaka M, Okumura M, Fujita K, et al. Effects of pomegranate juice on human cytochrome P450 3A (CYP3A) and carbamazepine pharmacokinetics in rats. Drug Metab Dispos 2005; 33:644–8. 134. Hidaka M, Fujita K, Ogikubo T, et al. Potent inhibition by star fruit of human cytochrome P450 3A (CYP3A) activity. Drug Metab Dispos 2004; 32:581–3. 135. Egashira K, Ohtani H, Itoh S, et al. Inhibitory effects of pomelo on the metabolism of tacrolimus and the activities of CYP3A4 and p-glycoprotein. Drug Metab Dispos 2004; 32:828–33.

304

DeLeve

136. Honda Y, Ushigome F, Koyabu N, et al. Efffects of grapefruit juice and orange juice components on p-glycoprotein- and MRP2-mediated drug efflux. Br J Pharmacol 2004; 143:856–64. 137. Paine MF, Widmer WW, Hart HL, et al. A furanocoumarin-free grapefruit juice establishes furanocoumarins as the mediators of the grapefruit juice–felodipine interaction. Am J Clin Nutr 2006; 83:1097–105. 138. Pessayre D, Wandscheer JC, Cobert B, et al. Additive effects of inducers and fasting on acetaminophen hepatotoxicity. Biochem Pharmacol 1980; 29:2219–23. 139. Whitcomb DC, Block GD. Association of acetaminophen hepatotoxicity with fasting and ethanol use. J Am Med Assoc 1994; 272:1845–50. 140. Price VF, Jollow DJ. Mechanism of decreased acetaminophen glucuronidation in the fasted rat. Biochem Pharmacol 1988; 37:1067–75. 141. Hong JY, Pan JM, Gonzalez FJ, Gelboin HV, Yang CS. The induction of a specific form of cytochrome P-450 (P-450j) by fasting. Biochem Biophys Res Commun 1987; 142:1077–83. 142. Soh Y, Rhee HM, Sohn DH, Song BJ. Immunological detection of CYP2E1 in fresh rat lymphocytes and its pretranslational induction by fasting. Biochem Biophys Res Commun 1996; 227:541–6. 143. Hu Y, Ingelman Sundberg M, Lindros KO. Induction mechanisms of cytochrome P450 2E1 in liver: interplay between ethanol treatment and starvation. Biochem Pharmacol 1995; 50:155–61. 144. Vuppalanchi R, Teal E, Chalasani N. Patients with elevated baseline liver enzymes do not have higher frequency of hepatotoxicity from lovastatin than those with normal baseline liver enzymes. Am J Med Sci 2005; 329:62–5. 145. Sanyal AJ, Mofrad PS, Contos MJ, et al. A randomized controlled pilot study of pioglitazone and vitaminm E versus vitamin E for nonalcoholic steatohepatitis. Clin Gastroenterol Hepatol 2004; 2:1107–15. 146. Chalasani NE, Teal E, Hall SD. Effect of rosiglitazone on serum liver biochemistries in diabetic patients with normal and baseline liver enzymes. Am J Gastroenterol 2005; 100:1317–21. 147. Ferna´ndez-Villar A, Sopena B, Ferna´ndez-Villar J, et al. The influence of risk factors on the severity of anti-tuberculosis drug-induced hepatotoxicity. Int J Tuberc Lung Dis 2004; 8:1499–505. 148. Patel PA, Voigt MD. Prevalence and interaction of hepatitis B and latent tuberculosis in Vietnamese immigrants to the United States. Am J Gastroenterol 2002; 97:1198–203. 149. Lee BH, Koh WJ, Choi MS, et al. Inactive hepatitis B surface antigen carrier state and hepatotoxicity during antituberculosis chemotherapy. Chest 2005; 127:1304–11. 150. Gordin FM, Simon GL, Wofsy CB, Mills J. Adverse reactions to trimethoprim–sulfamethoxazole in patients with the acquired immunodeficiency syndrome. Ann Intern Med 1984; 100:495–9. 151. Medina I, Mills J, Leoung G, et al. Oral therapy for Pneumocystis Carinii pneumonia in the acquired immunodeficiency syndrome. A controlled trial of trimethoprim–sulfamethoxazole versus trimethoprim–dapsone. N Engl J Med 1990; 323:776–82. 152. Alfirevic A, Stalford AC, Vilar FJ, Wilkins EG, Park BK. Slow acetylator phenotype and genotype in HIV-positive patients with sulphamethoxazole hypersensitivity. Br J Clin Pharmacol 2003; 55:158–65. 153. O’Neil WM, MacArthur RD, Farrough MJ, et al. Acetylator phenotype and genotype in HIV-infected patients with and without sulfonamide hypersensitivity. J Clin Pharmacol 2002; 42:613–9. 154. Eck HP, Gmu¨nder H, Hartmann M, Petzoldt D, Daniel V, Dro¨ge W. Low concentrations of acidsoluble thiol (cysteine) in the blood plasma of HIV-1 infected patients. Biol Chem 1989; 370:101–8. 155. Buhl R, Holroyd KJ, Mastrangeli A, Cantin AM, Jaffe HA, Wells FB, Crystal RG. Systemic glutathione deficiency in symptom-free HIV seropositive individuals. Lancet 1989; 2:1294–8. 156. Naisbitt DJ, Vilar FJ, Stalford AC, Wilkins EG, Pirmohamed M, Park BK. Plasma cysteine deficiency and decreased reduction of nitrosomethoxazole with HIV infection. AIDS Res Hum Retroviruses 2000; 16:1929–38. 157. Rieder MJ, Krause R, Bird IA, Dekaban GA. Toxicity of sulfonamide-reactive metabolites in HIVinfected, HTLV-infected, and noninfected cells. J Acquir Immune Defic Syndr Hum Reterovirol 1995; 8:134–40. 158. Carr A, Swanson C, Penny R, Cooper DA. Clinical and laboratory markers of hypersensitivity to trimethoprim–sulfamethoxazole in patients with Pneumocystis carinii pneumonia and AIDS. J Infect Dis 1993; 167:180–5. 159. Ozick LA, Jacob L, Comer GM, et al. Hepatotoxicity from isoniazid and rifampin in inner-city AIDS patients [see comments]. Am J Gastroenterol 1995; 90:1978–80. 160. Russell AS, Sturge RA, Smith MA. Serum transaminases during salicylate therapy. Br Med J 1971; 2:428–9. 161. Miller JJ, III, Weissman DB. Correlations between transaminase concentrations and serum salicylate concentration in juvenile rheumatoid arthritis. Arthritis Rheum 1976; 19:115–8. 162. Bernstein BH, Singsen BH, Koster King K, Hanson V. Aspirin-induced hepatotoxicity and its effect on juvenile rheumatoid arthritis. Am J Dis Child 1977; 131:659–63. 163. Iancu T. Serum transaminases and salicylate therapy. Br Med J 1972; 2:167.

Risk Factors

305

164. Seaman WE, Plotz PH. Effect of aspirin on liver tests in patients with RA or SLE and in normal volunteers. Arthritis Rheum 1976; 19:155–60. 165. Zimmerman HJ. Effects of aspirin and acetaminophen on the liver. Arch Intern Med 1981; 141:333–42. 166. Kromann-Andersen H. Adverse reactions during salazosulfapyridine (sulfasalazin) therapy in Denmark between 1968–1987. Ugeskrift For L*ger 1987; 149:3557–59. 167. Banks AT, Zimmerman HJ, Ishak KG, Harter JG. Diclofenac-associated hepatotoxicity: analysis of 180 cases reported to the Food and Drug Administration as adverse reactions. Hepatology 1995; 22:820–7.

15

Genomics, Proteomics, and Metabolomics in the Diagnosis and Mechanisms of Drug-Induced Liver Disease Andrew A. Stolz

Division of Gastrointestinal and Liver Diseases, Department of Medicine, Keck School of Medicine, University of Southern California, Los Angeles, California, U.S.A.

INTRODUCTION In the discipline of pharmacology, the “ome” revolution (the genome, transcriptome, proteome, and metabolome) has merged to create the toxicogenome and the pharmacogenome. The evolution of these disciplines has been made possible by the convergence of rapidly developing technological advances for the detection of gene expression levels and the identification of proteins and small molecules on automated platform arrays. These arrays are able to simultaneously evaluate the individual components of complex samples, whose identity is predicted by sophisticated bioinformatics software requiring powerful computational analysis. These technological advances in combination with the completion of the genomes of humans, and the traditional animals used in preclinical drug toxicity studies, mice and rats, are predicted to rapidly lead to major advances for both the diagnosis of drug-induced liver disease and for an enhanced understanding of their molecular mechanisms of toxicity. Future technological advances such as global analysis of lipids, carbohydrates, and lipoproteins will no doubt contribute to these fields. Drug mechanisms coupled with variations in host response due to genomic polymorphisms and environmental factors, such as coadministrations of drugs or concurrent conditions, will define how host factors may render individuals more susceptible to drug-induced liver disease. Besides providing new biomarkers, changes in gene expression profile hint at specific pathways, which gives clues to the mechanism of drug hepatotoxicity, opening new areas for investigation of both the mechanism and prevention of liver injury. In this chapter, the methodology of the transcriptomic, genomic, proteomic, and metabolomics profiling will be reviewed with pertinent examples that specifically relate to hepatotoxicity. The role of gene polymorphisms in drug metabolic enzymes will be reviewed followed by a summary of the anticipated future advances in the field and how this will impact the evaluation of drug-induced liver disease. Several excellent reviews are suggested for a more global overview of this exciting and rapidly evolving field of pharmacology (1–3). OVERVIEW OF PROFILING TECHNOLOGIES An overview of the principles used for assessment of global expression of genes, proteins, and metabolites are provided. Rapidly evolving technologies will continue to influence how many samples can be analyzed at the same time and how much material is required for these analyses. Personal Genomes Advances in DNA sequence technology have lead to the routine genome analysis of numerous species. As of July 2006, 405 genomes have been entirely sequenced including 41 eukaryotic genomes with the vast majority of the others being of bacterial origin. See Ref. 4 for a current list of completed genomes. This rapidly expanding number of completed genomes is a remarkable

308

Stolz

achievement as the human genome project was only initiated in 1990 and completed two years ahead of schedule in 2003, and the genome of the first living species, Haemophilus influenza, was only completed in 1995 (5). The routine completion of numerous genomes from other mammals such as cat, dog, horse, and ape now allows for comparison of genome organizations and predicted protein sequence homologies across diverse species, which will identify essential proteins responsible for basic cellular functions, such as those involved with cell cycle and DNA repair. Within the next three to five years, it is predicted that the cost of completely sequencing the genome of an individual will be reduced to $1000. This cost is low enough that genome analysis may be become part of routine health care, such as vaccinations or cancer screening. Currently, DNA sequence analysis of exons and promoter regions is routinely performed in individuals and family members with histories suggestive of hereditary cancer disorders, such as familial adenomatous polyposis or breast cancer. A major hurdle for advocating use of routine genome analysis will be how this information will be used for both health prevention and more specifically for prescribing drugs and predicting the risk for developing adverse events. The availability of an individual’s genome will also raise difficult social and ethical issues, which need to be carefully considered before proceeding with routine genome analysis (6). With the detailed completion of a prototypical human genome, the current focus of human genome research has shifted to understanding how variations in genome sequences can modify risk for disease and response to pharmaceutical agents (7–9). On average, humans share 99.5% to 99.8% identity in their DNA sequence between individuals. It is these subtle differences in gene sequence in combination with environmental factors, which may modify the risk for specific diseases and the individual’s ability to metabolize and respond to pharmaceutical agents. Genetic variations include single nucleotide polymorphisms, referred to as SNPs, deletions, insertions, and variations in length of repetitive elements. Besides DNA sequence variations, epigenetic factors, such as those due to methylation of DNA or changes in chromatin structure, are also important modifiers of gene expression (10). It has been estimated that 90% of the polymorphisms between humans are due to SNPs. The distribution of SNPs varies and those SNPs in the coding region of the gene, referred to as cSNPs, occur less frequently as variations in protein sequence may alter its function. Nonetheless, it is estimated that up to 60,000 cSNPs exist in the human genome. As with the human genome project, a common database is being established for studying the genetic variability between humans. In order to expedite this process, the international HapMap project was launched in October 2002 to create a public, genome-wide database of common human sequence variations as a necessary tool to guide genetic studies of different clinical phenotypes. The initial goal of the HapMap was to identify a SNP located within 5000 bps blocks throughout the genome allowing for association studies. Initially, SNP frequencies were determined in 269 genomes from four different ethnic groups. These DNA samples consisted of 30 trios—30 parents and offspring from Yoruba in Ibadan, Nigeria; 30 from the state of Utah, U.S.A., obtained from the Centre d’etudes du Polymorphisme Humain Collections; 45 Han Chinese in Beijing, China; and 44 Japanese residing in Tokyo, Japan. The initial findings of the HapMap were reported in October 2005 (11). From these initial studies, it was found that individuals that carry a particular SNP allele at one site often predictably carry specific alleles at other nearby variant sites. This correlation is referred to as linkage disequilibrium in which regions of DNA are shared as a unit. Combination of alleles along a chromosome is referred to as a haplotype. Regions are shared because of the ancestry of contemporary chromosomes. When a mutation occurs within a chromosome, it is initially tethered to a unique chromosome on which it occurred, marked by a distinct combination of genetic variability. Future recombinations and mutations eventual erode this association but this is a very slow process as mutations occur at rates of 1!108 as compared to number of generations, typically 104 to 105, since the mutational event occurred. Determination of SNP haplotypes across genomes can rapidly identify shared regions of chromosomes based on their haplotype pattern (11). The HapMap will be a very useful tool for future genetic association studies, and to identify loci in the human genome that have been subjected to natural selection during human evolution. With the HapMap available, a chip-based assay has been developed allowing for

Global Profiling in Drug-Induced Hepatotoxicity

309

rapid assessment of an individual’s haplotype across their genome facilitating identification of shared regions between different individuals (12). This can be particularly powerful when dealing with a relative homogeneous population that has both detailed genealogic data and comprehensive health records, as illustrated by recent studies in Iceland identifying chromosomal regions associated with common diseases (13–15). In these studies, the SNP itself may be responsible for the specific phenotype or the mutation responsible for that phenotype may be in linkage with that SNP, which is co-inherited with the mutation. In order to perform these association studies, it is very important to separate individuals into homogeneous phenotypes so that specific genotypes can be associated with that phenotype. An inherent problem with drug-induced liver disease is that it may be very difficult to separate out environmental factors from genetic variability. However, the availability of whole genome haplotypes will advance association studies within well-defined populations. To perform these studies, array techniques as described for analysis of the transcriptome or PCR-based genotyping can be utilized. Real-time PCR amplifies the DNA region of interest with PCR primers in combination with SNP-specific oligonucleotide probes. These probes are distinguished from each other by fluorescent dyes located at the 5 0 end of the probe. The fluorescence of these dyes is severely reduced when they come in close proximity to a fluorescence quencher located at the 3 0 end of the primer due to the fluorescence energy transfer phenomenon. The 5 0 exonuclease activity of the Taq DNA polymerase cleaves the 5 0 -flanking dye from the primer after the probe binds to the amplified region of DNA thereby releasing it from its quencher. The genotype at the site is determined by the relative amounts of the two fluorescent dyes. This approach is suitable for large-scale screening of multiple SNPs. Other technologies such as the Illuminae bead system are also capable of simultaneously monitoring up to 256 SNPs on one sample analysis based on their proprietary technology. Global Gene Expression Profiling: The “Transcriptome” The global assessment of relative levels of gene transcription is routinely now available with the advent of microarray chips performed on different platforms. The principal for all these techniques is based on the thermal stability of complementary hybridization between single stranded reverse complemented DNA bound to a platform and labeled cRNA or its converted cDNA isolated from the source of interest (16,17). These DNA sequences correspond to genes identified by cloning, predicted transcript identified by sequencing of the human genome, as well as the partial sequence of cDNA fragments, determined by random sequencing of cDNA libraries. For the Affymetrixe microarrays, the synthesis of unique oligonucleotides on the platform is achieved by using the technique of photolithography, which was developed to build multilayer circuits on silicon wafers used in computer chips. Other technologies are also available for synthesizing oligomers on these platforms. These microarray chips may have up to 40,000 30- to 40-mer oligonucleotides as targets for hybridization to the fluorescently labeled cRNA or cDNA. Typically, between 11 and 20 pairs of oligomer probe a targeted RNA sequence, which is complementary to unique gene-specific sequences distributed throughout a gene. One of the pair of oligonucleotides has a single base pair mutation in its middle, which provides a means to monitor nonspecific stray hybridization as compared to its perfect match. For these studies, RNAs or cDNAs are labeled by incorporation of a fluorescently labeled dUTP base pair—either Cy3 or Cy5, which provides a means to detect hybridization to the oligomer attached to the platforms. These labeled nucleic acids are fragmented prior to hybridization to facilitate binding to the oligonucleotides. After hybridization to these oligomers on the microarray, a scanning laser records the intensity of the fluorescence at the site corresponding to a specific oligomer. Presence of alternatively processed mRNA can also be detected if significant variation in hybridization occurs within different regions of the same gene. Chip sets are also available for global assessment of mouse, rat, and yeast genomes. Other techniques use spotted cDNAs onto glass slides for evaluation of a more discreet set of genes, which is more economical than the high-density microarray chips. For these arrays, only one sequence is typically used. RNA pools from different conditions labeled with unique fluorescent markers can be mixed and relative intensity of each fluoroprobe determined at the same time. Confirmatory tests such as real-time PCR are needed to verify changes in gene

310

Stolz

expression. Little correlation may exist between these two different techniques emphasizing the difference in these two technologies (18). Non-transformed datasets for these microarray studies are typically deposited in common databases to allow other investigators to evaluate the raw data. Differences in hybridization can be represented graphically in so-called heat maps in which relative expression in genes that are over-expressed are depicted in red. In general, gene array studies can be used in two different ways. First, analysis of global gene expression can identify a pattern of gene expression, which may be predictive of a specific outcome (3). Numerous examples exist of this type of analysis in the oncologic literature. For example, treatment of acute lymphocyte leukemia in the pediatric population has an 80% long-term disease free survival, but currently no parameters exist that can consistently predict the response of these patients prior to their chemotherapy treatment. In one study, gene expression profile was correlated with the de novo sensitivity of leukemic cells to the four chemotherapeutic agents used for treatment and 46 genes that were differentially expressed in these cells were identified (19). This gene expression pattern was able to accurately predict response to treatment with these agents in a different cohort of patients. The other major use of transcriptome analysis is to identify mechanistic aspects by determining changes in expression patterns in genes involved with specific pathways. Mechanistic analysis of drug-induced liver disease has been examined by feeding mice with prototypic hepatotoxins, such as benzene, which causes cellular necrosis, or peroxisomal proliferators, which cause steatosis, to identify patterns of gene expression, which underlie specific pathways associated with these two different classes of hepatic injury (20). The hope with use of microarray analysis early in the drug-development stage is identification of potentially toxic compounds during their initial screening thereby eliminating those compounds from further development. With prototypical patterns of liver gene expression already established in animals, these signatures of expression patterns could be used to characterize liver injury in patients with drug-induced liver disease if hepatic RNA is available for analysis. The ability to monitor expression profile of up to 10,000 genes is relatively easy compared with the complexity of analyzing the results. The interassay variability is lower than with other detection methods such as northern blot. However, many factors influence the ability to compare findings in different experiments and to assess the potential significance of relative changes in gene expression. The background fluorescence needs to be determined for each run, which introduces a significant potential bias to the analysis. Low levels of fluorescence are inherently more difficult to accurately measure and have greater variability compared to higher levels. As most genes are expressed at relatively low levels, it is difficult to accurately detect subtle, but potentially important differences in their relative gene expression under different conditions. Background “noise” in these experiments, such as efficiency of reverse transcription and subsequent amplification, can also have a significant impact on overall ability to discern subtle differences. It is also important to note that 500 potential false-positive genes may be identified when using the statistically significant P value of 0.05 when analyzing 10,000 genes. When comparing results from different sources, these concerns as well as differences in experiments design, timing, and dose used for studies, expression differences due to species and within strains of the same species makes it difficult to assign changes in gene expression due to drug treatment as compared to other factors. Other considerations such as differences in diet composition and exposure to environmental agents are other potential confounding factors as well as the time of day when the RNA is harvested, when differences in cortisol or growth hormone levels can have independent effects and contribute to the biological noise in these studies. When using whole organs, minor differences in specific cell populations between samples can skew the analysis, if these cells are contributing the bulk of the differentially expressed genes. However, sophisticated software and statistical analysis have begun to identify different patterns of expression and now these genes are also linked to specific metabolic pathways. Databases are now also culled to examine which genes are co-regulated in parallel so as to identify potential pathways. A great deal of experience has been gained in both the study design and pitfalls of global expression profiling. Experience gained from these large-scale analyses combined with the use of uniform chips and availability of raw data allows for continuing improvement of the mathematical and statistical algorithms used to identify changes in gene expression within single and across multiple experiments.

Global Profiling in Drug-Induced Hepatotoxicity

311

Besides transcriptome analysis, this platform can also be used for sequencing genes or SNP analysis. Genes can be sequenced in their entirety by generating oligonucleotides that interrogate every possible base pair sequence at any given nucleotide position and the technique is sensitive enough to determine the DNA sequence for each allele by comparing their relative hybridization intensity. A P53 sequencing chip is currently available using this technology. This same platform has been also adapted for genome-wide evaluation of SNPs, which can be used for linkage or haplotype mapping studies (16). Global Assessment of Protein Expression: The “Proteome” Array-based genotyping and mRNA expression profiles are relatively simple compared to the complexity of characterizing global expression patterns of proteins. Three major reasons exist for monitoring the proteome expression profile. First, temporal and spatial expression of proteins is not apparent from their gene expression profiling. Second, post-translational modifications can only be determined by direct examination, such as protein phosphorylation or disulfide formation, which are important modifiers of protein function. Also, posttranslational modifications of proteins are a critical feature of all cellular signaling pathways. Third, protein–protein interactions cannot be deduced by gene array studies and these interactions are known to play an important role in signal transduction mechanisms (21). For hepatotoxicity studies, identification of proteins that may be modified as a consequence of free radical formation or reactive drug metabolites may be useful for both diagnostic testing and to discern potential targets for these hepatotoxins. Use of proteome analysis allows a more detailed understanding of the pathways responsible for patterns of liver injury. For example, proteome analysis of a compound which caused steatosis in rat liver identified 22 proteins in the liver of which a large number were involved in pathways that lead to accumulation of acetyl-CoA and triglycerides (3). Proteome analyses are broadly subdivided into three major categories. First, a global expression profile identifies proteins that are either over- or under-expressed under different conditions. Routinely, proteins are resolved by two-dimensional gel electrophoresis in which they are first separated by their charge followed by their molecular weight. Proteins are next transferred to membrane matrix or those that are differentially expressed are extracted and undergo peptide sequence analysis as described below. Second, organelle-based proteomics aims to identify all proteins within a compartment to provide detailed characterization of their physiological functions at that site. Finally, modular proteomics seeks to identify those proteins that constitute a functional unit in order to gain insight into its mechanism, such as gene transcription. With the completion of the human genome, the peptide sequence of all genes can be deduced assuming that the exonic regions for all genes are known. To determine the peptide sequence, proteins typically undergo proteolytic cleavage, which generates protein fragments whose terminal amino acid sequence is known. These peptide fragments are then separated by sequential liquid chromatography systems separating the proteins by their charge and hydrophobicity. Protein peaks are collected and then subjected to sequential mass spectroscopy analysis. First, proteins are separated by their mass by a time of flight mass spectroscopy. The protein peaks are collected, and next applied to another mass spectroscopy instrument in which peptide fragments are randomly generated during their ionization. In this process, the molecular mass of the smaller peptides fractured at their amide bounds are determined. Amino acid sequence is deduced by comparing the molecular mass of individual fragments with databases allowing reconstruction of the peptide sequence. Post-translational modifications of amino acids are identified by comparing the mass of individual peptides with either normal or modified amino acids. The rapidity of mass spectroscopy and its accuracy are essential components for its use in proteome analysis (21). These studies generate large datasets, which undergo extensive sequence alignments and requires sophisticated computational analysis. Besides direct sequence analysis, protein chip technology can also be used to rapidly identify protein constituents of cells. Proteins of up to 30 amino acids may be immobilized on matrix material, such as cellulose membranes or glass. These peptides are short enough

312

Stolz

to allow interactions between enzymes and substrates, antigens and antibodies, receptors and hormones, and nucleic acids and those transcription factors that bind to specific sequences (22). Technology has also been developed to directly synthesize peptides on the membrane with different amino acid sequences. These libraries of short peptides can be especially useful for identifying the epitope of antibodies. Antibody arrays have also been developed to identify proteins in which protein extracts are labeled with different fluorescent probes and relative protein expression compared between groups. Metabolomics Metabolites are the endpoint of an organism’s response to stimuli and profiling provides a direct measure of physiological function. With metabolomics, the fluxes of metabolites are determined within a cell and metabonomics refers to a global, system-wide evaluation of metabolites in biological fluids, such as serum and urine, along with tissue samples (3). Integration of the transcriptome, proteome, genome, and metabonome provides a global view of cellular function. For example, profiling of metabolic pathways in different plant species has been correlated with gene expression profiles and metabolomic analysis has successfully differentiated strains of yeast. This technology is also being used to evaluate urinary metabolites in animals treated with hepatic or renal toxins. As with other “omic” technologies, high-throughput platforms are being developed for these studies. Typically, mass spectroscopy is being used for detection of metabolites and it has an advantage over NMR technique because of its increased sensitivity. However, larger magnets are being developed to increase the sensitivity of NMR and it potentially can be used noninvasively for detecting changes in flux of metabolites in vivo (23). Analysis of large datasets requires pattern recognition software. With metabolomic analysis, subtle differences in drug metabolism may be apparent and linking these changes with expression data and genome analysis is expected to further elucidate the metabolic pathways for drugs. At this stage of development, datasets are being developed with model hepatotoxins for developing the analytical tools necessary to integrate and compare global changes with the anticipation that profiles will be developed that will help to elucidate mechanisms of drug-induced injury. As with gene expression studies, new techniques are being developed to monitor the flux of metabolites. To achieve this sort of automated separation and identification, combination of separation techniques such as gas chromatography, liquid chromatography, and capillary and microchip electrophoreses are combined with each other and then coupled with mass spectroscopy or NMR analysis to identify the metabolites. Tandem GC is being used to separate metabolites in which the first separation is typically performed over a one- to two-hour period in a 15- to 30-m column using a nonpolar stationary phase, which supports boiling point-based separation. Peaks are then passed through a modulator, which refocuses the content and then injects its constituents onto a shorter GC column with a polar stationary phase run over a much shorter period of time. Peaks separated by GC are then injected into a mass spectrometry, and mass is determined by time of flight mass spectroscopy. In addition to GC separation, tandem HPLC methodology as used for proteome analysis can resolve metabolites from each other. These instruments require computer-controlled valves to separate the peaks and then transfer them to the other HPLC. Constituents can be detected by fluorescence, MS, photodiode array, in which an absorption spectrum is monitored simultaneously, evaporative light scatter and more recently NMR. Finally, metabolites can be separated by capillary electrophoresis and capillary electrochromatography that utilize narrow bore capillaries to separate molecules by electrical field. These techniques offer rapid separation with minimal amount of buffers. These technologies make it feasible to continuously monitor changes in metabolites during toxicity studies. ROLE OF GLOBAL PROFILING TECHNOLOGIES IN DRUG-INDUCED LIVER DISEASE Adverse drug reactions account for over 2 million hospitalizations and are estimated to account for 100,000 deaths per year in the United States thereby establishing them as a major health concern (24). Drug-induced liver injury is a significant contributor to these adverse events and

Global Profiling in Drug-Induced Hepatotoxicity

313

carries a particular poor prognosis when jaundice is the presenting symptom of hepatocellular necrosis. Besides contributing to adverse drug events, drug-induced hepatic injury is the leading cause for withdrawal of newly approved pharmaceutical agents. Until recently, preclinical testing of drug-induced liver injury typically consisted of monitoring serum markers of hepatic necrosis, AST and ALT, synthetic function, and cholestatic markers such as bilirubin and alkaline phosphatase with corresponding histologic evaluation in rodent models. Patients in early-phase testing are prospectively evaluated with routine serum chemistries and clinical evaluation. Patterns of drug-induced liver injury are defined by their relative ratio of serum transaminase, markers of cholestasis, and the presence of clinical manifestation suggestive of a hypersensitivity reaction (fever, eosinophilia, or other organ involvement) in the absence of another etiology for the liver injury. To gain insight into the molecular mechanisms and what constitutes the different patterns of drug-induced liver injury, signature profiles for genes, proteins, and metabolites in rodent models are being generated by using prototypical hepatotoxins in combination with the standard serum and histologic evaluation. The anticipated benefit of global profiling is that critical pathways associated with the development of a specific pattern of liver injury will be revealed. By identifying specific pathways, polymorphisms in those genes that are components of that pathway or regulate its activity are candidates for extensive genotyping to identify haplotype associated with a specific pattern of liver injury. Profiling may also identify a potential hepatotoxin prior to the actually development of overt toxicity during the extensive preclinical testing that occurs during the initial phase of drug development. By linking profiling with histologic features, the knowledge gained from these studies can be used to revaluate previous studies on drug-induced liver injury. This approach is only in its infancy but is anticipated to rapidly advance our understanding of the pathophysiology of drug-induced liver disease as well as influence how new drugs are developed in the future. This enthusiasm must be tempered by the realization that preclinical studies in animal models are poor predictors of idiosyncratic hepatotoxicity observed in humans. A few examples of global profiling in drug-induced liver disease are presented. Expression Profiling for Hepatotoxicity Gene profiling is an extremely attractive tool for the screening of potential hepatotoxins and to extend the classification of drug-induced liver injury beyond currently available histologic or serum markers of liver injury. With completion of the mouse and rat genomes, the ease of monitoring the transcriptome using microarray platforms has spurred the development of expression profiling for different hepatotoxins. Waring et al. (25) treated primary rat hepatocytes with 15 different hepatotoxins, which cause hepatic injury by DNA damage, generation of reactive oxygen species, cytochrome (CYP) P450 induction as well as other mechanisms. When the expression profiles were analyzed by hierarchical cluster analysis, many of the expression profiles clustered together according to their mechanism of hepatotoxicity, demonstrating the predictive value of expression profiling. Applicability of this approach is limited as only primary rat hepatocytes were evaluated, but these findings demonstrate the feasibility of identifying expression profiles for specific classes of hepatotoxins. In another study, a training set of rat liver expression profiles was determined after treatment with a prototypical CYP450 inducer or peroxisomal proliferators. This dataset was then able to accurately identify the hepatic transcriptome profile caused by drugs of a similar class as well as correctly distinguishing the profiles induced by another class of agents (26). In a third study, prototypical agents were used to generate a toxicogenomic database to compare expression profiles of two highly related 5-HT6 receptor antagonists, one of which was a known steatotic hepatotoxin. The hepatic histology and routine serum tests as well as biochemical tests were used to define the following categories of hepatic injury (27): (1) Steatosis: Mitochondrial injury has been implicated as a cause of the microvesicular form of steatosis due to reduced capacity to catalyze b-oxidation of lipids with their consequential accumulation within hepatocytes. More extensive mitochondrial injury can also lead to metabolic acidosis, hepatic coma, liver failure, and death.

314

Stolz

(2) Cholestasis: Cholestasis defined as impairment of bile flow is typically due to reduced transport capacity of hepatic canalicular transporters that are responsible for the efflux of cholesterol, phospholipids, or bile salts. Intracellular retention of these noxious biliary components leads to hepatocellular injury. Cholangiocytes may also be the cellular targets of injury, which can lead to a cholestatic pattern of liver injury. It is important to note that inflammatory cytokines can cause a cholestatic pattern of liver injury, defined by predominant elevation of serum alkaline phosphatase, as compared to serum transaminases, in association with elevation of the serum bilirubin. (3) Some cellular hepatotoxins are direct acting toxins that characteristically are free radicals generated by their hepatic metabolism. These metabolites or reactive oxygen species can interact with proteins, membrane, or DNA resulting in hepatocellular necrosis, presenting with marked elevation in serum transaminases and impairment of hepatic function. The importance of the immune system is now being increasingly recognized as being an integral component of this type of liver injury. (4) Peroxisomal proliferators and non-genotoxic agents are associated with hepatic enlargement and increased risk for hepatocellular carcinogenesis in rodent models. Analysis of these microarray results by a supervised learning method and other bioinformatics strategies were compared to histological and serum chemistry to generate characteristic expression profile, or signatures for these different classes of hepatotoxicity with the assumption that toxins that have similar histology and serum patterns of liver injury will also share a comparable gene expression profile. Gene expression of known steatotic agents exhibited a comparable expression to other steatotic hepatotoxins as early as 24 hours after administration of the agent. In subchronic dosing studies, gene expression profile correlated with the histologic appearance of steatosis after achieving a threshold dose. These findings suggest that transcriptional profiling can accurately identify toxicity prior to changes in histology. Although these studies demonstrate the feasibility of gene expression profiling to identify a class of toxin, identifying the mechanism for injury has been much more elusive. For example, multiple studies have reported the toxicogenomic characterization of acetaminophen hepatotoxicity as it reproducibly produces a severe centrilobular hepatic necrosis in humans and experimental rodents. Comparison of differential gene expression between major categories of biological activities is one of the simplest ways of comparing results among these studies. Changes in gene expression of stress response and transcriptional control were consistently altered in only two of five studies (1). The inconsistency in the findings highlights the importance of study design and species differences as well as the difficulty of comparing data from different sources. Measurement of both genes and proteins at the same time revealed that protein levels changed rapidly where as gene expression lagged behind (28). Using a sublethal and lethal dose identified shared gene targets, which were related to the generation of the toxic metabolite. However, the ability to recover from the insults depended on levels of cellular damage, which may not be apparent from changes in gene expression or proteomics. These studies demonstrate how using expression profiling of proteins and genes begins to reveal the mechanism of hepatotoxicity, but identifying consistent pathways may be difficult given the importance of these other factors. Variation in gene expression and protein in these studies despite consistent histologic appearance reveals that pathologic features are not necessarily predicted by transcriptome or proteome analysis. Integrative analyses of findings from transcriptome, proteome, and metabonomic profiling are beginning to dissect the causes of the initial drug-induced liver insult from the cellular response to the injury. An increasing number of studies are now utilizing global profiling to evaluate and uncover molecular mechanisms of hepatotoxicity and the liver’s response to these toxin-induced injuries. In rat models, methapyrilene is known to induce periportal necrosis associated with bile duct hyperplasia and inflammation. In isolated hepatocytes, glutathione depletion is observed, but not in the whole animal. Global analysis was performed to understand the mechanism of toxicity in conjunction with histologic evaluation (29). In these studies, timed liver samples were collected for RNA, protein, and

Global Profiling in Drug-Induced Hepatotoxicity

315

metabonomics studies along with histology and corresponding serum and urinary samples and were compared with vehicle control. Dose-dependent changes in gene expression occurred with about 60% of genes upregulated at low exposure levels, which were also increased in animals treated with the higher dose of the hepatotoxin. Dose escalation revealed new groups of genes, which were induced by the drug. Induction of genes along with proteins and metabolic by-products of a specific pathway strengthen the significance of the observed changes in the gene profile analysis. Thus, analysis of a profile in one field if associated with changes in another field validates the significance of the findings in both fields. Combination of biomarkers identified by this approach may thus be more specific for the hepatotoxin than biomarkers from only one type of analysis. Using this approach, biomarkers from two different profiles such as transcriptomics and proteomics may identify a unique, drug-specific hepatotoxic signature. Metabolic analysis may be most relevant for assessing the functional significance of changes in protein and gene expression patterns. In comparing expression level between a nontoxic and toxic dose, the physiological stress became apparent when genes involved with apoptosis and stress response were upregulated as well as activation of metabolic pathways involving choline, glucose, and lipid metabolism. Other toxins elicited similar changes in metabolic pathways suggesting a link between hepatocellular necrosis and perturbations in energy metabolism. Linking these observations provides insight into the cellular response associated with hepatocellular necrosis. An example of the value of this approach is shown with a study of bromobenzene hepatotoxicity, which is a well-recognized cause of centrilobular necrosis due to its cytochrome P450-mediated metabolism that generates electrophilic epoxide intermediates, which bind to macromolecules. These reactive metabolites also generate oxidant stress. Timed samples after the exposure to bromobenzene provide a means of identifying the initial injury followed by a secondary response to the toxin by monitoring flux of metabolites out of the liver into the serum and their eventual deposition into urine (30). Metabonomic evaluation of timed samples from liver, serum, and urine of rats compared to vehicle-treated controls demonstrated elevation of 5-oxoproline, suggesting a defect in glutathione synthesis as a result of oxidant injury. In another study, oxidative stress induced by acetaminophen ingestion lead to the progressive appearance of ophthalmic acid in the liver and eventually the serum. In acetaminophen injury, hepatic glutathione depletion is a key event and was temporally associated with an increase in the serum ophthalmic acid levels (31). From this model, a promising new biological marker for glutathione depletion was identified, which may be utilized in future drug-induced liver injury models to assess for oxidative injury (31). Changes in the transcriptome and proteome can be a direct response to the drug or a secondary response to the injury induced by that agent. Once a pathway is identified with a class of hepatotoxins, its role can be examined in other model species, such as mice or yeast. Advantages of using yeast as test species are that specific genes can be selectively eliminated with ease, thereby allowing their function to be directly tested. The findings from these studies can provide insight into basic mechanisms of hepatotoxicity in humans because of the high degree of shared sequence homology between these two species for genes involved with basic cellular functions. For rodent models, there are already a large number of genetically engineered mice that are available to study critical pathways. Strain differences in the sensitivity of a drug-induced adverse event can also be exploited to identify areas of genome divergence that may harbor genes responsible for the enhanced sensitivity to a specific hepatotoxin. The hope with this approach is that identification of strain-dependent modifier genes or specific polymorphisms will reveal the pathophysiological mechanism for the hepatotoxicity or host susceptibility factors. This potential enthusiasm has to be balanced by the difficulty in analyzing and characterizing the complexity inherent in such a large number of parameters and the growing recognition that expression profiles are significantly modified by variables which had received little attention in the past. Finally, how applicable these results will be to another species remains to be determined. This approach will provide a set of genes and proteins that are logical candidates to examine in humans who present with a similar pattern of drug-induced liver injury. Eventually, the use of global gene, protein, and metabolic profiling to assess hepatotoxicity is predicted to supplant the classical markers of histology

316

Stolz

and serum markers in both preclinical assessment and the evaluation of patients with a suspected drug-induced liver injury.

ROLE OF HUMAN VARIABILITY IN TOXICOGENOMICS With completion of the human genome project, we are now entering the post-genomic period, in which the role of inherited polymorphisms can now be compared between individuals and their potential role in hepatotoxicity determined. Genetic makeup and environmental exposure to different agents are predicted to be important factors for modifying the risk for development of hepatotoxicity. In order to have a better understanding of the differences in the genetic makeup within different ethnic groups, the NIES has developed the Environmental Genome Project given in Ref. 32 to identify common sequence polymorphisms in 200 genes predicted to be important in environmental disease and likely to also play an important role in druginduced liver injury. The risk of developing hepatotoxicity in any given individual has been attributed to both environmental factors, such as coadministered drugs, effect of diet, or comorbid conditions as well as their unique genotype. Currently, those gene products identified as potential modifiers of hepatotoxicity include polymorphisms in enzymes involved in drug metabolism or detoxification of toxic by-products of their metabolism. Membrane transporters that either uptake or efflux drugs are now also recognized as important modifier for a drug’s pharmacokinetics and exposure, which may be directly responsible for its associated toxicity. This may be especially important for drugs with a narrow therapeutic range and with predictable, dose-dependent adverse events. Other genes are also implicated in modifying drug metabolism or sensitivity to hepatic injury but are not as clearly established. Prime candidates include genes of the immune system, especially for those immune-mediated adverse drug reactions that also involve the liver. With the recognition that most drugs are metabolized by only a small number of enzymes, determining their genotype has gained considerable interest as a means for potentially identifying those individuals at increased risk for a specific drug-induced liver injury. The most obvious candidates are members of the cytochrome P450 (CYP) family; of which 57 are predicted to exist based on DNA sequence analysis of the human genome. Of these, 15 are implicated in metabolism of xenobiotics (33). Of these, CYP3A4, CYP2D6, and CYP2C9 are the major CYPs metabolizing pharmaceutical agents with CYP2C19, CYP2B6, and CYP1A2 also contributing. These specific CYPs are responsible for the metabolism of approximately 90% of all drugs and therefore functional polymorphisms in these genes have been the initial focus for genetic causes of drug-induced liver disease. Expression levels of CYPs can be highly variable between individuals and for CYP2D6, its pharmacokinetic properties can vary by a factor of 10,000 between individuals (34). In order to understand how a particular polymorphism may influence drug disposition, the significance of these polymorphisms depends on what effect it has on enzymatic activity for a specific drug. CYP polymorphisms have previously been referred to by specific alleles, but this nomenclature is now being replaced by their specific genotype, which can be determined by DNA sequencing or SNP analysis. Websites such as Ref. 35 list the genotype for specific CYP alleles. Polymorphisms in other pathways can also modify drug-induced liver injury by either modifying detoxification of reactive metabolites or promotion of hepatotoxicity by noxious stimuli (34). Transcriptional profiling of prototypical hepatotoxins coupled with well-conducted genome association studies are expected to identify genetic modifiers of drug-induced liver disease, which are not related to drug metabolism or detoxification. A review of the polymorphisms in the major CYPs follows. CYP2D6 Wide ethnic variations in CYP2D6 are recognized with 5% to 10% of the Caucasian population having little enzyme activity in contrast to populations of Mongolian origin in which less than 1% of individuals have reduced levels. In patients who have reduced enzyme activity, increased serum levels of drugs predominately metabolized by this enzyme may be associated with

Global Profiling in Drug-Induced Hepatotoxicity

317

higher incidence of side effects. As an example, patients on tricyclic antidepressants who lack CYP2D6 activity have a higher incidence of cardiac arrhythmias. As with other CYPs, CYP2D6 polymorphisms are grouped by their effect on metabolism and are arbitrarily divided into poor, extensive, and ultrarapid metabolizers. To date, more than 70 polymorphisms have been identified within exonic or promoter regions of the gene, which tend to reduce its enzyme activity. As an example, perhexiline, an effective agent for the prevention of angina that is predominately metabolized by CYP2D6, has a dose-dependent hepatotoxicity, and serum levels are routinely used to monitor these (36). Individuals whose CYP2D6 genotype are associated with reduced enzyme activities have higher serum levels suggesting that genotyping prior to initiation therapy may identify those at increased risk for dose-dependent liver injury. Polymorphisms in intronic regions may also influence enzyme activity by reducing mRNA stability (37). In some patients, a duplication of CYP2D6 renders drug substrates ineffective because of rapid metabolism. In one report, duplication was reported to cause a life-threatening intoxication with a small dose of codeine due to excess production of active metabolites. Although a relatively rare occurrence among Northern Europeans, the frequency of this duplication was present in 29% of an East African population. It has been postulated that wide variations in enzyme activity due to CYP2D6 polymorphisms may contribute to hepatotoxicity, although it has not been directly implicated in a cause for a specific drug. Development of pharmaceutical agents that are predominantly metabolized by CYP2D6 is now being avoided because of its highly variable enzymatic activity in different populations. CYP2C9 CYP2C9 catalyzes the metabolism of warfarin and the antiepileptic, phenytoin, and frequency of its polymorphisms vary in different populations. These polymorphisms can reduce catalytic activity, which therefore may require significantly lower dosages in those individuals that harbor these polymorphisms. For example, individuals with alleles CYP2C9*2 (Arg 144 Cys) and CYP2C9*3 (Ile 359 Leu) have a lower daily requirement for warfarin and are more susceptible to its adverse events (38). In combination with these CYP2C9 functional alleles, genetic polymorphisms in its pharmaceutical target, the vitamin K epoxide reductase multiprotein complex, can predict up to 60% of the warfarin dose for an individual. The causes for the remaining 40% variability of warfarin’s dosage are as yet unaccounted for. It is important to recognize that functional polymorphisms within a CYP may also be specific for certain drugs. For example, a CYP2C9 mutation can affect the 6-hydroxylation of R-warfarin but not the 4 0 hydroxylation of diclofenac. Other polymorphisms, which reduce or increase the levels of mRNA expression, will equally affect all substrates for that particular CYP (33). CYP2C19 Major classes of drugs metabolized by CYP2C19 are the antiepileptic agent, mephenytoin, and the proton pump inhibitor, omeprazole. Widely variable enzymatic activities have been found in different populations with approximately 20% of Japanese having reduced activity when compared with only 3% of Caucasians. Those with polymorphisms causing reduced enzyme activity have significantly higher serum levels 10 hours post-administration when compared with peak level of those with the wild-type gene. These polymorphisms in CYP2C19 are clinically significant as those individuals who are poor CYP2C19 metabolizers have a higher rate of clearance of Heliobacter pylori gastric infection when omeprazole is used in the regimen as compared to those with the wild-type gene. Non-P450 Genes Besides cytochrome P450, polymorphisms in other metabolic enzymes are associated with increased risk for adverse drug reactions. N-Acetyltransferase 2 Variations in the N-acetyltransferase 2 (NAT2) enzyme activity, which acetylates drugs including isoniazid and some of its metabolites, have been recognized for over 50 years.

318

Stolz

Slow acetylator status has been implicated with increased risk for isoniazid hepatotoxicity. Now, with the relative ease of genotyping, NAT2 polymorphisms, whose variations are responsible for differences in the pharmacokinetics of diverse drugs, such as hydralazine, procainamide, and isoniazid, can be easily identified as compared to using a pharmacokinetic probe. A few clinical studies have demonstrated a clear association between NAT2 genotypes and the risk for developing hepatocellular necrosis with isoniazid therapy. To determine if NAT2 genotype could influence the half-life of isoniazid, isoniazid pharmacokinetics after oral or intravenous administration was determined in 18 health Caucasian individuals and compared to their genotype for functionally significant polymorphisms (39). Highly active NAT2 phenotype was strongly associated with the non-compartmental area under the curve values and a linear relationship existed between the number of high activity NAT2 genotypes and overall variability in pharmokinetic parameters. Mode of isoniazid administration or body weight had minimal influence on these parameters and there was no significant gender difference. Slower NAT2-metabolizing genotypes were associated with four- to six-fold greater serum levels of isoniazid than rapid acetylators. Implications of these findings are that lower dosages of isoniazid may be required for those with the slow acetylator status, which may reduce the incidence and severity of isoniazid-induced hepatotoxicity. Therefore, determining an individual’s genotype before initiation of treatment may reduce the incidence of isoniazidinduced liver injury. This may explain the findings in another study of 102 Japanese patients receiving isoniazid alone, in which six patients had adverse events and in those with reduced NAT2 metabolism, 83.3% experienced an adverse event, such as nausea, fever, visual disturbance, or peripheral neuropathy, which responded to either reduction or elimination of the drug (40). In clinical studies comparing genotype frequency with isoniazid-associated hepatotoxicity, it is important to determine if isoniazid was coadministered with rifampin, a well-recognized inducer of the hydrolysis of isoniazid that accelerates the production of monoacetylhydrazine from acetylisoniazid, or hydrazine from isoniazid. Hydrazine has been incriminated in the hepatotoxicity of isoniazid (41). In another study, 224 newly diagnosed Taiwanese patients treated with isoniazid and rifampin were closely followed for development of isoniazid-associated hepatotoxicity, which developed in 14.7% of the population and was confirmed by rechallenge with the drug (41). Three known slow acetylator genotypes were evaluated and individuals with two or more of these genotypes were defined as slow acetylators. The odds ratio for having isoniazid-associated hepatitis was 2.87 for those with slow acetylator status when compared with rapid acetylators. These patients had higher mean AST and ALT levels and were more likely to increase their serum enzymes upon being rechallenged with the drug. In a multivariant analysis of this study, NAT2 acetylator status (OR 3.66) and old age were the only risk factors identified for isonazid-associated hepatotoxicity. Membrane Transporters In addition to metabolic enzymes, apical and sinusoidal/basolateral transporters and members of the ATP-binding cassette family of active transporters are now recognized as important modifiers for the pharmacokinetics of drugs. One of the best-characterized transporters is the MDR1, also known as ABCB1, which transports a wide range of drugs. To date, 28 SNPs have been reported; one SNP (C3435T), despite not changing the coding sequence of the transporter, is associated with increased risk for Parkinson’s disease, renal epithelial tumors, inflammatory bowel disease, and drug-resistant epilepsy (37). This synonymous polymorphism is associated with different levels of protein expression demonstrating that polymorphisms that do not alter the amino acid sequence or that are located in the non-promoter region of the gene, may still exert significant effects on protein expression. Polymorphisms in these genes in combination with genes involved with drug metabolism could significantly alter the pharmacokinetics of drugs leading to potentially toxic drug levels. HLA Genotypes Beside genes involved with drug metabolism and transport, polymorphisms in critical components of the immune systems have been shown to be highly associated with immunemediated adverse drug reactions, which may also involve the liver. For example, in a Chinese

319

Global Profiling in Drug-Induced Hepatotoxicity

Han population, specific phenotypic presentations of carbamazepine-induced cutaneous adverse drug reactions are strongly associated with specific genotypes of HLA molecules. Figure 1 illustrates the association of specific SNPS in the enzyme responsible for the metabolism of carbamazepine, epoxide hydrolase 1, or cytochromes (CYPs) as well as major histocompatibility complex (MHC) region with different cutaneous manifestation of carbamazepine adverse reactions. Remarkably, the most significant associations lie within polymorphism of the MHC locus, confirming the importance of the immune system in adverse skin reactions associated with carbamazepine. An extended haplotype B*1502 was associated with an increased relative risk of over 1300-fold for the development of

40 35 30 25 20 15 10 5 (A) 0

EPHX1

CYPs

MHC region Alleles Genotypes

5 4 3

–log10 (P value)

2 (B)

1 0 5 4 3 2 1

(C) 0 5 4 3 2 1 (D)

0

1

21

41

61

81

101

121

141

161

181

201

221

241

261

SNP markers

FIGURE 1 Association of SNPs with screening of 278 candidate SNPs for association with (A) CBZ-induced Stevens Johnson syndrome/toxic epidermal necrolysis, (B) CBZ-induced MPE/HSS, (C) CBZ-induced HSS, and (D) CBZ-induced MPE. On the x-axis, 278 SNPs are ordered by their chromosomal positions, including 58 SNPs of EPHX1 on chromosome 1, 190 SNPs in the MHC region on chromosome 6, and 30 SNPs of CYP3A4, 2C8, 2C9, 1A2, and 2B6. On the y-axis, the Klog10 P values were calculated by comparison of the allele (diamond symbols) or genotype (triangle symbols) frequencies between the patients and tolerant group using the Cochran-Armitage exact trend test. Significant association of CBZ cutaneous adverse events defined as a log greater than 3 are predominately associated with SNPs in the MHC loci and not for genes that encode for CBZ metabolic enzymes. Abbreviations: SNPs, single nucleotide polymorphisms; CBZ, carbamazepine; MPE, maculopapular eruption; HSS, hypersensitivity syndrome; EPHX1, epoxide hydrolase 1; MHC, major histocompatibility complex. Source: From Ref. 42.

320

Stolz

a Stevens–Johnson cutaneous adverse event. In contrast, no significant association was found for genotypes in enzymes responsible for generation of carbamazepine metabolites. Alleles in the HLA E region in combination with a HLA A*3103 allele had an increased odds ratio of 17-fold for development of a maculopapular eruption (42). Another example of a genotype associated with increased hypersensitivity in the HLA-B loci is HLA B*5701. Adverse hypersensitivity reactions to abacavir, a reverse transcriptase inhibitor used for the treatment of HIV, are closely associated with this specific genotype. Studies have demonstrated that prescreening individuals with HLA B*5701 genotype eliminated hypersensitivity reactions for abacavir and another study concluded that pre-prescription pharmacogenetics testing for HLA B*5701 allele is a cost-effective strategy (43,44). Another HLA B haplotype, HLA B*5801, is associated with a skin hypersensitivity syndrome with allopurinol in a Taiwanese population (45). These reports demonstrate that specific genotypes in HLA molecules are highly associated with increased risk for particular patterns of immune-mediated adverse events. The applicability of these findings for an individual patient depends on how similar the patient compares with the specific population in which the association was identified. Nonetheless, such high odds ratios for specific alleles suggest that a specific genotype plays a prominent role in mediating that specific type of adverse events and emphasizes the importance of the immune system in drug-associated adverse events. These studies exemplify the power of these kinds of association studies and are predicted to identify genes closely associated with drug-induced liver injury if used with the appropriate populations of patients. Role of Genotyping in Drug-Induced Liver Disease Currently, no recommendations exist for genotyping prior to initiation of a drug to reduce the incidence of a drug-induced liver injury. To reduce the incidence of adverse event prior to initiation of therapy, routine genotyping is recommended for TMPT genetic polymorphism prior to treatment with azathioprine or 6-mercaptopurine and UGT1A1 polymorphisms for those patients treated with an irinotecan-containing regiment. Although these examples do not involve hepatotoxicity, they illustrate how genotyping may be used in the future to reduce adverse events involving the liver. In both these cases, deficiency in metabolism of the chemotherapeutic drug due to polymorphisms that reduce enzyme activity leads to toxic accumulation of these agents. For thiopurine therapy, the TPMT*3A allele, in which nonsynonymous mutations (Ala154Thr and Tyr240Cys) lead to rapid degradation of the protein by ubiquination, effectively resulting in an almost complete absence of enzyme activity. Currently, approximately 11 variant alleles have been associated with low TPMT enzymatic activity in humans (46). In one study, no complications were reported at 39 weeks for those undergoing treatment for leukemia with wild-type genotype whereas adverse events were found in heterozygotes after two weeks of treatment. Homozygous carriers of low TPMT activity alleles require only 10% to 20% of the typical dose of 6-mercaptopurine or azathioprine and more diligent follow-up is required to monitor for bone marrow suppression (47). Polymorphisms of UGT1A1, including the relatively common promoter polymorphism responsible for Gilbert’s syndrome, can significantly influence the toxicity of the potent chemotherapeutic agent, CPT11. The active component of CPT11, SN28 is glucuronidated to an inactive metabolite by UGT1A1. Polymorphisms causing reduced activity or lower expression therefore lead to accumulation of the toxic metabolite with consequential increased toxicity. In both these examples, known polymorphisms that are responsible for decreased activity of these detoxification pathways can result in significant increases in drug levels and dose-dependent toxicities. Unfortunately, no known genotypes are so reliably associated with hepatotoxicity and currently there are no recommendations for genotyping of patients to reduce incidence of drug-induced liver injury (48). The ease of genotyping as compared to determining the actual pharmacokinetics of a specific CYP by using a model substrate to assess its enzymatic activity will greatly advance the ability to identify the significance of polymorphisms in drug-induced liver injury. Using microarray technology, the Amplichip CYP450 is a commercial and FDA-approved chip that genotypes 29 polymorphisms in the CYP2D6 gene including deletions and 2 polymorphisms of the CYP2C19 gene (49). Although it currently has no clinical applications, it may be used in drug

Global Profiling in Drug-Induced Hepatotoxicity

321

development for evaluation of a pharmacokinetic profile of a drug that is predominantly metabolized by these CYP family members. Future genotyping chip-based arrays could be developed if highly predictive genotypes are associated with specific adverse events involving the liver. Profiling of NAT2 polymorphisms prior to treatment with isoniazid would be a potential candidate. Drug-induced toxicity is likely to result from the integration of subtle variations in multiple pathways that ultimately modify pharmacokinetics and metabolism of the drug coupled with gene-associated susceptibility factors such as response to injury and repair processes. Environmental factors such as comorbid conditions, dietary factors, and coadministered drugs are also likely to be important modifiers. The role of genotyping in the reduction of the occurrence of rare drug-induced liver injury is not clear at this time. When whole genome sequencing for an individual becomes routine, bioinformatics analysis will encounter major hurdles when evaluating the potential significance of polymorphisms that do not change the amino acid sequence or are located within intronic and non-promoter regions of a gene.

FUTURE APPLICATION OF GLOBAL PROFILING ON EVALUATION OF DRUG-INDUCED LIVER INJURY Idiosyncratic drug-induced liver injury by definition is difficult to predict. The assessment of a drug’s ability to cause an adverse hepatic event in humans during drug development relies on a relatively small population as compared to the total number of individuals who will be eventually exposed to that drug. This initial evaluation in a restricted population limits the ability to identify an idiosyncratic drug-induced liver injury. Harvesting of DNA, metabonomic studies in the urine and serum along with proteomic analysis of serum samples from affected individuals may identify a unique signature associated with that toxicity. Haplotype analysis of those individuals as compared to non-affected individual could be used to identify individuals at greater risk for a particular pattern of hepatotoxicity based on their genotype (50). This approach avoids use of candidate genes and only identifies shared common regions of the genome between individual, whose genes can be further evaluated based on their potential physiological function. The significance of these genes could then be evaluated in different experimental models of toxicity. Other nongenomic approaches for predicting the risk for druginduced liver injury include metabonomic studies of urine and serum to characterize an individual’s metabolic capacity. In one study, pre- and post-urinary metabonomics were performed and compared to the risk for liver injury due to acetaminophen in rats given at a nontoxic threshold dose, as assessed by histology and serum transaminases (51). Relative ratio of conjugated acetaminophen metabolites in rats could be predicted based on pretreatment urinary levels of taurine, trimethylamine-N-oxide, and betaine, suggesting that metabolic capacity as evidenced by a specific pattern of urinary metabolites was associated with increased resistance to acetaminophen-induced injury. The potential to utilize global profiling earlier in the drug-development process has major implications for future drug development. Figure 2 summarizes how studies performed in rodent models can now be complemented by “omic” profiling, which are ultimately dependent on genomic databases. Information from individual toxins coupled with literature mining using iterative analysis will eventually support the development of integrative system–based toxicology of pharmaceutical agents. Genes identified in these pathways could be used to screen individuals with genotypes associated with increased risk for drug-induced liver injury, who would then be eliminated from preclinical studies with the anticipation of reducing incidence of drug-induced liver injury. This approach would eventually require that only patients lacking that genotype could be candidates for that drug. The feasibility of this approach would be dependent on multiple factors including the significance of the illness and the availability of other agents. Genotyping of the therapeutic target could also be used to eliminate those who are unlikely to respond to a specific agent. For example, a polymorphism in the b1-adrenergic receptor predicts responsiveness to b-receptor blockers (52). If one considers both polymorphisms in the therapeutic target in combination with genotypes associated with increased risk for drug-induced liver injury, initial evaluation of a new drug could be limited to those individuals with favorable genotypes. The sample size calculation for

322

Stolz

100 10110 0101001 101011010 01001010110 0101110111101 001010010010100

Treatment

-Omics database(s)

Gene, protein or metabolite-expression profiles Histopathology

'Sequence anchoring' of molecular expression

Literature mining

Computational analysis

Identify gene/protein functional groups, pathways and networks

Clinical chemistry

'Phenotypic anchoring' of molecular expression

Weight, physiology

Absorbtion,distribution, metabolism,excretion o

Genome database(s)

N

Multi-domain multi-genome knowledgebase

Iterative biological modelling Integrated systems toxicology

0 Toxicology database(s)

FIGURE 2 Overview of system-based toxicology. This figure outlines the integration of information from the initial observation in an animal model (rat in upper left) with routine histological, clinical chemistry, and pharmacokinetic parameters. These classic observations are now complemented by omic profiling, which are ultimately supported by genome databases. These findings are integrated into a toxicogenomics knowledge base (cylinder). These databases are continuingly refined and complemented by knowledge base integrating both classical and omic data streams, along with the literature-based knowledge, which will eventually lead to an integrated systems toxicology understanding. The framework combines both classical phenotypic characterization coupled to evolving sequencedependent analysis. Source: From Ref. 1.

these studies would be complex and would need to take into account the significance of the SNP effect on the drug and its frequency in the population. If multiple genes were responsible for the effect, this would further complicate the analysis. Post-marketing assessment could include collection of blood samples in anticipation of DNA analysis. In all studies involving collection of DNA and genotyping, care must be taken that patients are wellinformed and that consent is obtained as there are potential unforeseen social and legal implications for these studies (50). Issues such as privacy and confidentiality as well as access to the data and the eventual commercial or intellectual property rights must also be considered. The promise of discovery of genetic factors associated with an increased risk for drug-induced liver injury will hopefully usher in the era of personalized medicine, so that patients will be selected for specific drugs based on their genotype. REFERENCES 1. Waters MD, Fostel JM. Toxicogenomics and systems toxicology: aims and prospects. Nat Rev Genet 2004; 5(12):936–48. 2. Suter L, Babiss LE, Wheeldon EB. Toxicogenomics in predictive toxicology in drug development. Chem Biol 2004; 11(2):161–71. 3. Ekins S, Nikolsky Y, Nikolskaya T. Techniques: application of systems biology to absorption, distribution, metabolism, excretion and toxicity. Trends Pharmacol Sci 2005; 26(4):202–9. 4. http://www.genomesonline.org 21 September 2006. 5. International Human Genome Sequencing Consortium. Finishing the euchromatic sequence of the human genome. Nature 2004; 431(7011):931–45. 6. Robertson JA. The $1000 genome: ethical and legal issues in whole genome sequencing of individuals. Am J Bioeth 2003; 3(3):W–F1.

Global Profiling in Drug-Induced Hepatotoxicity

323

7. Shastry BS. Pharmacogenetics and the concept of individualized medicine. Pharmacogenomics J 2006; 6(1):16–21. 8. Walgren RA, Meucci MA, McLeod HL. Pharmacogenomic discovery approaches: will the real genes please stand up? J Clin Oncol 2005; 23(29):7342–9. 9. Marsh S, McLeod HL. Pharmacogenomics: from bedside to clinical practice. Hum Mol Genet 2006; 15(Spec No. 1):R89–93. 10. Jones PA, Takai D. The role of DNA methylation in mammalian epigenetics. Science 2001; 293(5532):1068–70. 11. The International HapMap Consortium. A haplotype map of the human genome. Nature 2005; 437(7063):1299–320. 12. Hinds DA, Stuve LL, Nilsen GB, et al. Whole-genome patterns of common DNA variation in three human populations. Science 2005; 307(5712):1072–9. 13. Karason A, Gudjonsson JE, Jonsson HH, et al. Genetics of psoriasis in Iceland: evidence for linkage of subphenotypes to distinct Loci. J Invest Dermatol 2005; 124(6):1177–85. 14. Helgadottir A, Manolescu A, Helgason A, et al. A variant of the gene encoding leukotriene A4 hydrolase confers ethnicity-specific risk of myocardial infarction. Nat Genet 2006; 38(1):68–74. 15. Foltynie T, Hicks A, Sawcer S, et al. A genome wide linkage disequilibrium screen in Parkinson’s disease. J Neurol 2005; 252(5):597–602. 16. Nuwaysir EF, Bittner M, Trent J, et al. Microarrays and toxicology: the advent of toxicogenomics. Mol Carcinog 1999; 24(3):153–9. 17. Gant TW. Classifying toxicity and pathology by gene-expression profile—taking a lead from studies in neoplasia. Trends Pharmacol Sci 2002; 23(8):388–93. 18. Butte A. The use and analysis of microarray data. Nat Rev Drug Discov 2002; 1(12):951–60. 19. Lugthart S, Cheok MH, den Boer ML, et al. Identification of genes associated with chemotherapy crossresistance and treatment response in childhood acute lymphoblastic leukemia. Cancer Cell 2005; 7(4):375–86. 20. Huang Q, Jin X, Gaillard ET, et al. Gene expression profiling reveals multiple toxicity endpoints induced by hepatotoxicants. Mutat Res 2004; 549(1–2):147–67. 21. Corthals GL, Nelson PS. Large-scale proteomics and its future impact on medicine. Pharmacogenomics J 2001; 1(1):15–19. 22. Maercker C. Protein arrays in functional genome research. Biosci Rep 2005; 25(1-2):57–70. 23. Griffin JL. Metabonomics: NMR spectroscopy and pattern recognition analysis of body fluids and tissues for characterisation of xenobiotic toxicity and disease diagnosis. Curr Opin Chem Biol 2003; 7(5):648–54. 24. Lazarou J, Pomeranz BH, Corey PN. Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA 1998; 279(15):1200–5. 25. Waring JF, Ciurlionis R, Jolly RA, et al. Microarray analysis of hepatotoxins in vitro reveals a correlation between gene expression profiles and mechanisms of toxicity. Toxicol Lett 2001; 120(1–3):359–68. 26. Morgan KT. Gene expression analysis reveals chemical-specific profiles. Toxicol Sci 2002; 67(2):155–6. 27. Ruepp S, Boess F, Suter L, et al. Assessment of hepatotoxic liabilities by transcript profiling. Toxicol Appl Pharmacol 2005; 207(Suppl. 2):161–70. 28. Ruepp SU, Tonge RP, Shaw J, et al. Genomics and proteomics analysis of acetaminophen toxicity in mouse liver. Toxicol Sci 2002; 65(1):135–50. 29. Craig A, Sidaway J, Holmes E, et al. Systems toxicology: integrated genomic, proteomic and metabonomic analysis of methapyrilene induced hepatotoxicity in the rat. J Proteome Res 2006; 5(7):1586–601. 30. Waters NJ, Waterfield CJ, Farrant RD, et al. Integrated metabonomic analysis of bromobenzeneinduced hepatotoxicity: novel induction of 5-oxoprolinosis. J Proteome Res 2006; 5(6):1448–59. 31. Soga T, Baran R, Suematsu M, et al. Differential metabolomics reveals ophthalmic acid as an oxidative stress biomarker indicating hepatic glutathione consumption. J Biol Chem 2006; 281(24):16768–76. 32. http://www.niehs.nih.gov/envgenom/home.htm 07 May 2004. 33. Guengerich FP. Cytochromes P450, drugs, and diseases. Mol Interv 2003; 3(4):194–204. 34. Hiratsuka M, Sasaki T, Mizugaki M. Genetic testing for pharmacogenetics and its clinical application in drug therapy. Clin Chim Acta 2006; 363(1–2):177–86. 35. http://www.cypalleles.ki.se/index.htm 27 March 2007. 36. Barclay ML, Sawyers SM, Begg EJ, et al. Correlation of CYP2D6 genotype with perhexiline phenotypic metabolizer status. Pharmacogenetics 2003; 13(10):627–32. 37. Shastry BS. Genetic diversity and new therapeutic concepts. J Hum Genet 2005; 50(7):321–8. 38. Evans WE, Johnson JA. Pharmacogenomics: the inherited basis for interindividual differences in drug response. Annu Rev Genomics Hum Genet 2001; 2:9–39. 39. Kinzig-Schippers M, Tomalik-Scharte D, Jetter A, et al. Should we use N-acetyltransferase type 2 genotyping to personalize isoniazid doses? Antimicrob Agents Chemother 2005; 49(5):1733–8.

324

Stolz

40. Hiratsuka M, Kishikawa Y, Takekuma Y, et al. Genotyping of the N-acetyltransferase2 polymorphism in the prediction of adverse drug reactions to isoniazid in Japanese patients. Drug Metab Pharmacokinet 2002; 17(4):357–62. 41. Huang YS, Chern HD, Su WJ, et al. Polymorphism of the N-acetyltransferase 2 gene as a susceptibility risk factor for antituberculosis drug-induced hepatitis. Hepatology 2002; 35(4):883–9. 42. Hung SI, Chung WH, Jee SH, et al. Genetic susceptibility to carbamazepine-induced cutaneous adverse drug reactions. Pharmacogenet Genomics 2006; 16(4):297–306. 43. Hughes DA, Vilar FJ, Ward CC, et al. Cost-effectiveness analysis of HLA B*5701 genotyping in preventing abacavir hypersensitivity. Pharmacogenetics 2004; 14(6):335–42. 44. Rauch A, Nolan D, Martin A, et al. Prospective genetic screening decreases the incidence of abacavir hypersensitivity reactions in the Western Australian HIV cohort study. Clin Infect Dis 2006; 43(1):99–102. 45. Hung SI, Chung WH, Liou LB, et al. HLA-B*5801 allele as a genetic marker for severe cutaneous adverse reactions caused by allopurinol. Proc Natl Acad Sci USA 2005; 102(11):4134–9. 46. Krynetski E, Evans WE. Drug methylation in cancer therapy: lessons from the TPMT polymorphism. Oncogene 2003; 22(47):7403–13. 47. Koo SH, Lee EJ. Pharmacogenetics approach to therapeutics. Clin Exp Pharmacol Physiol 2006; 33(5-6):525–32. 48. Evans WE. Pharmacogenomics: marshalling the human genome to individualise drug therapy. Gut 2003; 52(Suppl. 2)ii10–18. 49. Juran BD, Egan LJ, Lazaridis KN. The AmpliChip CYP450 test: principles, challenges, and future clinical utility in digestive disease. Clin Gastroenterol Hepatol 2006; 4(7):822–30. 50. Brazell C, Freeman A, Mosteller M. Maximizing the value of medicines by including pharmacogenetic research in drug development and surveillance. Br J Clin Pharmacol 2002; 53(3):224–31. 51. Clayton TA, Lindon JC, Cloarec O, et al. Pharmaco-metabonomic phenotyping and personalized drug treatment. Nature 2006; 440(7087):1073–7. 52. Liggett SB, Mialet-Perez J, Thaneemit-Chen S, et al. A polymorphism within a conserved beta(1)adrenergic receptor motif alters cardiac function and beta-blocker response in human heart failure. Proc Natl Acad Sci USA 2006; 103(30):11288–93.

16

Causality Assessment Ma Isabel Lucena

Departmento de Farmacologia Clinica, Hospital Virgen de la Victoria, School of Medicine, MÆlaga, Spain

Raœl J. Andrade, Raquel Camargo, and Miren Garc a-CortØs

Liver Unit, Hospital Virgen de la Victoria, MÆlaga, Spain

INTRODUCTION Liver disease caused by drugs or toxins remains a challenge of modern hepatology. This is not only because the pathogenesis and susceptibility factors for idiosyncratic toxic liver damage are still poorly understood, but also because of the lack of reliable and standardized markers for toxic liver damage (1). The diagnosis of drug hepatotoxicity then relies upon circumstantial evidence of exposure to a potential hepatotoxin as well as the exclusion of other causes of liver injury (2). This complex process requires a systematic approach and the use of causality assessment methods, in order to add consistency to the diagnostic process by translating the suspicion into a quantitative score and by providing a framework which emphasizes the features that merit attention in cases of suspected hepatic adverse reaction. A widely acceptable diagnostic algorithm would allow unified criteria among clinicians, regulatory agencies, and pharmaceutical companies. This issue is critical because the consequences of hepatotoxicity may be catastrophic, leading to acute liver failure and death. Rapid decisions (including warnings and withdrawals), especially if based on high accuracy and completeness of data, may prevent further severe cases. However, none of the available assessment methods has been validated as yet due to the absence of an accepted “gold standard” for the diagnosis of hepatotoxicity, and currently neither eliminates nor quantifies uncertainty and has only limited scientific value (3). DIAGNOSIS IN THE CLINICAL SETTING A straightforward diagnosis of hepatotoxicity in clinical practice can only be made in rare circumstances. An example is when symptoms of hepatitis rapidly ensue after the obvious exposure to overdose of agents capable of inducing intrinsic hepatic damage, including drugs such as acetaminophen or aspirin as well as herbal agents (e.g., Amanita phalloides) or industrial hepatotoxins (carbon tetrachloride). In most of these cases, there is close relationship between the severity of the liver injury and blood level of the compound, which may be used to confirm the suspicion. In other cases, the diagnosis of drug-induced hepatotoxicity may also be easily made because evidence of liver damage becomes apparent after re-exposure to a drug suspected to be the cause of previous hepatitis. The topic of re-challenge is discussed in more detail below. Direct evidence for idiosyncratic hepatotoxicity is rarely available. This includes the detection of serum circulating autoantibodies to specific forms of cytochrome P450 for certain drugs (Table 1), most of which have been withdrawn from the market. In addition to uncertainty with regard to sensitivity and specificity of this test, the fact that it can be applied only to The authors of this chapter have relationships with the following corporations: Raul J. Andrade is advisor for Schering Plough and Bristol Myers Squibb and consultant for Novartis and Merck Sharp and Domme, and has received research grants from Roche Pharma and Schering Plough. Raquel Camargo has received research grants from Roche. Miren Garcı´a-Corte´s has received research grants from Roche and ScheringPlough.

326

Lucena et al.

TABLE 1 Specific Targets for Diagnosis of Drug-Induced Liver Injury Antimitochondrial (anti-M6) autoantibody Anti-liver kidney microsomal 2 antibody (anti-LKM2) Anti CYP 1A2 Anti CYP 2E1 Anti-liver microsomal autoantibody Anti-microsomal epoxide hydrolase

Iproniazid Tienilic acid Dihydralazine Halothane Carbamazepine Germander

Abbreviation: CYP, cytochrome-P450.

the above-mentioned drugs makes it nowadays irrelevant in clinical practice relegating it to an investigational tool (4). The in vitro lymphocyte-stimulation test is another tool that has been used in the search for evidence of drug allergy. In this test, lymphocyte proliferation is counted after exposure of peripheral blood mononuclear cells from the patient to the suspected drug by the use of radiolabeled thymidine incorporation in the presence of a prostaglandin inhibitor (such as indomethacin) to prevent the suppressive influence of activated monocytes on T-cells (5,6). Although a positive response has long been considered specific evidence that the drug is responsible for the injury (7), and recent studies in large series of patients have been promising (5), such positive results only indicate sensitization towards a certain drug but cannot actually be related to effectors mechanism (symptoms) and a negative test does not exclude drug allergy (8). Ultimately, in vitro tests are difficult to standardize, poorly reproducible between laboratories, and have not gained clinical acceptance (6,8). Nevertheless, Kaplowitz in a recent review has encouraged a global effort to validate and test this approach (9) and we believe that all would agree. In the remaining situations, a lesser degree of uncertainty cannot be avoided. It is important a look for the presence of risk factors that modify susceptibility for drug-induced liver disease (chap. 14), since some of them (older age, alcohol consumption, pregnancy) are taken into account in one of the diagnostic scales now being used for causality assessment (see below). Several items should be taken into account in a “step-by-step” approach (Fig. 1).

Liver disease Suspicion

Drug exposure data and chronology

Not compatible

If compatible Hepatotoxic potential

Search for an alternative diagnosis Not found Assess presence of incriminating criteria Allergy drug features Course on “dechallenge” Course on rechallenge Liver biopsy findings and Biochemical signature

Found

Specific management

FIGURE 1 Schematic algorithm for the diagnosis of drug-induced liver injury.

Causality Assessment

327

Drug Exposure Data and Chronology A thorough drug and chemical history is essential. Prescribed and over-the-counter medications need to be recorded. Consumption of illicit drugs should also be assessed. Careful questioning is crucial because sometimes patients have taken drugs or complementary medicines provided by friends or neighbors. Requesting medication containers or a written medication plan, when available, may reduce the possibility of recall errors by the patient (10). A special difficulty occurs in unconscious or confused patients or in those who are not able to cooperate. In such cases, relatives or caregivers should be interviewed. After checking for exposure to a medication or herbal product, the question is whether treatment was clearly begun before presentation or in the early phase of hepatitis. This obvious point is often overlooked, as a mistake may easily arise when nonspecific gastrointestinal symptoms antedate clinically overt jaundice or even dark urine, hampering the recognition of the true onset of the liver dysfunction. The suspected drug could actually have been prescribed to alleviate these first symptoms of hepatitis. After excluding this possibility, duration of therapy with the suspected drug must be assessed. The usual time interval between the beginning of treatment and the onset of liver injury varies widely among drugs yet it appears to have a somewhat consistent “signature” regarding latency period for each drug, which is linked to the involved mechanism of damage. For instance, intrinsic hepatotoxins induce overt liver damage a few hours after exposure. In most idiosyncratic cases, the latency period is roughly between one week and three months. A shorter period (one or two days) may occur in patients who have been previously treated with the drug and have become sensitized (allergic hepatotoxicity). In general, allergic hepatic reactions are likely to occur during the first one to five weeks of taking the drug. A delay longer than three months is typically seen with compounds that act by nonallergic mechanism, the so-called “metabolic idiosyncrasy.” On the other hand, drug-induced acute hepatitis seldom occurs after 12 months of exposure, these long-latency periods are still possible in unusual forms of chronic liver damage (such as steatohepatitis, fibrosis, and chronic hepatitis) in which the expression of hepatotoxicity is symptom-less, enabling treatment to persist (11–17), or simply because the type of lesion requires prolonged exposure to develop (e.g., vascular lesions and tumors) (18,19). In some instances, the role of a drug is difficult to recognize because of a considerable delay between the discontinuation of therapy and clinical presentation. A classical example is hepatitis associated with halothane and its derivatives, which typically occurs three weeks after the first exposure. The phenomenon is extreme with amoxicillin–clavulanate that may cause hepatitis several weeks after cessation of the drug (20), but has also recently been reported with midecamycin (21) and trovafloxacin (22)—an antibacterial agent removed from the European market and remaining in the U.S.A. under very restricted indications. The reasons for this are unclear, but such an unusual time might be a combination of a late immune response to the drug with its prolonged retention in the body (23). There is not a clear rule to incriminate one drug versus another one in patients taking various medications simultaneously. Attention should be paid to the latest one introduced, which will usually be the responsible one. However, when a known hepatotoxic drug antedates the latest introduced medication, it seems reasonable to ascribe the picture to the combination of them because of the possibility of pharmacokinetic interactions (10,24). Hepatotoxic Potential Although almost any marketed drug has been incriminated in incidences of hepatotoxicity (25), their potential for causing liver damage is not the same. For instance, whereas some drugs like isoniazid, diclofenac, or amoxicillin—clavulanate, which has been recognized as the drug most frequently implicated in hepatotoxicity in a recent large case-series (10), are wellknown hepatotoxic agents, others rank very low or not at all in the list of hepatotoxins. Thus, there is no convincing evidence for implicating medications such as digoxin, insulin, or streptomycin, which have been extensively used for decades, in toxic liver damage (7).

328

Lucena et al.

Information about the likelihood of liver damage from many other drugs is scanty or vague. In the summary of product characteristic, phrases such as “can produce transient increase in transaminases” or “may cause hepatitis rarely” are common. A search of databases of hepatotoxic drugs such as HEPATOX from France (25) or lists in reference books (7,26,27) may provide valuable information. The Medline–PubMed database of the National Library of Medicine is a more up-to-date resource, in which crossing the name of the drug together with the terms “hepatotoxicity,” “hepatitis,” “drug-induced hepatotoxicity,” or just “liver” if the former do not yield details, can be useful. A special difficulty arises when the patient has been taking a newly marketed drug. In these cases, data regarding the medication’s hepatotoxic potential, if known, are only available in pre-approval clinical trials (see chap. 36 for further details). To detect significant (clinically overt) liver disease due to a drug with 95% confidence in these studies, the number of patients exposed must be at least three times the true incidence (“rule of threes”) (28). Finding hepatotoxicity with an incidence of 1:10,000 (the approximate incidence of most idiosyncratic reactions) would therefore require 30,000 treated patients. Phase III clinical trials usually involve 1500 to 2500 patients. As a result, the occurrence of overt hepatitis during clinical development of a drug is very unlikely and in any case usually indicates an unacceptably high hepatotoxic potential as well as the need to stop further development. Nevertheless, because each phase of clinical testing before approval includes close monitoring of liver tests, the appearance of less prominent signals of liver damage in preapproval studies should be carefully sought. These include the incidence of asymptomatic alanine aminotransferase (ALT) and bilirubin elevations. The magnitude of the increase in transaminases is especially meaningful. An ALTReightfold the maximum normal value or an increase in direct bilirubin (R1.5-fold), especially if it is accompanied by a raised ALT, deserves attention since this rarely occurs in the control population. Clinical development of troglitazone serves to illustrate the relevance of identification of hepatotoxicity signals in predicting further post-marketing incidents (29). In 2500 patients participating in phase III studies, ALTOeightfold occurred in 0.6% of the treated patients (vs. none of the control patients), and two cases of acute cytolytic hepatitis were observed (30). Thus, it could be anticipated that at least 1 out of 1250 treated patients would suffer from significant troglitazone hepatotoxicity. Moreover, this issue is critical since Zimmerman predicted that roughly 10% jaundiced patients with acute hepatocellular drug-induced liver injury will ultimately go into fulminant hepatic failure, “Hy’s Rule” (7). Such a figure has been recently validated in a large cohort of patients with drug-induced idiosyncratic hepatotoxicity, in which mortality (or its surrogate, liver transplantation) of patients with hepatocellular jaundice was 11.7% (10). This was, actually, the case with troglitazone, which was withdrawn after its implication in instances of severe hepatitis and acute liver failure leading to death or liver transplantation (29). Exclusion of Alternative Causes Exposure to one or more drug does not in itself prove causality (31). Indeed, paraphrasing Lee “as with hardened criminals, the fact that a drug is capable of causing an adverse effect does not mean that it did” (32). Therefore, the exclusion of alternative etiologies is mandatory to incriminate a specific drug as the cause of liver dysfunction (Table 2). Diagnostic evaluation of any patient with acute liver disease of unknown origin should include a careful history to exclude alcohol abuse, recent episodes of hypotension (ischemic hepatitis), epidemiological risk factors of infectious hepatitis (recent travel to areas of endemic hepatitis, intravenous drug addiction, blood transfusion, or recent surgery), specific serology and molecular biology studies for common viruses involved in viral hepatitis, as well as screening for autoimmune hepatitis. Although hepatitis C infection less commonly presents as overt hepatitis, RNA test by polymerase chain reaction (PCR) in addition to anti-HCV by ELISA should be used to screen this possibility, because the “window” for detectable antibodies may be as long as eight weeks in acute hepatitis C. All patients should also have an abdominal ultrasound to exclude mechanical biliary obstruction.

329

Causality Assessment

TABLE 2 Workup for Exclusion of Alternative Causes of Liver Disease Test interpretation

Diseases

Viral serology: IgM anti-HAV, IgM anti-HBc, anti-HCV, RNA-HCV(RT-PCR), IgM-CMV, IgM-EBV, herpes virus

Viral hepatitis

Bacterial serology: Salmonella, Campylobacter, Listeria, Coxiella Serology for syphilis

Bacterial hepatitis

Autoimmunity (ANA, ANCA, AMA, ASMA, anti-LKM-1)

Autoimmune hepatitis, primary biliary cirrhosis

AST/ALT ratioO2

Alcoholic hepatitis

Ceruloplasmin, urine cooper a-1-Antitrypsin Transferrin saturation

Wilson’s disease Deficit of a-1antitrypsin Hemochromatosis

Hyperechoic liver (ultrasound)

Nonalcoholic steatohepatitis

Very high transaminase levels

Ischemic hepatitis

Dilated bile ducts by image procedures (AU, CT, MRCP, ERCP)

Biliary obstruction

Secondary syphilis

Clinical features commentary Less frequent in older patients, especially hepatitis A. Search for epidemiologic risk factors. Outcome may be similar to that of drug-induced idiosyncratic liver injury after de-challenge If persistent fever and/or diarrhea Multiple sexual partners. Disproportionately high serum AP levels Women, ambiguous course after de-challenge. Other autoimmunity features Alcoholics. Moderate increases in transaminases despite severity at presentation Patients!40 years Pulmonary disease Anicteric hepatocellular damage. Middleaged men and older women In anicteric hepatocellular damage. Obesity, metabolic syndrome Disproportionately high AST levels. Hypotension, shock, recent surgery, heart failure, antecedent vascular disease, elderly Abdominal pain, cholestatic/mixed pattern

Abbreviations: ALT, alanine aminotransferase; AP, alkaline phosphatase; AST, aspartate aminotransferase; AU, abdominal ultrasonographic examination; Anti-HAV, hepatitis A antibody; Anti-HBc, hepatitis B core antibody; Anti-HCV, hepatitis C antibody; anti-LKM-1, Liver kidney microsomal antibody type 1; AMA, antimitochondrial antibody; ANA, antinuclear antibody; ANCA, perinuclear antineutrophil cytoplasmic antibody; ASMA, anti-smooth muscle antibody; CMV, citomegalovirus; CT, computed tomography; EBV, Epstein Barr virus; ERCP, endoscopic retrograde cholangiography; MRCP, magnetic resonance cholangiography.

The appropriateness of additional studies relies on the presence of particular symptoms or analytical features. Patients who exhibit prominent gastrointestinal and systemic manifestations such as diarrhea or persistent fever should also be screened for bacterial infections including Campylobacter, Salmonella, and Listeria. In patients with lymphadenopathy, unusual viruses leading to hepatitis such as Epstein–Barr, cytomegalovirus, or herpes virus should also be tested for. On the other hand, the presence of high titers of antinuclear or anti-smooth muscle antibodies and hyperglobulinemia do not in itself allow the diagnosis of autoimmune hepatitis as many toxic hepatic reactions to drugs are accompanied by laboratory features of autoimmunity, and actually drugs may occasionally trigger an autoimmune liver disease. Response to de-challenge can tentatively distinguish between a true autoimmune hepatitis and a toxic hepatitis reaction. However, in cases with ambiguous de-challenge a liver biopsy may help to distinguish between these two clinical situations (Table 3). Liver biopsy cannot reliably distinguish what features are being referred to. In younger patients, metabolic liver disease such as Wilson’s disease should also be considered and excluded by assessing serum ceruloplasmin levels and urinary copper excretion. Chronic abnormalities of liver tests may also require either a screen for hemochromatosis or even a liver biopsy to exclude nonalcoholic steatohepatitis, sometimes coexisting with toxic liver damage because both of them are frequent in the general population. Patients with the cholestatic or mixed pattern of hepatic injury may require imaging by magnetic resonance cholangiography or endoscopic retrograde cholangiography, despite normal abdominal ultrasound findings, to exclude benign or malignant obstruction of the biliary tract. Rarely, secondary syphilis may present with prominent hepatic cholestasis featuring disproportionately high serum levels of alkaline phosphatase (33). Therefore, patients at risk for

330

Lucena et al.

TABLE 3 Key Concepts in the Diagnosis of Drug-Induced Idiosyncratic Liver Injury (DILI) DILI mimics all forms of acute and chronic hepatobiliary diseases There are no specific markers or tests for DILI and the diagnosis requires high degree of suspicion Importance of recording baseline liver tests for a proper definition of hepatotoxicity Exclusion of other causes of liver injury according to age and type of damage is of paramount importance Recognition of accompanying hypersensitivity features should not be overlooked. Determination of the lymphocyte-stimulation tests may be promising if procedures are standardized and validated Routine liver biopsy should not be indicated for diagnostic confirmation, but for prognostic reasons and for clarifying ambiguous cases where an alternative explanation is possible The importance of an accurate pharmacological history is crucial. Any new or even old drug not previously incriminated in hepatotoxicity may be responsible Last drug attributable due to temporal sequence may not necessarily be the culprit Increased consumption of complementary and alternative medicines that are not necessarily safe may go unrecognized A drug signature (clinical, pathologic, and latency presentation) may vary Idiosyncratic hepatotoxicity may be dose related: search for changes in the pharmacokinetic properties due to decline in renal function or a drug drug metabolism or transporter interaction Underlying liver disease must not prevent physicians in recognizing DILI Recognize the likelihood of drug toxicity among individual agents and drug classes ( cross-reactivity ) Foster patient education directed at early recognition of the symptoms

sexually transmitted diseases should also have appropriate serological testing for syphilis. Finally, it should also be remembered that the mixed hepatocellular–cholestatic pattern of injury is far more characteristic of drug-induced hepatotoxicity than of viral hepatitis (7). Incriminating Criteria Once alternative causes of liver injury have been ruled out, the role of one or more drugs can be reinforced by a careful scrutiny of the presence of drug-allergy features, the course after drug withdrawal and after a challenge dose, as well as by typical biopsy findings or biochemical patterns of liver damage. Drug-Allergy Features As clinical or laboratory features of drug-induced hepatic injury are indistinguishable from those from other etiologies, focusing on unusual manifestations may be helpful in the diagnostic evaluation. Features that suggest drug-allergy include, skin rash, fever, peripheral eosinophilia, short-latency period (one month or less), and rapid symptoms recurrence on re-exposure to the drug. Hematological features including granulocytopenia, thrombopenia, or hemolytic anemia as well as renal and pancreatic involvement may also accompany some instances of drug-induced immunoallergic hepatic injury (34,35). In rare cases, the extreme skin involvement of Stevens-Johnson syndrome or the Lyell syndrome are strong clues to drug hypersensitivity as the cause of liver disease (7). However, because these manifestations occur in a minority of cases of hepatotoxicity, their absence is not a helpful sign. Actually, any of the hallmarks of hypersensitivity (fever, rash, eosinophilia, cytopenia) were present only in 106 among 446 cases (23%) with idiosyncratic hepatotoxicity submitted to the Spanish Registry (10). Drug-allergy manifestations are associated with hepatotoxicity in a widely variable rate, mainly depending on the drug class. Some drugs that act by an allergic mechanism may induce hepatic injury in only very rare instances (e.g., penicillin) (36). Other medications such as phenytoin frequently cause hepatic damage in the context of a syndrome of generalized hypersensitivity. For other types of drugs, systemic manifestations of drug allergy are present in approximately half of the patients. This is the case, for example, in halothane or chlorpromazine allergy. Finally, in some instances, typical hypersensitivity symptoms are absent but clues pointing toward an immunoallergic reaction might come from the presence of more subtle features such as detectable serum autoantibodies, antinuclear, and anti-smooth muscle antibody (17,37). In fact, it is likely that immunologic and metabolic idiosyncrasy operate concurrently in many cases of drug-induced hepatic injury (4,7).

Causality Assessment

331

Course After Drug Withdrawal (De-challenge) Rapid improvement after withdrawal of pharmacological therapy speaks in favor of a toxic etiology of the liver disease, even though this may be also seen in viral hepatitis. Nevertheless, in cases of hepatocellular injury, the role of a drug is strongly supported by a decrease of at least 50% in the levels of liver enzymes in the first eight days after stopping therapy (7). Although less conclusive, an expert consensus still considered suggestive of drug etiology (and positively weighted it in a clinical scale), a decrease of at least half in the levels of liver enzymes if it occurs within 30 days after cessation of exposure (38). Physicians who are unfamiliar with drug hepatotoxicity should be aware, however, that other atypical outcomes make laboratory scrutiny unhelpful after drug withdrawal; in general, cholestatic reactions subside more slowly, with abnormal enzyme levels persisting for long periods, even more than one year in some instances (36,39). Particularly confusing is also the evolution of some severe cases in which the injury may progress over several days despite drug cessation, or even evolve to fulminant hepatic failure (40,41). Unfortunately, the bigger the need, the lesser the certainty. This outcome can have a great impact in the adoption of regulatory measures, but it poses a special impediment to attempts to determine etiology. Conversely, a phenomenon of “adaptation” to injury occurs with some drugs (e.g., statins) and is responsible for the spontaneous improvement in liver tests despite treatment continuation (28). The Response to Re-administration (Re-challenge) At present, the only way to confidently confirm the role of a drug in the development of liver disease is to demonstrate a recrudescence of liver injury after re-exposure to the suspected agent. A positive response after exposure can be defined as doubling ALT and alkaline phosphatase (AP) values for hepatocellular and cholestatic reactions, respectively. From a practical standpoint, however, it is hard to demonstrate this in most circumstances in which unintentional re-exposure occurs. Furthermore, a history of inadvertent re-challenge may need careful inquiry because symptoms of overt liver disease may not accompany the index episode. In these particular cases, typically jaundice did not occur at the first episode, there being only nonspecific symptoms (malaise, fever) that could incriminate the drug. Besides, voluntary re-challenge carries several practical problems. First, re-challenge can be hazardous, especially in drug-induced hepatocellular hepatitis with associated hypersensitivity features, since the provoked reaction is often more severe than the initial episode. Hence, this approach is strictly contraindicated in such cases. Second, the amount of drug required to provoke the reaction is unknown. Arbitrarily, a single dose would be chosen; however, more than a few doses may be necessary to reproduce liver damage in hepatotoxicity caused by the progressive accumulation of toxic derivatives. This false negative response to re-challenge has been demonstrated for isoniazid (7), but is probably applicable for many other drugs that cause liver damage by an idiosyncratic, nonallergic mechanism. Finally and most importantly, re-exposure of the patient to the suspected drug cannot be used for pure diagnostic reasons. It should be considered only when the drug is deemed essential, as in the treatment of tuberculosis with isoniazid, and after written informed consent is obtained from the patient. Nevertheless, we think that because of the consequences of overlooking hepatotoxicity in pre-approval clinical trials (37), testing clinical or sub clinical hepatitis in this setting might be an additional indication for re-challenge. The Role of Liver Biopsy and the Drug “Signature” A common misconception among practicing physicians is the belief that a diagnosis of drug-induced hepatotoxicity cannot confidently be established without a liver biopsy. In fact, there are no histologic findings that can be considered absolutely specific for the diagnosis of drug-induced hepatotoxicity (chap. 13 for further details), and hence liver biopsy should not be routinely performed for this indication (42,43). Indeed, in many cases in which chronological criteria and exclusion of alternative causes seem to incriminate a drug, a liver biopsy specimen, often taken several days after presentation as pathological features begin to wane, does not add any useful information, generating on the contrary perplexity and confusion.

332

Lucena et al.

At present, the reasons for performing a liver biopsy in patients with suspected druginduced hepatotoxicity may be restricted to the following (44). First, when a toxic etiology is less likely and other alternative causes should be ruled out. In practical terms, this occurs when the patient may have an underlying liver disease and it is difficult to ascribe the picture to the candidate drug or to a recrudescence of the disease. Alternatively, biopsy is used to characterize the pattern of injury with drugs not previously incriminated in hepatotoxicity (16,22,37). We think that another important reason to perform a liver biopsy in patients with suspected drug-induced hepatotoxicity is to identify more severe or residual lesions (e.g., fibrosis) that could have prognostic significance. For instance, in some chronic variants of hepatotoxicity, clinical and laboratory features are poorly indicative of the severity of underlying liver damage (16,36). If one of these situations is suspected, a liver biopsy is helpful to identify the true magnitude of the hepatic lesion. Furthermore, severe bile duct injury during cholestatic hepatitis has been shown to be predictive of evolution to chronic cholestasis (45), as was also shown in a retrospective study (46) for the presence of fibrosis in the index liver biopsy for the development of chronic liver disease (46). Although not a reason to do a biopsy, finding eosinophilic infiltration or granuloma may be helpful in supporting drug-induced idiosyncratic liver injury. Because a liver biopsy is not available in most cases, focus on the biochemical expression of hepatic damage may help in incriminating a drug. It has been emphasized that each drug has its own pattern of liver injury “signature” (28,47). Although this is true for some drugs (e.g., estrogens induce cholestatic injury and seldom any other pattern of damage), for most others drugs, consistency is not so straightforward. For instance, amoxicillin–clavulanate tends to produce cholestatic or mixed damage, although hepatocellular damage has also been reported frequently (48). Hepatocellular and cholestatic or mixed injury have been noted with nimesulide (33,49) or troglitazone (50,51), to mention only a few examples. Therefore, it is important for physicians to view a suspicion of hepatotoxicity with the awareness that a given drug may produce diverse types of injury (52).

CAUSALITY ASSESSMENT Regarding hepatotoxicity in research settings, diagnostic precision and objectivity are obviously essential and accuracy has become paramount. But this is not less important from the perspective of the practicing clinicians, who must decide whether to continue or stop a therapy thought to be the most appropriate and that might induce new events in the future if it is not correctly identified. Except for the very rare circumstances in which an unintentional positive re-challenge may confirm the putative role of a drug, the evidence collected is circumstantial, often uses subjective impressions based on previous experiences and can frequently lead to inaccurate diagnosis (53). Several facts complicate the issue of establishing whether there is a causal relationship between a liver disease and the exposure to a given drug. The absence of an accepted “gold standard” (a definitive cause–effect test to define truth), but also the complex decisions required to attribute causality that are not made in a systematic manner or a fulminant course of the disease where a de-challenge course is missing. Clinical judgment is a necessary first step in the identification of any hepatic disease suspected to be caused by a drug or other toxin. The fact that the diagnosis of hepatotoxicity, in contrast to other diseases, has not yet benefited from the advances in genomics, proteomics, and imaging tests and still relies upon careful history taking and a correct interpretation of the patient’s clinical manifestations and laboratory data, should not be forgotten (Table 3). An approach to diagnosis that does not follow objective guidelines, which complements standard clinical practice, results in categories of causation that can be defined as drug related (e.g., acetaminophen over dosage, cases of positive re-challenge), clearly nondrug– related (an alternative explanation found) or possibly drug related (54). This last judgment in which subjectivity prevails and in which there is no strong argument for confirming or ruling out drug causality represents the bulk of situations in clinical practice. In addition, this judgment closely depends on physician’s skill and attitude toward the matter. Consequently,

333

Causality Assessment

agreement among physicians who evaluate a given case of drug-suspected hepatotoxicity may strongly differ. Variations in data consistency, completeness, and subjective weighting of causality arguments would presumably result in these differences. Algorithms or Clinical Scales In the last three decades, several groups have attempted to develop methods to improve consistency, accuracy, and objectivity in causality assessment of adverse drug reactions (Table 4). The qualities usually required of any scoring system are reproducibility and validity (55). Reproducibility ensures an identical result regardless of who the user is and when the scale is used. Validity refers to the capacity to consistently give a fair approximation of the truth, thus the method is able to distinguish between cases where the drug is responsible and cases where it is not. There exist two possible categories of approach. Probabilistic approaches (56) based on Bayesian statistics are rarely used in routine clinical practice since they require precisely quantified information to model probability distributions for each parameter. Instead, algorithms or clinical scales are widely used (57). They have attempted to identify the key features of an adverse reaction and to integrate them into an objective rating scale, based on the sum of weighted numerical values assigned to individual axes of decision strategy. The scores are translated into categories of suspicion. The different causality assessment methods developed can produce diverse numerical scales that may not be superimposable and with categories of suspicion not identical in name, making difficult the comparisons among them (57). We shall comment on the most relevant ones applied to the field of hepatotoxicity. The first to develop operational diagnostic criteria for the identification of adverse drug reactions were Karch and Lasagna (1977) (58). The major limitations of this approach are the impossibility of identifying the first case of new adverse reactions (they fall into the category of conditional), and the need for a subjective judgment in many of the steps (disagreement is easier). Because of its implicit mode of making decisions, this method has been criticized for being uncontrolled, imprecise, and prone to subjective bias. An alternative algorithm, the Naranjo Adverse Drug Reactions Probability Scale (1981) offers the advantage of simplicity and wide applicability (59). This scale involves 10 “yes,” “no,” or “unknown or inapplicable” questions concerning the areas: temporal relationship, competing causes, de-challenge/re-challenge results, knowledge of drug’s reactions, and in addition to these universally accepted criteria: placebo challenge, drug concentrations, and objective measurement of adverse drug reaction (Table 5). The adverse drug reaction is assigned to a probability category on the basis of the total score as follows: definiteR9, likely 5 to 8, possible 1 to 4, doubtful %0. The scale was validated, and led to improved reproducibility in assessments. The main source of inter-observer disagreement was the question about alternative causes, perhaps because of the complex clinical situations and differences in training between observers. The scale is being used to evaluate adverse drugrelated events prior to publication in some medical journals, as well as to assess reports submitted to national drug-monitoring centers.

TABLE 4 Advantages and Limitations of the Standardized Causality Assessment Methods in the Evaluation of Cases of Suspected Hepatotoxicity Advantages Improve the consistency of judgments Offer objectivity, specificity, and reproducibility Mechanism to grade the strength of the final designation in broad categories Excellent teaching tool Greatest clinical value in the assessment of complex patients and in research settings Signal detection and adoption of regulatory measures

Drawbacks Complexity and time consuming Do not provide certainty Do not substitute common sense clinical judgment Need cases with enough relevant information and follow-up Do not discriminate among concomitant potentially hepatotoxic drugs Evaluations of fatal cases or atypical presentation remain challenging

From drug withdrawal until onset event Compatible Course of the reaction Highly suggestive Suggestive Compatible Against the role of the drug Inconclusive or not available Risk factors Age(R 55 y) Alcohol or Pregnancy (cholestatic)

Chronological criteria (according to type of liver damage: hepatocellular or cholestatic/mixed) Time to onset of the event Suggestive Compatible

Axis

CIOMS/RUCAM

C1 to 0 C1 to 0 C1 to 0

C3 C2 C1 K2 0

C1

C2 C1

Score

Other causes Complete exclusion Partial exclusion Possible alternative causes Probable alternative causes

C3 0 K1 K3

0

C3

C1 to 0 C1 to 0

Previous reaction to the same or related drugs

C1 to 0

C1 to K1

C2 to K1

C2 to K1

Reaction dosed-related

Toxic drug levels in biological fluids

Positive re-challenge with placebo

Presence of alternative causes

Positive de-challenge

C1 to 0

Improvement of ADR after drug withdrawal or antagonist

C3 0 K3

C2 to K1

Onset of the ADR after drug

C3 C1

Score C1 to 0

From drug intake until onset event Four days to eight weeks Less than four days or more than eight weeks From drug withdrawal until onset event Zero to seven days 8 to 15 days More than 15 days Normalization of laboratory values Less than six months (cholestatic) or Two months (hepatocellular) More than six or two months

Axis Previous reports on ADR

Score

Naranjo

Chronological criteria

Axis

Maria & Victorino/CDS

TABLE 5 Comparison of the Scores for Individual Axes of the CIOMS, Maria & Victorino and Naranjo Diagnostic Scales

334 Lucena et al.

C3 C1 K2 0

C2 C1 0

C2 C1 0 K2 K3

0 0 K1 K2 K3 0

C3 C2 C1 0

C3 0

C2 0 K3

Extrahepatic manifestations Four or more Two or three 1 None Re-challenge Positive Negative or not performed

Known reaction Yes No (drugs marketed less than five years) No (drugs marketed more than five years)

ADR confirmed by any objective evidence

C1 to 0

With CIOMS, score %0, diagnosis of excluded; 1 to 2 unlikely; 3 to 5, possible; 6 to 8 probable; O8 highly probable or definite. With Maria & Vitorino clinical scale, scoreO17, definite; 14 to 17, probable; 10 to 13, possible; 6 to 9, unlikely; !6, drug hepatotoxicity excluded. With Naranjo adverse drug reaction probability scale, scoreR9 diagnosis of definite; 5 to 8, likely; 1 to 4 possible; %0, doubtful. Abbreviations: CIOMS, Council for International Organizations of Medical Sciences; RUCAM, Roussel-Uclaf Causality Assessment Method; CDS, clinical diagnostic scale; SPC, summary product characteristic; ADR, adverse drug reaction.

Concomitant drug(s) Absence Time to onset incompatible Compatible but unknown reaction Compatible and known reaction Role proved in this case None or information not available Exclusion non-drug-related causes Rule out all causes Six causes excluded Four or five causes excluded Less than four causes excluded Probable hepatic etiology Previous information on hepatotoxicity Reaction known Reaction published but unlabelled Reaction labelled in the SPC Re-challenge Positive Compatible Negative Not available or not interpretable

Causality Assessment

335

336

Lucena et al.

However, both methods lack specificity for hepatotoxicity. Therefore, an etiologic– diagnostic assessment that could weight relevant criteria according to their particular chronological and clinical characteristics would be more appropriate than a routine general system and would also allow increasing the number of well-documented cases. The first method of causality assessment for drug-induced liver injury was the decision tree developed by Stricker (27) in 1992. This method consists of an analytical scheme based on three points: specificity of the “clinico-pathological pattern” (defined as the combination of clinical, histological, biochemical, immunological, toxicological, experimental, and other variables) and its course, the temporal relationship between these findings, and the exclusion of other possible causes for the observed patterns (the less specific pattern is, the more important this last factor becomes). The model assesses the degree of certainty on a scale of several levels. Unfortunately, Stricker’s decision tree is a complex and perhaps overly subjective method for use in routine clinical practice. In addition, the scheme is based on current knowledge and would need regular updating. In 1992, under the auspices of the Council for International Organizations of Medical Sciences (CIOMS), a working group developed and implemented a standardized method for drug causality assessment termed CIOMS or Roussel-Uclaf Causality Assessment Method (RUCAM in reference to the organizers of the event) scale (60,61). This method provides a standardized scoring system in which the limits and contents of most criteria were decided by consensus among experts on the basis of organ-oriented characteristics. The parameters are (Table 5): time to onset and course or time to 50% improvement in laboratory values (strictly defined, improving reproducibility), risk factors (age of the patientR55 years, alcohol consumption, pregnancy), concomitant drugs, search for nondrug causes, previous information on hepatotoxicity of the drug, and response to re-administration. The time to onset and course are evaluated separately for hepatocellular versus cholestatic/mixed reactions, because the latter can occur after a longer post-cessation interval and resolve much more slowly. The CIOMS/RUCAM scale provides a scoring system for six axes in the decision strategy. The categories of suspicion are definite or highly probable (scoreO8), probable (score 6–8), possible (score 3–5), unlikely (score 1–2), and excluded (score!/Z0). This method was originally validated using cases of drug-induced liver disease with known positive re-challenge (the major type of evidence recognized as demonstrating the role of a drug), and the system performed well when these cases were assessed based on data prior to re-challenge or when concomitant drugs were included. Among the advantages of this system that could be pinpointed are the minimization of questions requiring subjectivity. A definite diagnosis of drug-induced hepatotoxicity is still possible in patients without re-challenge data and it can easily identify the first case for a new drug or for a previously unreported liver injury of an older drug. The major drawback of this scale is certainly its complexity. It requires training in its performance until a user becomes familiar with the format; the scale may seem cumbersome, and reading across needs to be taken into account if you do not want questions to be misunderstood and careless errors to be made. More recently, Maria and Victorino (M&V) from Portugal, developed a simplified scoring system in an attempt to overcome the above-mentioned problems, called the clinical diagnostic scale (CDS) (62) (also termed the M&V scale), which uses several features of the CIOMS/ RUCAM scale but omits and adds others (Table 5). Five components were selected for inclusion in the scale: temporal relationship between drug intake and the onset of clinical symptoms, exclusion of alternative causes, and presence of extra-hepatic manifestations (rash, fever, arthralgia, eosinophiliaO6%, and cytopenia), intentional or accidental re-exposure to the drug, and previous report in the literature. The sum of the points for each parameter can vary from K6 to 20. Correspondence with the five classic degrees of probability of adverse drug reactions is established on the basis of the score as follows: definite (scoreO17), probable (score 14–17), possible (score 10–13), unlikely (score 6–9), and excluded (score!6). The M&V clinical scale was validated using real and fictitious cases of immunoallergic drug-induced liver injury (high percentages of positive immunological tests in cases classified as definite or probable), and was compared with the external standard or experts’ classification. The authors have noted some limitations of the scale: The instrument performs poorly in atypical cases of drugs with unusually long-latency periods or chronic evolution after withdrawal. Specifying more clearly

Causality Assessment

337

the clinical conditions to be excluded and including detailed criteria for exclusion could improve the exclusion of alternative causes of liver injury. The main advantage of the M&V scale is its ease of application in clinical practice. Comparison of Assessment Methods in Hepatotoxicity The merits of these two organ-specific scales and their degree of concordance were compared in a population of 215 patients reported to a registry of hepatotoxicity. Causality in this population was verified by three experts as drug-induced (185) or as nondrug caused (30 cases) (63). Complete agreement between the M&V and the CIOMS scales was obtained in only 42 cases (18%). Discrepancies in the assessment of causality occurred in 186 ratings (out of 228 ratings generated because of the use of multiple drugs); in each of these cases, the CIOMS scale ascribed a higher level of certainty than the M&V scale. The M&V system classified only about one-third of the cases as probable or definite, and tended to underestimate the probability of causality. The lowest agreements were observed for patients with cholestatic injury but of 32 classified as unlikely hepatocellular group by M&V 23 were probably by CIOMS (Fig. 2). In cases whose outcome was death, no agreement was found. In sum, the performance of the M&V scale was poor in reactions with long-latency periods (more than 15 days; i.e., amoxicillin/clavulanic acid), evolution to chronicity after withdrawal (cholestatic pattern), or death. The consistency of assessment was lowered because the two methods assigned different weights to relevant assessment criteria and so the reasons for disagreement could be clearly identified. Thus, a delay of more than 15 days from drug withdrawal until onset of event was handled by questioning the causal link between the drug and the event by subtracting three points from the score in the M&V scale. Furthermore, time from drug withdrawal until normalization of laboratory values longer than six months (in cholestatic or mixed type of injury) or two months (in hepatocellular damage) precluded a definite or probable diagnosis. In addition, unknown reactions to drugs marketed for more than five years preclude a diagnosis of certainty. On the contrary, the best correlation between the two scales was found for drug-induced liver injury with a suggested immunoallergic mechanism (Fig. 3). The reason for this is that the M&V scale includes questions that apply only to cases with extra-hepatic features. It appears, therefore, that the CIOMS instrument shows better agreement with “common sense” clinical judgment and remains closer to the scenario that may be found in real life. Aside from its clinical validity, the usefulness of the CIOMS scale is that it provides a framework which emphasizes topics that need to be addressed in cases of suspected hepatic adverse reaction in order to improve the consistency of judgments (63). The CDS (M&V) was further evaluated in the causality assessment of 135 hepatotoxic adverse drug reaction reports (64). First, the CIOMS criteria were used to classify reactions as “drug related,” “drug unrelated,” and “indeterminate.” Reports classified as drug related (49 reactions) scored higher on the clinical scale, with a median score of 12 (range 8–15). Of those, no reactions were classified as “definite,” 20 were classified as probable and 23 as possible. It is important to note that six patients were classified as unlikely. The authors concluded that the clinical scale correlated well with the consensus classification of suspected hepatic drug reactions, and suggested that a cut-off scoreO9 could identify all drug-related liver injuries unless alternative diagnoses are strongly suspected. Therefore, they suggested that a cut-off scoreO9 be used in clinical decision making. This seems to be a risky conclusion since the cutoff value proposed falls into the category of possible in the CDS. It is worth noting that possible is a fairly low category adjacent to unlikely which makes this cut-off score quite unreliable for decision making. Besides, the authors did not assess or compare the merits of the two systems in any detail (32). Six patients whose hepatotoxicity was considered drug related on the basis of the consensus classification (four of these patients having a positive re-challenge) scored!10 (unlikely). Two patients had flucloxacillin-induced cholestasis which first appearedO15 days after drug withdrawal, and in two other patients, the reactions were fatal, complicating accurate assignment of the cause. Two other patients with a long-latency period scored only one point each for the onset of the reactions. These cases confirm the limitations of the clinical

338

Lucena et al.

Hepatocellular damage (Absolute agreement: 9.2%; discrepancies for 3 categories of rank: 2.2%) 40

35 28

30

23

26

CIOMS 20 12 10

6

3

0 0 1

0

0

0 Unlikely

Possible

Probable

Definite

Possible

6

0

0

0

Probable

23

28

12

0

Definite

3

26

35

1

M&V

(A)

10

Cholestatic damage (Absolute agreement: 6%; discrepancies for 3 categories of rank: 9.1%) 10

8

7 6

6 CIOMS

4

4

2

2 0

(B)

1

1 0

1

1

0

0

Excluded

Unlikely

Possible

Probable

Possible

0

1

1

0

Probable

1

4

7

1

Definite

0

2

6

10

M&V

FIGURE 2 Categories of suspicion obtained with the M&V scale and the corresponding figures with the CIOMS scale according to type of liver damage, (A) hepatocellular or (B) cholestatic, when applied to 215 cases of DILI. Source: From Ref. 63.

scale as pointed out by its own authors and seem to be coincidental with those reached in the study by Lucena et al. (63). Indeed, these two studies could be evaluated in terms of sensitivity and specificity and predictive values, indexes currently used to evaluate the reliability of diagnostic tests. To compare the two diagnostic tests, a CIOMS score of at least 6 (“drugrelated”) was chosen as the critical value to determine that the hepatotoxicity was more likely caused by the suspected drug. A M&V/CDS score of at least 14 was chosen as the index score for this diagnostic aid. A 2!2 table based on these criteria shows that the specificity of the M&V/CDS scale was 100% in both studies, and the sensitivity was 37% in the study by Lucena et al. (63), while the corresponding figure was 41% in the study by Aithal et al. (64). This data further

339

Causality Assessment

Inmunoallergic Idiosyncrasy (Absolute agreement: 12%; discrepancies for 3 categories of rank: 2%) 20

18

15 10 CIOMS

11

10 5 5

3 1

0

1

Unlikely

0

0

0 Possible

Probable

0

Definite

Possible

1

0

0

0

Probable

3

10

5

0

Definite

1

11

18

1

(A)

1

M&V

Metabolic Idiosyncrasy (Absolute agreement: 9.1%; discrepancies for 3 categories of rank: 3.5 %) 33

35 30

27 24

25 CIOMS

20 15

12

10 5 0

(B)

33

7 4 1

0

Excluded

1

0 Unlikely

0

Possible

Probable

Possible

0

7

1

0

Probable

1

24

33

12

Definite

0

4

27

33

M&V

FIGURE 3 Categories of suspicion obtained with the M&V scale and the corresponding figures with the CIOMS scale according to the presumed mechanism of lesion, (A) metabolic idiosyncrasy or (B) immunoallergic idiosyncrasy, when applied to 215 cases of DILI. Source: From Ref. 63.

340

Lucena et al.

support the conclusions already reached. M&V/CDS scale is able to recognize those cases of liver damage not related to drugs, but its capacity to ascertain a diagnosis of drug-related hepatotoxicity with confidence is very poor. Indeed, this instrument tends to underestimate the probability of causality. We have also pointed out the lack of agreement found between the authors and referees when applying the Naranjo Adverse Drug Reactions Probability Scale (NADRPS) (59) to cases of hepatotoxicity prior to their publication in a peer reviewed journal (65). We should keep in mind that the policy of some journals is the requirement of use of this scale prior to publication of a potential adverse drug reaction. Furthermore, the assessment of these cases by the CIOMS scale yielded score values closer to those of the authors’ clinical judgment. We could argue that the disagreements found between raters and scales were related to the use of judgment in the evaluation of alternative causes, as a standardized search is not provided. Furthermore, as this scale is not specific for hepatotoxicity, strict chronological criteria upon type of liver injury are not defined and associated risk factors are not considered. Question number 10 in the Naranjo scale, which refers to the confirmation of an adverse reaction by objective evidence is left at best to individual interpretation of the relevance of the findings of a liver biopsy with regard to hepatotoxicity. Considering that only suspected hepatic injuries with an idiosyncratic mechanism were studied, some of the criteria (axis 6, 7, 8, and 9 in Table 5) used in the NADRPS are not assessable in the evaluation of an adverse hepatic reaction since they are formulated to evaluate predictable adverse drug reactions, related to the pharmacological actions of the drugs. Recently, the performance and consistency of this scale to ascertain causality in drug-induced hepatotoxicity in comparison with the CIOMS scale has been evaluated. The results obtained confirm that the Naranjo scale lacks validity and reproducibility in the attribution of causality in hepatotoxicity, along with a low discriminative power and a limited capability to distinguish between adjacent categories of probability (66). Taken together, these comparative studies among the different clinical scales clearly show that the CIOMS scale, although far from being a perfect instrument, provides a uniform basis for developing a more precise diagnostic approach in drug-induced hepatotoxicity. Certainly, raters would benefit from training with the CIOMS scale to improve the quality of the data collected so as to provide the relevant information for interpretation of signal detection in the pharmacovigilance setting that may lead to the adoption of regulatory measures (37) or even the publication of information in a medical journal. Indeed, medical journals should require the application of the scale as a quality control previous to the acceptance of a case of hepatotoxicity for evaluation. However, rules for assigning causality in drug-induced liver injury do not substitute for clinical judgment. For instance, when more than one drug could be the culprit, a “blinded” application of the scale can lead to a somewhat misleading causality assessment if only chronological criteria are taken into account (24). To avoid this, attention should be paid to major drug-metabolism mechanisms and the potential for pharmacokinetic drug interactions. PROSPECTS FOR THE FUTURE We realize that the time has come to consider some refinements in order to make the scale more realistic, accommodating relevant data within the scoring system and deleting low-impact items. This task can now be carried out using large databases of “bona fide” cases of hepatotoxicity. At present, for example, maintaining alcohol as a general risk factor for hepatotoxicity (taking into account that in fact it has only demonstrated its role for very few drugs) while obesity, a prevalent condition that is associated with an increased expression of CYP2E1 (1) is disregarded, may be a gross oversimplification. At the same time, the cut-off age point does not consider pediatric age as a risk factor for toxicity with some drugs. Other known risk factors to individual drugs when present in the appropriate setting could be incorporated (e.g., HIV infection in sulfonamide users, coinfection with hepatitis B/hepatitis C virus and antiretroviral drugs, female sex for diclofenac). Among the aspects that deserve commentary is the actual gap between usual clinical data that merit attention by physicians and that are not assessed in CIOMS scoring. In our opinion, the major item of controversy is the relevance that clinicians give (and the scale does not) to

Causality Assessment

341

a biochemical pattern of injury that fits with previous knowledge about the hepatotoxic profile of the candidate drug. For instance, physicians would tend to doubt the responsibility of amoxicillin–clavulanate if the damage is of the hepatocellular type whereas CIOMS scale does not take into consideration this apparent discrepancy. Besides, clinicians attribute a crucial value to the presence of hypersensitivity features to make a definite diagnosis of DILI whereas these manifestations are not weighted at all in the CIOMS scale (67). On the other hand, the way in which hepatotoxicity is defined is fundamental to the incidence of hepatotoxicity in each study. Thus, in the original criteria (60) when the reaction occurred more than 15 days after stopping the treatment in cases of hepatocellular injury, and more than 30 days in cases of cholestatic or mixed type, cases were assessed as “unrelated.” This aspect should be modified to take into account the known data of delayed appearance of hepatotoxicity for certain drugs (e.g., amoxicillin–clavulanate, minocycline). Otherwise diagnosis may be hindered as actually was the case for amoxicillin–clavulanate (20). Beyond these important questions, there is the need to validate a new instrument, an abridged scale that may give a fair approximation to the truth—the likelihood that a given case of hepatitis is due to a drug—at the very beginning of the process of evaluation of the patient when clinical decisions have to be taken. The diagnosis achieved should be confidently made on admission and maintained as further information is collected. Hence, this certainly would be the goal of a clinical assessment method in hepatotoxicity as a clinical tool. In summary, after optimization and validation of the critical items a new consensus should be encouraged to accommodate all data relevant to DILI within the scoring system so that an unclear diagnosis can be upgraded whereas a secure diagnosis can remain unchallenged. This would presumably aid in its definitive integration into the set of diagnostic tools for drug-induced hepatotoxicity. In fact, one of the challenges faced by post-marketing surveillance and risk assessment systems is to translate scientific accuracy into routine clinical practice. REFERENCES 1. Bissel DM, Gores GJ, Laskin DL, et al. Drug-induced liver injury: mechanisms and test systems. Hepatology 2001; 33(4):1009–13. 2. Nirenberg D. “Did this drug cause my patient’s hepatitis?” and related questions Ann Intern Med 2002; 136(6):480–3. 3. Meyboom RHB, Hekster YA, Egberts ACG, et al. Causal or casual? The role of causality assessment in Pharmacovigilance Drug Saf 1997; 17(6):374–89. 4. Kaplowitz N. Drug-induced liver disorders: introduction and overview. In: Kaplowitz N, Deleve LD, eds. Drug-Induced Liver Disease. New York: Marcel Dekker, Inc., 2003:1–14. 5. Maria VA, Victorino RM. Diagnostic value of specific T-cell reactivity to drugs in 95 cases of drug induced liver injury. Gut 1997; 41(4):534–40. 6. Maria VA, Victorino RM. Immunological investigation in hepatic drug reactions. Clin Exp Allergy 1998; 28(Suppl. 4):71–7. 7. Zimmerman HJ. Hepatotoxicity. The Adverse Effects of Drugs and Other Chemicals on the Liver. 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins, 1999. 8. Berg PA, Becker EW. The lymphocyte transformation test: a debated method for the evaluation of drug allergic hepatic injury. J Hepatol 1995; 22(1):115–8. 9. Kaplowitz N. Idiosyncratic drug hepatotoxicity. Nat Rev Drug Discov 2005; 4(6):489–99. 10. Andrade R, Lucena MI, Fernandez MC, et al. Drug-induced liver injury: an analysis of 461 incidences submitted to the Spanish registry over a 10-year period. Gastroenterology 2005; 129(2):512–21. 11. Lewis JH, Ranard RC, Caruso A, et al. Amiodarone hepatotoxicity. Prevalence and clinicopathologic correlation among 104 patients. Hepatology 1989; 9(5):679–85. 12. Oien KA, Moffat D, Curry GW, et al. Cirrhosis with steatohepatitis after adjuvant tamoxifen. Lancet 1999; 353(9146):36–7. 13. Lewis JH, Schiff ER. Methotrexate-induced chronic liver injury: guidelines for detection and prevention. Am J Gastroenterol 1988; 88(12):1337–45. 14. Sharp JR, Ishak KG, Zimmerman HJ. Chronic active hepatitis and severe hepatic necrosis associated with nitrofurantoin. Ann Intern Med 1980; 92(1):14–19. 15. Sturkenboom M, Meier CR, Jick H, et al. Minocycline and lupus-like syndrome in acne patients. Arch Intern Med 1999; 159(5):493–7. 16. Andrade RJ, Lucena MI, Alcantara R, et al. Bentazepam-associated chronic liver disease. Lancet 1994; 343(8901):860.

342

Lucena et al.

17. Andrade RJ, Lucena MI, Aguilar J, et al. Chronic liver injury related with the use of bentazepam: an unusual instance of benzodiazepine hepatotoxicity. Dig Dis Sci 2000; 45(7):1400–4. 18. Ishak KG, Zimmerman HJ. Hepatotoxic effects of anabolic androgenic steroids. Semin Liver Dis 1989; 7(3):230–9. 19. Ishak KG, Zimmerman HJ. Morphologic spectrum of drug-induced hepatic disease. Gastroenterol Clin North Am 1995; 24(3):759–86. 20. Andrade RJ, Lucena MI, Ferna´ndez MC, et al. Hepatotoxicity in patients with cirrhosis, an often unrecognized problem. Lessons from a fatal case related with amoxicillin/clavulanic acid. Dig Dis Sci 2001; 46(7):1416–9. 21. Pe´rez-Moreno JM, Saldana FJ, Puertas M, et al. Cholestatic hepatitis caused by midecamycin. Gastroenterol Hepatol 1996; 19(9):459–61. 22. Lucena MI, Andrade RJ, Rodrigo L, et al. Trovafloxacin-induced acute hepatitis. Clin Infect Dis 2000; 30(2):400–1. 23. Kaplowitz N. Drug-induced liver injury. Clin Infect Dis 2004; 38(Suppl. 2):S44–8. 24. Lucena MI, Andrade RJ, Vicioso L, et al. Prolongued cholestasis after raloxifene and fenofibrate interaction. World J Gastroenterol 2006; 12(32):5244–6. 25. Biour M, Poupon R, Grange´ J-D, et al. He´patotoxicite´ des me´dicaments. 13e mise a´ jour du fichier bibliographique des atteintes he´patiques et des me´dicaments reponsables. Gastroenterol Clin Biol 2000; 24(11):1052–91. 26. Farrell GC. Drug Induced Liver Disease. London: Churchill-Livingstone, 1994. 27. Stricker BH. Drug-Induced Hepatic Injury. 2nd ed. Amsterdam: Elsevier, 1992. 28. Kaplowitz N. Drug-induced liver disorders: implications for drug development and regulation. Drug Saf 2001; 24(7):483–90. 29. Shah RR. Drug-induced hepatotoxicity: pharmacokinetic perspectives and strategies for risk reduction. Adverse Drug React Toxicol Rev 1999; 18(4):181–233. 30. Watkins P, Withcomb R. Hepatic dysfunction associated with troglitazone. N Engl J Med 1998; 338(13):916–7. 31. Andrade RJ, Camargo R, Lucena MI, et al. Causality assessment in drug-induced hepatotoxicity. Expert Opin Drug Saf 2004; 3(4):329–3444. 32. Lee WM. Assessing causality in drug-induced liver injury. J Hepatol 2000; 33(6):1003–5. 33. Relvas S, Carreira F, Castro B. Liver involvement in secondary syphilis. Am J Gastroenterol 1992; 87(10):1528. 34. Carrillo-Jime´nez R, Nurnberger M. Celecoxib-induced acute pancreatitis and hepatitis. Arch Intern Med 2000; 160(4):553–4. 35. Schattner A, Sokolovskaya N, Cohen J. Fatal hepatitis and renal failure during treatment with nimesulide. J Intern Med 2000; 247(1):153–5. 36. Andrade RJ, Lucena MI, Kaplowitz N, et al. Outcome of acute idiosyncratic drug-induced liver injury: Long-term follow-up in a hepatotoxicity registry. Hepatology 2006; 44(6):1581–8. 37. Andrade RJ, Lucena MI, Martı´n-Vivaldi R, et al. Acute liver injury associated with the use of ebrotidine, a new H2- receptor antagonist. J Hepatol 1999; 31(4):641–6. 38. Be´nichou C. Report of an International Consensus Meeting. Criteria of drug-induced liver disorders. J Hepatol 1990; 11(2):272–6. 39. Andrade RJ, Lucena MI, Fernandez C, et al. Cholestatic hepatitis related to use of irbesartan. A case report and a literature review of angiotensin II antagonists associated hepatotoxicity. Eur J Gastroenterol Hepatol 2002; 14(8):887–90. 40. Andrade RJ, Lucena MI, Fernandez MC, Gonzalez M. Fatal hepatitis associated with nimesulide. J Hepatol 2000; 32(1):174. 41. Lucena MI, Andrade RJ, Gomez-outes A, et al. Acute liver failure after treatment with nefazodone. Dig Dis Sci 1999; 44(12):2577–9. 42. Goodman ZD. Drug hepatotoxicity. Clin Liver Dis 2002; 6(2):381–97. 43. Bianchi L. Liver biopsy in elevated liver function tests? An old question revisited. J Hepatol 2001; 35(2):290–4. 44. Larrey D. Drug-induced liver diseases. J Hepatol 2000; 32(Suppl. 1):77–88. 45. Degott C, Feldmann G, Larrey D, et al. Drug-induced prolonged cholestasis in adults: a histological semiquantitative study demonstrating progressive ductopenia. Hepatology 1992; 15(2):244–51. 46. Aithal PG, Day CP. The natural history of histologically proved drug induced liver disease. Gut 1999; 44(5):731–5. 47. Lee WM. Drug-induced hepatotoxicity. N Engl J Med 2003; 349(5):474–85. 48. Lucena MI, Andrade RJ, Ferna´ndez MC, et al. Determinants of the clinical expression of amoxicillinclavulanate hepatotoxicity: a prospective series from Spain. Hepatology 2006; 44(4):850–6. 49. Van Steenbergen W, Peeters P, De Bondt J, et al. Nimesulide-induced acute hepatitis: evidence from six cases. J Hepatol 1998; 29(1):135–41. 50. Herrine SK, Keeffe EB. Severe hepatotoxicity associated with troglitazone. Ann Intern Med 1999; 130(2):163–4.

Causality Assessment

343

51. Bonkovsky HL, Azar R, Bird S, et al. Severe cholestatic hepatitis caused by thiazolidinediones: risks associated with substituting rosiglitazone for troglitazone. Dig Dis Sci 2002; 47(7):1632–7. 52. Andrade RJ, Lucena MI. Drug-induced hepatotoxicity. N Engl J Med 2003; 349(20):1974–6. 53. Aithal GP, Rawlins MD, Day CP. Accuracy of hepatic adverse drug reaction reporting in one English health region. BMJ 1999; 319(7224):1541. 54. Kaplowitz N. Causality assessment versus guilt-by-association in drug hepatotoxicity. Hepatology 2001; 33(1):308–10. 55. Hutchinson TA, Lane DA. Assessing methods for causality assessment of suspected adverse drug reactions. J Clin Epidemiol 1989; 42(1):5–16. 56. Bate A, Lindquist M, Edwards IR, et al. A Bayesian neural network method for adverse drug reaction signal generation. Eur J Clin Pharmacol 1998; 54(4):315–21. 57. Pere JC, Begaud B, Haramburu F, et al. Computarized comparison of six adverse drug reaction assessment procedures. Clin Pharmacol Ther 1986; 40(4):451–61. 58. Karch FE, Lasagna L. Toward the operational identification of adverse drug reaction. Clin Pharmacol Ther 1977; 21(3):247–54. 59. Naranjo CA, Busto U, Sellers EM, et al. A method for estimating the probability of adverse drug reactions. Clin Pharmacol Ther 1981; 30(2):239–45. 60. Danan G, Be´nichou C. Causality assessment of adverse reactions to drugs I. A novel method based on the conclusions of international consensus meetings: application to drug-induced liver injuries. J Clin Epidemiol 1993; 46(1):1323–30. 61. Benichou C, Danan G, Flahault A. Causality assessment of adverse reactions to drugs. II. An original model for validation of drug causality assessment methods: case reports with positive rechallenge. J Clin Epidemiol 1993; 46(11):1331–6. 62. Maria V, Victorino R. Development and validation of a clinical scale for the diagnosis of drug-induced hepatitis. Hepatology 1997; 26(3):664–9. 63. Lucena MI, Camargo R, Andrade RJ, et al. Comparison of two clinical scales for causality assessment in hepatotoxicity. Hepatology 2001; 33(1):123–30. 64. Aithal G, Rawlins M, Day C. Clinical diagnostic scale: a useful tool in the evaluation of suspected hepatotoxic adverse drug reactions. J Hepatol 2000; 33(6):949–52. 65. Garcı´a-Corte´s M, Lucena MI, Andrade RJ, et al. Is the Naranjo probability scale accurate enough to ascertain causality in drug-induced hepatotoxicity? Ann Pharmacother 2004; 38(9):1540–1. 66. Garcı´a-Cortes M, Lucena MI, Pachkoria K, et al. Evaluation of the Naranjo adverse drug reactions probability scale in the causality assessment of drug-induced liver injury. Am J Gastroenterol 2007 (submitted). 67. Takamori Y, Takikawa H, Kumagi T, et al. Assessment of the diagnostic scale for drug-induced liver injury by the international consensus meeting and a proposal of its modifications. Hepatology 2003; 38(4, Suppl. 1):703A.

17

Management of the Patient with Drug-Induced Liver Disease Thomas D. Boyer

Liver Research Institute, University of Arizona, Tucson, Arizona, U.S.A.

INTRODUCTION Drug-induced liver disease is a common clinical problem. Although adverse reactions appear rare, occurring in 1:10,000 to 1:100,000 individuals exposed to a given drug (1), incidence figures in a study from France for all types of hepatic drug reactions were 13.9G2.4 per 100,000, with 12% being hospitalized and 6% dying from the adverse reaction. The author believes that this is an underestimate of the true incidence of drug-induced liver disease (2). Similarly, in a report from the United States, the medical records of an HMO were examined for elevated liver tests. They found an incidence of 40.6 cases per 100,000 for drug-associated liver test abnormalities when compared with an incidence of 15.4 per 100,000 for alcohol as a cause of the elevated liver tests (3). Adverse hepatic drug reactions are common in the outpatient setting as well. In one report, 84% of adverse hepatic drug reactions occurred in outpatients (2). Identification of hepatotoxic drug reactions in the outpatient setting is difficult as multiple healthcare providers may prescribe medications for a given patient and if the reaction is delayed, as with clavulanic acid–containing drugs, cause and effect may be difficult to establish. In addition, the spontaneous reporting of adverse drug reactions is more likely to occur with hospital-based physicians than with nonhospital-based physicians, leading to an underestimate of the frequency of drug-induced hepatotoxicity in outpatients (4). Thus, the clinician is likely to see a number of patients with drug-induced liver injury during the course of their practice and the management of these patients will be a common issue. This chapter will focus on how to manage the patient with drug-induced liver injury. A brief discussion of how the diagnosis is established is followed by a discussion of whether or not discontinuation of the offending medication is required and whether a rechallenge or use of a medication of the same class is safe. Treatment of complications of the adverse reactions will also be discussed and, lastly, the criteria for referral to a liver transplant center will be reviewed. Although the author will attempt to use data from studies to support most of the recommendations provided in this chapter, on occasion the recommendations will be based on the author’s experience in managing these types of patients.

DIAGNOSIS OF DRUG-INDUCED LIVER INJURY AND MONITORING OF LIVER TESTS The diagnosis of an acetaminophen overdose is not difficult as a blood test is available and usually the time of exposure is known. Idiosyncratic drug reactions are much more difficult to diagnose as the patients are frequently taking several medications and there is no specific test available to establish the diagnosis. Given these difficulties, criteria for the diagnosis of druginduced liver disease have been established. Table 1 shows one such scoring system. Using this scoring system, it is possible to come up with an estimate of whether or not the suspected drug actually caused liver injury: %0, relationship with the drug excluded; 1 to 2, unlikely; 3 to 5, possible; 6 to 8, probable; O8, highly probable. The system of Maria and Victorino (5) uses slightly different criteria and with their system the following scores are used: O17, definite; 14 to 17, probable; 10 to 13, possible; 6 to 9, unlikely; !6, drug hepatotoxicity excluded. Short of using such a scoring system, the development of liver test abnormalities within 12 weeks

346

Boyer

TABLE 1 Criteria for Adverse Hepatic Drug Reactions Criteria

Range of scores

Time to onset of reaction Course of reaction Risk factors for reaction Concomitant drugs Nondrug causes Known potential hepatotoxin Rechallenge Toxic plasma concentration

0 to C3 0 to C3 0 to C1 0 to K3 K3 to C2 0 to C2 K2 to C3 K3 to C3

Source: Adapted from Ref. 1.

of starting a new drug, improvement in liver tests with drug withdrawal, and a recurrence of liver injury upon reexposure to the drug are the most convincing evidence that the liver injury is related to the new medication. If several medications have been started recently, then withdrawal of all medications that are not essential to the health of the patient is frequently the only course to follow. A common clinical issue is the development of mild increases in liver tests in a patient taking a drug known to cause hepatotoxicity. The risk of a serious adverse reaction with these types of drugs may be great or small. Perhaps, the best known drug where this problem is seen is with the use of isoniazid to prevent the development of active pulmonary tuberculosis. Elevated levels of aspartate and alanine aminotransferases (AST and ALT) are seen in 10% to 20% of patients taking isoniazid and yet only w1% developed severe hepatitis. Should isoniazid be withdrawn when elevated liver tests are noted? Probably not, as the development of elevated liver tests is not predictive of a severe reaction, but careful monitoring of liver tests is required and if a steady increase is noted, the drug should be stopped. Waiting until the patient is jaundiced is not appropriate as the development of jaundice is associated with a significant risk of acute hepatic failure. The risk of isoniazid hepatitis can be reduced by not using the drug in patients (older individuals, African-Americans) known to be at increased risk for an adverse reaction or using alternative therapies, such as two months of rifampicin– pyrazinamide therapy. However, this combination also requires monitoring for hepatic injury (6,7). Perhaps, the best approach when using isoniazid is to monitor only those at increased risk for an adverse reaction (7). The other common situation is when a patient is one taking a drug, like a statin, that is known to cause elevated liver tests but rarely causes a severe reaction. The incidence of increased liver tests in patients using a statin is 2% to 5% and probably dose related and yet with lovastatin, the rate of severe hepatitis is 1/1.14 million (8). In addition, even the presence of liver disease does not increase the risk of an adverse reaction with statins (9). Should the statin be withdrawn when elevated liver tests are observed? Based on current data probably not, as the benefit of the statin may outweigh the small risk of a serious drug reaction (8).

CLASS SENSITIZATION AND RECHALLENGE One of the criteria listed for the diagnosis of drug-induced liver disease is the recurrence of the disease upon reexposure of the patient to the drug. This rechallenge can be elective (rare) or involuntary. Elective rechallenge is performed either when the cause of the adverse reaction is not clear or, more commonly, when the drug is essential to the well-being of the patient. Diagnostic rechallenge is rarely performed because of the risk to the patient, especially if the original reaction was hepatocellular in nature (10). Continued use of a drug that is thought to be causing significant liver damage is also uncommon but can occur with drugs, such as amiodarone, which may be the only agent available to prevent a lethal event, such as ventricular tachycardia (11). In this latter situation, the risk–benefit to the patient of a given drug must be decided upon by the treating physician.

Management of the Patient

347

The far more common event is the inadvertent reexposure of the patient to a drug that has caused previously undiagnosed hepatotoxicity or the prescribing physicians were unaware of the previous adverse event. One example of the latter is when the patient has received a combination drug, in which one of the drugs but not the other is the hepatotoxin. An example of this is clavulanic acid, which is in both Augmentinw and Timentinw. The author saw a patient who developed jaundice while taking Augmentin and the patient was told he was allergic to amoxicillin. Several months later, the patient received Timentin and again became jaundiced, as it was the clavulanic acid in both of these drugs and not the antibiotic that was responsible for the drug hepatotoxicity (12). Perhaps, the best known example of inadvertent reexposure to a drug leading to hepatotoxicity is halothane. More than 80% of cases of halothane hepatitis are described in patients who had more than one exposure to the anesthetic. In addition, the time from exposure to onset of jaundice is shorter with repeated exposures (13). Fever and elevated liver tests without jaundice may be seen following the first exposure to halothane and upon repeated exposure, the patient may become jaundiced and develop acute hepatic failure. The repeated exposure may also occur in the absence of the use of halothane due to the contamination by halothane of the equipment used to administer the anesthetic from a preceding case (14). Once a patient has manifested drug-induced liver injury to a specific drug, care must be exercised in preventing reexposure to that specific drug, but must the patient also not be exposed to other agents of the same class? The generally held belief is that most cases of hepatotoxicity are not due to the original drug but rather to a breakdown product of the drug. Liver injury from acetaminophen is due to the formation of a reactive product during metabolism of the drug (15). Acetaminophen is a predictable hepatotoxin because all individuals if exposed to enough of the drug will develop liver injury; however, there is individual susceptibility (14). Similarly, the development of liver injury from an unpredictable hepatotoxin such as isoniazid, halothane, and troglitazone is due to the formation of a breakdown product of the drug that causes the liver injury most likely via an immune mechanism or by causing intracellular stress (16,17). The formation of this toxic product and the immune reaction to it is individual specific, although family members may also be susceptible presumably due to the presence of the same type of drug-metabolizing enzymes and immune response (18). Thus, in contrast to most drug allergies, unpredictable hepatic drug reactions due to one agent may not mean that all drugs in the same class must be avoided. Halothane is converted to a reactive metabolite via cytochrome P450 2E1. The reactive product then binds covalently to proteins forming a hapten, leading to an immune response and liver injury (13). If another halogenated anesthetic is to cross-react with halothane, then the other agent would have to form a covalent product with intracellular proteins that had epitopes similar to the haptens formed from the detoxification of halothane for this cross-reactivity to occur. Experimental studies have suggested that antibodies formed against halothane adducts may cross-react with adducts formed from the metabolism of enflurane (19,20). In support of these experimental observations are reports of cross-reactivity between halothane and enflurane (21,22). Although drugs such as enflurane are less likely to cause toxic hepatitis when compared with halothane because of their less extensive metabolism, it would appear wise to use a nonhalogenated anesthetic agent in a patient with a previous history of a halogenated anesthetic-induced liver injury. Treatment of epilepsy also requires the use of drugs known to cause liver injury. The most commonly used agent is diphenylhydantoin (phenytoin), which rarely causes acute hepatitis associated with a skin rash and, if the reaction is severe, the patient can develop hepatic failure or the Steven–Johnson’s syndrome (23). Cross-sensitization with phenobarbital or diphenylhydantoin carbamazepine precludes their use when phenytoin is implicated (3). However, valproic acid, another agent known to cause liver injury, does not cross-react with phenytoin. Diphenylhydantoin-induced liver injury is thought to occur via conversion of the drug to an intermediate that forms a hapten with cellular proteins eliciting an immune response. In contrast, valproic acid appears to interfere with fatty acid metabolism, leading to fatty liver and liver failure (24). Thus, a patient who had an adverse hepatic reaction to hydantoin is at no increased risk of liver injury from valproic acid. Coadministration of other

348

Boyer

antiseizure medications with valproic acid may, however, increase the risk of valproic acid hepatotoxicity (23,24). One of the most recent agents to cause hepatotoxicity leading to its withdrawal from the market is troglitazone. This drug caused severe liver injury leading to hepatitis and liver failure in a number of patients (25). Although other thiazolidinediones have a lower incidence of hepatotoxicity, they are not without risk (26,27). The structures of all of the thiazolidinediones are quite similar; however, in the absence of knowing exactly how they cause liver injury, it is difficult to judge the risk of toxicity of each. The three glitazones in use do not share a common degradative pathway and if the liver toxicity is due to the generation of a unique metabolic product, then cross-reaction should be rare (28). In a single case report, a patient developed mild liver injury while on troglitazone and was then switched to rosiglitazone and developed cholestatic hepatitis (29). It is difficult to judge the significance of a single case report, but pending further studies, the use of another glitazone in a patient with hepatitis secondary to troglitazone should be discouraged. In conclusion, the decision to use an agent in a patient who has had an adverse reaction to another drug in the same class should be based on the following factors. The similarity of the new drug to the toxic drug should be known as far as their chemical structures are concerned, so that some estimate of the risk of cross-reaction can be determined, i.e., none for the patient with history of diphenylhydantoin hepatitis in whom valproic acid is to be used versus some for the patient who had halothane hepatitis and now is going to receive enflurane. In the latter situation, an alternative agent should be used. The risk to the patient must always be balanced against the need for the potentially toxic drug. If the first drug caused cholestasis, then the risk of hepatic failure is less than if a hepatocellular-type reaction occurred. In most situations where a patient has suffered a hepatotoxic drug reaction, it is best to use a drug of a different class or the one that has a very dissimilar structure. TREATMENT OF ADVERSE HEPATIC DRUG REACTIONS AND THEIR COMPLICATIONS Antidotes and Steroids for Hepatitis-Like Reactions N-Acetylcysteine (NAC) is the best example of an agent that truly prevents an adverse drug reaction. This agent limits the injury by the toxic metabolic products of acetaminophen by generating glutathione. In one report of 71 cases of acetaminophen overdose, either accidental or deliberate, 76% to 80% of patients received NAC and 10 patients developed acute liver failure and only five died. Without NAC, this death rate would have been much higher (30). L-Carnitine has been recommended for the treatment of valproate hepatotoxicity, but its efficacy is less clear (24). Most other adverse drug reactions have no specific therapy, and this leads to the use of less specific agents such as steroids and ursodeoxycholic acid in the hope of hastening recovery from the toxic hepatitis. There are no randomized controlled trials in the use of steroids in the treatment of hepatitis-like drug reactions. Steroids have been used to treat the type of reactions seen with diphenylhydantoin because of the hypersensitivity appearance of the reaction, i.e., eosinophilia, skin rash, etc. However, there is no evidence that the use of steroids reduces the severity of the disease. Case reports of successful treatment (31) cannot be used as a justification for the use of steroids, as many patients will recover spontaneously. One exception is when the adverse hepatic reaction is autoimmune hepatitis as is seen following the use of alpha-methyldopa or nitrofurantoin (32,33). These patients may benefit from the use of steroids. Ursodeoxycholic Acid and Other Agents for Management of Cholestasis The development of cholestasis is a common reaction to a variety of different drugs. Ursodeoxycholic acid is commonly used to treat a variety of cholestatic disorders, such as primary biliary cirrhosis, cholestasis of pregnancy, and sclerosing cholangitis. There are limited reports on the use of ursodeoxycholic acid to treat toxic hepatitis. All of the reports are anecdotal and it is difficult to be certain whether the drug actually led to an improvement in the liver

349

Management of the Patient

TABLE 2 Variables Predictive of Death or Need for Transplantation with Drug-Induced Liver Disease Variable Age (years) Duration Rx (days) Bilirubin (!ULN) AST (!ULN) ALT (!ULN)

Died/transplantation

Recovered

p value

65 (47 77) 25 (10 94) 18.7 (12.60 25.3) 34 (14 59) 31 (15.7 56)

58 (41 74) 21 (10 49) 5.5 (3.3 9.5) 6.7 (3.4 17.1) 11.4 (5.7 24)

0.04 NS !0.0001 !0.0001 !0.0001

Abbreviations: Rx, treatment; ULN, upper limit normal; AST, aspartate aminotransferase; ALT, alanine aminotransferase; NS, not significant. Source: Adapted from Ref. 37.

disease. For example, one report described the successful treatment of terbinafine cholestasis with ursodeoxycholic acid, whereas in another report use of ursodeoxycholic acid was only partially effective in treating the cholestasis that developed due to the same drug (34,35). Ursodeoxycholic acid may well help the pruritus seen with cholestasis and its use is warranted because of this beneficial effect. Other drugs used to treat the pruritus are cholestyramine and rifampin. The author has found rifampin to be quite effective in the control of pruritus from drug-induced cholestasis, but careful monitoring of liver tests is required as w10% of patients taking rifampin will develop hepatotoxicity (36). ACUTE LIVER FAILURE The most dreaded complication of drug hepatotoxicity is the development of severe hepatitis leading to acute liver failure. Liver failure may be seen from the injury of a predictable hepatotoxin such as acetaminophen or from an idiosyncratic reaction such as has been seen with halothane, isoniazid, hydantoin, valproic acid, and troglitazone to name just a few. In the patient with drug-induced liver disease, the development of jaundice in association with high aminotransferases (Hy’s rule) is predictive of an increased risk of death or liver transplantation when compared with those who fail to manifest these features (Table 2) (37). Acute liver failure is defined as a prolonged prothrombin time (international normalized ratio, INRO1.5) and hepatic encephalopathy. Using this definition, acetaminophen is the most common cause of acute liver failure in the United States and idiosyncratic drug reactions are the second most common cause (Fig. 1) (39). Hence, one must be aware that a patient with drug-induced liver injury is at risk for developing hepatic failure, especially if they become jaundiced, and early rather than late referral to a liver transplant center is essential if the patient is to survive.

140

Number of patients

120 100 80 60 40 20

Acetaminophen Drug HBV Shock HAV Autoimmune Wilson Pregnancy Budd-Chiari Cancer Other

0

Unknown

FIGURE 1 Causes of acute hepatic failure in United States. Source: Adapted from Ref. 39.

350

Boyer

TABLE 3 Factors Predictive of Outcome with Acute Liver Failure Etiology of acute liver failure Acetaminophen Nonacetaminophen

Criteria for liver transplantation Arterial pH!7.0 or all of the following PTO100 sec, creatinineO3.4 mg/dL, and Grade 3/4 encephalopathy PTO100 sec (INRO6.5) or any three of the following NANB/drug/halothane as cause, jaundice to encephalopathyO7 days, age!10 yearsO40 years, PTO50 sec, bilirubinO17.4 mg/dL

Abbreviations: NANB, non-A, non-B hepatitis; PT, prothrombin time; INR, international normalized ratio. Source: Adapted from Ref. 40.

Acetaminophen-induced liver injury is a special case, as the time of exposure and dose is usually known and an effective antidote is available if administered in time. Very few patients with acetaminophen overdose require transfer to a liver transplant center. Those who are treated early with NAC and who develop minimal elevation of liver tests can be managed at a local hospital. Even those who develop acute liver failure due to acetaminophen have a good prognosis with approximately 68% of patients surviving without a liver transplant (39). Patients with a suicidal attempt may have a better prognosis than those with an accidental overdose of acetaminophen (30), although in a recent report transplant-free survival was 67% for those with accidental overdose when compared with 71% for those with suicidal overdose following development of acute hepatic failure (39). Criteria for liver transplant following acetaminophen toxicity have been developed (Table 3) but are of poor sensitivity and specificity (39,40). Patients with a suicidal overdose of acetaminophen with acidosis or who develop a rising creatinine or prothrombin time should be referred to a liver transplant center. The patient who presents with an accidental overdose may be at greater risk of acute liver failure (30) and those with a prolonged prothrombin time (INRO2.0) should be referred to a transplant center. Patients who develop acute liver failure have a very small window in which to receive a liver transplant and waiting for the development of hepatic encephalopathy before referral to a transplant center is a mistake. Drug-induced liver injury other than acetaminophen is the second most common cause of acute liver failure in the United States. These patients have only about a 25% chance of recovering without a liver transplant (39). The most common drugs causing acute liver failure associated with death or liver transplantation in a recent report from Europe covering the years 1970–2004 included flucloxacillin, halothane, diclofenac, naproxen, isoniazid, disulfiram, chlorpromazine, ciprofloxacin, enalapril, and trimethoprim/sulfamethoxazol (37). In the United States, isoniazid, bromfenac, and troglitazone were the most common causes of acute liver failure (39). The critical steps in treating these patients include making the association between the liver injury and the drug (Table 1) and then stopping the drug. If the drug can be stopped before the patient becomes jaundiced, then their prognosis is much improved. Once the bilirubin exceeds twice the upper limit of normal, then the mortality/transplantation rate is 9.2% (37). Patients with hepatocellular-type injury have the highest death/transplantation rates (12.7%), but even patients with a predominately cholestatic pattern are at risk for death/ transplantation (7.8%) once they become jaundiced (37). Thus, the development of jaundice in a patient with a hepatotoxic drug reaction should prompt a call to a transplant center and if the patient develops worsening coagulopathy despite the administration of vitamin K, then a referral to a center is essential. LIVER TRANSPLANTATION It is beyond the scope of this chapter to discuss liver transplantation for the patient with acute drug-induced liver failure and only a few points will be mentioned. Outcomes following liver transplant for acute hepatic failure are not as good as is seen for chronic liver disease. The poor outcomes reflect the acuity of the illness and the risk of cerebral edema and infection that are not seen in patients transplanted for chronic liver disease. Overall survival rates are 77%, 67%, and 60% at 1, 5, and 10 years, respectively, for patients undergoing liver transplant for acute liver failure. Etiology, including drug-induced liver failure, is not associated with survival. Using

Management of the Patient

351

multivariate analysis, body mass indexR30 kg/m2, serum creatinineO2 mg/dL, ageO50 years, and history of life support are associated with decreased five-year survival (38). As the patient’s coma deepens, their likelihood of survival declines (Table 3). For idiosyncratic drug reactions, transplant-free survival is only 25%, hence clearly for this group of patients liver transplantation is lifesaving compared with no intervention (39).

SUMMARY Some drugs, like statins, commonly cause elevated liver tests but rarely cause serious liver injury, so their use is associated with a large margin of safety, and continued use of the drug despite elevated liver test may be warranted. Drugs such as isoniazid also commonly cause elevated liver tests but may also cause serious liver injury leading to liver failure. Thus, the margin of safety is less with these types of drugs and more careful monitoring of liver tests is required. Lastly, drugs such as phenytoin or halothane cause liver test elevations that may rapidly process to jaundice and serious liver injury when there is an immune allergic picture, and thus the drugs must be discontinued when they are believed to be the causative agent. Use of drugs in the same chemical class of a drug known to cause liver injury in a specific patient is not encouraged if there is a suitable alternative. Use of drugs of different chemical structure but having the same medicinal effect as a drug that has caused a toxic hepatitis, i.e., phenytoin and valproic acid, is quite safe, whereas use of enflurane in a patient with a history of halothane hepatitis is not. The key to the management of the patient with drug-induced liver injury is to make the correct diagnosis, stop the offending agent, and administer an antidote if available. Once the patient becomes jaundiced, the risk of liver failure increases and early referral to a liver transplant center is essential. Lastly, new drugs will continue to enter the market and a few will be associated with liver injury. The clinician must always be aware of this risk and if a patient on a new drug develops liver test abnormalities, there should be concern that the drug is the cause of the liver injury. Reporting all possible cases of drug-induced liver injury to the Food and Drug Administration is essential if we are to minimize the risk of serious hepatic injury from newly developed drugs. REFERENCES 1. Larrey D. Epidemiology and individual susceptibility to adverse drug reactions affecting the liver. Semin Liver Dis 2002; 22:145–55. 2. Sgro C, Clinard F, Ouazir K, et al. Incidence of drug-induced hepatic injuries: a French populationbased study. Hepatology 2002; 36:451–5. 3. Duh M-S, Walker AM, Kronlund KH. Descriptive epidemiology of acute liver enzyme abnormalities in the general population of central Massachusetts. Pharmacoepidemiol Drug Saf 1999; 8:275–83. 4. Lemozit JP, Petit de la Rhodiere G, Lapeyre-Mestre M, et al. A comparative study of adverse drug reactions reported through hospital and private medicine. Br J Clin Pharmacol 1966; 41:166–8. 5. Maria V, Victorino R. Development and validation of a clinical scale for the diagnosis of drug-induced hepatitis. Hepatology 1997; 26:664–9. 6. Zimmermann HJ. Hepatic injury from the treatment of infectious and parasitic diseases. In: Zimmerman HJ, ed. Hepatotoxicity. 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins, 1999:613–9. 7. Brown SJ, Desmond RV. Hepatotoxicity of antimicrobial agents. Semin Liver Dis 2002; 22:157–67. 8. Tolman KG. The liver and lovastatin. Am J Cardiol 2002; 890:1374–80. 9. Chalasani N, Aljadhey H, Kesterson J, Murray MD, Hall SD. Patients with elevated liver enzymes are not at higher risk for statin hepatotoxicity. Gastroenterology 2004; 126:1287–92. 10. Boyer TD, Sun N, Reynolds TB. Allopurinol hypersensitivity vasculitis and liver damage. West J Med 1977; 126:143–7. 11. Lewis JH, Ranard RC, Caruso A, et al. Amiodarone hepatotoxicity: prevalence and clinicopathologic correlations among 104 patients. Hepatology 1989; 9:679–85. 12. Larrey D, Vial T, Miclaeff A, et al. Hepatitis associated with amoxycillin–clavulanic acid combination: report of 15 cases. Gut 1992; 33:368–71. 13. Zimmermann HJ. Anesthetic agents. In: Zimmerman HJ, ed. Hepatotoxicity. 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins, 1999:457–82.

352

Boyer

14. Varma RR, Whitesell RC, Iskandarani MM. Halothane hepatitis without halothane: role of inapparent circuit contamination and its prevention. Hepatology 1985; 5:1159–62. 15. Navarro VJ, Senior JR. Drug-related hepatotoxicity. N Engl J Med 2006; 354:731–9. 16. Kaplowitz N. Biochemical and cellular mechanisms of toxic liver injury. Semin Liver Dis 2002; 22:137–44. 17. Lee WM. Drug-induced hepatotoxicity. N Engl J Med 2003; 349:474–85. 18. Spielberg SP, Gordon GB, Blake DA, et al. Predisposition to phenytoin hepatotoxicity assessed in vitro. N Engl J Med 1981; 305:722–7. 19. Christ DD, Satoh H, Kenna JG, et al. Potential metabolic basis for enflurane hepatitis and the apparent cross-sensitization between enflurane and halothane. Drug Metab Dispos 1988; 16:135–40. 20. Njoku D, Laster MJ, Gong DH, et al. Biotransformation of halothane, enflurane, isoflurane, and desflurane to trifluoroacetylated liver proteins: association between protein acylation and hepatic injury. Anesth Analg 1977; 84:173–8. 21. Sigurdsson AB, Hreidarsson AB, Thiodleifsson B. Enflurane hepatitis. A report of a case with a previous history of halothane hepatitis. Acta Anaesthesiol Scand 1985; 29:495–6. 22. Lewis JH, Zimmerman HJ, Ishak KG, et al. Enflurane hepatotoxicity. A clinicopathologic study of 24 cases. Ann Intern Med 1983; 98:984–92. 23. Zimmermann HF. Psychotropic and anticonvulsant agents. In: Zimmerman HJ, ed. Hepatotoxicity. 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins, 1999:483–516. 24. Konig SA, Schenk M, Sick C, et al. Fatal liver failure associated with valproate therapy in a patient with Friedreich’s disease: review of valproate hepatotoxicity in adults. Epilepsia 1999; 40:1036–40. 25. Maddrey WC. Drug-induced hepatotoxicity. J Clin Gastroenterol 2005; 39:S83–9. 26. Forman LM, Simmons DA, Diamond RH. Hepatic failure in a patient taking rosiglitazone. Ann Intern Med 2000; 132:118–21. 27. Menees SB, Anderson MA, Chensue SW, et al. Hepatic injury in a patient taking rosiglitazone. J Clin Gastroenterol 2005; 39:638–40. 28. Scheen AJ. Hepatotoxicity with thiazolidinediones. Is it a class effect? Drug Saf 2001; 24:873–88. 29. Bonkovsky HL, Azar R, Bird S, et al. Severe cholestatic hepatitis caused by thiazolidinediones: risks associated with substituting rosiglitazone for troglitazone. Dig Dis Sci 2002; 47:1632–7. 30. Schiodt FV, Rochling FA, Casey DL, et al. Acetaminophen toxicity in an urban county hospital. N Engl J Med 1997; 337:1112–7. 31. Prendergast KA, Berg CL, Wisniewski R. Troglitazone-associated hepatotoxicity teated successfully with steroids. Ann Intern Med 2000; 133:751. 32. Sharp JR, Ishak KG, Zimmerman HF. Chronic active hepatitis and severe heaptic necrosis associated with nitrofurantoin. Ann Intern Med 1980; 92:14–9. 33. Rodman JS, Deutsch DJ, Gutman SI. Methyldopa hepatitis: a report of six cases and review of the literature. Am J Med 1976; 60:941–8. 34. Ajit C, Suvannasankha A, Zaeri N, et al. Tebinafine-associated hepatotoxicity. Am J Med Sci 2003; 325:292–5. 35. Agca E, Akcay A, Simsek H. Ursodeoxycholic acid for terbinafine-induced toxic hepatitis. Ann Pharmacol 2004; 38:1088–9. 36. Prince MI, Burt AD, Jones DE. Hepatitis and liver dysfunction with rifampicin therapy for pruritus in primary biliary cirrhosis. Gut 2002; 50:436–9. 37. Bjornsson E, Olsson R. Outcome and prognostic markers in severe drug-induced liver disease. Hepatology 2005; 42:481–9. 38. Barshes NR, Lee TC, Balkrishnam R, et al. Risk stratification of adult patients undergoing orthotopic liver transplantation for fulminant hepatic failure. Transplantation 2006; 81:195–201. 39. Lee WM. Acute liver failure in the United States. Semin Liver Dis 2003; 23:217–26. 40. Sanyal AJ, Stravitz RT. Acute liver failure. In: Zakim D, Boyer TD, eds. Hepatology: A Textbook of Liver Disease. 4th ed. Philadelphia, PA: W.B. Saunders, 2003:445–96.

PART III HEPATOTOXICITY OF SPECIFIC DRUGS

18

Mechanisms of Acetaminophen-Induced Liver Disease Sidney D. Nelson and Sam A. Bruschi

Department of Medicinal Chemistry, University of Washington School of Pharmacy, Seattle, Washington, U.S.A.

INTRODUCTION Acetaminophen is the generic name in the United States for 4 0 -hydroxyacetanilide, the N-acetylated derivative of p-aminophenol (paracetamol is the generic name used in Great Britain and several other countries). This nonnarcotic analgesic/antipyretic, available over the counter in the United States since 1960, is one of the most widely used drugs in the world, and is available alone and in combination with many other drugs (1). In recommended doses, acetaminophen is considered to be efficacious and safe, and is not associated with the high incidence of gastrointestinal bleeding caused by aspirin and nonsteroidal anti-inflammatory drugs, or with the development of Reye’s syndrome. Although many different kinds of toxic effects have been attributed to acetaminophen use and abuse, except for hepatotoxicity and nephrotoxicity, their incidence is very low and in many cases considered insignificant (for reviews see Refs. 1,2). Even the incidence of nephrotoxicity is low, with an estimated occurrence of acute tubular necrosis of less than 2% of all acetaminophen poisonings (3), and there is an insignificant association between acetaminophen use and chronic renal injury, such as nephropathy (4,5). Thus, the major toxicity caused by acetaminophen is hepatotoxicity characterized by acute hepatocellular damage primarily in zone 3, the centrilobular region of the liver. It is not the intent of this chapter to describe the clinical and morphological characteristics of acetaminophen hepatotoxicity; readers are referred to several excellent reviews on acetaminophen for further information in this area (1,6–9) and chapter 19 which follows. Suffice it to say that acetaminophen is a major cause of acute liver failure in the United States, United Kingdom, and Australia (10–13), and accounts for a high percentage of inquiries to poison centers (O100,000 cases/yr) and deaths from poisonings (14,15). In the United States, acetaminophenassociated overdoses account for w56,000 emergency room visits and 26,000 hospitalizations yearly, with an annual mortality of w450 deaths, 100 of these unintentional (16). Initial results from a multicenter study suggest that approximately one-half of cases of acute liver failure are caused by acetaminophen, with 44% of those cases due to intentional overdose, and 56% classified as accidental or therapeutic misadventures where high therapeutic doses are taken by individuals who take more than one product containing acetaminophen, who are alcoholics, have other severe illness, and/or are malnourished (12). There is considerable controversy concerning both the acetaminophen dose exposure in these cases of “therapeutic misadventure,” and the importance of various factors in the susceptibility of individuals to acetaminophen hepatotoxicity, even at recommended therapeutic doses of the drug (17,18), where it now has been established that daily intake of 4 g of acetaminophen in healthy adults can lead to serum aminotransferase elevations of greater than three times the upper limit of normal in some individuals (19). There is evidence that restricting the availability of

354

Nelson and Bruschi

acetaminophen in the United Kingdom in 1998 by requiring blister packs of only 32 tablets has decreased the incidence of severe acetaminophen poisoning (20). The remainder of this chapter will focus on mechanisms of acetaminophen hepatotoxicity with discussion of events involved in initiation or early stages of hepatotoxicity followed by discussion of events involved in progression of the injury to hepatic necrosis. Some limited discussion of factors (e.g., age, diet, other drugs) that may influence the metabolism and disposition of acetaminophen will be included as they relate to the mechanisms proposed. EARLY EVENTS IN ACETAMINOPHEN-INDUCED HEPATOTOXICITY Pathways of Acetaminophen Metabolism Introduction and General Scheme After hepatotoxicity caused by acetaminophen was reported in animals (21,22) and humans (23,24) in the mid-1960s, several investigations commenced on the mechanism. In 1973, Mitchell and colleagues published a series of classic papers (25–27) that outlined a scheme for the metabolic activation of acetaminophen to an electrophilic quinone imine, N-acetyl-pbenzoquinone imine (NAPQI), that covalently bound to hepatic proteins, mostly in centrilobular liver cells that became necrotic. The role of glutathione (GSH) in protecting liver cells from injury also was elucidated (28), and this led to the development of N-acetylcysteine as an effective antidote that is still widely successful today in the management of acetaminophen toxicity (29–32). A scheme for the metabolism of acetaminophen is shown in Figure 1 (33). Work from several laboratories has contributed to the development of this and related metabolic schemes as reviewed in more detail elsewhere (1,2,7,34), and pathways and enzymes primarily related to metabolism in humans will be highlighted in the following discussion. With advances in analytical technologies, such as microbore LC columns and exact mass TOF MS instruments, additional minor metabolites of acetaminophen have been detected, but not yet fully characterized (35). Major Non-P450 Pathways The two major pathways of acetaminophen metabolism in all species are glucuronidation and sulfation of the phenolic group. After therapeutic doses of acetaminophen, humans excrete approximately 50% of the dose as the phenolic O-glucuronide and approximately 30% as the O-sulfate. Based on studies with expressed UGTs (UDP-glucuronosyltransferases), isoforms 1A1, 1A6, 1A9, and 2B15 are major contributors to acetaminophen glucuronidation (36–38), and members of the SULT1 family of sulfotransferases are major contributors to acetaminophen sulfation (39). Decreased elimination of acetaminophen as its glucuronide in congenic rats with a hereditary deficiency in UDP-glucuronosyltransferase activity (40) or in cats with a similar deficiency (41) makes these animals significantly more susceptible to acetaminophen-induced hepatotoxicity, and in humans, increased susceptibility to acetaminophen-induced hepatotoxicity may occur in Gilbert’s disease (42,43). However, except for one case report of a possible Zidovudine interaction (44), there is no clinical evidence that patients who receive drugs that inhibit the glucuronidation of acetaminophen are at increased risk of acetaminophen-induced hepatotoxicity (1,33). Moreover, there have been no reports of effects on the sulfation pathway where increased toxicity to acetaminophen has been observed. However, recent reports (37,38) provided data demonstrating that inhibition of UGTs by phenobarbital and phenytoin enhances cytotoxicity caused by acetaminophen in cultured human hepatocytes, and the authors suggest that this interaction may explain in part the hepatotoxicity observed in some patients on combinations of these drugs (45–48). It should be noted that both phenobarbital and phenytoin also are inducers of cytochrome P450s (CYPs) that oxidize acetaminophen to its toxic metabolite, and that the concentrations of acetaminophen used in the human hepatocyte experiments were high and may not be relevant to the situation in vivo. Carboxylesterases of the CES1 family and N-acetyltransferases (both NAT1 and NAT2) likely contribute to the hydrolysis of approximately 10% of a dose of acetaminophen to p-aminophenol and its reacetylation back to acetaminophen, a “futile cycling” that has been

355

Acetaminophen-Induced Liver Disease

NH2

OH p-Aminophenol Esterase/ Amidase

O HN C CH3

O S OH O

O

N-Acetyltransferases O HN C CH3

PAPS Sulfotransferases

Acetaminophen-O-Sulfate Cytochrome P450

UDPGA Glucuronosyltransferases

OH

Acetaminophen Cytochromes P450

NAD(P)H and reductases

O OH

O

N C CH3 H 2O

SG OH

O p-Benzoquinone reductases

Cysteine, Mercapturate, + other thiol conjugates

OH

O HN C CH3

O N-Acetyl-p-benzoquinone imine (NAPQI)

O

HO

O O OH OH OH

Acetaminophen-O-Glucuronide Cytochromes P450 O HN C CH3 Glucuronides and sulfates (+ small OH amounts of thiol OH conjugates) 3´-Hydroxyacetaminophen COMT O HN C CH3

Glutathione-Stransferases O HN C CH3

GSH

OH

OH Hydroquinone

O

OCH3

Glucuronides and sulfates (+ small amounts of thiol conjugates)

3´-Methoxyacetaminophen

SG OH 3´-(S-Glutathionyl)acetaminophen

Glucuronides and sulfates O HN C CH3

OH

S CH2

H O C C OH HN C CH3

O 3´-(S-Mercapto)acetaminophen

O HN C CH3

O HN C CH3

OH

S CH2

H O C C OH NH2

3´-(S-Cysteinyl)acetaminophen

OH

(+ Sulfoxide and sulfone) S CH3

3´-Thiomethylacetaminophen

FIGURE 1 Metabolic pathways of acetaminophen. Bold arrows indicate major pathways, normal arrows indicate intermediate pathways, and broken arrows indicate minor pathways. Benzoquinone metabolites have been detected only in mice, whereas all other pathways have been detected in several species, including humans.

356

Nelson and Bruschi

documented in rats (49). The extent to which this occurs in humans is unknown, but may be more related to the risk of nephrotoxicity from acetaminophen than hepatotoxicity, since p-aminophenol is a known nephrotoxicant that has been implicated in nephrotoxicity caused by acetaminophen in rats (50,51). With regard to drug interactions, it has been found that therapeutic doses of acetaminophen in humans can inhibit the polymorphic NAT2 (52). CYP Oxidation Pathways CYP P450-catalyzed oxidation of acetaminophen to NAPQI is the major metabolic pathway leading to hepatotoxicity (27,53). Originally, N-hydroxyacetaminophen was proposed as a precursor to NAPQI (29), but kinetic (54) and carrier-trapping (55) experiments both showed that if this metabolite is formed, it decomposes before it leaves the CYP active site. NAPQI has been synthesized (56,57), and has the chemical, biochemical, and toxicological properties consistent with its role as the major ultimate toxic metabolite of acetaminophen (56–61). Because of its short half-life in biological systems [w0.7 seconds (61)], NAPQI is usually detected as thioether conjugates (62,63), but it has been detected directly in incubations of acetaminophen with purified rat CYP1A1 (64). Also, because of its reactivity, NAPQI causes necrosis of cells in the periportal region of the livers of rats infused with NAPQI through the portal vein (Fig. 2), rather than in the centrilobular region as occurs when acetaminophen is administered (see Ref. 1 for a review and photomicrographs). Thiol ether metabolites of acetaminophen are excreted into urine as an indicator of NAPQI formation (Fig. 1), and represent 5% to 10% of normal therapeutic doses in humans (1,65). However, this is probably an underestimation of the extent of oxidation of acetaminophen to NAPQI because this highly reactive quinone imine can be reduced back to acetaminophen by several reductases and their reducing cofactors including NADPH–CYP reductase and NADPH (53,66,67) and via ipso adduct decomposition reactions (68,69). A catechol metabolite of acetaminophen, 3-hydroxyacetaminophen (Fig. 1), is apparently formed in a classical CYP monooxygenase reaction, inasmuch as the hydroxyl group oxygen is

FIGURE 2 Section of rat liver obtained 5 hours after intraportal infusion of a 20 mg/kg dose of NAPQI in FC-43 emulsion (Oxypherol). Hematoxylin- and eosin-stained micrograph shows necrosis of cells proximal to the periportal vein and portal triad. The FC-43 vehicle does not cause cellular damage. Abbreviation: NAPQI, N-acetylp-benzoquinone imine.

357

Acetaminophen-Induced Liver Disease

derived from molecular oxygen (70). This catechol is nontoxic based on studies in mice (71), though small amounts of thioether metabolites of the catechol and its 3-O-methylated metabolite are formed (!0.5%) indicating their further oxidation to electrophilic quinones and quinone imines. Overall, the catechol and catechol-derived metabolites account for 4% to 8% of therapeutic doses of acetaminophen in humans (1,65). There are several reports concerning the specific CYP isoforms that are involved in the oxidation of acetaminophen to NAPQI in laboratory animals and humans, and the reader is referred to Refs. 1,2 for reviews. The following discussion will focus on relative efficiencies of expressed and, in some cases, purified human CYP isoforms with a comparison to information obtained with human liver microsomes and in vivo studies in humans. The data in Table 1 compare data obtained in our laboratory with human liver CYP isoforms purified to homogeneity from baculovirus expression systems (72–76). Other human CYP isoforms (1B1,2B6,2C19) were obtained as Supersomes (Gentest Corp.) and showed no detectable formation of either NAPQI or the catechol metabolite, whereas human CYP 1A1 was nearly as active as CYP 1A2. However, CYP 1A1 is not normally expressed in human liver, and is therefore unlikely to play a role in bioactivation of acetaminophen to its proximal reactive intermediate in hepatocytes. In assessing the data in Table 1, one should consider the following. First, as discussed previously, the kinetic parameters are only apparent for NAPQI formation since GS-APAP is the product of a very reactive metabolite that can undergo other undetectable reactions. Second, the values obtained with the purified enzymes are not the same as those obtained by others using partially purified preparations and transfected cells (77,78). However, the relative importance of these human isoforms in forming the major hepatotoxicant from acetaminophen is fairly consistent among the various studies with the exception of CYP 1A2. The possible significance of each major isoform is discussed below. CYP 1A2

Although purified CYP 1A2 is one of the most efficient catalysts of NAPQI formation, its importance in acetaminophen metabolism and toxicity is unclear. Induction of this isoform in humans (e.g., charcoal-broiled meat, cigarette smoking, omeprazole) does not increase the formation of NAPQI based on amounts of thioether metabolites formed (79–82). Furthermore, Cyp 1a2 null mice are not significantly less susceptible to acetaminophen-mediated hepatotoxicity than wild-type mice (83). However, CYP 1A2 is involved in acetaminophen oxidation to NAPQI in human liver microsomes (78), and when mice have deletions of both the Cyp 2e1 and Cyp 1a2 genes, they are less susceptible to acetaminophen hepatotoxicity than just Cyp 2e1 null mice (84,85). These results suggest that some CYPs may function differently in intact cells, or TABLE 1 Kinetic Parameters for Purified Human CYP Isoforms Involved in Acetaminophen Oxidation to Its Toxic Metabolite, NAPQI, Measured as Its Glutathione Conjugate, GS-APAP, and the Nontoxic Catechol Metabolite, 3-OH-APAP GS-APAP Purified human isoform CYP CYP CYP CYP CYP CYP CYP a

1A2 2A6 2C8 2C9 2D6 2E1 3A4

3-OH-APAP

Km (mM)

Vmax (nmol/min per nmolP450)

V/K

1.4 4.6 1.0 1.1 1.8 1.3 0.14

14.4 7.9 0.2 0.1 3.0 6.9 1.5

10.3 1.7 0.2 0.1 1.7 5.2 10.5

Km (mM) ND 2.2 ND ND ND 4.0 ND

Vmax (nmol/min per nmolP450) 0.1 14.2 0.1 ND ND 2.5 0.1

V/K NDa 6.5 NDa NDb NDb 0.6 NDb

Although CYP 1A2, CYP 2C8, and CYP 3A4 did form measurable amounts of the catechol metabolite (Vmax, z0.1 nmol/min per nmol P450), limits of detection of the HPLC/EC assay were not sufficient to accurately determine Km values. CYP 2C9 and CYP 2D6 did not form detectable amounts of the catechol metabolite. Abbreviations: NAPQI, N-acetyl-p-benzoquinone imine; GS-APAP, 3-(Glutathion-S-yl) Acetaminophen; 3-OH-APAP, 3-Hydroxyacetaminophen; CYP, cytochrome P450. b

358

Nelson and Bruschi

even in their membrane-bound states, which is consistent with much higher Km (3–4 mM) and lower Vmax (0.3–3 nmol/min/nmol P450) values for CYP 1A2 in these preparations (72,77,78). Whether it is altered reductase interactions or other factors that modulate CYP 1A2 activity in membranes is unknown. CYP 2A6

CYP 2A6 appears to be the major human CYP isoform that catalyzes the oxidation of acetaminophen to its catechol metabolite [Table 1, (73)]. It also forms NAPQI, but the relatively high Km for this reaction suggests that this would be important only at relatively high hepatotoxic doses of acetaminophen. There is one report that methoxsalen, a selective inhibitor of CYP 2A6, decreases the formation of NAPQI from acetaminophen in humans (86). Methoxsalen does decrease hepatotoxicity caused by acetaminophen in mice (87). CYP 2D6

Overall CYP 2D6 plays a minor role in the oxidation of acetaminophen to NAPQI (ranging from 4.5% to 22% of total thioether conjugates formed in a panel of human liver microsomes) as would be expected based on kinetic parameters [Table 1, (74)]. However, CYP 2D6 is polymorphically expressed in humans and it could contribute significantly in CYP 2D6 ultrarapid and extensive metabolizers (88), and put some populations at greater risk of acetaminophen hepatotoxicity (89). CYP 2E1

Although CYP 2E1 only accounts for about 5% to 10% of total P450 in most human livers (90), this isoform appears to be the major isoform that catalyzes the oxidation of acetaminophen to NAPQI in humans (73,77,78,91). It also forms significant amounts of the catechol metabolite [Table 1, (73)]. CYP 2E1 will be discussed further below. CYP 3A4

CYP 3A4 is the most efficient P450 in the oxidation of acetaminophen to NAPQI [Table 1, (76,77)]. Its high abundance in human liver (about 30% of total hepatic P450) (90) and relatively low Km would suggest that CYP 3A4 should be the most important contributor to NAPQI formation from acetaminophen given at therapeutic doses in humans. However, studies in vitro in human liver microsomes using troleandomycin as an inhibitor of CYP 3A4 (76), and in vivo in humans using rifampin to induce CYP 3A4 (91), indicate that CYP 3A isoforms only contribute to about 10% of the oxidation of acetaminophen to NAPQI. Because of its low Km and low Vmax, it is unlikely that CYP 3A4 is a major contributor at concentrations of acetaminophen (1 mM) that are normally achieved in cases of hepatotoxicity (92). Factors that Influence Acetaminophen Metabolism and Hepatotoxicity This section will primarily focus on a brief discussion of those cases where there is substantial evidence for modulating the hepatotoxicity caused by acetaminophen in humans. A much more detailed discussion is provided in Ref. 1. However, some discussion of the substantial work carried out (mostly over the past five years) in nuclear receptor and drug transporter gene knockout mice will be included. Age One of the few factors that has been reasonably well documented in humans that affects the incidence of hepatotoxicity is age, and only in the sense that young children appear to be less susceptible to hepatotoxicity caused by acetaminophen than adults (93). This has been attributed to the increased rates of sulfation of acetaminophen in children (94), and to the increased rates of GSH resynthesis when the liver is challenged based on studies in rats (95). When used properly in children, acetaminophen appears to be a very safe drug (96), but unintentional multiple overdoses of acetaminophen in children have led to several cases of hepatotoxicity (97–102). A recommendation is that 60 to 90 mg/kg per day is a reasonable

Acetaminophen-Induced Liver Disease

359

therapeutic dose, but that repeated supratherapeutic dosing of greater than 90 mg/kg per day is inappropriate (103). Drug–Drug Interactions Although several chemicals, including many drugs, can induce CYPs that oxidize acetaminophen to NAPQI (29) (Table 14.1 in Ref. 1), in only a few cases in humans on anticonvulsant drugs has the induction apparently led to hepatotoxicity caused by acetaminophen (45–48,104– 107). However, some anticonvulsants and other drugs that induce CYPs, have additional effects on acetaminophen metabolism (37,81,108) that may either potentiate or ameliorate toxicity. Only recently have mechanisms for induction by many of these drugs been elucidated that have begun to reveal the complex nature of some drug–drug interactions, such as with acetaminophen. In the past few years, a group of orphan nuclear receptors [constitutive androstane receptor (CAR), pregnane X receptor (PXR) and retinoid X receptor (RXR)] have been found to regulate the expression of several drug-metabolizing enzymes and drug transporters, and these receptors can be activated by many known drug inducers such as phenobarbital (109,110). RXRalpha is an obligatory partner of CAR and PXR, and RXRalphadeficient mice are resistant to acetaminophen-induced hepatotoxicity, apparently by expressing lower levels of Cypla2 and Cyp3a11 (111). CAR knockout mice also are protected from the hepatotoxic effects of acetaminophen due to lowered expression of Cyp1a2, Cyp3a11, and GST pi, and humanized CAR mice pretreated with phenobarbital showed increased hepatotoxicity compared with untreated mice (112). Results with PXR null mice have been discordant with some studies showing increased susceptibility (113) and others showing either no difference (114) or decreased susceptibility (115) to acetaminophen-induced hepatotoxicity. Results of studies reported in the latter study demonstrated that despite an approximate threefold increase in CYP3A protein and activity in livers of PXR null mice, they were less susceptible to acetaminophen toxicity most likely because of threefold higher levels of Cyp1a2 in wild-type mice along with 1.5-fold increases in acetaminophen absorption (115). The difference in effects of knocking out these nuclear receptors in mice of different genetic backgrounds may provide a partial explanation for the finding that induction of CYP3As in humans has not been shown to have any significant effect on acetaminophen toxic metabolite formation (91). Murine and human nuclear receptors appear to differ substantially in their response to drugs and other regulatory factors, so that their significance in modulating acetaminophen hepatotoxicity in most humans is questionable (116). Additional complicating factors are the unknown effects of most drugs on drug transporters. Acetaminophen glucuronide is excreted into the bile by a canalicular multidrug resistance–associated protein (Mrp2) and into blood by another related protein, Mrp3 (117). Acetaminophen sulfate is excreted into the bile by both Mrp2 and a breast cancer–resistance protein, BRCP (118). However, the sulfate metabolite appears to be transported into blood via transporters different from Mrp3 (118). Mrp3 knockout mice are less susceptible to hepatotoxicity caused by acetaminophen (119), and acetaminophen pretreatment of rats induces Mrp3 selectively, leading to decreases in hepatotoxicity in this species as well, by reducing liver exposure to acetaminophen through decreased enterohepatic recirculation (120). How drugs and other factors affect these transporters in humans is unknown. Another complication in drug–drug interactions with acetaminophen is that some drugs can both inhibit and induce CYPs. For example, propoxyphene is an analgesic often used in combination with acetaminophen that can cause deaths in humans. This has been attributed to respiratory depression caused by propoxyphene although some cases may involve induction of acetaminophen metabolism and hepatotoxicity (121). In contrast, propoxyphene forms a tightbinding metabolite inhibitory complex with P450 heme (122) that may have protected against hepatotoxicity after ingestion of a normally hepatotoxic dose of acetaminophen by an individual (123). Other drugs known to inhibit and induce CYP isoforms involved in acetaminophen oxidation in humans are isoniazid (124,125) and ethanol (126,127). A model of time-dependent induction of CYP 2E1 by ligand stabilization, such that inhibition is observed while the inducer is present and enhancement of activity occurs with removal of the inducer, has been proposed (128) and was very predictive of the ethanol–acetaminophen interaction in humans (127).

360

Nelson and Bruschi

Also, isoniazid was found to inhibit the oxidation of acetaminophen to NAPQI in humans for nearly 24 hours after coadministration (125,129), but between 24 and 48 hours after administration, significant increases in its formation were observed that may have contributed to hepatotoxicity and nephrotoxicity in some patients who received acetaminophen after taking isoniazid (130–132). Although several studies suggest that chronic ingestion of alcohol puts humans at greater risk of acetaminophen-mediated hepatotoxicity (for reviews see Refs. 1,133–135), the majority of cases appear to involve accidental overdoses of acetaminophen in chronic alcohol abusers (136,137). One study has even shown that maximal daily doses of acetaminophen have no significant effect on the livers of alcoholic patients (138). This, and other information (139) has led Rumack (18) to suggest that consumption of alcohol leads to hepatotoxicity only in highrisk individuals whose acetaminophen levels are above 300 mg/mL, or approximately a dose of 8 g at one time. However, as pointed out by Kaplowitz (17), if some cases of acetaminophen poisoning are idiosyncratic in nature, then additional factors may lower the dose required to elicit toxicity in some individuals. As previously discussed (126,127), induction of CYP 2E1 by ethanol stabilization of this isoform is likely to increase NAPQI formation only two to three fold, and additional factors such as other mechanisms of CYP 2E1 induction (139), decreases in hepatic GSH status, particularly mitochondrial GSH, (140–142), and poor nutritional status (143–146) are likely to be important contributors to hepatotoxicity. Based on studies in rats, there is a relatively narrow window of 12 to 24 hours during which ethanol, through its effects on CYP2E1 and GSH transport into mitochondria, would increase the risk of acetaminophenmediated hepatotoxicity three to four fold (147). Furthermore, induction of other CYP isoforms (such as CYP 3A isoforms) by ethanol and other alcohols present in alcoholic beverages may play a role in increased risk of hepatotoxicity caused by chronic alcohol ingestion based on studies in animals (148,149), although induction of CYP 3A isoforms in humans does not appear to significantly increase acetaminophen oxidation to NAPQI (91). Acetaminophen has been safely used for many years as the analgesic/antipyretic of choice in patients taking the narrow therapeutic index anticoagulant warfarin, with only a few case reports of enhanced anticoagulant effect (150–156). Despite some prospective studies that have shown no clinically significant changes in anticoagulation status of patients on warfarin who are administered acetaminophen (157–159), other studies (160,161) have suggested that acetaminophen is an under-recognized cause of excessive anticoagulation, especially in elderly patients. Although acetaminophen does not significantly inhibit CYP 2C9 hydroxylation of S-warfarin (162), unrecognized genetic and/or nongenetic factors may put some individuals at risk of hypocoagulation as a result of acetaminophen therapy with warfarin. For example, recent studies have shown that acetaminophen can reduce the function of blood clotting factor VII (163), and the reactive metabolite of acetaminophen, NAPQI, can inhibit enzymes of the vitamin K cycle (164). Mechanisms of Reactive Metabolite Formation Mechanisms of the oxidation of acetaminophen by CYPs to NAPQI, 3-hydroxyacetaminophen, and p-benzoquinone have been reviewed elsewhere (2,165). Briefly, differential CYP isoform selectivity in the formation of NAPQI and the catechol metabolite strongly implies that acetaminophen orients differently in the active sites of different CYP isoforms, such that hydrogen atom removal by a high valency heme iron–oxo complex, usually depicted as FeO3C or Fev aO, occurs to generate either an amide nitrogen radical that rebounds to form a ternary enzyme–oxy–acetaminophen complex that essentially dehydrates to yield NAPQI, or a phenoxy radical that rebounds via generation of a stabilized semiquinone aryl radical to form the catechol metabolite. Although there is no direct evidence to support these reactions, indirect evidence from NMR paramagnetic relaxation studies has demonstrated that acetaminophen binds to the resting, ferric form of different rat CYPs consistent with the selective formation of NAPQI by CYP 1A1 and 3-hydroxyacetaminophen by CYP 2B1 (166). Similar results were observed with two purified human CYPs (72). The results show that acetaminophen preferentially orients in CYP 2E1, which selectively forms NAPQI (73), with the amide group significantly closer to the

361

Acetaminophen-Induced Liver Disease

H H COOH

H

O

O N

>4.5A

H

H

CH3

H

~3.8 A

N

N Fe

N

N

HOOC (A)

H

H H COOH

O ~3.8A

H

H

O N H

CH3

~4.1A

N

Fe

N HOOC (B)

N N

FIGURE 3 (A) The favored orientation of acetaminophen in relation to the heme iron at the active site of CYP 2E1 based on 1H-NMR relaxation studies. (B) The favored orientation of acetaminophen in relation to heme iron at the active site of CYP 2A6 based on 1H-NMR relaxation studies. Abbreviation: CYP, cytochrome P450.

heme iron than the phenolic group (Fig. 3A), whereas the reverse is true with CYP 2A6 (Fig. 3B), which selectively forms the catechol metabolite (73). It should be emphasized that these results do not confirm mechanism, since it is known from time-resolved crystallographic studies of CYP 101 that changes in its active site structure occur at every step of the reaction (167), and structural changes upon reduction and oxygenation occur with other CYP active sites as well (168,169). Nonetheless, based on the distances from the heme iron obtained for acetaminophen in the NMR paramagnetic relaxation studies and a CYP 2E1 homology model, a possible mode of interaction of acetaminophen in the active site of human CYP 2E1 has been described (170), and awaits testing and refining. Reactive Metabolite Disposition The discussion in this section will focus on the reactions of the major reactive and toxic metabolite of acetaminophen, NAPQI. Although small amounts of quinones and quinoneimines of the catechol and its methylated metabolite are formed in mice (Fig. 1), they are formed in larger amounts from a nonhepatotoxic regioisomer of acetaminophen, 3-hydroxyacetanilide (171–173), and therefore, are unlikely to contribute significantly to hepatotoxicity. Since NAPQI is a quinone imine that is both a strong oxidant and electrophile, it can react in more than one way leading to both covalent and noncovalent modifications of cellular constituents (1,2,6,7,34,68,165,174–176). A major unresolved question is how important each of these modifications is in the pathogenesis of liver cell injury. Covalent Binding to Hepatocellular Proteins At the organ level, covalent binding of acetaminophen to the liver of mice occurs in hepatocytes in the centrilobular region, and this binding precedes cellular necrosis of those hepatocytes (25,177,178). One study (177) showed colocalization of CYP 2E1 in necrotic cells in the liver and

362

Nelson and Bruschi

other necrotic tissues in mice that contained protein-bound acetaminophen. Studies of postmortem human livers from acetaminophen overdose cases have shown acetaminophen adducts to proteins in the necrotic centrilobular regions (179), and a strong correlation between plasma ALT and 3-(cysteinyl-S-yl) acetaminophen protein adducts has also been observed in human acetaminophen overdose cases (180–182). NAPQI is a soft electrophile that appears to form covalent adducts primarily, though not exclusively, with cysteine residues on proteins (68,183–186). One study has provided evidence that oxidation of acetaminophen by mouse liver microsomes yields lysine amino adducts with some proteins (187), and minor amounts of noncysteinyl adducts of acetaminophen have been detected in mouse liver (188). Acetaminophen has been found to bind covalently in a stable enough form to several hepatic proteins in the mouse that has allowed for the identification of the proteins (for reviews, see Refs. 1,2,189–192). For many of the proteins, investigations have not been carried out to determine whether their function has been altered as a result of adduction by NAPQI, but for a few, decreased catalytic activity or changes in function have been measured (Table 2). The first protein, or group of proteins, that were well characterized as forming adducts with reactive metabolites of acetaminophen were 55 to 58-kDa cytosolic proteins that are very similar to cytosolic selenium-binding proteins whose function is not well understood (2,193,194). These proteins have also been found to decrease in concentration in mouse liver after acetaminophen treatment, but not after treatment with the nonhepatotoxic regioisomer 3-hydroxyacetanilide (Table 2) (202). This regioisomer does form reactive metabolites that covalently bind to the same set of selenium-binding proteins, but the adducts seem to be less stable (203,204). Preliminary evidence has been presented that the acetaminophen-adducted protein translocates to the nucleus as a possible signal of electrophile damage (201). A recent report also implicates a 56-kDa selenium-binding protein in intra-Golgi protein transport (205), and it will be of interest to determine whether acetaminophen affects this function. Acetaminophen-Induced Liver Disease

From Table 2, it can be seen that several cellular dehydrogenases form adducts with acetaminophen reactive metabolites, and where measured, their activities are significantly decreased. Such reactions would favor an oxidant state in cells, and ATP synthesis could be impaired. It is interesting that ipso adduct forms of NAPQI (see discussion below) resemble the NADH and NADPH products of cofactor reduction, which may explain in part the selectivity of NAPQI for these enzymes. Moreover, a mitochondrial housekeeping protein that is a precursor to thioredoxin reductase, a selenoprotein requiring NADPH for its activity (206–208), is also adducted [Table 2, (192)]. In general, mitochondria appear to be an important target for the pathogenesis of acetaminophen hepatotoxicity. This organelle sustains high levels of protein adduct formation after hepatotoxic doses of acetaminophen in mice. In comparison, there is little binding to mitochondrial proteins with doses of the relatively nonhepatotoxic analog 3-hydroxyacetanilide, despite similar amounts of overall cellular adduct formation (172,176,197,209). However, when mitochondrial GSH is depleted in mice, 3-hydroxyacetanilide does bind to hepatic mitochondrial proteins, alters mitochondrial function, and hepatotoxicity ensues (210). Mitochondria are also one of the earliest organelles to undergo both morphological (211,212) and functional (173,176,213–219) changes in hepatocytes after hepatotoxic doses of acetaminophen. Significant decreases in rates of ATP synthesis (219) and ATP concentrations (176,218–221) occur, and mitochondrial ATP synthetase appears to be a target of acetaminophen-mediated damage (192,222). Several other enzymes, and a few receptor and structural proteins, form covalent adducts with reactive metabolites of acetaminophen (Table 2), but it is not yet known if their activities or functions are affected. Furthermore, the time courses of adduct formation, repair, and/or degradation, and of activity and function are unknown for most of the proteins identified (see below). Finally, the effects of acetaminophen on human liver proteins and their activities have not been investigated. The use of human liver tissue, liver slices, and hepatocytes, and functional proteomics approaches should yield substantial new information in all of these areas.

363

Acetaminophen-Induced Liver Disease

TABLE 2 Modifications to Proteins in Mouse Liver After In Vivo Administration of Hepatotoxic Doses of Acetaminophen Proteins and subcellular localization Cytoplasm Selenium-binding proteins N-10-formyl tetrahydrofolate dehydrogenase Glyceraldehyde-3-phosphate dehydrogenase Albumin Glutathione peroxidases Thioether S-methyltransferase Aryl sulfotransferase Inorganic pyrophosphatase Proteasome subunit C8 Methionine adenosyl transferase (synthetases) Aldehyde dehydrogenases Osteoblast-specific factor 3 Glutathione S-transferase Pi Glutathione S-transferase Alpha Carbonic anhydrase III Sorbitol dehydrogenase (fragment) Glycine N-methyl transferase 3-Hydroxyanthranilate 3,4-dioxygenase Adenosine kinase Phosphoenolpyruvate carboxykinase Superoxide dismutase [Cu Zn] Thioredoxin peroxidase 1 Endoplasmic Reticulum Glutamine synthetase Calreticulin precursor (crp 55, erp 60) A probable protein disulfide isomerase (er-60) Mitochondria Aldehyde dehydrogenase Glutamate dehydrogenase Carbamyl phosphate synthetase-I ATP synthetase a-subunit Housekeeping protein Matrix protein P1 (hsp-60, GroEl protein) Mitochondrial stress-70 protein precursor (grp 75) Thioredoxin-dependent peroxide reductase 2 Nucleus Lamin A Cytoskeleton Tropomyosin 5 Actin Peroxisomes Likely 2,4-dienoyl-CoA reductase Urate oxidase Catalase Other Tumor necrosis factor, type 1 receptor associated protein Senescence marker protein-30 (smp-30)

Adducts (Ref.) Yes (190,192 194) Yes (195) Yes (183) Yes (192) Yes (192) Yes (192) Yes (192) Yes (192) Yes (192) Yes (192) Yes (192) Yes (192)

Activity (Ref.)a

YY (195) YY (183) YY (176,196)

YY Y [[ Y Y Y YY Y Y YY [c YYd

Yes (192) Yes (192) Yes (192) Yes (192)

Yes (197) In vitro only (187) In vitro only (187)

Protein levelb

Y Y YYe YY Yd YY (197,198)

Yes (192,199) YY (199) Yes (200) YY (200) Indirect evidence only (198) YY (198) Yes (192) Yes (192)

[ YYd Yd Yd Y YY YY YY

Yes (201) Yes (192) Yes (192) Yes (192)

Yd Yd Y Yd YY YY

See individual references for doses used and when livers were obtained for measurements. Double arrows indicate significant decreases in enzyme activity were observed. Note: The relatively nonhepatotoxic regioisomer, 3 0 hydroxyacetanilide, caused significantly less decrease in the activity of glyceraldehyde-3-phosphate dehydrogenase (183) and glutathione peroxidase (176). The effects of this isomer on other enzyme activities have not yet been determined. b See Ref. 202 for more information. A double arrow indicates significant decreases (YY) or increases ([[) in levels of proteins 8 hours after doses of 300 mg/kg in mice. A single arrow indicates a trend towards a decrease (Y) or increase ([). Note: The relatively nonhepatotoxic regioisomer, 3 0 -hydroxyacetanilide, did not significantly change protein expression levels except as indicated belowc,d. c The relatively nonhepatotoxic regioisomer, 3 0 -hydroxyacetanilide, caused a small increase in protein expression level. d The relatively nonhepatotoxic regioisomer, 3 0 -hydroxyacetanilide, caused a small decrease in protein expression level. e The relatively nonhepatotoxic regioisomer, 3 0 -hydroxyacetanilide, caused a significant decrease in protein expression level. a

364

Nelson and Bruschi

“Noncovalent” Interactions with Cellular Proteins As evident from Table 2, there are some hepatic proteins whose basal levels are decreased after acetaminophen administration to mice, yet no protein adducts have been identified. Additionally, it is known that hepatotoxic doses of acetaminophen affect some hepatocyte enzyme activities early in the pathogenesis of toxicity without detectable adduct formation. For example, there is evidence that acetaminophen, through its reactive metabolite NAPQI, inhibits calcium-dependent ATPases (173,176,223–227), but no adducts to these ATPases have been identified. Hepatocyte plasma membrane NaC/KC-ATPase activity is also significantly inhibited after hepatoxic doses of acetaminophen in rats (228), and despite attempts to detect adducts, none have been detected to this ATPase (191). Xanthine dehydrogenase in mouse liver is converted to its oxidase form after administration of hepatotoxic doses of acetaminophen (176,220) and may represent another indicator and mediator of oxidant stress. Finally, protein phosphatase activity is decreased prior to the evidence of cytotoxicity in mouse hepatocytes after exposure to toxic concentrations of acetaminophen, with no evidence of acetaminophen binding to the phosphatase enzymes (229). Although there are several possible explanations for these observations, including activation of proteases and/or signal transduction pathways, another possibility is that protein ipso adducts of NAPQI may form that react with cellular thiols to form S-thiolated proteins with modified activities. Glutathionyl and cysteinyl ipso adducts of NAPQI have been characterized, and the evidence for protein ipso adducts of NAPQI with inhibition of protein activity has been presented (68,69,226). It is unlikely that these ipso adducts are stable enough to be detected as protein adducts under the conditions of gel electrophoresis that have been used to separate adducted proteins prior to their identification by either mass spectral or immunochemical methods. However, there is evidence that S-glutathionylated proteins formed from these ipso adducts contribute to hepatocellular damage. First, both hepatic nonprotein and protein thiols are oxidized in mice within the first few hours after administration of hepatotoxic doses of acetaminophen (176,220). Second, some thiol compounds protect against hepatocellular injury caused by acetaminophen and NAPQI without significantly decreasing the levels of acetaminophen protein adducts (58,230–232). Third, 3-hydroxyacetaminophen covalently binds to proteins, but causes very little protein thiol oxidation and is relatively nonhepatotoxic (176,231). Finally, the time course of increases in glutathione disulfide (GSSG) concentrations in hepatic tissue in the pathogenesis of acetaminophen toxicity is consistent with the resynthesis of GSH a few hours after its initial rapid depletion by both covalent interaction with NAPQI and formation of protein ipso adducts of NAPQI (Fig. 4), and is consistent with reaction of the newly synthesized GSH with the ipso adducts to form GSSG and S-glutathionylated proteins (68,173,231). Alternatively, increases in GSSG in hepatocytes, observed particularly in mitochondria, does occur in the 4 to 6-hour time period after acetaminophen administration when hepatocytes begin to show overt signs of pathophysiological damage, and may simply be a result of tissue injury (175,233–235). Lipid Peroxidation and Related Oxidative Events Although some products of lipid peroxidation have been observed in mice and rats after hepatotoxic doses of acetaminophen (236–238), the products generally appear after hepatic damage has commenced, whereas they occur in the initiation stages of liver injury caused by such agents as carbon tetrachloride (239–242). In humans, F2-isoprostanes were measured as a sensitive marker of lipid peroxidation and were significantly (approx. ninefold) elevated in the plasma of 10 patients with acute liver and renal failure associated with acetaminophen overdose (243), but this was most likely late in the course after severe injury since serum F2-isoprostanes are not elevated in the first 6 hours after hepatotoxic doses of acetaminophen in rats (241). Other products of oxidative stress potentially caused by acetaminophen radicals and their secondary oxyradical products include protein carbonyls. However, these products were not increased in rats administered hepatotoxic doses of acetaminophen, though they were increased with the hepatotoxic redox cycling compound diquat (244). Protein carbonyls were modestly elevated in livers of mice administered hepatotoxic doses of acetaminophen, but only when the mice were pretreated with ferrous sulfate (245). It is also noteworthy that

365

Acetaminophen-Induced Liver Disease

O N C CH3

O H N C CH3

Protein S Protein−SH

O

O HN C CH3

G SH

O

OH +

H

+

Protein SH

Protein S S G

O HN C CH3

OH

S Protein

GSH Protein SH

+

GSSG

FIGURE 4 Possible reactions of NAPQI with protein thiols to generate a stable 3-protein thiol adduct and unstable ipso protein thiol adduct that can form S-glutathionylated proteins which can further react with GSH to regenerate the free protein thiol and oxidized glutathione (GSSG). Abbreviations: GSH, glutathione; GSSG, glutathione disulfide.

agents that protect against radical-mediated oxidant stress, such as the iron chelator deferoxamine, can delay the development, but not decrease the extent of hepatotoxicity caused by acetaminophen (246–248). Again, the data suggest that oxyradical-mediated oxidant events are not initiating events in the course of acetaminophen hepatotoxicity, but rather occur later, likely as a result of Kupffer cell activation (249,250), and protection afforded against acetaminophen hepatotoxicity by such agents as liposome-encapsulated superoxide dismutase in vivo in rats (251) is most likely related to scavenging of superoxide generated by phagocytic cells. One mechanism for radical-mediated oxidation would be redox cycling of NAPQI. Although the semiquinone imine radical of NAPQI can be formed either by one-electron oxidation of acetaminophen by cyclooxygenases (252,253) and peroxidases (254,255), or by oneelectron reduction of NAPQI (67,256), the semiquinone imine radical has a relatively high redox potential that would not favor direct reduction of oxygen to the superoxide radical anion (257). However, superoxide anion may be generated in coupled reactions of the semiquinone imine with GSH or NAD(P)H (258,259) or the ferrous-oxy form of CYP (260). Despite the preponderance of evidence which suggests that oxidative stress does not play a major role in the early (0–2 hours) metabolic phase of hepatocellular injury, there is substantial evidence both that proteins involved in mitigating oxidant stress are targets of the reactive metabolite of acetaminophen [Table 2, (261)], and that reactive nitrogen and oxygen species generated in mitochondria and from tissue macrophages (e.g., activated Kupfer cells) are critical in an oxidative, progression phase of injury (190,262–264). This topic will be described in more detail in the following section. LATE EVENTS IN ACETAMINOPHEN-INDUCED HEPATOTOXICITY Introduction The events following the metabolism and bioactivation of acetaminophen are generally considered to be those which are obligatory for the eventual expression of liver injury. In particular, there has been considerable interest in the generation of electrophilic metabolites, which are capable of covalently modifying cellular macromolecules. A widespread belief is that such covalent modifications of cellular constituents produces pathological consequences as a direct result of alterations in the function(s) of the species modified. This process has been referred to as the “covalent binding hypothesis” of cell death and organ damage (174,189,245,265).

366

Nelson and Bruschi

In the case of the liver injury produced by acetaminophen, an unequivocal and causal relationship between the loss of function of a specific protein “target” (for example) and eventual cell death has not yet been demonstrated. A complete prevention of injury has been demonstrated in mice deficient in particular CYP isoforms (85,266) but CYP isoforms are not targets for acetaminophen-generated reactive metabolites per se, i.e., they are not arylated by NAPQI. Despite being of fundamental scientific interest, these types of studies are far removed from clinical situations of acetaminophen overdose. From a clinical perspective, the cellular and organ sequelae which follow acetaminophen bioactivation by CYP enzymes are of greater interest. Consequently, the molecular and cellular mechanisms of acetaminophen-induced hepatotoxicity represent active areas of research and also have broader relevance to other drugand chemically induced injuries. In this section, the more recent studies examining the biochemical, cellular, and organ consequences of acetaminophen overdose will be reviewed and perspectives will be given as to relevance and potential clinical applicability in the future. Where possible, these will also be placed into perspective with the considerable body of work available on acetaminophen-mediated hepatotoxicity. For ease of study, the organ damage produced by acetaminophen can be subdivided into two components, namely an initial (or intrinsic) cell death phase followed by a delayed (or extrinsic) phase. These correlate approximately with the stage 1 and stage 2 classifications of Bessems and Vermeulen (2). For the purposes of this discussion, the two phases should correlate with a period before (intrinsic) and after (extrinsic) the loss of hepatocyte integrity within the region of liver injury. Such a reductionist perspective facilitates the analysis in time of acetaminophen’s toxicological actions and this is particularly important given the complexity of biochemical, cellular, and organism effects elicited by this drug. In many cases, it has been difficult to distinguish between potentially important events in acetaminophen’s toxic actions from mere “epiphenomena.” Consequently, future studies may more accurately define the cause–effect relationships that are important in the development of acetaminophen-induced hepatotoxicity. This classification is distinct from the clinical phases of acetaminophen toxicity as previously outlined (267). Initial clinical signs (which include anorexia, malaise, pallor, diaphoresis, nausea, and vomiting) are first evident within 24 hours and this is typically followed by upper-quadrant pain. Liver function abnormalities become apparent from between 24 and 48 hours if left untreated. Late-stage hepatotoxicity indicators (48–96 hours postingestion) include peak serum aminotransferases levels (which are released from within the dying hepatocyte), increased prothrombin time, encephalopathy, and coagulopathy. It is of interest to note that many diagnostic criteria—either alone or in combination—are insufficient to provide good prognostic indicators of outcome in acetaminophen overdose cases (268). Unexpectedly, perhaps, absolute or relative serum aminotransferase levels are also insufficient to predict outcome. Recoveries are even possible with late coagulopathy and encephalopathy present on the fourth day. The possible exception is development of an early and severe metabolic acidosis (268,269) which may have relevance to observations from more basic studies of acetaminophen-induced mitochondrial toxicities (190,270,271). (see section “The Intrinsic Cell Death Pathway: Macromolecular and Organellar Events” for a more detailed discussion on this topic.) Figure 5 represents a first attempt at ordering some of the significant events observed after toxic acetaminophen exposure. We have included potentially noteworthy cellular phenomena both from the exhaustive literature base and our own findings. This figure is inherently an approximation and limited by the numerous experimental models that have been utilized over the last 30 or so years. Nonetheless, Figure 5 does allow broad assessments of acetaminophen-induced pathological events and with refinement over time may come to serve as an aid in the critical evaluation of more data intensive proteomic and toxicogenomic methodologies [i.e., “phenotypic anchoring” of data, (272)]. The prospect remains that with further advances in our understanding of the cause–effect relationships of acetaminopheninduced liver pathology, more effective clinical strategies will become evident and increase the available treatment options—especially for late presenters.

367

Acetaminophen-Induced Liver Disease "Intrinsic" injury Phase 1 2h

4h

6h

"Extrinsic" injury Phase 2

loss of hepatocyte integrity 1d 8h 10h 12h

CYP-mediated bioctivation depletion { GSH, glycogen ATP protein arylation BAX translocation intraorganellar O2 stress metabonomic changes MAPK & JNK1 activation Ca2+ perturbations nitrotyrosine formation lipid peroxidation cell congestion at injury site* CyD-inhibitable MPT CyD-uninhibitable MPT DNA fragmentation chromation condensation protein synthesis mitochondrial fission mitochondria surrounded by ER steatosis protein breakdown chemokine/cytokine production oval cell activation mitochondrial structure changes ER & golgi remodelling/distruption advanced mitoch. autophagy Kupffer/macrophage activation mitosis, regeneration late stage sequelae**

2d

3d

4h

molecular & biochemical

macromolecular & organellar

cellular, tissue & organism

* incl. initial macrophage migration, rbc, PMNs, increased liver size. ** e.g.peak plasma ALT, AST; encephalopathy & coagulopathy

FIGURE 5 Selected molecular, biochemical, cellular, and tissue events observed following toxic acetaminophen administration have been categorized within the limitations of interspecies and inter-experimental model comparisons. The magnitude of the response to the parameter listed on the left is approximated graphically from lowest response (open circle) to highest (solid bar).

The Intrinsic Cell Death Pathway Molecular and Biochemical Events The earliest molecular and biochemical changes initiated by the reactive metabolite of acetaminophen (NAPQI)—and partially represented as upper group in Figure 5—can be grouped into the following broad categories: 1. 2. 3. 4.

Direct protein arylation events, Intracellular ion and intermediary metabolism alterations, Activation of intracellular signal transduction pathways, Generation of reactive oxygen (ROS) and nitrogen species (RNS)—either proteinaceous or nonproteinaceous.

Acetaminophen appears to be distinctive in its ability to generate early cellular perturbations through a combination of direct protein arylation events and ROS/RNS stress. Despite considerable attention, the relative contribution of protein covalent modifications in comparison to cellular redox alterations is still not well defined for acetaminophen-induced cytotoxicity or hepatotoxicity. This remains as a challenge for future studies and new technologies. Conceptually, however, the contributions from both direct protein arylation damage and redox changes to acetaminophen-induced cell death are plausible given the chemical properties of the toxic reactive intermediate. NAPQI—as a quinone imine—is both a strong oxidant per se and an electrophile capable of modifying proteins some of which also function as redox regulators (e.g., thioredoxin reductase & glutathione peroxidase) (176,192,273–277). In comparison, other compounds are known to initiate cell death exclusively or primarily by either covalently modifying proteins (e.g., chloramphenicol, tetrafluoroethylcysteine [TFEC], thioacetamide) or by altering the cellular redox state (i.e., “oxidative stress,” e.g., adriamycin, paraquat and tert-butyl hydroperoxide) (278,279).

368

Nelson and Bruschi

In short, although an excellent correlation has generally been observed between the extent of damage and the magnitude and location of covalent binding, some older studies have dissociated the two phenomena (280–282). There are no reported examples where cell death or liver injury has been caused by acetaminophen in the absence of covalent binding of the reactive intermediate (190). This can be interpreted in one of two ways. First, there exists a set of critical protein targets in the ultimate expression of acetaminophen-induced cell death. Alternatively, a threshold level of proteotoxicity may be necessary to manifest cellular injury. Lethal cellular injury to sublethal proteotoxic insult may thus be possible in circumstances of compromised cellular energetics and low ATP levels and this has been discussed at length previously elsewhere (283). As detailed earlier, and also in Table 2, proteomic and more traditional protein identification methods have identified a complete complement of cellular proteins modified by NAPQI. Their identities provide support for the perturbation of cellular homeostasis by redox and nonredox mechanisms (refer, for example, to arylation of the proteasome C8 subunit, ATP synthetase a subunit, thioredoxin reductase, and GSH peroxidase) (192,195,274,275). Significantly, some of the most prominently modified proteins either have poorly defined biological function (i.e., selenium- or acetaminophen-binding protein) or have many functions (glyceraldehyde-3-phosphate dehydrogenase) (284–286). Moreover, many in vitro and in vivo studies have examined the role of oxidative stress per se in acetaminophen-induced toxicity and have concluded that cellular oxidative stress plays an important role (276,287–289). Importantly, these in vitro studies also invariably indicate that injury progression can only be delayed and not totally prevented. Macromolecular and Organellar Events Acetaminophen-induced damage to parenchymal cell mitochondria is a prominent feature of the phase of injury (Fig. 5; middle group) and is increasingly recognized as an important, and perhaps even an obligatory, step in liver injury. Many of the observations concerning alterations to the mitochondria can be interpreted as either pathophysiological or reparative in nature. Both viewpoints are relevant and are dependent on factors such as the dose of acetaminophen, the experimental model used, and the time frame of observation. That mitochondria appear as central organelles is of interest as there are immediate parallels to research in the broader areas of apoptosis and autophagy. Earlier studies have defined the locus of acetaminophen action by indicating a preferential depletion of intramitochondrial GSH levels, adenine nucleotide pools, and redistributions of organellar calcium [e.g., (290–293)]. More recently, using metabonomic approaches, very early increases have been found in metabolic rates and mitochondrial fatty acyl chain resonances and these findings are consistent with toxicogenomic studies utilizing sublethal acetaminophen concentrations (270,294–296). Collectively these studies—from different laboratories and using various platforms—confirm that the arylation of important intramitochondrial energy generating proteins (e.g., ATP synthetase a subunit) has important metabolic and biochemical repercussions. Interestingly, metabonomic approaches have been used to phenotype rats prior to dosing with acetaminophen, and the phenotypes that are often segregated based on mitochondrial metabolites predict the susceptibility to acetaminopheninduced liver injury (297). Subcellular molecular changes often overlap both spatially and temporally with macromolecular and organellar responses (Fig. 5, orange). Recent proteomic and toxicogenomic studies point to these responses as either bona fide or truncated attempts at maintenance of homeostasis rather than as predefined and degenerative steps leading to cell death (270,275,277,298–303). The de novo upregulation of several common cytoprotective genes and proteins are found in response to acetaminophen exposure and include prominent examples such as metallothionein 1, heat shock proteins 40/60/70/105 kDa, several GSH transferase isoforms, heme oxygenase 1 and activating transcription factor 3. Although it is perhaps unrealistic to expect that a definitive meta-analysis of all acetaminophen proteomic and toxicogenomic studies can be undertaken with present technologies, these newly released data from several different laboratories add credibility to the concept of “genes/proteins of influence” in the process of acetaminophen cellular injury and hepatotoxicity. The injury produced by acetaminophen is morphologically akin to necrosis (304). Consequently, it was unexpected to find that a proapoptotic member of the BCL-2 family of

Acetaminophen-Induced Liver Disease

369

proteins known as BAX redistributes from the cytosol to mitochondria as a very early event in acetaminophen-induced hepatotoxicity in vivo (271). This finding has now been confirmed by another group (305,306). Moreover, we and others have reported that acetaminophen injury can be mostly prevented by either active sequestration of BAX with its anti-apoptotic binding partner BCL-xL (265) or by genetic ablation of the bax gene (307). It is also interesting to note that BAX is generally accepted to function in the direct control of apoptosis—via calcium ion sequestration—at the level of the mitochondrion and the endoplasmic reticulum (308–310). The significance of these findings is clear. Acetaminophen-induced BAX translocation to mitochondria is one of the earliest nonphysiological alterations to cellular function after metabolic activation and covalent binding. It is also amenable to external (i.e., pharmacological) manipulation. Specific inhibitors have been identified, or are being developed, to BAX based on natural equivalents [Ku70 (311–313), humanin (314,315) or BI-1 (316,317)]. Cell permeable pentapeptides designed from the BAX binding domain of the Ku70 component of a DNA strand break and repair complex have been reported and show early promise (313). The successful clinical development of BAX-specific inhibitors, analogous to those already identified and proven effective for the BH3-binding domain of BCL-xL (318), may help in preventing poisonings at a very early stage of acetaminophen intoxication. To be clinically useful, however, such inhibitors will need to be incorporated directly into formulations and have a level of safety higher than acetaminophen itself. The functional corollary of BAX translocation is an alteration to the permeability of the outer mitochondrial membrane (MPT or mitochondrial permeability transition) and this can be linked to a collapse of the inner mitochondrial membrane potential (Dfm collapse) (319,320). Both of these phenomena have been observed in response to acetaminophen either in vitro or in vivo [e.g., MPT in vitro (306,321–324) or in vivo (322,324) and Dfm in vitro (265,324)]. Acetaminophen cyto- and hepatotoxicity has also been shown to be variably inhibited by the cyclophilin D–binding compound, cyclosporin A (CsA) (321–325). Furthermore, most studies indicate that CsA only partially protects against damage (321,322,324)—if at all (306). Very high in vitro concentrations of acetaminophen (e.g., 10–20 mM) appear to be particularly resistant to CsA inhibition altogether (306). These observations can now be rationalized by an outstanding series of reports which confirm that mice deficient in cyclophilin D have both a CsA-sensitive and CsA-insensitive MPT (326–329). Furthermore, these reports collectively indicate that cyclophilin D-deficient mice are protected against necrotic (i.e., CsA-sensitive) forms of cell death but not against BAXmediated and mitochondrially-directed cell death. As a result, we can conclude that a severe insult from elevated concentrations or doses of acetaminophen will shift the mechanism of cell killing to a high calcium, BAX driven, and CsA-insensitive form of injury. Other “downstream” events in acetaminophen injury are not as well defined at present. Despite the involvement of the BCL-2 family of proteins, an expected contribution of caspases (a group of cysteine proteases traditionally considered a hallmark of cell death by apoptosis—at least from a biochemical perspective) is less clear (330). Although still contentious, it appears that the majority (271,331,332) but not all (305,306) reports indicate a failure to activate caspases during acetaminophen toxicity. At best only a weak response is observed. It would then appear that the bioenergetic collapse observed with acetaminophen toxicity contributes to intracellular ADP/dATP/ATP depletion which is required for caspase activation and that this likely goes unfulfilled (176,333,334). As a result—and for most circumstances—acetaminophen can be considered to initiate an apoptotic version of cell death which progresses in the longer term to a form of cellular decay that is morphologically more similar to necrosis. These admittedly subjective considerations are dealt with in greater detail later in the text as we discuss our observations of acetaminophen-induced damage from an ultrastructural viewpoint (and likely the only perspective of relevance) (335). The multifaceted nature of the intracellular events occurring in response to acetaminophen exposure is represented diagrammatically in Figure 6 and is not intended to be exhaustive. Previously, we examined the biological significance of some of the major protein players, including the contributions from metallothionein, glutathione peroxidase and synthetase, superoxide dismutase, Jun N-terminal kinase, and NF-E2-related factor 2 (NRF2) (283). In the intervening period, several significant proteomic and toxicogenomic studies have

370

Nelson and Bruschi Centrilobular hepatocyte

Alterations to Cellular homeostasis Subcellularion redistributions phosphatase inhibition extended JNK activation transient proteasome inhibition (in vivo)

via ER

Extended proteasome inhibition (in vivo) Rebound proteasome or other protease activation ∆ψ collapse (in vitro & in vivo) Chromatin margination and DNA fragmentation (in vitro & in vivo) Do novo protein synthesis - in vivo

ATF3; HSP70i; HSP25; NRF2 -in vitro ATF3; GADD753/CHOP U1 70 kDa snRNP

Ca2+ ATPase

Ca2+ waves

Mitochondrial proliferation

BAX NAPQI

BAX* (activated)

?

Ca2+

ER

?

Direct REDOX alterations

NAPQI

Bcl-XL

Mitochondria Bongkrekic acid

?

Golgi

Cytochrome c P65 ICAD Little or no caspase activation

Ca2+ Indirect REDOX alterations

?

? AIF DNase1 EndoG

Target proteins (for example) GAPDH GSH peroxidase proteasome C8 GST Pi ATP synthetase α Se-binding protein(s)

Nucleus CAD

Na+/K+ ATPase

CAD NF κB translocation

NitrosoGAPDH Translocation of modified and unmodified proteins to the nucleus

ICAD

FIGURE 6 Depiction of selected biochemical pathways and cellular processes that appear to be affected in the hepatocyte as a result of metabolic activation of acetaminophen to NAPQI and which may be involved in the pathogenesis of liver injury. The scheme does not include events and pathways extrinsic to the hepatocyte (see text for discussion). Abbreviation: NAPQI, N-acetyl-p-benzoquinone imine.

appeared which confirm the prominent positions these genes (and others) have in acetaminophen-induced cyto- and hepatotoxicity (270,275,277,296,302,336,337). Activation of c-Jun N-terminal kinase by acetaminophen in BAX translocation appears to be a particularly significant pathway in cytotoxicity (338) and hepatotoxicity caused by acetaminophen in mice (339). However, there are still many unanswered questions regarding the mechanism of acetaminophen-induced intrinsic cell death. The most prominent of these include: 1. What are the biological functions—and how are they perturbed—of the major protein targets of NAPQI, namely selenium-binding- and acetaminophen-binding proteins? Are mice deficient in either or both of these proteins resistant to the damage produced by acetaminophen? 2. Which is the major protein responsible for the nuclear morphologies observed during acetaminophen-mediated hepatocyte death? Is it DNase 1 (340), AIF (341), or Endo G (341,342)? 3. Which HSP70 isoforms are protective and which are not (296,301)? 4. Can the recent identification of senescence marker protein-30 (SMP30) as a protective gene product (275) be unequivocally connected to any of its known functions, e.g., in calcium ion regulation (343–345), signal transduction (346), or as a gluconolactase (347,348)? 5. What is the contribution of the recently identified S-nitrosylated product of GAPDH (a major arylation target of NAPQI) in the early apoptotic signals following overdose to acetaminophen (284,285)? Does NAPQI-mediated arylation of an active site cysteine inhibit or enhance early apoptotic signaling by GAPDH (183)? What is the exact relationship between initial S-nitrosylation events and protein nitrotyrosine formation after acetaminophen overdose (190,349–351)? 6. What is the function of poly(ADP-ribose) polymerase (PARP) in detecting the bioenergetic collapse following acetaminophen and in modulating the resultant severity of injury (352,353)?

Acetaminophen-Induced Liver Disease

371

7. Can early injury to hepatic oval cells (the hepatic progenitor cell type) be prevented or otherwise modulated so as to alleviate subsequent injury (354–356)? The challenge for the future will be to separate the critical intrinsic pathways from noncritical epiphenomena amid this considerable complexity and, in so doing, ultimately determine cell death susceptibility factors. Future studies, and especially those using toxicogenomic and proteomic strategies, must also relate any observed findings to the biology of acetaminophen-induced cyto- and hepatotoxicity. This is due to the “serious limitations” of such technologies with regard to biological inference, i.e., cause and effect can rarely be determined (272). Consequently, studies which exploit known factors in acetaminophen biology, such as those recently described (275,296), can prove powerful in their scope. For example, whereas several studies previously observed either no change or an absolute decline in SMP30 protein and gene expression levels, Welch and colleagues (275,296) have been able to show that sensitive mice (Black6 genetic background) are susceptible—at last in part—because of an inability to induce this putative protective gene in response to acetaminophen.

Cellular, Tissue, and Whole Body Factors in Acetaminophen Liver Injury Acetaminophen-induced toxicity can also be shown to have effects at a greater level of organizational complexity than that of individual genes and proteins (Fig. 5; lower group). Previously, we reported changes in the ultrastructural morphology of mitochondria in response to acetaminophen both in vivo and in vitro. As early as 6 hours hepatocytes respond to acetaminophen by a prominent increase in the number of mitochondria (i.e., mitochondrial fission) (271,332). Ultrastructural studies further indicate the close interrelationships between at times densely-packed mitochondria in the cytoplasm and double-membrane structures of apparent endoplasmic reticulum origin (265,271). At later time points, (e.g., 24 hours) atypical mitochondria are completely surrounded by multi-laminated double-membrane vacuoles and whorls (265), which are distinctive of a regulated process of degradation and recycling of cellular constituents known as macroautophagy (357–359). Over the last few years, anti-apoptotic BCL-2 members have been recognized as playing important regulatory roles in this process of macroautophagy. Autophagy is generally considered to recycle nutrients, maintain energy homeostasis, and promote survival in response to the bioenergetic collapse that results from external starvation. BCL-2 anti-apoptotic proteins are able to inhibit autophagy through a specific interaction with Beclin 1, the first identified mammalian autophagy gene product (360–362). The current thinking is that BCL-2 coordinates levels of cytoprotective autophagy and that if this process is inhibited cells will die preferentially by apoptosis unless a functional apoptotic machinery is not present (i.e., caspaseindependent and Dfm collapse). In the latter case cell death becomes necrotic (363–366). It is also interesting to note here that the failure to activate caspases has recently been linked to the induction of autophagy by the selective degradation of catalase—an acetaminophen target protein (192,288)—and with the absence of this protein an accumulation of intracellular ROS (367). Furthermore, protein turnover and degradation processes more generally have been linked to the activation of macroautophagy and membrane-remodeling phenomena (368,369). Our own ultrastructural studies with acetaminophen-mediated hepatotoxicity in a transgenic mouse model overexpressing BCL-2 protein suggest that the BCL-2 family of proteins can modulate these types of cell death processes (271) (Fig. 7). Overexpression of BCL-2 was found to promote the early formation of autophagic vacuole formation in comparison to wild-type animals treated with acetaminophen (Fig. 7). Further studies will be required to determine if these autophagic vacuole-like structures are bona fide structures of macroautophagy. In comparison, we have observed well-organized and advanced stacked cisternae of endoplasmic reticulum origin from 12 to 24 hours after acetaminophen treatment in wild-type animals (265). The literature has proposed that macroautophagy and apoptosis are mutually exclusive cell death mechanisms. For example, autophagy occurs only in apoptoticincompetent cells (358,362) or, alternatively, inhibition of autophagy promotes cell death by apoptosis (363). These considerations are likely to be applicable to acetaminophen-induced

372

Nelson and Bruschi

m nuc

control

m lip

m

m

rer

m lip

nuc

m bcl-2

m

m m m

rer m

m

(A)

(B) m

APAP 6 hr

nuc

m m

nuc

(C)

(D)

nuc

APAP + bcl-2 6 hr

nuc

FIGURE 7 Ultrastructural examinations of hepatic parenchymal cell morphologies from wild-type or bcl-2 (K/C) transgenic mice inducibly overexpressing BCL-2 protein under control of metallothionein promoter treated with either acetaminophen (500 mg/kg, ip.) or a saline vehicle control (271). Control: Typical hepatocyte ultrastructures were observed in saline vehicle-treated wld-type animals (control; left) with the presence of normal nuclear architecture (nuc), numerous mitochondria (m), primary (arrowhead) and secondary lysosomes (arrow), lipid droplets (lip) and electrondense glycogen granules (*). (Magnification 6000!). At higher magnification (control; right) there was an unremarkable association of typical mitochondria (m) with rough endoplasmic reticulum (rer) in the presence of occasional lysosomes (arrow). (Magnification 12,000!). bcl-2: Comparable morphologies were evident in bcl-2 (K/C) mice treated with zinc in the drinking water to induce expression of BCL-2 protein (271) (bcl-2). Representative fields are shown at 6,000! (left) and 12,000! (right) magnification and indicate loose mitochondrial and endoplasmic associations with an apparently increased presence of glycogen granules (*). APAP: Acetaminophen treatment of wild-type mice for 6 hours was observed to alter both the morphology and density of mitochondria (m) in the cytoplasm (APAP, left). Closely packed ring-shaped or elongated mitochondria (pointer hand) were more prominent and occasionally partially surrounded by double-membrane structures of endoplasmic reticulum (rer) origin (Magnification 6,000!). At higher magnification (APAP, right) periodic intramitochondrial crystalline inclusions were observed (pointer) with occasional double membranes in close by disorganized proximity to mitochondria (A). Mitochondria (m) with clear variations in matrix swelling were observed often in the same field (B) (Magnification 12,000!). APAPCbcl-2: Acetaminophen treatment of zinc-induced bcl-2 (K/C) transgenic mice (APAPCbcl-2, left) revealed many of the morphologies observed with acetaminophen above including ring-shaped or elongated mitochondria (pointer hand) but with closer associations with membranes and the additional presence of electron-dense lipid-like structures (not membrane bound). (Magnification 6000!). At higher magnification (APAPKbcl-2, right) atypical mitochondria are observed to surround cytosol or even lipid in unusual autophagic vacuole-like structures (pointer hand) (C). Alternatively multilamellar double-membrane whorls are observed to surround ringed mitochondria or cytoplasm (pointer hand) (D) in autophagic structures not observed in acetaminophen-treated wild-type mice at this time point. (Magnification 12,000!). Abbreviation: APAP, acetaminophen.

Acetaminophen-Induced Liver Disease

373

liver injury in animal models, but further work is necessary to establish relevance to a clinical setting. The Extrinsic Cell Death Pathway The loss of cellular integrity during the intrinsic phase of acetaminophen-induced liver injury releases intracellular contents (including proteins adducted by NAPQI) (370), which have proinflammatory capabilities and/or protein hydrolytic capability (371). Consequently, the extrinsic phase of acetaminophen-induced hepatotoxicity equates with the recruitment of immune surveillance. This undoubtedly adds to the complexity of the initial intracellular damage as numerous cytokine and chemokine mediators are liberated from nonparenchymal and nonhepatic cell types. The liver is an organ capable of mounting a major inflammatory response. Up to 35% of all liver cells are nonparenchymal (constituting the sinusoidal endothelium, Kupffer cells, and stellate cells). These cell types, in addition to infiltrating cells such as macrophages, are implicated in both the protective and the pathological (i.e., necrosis/ fibrosis) responses of immune activation (250,354,372,373). With such increased complexity, however, comes the probability that simple therapeutic interventions may be difficult to achieve. The therapeutic modulation of immune responses has been an effective clinical strategy for many years, but is not without inherent risks (374,375). Although many cytokines/chemokines have been proposed to participate in the responses to acetaminophen exposure, the proinflammatory mediators have come under the most scrutiny. The role of inflammation in acetaminophen-mediated liver injury is complex with many proinflammatory mediators being implicated (e.g., TNF, Fas, interleukin 6, 8, and 11, leukemia inhibitory factor, oncostatin M, and macrophage migration inhibitory factor, MIP-2, MCP-1). With the influx of activated cells such as macrophages, neutrophils, and monocytes comes the involvement of nitric oxide and RNS, superoxide anions, and other ROS. The release of proteases from degranulation events further adds to the complexity. Nonetheless, good evidence points to the potential for increased injury after exposure as a result of such proinflammatory processes (376,377). As a consequence, the possibility exists for the prevention of injury with treatments designed to inhibit extrinsic injury. To this end, some experimental models have shown good success in preventing mortality after acetaminophen treatment by transgenically overexpressing redox-active enzymes (289). Many other studies have examined the feasibility of preventing or reducing acetaminophen-induced injury by modulating immune responses during the extrinsic phase. These have included studies with TNF (372,378–383), various chemokines and cytokines (372,377,384–386), nitric oxide and peroxynitrite and related enzymes (190,263,323,325,387–390), antisense oligonucleotides to Fas/CD95 (391), the archetypal inflammatory transcription factor NF-kB (372,376,392), and other extrinsic factors (for a review see Ref. 263). Many of these reports point to the possibility of therapeutic intervention during the extrinsic phase of acetaminopheninduced hepatotoxicity. Considerably more work will be required, however, as many reports are at times contradictory and our understanding of the exact contributions and interrelationships of the numerous players in the extrinsic response is still in its infancy. Nonetheless, research in the area of extrinsic factors of acetaminophen-induced hepatotoxicity has progressed significantly over the last few years. There is even the prospect of a clinical alternative to acetaminophen which takes into account some aspects of the biology of this phase of injury. Phase III evaluations have recently been reported for a novel, nitric oxidereleasing, nitroxybutyl ester derivative of acetaminophen (NCX-701 or “nitroparacetamol”) which is also a potent analgesic (325,393–395). Ultimately, the aim will be to either provide safer and more effective alternatives to acetaminophen itself or to the treatment strategies for overdose situations currently available (e.g., N-acetylcysteine) (396,397). CONCLUSIONS Recent work has raised the possibilities of novel and more effective strategies to treat accidental or intentional acetaminophen-induced hepatotoxicity. These are based on new and powerful techniques, which promise to delineate the critical steps necessary for expression of

374

Nelson and Bruschi

acetaminophen liver injury at either the cellular or the extracellular level. Toxicogenomic, transgenic and proteomic studies which are based on the extensive literature dealing with the biology of acetaminophen liver injury will yield the most valuable information. Finally, determining the significant cell biological aspects of acetaminophen-induced hepatotoxicity remains a formidable problem. How the hepatocyte responds to the initiation of damage via protein adduction will undoubtedly reveal fundamental aspects of hepatocyte biology and its response to external stress. In addition, the elucidation of functions for the protein targets of NAPQI will provide information with respect to basic processes in cell biology (e.g., protein processing, transport, and compartmentation). Acetaminophen-induced hepatotoxicity, now considered an excellent model for chemically induced cell death, is still an enigma and remains a challenge despite over 20 years of intensive study. REFERENCES 1. Prescott LF. Paracetamol (Acetaminophen): A Critical Bibliographic Review. London: Taylor and Francis, 1996. 2. Bessems JG, Vermeulen NP. Paracetamol (acetaminophen)-induced toxicity: molecular and biochemical mechanisms, analogues and protective approaches. Crit Rev Toxicol 2001; 31:55–138. 3. Blakely P, McDonald BR. Acute renal failure due to acetaminophen ingestion: a case report and review of the literature. J Am Soc Nephrol 1995; 6:48–53. 4. De Broe ME, Elseviers MM. Analgesic nephropathy. N Engl J Med 1998; 338:446–52. 5. Feinstein AR, Heinemann LAJ, Curham GC, et al. Relationship between nonphenacetin combined analgesics and nephropathy: a review. Kidney Int 2000; 58:2259–64. 6. Bromer MQ, Black M. Acetaminophen hepatotoxicity. Clin Liver Dis 2003; 7:351–67. 7. Hinson JA. Biochemical toxicology of acetaminophen. Rev Biochem Toxicol 1980; 2:103–30. 8. Josephy DP. The molecular toxicology of acetaminophen. Drug Metab Rev 2005; 37:581–94. 9. Rumack BH. Acetaminophen hepatotoxicity: the first 35 years. J Toxicol Clin Toxicol 2002; 40:3–20. 10. Atcha Z, Majeed A. Paracetamol related deaths in England and Wales. Health Stat Q 2000; 7:5–9. 11. Gow PJ, Jones RM, Dobson JL, Angus PW. Etiology and outcome of fulminant hepatic failure managed at an Australian liver transplant unit. J Gastroenterol Hepatol 2004; 19:154–9. 12. Larson AM, Polson J, Fontana RJ, et al. Acetaminophen-induced acute liver failure: results of a United States multicenter, prospective study. Hepatology 2005; 42:1364–72. 13. Sheen CL, Dillon JF, Bateman DN, Simpson KJ, MacDonald TM. Paracetamol-related deaths in Scotland, 1994–2000. Br J Clin Pharmacol 2002; 54:430–2. 14. Litovitz TL, Klein-Schwartz W, Dyer KS, Shannon M, Lee S, Powers M. 1997 annual report of the American Association of Poison Control Centers Toxic Exposure Surveillance System. Am J Emerg Med 1998; 16:443–97. 15. Vale JA, Proudfoot AT. Paracetamol (acetaminophen) poisoning. Lancet 1995; 346:547–52. 16. Nourjah P, Ahmad SR, Karwoski C, Willy M. Estimates of acetaminophen (paracetomal)-associated overdoses in the United States. Pharmacoepidemiol Drug Saf 2006; 15:398–405. 17. Kaplowitz N. Acetaminophen hepatoxicity: what do we know, what don’t we know, and what do we do next? Hepatology 2004; 40:23–6. 18. Rumack BH. Acetaminophen misconceptions. Hepatology 2004; 40:10–15. 19. Watkins PB, Kaplowitz N, Slattery JT, Colonese CR, Colucci SV, Stewart PW, Harris SC. Aminotransferase elevations in healthy adults receiving 4 grams of acetaminophen daily: a randomized controlled trial. JAMA 2006; 296:87–93. 20. Hawton K, Townsend E, Deeks J, Appleby L, Gunnell D, Bennewith O, Cooper J. Effects of legislation restricting pack sizes of paracetamol and salicylate on self poisoning in the United Kingdom: before and after study. Br Med J 2001; 322:1203–7. 21. Boyd EM, Bereczky GM. Liver necrosis from paracetamol. Br J Pharmacol Chemother 1966; 26:606–14. 22. Eder H. Chronic toxicity studies on phenacetin. N-acetyl-p-aminophenol (NAPA) and acetylsalicylic acid on cats. Acta Pharmacol Toxicol (Copenh) 1964; 21:197–204. 23. Davidson DG, Eastham WN. Acute liver necrosis following overdose of paracetamol. Br Med J 1966; 5512:497–9. 24. Thomson JS, Prescott LF. Liver damage and impaired glucose tolerance after paracetamol overdosage. Br Med J 1966; 5512:506–7. 25. Jollow DJ, Mitchell JR, Potter WZ, Davis DC, Gillette JR, Brodie BB. Acetaminophen-induced hepatic necrosis. II. Role of covalent binding in vivo. J Pharmacol Exp Ther 1973; 187:195–202. 26. Mitchell JR, Jollow DJ, Potter WZ, Davis DC, Gillette JR, Brodie BB. Acetaminophen-induced hepatic necrosis. I. Role of drug metabolism. J Pharmacol Exp Ther 1973; 187:185–94.

Acetaminophen-Induced Liver Disease

375

27. Potter WZ, Davis DC, Mitchell JR, Jollow DJ, Gillette JR, Brodie BB. Acetaminophen-induced hepatic necrosis. III. Cytochrome P-450-mediated covalent binding in vitro. J Pharmacol Exp Ther 1973; 187:203–10. 28. Mitchell JR, Jollow DJ, Potter WZ, Gillette JR, Brodie BB. Acetaminophen-induced hepatic necrosis. IV. Protective role of glutathione. J Pharmacol Exp Ther 1973; 187:211–7. 29. Mitchell JR, Thorgeirsson SS, Potter WZ, Jollow DJ, Keiser H. Acetaminophen-induced hepatic injury: protective role of glutathione in man and rationale for therapy. Clin Pharmacol Ther 1974; 16:676–84. 30. Polson J, Lee WM. AASLD position paper: the management of acute liver failure. Hepatology 2005; 41:1179–97. 31. Routledge P, Vale JA, Bateman DN, et al. Paracetamol (acetaminophen) poisoning. No need to change current guidelines to accident departments. Br Med J 1998; 317:1609–10. 32. Smilkstein MJ, Knapp GL, Kulig KW, Rumack BH. Efficacy of oral N-acetylcysteine in the treatment of acetaminophen overdose. Analysis of the national multicenter study (1976 to 1985). N Engl J Med 1988; 319:1557–62. 33. Nelson SD. Analgesic-antipyretics. In: Levy RH et al, ed. Metabolic Drug Interactions. Philadelphia, PA: Lippincott Williams and Williams, 2000:447–55. 34. Nelson SD. Molecular mechanisms of the hepatotoxicity caused by acetaminophen. Semin Liver Dis 1990; 10:267–78. 35. Johnson KA, Plumb R. Investigating the human metabolism of acetaminophen using UPLC and exact mass oa-TOF MS. J Pharm Biomed Anal 2005; 39:805–10. 36. Court MH, Duan SX, von Moltke LL, Greenblatt DJ, Patten CJ, Miners JO, Mackenzie PI. Interindividual variability in acetaminophen glucuronidation by human liver microsomes: identification of relevant acetaminophen UDP-glucuronosyltransferase isoforms. J Pharmacol Exp Ther 2001; 299:998–1006. 37. Kostrubsky SE, Sinclair JF, Strom SC, et al. Phenobarbital and phenytoin increased acetaminophen hepatotoxicity due to inhibition of UDP-glucuronosyltransferases in cultured human hepatocytes. Toxicol Sci 2005; 87:146–55. 38. Mutlib AE, Goosen TC, Bauman JN, Williams JA, Kulkarni S, Kostrubsky SE. Kinetics of acetaminophen glucuronidation by UDP-glucuronosyltransferases 1A1, 1A6, 1A9 and 2B15. Potential implications in acetaminophen induced toxicity. Chem Res Toxicol 2006; 19:701–9. 39. Duffel MW. Sulfotransferases. In: Guengerich FP, ed. Comprehensive Toxicology, Vol. 3. Oxford: Elsevier, 1997:365–83. 40. de Morais SM, Chow SY, Wells PG. Biotransformation and toxicity of acetaminophen in congenic RHA rats with or without a hereditary deficiency in bilirubin UDP-glucuronosyltransferase. Toxicol Appl Pharmacol 1992; 117:81–7. 41. Court MH, Greenblatt DJ. Molecular basis for deficient acetaminophen glucuronidation in cats. An interspecies comparison of enzyme kinetics in liver microsomes. Biochem Pharmacol 1997; 53:1041–7. 42. de Morais SM, Uetrecht JP, Wells PG. Decreased glucuronidation and increased bioactivation of acetaminophen in Gilbert’s syndrome. Gastroenterology 1992; 102:577–86. 43. Rodriguez E, Regadera P. The heterogeneity of paracetamol metabolism in Gilbert’s disease. Gastroenterol Hepatol 1999; 22:265–9. 44. Richman DD, Fischl MA, Grieco MH, et al. The toxicity of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex. A double-blind, placebo-controlled trial. N Engl J Med 1987; 317:192–7. 45. Bray GP, Harrison PM, O’Grady JG, Tredger JM, Williams R. Long-term anticonvulsant therapy worsens outcome in paracetamol-induced fulminant hepatic failure. Hum Exp Toxicol 1992; 11:265–70. 46. Lystback BB, Norregaard P. A case of paracetamol retard poisoning with fatal outcome. Ugeskr Laeger 1995; 157:899–900. 47. Minton NA, Henry JA, Frankel RJ. Fatal paracetamol poisoning in an epileptic. Hum Toxicol 1988; 7:33–4. 48. Pirotte JH. Apparent potentiation of hepatotoxicity from small doses of acetaminophen by phenobarbital. Ann Intern Med 1984; 101:403. 49. Nicholls AW, Caddick S, Wilson ID, Farrant RD, Lindon JC, Nicholson JK. High resolution NMR spectroscopic studies on the metabolism and futile deacetylation of 4-hydroxyacetanilide (paracetamol) in the rat. Biochem Pharmacol 1995; 49:1155–64. 50. Mugford CA, Tarloff JB. The contribution of oxidation and deacetylation to acetaminophen nephrotoxicity in female Sprague–Dawley rats. Toxicol Lett 1997; 93:15–22. 51. Newton JF, Kuo CH, DeShone GM, Hoefle D, Bernstein J, Hook JB. The role of p-aminophenol in acetaminophen-induced nephrotoxicity: effect of bis(p-nitrophenyl) phosphate on acetaminophen and p-aminophenol nephrotoxicity and metabolism in Fischer 344 rats. Toxicol Appl Pharmacol 1985; 81:416–30.

376

Nelson and Bruschi

52. Rothen JP, Haefeli WE, Meyer UA, Todesco L, Wenk M. Acetaminophen is an inhibitor of hepatic N-acetyltransferase 2 in vitro and in vivo. Pharmacogenetics 1998; 8:553–9. 53. Dahlin DC, Miwa GT, Lu AY, Nelson SD. N-acetyl-p-benzoquinone imine: a cytochrome P-450mediated oxidation product of acetaminophen. Proc Natl Acad Sci USA 1984; 81:1327–31. 54. Hinson JA, Pohl LR, Gillette JR. N-Hydroxyacetaminophen: a microsomal metabolite of N-hydroxyphenacetin but apparently not of acetaminophen. Life Sci 1979; 24:2133–8. 55. Nelson SD, Forte AJ, Dahlin DC. Lack of evidence for N-hydroxyacetaminophen as a reactive metabolite of acetaminophen in vitro. Biochem Pharmacol 1980; 29:1617–20. 56. Blair IA, Boobis AR, Davies DJ, Cresp TM. Paracetamol oxidation: synthesis and reactivity of N-acetyl-p-benzoquinone imine. Tetrahedron Lett 1980; 21:4947–50. 57. Dahlin DC, Nelson SD. Synthesis, decomposition kinetics, and preliminary toxicological studies of pure N-acetyl-p-benzoquinone imine, a proposed toxic metabolite of acetaminophen. J Med Chem 1982; 25:885–6. 58. Albano E, Rundgren M, Harvison PJ, Nelson SD, Moldeus P. Mechanisms of N-acetyl-p-benzoquinone imine cytotoxicity. Mol Pharmacol 1985; 28:306–11. 59. Corcoran GB, Mitchell JR, Vaishnav YN, Horning EC. Evidence that acetaminophen and N-hydroxyacetaminophen form a common arylating intermediate, N-acetyl-p-benzoquinoneimine. Mol Pharmacol 1980; 18:536–42. 60. Holme JA, Dahlin DC, Nelson SD, Dybing E. Cytotoxic effects of N-acetyl-p-benzoquinone imine, a common arylating intermediate of paracetamol and N-hydroxyparacetamol. Biochem Pharmacol 1984; 33:401–6. 61. Miner DJ, Kissinger PT. Evidence for the involvement of N-acetyl-p-quinoneimine in acetaminophen metabolism. Biochem Pharmacol 1979; 28:3285–90. 62. Hinson JA, Monks TJ, Hong M, Highet RJ, Pohl LR. 3-(glutathion-S-yl)acetaminophen: a biliary metabolite of acetaminophen. Drug Metab Dispos 1982; 10:47–50. 63. Nelson SD, Vaishnav Y, Kambara H, Baillie TA. Comparative electron impact, chemical ionization and field desorption mass spectra of some thioether metabolites of acetaminophen. Biomed Mass Spectrom 1981; 8:244–51. 64. Harvison PJ, Guengerich FP, Rashed MS, Nelson SD. Cytochrome P-450 isozyme selectivity in the oxidation of acetaminophen. Chem Res Toxicol 1988; 1:47–52. 65. Slattery JT, Wilson JM, Kalhorn TF, Nelson SD. Dose-dependent pharmacokinetics of acetaminophen: evidence of glutathione depletion in humans. Clin Pharmacol Ther 1987; 41:413–8. 66. Powis G, See KL, Santone KS, Melder DC, Hodnett EM. Quinoneimines as substrates for quinone reductase (NAD(P)H: (quinone-acceptor)oxidoreductase) and the effect of dicumarol on their cytotoxicity. Biochem Pharmacol 1987; 36:2473–9. 67. Powis G, Svingen BA, Dahlin DC, Nelson SD. Enzymatic and non-enzymatic reduction of N-acetylp-benzoquinone imine and some properties of the N-acetyl-p-benzosemiquinone imine radical. Biochem Pharmacol 1984; 33:2367–70. 68. Chen W, Shockcor JP, Tonge R, Hunter A, Gartner C, Nelson SD. Protein and nonprotein cysteinyl thiol modification by N-acetyl-p-benzoquinone imine via a novel ipso adduct. Biochemistry 1999; 38:8159–66. 69. Coles B, Wilson I, Wardman P, Hinson JA, Nelson SD, Ketterer B. The spontaneous and enzymatic reaction of N-acetyl-p-benzoquinonimine with glutathione: a stopped-flow kinetic study. Arch Biochem Biophys 1988; 264:253–60. 70. Hinson JA, Pohl LR, Monks TJ, Gillette JR, Guengerich FP. 3-Hydroxyacetaminophen: a microsomal metabolite of acetaminophen. Evidence against an epoxide as the reactive metabolite of acetaminophen. Drug Metab Dispos 1980; 8:289–94. 71. Forte AJ, Wilson JM, Slattery JT, Nelson SD. The formation and toxicity of catechol metabolites of acetaminophen in mice. Drug Metab Dispos 1984; 12:484–91. 72. Chen W. Mechanistic studies on the formation of oxidative metabolites of acetaminophen. Medicinal Chemistry. Seattle: University of Washington, 1998. 73. Chen W, Koenigs LL, Thompson SJ, Peter RM, Rettie AE, Trager WF, Nelson SD. Oxidation of acetaminophen to its toxic quinone imine and nontoxic catechol metabolites by baculovirusexpressed and purified human cytochromes P450 2E1 and 2A6. Chem Res Toxicol 1998; 11:295–301. 74. Dong H, Haining RL, Thummel KE, Rettie AE, Nelson SD. Involvement of human cytochrome P450 2D6 in the bioactivation of acetaminophen. Drug Metab Dispos 2000; 28:1397–400. 75. Regal KM. Caffeine as an active site probe of cytochrome P4501A2. School of Pharmacy. Seattle: University of Washington, 1998. 76. Thummel KE, Lee CA, Kunze KL, Nelson SD, Slattery JT. Oxidation of acetaminophen to N-acetyl-paminobenzoquinone imine by human CYP3A4. Biochem Pharmacol 1993; 45:1563–9. 77. Patten CJ, Thomas PE, Guy RL, Lee M, Gonzalez FJ, Guengerich FP, Yang CS. Cytochrome P450 enzymes involved in acetaminophen activation by rat and human liver microsomes and their kinetics. Chem Res Toxicol 1993; 6:511–8.

Acetaminophen-Induced Liver Disease

377

78. Raucy JL, Lasker JM, Lieber CS, Black M. Acetaminophen activation by human liver cytochromes P450IIE1 and P450IA2. Arch Biochem Biophys 1989; 271:270–83. 79. Anderson KE, Schneider J, Pantuck EJ, et al. Acetaminophen metabolism in subjects fed charcoalbroiled beef. Clin Pharmacol Ther 1983; 34:369–74. 80. Bock KW, Wiltfang J, Blume R, Ullrich D, Bircher J. Paracetamol as a test drug to determine glucuronide formation in man. Effects of inducers and of smoking. Eur J Clin Pharmacol 1987; 31:677–83. 81. Miners JO, Attwood J, Birkett DJ. Determinants of acetaminophen metabolism: effect of inducers and inhibitors of drug metabolism on acetaminophen’s metabolic pathways. Clin Pharmacol Ther 1984; 35:480–6. 82. Sarich T, Kalhorn T, Magee S, et al. The effect of omeprazole pretreatment on acetaminophen metabolism in rapid and slow metabolizers of S-mephenytoin. Clin Pharmacol Ther 1997; 62:21–8. 83. Tonge RP, Kelly EJ, Bruschi SA, Kalhorn T, Eaton DL, Nebert DW, Nelson SD. Role of CYP1A2 in the hepatotoxicity of acetaminophen: investigations using Cyp1a2 null mice. Toxicol Appl Pharmacol 1998; 153:102–8. 84. Lee SS, Buters JT, Pineau T, Fernandez-Salguero P, Gonzalez FJ. Role of CYP2E1 in the hepatotoxicity of acetaminophen. J Biol Chem 1996; 271:12063–7. 85. Zaher H, Buters JT, Ward JM, et al. Protection against acetaminophen toxicity in CYP1A2 and CYP2E1 double-null mice. Toxicol Appl Pharmacol 1998; 152:193–9. 86. Amouyal G, Larrey D, Letteron P, Geneve J, Labbe G, Belghiti J, Pessayre D. Effects of methoxsalen on the metabolism of acetaminophen in humans. Biochem Pharmacol 1987; 36:2349–52. 87. Letteron P, Descatoire V, Larrey D, DeGott C, Tinel M, Geneve J, Pessayre D. Pre- or post-treatment with methoxsalen prevents the hepatotoxicity of acetaminophen in mice. J Pharmacol Exp Ther 1986; 239:559–67. 88. Johansson I, Lundqvist E, Bertilsson L, Dahl ML, Sjoqvist F, Ingelman-Sundberg M. Inherited amplification of an active gene in the cytochrome P450 CYP2D locus as a cause of ultrarapid metabolism of debrisoquine. Proc Natl Acad Sci USA 1993; 90:11825–9. 89. Aklillu E, Persson I, Bertilsson L, Johansson I, Rodrigues F, Ingelman-Sundberg M. Frequent distribution of ultrarapid metabolizers of debrisoquine in an ethiopian population carrying duplicated and multiduplicated functional CYP2D6 alleles. J Pharmacol Exp Ther 1996; 278:441–6. 90. Shimada T, Yamazaki H, Mimura M, Inui Y, Guengerich FP. Interindividual variations in human liver cytochrome P-450 enzymes involved in the oxidation of drugs, carcinogens and toxic chemicals: studies with liver microsomes of 30 Japanese and 30 Caucasians. J Pharmacol Exp Ther 1994; 270:414–23. 91. Manyike PT, Kharasch ED, Kalhorn TF, Slattery JT. Contribution of CYP2E1 and CYP3A to acetaminophen reactive metabolite formation. Clin Pharmacol Ther 2000; 67:275–82. 92. Prescott LF, Howie D, Darrien I, Adriaenssens P. Paracetamol hepatotoxicity in man. In: Bundgaard H, Juul P, Kofod H, eds. Drug Design and Adverse Reactions. Cophenhagen: Munksgaard, 1977:99–108. 93. Rumack BH. Acetaminophen overdose in young children. Treatment and effects of alcohol and other additional ingestants in 417 cases. Am J Dis Child 1984; 138:428–33. 94. Miller RP, Roberts RJ, Fischer LJ. Acetaminophen elimination kinetics in neonates, children, and adults. Clin Pharmacol Ther 1976; 19:284–94. 95. Lauterburg BH, Vaishnav Y, Stillwell WG, Mitchell JR. The effects of age and glutathione depletion on hepatic glutathione turnover in vivo determined by acetaminophen probe analysis. J Pharmacol Exp Ther 1980; 213:54–8. 96. Lesko SM, Mitchell AA. The safety of acetaminophen and ibuprofen among children younger than two years old. Pediatrics 1999; 104:e39. 97. Alonso EM, Sokol RJ, Hart J, Tyson RW, Narkewicz MR, Whitington PF. Fulminant hepatitis associated with centrilobular hepatic necrosis in young children. J Pediatr 1995; 127:888–94. 98. Caravati EM. Unintentional acetaminophen ingestion in children and the potential for hepatotoxicity. J Toxicol Clin Toxicol 2000; 38:291–6. 99. Heubi JE, Barbacci MB, Zimmerman HJ. Therapeutic misadventures with acetaminophen: hepatoxicity after multiple doses in children. J Pediatr 1998; 132:22–7. 100. Hynson JL, South M. Childhood hepatotoxicity with paracetamol doses less than 150 mg/kg per day. Med J Aust 1999; 171:497. 101. Miles FK, Kamath R, Dorney SF, Gaskin KJ, O’Loughlin EV. Accidental paracetamol overdosing and fulminant hepatic failure in children. Med J Aust 1999; 171:472–5. 102. Rivera-Penera T, Gugig R, Davis J, et al. Outcome of acetaminophen overdose in pediatric patients and factors contributing to hepatotoxicity. J Pediatr 1997; 130:300–4. 103. Daly FF, Dart RC, Prescott LF. Accidental paracetamol overdosing and fulminant hepatic failure in children. Med J Aust 2000; 173:558–60. 104. Parikh S, Dillon LC, Scharf SL. Hepatotoxicity possibly due to paracetamol with carbamazepine. Intern Med J 2004; 34:441–2.

378

Nelson and Bruschi

105. Wilson JT, Kasantikul V, Harbison R, Martin D. Death in an adolescent following an overdose of acetaminophen and phenobarbital. Am J Dis Child 1978; 132:466–73. 106. Wright N, Prescott LF. Potentiation by previous drug therapy of hepatotoxicity following paracetamol overdosage. Scott Med J 1973; 18:56–8. 107. Young CR, Mazure CM. Fulminant hepatic failure from acetaminophen in an anorexic patient treated with carbamazepine. J Clin Psychiatry 1998; 59:622. 108. Prescott LF, Critchley JA, Balali-Mood M, Pentland B. Effects of microsomal enzyme induction on paracetamol metabolism in man. Br J Clin Pharmacol 1981; 12:149–53. 109. Dai G, Chou N, He L, et al. Retinoid X receptor alpha regulates the expression of glutathione S-transferase genes and modulates acetaminophen-glutathione conjugation in mouse liver. Mol Pharmacol 2005; 68:1590–6. 110. Maglich JM, Stoltz CM, Goodwin B, Hawkins-Brown D, Moore JT, Kliewer SA. Nuclear pregnane X receptor and constitutive androstane receptor regulate overlapping but distinct sets of genes involved in xenobiotic detoxification. Mol Pharmacol 2002; 62:638–46. 111. Wu Y, Zhang X, Bardag-Gorce F, et al. Retinoid X receptor alpha regulates glutathione homeostasis and xenobiotic detoxification processes in mouse liver. Mol Pharmacol 2004; 65:550–7. 112. Zhang J, Huang W, Chua SS, Wei P, Moore DD. Modulation of acetaminophen-induced hepatotoxicity by the xenobiotic receptor CAR. Science 2002; 298:422–4. 113. Guo GL, Moffit JS, Nicol CJ, et al. Enhanced acetaminophen toxicity by activation of the pregnane X receptor. Toxicol Sci 2004; 82:374–80. 114. Zhang J, Huang W, Qatanani M, Evans RM, Moore DD. The constitutive androstane receptor and pregnane X receptor function coordinately to prevent bile acid-induced hepatotoxicity. J Biol Chem 2004; 279:49517–22. 115. Wolf KK, Wood SG, Hunt JA, Walton-Strong BW, Yasuda K, Lan L, Duan SX, Hao Q, Wrighton SA, Jeffery EH, et al. Role of the nuclear receptor pregnane X receptor in acetaminophen hepatotoxicity. Drug Metab Dispos 2005; 33:1827–36. 116. Nelson SD, Slattery JT, Thummel KE, Watkins PB. CAR unlikely to significantly modulate acetaminophen hepatotoxicity in most humans. Hepatology 2003; 38:254–7. 117. Xiong H, Turner KC, Ward ES, Jansen PL, Brouwer KL. Altered hepatobiliary disposition of acetaminophen glucuronide in isolated perfused livers from multidrug resistance-associated protein 2-deficient TR(K) rats. J Pharmacol Exp Ther 2000; 295:512–8. 118. Zamek-Gliszczynski MJ, Hoffmaster KA, Tian X, et al. Multiple mechanisms are involved in the biliary excretion of acetaminophen sulfate in the rat: role of Mrp2 and Bcrp1. Drug Metab Dispos 2005; 33:1158–65. 119. Manautou JE, de Waart DR, Kunne C, Zelcer N, Goedken M, Borst P, Elferink RO. Altered disposition of acetaminophen in mice with a disruption of the Mrp3 gene. Hepatology 2005; 42:1091–8. 120. Ghanem CI, Ruiz ML, Villanueva SS, et al. Shift from biliary to urinary elimination of acetaminophen-glucuronide in acetaminophen-pretreated rats. J Pharmacol Exp Ther 2005; 315:987–95. 121. Robinson AE, Sattar H, McDowall RD, Holder AT, Powell R. Forensic toxicology of some deaths associated with the combined use of propoxyphene and acetaminophen (paracetamol). J Forensic Sci 1977; 22:708–17. 122. Peterson GR, Hostetler RM, Lehman T, Covault HP. Acute inhibition of oxidative drug metabolism by propoxyphene (Darvon). Biochem Pharmacol 1979; 28:1783–9. 123. Pond SM, Tong TG, Kaysen GA, Menke DJ, Galinsky RE, Roberts SM, Levy G. Massive intoxication with acetaminophen and propoxyphene: unexpected survival and unusual pharmacokinetics of acetaminophen. J Toxicol Clin Toxicol 1982; 19:1–16. 124. Chien JY, Peter RM, Nolan CM, et al. Influence of polymorphic N-acetyltransferase phenotype on the inhibition and induction of acetaminophen bioactivation with long-term isoniazid. Clin Pharmacol Ther 1997; 61:24–34. 125. Zand R, Nelson SD, Slattery JT, Thummel KE, Kalhorn TF, Adams SP, Wright JM. Inhibition and induction of cytochrome P4502E1-catalyzed oxidation by isoniazid in humans. Clin Pharmacol Ther 1993; 54:142–9. 126. Slattery JT, Nelson SD, Thummel KE. The complex interaction between ethanol and acetaminophen. Clin Pharmacol Ther 1996; 60:241–6. 127. Thummel KE, Slattery JT, Ro H, Chien JY, Nelson SD, Lown KE, Watkins PB. Ethanol and production of the hepatotoxic metabolite of acetaminophen in healthy adults. Clin Pharmacol Ther 2000; 67:591–9. 128. Chien JY, Thummel KE, Slattery JT. Pharmacokinetic consequences of induction of CYP2E1 by ligand stabilization. Drug Metab Dispos 1997; 25:1165–75. 129. Epstein MM, Nelson SD, Slattery JT, Kalhorn TF, Wall RA, Wright JM. Inhibition of the metabolism of paracetamol by isoniazid. Br J Clin Pharmacol 1991; 31:139–42. 130. Moulding TS, Redeker AG, Kanel GC. Acetaminophen, isoniazid, and hepatic toxicity. Ann Intern Med 1991; 114:431.

Acetaminophen-Induced Liver Disease

379

131. Murphy R, Swartz R, Watkins PB. Severe acetaminophen toxicity in a patient receiving isoniazid. Ann Intern Med 1990; 113:799–800. 132. Nolan CM, Sandblom RE, Thummel KE, Slattery JT, Nelson SD. Hepatotoxicity associated with acetaminophen usage in patients receiving multiple drug therapy for tuberculosis. Chest 1994; 105:408–11. 133. McClain CJ, Kromhout JP, Peterson FJ, Holtzman JL. Potentiation of acetaminophen hepatotoxicity by alcohol. JAMA 1980; 244:251–3. 134. Seeff LB, Cuccherini BA, Zimmerman HJ, Adler E, Benjamin SB. Acetaminophen hepatotoxicity in alcoholics. A therapeutic misadventure. Ann Intern Med 1986; 104:399–404. 135. Zimmerman HJ, Maddrey WC. Acetaminophen (paracetamol) hepatotoxicity with regular intake of alcohol: analysis of instances of therapeutic misadventure. Hepatology 1995; 22:767–73. 136. Makin A, Williams R. Paracetamol hepatotoxicity and alcohol consumption in deliberate and accidental overdose. QJM 2000; 93:341–9. 137. Prescott LF. Paracetamol, alcohol and the liver. Br J Clin Pharmacol 2000; 49:291–301. 138. Kuffner EK, Dart RC, Bogdan GM, Hill RE, Casper E, Darton L. Effect of maximal daily doses of acetaminophen on the liver of alcoholic patients: a randomized, double-blind, placebo-controlled trial. Arch Intern Med 2001; 161:2247–52. 139. Takahashi T, Lasker JM, Rosman AS, Lieber CS. Induction of cytochrome P-4502E1 in the human liver by ethanol is caused by a corresponding increase in encoding messenger RNA. Hepatology 1993; 17:236–45. 140. Garcia-Ruiz C, Morales A, Ballesta A, Rodes J, Kaplowitz N, Fernandez-Checa JC. Effect of chronic ethanol feeding on glutathione and functional integrity of mitochondria in periportal and perivenous rat hepatocytes. J Clin Invest 1994; 94:193–201. 141. Hirano T, Kaplowitz N, Tsukamoto H, Kamimura S, Fernandez-Checa JC. Hepatic mitochondrial glutathione depletion and progression of experimental alcoholic liver disease in rats. Hepatology 1992; 16:1423–7. 142. Zhao P, Kalhorn TF, Slattery JT. Selective mitochondrial glutathione depletion by ethanol enhances acetaminophen toxicity in rat liver. Hepatology 2002; 36:326–35. 143. Hu Y, Ingelman-Sundberg M, Lindros KO. Induction mechanisms of cytochrome P450 2E1 in liver: interplay between ethanol treatment and starvation. Biochem Pharmacol 1995; 50:155–61. 144. Lauterburg BH, Velez ME. Glutathione deficiency in alcoholics: risk factor for paracetamol hepatotoxicity. Gut 1988; 29:1153–7. 145. Lee TD, Sadda MR, Mendler MH, Bottiglieri T, Kanel G, Mato JM, Lu SC. Abnormal hepatic methionine and glutathione metabolism in patients with alcoholic hepatitis. Alcohol Clin Exp Res 2004; 28:173–81. 146. Whitcomb DC, Block GD. Association of acetaminophen hepatotoxicity with fasting and ethanol use. JAMA 1994; 272:1840–5. 147. Zhao P, Slattery JT. Effects of ethanol dose and ethanol withdrawal on rat liver mitochondrial glutathione: implication of potentiated acetaminophen toxicity in alcoholics. Drug Metab Dispos 2002; 30:1413–7. 148. Sinclair JF, Szakacs JG, Wood SG, et al. Acetaminophen hepatotoxicity precipitated by short-term treatment of rats with ethanol and isopentanol: protection by triacetyloleandomycin. Biochem Pharmacol 2000; 59:445–54. 149. Sinclair JF, Szakacs JG, Wood SG, et al. Short-term treatment with alcohols causes hepatic steatosis and enhances acetaminophen hepatotoxicity in Cyp2e1(K/K) mice. Toxicol Appl Pharmacol 2000; 168:114–22. 150. Andrews FJ. Retroperitoneal haematoma after paracetamol increased anticoagulation. Emerg Med J 2002; 19:84–5. 151. Antlitz AM, Mead JA, Jr., Tolentino MA. Potentiation of oral anticoagulant therapy by acetaminophen. Curr Ther Res Clin Exp 1968; 10:501–7. 152. Bartle WR, Blakely JA. Potentiation of warfarin anticoagulation by acetaminophen. JAMA 1991; 265:1260. 153. Boeijinga JJ, Boerstra EE, Ris P, Breimer DD, Jeletich-Bastiaanse A. Interaction between paracetamol and coumarin anticoagulants. Lancet 1982; 1:506. 154. Gebauer MG, Nyfort-Hansen K, Henschke PJ, Gallus AS. Warfarin and acetaminophen interaction. Pharmacotherapy 2003; 23:109–12. 155. Mahe´ I, Bertrand N, Drouet L, et al. Paracetamol: a haemorrhagic risk factor in patients on warfarin. Br J Clin Pharmacol 2005; 59:371–4. 156. Rubin RN, Mentzer RL, Budzynski AZ. Potentiation of anticoagulant effect of warfarin by acetaminophen (Tylenol). Clin Res 1984; 32:698a. 157. Gadisseur AP, Van Der Meer FJ, Rosendaal FR. Sustained intake of paracetamol (acetaminophen) during oral anticoagulant therapy with coumarins does not cause clinically important INR changes: a randomized double-blind clinical trial. J Thromb Haemost 2003; 1:714–7.

380

Nelson and Bruschi

158. Kwan D, Bartle WR, Walker SE. The effects of acetaminophen on pharmacokinetics and pharmacodynamics of warfarin. J Clin Pharmacol 1999; 39:68–75. 159. Shek KL, Chan LN, Nutescu E. Warfarin-acetaminophen drug interaction revisited. Pharmacotherapy 1999; 19:1153–8. 160. Hylek EM, Heiman H, Skates SJ, Sheehan MA, Singer DE. Acetaminophen and other risk factors for excessive warfarin anticoagulation. JAMA 1998; 279:657–62. 161. Mahe´ I, Caulin C, Bergmann JF. Does paracetamol (acetaminophen) potentiate the effects of oral anticoagulants? A literature review Drug Saf 2004; 27:283–352. 162. Takigawa T, Tainaka H, Mihara K, Ogata H. Inhibition of S-warfarin metabolism by nonsteroidal antiinflammatory drugs in human liver microsomes in vitro. Biol Pharm Bull 1998; 21:541–3. 163. Whyte IM, Buckley NA, Reith DM, Goodhew I, Seldon M, Dawson AH. Acetaminophen causes an increased International Normalized Ratio by reducing functional factor VII. Ther Drug Monit 2000; 22:742–8. 164. Thijssen HH, Soute BA, Vervoort LM, Claessens JG. Paracetamol (acetaminophen) warfarin interaction: NAPQI, the toxic metabolite of paracetamol, is an inhibitor of enzymes in the vitamin K cycle. Thromb Haemost 2004; 92:797–802. 165. Nelson SD. Mechanisms of the formation and disposition of reactive metabolites that can cause acute liver injury. Drug Metab Rev 1995; 27:147–77. 166. Myers TG, Thummel KE, Kalhorn TF, Nelson SD. Preferred orientations in the binding of 4 0 hydroxyacetanilide (acetaminophen) to cytochrome P450 1A1 and 2B1 isoforms as determined by 13C- and 15N-NMR relaxation studies. J Med Chem 1994; 37:860–7. 167. Schlichting I, Berendzen J, Chu K, et al. The catalytic pathway of cytochrome p450cam at atomic resolution. Science 2000; 287:1615–22. 168. Hasemann CA, Kurumbail RG, Boddupalli SS, Peterson JA, Deisenhofer J. Structure and function of cytochromes P450: a comparative analysis of three crystal structures. Structure 1995; 3:41–62. 169. Modi S, Primrose WU, Boyle JM, Gibson CF, Lian LY, Roberts GC. NMR studies of substrate binding to cytochrome P450 BM3: comparisons to cytochrome P450 cam. Biochemistry 1995; 34:8982–8. 170. Lewis DF, Bird MG, Dickins M, Lake BG, Eddershaw PJ, Tarbit MH, Goldfarb PS. Molecular modelling of human CYP2E1 by homology with the CYP102 haemoprotein domain: investigation of the interactions of substrates and inhibitors within the putative active site of the human CYP2E1 isoform. Xenobiotica 2000; 30:1–25. 171. Rashed MS, Myers TG, Nelson SD. Hepatic protein arylation, glutathione depletion, and metabolite profiles of acetaminophen and a nonhepatotoxic regioisomer, 3 0 -hydroxyacetanilide, in the mouse. Drug Metab Dispos 1990; 18:765–70. 172. Roberts SA, Price VF, Jollow DJ. Acetaminophen structure-toxicity studies: in vivo covalent binding of a nonhepatotoxic analog, 3-hydroxyacetanilide. Toxicol Appl Pharmacol 1990; 105:195–208. 173. Tirmenstein MA, Nelson SD. Subcellular binding and effects on calcium homeostasis produced by acetaminophen and a nonhepatotoxic regioisomer, 3 0 -hydroxyacetanilide. J Biol Chem 1989; 264:9814–9. 174. Nelson SD, Pearson PG. Covalent and noncovalent interactions in acute lethal cell injury caused by chemicals. Annu Rev Pharmacol Toxicol 1990; 30:169–95. 175. Rogers LK, Valentine CJ, Szczpyka M, Smith CV. Effects of hepatotoxic doses of acetaminophen and furosemide on tissue concentrations of CoASH and CoASSG in vivo. Chem Res Toxicol 2000; 13:873–82. 176. Tirmenstein MA, Nelson SD. Acetaminophen-induced oxidation of protein thiols. Contribution of impaired thiol-metabolizing enzymes and the breakdown of adenine nucleotides. J Biol Chem 1990; 265:3059–65. 177. Hart SG, Cartun RW, Wyand DS, Khairallah EA, Cohen SD. Immunohistochemical localization of acetaminophen in target tissues of the CD-1 mouse: correspondence of covalent binding with toxicity. Fundam Appl Toxicol 1995; 24:260–74. 178. Roberts DW, Bucci TJ, Benson RW, Warbritton AR, McRae TA, Pumford NR, Hinson JA. Immunohistochemical localization and quantification of the 3-(cystein-S-yl)-acetaminophen protein adduct in acetaminophen hepatotoxicity. Am J Pathol 1991; 138:359–71. 179. Bartolone JB, Beierschmitt WP, Birge RB, Hart SG, Wyand S, Cohen SD, Khairallah EA. Selective acetaminophen metabolite binding to hepatic and extrahepatic proteins: an in vivo and in vitro analysis. Toxicol Appl Pharmacol 1989; 99:240–9. 180. Hinson JA, Roberts DW, Benson RW, Dalhoff K, Loft S, Poulsen HE. Mechanism of paracetamol toxicity. Lancet 1990; 335:732. 181. James LP, Farrar HC, Sullivan JE, et al. Measurement of acetaminophen–protein adducts in children and adolescents with acetaminophen overdoses. J Clin Pharmacol 2001; 41:846–51. 182. Muldrew KL, James LP, Coop L, McCullough SS, Hendrickson HP, Hinson JA, Mayeux PR. Determination of acetaminophen–protein adducts in mouse liver and serum and human serum after hepatotoxic doses of acetaminophen using high-performance liquid chromatography with electrochemical detection. Drug Metab Dispos 2002; 30:446–51.

Acetaminophen-Induced Liver Disease

381

183. Dietze EC, Schafer A, Omichinski JG, Nelson SD. Inactivation of glyceraldehyde-3-phosphate dehydrogenase by a reactive metabolite of acetaminophen and mass spectral characterization of an arylated active site peptide. Chem Res Toxicol 1997; 10:1097–103. 184. Hoffmann KJ, Streeter AJ, Axworthy DB, Baillie TA. Identification of the major covalent adduct formed in vitro and in vivo between acetaminophen and mouse liver proteins. Mol Pharmacol 1985; 27:566–73. 185. Potter DW, Pumford NR, Hinson JA, Benson RW, Roberts DW. Epitope characterization of acetaminophen bound to protein and nonprotein sulfhydryl groups by an enzyme-linked immunosorbent assay. J Pharmacol Exp Ther 1989; 248:182–9. 186. Streeter AJ, Dahlin DC, Nelson SD, Baillie TA. The covalent binding of acetaminophen to protein. Evidence for cysteine residues as major sites of arylation in vitro. Chem Biol Interact 1984; 48:349–66. 187. Zhou L, McKenzie BA, Eccleston ED, Jr., Srivastava SP, Chen N, Erickson RR, Holtzman JL. The covalent binding of [14C]acetaminophen to mouse hepatic microsomal proteins: the specific binding to calreticulin and the two forms of the thiol:protein disulfide oxidoreductases. Chem Res Toxicol 1996; 9:1176–82. 188. Matthews AM, Roberts DW, Hinson JA, Pumford NR. Acetaminophen-induced hepatotoxicity. Analysis of total covalent binding vs. specific binding to cysteine. Drug Metab Dispos 1996; 24:1192–6. 189. Cohen SD, Khairallah EA. Selective protein arylation and acetaminophen-induced hepatotoxicity. Drug Metab Rev 1997; 29:59–77. 190. Hinson JA, Reid AB, McCullough SS, James LP. Acetaminophen-induced hepatotoxicity: role of metabolic activation, reactive oxygen/nitrogen species, and mitochondrial permeability transition. Drug Metab Rev 2004; 36:805–22. 191. Pumford NR, Halmes NC. Protein targets of xenobiotic reactive intermediates. Annu Rev Pharmacol Toxicol 1997; 37:91–117. 192. Qiu Y, Benet LZ, Burlingame AL. Identification of the hepatic protein targets of reactive metabolites of acetaminophen in vivo in mice using two-dimensional gel electrophoresis and mass spectrometry. J Biol Chem 1998; 273:17940–53. 193. Bartolone JB, Birge RB, Bulera SJ, Bruno MK, Nishanian EV, Cohen SD, Khairallah EA. Purification, antibody production, and partial amino acid sequence of the 58-kDa acetaminophen-binding liver proteins. Toxicol Appl Pharmacol 1992; 113:19–29. 194. Pumford NR, Martin BM, Hinson JA. A metabolite of acetaminophen covalently binds to the 56 kDa selenium binding protein. Biochem Biophys Res Commun 1992; 182:1348–55. 195. Pumford NR, Halmes NC, Martin BM, Cook RJ, Wagner C, Hinson JA. Covalent binding of acetaminophen to N-10-formyltetrahydrofolate dehydrogenase in mice. J Pharmacol Exp Ther 1997; 280:501–5. 196. Lores Arnaiz S, Llesuy S, Cutrin JC, Boveris A. Oxidative stress by acute acetaminophen administration in mouse liver. Free Radic Biol Med 1995; 19:303–10. 197. Bulera SJ, Birge RB, Cohen SD, Khairallah EA. Identification of the mouse liver 44-kDa acetaminophen-binding protein as a subunit of glutamine synthetase. Toxicol Appl Pharmacol 1995; 134:313–20. 198. Gupta S, Rogers LK, Taylor SK, Smith CV. Inhibition of carbamyl phosphate synthetase-I and glutamine synthetase by hepatotoxic doses of acetaminophen in mice. Toxicol Appl Pharmacol 1997; 146:317–27. 199. Landin JS, Cohen SD, Khairallah EA. Identification of a 54-kDa mitochondrial acetaminophenbinding protein as aldehyde dehydrogenase. Toxicol Appl Pharmacol 1996; 141:299–307. 200. Halmes NC, Hinson JA, Martin BM, Pumford NR. Glutamate dehydrogenase covalently binds to a reactive metabolite of acetaminophen. Chem Res Toxicol 1996; 9:541–6. 201. Khairallah EA, Bruno MK, Hong M, Cohen SD. Cellular consequences of protein adduct formation. Toxicologist (Abs) 1995; 15:86. 202. Fountoulakis M, Berndt P, Boelsterli UA, Crameri F, Winter M, Albertini S, Suter L. Two-dimensional database of mouse liver proteins: changes in hepatic protein levels following treatment with acetaminophen or its nontoxic regioisomer 3-acetamidophenol. Electrophoresis 2000; 21:2148–61. 203. Myers TG, Dietz EC, Anderson NL, Khairallah EA, Cohen SD, Nelson SD. A comparative study of mouse liver proteins arylated by reactive metabolites of acetaminophen and its nonhepatotoxic regioisomer, 3 0 -hydroxyacetanilide. Chem Res Toxicol 1995; 8:403–13. 204. Qiu Y, Benet LZ, Burlingame AL. Identification of hepatic protein targets of the reactive metabolites of the nonhepatotoxic regioisomer of acetaminophen, 3 0 -hydroxyacetanilide, in the mouse in vivo using two-dimensional gel electrophoresis and mass spectrometry. Adv Exp Med Biol 2001; 500:663–73. 205. Porat A, Sagiv Y, Elazar Z. A 56-kDa selenium-binding protein participates in intra-Golgi protein transport. J Biol Chem 2000; 275:14457–65. 206. Gasdaska PY, Gasdaska JR, Cochran S, Powis G. Cloning and sequencing of a human thioredoxin reductase. FEBS Lett 1995; 373:5–9.

382

Nelson and Bruschi

207. Gladyshev VN, Jeang KT, Stadtman TC. Selenocysteine, identified as the penultimate C-terminal residue in human T-cell thioredoxin reductase, corresponds to TGA in the human placental gene. Proc Natl Acad Sci USA 1996; 93:6146–51. 208. Luthman M, Holmgren A. Rat liver thioredoxin and thioredoxin reductase: purification and characterization. Biochemistry 1982; 21:6628–33. 209. Matthews AM, Hinson JA, Roberts DW, Pumford NR. Comparison of covalent binding of acetaminophen and the regioisomer 3 0 -hydroxyacetanilide to mouse liver protein. Toxicol Lett 1997; 90:77–82. 210. Tirmenstein MA, Nelson SD. Hepatotoxicity after 3 0 -hydroxyacetanilide administration to buthionine sulfoximine pretreated mice. Chem Res Toxicol 1991; 4:214–7. 211. Dixon MF, Dixon B, Aparicio SR, Loney DP. Experimental paracetamol-induced hepatic necrosis: a light- and electron-microscope, and histochemical study. J Pathol 1975; 116:17–29. 212. Walker RM, Racz WJ, McElligott TF. Acetaminophen-induced hepatotoxicity in mice. Lab Invest 1980; 42:181–9. 213. Burcham PC, Harman AW. Acetaminophen toxicity results in site-specific mitochondrial damage in isolated mouse hepatocytes. J Biol Chem 1991; 266:5049–54. 214. Donnelly PJ, Walker RM, Racz WJ. Inhibition of mitochondrial respiration in vivo is an early event in acetaminophen-induced hepatotoxicity. Arch Toxicol 1994; 68:110–8. 215. Katyare SS, Satav JG. Impaired mitochondrial oxidative energy metabolism following paracetamolinduced hepatotoxicity in the rat. Br J Pharmacol 1989; 96:51–8. 216. Meyers LL, Beierschmitt WP, Khairallah EA, Cohen SD. Acetaminophen-induced inhibition of hepatic mitochondrial respiration in mice. Toxicol Appl Pharmacol 1988; 93:378–87. 217. Ramsay RR, Rashed MS, Nelson SD. In vitro effects of acetaminophen metabolites and analogs on the respiration of mouse liver mitochondria. Arch Biochem Biophys 1989; 273:449–57. 218. Strubelt O, Younes M. The toxicological relevance of paracetamol-induced inhibition of hepatic respiration and ATP depletion. Biochem Pharmacol 1992; 44:163–70. 219. Vendemiale G, Grattagliano I, Altomare E, Turturro N, Guerrieri F. Effect of acetaminophen administration on hepatic glutathione compartmentation and mitochondrial energy metabolism in the rat. Biochem Pharmacol 1996; 52:1147–54. 220. Jaeschke H. Glutathione disulfide formation and oxidant stress during acetaminophen-induced hepatotoxicityinmiceinvivo: the protectiveeffect ofallopurinol. J Pharmacol ExpTher 1990;255:935–41. 221. Martin FL, McLean AE. Adenosine triphosphate (ATP) levels in paracetamol-induced cell injury in the rat in vivo and in vitro. Toxicology 1995; 104:91–7. 222. Parmar DV, Ahmed G, Khandkar MA, Katyare SS. Mitochondrial ATPase: a target for paracetamolinduced hepatotoxicity. Eur J Pharmacol 1995; 293:225–9. 223. Burcham PC, Harman AW. Effect of acetaminophen hepatotoxicity on hepatic mitochondrial and microsomal calcium contents in mice. Toxicol Lett 1988; 44:91–9. 224. Lauterburg BH. Early disturbance of calcium translocation across the plasma membrane in toxic liver injury. Hepatology 1987; 7:1179–83. 225. Moore M, Thor H, Moore G, Nelson S, Moldeus P, Orrenius S. The toxicity of acetaminophen and N-acetyl-p-benzoquinone imine in isolated hepatocytes is associated with thiol depletion and increased cytosolic Ca2C. J Biol Chem 1985; 260:13035–40. 226. Nicotera P, Rundgren M, Porubek DJ, Cotgreave I, Moldeus P, Orrenius S, Nelson SD. On the role of Ca2C in the toxicity of alkylating and oxidizing quinone imines in isolated hepatocytes. Chem Res Toxicol 1989; 2:46–50. 227. Tsokos-Kuhn JO, Todd EL, McMillin-Wood JB, Mitchell JR. ATP-dependent calcium uptake by rat liver plasma membrane vesicles. Effect of alkylating hepatotoxins in vivo. Mol Pharmacol 1985; 28:56–61. 228. Corcoran GB, Chung SJ, Salazar DE. Early inhibition of the NaC/KC-ATPase ion pump during acetaminophen-induced hepatotoxicity in rat. Biochem Biophys Res Commun 1987; 149:203–7. 229. Bruno MK, Khairallah EA, Cohen SD. Inhibition of protein phosphatase activity and changes in protein phosphorylation following acetaminophen exposure in cultured mouse hepatocytes. Toxicol Appl Pharmacol 1998; 153:119–32. 230. Labadarios D, Davis M, Portmann B, Williams R. Paracetamol-induced hepatic necrosis in the mouse-relationship between covalent binding, hepatic glutathione depletion and the protective effect of alpha-mercaptopropionylglycine. Biochem Pharmacol 1977; 26:31–5. 231. Nelson SD, Tirmenstein MA, Rashed MS, Myers TG. Acetaminophen and protein thiol oxidation. In: Witmer CM et al, ed. Biological Reactive Intermediates IV. New York: Plenum, 1991:579–88. 232. Tee LB, Boobis AR, Huggett AC, Davies DS. Reversal of acetaminophen toxicity in isolated hamster hepatocytes by dithiothreitol. Toxicol Appl Pharmacol 1986; 83:294–314. 233. Lauterburg BH, Smith CV, Hughes H, Mitchell JR. Biliary excretion of glutathione and glutathione disulfide in the rat. Regulation and response to oxidative stress. J Clin Invest 1984; 73:124–33. 234. Smith CV. Correlations and apparent contradictions in assessment of oxidant stress status in vivo. Free Radic Biol Med 1991; 10:217–24.

Acetaminophen-Induced Liver Disease

383

235. Smith CV, Mitchell JR. Acetaminophen hepatotoxicity in vivo is not accompanied by oxidant stress. Biochem Biophys Res Commun 1985; 133:329–36. 236. Amimoto T, Matsura T, Koyama SY, Nakanishi T, Yamada K, Kajiyama G. Acetaminophen-induced hepatic injury in mice: the role of lipid peroxidation and effects of pretreatment with coenzyme and alpha-tocopherol. Free Radic Biol Med 1995; 19:169–76. 237. Fairhurst S, Barber DJ, Clark B, Horton AA. Studies on paracetamol-induced lipid peroxidation. Toxicology 1982; 23:249–59. 238. Wendel A, Feuerstein S, Konz KH. Acute paracetamol intoxication of starved mice leads to lipid peroxidation in vivo. Biochem Pharmacol 1979; 28:2051–5. 239. Burk RF, Lane JM. Ethane production and liver necrosis in rats after administration of drugs and other chemicals. Toxicol Appl Pharmacol 1979; 50:467–78. 240. Kamiyama T, Sato C, Liu J, Tajiri K, Miyakawa H, Marumo F. Role of lipid peroxidation in acetaminophen-induced hepatotoxicity: comparison with carbon tetrachloride. Toxicol Lett 1993; 66:7–12. 241. Morrow JD, Awad JA, Kato T, Takahashi K, Badr KF, Roberts LJ, II, Burk RF. Formation of novel noncyclooxygenase-derived prostanoids (F2-isoprostanes) in carbon tetrachloride hepatotoxicity. An animal model of lipid peroxidation. J Clin Invest 1992; 90:2502–7. 242. Walker RM, Massey TE, McElligott TF, Racz WJ. Acetaminophen toxicity in fed and fasted mice. Can J Physiol Pharmacol 1982; 60:399–404. 243. Morrow JD, Chen Y, Brame CJ, et al. The isoprostanes: unique prostaglandin-like products of freeradical-initiated lipid peroxidation. Drug Metab Rev 1999; 31:117–39. 244. Gupta S, Rogers LK, Smith CV. Biliary excretion of lysosomal enzymes, iron, and oxidized protein in Fischer-344 and Sprague–Dawley rats and the effects of diquat and acetaminophen. Toxicol Appl Pharmacol 1994; 125:42–50. 245. Gibson JD, Pumford NR, Samokyszyn VM, Hinson JA. Mechanism of acetaminophen-induced hepatotoxicity: covalent binding versus oxidative stress. Chem Res Toxicol 1996; 9:580–5. 246. Schnellmann JG, Pumford NR, Kusewitt DF, Bucci TJ, Hinson JA. Deferoxamine delays the development of the hepatotoxicity of acetaminophen in mice. Toxicol Lett 1999; 106:79–88. 247. Younes M, Sause C, Siegers CP, Lemoine R. Effect of deferrioxamine and diethyldithiocarbamate on paracetamol-induced hepato- and nephrotoxicity. The role of lipid peroxidation. J Appl Toxicol 1988; 8:261–5. 248. Younes M, Siegers CP. The role of iron in the paracetamol- and CCl4-induced lipid peroxidation and hepatotoxicity. Chem Biol Interact 1985; 55:327–34. 249. ElSisi AE, Hall P, Sim WL, Earnest DL, Sipes IG. Characterization of vitamin A potentiation of carbon tetrachloride-induced liver injury. Toxicol Appl Pharmacol 1993; 119:280–8. 250. Laskin DL. Sinusoidal lining cells and hepatotoxicity. Toxicol Pathol 1996; 24:112–8. 251. Nakae D, Yamamoto K, Yoshiji H, Kinugasa T, Maruyama H, Farber JL, Konishi Y. Liposomeencapsulated superoxide dismutase prevents liver necrosis induced by acetaminophen. Am J Pathol 1990; 136:787–95. 252. Harvison PJ, Egan RW, Gale PH, Christian GD, Hill BS, Nelson SD. Acetaminophen and analogs as cosubstrates and inhibitors of prostaglandin H synthase. Chem Biol Interact 1988; 64:251–66. 253. Potter DW, Hinson JA. The 1- and 2-electron oxidation of acetaminophen catalyzed by prostaglandin H synthase. J Biol Chem 1987; 262:974–80. 254. Potter DW, Miller DW, Hinson JA. Identification of acetaminophen polymerization products catalyzed by horseradish peroxidase. J Biol Chem 1985; 260:12174–80. 255. West PR, Harman LS, Josephy PD, Mason RP. Acetaminophen: enzymatic formation of a transient phenoxyl free radical. Biochem Pharmacol 1984; 33:2933–6. 256. Fischer V, West PR, Nelson SD, Harvison PJ, Mason RP. Formation of 4-aminophenoxyl free radical from the acetaminophen metabolite N-acetyl-p-benzoquinone imine. J Biol Chem 1985; 260:11446–50. 257. Bisby RH, Tabassum N. Properties of the radicals formed by one-electron oxidation of acetaminophen—a pulse radiolysis study. Biochem Pharmacol 1988; 37:2731–8. 258. Keller RJ, Hinson JA. Mechanism of acetaminophen-stimulated NADPH oxidation catalyzed by the peroxidase-H2O2 system. Drug Metab Dispos 1991; 19:184–7. 259. Ross D, Albano E, Nilsson U, Moldeus P. Thiyl radicals—formation during peroxidase-catalyzed metabolism of acetaminophen in the presence of thiols. Biochem Biophys Res Commun 1984; 125:109–15. 260. van de Straat R, de Vries J, Vermeulen NP. Role of hepatic microsomal and purified cytochrome P-450 in one-electron reduction of two quinone imines and concomitant reduction of molecular oxygen. Biochem Pharmacol 1987; 36:613–9. 261. Senter PD, Al-Abed Y, Metz CN, et al. Inhibition of macrophage migration inhibitory factor (MIF) tautomerase and biological activities by acetaminophen metabolites. Proc Natl Acad Sci USA 2002; 99:144–9. 262. Jaeschke H, Knight TR, Bajt ML. The role of oxidant stress and reactive nitrogen species in acetaminophen hepatotoxicity. Toxicol Lett 2003; 144:279–88.

384

Nelson and Bruschi

263. James LP, Mayeux PR, Hinson JA. Acetaminophen-induced hepatotoxicity. Drug Metab Dispos 2003; 31:1499–506. 264. Knight TR, Ho YS, Farhood A, Jaeschke H. Peroxynitrite is a critical mediator of acetaminophen hepatotoxicity in murine livers: protection by glutathione. J Pharmacol Exp Ther 2002; 303:468–75. 265. Bruschi SA. Methods and approaches to study metabolism and toxicity of acetaminophen. In: Lash LH, ed. Drug Metabolism and Transport. Molecular Methods and Mechanisms. Totowa, NJ: Humana, 2005:387. 266. Gonzalez FJ. The use of gene knockout mice to unravel the mechanisms of toxicity and chemical carcinogenesis. Toxicol Lett 2001; 120:199–208. 267. Mokhlesi B, Leikin JB, Murray P, Corbridge TC. Adult toxicology in critical care: part II: specific poisonings. Chest 2003; 123:897–922. 268. O’Grady JG. Broadening the view of acetaminophen hepatotoxicity. Hepatology 2005; 42:1252–4. 269. Schmidt LE, Dalhoff K. Serum phosphate is an early predictor of outcome in severe acetaminopheninduced hepatotoxicity. Hepatology 2002; 36:659–65. 270. Heinloth AN, Irwin RD, Boorman GA, et al. Gene expression profiling of rat livers reveals indicators of potential adverse effects. Toxicol Sci 2004; 80:193–202. 271. Adams ML, Pierce RH, Vail ME, et al. Enhanced acetaminophen hepatotoxicity in transgenic mice overexpressing BCL-2. Mol Pharmacol 2001; 60:907–15. 272. Freeman K. From point B to point A: applying toxicogenomics to biological inference. Environ Health Perspect 2005; 113:A388–93. 273. Kitteringham NR, Powell H, Jenkins RE, et al. Protein expression profiling of glutathione S-transferase pi null mice as a strategy to identify potential markers of resistance to paracetamolinduced toxicity in the liver. Proteomics 2003; 3:191–207. 274. Ruepp SU, Tonge RP, Shaw J, Wallis N, Pognan F. Genomics and proteomics analysis of acetaminophen toxicity in mouse liver. Toxicol Sci 2002; 65:135–50. 275. Welch KD, Wen B, Goodlett DR, et al. Proteomic identification of potential susceptibility factors in drug-induced liver disease. Chem Res Toxicol 2005; 18:924–33. 276. Zhu JH, Lei XG. Double null of selenium-glutathione peroxidase-1 and copper, zinc-superoxide dismutase enhances resistance of mouse primary hepatocytes to acetaminophen toxicity. Exp Biol Med (Maywood) 2006; 231:545–52. 277. Yamamoto T, Kikkawa R, Yamada H, Horii I. Investigation of proteomic biomarkers in in vivo hepatotoxicity study of rat liver: toxicity differentiation in hepatotoxicants. J Toxicol Sci 2006; 31:49–60. 278. Ho HK, White CC, Fernandez C, Fausto N, Kavanagh TJ, Nelson SD, Bruschi SA. Nrf2 activation involves an oxidative-stress independent pathway in tetrafluoroethylcysteine-induced cytotoxicity. Toxicol Sci 2005; 86:354–64. 279. Suntres ZE. Role of antioxidants in paraquat toxicity. Toxicology 2002; 180:65–77. 280. Devalia JL, McLean AE. Covalent binding and the mechanism of paracetamol toxicity. Biochem Pharmacol 1983; 32:2602–3. 281. Devalia JL, Ogilvie RC, McLean AE. Dissociation of cell death from covalent binding of paracetamol by flavones in a hepatocyte system. Biochem Pharmacol 1982; 31:3745–9. 282. Streeter AJ, Bjorge SM, Axworthy DB, Nelson SD, Baillie TA. The microsomal metabolism and site of covalent binding to protein of 3 0 -hydroxyacetanilide, a nonhepatotoxic positional isomer of acetaminophen. Drug Metab Dispos 1984; 12:565–76. 283. Nelson SD, Bruschi S. Mechanisms of acetaminophen-induced liver disease. In: Kaplowitz N, DeLeve LD, eds. Drug Induced Liver Disease. 1st ed. New York: Marcel Dekker, 2003:773. 284. Hara MR, Thomas B, Cascio MB, et al. Neuroprotection by pharmacologic blockade of the GAPDH death cascade. Proc Natl Acad Sci USA 2006; 103:3887–9. 285. Hara MR, Agrawal N, Kim SF, et al. S-nitrosylated GAPDH initiates apoptotic cell death by nuclear translocation following Siah1 binding. Nat Cell Biol 2005; 7:665–74. 286. Sirover MA. New insights into an old protein: the functional diversity of mammalian glyceraldehyde-3-phosphate dehydrogenase. Biochim Biophys Acta 1999; 1432:159–84. 287. Gerson RJ, Casini A, Gilfor D, Serroni A, Farber JL. Oxygen-mediated cell injury in the killing of cultured hepatocytes by acetaminophen. Biochem Biophys Res Commun 1985; 126:1129–37. 288. Kyle ME, Miccadei S, Nakae D, Farber JL. Superoxide dismutase and catalase protect cultured hepatocytes from the cytotoxicity of acetaminophen. Biochem Biophys Res Commun 1987; 149:889–96. 289. Mirochnitchenko O, Weisbrot-Lefkowitz M, Reuhl K, Chen L, Yang C, Inouye M. Acetaminophen toxicity. Opposite effects of two forms of glutathione peroxidase. J Biol Chem 1999; 274:10349–55. 290. Boobis AR, Seddon CE, Nasseri-Sina P, Davies DS. Evidence for a direct role of intracellular calcium in paracetamol toxicity. Biochem Pharmacol 1990; 39:1277–81. 291. Farber JL. The biochemical pathology of toxic cell death. Monogr Pathol 1985;19–31. 292. Tirmenstein MA, Nelson SD. Subcellular binding and effects on calcium homeostasis produced by acetaminophen and a nonhepatotoxic regioisomer, 3 0 -hydroxyacetanilide, in mouse liver. J Biol Chem 1989; 264:9814–9.

Acetaminophen-Induced Liver Disease

385

293. Bruschi SA, Priestly BG. Implications of alterations in intracellular calcium ion homeostasis in the advent of paracetamol-induced cytotoxicity in primary mouse hepatocyte cultures. Toxicol In Vitro 1990; 4:743–9. 294. Coen M, Lenz EM, Nicholson JK, Wilson ID, Pognan F, Lindon JC. An integrated metabonomic investigation of acetaminophen toxicity in the mouse using NMR spectroscopy. Chem Res Toxicol 2003; 16:295–303. 295. Coen M, Ruepp SU, Lindon JC, Nicholson JK, Pognan F, Lenz EM, Wilson ID. Integrated application of transcriptomics and metabonomics yields new insight into the toxicity due to paracetamol in the mouse. J Pharm Biomed Anal 2004; 35:93–105. 296. Welch KD, Reilly TP, Bourdi M, et al. Genomic identification of potential risk factors during acetaminophen-induced liver disease in susceptible and resistant strains of mice. Chem Res Toxicol 2006; 19:223–33. 297. Clayton TA, Lindon JC, Cloarec O, et al. Pharmaco-metabonomic phenotyping and personalized drug treatment. Nature 2006; 440:1073–7. 298. Kikkawa R, Fujikawa M, Yamamoto T, Hamada Y, Yamada H, Horii I. In vivo hepatotoxicity study of rats in comparison with in vitro hepatotoxicity screening system. J Toxicol Sci 2006; 31:23–34. 299. Minami K, Saito T, Narahara M, et al. Relationship between hepatic gene expression profiles and hepatotoxicity in five typical hepatotoxicant-administered rats. Toxicol Sci 2005; 87:296–305. 300. Reilly TP, Bourdi M, Brady JN, Pise-Masison CA, Radonovich MF, George JW, Pohl LR. Expression profiling of acetaminophen liver toxicity in mice using microarray technology. Biochem Biophys Res Commun 2001; 282:321–8. 301. Tolson JK, Dix DJ, Voellmy RW, Roberts SM. Increased hepatotoxicity of acetaminophen in Hsp70i knockout mice. Toxicol Appl Pharmacol 2006; 210:157–62. 302. Williams DP, Garcia-Allan C, Hanton G, et al. Time course toxicogenomic profiles in CD-1 mice after nontoxic and nonlethal hepatotoxic paracetamol administration. Chem Res Toxicol 2004; 17:1551–61. 303. Yamamoto T, Kikkawa R, Yamada H, Horii I. Identification of oxidative stress-related proteins for predictive screening of hepatotoxicity using a proteomic approach. J Toxicol Sci 2005; 30:213–27. 304. Gujral JS, Knight TR, Farhood A, Bajt ML, Jaeschke H. Mode of cell death after acetaminophen overdose in mice: apoptosis or oncotic necrosis? Toxicol Sci 2002; 67:322–8. 305. El-Hassan H, Anwar K, Macanas-Pirard P, et al. Involvement of mitochondria in acetaminopheninduced apoptosis and hepatic injury: roles of cytochrome c, Bax, Bid, and caspases. Toxicol Appl Pharmacol 2003; 191:118–29. 306. Macanas-Pirard P, Yaacob NS, Lee PC, Holder JC, Hinton RH, Kass GE. Glycogen synthase kinase-3 mediates acetaminophen-induced apoptosis in human hepatoma cells. J Pharmacol Exp Ther 2005; 313:780–9. 307. Bajt ML, Lemasters JJ, Jaeschke H. Role of mitochondrial acetaminophen bax translocation in acetaminophen-induced hepatic necrosis in mice. Society of Toxicology. New Orleans: Oxford University Press, 2005. 308. Bassik MC, Scorrano L, Oakes SA, Pozzan T, Korsmeyer SJ. Phosphorylation of BCL-2 regulates ER Ca2C homeostasis and apoptosis. EMBO J 2004; 23:1207–16. 309. Oakes SA, Scorrano L, Opferman JT, Bassik MC, Nishino M, Pozzan T, Korsmeyer SJ. Proapoptotic BAX and BAK regulate the type 1 inositol trisphosphate receptor and calcium leak from the endoplasmic reticulum. Proc Natl Acad Sci USA 2005; 102:105–10. 310. Scorrano L, Oakes SA, Opferman JT, Cheng EH, Sorcinelli MD, Pozzan T, Korsmeyer SJ. BAX and BAK regulation of endoplasmic reticulum Ca2C: a control point for apoptosis. Science 2003; 300:135–9. 311. Sawada M, Hayes P, Matsuyama S. Cytoprotective membrane-permeable peptides designed from the Bax-binding domain of Ku70. Nat Cell Biol 2003; 5:352–7. 312. Sawada M, Sun W, Hayes P, Leskov K, Boothman DA, Matsuyama S. Ku70 suppresses the apoptotic translocation of Bax to mitochondria. Nat Cell Biol 2003; 5:320–9. 313. Yoshida T, Tomioka I, Nagahara T, Holyst T, Sawada M, Hayes P, Gama V, Okuno M, Chen Y, Abe Y, et al. Bax-inhibiting peptide derived from mouse and rat Ku70. Biochem Biophys Res Commun 2004; 321:961–6. 314. Guo B, Zhai D, Cabezas E, Welsh K, Nouraini S, Satterthwait AC, Reed JC. Humanin peptide suppresses apoptosis by interfering with Bax activation. Nature 2003; 423:456–61. 315. Zhai D, Luciano F, Zhu X, Guo B, Satterthwait AC, Reed JC. Humanin binds and nullifies Bid activity by blocking its activation of Bax and Bak. J Biol Chem 2005; 280:15815–24. 316. Chae HJ, Kim HR, Xu C, Bailly-Maitre B, Krajewska M, Krajewski S, Banares S, Cui J, Digicaylioglu M, Ke N, et al. BI-1 regulates an apoptosis pathway linked to endoplasmic reticulum stress. Mol Cell 2004; 15:355–66. 317. Huckelhoven R. BAX Inhibitor-1, an ancient cell death suppressor in animals and plants with prokaryotic relatives. Apoptosis 2004; 9:299–307.

386

Nelson and Bruschi

318. Oltersdorf T, Elmore SW, Shoemaker AR, Armstrong RC, Augeri DJ, Belli BA, Bruncko M, Deckwerth TL, Dinges J, Hajduk PJ, et al. An inhibitor of Bcl-2 family proteins induces regression of solid tumours. Nature 2005; 435:677–81. 319. Bernardi P. The permeability transition pore. Control points of a cyclosporin A-sensitive mitochondrial channel involved in cell death. Biochim Biophys Acta 1996; 1275:5–9. 320. Bouchier-Hayes L, Lartigue L, Newmeyer DD. Mitochondria: pharmacological manipulation of cell death. J Clin Invest 2005; 115:2640–7. 321. Kon K, Kim JS, Jaeschke H, Lemasters JJ. Mitochondrial permeability transition in acetaminopheninduced necrosis and apoptosis of cultured mouse hepatocytes. Hepatology 2004; 40:1170–9. 322. Masubuchi Y, Suda C, Horie T. Involvement of mitochondrial permeability transition in acetaminophen-induced liver injury in mice. J Hepatol 2005; 42:110–6. 323. Reid AB, Kurten RC, McCullough SS, Brock RW, Hinson JA. Mechanisms of acetaminopheninduced hepatotoxicity: role of oxidative stress and mitochondrial permeability transition in freshly isolated mouse hepatocytes. J Pharmacol Exp Ther 2005; 312:509–16. 324. Haouzi D, Cohen I, Vieira HL, et al. Mitochondrial permeability transition as a novel principle of hepatorenal toxicity in vivo. Apoptosis 2002; 7:395–405. 325. Fiorucci S, Antonelli E, Distrutti E, Mencarelli A, Farneti S, Del Soldato P, Morelli A. Liver delivery of NO by NCX-1000 protects against acute liver failure and mitochondrial dysfunction induced by APAP in mice. Br J Pharmacol 2004; 143:33–42. 326. Basso E, Fante L, Fowlkes J, Petronilli V, Forte MA, Bernardi P. Properties of the permeability transition pore in mitochondria devoid of Cyclophilin D. J Biol Chem 2005; 280:18558–61. 327. Baines CP, Kaiser RA, Purcell NH, et al. Loss of cyclophilin D reveals a critical role for mitochondrial permeability transition in cell death. Nature 2005; 434:658–62. 328. Schinzel AC, Takeuchi O, Huang Z, et al. Cyclophilin D is a component of mitochondrial permeability transition and mediates neuronal cell death after focal cerebral ischemia. Proc Natl Acad Sci USA 2005; 102:12005–10. 329. Nakagawa T, Shimizu S, Watanabe T, et al. Cyclophilin D-dependent mitochondrial permeability transition regulates some necrotic but not apoptotic cell death. Nature 2005; 434:652–8. 330. Lavrik IN, Golks A, Krammer PH. Caspases: pharmacological manipulation of cell death. J Clin Invest 2005; 115:2665–72. 331. Jaeschke H, Cover C, Bajt ML. Role of caspases in acetaminophen-induced liver injury. Life Sci 2006; 78:1670–6. 332. Pierce RH, Franklin CC, Campbell JS, et al. Cell culture model for acetaminophen-induced hepatocyte death in vivo. Biochem Pharmacol 2002; 64:413–24. 333. Liu X, Kim CN, Yang J, Jemmerson R, Wang X. Induction of apoptotic program in cell-free extracts: requirement for dATP and cytochrome c. Cell 1996; 86:147–57. 334. Leist M, Single B, Castoldi AF, Kuhnle S, Nicotera P. Intracellular adenosine triphosphate (ATP) concentration: a switch in the decision between apoptosis and necrosis. J Exp Med 1997; 185:1481–6. 335. Kroemer G, El-Deiry WS, Golstein P, et al. Classification of cell death: recommendations of the Nomenclature Committee on Cell Death. Cell Death Differ 2005; 12(Suppl. 2):1463–7. 336. Goldring CE, Kitteringham NR, Elsby R, et al. Activation of hepatic Nrf2 in vivo by acetaminophen in CD-1 mice. Hepatology 2004; 39:1267–76. 337. Ruepp S, Boess F, Suter L, et al. Assessment of hepatotoxic liabilities by transcript profiling. Toxicol Appl Pharmacol 2005; 207:161–70. 338. Bae MA, Pie JE, Song BJ. Acetaminophen induces apoptosis of C6 glioma cells by activating the c-Jun NH(2)-terminal protein kinase-related cell death pathway. Mol Pharmacol 2001; 60:847–56. 339. Gunawan BK, Liu ZX, Han D, Hanawa N, Gaarde WA, Kaplowitz N. c-Jun N-terminal kinase plays a major role in murine acetaminophen hepatotoxicity. Gastroenterology 2006; 131:165–78. 340. Napirei M, Basnakian AG, Apostolov EO, Mannherz HG. Deoxyribonuclease 1 aggravates acetaminophen-induced liver necrosis in male CD-1 mice. Hepatology 2006; 43:297–305. 341. Jaeschke H, Bajt ML. Intracellular signaling mechanisms of acetaminophen-induced liver cell death. Toxicol Sci 2006; 89:31–41. 342. Bajt ML, Cover C, Lemasters JJ, Jaeschke H. Nuclear translocation of endonuclease G and apptosisinducing factor during acetaminophen hepatotoxicity. Society of Toxicology. San DIego: Oxford University Press, 2006. 343. Fujita T, Shirasawa T, Maruyama N. Expression and structure of senescence marker protein-30 (SMP30) and its biological significance. Mech Ageing Dev 1999; 107:271–80. 344. Fujita T. Senescence marker protein-30 (SMP30): structure and biological function. Biochem Biophys Res Commun 1999; 254:1–4. 345. Fujita T, Inoue H, Kitamura T, Sato N, Shimosawa T, Maruyama N. Senescence marker protein-30 (SMP30) rescues cell death by enhancing plasma membrane Ca(2C)-pumping activity in Hep G2 cells. Biochem Biophys Res Commun 1998; 250:374–80. 346. Matsuyama S, Kitamura T, Enomoto N, et al. Senescence marker protein-30 regulates Akt activity and contributes to cell survival in Hep G2 cells. Biochem Biophys Res Commun 2004; 321:386–90.

Acetaminophen-Induced Liver Disease

387

347. Kondo Y, Inai Y, Sato Y, et al. Senescence marker protein 30 functions as gluconolactonase in L-ascorbic acid biosynthesis, and its knockout mice are prone to scurvy. Proc Natl Acad Sci USA 2006; 103:5723–8. 348. Kondo Y, Ishigami A, Kubo S, et al. Senescence marker protein-30 is a unique enzyme that hydrolyzes diisopropyl phosphorofluoridate in the liver. FEBS Lett 2004; 570:57–62. 349. Hess DT, Matsumoto A, Kim SO, Marshall HE, Stamler JS. Protein S-nitrosylation: purview and parameters. Nat Rev Mol Cell Biol 2005; 6:150–66. 350. Hinson JA, Pike SL, Pumford NR, Mayeux PR. Nitrotyrosine–protein adducts in hepatic centrilobular areas following toxic doses of acetaminophen in mice. Chem Res Toxicol 1998; 11:604–7. 351. Kim SF, Huri DA, Snyder SH. Inducible nitric oxide synthase binds, S-nitrosylates, and activates cyclooxygenase-2. Science 2005; 310:1966–70. 352. Cover C, Fickert P, Knight TR, Fuchsbichler A, Farhood A, Trauner M, Jaeschke H. Pathophysiological role of poly(ADP-ribose) polymerase (PARP) activation during acetaminophen-induced liver cell necrosis in mice. Toxicol Sci 2005; 84:201–8. 353. Zong WX, Ditsworth D, Bauer DE, Wang ZQ, Thompson CB. Alkylating DNA damage stimulates a regulated form of necrotic cell death. Genes Dev 2004; 18:1272–82. 354. DeLeve LD, Wang X, Kaplowitz N, Shulman HM, Bart JA, van der Hoek A. Sinusoidal endothelial cells as a target for acetaminophen toxicity. Direct action versus requirement for hepatocyte activation in different mouse strains. Biochem Pharmacol 1997; 53:1339–45. 355. Kofman AV, Morgan G, Kirschenbaum A, Osbeck J, Hussain M, Swenson S, Theise ND. Dose- and time-dependent oval cell reaction in acetaminophen-induced murine liver injury. Hepatology 2005; 41:1252–61. 356. Donahower B, McCullough SS, Kurten RC, Lamps LW, Simpson PM, Hinson JA, James LP. Vascular endothelial growth factor and hepatocyte regeneration in acetaminophen toxicity. Am J Physiol Gastrointest Liver Physiol 2006; 291:102–9. 357. Klionsky DJ. The molecular machinery of autophagy: unanswered questions. J Cell Sci 2005; 118:7–18. 358. Levine B, Yuan J. Autophagy in cell death: an innocent convict? J Clin Invest 2005; 115:2679–88. 359. Shintani T, Klionsky DJ. Autophagy in health and disease: a double-edged sword. Science 2004; 306:990–5. 360. Liang XH, Jackson S, Seaman M, Brown K, Kempkes B, Hibshoosh H, Levine B. Induction of autophagy and inhibition of tumorigenesis by beclin 1. Nature 1999; 402:672–6. 361. Pattingre S, Tassa A, Qu X, et al. Bcl-2 antiapoptotic proteins inhibit beclin 1-dependent autophagy. Cell 2005; 122:927–39. 362. Shimizu S, Kanaseki T, Mizushima N, Mizuta T, Arakawa-Kobayashi S, Thompson CB, Tsujimoto Y. Role of Bcl-2 family proteins in a non-apoptotic programmed cell death dependent on autophagy genes. Nat Cell Biol 2004; 6:1221–8. 363. Boya P, Gonzalez-Polo RA, Casares N, et al. Inhibition of macroautophagy triggers apoptosis. Mol Cell Biol 2005; 25:1025–40. 364. Edinger AL, Thompson CB. Death by design: apoptosis, necrosis and autophagy. Curr Opin Cell Biol 2004; 16:663–9. 365. Elmore SP, Qian T, Grissom SF, Lemasters JJ. The mitochondrial permeability transition initiates autophagy in rat hepatocytes. FASEB J 2001; 15:2286–7. 366. Lemasters JJ, Qian T, He L, Kim JS, Elmore SP, Cascio WE, Brenner DA. Role of mitochondrial inner membrane permeabilization in necrotic cell death, apoptosis, and autophagy. Antioxid Redox Signal 2002; 4:769–81. 367. Yu L, Wan F, Dutta S, et al. Autophagic programmed cell death by selective catalase degradation. Proc Natl Acad Sci USA 2006; 103:4952–7. 368. Massey AC, Kaushik S, Sovak G, Kiffin R, Cuervo AM. Consequences of the selective blockage of chaperone-mediated autophagy. Proc Natl Acad Sci USA 2006; 103:5805–10. 369. Snapp EL, Hegde RS, Francolini M, et al. Formation of stacked ER cisternae by low affinity protein interactions. J Cell Biol 2003; 163:257–69. 370. Davern TJ, II, James LP, Hinson JA, et al. Measurement of serum acetaminophen–protein adducts in patients with acute liver failure. Gastroenterology 2006; 130:687–94. 371. Limaye PB, Apte UM, Shankar K, Bucci TJ, Warbritton A, Mehendale HM. Calpain released from dying hepatocytes mediates progression of acute liver injury induced by model hepatotoxicants. Toxicol Appl Pharmacol 2003; 191:211–26. 372. Dambach DM, Durham SK, Laskin JD, Laskin DL. Distinct roles of NF-kappaB p50 in the regulation of acetaminophen-induced inflammatory mediator production and hepatotoxicity. Toxicol Appl Pharmacol 2006; 211:157–65. 373. Dambach DM, Watson LM, Gray KR, Durham SK, Laskin DL. Role of CCR2 in macrophage migration into the liver during acetaminophen-induced hepatotoxicity in the mouse. Hepatology 2002; 35:1093–103.

388

Nelson and Bruschi

374. Marshall E. Clinical medicine. Accident prompts a closer look at antibody trials. Science 2006; 312:172. 375. James MJ, Cleland LG. Cyclooxygenase-2 inhibitors: what went wrong? Curr Opin Clin Nutr Metab Care 2006; 9:89–94. 376. Blazka ME, Germolec DR, Simeonova P, Bruccoleri A, Pennypacker KR, Luster MI. Acetaminopheninduced hepatotoxicity is associated with early changes in NF-kB and NF-IL6 DNA binding activity. J Inflamm 1995; 47:138–50. 377. Blazka ME, Wilmer JL, Holladay SD, Wilson RE, Luster MI. Role of proinflammatory cytokines in acetaminophen hepatotoxicity. Toxicol Appl Pharmacol 1995; 133:43–52. 378. Ishida Y, Kondo T, Tsuneyama K, Lu P, Takayasu T, Mukaida N. The pathogenic roles of tumor necrosis factor receptor p55 in acetaminophen-induced liver injury in mice. J Leukoc Biol 2004; 75:59–67. 379. Gardner CR, Laskin JD, Dambach DM, et al. Exaggerated hepatotoxicity of acetaminophen in mice lacking tumor necrosis factor receptor-1. Potential role of inflammatory mediators. Toxicol Appl Pharmacol 2003; 192:119–30. 380. Chiu H, Gardner CR, Dambach DM, Brittingham JA, Durham SK, Laskin JD, Laskin DL. Role of p55 tumor necrosis factor receptor 1 in acetaminophen-induced antioxidant defense. Am J Physiol Gastrointest Liver Physiol 2003; 285:G959–66. 381. Bernal W, Donaldson P, Underhill J, Wendon J, Williams R. Tumor necrosis factor genomic polymorphism and outcome of acetaminophen (paracetamol)-induced acute liver failure. J Hepatol 1998; 29:53–9. 382. Boess F, Bopst M, Althaus R, Polsky S, Cohen SD, Eugster HP, Boelsterli UA. Acetaminophen hepatotoxicity in tumor necrosis factor/lymphotoxin-alpha gene knockout mice. Hepatology 1998; 27:1021–9. 383. Blazka ME, Elwell MR, Holladay SD, Wilson RE, Luster MI. Histopathology of acetaminopheninduced liver changes: role of interleukin 1 alpha and tumor necrosis factor alpha. Toxicol Pathol 1996; 24:181–9. 384. Bourdi M, Masubuchi Y, Reilly TP, et al. Protection against acetaminophen-induced liver injury and lethality by interleukin 10: role of inducible nitric oxide synthase. Hepatology 2002; 35:289–98. 385. Hogaboam CM, Simpson KJ, Chensue SW, et al. Macrophage inflammatory protein-2 gene therapy attenuates adenovirus- and acetaminophen-mediated hepatic injury. Gene Ther 1999; 6:573–84. 386. Hogaboam CM, Bone-Larson CL, Steinhauser ML, et al. Novel CXCR2-dependent liver regenerative qualities of ELR-containing CXC chemokines. FASEB J 1999; 13:1565–74. 387. Liu J, Li C, Waalkes MP, Clark J, Myers P, Saavedra JE, Keefer LK. The nitric oxide donor, V-PYRRO/NO, protects against acetaminophen-induced hepatotoxicity in mice. Hepatology 2003; 37:324–33. 388. Gardner CR, Laskin JD, Dambach DM, et al. Reduced hepatotoxicity of acetaminophen in mice lacking inducible nitric oxide synthase: potential role of tumor necrosis factor-alpha and interleukin10. Toxicol Appl Pharmacol 2002; 184:27–36. 389. Hinson JA, Bucci TJ, Irwin LK, Michael SL, Mayeux PR. Effect of inhibitors of nitric oxide synthase on acetaminophen-induced hepatotoxicity in mice. Nitric Oxide 2002; 6:160–7. 390. Michael SL, Mayeux PR, Bucci TJ, Warbritton AR, Irwin LK, Pumford NR, Hinson JA. Acetaminophen-induced hepatotoxicity in mice lacking inducible nitric oxide synthase activity. Nitric Oxide 2001; 5:432–41. 391. Zhang H, Cook J, Nickel J, Yu R, Stecker K, Myers K, Dean NM. Reduction of liver Fas expression by an antisense oligonucleotide protects mice from fulminant hepatitis. Nat Biotechnol 2000; 18:862–7. 392. Boulares AH, Giardina C, Inan MS, Khairallah EA, Cohen SD. Acetaminophen inhibits NF-kappaB activation by interfering with the oxidant signal in murine Hepa 1-6 cells. Toxicol Sci 2000; 55:370–5. 393. Joshi GP. NCX-701. NicOx. Curr Opin Investig Drugs 2004; 5:755–9. 394. Fiorucci S, Antonelli E, Mencarelli A, et al. A NO-releasing derivative of acetaminophen spares the liver by acting at several checkpoints in the fas pathway. Br J Pharmacol 2002; 135:589–99. 395. Moore PK, Marshall M. Nitric oxide releasing acetaminophen (nitroacetaminophen). Dig Liver Dis 2003; 35(Suppl. 2):S49–60. 396. Lyons L, Studdiford JS, Sommaripa AM. Treatment of acetaminophen overdosage with N-acetylcysteine. N Engl J Med 1977; 296:174–5. 397. Piperno E, Berssenbruegge DA. Reversal of experimental paracetamol toxicosis with N-acetylcysteine. Lancet 1976; 2:738–9.

19

Acetaminophen: Pathology and Clinical Presentation of Hepatotoxicity William M. Lee and George Ostapowicz

University of Texas Southwestern Medical Center, Dallas, Texas, U.S.A.

INTRODUCTION Acetaminophen (called paracetamol outside the United States) is a popular and widely used analgesic and antipyretic agent. First synthesized in 1893, it was introduced for prescription use in the United States in 1955 and approved for over-the-counter use in 1960 (1). It is frequently combined with codeine or other analgesic agents, decongestants, and antihistamines. Over 300 different preparations are now available in the United States with more than one billion pills sold annually. Acetaminophen’s popularity has in part arisen from its apparent lack of side effects. Unlike aspirin and other nonsteroidal anti-inflammatory drugs, it does not cause gastric inflammation, ulcers, or coronary ischemia. Although it is remarkably safe when used at usual therapeutic doses, it has a relatively narrow therapeutic window. Acetaminophen overdose is the leading cause for calls to Poison Control Centers (O100,000/yr), and accounts for more than 56,000 emergency room visits, 2600 hospitalizations and an estimated 458 deaths due to acute liver failure (ALF) each year (1). Data from the U.S. Acute Liver Failure Study Group registry of more than 1100 patients with acute liver failure (ALF) from across the United States, suggests that acetaminophen poisoning alone currently constitutes nearly 50% of all ALF [Fig. 1; (2)]. Available in single or combination products, acetaminophen produces more than a billion dollars in annual sales and is heavily marketed for its safety compared to nonsteroidal analgesics. By enabling self-diagnosis and treatment of minor aches and pains, its benefits are said by the United States Food and Drug Administration (FDA) to outweigh its risks (3). Reports of fatal and nonfatal hepatic necrosis following suicide attempts first began to appear in the mid-1960s (4,5) in Great Britain and in the mid-1970s in the United States (6). Despite a recent decline in incidence in the United Kingdom, the result of limiting package size, acetaminophen self-poisoning remains a popular means of attempting suicide in Great Britain (7,8). More recently, it has also become evident that even therapeutic doses can be hepatotoxic in some individuals, especially in the presence of chronic alcohol consumption and fasting (9–11). An important element of the problem is the public’s perception of acetaminophen as a safe drug and the lack of awareness of the potential dangers of its ingestion in just above therapeutic doses. The pervasive inclusion of acetaminophen in a variety of products (narcotic analgesics, liquid flu medications, headache combinations, sedatives, etc.) results in many patients inadvertently consuming more than one acetaminophen-containing preparation, simultaneously. Narcotic–acetaminophen combinations are found in a large proportion (62%) of unintentional overdoses leading to ALF (3). This fact suggests that patients with chronic pain become addicted to the narcotic, continuing to increase the dose until a threshold level is reached, after which significant toxicity ensues. It seems likely that patients can develop tolerance to increasing acetaminophen doses, based on clinical experience and experimental studies (12). Concomitant use of other substances (including cocaine or marijuana) is often observed. This may cloud the sensorium and increase risky behavior such as repeated dosing. Unlike most other causes of ALF, however, timely use of the antidote N-acetylcysteine (NAC) will prevent its development, or at least lessen the severity of hepatic damage. Liver The authors of this chapter have relationships with the following corporations: consultation agreements with Astra Zeneca, Lilly. Research Support from BMS, Schering, Vertex, and Roche.

390

Lee and Ostapowicz

Etiology of ALF in the USA: Adult Registry (n = 1033) 46% More than half of all US ALF is drug-related

12%

151

8 Pregnancy

50 Indeter

9

Other

17

Budd-Chiari

43 Ischemic

55 Autoimm

31 Hep A

Hep B

75

Wilson's

119

Drug

ACM

n 480 475 450 420 390 360 330 300 270 240 210 180 150 120 90 60 30 0

FIGURE 1 Number of patients enrolled in the Acute Liver Failure Study by etiology as determined by the site investigator using standard definitions. Acetaminophen poisoning accounted for nearly 50% of cases over an eight year period 1998 2005, with an increase on an annual basis from 28% in 1998 to 51% over this period.

damage results from the metabolism of acetaminophen by the cytochrome P450 (CYP) system and the production of the highly reactive metabolite N-acetyl-benzoquinone-imine (NAPQI). This is discussed in detail in another chapter. TOXIC DOSES The recommended maximum “safe” dose of acetaminophen that can be ingested over 24 hours is 4 g in adults and 60 mg/kg in children. While acetaminophen is the classical example of a dose-dependent hepatotoxic drug, there is no definite threshold dose for hepatic injury. The amount of acetaminophen ingested as a single dose required to produce injury is relatively variable. Single doses of acetaminophen exceeding 7 to 10 g in adults or 150 mg/kg body weight in children are enough to cause significant hepatocellular necrosis; however, this is not inevitable. Severe liver injury, defined as alanine aminotransferase (ALT) or aspartate aminotransferase (AST) greater than 1000 IU/L, or fatal cases usually involve doses of at least 15 to 25 g (13). Ingested doses have exceeded 15 g (O200 mg/kg) in 80% of serious and fatal cases (14). As acetaminophen metabolism and susceptibility to toxicity differ between individuals, survival is possible even after ingestion of massive doses as large as 75 g. Among subjects with significant acetaminophen overdose who did not undergo treatment, severe liver injury was reported in only 20%, and among those with severe liver injury, the mortality was 20% (13). On the other hand, daily doses as low as 2 to 6 g have been associated with fatal hepatotoxicity in heavy drinkers (10). It must be remembered, however, that calculation of an accurate ingested dose has often been made difficult by the failure of patients to provide exact information, either because of intoxication or drowsiness caused by ingested substances (including narcotic or antihistamine containing compound formulations), or through lack of patient cooperation. Early vomiting of the ingested dose will also interfere with accurately determining the real amount ingested. RISK FACTORS FOR ACETAMINOPHEN-INDUCED HEPATOTOXICITY While the dose of acetaminophen ingested is clearly important in the development of hepatotoxicity, a number of other risk factors predispose to liver damage. Children seem to be relatively resistant to acetaminophen hepatotoxicity (15), but whether this is related to vomiting part of the ingested dose, or to biological resistance, or both is unclear. Children have a different relative importance of the metabolic pathways involved. Their lower ratio of glucuronidation to sulfation, may protect them against liver damage (16). A number of studies have now established that chronic alcohol use increases the individual’s susceptibility to acetaminophen-induced hepatotoxicity (9–11,17–23). The mechanisms appear to include both the induction of the CYP system and glutathione depletion. Results suggesting enhanced acetaminophen metabolism have been reported; disappearance

Acetaminophen-Induced Hepatotoxicity

391

of acetaminophen from plasma after ingestion of 1 g was faster in chronic alcoholics than in normal individuals (22). This has been confirmed by studies that show competition of the acetaminophen parent compound for cytochrome P450 2E1 (CYP2E1) during ingestion of alcohol but induction of the enzyme that forms the toxic intermediate in the hours following alcohol ingestion (24). The alcohol consumption threshold that predisposes to hepatic injury is uncertain, but may be relatively low. A retrospective analysis of patients with severe acetaminophen-induced liver injury found a higher mortality in men consuming more than 24 g alcohol/day and women consuming more than 16 g/day compared to those who drank less (23). Hepatic injury may occur even at therapeutic doses of acetaminophen in alcoholics. A study assessing therapeutic misadventures in a group of individuals, most of whom drank more than 60 g alcohol/day, found that 60% had ingested less than 6 g acetaminophen/day and 40% less than 4 g/day (11). Of these patients, 95% developed transaminase elevations greater than 1000 IU/L and 18% died. Other studies have reported similar findings with daily acetaminophen doses ranging from 2.6 to 16.5 g/day (9,16). The association between chronic alcohol ingestion and enhanced acetaminophen hepatotoxicity has been questioned however (25). Fasting or malnourished patients have an increased susceptibility to acetaminophen hepatotoxicity presumed due to depressed levels of hepatic glutathione and induction of CYP2E1 (9). While the chronically malnourished alcoholic is a classic example of this scenario, the average individual may become at risk after a flu-like illness if nausea and vomiting results in decreased food intake. Chronic cardiopulmonary insufficiency has also been reported to predispose to liver damage after short-term ingestion of therapeutic doses of acetaminophen (!9 g over three days) (26). Consequent drug metabolism studies indicated that this patient had decreased rates of hepatic metabolism of acetaminophen to its primary nontoxic metabolites. It was speculated that limited hepatic blood flow and nutrient supply, resulting in reduced glutathione levels, might have played a role. Concurrent use of other medications such as phenobarbital (27), phenytoin (28,29), isoniazid (30,31), and zidovudine (32) is also a risk factor for hepatotoxicity. These agents induce CYP or compete with glucuronidation pathways resulting in the increased production of NAPQI, the reactive intermediate which must be detoxified by glutathione. The presence of underlying acute or chronic liver disease or cirrhosis does not in and of itself predispose to acetaminophen hepatotoxicity. However, the clinical presentation may be more severe in persons with underlying liver damage because of limited hepatic reserve. SUICIDAL POISONING Historically, the majority of cases of acetaminophen-induced hepatic damage have occurred as a result of a single large ingestion taken by an individual attempting suicide, or as a parasuicidal gesture—“a cry for help.” Ingestion of moderate to large amounts of alcohol around the time of the overdose is common. Such persons often take the drug on impulse and most do not want to actually die. They later regret their actions when they present to the emergency room to receive appropriate treatment. Fortunately, most patients do seek help in the appropriate time frame to benefit from receiving the antidote, NAC. UNINTENTIONAL POISONING Significant hepatic injury and death have also been reported in persons taking acetaminophen with therapeutic intent (3,33–36). These unintentional poisonings, termed in the past “therapeutic misadventures,” were initially associated with alcohol use or abuse but more recently have been shown to encompass a wider group of settings. In contrast to suicidal poisoning, these individuals ingest smaller amounts of acetaminophen for periods ranging from a few days to a few weeks in the setting of an acute or subacute painful illness or condition, including febrile illnesses with severe myalgias, dental or traumatic pain, headache, acute exacerbation of chronic back pain, postsurgical pain, pancreatitis, and even hangover. The typical presentation involves the use of higher-than-recommended doses, with median dose ingested of 7.5 g over two to three days. Nearly 10% of patients claimed to have taken less than 4 g/day (3). While

392

Lee and Ostapowicz

TABLE 1 Characteristics of Accidental and Suicidal Acetaminophen Ingestions Feature Presentation Dose ingested Association with alcohol abuse Blood acetaminophen levels Key to diagnosis N-acetylcysteine effective Severity Mortality Typical hospital stay

Accidental

Suicidal

Late Small Frequent Seldom elevated High aminotransferases, history Usually of value O60% Severe High Long

Early Large Present but less frequent Usually elevated History, toxic blood levels Definitely effective Occasionally (20%) severe Low Short

Source: Adapted from Ref. 33.

most ingestions occur over less than seven days, up to one-third of patients have reported using acetaminophen preparations for at least 30 days (3). This scenario is typically seen in patients dealing with chronic pain, many of whom are significant alcohol users. A summary of the clinical characteristics of accidental and suicidal poisonings is given in Table 1. Patients with unintentional overdoses are frequently exposed to multiple preparations under different trade names without realizing their combined toxic potential. In the most recent study from the U.S. ALF Study, 38% of unintentional patients reported taking more than one acetaminophen preparation, while 63% were taking a narcotic combination (Table 2). Because they are unaware of the self-harm they have inflicted, they present to hospital late, after symptomatic hepatic injury has occurred and this is associated with increased mortality vis-avis suicidal patients. Patients who develop jaundice on the background of significant alcohol use may be misdiagnosed as having alcoholic hepatitis (37). The key to diagnosis is the extreme aminotransferase elevations in the presence of relatively low bilirubin levels. These findings are counter to those seen with idiosyncratic drug reactions or alcoholic hepatitis where the aminotransferase levels are usually under 1000 IU/L (under 300 for alcoholic hepatitis) and the bilirubin levels more than 20 mg/dL, suggesting a more subacute course. However, aminotransferase levels greater than 5000 IU/L may occasionally be seen in severe ischemic injury and rarely seen in viral hepatitis. A high index of suspicion is warranted as even patients who present late may benefit to some extent from treatment with NAC. Recently, accidental poisoning has also been reported in children ranging from five weeks to over 12 years of age (38–44). The typical setting is the administration of too frequent doses over a few days by parents or on occasion by hospital staff, to children who have a febrile acute or subacute illness. As in the adult cases, supratherapeutic doses (greater than 150 mg/kg per 24 hr) are often used, although smaller doses have frequently been reported when used over several days. Doses used have ranged from 20 to over 600 mg/kg per 24 hr taken over periods of one day to six weeks (40,42). The association of a febrile illness with or without vomiting, resulting in a decreased oral intake, is likely to have made the children acutely malnourished TABLE 2 Suicidal Versus Unintentional APAP Cases nZ253 Female (%) ACM dose (g) Dose per day Coma (% R3) ALT (IU/L) Spont surv (%) Antidepressant History of depression Narcotic cpd (%) Multiple preps

Intentional (nZ122)

Unintentional (nZ131)

p-Value

74 25 25 39 5326 66 38 45 18 5

73 NS 20 NS 7.5 55 3129 64 NS 37 NS 24 63 38

NS NS 0.001 0.026 0.001 NS NS 0.001 0.001 0.001

Abbreviation: ACM, aclacinomycin A; ALT, alanine aminotransferase; APAP, N-acetyl-para-aminophenol; NS, not significant.

Acetaminophen-Induced Hepatotoxicity

393

and predisposed them to acetaminophen hepatotoxicity at lower-than-expected doses. In children with significant acetaminophen hepatotoxicity, progression to ALF has been reported in over 20% of cases with a significant mortality rate (39,40). Data from the Pediatric Acute Liver Failure Study suggests that most recent pediatric cases of acetaminophen poisoning resulting in liver failure are due to intentional overdoses in teenagers (45). Dosing errors most frequently made by parents result from a variety of mistakes, including: adult preparations used instead of infant or children’s preparations; dosing instructions misread or misunderstood; or extra doses given when symptoms persist despite recommended dosing (40). Failure to recognize the potential danger and the resulting delays in presentation and management further predispose to hepatotoxicity. It remains worrisome that some children appear to develop significant hepatic injury after receiving what are thought to be safe, though repetitive, doses of acetaminophen. Dose under-calculation cannot be discounted in at least some of these cases. The associated underlying febrile illness in addition to causing acute malnutrition also raises the possibility of viral cofactors playing a role in the hepatic damage. The typical course and outcome of this scenario is, however, more in keeping with acetaminophen-induced hepatotoxicity rather than fulminant viral or indeterminate hepatitis (46). In summary, significant hepatocellular injury from acetaminophen in children is relatively uncommon, occurring in less than 10% of cases in which potentially toxic doses of drug had been ingested (44). Repetitive doses of acetaminophen, even when in the therapeutic range, need to be given with caution to children suffering from a febrile illness associated with nausea and anorexia.

CHRONIC TOXICITY Possible chronic hepatotoxicity has been reported in one study of a person who had ingested therapeutic or near-therapeutic doses (2–6 g/day) over a prolonged period (47). Significant histological damage, including fibrosis, was found in a male who had ingested 4 g acetaminophen/day for one year. On stopping the drug, the elevated aminotransferase levels returned to near normal. Rechallenge with a dose of 1325 mg was associated with a rise in enzymes within 18 hours. This scenario is similar to the classic therapeutic misadventure, although the duration of ingestion was significantly longer. While some persons may be predisposed to developing liver damage, it remains unclear why the damage did not occur earlier or whether other drugs, toxins, or intercurrent illnesses may have made the liver more susceptible to acetaminophen-induced hepatotoxicity or, indeed, were responsible for some of the injury. Chronic acetaminophen hepatotoxicity remains unproven; no recent reports have been forthcoming. In contrast, some individuals ingest otherwise toxic amounts of acetaminophen over prolonged periods without developing any obvious signs of liver injury. Doses ranging from 3 g to 65 g/day have been reported (12,47–49). A recent study demonstrated that repeat exposure to incremental acetaminophen doses provides protection in an animal model, as a possible explanation for this scenario (12).

EPIDEMIOLOGY OF ACETAMINOPHEN POISONING Acetaminophen poisoning is predominantly a Western phenomenon, although, even here, the incidence varies between countries. Intentional overdose with acetaminophen is one of the most popular means of attempting suicide in the United Kingdom (50). It is of further concern that acetaminophen overdoses and toxicity have increased substantially in the United States (1–3). Over 110,000 cases of acetaminophen poisoning are reported per year in the United States (51) and an estimated 70,000 cases in the United Kingdom (52). Some 150 to 200 deaths from acetaminophen are thought to occur annually in the United Kingdom compared to approximately 450 in the United States. Case fatality rates have been estimated to be 0.4% in the United Kingdom (8), and 0.1% in the United States (51) and France (8). Unintentional overdoses have

394

Lee and Ostapowicz

been reported predominantly from the United States and comprise 50% of cases of severe hepatic injury; they have also been reported from Australia. While a number of social and psychological factors interact in predisposing individuals to take overdoses, there is a clear correlation between acetaminophen availability and sales, and the incidence of overdose (8,53). Countries such as Australia and the United Kingdom that have put in place limits on the number of acetaminophen tablets available per packet and mandated blister packaging have lowered significantly the incidence of severe hepatic injury. Legislation in effect in the United Kingdom since 1998 has made the impulsive large overdose less likely (54–59), although the drug may still be prescribed in larger quantities. Comparison of the incidence of serious events related to paracetamol prior to these laws with comparable periods since the legislation indicate a decline in hospital admissions (10%), deaths (19%), and liver transplants (56%), although a few studies are less clear. One large referral hospital in London, however, saw a 74% decline in paracetamol-related deaths or transplants comparing the pre- and post-legislation periods (59,60). Despite these efforts elsewhere, there has been little will in the United States for change. Widespread public education campaigns warning of the dangers of acetaminophen and chronic alcohol use are needed, in addition to specific medication label warnings, to decrease the prevalence of unintentional cases but this has not taken place in the United States. To date, FDA has not considered unbundling of the narcotic compounds, use of blister packs or limitation of package size. CLINICAL FEATURES The clinical characteristics of acetaminophen hepatotoxicity can be divided into four stages (61). Stage 1 occurs within a few hours of ingestion. It consists of acute gastrointestinal symptoms including anorexia, abdominal pain, nausea, and vomiting as well as general malaise, diaphoresis, and occasional vascular collapse. Drowsiness is more likely to occur with concomitant ingestion of narcotics or sedatives. Some patients may experience little or no symptoms. Stage 2 is seen 24 to 48 hours after ingestion. During this period, the patients are relatively well and may, in fact, be totally asymptomatic. Stage 3 is seen only in those patients who develop significant liver injury. Clinical signs of liver injury are usually present two to five days after ingestion. In addition to the gastrointestinal symptoms outlined above, patients develop lethargy, dark urine, jaundice, and hepatic tenderness. In more severe cases, overt ALF is seen with the development of hepatic encephalopathy. While this generally occurs three to four days after poisoning, occasionally it may be delayed for up to six days. ALF may be accompanied by renal failure, multiple organ failure, sepsis, and cerebral edema leading to death. Cerebral edema is particularly common in acetaminophen-related ALF when compared with other causes. Stage 4, the recovery phase, is usually seen 5 to 10 days after ingestion in those who survive the hepatic insult. As full recovery takes place, all symptoms gradually resolve. In those with more severe hepatic injury, jaundice, and renal failure may be more prolonged. Permanent liver damage is seen very infrequently after acetaminophen overdose (62) and liver function returns to normal in those without underlying liver disease. BIOCHEMICAL AND OTHER LABORATORY FEATURES A summary of laboratory values seen in acetaminophen hepatotoxicity is given in Table 2. Patients who are treated with NAC within 8 to 12 hours of ingestion will often display few or no laboratory abnormalities, even in the presence of a “high risk” serum acetaminophen level. Those treated later, or not at all, develop a number of typical biochemical and hematological abnormalities. These laboratory features in themselves are, however, not diagnostic of acetaminophen hepatotoxicity. Serum levels of ALT and AST, which are markers of hepatocellular injury, can become markedly elevated, occasionally reaching over 10,000 IU/L (3). Aminotransferase levels generally begin to increase 24 to 36 hours after ingestion, occasionally as early as eight hours, and peak at day two or three (33). Levels usually return to normal, or near normal, within 7 to 10 days of the insult. Aminotransferase levels of the magnitude seen in

395

Acetaminophen-Induced Hepatotoxicity

TABLE 3 Biochemical Characteristics of Acetaminophen Toxicity Laboratory value AST/ALT

Minor to moderate toxicity

INR

Normal to mildly elevated (up to 1000 IU/L) Normal to moderately elevated (up to 100 mol/L or 5.9 mg/dL) Normal to moderately prolonged up to 2

Creatinine Acid-base balance

Usually normal Respiratory alkalosis, pH usually !7.60

Platelets Blood glucose

Usually normal Usually normal or slightly decreased

Bilirubin

Severe toxicity Often very elevated; O3000 IU/L; typically peak at day 2 to 3 after ingestion Continues to increase, even after transaminases improve Very prolonged; may be O10 elevation out of proportion to level of encephalopathy Acute renal failure occurs frequently in patients with ALF Metabolic acidosis; admission arterial pH!7.30 is a poor sign Thrombocytopenia (!100!109) common in patients ALF Severe hypoglycemia may be seen in ALF

Abbreviations: ALF, acute liver failure; AST, aspartate aminotransferase; ALT, alanine aminotransferase; INR, International normalized ratio. Source: Adapted from Ref. 64.

acetaminophen hepatotoxicity are rarely seen in other causes of hepatic injury with the exception of ischemic hepatitis or heat stroke and rare cases of viral hepatitis. Serum bilirubin elevation occurs in more severe cases. Typically, bilirubin levels peak later than transaminases and may continue to increase after ALT/AST levels have begun to improve. Maximum bilirubin elevations are lower than that seen in viral hepatitis or idiosyncratic drug reactions. Very high bilirubin levels may be seen in the presence of acute renal failure. Hepatic production of clotting factors becomes reduced with a resulting prolongation in the prothrombin time and International normalized ratio (INR) (63). In severe cases, the INR may be greater than 10. The INR prolongation may be evident before aminotransferases are increased, and is often out of proportion to the level of encephalopathy seen in other forms of ALF (Table 3). Serum creatinine may be elevated in cases of severe poisoning, presumably from direct acetaminophen nephrotoxicity (3). As with hepatotoxicity, chronic alcohol ingestion has been reported to predispose to the development of renal damage and failure (65–67). Acute renal failure occurs in approximately 50% of patients with ALF with grade III or IV encephalopathy as part of the multisystem failure syndrome (67). Acid-base disturbances are frequently seen (68). Phosphate levels have been noted to be remarkably low in most patients with ALF and particularly in acetaminophen cases. Reasons for this have not been forthcoming but include starvation, intracellular shifts of phosphate to supply ATP for cellular regeneration or increased phosphaturia (69,70). High plasma or serum phosphate levels are considered an unfavorable prognostic sign probably reflecting renal failure superimposed (71,72) these findings could not be confirmed in another study (73). While respiratory alkalosis secondary to hyperventilation occurs in the milder cases, lactic acidosis is a poor prognostic sign in hepatic encephalopathy (74). Severe hypoglycemia requiring concentrated glucose infusions may be seen in ALF. Moderate thrombocytopenia can also occur with hepatic failure. Acetaminophen levels are measured in serum and are often still elevated in patients with suicidal overdoses, but are infrequently detectable at time of presentation in unintentional overdoses, because of their later presentation and chronic dosing pattern (3). Acetaminophen levels may be falsely elevated in the presence of high bilirubin levels as described in the section concerning diagnosis. A new test, measurement of acetaminophen adducts in serum of patients with hepatotoxicity, represents a step forward in our understanding and ability to diagnose this condition. The highly reactive intermediate NAPQI is bound to cell proteins via sulfhydryl groups to initiate the toxic process. When necrosis of hepatocytes occurs, acetaminopencysteine adducts appear in plasma and can be quantitated via high-pressure liquid chromatography with electrochemical detection (75–77). A recent clinical study showed that 100% of known suicidal overdoses contained significant levels of adducts while none were found in patients with other forms of ALF or in those with overdoses who entered the hospital rapidly and received the antidote, NAC, in a timely fashion (Fig. 2). However, adducts also were present in similar quantities in 19% of patients with indeterminate ALF, signifying

396

Lee and Ostapowicz

nmol APAP-CYS / mg protein

3.0 2.5 2.0 1.5 1.0 0.5 0.0 A

B

C Patient group

D

E

FIGURE 2 Serum levels of acetaminophen-CYS adducts in patient groups. (A) Patients with ALF secondary to known acetaminophen overdose. (B) Patients with ALF due to non-acetaminophen causes. (C) Patients with acetaminophen overdose but no ALF. (D) Patients with ALF of indeterminate etiology and detectable serum adducts. (E) Patients with ALF of indeterminate etiology and negative adducts. The boxes represent the 25th to 75th interquartile range and the horizontal line represents the median. The extremes of the population are represented by the endmarks. Abbreviation: ALF, acute liver failure.

that this group had unrecognized acetaminophen toxicity, either because of patient lack of understanding, denial or encephalopathy precluding adequate history taking. Similar findings were observed in pediatric patients: 12.5% of indeterminate pediatric ALF patients demonstrated the presence of adducts, while 9% of patients with other causes also were adduct-positive, suggesting a role for acetaminophen in many cases where toxicity was not suspected (78). A more recent study has identified the presence of acetaminophen adducts in serum of 9 of 73 patients with ALF due to either hepatitis A or B. Eight of the nine admitted to taking therapeutic doses (only) and indeed had lower adduct levels than those with suicidal ingestions, suggesting that even therapeutic doses of acetaminophen may enhance liver injury in viral hepatitis under certain circumstances (79). Outcomes were worse in the acetaminophen ingesting group: 62% died or were transplanted compared to 26% of those who did not have detectable adducts. PATHOLOGICAL FEATURES The characteristic histological changes seen with acetaminophen poisoning are centrilobular (zone III) hepatic necrosis and sinusoidal congestion (80). In severe cases, submassive (bridging) or panacinar (massive) necrosis is seen. This pattern of necrosis reflects the role of CYP2E1, which is concentrated in this part of the hepatic acinus, as well as the fact that glutathione levels are lower in zone III. Inflammation is not a significant feature. On recovery, complete resolution without fibrosis occurs. In the kidney, necrosis of the proximal and distal tubules is the most prominent finding. Myocardial necrosis and pancreatitis have also been reported (81–85) and plasma troponin levels are elevated in at least 25% of patients (unpublished observations). ACETAMINOPHEN-INDUCED ACUTE LIVER FAILURE Significant hepatic injury, defined as an aminotransferase level greater than 1000 IU/L, is seen relatively frequently in acetaminophen poisoning if treated late or untreated. The percentage of patients who develop ALF is, however, actually quite small (33,86). In a series of 71 hospitalized patients from a county hospital in the United States, 14% developed ALF and 7% died (33); all but one patient was in the unintentional poisoning group. Series from Australia have reported even lower morbidity and mortality. ALF was seen in 7% of 306 hospitalized patients, with no

397

Acetaminophen-Induced Hepatotoxicity

TABLE 4 Etiology of ALF in Different Countries

U.K.: 1993 1994 (nZ342) France: 1972 1990 (nZ502) India: 1987 1993 (nZ423) Denmark: 1973 1990 (nZ160) U.S.A.: 1994 1996 (nZ295)

HAV (%)

HBV (%)

Drug reactions (%)

Acetaminophen (%)

Othera (%)

2

2

2

73

21

4

32

17

2

45

2

31

5

0

62

1

7

10

45

37

7

10

12

20

51

a Includes NANB, HEV, and miscellaneous. Abbreviations: ALF, acute liver failure; HAV, hepatitis A virus; HBV, hepatitis B virus; HEV, hepatitis E virus; NANB, non-A non-B hepatitis. Source: Adapted from Ref. 90.

deaths (87). Two more recent studies found ALF in 9% of 151 patients with one death (0.6%) (88), and only two deaths (0.2%) among 981 patients (89). Factors associated with more severe liver injury and the development of ALF included late presentation and treatment, ingestion of large amounts of acetaminophen, chronic alcohol ingestion, and accidental overdose. The epidemiology of acetaminophen-induced ALF varies widely in countries around the world (Table 4).

DIAGNOSIS The diagnosis is straightforward in those who present with a clear history of ingestion including the amount and time taken. Such a history is not always available, however. Individuals may have a depressed level of consciousness due to alcohol intoxication or coingestion of sedatives or other drugs. Others may be simply uncooperative. A high index of suspicion is required in all potential suicidal overdose cases. Acetaminophen poisoning must also be suspected in all persons with elevated transaminases, especially if the levels are greater than 1000 IU/L. Of interest, the indeterminate patients identified as having acetaminophen adducts in the study mentioned above had a median AST of 8800 IU/L and a median bilirubin of 6.5 mg/dL, while the remaining indeterminate cases had a median AST of 888 IU/L and a median bilirubin level of 22.4 mg/dL (77). It must be remembered that aminotransferases will still be in the normal range soon after ingestion. In those with unintentional poisoning, careful questioning regarding all medications taken is imperative, as these persons will not be aware of the potential implications of their analgesic use. The reported dose may be inaccurate owing to either underreporting or partial vomiting, so this cannot be relied on completely. Acetaminophen plasma levels checked at four hours or later after ingestion can be very useful in assessing the potential for significant hepatotoxicity, but have their limitations especially when the time interval between ingestion and measurement is uncertain. In some cases, repeating the measurement in another four hours, if this is within eight hours of ingestion, and calculation of the acetaminophen half-life may be of help. Those who have a half-life of less than four hours are unlikely to develop significant hepatotoxicity. As noted, acetaminophen levels may be normal in accidental poisoning because of both later presentations and smaller individual doses. Aminotransferase levels may be of more help in this setting as the unintentional cases that reach the threshold of ALF demonstrate similarly high AST and ALT levels. The clinical picture of acetaminophen poisoning reaching the threshold is particularly distinct when compared with idiosyncratic drug reactions producing ALF. The acetaminophen cases all are of very rapid onset and offset if the patient recovers with days between onset of jaundice of ca. 1 day versus 8 to 10 days for idiosyncratic drug reactions. As might be expected,

398

Lee and Ostapowicz

TABLE 5 Comparison of Different ALF Etiology Groups ACM (nZ407) Age (median) Sex (% F) Jaundice (days) (median) Coma 3/4 (%) ALT (median) Bili (median) Tx (%) Spontaneous survival (%) Overall survival (%)

Drug (nZ111)

Indeterminate (nZ131)

HepA/HepB (nZ29/69)

All others (nZ159)

36 74 0

42 67 8

38 57 8

48.0/41.0 48/48 3.0/6.0

42 78 6

52 4248 4.5 9 63

39 586 20.9 42 25

49 899 22.7 41 26

52/50 2622/1740 11.8/19.7 31/49 55/26

42 674 15.8 35 30

71

64

63

83/68

60

Abbreviation: ALT, alanine aminotransferase.

the aminotransferase levels of the idiosyncratic reactions are considerably lower and bilirubin levels higher as befits their subacute evolution (Table 5). The Matthew–Rumack nomogram (Fig. 3) (91), often used to decide on the need for NAC treatment, was developed after assessment of a group of untreated patients for hepatotoxicity, defined as ALT or AST above 1000 IU/L. Plasma acetaminophen level is plotted on the Y-axis and time from ingestion on the X-axis. When plotted above the line starting at a plasma level of 300 mg/L at 4 hours, there is a 90% chance of developing hepatotoxicity. Treatment is usually instituted when values fall above the standard treatment line starting at a plasma level of 200 mg/L at 4 hours. The nomogram is, however, of limited use in predicting hepatotoxicity under several circumstances. It is of benefit only in a single time-point ingestion, not in repeated doses as occurs in accidental poisoning. Also, the absorption of extended-release acetaminophen preparations may be more prolonged and thus the original “safe” values may be misleading (92). The nomogram cannot be safely applied in patients at high risk for acetaminophen hepatotoxicity. Consequently, a treatment line starting at a plasma level of 100 mg/L at four hours has been suggested for high-risk patients (93). Finally, the interval since 300 Plasma acetaminophen (mg/L)

90% chance hepatotoxicity 250

25

200 150 100

–3

0%

ch

an

ce

he

Standard treatment line pa

tot

ox ic

USA treatment line

ity

hepatotoxicity unlikely

Suggested treatment line for high-risk patients

50 0

0

4

8 12 Time post overdose (h)

16

20

FIGURE 3 Matthew Rumack nomogram for plasma acetaminophen after single acetaminophen ingestion. Plots in the area between the standard treatment line (solid) and the slashed line represent 25% to 30% risk of hepatotoxicity, while those above the slashed line represent a 90% risk of hepatotoxicity. Treatment with NAC is instituted when a result is above either the standard treatment line or the U.S. treatment line. It has been recommended that treatment be started at lower plasma levels for those predisposed to acetaminophen hepatotoxicity, hence the treatment line for high-risk patients. Abbreviation: NAC, N-acetylcysteine.

Acetaminophen-Induced Hepatotoxicity

399

ingestion may be uncertain or unknown. In any case, when doubt exists about the potential toxicity of the ingestion, treatment must be instituted. Acetaminophen levels are now often used as a toxin screen to determine whether there might be any acetaminophen ingested regardless of the history. However, low levels of acetaminophen may be confusing for two reasons: first, therapeutic doses will give detectable but not toxic levels that have little clinical significance; second, false positive acetaminophen values are observed in the presence of elevated levels of bilirubin, apparently due to interference with the colorimetric assay system (94–96). Certain assays are not affected, while the majority shows interference at levels of bilirubin above 10 mg/dL. Thus, caution should be used in interpreting low levels of acetaminophen found in the presence of high serum bilirubin. Patients with bilirubin levels above 10 mg/dL are very unlikely to have acute acetaminophen toxicity in any case and a search for other causes such as viral hepatitis or drug-induced liver injury is likely to be more fruitful.

TREATMENT The principles of treatment of acetaminophen poisoning are interruption of drug absorption, use of a specific antidote—NAC—and supportive care. Prompt treatment with NAC is, however, central to the successful management of acetaminophen toxicity. An effort to decrease acetaminophen absorption should be made unless it is clear that ingestion occurred more than 24 hours earlier. Gastric lavage can be useful in patients who present within four hours of ingestion; many, however, present later. Oral activated charcoal has been shown to reduce acetaminophen absorption by binding to it in the lumen of the stomach. It appears to be superior to both gastric lavage and ipecac (97). Charcoal is unlikely to interfere with the efficacy of oral NAC (98). NAC is the established antidote for acetaminophen poisoning. It is the N-acetylated derivative of the sulfhydryl amino acid L-cysteine and its sulfhydryl group is thought to be essential for its early antidote effects. Its most important action is the replenishment of glutathione stores, which inhibits the damaging effects of the breakdown product NAPQI (99,100). Other actions of NAC include vasodilatation, increased tissue oxygen uptake, antioxidant effects, and suppression of Tumor necrosis factor-alpha (101–104). Whether there is continued benefit of NAC once hepatic injury and ALF have developed is unclear but the relative safety of the product has resulted in widespread use in this situation. A number of different NAC regimens exist in clinical practice. Oral NAC is used in the United States but the intravenous form has recently been approved by the FDA. The oral regimen involves a loading dose of 140 mg/kg and is followed by maintenance doses of 70 mg/kg, every 4 hours for 72 hours (104). Total NAC dose given is 1330 mg/kg. Many patients have significant nausea with or without vomiting after acetaminophen poisoning, and the oral preparation of NAC with its strong unpleasant odor is difficult to ingest. Intravenous NAC has been available in the United States as Acetadotew since 2003. The standard intravenous protocol is given as an infusion in 5% dextrose over 20 hours (105). This includes a loading dose of 150 mg/kg given over 15 minutes to 1 hour, followed by 50 mg/kg over 4 hours, then 100 mg/kg over 16 hours. Total NAC dose given is 300 mg/kg and the volume of dextrose required is about 1800 mL. This protocol may be extended in patients presenting late with significant hepatotoxicity. Infusions of 150 mg/kg over 24 hours can be given for a further 24 to 48 hours. Another intravenous regimen involves the intermittent doses based on the oral version (106). Each dose is given over one hour, every four hours, for 48 hours. Total NAC dose given is 980 mg/kg. Oral versus intravenous regimens have never been compared directly. A recent meta-analysis of seven studies reported that prevention of hepatotoxicity was similar in both groups (107). The authors concluded that intravenous NAC may be preferable because of a shorter hospital stay, patient convenience, and concerns over the bioavailability of oral NAC in the presence of nausea and vomiting. While none of the original NAC studies were of optimal design, lacking randomization and using only historical controls, the use of NAC within 10 hours of ingestion is associated with a definite protective effect. Less than 10% of patients developed hepatic damage if given

400

Lee and Ostapowicz

NAC within 10 hours, and almost complete protection is seen within eight hours (104). In patients who received NAC 10 to 24 hours after ingestion, 26% to 63% developed hepatic damage with a mortality of 2% to 7% (104). This contrasts with the development of hepatic injury in more than 80% of historical controls. Some benefit is said to be seen even in individuals given NAC 72 hours after overdose with less frequent progression to grade III or IV encephalopathy and a lower mortality (108). Furthermore, a randomized controlled trial in patients with acetaminophen-induced ALF found that those given NAC were less likely to develop hypotension and cerebral edema, and had a higher rate of survival (109). NAC is a safe treatment. Adverse effects have been reported in about 5% to 14% of patients (110). They occur more frequently with intravenous infusions than oral regimens. Most represent a mild anaphylactoid reaction and include pruritus, mild urticaria, flushing, or wheezing. Rare serious adverse reactions including arrhythmias, angioedema, hypotension, and death have been reported. It is very likely that factors other than NAC played important roles in the development of these events. Treatment includes slowing or stopping the infusion and/or antihistamine or, rarely, corticosteroid injection. Invariably, the infusion is continued without further adverse effects. Most reactions occur early in the front-loaded infusion and giving the first dose over 60 minutes, instead of 15 minutes, decreases the incidence of reactions (89). In the patients who develop ALF, supportive treatment in an intensive care unit is of paramount importance (111,112). Supportive care includes the close monitoring and correction of blood glucose levels and electrolyte abnormalities. Any signs of sepsis are treated aggressively with broad-spectrum antibiotics until results of cultures are known. Some units use prophylactic antibiotics. Sedation or analgesics are generally contraindicated. Dialysis may be required in patients with acute renal failure who develop serious electrolyte disturbances or fluid overload. Signs of increased intracranial pressure are treated immediately with intravenous mannitol. Patients with grade III or IV encephalopathy are usually intubated and ventilated. Transfer to a specialist liver unit must be considered to allow for urgent liver transplantation in deteriorating individuals before irreversible cerebral edema or multisystem failure occurs. Unfortunately, many of these patients are not transplant candidates because of underlying psychiatric histories or substance abuse.

OUTCOME The outcome of acetaminophen poisoning is largely related to the absolute amount of drug taken and the time interval between ingestion and administration of the antidote. While the overall estimated case fatality rates are very small, ranging from 0.1% to 0.4%, this figure does not fully represent the morbidity and overall impact on both individuals and Western societies of acetaminophen poisoning. In those persons who develop ALF, the spontaneous survival of 65% is significantly better than in ALF from other etiologies (25%) (35). Only a small percentage undergo liver transplantation (w8%); however, significant numbers still die (27%) either on the transplant list or after exclusion from transplantation for psychosocial reasons. Those who undergo transplantation have only a 71% likelihood of early survival (three weeks) and are faced with lifelong immunosuppression for what was potentially a reversible condition. A number of subsequent suicides have been reported in this group. Clinical features on presentation have been found to be associated with a poor outcome, which include acidosis, severe coagulopathy, low plasma coagulation factors V and VIII, high creatinine and bilirubin, and the presence of grade III or IV encephalopathy. Accurate prognostic criteria are important to predict who is likely to require a liver transplant or, conversely, survive without the need for transplantation. The King’s College criteria from London (Table 6) are among the most commonly used (74). Fulfillment of the criteria was originally associated with an 80% probability to require liver transplantation. As these criteria were developed some years ago, they may not be totally applicable currently in units in other countries. A recent report concluded that fulfillment of King’s criteria usually predicts a poor outcome (good positive predictive value) but lack of fulfillment of the criteria does not predict survival (low negative predictive value) (113). Analysis of the U.S. Acute Liver Failure Study

Acetaminophen-Induced Hepatotoxicity

401

TABLE 6 Indications for Liver Transplantation in Acetaminophen-Induced ALF Patients Developed at King’s College Hospital, London King’s College Hospital Criteria pH!7.3 (irrespective of encephalopathy grade) or Prothrombin time O100 sec (INRO7.0) and serum creatinine O300 mmol/L (O3.4 mg/dL) in patients with grade III or IV encephalopathy Abbreviation: INR, international normalized ratio.

Group data (275 cases of acetaminophen toxicity), on the other hand, revealed a poor positive predictive value (63%) and negative predictive value (69%) with a sensitivity and specificity of 26% and 92%, respectively. In this study, the APACHE II score performed slightly better with positive and negative predictive values of 77% and 81% and a sensitivity and specificity of 68% and 87% (2). A recent meta-analysis comparing various prognostic criteria in patients with ALF due to acetaminophen found that the King’s College Hospital criteria outperformed other prognostic indicators but still had an average sensitivity of only 69% (107). Clearly, a need exists for the development of better prognostic criteria. CONCLUSION Acetaminophen is one of the most readily available and widely used analgesics. Although sideeffect-free and relatively safe at recommended doses, it has a narrow therapeutic window. Both suicidal overdose and the more recently recognized unintentional poisoning are major causes of hepatotoxicity in Western countries. Acetaminophen hepatotoxicity is the most common cause of ALF in both the United States and United Kingdom and is associated with a significant number of deaths. The timely use of NAC prevents significant hepatotoxicity in the majority of cases. More importantly, measures should be undertaken to prevent the large number of poisonings that continue to occur in the developed world. Limiting the availability of acetaminophen in countries such as the United States is central to these measures. Alerting the public to the potential dangers of acetaminophen use, especially with chronic alcohol ingestion, without emphasizing its suicidal potential, is another challenge. REFERENCES 1. Proceedings of the FDA NDAC meeting September 19, 2002. Testimony of Parivash Nourjah, Ph.D. www.fda.gov/ohrms/dockets/ac/cder02.htm# 2. Larson AM, Fontana RJ, Davern TJ, et al. Acetaminophen-induced acute liver failure: results of a United States multicenter, prospective study. Hepatology 2005; 42:1364–72. 3. Proceedings of the NDAC meeting: Introduction by Charles E. Ganley, M.D. www.fda.gov/ohrms/ dockets/ac/cder02.htm#Nonprescription Drugs 4. Davidson DG, Eastham WN. Acute liver necrosis following overdose of paracetamol. Br Med J 1966; 2:497–9. 5. Thompson JS, Prescott LF. Liver damage and impaired glucose tolerance after paracetamol overdose. Brit Med J 1966; 2:506–7. 6. McJunkin B, Barwick KW, Little WC, Winfield JB. Fatal massive hepatic necrosis following acetaminophen overdose. JAMA 1976; 236:1874–5. 7. Bernal W, Wendon J, Rela M, Heaton N, Williams R. Use and outcome of liver transplantation in acetaminophen-induced acute liver failure. Hepatology 1998; 27:1050–5. 8. Gunnell D, Hawton K, Murray V, et al. Use of paracetamol for suicide and non-fatal poisoning in the U.K. and France: are restrictions on availability justified? J Epidemiol Community Health 1997; 51:175–9. 9. Whitcomb DC, Block GD. Association of acetaminophen hepatotoxicity with fasting and alcohol use. JAMA 1994; 272:1845–50. 10. Denison H, Kaczynski J, Wallerstedt S. Paracetamol medication and alcohol abuse: a dangerous combination for the liver and the kidney. Scand J Gastroenterol 1987; 22:701–4. 11. Zimmermann HJ, Maddrey WC. Acetaminophen (paracetamol) hepatotoxicity with regular intake of alcohol: analysis of instances of therapeutic misadventure. Hepatology 1995; 22:767–73.

402

Lee and Ostapowicz

12. Shayiq RM, Roberts DW, Rothstein K, et al. Repeat exposure to incremental doses of acetaminophen provides protection against acetaminophen-induced lethality in mice: an explanation for high acetaminophen dosage in humans without hepatic injury. Hepatology 1999; 29:451–63. 13. Prescott LF, Critchley JA. The treatment of acetaminophen poisoning. Annu Rev Pharmacol Toxicol 1983; 23:87–101. 14. Hamlyn AN, Douglas AP, James O. The spectrum of paracetamol (acetaminophen) overdose: clinical and epidemiological studies. Postgrad Med J 1978; 54:400–4. 15. Rumack BH. Acetaminophen overdose in young children. Am J Dis Child 1984; 138:428–33. 16. Miller RP, Roberts RJ, Fischer LJ. Acetaminophen elimination kinetics in neonates, children, and adults. Clin Pharmacol Ther 1976; 19:284–94. 17. McClain CJ, Kromhout JP, Peterson FJ, Holtzman JL. Potentiation of acetaminophen hepatotoxicity by alcohol. JAMA 1980; 244:251–3. 18. Seeff LB, Cuccherini BA, Zimmerman HJ, Adler E, Benjamin SB. Acetaminophen hepatotoxicity in alcoholics: a therapeutic misadventure. Ann Intern Med 1986; 104:399–404. 19. Lesser PB, Vietti MM, Clark WD. Lethal enhancement of therapeutic doses of acetaminophen by alcohol. Dig Dis Sci 1986; 3:103–5. 20. Wootton FT, Lee WM. Acetaminophen hepatotoxicity in the alcoholic. South Med J 1990; 83:1047–9. 21. Johnston SC, Pelletier LL. Enhanced hepatotoxicity of acetaminophen in the alcoholic patient: two case reports and a review of the literature. Medicine 1997; 76:185–91. 22. Girre C, Hispard E, Palombo S, N’Guyen C, Dally S. Increased metabolism of acetaminophen in chronically alcoholic patients. Alcohol Clin Exp Res 1993; 17:170–3. 23. Bray GP, Mowat C, Muir DF, Tredger JM, Williams R. The effect of chronic alcohol intake on prognosis and outcome in paracetamol overdose. Hum Exp Toxicol 1991; 10:435–8. 24. Thummel KE, Slattery JT, Ro H, et al. Ethanol exposure and acetaminophen hepatotoxicity: inhibition and induction of hepatotoxic metabolite formation. Clin Pharmacol Ther 2000; 67:5991–9. 25. Prescott LF. Paracetamol, alcohol and the liver. Br J Clin Pharmacol 2000; 49:291–301. 26. Bonkovsky HL, Kane RE, Jones DP, Galinsky RE, Banner B. Acute hepatic and renal toxicity from low doses of acetaminophen in the absence of alcohol abuse or malnutrition: evidence for increased susceptibility to drug toxicity due to cardiopulmonary and renal insufficiency. Hepatology 1994; 19:1141–8. 27. Pirotte JH. Apparent potentiation of hepatotoxicity from small doses of acetaminophen by phenobarbital. Ann Intern Med 1984; 101:403. 28. McClements BM, Hyland M, Callender ME, Blair TL. Management of paracetamol poisoning complicated by enzyme induction due to alcohol or drugs. Lancet 1990; 335:1526. 29. Bray GP, Harrison PM, O’Grady JG, Tredger JM, Williams R. Long-term anticonvulsant therapy worsens outcome in paracetamol-induced fulminant hepatic failure. Hum Exp Toxicol 1992; 11:265–70. 30. Murphy R, Swartz R, Watkins PB. Severe acetaminophen toxicity in a patient receiving isoniazid. Ann Intern Med 1990; 113:799–800. 31. Crippin JS. Acetaminophen hepatotoxicity: potentiation by isoniazid. Am J Gastroenterol 1993; 88:590–2. 32. Shriner K, Goetz MB. Severe hepatotoxicity in a patient receiving both acetaminophen and zidovudine. Am J Med 1992; 93:94–6. 33. Schiødt FV, Rochling FA, Casey DL, Lee WM. Acetaminophen toxicity in an urban county hospital. N Engl J Med 1997; 337:1112–7. 34. Schiødt FV, Atillasoy E, Shakil AO, et al. Etiology and outcome for 295 patients with acute liver failure in the United States. Liver Transpl Surg 1999; 5:29–34. 35. Ostapowicz GA, Fontana RJ, Schiodt FV, et al. Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States. Ann Intern Med 2002; 137:945–54. 36. Lee WM. Sounding board: acetaminophen hepatotoxicity and the U.S. acute liver failure study: lowering the risks of hepatic failure. Hepatology 2004; 40:6–9. 37. Kumar S, Rex DK. Failure of physicians to recognize acetaminophen hepatotoxicity in chronic alcoholics. Arch Intern Med 1991; 151:1189–91. 38. Rumack BH, Peterson RG. Acetaminophen overdose: incidence, diagnosis and management in 416 patients. Pediatrics 1978; 62:898–903. 39. Rivera-Penera T, Gugig R, Davis J, et al. Outcome of acetaminophen overdose in pediatric patients and factors contributing to hepatotoxicity. J Pediatr 1997; 130:300–4. 40. Heubi JE, Barbacci MB, Zimmerman HJ. Therapeutic misadventures with acetaminophen: hepatotoxicity after multiple doses in children. J Pediatr 1998; 132:22–7. 41. Hynson JL, South M. Childhood hepatotoxicity with paracetamol doses less than 150 mg/kg per day. Med J Aust 1999; 171:497. 42. Miles FK, Kamath R, Dorney SF, Gaskin KJ, O’Loughlin EV. Accidental paracetamol overdosing and fulminant hepatic failure in children. Med J Aust 1999; 171:472–5.

Acetaminophen-Induced Hepatotoxicity

403

43. Pershad J, Nichols M, King W. “The silent killer”: chronic acetaminophen toxicity in a toddler. Pediatr Emerg Care 1999; 15:43–6. 44. Alander SW, Dowd MD, Bratton SL, Kearns GL. Pediatric acetaminophen overdose: risk factors associated with hepatocellular injury. Arch Pediatr Adolesc Med 2000; 154:346–50. 45. Squires RH, Shneider BL, Bucuvalas J, et al. Acute liver failure in children: the first 348 patients in the pediatric acute liver failure study group. J Pediatr 2006; 148:652–8. 46. Alonso EM, Sokol RJ, Hart J, Tyson RW, Narkewicz MR, Whitington PF. Fulminant hepatitis associated with centrilobular hepatic necrosis in young children. J Pediatr 1995; 127:888–94. 47. Bonkowsky HL, Mudge GH, McMurtry RJ. Chronic hepatic inflammation and fibrosis due to low doses of paracetamol. Lancet 1978; 1:1016–8. 48. McBride AJ, Meredith-Smith P. Compound opioid/paracetamol analgesics: misuse and dependence. Br J Clin Pract 1995; 49:268–9. 49. Tredger JM, Thuluvath P, Williams R, Murray-Lyon IM. Metabolic basis for high paracetamol dosage without hepatic injury: a case study. Hum Exp Toxicol 1995; 14:8–12. 50. Hawton K, Fagg J. Trends in deliberate self poisoning and self injury in Oxford, 1976–90. Br Med J 1992; 304:1409–11. 51. Litovitz TL, Klein-Schwartz W, White S, et al. 1999 Annual report of the American Association of Poison Control Centers Toxic Exposure Surveillance System. Am J Emerg Med 2000; 18:517–74. 52. Fagan E, Wannan G. Reducing paracetamol overdoses. Br Med J 1996; 313:1417–8. 53. Gilbertson RJ, Harris E, Pandey SK, Kelly P, Myers W. Paracetamol use, availability, and knowledge of toxicity among British and American adolescents. Arch Dis Child 1996; 75:194–8. 54. Prince MI, Thomas SH, James OF, Hudson M. Reduction in incidence of severe paracetamol poisoning. Lancet 2000; 355:2047–8. 55. Turvill JL, Burroughs AK, Moore KP. Change in occurrence of paracetamol overdose in U.K. after introduction of blister packs. Lancet 2000; 355:2048–9. 56. Bernal W. Changing patterns of causation and the use of transplantation in the United Kingdom. Semin Liver Dis 2003; 23:227–37. 57. Hawton K, Townsend E, Deeks J, et al. Effects of legislation restricting packsizes of paracetamol and salicylate poisoning in the United Kingdom before and after stud. Br Med J 2001; 322:1203–7. 58. Sheen C, Dillon J, Bateman D, et al. Paracetamol pack size restriction: the impact on paracetamol poisoning and over-the-counter supply of paracetamol, aspirin and ibuprofen. Pharmacoepidemiol Drug Saf 2002; 11:329–31. 59. Robinson D, Smith AM, Johnston GD. Severity of overdose after restriction of paracetamol availability: retrospective study. Br Med J 2000; 321:926–7. 60. Hawton K, Simkin S, Deeks J, et al. U.K. legislation on analgesic packs: before and after study of long term effect on poisonings. Br Med J 2004; 329:1076. 61. Rumack BH. Acetaminophen overdose. Am J Med 1983; 75:104–12. 62. Portmann B, Talbot IC, Day DW, Davidson AR, Murray-Lyon IM, Williams R. Histopathological changes in the liver following a paracetamol overdose: correlation with clinical and biochemical parameters. J Pathol 1975; 117:169–81. 63. Harrison PM, O’Grady JO, Keays RT, Alexander GJ, Williams R. Serial prothrombin time as prognostic indicator in paracetamol induced fulminant hepatic failure. Br Med J 1990; 301:964–6. 64. Schiødt FV, Lee WM. Management of acetaminophen toxicity. In: Krawitt EL, ed. Medical Management of Liver Disease. New York: Marcel Dekker, 1999:325–77. 65. Cobden I, Record CO, Ward MK, Kerr DN. Paracetamol-induced acute renal failure in the absence of fulminant liver damage. Br Med J 1982; 284:21–2. 66. Kaysen GA, Pond SM, Roper MH, Menke DJ, Marrama MA. Combined hepatic and renal injury in alcoholics during therapeutic use of acetaminophen. Arch Intern Med 1985; 145:2019–23. 67. Blakely P, McDonald BR. Acute renal failure due to acetaminophen ingestion: a case report and review of the literature. J Am Soc Nephrol 1995; 6:48–53. 68. Wendon JA, Ellis AJ. Circulatory derangements, monitoring and management: heart, kidney and brain. In: Lee WM, Williams R, eds. Acute Liver Failure. Cambridge, MA: Cambridge University Press, 1997:132–43. 69. Dawson DJ, Babbbs C, Warnes TW, Neary RH. Hypophosphatemia in acute liver failure. Br Med J 1987; 295:1312–3. 70. Jones AF, Harvey JM, Vale JA. Hypophosphatemia and phosphaturia in paracetamol poisoning. Lancet 1989; 2:608–9. 71. Schmidt LE, Dalhoff K. Serum phosphate is an early predictor of outcome in severe acetaminopheninduced hepatotoxicity. Hepatology 2002; 36:659–65. 72. Chung PY, Sitrin MD, Te HS. Serum phosphorus levels predict clinical outcome in fulminant hepatic failure. Liver Transpl 2003; 9:248–53. 73. Davern TJ, Polson J, Lalani E, Lee WM. The U.S. ALF Study Group. Serum phosphate levels as a predictor of clinical outcome in acetaminophen induced acute liver failure. Hepatology 2003; 38:212A.

404

Lee and Ostapowicz

74. O’Grady J, Alexander GJM, Hayallar KM, Williams R. Early indicators of prognosis in fulminant hepatic failure. Gastroenterology 1989; 97:439–45. 75. Muldrew KL, James LP, Coop L, et al. Determination of acetaminophen-protein adducts in mouse liver and serum and human serum after hepatotoxic doses of acetaminophen using highperformance liquid chromatography with electrochemical detection. Drug Metab Dispos 2002; 30:446–51. 76. James LP, Mayeux PR, Hinson JA. Acetaminophen-induced hepatotoxicity. Drug Metab Dispos 2003; 31:1499–506. 77. Davern TJ, II, James LP, Hinson JA, et al. Measurement of serum acetaminophen-protein adducts in patients with acute liver failure. Gastroenterology 2006; 130:687–94. 78. James LP, Alonso E, Hynan L, et al. Detection of acetaminophen adducts in children with acute liver failure of indeterminate etiology. Pediatrics 2006; 118:e676–81. 79. Polson J, James LP, Davern TJ, et al. Acetaminophen as a co-factor in acute liver failure due to viral hepatitis determined by measurement of acetaminophen-protein adducts. Gastroenterology 2006; 130(Suppl. 4):772. 80. Clark R, Borirakchanyavat V, Davidson AR, et al. Hepatic damage and death from overdose of paracetamol. Lancet 1973; 1:66–70. 81. Wakeel RA, Davies HT, Williams JD. Toxic myocarditis in paracetamol poisoning. Br Med J 1987; 295:1097. 82. Mann JM, Peirre-Louis M, Kragel PJ, Kragel AH, Roberts WC. Cardiac consequences of massive acetaminophen overdose. Am J Cardiol 1989; 63:1018–21. 83. Weston MJ, Talbot IC, Howorth PJN, Mant AK, Capildeo R, Williams R. Frequency of arrhythmias and other cardiac abnormalities in fulminant hepatic failure. Br Heart J 1976; 38:1179–88. 84. Amour A, Slater SD. Paracetamol cardiotoxicity. Postgrad Med J 1993; 69:52–4. 85. Jones AL. Unusual complications of paracetamol poisoning. Q J Med 1997; 90:161–8. 86. Broughan TA, Soloway RD. Acetaminophen hepatotoxicity. Dig Dis Sci 2000; 45:1553–8. 87. Brotodihardjo AE, Batey RG, Farrell GC, Byth K. Hepatotoxicity from paracetamol self-poisoning in western Sydney: a continuing challenge. Med J Aust 1992; 157:382–5. 88. Gow PJ, Smallwood RA, Angus PW. Paracetamol overdose in a liver transplant centre: an 8-year experience. J Gastroenterol Hepatol 1999; 14:817–21. 89. Buckley NA, Whyte IM, O’Connell DL, Dawson AH. Oral or intravenous N-acetylcysteine: which is the treatment of choice for acetaminophen (paracetamol) poisoning? Clin Toxicol 1999; 37:759–67. 90. Lee WM, Schiødt FV. Fulminant hepatic failure. In: Schiff ER, Sorrell MF, Maddrey WC, eds. Schiff’s Textbook of Liver Diseases. New York: Lippincott-Raven, 1999:879–98. 91. Rumack BH, Matthew H. Acetaminophen poisoning and toxicity. Pediatrics 1975; 55:871–6. 92. Underhill TJ, Greene MK, Dove AF. A comparison of the efficacy of gastric lavage, ipecacuanha and activated charcoal in the emergency management of paracetamol overdose. Arch Emerg Med 1990; 7:148–54. 93. Bridger S, Henderson K, Glucksman E, Ellis AJ, Henry JA, Williams R. Deaths from low dose paracetamol poisoning. Br Med J 1998; 316:1724–5. 94. Bertholf RL, Johannsen LM, Bazooband A, Mansouri V. False positive acetaminophen results in a hyperbilirubinemic patient. Clin Chem 2003; 49:695–8. 95. Beuhler MC, Curry SC. False positive acetaminophen levels associated with hyperbilirubinemia. Clin Toxicol 2005; 43:167–70. 96. Polson J, Orsulak P, Wians F, et al. Elevated bilirubin may cause false positive acetaminophen levels in hepatitis patients. Hepatology 2004; 40:496A. 97. Spiller HA, Krenzelok EP, Grande GA, Safir EF, Diamond JJ. A prospective evaluation of the effect of activated charcoal before oral N-acetylcysteine in acetaminophen overdose. Ann Emerg Med 1994; 23:519–23. 98. Lauterburg BH, Corcoran GB, Mitchell JR. Mechanism of action of N-acetylcysteine in the protection against the hepatotoxicity of acetaminophen in rats in vivo. J Clin Invest 1983; 71:980–91. 99. Huggett A, Blair IA. The mechanism of paracetamol-induced hepatotoxicity: implications for therapy. Hum Toxicol 1983; 2:399–405. 100. Harrison PM, Wendon JA, Gimson AE, Alexander GJ, Williams R. Improvement by acetylcysteine of hemodynamics and oxygen transport in fulminant hepatic failure. N Engl J Med 1991; 324:1852–7. 101. Peristeris P, Clark BD, Gatti S, et al. N-Acetylcysteine and glutathione as inhibitors of tumor necrosis factor production. Cell Immunol 1992; 140:390–9. 102. Henderson A, Hayes P. Acetylcysteine as a cytoprotective antioxidant in patients with severe sepsis: potential new use for an old drug. Ann Pharmacother 1994; 28:1086–98. 103. Harrison P, Wendon J, Williams R. Evidence of increased guanylate cyclase activation by acetylcysteine in fulminant hepatic failure. Hepatology 1996; 23:1067–72. 104. Smilkstein MJ, Knapp GL, Kulig KW, Rumack BH. Efficacy of oral N-acetylcysteine in the treatment of acetaminophen overdose: analysis of the national multicenter study (1976 to 1985). N Engl J Med 1988; 319:1557–62.

Acetaminophen-Induced Hepatotoxicity

405

105. Prescott LF, Illingworth RN, Critchley JA, Stewart MJ, Adam RD, Proudfoot AT. Intravenous N-acetylcysteine: the treatment of choice for paracetamol poisoning. Br Med J 1979; 2:1097–100. 106. Smilkstein MJ, Bronstein AC, Linden C, Augenstein WL, Kulig KW, Rumack BH. Acetaminophen overdose: a 48-hour intravenous N-acetylcysteine treatment protocol. Ann Emerg Med 1991; 20:1058–63. 107. Bailey B, Amre DK, Gaudreault P. Fulminant hepatic failure secondary to acetaminophen poisoning: a systematic review and meta-analysis of prognostic criteria determining the need for liver transplantation. Crit Care Med 2003; 31:299–305. 108. Harrison PM, Keays R, Bray GP, Alexander GJ, Williams R. Improved outcome of paracetamolinduced fulminant hepatic failure by late administration of acetylcysteine. Lancet 1990; 335:1572–3. 109. Keays R, Harrison PM, Wendon JA, et al. Intravenous acetylcysteine in paracetamol induced fulminant hepatic failure: a prospective controlled trial. Br Med J 1991; 303:1026–9. 110. Dawson AH, Henry DA, McEwen J. Adverse reactions to N-acetylcysteine during treatment for paracetamol poisoning. Med J Aust 1989; 150:329–31. 111. Lee WM. Medical management of acute liver failure. In: Lee WM, Williams R, eds. Acute Liver Failure. Cambridge, MA: Cambridge University Press, 1997:115–31. 112. Polson J, Lee WM. AASLD position paper: acute liver failure. Hepatology 2005; 41:1179–97. 113. Shakil AO, Kramer D, Mazariegos GV, Funn JJ, Rakela J. Acute liver failure: clinical features, outcome analysis, and applicability of prognostic criteria. Liver Transpl 2000; 6:163–9.

20

Mechanisms Underlying the Hepatotoxicity of Nonsteroidal Anti-inflammatory Drugs Urs A. Boelsterli

Department of Pharmaceutical Sciences, School of Pharmacy, University of Connecticut, Storrs, Connecticut, U.S.A.

HEPATIC TOXICITY OF NSAIDs—A “CLASS EFFECT”? Nonsteroidal anti-inflammatory drugs (NSAIDs) belong to a group of therapeutic agents that are frequently prescribed because of their analgesic and antipyretic properties. Many NSAIDs are also available without a prescription and are generally considered safe. However, because a large population of patients is exposed to these drugs, it is not surprising that a relatively large number of adverse effects have been reported. The extreme worldwide use has led to an extensive literature on the incidence and types of major and minor adverse effects. Although the most frequent adverse effects associated with the use of NSAIDs clearly occur in the gastrointestinal tract, and also cardiovascular safety has become an issue of increasing concern, other target organs including the liver, have been identified. Comprehensive reviews are available that summarize the risk and the clinical manifestations of NSAID-related hepatic disorders (1–19). Although new NSAIDs have been designed and are becoming increasingly important (e.g., NO-releasing compounds; highly selective cyclooxygenase (COX-2) inhibitors; structural modifications of older compounds with a better safety profile and lack of reactive metabolite formation), many of the “classical” NSAIDs, including oxicams, are still extensively used (5). In spite of the fact that in the last few decades further clinical development of a significant number of NSAIDs had to be stopped, or drugs even withdrawn from the market because of hepatic toxicity (16), new cases of liver injury are being increasingly reported (4,20), usually occurring with a higher frequency in women (18). Therefore, the question arises whether this class of drugs has some common feature that would make them particularly prone to hepatic liability through the action of similar mechanisms. However, if one analyzes the relative incidence of acute or chronic liver injury caused by NSAIDs, it becomes apparent that the incidence is low, and not higher than that observed for other drugs (21). An exception may be sulindac: the unusually high incidence of approximately 150 cases of liver injury per 100,000 sulindac users suggests that there is a causal link between the use of this NSAID and an increased risk of hepatotoxicity (14,15). For other NSAIDs, the incidence is lower. For example, it is intermediate for mefenamic acid (2.5 per 100,000), diclofenac (3.6), or naproxen (3.8), and lowest for ibuprofen (1.6) and other NSAIDs (14,15,21). However, the “real” incidence is not known; it has been estimated to be 10- to 20-fold higher than the reported cases, due to notorious underreporting (22). In general, the clinical manifestations of NSAID toxicity in the liver can present as two distinct forms. On the one hand, mild hepatic changes, evident as minor increases in liver enzymes in the plasma, are relatively frequent and have been estimated to range between 1% and 15% (13). They are usually observed in phase III clinical trials prior to marketing. In contrast the clinically more significant hepatic injuries are of greater concern which The author of this chapter has relationships with the following corporations: consultation agreements with Biota Holdings Ltd. (Australia), Ono Pharmaceutical Co Ltd (Japan), and Novo Nordisk A/S (Denmark). He has research collaborations with Pfizer, Inc. (USA).

408

Boelsterli

become evident from case reports in the literature and sometimes can have a fatal outcome (8,16). These cases are very rare, but, in view of the great number of patients treated worldwide, the absolute numbers may be high. For example, in Denmark between 1978 and 1987, about 9% of all hepatic drug reactions could be attributed to NSAIDs (23). Whether the mild and severe forms of liver injury are causally linked to each other is currently not known. Awareness of severe hepatic adverse effects of NSAIDs became more widespread when benoxaprofen was introduced into clinical use. Almost immediately after its introduction in 1982, the drug was withdrawn from the market because of hepatic toxicity. Subsequently, the FDA Arthritis Advisory Committee issued the statement that hepatotoxicity is a class characteristic of NSAIDs (24). NSAIDs are, however, chemically heterogeneous, comprising several distinct classes (e.g., aspirin and other salicylates, phenylacetic acid derivatives, propionic acid derivatives, indol acetic acid derivatives, pyrazolone derivatives, oxicams, anthranilic acid derivatives, and the more recently introduced coxibs). Meanwhile, it has been recognized that this statement is an oversimplification for a number of reasons (25). For example, rates and types of injury vary within and between chemical classes. In addition, there is no consistent mechanism underlying all NSAIDinduced liver injuries. A likely reason why hepatotoxicity has been attributed to the entire therapeutic class may simply be that these drugs are among the most widely used medications in the world and that, therefore, hepatic adverse effects seem common. Nevertheless, most of those NSAIDs that have been associated with liver injury share some common structural features and follow similar pathways of hepatic metabolism and disposition. For example, many NSAIDs are weak acids (pKaw3.5–5.5), share a carboxylic acid moiety, and many of them feature lipophilic ring structures. They are often metabolized to acyl glucuronides, excreted, at least in part, via bile, and undergo enterohepatic circulation. Although these features may be general and shared by many other drugs, they are highlighted here for a better understanding of some of the molecular mechanisms underlying the hepatobiliary toxicity of these drugs. Only few reviews exist that describe the possible mechanisms underlying the hepatic toxicity of NSAIDs (25–28). Apart from in vitro data, the paucity of mechanistic data may reflect the fact that for most NSAIDs, mechanisms responsible for hepatic toxicity in vivo have remained largely enigmatic and speculative. In search of keys to unravel these mechanisms, the observed hepatic adverse effects of NSAIDs have often been grouped into two categories (often referred to as “mechanisms,” which they are not). In some cases, hepatic injury seems to be driven by a clear dose-dependent intrinsic toxicity of the compound (e.g., aspirin-induced hepatotoxicity). In most other cases, however, the hepatic reaction is idiosyncratic; that is, the toxicity largely depends on a host-dependent component and occurs in selected individuals only who feature a genetic and/or acquired predisposition. Since it is difficult at present to analyze all the individual susceptibility factors leading to idiosyncratic toxicity, efforts have concentrated on identifying the toxic risk of a compound. This risk is not only determined by the drug’s inherent toxic potential on the cellular or molecular level (hazard), but also driven by pharmacokinetic factors governing its disposition and metabolism. DISPOSITION AND METABOLISM OF NSAIDs—IMPLICATIONS FOR HEPATIC ADVERSE EFFECTS Plasma Protein Binding An important feature of NSAIDs is their high degree of reversible plasma protein binding, which usually is higher than 99% (29). Although the overall bound fraction is high, individual NSAIDs can exhibit marked differences in their unbound fractions (29). This becomes important for risk assessment in humans or for a critical extrapolation from in vitro studies in microsomes, isolated mitochondria, or cellular systems. Often, the incubation media do not contain exogenous albumin or other plasma proteins and results from in vitro assays can therefore be easily overestimated. It is not only the parent compounds, but also the metabolites (including the glucuronoconjugates) that are highly protein bound. However, for the glucuronides, the fraction of the

409

Nonsteroidal Anti-inflammatory Drugs

free (nonbound) metabolite can be considerably higher than that of the free parent compound. For example, although more than 90% of naproxen acyl glucuronide is bound to plasma proteins (30), the concentration of the free acyl glucuronide was approximately 10-fold higher than that of the free parent compound. Furthermore, the iso-glucuronides (spontaneously formed positional isomers of the acyl glucuronide) can exhibit an even lower degree of protein binding (e.g., 66% for naproxen) (30). As glucuronides play a putative role as pivotal mediators of NSAID hepatotoxicity, the sharp increase in concentration of the free circulating conjugates can become important. Lipophilicity Many NSAIDs are highly lipophilic compounds. Therefore, one predominant route of their uptake from blood into hepatocytes is transmembrane diffusion. Indeed, for certain NSAIDs (e.g., benzoic acid-containing and phenylacetic acid-containing NSAIDs), a correlation between the degree of lipophilicity and their acute cytotoxicity (potency) has been established by quantitative structure–activity relationships (31). The degree of lipophilicity can even have a greater influence on tissue penetration than protein binding (as the drug–plasma protein complex rapidly dissociates, leading to a new equilibrium); nevertheless, many NSAIDs still have a relatively low overall hepatic extraction rate. Besides its role in cellular uptake, lipophilicity may also play a critical function in mitochondrial toxicity (see below). Bioactivation to Reactive Metabolites Some NSAIDs (e.g., diclofenac, piroxicam, ketoprofen) have been shown to form reactive metabolites (32), causing acute lethal cell injury in cultured rat hepatocytes. Unfortunately, there is no clear-cut correlation between the potential to form such reactive metabolites in vitro and the relative incidence of liver injury in patients. Acute cell killing triggered by high drug concentrations in cell culture systems (33) cannot be directly translated into the human situation. However, such data may provide qualitative evidence for the generation of a reactive metabolite in the liver and can help to explain certain interactions of these intermediates with cellular macromolecules. Cytochrome P450-Mediated Bioactivation Some NSAIDs are ring-hydroxylated by selective cytochrome P450 (CYP) forms, which can lead to the formation of reactive intermediates. For example, diclofenac has been shown to form distinct p-benzoquinone imines (34,35), which are thiol-reactive electrophilic species (36,37). In rats, these intermediates are readily conjugated with glutathione and excreted in bile as S-glutathionyl adducts (Fig. 1). Interestingly, the same adducts were detected in human hepatocytes (37,38). Alternatively, aromatic ring structures can be bioactivated to an epoxide; for example, diclofenac arene oxides have been identified as novel metabolites of diclofenac (39). Again, these reactive intermediates can react with glutathione. If glutathione levels are depleted or if a metabolite is highly reactive, one can surmise that the electrophilic intermediate will also arylate cellular proteins. Indeed, in rats, one of the metabolites of diclofenac generated by CYP2C11-catalyzed reactions is so reactive that it forms a covalent adduct with P450 itself at the site of its generation (40). The toxicological implications of these reactions, however, are not clear.

O O NH Cl

Cl

4’-OH

O CYP2C9

N Cl

O

O

FIGURE 1 P450-mediated ring hydroxylation of diclofenac can result in the formation of quinone Cl imines. For example, CYP2C9-catalyzed 4’-hydroxylation leads to the 1’,4’-quinone imine, which exhibits electrophilic centers (arrows) that GSH can react with glutathione or nucleophilic residues of proteins.

410

Boelsterli

Bioactivation by Peroxidases to Pro-oxidant Radicals Although CYPs are the predominant NSAID-bioactivating enzymes, other less well-recognized enzymes may also participate. For example, certain NSAIDs can be metabolized and bioactivated by peroxidases (41). Peroxidases are hemoproteins that, in the presence of a peroxide (including hydrogen peroxide), are able to catalyze a one-electron oxidation of a substrate (NSAID). Typically, phenolic moieties can be oxidized to phenoxyl radicals, or amines can be converted to a nitrogen cation or nitroxide radical. An example for the latter is the diphenylamine moiety (which is common to a number of NSAIDs, e.g., diclofenac and mefenamic acid, Fig. 2). Peroxidases (including myeloperoxidase) are most abundant in granulocytes and macrophages. Since the liver harbors a vast number of Kupffer cells, hepatic peroxidase activity may thus bioactivate certain NSAIDs in the liver, and this could even be enhanced under inflammatory conditions. The ensuing NSAID radicals, if not scavenged by glutathione or ascorbate, could then induce oxidative stress in hepatocytes and lead to cell injury (28). However, the role of this alternative bioactivation pathway has not been fully explored in cellular systems or in vivo, and its role in humans is unclear. Interestingly, the immune cells in the liver also contain high activity of prostaglandin endoperoxide synthase, an enzyme which exhibits peroxidase activity. Although this enzyme itself is a major pharmacological target of NSAIDs, the drugs only inhibit the COX activity of prostaglandin endoperoxide synthase, while leaving the peroxidase moiety of the protein complex intact (42). Activation by Coenzyme A to Acyl-CoA Thioesters Carboxylic acid containing NSAIDs can be activated by acyl-coenzyme A synthetase (ACS1) to form acyl-CoA thioesters (Fig. 3) (43). The 2-arylpropionic acids (profens) are particularly prone to undergo this reaction. The liver is quantitatively the most important site of activation of profens to CoA thioesters (44). If large amounts of a drug are converted to CoA thioesters, this conjugation reaction in hepatocytes can have several consequences. First, it can lead to a depletion of the cytosolic CoA pool. Furthermore, similar to activated fatty acids, these NSAID–CoA thioesters can enter the pathways of lipid biochemistry. For example, they can be conjugated with cholesterol, bile acids, carnitine, or other amino acids (45) or, alternatively, incorporated into glycerolipids or phospholipids (46,47). Finally, acyl-CoA thioesters are protein-reactive intermediates that can directly acylate proteins (48). Thus, bioactivation by CoA is an important step in NSAID metabolism with possible toxicological implications. Interestingly, activation of profens, which exhibits a chiral carbon atom at the a-carboxy position, to their thioesters is stereoselective. Since there is a great variability in the nature Hepatocyte

Kupffer cell

H2O2 Peroxidase

R

NSAID

NSAID radical R

NH R’

Oxidative stress Covalent binding

GSH Ascorbate

+ R O NH or N R’ R’

FIGURE 2 Peroxidase-mediated bioactivation of certain NSAIDs (containing a diphenylamine or phenol moiety) to reactive intermediates. In the presence of a peroxide (e.g., H2O2), peroxidases in granulocytes or macrophages can oxidize a substrate to a prooxidant radical, which in turn may disrupt hepatocellular function. Abbreviation: NSAIDs, nonsteroidal anti-inflammatory drugs.

411

Nonsteroidal Anti-inflammatory Drugs

R

O

Depletion of cytosolic CoA pool

Acyl CoA R synthetase

O

CoA-SH

O

S-CoA

Incorporation into “mixed lipids” Acylation of proteins

FIGURE 3 Some carboxylic acid-containing NSAIDs (e.g., 2-arylpropionic acids) can be biotransformed by acyl-CoA synthetase to acyl-CoA thioesters. This bioactivation step can entail a number of toxicological consequences.

and extent of this stereoselectivity, both across species and different NSAIDs (49), one can expect that the extent of interference of profens with lipid metabolism and their protein reactivity is subject to considerable variability. The quantitative contribution of acyl-CoA synthetase in bioactivating NSAIDs is probably underestimated and could be substantial. For example, a comparative study with profens has revealed that the extent of covalent protein adduct formation from acyl-CoA thioesters is considerably greater than that from acyl glucuronides (50), another quantitatively important pathway of bioactivation. Activation by UDP-Glucuronosyltransferase to Reactive Acyl Glucuronides and Iso-Glucuronides Many carboxylic acid-containing NSAIDs are glucuronidated by members of the UDPglucuronosyltransferase (UGT) superfamily (reviewed in Ref. 51) to their corresponding acyl (or ester) glucuronides. In particular, UGT2B7 has been implicated in catalyzing the glucuronidation of many NSAIDs in humans (Fig. 4) (52,53), but other forms including UGTs 1A1, 1A9, and 2B4 are involved as well (54). Because both acyl glucuronides and their positional isomers, generated after spontaneous acyl migration, are protein-reactive metabolites, the glucuronoconjugation cannot be considered a mere detoxication step, but must also be considered a bioactivating pathway. UGT expression can vary interindividually due to both environmental and genetic factors. On the one hand, UGTs can be selectively induced by a number of drugs and other chemicals (55). On the other hand, there exists genetic polymorphisms in the genes coding for UGT (56,57). It is likely that additional genetic variations in the human population will be detected that could perhaps help to explain the variability in the plasma or urine concentrations of NSAID glucuronides in the human population (57). Although acyl glucuronidation is a common pathway, not all carboxylic acid-containing NSAIDs form these conjugates. For example, bromfenac metabolism by human microsomes

R

UGT2B7 O

R

UDP-glucuronic O acid

Hydrolysis (pH-dependent) O

Acyl migration

O-Glucuronide Protein adduct formation R

O S-glutathione

FIGURE 4 Carboxylic acid-containing NSAIDs can be bioactivated by UGT to b-1-O-glucuronides. These acyl (ester) glucuronides are reactive metabolites that can engage in a number of reactions with potential toxicological consequences. Furthermore, an acyl glucuronide (e.g., diclofenac glucuronide) can be converted to an S-acyl glutathione metabolite, which is highly protein reactive. Abbreviation: UGT, UDP-glucuronosyltransferase.

412

Boelsterli

does not produce an acyl glucuronide. Similarly, rats administered bromfenac unexpectedly produced an acid-labile N-glucoside conjugate (58). It is not known whether this unusual conjugation step with glucose is related to the hepatic toxicity associated with bromfenac, which was recently withdrawn from the market (59–61). Finally, NSAID acyl glucuronides can be further bioactivated, in a second step, and this may lead to even more reactive intermediates. For example, diclofenac acyl glucuronide can transacylate glutathione which will result in the formation of diclofenac-S-acyl-glutathione (Fig. 4) (62,63). This new conjugate is then excreted from the hepatocytes into bile. Interestingly, this diclofenac–glutathione thioester has been found to be approximately 200-fold more reactive than the parent glucuronide, implicating that this unusual glutathione conjugate could contribute to the covalent binding of NSAIDs to canalicular proteins. Hydrolysis and Systemic Cycling Acyl glucuronides are readily hydrolyzed, particularly in vivo (64). Serum albumin, to which most NSAIDs bind, can increase this process by interaction of the aglycone moiety with basic amino acids (65). Thus, the reversal of glucuronidation, coupled with reglucuronidation, can lead to a systemic cycling resulting in increased exposure and reduced renal elimination (66). Biliary Excretion Biliary excretion is an important pathway of NSAID clearance. In particular, the glucuronoconjugates and glutathione S-conjugates are eliminated across the canalicular membrane into the biliary tree. In the cases where NSAIDs contain a chiral center (e.g., the 2-arylpropionic acids), the hepatocanalicular export can be stereoselective. For example, the S-diastereomer of naproxen glucuronide has a higher export rate than the R-diastereomer (67). NSAID acyl glucuronides are selectively exported into bile by the canalicular isoform of the multidrug resistance-associated protein, Mrp2 (Fig. 5) (68). The hepatic expression of Mrp2 is inducible by a number of drugs (69). This may represent an adaptive response aimed at enhancing biliary elimination of the inducing drug and/or its metabolites. In contrast, under conditions of cholestasis, Mrp2 may disappear from the canalicular membrane, or be relocalized to other subcellular compartments (70). Genetic defects in the gene coding for the canalicular form of Mrp2, such as in the Dubin–Johnson syndrome in humans or in a number of transport-deficient rat models, have been described (71). These altered expression patterns of the canalicular conjugate export pump may have severe toxicokinetic and toxicodynamic consequences. Carboxylic acid NSAID Hepatocyte NSAID

Mrp2

Mrp2 • Inducible by drugs • Downregulated by cholestasis

Acyl glucuronides Iso-glucuronides MRP1

?

• Genetic defects

MRP3

FIGURE 5 Acyl glucuronides or their positional isomers are transported across the canalicular plasma membrane by the canalicular isoform of the multidrug resistance-associated protein (Mrp2) into bile. Other Mrp isoforms can export the glucuronides across the basolateral membrane into blood. The expression of Mrp is highly regulated, and genetic variations exist.

413

Nonsteroidal Anti-inflammatory Drugs

When biliary elimination is impaired by drugs or obstruction of the bile duct, other pathways may be activated. For example, while under normal conditions, acyl glucuronides and its isomers are excreted into bile (in the rat) and only a minor amount is excreted into the urine, experimental bile duct ligation caused the acyl glucuronide of zomepirac to be shunted into blood (72). These altered pathways can cause increased systemic exposure to the drug. Enterohepatic Circulation NSAID glucuronides that are excreted via the biliary tree into the gut can be deconjugated by several mechanisms. First, the slightly alkaline pH in bile and gastrointestinal tract can favor hydrolysis of the alkali-labile acyl glucuronide. In addition, acyl glucuronides, but not the iso-glucuronides, can be cleaved by bacterial b-glucuronidase and nonspecific esterases. This leads to a rapid reuptake of the free parent drug and to enterohepatic circulation (Fig. 6) (73– 76). Possible consequences of this repeated cycling are increased hepatic exposure and sustained interaction with hepatocanalicular transport, which is the rate-limiting site in drug elimination. That enterohepatic circulation of NSAIDs is an important factor which determines their hepatic toxic potential can be illustrated by therapeutic studies in dogs. NSAIDs undergo extensive enterohepatic circulation in dogs and are therefore eliminated slowly. Carprofen administration has caused hepatopathy, manifested by increases in aminotransferase activity, cholestasis, hepatocellular degeneration, and necrosis (77,78). Similarly, dogs that received naproxen at therapeutic doses developed toxic side effects including increases in hepatic enzyme markers (79). Cholehepatic Circulation Following excretion into the biliary tree, some NSAIDs can be readily reabsorbed. For example, it has long been suspected that sulindac (the aglycone, not the conjugated form) can undergo such cholehepatic circulation in humans (80). This has been confirmed in rats: following its canalicular secretion via the bile salt-exporting protein (cBsep), sulindac is reabsorbed across the bile duct epithelium. This results in decreased overall biliary excretion and higher blood levels during long-term administration of this drug (81). The possible consequences of cholehepatic circulation include increased exposure and possible competitive interaction at the hepatocanalicular transport site of bile salts and other substrates of the cBsep (Fig. 7).

Hepatocyte

Enterohepatic circulation

GSH, GSSG GS conjugates organic anions NSAID acyl glucuronide

Toxicological consequences: • Prolonged exposure • Inhibition of conjugate export • Retention of NSAID glucuronides in hepatocytes due to competition at the transporter site

Mrp2

Aglycone

Small intestine

FIGURE 6 Biliary excretion of the glucuronides, hydrolysis in the biliary tree and/or small intestine, and reabsorption of the aglycone lead to extensive enterohepatic circulation of a number of acidic NSAIDs. This condition can have toxicological consequences, as indicated.

414

Boelsterli

Hepatocyte Cholehepatic circulation

Bile salts

Sulindac

Bsep

Toxicological consequences: • Prolonged exposure • Inhibition of bile salt export • Retention of NSAID in hepatocytes due to competition at the transport site

FIGURE 7 Hepatobiliary excretion of sulindac (parent compound) across the canalicular membrane via the bile saltexporting protein (cBsep) and subsequent reabsorption in bile ducts leads to extensive cholehepatic circulation of this compound. This condition can have toxicological consequences, as indicated.

Renal Excretion If the renal elimination of NSAID metabolites, e.g., acyl glucuronides, is impaired, this can lead to higher levels of circulating conjugates and hence increased exposure to these potentially reactive metabolites. This was illustrated by experimental inhibition by probenecid of the renal elimination of zomepirac, which caused increased formation of drug adducts to plasma proteins (82).

CELLULAR AND MOLECULAR MECHANISMS OF NSAID-INDUCED HEPATOTOXICITY The “mechanisms” of NSAID hepatotoxicity are often classified as either intrinsic or idiosyncratic (4,12,16). However, this distinction is merely a phenotypical classification and reflects our lack of a clear understanding of the underlying mechanisms on a molecular and cellular basis. For overt dose-dependent adverse drug effects (e.g., acetaminophen), the mechanisms of toxicity can be more readily identified and taken into account for risk assessment. In contrast, for idiosyncratic drug toxicity, where the drug-induced liver injury may be cryptic or latent and clearly host-dependent, the mechanisms are much less easily determined because the “chemotype” of a drug may be less important than the recipient’s genotype and other determinants of susceptibility (disease, acquired factors). Both types of drug toxicity, however, harbor an intrinsic component, and some of the underlying cellular and molecular mechanisms may be similar. In search of these mechanisms, a number of key pathways have been emerging. Apart from possible effects related to the pharmacological targets, four major molecular mechanisms implicated in NSAID hepatotoxicity have been identified. These mechanisms are: (1) mitochondrial toxicity and activation of cell death-signaling pathways, (2) induction of oxidant stress and apoptosis, (3) protein binding of a reactive metabolite and subsequent hapten formation, and (4) interference of NSAIDs with hepatobiliary transport of cholephilic compounds, leading to intracellular accumulation of endogenous and/or exogenous compounds. Pharmacological Targets—Are They Related to Hepatotoxic Side Effects? The effectiveness of NSAIDs has been mostly attributed to COX inhibition, but other receptors, including the peroxisome proliferator-activated receptors (PPARs) have recently gained attention. Experimental evidence relating NSAID hepatotoxicity to these pharmacological targets is still limited but has promoted some novel concepts.

Nonsteroidal Anti-inflammatory Drugs

415

COX Inhibition NSAIDs are selective inhibitors of COX. This enzyme subfamily catalyzes the metabolic conversion of arachidonic acid to prostaglandins. Two major isoforms of COX have been identified in mammalian cells (a third form has recently been found). They are encoded from two different genes and exhibit tissue-specific expression. COX-1 is constitutively expressed and is a house-keeping enzyme, whereas COX-2 is an inducible form that is normally expressed at very low levels in many tissues but is upregulated by inflammatory mediators (83). It is possible that inhibition of the degradation of arachidonic acid via the COX pathway by NSAIDs will in turn stimulate the alternative pathway of arachidonic acid metabolism, that is, activation of the lipoxygenase pathway. This would increase the formation of leukotrienes and could therefore alter microsomal membranes (84) via generation of hydroperoxy derivatives (85) and inflammatory responses (86,87). However, there is no experimental evidence that this pathway may be related to the hepatic liability associated with NSAIDs. Another consequence of the COX-inhibitory effects of NSAIDs could, however, become important as it would make an individual more susceptible to the potential toxicity of other drugs administered concurrently. This is based on the following rationale: COX-1/2 have beneficial effects in that they catalyze the production of prostaglandins, which confer cytoprotection against a number of hepatotoxicants including acetaminophen. In line with this, COX-2 gene deficiency, or pretreatment with a COX-2 inhibitor, greatly potentiated acetaminophen-induced liver injury in mice (88). The underlying mechanism has been suggested to be an insufficient stress response (prostaglandins upregulate heat shock proteins, and heat shock protein expression is reduced when COX is inhibited) (88). Thus, although not yet sufficiently explored, the intrinsic pharmacologic effect of NSAIDs via COX inhibition could be one of the risk factors for hepatic adverse effects by dysregulating protective pathways. Peroxisome Proliferator-Activated Receptors NSAIDs can bind to and activate PPARs, which are ligand-activated nuclear receptors that are involved in the regulation of lipid homeostasis. Among the several isoforms of PPAR, PPARa is abundant in liver and plays a key role in peroxisomal fatty acid b-oxidation and peroxisome proliferation. In fact, earlier observations had revealed that certain NSAIDs (e.g., ibuprofen and flurbiprofen) are inducers of peroxisomal b-oxidation (89). Although these compounds are structurally similar to clofibric acid, a powerful inducer of peroxisomal b-oxidation, NSAIDs are not as potent inducers as clofibric acid. In contrast to PPARa, PPARg is normally expressed at low levels in the liver. However, obesity and nutrition can upregulate PPARg expression in the liver (90,91). PPARg plays a key role in decreasing mitochondrial b-oxidation and increasing fatty acid incorporation into storage lipid. In cell cultures, indomethacin, fenoprofen, ibuprofen, and flufenamic acid all have been shown to bind and activate PPARg and to induce lipogenesis (92). Interestingly, at low (nanomolar) concentrations, indomethacin blocked only COX activity, thus inhibiting the formation of prostaglandin derivatives that are activators of PPARg, without directly binding to PPARg. At higher (micromolar) concentrations, however, indomethacin not only inhibited COX activity but also activated PPARg. Thus, depending on the concentration, NSAIDs may function as either inhibitors or activators of PPARg-mediated processes (92). One of these PPARg-mediated processes is apoptosis. NSAIDs have been shown to induce apoptosis in cell lines by a COX-independent pathway (93). This has been causally linked with the compounds’ well-known chemopreventive activity against intestinal tumorigenesis (94–97). In spite of these well-known interactions of NSAIDs with PPARs, the relevance of these findings for the liver has remained unclear. Disruption of Mitochondrial Function Mitochondria have been identified as a potential pivotal target of NSAID toxicity in the liver. The underlying mechanisms of mitochondrial damage include uncoupling of oxidative phosphorylation, opening of the mitochondrial permeability transition pore, and inhibition

416

Boelsterli

of mitochondrial b-oxidation. These mechanisms not only can lead to impairment of bioenergetics but can also alter cell signaling and entail cell death. Uncoupling of Oxidative Phosphorylation Mitochondrial uncoupling of oxidative phosphorylation is one of the most widely discussed mechanisms underlying the toxicity of NSAIDs (98–102). This effect can be explained by the chemical structure; many NSAIDs are monocarboxylic acids with one or more aromatic rings and most of them are lipophilic. These features are typical for uncoupling agents. Uncoupling compounds short-circuit the proton gradient that is normally built up in the intermembraneous space during electron transport, by reversing the proton flux from the intermembraneous space back into the matrix (Fig. 8). The resulting dissipation of the proton gradient precludes the oxidative phosphorylation of ADP by the ATP synthetase. This ultimately leads to release of Ca2C and to an energy crisis and cell demise (103–106). At least three major mechanisms have been identified that form the basis of this uncoupling effect. First, the protonophoric activity of acidic NSAIDs shuttles the protons back into the matrix. Evidence indicates that the carboxylic acid moiety, the diphenylamine structure, and the sulfonamide amine all may contribute to this effect (105,107). Second, NSAIDs can directly cause opening of the mitochondrial permeability transition pore, which allows and further promotes rapid influx of protons into the matrix. Finally, there is evidence that the hydrophobic NSAIDs accumulate nonspecifically in the mitochondrial membranes and cause membrane disordering, which also could contribute to the uncoupling effects (101). Most of the experiments demonstrating an uncoupling effect of NSAIDs were carried out either in isolated mitochondria (101) or in isolated or cultured hepatocytes. The concentrations needed to induce such mitochondrial changes ranged from low micromolar (for e.g., mefenamic acid, flufenamic acid, diflunisal, or nimesulide) (98,108), to high micromolar (100–250 mM e.g., for diclofenac), to the millimolar range (e.g., for aspirin). Some NSAIDs did not inhibit hepatic mitochondrial ATP synthesis at all (e.g., indomethacin or sulindac), even at concentrations exceeding 5 mM (98). A comparison among NSAIDs of the IC50 values for inhibition of ATP synthesis and the relative incidence of inducing liver dysfunction in humans readily reveals that there is no positive correlation between the in vitro data and the hepatotoxic potential in humans. For example, sulindac, which is more frequently associated with hepatotoxicity than all other NSAIDs (109), was not cytotoxic and did not deplete R Uncoupling of oxidative phosphorylation by protonophores

NH R

H+

H+

H+

O

R

,

O

e–

R

H+

H H+ H+

ATP synthase

Opening of the mPT

HN

SO2

R

,

FIGURE 8 Acidic NSAIDs dissipate the mitochondrial membrane potential by facilitating proton reflux across the inner mitochondrial membrane. This can occur either by the protonophore activity of the carboxylic acid moiety, the diphenylamine structure, or a sulfonamide moiety of an NSAID, causing uncoupling of electron transport from ATP synthesis. Uncoupling can result in mitochondrial permeability transition (mPTP) pore opening, which leads to a collapse of the mitochondrial transmembrane potential.

417

Nonsteroidal Anti-inflammatory Drugs

hepatocellular ATP in rat hepatocyte cultures (105). In contrast, nimesulide and meloxicam do not have a higher incidence of hepatic side effects than other NSAIDs, yet are among the most potent uncouplers (102,108). Thus, uncoupling is not a general property of all NSAIDs and this hazard alone is unlikely to be the only underlying mechanism of toxicity in vivo. In addition, most of the effects observed with isolated mitochondria or cell cultures were caused by high drug concentrations only. Since most NSAIDs are highly bound to plasma proteins in vivo (29), the concentration of “free” (unbound) drug is approximately two to three orders of magnitude smaller than the concentrations used in vitro. Induction of the Membrane Permeability Transition in Mitochondria In recent years, the mitochondrial membrane permeability transition (mPT) pore, a tightly regulated, cyclosporine A-sensitive megachannel associated with Ca2C-dependent increases in the permeability of ions and solutes with molecular masses %1.5 kDa, has gained much attention [there is a nonregulated, cyclosporine A-insensitive mechanism, too (110)]. The mPT has been implicated in mitochondrial uncoupling, release of proapoptotic factors into the cytosol, and induction of cell demise. Permeability transition can be triggered by oxidant stress, increased [Ca2C]i, and drugs including NSAIDs (Fig. 9). In fact, a recent report indicates that induction of the mPT is a general response to short-chain carboxylic acids having a pKa of 4 to 5 (111). NSAIDs are able to differentially induce the mitochondrial mPT, leading to a collapse of the proton gradient and an energy crisis (107,108,112–114). This can be induced through several pathways but generally is mediated through ROS originating from oxidative metabolites or as a result of uncoupling (Fig. 9). While some NSAIDs (e.g., piroxicam, aspirin) exert this effect only at relatively high concentrations (500 mM), others (e.g., diclofenac, mefenamic acid, nimesulide) cause opening of the mPT pore at low micromolar concentrations (108,114). The mechanism of mPT induction caused by NSAIDs is not well understood, but it has been suggested that oxidation of pyridine nucleotides or protein thiols may play a role. Alternatively, because many R

O O Caspase-8 activation

R NH R

,

Oxidative metabolite Uncoupling ∆ψm

Truncated Bid

Bid

ROS

mPT

Cytochrome c

Apoptosis

Activation of Caspase-9 and 3

FIGURE 9 NSAIDs can produce ROS via oxidoreductive metabolism to prooxidant intermediates or as a result of uncoupling of ox-phos. This can lead to mitochondria-dependent induction of apoptosis. For example, opening of the mPT pore makes the mitochondria permeable to solutes R1.5 kDa and leads to mitochondrial swelling and release of proapoptotic factors. Abbreviation: mPT, membrane permeability transition.

418

Boelsterli

of the NSAIDs investigated are also powerful uncouplers of oxidative phosphorylation, and because uncoupling can lead to the opening of the mPT pore, mPT induction could be a secondary event. However, mPT induction is Ca2C-dependent and can only be induced experimentally by NSAIDs in the presence of exogenously added Ca2C; therefore, the role of mPT in vivo is still under investigation. In vitro, certain NSAIDs alone may cause only minimal killing of cultured hepatocytes even at very high (millimolar) concentrations, making this an unlikely mechanism in vivo. However, these compounds may potentiate necrosis and/or apoptosis induced by other drugs or toxicants by lowering the threshold for mPT. For example, salicylate (the active metabolite of aspirin) has been shown to trigger mPT-dependent apoptosis induced by nontoxic concentrations of protoxicants (115). Inhibition of Mitochondrial b-Oxidation Inhibition of mitochondrial b-oxidation has been discussed as one of the mechanisms involved in NSAID hepatotoxicity, in particular that associated with 2-arylpropionic acid derivatives (Fig. 10) (116–121). Mitochondrial b-oxidation is the process by which nonesterified fatty acids (NEFAs) are oxidized and shortened into acetyl-CoA fragments, which in turn are either condensed to ketone bodies or further metabolized by entry into the citrate cycle. In contrast to short- and medium-chain NEFAs, which can readily penetrate into mitochondria, long-chain NEFAs have to be activated to an acyl-CoA intermediate prior to being transported across the inner mitochondrial membrane by the carnitine shuttle system. Inhibition of b-oxidation, which is an important sink for NEFAs in the liver, leads not only to decreased ATP production but also to an accumulation of fatty acids. Ultimately, this can develop into microvesicular steatosis. The mechanism underlying NSAID-induced inhibition of b-oxidation has both a stereoselective and a nonstereoselective basis (117). The first mechanism can be explained by the stereoselective formation of 2-arylpropionic acid-CoA thioester formation, which can lead to extramitochondrial CoA sequestration. As a consequence, activation of long-chain NEFAs is inhibited and, hence, b-oxidation is decreased (120,122). Second, for those 2-arylpropionic acids that do not form CoA intermediates (e.g., flurbiprofen), a nonstereoselective, CoA-independent b-oxidation pathway has been described (99). The exact mechanism is not known but, because these drugs are also uncouplers, it has been suggested that the NSAID may enter mitochondria and directly inhibit b-oxidation (99). Finally, as many NSAIDs are activators of PPARg, which is associated with downregulation of the mitochondrial b-oxidation pathway and accumulation

R

O

R

O

O S-CoA

CoA-SH Long-chain fatty acids

Role of PPARγ activation?

Acylcarnitine

Acyl-CoA

H

β-oxidation

Toxicological consequences: • Microvesicular steatosis

FIGURE 10 Inhibition by 2-arylpropionic acids of long-chain fatty acyl b-oxidation in mitochondria. CoA is sequestered by the acidic NSAIDs and is less available to activate long-chain fatty acids prior to their carnitinemediated transport across the inner mitochondrial membrane. The ensuing accumulation of fatty acids in hepatocytes may lead to microvesicular steatosis.

Nonsteroidal Anti-inflammatory Drugs

419

of NEFAs, it is possible that the observed microvesicular steatosis in hepatic parenchymal cells could be attributed, at least in part, to activation of upregulated PPARg in the liver. The concentrations required to achieve a significant inhibition of b-oxidation in vitro are usually much higher than the therapeutic plasma concentrations, considering again that the nonbound (“free”) drug is less than 1% of the total. Therefore, it is not likely that these effects will become important in the vast majority of patients. However, the data can explain a mechanism that may become relevant in compromised cells, in genetically altered mitochondrial b-oxidation (123), or at exceedingly high intracellular NSAID concentrations. Indeed, microvesicular steatosis has been described in patients who received pirprofen (124), naproxen (125), ibuprofen (126), or ketoprofen (127). However, because microvesicular steatosis can be nonspecific and because it is more prevalent than previously suspected, a causal link with NSAIDs is often difficult to establish. A causal link can only be made with the use of aspirin (128). Oxidant Stress and Induction of Apoptosis NSAIDs are well known to induce apoptosis under certain conditions. In fact, although conceived as a hazard at first, the apoptosis-inducing properties of NSAIDs has given some of them (in particular, the COX-2 inhibitors) potential therapeutic application against colon cancer (129). In liver, diclofenac has been particularly well studied; this drug causes DNA fragmentation, induces release of cytochrome c from mitochondria, activates the caspase cascade, and induces the mitochondrial permeability transition (mPT) in hepatocytes and other cells (130–132). The molecular mechanisms underlying diclofenac-induced apoptosis likely include oxidant stress as an early inducing event (130,131,133) and mitochondria as a target of toxicity (114). It has been suggested that reactive oxygen species (likely produced from diclofenac reactive metabolites) might suppress Akt activity, thereby activating caspase-8, which in turn stimulates Bid cleavage and causes release of cytochrome c from the mitochondrial intermembrane space. Eventually, cytochrome c activates executor caspases (132). A pivotal role for ROS has been inferred primarily from the observation that a number of antioxidants can protect from diclofenac-induced apoptosis in vitro; however, it is not always clear whether these antioxidants inhibit the initial signal-triggering steps (drug specific) or they interfere with some of the apoptosis-executing steps further downstream (generic pathways). Most of the mechanistic studies have been performed in cultured cells or in laboratory animals receiving very high (supratherapeutic) doses of NSAIDs. Therefore, these results clearly indicate a hazard (potential to induce oxidant stress and/or apoptosis); however, the extrapolation to the actual risk in humans receiving these drugs has been proven difficult. A recent study, however, has used an animal model in which low doses of the nitroaromatic NSAID nimesulide (comparable to human therapeutic doses and resulting in similar peak plasma levels as in patients) were repeatedly administered to mice (134). While normal healthy mice did not exhibit any signs of mitochondrial changes upon drug treatment, the potential of nimesulide to induce oxidant stress in liver mitochondria became unmasked in heterozygous superoxide dismutase 2-deficient (Sod2C/K) mice where the drug was superimposed on an underlying genetic mitochondrial abnormality. In these animals, exhibiting an underlying silent mitochondrial stress, nimesulide caused a sharp decrease in mitochondrial aconitase activity and increased levels of mitochondrial protein carbonyls, both specific indicators of oxidant stress. Furthermore, liver cytosolic cytochrome c levels and caspase-3 activity were increased, and the number of TUNEL-positive hepatocytes was significantly enhanced, clear indicators of mitochondria-mediated proapoptotic events. Finally, nimesulide clearly increased the production of superoxide anion in isolated mitochondria from Sod2C/K mice (but not normal wild-type mice). Thus, in this animal model, an oxidant stress combined with proapoptotic events became unleashed only when genetically altered mitochondria were exposed to the drug. This is commensurate with the clinical data—in the overwhelming majority of patients, nimesulide and other NSAIDs do not cause liver injury.

420

Boelsterli

Protein Adduct Formation and Immune-Mediated Toxicity Mechanisms and Molecular Targets of NSAID Acyl Glucuronides and Iso-Glucuronides Carboxylic acid-containing NSAIDs are biotransformed both in the liver and extrahepatically to reactive metabolites. Among these, acyl glucuronides have gained special attention because they are quantitatively important and they are protein-reactive. As a consequence of their moderate reactivity, they often do not react with target molecules in their immediate vicinity, but leave the site of their formation and reach the blood or biliary compartment. In the vascular system, acyl glucuronides can react with plasma proteins (135–138). Protein binding to e.g., albumin, has also been demonstrated in vitro (139–146) and used to investigate the molecular mechanisms of binding. For human serum albumin, the most prominent binding site for covalent interactions with NSAIDs has been determined. Following exposure to tolmetin acyl glucuronide, the intramolecular target has been identified as Lys-199 (143,147). This is particularly interesting, as Lys-199 is a lysine 3-amino group located in a hydrophobic region of the protein that is a target for covalent binding of penicillin derivatives, well-known immunogenic drugs (148). One mechanism of covalent binding of an NSAID acyl glucuronide to protein is nucleophilic displacement or transacylation, whereby the NSAID acylates a nucleophilic amino acid residue of a target protein and the glucuronic acid moiety is released (Fig. 11). A second mechanism of binding is protein glycation by which the glucuronic acid moiety is retained in the adduct. This reaction is made possible after intracellular rearrangement (acyl migration), where the acyl group migrates from the 1-O-position to the 2-, 3-, or 4-position of the sugar ring, exposing a new electrophilic center in the resulting iso-glucuronide (Fig. 12). Protein glycation by this mechanism is quite common, in particular by 2-arylpropionic acids (67,149). Protein adducts with iso-glucuronides may be quantitatively as important as those derived from the original acyl glucuronide (150). They may even become more relevant in toxicology as they are more persistent than the less stable adducts formed from the 1-O-acyl glucuronide. There is a clear correlation between the stability of a formed acyl glucuronide and its reactivity (covalent binding) (151). For a number of acyl glucuronides, the apparent first-order disappearance rate constant in buffer correlated in a linear fashion with maximum irreversible binding to albumin in vitro. The degree of substitution at the a-carboxy atom is crucial in determining the stability; unsubstituted acetic acid derivatives are relatively unstable and exhibit high covalent binding. More stable glucuronides are mono-a-substituted acetic acids, which exhibit intermediate covalent binding. Finally, stable acids fully substituted at the a-carbon exhibit only little covalent binding (151). It has therefore been suggested that steric factors hindering the reaction may play a role in this differential reactivity. However, the predictability of covalent binding becomes less accurate for the in vivo situation: besides the degradation rate constant of a given glucuronide, many other factors can modify the site

X = NH O

XH-Protein

S

OH

R O

HO

OH O

COOH

O

FIGURE 11 Covalent adduct formation of a NSAID acyl glucuronide (unsubstituted at the a-carbon atom) to a nucleophilic amino acid residue of a target protein by the transacylation mechanism.

421

Nonsteroidal Anti-inflammatory Drugs

OH HO O

R O

O

OH COOH R

1-O-β-Glucuronide

O R

R O HO

OH O

O

O OH COOH

HO O

HO

OH COOH

3-O-β-Glucuronide

2-O-β-Glucuronide

OH HO O

O

O O COOH

4-O-β-Glucuronide

H2N-Protein

FIGURE 12 Covalent adduct formation of a NSAID iso-glucuronide to a target protein by the glycation mechanism. Following intracellular rearrangement (acyl migration of the aglycone along the sugar ring), positional isomers can be formed. These iso-glucuronides (in this example, the 3-O-glucuronide of an NSAID) can exist transiently in the open ring form, where the exposed aldehyde group is attacked by a nucleophilic amino acid residue.

and extent of binding considerably, including differential bioactivation pathways and pharmacokinetic variables. Stability and, hence, protein adduct formation of acyl glucuronides are pH dependent. Acyl glucuronides are unstable at pHO8, a pH that may occur in the biliary tree and the lower intestine. Therefore, covalent binding in the biliary tree will occur more extensively than under conditions of lower pH, as is found in blood. In the liver, covalent protein binding does not occur at random, but the electrophilic NSAIDs target selective proteins. A number of experimental studies have characterized or identified proteins that are preferentially alkylated or acylated by reactive NSAID metabolites (40,68,152–157). Which proteins are detected in Western blots or by fluorography depends on a number of factors, including the duration of drug administration and the time that has elapsed between the dose and the sacrificing of the animal. In most studies, persistent adducts were found to be associated with plasma membrane proteins (68,154,156–158). Subcellular fractionation studies and immunohistochemical analysis have revealed that one (or one group of) particularly abundant adduct(s) is associated with proteins of the canalicular plasma membrane domain. Evidence suggests that the proteins targeted by a number of NSAIDs including diclofenac (68,154,157), sulindac (157), or benoxaprofen (159) may be identical. Specifically, they all are in the molecular mass range of 110 to 126 kDa, and one common target that has been identified is dipeptidyl peptidase IV (DPPIV) (Fig. 13) (156,160,161).

Hepatocyte NSAID Acyl glucuronides Iso-glucuronides Intracellular targets?

DPPIV Other canalicular proteins ? • pH ~8.0 • High concentration of protein-reactive metabolite • Primary interaction with target protein

FIGURE 13 NSAID metabolites covalently bind to one or several 110 to 118 kDa plasma membrane proteins exposed to the canalicular lumen. One identified common target is dipeptidyl peptidase IV (DPPIV).

422

Boelsterli

Adducts to canalicular membrane proteins are only generated when the protein-reactive metabolite (the acyl glucuronide or isomers) is transported from the hepatocyte across the canalicular membrane into the bile canalicular lumen (68). This was inferred from the observation that in mutant transport-deficient (TRK) rats, which do not express a functional export pump for the glucuronides (Mrp2), no membrane protein adducts were detectable after treatment with diclofenac. The reason for this compartment-selective reaction is several fold. First, the reactive metabolite that is exported from the cell by an ATP-dependent transporter is upconcentrated in the canalicular lumen and reaches high concentrations. Second, the reactivity of the acyl glucuronides is increased in a slightly alkaline compartment, such as the biliary tree (162,163). Finally, the nature of the target protein(s) may favor a primary interaction with the glucuronide metabolite. The mechanistic insights in NSAID protein adduct formation and any possible downstream consequences all stem from experimental animal studies. Interestingly, there is evidence that in humans (patients treated with diclofenac), similar adducts can be formed (164). These adducts will alter the protein, and, if presented to the immune system, an antibody response may be directed against the “foreign” peptide. For example, anti-diclofenac antibodies were found in the serum of all patients who had developed diclofenac-related liver injury, but not in healthy control subjects. However, because circulating anti-diclofenac antibodies were equally found in 60% of patients on diclofenac who did not exhibit any signs of liver toxicity, adduct formation and formation of anti-diclofenac antibodies does not seem to be sufficient to precipitate liver injury. Collectively, the role of covalent protein binding in NSAID hepatotoxicity is not clear. However, theoretically, adduct formation can have toxicological consequences, including inactivation of a critical target protein or hapten formation and an immunogenic response against the drug-altered protein. Inactivation of Critical Protein Function For some proteins covalently modified by NSAID metabolites, a clear functional impairment has been demonstrated. For example, the activity of DPPIV, which is an abundant ectoenzyme residing in the canalicular plasma membrane and a major target, was reduced by 22% in rats treated with diclofenac (156). The biological significance of this functional alteration has remained unclear, as there is no obvious relationship between the decrease in DPPIV activity and hepatotoxicity. In fact, NSAIDs that produce focal necrosis and apoptosis in rat liver (diclofenac and sulindac), as well as an NSAID that produces much less overt injury (ibuprofen), caused equal inhibition of DPPIV expression and activity (160). The function of a number of other proteins covalently modified by NSAIDs was shown to be similarly compromised. For example, acute administration of diclofenac to rats resulted in covalent binding to CYP2C11 and a 72% decrease in the catalytic activity of this P450 (40). Furthermore, incubation of suprofen acyl glucuronide with human serum albumin or superoxide dismutase significantly inhibited both the binding capacity of albumin for other drugs and the catalytic activity of superoxide dismutase (165). Finally, the reactive acyl glucuronide of zomepirac covalently modified tubulin and inhibited its assembly into microtubules (166). The relevance of these findings for NSAID hepatotoxicity is not clear. Acute damage to and rapid replacement of a protein may not be so critical. Thus, for proteins featuring a relatively short half-life, such as albumin, binding may be less relevant than for proteins featuring a much slower turnover rate (e.g., collagen), where accumulating adducts may be more severe (165). Hapten Formation and Immune-Mediated Hepatic Injury Covalent binding of a reactive NSAID metabolite to hepatocellular proteins may lead to the formation of a hapten that could play a role in a possible immune response against the drugaltered self-protein (167). The hapten itself, or conformational changes of the target protein, could then result in the formation of structural and conformational epitopes, respectively. Direct evidence for the involvement of such a mechanism is, however, scanty. Several requirements have to be fulfilled for an adduct to become immunogenic. First, the adduct density (mol adduct bound per mol protein) is an important determinant.

Nonsteroidal Anti-inflammatory Drugs

423

Chronic treatment will increase the adduct density; indeed, multiple dosing, as opposed to single administration of an NSAID, can lead to accumulation of the protein adducts (138,168). Second, the half-life (stability) of the adduct itself may also determine its potential immunogenicity. The stability of an adduct is partially determined by the mechanism of covalent binding: for many drugs, adducts arising from the reactive iso-glucuronides are more stable than those arising from the acyl glucuronide. Often the adducts actually persist in plasma far beyond the period when concentrations of the parent compound and/or its glucuronides are measurable (138,168). For example, in human volunteers (six-day oral study), the terminal halflife of covalently bound diflunisal to plasma proteins was calculated to be 10 days (169). Longlived adducts may lead to enhanced uptake by macrophages, resulting in greater processing and presentation of antigenic peptides compared with unaltered albumin (170). Adduct formation alone however is not sufficient to trigger an immune response. In fact, most, if not all, recipients receiving NSAlDs form adducts and the formation of drug-altered peptides induces tolerance rather than an immune response. To stimulate B cells or T cells, peptides require presentation by major histocompatibility complex (MHC) molecules. This process is limited in the liver: hepatocytes express MHC class I molecules at a very low level, and class II molecules are not expressed under normal conditions (171). MHC class I molecules can, however, be upregulated under pathophysiological conditions, including viral hepatitis, autoimmune liver disease, or cholestasis (172–174). Similarly, MHC II antigen can become expressed in hepatocytes following stimulation by cytokines or under pathophysiological conditions such as hepatitis (171,175–177). By implication, one can speculate that alkylated hepatic proteins (“nonself” peptides) are released during cell turnover or after toxic injury, phagocytosed by Kupffer cells, degraded, and presented in conjunction with MHC II, thus activating a particular Th cell clone bearing an appropriate T-cell receptor. The activated T-cell clone may begin to express IL-2 receptors and secrete various immune modulators including 1L2 and IL-4, which will activate other immune cells, including B cells and cytotoxic Tcells (Fig. 14). Immune-mediated liver injury could be mediated by a number of effector mechanisms. First, antibodies may bind to alkylated membrane proteins on the cell surface and induce cell lysis through complement or killer cells. Although, in some cases, such antibodies against NSAID-altered peptides have indeed been found, the pathophysiological significance of antibodies is not clear. In particular, it is not clear whether they are causally involved in the killing of target cells or merely markers for antigenicity. For example, two independent experiments have shown that immunization with a syngeneic serum albumin conjugate of a NSAID (diflunisal and tolmetin) acyl glucuronide stimulated an antibody response in rodents (178,179). These studies show that self-proteins covalently modified by incubation with a reactive NSAID glucuronide can be immunogenic. They do not, however, provide evidence that these antibodies are indeed causally involved in an aberrant immune response in the liver. It is only in few cases that there is evidence that these antibodies may play a causative role in cell-mediated toxicity. For example, the addition of sera from patients with clometacine hepatitis to cultured human hepatocytes that had been exposed to clometacine resulted in hepatocellular injury when autologous lymphoid cells were added (180). The antiserum from such patients did not, however, recognize drug-modified proteins, but only native proteins. Therefore, it is likely that clometacine, and possibly other NSAIDs, can in rare cases induce autoimmune-type hepatitis. This has been inferred from the presence of accompanying high titers of anti-DNA or anti-smooth muscle antibodies, typical for autoimmune hepatitis (181). It is possible that the role of the drug, being one of the factors that add to the risk of developing autoimmune hepatitis, would be in revealing a latent autoimmune disease (182). One hypothesis involves autoreactive B cells that are normally quiescent but would present both the drug-altered protein and native peptides, triggering a Th response against the native protein (183). Recent evidence also indicates that drugs can disrupt the process of positive selection of immature Tcells in the thymus that occurs throughout life, breaking the tolerance to self, and that this process can lead to autoimmune disease (184). The second effector mechanism involves cytolytic T cells, following appropriate stimulation by Th cells. Cytolytic T cells would recognize the drug-modified peptides presented in conjunction with MHC I and destroy the target cells by Fas- or perforin-mediated mechanisms. Such a mechanism has not been proven in vivo, and to date there is no animal model available

424

Boelsterli

NSAID Th Cytokine release

MHCII

Kupffer cell Activated B cell

Armed MF

MHCII Th

MHCI

ROS C3a TNFα

B cell activation, proliferation, plasma cell

CytolyticT cell precursor activation Proliferation maturation

NSAID Activated MF /NK cells

B cell activation

Cytolytic T cell FasL Perforin

NSAID

Hepatocyte destruction

FIGURE 14 Putative pathways of immune-mediated hepatocyte injury induced by NSAIDs. Covalent adducts of reactive metabolites of NSAIDs will be accessible to cells of the immune system following degradation of hepatocytes. Adducts are primarily formed in the endoplasmic reticulum and at the canalicular plasma membrane. Internalization of NSAID-modified proteins by antigen-presenting cells (e.g., Kupffer cells) is followed by processing and presentation of the peptides in conjunction with MHC class II. The hapten or conformational epitope of the peptides is recognized by the T-cell receptor of T helper cells. Cytokine release and subsequent activation of B cells and cytolytic T-cell precursors results in clonal expansion and maturation to cytolytic T cells. These T cells recognize the antigen presented in conjunction with MHC I on the hepatocytes and destroy the target cell. Activated macrophages or natural killer cells may also lyse target cells by various mediators. Alternatively, NSAID-altered proteins may interact with a B-cell receptor, followed by internalization, processing, and presentation by MHC II, which leads to activation and clonal expansion of B cells, maturation to plasma cells, and secretion of antibodies. These antibodies may bind to epitopes on the plasma membrane of hepatocytes (exposed protein adducts). This may result in nonspecific recognition and binding via the Fc receptor of killer cells and macrophages and killing of the target hepatocytes.

with which one could study these pathways. Limited evidence stems from a mouse study which has shown that in vivo activated T cells derived from diclofenac-immunized mice are able to kill hepatocytes that had been exposed in vitro to diclofenac (Fig. 15) (185). In this ex vivo/in vitro model, C57BL/6 mice were immunized with diclofenac coupled to the carrier protein, keyhole limpet hemocyanin (KLH). To explore the role of a T-cell-mediated response directed against diclofenac-modified peptides, splenocytes from KLH-diclofenac (and KLH only) immunized mice were harvested and either used as a crude splenocyte fraction (containing Tcells, B cells, NK cells, and macrophages) or further purified to a T-cell-enriched fraction. These splenocytes were then combined with isolated and precultured syngeneic hepatocytes and kept in coculture for several days. Prior to being combined with splenocytes, the hepatocytes were exposed to high but nontoxic concentrations (100 mM) of diclofenac, which was biotransformed to a reactive metabolite and subsequently formed covalent adducts to hepatocellular proteins. Upon contact with these diclofenac-modified hepatocytes, the primed lymphocytes responded with a proliferative burst and an upregulation of interleukin-2 receptor expression, both specific markers of T-cell activation. Furthermore, the activated T cells were able to kill the diclofenac-pretreated hepatocytes as demonstrated by a delayed increase in ALT release from injured hepatocytes. Prior incubation of diclofenacexposed hepatocytes with an anti-MHC I antibody afforded partial protection against T-cellmediated cell killing. This indicates that hepatocyte injury was, at least in part, dependent on the T-cell receptor that recognized MHC class I-associated diclofenac-modified peptides, rather

425

Nonsteroidal Anti-inflammatory Drugs

Syngeneic hepatocytes exposed to

Diclofenac-KLH conjugate

Vehicle

Diclofenac Cytotoxic T cells added to

Diclofenac + anti-MHC I antibody

Hepatocyte lysis NO

YES

Partial protection

FIGURE 15 Cytotoxic T cells from diclofenac-immunized mice kill hepatocytes previously exposed to diclofenac in vitro. Immunization alone or exposure to diclofenac alone did not cause hepatocyte injury.

than being a nonspecific effect. A number of other observations also suggest that in this model cytolytic T cells were primarily involved as effectors of hepatocyte injury. For example, the highest degree of hepatocyte lysis was achieved when a T-cell-enriched fraction was used, as opposed to a crude splenocyte fraction. Furthermore, the supernatant from an activated lymphocyte culture, containing cytokines and other soluble mediators, alone had no damaging effect on hepatocytes unless effector cells were added. Importantly, control experiments revealed that immunization with diclofenac alone or exposure of hepatocytes to diclofenac alone was not sufficient to induce hepatocyte injury. However, because high effector cell/target cell ratios were required to elicit the cytotoxic effects, and because both the extent and molecular identity of the diclofenac adducts (or other forms of intracellular stress) generated by high concentrations of diclofenac in vitro are not directly comparable with those following in vivo exposure to diclofenac, caution has to be exerted in interpreting the data as a general mechanism. Finally, another possible effector mechanism of NSAID-induced immune-mediated liver injury could involve cytokines. For example, lymphokine release by activated Th cells, which would recruit and activate macrophages, could result in tissue damage and inflammation. Furthermore, immunoregulatory cytokines (e.g., IL-10 and IL-4) could be involved in sensitizing individuals to immune-mediated liver injury, because there are polymorphisms and great interindividual variation in the expression of these cytokines. For example, low IL-10 levels could increase antigen presentation and thus activate T cells. Alternatively, high IL-4 levels could promote Th2-mediated immune responses and induce B-cell differentiation (164). That the expression levels of some of these immunoregulatory and pro-/anti-inflammatory cytokines can indeed greatly influence the susceptibility to hepatotoxicants was demonstrated in IL-10-deficient mice treated with acetaminophen, which resulted in potentiation of liver damage (186). Collectively, the evidence for immune-mediated reactions that are involved in NSAIDinduced liver toxicity stems primarily from clinical criteria (delayed onset of disease, appearance of rash, fever, eosinophilia, rapid onset of symptoms after rechallenge with drug). When these criteria are employed, a number of NSAIDs, including clometacine, phenylbutazone, diclofenac, naproxen, piroxicam, and tolmetin, all exhibit, at least to some degree, signs of hypersensitivity. The molecular pathways that can trigger such an immune response and, even more important, the mechanisms underlying breaking of immune tolerance, are, however, poorly understood. Impairment of Canalicular Transmembrane Transport of Endogenous Compounds and Retention of Toxic Bile Acids Some NSAIDs can interfere directly with the hepatobiliary transport of endogenous or exogenous cholephilic compounds, such as bile salts (Fig. 6). For example, sulindac at high

426

Boelsterli

doses not only inhibits the basolateral uptake of bile salts, but also inhibits the ATP-dependent canalicular export of conjugated bile acids (81). In rats, this drug can induce cholestasis, leading to retention of bile salts in the hepatocytes. If toxic (hydrophobic) bile salts accumulate, they can pose an oxidative stress (187) and/or induce apoptosis and necrosis (188) by Fas-mediated pathways (189). Alternatively, NSAIDs may impair the excretion of cholephilic compounds by pathways other than direct inhibition of the bile salt transporter, including secondary hepatic effects due to gastrointestinal injury. Since most NSAIDs affect the gastrointestinal epithelium and cause bleeding and ulceration in sensitive species even at low doses (76,190), intestinal injury could entail increases in the permeability of the small intestine. Increased release of bacterial endotoxin can not only cause cholestasis and impairment of liver function, but also downregulate some hepatobiliary carriers including the conjugate export pump, Mrp2; this can lead to an accumulation of the drug or drug conjugate in the hepatocyte (Fig. 16). Release of Lipopolysaccharide and Inflammatory Changes Interestingly, a systematic in vivo study on the overall gene expression pattern (transcriptional profile) involving a number of NSAIDs confirmed that certain NSAIDs exhibited an “LPS-like” fingerprint, most likely resulting from endotoxin (lipopolysaccharide, LPS) leakage following damage to the GI tract (191). Indeed, a single low dose (1.5 mg/kg of diclofenac) can induce small intestinal ulceration in laboratory rodents (Fig. 16) (76,190). In this context, a series of interesting experiments have recently identified mild endotoxemia (caused by administration of low, nonhepatotoxic doses of LPS) as a risk factor that potentiates the hepatotoxicity of a variety of drugs in rodents (192). It remains to be investigated whether clinically silent levels of LPS may play a role in NSAID-induced liver injury in susceptible patients. Other Mechanisms Formation of Mixed Lipids Since some carboxylic acids can be activated to acyl-CoA thioesters, they can enter the pathways of lipid biosynthesis, similar to endogenous fatty acids activated by CoA. For example, a number of 2-arylpropionic acids that are substrates for acyl-CoA ligase thus become incorporated into “hybrid” triacylglycerol (Fig. 17) (193,194). This incorporation and accumulation of NSAIDs in endogenous lipid is stereoselective; only the R-enantiomer is activated and incorporated into lipids (47,195,196). Because hybrid triacylglycerides have the potential to form long-lasting residues in adipose tissues and to be incorporated into biomembranes, where they may disturb membrane function, their formation has been NSAID NSAID/ NSAID metabolites Increased permeability Ulceration

LPS

Cytokine release Acute phase response/Inflammation Downregulation of carriers/cholestasis

FIGURE 16 NSAID-associated enteropathy can lead to increased release of bacterial lipopolysaccharide and trigger secondary hepatic effects including downregulation of hepatobiliary transporters.

427

Nonsteroidal Anti-inflammatory Drugs

R

O O-CH2

Palmitoyl-O-CH

Palmitoyl-O-CH2

Toxicological consequences?

FIGURE 17 Steroselective activation by CoA and incorporation of 2-arylpropionic acids into hybrid triacylglycerols.

considered a potential pathway for toxicity (45,197,198). However, no clear mechanistic link to the hepatic toxicity of NSAIDs has been established. Predisposition by Hepatic Changes Caused by Rheumatic Diseases for Which the NSAIDs Are Prescribed In search of mechanisms of toxicity, it is often overlooked that the treated individuals are patients whose liver function may already be altered due to the disease and prior to the intake of the drug. For example, in patients with acute rheumatic fever, NSAID disposition can be altered: it has been shown that the unbound fraction of the drug in plasma was higher than that in healthy persons. This condition was not only inversely related to the serum albumin concentration, but also showed a positive correlation with increased aminotransferases (199). In fact, liver tests in rheumatoid arthritis patients or in systemic lupus erythematosus (SLE) patients are often abnormal, featuring increased serum activities of alkaline phosphatase and other liver-selective enzymes (13,200–202). Similarly, in an experimental rat model of rheumatoic arthritis, isolated mitochondria from adjuvant-induced arthritic rats were more susceptible to the mitochondria-toxic response elicited by nimesulide than mitochondria from normal healthy rats (203). The underlying mechanisms are not understood.

IDIOSYNCRATIC LIVER TOXICITY CAUSED BY NSAIDS The frequently used (and often misused) term “idiosyncratic” liver injury implies that toxicity occurs in a very small population subset and that the etiology is unknown. More importantly, it also implies that host-dependent factors govern whether the drug is well tolerated or liver injury will ensue. Traditionally, NSAID-associated idiosyncratic reactions have been considered largely dose independent and have been subdivided into metabolic idiosyncrasy and hypersensitivity reactions (4,12,16). More recent concepts have, however, slightly altered the view about the mechanisms underlying these rare cases of toxicity (204). In particular, there is growing evidence that idiosyncratic liver toxicity is dose-dependent too. Since “idiosyncrasy” does not refer to a mechanism, it is proposed here not to use this term any more in the context of mechanistic explanations of drug effects. Most NSAIDs that rarely cause liver injury are intrinsically toxic. In the vast majority of individuals, however, the risk is small and toxicity does not manifest itself because of a powerful system that mediates tolerance and/or secures cellular defense and repair, or simply because the exposure is too small. It is only a small subset of patients who exhibit a specific genetically determined abnormality or an acquired proclivity for altered toxicokinetics or toxicodynamics or for abnormal immune responses. If, in these patients, unusually large amounts of reactive metabolites are generated, combined with low levels of detoxifying pathways or abnormalities in mitochondrial function, or if factors causing intracellular accumulation or impaired transport prevail, then NSAIDs will eventually precipitate hepatic toxicity by the mechanisms discussed above (Fig. 18).

428

Boelsterli

Genetic factors

Environmental factors

Drug-metabolizing enzymes Hepatobiliary transporters Cellular defense systems Immune system Mitochondrial abnormalities

Underlying disease Infections Age Co-medication Acquired cholestasis Regulation of gene expression

NSAID-induced “Danger Signal” High dose Cellular stress Mitochondrial toxicity High adduct levels Cytotoxicity

FIGURE 18 NSAID-induced idiosyncratic reactions are multifactorial and become manifested when several critical risk factors (both genetically determined and acquired) are simultaneously expressed in an individual (overlapping area). The intrinsic toxicity of NSAIDs does not normally become apparent, owing to immune tolerance and/or cellular defense and repair systems.

For these reasons, most of the data gathered on molecular and cellular mechanisms of NSAID-induced liver injury simply describe a molecular hazard, which, in the clinical setting and in the vast majority of patients will exhibit a low penetrance and small expressivity. This is also one of the reasons why it has proven extremely difficult to reproduce NSAIDinduced liver damage in animal models. Only if we can mimic some of the underlying risk factors and actually use animal models of human disease that exhibit the critical determinants of susceptibility, then it will be possible to study the underlying mechanisms and pathogenesis of NSAID-associated hepatotoxicity. One example towards this goal is the heterozygous superoxide dismutase 2- (Sod2C/K) deficient mouse, which exhibits mild mitochondrial functional abnormalities but otherwise appears phenotypically normal. Using this model, it was possible to induce mitochondrial oxidative damage and proapoptotic events in the liver of mice chronically treated with low doses of nimesulide, whereas normal healthy mice (wild-type littermates) did not exhibit mitochondrial oxidative injury upon exposure to nimesulide (134). Thus, translated to a (hypothetical) model, this would imply that, in normal healthy individuals, NSAIDs would not lead to any hepatic toxicity, while in other individuals who may feature compromised mitochondrial function (genetic abnormalities, acquired cumulative damage), the adverse effects of a mitochondria-targeting NSAID would be superimposed on the underlying mitochondrial damage and ultimately precipitate liver injury (Fig. 19). The most promising approach, however, will be to start from the patient side and to analyze the toxicogenetics and expression profile of critical genes in those patients who developed rare NSAID-associated liver disease, and to compare the results with those of patients who tolerate the drug well. Identification of the “critical” genes remains a major challenge for the future.

CONCLUSIONS NSAID-induced hepatotoxicity has become a paradigm for studying drug-induced liver injury. The reason for this is not necessarily that these drugs are more toxic, but they are more frequently used than other drugs, and that, therefore, the number of reported hepatic adverse effects has become significant.

429

Nonsteroidal Anti-inflammatory Drugs

Normal mitochondria (Sod2 +/+ mice)

Genetically abnormal mitochondria (Sod2 +/- mice) mtDNA

Oxidative damage to mtDNA

II I ETC

O III

IV

2

V Decreased ETC function Nimesulide (targeting mitochondria)

+

Mitochondrial injury

No manifestation of mitochondrial damage Threshold (critical mass of damaged mitochondria)

• mt aconitase activity • mt protein carbonyls • cytosolic cytochrome c • caspase-3 activity • TUNEL-positive cells (DNA cleavage)

Delayed manifestation of toxicity ??

FIGURE 19 Nimesulide-induced hepatic mitochondrial injury in mice repeatedly treated with nimesulide (10 mg/kg per day, i.p., bid, for 28 days). Normal wild-type mice did not exhibit any mitochondrial changes, while mice with an underlying clinically silent impairment of mitochondrial function (heterozygous Sod2 deficiency) developed oxidative damage to mitochondria and proapoptotic changes in the liver. Abbreviations: ETC, electron transport chain; SOD2, superoxide dismutase 2. Source: From Ref. 134.

It has become clear that multiple mechanisms are involved in NSAID toxicity to the liver and not a single mechanism can be advocated for the adverse effects. Although mitochondrial injury, immune-mediated toxicity, and impaired hepatobiliary transport have all been discussed as potential mechanisms contributing to the toxicity, the evidence has remained circumstantial. On this basis, a toxic hazard has been recognized for many NSAIDs. Nevertheless, although the tools for both detecting mechanisms and calculating drug exposure have been improved, risk assessment and a reliable prediction of hepatic effects caused by NSAIDs in humans have remained difficult for a number of reasons. First, in spite of vigorous efforts, no animal model is available to study the hepatic toxicity of NSAIDs, in particular, immune-mediated mechanisms (205). Furthermore, there is no clear correlation between the in vitro toxicity (e.g., mitochondrial toxicity) and the reported incidence of hepatic toxicity in patients. Finally, and importantly, it is impossible to date to define and detect the individual risk factors in the patient population. Prediction of NSAID hepatotoxicity can be made at two different levels; at the drug level and patient level. Prediction for a chemical can only be made on the basis of a number of factors. These include (a) the formation of acyl glucuronides and iso-glucuronides with a relatively high protein reactivity (137,163), (b) other reactive metabolites, (c) the formation of canalicular membrane protein adducts in the molecular mass range of 110 to 120 kDa, and a high degree of entero-hepatic circulation, and (d) the potential to uncouple mitochondrial ox-phos and induce the mPT at low (mM) concentrations. Prediction at the patient level is much more difficult. However, the advent of genomics, technologies that will allow us to detect specific genetic abnormalities, as well as novel techniques to detect the regulation of certain key genes, will

430

Boelsterli

hopefully lead to a better understanding of how and why NSAIDs can precipitate liver injury in some susceptible patients. REFERENCES 1. O’Brien WM, Bagby GF. Rare adverse reactions to nonsteroidal antiinflammatory drugs. J Rheumatol 1985; 12:562–7. 2. Doube A. Hepatitis and non-steroidal anti-inflammatory drugs. Ann Rheum Dis 1990; 49:489–90. 3. Tolman KG. Hepatotoxicity of antirheumatic drugs. J Rheumatol 1990; 17(Suppl. 22):6–11. 4. Zimmerman HJ. Update of hepatotoxicity due to classes of drugs in common clinical use: nonsteroidal drugs, anti-inflammatory drugs, antibiotics, antihypertensives, and cardiac and psychotropic agents. Sem Liver Dis 1990; 10:322–38. 5. Garcia Rodriguez LA, Gutthann SP, Walker AM, et al. The role of non-steroidal anti-inflammatory drugs in acute liver injury. Br Med J 1992; 305:865–8. 6. Rabinovitz M, VanThiel DH. Hepatotoxicity of nonsteroidal anti-inflammatory drugs. Am J Gastroenterol 1992; 87:1696–704. 7. Carson JL, Strom BL, Duff A, et al. Safety of nonsteroidal anti-inflammatory drugs with respect to acute liver disease. Arch Intern Med 1993; 153:1331–6. 8. Farrell GC. Liver disease produced by non-steroidal anti-inflammatory drugs. In: Farrell GC, ed. Drug-Induced Liver Disease. Edinbourgh: Churchill Livingstone, 1994:371–88. 9. Garcia Rodriguez LA, Williams R, Derby LE, et al. Acute liver injury associated with nonsteroidal anti-inflammatory drugs and the role of risk factors. Arch Intern Med 1994; 154:311–6. 10. Johnson AG, Seidemann P, Day RO. NSAID-related adverse drug interactions with clinical relevance—an update. Int J Clin Pharmacol Ther 1994; 32:509–32. 11. Singh G, Ramey DR, Morfeld D, et al. Comparative toxicity of non-steroidal anti-inflammatory agents. Pharmacol Ther 1994; 62:175–91. 12. Zimmerman HJ. Hepatic injury associated with nonsteroidal anti-inflammatory drugs. In: Lewis AJ, Furst DE, eds. Nonsteroidal Anti-Inflammatory Drugs. New York: Marcel Dekker, 1994:171–94. 13. Fry SW, Seeff LB. Hepatotoxicity of analgesics and anti-inflammatory agents. Gastroenterol Clin North Am 1995; 24:875–905. 14. Manoukian AV, Carson JL. Nonsteroidal anti-inflammatory drug-induced hepatic disorders. Drug Saf 1996; 15:64–71. 15. Walker AM. Quantitative studies of the risk of serious hepatic injury in persons using nonsteroidal antiinflammatory drugs. Arthritis Rheum 1997; 40:201–8. 16. Zimmerman HJ. Hepatotoxicity. The Adverse Effects of Drugs and Other Chemicals on the Liver. 2nd ed. Philadelphia: Lippincott Williams & Wilkins, 1999. 17. Lewis JH. Nonsteroidal anti-inflammatory drugs: pathology and clinical presentation of hepatotoxicity. In: Kaplowitz N, DeLeve LD, eds. Drug-Induced Liver Disease. New York: Marcel Dekker, 2003:377–404. 18. Lacroix I, Lapeyre-Mestre M, Bagheri H, et al. Nonsteroidal anti-inflammatory drug-induced liver injury: a case-control study in primary care. Fundam Clin Pharmacol 2004; 18:201–6. 19. Rubenstein JH, Laine L. The hepatotoxicity of non-steroidal anti-inflammatory drugs. Aliment Pharmacol Ther 2004; 20:373–80. 20. Katsinelos P, Katsos I, Patsiaoura K, et al. Tenoxicam-associated hepatic injury: a case report and review. Eur J Gastroenterol Hepatol 1997; 9:403–6. 21. Kra¨henbu¨hl S, Reichen J. Drug hepatotoxicity. In: Bacon BR, DiBisceglie AM, eds. Liver Disease Diagnosis and Management. New York: Churchill Livingstone, 2000:294–309. 22. Sgro C, Clinard F, Ouazir K, et al. Incidence of drug-induced hepatic injuries: a French populationbased study. Hepatology 2002; 36:451–5. 23. Friis H, Andreasen PB. Drug-induced hepatic injury—an analysis of 1100 cases reported to the Danish Committee on Adverse Drug Reactions between 1978 and 1987. J Intern Med 1992; 232:133–8. 24. Paulus HE. FDA Arthritis Advisory Committee meeting. Arthritis Rheum 1982; 25:1124–5. 25. Tolman KG. Hepatotoxicity of non-narcotic analgesics. Am J Med 1998; 105:13S–19. 26. Boelsterli UA, Zimmerman HJ, Kretz-Rommel A. Idiosyncratic liver toxicity of nonsteroidal antiinflammatory drugs: molecular mechanisms and pathology. Crit Rev Toxicol 1995; 25:207–35. 27. Boelsterli UA. Mechanisms of NSAID-induced hepatotoxicity—focus on nimesulide. Drug Saf 2002; 25:633–48. 28. Galati G, Tafazoli S, Sabzevari O, et al. Idiosyncratic NSAID drug induced oxidative stress. Chem Biol Interact 2002; 142:25–41. 29. Borga O, Borga B. Serum protein binding of nonsteroidal antiinflammatory drugs: a comparative study. J Pharmacokinet Biopharm 1997; 25:63–77.

Nonsteroidal Anti-inflammatory Drugs

431

30. Vree TB, Van den Biggelaar-Martea M, Verwey-Van Wissen CPWGM, et al. The pharmacokinetics of naproxen, its metabolite O-desmethylnaproxen, and their acylglucuronides in humans. Effect of cimetidine. Br J Clin Pharmacol 1993; 35:467–72. 31. Siraki AG, Chevaldina T, O’Brien PJ. Application of quantitative structure-toxicity relationships for acute NSAID cytotoxicity in rat hepatocytes. Chem Biol Interact 2005; 151:177–91. 32. Jurima-Romet M, Crawford K, Huang HS. Comparative cytotoxicity of non-steroidal anti-inflammatory drugs in primary cultures of rat hepatocytes. Toxicol In Vitro 1994; 8:55–66. 33. Sorensen EMB, Acosta D. Relative toxicities of several nonsteroidal antiinflammatory compounds in primary cultures of rat hepatocytes. J Toxicol Environ Health 1985; 16:425–40. 34. Shen S, Marchick MR, Davis MR, et al. Metabolic activation of diclofenac by human cytochrome P450 3A4: role of 5-hydroxydiclofenac. Chem Res Toxicol 1999; 12:214–22. 35. Tang W. The metabolism of diclofenac—enzymology and toxicology perspectives. Curr Drug Metab 2003; 4:319–29. 36. Miyamoto G, Zahid N, Uetrecht JP. Oxidation of diclofenac to reactive intermediates by neutrophils, myeloperoxidase, and hypochlorous acid. Chem Res Toxicol 1997; 10:414–9. 37. Tang W, Stearns RA, Bandiera SM, et al. Studies on cytochrome P-450-mediated bioactivation of diclofenac in rats and in human hepatocytes: identification of glutathione conjugated metabolites. Drug Metab Dispos 1999; 27:365–72. 38. Tang W, Stearns RA, Wang RW, et al. Roles of human hepatic cytochrome P450s 2C9 and 3A4 in the metabolic activation of diclofenac. Chem Res Toxicol 1999; 12:192–9. 39. Yan Z, Li JL, Huebert N, et al. Detection of a novel reactive metabolite of diclofenac: evidence for CYP2C9-mediated bioactivation via arene oxides. Drug Metab Dispos 2005; 33:706–13. 40. Shen SJ, Hargus SJ, Martin BM, et al. Cytochrome P4502C11 is a target of diclofenac covalent binding in rats. Chem Res Toxicol 1997; 10:420–3. 41. Tafazoli S, O’Brien PJ. Peroxidases: a role in the metabolism and side effects of drugs. Drug Discov Today 2005; 10:617–25. 42. O’Brien PJ. Peroxidases. Chem Biol Interact 2000; 129:113–9. 43. Sevoz C, Benoit E, Buronfosse T. Thioesterification of 2-arylpropionic acids by recombinant acylcoenzyme A synthetases (ACS1 and ACS2). Drug Metab Dispos 2000; 28:398–402. 44. Hall SD, Qian XT. The role of coenzyme A in the biotransformation of 2-arylpropionic acids. Chem Biol Interact 1994; 90:235–51. 45. Mayer JM, Roy-DeVos M, Audergon C, et al. Interactions of antiinflammatory 2-arylpropionates (Profens) with the metabolism of fatty acids-in vitro studies. Int J Tissue React-Exp Clin Aspects 1994; 16:59–72. 46. Caldwell J, Marsh MV. Interrelationships between xenobiotic metabolism and lipid biosynthesis. Biochem Pharmacol 1983; 32:1667–72. 47. Mayer JM, Roy-De Vos M, Audergon C, et al. Interactions of anti-inflammatory 2-arylpropionates (profens) with the metabolism of fatty acids: in vitro studies. Int J Tissue React-Exp Clin Aspects 1994; 16:59–72. 48. Sallustio BC, Nunthasomboon S, Drogemuller CJ, et al. In vitro covalent binding of nafenopin-CoA to human liver proteins. Toxicol Appl Pharmacol 2000; 163:176–82. 49. Hayball PJ. Chirality and nonsteroidal anti-inflammatory drugs. Drugs 1996; 52:47–58. 50. Li C, Grillo MP, Benet LZ. In vivo mechanistic studies on the metabolic activation of 2-phenylpropionic acid in rat. J Pharmacol Exp Ther 2003; 305:250–6. 51. Mackenzie PI, Owens IS, Burchell B, et al. The UDP glycosyltransferase gene superfamily: recommended nomenclature update based on evolutionary divergence. Pharmacogenetics 1997; 7:255–69. 52. Jin C, Miners JO, Lillywhite KJ, et al. Complementary deoxyribonucleic acid cloning and expression of a human liver uridine diphosphate-glucuronosyltransferase glucuronidating carboxylic acidcontaining drugs. J Pharmacol Exp Ther 1993; 264:475–9. 53. Burchell B, McGurk K, Brierley CH, et al. UDP-glucuronosyltransferases. In: Sipes IG, McQueen CA, Gandolfi AJ, eds. Comprehensive Toxicology. New York: Elsevier, 1997:401–35. 54. Kuehl GE, Lampe JW, Potter JD, et al. Glucuronidation of nonsteroidal anti-inflammatory drugs: identifying the enzymes responsible in human liver microsomes. Drug Metab Dispos 2005; 33:1027–35. 55. Le HT, Franklin MR. Selective induction of phase II drug metabolizing enzyme activities by quinolines and isoquinolines. Chem Biol Interact 1997; 103:167–78. 56. Patel M, Tang BK, Grant DM, et al. Interindividual variability in the glucuronidation of (S) oxazepam contrasted with that of (R) oxazepam. Pharmacogenetics 1995; 5:287–97. 57. Coffman BL, King CD, Rios GR, et al. The glucuronidation of opioids, other xenobiotics, and androgens by human UGT2B7Y(268) and UGT2B7(268). Drug Metab Dispos 1998; 26:73–7. 58. Kirkman SK, Zhang MY, Horwatt PM, et al. Isolation and identification of bromfenac glucoside from rat bile. Drug Metab Dispos 1998; 26:720–3.

432

Boelsterli

59. Fontana RJ, McCashland TM, Brenner KG, et al. Acute liver failure associated with prolonged use of bromfenac leading to liver transplantation. Liver Transpl Surg 1999; 5:480–4. 60. Moses PL, Schroeder B, Alkhatib O, et al. Severe hepatotoxicity associated with bromfenac sodium. Am J Gastroenterol 1999; 94:1393–6. 61. Skjodt NM, Davies NM. Clinical pharmacokinetics and pharmacodynamics of bromfenac. Clin Pharmacokinet 1999; 36:399–408. 62. Grillo MP, Knutson CG, Sanders PE, et al. Studies on the chemical reactivity of diclofenac acyl glucuronide with glutathione: identification of diclofenac-S-acyl-glutathione in rat bile. Drug Metab Dispos 2003; 31:1327–36. 63. Grillo MP, Hua F, Knutson CG, et al. Mechanistic studies on the bioactivation of diclofenac: identification of diclofenac-S-acyl-glutathione in vitro in incubations with rat and human hepatocytes. Chem Res Toxicol 2003; 16:1410–7. 64. Watt JA, King AR, Dickinson RG. Contrasting systemic stabilities of the acyl and phenolic glucuronides of diflunisal in the rat. Xenobiotica 1991; 21:403–15. 65. Georges H, Presle N, Buronfosse T, et al. In vitro stereoselective degradation of carprofen glucuronide by human serum albumin. Characterization of sites and reactive amino acids. Chirality 2000; 12:53–62. 66. Hayball PJ. Formation and reactivity of acyl glucuronides: the influence of chirality. Chirality 1995; 7:1–9. 67. Iwaki M, Bischer A, Nguyen AC, et al. Stereoselective disposition of naproxen glucuronide in the rat. Drug Metab Dispos 1995; 23:1099–103. 68. Seitz S, Kretz-Rommel A, Oude Elferink RPJ, et al. Selective protein adduct formation of diclofenac glucuronide is critically dependent on the rat canalicular conjugate export pump (Mrp2). Chem Res Toxicol 1998; 11:513–9. 69. Kauffmann HM, Keppler D, Gant TW, et al. Induction of hepatic mrp2 (cmrp/cmoat) gene expression in nonhuman primates treated with rifampicin or tamoxifen. Arch Toxicol 1998; 72:763–8. 70. Rost D, Kartenbeck J, Keppler D. Changes in the localization of the rat canalicular conjugate export pump Mrp2 in phalloidin-induced cholestasis. Hepatology 1999; 29:814–21. 71. Paulusma CC, Bosma PJ, Zaman GJR, et al. Congenital jaundice in rats with a mutation in a multidrug resistance-associated protein gene. Science 1996; 271:1126–8. 72. Wang M, Dickinson RG. Bile duct ligation promotes covalent drug–protein adduct formation in plasma but not in liver of rats given zomepirac. Life Sci 2000; 68:525–37. 73. Yesair DW, Callahan M, Remington L, et al. Role of the entero-hepatic cycle of indomethacin on its metabolism, distribution in tissues and its excretion by rats, dogs and monkeys. Biochem Pharmacol 1970; 19:1579–90. 74. Brune K. Is there a rational basis for the different spectra of adverse effects of nonsteroidal antiinflammatory drugs (NSAIDs)? Drugs 1990; 40(Suppl. 5):12–15. 75. Reuter BK, Davies NM, Wallace JL. Nonsteroidal anti-inflammatory drug enteropathy in rats: role of permeability, bacteria, and enterohepatic circulation. Gastroenterology 1997; 112:109–17. 76. Seitz S, Boelsterli UA. Diclofenac acyl glucuronide, a major biliary metabolite, is directly involved in small intestinal injury in rats. Gastroenterology 1998; 115:1476–82. 77. MacPhail CM, Lappin MR, Meyer DJ, et al. Hepatocellular toxicosis associated with administration of carprofen in 21 dogs. J Am Vet Med Assoc 1998; 212:1895–901. 78. Priymenko N, Garnier F, Ferre JP, et al. Enantioselectivity of the enterohepatic recycling of carprofen in the dog. Drug Metab Dispos 1998; 26:170–6. 79. Khan SA, Khan MA, Albretsen JC, et al. Naproxen toxicosis in dogs: a review of fifty-seven cases. Toxicol Sci 1999; 48(Suppl.):32 (abstract). 80. Dobrinska MR, Furst DE, Spiegel T, et al. Biliary secretion of sulindac and metabolites in man. Biopharm Drug Dispos 1983; 4:347–58. 81. Bolder U, Trang NV, Hagey LR, et al. Sulindac is excreted into bile by a canalicular bile salt pump and undergoes a cholehepatic circulation in rats. Gastroenterology 1999; 117:962–71. 82. Smith PC, McDonagh AF, Benet LZ. Irreversible binding of zomepirac to plasma protein in vitro and in vivo. J Clin Invest 1986; 77:934–9. 83. Needleman P, Isakson PC. The discovery and function of COX-2. J Rheumatol 1997; 24:6–8. 84. Pessayre D, Mazel P, Descatoire V, et al. Inhibition of hepatic drug-metabolizing enzymes by arachidonic acid. Xenobiotica 1979; 9:301–10. 85. Gasperini R, Leone R, Velo GP, et al. The inhibition of hepatic microsomal drug metabolism in rats by non-steroidal anti-inflammatory drugs. Pharmacol Res 1990; 22(Suppl. 3):115–6. 86. Hagmann W, Kirn A, Keppler D. Role of leukotrienes in acute inflammatory liver disease. In: Reutter W, Heinrich PC, Popper H, Keppler D, Arias IM, Landmann L, eds. Modulation of Liver Cell Expression. Lancaster: MTP Press, 1987:423–33. 87. Trudell JR, Gut J, Costa AK. Leukotrienes in hepatocyte injury. In: Reutter W, Heinrich PC, Popper H, Keppler D, Arias IM, Landmann L, eds. Modulation of Liver Cell Expression. Lancaster: MTP Press, 1987:411–21.

Nonsteroidal Anti-inflammatory Drugs

433

88. Reilly TP, Brady JN, Marchik MR, et al. A protective role for cyclooxygenase-2 in drug-induced liver injury in mice. Chem Res Toxicol 2001; 14:1620–8. 89. Foxworthy PS, Perry DN, Eacho PI. Induction of peroxisomal b-oxidation by nonsteroidal antiinflammatory drugs. Toxicol Appl Pharmacol 1993; 118:271–4. 90. Vidal-Puig AJ, Considine RV, Jimenez-Lin˜an M, et al. Peroxisome proliferator-activated receptor gene expression in human tissues. Effects of obesity, weight loss, and regulation by insulin and glucocorticoids. J Clin Invest 1997; 99:2416–22. 91. Bedoucha M, Atzpodien E, Boelsterli UA. Diabetic KKAy mice exhibit increased hepatic PPARg1 gene expression and develop hepatic steatosis upon chronic treatment with antidiabetic thiazolidinediones. J Hepatol 2001; 35:17–23. 92. Lehmann JM, Lenhard JM, Oliver BB, et al. Peroxisome proliferator-activated receptors a and g are activated by indomethacin and other non-steroidal anti-inflammatory drugs. J Biol Chem 1997; 272:3406–10. 93. Rahman MA, Dhar DK, Masunaga R, et al. Sulindac and exisulind exhibit a significant antiproliferative effect and induce apoptosis in human hepatocellular carcinoma cell lines. Cancer Res 2000; 60:2085–9. 94. Shiff SJ, Qiao L, Tsai L, et al. Sulindac sulfide, an aspirin-like compound, inhibits proliferation, causes cell cylcle quiescence, and induces apoptosis in HT-29 colon adenocarcinoma cells. J Clin Invest 1995; 96:491–503. 95. Hanif R, Pittas A, Feng Y, et al. Effects of nonsteroidal anti-inflammatory drugs on proliferation and on induction of apoptosis in colon cancer cells by a prostaglandin-independent pathway. Biochem Pharmacol 1996; 52:237–45. 96. Shiff SJ, Koutsos MI, Qiao L, et al. Nonsteroidal anti-inflammatory drugs inhibit the proliferation of colon adenocarcinoma cells: effects on cell cycle and apoptosis. Exp Cell Res 1996; 222:179–88. 97. Zhu GH, Wong BCY, Ching CK, et al. Differential apoptosis by indomethacin in gastric epithelial cells through the constitutive expression of wild-type p53 and/or upregulation of c-myc. Biochem Pharmacol 1999; 58:193–200. 98. McDougall P, Markham A, Cameron I, et al. The mechanism of inhibition of mitochondrial oxidative phosphorylation by the non-steroidal anti-inflammatory agent diflunisal. Biochem Pharmacol 1983; 32:2595–8. 99. Browne GS, Nelson C, Nguyen T, et al. Stereoselective and substrate-dependent inhibition of hepatic mitochondrial b-oxidation and oxidative phosphorylation by the non-steroidal anti-inflammatory drugs ibuprofen, flurbiprofen, and ketorolac. Biochem Pharmacol 1999; 57:837–44. 100. Mahmud T, Scott DL, Bjarnason I. A unifying hypothesis for the mechanism of NSAID related gastrointestinal toxicity. Ann Rheum Dis 1996; 55:211–3. 101. Petrescu I, Tarba C. Uncoupling effects of diclofenac and aspirin in the perfused liver and isolated hepatic mitochondria of rat. Biochim Biophys Acta (Bioenergetics) 1997; 1318:385–94. 102. Moreno-Sanchez R, Bravo C, Vasquez C, et al. Inhibition and uncoupling of oxidative phosphorylation by nonsteroidal anti-inflammatory drugs. Biochem Pharmacol 1999; 57:743–52. 103. Tokomitsu Y, Lee S, Ui M. In vitro effects of nonsteroidal anti-inflammatory drugs on oxidative phosphorylation in rat liver mitochondria. Biochem Pharmacol 1977; 26:2101–6. 104. Knights KM, Drew R. The effects of ibuprofen enantiomers on hepatocyte intermediary metabolism and mitochondrial respiration. Biochem Pharmacol 1992; 44:1291–6. 105. Masubuchi Y, Saito H, Horie T. Structural requirements for the hepatotoxicity of nonsteroidal antiinflammatory drugs in isolated rat hepatocytes. J Pharmacol Exp Ther 1998; 287:208–13. 106. Ponsoda X, Bort R, Jover R, et al. Molecular mechanism of diclofenac hepatotoxicity: association of cell injury with oxidative metabolism and decrease in ATP levels. Toxicol In Vitro 1995; 9:439–44. 107. Masubuchi Y, Yamada S, Horie T. Possible mechanism of hepatocyte injury induced by diphenylamine and its structurally related nonsteroidal anti-inflammatory drugs. J Pharmacol Exp Ther 2000; 292:982–7. 108. Tay VKS, Wang AS, Leow KY, et al. Mitochondrial permeability transition as a source of superoxide anion induced by the nitroaromatic drug nimesulide in vitro. Free Radic Biol Med 2005; 39:949–59. 109. Bjorkman D. Nonsteroidal anti-inflammatory drug-associated toxicity of the liver, lower gastrointestinal tract, and esophagus. Am J Med 1998; 105:17S–21. 110. He L, Lemasters JJ. Regulated and unregulated mitochondrial permeability transition pores: a new paradigm of pore structure and function? FEBS Lett 2002; 512:1–7. 111. Palmeira CM, Rana MI, Frederick CB, et al. Induction of the mitochondrial permeability transition in vitro by short-chain carboxylic acids. Biochem Biophys Res Commun 2000; 272:431–5. 112. Uyemura SA, Santos AC, Mingatto FE, et al. Diclofenac sodium and mefenamic acid: potent inducers of the membrane permeability transition in renal cortex mitochondria. Arch Biochem Biophys 1997; 342:231–5. 113. Al-Nasser IA. Ibuprofen-induced liver mitochondrial permeability transition. Toxicol Lett 2000; 111:213–8.

434

Boelsterli

114. Masubuchi Y, Nakayama S, Horie T. Role of mitochondrial permeability transition in diclofenacinduced hepatocyte injury in rats. Hepatology 2002; 35:544–51. 115. Oh KW, Qian T, Brenner D, et al. Salicylate enhances necrosis and apoptosis mediated by the mitochondrial permeability transition. Toxicol Sci 2003; 73:44–52. 116. Geneve J, Hayat-Bonan B, Labbe G, et al. Inhibition of mitochondrial b-oxidation of fatty acids by pirprofen. Role in microvesicular steatosis due to this nonsteroidal anti-inflammatory drug. J Pharmacol Exp Ther 1987; 242:1133–7. 117. Freneaux E, Fromenty B, Berson A, et al. Stereoselective and nonstereoselective effects of ibuprofen enantiomers on mitochondrial b-oxidation of fatty acids. J Pharmacol Exp Ther 1990; 255:529–35. 118. Zhao B, Geisslinger G, Hall I, et al. The effect of the enantiomers of ibuprofen and flurbiprofen on the b-oxidation of palmitate in the rat. Chirality 1992; 4:137–41. 119. Fraser JL, Antonioli DA, Chopra S, et al. Prevalence and nonspecificity of fatty change in the liver. Mod Pathol 1995; 8:65–70. 120. Fromenty B, Pessayre D. Inhibition of mitochondrial beta-oxidation as a mechanism of hepatotoxicity. Pharmacol Ther 1995; 67:101–54. 121. Fromenty B, Pessayre D. Impaired mitochondrial function in microvesicular steatosis. J Hepatol 1997; 26:43–53. 122. Deschamps D, Fisch C, Fromenty B, et al. Inhibition by salicylic acid of the activation and thus oxidation of long-chain fatty acids. Possible role in the development of Reye’s syndrome. J Pharmacol Exp Ther 1991; 259:894–904. 123. Vianey-Liaud C, Divry N, Gregersen N, et al. The inborn errors of mitochondrial fatty acid oxidation. J Inherit Metab Dis 1987; 10(Suupl. 1):159–98. 124. Danan G, Trunet P, Bernuau J, et al. Pirprofen-induced fulminant hepatitis. Gastroenterology 1985; 89:210–3. 125. Victorino RMM, Silveira JCB, Baptista A, et al. Jaundice associated with naproxen. Postgrad Med J 1980; 56:368. 126. Bravo JF, Jacobson MP, Mertens BF. Fatty liver and pleural effusion with ibuprofen therapy. Ann Intern Med 1977; 87:200–1. 127. Dutertre JP, Bastides F, Jonville AP, et al. Microvesicular steatosis after ketoprofen administration. Eur J Gastroenterol Hepatol 1991; 3:953–4. 128. Fraser JL, Antonioli DA, Chopra S, et al. Prevalence and nonspecificity of microvesicular fatty change in the liver. Mod Pathol 1995; 8:65–70. 129. Thun MJ, Henley SJ, Patrono C. Nonsteroidal anti-inflammatory drugs as anticancer agents: mechanistic, pharmacologic, and clinical issues. J Natl Cancer Inst 2002; 94:252–66. 130. Gomez-Lechon MJ, Ponsoda X, O’Connor E, et al. Diclofenac induces apoptosis in hepatocytes. Toxicol In Vitro 2003; 17:675–80. 131. Gomez-Lechon MJ, Ponsoda X, O’Connor E, et al. Diclofenac induces apoptosis in hepatocytes by alteration of mitochondrial function and generation of ROS. Biochem Pharmacol 2003; 66:2155–67. 132. Inoue A, Muranaka S, Fujita H, et al. Molecular mechanisms of diclofenac-induced apoptosis of promyelocytic leukemia: dependency on reactive oxygen species, Akt, Bid, cytochrome c, and caspase pathway. Free Radic Biol Med 2004; 37:1290–9. 133. Cantoni L, Valaperta R, Ponsoda X, et al. Induction of hepatic heme oxygenase-1 by diclofenac in rodents: role of oxidative stress and cytochrome P-450 activity. J Hepatol 2003; 38:776–83. 134. Ong MMK, Wang AS, Leow KY, et al. Nimesulide-induced hepatic mitochondrial injury in heterozygous Sod2C/K mice. Free Radic Biol Med 2006; 40:420–9. 135. Hyneck ML, Smith PC, Munafo A, et al. Disposition and irreversible plasma protein binding of tolmetin in humans. Clin Pharmacol Ther 1988; 44:107–14. 136. Smith PC, Benet LZ, McDonagh AF. Covalent binding of zomepirac glucuronide to proteins: evidence for a Schiff base mechanism. Drug Metab Dispos 1990; 18:639–44. 137. King AR, Dickinson RG. Studies on the reactivity of acyl glucuronides—IV. Covalent binding of diflunisal to tissues of the rat. Biochem Pharmacol 1993; 45:1043–7. 138. Zia-Amirhosseini P, Ojingwa JC, Spahn-Langguth H, et al. Enhanced covalent binding of tolmetin to proteins in humans after multiple dosing. Clin Pharmacol Ther 1994; 55:21–7. 139. Van Breemen RB, Fenselau C. Acylation of albumin by 1-O-acyl glucuronides. Drug Metab Dispos 1985; 13:318–20. 140. Iwakawa S, Spahn H, Benet LZ, et al. Carprofen acyl glucuronides: stereoselective degradation and interaction with human serum albumin. Pharm Res 1988; 5(Suppl.):S-214. 141. Dickinson RG, King AR. Studies on the reactivity of acyl glucuronides—II. Interaction of diflunisal acyl glucuronide and its isomers with human serum albumin in vitro. Biochem Pharmacol 1991; 42:2301–6. 142. Smith PC, Song WQ, Rodriguez RJ. Covalent binding of etodolac acyl glucuronide to albumin invitro. Drug Metab Dispos 1992; 20:962–5.

Nonsteroidal Anti-inflammatory Drugs

435

143. Ding A, Ojingwa JC, McDonagh AF, et al. Evidence for covalent binding of acyl glucuronides to serum albumin via an imine mechanism as revealed by tandem mass spectrometry. Proc Natl Acad Sci USA 1993; 90:3797–801. 144. Dubois N, Lapicque F, Maurice MH, et al. In vitro irreversible binding of ketoprofen glucuronide to plasma proteins. Drug Metab Dispos 1993; 21:617–23. 145. Smith PC, Liu JH. Covalent binding of suprofen acyl glucuronide to albumin in vitro. Xenobiotica 1993; 23:337–48. 146. Smith PC, Liu JH. Covalent binding of suprofen to renal tissue of rat correlates with excretion of its acyl glucuronide. Xenobiotica 1995; 25:531–40. 147. Ding A, Zia-Amirhosseini P, Mcdonagh AF, et al. Reactivity of tolmetin glucuronide with human serum albumin—identification of binding sites and mechanisms of reaction by tandem mass spectrometry. Drug Metab Dispos 1995; 23:369–76. 148. Yvon M, Wal JM. Identification of lysine residue 199 of human serum albumin as a binding site for benzylpenicilloyl groups. FEBS Lett 1988; 239:237–40. 149. Akira K, Hasegawa H, Shinohara Y, et al. Stereoselective internal acyl migration of 1b-O-acyl glucuronides of enantiomeric 2-phenylpropionic acids. Biol Pharm Bull 2000; 23:506–10. 150. Iwaki M, Ogiso T, Inagawa S, et al. In vitro regioselective stability of b-1-O and 2-O-acyl glucuronides of naproxen and their covalent binding to human serum albumin. J Pharm Sci 1999; 88:52–7. 151. Benet LZ, Spahn-Langguth H, Iwakawa S, et al. Predictability of the covalent binding of acidic drugs in man. Life Sci 1993; 53:PL141–6. 152. Spahn H, Na¨thke I, Mohri K, et al. Preliminary characterization of proteins to which benoxaprofen glucuronide binds irreversibly. Pharm Res 1990; 7(Suppl.):S-257. 153. Pumford NR, Myers TG, Davila JC, et al. Immunochemical detection of liver protein adducts of the nonsteroidal antiinflammatory drug diclofenac. Chem Res Toxicol 1993; 6:147–50. 154. Hargus SJ, Amouzedeh HR, Pumford NR, et al. Metabolic activation and immunochemical localization of liver protein adducts of the nonsteroidal anti-inflammatory drug diclofenac. Chem Res Toxicol 1994; 7:575–82. 155. Kretz-Rommel A, Boelsterli UA. Selective protein adducts to membrane proteins in cultured rat hepatocytes exposed to diclofenac. Radiochemical and immunochemical analysis. Mol Pharmacol 1994; 45:237–44. 156. Hargus SJ, Martin BM, George JW, et al. Covalent modification of rat liver dipeptidyl peptidase IV (CD26) by the nonsteroidal anti-inflammatory drug diclofenac. Chem Res Toxicol 1995; 8:993–6. 157. Wade LT, Kenna JG, Caldwell J. Immunochemical identification of mouse hepatic protein adducts derived from the nonsteroidal anti-inflammatory drugs diclofenac, sulindac, and ibuprofen. Chem Res Toxicol 1997; 10:546–55. 158. Myers TG, Pumford NR, Davila JC, et al. Covalent binding of diclofenac to plasma membrane proteins of the bile canaliculi in the mouse. Toxicologist 1992; 12:253. 159. Caldwell J, Fakurakzi S, Ramsay LA, et al. Benoxaprofen forms protein adducts in the bile canaliculi of female CD1 mice. Toxicol Sci 2000; 54(Suppl.):46 (abstract). 160. Caldwell J, Somchit N, Ramsay LA, et al. Inhibition of canalicular plasma membrane dipeptidyl peptidase IV (DPP IV) is not associated with NSAID-induced hepatotoxicity in rat. In: Proceedings of the 5th International ISSX Meeting. Vol. 13. Cairns, Australia 1998:151. 161. Wang M, Gorrell MD, McCaughan GW, et al. Dipeptidyl peptidase IV is a target for covalent adduct formation with the acyl glucuronide metabolite of the anti-inflammatory drug zomepirac. Life Sci 2001; 68:785–97. 162. Boelsterli UA. Reactive acyl glucuronides: possible role in small intestinal toxicity induced by nonsteroidal anti-inflammatory drugs. Toxic Subst Mech 1999; 18:83–100. 163. Boelsterli UA. Xenobiotic acyl glucuronides and acyl CoA thioesters as protein-reactive metabolites with the potential to cause idiosyncratic drug reactions. Curr Drug Metab 2002; 3:439–50. 164. Aithal GP, Ramsay L, Daly AK, et al. Hepatic adducts, circulating antibodies, and cytokine polymorphisms in patients with diclofenac hepatotoxicity. Hepatology 2004; 39:1430–40. 165. Chiou YJ, Tomer KB, Smith PC. Effect of nonenzymatic glycation of albumin and superoxide dismutase by glucuronic acid and suprofen acyl glucuronide on their functions in vitro. Chem Biol Interact 1999; 121:141–59. 166. Bailey MJ, Worrall S, de Jersey J, et al. Zomepirac acyl glucuronide covalently modifies tubulin in vitro and in vivo and inhibits its assembly in an in vitro system. Chem Biol Interact 1998; 115:153–66. 167. De Weck AL. Immunopathological mechanisms and clinical aspects of allergic reactions to drugs. In: De Weck AL, Bungaard H, eds. Allergic Reactions to Drugs. Berlin: Springer, 1983:75–133. 168. Munafo A, Hyneck ML, Benet LZ. Pharmacokinetics and irreversible binding of tolmetin and its glucuronic acid esters in the elderly. Pharmacology 1993; 47:309–17.

436

Boelsterli

169. McKinnon GE, Dickinson RG. Covalent binding of diflunisal and probenecid to plasma protein in humans: persistence of the adducts in the circulation. Res Commun Chem Pathol Pharmacol 1989; 66:339–54. 170. Dohlman JG, Pillion DJ, Rokeach LA, et al. Identification of macrophage cell-surface binding sites for cationized bovine serum albumin. Biochem Biophys Res Commun 1991; 181:787–96. 171. Franco A, Barnaba V, Natali P, et al. Expression of class I and class II major histocompatibility complex antigens on human hepatocytes. Hepatology 1988; 8:449–54. 172. Harris HW, Gill TJ. Expression of class I transplantation antigens. Transplantation 1986; 42:109–16. 173. Calmus Y, Arvieux C, Gane P, et al. Cholestasis induces major histocompatibility complex class-I expression in hepatocytes. Gastroenterology 1992; 102:1371–7. 174. Arvieux C, Calmus Y, Gane P, et al. Immunogenicity of rat hepatocytes in vivo—effect of cholestasisinduced changes in major histocompatibility complex expression. J Hepatol 1993; 18:335–41. 175. Lobo-Yeo A, Senaldi G, Portmann B, et al. Class I and class II major histocompatibility complex antigen expression on hepatocyes: a study in children with liver disease. Hepatology 1990; 12:224–32. 176. Volpes R, Vandenoord JJ, Desmet VJ. Can hepatocytes serve as activated immunomodulating cells in the immune response. J Hepatol 1992; 16:228–40. 177. Chedid A, Mendenhall CL, Moritz TE, et al. Cell-mediated hepatic injury in alcoholic liver disease. Gastroenterology 1993; 105:254–66. 178. Worrall S, Dickinson RG. Rat serum albumin modified by diflunisal acyl glucuronide is immunogenic in rats. Life Sci 1995; 56:1921–30. 179. Zia-Amirhosseini P, Harris RZ, Brodsky FM, et al. Hypersensitivity to nonsteroidal anti-inflammatory drugs. Nat Med 1995; 1:2–4. 180. Sidrouphis L, Beaugrand M, Malledant Y, Brissot P, Guguen-Guillouzo C, Guillouzo A. Use of adult human hepatocytes in primary culture of the study of clometacin-induced immunoallergic hepatitis. Toxic In Vitro 1991; 5:529–34. 181. Toh BH. Anti-cytoskeletal autoantibodies: diagnostic significance for liver diseases, infections and systemic autoimmune diseases. Autoimmunity 1991; 11:119–25. 182. Hillon P, Bedenne L, Piard F, et al. Clometacine-induced hepatitis. Dig Dis Sci 1988; 33:1045–53. 183. Pessayre D. Physiopathologie des he´patopathies me´dicamenteuses. Gastroente´rol Clin Biol 1993; 17:H3–17. 184. Kretz-Rommel A, Rubin RL. Disruptioin of positive selectioin of thymocytes causes autoimmunity. Nat Med 2000; 6:298–305. 185. Kretz-Rommel A, Boelsterli UA. Cytotoxic activity of T cells and non-T cells from diclofenacimmunized mice against cultured syngeneic hepatocytes exposed to diclofenac. Hepatology 1995; 22:213–22. 186. Bourdi M, Masubuchi Y, Reilly TP, et al. Protection against acetaminophen-induced liver injury and lethality by interleukin 10: role of inducible nitric oxide synthase. Hepatology 2002; 35:289–98. 187. Sokol RJ, Devereaux M, Khandwala R, et al. Evidence for involvement of oxygen free radicals in bile acid toxicity to isolated rat hepatocytes. Hepatology 1993; 17:869–81. 188. Spivey JR, Bronk SF, Gores GJ. Glycochenodeoxycholate-induced lethal hepatocellular injury in rat hepatocytes. Role of ATP depletion and cytosolic free calcium. J Clin Invest 1993; 92:17–24. 189. Miyoshi H, Rust C, Roberts PJ, et al. Hepatocyte apoptosis after bile duct ligation in the mouse involves Fas. Gastroenterology 1999; 117:669–77. 190. Atchison CR, West AB, Balakumaran A, et al. Drug enterocyte adducts: possible causal factors for diclofenac enteropathy in rats. Gastroenterology 2000; 119:1537–47. 191. McMillian M, Nie AY, Parker JB, et al. Inverse gene expression patterns for macrophage activating hepatotoxicants and peroxisome proliferators in rat liver. Biochem Pharmacol 2004; 67:2141–65. 192. Roth RA, Harkema JR, Pestka JP, et al. Is exposure to bacterial endotoxin a determinant of susceptibility to intoxication from xenobiotic agents? Toxicol Appl Pharmacol 1997; 147:300–11. 193. Fears R, Baggaley KH, Alexander R, et al. The participation of ethyl 4-benzoyloxybenzoate (BRL 10894) and other aryl-substituted acids in glycerolipid metabolism. J Lipid Res 1978; 19:3–11. 194. Mayer JM. Ibuprofen enantiomers and lipid metabolism. J Clin Pharmacol 1996; 36:27S–32. 195. Williams K, Day R, Knihinicki R, et al. The stereoselective uptake of ibuprofen enantiomers into adipose tissue. Biochem Pharmacol 1986; 35:3403–5. 196. Mayer JM. Stereoselective metabolism of anti-inflammatory 2-arylpropionates. Acta Pharm Nord 1990; 2:197–216. 197. Sherrat HSA. Acyl-CoA esters of xenobiotic carboxylic acids as biochemically active intermediates. Biochem Soc Trans 1985; 13:856–8. 198. Dodds PF. Incorporation of xenobiotic carboxylic acids into lipids. Life Sci 1991; 49:629–49. 199. Gitlin N. Salicylate hepatotoxicity: the potential role of hypoalbuminemia. J Clin Gastroenterol 1980; 2:281.

Nonsteroidal Anti-inflammatory Drugs

437

200. Runyon BA, LaBrecque DR, Anuras S. The spectrum of liver disease in systemic lupus erythematosus: report of 33 histologically-proved cases and review of the literature. Am J Med 1980; 69:187–94. 201. Mills PR, Sturrock RD. Clinical associations between arthritis and liver disease. Ann Rheum Dis 1982; 41:295–307. 202. Weinblatt ME, Tesser J, Gilliam JH. The liver in rheumatic disease. Semin Arthritis Rheum 1982; 11:399–405. 203. Caparroz-Assef SM, Salgueiro-Pagadigorria CL, Bersani-Amado CA, et al. The uncoupling effect of the nonsteroidal anti-inflammatory drug nimesulide in liver mitochondria from adjuvant-induced arthritic rats. Cell Biochem Funct 2001; 19:117–24. 204. Uetrecht JP. New concepts in immunology relevant to idiosyncratic drug reactions: the “Danger Hypothesis” and innate immune system. Chem Res Toxicol 1999; 12:387–95. 205. Furst SM, Gandolfi AJ. Immunologic mediation of chemical-induced hepatotoxicity. In: Plaa GL, Hewitt WR, eds. Toxicology of the Liver. Taylor & Francis: Washington, DC, 1998:259–95.

21

Nonsteroidal Anti-inflammatory Drugs and Leukotriene Receptor Antagonists: Pathology and Clinical Presentation of Hepatotoxicity James H. Lewis

Georgetown University Medical Center, Washington, D.C., U.S.A.

INTRODUCTION Nonsteroidal anti-inflammatory drugs (NSAIDs) as a class are an important cause of druginduced toxic injury to several organ systems, including well-known injury to the gastrointestinal tract and kidneys, and more recently, cardiac toxicity associated with rofecoxib (Vioxx) and other COX-2 inhibitors (1). While perhaps less well appreciated, NSAIDs are a leading cause of drug-induced hepatotoxicity. For instance, in Denmark, NSAIDs accounted for approximately 9% (97 of 1100) of all drug-related liver injury reports between 1978 and 1987 (2) and they continue to be reported at a rate that often exceeds other drug classes (3). A number of NSAIDs are listed among drugs causing fulminant hepatic failure (FHF) (4,5), occasionally leading to emergency liver transplantation (6,7). The early history of NSAID-induced hepatic injury dates back more than 60 years (8). Cinchophen was one of the first agent to be associated with hepatotoxicity, with a case-fatality rate of nearly 50% that forced its withdrawal from clinical use (9,10). Over the ensuing decades, several other NSAIDs have been developed or introduced into practice only to be abandoned during pre- or post-market evaluation owing to serious liver injury (Table 1) (8,11). Early examples included glafenine (12), an NSAID similar to cinchophen; ibufenac (13), a precursor of ibuprofen; fenclozic acid, an early acetic acid derivative NSAID (14); and fluproquazone, the precursor compound of quinazolone derivatives (8). However, it was not until the withdrawal of benoxaprofen (Oraflex) in 1982 owing to reports of fatal jaundice in the United Kingdom (15) that attention was intensively focused on the hepatotoxicity of NSAIDs as a group (8,11,16). At that time, the Arthritis Advisory Committee of the Food and Drug Administration concluded that hepatic injury should be considered a class characteristic of NSAIDs (17). However, this uniform characterization of hepatic injury obscures the many individual differences and potential for hepatic injury found within and among the different NSAID classes (18). Newer NSAIDs continue to include agents associated with instances of FHF that have forced their withdrawal, as most recently occurred with bromfenac (19). For others, such as diclofenac, liver enzyme monitoring to detect hepatotoxicity is recommended to help prevent FHF (16). Several other agents, including sulindac, piroxicam, and mefenamic acid, also should be monitored closely for clinical signs of liver injury (16). This chapter will review the clinical presentation and pathological features of the acute hepatic injury associated with the currently available NSAIDs. The leukotriene receptor antagonists (LRAs) used in the prevention of asthma are also included, as numerous reports of hepatic injury, including instances of liver failure have appeared since the first agent in this class, zafirlukast, was introduced in 1996 (20,21). For many agents under discussion, liver injury has been well characterized; for several others, however, only limited data are available and clinical summaries are necessarily less complete.

440

Lewis

TABLE 1 Examples of NSAIDs of Various Classes Withdrawn or Abandoned Due to Hepatotoxicity Anthranilic acid derivatives Cinchophen Glafenine Acetic acid derivatives Amphenac Fenclozic acid Isoxepac Bromfenac Propionic acid derivatives Benoxaprofen Ibufenac Pirprofen Suprofen Fenbufen Pyrazolone derivatives Phenylbutazone Oxyphenbutazone Oxicams Isoxicam Sudoxicam Quinazolone derivatives Fluproquazone

INCIDENCE OF NSAID-INDUCED HEPATIC INJURY It has been suggested that the occurrence of serious overt hepatic injury due to NSAIDs as a group is well under 0.1% (22), although figures to determine the true incidence of NSAIDinduced hepatic damage are generally lacking. However, with upward of tens of millions of patients in the United States taking NSAIDs on a regular basis, even this very low incidence of injury may translate into a substantial number of affected individuals. For example, using Medicaid billing data from hospital admissions for acute liver disease in Michigan and Florida, Carson and colleagues (23) reported an annual incidence of acute hepatitis due to NSAIDs leading to hospitalization of 2.2 per 100,000 persons. However, when NSAID cases were compared with controls, none of the individual NSAIDs was associated with a statistically significant increased risk. In contrast, a large retrospective Canadian study involving nearly 230,000 patients and 650,000 person-years of NSAID exposure showed a risk of NSAIDassociated hospitalization for acute (mostly cholestatic) liver injury of 1.7, based on an excess risk of injury of 5 per 100,000 person-years (24). Walker et al. (25) reported that clinically evident liver injury occurred rarely, with a frequency of one case per 10,000 patient-years of use. Sulindac had a 5- to 10-fold higher incidence than other NSAIDs, including diclofenac in this populationbased study. A similar estimate of the relative risk of NSAID-associated liver injury was also reported in Denmark for sulindac and fenbufen, which were reported more often than other NSAIDs (2). Lacroix et al. from France (26) assessed the risk of NSAID use in a primary care setting during the years 1998–2000. Twenty-two agents were implicated in hepatotoxicity, although none caused fatal injury. Women were at much higher risk than men. Incidence rates for individual agents have been reliably estimated for only a few drugs, as will be presented. In general, spontaneous reports do not reliably reflect the results of large epidemiological studies (27,28). EFFECT OF RHEUMATIC DISEASES ON THE LIVER Any discussion of NSAID-induced hepatic injury must take into account the underlying rheumatic disease being treated that may also adversely affect the liver. Many rheumatic diseases may have hepatic-associated enzyme elevations that may mimic drug injury. For example, rheumatoid arthritis is associated with elevations of alkaline phosphatase

441

NSAIDs and Leukotriene Receptor Antagonists

TABLE 2 Effects of Some Rheumatological Diseases on the Liver Disease

Hepatic abnormalities

Rheumatoid arthritis

Elevated alkaline phosphatase, GGT in 25% to 50%, hepatomegaly 10%

Systemic lupus erythematosus Fety’s syndrome (RA, splenomegaly, neutropenia)

Elevated LAEs 20% to 50%, autoimmune (lupoid) hepatitis, hepatomegaly 20% to 25%, jaundice 4%, ascites 10% Elevated LAEs 33%, hepatomegaly 33% to 66%

Sjo€ gren’s syndrome Essential mixed cryoglobulinemia Polyarteritis nodosa

Elevated LAEs 5%, jaundice 2% Chronic hepatitis C in 40%

Psoriatic arthritis

Elevated LAEs

Hepatitis B, hepatomegaly, elevated LAEs, acalculous cholecystitis

Pathology Steatosis, nonspecific changes, mild portal inflammation Steatosis, cholestasis, CAH, granulomas, cirrhosis Nodular regenerative hyperplasia (up to 70%), portal fibrosis, portal hypertension CAH, may be part of PBC HBsAg-associated immune complexes, changes of vasculitis Steatosis, inflammation, hepatic necrosis, fibrosis cirrhosis (!1%)

Abbreviations: LAE, liver-associated enzymes (usually ALT, AST); CAH, chronic active hepatitis; PBC, primary biliary cirrhosis. Source: From Refs. 20 23.

levels in 25% to 50% of individuals not receiving drug treatment (29–32). Hepatic involvement in systemic lupus erythematosus (SLE) is present in as many as 20% of individuals with a twofold elevation in hepatic-associated enzymes (32,33). Hepatic abnormalities in biochemical testing as well as hepatic histology have also been observed in patients with Felty’s syndrome, € Sjogren’s syndrome, progressive systemic sclerosis, polyarteritis nodosa, essential mixed cryoglobulinemia (which may be associated with underlying chronic hepatitis C infection), polymyalgia rheumatica, Reiter’s syndrome, and occasionally even osteoarthritis (OA) (31,32) (Table 2). Data observed with agents such as diclofenac suggest that patients, of either gender, with OA are more susceptible to minor drug-induced hepatic injury compared to those with rheumatoid arthritis, and that susceptibility to clinically significant injury is enhanced even further in women (8,18). CLINICAL AND BIOCHEMICAL SPECTRUM OF NSAID-INDUCED HEPATIC INJURY While most NSAIDs that can produce overt hepatic injury (with jaundice) do so rarely, many agents are associated with mild abnormalities of hepatic enzymes such as aspartate aminotransferase (AST) and alanine aminotransferase (ALT). Elevations in AST and ALT values occur in 5% to 15% of patients taking NSAIDs as a class (22). Most of these levels remain less than three times the upper limits of normal (ULN), and in some cases may resolve despite continuation of the NSAID. In general, however, the higher the incidence of even mildly elevated aminotransferase levels (especially ALT), the more likely the risk of overt hepatic disease (17). The histological lesions produced by NSAIDs depend on the agent involved and on the mechanism of injury. Table 3 lists the predominant types of injury for the currently available NSAIDs. Acute hepatocellular injury involves hepatic degeneration or cell necrosis, while cholestatic injury is characterized mainly by the agents that arrest bile flow. Mixed injury refers to hepatocellular (cytotoxic) injury and cholestasis. Intrinsic hepatotoxins cause injury that is mainly cytotoxic with necrosis, degeneration, and/or steatosis, although a few can cause cholestasis. In contrast, idiosyncratic injury results in cholestatic or hepatocellular injury (8,16). The biochemical changes seen with NSAID-associated liver injury reflect the histological pattern of damage. Hepatocellular injury with necrosis resembles acute viral hepatitis with AST and ALT levels increased 10- to 100-fold or more and bilirubin levels that are variably increased. Serum alkaline phosphatase values are generally normal or only mildly elevated. Toxic microvesicular steatosis may resemble acute fatty liver of pregnancy or Reye’s syndrome (RS)

442

Lewis

TABLE 3 Clincopathological Features of Hepatotoxicity Due to NSAIDs Class agent Salicylates Aspirin Sodium, choline salicylates Diflunisal (Dolobid) Benorilate Salsalate (Disalcid) Acetic acid derivatives Diclofenac (Voltaren) Etodolac (Lodine) Ketorolac (Toradol) Bromfenac (Duract) Indomethacin (Indocin) Sulindac (Clinoril) Tolmetin (Tolectin) Nabumetone (Relafen) Clometacin Propionic acid derivatives Ibuprofen (Motrin et al.) Naproxen (Naprosyn et al.) Fenoprofen (Nalfon) Flurbiprofen (Ansaid) Oxaprozine (Daypro) Ketoprofen (Orudis) Benoxaprofen (Oraflex) Oxicams Piroxicam (Feldene) Droxicam Pyrazolone derivatives Phenylbutazone Oxyphenbutazone Fenamates Mefenamic acid (Ponstel) Meclofenamic acid (Meclomen) Cyclooxygenase-2 inhibitors Nimesulide Celecoxib Rofecoxib

Proposed mechanism

Type of injury Acute H-cell, CAH?, Reye’s syndrome H-cell (minor) Cholestatic, mixed Zone 3 necrosis H-cell (minor)

Intrinsic toxicity

Susceptibility factors JRA, SLE, RF

Hypersensitivity Hyposensitivity? Intrinsic a

Acute H-cell necrosis, autoimmune CAH-like

Metabolic idiosyncrasy

H-cell necrosis Not reported Massive necrosis H-cell necrosis, microvesicular steatosis, cholestasis (less often) Cholestasis or mixed, H-cell in 25% Jaundice, steatosis Cholestatic jaundice Autoimmune CAH, granulomas, cholestasis

Metabolic idiosyncrasy?

H-cell or mixed (rare), steatosis H-cell jaundice, cholestasis Cholestatic jaundice (rare)

Hypersensitivity? Hypersensitivity? Hypersensitivity?

H-cell jaundice (rare) H-cell H-cell jaundice (rare) Cholestatic jaundice

Hypersensitivity? Metabolic idiosyncrasy? Metabolic idiosyncrasy? Metabolic idiosyncrasy

H-cell necrosis, cholestasis Cholestasis

Hypersensitivity Hypersensitivity?

Elderly

H-cell necrosis, steatosis or cholestasis, granulomas H-cell necrosis, granulomas

Hypersensitivity (intrinsic toxicity in high doses)? Hypersensitivity

Adults, women

H-cell necrosis (rare) H-cell (minor) H-cell necrosis, cholestasis less often H-cell (rare) a

Elderly females with OA, cross-sensitivity with ibuprofen?

Metabolic idiosyncrasy Metabolic idiosyncrasy

Prolonged use Children

Hypersensitivity

JRA, SLE

Metabolic idiosyncrasy? Metabolic idiosyncrasy? Hypersensitivity

Elderly females Cross-sensitivity with diclofenac Cross-sensitivity with naproxen

Elderly females

a a

Metabolic idiosyncrasy (? hypersensitivity) ? a

a

Unknown, too little information available. Abbreviations: CAH, chronic active hepatitis; H-cell, hepatocellular; JRA, juvenile rheumatoid arthritis; SLE, systemic lupus erythematosus; RF, rheumatic fever; LFT, liver function tests.

with aminotransferase values 5 to 20 times normal and up to threefold elevations in alkaline phosphatase and bilirubin (34). Clinically, hepatocellular injury may cause anorexia, fatigue, nausea, malaise, and jaundice. Massive necrosis leading to fulminant hepatitis may result in hepatic coma, coagulopathy, ascites, and death (7). Drug-induced jaundice associated with hepatocellular

NSAIDs and Leukotriene Receptor Antagonists

443

injury must be regarded as a serious lesion since case-fatality rates are 10% or more, depending on the agent (4,5,35) (Hy’s Law). The prognosis of complete recovery is usually good for patients who survive the acute phase of injury (8). For example, Lacroix et al. described most patients having their elevated liver-associated enzymes (LAEs) normalize after the offending NSAID was discontinued, with no fatalities occurring from any of the 22 NSAIDs reported in their case–control study (26). FHF leading to transplant has only rarely been reported with NSAIDs; a review of the UNOS database of all drug-induced causes of acute liver failure leading to liver transplant in the United States over the years 1990 through 2002 found only one case each of three NSAIDs (ibuprofen, bromfenac, and naproxen) from 270 drug-induced cases among more than 51,000 transplants in the United States over that period (6). Other instances of FHF requiring liver transplant also have been reported separately for bromfenac (19), diclofenac (36), and ibuprofen (37). Cholestatic injury is characterized by elevated alkaline phosphatase levels (3 to 10 times normal) with parallel increases in g-glutamyl transpeptidase or 5 0 -nucleotidase and variable increases in serum bilirubin, with AST and ALT remaining normal or only modestly elevated. The predominant features of cholestatic injury are jaundice and pruritus. Some patients complain of abdominal pain that may mimic acute extrahepatic biliary obstruction. Rarely does intrahepatic cholestatic injury cause a fatal outcome, although prolonged jaundice may sometimes be seen (38). HEPATIC INJURY DUE TO INDIVIDUAL NSAIDS Salicylates Aspirin (Acetylsalicylic Acid) Several hundred cases of aspirin-related hepatic injury have been reported since the 1970s (39). However, this represented a delay of more than 75 years before the hepatotoxic potential of aspirin was truly appreciated. In part, this may have been due to the fact that the injury is often mild and anicteric and was overlooked in the era prior to routine enzymatic testing (8,39). Alternatively, the injury may have been attributed to the underlying rheumatic disease (30). In contrast, NSAIDs developed in the past two decades often have had hepatic injury observed during clinical trials or soon after initial marketing because of the routine use of biochemical testing (18). Aspirin injury is primarily hepatocellular, but in general is clinically mild and reversible with ALT/AST levels less than 10-fold elevated. Bilirubin levels usually remain normal or are minimally elevated with jaundice seen in fewer than 5% of cases (39). Liver biopsy characteristically shows areas of focal necrosis with a mild inflammatory response in the portal areas. In addition, cellular unrest, ballooning, and eosinophilic degeneration have been described (39). Ultrastructural changes include increased numbers of lysosomes, peroxisomes, and mitochondria with dilation of the smooth and the rough endoplasmic reticulum (40). Aspirin injury has been described as being both dose dependent and blood concentration dependent consistent with intrinsic toxicity (39), as seen in both animals and cell culture experiments (41,42). More recently, however, Singh et al. described elevated ALT levels in 22 of 50 children treated with high-dose salicylates for rheumatic fever; 12 of whom were symptomatic with ALT values 5- to 10-fold elevated. In contrast to earlier reports suggesting that hepatic injury occurred as a function of higher serum salicylates levels, these authors did not observe any specific relationship between serum salicylate concentrations and hepatocellular injury in these children (43). The major metabolites of aspirin are salicyluric and salicylphenolic glucuronide. It has been suggested that these metabolic pathways are readily saturated in children as well as adults leading to the accumulation of an otherwise minor nontoxic metabolite that may become responsible for hepatic injury (44). The exact mechanism of the cellular injury is unclear, although several possible modes of action have been postulated. These include lipid peroxidation, mitochondrial damage, hydroxyl radical scavenging, and injury to hepatocyte membranes (42,45). Hepatic damage as indicated by elevated AST and ALT levels is seen in up to 50% of patients taking sufficient aspirin to produce serum blood levels above 15 mg/dL, although

444

Lewis

hepatic injury has been noted with levels as low as 10 mg/dL (39), and a correlation between serum levels and ALT elevations is not always apparent (43). Toxicity appears to be a property of the salicylate molecule, since sodium and choline salicylate also lead to elevated aminotransferase values (11,46). Susceptibility to aspirin injury is reported to be greater in patients with juvenile rheumatoid arthritis (JRA), SLE, and rheumatic fever, perhaps because of the relatively higher doses taken for these disorders (30,39,46–50). The incidence of abnormal hepatic-associated enzymes detected in these patients ranges from 20% to 70% with children under age 12 having a higher incidence compared with adults (39,49). No gender differences in susceptibility have been observed, although a possible genetic predisposition for hepatotoxicity was reported for children with JRA who have the A2BW40 haplotype (50). Aspirin-associated injury, in general, has not been serious and resolves promptly when the drug is stopped. Severe injury occurs in less than 3% of patients and no convincing cases of fatal necrosis were reported in earlier reports (8,39). However, in the United Kingdom, deaths from suicidal overdoses of salicylates [as well as paracetamol (acetaminophen) and ibuprofen] appear to be much more commonplace than in the United States, prompting legislation that restricted the number of tablets that could be sold over the counter. Limiting the sale of salicylates reduced the number of instances of suicidal deaths from these compounds by 22% in the first year after the legislation was passed (which paralleled the reduction in liver unit admissions and liver transplants for paracetamol), and the number of nonfatal overdoses was also reduced, demonstrating the benefits of restricting the size of analgesic packages for OTC sale (51). Although a few instances of fatal illness associated with encephalopathy and coagulopathy in patients with JRA and SLE receiving high doses of salicylates were reported prior to 1980 (39,52), they may have represented early examples of RS, as will be discussed. Hypoalbuminemia has been reported to increase the risk of salicylate hepatotoxicity from decreased protein binding (53), as has chronic liver disease in general, where more salicylate is free to distribute to the tissues and injure the liver (54). Several reports have suggested that chronic active hepatitis may develop as a result of acute salicylate injury (55), although these cases antedated the availability of hepatitis C testing (17). Reports of aspirin-related chronic hepatitis are lacking in the hepatitis C era. Epidemiological studies in the 1980s demonstrated a strong association between aspirin and RS in children with influenza or varicella (chickenpox) (56–58). Adults were also affected as evidenced by several reports of older individuals developing RS after taking aspirin for a presumed viral infection (59–61). Convincing evidence for the association also comes from the striking decline in the incidence of RS in the United States that paralleled the decreased use of aspirin (62,63). In work by Pinsky and colleagues (64), an increased risk of RS was associated with an increased dose of aspirin, although doses as low as 15 mg/kg per day (the equivalent of two 325-mg tablets in a 40-kg child) were associated with a substantially increased risk. As a result, the use of aspirin continues to be strongly discouraged in acute febrile illnesses, especially in children. The mechanism by which aspirin acts with the viral illness to produce RS is unclear. Salicylate toxicity leads to mitochondrial injury resembling that of RS both in vivo and in vitro (65), although other antipyretics, including acetaminophen (not thought to be related to RS), have the potential in animal models to exacerbate the lethal effects of a viral infection by decreasing interferon-induced antiviral responses (66). Recently, the association between aspirin and RS in children has been challenged as more likely being the result of one of several inborn errors in mitochondrial metabolism that were first diagnosed in the 1980s. According to Orlowski (67), 69% of patients who survived a bout of RS in Australia were subsequently diagnosed as having medium-chain aryl-coenzyme A dehydrogenase deficiency or other now well-described metabolic disorders; none of their 49 original RS patients would be diagnosed as having definite RS by today’s criteria. Other Salicylates Nonacetylated salicylates also appear able to produce hepatic injury with sodium and choline salicylate resulting in elevated aminotransferase levels and jaundice in some instances (11,68,69). A rather severe but reversible hypersensitivity reaction with markedly elevated

NSAIDs and Leukotriene Receptor Antagonists

445

AST and ALT values accompanied the injury reported in a 66-year-old woman who took choline magnesium trisalicylate after just three days (68). Eosinophilia as part of a hypersensitivity reaction was also reported by Nadkarni (68). Diflunisal (Dolobid), a difluorophenyl derivative of salicylic acid, has been incriminated in cholestatic and mixed cholestatic hepatocellular jaundice in a few reports (70,71). Diflunisal does not undergo metabolism to salicylate, which may explain the relative absence of clinical hepatic injury (72), although experimentally, diflunisal causes cytotoxicity (73). Hypersensitivity has been suggested as the mechanism (66). Benorilate is an acetaminophen ester of acetylsalicylic acid cited as producing hepatic injury resembling that caused by acetaminophen (paracetamol) toxicity, namely zone 3 (centrilobular) necrosis, rather than degeneration and microvesicular steatosis more typical of aspirin-related injury (74). Salsalate also has been reported to cause elevated aminotransferase values, but serious injury appears unlikely (18). Acetic Acid Derivatives Diclofenac Diclofenac (Voltaren, Cataflam), a benzene–acetic acid derivative, is one of the most widely prescribed NSAIDs worldwide, having been introduced in the United States more than a decade and a half ago. There are two formulations, a delayed-released enteric-coated diclofenac sodium and an immediate-release potassium formulation, which will be considered together for purposes of hepatic injury. The drug has been implicated as the cause of approximately 250 cases of hepatocellular damage in published reports, with a case-fatality rate of approximately 10% (75–78), and FHF leading to death or liver transplantation (4,7,37). Abnormal aminotransferase values that may not progress develop in 15% to 20% of patients taking the drug (77). It has been estimated that there are approximately one to two cases per million prescriptions of hepatic injury due to delayed-released diclofenac (78,79), although Food and Drug Administration (FDA) data suggest an incidence that may be two to three times higher (75). Diclofenac is more likely to produce hepatic injury than are most other NSAIDs, exceeded only by sulindac (75,80), although not all cohorts have demonstrated an increased risk (81). Diclofenac injury is predominantly hepatocellular, resembling acute viral hepatitis. In the series by Banks and colleagues (75), 79% of affected individuals were women, most of whom were aged 60 or above, and two-thirds had underlying OA. A majority of cases (67%) were initially detected on the basis of hepatic symptoms with the remainder identified by abnormal LAEs. Latency periods were observed to be one month in 24% of cases, with cumulative rates of injury being 63% by three months and 85% by six months. Twelve percent of individuals had taken diclofenac for 6 to 12 months, and only 3% for more than 12 months, prior to the onset of hepatic injury. Hepatocellular injury was apparent in 97 of the 180 patients studied (54%), of whom 60% were jaundiced. Mixed injury was seen in 12%, indeterminate injury in 26%, and intrahepatic cholestasis in 8%. When the alkaline phosphatase was elevated greater than three times the ULN, the injury was invariably mixed or cholestatic. However, published reports of acute cholestatic hepatitis with diclofenac are less common outside of this series (82). Most patients present with jaundice, fatigue, anorexia, nausea, and vomiting. Fever, rash, and eosinophilia are uncommonly seen (75). Aminotransferase levels range from 10 to 100 times the ULN, and jaundice may be prominent. Based on the biopsy or autopsy material available for review in 21 of the FDA series cases (75), the main lesion was acute hepatic necrosis (predominantly zone 3), the severity of which often matched the marked elevations of aminotransferase levels. Other histological findings included granulomas in 1 of 21 patients and changes of chronic hepatitis in 6 individuals. About 50% of the 180 cases reported to the FDA and analyzed by Banks and colleagues were anicteric with only modestly elevated aminotransferase values, and occurred in mostly asymptomatic individuals with elevated enzymes found during routine biochemical testing (75). Females and patients with OA appear to have a significantly higher risk of hepatic injury than do males or rheumatoid arthritis patients. The average age of affected individuals has been 60 years, reflecting their underlying OA (75). Autoimmune chronic active hepatitis has been suspected in several patients reported and summarized by Scully et al. (83) and by Sallie (84) based on the presence of antinuclear or

446

Lewis

anti-smooth muscle antibodies. Histological findings in these cases ranged from periportal inflammation with mild fibrosis to panlobular hepatitis (83). The delayed onset of injury after taking diclofenac (up to 12 months) and a late response to rechallenge (as long as five weeks after re-administration) suggest metabolic idiosyncrasy as the likely mechanism (75). While 6 of 21 patients in the collected series reported by Scully et al. had features of a hypersensitivity reaction, including peripheral eosinophilia and rash, the other 15 patients had injury in keeping with a metabolic abnormality (83). None of the patients in the larger FDA series had hypersensitivity features (75). Several investigators have identified reactive metabolites of diclofenac that are concentrated in bile canaliculi, decrease cellular ATP, and are presumably responsible for experimental liver toxicity (85–92) as well as that seen in humans (93,94). Recovery is usually prompt after diclofenac is withdrawn, although as with other druginduced hepatocellular injury, massive necrosis with FHF and death is a feared complication that occurred in about 8% of icteric cases in the FDA series (75). A report from Japan cites the beneficial effects of an intravenous prostaglandin E infusion in combination with intravenous prednisolone that contributed to the recovery of a 56-year-old man who developed fulminant hepatitis from diclofenac (95). A few patients have received corticosteroids for presumed diclofenac-associated autoimmune hepatitis (83), although its value is unclear in this setting. Etodolac (Lodinew) This pyranocarboxylic acid derivative rarely caused hepatic injury in clinical trials. A rise in aminotransferase levels or bilirubin values greater than 1.5 times ULN was seen in only 10 of 3302 patients in doses from 50 to 600 mg daily for six weeks to as long as 88 months (96). However, a recent report offatal hepatitis underscores the possibility of serious hepatic injury (97). That case involved an obese 67-year-old woman who had taken etodolac 300 mg twice daily for about four months before developing a one-week prodrome of nausea, vomiting, weakness, anorexia, jaundice, and confusion, followed by liver failure. At autopsy, the liver revealed submassive bridging necrosis, early fibrosis, and microvesicular steatosis. The mechanism of the hepatocellular injury, similar to other members of this class, was presumably metabolic idiosyncrasy. Etodolac undergoes extensive enterohepatic circulation and its elimination is markedly inhibited in rats with either hepatic or renal failure producing high plasma levels (98). A false-positive test for urinary bilirubin may occur due to its phenolic metabolite (99). Ketorolac (Toradol w) I am not aware of any published report of hepatic injury with this agent although the manufacturer notes that patients with impaired hepatic function or other causes of hypoalbuminemia may be at risk for hepatic toxicity, including liver failure (100). Elevated AST and ALT values may be seen in patients with preexisting liver disease and the drug should be used with caution in this setting. Bromfenac This acetic acid derivative was introduced in 1997 as a nonnarcotic analgesic of the phenyl acetate class for short-term pain relief, but was removed from the market in 1998 owing to several instances of FHF leading to death or transplant that occurred after prolonged administration (19,101–103). While the drug did not appear to be hepatotoxic during limited short-term use (less than 10 days), reports of severe hepatotoxicity began to appear in patients who were treated for periods exceeding 30 to 90 days. In a case series reported by Fontana and colleagues (19), a prodrome of malaise and fatigue heralded severe hepatocellular injury progressing to FHF. The histological findings included massive or submassive centrizonal necrosis accompanied by a lymphocytic infiltrate. Two patients with a protracted clinical course developed nodular regeneration. Resolution of FHF within three months using supportive measures was seen in a patient reported by Moses and colleagues (101). Others have required liver transplantation (19,102), and deaths were reported (103). No evidence of a hypersensitivity reaction was apparent in any of the reported cases. The drug binds to plasma albumin and is extensively metabolized (104). As a result, the mechanism of injury was thought to be metabolic idiosyncrasy. The inability to identify

NSAIDs and Leukotriene Receptor Antagonists

447

individuals at risk from prolonged use forced its withdrawal soon after initial FHF cases were reported (19). Fenclozic Acid This early derivative of the arylakanoic acid group produced mixed or cholestatic jaundice in 10% of recipients and was withdrawn in 1970 (14). Indomethacin (Indocinw) This indole acetic acid derivative has been available in the United States since 1963. Despite its use being limited by side effects that appear in up to 50% of patients, only a few instances of jaundice have been reported (11). An analysis of adverse reactions to NSAIDs during a nineyear period in the United Kingdom showed indomethacin to be responsible for more than 1260 total reactions, of which 114 were fatal. However, only approximately 3% of all reactions and about 6% of fatalities involved the liver in that series (13). Indomethacin has produced mainly hepatocellular necrosis (massive or central), sometimes accompanied by microvesicular steatosis and striking cholestasis (105). A high case-fatality rate (approximately 15%) was estimated by Cuthbert in 1974 (13). Although there are few reports documenting indomethacin as the cause of fatal hepatic disease, children appear more vulnerable and the drug is not recommended in the pediatric age group based on several deaths involving hepatocellular necrosis in children with JRA (106–108). Indomethacin is converted to active metabolites, and since none appear to be intrinsically hepatotoxic, metabolic idiosyncrasy seems the most likely mechanism (22). Experimentally, indomethacin has completely prevented the mortality and hemorrhagic hepatic necrosis caused by the mushroom toxin phalloidin (109) and also protects against carbon tetrachloride–mediated injury, possibly increasing mitochondrial respiration (110). Although it has been suggested that individuals with underlying liver disease avoid this agent given the high case-fatality rate (8), a recent study in chronic hepatitis B virus–infected patients found that indomethacin in a dose of 25 mg t.i.d. was associated with a higher rate of e antigen seroconversion (111). Sulindac (Clinoril w) This indene derivative, which bears a structural similarity to indomethacin, was approved for use in the United States in 1978 after several years of study in Europe. The overall incidence of adverse side effects has been 25%, and toxicity requiring withdrawal of the drug was seen in 5% to 7% of patients (112–114). It is considered one of the most likely NSAIDs to produce hepatic injury (25,115,116). More than two dozen published case reports or case series of sulindacassociated jaundice have been published (11). Among 338 cases of suspected sulindac hepatic injury reported to the FDA, and analyzed by Tarazi and colleagues (117), 91 were considered probably or definitely related. This relatively high case number is consistent with an incidence of sulindac-associated hepatic injury that exceeds that for most other NSAIDs and is comparable to that of diclofenac (2). In the series by Tarazi and colleagues (117), the onset of illness usually occurred within eight weeks of starting the drug and in many cases (48%) developed in less than four weeks. Histological material was available for review in 15 of these 91 cases, and revealed that most were cholestatic or had mixed jaundice. However, one-quarter of cases were hepatocellular, and another 20% were indeterminate (117). Women outnumbered men by 3.5 to 1 and 69% were over age 50 with only 6% of individuals being under age 20. The biochemical features of injury mimicked the histological picture; those with cholestatic injury had an alkaline phosphatase elevated greater than two times ULN ranging up to 3500 mU/mL with a mean bilirubin elevation of 7 mg/dL (ranging up to 35 mg/dL). AST and ALT values were raised three to four times ULN in this group. Among those with hepatocellular injury, mean elevations in AST and ALT were 20 to 25 times ULN ranging to more than 100-fold; mean bilirubin was 5.4 mg/dL and alkaline phosphatase was normal. Since sulindac has also led to pancreatitis (118), ductal obstruction due to pancreatitis may have contributed to some of the jaundiced cases. Hypersensitivity features were present in most patients, with fever in 55%, rash in 48%, pruritus in 40%, and eosinophilia in 35% in the series reported by Tarazi et al. (117). These percentages are nearly identical to those reported in the literature and were uniformly

448

Lewis

distributed across the histological groups, except for eosinophilia, which was absent in patients whose ALT/AST values were elevated more than eightfold (117). Rechallenge with the drug after recovery has led to recurrence of hepatic injury within several days, as was seen in nearly one-quarter of cases overall, and supports drug allergy (immunological idiosyncrasy) as the mechanism of injury (117). Further support is found in cases of Stevens–Johnson syndrome reported with sulindac (119,120). Most patients recover within one to two months after stopping the drug (117), but recovery may be delayed for as long as seven months (121). Case-fatality rates are about 5%, and deaths appear to be due to severe generalized hypersensitivity, including toxic epidermal necrolysis and renal failure, rather than to hepatic disease alone, since the most common pattern is cholestatic injury. However, there have been a few cases of fatal hepatic necrosis (117). As with aspirin, children with rheumatoid arthritis and patients with SLE may be at increased risk of sulindac injury (115,122), although in general, factors associated with increased susceptibility include older age and female gender. While the use of NSAIDs (including sulindac) is not advised in patients with cirrhosis owing to the risk of renal toxicity and hepatorenal syndrome, sulindac is a less potent inhibitor of renal eicosanoids compared with other NSAIDs, such as ibuprofen, and is reported to have renal-sparing effects (123). Tolmetin (Tolectinw) This pyrrole acetic acid derivative has been in use in the United States for about a decade. A few reports have been submitted to the FDA including cases of jaundice, although the rate of hepatic injury appears to be somewhat less than for other NSAIDs (8). One published case report involved microvesicular steatosis as part of widespread multisystem organ failure in a 15-year-old girl who died having a markedly elevated tolmetin blood level (124). It has been reported that about 5% of individuals receiving tolmetin develop minor elevations in aminotransferase levels that do not progress (125). Nabumetone (Relafenw) This naphthyl acetic acid derivative includes mention of mild elevations in aminotransferases in the class labeling in the package insert based on less than 1% risk among 1677 patients in premarketing studies. Worldwide safety experience in nearly 38,000 patients did not include any reports of liver necrosis (126, 127). Clometacin This agent is an isomer of indomethacin and is used primarily in France. It is not available in the United States. Clometacin has been associated with a form of autoimmune hepatitis that is seen in women after a latent period of six months to several years (3,8,128–130). In a series of 30 cases reported by Islam and colleagues (129), a female predominance of 29 to 1 was seen, with an age range of 32 to 84 years. Acute hepatitis with centrilobular necrosis was present in 17 of 25 patients undergoing liver biopsy, and 8 showed chronic active hepatitis. Anti-smooth muscle (anti-actin) and antinuclear antibodies were found in 66% and 52%, respectively, in titers ranging to 1/2560. Seventy-three percent of these patients had hypergammaglobulinemia. Pariente et al. (130) reported similar clinical features in their series, including three fatalities. The syndrome seen with clometacin was noted to be similar to that produced by the laxative oxyphenisatin (131). Renal injury, rash, and eosinophilia may accompany the hepatic disease (8). In addition to acute and chronic autoimmune hepatitis, other histological forms of injury seen with clometacin include granulomatous injury, multinuclear giant cell hepatitis, cholestatic hepatitis, and cirrhosis (8,131). The acute syndrome with fulminant hepatitis has been fatal in some instances (130–132). Immunological idiosyncrasy is the presumed mechanism of injury, although intrinsic toxicity has also been suspected as the basis of injury seen in overdose settings (8). Propionic Acid Derivatives Ibuprofen (Motrinw, Advil w, and Others) This derivative of ibufenac [a drug withdrawn from use in the 1960s because of fatal hepatocellular injury (13)] has proven to be far less hepatotoxic (8,22). Figures from the 1970s

NSAIDs and Leukotriene Receptor Antagonists

449

indicated that adverse reactions to ibuprofen (4% overall) rarely involved the liver (13). More recent studies suggested that ibuprofen is safer compared to aspirin and oxaprozin in arthritis patients undergoing AST monitoring (133), consistent with its safety profile amassed over a 15-year period (134). In fact, the relatively few reports of hepatic injury in early United Kingdom and FDA series suggest that ibuprofen is among the least likely of the commonly used NSAIDs to produce hepatic injury (22). Occasional reports of acute hepatocellular or mixed cholestatic injury have appeared with ibuprofen (135–137), including a case of overlapping susceptibility with diclofenac (75), and rarely, reports of fulminant hepatitis contained in large case series spanning several decades of use (4–6,37). Fever and generalized hypersensitivity accompanied the injury and supported immunological idiosyncrasy as the mechanism. The occurrence of fatal steatosis in one patient (137) indicates that the injury may be metabolically mediated (22). Indeed, reactive metabolites have been demonstrated (85). A report of acute hepatocellular injury in a patient taking a large overdose (20 g) implies some intrinsic toxicity is also possible (138) and is supported by experimental studies of the relative hepatotoxicity of ibuprofen and related agents given in high doses (139). In vitro, ibuprofen alters mitochondrial membrane permeability (140) and induces hepatocellular hypertrophy and hyperplasia through an effect on peroxisomes (141). While most ibuprofen-associated liver toxicity is transient in nature, a report by Alam and colleagues (142) suggests that this agent should be added to the growing list of drugs causing prolonged cholestasis as part of the vanishing bile duct syndrome (38). They describe a 29-yearold man who presented with acute right-upper-quadrant pain, nausea, vomiting, hepatosplenomegaly, and jaundice three weeks after taking ibuprofen daily for body aches and headaches during hyposensitization treatment for common allergens. He had no prior exposure to ibuprofen. Values for LAEs included a peak bilirubin of 24 mg/dL, an alkaline phosphatase of nearly 4000 IU/L, and an ALT of 488 IU/L. These abnormalities persisted over the next year, with serial liver biopsies that evolved from a marked portal inflammatory process with bile duct proliferation by neutrophils and eosinophils, to one of increased cholestasis, obliteration, and eventual paucity of bile ducts. Progressive xanthomatosis and hypercholesterolemia also developed, and the syndrome defied treatment, as well as any other explanation. When the report was published, the current status of the patient was unknown. Riley and Smith (143) recently reported a provocative case series of possible increased susceptibility to ibuprofen-induced hepatotoxicity among three patients with chronic hepatitis C infection. Sudden rises in ALT and AST values to greater than 1000 IU/L were recorded after the brief use (up to one week) of ibuprofen for pain. These rises were reproducible in one of these patients on rechallenge. The transaminases slowly returned to baseline two to three months after ibuprofen was discontinued. I am currently unaware of any other reports that corroborate this observation, although caution and hepatic enzyme monitoring have been advised in this setting. Naproxen (Naprosynw, Anaproxw, and Others) Only a few cases of hepatic injury have been reported (8) for this arylacetic acid derivative, including hepatocellular jaundice, cholestatic jaundice, a case of indeterminate jaundice, and rare instances of FHF (4,6,144–146). Cuthbert (13) reported that 4% of 179 adverse reactions involved the liver in the United Kingdom in the 1970s. The onset of overt injury has been within 1 to 12 weeks of starting the drug and mild elevations in serum aminotransferases have regressed during continued therapy in some instances (11). Fever in one patient and hepatic eosinophilia in another indicated possible hypersensitivity (146). While there have been too few cases to precisely identify the mechanism, no evidence of intrinsic toxicity, following accidental overdose or experimental studies, has been seen (147,148). Fenoprofen (Nalfonw) This drug resembles ibuprofen in its structure and metabolism, and also rarely results in hepatic injury (149), having been incriminated in only two instances (150,151). One patient developed cholestatic jaundice while initially taking a form of naproxen (146). After it was discontinued and recovery from jaundice was complete, rechallenge with fenoprofen led to recurrence of the abnormality, implying cross-sensitivity. In the other case, jaundice occurred

450

Lewis

after seven weeks of therapy and resolved on discontinuation (151). Injury in animals has not been observed (149). A related agent, fenbufen, has produced AST/ALT elevations in 25% of recipients, possibly by a hypersensitivity and/or intrinsic mechanism (152). It is not available in the United States. Flurbiprofen (Ansaid w) There has been only one published report of jaundice to my knowledge, in which the injury was apparently hepatocellular, and was accompanied by symptoms of hypersensitivity (153). The onset in that case, however, was unusually delayed for hypersensitivity, occurring three months after the drug was started. Oxaprozin (Dayprow) This agent elevates aminotransferase levels in about 15% of patients, but only 1% had values exceeding a threefold elevation. In two-thirds of patients with abnormal ALT, the values decreased or remained essentially unchanged, despite continuation of the drug (154). Overt hepatitis appears to be rare, although at least two cases of symptomatic liver injury have been reported. In one, massive necrosis was fatal (155), and in the other, recovery from acute hepatocellular jaundice occurred (156). The mechanism is unknown, but is presumed to be a toxic metabolite. Ketoprofen (Orudisw) A few instances of jaundice have been reported to the manufacturer (157), but the overall incidence of injury appears quite low and I am unaware of any published reports. In contrast, pirprofen, an earlier phenyl propionic acid derivative, was withdrawn from use after fatal liver toxicity was reported (158). Most of the cases of severe necrosis occurred in older women who had taken the drug for 1.5 to 9 months. An hepatotoxic metabolite was suspected from clinical and experimental data (115,158). Benoxaprofen (Oraflexw) Benoxaprofen was approved in the United States in April 1982, only to be withdrawn four months later after several elderly patients in the United Kingdom had died with hepatic and renal disease (8,23,159). In Britain, about 500,000 patients took the drug, and the Committee on the Safety of Medicines received about 3500 reports of adverse reactions, most involving photosensitization of the skin (160,161). Other toxicity included gastrointestinal upset, a few cases of Stevens–Johnson-type reactions, and several instances of hepatic injury (11,15,160–162). Many of the fatalities had hepatic and renal involvement. While there were no published U.S. case reports, several hundred instances of hepatic disease were submitted to the Adverse Reaction Registry of the FDA (8). Although the validity of these reports has not been firmly established, the sheer number suggests hepatic injury occurred, and its historical importance warrants the inclusion of benoxaprofen in any discussion on NSAID-induced hepatotoxicity. Hepatic injury appeared after the drug had been taken for 1 to 12 months in a daily dose of 600 mg (15,162). The first prominent symptom was jaundice, although this was sometimes preceded by anorexia, nausea, and vomiting. Several patients had abdominal pain and hematemesis. Even after the drug had been stopped, some individuals continued to deteriorate with deepening jaundice, renal failure, and coagulopathy progressing to death. Peak bilirubin levels ranged up to 17 mg/dL, although most were below 8 mg/dL. Aminotransferase levels were modestly elevated, exceeding eight times ULN in only a few individuals. Alkaline phosphatase levels were elevated threefold or more in about half of the patients. The histological features consisted of marked cholestasis and slight to moderate necrosis. Cholestasis was particularly evident in zone 3, and both cholangioles and canaliculi showed characteristic inspissated bile casts (11,38). The lack of hypersensitivity hallmarks and the prolonged exposure to the drug prior to the injury suggest metabolic idiosyncrasy was responsible. About 50% of the drug is converted to a glucuronide, and concentrations of the drug found in bile canaliculi may have been composed of this or other poorly soluble metabolites. Precipitation of the drug in the small bile ducts apparently led to jaundice (38).

NSAIDs and Leukotriene Receptor Antagonists

451

All but one of the published fatal cases involved women, most of whom were over the age of 70. Elderly individuals metabolize the drug much more slowly than younger people, the half-life of the drug being almost four times greater in elderly patients than in those 40 years and younger (163). Prolonged metabolism may have led to higher blood levels and, ultimately, biliary excretion of larger amounts of higher concentrations of a poorly soluble product that precipitated in the bile canaliculi. In experimental studies using isolated rat hepatocytes, marked toxicity was seen independent of the P450-mediated metabolism of benoxaprofen (164). Injury, however, did correlate with both concentration and duration of exposure in this model. Injury was also demonstrated in a cultured rat hepatocyte model, where, again, dose-related injury was seen (165). More recent investigations have found that benoxaprofen has a molecular structure similar to clofibrate and may be a substrate for CYP4 resulting in hepatic peroxisomal proliferation. The structural similarity of benoxaprofen with psoralen explains its association to phototoxicity (166,167). Flunoxaprofen, a structural analog of benoxaprofen was found to be much less hepatotoxic in rats, likely the result of less reactive metabolites (168). The cause of death in the patients who developed benoxaprofen jaundice remains unclear. Drug-induced cholestasis is rarely associated with case fatalities (38), yet 11 of the 14 individuals with benoxaprofen-induced cholestasis died (15). The biochemical data and histological features suggested that the parenchymal liver injury was not severe enough in most of these individuals to have led to death. It has been proposed that the slower metabolism in these older individuals led to higher benoxaprofen blood and tissue levels and that it was a combination of cholestasis and, perhaps more importantly, renal failure that led to their demise. As a result, death seemed more likely due to a generalized drug toxicity and renal failure than to hepatic injury alone (17). Oxicams The first member of this class of benzothiazine derivatives to be studied, sudoxicam, was implicated in several cases of hepatocellular jaundice, including fatal hepatic necrosis, and was withdrawn from further clinical trials in 1977 (8). Piroxicam (Feldenew) is a carboxamide derivative used as a once-a-day treatment for rheumatoid arthritis and OA. It has been available since 1982 and a number of instances of severe hepatic necrosis and cholestatic jaundice have been reported (22,169–172). While most of these cases have occurred in patients over age 60, a report of transient hepatic dysfunction in a two-year-old child who ingested an inadvertent overdose has been published (173). Some of the cases were fatal from massive or submassive necrosis, and several other individuals have had prolonged cholestatic jaundice (O4 months). Hypersensitivity is suspected given the short latency period (as early as three days) as well as the clinical features. In the perfused rat liver, piroxicam has an effect on energy metabolism via its action to decrease mitochondrial ATP generation (174,175). Piroxicam has a possible protective effect on ethanol-induced glutathione depletion in rats that deserves further study (176). Meloxicam (Mobicw) is an enolic acid derivative that in clinical trials has not been associated with any hepatic abnormalities to date (177). Droxicam and tenoxicam are oxicams available outside of the United States that are both associated with hepatic injury, mostly cholestatic (178). Eosinophilia in some cases suggests an immunoallergic mechanism of injury similar to piroxicam. In the case of isoxicam, reports of toxic epidermal necrolysis in association with cholestatic injury forced its withdrawal from clinical use (179). Pyrazolone Derivatives Phenylbutazone (PBZ, Butazolidin) was introduced in 1949 in the United States for the treatment of rheumatoid arthritis and related disorders. At present, it is no longer being manufactured for human use but is still used in veterinary medicine (180). Side effects recorded in up to 45% of recipients and serious reactions in 10% to 15% of patients forced its withdrawal from the market several years ago (13,17). More than 100 cases of hepatic injury were described with an incidence of overt hepatotoxicity of 1% to 5%, depending on the series (22,181).

452

Lewis

Most patients who developed PBZ hepatotoxicity were adults who had taken the drug for one to six weeks. Men and women appeared to be affected equally; most were over age 30 and one-thirds were older than age 60. Nearly half had hypersensitivity hallmarks such as fever, rash, and eosinophilia. Hepatocellular injury predominated in two-thirds with cholestasis in one-third of cases. Hepatic granulomas were found in 30% of those who underwent liver biopsy (181). The relatively short, fixed latent period, a prompt response to rechallenge, and the high incidence of allergic manifestations and granulomatous hepatitis suggested an immunological mechanism, although intrinsic toxicity most likely explained the injury that was seen in children receiving an overdose, as well as in some experimental models (181). The prognosis from PBZ hepatotoxicity depended on the morphological form of injury. Those with cholestatic features or granulomas usually recovered within a few weeks or months, although one case evolved into chronic cholestasis. A case-fatality rate of 25% was recorded for those with severe hepatic necrosis (22,181). In mice, hepatic and renal cell tumors developed in long-term carcinogenicity studies (182). Oxyphenbutazone is the hydroxylated derivative of PBZ and one of its active metabolites. It shares a similar toxicity profile with the parent compound (13,183) and is not currently marketed. Other pyrazolone derivatives that have been developed and were chemically related to PBZ, include azapropazone and feprazone. They have side effect profiles similar to the parent compound and are not currently available (11,13). Proquazone, a quinazolone compound designed to replace fluproquazone, a parent compound that proved too hepatotoxic for clinical use (8), is not in use in the United States, although it produced a low incidence of elevated hepatic enzyme in initial clinical trials (184). Fenamates (Anthranilic Acids) Mefenamic acid (Ponstelw) has been incriminated in at least one instance of severe but nonfatal necrosis (185). A related agent, meclofenamic acid (Meclomenw), leads to minor elevation of aminotransferases in fewer than 5% of recipients who regress in many instances despite continued use of the drug (186). Cyclooxygenase-2-Selective Inhibitors Nimesulide This weakly acidic sulfonanilide derivative is currently available outside of the United States as one of the new selective COX-2 inhibitors that have fewer gastrointestinal side effects. In clinical trials, nimesulide rarely produced elevations in AST or ALT values (1.6% of patients treated for greater than three months), as reported by McCormick and colleagues (187). However, several recent reports of acute liver injury with this agent have subsequently appeared (5,187–193). The risk of hepatic injury was considered to be relatively low (relative risk 1.4) among 2 million prescriptions for nimesulide and other NSAIDs in 400,000 patients reported by Traversa et al. (194). Nimesulide was associated with the highest number of spontaneously reported adverse reactions among NSAIDs in Northern Italy during the years 1988–2000, with gastrointestinal toxicity significantly outweighing hepatic injury (193). Boelsterli (195) estimated the overall incidence of hepatic injury to be low; 0.1 case per 100,000 patients, with the likely mechanism being formation of reactive intermediates from the nitroarene group. The clinicopathological features of nimesulide hepatotoxicity can be drawn from several reports. Van Steenbergen and colleagues (188) from Belgium described four women with centrilobular or panlobular necrosis and two men with bland intrahepatic cholestasis, where jaundice was the presenting symptom in five of six. Two individuals (one man and one woman) had eosinophilia, suggesting drug allergy, but none of the others had any hypersensitivity signs or symptoms. All abnormal values returned to normal after the drug was discontinued, but this took between 6 and 17 months. One of their patients was diagnosed with pancreatic cancer and succumbed to that illness, which was considered unrelated to the drug. Other reports of acute cholestatic hepatitis from nimesulide have appeared (196), but are less common than hepatocellular injury.

NSAIDs and Leukotriene Receptor Antagonists

453

Several reports of fulminant hepatic necrosis, some fatal, have been described and continue to appear with nimesulide, prompting its withdrawal from a number of countries (194). McCormick and colleagues (187) described a middle-aged woman who developed FHF after restarting nimesulide for back pain. She developed mild to moderate increases in aminotransferase values from a normal baseline, and subsequently developed jaundice and markedly elevated aminotransferases (AST 2014, ALT 2857) at which time the drug was discontinued. Despite undergoing an emergency liver transplant, she died of primary graft nonfunction. The explanted liver revealed massive necrosis. Weiss and colleagues (189) reported six individuals (five of whom were women) with acute hepatitis that resolved when the drug was discontinued. Most had associated fatigue, nausea, and vomiting, with median ALT values elevated 15 times the ULN. These returned to normal within two to four months after the drug was stopped. One patient, however, continued the drug for two weeks after becoming symptomatic and developed a fatal sub-FHF including hepatorenal syndrome and died six weeks later. These authors recommended liver enzyme monitoring with this agent and caution to discontinue it immediately if any biochemical or clinical symptoms of hepatitis develop. Despite these warning, severe injury continues to be described (197,198). Celecoxib is a nonarylamine benzene sulfonamide derivative that was reported to have a very low potential for hepatotoxicity based on clinical trials with exposure of up to two years (199). Since it was approved, a number of reports of reversible cholestatic hepatitis have appeared (200–205), with a latency ranging from just 2 doses (203) to two years (202). Rofecoxib does not contain a sulfonamide moiety and had not been associated with liver toxicity in any published reports (206) prior to being withdrawn due to potential cardiotoxicity (1). However, a review of randomized controlled clinical trials in the literature and in FDA databases by Rostom et al. found a higher rate of aminotransferase values greater than three times ULN for rofecoxib compared with placebo and with other NSAIDs (except diclofenac) (207). Valdecoxib has not been associated with any published ports of hepatotoxicity (208). HEPATIC INJURY DUE TO DISEASE-MODIFYING ANTI-RHEUMATIC DRUGS Methotrexate (MTX) is discussed in Chapter 31. Other DMARDS Leflunomide Leflunomide has been associated with several instances of hepatic injury in the postmarketing setting, although many spontaneous reports lacked sufficient information to assess causality (209). Only a few published reports have appeared. Suissa et al. (210) did not find any increased rate of serious adverse events with leflunomide (RR 0.9), although nonserious hepatic events were found for the use of biological disease modifying anti-rheumatic drugs (DMARDs) (REF). Van Roon et al. (211) reported that grade 2 to 3 elevations in any LAEs occurred in 9% of 101 RA patients receiving leflunomide who were followed for 10 months. No grade 4 (severe) elevations were found in their series. Only 4% developed elevations in aminotransferases; all of whom were asymptomatic and reversible. One in eight patients taking leflunomide and MTX had hepatotoxicity develop; although other studies have not reported significant injury from this combination therapy (212). Sevilla-Mantilla et al. (213) described acute reversible hepatitis in a 67-year-old woman being treated for RA after a latency of 15 days. The mechanism is undefined, but may relate to a reactive metabolite. As a result, LAE monitoring is recommended (214,215). Gold Salts Jaundice occurring in chrysotherapy has been described for decades (216), although it is likely that early reports of acute hepatic injury may have been secondary to viral hepatitis transmitted unwittingly by contaminated needles or syringes (8,16). Cholestatic injury, that in some cases may be prolonged and associated with cholangitis progressing to ductopenia, appears to be the most characteristic hepatic lesion associated with gold treatment (217,218). Immunoallergic

454

Lewis

mechanisms are responsible for the hypersensitivity reaction that is typically seen, which may include eosinophilia and dermatitis. HEPATOTOXICITY OF LEUKOTRIENE RECEPTOR ANTAGONISTS (LRAS) These agents were developed for the treatment and prevention of asthma, based on the role that leukotrienes play in bronchoconstriction and pulmonary inflammation (219,220). Three oral LRAs are currently available, of which zafirlukast, the first of this class to be marketed, has been implicated in numerous cases of liver toxicity, including several instances of fulminant hepatitis leading to liver transplantation (20,21). Only isolated reports of hepatic injury have appeared to date with zileuton (221) and montelukast (222). Zafirlukast Zafirlukast (Accolatew) is a selective competitor of leukotriene D4 and E4 receptors that was not associated with any reports of severe hepatotoxicity in premarketing studies (223). Overall, clinically silent ALT elevations of two to three times normal were found in 1.5% of more than 4000 subjects receiving zafirlukast (often in doses higher than are currently recommended), compared to 1.1% of placebo recipients in clinical trials in North America and Europe. This difference was not statistically significant, the ALT increases were reversible after the drug was stopped, and no cases of jaundice were reported (224). In the postmarketing period, however, more than 100 reports of hepatotoxicity have appeared, including at least 14 instances of liver failure, some of which have resulted in the need for liver transplantation (6,20,21,225–230). I am also aware of at least one other case that I reviewed that may have been due to zafirlukast, but was initially attributed to rosiglitazone (231–233). Conversely, the patient described by Torres and Reddy (230) was also taking LipoKinetix, which has subsequently been linked to severe instances of hepatotoxicity (234). Most patients have been women who were receiving treatment with zafirlukast for a mean of about six months, several of whom developed eosinophilia and other hallmarks of a hypersensitivity reaction, including erythema multiforme or leukocytoclastic vasculitis (20,21,227,225). In one patient reported by Reinus et al. (21), a delayed response to rechallenge with flu-like symptoms and jaundice occurred. In cases of alleged zafirlukast liver injury in which hepatic histology is available (from biopsy or explants), submassive necrosis has been seen in association with an eosinophilic infiltrate (225). Corticosteroids have been administered to a few patients, each of whom recovered clinically (21,227). The mechanism of zafirlukast hepatotoxicity may be immune-mediated hypersensitivity, possibly part of the eosinophilic vasculitis (Churg–Strauss syndrome) that was described in some patients, and supported by at least one positive rechallenge case (21). However, not all reports suggested an allergic basis, and a reactive metabolite may be formed as well, accounting for the relatively long length of exposure prior to toxicity developing (223,224). Given the apparent rarity of severe hepatotoxicity, the FDA does not call for routine LAE monitoring during treatment with zafirlukast, but recommends ALT values be obtained only if there is clinical evidence of acute hepatitis or other symptoms, for which the drug may need to be stopped (235). It is not recommended to be used in individuals with preexisting hepatic dysfunction (20). Montelukast Clinical trials did not suggest a significant difference in reversible AST elevations among recipients of montelukast (1.6%) and those receiving placebo (1.2%), and in contrast to zafirlukast, postmarketing reports of acute hepatic injury are rare. Goldstein and Black (222) described a 37-year-old woman who developed jaundice, nausea, and intense pruritus three weeks after starting montelukast for asthma in the setting of aspirin sensitivity with nasal polyps and allergic rhinitis. Eosinophilia was present with AST 3!ULN, ALT 4.5!ULN, alkaline phosphatase 2.4!ULN, and bilirubin 10.6 mg/dL. Viral serology and autoantibodies were negative. The hyperbilirubinemia resolved within the next five months after the drug was

NSAIDs and Leukotriene Receptor Antagonists

455

stopped, but the aminotransferases remained elevated and a liver biopsy obtained 14 months after the original start of the therapy showed mild chronic portal inflammation without fibrosis. Only one other report of hepatitis has been published to date (236). The mechanism may also be hypersensitivity, as with zafirlukast. Zileuton Premarketing clinical trials of zileuton (Zyflow) found elevations in ALT of 1.9% versus 0.2% in placebo recipients; with one case of reversible hepatitis with jaundice that forced discontinuation of the drug (222). A postmarketing surveillance study of 2458 patients receiving zileuton over a 12-month period also found a higher frequency of reversible aminotransferase elevations greater than 3 times ULN compared to a usual care group (4.6% vs. 1.1%) (222). To date, no published case reports of hepatotoxicity have appeared, but would not be unexpected, with the potential for a toxic metabolite to form (221). MONITORING FOR NSAID AND OTHER ANTI-INFLAMMATORY DRUG-INDUCED HEPATIC INJURY NSAID Monitoring Although the FDA considers hepatic injury to be a class effect of NSAIDs, the agency stopped short of recommending mandatory enzyme monitoring during NSAID therapy (17). However, in the case of diclofenac, monitoring has been recommended by other authorities (237). Table 4 includes the information from monitoring statements included in manufacturers’ current package inserts. While class labeling varies somewhat, it generally mentions that for any given agent, liver abnormalities are possible and may progress, may remain unchanged, or may be transient with continued therapy. For many compounds, the labeling adds that cases of severe hepatic injury, including jaundice and even fatal fulminant hepatitis, have been reported. For these particular agents, physicians should be aware of the potential toxicity and remain alert for abnormal hepatic enzymes that persist or worsen, to clinical signs and symptoms of liver disease, or systemic manifestations such as eosinophilia, rash, or fever. For a few agents, mention is made that they should be used cautiously in individuals with underlying chronic liver disease. NSAIDs, in general, are usually best avoided in cirrhosis because of a risk of renal toxicity leading to hepatorenal syndrome, the possible exceptions being sulindac and the newer COX-2-selective agents. More intensive monitoring in patients with chronic hepatitis or cirrhosis seems prudent for NSAIDs as well as other agents, but uniform recommendations await additional study. It is recognized that liver enzyme monitoring is controversial from a clinical, as well as a cost, standpoint. For drugs whose incidence of injury is extremely low, routine monitoring is not warranted and, if prescribed, would probably not be diligently followed. For injury due to hypersensitivity, a rational case for no biochemical monitoring also could be made, since the drug toxicity announces itself when the hypersensitivity reaction develops. It is unlikely that monitoring would serve as a harbinger of impending hepatic injury in this setting (8). In contrast, NSAIDs capable of causing severe hepatic injury and that act through metabolic idiosyncrasy should be monitored on a periodic basis (11,18,238). For those agents where a greater than threefold rise in ALT from a normal baseline is seen, the specter of hepatotoxicity is raised and the monitoring frequency should be increased. If the abnormality does not subside or if it progresses, the drug should probably be stopped. In the event that clinical signs or symptoms of liver disease develop (i.e., nausea, fatigue, lethargy, pruritus, abdominal discomfort, in addition to jaundice), the drug should be discontinued immediately. If the biochemical abnormality resolves despite continuation of therapy, the NSAID can be continued. While the performance of biochemical testing, periodic symptom assessments and physical examinations do not guarantee the detection or prevention of NSAID-induced or any drug-induced hepatic injury, it is the expectation of any reasonably designed monitoring program that abnormalities detected early will prevent progression to more serious or irreversible toxicity (239).

456

Lewis

TABLE 4 Current Monitoring Recommendations to Prevent NSAID Hepatotoxicity Among Agents Available in the United States Class/agent Salicylates Aspirin Diflunisal (Dolobid) Sodium, choline salicylate Salsalate (Disalcid, salflex) Acetic acid derivatives Diclofenac (voltaren)

Manufacturer’s product information Transient elevation LAEs, hepatitis CL

Reye’s syndrome warning (for plain, enteric coated) increased toxicity at high doses during pediatric use Potentially life-threatening hypersensitivity involving the liver

CL

Reye’s syndrome warning, periodically monitor blood levels

CL, FFH, OLT mentioned LAEs

No change in metabolism or elimination in cirrhosis; measure AST and ALT in first 4 wk, and periodically to 24 wk. Inform patient of warning signs of hepatotoxicity

Nabumetone (Relafen)

!3!ULN in 15% O3! in 4% O8! in 1% CL; FFH mentioned CL; FFH-mentioned hypersensitivity including severe skin reactions CL; anaphylaxis, FFH mentioned CL; FFH mentioned

Etodolac (Lodine)

CL; FFH mentioned

Ketorolac (Toradol)

Modified CL; hepatitis, liver failure mentioned

Indomethacin (Indocin) Sulindac (Clinoril)

Tolmetin (Tolectin)

Propionic acid derivatives Ibuprofen

CL; FFH mentioned

Naproxen

CL; FFH mentioned

Fenoprofen (Nalfon)

CL; FFH mentioned

Flurbiprofen (Ansaid) Oxaprozin (Daypro) Ketoprofen (Orudis)

Oxicams Pkiroxicam (Feldene) Fenamates Mefenamic acid (Ponstel) Meclofenamic acid (Meclomen) Cyclooxygenase-2 inhibitors Celecoxib Rofecoxiba Valdecoxib

Manufacturer’s monitoring recommendations

D/C for signs or symptoms of liver disease Hypersensitivity reactions and cholestatic hepatitis may occur, monitor closely in patients with poor liver function; LFTs checked whenever a patient develops hypersensitivity features and D/C drug Discontinue for signs of livery injury, hypersensitivity Use caution in patients with severe liver impairment as metabolism may be reduced No change in compensated cirrhosis but decrease dose in severe hepatic failure; false-positive urine test for bilirubin due to phenolic metabolites Clearance not affected by low albumin in cirrhosis but preexisting liver dysfunction may lead to more severe hepatic reaction; D/C if abnormal LAEs occur Evaluate for evidence of more severe hepatic reaction if liver injury occurs; discontinue for signs of allergy Use caution in cirrhosis and in patients with renal impairment. D/C if LFTs worsen Use under strict observation in patients with impaired liver function; monitor LFTs periodically during long-term therapy

CL; may be risk of fatal hepatitis CL; serious hepatic reactions, jaundice reported

No dose adjustment necessary in compensated liver disease; use caution in severe liver disease; D/C if abnormal LFTs worsen No change in drug disposition in cirrhosis but carefully monitor and keep dose at a minimum since unbound biologically active fraction is doubled in hypoalbuminemia; use lower doses in patients with albumin !3.5 g/dL or hepatic impairment

CL; FFH mentioned

Discontinue if signs of liver injury or allergy develop

CL; FFH mentioned

Reduce dose in liver dysfunction; discontinue if signs of liver injury persist or worsen

CL

Reduced dose for moderate hepatic impairment; not recommended for use in severe hepatic disease; monitor carefully if abnormal LFTs occur Limited data in patients with hepatic impairment; monitor carefully if abnormal LFTs occur

CL

a Rofecoxib has been withdrawn due to the risk of cardiac toxicity. Abbreviations: CL, class labeling (see text); FFH, fatal fulminant hepatitis; OLT, liver transplantation; D/C, discontinue; LAE, liverassociated enzymes

NSAIDs and Leukotriene Receptor Antagonists

457

ALT values should be obtained at baseline and monitored monthly. If elevations exceed two to three times ULN, consideration should be given to liver biopsy and/or stopping the drug. Leukotriene Receptor Antagonists Given the apparent immunoallergic mechanism of injury, clinical symptomatology (rash, pruritus, and symptoms of acute hepatitis or jaundice) should be sought and the drug stopped. Concomitant elevations in ALT values are considered another indication to discontinue therapy. REFERENCES 1. Graham DJ, Campen D, Hui R, et al. Risks of acute myocardial infarction and sudden cardiac death in patients treated with cyclo-oxygenase 2 selective and non-selective non-steroidal anti-inflammatory drugs: nested case-control study. Lancet 2005; 365:475–81. 2. Friis H, Andreasen PB. Drug-induced hepatic injury: an analysis of 1,100 cases reported to the Danish Committee on Adverse Drug Reactions between 1978 and 1987. J Intern Med 1992; 232:133–8. 3. Koff RS. Liver disease induced by nonsteroidal anti-inflammatory drugs. In: Borda JT, Koff RS, eds. A Profile of Adverse Effects. St. Louis, MO: Mosby-Year Book, 1992:133–45. 4. Bjornsson E, Olsson R. Outcome and prognostic markers in severe drug-induced liver disease. Hepatology 2005; 42:481–9. 5. Andrade RJ, Lucena MI, Fernandez MC, et al. Drug-induced liver injury: an analysis of 461 instances submitted to the Spanish Registry over a 10-year period. Gastroenterology 2005; 129:512–21. 6. Russo MW, Galanko JA, Shrestha R, et al. Liver transplantation for acute liver failure from druginduced liver injury in the United States. Liver Transpl 2004; 10:1018–23. 7. O’Connor N, Dargan PI, Jones AL. Hepatocellular damage from non-steroidal anti-inflammatory drugs. Q J Med 2003; 96:787–91. 8. Zimmerman HJ. Hepatotoxicity: The Adverse Effects of Drugs and Other Chemicals on the Liver. 2nd ed. Philadelphia, PA: Lippincott, William & Wilkins, 1999. 9. Palmer WL, Woodall PS, Wang KC. Cinchophen and toxic necrosis of the liver, a survey of the problem. Trans Assoc Am Phys 1936; 51:381–93. 10. Heuper WE. Cinchophen (Atophan): a critical review. Medicine 1948; 27:43. 11. Lewis JH. Hepatic toxicity of nonsteroidal anti-inflammatory drugs. Clin Pharm 1984; 3:128–38. 12. Strieker BHC, Blok AP, Bronkhorst FB. Glafenine-associated hepatic injury: analysis of 38 cases and review of the literature. Liver 1986; 6:63–72. 13. Cuthbert MF. Adverse reactions to nonsteroidal anti-inflammatory drugs. Curr Med Res Opin 1974; 2:600–10. 14. Hart FD, Bain LS, Huskisson EC, et al. Hepatic effects of fenclozic acid. Ann Rheum Dis 1970; 29:684. 15. Taggart HM, Alderdice JM. Fatal cholestatic jaundice in elderly patients taking benoxaprofen. Br Med J (Clin Res Ed) 1982; 284:1372. 16. Lewis JH. NSAID-induced hepatotoxicity. Clin Liver Dis 1998; 2:543–61. 17. Paulus HE. FDA Arthritis Advisory Committee meeting. Arthritis Rheum 1981; 25:1124–5. 18. Lewis JH, Zimmerman HJ. NSAID hepatotoxicity. IM Intern Med 1996; 17:45–67. 19. Fontana RJ, McCashland TM, Benner KG, et al. Acute liver failure associated with prolonged use of bromfenac leading to liver transplantation: the acute liver failure study group. Liver Transpl Surg 1999; 5:480–4. 20. Wooltorton E. Asthma drug zafirlukast (Accolate): serious hepatic events. Can Med Assoc J 2004; 170:1668. 21. Reinus JF, Persky S, Burkiewicz JS, et al. Severe liver injury after treatment with the leukotriene receptor antagonist zafirlukast. Ann Intern Med 2000; 133:964–8. 22. Zimmerman HJ. Update of hepatotoxicity due to classes of drugs in common clinical use: nonsteroidal anti-inflammatory drugs, antibiotics, antihypertensives, and cardiac and psychotropic agents. Semin Liver Dis 1990; 10:322–8. 23. Carson JL, Strom BL, Duff A, et al. Safety of nonsteroidal anti-inflammatory drugs with respect to acute liver disease. Arch Intern Med 1993; 153:1331–6. 24. Garcia Rodriguez LA, Williams R, Derby LE, et al. Acute liver injury associated with nonsteroidal anti-inflammatory drugs and the role of risk factors. Arch Intern Med 1994; 154:311–6. 25. Walker AM. Quantitative studies of the risk of serious hepatic injury in persons using nonsteroidal anti-inflammatory drugs. Arthritis Rheum 1997; 40:201–8. 26. Lacroix I, Lapeyre-Mestre M, Bagheri H, et al. Nonsteroidal anti-inflammatory drug-induced liver injury: a case-control study in primary care. Fundam Clin Pharmacol 2004; 18:201–6.

458

Lewis

27. Singh G, Ramey DR, Morfeld D, et al. Comparative toxicity of non-steroidal anti-inflammatory agents. Pharmacol Ther 1994; 62:175–91. 28. Miwa LJ, Jones JK, Pathiyal A, et al. Value of epidemiologic studies in determining the true incidence of adverse events: the nonsteroidal anti-inflammatory drug story. Arch Intern Med 1997; 157:2129–36. 29. Mills PR, Sturrock RD. Clinical associations between arthritis and liver disease. Ann Rheum Dis 1982; 41:295–307. 30. Seaman WE, Plotz PH. Effect of aspirin on liver tests in patients with RA and SLE and in normal volunteers. Arthritis Rheum 1976; 19:155–60. 31. Weinblatt ME, Tesser JR, Gilliam JH, et al. The liver in rheumatic disease. Semin Arthritis Rheum 1982; 11:399–405. 32. Bailey M, Chapin W, Licht H, et al. The effects of vasculitis on the gastrointestinal tract and liver. Gastroenterol Clin North Am 1998; 98:747–82. 33. Runyon BA, LaBrecque DR, Anuras S. The spectrum of liver disease in systemic lupus erythematosus: report of 33 histologically proved cases and review of the literature. Am J Med 1980; 69:187–94. 34. Lewis JH. Drug-induced liver disease. Med Clin North Am 2000; 84:1275–311. 35. Lewis JH. “Hy’s Law”, the “Rezulin Rule” and other predictors of severe drug-induced hepatotoxicity: putting risk-benefit into perspective. Pharmacoepidemiol Drug Saf 2006; 15(4):221–9. 36. Jones AL, Latham T, Shallcross TJ, et al. Fulminant hepatic failure due to diclofenac treated successfully by orthotopic liver transplantation. Transplant Proc 1998; 30:192–4. 37. Rodriguez-Gonzalez FJ, Montero JL, Puente J, et al. Orthotopic liver transplantation after subacute liver failure induced by therapeutic doses of ibuprofen. Am J Gastroenterol 2002; 97:2476–7. 38. Lewis JH, Zimmerman JH. Drug- and chemical-induced cholestasis. Clin Liver Dis 1999; 3:433–64. 39. Zimmerman HJ. Effects of aspirin and acetaminophen on the liver. Arch Intern Med 1981; 141:333–42. 40. Tomoda T, Kurashige T, Hayashi Y, et al. Primary changes in liver damage by aspirin in rats. Acta Paediatr Jpn 1998; 40:593–6. 41. Janota J, Wincey CW, Sandiford M, et al. Effect of salicylate on the activity of plasma enzymes in the rabbit. Nature 1960; 185:935–6. 42. Tolman KG, Peterson P, Gray P, et al. Hepatotoxicity of salicylates in monolayer cell cultures. Gastroenterology 1978; 74:205–8. 43. Singh H, Chugh JC, Shembesh AH, et al. Hepatotoxicity of high dose salicylate therapy in acute rheumatic fever. Ann Trop Paediatr 1992; 12:37–40. 44. Levy G, Yaffe SJ. Clinical implications of salicylate-induced liver damage. Am J Dis Child 1975; 129:1385. 45. Ingleman-Sundberg M, Kaur H, Terelius Y, et al. Hydroxylation of salicylate by microsomal fractions and cytochrome P-450: lack of production of 2,3-dihydroxybenzoate unless hydroxyl radical formation is permitted. Biochem J 1991; 276:753–7. 46. O’Gorman T, Koff RS. Salicylate hepatitis. Gastroenterology 1977; 72:726–8. 47. Athreya BH, Moser G, Cecil HS, et al. Aspirin-induced hepatotoxicity in juvenile rheumatoid arthritis: a prospective study. Arthritis Rheum 1975; 18:347–52. 48. Rich RR, Johnson JJ. Salicylate hepatotoxicity in patients with juvenile rheumatoid arthritis. Arthritis Rheum 1973; 16:1–9. 49. Bernstein BH, Singsen GH, King KK, et al. Aspirin-induced hepatotoxicity and its effect on juvenile rheumatoid arthritis. Am J Dis Child 1977; 131:659–63. 50. Bell CL, Schur PH. Juvenile rheumatoid arthritis and salicylate related liver chemistry abnormalities: clinical and genetic considerations (abstr). Arthritis Rheum 1980; 22:592. 51. Hawton K, Simkin S, Deeks J, et al. U.K. legislation on analgesic packs: before and after study of long term effect on poisonings. Br Med J 2004; 329:1076. 52. Koff RS, Galdabini JJ. Fever, myalgias and hepatic failure in a 17 year old girl. N Engl J Med 1977; 296:1337–46. 53. Gitlin N. Salicylate hepatotoxicity: the potential role of hypoalbuminemia. J Clin Gastroenterol 1980; 2:281. 54. Okamura H, Ichikawa T, Obayashi K, et al. Studies on aspirin-induced hepatic injury. Recent Adv Gastroenterol 1967; 3:223. 55. Zimmerman HJ. Drug-induced chronic liver disease. Med Clin North Am 1979; 63:567–82. 56. Soller RW, Slander H. Association between salicylates and Reye’s syndrome. JAMA 1983; 249:883. 57. Starko KM, Ray CG, Dominques LB, et al. Reye’s syndrome and salicylate use. Pediatrics 1980; 66:859–64. 58. Arrowsmith JB, et al. National patterns of aspirin use and Reye syndrome reporting, United States, 1980–1985. Pediatrics 1987; 79:858–63. 59. Peters LJ, Wiener GJ, Gilliam J, et al. Reye’s syndrome in adults: a case report and review of the literature. Arch Intern Med 1986; 146:2401.

NSAIDs and Leukotriene Receptor Antagonists

459

60. Meythaler JM, Varma RR. Reye’s syndrome in adults: diagnostic considerations. Arch Intern Med 1987; 147:61–4. 61. Forsyth BW, Horwithz RI, Acampora D, et al. New epidemiologic evidence confirming that bias does not explain the aspirin/Reye’s syndrome association. JAMA 1989; 261:2517–24. 62. Remington PL, Rawley D, McGee H, et al. Decreasing trend in Reye’s syndrome and aspirin use in Michigan, 1979 to 1984. Pediatrics 1986; 77:93–8. 63. Belay ED, Bresee JS, Holman RC, et al. Reye’s syndrome in the United States from 1981 through 1997. N Engl J Med 1999; 340:1377–82. 64. Pinsky PF, Hurwitz ES, Schonberger LB, et al. Reye’s syndrome and aspirin: evidence for a dose– response effect. JAMA 1988; 260:657–61. 65. Martens ME, Change CH, Lee CP. Reye’s syndrome: mitochondrial swelling and calcium release induced by Reye’s plasma, allantoin, and salicylate. Arch Biochem Biophys 1986; 244:773–86. 66. Crocker JF, Digout SC, Lee SH, et al. Effects of antipyretics on mortality due to influenza B virus in a mouse model of Reye’s syndrome. Clin Invest Med 1998; 21:192–202. 67. Orlowski JP. Whatever happened to Reye’s syndrome? Did it ever really exist? Crit Care Med 1999; 27:1582–7. 68. Nadkarni MM, Peller CA, Retig J. Eosinophilic hepatitis after ingestion of choline magnesium trisalicylate. Am J Gastroenterol 1992; 87:151–3. 69. Cersoimo RJ, Matthews SJ. Hepatotoxicity associated with choline magnesium trisalicylate: case report and review of salicylate-induced hepatotoxicity. Drug Intell Clin Pharm 1987;(21):621–5. 70. Cook DJ, Achong MR, Murphy FR. Three cases of diflunisal hypersensitivity. Can Med Assoc J 1988; 138:1029–30. 71. Warren NS. Diflunisal-induced cholestatic jaundice. Br Med J 1978; 2:736–7. 72. Brogden RN, Heel RC, Pakes GE, et al. Diflunisal: a review of its pharmacological properties and therapeutic use in pain and musculoskeletal strains and sprains and pain in osteoarthritis. Drugs 1980; 19:84–106. 73. Masubuchi Y, Saito H, Horie T. Structural requirements for the hepatotoxicity of nonsteroidal antiinflammatory drugs in isolated rat hepatocytes. J Pharmacol Exp Ther 1998; 287:208–13. 74. Symon DN, Gray ES, Hanmer OJ, et al. Fatal paracetamol poisoning from benorylate therapy in child with cystic fibrosis. Lancet 1982; 2:1151–2. 75. Banks AT, Zimmerman HJ, Ishak KG, et al. Diclofenac-associated hepatotoxicity: analysis of 180 cases reported to the Food and Drug Administration. Hepatology 1995; 22:820–7. 76. Breen EG, McNicholl J, Cosgrove E, et al. Fatal hepatitis associated with diclofenac. Gut 1986; 27:1390–3. 77. Helfgott SM, Sanberg-Cook J, Zakim D, et al. Diclofenac-associated hepatotoxicity. JAMA 1990; 264:2660–2. 78. Purcell P, Henry D, Melville G. Diclofenac hepatitis. Gut 1991; 32:1381–5. 79. Ciucci AG. A review of spontaneously reported adverse drug reactions with diclofenac sodium (Voltarol). Rheumatol Rehabil 1979; 2(Suppl.):116–21. 80. Fry SW, Seeff LB. Hepatotoxicity of analgesics and anti-inflammatory agents. Gastroenterol Clin North Am 1995; 24:875–905. 81. Perez-Gutthann S, Garcia-Rodriguez LA, Duque-Oliart A, et al. Low-dose diclofenac, naproxen, and ibuprofen cohort study. Pharmacotherapy 1999; 19:854–9. 82. Hackstein H, Mohl W, Puschel W, et al. Diclofenac-associated acute cholestatis hepatitis. Z Gastroenterol 1998; 36:385–9. 83. Scully W, Clarke D, Barr RJ. Diclofenac induced hepatitis: three cases with features of autoimmune chronic hepatitis. Dig Dis Sci 1993; 38:744–51. 84. Sallie R. Diclofenac hepatitis (letter). J Hepatol 1990; 11:281. 85. Wade LT, Kenna JG, Caldwell J. Immunochemical identification of mouse hepatic protein adducts derived from the nonsteroidal anti-inflammatory drugs diclofenac, sulindac, and ibuprofen. Chem Res Toxicol 1997; 10:546–55. 86. Masubuchi Y, Yamada S, Horie T. Possible mechanism of hepatocyte injury induced by diphenylamine and its structurally related nonsteroidal anti-inflammatory drugs. J Pharmacol Exp Ther 2000; 292:982–7. 87. Bort R, Ponsoda X, Jover R, et al. Diclofenac toxicity to hepatocytes; a role for drug metabolism in cell toxicity. J Pharmacol Exp Ther 1999; 288:65–72. 88. Seitz S, Kretz-Rommel A, Oude Elferink RP, et al. Selective protein adduct formation of diclofenac glucuronide is critically dependent on the rat canalicular conjugate export pump (Mrp2). Chem Res Toxicol 1998; 11:513–9. 89. Miyamoto G, Zahid N, Uetrecht JP. Oxidation of diclofenac to reactive intermediates by neutrophils, myeloperoxidase, and hypochlorous acid. Chem Res Toxicol 1997; 10:414–9. 90. Tang W. The metabolism of diclofenac—enzymology and toxicology perspectives. Curr Drug Metab 2003; 4:319–29.

460

Lewis

91. Boelsterli UA. Diclofenac-induced liver injury: a paradigm of idiosyncratic drug toxicity. Toxicol Appl Pharmacol 2003; 192:307–22. 92. Grillo MP, Hua F, Knutson CG, et al. Mechanistic studies on the bioactivation of diclofenac: identification of diclofenac-S-acyl-glutathione in vitro in incubations with rat and human hepatocytes. Chem Res Toxicol 2003; 16:1410–7. 93. Bort R, Mace K, Boobis A, et al. Hepatic metabolism of diclofenac: role of human CYP in the minor oxidative pathways. Biochem Pharmacol 1999; 58:787–96. 94. Aithal GP, Ramsay L, Daly AK, et al. Hepatic adducts, circulating antibodies, and cytokine polymorphisms in patients with diclofenac hepatotoxicity. Hepatology 2004; 39:430–40. 95. Ohana M, Hajiro K, Takakuwa H, et al. Recovery from diclofenac-induced hypersensitive fulminant hepatitis and prostaglandins. Dig Dis Sci 1997; 42:2031–2. 96. Schattenkirchner M. An updated safety profile of etodolac in several thousand patients. Eur J Rheumatol Inflamm 1990; 10:56. 97. Mabee CL, Mabee SW, Baker PB, et al. Fulminant hepatic failure associated with etodolac use. Am J Gastroenterol 1995; 90:659. 98. Ogiso T, Kitagawa T, Iwaki M, et al. Pharmacokinetic analysis of enterohepatic circulation of etodolac and effect of hepatic and renal injury on the pharmacokinetics. Biol Pharm Bull 1997; 20:405–10. 99. Etodolac product information. Physicians’ Desk Reference. 54th ed. Montvale, NJ: Medical Economics Company, 2000. 100. Ketorolac product information. Physicians’ Desk Reference. 54th ed. Montvale, NJ: Medical Economics Company, 2000. 101. Moses PL, Schroeder B, Alkhatib O, et al. Severe hepatotoxicity associated with bromfenac sodium. Am J Gastroenterol 1999; 94:1393–6. 102. Hunter EB, Johnson PE, Tanner G, et al. Bromfenac (Duract)-associated hepatic failure requiring liver transplantation. Am J Gastroenterol 1999; 94:2299–301. 103. Goldkind L, Laine L. A systematic review of NSAIDs withdrawn from the market due to hepatotoxicity: lessons learned from the bromfenac experience. Pharmacoepidemiol Drug Saf 2006; 15(4):213–20. 104. Skjodt NM, Davies NM. Clinical pharmacokinetics and pharmacodynamics of bromfenac. Clin Pharmacokinet 1999; 36:399–408. 105. Fenech FF, Bannister WH, Grech JL. Hepatitis with biliberdinaemia in association with indomethacin therapy. Br Med J 1967; 3:155–6. 106. Jacobs JS. Sudden death in arthritic children receiving large doses of indomethacin. JAMA 1967; 199:93. 107. Kelsey WM, Scharyi M. Fatal hepatitis probably due to indomethacin. JAMA 1967; 199:154–5. 108. Boardman PL, Hart FD. Side-effects of indomethacin. Ann Rheum Dis 1967; 26:127–32. 109. Barriault C, Audet M, Yousef IM, et al. Protection of indomethacin against the lethality and hepatotoxicity of phalloidin in mice. Toxicol Lett 1994; 71:257–69. 110. Makogon NV, Lushnikova IV, Korneitchuk AN, et al. Effects of noridhydroguaiaretic acid and indomethacin on the viability and functional activities of normal and carbon tetrachloride-injured rat hepatocytes cultured alone and with Kupffer cells. Acta Physiol Pharmacol Bulg 1998; 23:33–8. 111. Bahrami H, Daryani NE, Haghpanah B, et al. Effects of indomethacin on viral replication markers in asymptomatic carriers of hepatitis B: a randomized, placebo-controlled trial. Am J Gastroenterol 2005; 100:856–61. 112. Smith FE, Lindberg PJ. Life-threatening hypersensitivity to sulindac. JAMA 1980; 244:269. 113. Brogden RN, Heel RC, Speight TM, et al. Sulindac: a review of its pharmacological properties and therapeutic efficacy in rheumatic diseases. Drugs 1978; 16:97–114. 114. Park GD, Spector R, Headstream T, et al. Serious adverse reactions associated with sulindac. Arch Intern Med 1982; 142:1292–4. 115. Garcia Rodriguez LA, Perez-Gutthann S, Walker AM, et al. The role of nonsteroidal antiinflammatory drugs in acute liver injury. Br Med J 1992; 305:865–8. 116. Kromann-Anderson H, Pedersen A. Reported adverse reactions and consumption of nonsteroidal anti-inflammatory drugs in Denmark over a 17-year period. Danish Med Bull 1988; 35:187–92. 117. Tarazi EM, Harter JG, Zimmerman HJ, et al. Sulindac-associated hepatic injury: analysis of 91 cases reported to the Food and Drug Administration. Gastroenterology 1992; 104:569–74. 118. Lerche A, Vyberg M, Kirkegaard E. Acute cholangitis and pancreatitis associated with sulindac (Clinoril). Histopathology 1987; 11:647. 119. Levitt L, Pearson RW. Sulindac-induced Stevens–Johnson toxic epidermal neurolysis syndrome. JAMA 1980; 243:1262. 120. Maguire FW. Stevens–Johnson syndrome due to sulindac: a case report and review of the literature. Del Med J. 1981; 53(4):193–7. 121. Whittaker SJ, Amar JN, Wanless IR, et al. Sulindac hepatotoxicity. Gut 1982; 23:875–7.

NSAIDs and Leukotriene Receptor Antagonists

461

122. Calabro J, Marchesano J, Partirdge R, et al. Sulindac in juvenile rheumatoid arthritis. Clin Pharmacol Ther 1979; 25:216. 123. Laffi G, Daskalopoulos G, Kronborg I, et al. Effects of sulindac and ibuprofen in patients with cirrhosis and ascites: an explanation for the renal-sparing effect of sulindac. Gastroenterology 1986; 90:182–7. 124. Shaw GR, Anderson WR. Multisystem failure and hepatic microvesicular fatty metamorphosis associated with tolmetin ingestion. Arch Pathol Lab Med 1991; 115:818–21. 125. O’Brien WM. Long-term efficacy and safety of tolmetin sodium in treatment of geriatric patients with rheumatoid arthritis and osteoarthritis: a retrospective study. J Clin Pharmacol 1983; 23:309–23. 126. Bernhard GC. Worldwide safety experience with nabumetone. J Rheumatol 1992; 19(Suppl.):48–57. 127. Hedner T, Samulesson O, Wahrborg P, et al. Nabumetone: therapeutic use and safety profile in the management of osteoarthritis and rheumatoid arthritis. Drugs 2004; 64:2315–43. 128. Pessayre D, Degos F, Feldmann G, et al. Chronic active hepatitis and giant multinucleated hepatocytes in adults treated with clometacin. Digestion 1982; 22:66–72. 129. Islam S, Mekhloufi F, Paul JM, et al. Characteristics of clometacin-induced hepatitis with special reference to the presence of anti-actin cable antibodies. Autoimmunuity 1989; 2:213–21. 130. Pariente EA, Hamoud A, Goldfain D, et al. Hepatitis caused by clometacin (Duperan). Retrospective study of 30 cases. A model of autoimmune drug-induced hepatitis? Gastroenterol Clin Biol 1989; 13:769–74. 131. Tordjmann T, Grimbert S, Genestie C, et al. Adult multi-nuclear cell hepatitis: a study of 17 patients. Gastroenterol Clin Biol 1998; 22:305–10. 132. Poitrine A, Poynard T, Naveau S, et al. A new fatal case of hepatitis caused by clometacin. Gastroenterol Clin Biol 1983; 7:99. 133. Freeland GR, Northington RS, Hedrich DA, et al. Hepatic safety of two analgesics used over the counter: ibuprofen and aspirin. Clin Pharmacol Ther 1988; 43:473–9. 134. Royer GL, Seckman CE, Welshman IR. Safety profile: fifteen years of clinical experience with ibuprofen. Am J Med 1984; 77:25–34. 135. Sternlieb P, Robinson RM. Stevens–Johnson syndrome plus toxic hepatitis due to ibuprofen. NY State J Med 1978; 78:1239–43. 136. Stempel DA, Miller JJ, III. Lymphopenia and hepatic toxicity with ibuprofen. J Pediatr 1977; 90:657–8. 137. Bravo JF, Jacobson MP, Mertens BF. Fatty liver and pleural effusion with ibuprofen therapy. Ann Intern Med 1977; 87:200–1. 138. Lee CY, Finkler A. Acute intoxication due to ibuprofen overdose. Arch Pathol Lab Med 1986; 110:747–9. 139. Castell JV, Larrauri A, Gomez-Lechon MJ. A study of the relative hepatotoxicity in vitro of the nonsteroid anti-inflammatory drugs ibuprofen, flurbiprofen and butibufen. Xenobiotica 1988; 18:737–45. 140. Al-Nasser IA. Ibuprofen-induced liver mitochondrial permeability. Toxicol Lett 2000; 111:213–8. 141. Bendele AM, Hulman JF, White S, et al. Hepatocellular proliferation in ibuprofen-treated mice. Toxicol Pathol 1993; 21:15–20. 142. Alam I, Ferrell LD, Bass NM. Vanishing bile duct syndrome temporally associated with ibuprofen use. Am J Gastroenterol 1996; 91:1626–30. 143. Riley TR, III, Smith JP. Ibuprofen-induced hepatotoxicity in patients with chronic hepatitis C: a case series. Am J Gastroenterol 1998; 93:1563–5. 144. Bass BG. Jaundice associated with naproxen (letter). Lancet 1974; 1:998. 145. Giarelli L, Falconieri G, Delender M. Fulminant hepatitis following naproxen administration. Hum Pathol 1986; 17:1079. 146. Law JP, Knight H. Jaundice associated with naproxen. N Engl J Med 1976; 295:1201. 147. Fredell EW, Strand LJ. Naproxen overdose (letter). JAMA 1997; 238:938. 148. Brogden RN, Heel RC, Speight TM, et al. Naproxen up to date: a review of its pharmacological properties and therapeutic efficacy and use in rheumatic diseases and pain states. Drugs 1979; 18:241–77. 149. Brogden RN, Pinder RM, Speight TM, et al. Fenoprofen: a review of its pharmacological properties and therapeutic efficacy in rheumatic diseases. Drugs 1977; 13:241–65. 150. Andrejak M, Davion T, Gineston JL, et al. Cross hepatotoxicity between nonsteroidal antiinflammatory drugs. Br Med J 1987; 295:180. 151. Stennett DJ, Simonson W, Hall CA. Fenoprofen-induced hepatotoxicity (letter). Am J Hosp Pharm 1978; 35:901. 152. Brogden RN, Heel RC, Speight TM, et al. Fenbufen: a review of its pharmacological properties and therapeutic use rheumatic disease and acute pain. Drugs 1981; 21:1–22. 153. Kotowski KE, Grayson MF. Side effects of non-steroidal anti-inflammatory drugs. Br Med J 1982; 285:377. 154. Zimmerman HJ. Hepatic effects of oxaprozin. Semin Arthritis Rheum 1986; 15:35–42.

462

Lewis

155. Purdum PE, Shelden SL, Boyd JW, et al. Oxaprozin-induced fulminant hepatitis. Ann Pharmacother 1994; 28:1159. 156. Kethu SR, Rukkannagari S, Lansford CL. Oxaprozin-induced symptomatic hepatotoxicity. Ann Pharmacother 1999; 33:942–4. 157. Ketoprofen product information. Physicians’ Desk Reference. 59th ed. Montvale, NJ: Medical Economics Company, 2005. 158. Depla ACTM, Vermeersch PH M J, Gorp LHM, et al. Fatal acute liver failure associated with pirprofen. Report of a case and a review of the literature. Neth J Med 1990; 37:32. 159. Anonymous. Arthritis drug withdrawn. FDA Consum Bull 1982; 16:4. 160. Halsey J, Cardoe N. Benoxaprofen: side effect profile in 300 patients. Br Med J (Clin Res Ed) 1982; 284:1385–8. 161. Prescott LE, Leslie PJ, Padfield P. Side effects of benoxaprofen. Br Med J 1982; 284:1782. 162. Goudie BM, Birnie GF, Watkinson G, et al. Jaundice associated with the use of benoxaprofen. Lancet 1982; 1:959. 163. Hamdy RC, Mumang B, Perera N, et al. The pharmacokinetics of benoxaprofen in elderly subjects. Eur J Rheumatol Inflamm 1982; 5:69. 164. Knights KM, Cassidy MR, Drew R. Benoxaprofen induced toxicity in isolated rat hepato-cytes. Toxicology 1986; 40:327–39. 165. Sorensen EM. Morphometric analysis of cultured hepatocytes exposed to benoxaprofen. Toxicol Lett 1986; 34:277–86. 166. Lewis DF, loannides C, Parke DV. A retrospective study of the molecular toxicology of benoxaprofen. Toxicology 1990; 65:33–47. 167. Ayrton AD, loannides C, Parke DV. Induction of the cytochrome P450 I and IV families and peroxisomal proliferation in the liver of rats treated with benoxaprofen. Possible implication in its hepatotoxicity. Biochem Pharmacol 1991; 42:109–15. 168. Dong JQ, Liu J, Smith PC. Role of benoxaprofen and flunoxaprofen acyl glucuronides in covalent binding to rat plasma and liver proteins in vivo. Biochem Pharmacol 2005; 70:937–48. 169. Lee SM, O’Brien CJ, Williams R, et al. Subacute hepatic necrosis induced by piroxicam. Br Med J 1986; 293:540–1. 170. Paterson D, Kerlin P, Walker N, et al. Piroxicam induced submassive necrosis of the liver. Gut 1992; 33:1436–8. 171. Planas R, De Leon R, Quer JC, et al. Fatal submassive necrosis of the liver associated with piroxicam. Am J Gastroenterol 1990; 85:468–70. 172. Hepps KS, Maliha GM, Estrada R, et al. Severe cholestatic jaundice associated with piroxicam. Gastroenterology 1991; 101:1737–40. 173. MacDougall LG, Taylor-Smith A, Rothberg AD, et al. Piroxicam poisoning in a 2-year-old child: a case report. S Afr Med J 1984; 66:31–3. 174. Salgueiro-Pagadigorria CL, Kelmer-Bracht AM, Bracht A, et al. Effects of nonsteroidal antiinflammatory drug piroxicam on rat liver mitochondria. Comp Biochem Physiol C Pharmacol Toxicol Endocrinol 1996; 113:85–91. 175. Salgueiro-Pagadigorria CL, Constantin J, Bracht A, et al. Effects of nonsteroidal anti-inflammatory drug piroxicam on energy metabolism in the perfused rat liver. Comp Biochem Physiol C Pharmacol Toxicol Endocrinol 1996; 113:93–8. 176. Zentella de Pina M, Corona S, Rocha-Hernandez AE, et al. Restoration by prioxicam of liver glutathione levels decreased by acute ethanol intoxication. Life Sci 1994; 54:1433–9. 177. Yocum D, Fleishmann R, Dalgin P, et al. Safety and efficacy of meloxicam in the treatment of osteoarthritis. Arch Intern Med 2000; 160:2947–54. 178. Garcia Gonzalez M, Sanroman AL, Herrero C, et al. Droxicam-induced hepatitis. Description of 3 new cases and review of the literature. Rev Clin Esp 1994; 194(3):170–2. 179. Ollagnon HO, Perpoint B, Decousus H, et al. Hepatitis induced by isoxicam. Hepatogastroenterology 1986; 33:109. 180. Carpenter SL, McDonnell WM. Misuse of veterinary phenylbutazone. Arch Intern Med 1995; 155:1229. 181. Benjamin SB, Ishak KG, Zimmerman HJ, et al. Phenylbutazone liver injury: a clinical pathologic survey of 23 cases and review of the literature. Hepatology 1981; 1:255–63. 182. Kari F, Bucher J, Haseman J, et al. Long-term exposure to the anti-inflammatory agent phenylbutazone induces kidney tumors in rats and liver tumors in mice. Jpn J Cancer Res 1995; 86:252–63. 183. Popper H, Rubin E, Gardio D, et al. Drug-induced liver disease: a penalty for progress. Arch Intern Med 1965; 115:128–36. 184. Gubler HU, Baggiolini M. Pharmacological properties of proquazone. Scand J Rheumatol 1978; 21(Suppl.):8–11. 185. Imoto S, Matzumoto H, Fujii M. Drug-related hepatitis (letter). Ann Intern Med 1979; 91:129. 186. Preston SN. Safety of sodium meclofenamate (meclomen). Curr Ther Res 1978; 23(Suppl.):5107–12.

NSAIDs and Leukotriene Receptor Antagonists

463

187. McCormick PA, Kennedy F, Curry M, et al. COX 2 inhibitor and fulminant hepatic failure. Lancet 1999; 353:40–1. 188. Van Steenbergen W, Peelers P, De Bondt I, et al. Nimesulide-induced acute hepatitis: evidence from six cases. J Hepatol 1998; 29:135–41. 189. Weiss P, Mouallem M, Bruck R, et al. Nimesulide-induced hepatitis and acute liver failure. Isr Med Assoc J 1999; 1:89–91. 190. Villa G. NSAIDs and hepatic reactions (letter). Lancet 1999; 353:846. 191. Rainsford KD. An analysis from clinico-epidemiological data of the principal adverse events from the COX-2 selective NSAID, nimesulide, with particular reference to hepatic injury. Inflammopharmacology 1998; 6:203–21. 192. Andrade RJ, Lucena MI, Fernandez MC, et al. Fatal hepatitis associated with nimesulide. J Hepatol 2000; 32:174. 193. Conforti A, Leone R, Moretti U, et al. Adverse drug reactions related to the use of NSAIDs with a focus on nimesulide: results of spontaneous reporting from a Northern Italian area. Drug Saf 2001; 24:1081–90. 194. Traversa G, Bianchi C, Da Cas R, et al. Cohort study of hepatotoxicity associated with nimesulide and other non-steroidal anti-inflammatory drugs. Br Med J 2003; 327:18–22. 195. Boelsterli UA. Mechanisms of NSAID-induced hepatotoxicity: focus on nimesulide. Drug Saf 2002; 25:633–48. 196. Romero-Gomez M, Nevado Santos M, Otero Fernandez MA, et al. Acute cholestatic hepatitis by nimesulide. Liver 1999; 19:164–5. 197. Papaionnides D, Korantzopoulos P, Athanassiou E, et al. Nimesulide-induced acute hepatotoxicity (letter). Indian J Gastroenterol 2003; 22:239. 198. Dumortier J, Borel I, Delafosse B, et al. Subfulminant hepatitis associated with nimesulide treatment requiring liver transplantation. Gastroenterol Clin Biol 2002; 26:415–6. 199. Maddrey WC, Maurath CJ, Verburg KM, et al. The hepatic safety and tolerability of the novel cyclooxygenase-2 inhibitor celecoxib. Am J Ther 2000; 7:153–8. 200. Carrillo-Jimenez R, Nurnberger M. Celecoxib-induced acute pancreatitis and hepatitis: a case report. Arch Intern Med 2000; 160:553–4. 201. Alegria P, Lebre L, Chagas C. Celecoxib-induced cholestatic hepatotoxicity in a patient with cirrhosis (letter). Ann Intern Med 2002; 137:75. 202. Zinsser P, Meyer-Wyss B, Rich P. Hepatotoxicity induced by celecoxib and amlodipine. Swiss Med Wkly 2004; 134:201. 203. Grieco A, Miele L, Giorgi A, et al. Acute cholestatic hepatitis associated with celecoxib. Ann Pharmacother 2002; 36:1887–9. 204. Nachimuthu S, Volfinzon L, Gopal L. Acute hepatocellular and cholestatic injury in a patient taking celecoxib. Postgrad Med J 2001; 77:548–50. 205. Galan MV, Gordon SC, Silverman AL. Celecoxib-induced cholestatic hepatitis. Ann Intern Med 2001; 134:254. 206. Fung HB, Kirschenbaum HL. Selective cyclooxygenase-2 inhibitors for the treatment of arthritis. Clin Ther 1999; 21:1131–57. 207. Rostom A, Goldkind L, Laine L. Nonsteroidal anti-inflammatory drugs and hepatic toxicity: a systematic review of randomized controlled trials in arthritis patients. Clin Gastroenterol Hepatol 2005; 3:489–98. 208. Chavez ML, DeKorte CJ. Valdecoxib: a review. Clin Ther 2003; 25:817–51. 209. FDA Arthritis Advisory Committee meeting March 5, 2003. Arava (leflunomide) hepatotoxicity. Accessed via www.fda.gov/ohrms/dockets/ac/03/briefing/3930b2.htm 210. Suissa S, Ernst P, Hudson M, et al. Newer disease-modifying antirheumatic drugs and the risk of serious hepatic adverse events in patients with rheumatoid arthritis. Am J Med 2004; 117:87–92. 211. van Roon EN, Jansen TL, Houtman NM, et al. Leflunomide for the treatment of rheumatoid arthritis in clinical practice: incidence and severity of hepatotoxicity. Drug Saf 2004; 27:345–52. 212. Sanders S, Harisdangkul V. Leflunomide for the treatment of rheumatoid arthritis and autoimmunity. Am J Med Sci 2002; 323:190–3. 213. Sevilla-Mantilla C, Ortega L, Agundez JA, et al. Leflunomide-induced acute hepatitis. Dig Liver Dis 2004; 36:82–4. 214. Cannon GW, Kremer JM. Leflunomide. Rheum Dis Clin North Am 2004; 30:295–301. 215. Li EK, Tam LS, Tomlinson B. Leflunomide in the treatment of rheumatoid arthritis. Clin Ther 2004; 26:447–59. 216. Favreau M, Tannenbaum H, Lough J. Hepatic toxicity associated with gold therapy. Ann Intern Med 1977; 87:717–9. 217. Basset C, Vadrot J, Denis J, et al. Prolonged cholestasis and ductopenia following gold salt therapy. Liver Int 2003; 23:89–93. 218. te Boekhorst PA, Barrera P, Laan RF, et al. Hepatotoxicity of parenteral gold therapy in rheumatoid arthritis: a case report and review of the literature. Clin Exp Rheumatol 1999; 17:359–62.

464

Lewis

219. Lipworth BJ. Leukotriene-receptor antagonists. Lancet 1999; 353:57–62. 220. Drazen JM, Israel E, O’Byrne PM. Treatment of asthma with drugs modifying the leukotriene pathway. N Engl J Med 1999; 340:197–206. 221. Joshi EM, Heasley BH, Chordia MD, et al. In vitro metabolism of 2-acetylbenzothiophene: relevance to zileuton hepatotoxicity. Chem Res Toxicol 2004; 17:137–43. 222. Goldstein MF, Black M. Montelukast-induced hepatitis (letter). Ann Intern Med 2004; 140:586–7. 223. Calhoun WJ. Summary of clinical trials with zafirlukast. Am J Respir Crit Care Med 1998; 157:S238–45. 224. Davern TJ, Bass NM. Leukotriene antagonists. Clin Liver Dis 2003; 7:501–12. 225. Danese S, De Vitis I, Gasbarrini A. Severe liver injury associated with zafirlukast (letter). Ann Intern Med 2001; 135:930. 226. Su CW, Wu JC, Huang YS, et al. Zafirlukast-induced acute hepatitis. Zhonghua Yi Xue Za Zhi 2002; 65:553–6. 227. Soy M, Ozer H, Canataroglu A, et al. Vasculitis induced by zafirlukast therapy. Clin Rheumatol 2002; 21:328–9. 228. Actis GC, Morgando A, Lagget M, et al. Zafirlukast-related hepatitis: report of a further case. J Hepatol 2001; 35:539–41. 229. Moles JR, Primo J, Fernandez JM, et al. Acute hepatocellular injury associated with zafirlukast. J Hepatol 2001; 35:541–2. 230. Torres M, Reddy KR. Severe liver injury (letter). Ann Intern Med 2001; 135:550. 231. Al-Salman J, Arjomand H, Kemp DG, et al. Hepatocellular injury in a patient receiving rosiglitazone. Ann Intern Med 2000; 132:121–4. 232. Freid J, Everitt D, Boscia J. Rosiglitazone and hepatic failure(letter). Ann Intern Med 2000; 132:164. 233. Correction. Liver injury and rosiglitazone. Ann Intern Med 2000; 133:237. 234. Favreau JT, Ryu ML, Braunstein G, et al. Severe hepatotoxicity associated with the dietary supplement LipoKinetix. Ann Intern Med 2002; 136:590–5. 235. Accolate (zafirlukast) product information. Physicians’ Desk Reference. 59th ed. Montvale, NJ: Thomson PDR, 2005. 236. Sass DA, Chopra KB, Wu T. A case of montelukast -induced hepatotoxicity. Am J Gastroenterol 2003; 98:704–5. 237. Simon LS. Toxicity of the nonsteroidal anti-inflammatory drugs. Curr Opin Rheumatol 1991; 3:341. 238. Schiff ER. Can we prevent nonsteroidal anti-inflammatory drug-induced hepatic failure? Gastrointest Dis Today 1994; 3:768–70. 239. Lewis JH. The rational use of potentially hepatotoxic medications in patients with underlying liver disease. Expert Opin Drug Saf 2002; 1:159–72.

22

Mechanism, Pathology, and Clinical Presentation of Hepatotoxicity of Anesthetic Agents J. Gerald Kenna

AstraZeneca Safety Assessment, R&D Alderley Park, Macclesfield, Cheshire, U.K.

INTRODUCTION Liver dysfunction may occur in patients exposed to any of the current generation of volatile anesthetic agents. These drugs are halogenated compounds that undergo metabolism in the liver by cytochromes P450 (CYP, particularly CYP2E1), generating chemically reactive intermediates that interact with cellular macromolecules. However, the liver damage that occurs in man following exposure to volatile anesthetics is not attributable to “conventional” cell cytotoxicity initiated within the liver by reactive intermediates (as has been described for many other halogenated compounds, e.g., the anesthetic chloroform) (1,2). Rather, in susceptible patients, covalent binding of reactive metabolites to liver proteins triggers specific adaptive immune responses that are believed to mediate the hepatotoxicity. This process was first described for halothane, and similar mechanisms have been proposed to explain liver injury caused by enflurane, isoflurane, and desflurane. Analogous mechanisms have also been proposed to explain cases of liver damage in humans exposed accidentally to high levels of halogenated hydrocarbon refrigerants, which are related chemically to the volatile anesthetics and undergo CYP-mediated metabolic bioactivation in the liver. The purpose of this chapter is to review the clinical features of liver damage caused by halothane and the recent generation of volatile anesthetics (enflurane, isoflurane, desflurane, and sevoflurane), to evaluate the data obtained to date on underlying biochemical and cellular mechanisms, and to discuss our current understanding of individual susceptibility factors. CLINICAL FEATURES OF VOLATILE ANESTHETIC-INDUCED HEPATOTOXICITY Halothane The clinical features and the pathology of halothane-induced liver injury are well documented and have been reviewed extensively (1,3–5). Typically, patients have no history of preexisting liver disease, do not abuse alcohol, and have no concurrent intake of drugs with known hepatotoxic potential. The incidence of clinically significant liver dysfunction due to halothane is not known precisely, but is low (between 1 in several thousands and 1 in 30,000 patients). For many years, this led to confusion concerning whether halothane was truly hepatotoxic in man, or other undiagnosed causative agents (e.g., acute viral infection) were responsible. Retrospective and prospective clinical studies were inconclusive (6), and the matter was only resolved when diagnostic antibody tests were devised (see below). A substantial proportion of patients who sustain halothane hepatitis are female and of late middle age, and obesity is common. However, many cases of halothane hepatitis in nonobese males have also been described (1,3–5), as have several well-documented cases in young children (7). In addition, cases of liver damage that can be attributed to occupational The author of this chapter is a paid employee of the pharmaceutical industry.

466

Kenna

halothane exposure in medical personnel have been reported (8). Typically, the first symptoms displayed by patients include malaise, anorexia, and nonspecific gastrointestinal symptoms (nausea and upper abdominal discomfort). The majority of patients exhibit delayed-onset pyrexia, while some develop nonspecific rash and/or arthralgia (3,9,10). These are followed by large elevations in serum transaminases and then by jaundice. The time to onset of jaundice is very variable but frequently prolonged, and may be greater than 28 days (11). In some patients, the liver damage is severe and fulminant hepatic failure develops, which may require liver transplantation. However, many patients with halothane hepatitis do not sustain liver failure (11). Nonfatal cases have been reported to resolve progressively and otherwise uneventfully and not to be associated with development of chronic liver disease, provided further exposure to halothane (and to structurally related compounds, as discussed below) is avoided. The pathological features of halothane hepatitis are not diagnostic and are similar to those described in cases of acute hepatitis due to various other causes. Centrilobular hepatic necrosis is a common histological feature and a spectrum of severity ranging from panlobular and multifocal necrosis to massive necrosis has been described, as has ballooning degeneration of hepatocytes, inflammatory infiltration, and fibrosis (1,9,12). Fatty infiltration has been reported in many patients, and in some cases granulomatous aggregates have been observed (12). Mitochondrial membrane abnormalities have been identified by electron microscopy in some instances (9,13), but not in others (14). The vast majority of patients who develop halothane hepatitis have been anesthetized with halothane on multiple occasions, and the severity of liver injury and the interval from final anesthesia to onset of jaundice tend to be inversely related to the interval between anesthetics (3,11). However, halothane hepatitis has been described in cases where the interval between anesthetic exposures has been for many years (11,15). A significant proportion of patients have exhibited adverse reactions (delayed-onset pyrexia or jaundice) following previous halothane exposure (3,11,16,17), therefore taking a careful clinical history is an essential precaution before exposing an individual patient to halothane. Patients with halothane hepatitis exhibit a high incidence of peripheral eosinophilia, circulating autoantibodies to tissue antigens, and also cellular and humoral immune sensitization to reactive metabolite-modified liver neoantigens (3–5,17). These features indicate that the liver injury has an immune component, which is discussed further below. A much milder, and clinically unimportant, form of halothane-induced liver injury in humans has also been described. This was identified from prospective studies, which revealed that up to 30% of patients who underwent surgical anesthesia with halothane developed mild liver injury that resolved asymptomatically (18,19). Serological investigations revealed no evidence of immune activation in these patients, including no detectable antibodies to metabolite-modified liver neoantigens (20,21). It is believed that this mild form of liver injury does not have an immune component. Other Volatile Agents (Enflurane, Isoflurane, Desflurane, and Sevoflurane) The incidence of liver injury caused by these other anesthetics is much lower than the incidence of halothane hepatitis and consequently, the clinical features are less well defined. A retrospective review of 24 cases of otherwise unexplained liver damage in patients anesthetized with enflurane has indicated that this agent can cause severe liver damage in humans that is not simply attributable to contamination of anesthetic circuits with halothane (22). The incidence of enflurane-induced liver injury has been estimated to be of the order of 1:800,000 exposed patients (23). The clinical, biochemical, and pathological features of patients with enflurane hepatitis are similar to those described for halothane hepatitis, including a delayed interval between anesthesia and onset of jaundice, frequently a previous history of exposure to enflurane or halothane, and a significant incidence of mortality due to liver failure (22). In addition, an association between prior exposure of certain patients to halothane and subsequent liver injury following anesthesia with enflurane has been reported (22,24,25). This has been attributed to immune cross-sensitization between the anesthetics (see below). Cases of liver injury that may be attributable to anesthesia with isoflurane have also been described (26–29), although the number reported to date is small when one considers

Hepatotoxicity of Anesthetic Agents

467

the widespread worldwide use of this agent since its introduction in 1984. It is reasonable to conclude that the incidence of hepatitis due to isoflurane is far lower than the incidence of halothane hepatitis. Although causality has been difficult to establish, at least some cases of possible isoflurane hepatitis exhibit clinical and pathological features that are consistent with a “halothane hepatitis-like” mechanism. These are associations with multiple exposures to volatile anesthetics, a preponderance of females, obesity, and severe centrilobular hepatocellular injury (30). Detection of antibodies to metabolite-modified protein neoantigens has been described in some cases (31,32), which is consistent with an immune pathogenesis. A few cases of postoperative liver injury in a patient exposed to desflurane has been described. In one patient, antibodies that recognized metabolite-modified liver neoantigens were detected in the patient’s serum and it was noted that the patient had a prior history of anesthesia with halothane (albeit 12 and 18 years previously) (33). By analogy with halothane hepatitis, the proposed mechanism of liver injury was immunological. Sevoflurane was introduced into anesthetic practice relatively recently in the Western world (in 1995), but prior to this was used widely as a surgical anesthetic for over 10 years in Japan. Cases of liver injury in patients anesthetized with sevoflurane have been described, including several reports of liver injury in children and one case of liver injury in a patient exposed on separate occasions to sevoflurane and desflurane (34–37). Halothane (and other volatile anesthetics) is/are highly soluble in the materials used in anesthetic circuits, and may be absorbed into the circuits and released during subsequent anesthesia with different anesthetic agents (38). Contamination of anesthetic circuits with halothane has been implicated as the cause of liver dysfunction in a patient with a past history of halothane hepatitis, who subsequently developed liver injury following anesthesia using an anesthetic circuit inadvertently contaminated with halothane (38). Such unapparent circuit contamination is an important issue that needs to be taken into account, since for all volatile anesthetics the mechanism of toxicity is believed to require metabolism, and the extents of metabolism differ markedly between the compounds (as discussed below). Whether it explains any of the cases of apparent sevoflurane- or desflurane-induced liver injury is unclear. METABOLISM OF VOLATILE ANESTHETICS Halothane Approximately 20% of an inhaled dose of halothane is metabolized in humans (39). This process is catalyzed by isozymes of hepatic CYP, and involves distinct oxidative and reductive pathways. These have been reviewed elsewhere (40) and are summarized in simplified form in Figure 1. The reductive pathway is favored at reduced oxygen tensions, but occurs only to a limited extent at normal oxygen tensions (i.e., during surgical anesthesia in humans) (41). The initial step is insertion of a single electron to form the trifluorochlorobromoethyl radical, which undergoes a well-defined further series of chemical reactions and biotransformations to other chemically reactive species. These initiate lipid peroxidation, bind covalently to cellular macromolecules (both lipids and proteins), and/or are excreted as urinary and volatile metabolites. Oxidative metabolism is the major pathway of metabolism in humans (41). This proceeds via insertion of oxygen, which is followed by spontaneous debromination to produce a reactive species (trifluoroacetyl chloride) that reacts with water to form the major urinary metabolite trifluoroacetate (41). In addition, a small fraction of trifluoroacetyl chloride binds covalently to 3-amino groups of cellular phospholipids (i.e., phosphatidylethanolamine) and proteins (42–44), thereby generating trifluoroacetylated lipid and protein adducts (Fig. 1). Several different CYP isozymes have been shown to catalyze reductive and oxidative metabolism of halothane (45). Oxidative metabolism is catalyzed preferentially by CYP2E1 and this is the major isozyme responsible for bioactivation of halothane in humans (46,47). Enflurane, Isoflurane, and Desflurane Neither enflurane nor isoflurane undergoes reductive metabolism. However, in common with halothane, both drugs are metabolized by hepatic CYP2E1 [via oxidative dehalogenation (48,49)]

468

Kenna

Halothane Oxidative O2 Br

F Br F C C H F Cl CYP 2E1 CYP

F O F C C Cl F

Reductive e– F Br– F C C H F Cl

Other reactive species

Glutathione conjucation covalent binding to lipids and proteins

F O F C C OH F

F Br N-Ac-cys C C H F Cl

F H C C Cl F

+

H

Br,F– Lipid peroxidation

F H F C C H F Cl

Urinary excretion Pulmonary excretion

FIGURE 1 Metabolism of halothane. Abbreviation: CYP, cytochrome P450.

to reactive intermediates that bind covalently to proteins. The extent of metabolism in humans is 2% to 4% for enflurane (50) and approximately 0.2% for isoflurane (51), and so is markedly lower than the extent of metabolism of halothane (52). Metabolism of enflurane proceeds via a reactive intermediate that either reacts with water to form an acid that is excreted in urine as a glucuronide conjugate (53) or binds covalently to liver proteins to form protein adducts (Fig. 2) (54). CYP2E1mediated oxidation of isoflurane results in renal excretion of inorganic fluoride and trifluoroacetic acid (55) and proceeds via reactive intermediates (believed to include trifluoroacetyl chloride) that also can covalently modify liver proteins and form protein adducts (Fig. 2) (54). Enfluence

Isoflurane F F H H C O C C Cl F F F

F F H H C O C C Cl F F F 02

CYP 2E1

02 F O H C O C F F F F

F O F H C O C C F F F

F F O Covalent binding or H C O C C to macromolecules F OH F UDP-GA

CYP 2E1

or

F O H C C Cl F

F O Covalent binding or H C C OH to macromolecules F

Desflurane F H F H C O C C F F F F 02

CYP 2E1

F O F H C O C C F F F

F O Covalent binding or H C C OH to macromolecules F

UGT

Glucuronide conjugate Urinary excretion

FIGURE 2 Metabolic bioactivation of enflurane, isoflurane, and desflurane. Abbreviation: UGT, UDP-glucuronyltransferase.

Hepatotoxicity of Anesthetic Agents

469

Desflurane is extremely resistant to biodegradation and the extent of metabolism in humans is no more than 10% of that described for isoflurane (i.e., !0.02%) (56). Trifluoroacetic acid has been identified as a urinary metabolite (56), which implies that metabolism of desflurane proceeds via pathways similar to those proposed for isoflurane and involves formation of protein adducts (Fig. 2) (33). However, experimental verification that protein adducts are indeed expressed in the liver following exposure to desflurane has not been described. Sevoflurane Sevoflurane is highly susceptible to CYP-mediated metabolism in vitro, and its rate of metabolism is markedly greater in vitro when compared with the other volatile anesthetics described above (including halothane) (57). However, its metabolism in vivo is limited by low blood and tissue solubility. Consequently, approximately 2% to 5% of an anesthetic dose of sevoflurane is metabolized in humans (58,59). In common with halothane, enflurane, isoflurane, and desflurane, the process is catalyzed in vivo by CYP2E1 (59). The first step is oxidation of the fluoromethyl group (Fig. 3), which results in liberation of inorganic fluoride and carbon dioxide and in formation of hexafluoroiso-propanol. The latter metabolite undergoes glucuronidation and is excreted in urine as the glucuronide conjugate (Fig. 3) (59). The chemistry involved in oxidation of the fluoromethyl group of sevoflurane has not been defined, but may involve generation of formyl fluoride as an intermediate. Since formyl fluoride would be a highly reactive species, this is a possible mechanism by which generation of novel liver protein adducts derived from sevoflurane might occur and could be linked to the rare case reports of liver injury following anesthesia with this anesthetic. However, whether such adducts are actually produced in livers of sevoflurane-exposed animals or humans is unknown. When used during surgical anesthesia, sevoflurane also can undergo chemical degradation followed by complex biotransformation, which may result in formation of reactive intermediates that have the potential to bind covalently to tissue proteins. The processes involved are complex and are outlined schematically in Figure 4. The first step is a chemical reaction between sevoflurane and the strong bases used as carbon dioxide adsorbents in anesthesia circuits, producing fluoromethyl-2,2,-difluoro-1-(trifluoromethyl)vinyl ether (FDVE; also termed “compound A”) (Fig. 4) (60). Subsequently, FDVE undergoes glutathione conjugation (61). Glutathione-S-conjugates derived from FDVE have been detected in bile in Sevoflurane F F F C F H C O C F H F C H F O2 CYP 2E1 CO2 FF F C F HO C H F C F UDP-GA

UGT

Glucuronide conjugate

Urinary excretion

FIGURE 3 Metabolism of sevoflurane. Abbreviations: CYP2E1, cytochrome P450 2E1; UGT, UDP-glucuronyltransferase; UDP-GA, UDP-glucuronic acid.

470

Kenna

Sevoflurance F

F F C F Chemical reduction H C O C H H F C F

H H

C F

F O

F FDVE

C

C F F

F

UGT

F GSH

Glutathione conjugates Enzymic processing

Billiary excretion NAT

Cysteine conjugates β-1 yase

N-Actyl crytein conjugates CYP3A

O F H2 C H C C C OH F TFMP H F C

Sulfoxide conjugates

Urinary excretion

Covalent binding to macromolecules

O H C OH HO C F TFLA F C F

FIGURE 4 Formation and biotransformation of FDVE from sevoflurane. Abbreviations: GST, glutathioneS-transferase; GSH, glutathione; NAT, N-acetyltransferase; b-lyase, cysteine conjugate b-lyase; FDVE, fluoromethyl-2, 2,-difluoro-1-(trifluoromethyl)vinyl ether; TFMP, 3,3,3-trifluro-2-(fluoromethoxy)propanoic acid; TFLA, 3,3,3-trifluorolactic acid.

the rat (61), although within the body they are processed enzymically to cysteine-S- and N-acetyl cysteine-S-conjugates. The cysteine-S-conjugates are further metabolized by cysteine conjugate b-lyase to unstable intermediates that undergo hydrolysis to 3,3,3-trifluro-2-(fluoromethoxy) propanoic acid and 3,3,3-trifluorolactic acid, which are excreted in urine (62,63). By analogy with other halogenated compounds, the unstable intermediates are believed to be reactive species that may also bind covalently to proteins. The N-acetyl cysteine-S-conjugates are excreted in urine (62,63), although recently their further metabolism (by CYP3A isoforms) to sulfoxide metabolites that also are present in urine has been reported (64,65). In view of the chemical reactivity of sulfoxide metabolites of other halogenated compounds, the sulfoxide metabolites derived from FDVE could themselves be reactive intermediates (64,65). ANIMAL MODELS OF ANESTHETIC-INDUCED LIVER INJURY Halothane Reproducible and dose-dependent hepatotoxicity has been produced by pretreatment of rats with polychlorinated biphenyls or phenobarbitone to induce CYPs, then exposure of the animals to halothane at reduced oxygen tensions to promote reductive metabolism (66–68). The liver injury produced in these models has been attributed to a combination of direct toxic effects of reductive reactive metabolites (which promote lipid peroxidation and bind covalently to hepatocellular lipids and proteins) plus ischemic injury caused by tissue hypoxia (69,70). Liver toxicity that has been attributed to hypoxia alone has been observed in rats exposed to halothane at normal oxygen tensions, following pretreatment with triiodothyronine (71). In addition, two animal models have been described that involve hepatotoxicity

Hepatotoxicity of Anesthetic Agents

471

mediated via oxidative metabolites of halothane. In the first of these models, Fischer 344 rats were exposed to halothane at normal oxygen tensions following pretreatment with isoniazid, which is a potent CYP2E1 inducer that increased oxidative metabolism of the anesthetic (72). In the second model, hepatotoxicity was produced in guinea pigs by exposure to halothane at normal oxygen tensions (73,74). Marked strain- and sex-dependent differences in susceptibility to halothane-induced liver injury were observed in the guinea pig model. These were not attributable to differences in oxidative metabolism of halothane, but could be a consequence of variable effects of halothane on hepatic blood flow (75) and/or due to variability in hepatic thiol status (which will affect capacity to detoxify electrophilic-reactive intermediates, as illustrated in Fig. 1) (76). Although the various animal model studies have demonstrated that both oxidative and reductive pathways of metabolism of halothane may mediate hepatotoxicity, the mechanisms underlying tissue injury remain quite poorly defined. Lipid peroxidation arising from increased free radical formation and oxidative stress has been observed in reductive rat models (77), as has induction of the stress response proteins heme oxygenase-1 (HO-1) and heat shock protein-70, and an increase in the (pro-oxidant) free heme pool in the liver (78). Moreover, pretreatment of rats with hemin prior to exposure to halothane (in the phenobarbital induction–hypoxia model) increased the expression of HO-1 and abrogated liver toxicity (78), indicating that this is a key protective response. Elevated hepatic calcium content has been observed in rat and guinea pig models and has been implicated in the mechanism of liver injury (79,80). In view of this, it is interesting that one of the major trifluoroacetylated proteins generated within the liver is the calcium-binding protein, calreticulin (see below). In the guinea pig model, impaired bile flow has also been described which is associated with increased permeability of tight junctions (81), increased serum levels of macrophage inhibitory factor (a pro-inflammatory signal) have been reported (82), and a correlation between levels of trifluoroacetylated liver protein adducts in individual animals and extent of liver injury has been observed (83). It is notable that none of the animal models require multiple exposures to halothane, all are reproducible and dose-dependent, and none are characterized by the selective immune stimulation that is characteristic of halothane hepatitis in humans. Consequently, it is considered that these models are not directly relevant to halothane hepatitis in humans. One or more of the models could explain why up to 30% of halothane-exposed patients sustain very mild and transient liver injury, however (40,70). Other Volatile Agents Hepatocellular liver injury has been produced by pretreatment of rats with triiodothyronine, then anesthesia with enflurane, isoflurane, or halothane (71). In addition, exposure of phenobarbitone-pretreated rats to halothane, enflurane, or isoflurane under markedly hypoxic conditions has been reported to result in liver damage (84–86). The toxicity observed in these studies has been attributed to ischemic liver damage caused by tissue hypoxia and not to specific toxic effects of the anesthetics themselves. This is because similar extents of liver injury were observed when rats that had been pretreated with phenobarbital were anesthetized intravenously with thiopental or fentanyl under hypoxic conditions, in place of the volatile agent (86). When rats were pretreated with phenobarbital and then anesthetized with halothane, enflurane, or isoflurane under mildly hypoxic conditions, liver injury was observed only in the group of rats that had received halothane (85). It was concluded that enflurane and isoflurane have minimal intrinsic hepatotoxic potential when compared with halothane. Hepatotoxicity has not been observed in the animal investigations undertaken with desflurane or sevoflurane. However, renal toxicity has been observed in rats treated with high doses of the sevoflurane degradation product FDVE, and has been attributed to metabolism of the compound to reactive species via the cysteine conjugate b-lyase and CYP3A pathways described previously and illustrated in Figure 4 (64,87–89). Renal toxicity has not been observed in humans following during surgical anesthesia with sevoflurane (90).

472

Kenna

THE IMMUNE HYPOTHESIS OF ANESTHETIC-INDUCED LIVER INJURY Halothane Specific immune responses to halothane metabolite-modified liver neoantigens in patients with halothane hepatitis were described initially by Vergani and co-workers. Cellular sensitization to the antigens in patients with halothane hepatitis, but not control individuals, was observed using the technique of in vitro leukocyte migration (91). Subsequently, antibody-dependent cytotoxic killing was demonstrated following incubation of hepatocytes from halothaneexposed rabbits with sera from patients with halothane hepatitis (which had been adsorbed to deplete antibodies to normal liver antigens) plus normal human leukocytes (92). The hepatocyte toxicity was not observed when experiments were undertaken using control hepatocytes, or hepatocytes from rabbits that had been anesthetized with ether, or when hepatocytes from halothane-treated rabbits were incubated with sera from various groups of control individuals (including anesthetists, patients exposed to halothane who did not sustain liver injury, patients with various other types of liver disease, and normal controls) (92,93). These findings indicated that antibodies in sera from patients with halothane hepatitis mediated the cytotoxicity, that the antibody response to the antigens was specific to the patients, and that the target antigens were halothane-modified molecules expressed on the surface of hepatocytes (92). Subsequent investigations, which have used a variety of other methods to detect the antibodies [including enzyme-linked immunosorbent assay (ELISA) and immunoblotting], have confirmed these findings and have shown that antibodies to halothane-induced neoantigens are predominantly of IgG class (94,95). The specificity of the immune response indicates that it is not simply a secondary consequence of halothane exposure and/or liver damage, while the finding of IgG-class antibodies implies that patients’ immune responses have been sensitized to halothane-induced antigens following previous halothane exposures. This is consistent with the known association of halothane hepatitis with multiple exposures to the anesthetic. The nature of the halothane-induced neoantigens has been explored using denaturing [SDS-polyacrylamide gel electrophoresis (SDS-PAGE), then immunoblotting] and nondenaturing (ELISA and immunoprecipitation) approaches. The immunoblotting studies identified a range of different liver microsomal polypeptide neoantigens that were expressed in halothaneexposed rabbits, humans, and rats, but not in livers from control animals or humans (95–97). The neoantigens were shown to be generated via oxidative CYP-mediated metabolism of halothane (97,98), which proceeds via a highly reactive intermediate (trifluoroacetyl chloride) that binds covalently to liver macromolecules (Fig. 1). The neoantigens recognized by antibodies from patients with halothane hepatitis were found to contain covalently bound trifluoroacetyl groups (linked to 3-amino groups on lysine residues of the carrier proteins) and to comprise the major trifluoroacetylated proteins expressed in livers of halothane-exposed animals (97–100). Antibodies from patients with halothane hepatitis were found to recognize epitopes that include the trifluoroacetyl group and also individual structural features (presumably amino acid residues adjacent to the covalently modified lysine residues) that are unique to each of the individual target proteins (97). ELISA and immunoprecipitation studies revealed the existence of additional halothaneinduced neoantigens that were recognized by antibodies from patients with halothane hepatitis, but could not be detected by SDS-PAGE/immunoblotting (101,102). Generation of these neoantigens was also found to require oxidative metabolism of halothane, and the immunoprecipitation studies established that the target proteins are trifluoroacetylated integral membrane proteins (101,102). In contrast, the neoantigens detected by immunoblotting predominantly comprise peripheral membrane proteins (103). Many of the trifluoroacetylated liver target proteins have been characterized by amino acid sequence analysis and/or cDNA cloning (Table 1) (102,104–114). The majority are highly abundant and relatively long-lived peripheral membrane proteins that reside in the lumen of the endoplasmic reticulum (103). The trifluoroacetylated forms of the proteins accumulate progressively in the liver over a period of about 24 hours and they persist for many days once

473

Hepatotoxicity of Anesthetic Agents

TABLE 1 Trifluoroacetylated Liver Neoantigens Derived from Halothane Molecular mass (kDa) determined by SDS-PAGE 29 50 52 57 58 59 63 80 82 100 170

Identity of protein Cytosolic glutathione-Stransferase Microsomal epoxide hydrolase CYP2E1 Protein disulfide isomerase Unknown function Microsomal carboxylesterase Calreticulin Erp72 BiP/GRP78 Endoplasmin/Erp99 /GRP94 UDP-glucose:glycoprotein glucosyltransferase

Method initially used to detect neoantigen

Recognition by antibodies from patients with halothane hepatitis

Protein purification

Not demonstrated

104

Immunoprecipitation Immunoprecipitation Immunoblotting Immunoblotting Immunoblotting Immunoblotting Immunoblotting Immunoblotting Immunoblotting Immunoprecipitation

Yes Yes ( autoantibodies) Yes Yes Yes Yes Yes Yes Yes Not demonstrated

102 105,106 107 108 109 110 111 112 113 114

Reference

Abbreviation: CYP2E1, cytochrome P450 2E1.

formed (98). Presumably, they become covalently modified in a relatively nonselective manner. Since the active site of CYP2E1 (the enzyme responsible for neoantigen formation) is located on the cytosolic face of the endoplasmic reticulum, yet the target proteins are lumenal, it has been proposed that their generation requires diffusion of trifluoroacetyl chloride across the lipid bilayer (103), as illustrated in Figure 5. The major conformation-dependent neoantigen detectable by immunoprecipitation comprises a trifluoroacetylated form of microsomal epoxide hydrolase (102), which is a highly abundant integral membrane protein. In addition, a trifluoroacetylated form of CYP2E1 has been detected in livers of halothane-treated rats by immunoprecipitation (105), and a high incidence of autoantibodies to a nontrifluoroacetylated form of CYP2E1 has been demonstrated in sera from patients with halothane hepatitis (105,106). The epitopes recognized by the patients’ autoantibodies are conformational, and have been localized to specific regions of the surface of the protein (115). It has been proposed that these CYP2E1 autoantibodies are generated because trifluoroacetylation of CYP2E1 leads to a selective breakdown in immune tolerance (105,106). Similar processes may explain the presence of autoantibodies to unmodified forms of other trifluoroacetylated neoantigens in the patients’ sera (107,108,111–113,116–118). Sera from patients with halothane hepatitis also contain autoantibodies that recognize various liver microsomal proteins that apparently are not trifluoroacetylated (119). The presence of these autoantibodies suggests that the immune response against trifluoroacetylated epitopes leads to a broad loss of immune tolerance to liver proteins in patients with halothane hepatitis. Whether autoantibodies to liver proteins contribute to the process of immune-mediated liver injury is unclear. The presence in serum of high levels of the autoantibodies is characteristic of patients with halothane hepatitis (105–108,111,117–119), and in vitro cytotoxicity studies have suggested that they could mediate hepatocyte injury (120). However, testing of sera from a large cohort of pediatric anesthesiologists by ELISA has demonstrated elevated levels of autoantibodies to CYP2E1 and Erp58 in some anesthetic personnel with normal liver function, especially female pediatric anesthetists (121). This implies that the autoantibody response (at least to these two proteins) is not pathogenic in patients per se, although it could contribute to exacerbation of immune-mediated liver injury targeted against trifluoroacetylated liver protein epitopes. With the exception of CYP2E1 and cytosolic glutathione-S-transferase (GST), the normal cellular functions of the proteins are unrelated to metabolism of halothane. However, most of the target proteins are key components of the quality control system responsible for “proofreading” newly synthesized proteins within the endoplasmic reticulum (122) [i.e., calreticulin, protein disulfide isomerase, UDP-glucose:glycoprotein glucosyltransferase (UGGT), and the stress inducible chaperones BiP and GRP94; Table 1]. A large fraction of the proteins

474

Kenna

Trifluoroacetylated target proteins in ER ER lumen ER membrane

CYP 2E1

mEH

Cystosol

Halothane

CF3COCl

Topological inversion? Membrane flow via golgi apparatus

Cytosol Plasma membrane CYP 2E1 Hepatocyte surface Trifluoraccetylated neoantigens on hepatocyde surface

FIGURE 5 Proposed mechanism of generation of trifluoroacetylated protein neoantigens derived from halothane, and of cell-surface expression of trifluoroacetylated CYP2E1. Abbreviation: ER, endoplasmic reticulum.

synthesized within the endoplasmic reticulum fail to fold and mature properly. The overall function of the quality control system is to ensure that only native conformers reach their final destinations within the cell (122). Hence the system plays a key role in ensuring fidelity of cellular functions and defective operation of it can be expected to result in toxicity. Whether the quality control system is impaired in vivo in the liver following exposure to halothane has not been explored, although reduced activity of one of the target proteins (UGGT) has been demonstrated in the rat (114). It is possible that reduced activity of the quality control system in the liver could contribute to liver injury caused by halothane in experimental animals, and perhaps might also play a role in initiation of halothane hepatitis in man. Expression of a small, but significant, fraction of total trifluoroacetylated CYP2E1 on the hepatocyte plasma membrane has been demonstrated (105). This is likely to occur via membrane flow through the Golgi apparatus, as has been described for other CYP isozymes (123), and could involve inversion of the normal topology of the protein in the endoplasmic reticulum (Fig. 5) (124). Cell surface expression of other trifluoroacetylated neoantigens [particularly microsomal epoxide hydrolase (125)] is also likely, but has not yet been demonstrated directly. Overall, the available data indicate that exposure of susceptible individuals to halothane “primes” immune effector mechanisms directed against trifluoroacetylated liver protein

475

Hepatotoxicity of Anesthetic Agents

Hepatocyte

Halothane Enflurane Isoflurane Desflurane

Neoantigen-modified hepatocyte

Bioactivation by CYP 2E1 Neoantigen-driven Immune response in susceptible individuals

Immunemediated toxicity

Activated immune effector mechanisms Hepatic necrosis

FIGURE 6 Proposed mechanism of immune-mediated hepatotoxicity of volatile anesthetics.

neoantigens, which mediate liver injury when the patients are rechallenged with halothane. This is illustrated in Figure 6. The mechanisms of immune-mediated liver injury remain undefined, and could involve T-cell-mediated and/or antibody-dependent processes. Enflurane, Isoflurane, and Desflurane Analysis of livers from rats treated with enflurane or isoflurane by immunoblotting has demonstrated the expression of novel microsomal protein neoantigens similar to those derived from halothane (Fig. 2) (54,126). These neoantigens were expressed at markedly lower levels than the trifluoroacetylated neoantigens derived from halothane and the rank order was halothaneOenfluraneOisoflurane (54,126,127). This is consistent with the relative extents of CYP-mediated metabolism of the anesthetics (52). The neoantigens derived from enflurane and isoflurane were shown to be recognized by an anti-(trifluoroacetyl-protein) rabbit antiserum that had been raised by immunizing rabbits with trifluoroacetylated rabbit albumin (54,127). The neoantigens derived from enflurane were also recognized by antibodies from patients with halothane hepatitis (126). In addition, antibodies to trifluoroacetylated protein neoantigens have been detected in sera from two patients with hepatitis presumed to be attributable to isoflurane (31,32). It was concluded that both enflurane and isoflurane have the potential to elicit liver injury in humans via immune processes similar to those implicated in halothane hepatitis (Fig. 5) and that this provides a mechanistic basis for clinical cases of apparent cross-sensitization between the anesthetics (54,126). It is important to note that many patients who have developed halothane hepatitis have been anesthetized uneventfully with isoflurane and/or enflurane (indicating that cross-sensitization had not occurred). Presumably, the relatively low levels of expression of liver neoantigens derived from the latter anesthetics are insufficient to trigger immune responsiveness in the majority of cases, regardless of whether or not sensitization to halothane (and development of halothane hepatitis) has occurred. Antibodies that recognize trifluoroacetylated liver neoantigens have also been detected in serum from a patient with presumed desflurane hepatitis (33). The proposed mechanism of liver injury was immunological, and similar to that proposed for halothane hepatitis (Fig. 6). In view of the chemical structure and predicted pathway of CYP-mediated metabolism of desflurane (Fig. 2), it seems likely that trifluoroacetylated neoantigens are generated in livers of animals and humans exposed to this anesthetic (33). However, this has not been demonstrated

476

Kenna

experimentally and the levels of the expression of neoantigens derived from desflurane should be extremely low, because of the very limited extent of metabolism of the compound (127). Consequently, the risk of immune-mediated liver injury in patients exposed to desflurane is likely to be extremely small. Sevoflurane It is unclear whether sevoflurane has the potential to elicit immune-mediated liver injury in humans. Cases of possible “sevoflurane hepatitis” have been reported in the Japanese literature, and in some instances data obtained using in vitro lymphocyte transformation tests have suggested that the patients had become sensitized to sevoflurane (34). However, the diagnostic value of the in vitro lymphocyte transformation studies is unclear. This is because lymphocytes have minimal CYP-dependent metabolic capability and the analyses were undertaken in the absence of liver proteins. As discussed previously, it is possible that liver protein neoantigens might be formed during metabolism of sevoflurane (via release of formyl fluoride), or as a consequence of metabolism of the degradation product FDVE (see above and Fig. 4). However, there is no experimental proof that these putative neoantigens are actually produced, in vitro or in vivo. Hydrochlorofluorocarbon Refrigerants The ozone-depleting chlorofluorocarbons are being replaced as industrial chemicals by hydrochlorofluorocarbons (HCFCs), which have little ozone-depleting potential. Some of the HCFCs are similar in structure to halothane and the other volatile anesthetic agents. It has been shown that several of these compounds are metabolized in the liver by CYP2E1 to reactive species that covalently bind to proteins, thereby generating protein adducts that are immunochemically cross-reactive with the trifluoroacetylated protein adducts derived from halothane (99,128). Moreover, hepatotoxicity has been reported in guinea pigs exposed to HCFC-123 (1,1-dichloro-2,2,2-trifluoroethane) (129) and in nine humans who received repeated accidental occupational exposure to very high concentrations of a mixture of HCFC-123 and HCFC-124 (1-chloro-2,2,2,2-tetrafluoroethane) (130). Trifluoroacetylated protein adducts were detected in liver from one of the occupationally exposed human cases, while autoantibodies to two of the target proteins implicated in the mechanism of halothane hepatitis (CYP2E1 and protein disulfide isomerase) were detected in five of the cases (130). It was concluded that humans exposed on multiple occasions to very high concentrations of certain HCFCs (most notably HCFC-123) are at risk of liver injury. Whether HCFC-induced liver injury in humans is immune mediated, or can be attributed to direct cytotoxicity, is not known. However, it is reasonable to presume that patients who have become sensitized to volatile anesthetics may become cross-sensitized to certain HCFCs. Immune Responses in Experimental Animals Various attempts have made to produce immune-mediated liver injury in preclinical species exposed to halothane. Generation of adduct-specific antibody responses that recognized the trifluoroacetyl hapten group (detected using serum albumin as carrier protein) have been demonstrated in rabbits and in guinea pigs following multiple exposures to halothane (131–133). In addition, cellular sensitization to trifluoroacetylated protein epitopes (on serum albumin) was demonstrated in the guinea pig (134). The antibody responses were of low titer when compared with that in patients with halothane hepatitis. It is unclear whether the antibody responses produced in the animals are specifically targeted against individual trifluoroacetylated liver proteins (as has been observed in patients with halothane hepatitis, and discussed above), because of the limitations of the antibody test system used to evaluate the animals. In rabbits, the halothane-induced antibody response was not accompanied by liver dysfunction (131,132). Small and transient elevations in plasma transaminase levels were observed in guinea pigs that paralleled peaks in antibody titers (133). These results indicate that the trifluoroacetylated proteins expressed in the liver are immunogenic in animals, but that

Hepatotoxicity of Anesthetic Agents

477

only a transient and low-level immune response occurs which is insufficient to elicit significant tissue injury. This would be consistent with activation of regulatory processes that prevent further amplification of the immune system. Attempts to overcome the putative immunoregulatory processes before exposure of animals to halothane have been undertaken, but to date have not resulted in liver damage (135–137). The approaches used were preimmunization of animals with exogenous trifluoroacetylated antigens (trifluoroacetylated serum albumin and other proteins, or trifluoroacetylated hepatocytes from animals exposed to halothane) (135,136), and treatment of animals with a reactive compound that trifluoroacetylated endogenous proteins (137). Low titers of antibodies that recognized the trifluoroacetyl hapten (expressed on rabbit serum albumin) have also been observed in guinea pigs exposed to the sevoflurane degradation product FDVE, although once more no liver injury was observed (138). In view of this finding, it was suggested that FDVE has the potential to trigger specific immune responses similar to those initiated by halothane, etc., and that this could be implicated in the liver injury described very rarely in patients anesthetized with sevoflurane. However, it is not clear whether the antibody response observed in guinea pigs exposed to FDVE was linked to generation of protein adduct neoantigens in the liver, or elsewhere in the animals. A likely site of adduct formation would be the kidney, since FDVE causes renal toxicity which is believed to occur due to metabolic bioactivation (Fig. 4) (65,87–89). Clearly expression of protein adducts in the kidney, but not the liver, would be inconsistent with liver injury (not renal injury) in sevoflurane-exposed individuals.

INDIVIDUAL SUSCEPTIBILITY FACTORS Why a very small fraction of individuals exposed to volatile anesthetics sustain hepatic injury, while the vast majority do not, has yet to be established. In view of the proposed mechanism of immune-mediated liver injury (Fig. 6), both metabolic and immune factors are likely to be involved. Metabolic Factors A key metabolic susceptibility factor may be the balance between CYP2E1-catalyzed metabolic bioactivation of the volatile anesthetics and detoxification of reactive intermediates by glutathione and other cellular nucleophiles (47). This could vary throughout the human population (due to individual differences in CYP2E1 activity, GST activity, and/or levels of hepatic glutathione), and thereby result in interindividual variability in levels of expression of metabolite-modified liver neoantigens (47,98). Studies undertaken with various haptenated autologous proteins have shown that the density of hapten groups and the concentrations of haptenated proteins play major roles in breaking immunological tolerance to self-proteins (139). Consequently, patients who express relatively high levels of neoantigens derived from volatile anesthetics are likely to be at greater risk of developing a neoantigen-driven immune response, and sustaining immune-mediated liver injury, than patients who express lower levels of the neoantigens. This would explain why the incidence of liver injury in patients exposed to enflurane or isoflurane (which are metabolized to a markedly lesser extent than halothane) is markedly lower than the incidence of halothane hepatitis. It may also explain why obesity [which can enhance metabolism by inducing CYP2E1 (140) and/or by enhancing distribution of volatile anesthetics into body fat (141)] has been identified as a risk factor in halothane hepatitis (17). In addition, existence of a metabolic susceptibility factor is supported by a report of abnormal sensitivity to electrophilic-reactive metabolites derived from phenytoin in lymphocytes from 11 patients with halothane hepatitis, when compared with control lymphocytes (142). This abnormality, which implies a defective ability to detoxify the reactive intermediates, was exhibited also by lymphocytes from 19 family members of four of the patients, indicating that it is genetically inherited (142). An inherited susceptibility factor is consistent with reports of cases of halothane hepatitis in pairs of closely related women (143).

478

Kenna

Immunological Factors A consistent association between susceptibility to halothane hepatitis and human leukocyte antigen phenotype has not been observed (144,145). Experiments undertaken by Gut and co-workers have shown that trifluoroacetylated epitopes on proteins are very similar (both structurally and immunochemically) to the lipoyl-lysine regions contained within the E2 subunits of mitochondrial pyruvate dehydrogenase and other 2-oxoacid dehydrogenase proteins (146). The molecular mimicry was found to extend to epitopes recognized by antibodies from patients with halothane hepatitis (147,148). This structural similarity could result in immunological tolerance to trifluoroacetylated protein epitopes and may help to explain why “normal individuals” do not develop neoantigen-induced immune responses and do not sustain immune-mediated liver damage when exposed to halothane and other volatile anesthetics, and such immunological tolerance might also be part of the immunoregulatory network that has confounded attempts to produce immune-mediated liver injury in experimental animals. Marked interindividual variability in levels of expression of the E2 subunit of pyruvate dehydrogenase was observed in a panel of 19 human liver samples, and abnormally low levels of expression of the protein were observed in liver biopsy samples from five out of seven patients with halothane hepatitis (149). In view of this, it has been proposed that susceptibility to halothane hepatitis (and to liver injury caused by other volatile anesthetics) may arise, at least in part, because certain individuals express unusually low levels of lipoylated E2 subunits of PDH and related proteins and so have defective immune tolerance to trifluoroacetylated protein epitopes (146,149). The “danger hypothesis” has been proposed by Matzinger, in an attempt to provide an explanation for the loss of immune tolerance that occurs in autoimmune diseases (150). According to this hypothesis, an immune response is triggered when an antigen is presented to the immune system in the context of a “danger signal,” such as cell damage, but not if a danger signal is absent. Other investigators have suggested that this concept could help to explain many immune-mediated adverse drug reactions, including liver damage due to anesthetic agents (151,152). This is an intriguing suggestion that merits further exploration, especially since several of the trifluoroacetylated neoantigens identified in livers of halothanetreated rats are stress proteins (Table 1). Another important factor will be the intrahepatic immune environment, which overall is associated with induction of tolerance and not immunogenicity (153–155). Local presentation of antigens to the immune system within the liver is generally associated with inactivation and apoptosis of T cells, and induction of tolerance (153–155). In contrast, antigen presentation within lymphoid tissues is linked to immune activation. In view of this, it is perhaps not surprising that neoantigen-driven immune responses are not observed in experimental animals or in normal humans following exposure to volatile anesthetics. A key question that needs to be addressed is why immune tolerance does not occur in certain susceptible patients, who mount immune responses to the neoantigens and develop liver injury. SUMMARY Liver damage that can be attributed to volatile anesthetic agents is rare, and may occur following exposure to all of the volatile agents in use currently. The incidence of liver injury in man caused by the individual anesthetic agents can be correlated with their extents of metabolic bioactivation. For the agents favored currently (especially, desflurane and sevoflurane, but also isoflurane), metabolic bioactivation occurs to a minimal extent and liver injury in man is an extremely rare event. This may be contrasted with the rather higher (albeit still very infrequent) incidence of halothane-induced liver damage, and is a major reason why halothane has been largely superceded by the newer agents in modern anesthetic practice. Liver damage due to halothane, enflurane, isoflurane, and desflurane is a consequence of their CYP2E1-mediated bioactivation to reactive intermediates that bind covalently to liver proteins to form neoantigens. The neoantigens elicit immune responses in susceptible patients and these immune responses have been implicated in the mechanism of anesthetic-induced liver damage.

Hepatotoxicity of Anesthetic Agents

479

A variety of antibody assays have been described that aid diagnosis. The available evidence indicates that a complex combination of risk factors, which include metabolic and immunological components, explains why certain anesthetic-exposed individuals sustain liver injury, whereas the vast majority of the human population does not. REFERENCES 1. Zimmerman HJ. Anesthetic Agents. Hepatotoxicity: The Adverse Effects of Drugs and Other Chemicals on the Liver. New York: Appleton-Century-Crofts, 1978:370-94. 2. Pohl LR. Biochemical toxicology of chloroform. Rev Biochem Toxicol 1979; 1:79–107. 3. Inman WH, Mushin WW. Jaundice after repeated exposure to halothane: an analysis of Reports to the Committee on Safety of Medicines. Br Med J 1974; 1:5–10. 4. Ray DC, Drummond GB. Halothane hepatitis. Br J Anaesth 1991; 67:84–99. 5. Kenna JG, Neuberger J. Immunopathogenesis and treatment of halothane hepatitis. Clin Immunother 1995; 3:108–24. 6. Bunker JP. Final report of the National Halothane Study. Anesthesiology 1968; 29:231–2. 7. Kenna JG, Neuberger J, Mieli-Vergani G, Mowat AP, Williams R. Halothane hepatitis in children. Br Med J (Clin Res Ed) 1987; 294:1209–11. 8. Neuberger J, Vergani D, Mieli-Vergani G, Davis M, Williams R. Hepatic damage after exposure to halothane in medical personnel. Br J Anaesth 1981; 53:1173–7. 9. Klion FM, Schaffner F, Popper H. Hepatitis after exposure to halothane. Ann Intern Med 1969; 71:467–77. 10. Moult PJ, Sherlock S. Halothane-related hepatitis: a clinical study of twenty-six cases. Q J Med 1975; 44:99–114. 11. Kenna JG, Neuberger J, Williams R. Specific antibodies to halothane-induced liver antigens in halothane-associated hepatitis. Br J Anaesth 1987; 59:1286–90. 12. Benjamin SB, Goodman ZD, Ishak KG, Zimmerman HJ, Irey NS. The morphologic spectrum of halothane-induced hepatic injury: analysis of 77 cases. Hepatology 1985; 5:1163–71. 13. Uzunalimoglu B, Yardley JH, Boitnott JK. The liver in mild halothane hepatitis. Light and electron microscopic findings with special reference to the mononuclear cell infiltrate. Am J Pathol 1970; 61:457–78. 14. Wills EJ, Walton B. A morphologic study of unexplained hepatitis following halothane anesthesia. Am J Pathol 1978; 91:11–32. 15. Martin JL, Dubbink DA, Plevak DJ, et al. Halothane hepatitis 28 years after primary exposure. Anesth Analg 1992; 74:605–8. 16. Inman WH, Mushin WW. Jaundice after repeated exposure to halothane: a further analysis of reports to the Committee on Safety of Medicines. Br Med J 1978; 2:1455–6. 17. Walton B, Simpson BR, Strunin L, Doniach D, Perrin J, Appleyard AJ. Unexplained hepatitis following halothane. Br Med J 1976; 1:1171–6. 18. Wright R, Eade OE, Chisholm M, et al. Controlled prospective study of the effect on liver function of multiple exposures to halothane. Lancet 1975; 1:817–20. 19. Trowell J, Peto R, Smith AC. Controlled trial of repeated halothane anaesthetics in patients with carcinoma of the uterine cervix treated with radium. Lancet 1975; 1:821–4. 20. Davis M, Eddleston AL, Neuberger JM, Vergani D, Mieli-Vergani G, Williams R. Halothane hepatitis. N Engl J Med 1980; 303:1123–4 (letter). 21. Sakaguchi Y, Inaba S, Irita K, Sakai H, Nawata H, Takahashi S. Absence of antitrifluoro-acetate antibody after halothane anaesthesia in patients exhibiting no or mild liver damage. Can J Anaesth 1994; 41:398–403. 22. Lewis JH, Zimmerman HJ, Ishak KG, Mullick FG. Enflurane hepatotoxicity: a clinicopathologic study of 24 cases. Ann Intern Med 1983; 98:984–92. 23. Brown BR, Jr., Gandolfi AJ. Adverse effects of volatile anaesthetics. Br J Anaesth 1987; 59:14–23. 24. Sigurdsson J, Hreidarsson AB, Thjodleifsson B. Enflurane hepatitis: a report of a case with a previous history of halothane hepatitis. Acta Anaesthesiol Scand 1985; 29:495–6. 25. Gogus FY, Toker K, Baykan N. Hepatitis following use of two different fluorinated anesthetic agents. Isr J Med Sci 1991; 27:156–9. 26. Brunt EM, White H, Marsh JW, Holtmann B, Peters MG. Fulminant hepatic failure after repeated exposure to isoflurane anesthesia: a case report. Hepatology 1991; 13:1017–21. 27. Sinha A, Clatch RJ, Stuck G, Blumenthal SA, Patel SA. Isoflurane hepatotoxicity: a case report and review of the literature. Am J Gastroenterol 1996; 91:2406–9. 28. Weitz J, Kienle P, Bohrer H, Hofmann W, Theilmann L, Otto G. Fatal hepatic necrosis after isoflurane anaesthesia. Anaesthesia 1997; 52:892–5. 29. Turner GB, O’Rourke D, Scott GO, Beringer TR. Fatal hepatotoxicity after re-exposure to isoflurane: a case report and review of the literature. Eur J Gastroenterol Hepatol 2000; 12:955–9.

480

Kenna

30. Zimmerman H. Even isoflurane. Hepatology 1991; 13:1251–3 (editorial). 31. Gunaratnam NT, Benson J, Gandolfi AJ, Chen M. Suspected isoflurane hepatitis in an obese patient with a history of halothane hepatitis. Anesthesiology 1995; 83:1361–4. 32. Meldrum DJ, Griffiths R, Kenna JG. Gallstones and isoflurane hepatitis. Anaesthesia 1998; 53:905–9. 33. Martin JL, Plevak DJ, Flannery KD, et al. Hepatotoxicity after desflurane anesthesia. Anesthesiology 1995; 83:1125–9. 34. Ogawa M, Doi K, Mitsufuji T, Satoh K, Takatori T. Drug induced hepatitis following sevoflurane anesthesia in a child. Masui 1991; 40:1542–5. 35. Shichinohe Y, Masuda Y, Takahashi H, et al. A case of postoperative hepatic injury after sevoflurane anesthesia. Masui 1992; 41:1802–5. 36. Jang Y, Kim I. Severe hepatotoxicity after sevoflurane anesthesia in a child with mild renal dysfunction. Pediatr Anesth 2005; 15:1140–4. 37. Chung PC, Chiou SC, Lien JM, Li AH, Wong CH. Reproducible hepatic dysfunction following separate anesthesia with sevoflurane and desflurane. Chang Gung Med J 2003; 26:357–62. 38. Varma RR, Whitesell RC, Iskandarani MM. Halothane hepatitis without halothane: role of inapparent circuit contamination and it prevention. Hepatology 1985; 5:1159–62. 39. Rehder K, Forbes J, Alter H, Hessler O, Stier A. Halothane biotransformation in man: a quantitative study. Anesthesiology 1967; 28:711–5. 40. Kenna JG, van Pelt FNAM. The metabolism and toxicity of inhaled anaesthetic agents. Anaesth Pharmacol Rev 1994; 2:29–42. 41. Cohen EN. Metabolism of the volatile anesthetics. Anesthesiology 1971; 35:193–202. 42. Gandolfi AJ, White RD, Sipes IG, Pohl LR. Bioactivation and covalent binding of halothane in vitro: studies with [3H]- and [14C]halothane. J Pharmacol Exp Ther 1980; 214:721–5. 43. Satoh H, Fukuda Y, Anderson DK, Ferrans VJ, Gillette JR, Pohl LR. Immunological studies on the mechanism of halothane-induced hepatotoxicity: immunohistochemical evidence of trifluoroacetylated hepatocytes. J Pharmacol Exp Ther 1985; 233:857–62. 44. Trudell JR, Ardies CM, Anderson WR. Antibodies raised against trifluoroacetyl-protein adducts bind to N-trifluoroacetyl-phosphatidylethanolamine in hexagonal phase phospholipid micelles. J Pharmacol Exp Ther 1991; 257:657–62. 45. Kharasch ED, Hankins DC, Fenstamaker K, Cox K. Human halothane metabolism, lipid peroxidation, and cytochromes P(450)2A6 and P(450)3A4. Eur J Clin Pharmacol 2000; 55:853–9. 46. Kharasch ED, Hankins D, Mautz D, Thummel KE. Identification of the enzyme responsible for oxidative halothane metabolism: implications for prevention of halothane hepatitis. Lancet 1996; 347:1367–71. 47. Eliasson E, Gardner I, Hume-Smith H, de Waziers I, Beaune P, Kenna JG. Interindividual variability in P450-dependent generation of neoantigens in halothane hepatitis. Chem Biol Interact 1998; 116:123–41. 48. Kharasch ED, Thummel KE, Mautz D, Bosse S. Clinical enflurane metabolism by cytochrome P450 2E1. Clin Pharmacol Ther 1994; 55:434–40. 49. Kharasch ED, Hankins DC, Cox K. Clinical isoflurane metabolism by cytochrome P450 2E1. Anesthesiology 1999; 90:766–71. 50. Chase RE, Holaday DA, Fiserova-Bergerova V, Saidman LJ, Mack FE. The biotransformation of ethrane in man. Anesthesiology 1971; 35:262–7. 51. Holaday DA, Fiserova-Bergerova V, Latto IP, Zumbiel MA. Resistance of isoflurane to biotransformation in man. Anesthesiology 1975; 43:325–32. 52. Fiserova-Bergerova V, Holaday DA. Uptake and clearance of inhalation anesthetics in man. Drug Metab Rev 1979; 9:43–60. 53. Burke TR, Jr., Branchflower RV, Lees DE, Pohl LR. Mechanism of defluorination of enflurane: identification of an organic metabolite in rat and man. Drug Metab Dispos 1981; 9:19–24. 54. Christ DD, Satoh H, Kenna JG, Pohl LR. Potential metabolic basis for enflurane hepatitis and the apparent cross-sensitization between enflurane and halothane. Drug Metab Dispos 1988; 16:135–40. 55. Hitt BA, Mazze RI, Cousins MJ, Edmunds HN, Barr GA, Trudell JR. Metabolism of isoflurane in Fischer 344 rats and man. Anesthesiology 1974; 40:62–7. 56. Sutton TS, Koblin DD, Gruenke LD, et al. Fluoride metabolites after prolonged exposure of volunteers and patients to desflurane. Anesth Analg 1991; 73:180–5. 57. Cook TL, Beppu WJ, Hitt BA, Kosek JC, Mazze RI. Renal effects and metabolism of sevoflurane in Fisher 3444 rats: an in-vivo and in-vitro comparison with methoxyflurane. Anesthesiology 1975; 43:70–7. 58. Holaday DA, Smith FR. Clinical characteristics and biotransformation of sevoflurane in healthy human volunteers. Anesthesiology 1981; 54:100–6. 59. Kharasch ED. Biotransformation of sevoflurane. Anesth Analg 1995; 81:S27–38. 60. Hanaki C, Fujii K, Mario M, Tashima T. Decomposition of sevoflurane by soda lime. Hiroshima J Med Sci 1987; 36:61–7.

Hepatotoxicity of Anesthetic Agents

481

61. Jin L, Davis MR, Kharasch ED, Doss GA, Baillie T. Identification in rat bile of glutathione conjugates of fluoromethyl 2,2-difluoro-1-(trifluoromethyl) vinyl ether, a nephrotoxic degradate of the anesthetic agent sevoflurane. Chem Res Toxicol 1996; 9:555–61. 62. Spracklin D, Kharasch ED. Evidence for the metabolism of fluoromethyl 2,2-difluoro-1-(trifluoromethyl) vinyl ether (Compound A) by cysteine conjugate b-lyase in rats. Chem Res Toxicol 1996; 9:696–702. 63. Iyer RA, Frink EJ, Jr., Ebert TJ, Anders MW. Cysteine conjugate b-lyase dependent metabolism of compound A (2-[fluoromethoxy]-1,1,3,3,3-pentafluoro-1-propene) in human subjects anesthetized with sevoflurane and in rats given compound A. Anesthesiology 1998; 88:611–8. 64. Altuntas TG, Park SB, Kharasch ED. Sulfoxidation of cysteine and mercapturic acid conjugates of the sevoflurane degradation product fluoromethyl 2,2-difluoro-1-(trifluoromethyl) vinyl ether (Compound A). Chem Res Toxicol 2004; 17:435–45. 65. Sheffels P, Schroeder JL, Altunatas TG, Liggitt HD, Kharasch ED. Role of cytochrome P4503A in cysteine S-conjugates sulfoxidation and the nephrotoxicity of the sevoflurane degradation product fluoromethyl 2,2-difluoro-1-(trifluoromethyl) vinyl ether (Compound A) in rats. Chem Res Toxicol 2004; 17:1177–89. 66. Reynolds ES, Moslen MT. Halothane hepatotoxicity: enhancement by polychlorinated biphenyl pretreatment. Anesthesiology 1977; 47:19–27. 67. McLain GE, Sipes IG, Brown BR, Jr. An animal model of halothane hepatotoxicity: roles of enzyme induction and hypoxia. Anesthesiology 1979; 51:321–6. 68. Cousins MJ, Sharp JH, Gourlay GK, Adams JF, Haynes WD, Whitehead R. Hepatotoxicity and halothane metabolism in an animal model with application for human toxicity. Anaesth Intens Care 1979; 7:9–24. 69. Pohl LR, Gillette JR. A perspective on halothane-induced hepatotoxicity. Anesth Analg 1982; 61:809–11 (letter). 70. Clarke JB, Gandolfi AJ. Volatile anesthetics: mechanisms of potential hepatotoxicity. Clin Anesth Updates 1992; 3:1–11. 71. Berman ML, Kuhnert L, Phythyon JM, Holaday DA. Isoflurane and enflurane-induced hepatic necrosis in triiodothyronine-pretreated rats. Anesthesiology 1983; 58:1–5. 72. Rice SA, Maze M, Smith CM, Kosek JC, Mazze RI. Halothane hepatotoxicity in Fischer 344 rats pretreated with isoniazid. Toxicol Appl Pharmacol 1987; 87:411–9. 73. Lunam CA, Cousins MJ, Hall PD. Guinea-pig model of halothane-associated hepatotoxicity in the absence of enzyme induction and hypoxia. J Pharmacol Exp Ther 1985; 232:802–9. 74. Lind RC, Gandolfi AJ, Hall PD. The role of oxidative biotransformation of halothane in the guinea pig model of halothane-associated hepatotoxicity. Anesthesiology 1989; 70:649–53. 75. Farrell GC, Frost L, Tapner M, Field J, Weltman M, Mahoney J. Halothane-induced liver injury in guinea-pigs: importance of cytochrome P450 enzyme activity and hepatic blood flow. J Gastroenterol Hepatol 1996; 11:594–601. 76. Lind RC, Gandolfi AJ, Hall PM. Glutathione depletion enhances subanesthetic halothane hepatotoxicity in guinea pigs. Anesthesiology 1992; 77:721–7. 77. de Groot H, Noll T. Halothane hepatotoxicity: relation between metabolic activation, hypoxia, covalent binding, lipid peroxidation and liver cell damage. Hepatology 1983; 3:601–6. 78. Odaka Y, Takahashi T, Yamasaki A, et al. Prevention of halothane-induced hepatotoxicity by hemin pretreatment. Protective role of heme oxygenase-1 induction. Biochem Pharmacol 2000; 59:871–80. 79. Goto T, Ohwan K, Matsumoto N, et al. Protective effect of calcium channel blockers on the liver against halothane hepatitis in rats. Masui 1990; 39:204–9. 80. Farrell GC, Mahoney J, Bilous M, Frost L. Altered hepatic calcium homeostasis in guinea pigs with halothane-induced hepatotoxicity. J Pharmacol Exp Ther 1988; 247:751–6. 81. Frost L, Mahoney J, Field J, Farrell GC. Impaired bile flow and disordered hepatic calcium homeostasis are early features of halothane-induced liver injury in guinea pigs. Hepatology 1996; 23:80–6. 82. Bourdi M, Reilly TP, Elkahloun AG, George JW, Pohl LR. Macrophage migration inhibitory factor in drug-induced liver injury: a role in susceptibility and stress responsiveness. Biochem Biophys Res Comm 2002; 294:225–30. 83. Bourdi M, Amouzadeh HR, Rushmore TH, Martin JL, Pohl LR. Halothane-induced liver injury in outbred guinea pigs: role of trifluoroacetylated protein adducts in animal susceptibility. Chem Res Toxicol 2001; 14:362–70. 84. Van Dyke RA. Hepatic centrilobular necrosis in rats after exposure to halothane, enflurane, or isoflurane. Anesth Analg 1982; 61:812–9. 85. Harper MH, Collins P, Johnson B, Eger EI II, Biava C. Hepatic Injury following halothane, enflurane, and isoflurane anesthesia in rats. Anesthesiology 1982; 56:14–7. 86. Shingu K, Eger EI II, Johnson BH, et al. Hepatic injury induced by anesthetic agents in rats. Anesth Analg 1983; 62:140–5.

482

Kenna

87. Gonsowski CT, Laster MJ, Eger EI II, Ferrell LD, Kerschmann RL. Toxicity of compound A in rats. Effect of increasing duration of administration. Anesthesiology 1994; 80:566–73. 88. Keller KA, Callan C, Prokocimer P, et al. Inhalation toxicology study of a haloalkene degradant of sevoflurane, Compound A (PIFE), in Sprague–Dawley rats. Anesthesiology 1995; 83:1220–32. 89. Kharasch ED, Hoffman GM, Thorning D, Hankins DC, Kilty CG. Role of the cysteine conjugate b-lyase pathway in inhaled compound A nephrotoxicity in rats. Anesthesiology 1998; 88:1624–33. 90. Obata R, Bito H, Ohmura M, et al. The effects of prolonged low-flow sevoflurane anesthesia on renal and hepatic function. Anesth Analg 2000; 91:1262–8. 91. Vergani D, Tsantoulas D, Eddleston AL, Davis M, Williams R. Sensitisation to halothane-altered liver components in severe hepatic necrosis after halothane anaesthesia. Lancet 1978; 2:801–3. 92. Vergani D, Mieli-Vergani G, Alberti A, et al. Antibodies to the surface of halothane-altered rabbit hepatocytes in patients with severe halothane-associated hepatitis. N Engl J Med 1980; 303:66–71. 93. Neuberger J, Gimson AE, Davis M, Williams R. Specific serological markers in the diagnosis of fulminant hepatic failure associated with halothane anaesthesia. Br J Anaesth 1983; 55:15–9. 94. Kenna JG, Neuberger J, Williams R. An enzyme-linked immunosorbent assay for detection of antibodies against halothane-altered hepatocyte antigens. J Immunol Methods 1984; 75:3–14. 95. Kenna JG, Neuberger J, Williams R. Identification by immunoblotting of three halothane-induced liver microsomal polypeptide antigens recognized by antibodies in sera from patients with halothane-associated hepatitis. J Pharmacol Exp Ther 1987; 242:733–40. 96. Kenna JG, Neuberger J, Williams R. Evidence for expression in human liver of halothane-induced neoantigens recognized by antibodies in sera from patients with halothane hepatitis. Hepatology 1988; 8:1635–41. 97. Kenna JG, Satoh H, Christ DD, Pohl LR. Metabolic basis for a drug hypersensitivity: antibodies in sera from patients with halothane hepatitis recognize liver neoantigens that contain the trifluoroacetyl group derived from halothane. J Pharmacol Exp Ther 1988; 245:1103–9. 98. Kenna JG, Martin JL, Satoh H, Pohl LR. Factors affecting the expression of trifluoroacetylated liver microsomal protein neoantigens in rats treated with halothane. Drug Metab Dispos 1990; 18:788–93. 99. Harris JW, Pohl LR, Martin JL, Anders MW. Tissue acylation by the chlorofuorocarbon substitute 2,2dichloro-1,1,1-trifluoroethane. Proc Natl Acad Sci USA 1991; 88:1407–10. 100. Heijink E, De Matteis F, Gibbs AH, Davies A, White IN. Metabolic activation of halothane to neoantigens in C57B1/10 mice: immunochemical studies. Eur J Pharmacol 1993; 248:15–25. 101. Knight TL, Scatchard KM, Van Pelt FN, Kenna JG. Sera from patients with halothane hepatitis contain antibodies to halothane-induced liver antigens which are not detectable by immunoblotting. J Pharmacol Exp Ther 1994; 270:1325–33 (published erratum appears in J Pharmacol Exp Ther 1995; 272(2):962). 102. Ramsay L, Barnes S, Gardner I, Eliasson E, Kenna JG. Identification by immunoprecipitation of novel halothane-induced protein antigens recognised by antibodies from patients with halothane hepatitis. Br J Clin Pharmacol 1998; 46:303P–4P. 103. Kenna JG, Martin JL, Pohl LR. The topography of trifluoroacetylated protein antigens in liver microsomal fractions from halothane treated rats. Biochem Pharmacol 1992; 44:621–9. 104. Brown AP, Gandolfi AJ. Glutathione-S-transferase is a target for covalent modification by a halothane reactive intermediate in the guinea pig liver. Toxicology 1994; 89:35–47. 105. Eliasson E, Kenna JG. Cytochrome P450 2E1 is a cell surface autoantigen in halothane hepatitis. Mol Pharmacol 1996; 50:573–82. 106. Bourdi M, Chen W, Peter RM, et al. Human cytochrome P450 2E1 is a major autoantigen associated with halothane hepatitis. Chem Res Toxicol 1996; 9:1159–66. 107. Martin JL, Kenna JG, Martin BM, Thomassen D, Reed GF, Pohl LR. Halothane hepatitis patients have serum antibodies that react with protein disulfide isomerase. Hepatology 1993; 18:858–63. 108. Martin JL, Reed GF, Pohl LR. Association of anti-58 kDa endoplasmic reticulum antibodies with halothane hepatitis. Biochem Pharmacol 1993; 46:1247–50. 109. Satoh H, Martin BM, Schulick AH, Christ DD, Kenna JG, Pohl LR. Human antiendoplasmic reticulum antibodies in sera of patients with halothane-induced hepatitis are directed against a trifluoroacetylated carboxylesterase. Proc Natl Acad Sci USA 1989; 86:322–6. 110. Butler LE, Thomassen D, Martin JL, Martin BM, Kenna JG, Pohl LR. The calcium-binding protein calreticulin is covalently modified in rat liver by a reactive metabolite of the inhalation anesthetic halothane. Chem Res Toxicol 1992; 5:406–10. 111. Pumford NR, Martin BM, Thomassen D, et al. Serum antibodies from halothane hepatitis patients react with the rat endoplasmic reticulum protein ERp72. Chem Res Toxicol 1993; 6:609–15. 112. Davila JC, Martin BM, Pohl LR. Patients with halothane hepatitis have serum antibodies directed against glucose-regulated stress protein GRP78/BiP. Toxicologist 1992; 12:255. 113. Thomassen D, Martin BM, Martin JL, Pumford NR, Pohl LR. The role of a stress protein in the development of a drug-induced allergic response. Eur J Pharmacol 1990; 183:1138–9.

Hepatotoxicity of Anesthetic Agents

483

114. Amouzadeh HR, Bourdi M, Martin JL, Martin BM, Pohl LR. UDP-glucose:glycoprotein glucosyltrasferase associates with endoplasmic reticulum chaperones and its activity is decreased in vivo by the inhalation anesthetic halothane. Chem Res Toxicol 1997; 10:59–63. 115. Vidali M, Hidestrand M, Eliasson E, et al. Use of molecular simulation for mapping conformational CYP2E1 epitopes. J Biol Chem 2004; 279:50949–55. 116. Pohl LR, Thomassen D, Pumford NR, et al. Hapten carrier conjugates associated with halothane hepatitis. Adv Exp Med Biol 1991; 283:111–20. 117. Martin JL, Kenna JG, Pohl LR. Antibody assays for the detection of patients sensitized to halothane. Anesth Analg 1990; 70:154–9. 118. Smith GC, Kenna JG, Harrison DJ, Tew D, Wolf CR. Autoantibodies to hepatic microsomal carboxylesterase in halothane hepatitis. Lancet 1993; 342:963–4. 119. Kitteringham NR, Kenna JG, Park BK. Detection of autoantibodies directed against human hepatic endoplasmic reticulum in patients with halothane-associated hepatitis. Br J Clin Pharmacol 1995; 40:379–86. 120. Neuberger J, Kenna JG. Halothane hepatitis: a model of immunoallergic hepatitis. In: Guillouzo A, ed. Liver Cells and Drugs Paris & London: John Libbey Eurotext Ltd, 1988:161–73. 121. Njoku DB, Greenberg RS, Bourdi M, et al. Antoantibodies associated with volatile anesthetic hepatitis found in the sera of a large cohort of pediatric anesthesiologists. Anesth Analg 2002; 94:243–9. 122. Ellgaard L, Helenius A. Quality control in the endoplasmic reticulum. Nat Rev Mol Cell Biol 2003; 4:181–91. 123. Robin MA, Descatoire V, Le Roy M, et al. Vesicular transport of newly synthesized cytochromes P4501A to the outside of rat hepatocyte plasma membranes. J Pharmacol Exp Ther 2000; 294:1063–9. 124. Neve EP, Ingelman-Sundberg M. Molecular basis for the transport of cytochrome P450 2E1 to the plasma membrane. J Biol Chem 2000; 275:17130–5. 125. Zhu Q, von Dippe P, Xing W, Levy D. Membrane topology and cell surface targeting of microsomal epoxide hydrolase. Evidence for multiple topological orientations. J Biol Chem 1999; 274:27898–904. 126. Christ DD, Kenna JG, Kammerer W, Satoh H, Pohl LR. Enflurane metabolism produces covalently bound liver adducts recognized by antibodies from patients with halothane hepatitis. Anesthesiology 1988; 69:833–8. 127. Njoku D, Laster MJ, Gong DH, Eger El II, Reed GF, Martin JL. Biotransformation of halothane, enflurane, isoflurane, and desflurane to trifluoroacetylated liver proteins: association between protein acylatioan and hepatic injury. Anesth Analg 1997; 84:173–8. 128. Harris JW, Jones JP, Martin JL, et al. Pentahaloethane-based chlorofluorocarbon substitutes and halothane: correlation of in vivo hepatic protein trifluoroacetylation and urinary trifluoroacetic acid excretion with calculated enthalpies of activation. Chem Res Toxicol 1992; 5:720–5. 129. Marit GB, Dodd DE, George ME, Vinegar A. Hepatotoxicity in guinea pigs following acute inhalation exposure to 1,1-dichloro-2,2,2-trifluoroethane. Toxicol Pathol 1994; 22:404–14. 130. Hoet P, Graf ML, Bourdi M, et al. Epidemic of liver disease caused by hydrochlorofluorocarbons used as ozone-sparing substitutes of chlorofluorocarbons. Lancet 1997; 350:556–9. 131. Callis AK, Brooks SD, Roth TP, Gandolfi AJ, Bown BR. Characterization of a halothane-induced humoral immune responses in rabbits. Clin Exp Immunol 1987; 67:343–51. 132. Roth TP, Hubbard AK, Gandolfi AJ, Brown BR. Chronology of halothane-induced antigen expression in halothane-exposed rabbits. Clin Exp Immunol 1988; 72:330–6. 133. Siadat-Pajouh M, Hubbard AK, Roth TP, Gandolfi AJ. Generation of halothane-induced immune response in a guinea pig model of halothane hepatitis. Anesth Analg 1987; 66:1209–14. 134. Furst SM, Luedke D, Gaw HH, Reich R, Gandolfi AJ. Demonstration of a cellular immune response in halothane-exposed guinea pigs. Toxicol Appl Pharmacol 1997; 143:245–55. 135. Mathieu A, Dipadua D, Kahan BD, Galdabini JJ, Mills J. Correlation between specific immunity to a metabolite of halothane and hepatic lesions after multiple exposures. Anesth Analg 1975; 54:332–9. 136. Neuberger J, Kenna JG, Williams R. Halothane hepatitis: attempt to develop an animal model. Int J Immunopharmacol 1987; 9:123–32. 137. Hastings KL, Thomas C, Brown AP, Gandolfi AJ. Trifluoroacetylation potentiates the humoral immune response to halothane in the guinea pig. Immunopharmacol Immunotoxicol 1995; 17:201–13. 138. Zheng XH, Begay C, Lind RC, Gandolfi AJ. Humoral immune response to a sevoflurane degradation product in the guinea pig following inhalation exposure. Drug Chemical Toxicol 2001; 24:339–46. 139. Allison AC. Theories of self tolerance and autoimmunity. In: Kammuller ME, Bloksma N, Seinen W, eds. Autoimmunity and Toxicology: Immune Dysregulation Induced by Drugs and Chemicals. Amsterdam: Elsevier, 1989:67–115. 140. O’Shea D, Davis SN, Kim RB, Wilkinson GR. Effect of fasting and obesity in humans on the 6-hydroxylation of chlorzoxazone: a putative probe of CYP2E1 activity. Clin Pharmacol Ther 1994; 56:359–67.

484

Kenna

141. Young SR, Stoelting RK, Peterson C, Madura JA. Anesthetic biotransformation and renal function in obese patients during and after methoxyflurane or halothane anesthesia. Anesthesiology 1975; 42:451–7. 142. Farrell G, Prendergast D, Murray M. Halothane hepatitis. Detection of a constitutional susceptibility factor. N Engl J Med 1985; 313:1310–4. 143. Hoft RH, Bunker JP, Goodman HI, Gregory PB. Halothane hepatitis in three pairs of closely related women. N Engl J Med 1981; 304:1023–4. 144. Otsuka S, Yamamoto M, Kasuya S, et al. HLA antigens in patients with unexplained hepatitis following halothane anesthesia. Acta Anaesthesiol Scand 1985; 29:497–501. 145. Eade OE, Grice D, Krawitt EL, et al. HLA A and B locus antigens in patients with unexplained hepatitis following halothane anaesthesia. Tissue Antigens 1981; 17:428–32. 146. Gut J, Christen U, Frey N, Koch V, Stoffler D. Molecular mimicry in halothane hepatitis: biochemical and structural characterization of lipoylated autoantigens. Toxicology 1995; 97:199–224. 147. Christen U, Quinn J, Yeaman SJ, et al. Identification of the dihydrolipoamide acetyltransferase subunit of the human pyruvate dehydrogenase complex as an autoantigen in halothane hepatitis. Molecular mimicry of trifluoroacetyl-lysine by lipoic acid. Eur J Biochem 1994; 223:1035–47. 148. Frey N, Christen U, Jeno P, et al. The lipoic acid containing components of the 2-oxoacid dehydrogenase complexes mimic trifluoroacetylated proteins and are autoantigens associated with halothane hepatitis. Chem Res Toxicol 1995; 8:736–46. 149. Gut J, Christen U, Huwyler J, Burgin M, Kenna JG. Molecular mimicry of trifluoroacetylated human liver protein adducts by constitutive proteins and immunochemical evidence for its impairment in halothane hepatitis. Eur J Biochem 1992; 210:569–76. 150. Matzinger P. Tolerance, danger, and the extended family. Annu Rev Immunol 1994; 12:991–1045. 151. Park BK, Pirmohamed M, Kitteringham NR. Role of drug disposition in drug hypersensitivity: a chemical, molecular, and clinical perspective. Chem Res Toxicol 1998; 11:969–88. 152. Uetrecht JP. New concepts in immunology relevant to idiosyncratic drug reactions: the “danger hypothesis” and innate immune system. Chem Res Toxicol 1999; 12:387–95. 153. Crispe IN. Hepatic T cells and liver tolerance. Nat Rev Immunol 2003; 3:51–62. 154. Bowen DG, Zen M, Holz L, Davis T, McCaughan GW, Bertolino P. The site of primary T cell activation is a determinant of the balance between intrahepatic tolerance and immunity. J Clin Invest 2004; 114:701–12. 155. Bowen DG, McCaughan GW, Bertolino P. Intrahepatic immunity: a tale of two sites? Trends Immunol 2005; 26:512–7.

23

Anticonvulsant Agents Munir Pirmohamed

Department of Pharmacology and Therapeutics, University of Liverpool, Liverpool, U.K.

Steven J. Leeder

Division of Pediatric Pharmacology and Medical Toxicology, Children’s Mercy Hospital and Clinics, Kansas City, Missouri, U.S.A.

INTRODUCTION Anticonvulsant agents are well known to cause severe liver injury. Agents such as phenacemide were removed from clinical use as a consequence of unacceptably high frequencies of liver toxicity. A clinically significant risk of hepatotoxicity also accompanies the use of phenytoin (Dilantinw, Parke-Davis, Morris Plains, NJ), carbamazepine (CBZ) (Tegretolw, Novartis Pharmaceuticals, Basel, Switzerland) and valproic acid (VPA) (Depakenew, Abbott Laboratories, Abbott Park, IL), and some of the newer anticonvulsants that have been introduced over the last decade. Evaluating which drug is causing liver injury in patients with seizure disorders is often difficult since (1) at least 30% of patients are refractory to treatment and are often on more than one anticonvulsant; (2) anticonvulsants may have complex interactions with other drugs used concomitantly in patients; and (3) the whole picture may be complicated by “environmental” factors such as alcohol misuse, which often accompanies or is a cause of the seizure disorder, and is also associated with liver injury. As such, it may be difficult to fully evaluate the hepatotoxic potential of individual agents; in many cases, and particularly for the newer agents, causality has to be presumed on the basis of case reports in the absence of any mechanistic laboratory studies. Nevertheless, the purpose of this chapter is to describe the clinical presentation, histopathology, mechanisms, and determinants of susceptibility of anticonvulsant-induced liver injury focusing on established and newer compounds. CARBAMAZEPINE CBZ is a widely used anticonvulsant, and is regarded as the drug of choice for partial epilepsies (1). It is also used in trigeminal neuralgia, neuropathic pain syndromes, and bipolar depression. Since its introducion in the 1960s, CBZ has been widely reported to adversely affect liver functions in one of three ways: 1. It leads to an increase in gamma glutamyl transferase (g-GT), and to lesser extent alkaline phosphatase (ALP), due to its enzyme inducing properties. A retrospective analysis showed that 64 and 14% of users had elevations of g-GT and ALP, respectively (2). Such an increase in liver enzymes is not an indication to stop the drug. 2. It leads to an asymptomatic mild to moderate increase in liver function tests, including transaminases. This has been observed in up to 22% of patients (3). The relationship to more severe forms of liver dysfunction is unclear. Munir Pirmohamed conducts research funded by sources that include charities (Wellcome Trust), research councils, UK Department of Health, and pharmaceutical companies including Astra Zeneca, Bristol Myers Squibb, GSK and Lilly.

486

Pirmohamed and Leeder

3. It leads to clinically symptomatic hepatic injury, which may occur in isolation or be part of a generalized hypersensitivity reaction. The exact incidence is unknown; an analysis of all adverse reactions to CBZ reported to the Swedish Regulatory Agency showed that liver disorders accounted for 10% of all reactions (4). The risk was estimated to be 16 cases per 100,000 treatment years. Furthermore, analysis of the reports of hepatotoxicity to the Danish Committee on Adverse Drug Reactions showed that CBZ rose from ninth in frequency between 1968 and 1978 (5) to second during 1978 and 1987 (6). The incidence of serious hypersensitivity reactions associated with liver involvement with the aromatic anticonvulsants has been estimated to be in the region of 1 in 5000 to 10,000 (7). In a systematic review of 165 published cases of CBZ hypersensitivity up to 1998, liver involvement was observed in 47% of cases (Pirmohamed et al., unpublished data). Clinical Manifestations Liver involvement by CBZ is often part of a hypersensitivity syndrome, although the liver can be affected on its own (3,8). In terms of severity, the effects on the liver range from an asymptomatic increase in liver enzymes (9), to fulminant hepatic failure, which has been reported to require liver transplantation (10) or lead to death (11). There is no obvious relationship with either the dose or serum levels of CBZ. The time for onset of symptomatic hepatotoxicity is about four weeks with a range of 1 to 16 weeks (12). Hepatic injury is often accompanied by fever, rash, and eosinophilia (13–17), typical features of a hypersensitivity reaction (8). This is sometimes termed the DRESS (drug rash with eosinophilia and systemic symptoms) syndrome (11). Occasionally, the hepatotoxicity is associated with hematological abnormalities (leukocytosis, agranulocytosis, pancytopenia, thrombocytopenia) (13,18,19), renal dysfunction (20) or pneumonitis (17). In some instances, the clinical picture resembles cholangitis, with jaundice, right upper quadrant abdominal pain, nausea, and vomiting being the predominant symptoms (21,22); the cholestasis may be prolonged in some cases (16,23). Rechallenge has been reported in a number of patients (12,13,21,24) and in accordance with an immune reaction, occurs sooner on re-exposure than on the initial challenge. Hepatic injury from CBZ usually recovers on drug withdrawal (12). However, the reactions can be fatal, with an estimated case fatality rate of 12% (25). Prognosis is worse in those with a predominantly hepatocellular pattern, than in those with cholestatic injury (12,25), although it is important to note that prognosis has been derived from individual case reports or small case series, and may thus be subject to reporting bias. Biochemical abnormalities in patients with CBZ-induced hepatic injury are variable; about 30% of cases have a cholestatic pattern with elevation of both ALP and g-GT, about 50% have a mixed pattern where elevation of ALP and g-GT is accompanied by an increase in transaminases, while the rest have a hepatocellular pattern where transaminases are grossly elevated with minimal changes in the cholestatic enzymes (26). There may be a rise in bilirubin levels, although the degree is variable, and is most prominent in those patients who present with cholangitis (21,22). A prolonged rise in bilirubin levels similar to that seen in primary biliary cirrhosis is observed in patients with vanishing bile duct syndrome (16,23). In patients with hepatocellular necrosis, the rise in bilirubin levels reflects the severity of damage (27) and may be accompanied by changes in clotting parameters. Pathology The histology like the biochemical picture is also variable. Granulomatous hepatitis is observed in up to three-quarters of patients (12,21,22). Granulomas, which are the predominant lesions, may be accompanied by tissue eosinophilia (26). Pericholangitis and bile duct injury are present occasionally; the vanishing bile duct syndrome has also been reported with CBZ therapy (16,23). This is characterized by disappearance of interlobular bile ducts, with or without inflammatory infiltration, and in more severe cases, cholestasis. Predominantly hepatocellular necrosis has also been reported; a case report of two children with acute liver failure demonstrated the presence of submassive necrosis on liver biopsy (10).

487

Anticonvulsant Agents

Diagnosis Diagnosis of CBZ-induced hepatotoxicity is largely clinical. The onset and offset of hepatic injury are important factors to consider; in most cases, onset occurs within 12 weeks of the start of drug therapy, while improvement in liver tests is seen within four weeks of stopping treatment (12). Rechallenge is often positive (12,24) but is usually not possible on ethical grounds. Clearly, patients with suspected CBZ-induced liver injury should have other causes excluded by the use of appropriate virological, immunological, and radiological investigations. There are no specific diagnostic laboratory tests. Positive cytotoxicity assays (8,24) and lymphocyte transformation tests (28,29), have been demonstrated in some patients, while others have been shown to have circulating autoantibodies (30–33). However, these tests are largely research tools at present; they are also labor-intensive, difficult to reproduce, and may be associated with a high false-negative rate. Susceptibility Factors Older patients may be more sensitive to hepatic reactions from CBZ, while there is no sex predilection (12,34). However, it is important to note that this is based on an analysis of adverse reaction reports, and is clearly liable to be biased by vagaries of any spontaneous adverse drug reaction reporting scheme. Indeed, severe reactions in children have also been reported (10). It is thought that susceptibility to CBZ hypersensitivity may be genetically determined (8,35); this has been borne out by a recent case report that described the occurrence of hypersensitivity in a pair of monozygotic twins (36). Genetic case control association studies have to date not shown any relationship to polymorphisms in genes coding for the metabolizing enzymes (37,38), although associations have been demonstrated in the major histocompatibility complex (MHC) on chromosome 6 (39)—this is discussed below.

Carbamazepine

Stable 10,11-CBZ Epoxide O CYP3A4/CYP2C8

N C O

N C O NH2

CYP3A4

NH2 Bioactivation CYP3A4

Unstable epoxide

Enzyme induction N C O

O

NH2

Bioinactivation

Detoxication

Microsomal epoxide hydrolase

Hapten formation

Immune response

Glutathione Tissue injury

Dihydrodiol

Glutathione conjugate

Carbamazepine hypersensitivity

FIGURE 1 Mechanism of CBZ hypersensitivity. Bioactivation of CBZ to an unstable arene oxide metabolite leads to hapten formation. Subsequent involvement of the immune systems results in tissue injury at the site(s) of hapten formation, including the liver. Abbreviation: CBZ, carbamazepine.

488

Pirmohamed and Leeder

Postulated Mechanisms Metabolism is thought to play an important role in the pathogenesis of CBZ hypersensitivity and hepatotoxicity (Fig. 1). Although the mechanism of toxicity is poorly understood it has been postulated that reactive metabolites (and not the parent drug) are the causal agents (8). The metabolism of CBZ in man and experimental animal, is complex. The major route of metabolism both in vitro and in vivo is 10,11-epoxidation to CBZ-10,11-epoxide (which is itself a pharmacologically active drug) (40–42). CBZ can also be metabolized to stable, cytotoxic, and protein-reactive metabolites (43). Several different types of reactive metabolites have been postulated (Fig. 2). Initial studies suggested that the reactive metabolite was an arene oxide (8,24). Indeed, detoxication products from the postulated arene oxide have been detected in rat bile (44) and suggested to be present in human urine (45). The arene oxide may also undergo further metabolism to catechols and quinones (46). Other reactive metabolites postulated include 9-acridine carboxaldehyde (47) and an iminoquinone intermediate derived from its precursor 2-hydroxyiminostilbene (48). The formation of these various reactive metabolites is dependent on oxidative metabolism by P450 enzymes. Epoxide formation (both stable and reactive) is, at least partly, dependent upon [cytochrome P450 (CYP)] CYP3A4 (49,50), while 2- and 3-hydroxycarbamazepine, which act as precursors for reactive metabolites, are formed by other P450s in addition to CYP3A4 (CYP2B6 for 3-hydroxy and CYP1A2, 2A6, 2B6, and 2E1 for the 2-hydroxy metabolite) (51). CYP3A4 is also important for the bioactivation of 2-hydroxycarbamazepine to the iminoquinone metabolite (52). The central role of CYP3A4 in the metabolism and bioactivation of CBZ is emphasized in vivo where it is known that CBZ induces CYP3A4, and thereby its own metabolism including the formation of the ringhydroxylated metabolites 2- and 3-hydroxy CBZ, (53). Interestingly, a recent study has shown that CBZ-10,11-epoxide, which is a stable and pharmacologically active metabolite of CBZ, when incubated with glutathione in the absence of microsomes and/or NADPH leads to 9-Acridine carboxaldehyde

Arene oxides other than carbamazepine 2,3-epoxide

Carbamazepine 2,3-epoxide

CHO

O N

N

H2N

O

O

3-Hydroxycarbamazepine

N

N

H2N Carbamazepine

O H2N Carbamazepine 10,11-epoxide

H2N

O

OH N

N H Iminostilbene

OH

N

O

OH

OH

N

O

?

OH

O

H2N H2N 2-Hydroxycarbamazepine 2, 3-Dihydroxycarbamazepine (carbamazepine catechol)

N H 2-Hydroxyiminostilbene

O O N Iminoquinone

O

N H2N

O

Carbamazepine o-quinone

FIGURE 2 Potential pathways of bioactivation of CBZ to reactive metabolites in man. Abbreviation: CBZ, carbamazepine.

Anticonvulsant Agents

489

the formation of glutathione adducts and can covalently bind to human plasma and liver proteins (54). The significance of this novel finding is unclear (see below). Whether and how the complex manifestations of CBZ idiosyncratic toxicity (including the different forms of liver toxicity) relate to the complexity of the metabolism and bioactivation of CBZ is unclear; it is possible that the metabolites that lead to liver toxicity are different from those involved in extrahepatic toxicity, but this needs further investigation. Based on the results of the in vitro cytotoxicity assay (8,24), a deficiency of microsomal epoxide hydrolase was thought to be responsible for predisposing to CBZ hypersensitivity. However, genetic analysis of the microsomal epoxide hydrolase gene has not identified specific mutations in patients with CBZ hypersensitivity (37,38). Furthermore, analysis of polymorphisms in glutathione transferases, catechol-O- methyl transferase, and quinone reductase has also not revealed any association with CBZ hypersensitivity (55). Hypersensitivity reactions to CBZ are thought to have an immune basis. This is evidenced by clinical manifestation of hypersensitivity such as rash, fever, and lymphadenopathy (8), as well as rapid recurrence on rechallenge (24). Furthermore, patients with CBZ hypersensitivity have been reported to have circulating autoantibodies (30–33), drug-reactive T cells (28,56), and positive patch tests (57), all of which support an immune-mediated pathogenesis. However, the mechanism by which CBZ induces an immune response has been the subject of intense debate recently. It has long been postulated that bioactivation of CBZ to the various reactive metabolites leads to an immune response through the formation of haptens (58). However, this fundamental concept that protein-conjugation is an obligatory step in the process of immune recognition of drugs has been challenged by the observation that T cell clones from patients hypersensitive to a number of drugs undergo proliferation in an antigen-processing-independent (but MHC-restricted) manner (59). The presence of T cell clones responsive to CBZ and its stable metabolites has now been shown in a number of in vitro studies of patients hypersensitive to CBZ (29,56,60); whether this pathway is important in the induction of the primary immune response in vivo however remains unclear. Consistent with the findings of the immunological studies, genetic studies have shown that individual susceptibility is mediated by various loci on the MHC. An initial study in Caucasians showed an association between serious hypersensitivity reactions (which comprised patients with varying forms of hepatotoxicity), but not mild skin reactions, and the K308 tumor necrosis factor a (TNFa) promoter gene polymorphism and the TNF2-DR3DQ2 haplotype (61). More recently, a strong association was observed with the heat shock protein locus in the class III region of the MHC in the same set of patients (62). Because of the high degree of linkage disequilibrium in the MHC, it is not clear whether the observed associations represent the causal variants, but the consistency of the evidence suggests that the predisposing gene in Caucasians lies on the ancestral haplotype 8.1 (39). A remarkable association between (HLA, human leukocyte antigen) HLA-B*1502 and CBZ-induced Stevens–Johnson syndrome has also recently been demonstrated in Han Chinese patients (63), but not in Caucasians (64). Furthermore, HLA-B*1502 is not associated with CBZ-induced hypersensitivity syndrome (65). Taken together, the recent genetic studies have demonstrated that the predisposing loci to CBZ-induced idiosyncratic reactions vary according to clinical manifestations, severity, and ethnicity. Where hepatotoxicity fits into the overall picture is unclear, but will require studies with adequate numbers of patients with different forms of liver injury.

OXCARBAZEPINE Oxcarbazepine is a keto-analogue of CBZ that has been available in Scandinavia for many years, and has only recently become licensed in the rest of Europe. It undergoes less oxidative metabolism than CBZ, and is a less potent enzyme inducer (66). There have been no literature reports of symptomatic hepatic injury with oxcarbazepine. However, there is cross-reactivity between CBZ and oxcarbazepine (24,67,68), with an estimated frequency of 25% (67). The immunological basis of this cross-reactivity has been demonstrated recently in an in vitro study, which showed that T cell response to CBZ is polyclonal, and that some

490

Pirmohamed and Leeder

subsets of T cells recognize both CBZ and oxcarbazepine (60). Therefore, it is possible that an individual who has suffered hepatic injury with CBZ will also develop similar injury with oxcarbazepine; in such patients, oxcarbazepine should either be used with caution (with close monitoring) or preferably not at all. PHENYTOIN Phenytoin, like CBZ, is an aromatic anticonvulsant. It has, however, been around much longer, and the first report of hypersensitivity to phenytoin appeared in 1941 (2). As with CBZ, phenytoin-induced hepatotoxicity is idiosyncratic in nature and is often accompanied by generalized hypersensitivity phenomena (69,70). Phenytoin, like CBZ, is also an enzyme inducer, and leads to asymptomatic elevation of g-GT in almost 100% of recipients (71). Serum transaminases may be mildly elevated, but often normalize despite continuation of therapy (72). More than 100 cases of symptomatic hepatic injury have been reported (25). The exact incidence is unknown, but has been estimated to be less than 1 in 10,000 patients. In a systematic review of 271 published cases of phenytoin hypersensitivity, liver involvement ranging from an elevation of liver enzymes to hepatic failure was present in 56% of cases (Pirmohamed et al., unpublished data). Clinical Manifestations The onset of hepatotoxicity after starting phenytoin ranges from a few days to eight weeks (26). There is no obvious relationship to dose or serum levels consistent with the idiosyncratic nature of the adverse reaction. An elevation of transaminases is the most common abnormality [Alanine amino tranferase (ALT)Oaspartate amino transferase (AST)], values ranging from 2 to 100 times the upper limit of normal (70). ALP may also be elevated, although less so than the transaminases, values ranging from two to eight times upper limit of normal. Hepatic injury often occurs as part of a hypersensitivity reaction (69,70), with hepatitis being second only to rash as the most common manifestation (73). The clinical features, in fact, are very similar to CBZ hypersensitivity, with rash, fever, eosinophilia, and leukocytosis often accompanying the hepatic injury. Jaundice is seen in nearly half of the patients with hepatitis (26). Lymphadenopathy and splenomegaly occur in 60% of cases (70), with the constellation of symptoms mimicking infectious mononucleosis. Cholestasis seems to be less common than observed with CBZ. Bilirubin is variably raised, and in severe cases there may be prolongation of the prothrombin time (26). Interstitial nephritis, pneumonitis, myositis, eosinophilic fasciitis, lupus-erythematosus-like syndrome, rhabdomyolysis, and pseudolymphoma have also been reported (69,70,74–76). Positive rechallenge has been reported in a number of patients (26). Early reports suggested a case fatality rate of 30% to 40% (2,77,78); however, this is probably an overestimate since liver involvement in most cases is mild, and recovers rapidly on drug withdrawal. Pathology Hepatocellular injury accompanied by a prominent inflammatory infiltrate (including eosinophil infiltration) is the most common histologic abnormality (70). The histological picture is similar to that seen in infectious mononucleosis apart from the eosinophil infiltration. Submassive or massive necrosis is seen in 15% of patients, and the necrosis tends to be mainly panacinar. Cholestasis has been reported in 10% of cases (79), but this is rarely the predominant lesion as it is often accompanied by hepatocellular injury producing a mixed pattern. Granulomatous hepatitis has also been reported although it is probably less common than witnessed with CBZ (70). Diagnosis The principles of diagnosis are similar to those mentioned above for CBZ. Thus, diagnosis is largely clinical, and requires the exclusion of non-drug causes. Although positive lymphocyte cytotoxicity (8,80) and transformation tests (81) have been reported, these cannot be routinely used as diagnostic tests.

491

Anticonvulsant Agents

Susceptibility Factors Phenytoin hepatotoxicity occurs predominantly in adults, with 80% being over the age of 20 years (25). However, phenytoin hepatotoxicity does occur in children, and indeed cholestatic hepatitis has been reported in a newborn infant (2). There appears to be no sex predilection (82). It has been suggested that American blacks are more susceptible to phenytoin reactions than Caucasians (70) including a cluster of three cases that was reported in an African American family (83). However, the apparent higher incidence of reactions in blacks may reflect inaccurate epidemiologic data, presumably reflecting the patient population served by inner city hospitals (82). Predisposition to phenytoin hypersensitivity is thought to be genetically determined: this has been suggested from the results of the in vitro cytotoxicity assay (8,80,84), and from a report of familial occurrence of phenytoin hypersensitivity (83), with two of the three affected siblings developing significant hepatitis. The nature of the genetic defect is unclear (discussed in more detail below). Postulated Mechanisms Evidence supporting that phenytoin reactions are immune-mediated include the clinical features, recurrence on rechallenge, positive lymphocyte transformation tests (81), and circulating antibodies to phenytoin (85). Chemically reactive metabolites produced through CYP-mediated metabolism of phenytoin are again thought to be important in the pathogenesis of phenytoin hypersensitivity. It has been suggested that phenytoin is metabolized to reactive arene oxides (86), and binding of these metabolites to endogenous macromolecules initiates an immune reaction (80). However, as with CBZ, Pichler, and coworkers have suggested that T cells from phenytoin hypersensitive patients may be able to recognize the parent drug itself in the absence of antigen processing (56). In human liver microsomes, the p-hydroxylated phenytoin metabolite, 5-(4’-hydroxyphenyl)-5-phenylhydantoin (p-HPPH), is more readily converted to a covalent adduct than the

NH O N H Phenytoin

O

CYP2C9 O

NH O N H Arene oxide

O

NH

HO

O N H r-hydroxy phenytoin O

CYP2C9 CYP2C19 Catechol CYP3A4

Hapten formation

NH

O O O

O N H Ortho-quinone

Phenytoin hypersensitivity

FIGURE 3 Mechanism of phenytoin hypersensitivity. The pathogenesis of phenytoin hypersensitivity is thought to proceed by mechanisms analogous to those described for CBZ. In addition to the postulated reactive arene oxide intermediate, recent evidence implicates a reactive ortho-quinone metabolite of phenytoin in hapten formation. Abbreviation: CBZ, carbamazepine.

492

Pirmohamed and Leeder

parent drug, phenytoin (87). Furthermore, the protein targets of the reactive species appear to be the members of the human CYP2C and CYP3A subfamilies that are responsible for their generation (88). CYP2C9, CYP2C19, and CYP3A4 are also responsible for the formation of a catechol metabolite of phenytoin (89) suggesting that the protein-reactive metabolites may be the o-quinone species derived from the catechol (Fig. 3). Antibodies in patient sera recognize members of the rat CYP2C and CYP3A subfamilies (30) suggesting that there may be a link between the bioactivation process and immune response in the pathogenesis of phenytoin idiosyncratic toxicity. While bioactivation in the liver may be mediated by CYP2C9, 2C19, and 3A4, a recent study has shown that CYP2C18 is capable of catalyzing the bioactivation of phenytoin more efficiently than CYP2C9, indicating that CYP2C18, which is poorly expressed in the liver but more highly expressed in other organs such as the skin, may be more particularly important in extrahepatic bioactivation (90). As with CBZ, cells taken from patients with phenytoin hypersensitivity are more sensitive to toxic metabolites of phenytoin generated by a murine microsomal system than cells from controls, suggesting a detoxification deficit (8,80). Although this has been postulated to be a deficiency of microsomal epoxide hydrolase, molecular analysis of the gene has not demonstrated the presence of specific mutations in patients with phenytoin hypersensitivity (37). It is known that CYP2C9 polymorphisms determine phenytoin dose requirements (91,92); one small study in 10 patients has suggested that CYP2C9 polymorphisms may also predispose to skin rashes (93), but these data need to be replicated in a large sample set. Whether there is a defect in the immune response genes, as demonstrated for CBZ hypersensitivity, is unknown. PHENOBARBITAL Phenobarbital is the oldest aromatic anticonvulsant, having been introduced in 1918. It is also an enzyme inducer, and thus can lead to asymptomatic increases in g-GT and ALP (71). Symptomatic hepatic injury has been reported with phenobarbital, although it is relatively rare (26). It is often associated with hypersensitivity manifestations such as rash, fever, and eosinophilia (94). Formation of chemically reactive metabolites and an inherited deficiency in the detoxication of these metabolites is thought to be involved in the pathogenesis of phenobarbital hypersensitivity (8). Cross-Sensitivity Between the Aromatic Anticonvulsants Given the common mechanisms of aromatic anticonvulsant hypersensitivity, it is not surprising that certain patients exhibit cross-sensitivity with these drugs. Shear et al., suggested that the rate of cross-sensitivity might be as high as 80% (8). A recent clinical study of 633 patients showed that 58% of patients who had a rash with phenytoin also developed a rash with CBZ, while 40% of those with a CBZ rash also developed a rash with phenytoin (95). The factors that determine whether a patient is going to exhibit cross-sensitivity are unclear. VALPROIC ACID VPA was introduced for clinical use in France in 1964, and eventually in the United States in 1978. It has a broad spectrum of anticonvulsant activity and is approved for the treatment of generalized absence seizures. It is also effective in the treatment of generalized tonic-clonic, myoclonic, atonic, and partial seizures with or without secondary generalization (96). In recent years, VPA has also become popular as adjunctive therapy for the treatment of bipolar disorders (96). Approval for this indication was received in 1995, and for the treatment of chronic headache in 1996. Initiation of VPA therapy is frequently associated with nausea, vomiting, and gastrointestinal disturbances that can be attenuated by gradual increases in dose, administration after meals, or use of sustained-release formulations. Excessive weight gain, hair changes, endocrinological effects, and neurological effects such as drowsiness, acute confusional states, irritability, and tremor may also be encountered (97). Most morbidity and mortality, however, is attributed to adverse events involving the liver.

Anticonvulsant Agents

493

Clinical Manifestations The clinical manifestations of VPA hepatotoxicity cover a broad spectrum. Dose-related elevations in hepatic transaminases may occur in approximately 40% of patients without attendant symptoms of hepatic dysfunction (98). These elevations are transient and generally abate with a reduction in dose (77). Occurring less frequently, but of greater clinical concern, is the potential for fulminant hepatic failure that is irreversible in most cases. Valproate often appears in the list of drugs most commonly associated with liver fatalities (99,100). Several retrospective reviews of fatal VPA hepatotoxicity have been published from which a consistent pattern of signs and systems has emerged (101–107). Severe hepatic damage initially manifests as nausea, vomiting, abdominal pain, increased seizure frequency, lethargy, and coma. An episode of status epilepticus in close temporal proximity to the appearance of symptoms has been reported in approximately 40% to 60% of patients (103,104,106). More recent reviews describe the frequent occurrence of febrile illness immediately prior to the onset of hepatic failure (106). The onset of symptoms occurs within the first six months of therapy in 95% of affected individuals (104,106), generally within the first two to three months (101–103). However, onset as early as day six (104) and as late as six years after initiation of therapy (105) has also been reported. AST, ALT and bilirubin may exceed three times the upper limit of normal in VPA-treated patients although fatal hepatotoxicity actually occurs in only a small number (!0.01%) of individuals (101). In fatal cases, high serum transaminase and bilirubin levels represent the consequences of extensive hepatocellular damage rather than being specific markers of VPAassociated hepatotoxicity. On the other hand, indicators of synthetic function, such as the prothrombin time and accompanying impairment of coagulation, likely provide more accurate assessment of residual hepatic function (77). Elevated ammonia levels are also present but are also observed in the absence of other indicators of compromised hepatic function. VPA has been reported to unmask latent heterozygous phenotypes of ornithine transcarbamylase deficiency leading to fatal hyperammonemic coma (106,107). Pathology The histopathology of VPA hepatotoxicity differs substantially from that associated with the aromatic anticonvulsants phenytoin, phenobarbital, and CBZ in that signs of immune involvement and eosinophilia are not present. In fatal cases, the most prominent findings reported by Zimmerman and Ishak consisted of microvesicular steatosis occurring primarily in periportal zone 1 together with zone 3 necrosis (108). A similar picture of hepatocellular damage, including microvesicular steatosis, cellular ballooning, and single cell or group necrosis was observed in a series of fatal pediatric cases (106). These authors also noted proliferation of bile ducts in two patients as has been reported by others (109). Susceptibility Factors A series of three retrospective studies of VPA-associated hepatotoxicity in the United States delineated, patient age less than two years, polytherapy with enzyme inducing anti-epileptic medications, developmental delay and coincident metabolic disorders as important risk factors for developing this adverse event (101–103). Although VPA hepatotoxicity may occur at any age, data collected between 1978 and 1986 indicated that the risk of fatal hepatotoxicity was highest in children less than two years of age receiving concurrent anticonvulsant therapy in whom the incidence was estimated to be approximately 1:500 (101,102). This represents a 16-fold increase in risk, relative to children of the same age on VPA monotherapy (1:8000). Comparative estimates of risk for older children aged three to 10 years were 1:11,000 on monotherapy and 1:6000 on polytherapy. The risk of fatal hepatotoxicity was essentially unchanged (1:600) in children less than two years of age on polytherapy between 1987 and 1993 despite a trend towards decreasing use of VPA in very young children—0.8% of the studied population from 1987 to 1993 (103) compared to 2.6% in the earlier studies (101,102). Other studies have confirmed polytherapy as a risk factor but found that there was little difference in risk between younger (less than three years of age) and older children (three to six years of age) (104,110).

494

Pirmohamed and Leeder

Patients with inborn errors of metabolism or reduced hepatic mitochondrial activity also appear to be at increased risk for fatal hepatotoxicity associated with VPA (110). Ko¨nig et al., pointed out that a similar picture of hepatic involvement may manifest as a consequence of several metabolic defects including medium chain acyl CoA dehydrogenase deficiency, ornithine transcarbamylase deficiency, carbamoyl phosphatase synthetase deficiency, pyruvate dehydrogenase deficiency, primary carnitine deficiency and methyl malonic academia, among others (106). The metabolic effects of VPA and its metabolites on mitochondrial b-oxidation (discussed in more detail below) may exacerbate the existing metabolic defects or unmask latent deficiencies in susceptible individuals. Epidemiologic studies have failed to demonstrate a relationship between VPA dose and hepatotoxicity (101–103). However, dose cannot be separated from polytherapy as a risk factor since, patients receiving concurrent antiepileptic drug therapy usually receive higher doses of VPA. The formation of putative toxic metabolites reportedly increases in proportion to serum VPA concentration (111) and with increasing VPA dose (112). It is most likely that the pathogenesis of severe VPA-associated hepatotoxicity is a multifactorial process with no single risk factor being a sole determinant of individual susceptibility. Rather, for example, low doses (and relatively low levels of toxic metabolite formation) may be sufficient for a hepatotoxic event in an individual with latent metabolic disorders whereas much larger doses would be required in less susceptible individuals. Postulated Mechanisms Two general mechanistic hypotheses for VPA-associated hepatotoxicity have emerged over the past few years. In each, VPA biotransformation appears to be intimately involved in the process leading to hepatotoxicity although the precise mechanism remains to be elucidated. There is no clinical or laboratory evidence to implicate the immune system in the hepatic injury caused by VPA, suggesting that it may be a form of metabolic idiosyncrasy. The first hypothesis focuses on VPA interference with b-oxidation of endogenous lipids. VPA forms an ester conjugate with carnitine (113) that may lead to secondary carnitine deficiency. Several lines of indirect evidence and in vitro studies (114) indicate that a thioester derivative of VPA and coenzyme A may exist as a metabolic intermediate in liver tissue. Depletion of coenzyme A or the VPA-CoA ester itself could be responsible for inhibition of mitochondrial metabolism (115). VPA undergoes b-oxidation to several products (2-ene-VPA, 3-hydroxy-VPA and 3-oxo-VPA; Fig. 4) and competes with endogenous lipids for enzymes in the b-oxidation pathway (116). Any or all of these effects could exacerbate existing latent deficiencies of mitochondrial function. The second important hypothesis focuses on hepatotoxic unsaturated VPA metabolites. This hypothesis is based on earlier observations that the microvesicular steatosis characteristic of VPA-associated hepatotoxicity bears considerable similarity to the clinical and histologic features of Jamaican vomiting sickness and Reye’s syndrome. An unsaturated metabolite of the u-oxidation pathway, 4-ene-VPA (Fig. 4), has received special attention because of its similarity to methylene cyclopropylacetic acid, the u-oxidation product of hypoglycin A, which is responsible for microvesicular steatosis in Jamaican vomiting sickness, and 4-ene-pentanoic acid, which is used to generate experimental models of Reye’s syndrome. In in vivo experimental models, 4-ene-VPA is more steatogenic than VPA in young rats (117) and is more potent as an inhibitor of b-oxidation (118). In vitro studies with cultured hepatocytes also indicate that 4-ene-VPA is more cytotoxic than the parent compound (119). Experimental evidence suggests that chemically reactive metabolites generated from 4-ene-VPA have the potential to inhibit enzymes in the b-oxidation pathway (120,121). Hepatic microsomal CYP isoforms are responsible for the formation of 4-ene-VPA and this activity is inducible by phenobarbital (122). Further work has specifically implicated CYP2C9 and to a lesser extent, CYP2A6 in the formation of 4-ene-VPA in humans (123). Administration of phenytoin and CBZ to adult epilepsy patients increases 4-ene-VPA formation approximately twofold under steady state conditions (124). Induction of CYP2A6 and CYP2C9 activities by anticonvulsants has not been rigorously evaluated in humans although modest increases in CYP2C immunoreactive proteins have been observed following phenobarbital treatment of cultured primary human hepatocytes (125,126). Given the

495

Anticonvulsant Agents

COO-glue

VPA-glucuronide

COOH COOH

VPA

COOH

CYP2C9 CYP2A6

β−Oxidation 4-ene-VPA

2,4-diene-VPA CoA

COOH

CO-S-CoA

2-ene-VPA

2,4-diene-VPA-CoA

GSH

COOH

OH

COOH

COOH

3-OH-VPA

COOH

NAC 5-NAC-3-ene-VPA

NAC 5-NAC-2-ene-VPA

Bioactivation pathway O 3-OXO-VPA β-Oxidation pathway

FIGURE 4 Major pathways for VPA biotransformation. Glucuronidation of the VPA carboxyl group to form an acyl glucuronide is quantitatively the most important pathway of VPA biotransformation in man. A significant proportion of VPA undergoes b-oxidation. Oxidation of VPA by CYP2C9 and CYP2A6 initiates metabolism down a bioactivation pathway. The extent of reactive metabolite formation can be estimated from glutathione conjugates of 2,4-diene-VPA detected in the urine as N-acetylcysteine adducts. Abbreviations: VPA, valproic acid; CYP, cytochrome P450.

comparative variabilities of CYP2A6 activity (30-fold) and CYP2C9 activity (!fivefold) in human liver microsomes (127), it has been proposed that CYP2C9 may be responsible for the majority of constitutive VPA 4-ene-desaturation while CYP2A6 plays a greater role during polytherapy with anticonvulsants (123). Interestingly, the CYP2C9 *2 and *3 polymorphisms decrease the formation of 4-ene-VPA, as well as hydroxylation at the 4- and 5- positions— whether this acts as a susceptibility factor for valproate hepatotoxicity requires further investigation (128).

496

Pirmohamed and Leeder

The implied relationship between coadministration of inducers, increased 4-ene-VPA formation and increased risk of hepatotoxicity in patients is analogous to the increased VPA toxicity observed in animal models following pretreatment with phenobarbital, an inducer of CYP2A, CYP2B, CYP2C, and CYP3A isoforms in rodents (129). Likewise, therapeutic drug monitoring studies of other drugs primarily metabolized by CYP2C9 (e.g., phenytoin) indicate that CYP2C9 activity in young children exceeds that of adults, gradually declining to adult levels during childhood (130). The fractional metabolism of VPA to the 4-ene metabolite (but not to the 2- or 3-ene metabolites) also has been shown to decline with increasing age (111,131) providing one possible explanation for the increased incidence of VPA-associated hepatotoxicity in young children. This has been confirmed in a recent study, which showed that the excretion of two thiol conjugates of the toxic 2,4-diene metabolite was increased in patients younger than 7.5 years, and in those on polytherapy with enzyme-inducing anticonvulsants (132). Furthermore, indirect evidence for a causal relationship between 4-ene-VPA and hepatotoxicity is derived from the observation of relatively high levels of 4-ene-VPA in case reports and studies of patients with fulminant hepatic failure (110,133) but not in VPA-treated children free of hepatic side effects (110). Despite these considerations, a direct causal relationship between 4-ene-VPA formation and hepatotoxicity has not been demonstrated unequivocally. For example, 4-ene-VPA has been detected in the plasma of patients with no or only minimal overt evidence of hepatic dysfunction (111), and in some studies, plasma levels of 4-ene-VPA did not appear to correlate with the degree of hyperammonemia (134) or hepatic involvement observed (135,136). Finally, in contrast to the studies cited earlier (111,131,132), Siemes et al., reported lower concentrations of 4-ene-VPA in children less than two years of age compared to older children (136). The apparent discrepancies in the literature regarding the role of 4-ene-VPA as a hepatotoxic metabolite and further resolution of toxic metabolites versus inhibition of b-oxidation as the primary mechanism of toxicity represent challenges for future research. The complexity of VPA metabolism obfuscates efforts to establish clear relationships between individual metabolites and the hepatotoxic process. Therefore, it may be useful to conceptualize the VPA hepatotoxic process as proceeded by two parallel processes (Fig. 5): VPA itself depletes the intra-mitochondrial pool of CoA and thus inhibits the mitochondrial b-oxidation of long-, medium- and short-chain natural fatty acids (137). Chemically reactive metabolites generated from 4-ene-VPA, such as 2,4-diene-VPA, have the potential to deplete mitochondrial glutathione pools (138) and through formation of conjugates with CoA (139), inhibit enzymes in the b-oxidation pathway (120,121). Identification of N-acetylcysteine conjugates of (E)-2, 4-diene-VPA in human urine provide evidence that metabolites sufficiently reactive to form thiol adducts have been formed (140). These conjugates are potentially useful markers for future investigations assessing the relationship between reactive metabolite exposure and hepatic damage.

FELBAMATE Felbamate was approved as an antiepileptic agent in the U.S. in July 1993 for use as monotherapy and adjunctive therapy for partial seizures (with and without generalization) in adults and as adjunctive therapy for generalized seizures associated with Lennox–Gastaut syndrome in children. While felbamate provided significant benefits to treated patients, reports of aplastic anemia attributed to felbamate started to appear in mid-1994 as well as cases of hepatic failure, including four deaths, by the fall of 1994. The clinical use of felbamate was severely curtailed after September 1994 when the Food and Drug Administration issued a warning of a higher than expected incidence of aplastic anemia and hepatic failure among patients treated with the drug (141). Clinical Manifestations and Pathology The risk of hepatic failure due to felbamate was initially estimated to be one per 26,000 to 34,000 exposures (141) and has been revised to approximately one per 18,500 to 25,000 exposures (142).

497

Anticonvulsant Agents

CYP2C9

VPA

4-ene-VPA

CYP2A6

CoA 2,4-diene-VPA

VPA-CoA

CoA Sequestation of VPA-CoA in mitochondria

2,4-diene-VPA-CoA

CoA available for fatty acid transport and subsequent b -oxidation

Inactivation of enzymes in b -oxidation pathway

Decreased b -oxidation

Latent Deficiencies in b -Oxidation or inborn errors metabolism

Hepatotoxity

FIGURE 5 Hypothesis for the mechanism of VPA hepatotoxicity. Both the parent compound, VPA, and reactive metabolites from the bioactivation pathway can lead to decreased b-oxidation by various mechanisms. In patients rendered susceptible by inborn errors of metabolism or latent deficiencies in mitochondrial metabolism, the effects of VPA or its reactive metabolites on b-oxidation may tip the balance towards the manifestations of hepatotoxicity. Abbreviation: VPA, valproic acid.

In one published case, a 61-year-old Caucasian female presented with a chief complaint of nausea, vomiting, and lethargy over the previous 3.5 weeks (143). On the first day of hospitalization (day 24 of felbamate therapy), evidence of hepatic dysfunction (AST 601 U/L and GGT 978 U/L) and eosinophilia were present. Hepatic function continued to decline over the ensuing two weeks ultimately progressing to multisystem organ failure. Massive to submassive necrosis without significant fibrosis was observed on microscopic sections and moderate inflammatory infiltrate consisting primarily of lymphocytes was present within portal tracts. Although few data are available to describe the typical presentation and clinical course, available information from seven cases of likely felbamate hepatotoxicity reveals a high incidence of females (6/7) and time to presentation 25 to 181 days. Two of the seven cases were patients less than 12 years of age and an aromatic anticonvulsant (primidone, phenobarbital, phenytoin or CBZ) was concomitantly administered in six cases (142). Postulated Mechanisms The mechanism of felbamate hepatotoxicity is unknown. However, considerable progress has been made over the past decade in identifying and characterizing potential reactive metabolites that may play a role in the pathogenesis of felbamate idiosyncratic toxicities.

498

Pirmohamed and Leeder

Evidence has been presented for an unstable aldehyde carbamate intermediate, 3-carbamoyl2-phenylproprionaldehyde (aldehyde monocarbamate), in the pathway leading to the formation of the major metabolite in humans, 3-carbamoyl-2-phenylproprionic acid (acid monocarbamate; Fig. 6). This aldehyde carbamate metabolite predominantly undergoes reversible cyclization to form a stable cyclic structure that may serve as a “reservoir” for transport to tissues distant to the site of formation, the liver. Alternatively, it may undergo elimination to form 2-phenylpropenal (commonly called atropaldehyde) a potent electrophile that is toxic to cells in culture (144). Atropaldehyde undergoes rapid conjugation with glutathione and can be observed in urine from felbamate-treated patients as mercapturate derivatives (145). Formation of the aldehyde carbamate appears to be a “commitment step” whereby the molecule is committed to a detoxication pathway leading to 3-carbamoyl-2-phenylproprionic acid, the major urinary metabolite, or the toxic pathway leading to atropaldehyde. The ratio of urinary mercapturate metabolites to the acid monocarbamate metabolite represents an estimate of the balance between bioactivation and detoxication, and may provide a marker for susceptibility to felbamate hepatotoxicity or aplastic anemia for future investigations (146). The mechanism by which atropaldehyde induces liver injury is unclear with both metabolic and immunologic mechanisms having been implicated. Atropaldehyde may impair hepatocyte detoxification mechanisms by binding to aldehyde dehydrogenase and glutathione transferases, and impairs hepatocyte viability which can be partially reversed by coincubation with

CH2OCONH2

CH2OCONH2 Felbamate HO

H N

Alcohol monocarbamate O

O

CHO

Cyclic carbanate "reservoir" CH2OCONH2

CHO

Atropaldehyde (Detected in urine as N-acetylcystein adducts)

Bio

a

Aldehyde monocarbanate "commitment step" n De o i tox at id ctiv

atio

n

COOH

CH2OCONH2 Acid carbamate (Major urinary metabolite)

FIGURE 6 Proposed mechanism of felbamate bioactivation. The initial step toward felbamate bioactivation is thought to involve hydrolysis to form and alcohol monocarbamate metabolite that is further oxidized to an aldehyde monocarbamate. The aldehyde monocarbamate metabolite can cyclize to form a structure that has been proposed to function as a relatively stable reservoir for transport throughout the body. More importantly, the aldehyde monocarbamate metabolite represents a commitment step with commitment down a detoxication pathway to form the major urinary acid carbamate metabolite, or down a bioactivation pathway leading to the reactive metabolite, atropaldehyde, which can be detected in urine as N-acetylcysteine adducts. Individual susceptibility may be determined, in part, by the relative contributions of each competing pathway.

Anticonvulsant Agents

499

glutathione (147). By contrast, Uetrecht and coworkers were able to show that atropaldehyde is a potent immunogen in a popliteal lymph node assay (148).

LAMOTRIGINE Lamotrigine, a phenyltriazine, is a broad-spectrum anticonvulsant that has been in use for about a decade. The main idiosyncratic adverse effects associated with lamotrigine use are skin rashes, which occur in between 3% and 10% of patients (149). Children seem to be more susceptible to cutaneous adverse reactions than adults (150). Such rashes however are often only one component of a generalized hypersensitivity reaction, which is also accompanied by fever and eosinophilia (151). In such cases, liver involvement is characterized by an abnormality of liver function without clinical symptoms of hepatitis (151). However, more severe liver damage by lamotrigine has also been reported; for example, there are two case reports of fulminant hepatic failure (152,153). In both patients, hepatic failure developed after introduction of lamotrigine, and was characterized by jaundice, an increase in transaminases, and coagulopathy that proved fatal in one case. Postmortem examination in one patient showed massive hepatic necrosis and extensive bile duct proliferation (154). The mechanism of lamotrigine-induced hypersensitivity and hepatotoxicity is unclear. Patients started on high doses of lamotrigine and those on concomitant therapy with sodium valproate seem to be at higher risk of lamotrigine rashes; the strategy of starting at low doses and escalating the dose slowly seems to reduce the risk. The clinical symptomatology is suggestive of an immune-mediated pathogenesis, and is consistent with reports of positive lymphocyte transformation tests (155). As with the aromatic anticonvulsants, metabolism is likely to be important in the pathogenesis of the reactions. Lamotrigine largely undergoes N-glucuronidation with little oxidative metabolism (156,157). A recent study in a rat model has shown that lamotrigine can undergo bioactivation to an arene oxide (158), which may be important in the pathogenesis of the hypersensitivity reactions. No pharmacogenetic studies have been performed so far to investigate whether there is genetic predisposition to lamotrigine hypersensitivity.

MANAGEMENT The first and essential step in management of anticonvulsant-associated hepatotoxicity is the recognition that the drug is responsible for the hepatic injury. This is essentially a diagnosis of exclusion, and will require measurement of biochemical, immunological, and virological markers to exclude non-drug induced diseases. If the drug is suspected, discontinuation of the offending agent is important. Subsequently, the management of severe hepatic toxicity attributed to anticonvulsant therapy is essentially supportive. There is little evidence to support the use of steroids in treatment, even when the hepatic injury is thought to be immunemediated. As a result, prevention, and monitoring are more effective means of minimizing the impact of anticonvulsant-associated hepatotoxicity. In the case of VPA, risk factors are reasonably well characterized and the drug should be avoided in children less than three years of age and those treated with CYP-inducing aromatic anticonvulsants. Likewise, extreme caution should be exercised if a family history of fatty acid oxidation defects or urea cycle defects is present. If there has been a hepatic reaction to one aromatic anticonvulsant, then given the possibility of cross-sensitivity, the other aromatic anticonvulsants should be avoided. There is no evidence of cross-reactivity between the aromatic anticonvulsants and VPA, and this may be used for future control of seizures, although the drug should be introduced cautiously while liver function is still impaired. Routine monitoring of liver function tests is recommended by several manufacturers of the anticonvulsants reviewed; however, there is little evidence to support the predictive value of such monitoring. Therefore, as a general rule, clinicians should suspect and rule out hepatotoxicity in any patient who becomes ill in the first six months of anticonvulsant treatment.

500

Pirmohamed and Leeder

REFERENCES 1. Chadwick D. Safety and efficacy of vigabatrin and carbamazepine in newly diagnosed epilepsy: a multicentre randomised double-blind study. Vigabatrin European Monotherapy Study Group [see comments]. Lancet 1999; 354:13–19. 2. Stricker BHC. Drug-Induced Hepatic Injury. Amsterdam: Elsevier Science Publishers, 1992. 3. Pellock JM. Carbamazepine side effects in children and adults. Epilepsia 1987; 28:S64–70. 4. Askmark H, Wiholm B. Epidemiology of adverse drug reactions to carbamazepine as seen in a spontaneous reporting system. Acta Neurol Scand 1990; 81:131–40. 5. Døssing M, Andreasen PB. Drug-induced liver disease in Denmark. An analysis of 572 cases of hepatotoxicity reported to the Danish Board of Adverse Reactions to Drugs. Scand J Gastroenterol 1982; 17:205–11. 6. Friis H, Andreasen PB. Drug-induced hepatic injury. An analysis of 1100 cases reported to the Danish Committee on Adverse Drug Reactions between 1978 and 1987. J Intern Med 1992; 232:133–8. 7. Tennis P, Stern RS. Risk of serious cutaneous disorders after initiation of use of phenytoin, carbamazepine, or sodium valproate: a record linkage study. Neurology 1997; 49(2):542–6. 8. Shear NH, Spielberg SP. Anticonvulsant hypersensitivity syndrome. In vitro assessment of risk. J Clin Invest 1988; 82:1826–32. 9. Pirmohamed M, Park BK. Cytochromes P450 and liver injury. In: Cameron RG, Feuer G, de la Iglesia F, eds. Drug-Induced Hepatotoxicity. Berlin: Springer, 1996:341–66. 10. Hadzic N, Portmann B, Davies ET, Mowat AP, Mieli-Vergani G. Acute liver failure induced by carbamazepine. Arch Dis Child 1990; 65:315–7. 11. Syn WK, Naisbitt DJ, Holt AP, Pirmohamed M, Mutimer DJ. Carbamazepine-induced acute liver failure as part of the DRESS syndrome. Int J Clin Pract 2005; 59(8):988–91. 12. Williams RJ, Ruppin DC, Grierson JM, Farrell GC. Carbamazepine hepatitis. The clinicopatholoical spectrum. J Gastroenterol Hepatol 1986; 1:159–68. 13. Levander HG. Granulomatous hepatitis in a patient receiving carbamazepine. Acta Med Scand 1980; 208:333–5. 14. Hopen G, Nesthus I, Laerum OD. Fatal carbamazepine-associated hepatitis. Report of 2 cases. Acta Med Scand 1981; 210:333–5. 15. Soffer EE, Taylor RJ, Bertram PD, Hagitt RC, Levinson MJ. Carbamazepine-induced liver injury. South Med J 1983; 76:681–3. 16. Larrey D, Hadengue A, Pessayre D, Choudat L, Degott C, Benhamou JP. Carbamazepine-induced acute cholangitis. Dig Dis Sci 1987; 32:554–7. 17. Cox NH, Johnston SRD, Marks J, Bates D. Extensive carbamazepine eruption with eosinophilia and pulmonary infiltrate. Postgrad Med J 1988; 64:249–50. 18. Fellows WR. A case of aplastic anemia and pancytopenia with tegretol therapy. Headache 1969; 9:92–5. 19. Ponte CD. Carbamazepine-induced thrombocytopenia, rash, and hepatic dysfunction. Drug Intell Clin Pharmcol 1983; 17:642–4. 20. Imai H, Nakamoto Y, Hirokawa M, Akihama T, Miura AB. Carbamazepine-induced granulomatous necrotizing angiitis with acute renal failure. Nephron 1989; 51:405–8. 21. Levy M, Goodman MW, Van Dyre J, Summer HW. Granulomatous hepatitis secondary to carbamazepine. Ann Intern Med 1981; 95:64–5. 22. Mitchell MC, Boitnott JK, Arregui A, Maddrey WC. Granulomatous hepatitis associated with carbamazepine therapy. Am J Med 1981; 71:733–5. 23. Forbes GM, Jeffrey GP, Shilkin KB, Reed WD. Carbamazepine hepatotoxicity. Another case of the vanihing bile duct syndrome. Gastroenterology 1992; 102:1385–8. 24. Pirmohamed M, Graham A, Roberts P, et al. Carbamazepine-hypersensitivity: assessment of clinical and in vitro chemical cross-reactivity with phenytoin and oxcarbazepine. Br J Clin Pharmacol 1991; 32:741–9. 25. Zimmerman HJ, Ishak KG. Antiepileptic drugs. In: Cameron RG, Feuer G, de la Iglesia F, eds. DrugInduced Hepatotoxicity. Berlin: Springer, 1996:637–62. 26. Farrell GC. Drug-Induced Liver Disease. Edinburgh: Churchill Livingstone, 1994. 27. Murphy JV, Francisco CB, Roberts C. Fatal hepatorenal failure in 3 patients receiving carbamazepine. Ann Neurol 1991; 30:276–7. 28. Zakrzewska JM, Ivanyi L. In vitro lymphocyte proliferation by carbamazepine, carbamazepine10,11-epoxide, and oxcarbazepine in the diagnosis of drug-induced hypersensitivity. J Allergy Clin Immunol 1988; 82:110–5. 29. Naisbitt DJ, Britschgi M, Wong G, et al. Hypersensitivity reactions to carbamazepine: characterization of the specificity, phenotype, and cytokine profile of drug-specific T cell clones. Mol Pharmacol 2003; 63(3):732–41. 30. Leeder JS, Riley RJ, Cook VA, Spielberg SP. Human anti-cytochrome P450 antibodies in aromatic anticonvulsant-induced hypersensitivity reactions. J Pharmacol Exp Ther 1992; 263:360–7.

Anticonvulsant Agents

501

31. Pirmohamed M, Kitteringham NR, Breckenridge AM, Park BK. Detection of an autoantibody directed against human liver microsomal protein in a patient with carbamazepine hypersensitivity. Br J Clin Pharmacol 1992; 33:183–6. 32. Riley RJ, Smith G, Wolf CR, Cook VA, Leeder JS. Human anti-endoplasmic reticulum antibodies produced in aromatic anticonvulsant hypersensitivity reactions recognise rodent CYP3A proteins and a similarly regulated human P450 enzyme(s). Biochem Biophys Res Commun 1993; 191:32–40. 33. Leeder JS, Gaedigk A, Lu X, Cook VA. Epitope mapping studies with human anti-cytochrome P450 3A antibodies. Mol Pharmacol 1996; 49:234–43. 34. Horowitz S, Patwardhan R, Marcus E. Hepatotoxic reactions associated with carbamazepine therapy. Epilepsia 1988; 29:149–54. 35. Shear NH, Bhimji S. Pharmacogenetics and cutaneous drug reactions. Semin Dermatol 1988; 8:219–26. 36. Edwards SG, Hubbard V, Aylett S, Wren D. Concordance of primary generalised epilepsy and carbamazepine hypersensitivity in monozygotic twins. Postgrad Med J 1999; 75:680–1. 37. Gaedigk A, Spielberg SP, Grant DM. Characterization of the microsomal epoxide hydrolase gene in patients with anticonvulsant adverse drug reactions. Pharmacogenetics 1994; 4:142–53. 38. Green VJ, Pirmohamed M, Kitteringham NR, et al. Genetic analysis of microsomal epoxide hydrolase in patients with carbamazepine hypersensitivity. Biochem Pharmacol 1995; 50:1353–9. 39. Pirmohamed M. Genetic factors in the predisposition to drug-induced hypersensitivity reactions. AAPS J 2006; 8(1):E20–6. 40. Tybring G, von Bahr C, Bertilsson L, Collste H, Glaumann H, Solbrand M. Metabolism of carbamazepine and its epoxide metabolite in human and rat liver in vitro. Drug Metab Dispos 1981; 9:561–4. 41. Eichelbaum M, Thomson T, Tybring G, Bertilsson L. Carbamazepine metabolism in man. Induction and pharmacogenetic aspects. Clin Pharmacokinet 1985; 10:80–90. 42. Kroetz DL, Kerr BM, McFarland LV, Loiseau P, Wilensky AJ, Levy RH. Measurement of in vivo microsomal epoxide hydrolase activity in white subjects. Clin Pharmacol Ther 1993; 53:306–15. 43. Pirmohamed M, Kitteringham NR, Guenthner TM, Breckenridge AM, Park BK. An investigation of the formation of cytotoxic, protein-reactive and stable metabolites from carbamazepine in vitro. Biochem Pharmacol 1992; 43:1675–82. 44. Madden S, Maggs JL, Park BK. Bioactivation of carbamazepine in the rat in vivo. Evidence for the formation of reactive aren oxide(s). Drug Metab Dispos 1996; 24:469–79. 45. Lertratanangkoon K, Horning MG. Metabolism of carbamazepine. Drug Metab Dispos 1982; 10:1–10. 46. Lillibridge JH, Amore BM, Slattery JT, et al. Protein-reactive metabolites of carbamazepine in mouse liver microsomes. Drug Metab Dispos 1996; 24:509–14. 47. Furst SM, Sukhai P, McClelland RA, Uetrecht JP. Covalent binding of carbamazepine oxidative metabolites to neutrophils. Drug Metab Dispos 1995; 23(5):590–4. 48. Ju C, Uetrecht JP. Detection of 2-hydroxyiminostilbene in the urine of patients taking carbamazepine and its oxidation to a reactive iminoquinone intermediate. J Pharmacol Exp Ther 1999; 288:51–6. 49. Pirmohamed M, Kitteringham NR, Breckenridge AM, Park BK. The effect of enzyme induction on the cytochrome P450-mediated bioactivation of carbamazepine by mouse liver microsomes. Biochem Pharmacol 1992; 44:2307–14. 50. Kerr BM, Thummel KE, Wurden CJ, et al. Human liver carbamazepine metabolism. Role of CYP3A4 and CYP2C8 in 10,11-epoxide formation. Biochem Pharmacol 1994; 47:1969–79. 51. Pearce RE, Vakkalagadda GR, Leeder JS. Pathways of carbamazepine bioactivation in vitro I. Characterization of human cytochromes P450 responsible for the formation of 2- and 3-hydroxylated metabolites. Drug Metab Dispos 2002; 30(11):1170–9. 52. Pearce RE, Uetrecht JP, Leeder JS. Pathways of carbamazepine bioactivation in vitro: II. The role of human cytochrome P450 enzymes in the formation of 2-hydroxyiminostilbene. Drug Metab Dispos 2005; 33(12):1819–26. 53. Bernus I, Dickinson RG, Hooper WD, Eadie MJ. Early-stage autoinduction of carbamazepine metabolism in humans. Eur J Clin Pharmacol 1994; 47:355–60. 54. Bu HZ, Kang P, Deese AJ, Zhao P, Pool WF. Human in vitro glutathionyl and protein adducts of carbamazepine-10,11-epoxide, a stable and pharmacologically active metabolite of carbamazepine. Drug Metab Dispos 2005; 33(12):1920–4. 55. Leeder JS. Mechanisms of idiosyncratic hypersensitivity reactions to antiepileptic drugs. Epilepsia 1998; 39(Suppl. 7):S8–16. 56. Mauri-Hellweg D, Bettens F, Mauri D, Brander C, Hunziker T, Pichler WJ. Activation of drug-specific CD4C and CD8C T cells in individuals allergic to sulfonamides, phenytoin, and carbamazepine. J Immunol 1995; 155:462–72. 57. Friedmann PS, Strickland I, Pirmohamed M, Park BK. Investigation of mechanisms in toxic epidermal necrolysis induced by carbamazepine. Arch Dermatol 1994; 130:598–604.

502

Pirmohamed and Leeder

58. Park BK, Pirmohamed M, Kitteringham NR. Role of drug disposition in drug hypersensitivity: a chemical, molecular, and clinical perspective. Chem Res Toxicol 1998; 11(9):969–88. 59. Pichler WJ. Delayed drug hypersensitivity reactions. Ann Intern Med 2003; 139(8):683–93. 60. Wu Y, Sanderson JP, Farrell J, et al. Activation of T cells by carbamazepine and carbamazepine metabolites. J Allergy Clin Immunol 2006; 118(1):233–41. 61. Pirmohamed M, Lin K, Chadwick D, Park BK. TNF-alpha promoter region gene polymorphisms in carbamazepine hypersensitive patients Neurology 2001; 56:890–6. 62. Alfirevic A, Mills T, Harrington P, et al. Serious carbamazepine-induced hypersensitivity reactions associated with the HSP70 gene cluster. Pharmacogenet Genomics 2006; 16(4):287–96. 63. Chung WH, Hung SI, Hong HS, et al. Medical genetics: a marker for Stevens–Johnson syndrome. Nature 2004; 428(6982):486. 64. Lonjou C, Thomas L, Borot N, et al. A marker for Stevens–Johnson syndrome ethnicity matters. Pharmacogenomics J 2006; 6(4):265–8. 65. Hung SI, Chung WH, Jee SH, et al. Genetic susceptibility to carbamazepine-induced cutaneous adverse drug reactions. Pharmacogenet Genomics 2006; 16(4):297–306. 66. Isojarvi JIT, Pakarinen AJ, Rautio A, Pelkonen O, Myllyla VV. Liver enzyme induction and serum lipid levels after repacement of carbamazepine with oxcarbazepine. Epilepsia 1994; 35:1217–20. 67. Jensen NO, Dam M, Jakobsen K. Oxcarbazepine in patients hypersensitive to carbamazepine. Ir J Med Sci 1986; 155:297. 68. Beran RG. Cross-reactive skin eruption with both carbamazeepine and oxcarbazepine. Epilepsia 1993; 34:163–5. 69. Haruda F. Phenytoin hypersensitivity: 38 cases. Neurology 1979; 29:1480–5. 70. Mullick FG, Ishak KG. Hepatic injury associated with diphenylhydantoin therapy. A clinicopathological study of 20 cases. Am J Clin Pathol 1980; 74:442–52. 71. Park BK, Wilson AC, Kaatz G, Ohnhaus EE. Enzyme induction by phenobarbitone and vitamin K1 disposition in man. Br J Clin Pharmacol 1984; 18:94–7. 72. Aiges HW, Daum F, Olson M, Kahn E, Teichberg S. The effects of phenobarbital and diphenylhydantoin on liver function and morphology. J Pediatr 1980; 97:22–6. 73. Flowers FP, Araujo OE, Hamm KA. Phenytoin hypersensitivity syndrome. J Emerg Med 1987; 5:103–8. 74. Egerton-vernon JM, Fisk MJ, Snell AP. Phenytoin-induced hepatotoxicity. N Z Med J 1983; 96:467–9. 75. Masso JFM, Carrera N, Zabalza R, Bengoechea MG. Phenytoin (diphenylhydantoin) induced hepatotoxicity. Med Clin 1983; 81:320–1. 76. Smythe MA, Umstead GS. Phenytoin hepatotoxicity. A review of the literature. Drug Intell Clin Pharmcol 1989; 23:13–18. 77. Dreifuss FE, Langer DH. Hepatic considerations in the use of antiepileptic drugs. Epilepsia 1987; 28(Suppl. 2):S23–9. 78. Bryant AE, Dreifuss FE. Hepatotoxicity associated with antiepileptic drug therapy—avoidance, identification and management. CNS Drugs 1995; 4:13–9. 79. Spechler SJ, Sperber H, Doos WG, Koff RS. Cholestasis and toxic epidermal necrolysis associated with phenytoin sodium. The role of bile duct injury. Ann Intern Med 1981; 95:455–6. 80. Spielberg SP, Gordon GB, Blake DA, Goldstein DA, Herlong HF. Predisposition to phenytoin hepatotoxicity assessed in vitro. N Engl J Med 1981; 305:722–7. 81. Tomsick RS. The phenytoin syndrome. Cutis 1983; 32:535–7. 82. Vittorio CC, Muglia JJ. Anticonvulsant hypersensitivity syndrome. Arch Intern Med 1995; 155:2285–90. 83. Gennis MA, Vemuri R, Burns EA, Hill JV, Miller MA, Spielberg SP. Familial occurrence of hypersensitivity to phenytoin. Am J Med 1991; 91:631–4. 84. Spielberg SP. In vitro assessment of pharmacogenetic susceptibility to toxic drug metabolites in humans. Fed Proc 1984; 43:2308–13. 85. Kleckner HB, Yakulis V, Heller P. Severe hypersensitivity to diphenylhydantoin with circulating antibodies to the drug. Ann Intern Med 1975; 83:522–3. 86. Spielberg SP, Gordon GB, Blake DA, Mellits ED, Bross DS. Anticonvulsant toxicity in vitro: possible role of arene oxides. J Pharmacol Exp Ther 1981; 217:386–9. 87. Munns AJ, De Voss JJ, Hooper WD, Dickinson RG, Gillam EMJ. Bioactivation of phenytoin by human cytochrome P450: characterization of the mechanism and targets of covalent adduct formation. Chem Res Toxicol 1997; 10:1049–58. 88. Cuttle L, Munns AJ, Hogg NA, et al. Phenytoin metabolism by human cytochrome P450: involvement of P450 3A and 2C forms in secondary metabolism and drug–protein adduct formation. Drug Metab Dispos 2000; 28:945–50. 89. Komatsu T, Yamazaki H, Asahi S, et al. Formation of a dihydroxy metabolite of phenytoin in human liver microsomes/cytosol: roles of cytochromes P450 2C9, 2C19, and 3A4. Drug Metab Dispos 2000; 28:1362–8.

Anticonvulsant Agents

503

90. Kinobe RT, Parkinson OT, Mitchell DJ, Gillam EM. P450 2C18 catalyzes the metabolic bioactivation of phenytoin. Chem Res Toxicol 2005; 18(12):1868–75. 91. Allabi AC, Gala JL, Horsmans Y. CYP2C9, CYP2C19, ABCB1 (MDR1) genetic polymorphisms and phenytoin metabolism in a Black Beninese population. Pharmacogenet Genomics 2005; 15(11):779–86. 92. Tate SK, Depondt C, Sisodiya SM, et al. Genetic predictors of the maximum doses patients receive during clinical use of the anti-epileptic drugs carbamazepine and phenytoin. Proc Natl Acad Sci USA 2005; 102(15):5507–12. 93. Lee AY, Kim MJ, Chey WY, Choi J, Kim BG. Genetic polymorphism of cytochrome P450 2C9 in diphenylhydantoin-induced cutaneous adverse drug reactions. Eur J Clin Pharmacol 2004; 60(3):155–9. 94. Evans WE, Self TH, Weisburst MR. Phenobarbital-induced hepatic dysfunction. Drug Intell Clin Pharmcol 1976; 10:439–43. 95. Hyson C, Sadler M. Cross sensitivity of skin rashes with antiepileptic drugs. Can J Neurol Sci 1997; 24:245–9. 96. Bourgeois BFD. Valproic acid. Clinical use. In: Levy RH, Mattson RH, Meldrum BS, eds. Antiepileptic Drugs. 4th ed. New York: Raven Press Ltd, 1995:633–40. 97. Dreifuss FE. Valproic acid. Toxicity. In: Levy RH, Mattson RH, Meldrum BS, eds. Antiepileptic Drugs. 4th ed. New York: Raven Press Ltd, 1995:641–8. 98. Sussman NM, McLain LW. A direct hepatotoxic effect of valproic acid. J Am Med Assoc 1979; 242:1173–4. 99. Russo MW, Galanko JA, Shrestha R, Fried MW, Watkins P. Liver transplantation for acute liver failure from drug induced liver injury in the United States. Liver Transpl 2004; 10(8):1018–23. 100. Bjornsson E, Olsson R. Suspected drug-induced liver fatalities reported to the WHO database. Dig Liver Dis 2006; 38(1):33–8. 101. Dreifuss FE, Santilli N, Langer DH, Sweeney KP, Moline KA, Menander KB. Valproic acid hepatic fatalities: a retrospective review. Neurology 1987; 37:379–85. 102. Dreifuss FE, Langer DH, Moline KA, Maxwell JE. Valproic acid hepatic fatalities. II. U.S. experience since 1984. Neurology 1989; 39:201–7. 103. Bryant AE, Dreifuss FE. Valproic acid hepatic fatalities. III. U.S. experience since 1986. Neurology 1996; 46:465–9. 104. Kay JDS, Hilton-Jones D, Hyman N. Valproate toxicity and ornithine carbamoltransferase deficiency. Lancet 1986; ii:1283–4. 105. Eadie MJ, McKinnon GE, Dunston PR, McLaughlin D, Dickinson RG. Valproate metabolism during hepatotoxicity associated with the drug. Q J Med 1990; 284:1229–40. 106. Tokatli A, Coskun S, Cataltepe S, Ozalp I. Valproate-induced lethal hyperammonemic coma in a carrier of ornithine transcarbamylase deficiency. J Inherit Metab Dis 1991; 14:836–7. 107. Honeycutt D, Callahan K, Rutledge L, Wans B. Heterozygotic ornithine transcarbamylase deficiency presenting as symptomatic hyperammonemia durining initiation of valproate treatment. Neurology 1992; 42:666. 108. Zimmerman HJ, Ishak KG. Valproate-induced hepatic injury: analysis of 23 fatal cases. Hepatology 1982; 2:591–7. 109. Suchy FJ, Balistreri WF, Buchino JJ, et al. Acute hepatic failure associated with the use of sodium valproate. N Engl J Med 1979; 300:962–6. 110. Scheffner D, Ko¨nig S, Rauterberg-Ruland I, Kochen W, Hofmann WJ, Unkelbach S. Fatal liver failure in 16 children with valproate therapy. Epilepsia 1988; 29:530–42. 111. Kondo T, Kaneko S, Otani K, et al. Associations between risk factors for valproate hepatotoxicity and altered valproate metabolism. Epilepsia 1992; 33:172–7. 112. Anderson GD, Acheampong AA, Wilensky AJ, Levy RH. Effect of valproate dose on formation of hepatotoxic metabolites. Epilepsia 1992; 33:736–42. 113. Millington DS, Bohan TP, Roe CR, Yergey AL, Liberato DJ. Valproylcarnitine: a novel drug metabolite identified by fast atom bombardment and thermospray liquid chromatography-mass spectrometry. Clin Chim Acta 1985; 145:69–76. 114. Li J, Norwood DL, Mao L-F, Schulz H. Mitochondrial metabolism of valproic acid. Biochemistry 1991; 30:388–94. 115. Ponchaut S, Draye JP, Veitch K. In vitro effects of valproate and valproate metabolites on mitochondrial oxidations: relevance of CoA sequestration to the observed inhibitions. Biochem Pharmacol 1992; 43:2435–42. 116. Bjorge SM, Baillie TA. Inhibition of medium-chain fatty acid b-oxidation in vivo by valproic acid and its unsaturated metabolite, 2-n-propyl-4-pentenoic acid. Biochem Biophys Res Commun 1985; 132:245–52. 117. Kesterson JW, Granneman GR, Machinist JM. The hepatotoxicity of valproic acid and its metabolites in rats I. Toxicologic, biochemical and histopathologic studies. Hepatology 1984; 4:1143–52. 118. Granneman GR, Wang SI, Kesterson JW, Machinist JM. The hepatotoxicity of valproic acid and its metabolites in rats. II. Intermediary and valproic acid metabolism. Hepatology 1984; 4:1153–8.

504

Pirmohamed and Leeder

119. Kingsley E, Gray PD, Tolman KG, Tweedale R. The toxicity of the metabolites of sodium valproate in cultured hepatocytes. J Clin Pharmacol 1983; 23:178–85. 120. Rettenmeier AW, Prickett KS, Gordon WP, et al. Studies on the biotransformation in the perfused rat liver of 2-n-propyl-4-pentenoic acid, a metabolite of the antiepileptic drug valproic acid. Evidence for the formation of chemically reactive intermediates. Drug Metab Dispos 1985; 13:81–96. 121. Baillie TA. Metabolic activation of valproic acid and drug-mediated hepatotoxicity. Role of the terminal olefin, 2-n-propyl-4-pentenoic acid. Chem Res Toxicol 1988; 1:195–9. 122. Rettie AE, Rettenmeier AW, Howald WN, Baillie TA. Cytochrome P-450-catalyzed formation of D4-VPA, a toxic metabolite of valproic acid. Science 1987; 235:890–3. 123. Sadeque AJM, Fisher MB, Korzekwa KR, Gonzalez FJ, Rettie AE. Human CYP2C9 and CYP2A6 mediate formation of the hepatotoxin 4-ene-valproic acid. J Pharmacol Exp Ther 1997; 283:698–703. 124. Levy RH, Rettemeier AW, Anderson GD, et al. Effect of polytherapy with phenytoin, carbamazepine, and stirpentol on formation of 4-ene-valproate, a hepatotoxic metabolite of valproic acid. Clin Pharmacol Ther 1990; 48:225–35. 125. Donato MT, Go´mez-Lecho´n MJ, Castell JV. Effect of model inducers on cytochrome P450 activities of human hepatocytes in primary culture. Drug Metab Dispos 1995; 23:553–8. 126. Chang TKH, Yu L, Maurel P, Waxman DJ. Enhanced cyclophosphamide and ifosfamide activation in primary human hepatocyte cultures: response to cytochrome P-450 inducers and autoinduction by oxazaphosphorines. Cancer Res 1997; 57:1946–54. 127. Wrighton SA, Brian WR, Sari M-A, et al. Studies on the expression and metabolic capabilities of human liver cytochrome P450IIIA5 (HLp3). Mol Pharmacol 1990; 38:207–13. 128. Ho PC, Abbott FS, Zanger UM, Chang TK. Influence of CYP2C9 genotypes on the formation of a hepatotoxic metabolite of valproic acid in human liver microsomes. Pharmacogenomics J 2003; 3(6):335–42. 129. Waxman DJ, Azaroff L. Phenobarbital induction of cytochrome P-450 gene expression. Biochem J 1992; 281:577–92. 130. Leeder JS, Kearns GL. Pharmacogenetics in pediatrics. Implications for practice. Pediatr Clin North Am 1997; 44:55–77. 131. Shen DD, Pollack GM, Cohen ME, Duffner P, Lacey D, Ryan-Dudek P. Effect of age on the serum metabolite pattern of valproic acid in epileptic children. Epilepsia 1984; 25:674. 132. Gopaul S, Farrell K, Abbott F. Effects of age and polytherapy, risk factors of valproic acid (VPA) hepatotoxicity, on the excretion of thiol conjugates of (E)-2,4-diene VPA in people with epilepsy taking VPA. Epilepsia 2003; 44(3):322–8. 133. Kochen W, Schneider A, Ritz A. Abnormal metabolism of valproic acid in fatal hepatic failure. Eur J Pediatr 1983; 141:30–5. 134. Paganini M, Zaccara G, Moroni F, et al. Lack of relationship between sodium valproate-induced adverse effects and the plasma concentration of its metabolite 2-propylpenten-4-oic acid. Eur J Clin Pharmacol 1987; 32:219–22. 135. Tennison MB, Miles MV, Pollack GM, Thorn MD, Dupuis RE. Valproate metabolites and hepatotoxicity in an epileptic population. Epilepsia 1988; 29:543–7. 136. Siemes H, Nau H, Schultze K, et al. Valproate (VPA) metabolites in various clinical conditions of probable VPA-associated hepatotoxicity. Epilepsia 1993; 34:332–46. 137. Fromenty B, Pessayre D. Inhibition of mitochondrial b-oxidation as a mechanism of hepatotxicity. Pharmacol Ther 1995; 67:101–54. 138. Kassahun K, Farrell K, Abbott FS. Identification and characterization of the glutathione and N-acetylcysteine conjugates of (E)-2-propyl-2,4-pentadienoic acid, a toxic metabolite of valproic acid, in rats and humans. Drug Metab Dispos 1991; 19:525–35. 139. Kassahun K, Abbott FS. In vivo formation of the thiol conjugates of reactive metabolites of 4-eneVPA and its analogue 4-pentenoic acid. Drug Metab Dispos 1993; 21:1098–106. 140. Gopaul SV, Farrell K, Abbott FS. Identification and characterization of N-acetylcysteine conjugates of valproic acid in humans and animals. Drug Metab Dispos 2000; 28:823–32. 141. Pellock JM, Brodie MJ. Felbamate: 1997 update. Epilepsia 1997; 38:1261–4. 142. Pellock JM. Felbamate. Epilepsia 1999; 40(Suppl. 5):S57–62. 143. O’Neil MG, Perdun CS, Wilson MB, McGown ST, Patel S. Felbamate-associated fatal acute hepatic necrosis. Neurology 1996; 46:1457–9. 144. Thompson CD, Kinter MT, Macdonald TL. Synthesis and in vitro reactivity of 3-carbamoyl-2phenylproprionaldehyde and 2-phenylpropenal: putative reactive metabolites of felbamate. Chem Res Toxicol 1996; 9:1225–9. 145. Thompson CD, Gulden PH, Macdonald TL. Identification of modified atropaldehyde mercapturic acids in rat and human urine after felbamate administration. Chem Res Toxicol 1997; 10:457–62. 146. Thompson CD, Barthen MT, Hopper DW, et al. Quantification in patient urine samples of felbamate and three metabolites: acid carbamate and two mercapturic acids. Epilepsia 1999; 40:769–76.

Anticonvulsant Agents

505

147. Kapetanovic IM, Torchin CD, Strong JM, et al. Reactivity of atropaldehyde, a felbamate metabolite in human liver tissue in vitro. Chem Biol Interact 2002; 142(1–2):119–34. 148. Popovic M, Nierkens S, Pieters R, Uetrecht J. Investigating the role of 2-phenylpropenal in felbamate-induced idiosyncratic drug reactions. Chem Res Toxicol 2004; 17(12):1568–76. 149. Buchanan N. Lamotrigine: clinical experience in 200 patients with epilepsy with follow-up to four years. Seizure 1996; 5:209–14. 150. Dooley J, Camfield P, Gordon K, Camfield C, Wirrell E, Smith E. Lamotrigine-induced rash in children. Neurology 1996; 46:240–2. 151. Schlienger RG, Knowles SR, Shear NH. Lamotrigine-associated anticonvulsant hypersensitivity syndrome. Neurology 1998; 51:1172–5. 152. Makin AJ, Fitt S, Williams R, Duncan JS. Fulminant hepatic failure induced by lamotrigine. Br Med J 1995; 311:292. 153. Arnon R, DeVivo D, Defelice AR, Kazlow PG. Acute hepatic failure in a child treated with lamotrigine. Pediatr Neurol 1998; 18:251–3. 154. Overstreet K, Costanza C, Behling C, Hassanin T, Masliah E. Fatal progressive hepatic necrosis associated with lamotrigine treatment: a case report and literature review. Dig Dis Sci 2002; 47(9):1921–5. 155. Sachs B, Ronnau AC, von Schmiedeberg S, Ruzicka T, Gleichmann E, Schuppe HC. Lamotrigineinduced Stevens–Johnson syndrome: demonstration of specific lymphocyte reactivity in vitro. Dermatology 1997; 195:60–4. 156. Yuen AWC, Land G, Weatherley BC, Peck AW. Sodium valproate acutely inhibits lamotrigine metabolism. Br J Clin Pharmacol 1992; 33:511–3. 157. Brody MJ. Lamotrigine. Lancet 1992; 339:1397–400. 158. Maggs JL, Naisbitt DJ, Tettey JNA, Pirmohamed M, Park BK. Metabolism of lamotrigine to a reactive arene oxide intermediate. Chem Res Toxicol 2000; 13:1075–81.

24

Hepatotoxicity of Psychotropic Drugs and Drugs of Abuse Dominique Larrey

^ Service d’HØpato-GastroentØrologie et Transplantation, Hopital Saint Eloi, Montpellier, France

INTRODUCTION Adverse drug reactions affecting the liver represent an important challenge for physicians, health authorities, and pharmaceutical companies. A recent prospective inquiry performed in a liver and gastrointestinal unit revealed that 9% of admission was related to drug-induced adverse effects, liver toxicity representing the major problem (1). Despite improvements in toxicological studies and in the safety analysis of clinical trials, the early detection of drug hepatotoxicity remains very difficult. This view is reinforced by the fact that the overall frequency of hepatotoxicity for all drugs has not decreased in the last 10 years (2). Practically, all cells present in the liver can be affected by drugs (1,3–6). The types of lesions may also vary in their clinicopathological manifestations, according to the mechanism of hepatotoxicity, the drug itself, details of the treatment course (low or high dose and duration of treatment), and the susceptibility of the person taking the agent. This explains how drug hepatotoxicity can reproduce a wide range of “natural” (non-iatrogenic) liver diseases (3–6). Acute hepatitis is the most common syndrome. Acute liver failure is the most severe expression and represents the major cause of fatalities related to drugs (7,8). As a consequence, it is also the major cause of drug withdrawal from the pharmaceutical market. Besides the large number of “classical” drugs (more than 1200) reported to exhibit potential hepatotoxicity (2), other agents should also be taken into consideration. These include the excipients (9,10) present in formulations of drugs, herbal medicines which are increasingly consumed and often undisclosed (11–15), recreational and illegal compounds (16–18). These general aspects are particularly well illustrated by psychotropic agents. EPIDEMIOLOGICAL ASPECTS OF PSYCHOTROPIC HEPATOTOXICTY The prevalence and incidence of drug hepatotoxicity are still only partially known (1). The majority of data is provided by retrospective studies of database from pharmacovigilance centers and/or pharmaceutical companies, aimed to describe what are the most frequently hepatotoxic drugs and their clinical characteristics. These studies also have tried to estimate the prevalence of hepatotoxicity of a given drug by comparing the apparent number of hepatic adverse events after spontaneous reports and published cases and the consumption of the drug in the corresponding population (1). Because of this imperfect process, it is clear that many events remain unknown and that we are only aware of the “tip of the iceberg” (1). Finally, our knowledge of the incidence of drug-induced liver adverse effects in a general population is based on a single recent prospective study performed over three years in a population comprising 81,000 inhabitants (19). It was found to be 14/100,000 persons and The author of this chapter has relationships with the following corporations. Clinical research: Roche and Shering-Plough. Consulting agreements: Shering-Plough . Member of a Drug Safety Monitoring Board specializing in liver safety in products being tested by : Astra Zeneca, Bayer, Novartis, Intermune, Teva, Sanofi Aventis / Sanofi Synthelabo Recherche, Gsk, Merck Sante, Negma-Lerads, Pierre Fabre, Bial, Helsinn, and Cytheris Sponsored lecture agreements with Roche, Shering-Plough, Novartis, and Astrazeneca.

508

Larrey

corresponded to a number of events 16 times higher than that collected by pharmacovigilance centers. This study shows that one should be very careful in the interpretation of studies trying to compare the hepatotoxicity risks of different drugs, on the basis of spontaneously reported cases analyzed retrospectively. Epidemiological studies regarding drug-induced acute liver failure exhibit similar limitations. Retrospective studies suggest that drugs may have caused around 10% to 20% of all cases of fulminant and subfulminant hepatitis (20,21). Recently, the first prospective study that focused on the causes of acute liver failure has produced very interesting information (7). This study was performed on the basis of 308 consecutive patients admitted over a 41-month period to 17 tertiary care centers, in the United States (7). Paracetamol overdose was the most common apparent cause of acute liver failure, accounting for 39% of cases. Idiosyncratic drug reactions were the presumptive cause in 13% of cases, in the same range as viral hepatitis A and B combined, which was incriminated in 12% of cases (7). The large contribution of paracetamol to liver failure was further supported by the extension of this study, which included 662 patients studied over six years (22). These data strongly support that drugs have replaced viral hepatitis as the most frequent cause of acute liver failure. IDENTIFICATION OF DRUG-INDUCED LIVER INJURY CAUSED BY PSYCHOTROPIC AGENTS Overall, as for the large majority of drugs, there are generally no specific markers or tests for the diagnosis of liver injuries induced by psychotropic agents (1). Several methods with scoring systems have been developed to address this question as recently reviewed and are used by physicians and heath authorities in some countries and some pharmaceutical firms (23), but none is universally recognized. Consequently, the diagnosis largely relies on circumstantial evidence (1,24–31). Drug toxicity should be considered when there is a close time relationship between the beginning of a psychotropic treatment and the onset of liver injury. Other indications of drug toxicity are when liver injury is well known for the given drug and/or when the clinical presentation is suggestive of drug-induced liver injury. For instance, acute hepatocellular or cholestatic hepatitis may be associated with hypersensitivity reactions, such as fever, skin rash, or eosinophilia. Liver biopsy may further support this suspicion by showing an inflammatory infiltration mainly containing eosinophils. Rarely, the patient exhibits a relapsing hepatitis after accidental reexposure to the causative drug. There is a single instance in which drug hepatotoxicity with this group of drugs is associated with a specific marker: iproniazid for which liver injury, generally acute hepatocellular hepatitis, is associated with the presence of a particular autoantibody, antimitochondrial type 6 (1). This point will be developed with more details in the paragraph regarding this compound. In some circumstances, it may be particularly difficult to attribute liver injury to the intake of a psychotropic agent. The major difficulties are summarized in Table 1. The first difficulty is the absence of specificity of the clinical presentation so that many causes may be considered. The second difficulty is related to the profile of patients using psychotropic agents. Indeed, in this TABLE 1 Major Difficulties in the Diagnosis of Psychotropic Drug-Induced Liver Injuries Nonspecific clinical features in most cases Previous chronic liver disease, in particular past or still present alcoholic liver disease, chronic hepatitis related to infection by HCV, HBV, HBV HVD coinfection Patients with HIV/AIDS and already taking many treatments potentially toxic to the liver (e.g., antiretroviral, antibacterial, antituberculosis agents) Compounds considered safe (herbal remedies) Drug prescription difficult to ascertain Self-medication Undisclosed information (illicit compounds, e.g., cocaine, amphetamines) Forgotten information (elderly), especially for drugs indicated for memory dysfunction (tacrine, etc.) Drug or herbal products bought through Internet Fulminant and subfulminant hepatitis (insufficient time to assess chronology of reaction)

Hepatotoxicity of Psychotropic Drugs

509

population, some factors predispose to several major other liver diseases: more frequent excessive alcohol intake; more frequent abnormal food intake with risk of obesity or, in contrast, malnourishment; more frequent present or past IV or nasal drug addiction or sexual behavior predisposing to viral infections, such as hepatitis virus C, B, coinfection HBV–HDV. Furthermore, it is well known that patients with HIV infection or AIDS are more likely to require psychotropic drugs. Therefore, the prevalence of preexisting liver disease (acute and chronic alcoholic liver disease, nonalcoholic fatty liver diseases, chronic hepatitis, and cirrhosis) is higher before starting the treatment. Thus, this may lead to some diagnostic confusion when liver injury appears after the onset of treatment, particularly when baseline levels of liver tests are unknown. Furthermore, some patients are particularly predisposed to ingestion of other potentially hepatotoxic drugs. This point is well illustrated by patients with HIV infection who are treated by several antiretroviral, antituberculosis, and antibacterial agents. Another difficulty is related to the use of herbal medicines, all the more since automedication is frequent for this type of compound. Some are ingested to feel better and are not considered as “true drugs” so that the patient may consider that there is no risk. In case of liver event, such patients may not consider it important to inform the physician of this consumption. Sometimes, the information is undisclosed because the use of the compound is illegal (e.g., cocaine) or the drug is used in an inappropriate way (e.g., buprenorphine used by IV instead of the oral route) (32). The information may be omitted because of a disease associated with memory dysfunction, which does not allow the patient to describe its treatment correctly, e.g., elderly with Alzheimer’s disease (1). Another source of difficulty may be the sale of the drug or herbal products through the Internet, which does not allow control of the quality of the product (33). The final cause of difficulty is a patient already in fulminant or subfulminant hepatitis with encephalopathy so that it is impossible to question the patient about drug consumption. In this situation, an additional difficulty is the absence of the “dechallenge” criteria since the patient may be eligible for emergency liver transplantation or die of liver failure very quickly (1). FACTORS CONTRIBUTING TO HEPATOTOXICITY OF PSYCHOTROPIC AGENTS The risk of developing hepatotoxic drug reactions depends on two groups of factors. The first concerns the characteristics of the drug: how is it eliminated (by the liver or by other routes?); is it metabolized, and if so, by which enzymes; what type of hepatotoxic mechanisms are involved (1)? The second group of factors concerns the susceptibility of the individual and the context or environment of drug ingestion. Thus, genetic and acquired factors influencing drug hepatotoxicity can be distinguished. Acquired Factors Age above 60 years increases the risk for drug hepatotoxicity because of a higher consumption of drugs, and thereby a higher exposure to additional risks due to each drug and drug– drug interactions (Table 2). Quality of nutrition can affect hepatotoxicity in different ways. For instance, obesity may promote the hepatotoxicity of sulpiride. Chronic alcohol abuse may potentiate hepatotoxicity of amphetamines, probably by complex mechanisms that involve both the induction of critical enzymes such as cytochrome P450s and the lowered resistance to potential toxic metabolites because of glutathione depletion (1). Drug interactions can also contribute to psychotropic agent hepatotoxicity. Enzyme induction can lead to increased formation of toxic metabolites from another drug; for instance, enzyme induction by phenobarbital can induce the hepatotoxicity of antidepressants (1). Extrahepatic and hepatic diseases can also contribute to hepatotoxicity of some drugs (1). The role on psychotropic agents is not clearly illustrated.

510

Larrey

TABLE 2 Acquired and Genetic Factors Contributing to Hepatotoxicity of Psychotropic Agents Acquired factors Nutrition Chronic alcohol abuse Other drugs Disease

Genetic factors contributing to drug hepatotoxicity Deficiency in CYP 2D6 Deficiency in CYP 2C19 Deficiency in sulfoxidation Deficiency in glutathione S-transferase type T Deficiency in detoxication of reactive metabolites (unknown mechanisms) HLA system A11 DR6 DRB1 1S01 DROBL 0602

Obesity Enzyme induction HCV and HBV infections with liver fibrosis? Nonalcoholic - Steatohepatitis

Sulpiride? Amphetamine Phenobarbital antidepressants Sulpiride?

Perhexiline No correlation with amlneptine, amitriptyline and chlorpromazine Atrium (phenobarbital, febarbamate, and difebarbamate) Chlorpromazine? Tacrine? Amineptine Halothane, anti-convulsants, phenytoine, carbamazepine Tricyclic antidepressants, diclofenac, Halothane Chlorpromazine Amoxicillin-clavulanic acid

Genetic Factors Decreased expression of CYP2D6. Debrisoquine slow metabolizer phenotype or CYP2D6 deficiency is a common pharmacogenetic polymorphism present in 6% to 8% of Caucasian populations; it is transmitted as an autosomal recessive trait, with more than 10 mutations of the CYP2D6 gene (1,34). This deficiency is involved in perhexiline hepatotoxicity (34). However, it does not appear to be involved in the hepatotoxicity of amineptine metabolized by this enzyme (34) (Table 2). Deficiency of CYP2C19 may be involved in Atriumw hepatotoxicity (35). Atrium is a complex drug combining febarbamate, difebarbamate, and phenobarbital. A study comprising a small number of patients with a previous history of Atrium hepatitis showed that all of them had partial or complete deficiency in CYP2C19; the prevalence of this polymorphism in the control population is w3% to 5% in Caucasians and w20% in Asians (34,35). This example, however, needs to be confirmed by the study of a larger number of patients. Sulfoxidation deficiency was implicated in one study assessing patients with chlorpromazine hepatitis (36). However, the method used to analyze sulfoxidation is not reproducible, making this conclusion questionable (34). Further studies using reproducible tests for sulfoxidation polymorphism are required. Deficiencies in reactive metabolite detoxication. By using the in vitro lymphotoxicity assay developed by Spielberg et al. (36), several groups have shown deficiencies in the detoxification capacity of reactive metabolites in patients with drug-induced hepatitis (34). A deficiency could also be observed in some members of their family, suggesting a genetic defect. Such observations have been made for several drugs including anticonvulsants and psychotropic agents: phenytoin, carbamazepine, and amineptine (34,37,38). The precise defects involved in these mechanisms have still not been identified. Glutathione S-transferases. It has been proposed that tacrine hepatotoxicity could be linked to a deficiency in glutathione S-transferase type T (39). However, another study (40) did not confirm this. Genetic variations in the immune system could also be involved in some psychotropic drugs (41).

Hepatotoxicity of Psychotropic Drugs

511

ANTIDEPRESSANTS Tricyclic Antidepressants Tricyclic antidepressant compounds are commonly used. Hepatotoxicity has been well documented for at least three compounds: imipramine, amitriptyline, and particularly amineptine (3–6,42). These drugs require hepatic biotransformation to be eliminated, mainly by oxidative pathways involving cytochrome P450s (3,42). In addition, amineptine and tianeptine structures comprise a carboxylic chain that undergoes a mitochondrial beta-oxidation (3,42). Genetic variations in the metabolism of these drugs have been demonstrated, related to the role of CYP2D6 (34). Amineptine Amineptine, once largely prescribed in several European countries, has been responsible for the majority of liver injury caused by antidepressants (3,42). The most typical clinical form was an acute cholestatic or mixed pattern hepatitis associated with fever, abdominal pain, mimicking gallstone migration, occurring 4 to 12 weeks after starting the treatment (3,42). The frequently associated eosinophilia and the rapid relapse of hepatitis after reexposure to the drug strongly support an immunoallergic reaction. The prognosis of this liver injury has been very good, with recovery after the discontinuation of treatment (3,42). Several lines of evidence indicate that the drug is oxidized by CYP3A4 into reactive metabolites, which seem to be involved in an immunoallergic reaction (3,43,44). Rarely, acute hepatitis has been observed after massive overdose for a suicidal purpose (3,42). The clinical picture is that of acute toxic liver failure without immunoallergic symptoms (3,42). Additionally, amineptine can produce microvesicular steatosis as shown by liver biopsy (42). This lesion has been shown to be related to the inhibitory effect of amineptine on the mitochondrial fatty acid beta-oxidation (3,45). Tianeptine is chemically and metabolically very close to amineptine. Prospective experimental studies have shown that in both animal models and human tissue, the drug was metabolized to a reactive metabolite by CYP3A4, which could inhibit mitochondrial fatty acid beta-oxidation (3,46,47). This might be predictive of clinical liver reactions similar to those caused by amineptine. Interestingly, this was confirmed thereafter by the observation of two patients with cholestatic hepatitis with hypersensitivity symptoms and microvesicular steatosis as shown by biopsy (48,49). This is a very uncommon example showing the prediction of druginduced liver injury and its characteristics by prospective experimental studies. Imipramine The administration of imipramine is associated with a moderate and frequently transient elevation of alanine aminotransferase (ALT) in 20% of patients (3–6,42). The frequency of clinical hepatitis has been estimated to be around 0.5% to 1% in some reports (3–6,42). However, the number of published cases is rather low (42). The clinical picture is variable with acute hepatocellular, cholestatic, or mixed pattern hepatitis. Rare cases of fulminant hepatitis have been recorded (3–6,42). Cholangitis with prolonged cholestasis for several years has very rarely been observed (50,51). Clinical and histological features are suggestive of an immune-mediated reaction. Amitriptyline The administration of amitriptyline has been associated with a transaminase elevation in 10% of subjects and eosinophilia in 38% (52). The prevalence of clinical hepatitis seems to be low and reported cases are scant (3–5,42). The clinical and histological spectrum is rather broad: acute cholestatic, hepatocellular, or mixed pattern hepatitis (3–5,42). Rare cases of fulminant hepatitis have been described (3–5); less frequently, acute and chronic cholangitis have occurred (51,53). An immunoallergic mechanism has been suggested. Other Tricyclic Antidepressants Rare cases of acute hepatitis have been observed during the administration of other tricyclic antidepressants (3–6). Clomipramine by itself has also been implicated as a cause of hepatotoxicity (3–6). The listing of these drugs is shown in Table 3.

C

C

C C

C

C

C

C

FH

C

C

I

C

Cirrhosis

Steatosis

Steatosis Steatosis

AC, CC

Associated lesions

DDT, RM IA IA

C

CC CC C C C C

CC

IA, RM I I I I I I

C 0/C CC C C C C CC 0/C 0/C 0/C C 0/C

Causality

IA I IA DDT IA IA, RM RM IA, RM ? ? ? ? ?

Mechanisms

1 day to 1.8 years 1 week to 2 years 5 days to 1.5 years

5 days to 1 year 1 5 months 1 6 months 3 days to 2 months 12 days to 6 months 1 6 months

3 days to 1 year

1 week to 4 years 4 18 days 1 week to 10 months 2 days to 4 years 1 week to 1 months 1 week to 1 year 26 days to 6 months 0.5 8 months 6 weeks ? 6 weeks 2 4 weeks 2 3 months

Time to onset

Abbreviations: Cholest., cholestatic; FH, fulminant hepatitis; CH, chronic hepatitis; IA, immunoallergic reaction; DDT, dose-dependent toxicity; I, idiosyncratic reaction; RM, reactive metabolite; AC, acute cholangitis; CC, chronic cholangitis/ductopenia; MAOI, monoamine oxidase inhibitor; SSRI, selective serotonin reuptake inhibitor.

C

C

C C C

C C C C C

C C C C C

C

C

C

C C

C

C C

C

C

Mixed

C

Cholest.

CH

C C C C C C

C C C C C C

C C C C C C

Hepatocellular

Acute hepatitis

Characteristics of Liver Injury Induced by Antidepressants

Tricyclic antidepressants Clomipramine Amoxapine Amitriptyline Maprotiline Desipramine Imipramine Tianeptine Amineptine Demexiptiline Dibenzepine Dothiepine Iprindole Nortriptyline MAOI Iproniazide SSRI Fluoxetine Paroxetine Citralopram Sertraline Venlafaxine Fluvoxamine Other antidepressants Mianserine Nefazodone Trazodone

Drugs

TABLE 3

512 Larrey

Hepatotoxicity of Psychotropic Drugs

513

Cross-Sensitization Cross-sensitization between tricyclic antidepressants has been observed with amineptine and clomipramine (54), as well as between desipramine and imipramine (3). Another case report has outlined cross-sensitization between desipramine, trimipramine, and the antipsychotic cyamemazine (55). All three drugs have a similar molecular structure. Monoamine Oxidase Inhibitor Iproniazid Iproniazid has been withdrawn from the market of several countries including the United States, but is still used in many countries (3–6). This compound can lead to various types of liver injury. The most common is acute hepatocellular hepatitis typically occurring within the first three months of treatment (56). Less frequently, liver injury may occur up to four weeks after the discontinuation of the treatment (3–6). The majority of patients recovered but cases of fulminant hepatitis have been documented (3–6). Interestingly, iproniazid hepatotoxicity is usually associated with the presence of antimitochondrial antibody anti-M6 in the serum, which is different from that found in primary biliary cirrhosis (anti-M2) (57,58). This specific antibody acts as a serological diagnostic marker and is very useful for isoniazid causality assessment (57,58). Chronic liver injury mimicking autoimmune hepatitis has been reported. Iproniazid has been shown to be transformed into a reactive isopropyl radical (3). It is also activated to reactive metabolites by monoamine oxidase, which may account for the formation of the anti-M6 antibody (3). The pathogenic role of these antibodies needs to be documented. Nevertheless, an autoimmune contribution might explain the progression of liver injury despite the discontinuation of drug administration (58). Selective Serotonin Reuptake Inhibitor The selective serotonin reuptake inhibitors (SSRIs) now make up a major drug family for treating depression. The overall incidence of their side effects is estimated to be around 0.7% of exposed patients. Of these side effects, liver injuries have rarely been reported. Indeed, a recent review disclosed only less than 20 cases, comprising a few cases for fluoxetine (59–65), paroxetine (66–69), sertraline (70), and venlafaxine (71–73). This led to a recent retrospective study based on the data bank of French pharmacovigilance centers (74). One hundred fiftyeight cases of liver injury were collected, comprising 97 females and 61 males aged from 15 to 94 years (74). Liver injuries accounted for 11.8% of total side effects collected in the data bank for paroxetine, 13.1% for fluoxetine, 10.8% for citaprolam, 11.3% for sertraline, and 11.2% for fluvoxamine. The number of collected cases were as follows: paroxetine, 63 cases; fluoxetine, 45 cases; citaprolam, 30 cases; sertraline, 18 cases; and fluvoxamine, 2 cases. The spectrum of liver injury was as follows: hepatocellular injury, 65 cases; cholestatic injury, 45 cases; mixed pattern hepatitis, 10 cases; and transaminase elevation, 9 cases. Liver injury remained moderate in 75 cases, serious in 75 cases, associated with risk of fatality in 4 cases, and associated with death in 4 cases. The usual delay to onset of hepatitis ranged from one to six months. In most cases, liver injury occurred at therapeutic doses without associated hypersensitivity manifestations. Therefore, this suggests an idiosyncratic metabolic reaction. The underlying hepatotoxicity mechanisms remain unknown. Identified potential risk factors analyzed in the severe cases were alcohol abuse, age over 70 years, and coadministration of other potentially hepatotoxic drugs. This study strongly supports that SSRI hepatotoxicity is much more frequent than that previously reported. Nevertheless, it remains relatively low compared with the very large number of prescriptions of these compounds. The clinical picture is variable, acute hepatocellular hepatitis appearing to be the most frequent event. Chronic hepatitis has been reported in isolated cases associated with fluoxetine (65) and paroxetine (67,68). These cases are too scant to set a firm relationship between these compounds and this lesion.

514

Larrey

Other Antidepressants Mianserine Rare cases of hepatitis have been recorded after the administration of mianserine, a tetracyclic antidepressant, for a very variable duration from one day to more than 1.5 year but generally within 3 to 12 weeks (75–77). The clinical picture was mainly that of cholestatic or mixed hepatitis (75–77). One case was associated with a positive rechallenge (77). Hypersensitivity manifestations were suggestive of an immunoallergic reaction (75). A case of cross-sensitization with a tricyclic antidepressant, dothiepin, has been reported (78). Nefazodone About 15 reports of hepatitis have been recorded including three cases of subfulminant hepatitis (2,79,80). The delay to onset ranged from two days to one year (2). Trazodone Rare cases have been reported (2,75,81–83). One of them exhibited a fatal course. Trazodone administration was associated with other drugs, lithium and trifluoperazine (83). Other rare cases comprise jaundice (2,75,81) or chronic hepatitis (82). NEUROLEPTIC AGENTS Phenothiazines Chlorpromazine Chlorpromazine is the most representative compound of this family. This drug undergoes various biotransformations in the liver: the tricyclic ring can be sulfoxidized, hydroxylated, or N-oxidized after demethylation of the lateral chain (42). Chlorpromazine hepatotoxicity was described more than 50 years ago. Hundreds of cases have been reported which has lead to several general reviews (3–5,42,84). Chlorpromazine administration is associated with liver enzyme abnormalities in up to 42% of subjects (85). The frequency of liver injury has been estimated to be around 0.5% to 1% (5). The mean duration of treatment before the onset of first signs is usually around two to five weeks (5,42). A majority of cases are associated with acute cholestatic hepatitis with jaundice and pruritus (51,42,84). The prodromic phase is frequently marked by fever, abdominal discomfort, or even abdominal pain mimicking abdominal emergency (51,42,84). Liver test abnormalities mainly consist of biological cholestasis whereas transaminases are generally moderately increased or normal. Hypercholesterolemia is reported in 60% of cases (3–6,51,42). Eosinophilia appears particularly frequent with an incidence of 10% to 40% (5,85). Thrombocytopenia and agranulocytosis may be occasionally associated with liver injury (5). Antimitochondrial antibodies have been noted in a few patients (86). The suspicion of an immunoallergic reaction is further supported by the induction of lupus erythematosis and the association with antinuclear or anti-DNA antibodies in up to 40% of patients treated for more than six months (87). Histological findings are predominantly marked by centrilobular cholestasis with some hepatocyte necrosis and portal inflammation. Infiltration by eosinophils may be seen. Histological cholangitis is seen in 20% of cases (42,88). More rarely, portal triads contain inflammatory infiltration and/or discrete fibrosis. After discontinuation of the treatment, the course of liver injury is usually favorable with complete recovery within a couple of weeks. In some cases, the course is marked by the persistence of cholestasis for more than one year (5,51). This prolonged course has been reported to exhibit a prevalence of 7% (5,51). In most cases, this appears to be limited to moderate anicteric cholestasis (5,51). However, in other cases, the course is more severe with long-lasting jaundice of up to six years (5,51,89). Liver biopsy reveals cholestasis and chronic cholangitis (84). Later, marked paucity of bile ducts develops (5,52,78). Despite a concerning presentation, resembling primary biliary cirrhosis, the long-term prognosis may be favorable with late decrease or, even, disappearance of jaundice (5,51). Nevertheless, in about a quarter of the analyzed cases, biliary cirrhosis can develop with complications of portal hypertension and liver failure (5,51,89,90). Genetic susceptibility to chlorpromazine toxicity has also been evaluated (36,91). No correlation

Hepatotoxicity of Psychotropic Drugs

515

has been proved with deficiency in CYP2D6, which is involved in the metabolism of this compound (91). Another study postulated that individuals who were poor sulfoxidizers would be more susceptible to chlorpromazine-induced jaundice (36). However, no confirmation has been provided and the method used to assess sulfoxidation has raised some concerns (34). A promoting influence of HLA DR6 phenotype has been described (41). Other Phenothiazines Other phenothiazines have been incriminated in the occurrence of liver injuries characterized by clinical, biological, and histological features similar to those observed with chlorpromazine, but with a lower frequency (3–5,42). The implicated drugs are listed in Table 4. Cyamemazine Cyamemazine has rarely been associated with hepatotoxicity (5,51,92,93), one case being an overdose with subsequent pruritus and jaundice (92). Liver biopsy showed moderate PMN infiltration of portal tracts, including eosinophils, mild hepatocyte necrosis, and centrilobular cholestasis (92). The patient fully recovered. Thioridazine Thioridazine is associated with a very low incidence of jaundice and/or hepatitis; less than 10 cases have been reported (5,51,94–96). Trifluoperazine Very rare cases of jaundice have been reported (97–99). Cross-Sensitization Occasional cases of cross-sensitization have been recorded between phenothiazines, in particular between chlorpromazine and promazine (5,42,51). Another case report has outlined cross-sensitization between cyamemazine and the tricyclic antidepressants desipramine and trimipramine (55). All three drugs have a rather similar molecular structure containing a tricyclic moiety. Thioxanthenes Rare cases of cholestatic jaundice and moderate elevation of transaminases have been reported with chlorprothixene and clopenthixol (2–6,100,101). Benzamides Sulpiride Very rare cases of acute hepatitis followed by recovery have been reported within one week to three months after starting the treatment (102–104). Other compounds of this family, amisulpiride, sultopride, and tiapride do not appear to cause liver injury (2). Butyrophenones Haloperidol Very rare cases of acute hepatitis with jaundice have been observed with haloperidol (2–6, 105,106). Skin rash, fever, and hypereosinophilia were present in some cases (105,106). Very uncommon cases of cholangitis with prolonged cholestasis have been reported (51,106). Bromperidol Bromperidol is almost identical to haloperidol. One case has also been noted of reversible hepatitis with eosinophilia after a short duration of treatment (107).

IA IA I ? ? ? ? IA IA ?

Mechanisms

C

C

Steatosis

AC

CC

I I

IA

IA

IA IA

C 0/C

CC 0/C 0/C

0/C 0/C

C/0

C/0 C/0

CC C C C/0 C/0 C/0 C/0 C C/0 C/0

Causality

2 days to 1 year 4 weeks

1 8 weeks 3 weeks 12 days to 5 months

2 weeks to 5 months

1 week to 3 months

1 week ?

1 day to 14 months 2 days to 7 weeks 1 week to 1 month 2 weeks 1 month 2 months 1 week to 4 months 2 days to 2 months 3 4 weeks 2 weeks to 8 years

Time to onset

Abbreviations: Cholest., cholestatic; FH, fulminant hepatitis; CH, chronic hepatitis; IA, immunoallergic reaction; DDT, dose-dependent toxicity; I, idiosyncratic reaction; RM, reactive metabolite; AC, acute cholangitis; CC, chronic cholangitis/ductopenia.

C C

C

C C C C

C C

C

CC

CC

AC, CC AC, CC

Associated lesions

I

C

Cirrhosis

C C

C

C

I

? ?

C

C

C

C

C

FH

C

Mixed

C C

C C C C C C C C C C

Cholest.

CH

C

C C

C C C

Hepatocellular

Phenothiazines Chlorpromazine Cyamemazine Thioridazine Fluphenazine Metopimazine Perphenazine Prochlorperazine Promazine Thioproperazine Trifluoperazine Thioxanthenes Chlorprothixene Chlopenthixol Benzamides Sulpiride Butyrophenones Haloperidol Bromperidol Diazepines oxazepines Clozapine Loxapine Olanzapine Other neuroleptic agents Risperidone Molindone

Drugs

Acute hepatitis

TABLE 4 Characteristics of Liver Injury Induced by Neuroleptic Agents

516 Larrey

Hepatotoxicity of Psychotropic Drugs

517

Diazepines and Oxazepines Clozapine Clozapine is a frequent cause of transient liver test abnormalities (108–110). In an investigation comparing hepatic tolerance of clozapine with that of haloperidol, 37.3% of patients with clozapine had elevated ALT versus 16.6% with haloperidol (105). In contrast, clinically significant hepatotoxicity is uncommon. There are reported cases of clozapine-induced hepatocellular hepatitis (110–115). In one case, drug rechallenge resulted in recurrence of the abnormality (110). One case of hepatocellular damage occurred in a patient with chronic hepatitis C (113). A case of fatal fulminant hepatic failure has been reported (115). Cholestasis with eosinophilia can also be observed (116). The usual delay to onset of hepatitis varies from one to eight weeks (114). Loxapine Loxapine has been implicated in one case of hepatocellular injury during the first three weeks of therapy (6). Olanzapine Olanzapine is associated with transient, asymptomatic increase in hepatic enzymes (117). Rarely, olanzapine has been implicated in clinical hepatic toxicity (114,117). The delay to onset varies from 12 days to 5 months (114). However, in one case, the onset was very late, up to three years after the beginning of treatment (117). The potential hepatotoxicity mechanism remains unknown. Other Neuroleptic Agents Risperidone Risperidone is a new neuroleptic. Several cases of liver injury have been ascribed to this drug (118–128). In some cases, the clinical picture is acute jaundice (118,119,122,125). Other reports have described steatosis and/or steatohepatitis associated with weight gain and obesity, and abnormal liver enzymes (120–123). These cases have lead to the recommendation to assess liver tests at baseline and carefully monitor them along with body weight for long-lasting treatment (128). Molindone Molindone, an antipsychotic drug with a unique structure, has been associated with asymptomatic transaminase elevations. It has also been documented by rechallenge as the cause of hepatocellular injury in a 17-year-old schizophrenic patient four weeks after the initiation of the drug (75). Other compounds such as aripiprazole and pimozide do not appear to be hepatotoxic (2). ANXIOLYTIC AGENTS Benzodiazepines Benzodiazepines have been used for several decades. Liver enzyme increase is rare (3–6,75). More significant hepatotoxicity appears very uncommon since only a very few cases have been reported (3–6,75). This comprises isolated reports of acute liver injury with alprazolam, camazepam, chlordiazepoxide, clorazepate, clotiazepam, diazepamflurazepam and triazolam, and recently bentazepam (75,129–141). The clinical pattern is frequently cholestatic. Nevertheless, alprazolam and diazepam have also been associated with hepatocellular hepatitis (75,136). The mechanism underlying liver damage remains unclear. Buspirone Hepatotoxicity is extremely uncommon. Two cases of liver injury have been recorded: one cholestatic and the other hepatocellular (2,75).

518

Larrey

OTHER DRUGS Fipexide Use of fipexide, a cognition activator, is proposed in asthenia and memory disorders. This drug has been incriminated in rare cases of liver injury, including three cases of fulminant hepatitis after two weeks to three months of treatment (75,142,143). Tacrine Tacrine, a reversible cholinesterase inhibitor, is used in the symptomatic treatment of Alzheimer’s disease. Careful monitoring of serum ALT levels and tolerance-dependent, stepwise escalation of the doses are recommended, because the drug increases ALT levels in approximately 50% of recipients (3–6,144). In most patients, the rise in ALT occurs abruptly, usually at about six weeks of treatment, without an upward drift in the preceding weeks (3–6). Most affected subjects have mild ALT increases, without clinical manifestations. This extremely high incidence of liver dysfunction suggests direct toxicity rather than immunoallergy. The weak base tacrine is taken up by mitochondria, where it may cycle back and forth across the mitochondrial inner membrane, uncoupling respiration and wasting energy without ATP production (145). These effects are initially compensated for by adaptative mitochondrial responses leading to enhanced respiratory rates. However, the drug accumulates within hepatic lysosomes, and exhibits time-dependent accumulation and toxicity in cultured hepatocytes. This may account for the delayed liver dysfunction in vivo. First-pass metabolism in the liver may spare other organs, explaining why the liver is selectively injured (145). After interruption of the treatment in patients with ALT values above three times the upper limit of normal, and recovery, a precocious rechallenge often produces an almost immediate recurrence of the increase in ALT (possibly due to the persistence of the drug in lysosomal stores and the more rapid achievement of hepatotoxic intrahepatic concentrations). The elevation in ALT is often lower on rechallenge than during the initial exposure to the drug, and most patients subsequently show normalization of their ALT levels despite continuation, and even escalation, of the dose (3–6,144). This noticeable exception to the general rule that rechallenge should be avoided further supports the direct toxicity of tacrine rather than a metabolite-mediated immunoallergic phenomenon. Although tacrine-induced liver injury remains mild in most patients, 2% of recipients develop ALT values greater than 20 times the upper limit of normal and a few liver biopsies have shown liver cell necrosis (144,146). Fever and rash are infrequent, but eosinophilia is present in 35% of those with ALT more than 20 times the upper limit of normal (3,144). Tacrine is transformed by cytochrome P450 1A2 into reactive metabolites (147). Through the immunization of some subjects, these metabolites might be involved in this severe form of hepatitis. Riluzole Riluzole is the first drug available to treat amyotrophic lateral sclerosis. Clinical trials were associated with elevation of ALT above three times the upper limit of normal in 12% of patients in a dose-dependent manner (2). In some cases, ALT returned to normal despite the continuation of the treatment. This frequency of abnormal liver tests contrasts with the scarcity of clinical events. Indeed, after more than 10 years of marketing, only five cases of liver injury have been reported (2,148). Hepatitis occurs generally within two to eight weeks. In one case, microvesicular steatosis was revealed by biopsy (148). The mechanism of riluzole hepatotoxicity remains unknown but appears to be dose dependent (2,148). Tolcapone Tolcapone is a reversible inhibitor of catechol-o-methyltransferase used as an adjunct to Parkinson’s disease treatment. Clinical trials were associated with an increased incidence of ALT above three times the upper limit of normal (1.3–3.7% of patients) (6). During postmarketing follow-up, four cases of liver failure were reported (6,149–151). The liver event occurs within the first six months of treatment. This led to the recommendation of careful enzyme monitoring during treatment (151).

Hepatotoxicity of Psychotropic Drugs

519

HERBAL MEDICINES Kava–Kava (Piper methysticum rhizoma) The kava–kava plant has been used for a long time in Pacific islands in a ceremonial setting and, more recently, in Germany and other European countries for its relaxing and anxiolytic properties (152). More than 70 cases of liver injury have been recorded, including 15 with liver failure leading to 11 transplants and 4 deaths (152–155). Generally, liver injury occurs three to six weeks after starting the ingestion of the product (152–155). Nevertheless, late occurrence beyond one year has been noted. Recent studies suggest various candidates for explaining the hepatotoxicity mechanisms. Among these candidates, kavalactones are suspected to be involved (152). These compounds are metabolized by cytochrome P450s, in particular CYP2D6 (152). Interestingly, two isolated cases suggest a possible association with the genetic deficiency of the CYP2D6 (156). This point deserves confirmation in a larger number of patients. An immune-mediated mechanism has been proposed (157), as well an inhibition of P-glycoprotein by the crude extracts and kavalactones (158). Finally, cytotoxic effects of flavokavain b present in the root have been suggested (159). Chewing kava leafs in Pacific islands has not been associated with toxicity, which suggests that the method of extraction (e.g., ethanol) in proprietary forms leads to enrichment in a toxic lactone. Cannabis The cannabis plant is used worldwide and smoked as tobacco. A recent study suggests that regular smoking may increase fibrosis in patients with chronic hepatitis C (160). Valerian Several cases of acute hepatitis have been reported in patients taking herbal preparations for relieving stress in the form of tablets containing various extracts of plants, in particular valerian and, possibly, skullcap (15). In one patient, extensive fibrosis and liver failure with encephalopathy were reported (15). However, there are no experimental data supporting the toxicity of valerian and the follow-up for the last five years does not confirm its hepatotoxicity (15).

ILLEGAL AND RECREATIONAL COMPOUNDS Cocaine Cocaine abuse is a worldwide problem with medical, social, economic, and legal issues. It has been reported that 22 million Americans have used cocaine at least once and 5 million used it regularly (161). Nonparenteral cocaine users exhibit mild liver enzyme elevation in 15% of hospitalized patients (3). Acute cocaine intoxication may lead to a severe syndrome characterized by fever, arterial hypertension, disseminated intravascular coagulation, renal failure, rhabdomyolysis, and severe liver injury (162). Acute hepatitis occurs within two days and is characterized by marked transaminase elevation and a liver lesion consisting of pericentral coagulative necrosis and peripheral microvesicular steatosis (162). A direct cocaine toxicity has been demonstrated, although hyperthermia and hypotension might contribute to the liver lesion (163). The mechanism of cocaine has been intensively investigated. Animal models have shown that cocaine toxicity is dose dependent and involves an oxidative reaction mediated by the cytochromes P450 (163). Cocaine is transformed by cytochrome P450 into norcocaine, which is further transformed to N-hydroxynorcocaine, norcocaine nitroxide, and the norcocaine nitrosonium ion. These metabolites may cause oxidative stress and lipid peroxidation in hepatocytes (163). In animals, toxicity is increased by inducers (phenobarbital and ethanol) and is prevented by inhibitors of cytochrome P450 (163). Male gender may promote cocaine hepatotoxicity (164).

520

Larrey

Amphetamines and Derivatives Hepatotoxicity of amphetamines and methamphetamine is uncommon and mainly appears to be related to hyperthermia. Ecstasy (3,4-methylenedioxymethamphetamine) is an illicit synthetic amphetamine derivative increasingly used as a recreational drug. Its use combined with vigorous exercise at all-night dance sessions can produce a syndrome reminiscent of that caused by cocaine intoxication with hyperthermia, hypotension, disseminated intravascular coagulation, rhabdomyolysis, acute renal failure, and death (18,165–167). Ecstasy-induced liver injury may also present independently, a few days to four weeks after ingestion of the drug, as mixed hepatitis in young patients (18,167,168). Ecstasy use should be considered specifically in all young patients with unexplained jaundice. Methylphenidate Some cases of hepatocellular injury have been observed with both oral and IV use of methylphenidate followed by recovery (169,170). Phencyclidine Phencyclidine, also named “angel dust,” is used for its potent psychedelic properties. It may result in hyperthermia, and thereby promotes acute hepatocellular hepatitis (6). Buprenorphine Buprenorphine, a morphine analog, has been used as a substitute drug to help patients stop drug addiction. When used by oral route at recommended doses, the risk of hepatotoxicity appears extremely low (171). Very rare cases have been reported (3,32,171). In contrast, when used by IV route, this compound is responsible for hepatotoxic reactions (171). Experimental studies argue for the involvement of mitochondria dysfunction as the possible mechanism (32). CONCLUSION All families of psychotropic agents can sometimes cause liver injury. However, considering the large consumption of these compounds, the incidence of liver damage is relatively low. The most common type of liver event is acute cholestatic or mixed hepatitis. The number of drugs causing liver failure is limited. The causality evaluation may be particularly difficult since the intake of the responsible compounds may be forgotten or masked by the patients. Noteworthy among the incriminated psychotropic compounds are herbal medicines and illegal substances, which further increase diagnostic difficulties. Psychotropic agents are among the main causes of cholangitis and ductopenia that mimic chronic biliary disease, but with a better prognosis. Particularly involved compounds include tricyclic antidepressants and phenothiazines. Another interesting pattern for prescribers is the observation of cross-sensitivity to hepatotoxicity between some compounds and even in different therapeutic families. Indeed, cross-sensitization leading to hepatotoxicity has been observed not only between tricyclic antidepressants or between phenothiazines, but also between these families and other drugs containing a tricyclic structure. Despite intensive investigation, hepatotoxicity mechanisms remain unknown for the majority of psychotropic agents. REFERENCES 1. Larrey D. Epidemiology and individual susceptibility to adverse drug reactions affecting the liver. Sem Liv Dis 2002; 22:145–55. 2. Biour M, Ben Salem C, Chazouilleres O, Grange JD, Serfaty L, Poupon R. Drug induced liver injury: fourteenth updated edition of the bibliographic database of liver injuries and related drugs. Gastroenterol Clin Biol 2004; 28:720–59. 3. Farrell OC. Drug-Induced Liver Disease. London: Churchill Livingstone, 1994. 4. Pessayre O, Larrey D, Biour M. Drug-induced liver injury. In: Bircher J, Benhamou JP, McIntyre N, Rizzetto M, Rode´s J, eds. Oxford Textbook of Clinical Hepatology. 2nd ed. Vol. 2. Oxford: Oxford University Press, 1999:1261–315.

Hepatotoxicity of Psychotropic Drugs

521

5. Zimmerman HJ. The Adverse Effects of Drugs and Other Chemicals on the Liver. 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins, 1999. 6. Kaplowitz N, DeLeve LD. Drug-Induced Liver Disease. New York (Basel): Marcel Dekker, 2003. 7. Ostapowicz G, Fontana RJ, Schiodt SV, et al. Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States. Ann Intern Med 2002; 137:947–54. 8. Larrey D, Pageaux GP. Drug-induced liver failure. Eur J Gastroenterol Hepatol 2005; 17:141–3. 9. Negro F, Mondardini A, Palmas E. Hepatotoxicity of saccharin. N EngI J Med 1994; 331:134–5. 10. Larrey D. Drug-induced liver diseases. J Hepatol 2000; 32:77–88. 11. Schuppan D, Jia JD, Brinkhaus B, Hahn EG. Herbal products for liver diseases: a therapeutic challenge for the new millennium. Hepatology 1999; 30:1099–104. 12. Seeff LB, Lindsay KL, Bacon BR, Kresina TF, Hoofnagle JH. Complementary and alternative medicine in chronic liver disease. Hepatology 2001; 34:595–603. 13. Stedman C. Herbal hepatoxicity. Sem Liv Dis 2002; 22:195–206. 14. Stickel F, Baumuller HM, Seitz K, Vasilakis D, Seitz G, Seitz H, Schuppan D. Hepatitis induced by kava (Piper methysticum rhizoma). J Hepatol 2003; 39:62–7. 15. Pageaux G, Larrey D. Alternative medecine, vitamins and natural hepatotoxins. In: Kaplowitz N, DeLeve LD, eds. Drug-Induced Liver Disease. New York (Basel): Marcel Dekker, 2003. 16. Selim K, Kaplowitz N. Hepatotoxicity of psychotrpic drugs. Hepatology 1999; 29:1347–51. 17. VanThiel DH, Perper JA. Hepatotoxicity associated with cocaine abuse. Recent Dev Alcoholism 1992; 10:335–41. 18. Jones AL, Simpson KJ. Mechanisms and management of hepatotoxicity in ecstasy (MDMA) and amphetamine. Aliment pharmacol Ther 1999; 13:129–33. 19. Sgro C, Clinard F, Ouazir K, et al. Incidence of drug induced hepatic injuries. A french populationbased study. Hepatology 2002; 36:451–5. 20. Bernuau J, Benhamou JP. Fulminant and subfulminant liver failure. In: Bircher J, Benhamou JP, McIntyre N, Rizzetto M, Rodes J, eds. Oxford Texbook of Clinical Hepatology. 2nd ed., Vol. 2. Oxford: Oxford University Press, 1999:1341–72. 21. Takahashi Y, Okuda K. Fulminant hepatitis in Japan: etiology and prognostic prediction. Chin J Gastroenterol 1993; 10:337–50. 22. Larson AM, Polson J, Fontana RJ, et al. Acute liver failure study group. Hepatology 2005; 42:1364–72. 23. Lee W, Larrey D, Olsson R, lewis JH, Keisu M, Auclert L. Sheth. Hepatic findings in long-term clinical trials of ximelagatran. Drug Saf 2005; 28:351–70. 24. Lee WM. Assessing causality in drug-induced liver injury. J Hepatol 2000; 33:1003–5. 25. Criteria of drug-induced liver disorders. Report of an International Consensus Meeting. J Hepatol 1990; 11:272–6. 26. Danan G, Benichou C. Causalty assessment of adverse reactions to drugs—I. A novel method based on the conclusions of international consensus meetings: application to drug-induced liver injuries. J Clin Epidemiol 1993; 46:1331–6. 27. Benichou C, Danan G, Flahault A. Causality assessment of adverse reactions to drug—II. An original model for validation of drug causality assessment methods: case reports with positive rechallenge. J Clin Epidemiol 1993; 46:1331–6. 28. Maria VAJ, Victorino RMM. Development and validation of a clinical scale for the diagnosis of druginduced hepatitis. Hepatology 1997; 26:664–9. 29. Aithal GP, Rawlins MD, Day CP. Clinical diagnostic scale: a useful tool in the evaluation of suspected hepatotoxic adverse drug reactions. J Hepatol 2000; 33:949–52. 30. Lucena I, Camargo R, Andrade J, Perez-Sanchez CJ, Sanchez De La Questa F. Comparison of two clinical scales for causality assessment in hepatotoxicity. Hepatology 2001; 33:123–30. 31. Kaplowitz N. Causality assessment versus guilt-by-association in drug hepatotoxicity. Hepatology 2001; 33:308–9. 32. Berson A, Fau D, Fornacciari R, et al. Mechanisms for experimental buprenorphine hepatotoxicity: major role of mitochondrial dysfunction versus metabolic activation. J Hepatol 2001; 34:261–9. 33. Cantrell FL. Look what I found! Poison hunting on eBay. Clin Toxicol 2005; 43:375–9. 34. Larrey D, Pageaux GP. Genetic predisposition to drug-induced hepatotoxicity. J Hepatol 1997; 26:12–21. 35. Horsmans Y, Lannes B, Pessayre D, Larrey D. Possible association between poor metabolism of mephenytoin and hepatotoxicity caused by Atriumw. A fixed combination preparation containing phenobarbital, febarbamate and difebarbamate. J Hepatol 1994; 21:1075–9. 36. Watson RGP, Olomu A, Clements D, Waring RH, Mitchell S, Elias E. A proposed mechanism for chlorpromazine jaundice—defective hepatic sulphoxidation combined with rapid hydroxylation. J Hepatol 1988; 7:72–8. 37. Spielberg SP. In vitro assessment of pharmacogenetic susceptibility to toxic drug metabolites in humans. Fed Proc 1984; 43:2308–13. 38. Larrey D, Berson A, Hebersetzer F, et al. Genetic predisposition to drug hepatotoxicity. Role in hepatitis causes by amineptine, a tricyclic antidepressant. Hepatology 1989; 10:168–73.

522

Larrey

39. Bequemont L, Lecoeur S, Simon T, Beaune P, Funck-Brentano C, Jaillon P. Glutathione S-transferase q genetic polymorphism might influence tacrine hepatotoxicity in Alzheimer’s patients. Pharmacogenetics 1997; 7:251–3. 40. De Sousa M, Pirmohamed M, Kitteringham NR, Woolf T, Park BK. No association between tacrine transaminitis and the glutathione transferase q genotype in patients with Alzheimers disease, Pharmacogenetics 1998; 8:353–5. 41. Berson A, Fre´neaux E, Larrey D, et al. Possible role of HLA in hepatotoxicity. An exploratory study in 71 patients with drug-induced idiosyncratic hepatitis. J Hepatol 1994; 20:336–42. 42. Geneve J, Larrey D, Pessayre D, Benhamou JP. Structure tricyclique des me´dicaments et he´patotoxicite´. Gastroenterol Clin Biol 1987; 11:242–9. 43. Geneve J, Larrey D, Letteron P, Descatoire V, Tinel M, Amouyal G, Pessayre D. Metabolic activation of the tricyclic antidepressant amineptine I. Cytochrome P-450-mediated in vitro covalent binding. Biochem Pharmacol 1987; 36:323–9. 44. Geneve J, Degott C, Letteron P, et al. Metabolic activation of the tricyclic antidepressant amineptine II. Protective role of glutathione against in vitro and in vivo covalent binding. Biochem Pharmacol 1987; 36:331–7. 45. Le Dinh T, Frenaux E, Labbe G, et al. Amineptine, a tricyclic antidepressant, inhibits the mitochondrial oxidation of fatty acids and produces microvesicular steatosis of the liver in mice. J Pharmacol Exp Ther 1988; 247:745–50. 46. Letteron P, Labbe G, Descatoire V, et al. Metabolic activation of the antidepressant tianeptine II. In vivo covalent binding and toxicological studies at sublethal doses. Biochem Pharmacol 1989; 38:3247–51. 47. Pessayre D, Mansouri A, Haouzi D, Fromenty B. Hepatotoxicity due to inhibition of mitochondrial dysfunction. Cell Biol Toxicol 1999; 15:367–73. 48. Le Bricquir Y, Larrey D, Blanc P, Pageaux GP, Michel H. Tianeptine—an instance of drug-induced hepatotoxicity predicted by prospective experimental. J Hepatol 1994; 21:771–3. 49. Balleyguier C, Sterin D, Ziol M, Trinchet JC. Acute mixed hepatitis caused by tianeptine. Gastroenterol Clin Biol 1996; 20:607–8. 50. Horst DA, Norman DG, Lecompte PM. Prolonged cholestasis and progressive hepatic fibrosis following imipramine therapy. Gastroenterology 1980; 79:550–4. 51. Larrey D, Erlinger S. Drug-induced cholestasis. Bailliere’s Clin Gastroenterol 1988; 2:423–52. 52. Holmberg MB, Jaussen B. A study of blood count and serum transaminase in prolonged treatment with amitriptyline. J New Drugs 1962; 2:361–5. 53. Larrey D, Amouyal G, Pessayre D, et al. Amitriptyline-induced prolonged cholestasis. Gastroenterology 1988; 94:200–3. 54. Larrey D, Rueff B, Pessayre D, Danan G, Algard M, Geneve J, Benhamou JP. Cross hepatotoxicity between tricyclic antidepressants. Gut 1986; 27:726–7. 55. Remy AJ, Larrey D, Pageaux GP, Desprez D, Ramos J, Michel H. Cross hepatotoxicity between tricyclic antidepressants and phenothiazines. Eur J Gastroenterol Hepatol 1995; 7:373–6. 56. Rosenblum LE, Korn RJ, Zimmerman HJ. Hepatocellular jaundice as a complication of iproniazid therapy. Arch Intern Med 1960; 105:583–93. 57. Homberg JC, Stelly M, Andreis I, Abuaf N, Saadoun F, Andre J. A new antimitochondrial antibody (anti-M6) in iproniazid-induced hepatitis. Clin Exp Immunol 1982; 47:93–103. 58. Danan G, Homberg JC, Bernuau J, Roche-Sicaud J, Pessayre D. Hepatite a` l’iproniazide. Inte´reˆt diagnostique d’un nouvel anticorps anti-mitochondrial, l’anti-M6. Gastroenterol Clin Biol 1983; 7:529–32. 59. Cooper GL. The safety of fluoxetine—an update. Br J Psychiatry 1988; 153:77–86. 60. Gram LF. Fluoxetine. N Engl J Med 1995; 332:960–1. 61. Capella D, Bruguera M, Figueras A, Lapone JR. Fluoxetine-induced hepatitis: why is postmarking surveillance needed? Eur J Clin Pharmacol 1999; 55:545–6. 62. Cai Q, Benson MA, Talbot TJ, Devadas G, Swanson HJ, Oison JL, Kirchner JP. Acute hepatitis due to fluoxetine therapy. Mayo Clin Proc 1999; 74:692–4. 63. Bobichon R, Bernard G, Mion F. Acute hepatitis during treatment with fluoxetine. Gastroenterol Clin Biol 1993; 17:406–7. 64. Friedenberg FK, Rothstein KD. Hepatitis secondary to fluoxetine treatment. Am J Psychiatry 1996; 153:580. 65. Johnston DE, Wheeler DE. Chronic hepatitis related to use of fluoxetine. Am J Gastroenterol 1997; 92:1225–6. 66. Helmchen C, Boerner RJ, Meyendorf R, Hegerl U. Reversible hepatotoxicity of paroxetine in a patient with major depression. Pharmacopsychiatry 1996; 29:223–6. 67. Benbow SJ, Gill G. Drug points: paroxetine and hepatotoxicity. Br Med J 1997; 314:1387. 68. deMan RA. Severe hepatitis attributed to paroxetine (Seroxat). Ned Tijdschr Geneeskund 1997; 141:540–2.

Hepatotoxicity of Psychotropic Drugs

523

69. Cadranel JF, DiMartino V, Cazier A, et al. Atrium and paroxetine-related severe hepatitis. J Clin Gastroenterol 1999; 28:52–5. 70. Hautekeete ML, Cole L, VanVlieberg H, Elewant A. Symptomatic liver injury probably related to sertraline. Gastroenterol Clin Biol 1998; 22:364. 71. Horsrnans Y, De Clercq M, Sernpoux C. Venlafaxine-associated hepatitis. Ann Intern Med 1999; 130:944. 72. Cardona X, Avila A, Castellanos P. Venlafaxine-associated hepatitis. Ann Intern Med 2000; 132:417. 73. Kim KY, Hwang W, Narendran R. Acute liver darnage possibly related to sertraline and venlafaxine ingestion. Ann Pharmacother 1999; 33:381–2. 74. Pinzani V, Peyriere H, Hillaire-Buys D, Pageaux GP, Blayac BP, Larrey D. Specific serotonine recapture inhibitor (SSRI) antidepressants: hepatotoxicity assessment in a large cohort in France. J Hepatol 2006; 44:S256. 75. Stricker BHCH. Drug-Induced Disorders. 2nd ed. Amsterdam: Elsevier, 1992. 76. Zarsky JP, Aubert H, Rachail M. Hepatotoxicity of new antidepressive drugs. A propos of a case. Gastroenterol Clin Biol 1983; 7:220–1. 77. Bazin N, Beaufils B, Feline A. A severe allergic reaction to mianserin. Am J Psychiatry 1991; 148:1088–9. 78. Rasmussen S, Quedens JH. Possible cross hepatotoxicity between tricyclic and tetracyclic antidepressive agents. Ugeskr Laeger 1991; 153:3020–2. 79. Aranda-Michel J, Koehler A, Bejarano PA, et al. Nefazodone-induced liver failure: report of three cases. Ann Intern Med 1999; 130:285–8. 80. Eloubeidi MA, Gaede JT, Swaim MW. Reversible nefazodone-induced liver failure. Dig Dis Sci 2000; 45:1036–8. 81. Chu AG, Gunsolly B, Summers RW, et al. Trazadone and liver toxicity. Ann Intern Med 1983; 99:128. 82. Beck PL, Bridges RJ, Demetrick DJ, et al. Chronic active hepatitis associated with Trazadone therapy. Ann Intern Med 1993; 118:791. 83. Hull M, Jones R, Bendall M. Drug points: fatal hepatic necrosis associated with trazodone and neuroleptic drugs. Br Med J 1994; 309:378. 84. Ishak KG, Irey NS. Hepatic injury associated with the phenothiazines. Ann Intern Med 1972; 93:283–304. 85. Dickes R, Schenker V, Deutsch L. Serial liver function and blood studies in patients receiving chlorpromazine. N Engl J Med 1957; 256:1–7. 86. Rodriguez M, Paronotto F, Schaffner F, Popper H. Antimitochondrial antibodies in jaundice following drug administration. JAMA 1968; 208:148–50. 87. Alarcon-Segovia D, Fishbein E, Cetina JA, Raya RJ, Bancra E. Antigenic specificity of chlorpromazine-induced antinuclear antibodies. Clin Exp Immunol 1973; 15:543–8. 88. Zelman S. Liver cell necrosis in chlorpromazine jaundice (allergic cholangitis). A serial study of 26 needle biopsy specimens in nine patients. Am J Med 1959; 27:708–28. 89. Read AE, Harrisson CV, Sherlock S. Chronic chlorpromazine jaundice with particular reference to its relationship to primary biliary cirrhosis. Am J Med 1961; 31:249–58. 90. Walker CD, Combes B. Biliary cirrhosis induced by chlorpromazine. Gastroenterology 1966; 51:253–62. 91. Larrey D, Tinel M, Amouyal G, et al. Genetically determined oxidation polymorphisn and drug hepatotoxicity. Study of 51 patients. J Hepatol 1989; 8:158–64. 92. Cadranel JF, Bonnard P, Cazier A, Di Martino V, Pras V, Devergie B, Biour B. Cyamemazine-induced acute hepatitis after unique massive intake: a case report. Eur J Gastroenterol Hepatol 1999; 11:451–3. 93. Rager P, Cosculluela D, Deviers D. Drug hepatitis: possible role of cyamemazine. Presse Med 1983; 12:1941. 94. Baraacik M, Brandbor LL, Albion MJ. Thorazine-induced of the cholestasis. JAMA 1967; 200:69–70. 95. Urberg M. Thioridazine-induced non-icteric hepatotoxicity. J Fam Pract 1990; 30:342–3. 96. Weiden PL, Buckner CD. Thioridazine toxicity. Agranulocytosis and hepatitis with encephalopathy. JAMA 1973; 224:518–20. 97. Smaga M. On the etiology of hepatitis developing during treatment with Stelazine. Vrachebnoe Delo 1967; 10:126–8. 98. Margulies AI, Berris B. Jaundice associated with administration of trifluoperazine. Can Med Assoc J 1968; 98:1063–4. 99. Dimova S, Koleva M, Rangelova D, Stoythchev T. Effective nifedipine, verapamil, diltiazem and trifluoperazine on acetaminophen toxicity in mice. Arch Toxicol 1995; 70:112–8. 100. Ruddock DGS, Hoenig J. Chlorprothixene an obstructive jaundice. Br Med J 1973; 1:231. 101. Yaryura-Tobias JA, Woipert A, White L, et al. A clinical evaluation of clopenthixol. Curr Ther Res 1970; 12:271. 102. Ohmoto K, Yamamoto S, Hirokawa M. Symptomatic primary biliary cirrhosis triggered by administration of sulpiride. Am J Gastroenterol 1999; 94:3660–1. 103. Sarfraz A, Cook M. Sulpiride-induced cholestalic jaundice. Aust NZ J Psychiatry 1996; 30:701–2.

524

Larrey

104. Villari D, Rubino F, Corica F, Spinella S, Di Cesare E, Longo G, Raimonde G. Bile ductopenia following therapy with sulpiride. Virchows Arch 1995; 427:223–6. 105. Fuller CM, Yassinger S, Donlon P, Imperato TJ, Ruebner B. Haloperidol-induced liver disease. West J Med 1977; 127:515–8. 106. Dincsoy HP, Saelinger DA. Haloperidol-induced chronic liver disease. Gastroenterology 1982; 83:694–700. 107. Van Bellinghen M, Peuskens J, Appelmans A. Hepatotoxicity following treatment with bromperidol. J Clin Psychopharmacol 1989; 9:389–90. 108. Lieberrnan JA. Maximizing clozapine therapy: managing side effects. J Clin Psychiatry 1998; 59(Suppl. 3):38–43. 109. Hummer M, Kurz M, Kurzthaler I, Oberhauer H, Miller C, Fleischhacker WW. Hepatotoxicity of clozapine. J Clin Pharmacol 1997; 17:314–7. 110. Markowitz JS, Grinberg R, Jackson C. Marked liver enzyme elevations with clozapine. J Clin Psychopharmacol 1987; 17:70–1. 111. Thatcher GW, Cates M, Bair B. Clozapine-induced toxic hepatitis. Am J Psychiatry 1995; 152:296–7. 112. Kellner M, Wiedernann K, Krieg JC, Berg PA. Toxic hepatitis by clozapine treatrnent. Am J Psychiatry 1983; 150:985–6. 113. Worrall R, Wilson A, Cullen M. Dystonia and drug-induced hepatitis in a patient clozapine. Am J Psychiatry 1995; 152:647–8. 114. Dumortier G, Cabaret W, Stamatiadis L, et al. Hepatic tolerance to atypical antipsychotic drugs. Enceplale 2002; 28:542–51. 115. MacFarlane B, Davies S, Mannen K, et al. Fatal acute fulminant liver failure due to clozapine: a case report and review of clozapine induced hepatoxicity. Gastroenterology 1997; 112:170–7. 116. Thompson J, Chengappa KNR, Good CB, Baker RW, Kiewe RP, Bezner J, Schooler NR. Hepatitis, hyperglycernia, pleural effusion, eosinophilia, hematuria and proteinuria occurring early in clozapine treatment. Int Clin Psychopharmacol 1998; 13:95–8. 117. Ozcanli T, Erdogan A, Ozdemir S, Onen B, Ozmen M, Doksat K, Sosuz A. Severe liver enzyme elevations after three years of olanzapine treatment: a case report and review of olanzapine associated hepatotoxicity. Prog Neuropsychopharmacol Biol Psychiatry 2006; 30:1163–6. 118. Fuller MA, Simon MR, Freedman L. Risperidone-associated hepatotoxicity. J Clin Psychopharmacol 1986; 16:84–5. 119. Phillips EJ, Liu BA, Knowles SR. Rapid onset of risperidone-induced hepatotoxicity. Ann Pharmacother 1998; 32:843. 120. Benazzi F. Risperidone-induced hepatotoxicity. Pharmacopsychiatry 1998; 31:241. 121. Kumra S, Herion D, Jacobsen LK, Briguglia C, Grothe D. Case study: risperidone-induced hepatotoxicity in pediatric patients. J Am Acad Child Adolesc Psychiatry 1997; 36:701–5. 122. Landau J, Martin A. Is liver function monitoring warranted during risperidone treatment? J Am Acad Child Adolesc Psychiatry 1998; 37:1007–8. 123. Llinares Tello F, Hernandez Prats C, Bosacoma Ros N, Perez Martinez E, Climent Grana E, Navarro Polo JN, Ordovas Baines JP. Acute cholestatic hepatitis probably associated with risperidone. J Psychiatry Med 2005; 35:199–205. 124. Holtmann M, Kopf D, Mayer M, Bechtinger E, Schmidt MH. Risperidone-associated steatohepatitis and excessive weight-gain. Pharmacopsychiatry 2003; 36:206–7. 125. Krebs S, Dormann H, Muth-Selbach U, Hahn EG, Brune K, Schneider HT. Risperidone-induced cholestatic hepatitis. Eur J Gastroenterol Hepatol 2001; 13:67–9. 126. Szigethy E, Wiznitzer M, Branicky LA, Maxwell K, Findling RL. Risperidone-induced hepatotoxicity in children and adolescents? A chart review study J Child Adolesc Psychopharmacol 1999; 9:93–8. 127. Geller WK, Zuiderwijk PB. Risperidone-induced hepatotoxicity? J Am Acad Child Adolesc Psychiatry 1998; 37:246–7. 128. Perry R, Pataki C, Munoz-Silva DM, Armenteros J, Silva RR. Risperidone in children and adolescents with pervasive developmental disorder: pilot trial and follow-up. J Child Adolesc Psychopharmacol 1997; 7:167–79. 129. Noyes R, DuPont RL, Pecknold JC, et al. Alprazolam in panic disorder and agoraphobia: results from a multicenter trial. Arch Gen Psychiatry 1988; 45:423–8. 130. Roy-Byrne P, Vittone BM, Uhde TW. Alprazolam-related hepatotoxicity. Lancet 1983; 2:786. 131. Kratzsch KH, Buttner W, Reinhardt G. Intrahepatic cholestasis following chlordiazepoxide-contribution to the differential diagnosis of drug jaundice. Zeitschr Gesamte lnnere Med Jhre Grenzgebiete 1972; 27:408–11. 132. Lo LU, Eastwood JR, Eidelman S. Cholestatic jaundice associated with chlordiazepoxide hydrochloride (librium) therapy. Am J Dig Dis 1967; 12:845–9. 133. Tedesco FJ, Milis LR. Diazepam (valium) hepatitis. Dig Dis Sci 1982; 27:470–2. 134. Fang MH, Ginsberg AL, Dobbins W. Cholestatic jaundice associated with flurazepam hydrochloride. Ann Intern Med 1978; 89:363–4.

Hepatotoxicity of Psychotropic Drugs

525

135. Cobden J, Record CO, White RWB. Fatal intrahepatic cholestasis associated with triazolam. Postgrad Med J 1981; 57:730. 136. Parker JLW. Potassium clorazepate (tranxene)-induced jaundice. Postgrad Med J 1979; 55:908. 137. Moulin CH, Roiachon A, Cohard M, et al. Fulminant hepatitis secondair to alprazolam. Therapie 1994; 49:362–3. 138. Hebersetzer F, Larrey D, Babany G, Degott C, Corbie M, Pessayre D, Benhamou JP. Clotiazepaminduced acute hepatitis. J Hepatol 1989; 9:256–9. 139. Roberts RK, Wilkinson GR, Branch RA, Schenker S. Effect of age and parenchymal liver disease on the disposition and elimination of chlordiazepoxide (librium). Gastroenterology 1978; 75:479–85. 140. Wilkinson GR. The effects of liver disease and aging on the disposition of diazepam, chlordiazepoxide, oxazepam and lorazepam in man. Acta Psychiatric Scand 1978; 274:56–74. 141. Andrade RJ, Lucena MI, Aguilar J, et al. Dig Dis Sci 2000; 45:1400–4. 142. Mion F, Descos L, Ge´rard F, Vial T. Acute-drug-induced hepatitis: a case of acute cytolysis after ingestion of fipexide. Gastroenterol Clin Biol 1990; 14:513–4. 143. Durand F, Samuel D, Bernau J, et al. Fipexide induced fulminant hepatitis report of three cases with emergency liver transplantation. J Hepatol 1992; 15:144. 144. Watkins P, Zimmerman H, Knapp M, Gracon S, Lewis K. Hepatotoxic effects of tacrine administration in patients with Alzheimer’s disease. JAMA 1994; 271:992–8. 145. Berson A, Renault S, Letteron P, Robin MA, Fromenty B, Fau D, LeBot MA. Uncoupling of rat and human mitochondria: a possible explanation for tacrine-induced liver dysfunction. Gastroenterology 1996; 110:1878–90. 146. Hammel P, Larrey D, Bernuau J, et al. Acute hepatitis after tetrahydroaminoacridine administration for Alzheimer’s disease. J Ciin Gastroenterol 1990; 12:329–31. 147. Madden S, Woolf T, Pool W, Park BK. An investigation into the formation of stable, protein reactive and cytotoxic metabolites from tacrine in vitro. Biochem Pharmacol 1993; 46:13–20. 148. Remy A, Camu W, Ramos J, Blanc P, Larrey D. Acute hepatitis after riluzole administration. J Hepatol 1999; 30:527–30. 149. Assai F, Spahr L, Hadengue A, Rubbia-Brandt L, Burkhard PR, Rubbici-Brandt L. Tolcapone and fulminant hepatitis. Lancet 1998; 352:958. 150. New Warnings for Parkinson’s Drug Tasmar. Rockville, MD: Food and Drug Administration, November 16, 1998. 151. Olanow CW. Tolcapone and hepatotoxic effects: Tasmar Advisory Panel. Arch Neurol 2000; 57:263–7. 152. Peyrin-Biroulet L, Barraud H, Petit-Laurent F, et al. Hepatotoxicite´ de la phytothe´rapie: donne´es cliniques, biologiques, histologiques et me´canismes en cause pour quelques exemples caracte´ristiques. Gastroenterol Clin Biol 2004; 28:540–50. 153. Campo JV, Mc Nabb J, Perel JM, Mazariegos GV, Hasegawa SL, Reyes J. Kava-induced fulminant hepatic failure. J Am Acad Child Adolesc Psychiatric 2002; 41:631–2. 154. Humberston CL, Akhtar J, Krenzelok EP. Acute hepatitis induced by kava kava. J Toxicol Clin Toxicol 2003; 41:109–13. 155. Thiolet C, Mennecier D, Bredin C, et al. Cytolyse aigue¨ apre`s prise d’un the´ chinois. Gastroenterol Clin Biol 2002; 26:939–40. 156. Russmann S, Lauterburg BH, Helbling A. Kava hepatotocicity. Ann Intern Med 2001; 137:68–9. 157. Musch E, Chrissafidou A, Malek M. Acute hepatitis due to kava–kava and St John’s Wort: an immune-mediated mechanism? Dtsch Med Wochenschr 2006; 131:1214–7. 158. Weiss J, Sauer A, Frank A, Unger M. Extracts and kavalactones of Piper methysticum G. Forst (kava– kava) inhibit P-glycoprotein in vitro. Drug Metab Dispos 2005; 33:1580–3. 159. Jhoo JW, Freeman JP, Heinze TM, et al. In vitro cytotoxicity of nonpolar constituents from different parts of kava plant (Piper methysticum). J Agric Food Chem 2006; 54:3157–62. 160. Hezode C, Roudot-Thoraval F, Nguyen S. Daily cannabis smoking as a risk factor for fibrosis progression in chronic hepatitis C. Hepatology 2004; 40(Suppl. 1):192A. 161. Kothur R, Marsh F, Posner G. Liver function tests in nonparenteral cocaine users. Arch Intern Med 1991; 151:1126–8. 162. Silva MO, Roth D, Reddy KR, Fernandez JA, Albores-Saavedra J, Schiff ER. Hepatic dysfunction accompanying acute cocaine intoxication. J Hepatol 1991; 12:312–5. 163. Mallat A, Dhumeaux D. Cocaine and the liver. J Hepatol 1991; 12:275–8. 164. Visalli T, Turkall R, Abdel-Rahman MS. Influence of gender on cocaine hepatotoxicity in CF-1 mice. Int J Toxicol 2005; 24:43–50. 165. Brown C, Osterloh J. Multiple severe complications from recreational ingestion of MDMA (Ecstasy). JAMA 1987; 258:780–1. 166. Garbino J, Henry JA, Mentha G, Romand JA. Ecstasy ingestion and fulminant hepatic failure: liver transplantation to be considered as a last therapeutic option. Vet Hum Toxicol 2001; 43:99–102. 167. Nunez O, Banares R, Barrio J, Menchen L, Diego A, Salinero E, Clemente G. Variability of the clinical expression of Ecstasy-induced hepatotoxicity. Gastroenterol Hepatol 2002; 25:497–500.

526

Larrey

168. Dykhuizen RS, Brunt PW, Atkinson P, Simpson JG, Smith CC. Ecstasy induced hepatitis mimicking viral hepatitis. Gut 1995; 36:939–41. 169. Mehta H, Murray B, Loludice TA. Hepatic dysfunction due to intravenous abuse of rnethylphenidate hydrochioride. J Clin Gastroenterol 1984; 6:149–51. 170. Stecyk O, Loludice TA, Demeter S, Jacobs J. Multiple organ failure resulting from intravenous abuse of methylphenidate hydrochloride. Ann Emerg Med 1985; 14:597–9. 171. Herve S, Riachi G, Noblet C, et al. Acute hepatitis due to buprenorphine administration. Eur J Gastroenterol Hepatol 2004; 16:1033–7.

25

Antibacterial and Antifungal Agents Richard H. Moseley

Ann Arbor Veterans Affairs Healthcare System and Department of Internal Medicine, University of Michigan Health System, Ann Arbor, Michigan, U.S.A.

INTRODUCTION Antibacterial and antifungal agents are among the most frequently prescribed medications, yet clinically significant liver injury is a relatively rare event. In addition, the assessment of the hepatotoxicity of antibacterial and antifungal agents is confounded by the effects and the severity of the underlying infection being treated. A cholestatic liver injury pattern is a welldescribed feature of sepsis, in part, reflecting adaptive responses in bile acid transport proteins that serve to protect the liver from the effects of bile acid retention and facilitate extrahepatic routes of bile acid excretion (1,2). Nevertheless, there are well-described hepatic reactions to antibacterial and antifungal agents, and the incidence of clinically significant liver injury associated with the use of these agents has been estimated from several different surveys. Using Medicaid billing data from Michigan and Florida from the years 1980 to 1987, the number of patients developing acute symptomatic liver disease resulting in hospitalization from a 10-day course of sulfonamides, erythromycin, and tetracycline was determined to be 4.8, 2.28, and 1.56 cases for each million patients treated, respectively (3). In a French study of 81,301 patients who could not go elsewhere for medical care, 6 out of 34 cases of reported drug-induced liver injury over a three-year period were attributed to antibiotics (four to amoxicillin–clavulanic acid, one to cloxacillin, and one to amoxicillin) (4). In a population-based case-control study using the U.K.based General Practice Research Database as the source of information that examined 1,636,792 persons aged 5 to 75 years old registered in the database from 1 January, 1994, to 31 December, 1999, and followed up for a total of 5,404,705 person-years, strong associations with acute and clinically relevant drug-induced liver injury were found for amoxicillin–clavulanic acid [adjusted odds ratio (AOR); 95% confidence interval (CI)Z94.8; 27.8, 323], flucloxacillin (AORZ17.7; 4.4, 71.0), macrolides (AORZ6.9; 2.3, 21.0), and tetracyclines (AORZ6.2; 2.4, 15.8), exceeded only by chlorpromazine (5). In a review of 4039 outpatients referred to a community-based hepatology practice in the United States, antibiotics (amoxicillin–clavulanic acid, minocycline, nitrofurantoin, telithromycin, trimethoprim–sulfamethoxazole (TMP/SMX), and trovafloxacin) were the class of drugs most frequently implicated in self-limited acute druginduced liver injury (6). The incidence of liver injury associated with antibacterial and antimycotic use in a hospitalized patient population was found to be 1.5% and 1.1%, respectively (7). Antimicrobial agents were the most frequently implicated drugs in an analysis from Spain of reports of drug-induced liver injury; amoxicillin–clavulanic acid accounted for 12.8% of the reports (8). Finally, flucloxacillin and TMP/SMX were among the five most common drugs associated with a fatal outcome in a Swedish analysis of drug-induced liver injury (9,10). The majority of hepatic reactions associated with antibacterial and antifungal agents are idiosyncratic. High doses have only been reported as a risk factor for liver injury from the tetracyclines and the oxypenicillins. Richard H. Moseley served as a consultant and an expert witness for law firms representing Pfizer Inc as a defendant in troglitazone (Rezulin) litigation. He has no current relationship with the pharmaceutical industry.

528

Moseley

ANTIBACTERIAL AGENTS Beta-Lactam Antibiotics Penicillins Liver injury from the penicillins as a class is extremely rare, asymptomatic, and more frequently displays a hepatocellular rather than a cholestatic pattern (11). There are limited reports of liver injury associated with the use of the natural penicillins, benzylpenicillin, and phenoxymethyl penicillin (11–15). Both hepatocellular and cholestatic patterns of injury have been reported. Similarly, ampicillin is rarely associated with liver injury. Two case reports describe cholestatic injury attributed to ampicillin; one patient had a protracted course in association with ductopenia (16,17). Cholestasis has also been attributed to ampicillin in combination with the b-lactamase inhibitor, sulbactam (18). In a retrospective cohort study in the United Kingdom, the incidence of reported acute liver injury associated with amoxicillin was 0.3 (0.02–0.5) per 10,000 prescriptions (19). A patient with prolonged cholestasis with ductopenia has also been reported with amoxicillin (20). Cholestatic liver injury has rarely been reported with nafcillin, a penicillinase-resistant semisynthetic penicillin (21–24). Mild anicteric hepatitis and liver biopsy findings of spotty liver cell necrosis and no cholestasis have been described with the use of high doses of carbenicillin (25). Cholestatic hepatitis is the predominant injury pattern reported with the use of the oxypenicillins, dicloxacillin, and cloxacillin, although granulomatous hepatitis has also been described in a patient treated with dicloxacillin (26–33). Liver injury related to oxacillin is also considered rare. When reported, both hepatocellular and cholestatic injury patterns have been described (34–42). Oxacillin use in children was associated with a higher incidence of hepatotoxicity (22%) when compared with the use of nafcillin and other intravenous antimicrobials (43). Reported risk factors for oxacillin-associated liver injury include high-dose intravenous therapy (O6 g/day) and HIV infection (44). Reduced scavenging by glutathione of the hydroxylamine derivative of the isoxazole ring of oxacillin in the setting of systemic glutathione deficiency was proposed as a possible explanation for the greater susceptibility of HIV-infected patients to oxacillin-related liver injury (44). Rapid recurrence of liver injury upon (inadvertent) rechallenge, the presence of serum and tissue eosinophilia, and the results of immunoallergic tests, including mast-cell degranulation and macrophage inhibition factor tests, suggests that liver injury associated with the oxypenicillins is an immunoallergic process, albeit idiosyncratic, given the rarity of liver injury, the absence of a true dose-dependency, and the significant variability in time before onset of liver dysfunction (30,45). However, cloxacillin has also been shown to inhibit ATP-dependent canalicular bile acid transport and exhibits a greater inhibitory effect on transport in human canalicular membrane vesicles when compared with rats (46). These findings suggest the possible involvement of the canalicular bile acid transporter, BSEP or ABCBII, in cloxacillinassociated cholestasis. From numerous case reports and series (47–51), a well-defined clinical picture has emerged for liver injury associated with flucloxacillin, a penicillinase-resistant semisynthetic penicillin used outside of the United States as an alternative to dicloxacillin for the oral antibiotic treatment of soft tissue infections caused by Staphylococcus aureus. Liver injury typically presents with prolonged painless jaundice and elevation of cholestatic liver enzymes, within two to six weeks of beginning treatment but as much as three weeks after the drug is stopped. Ursodeoxycholate treatment was reported to be effective in the management of two patients with flucloxacillin-related cholestasis (52). Although most patients eventually recover within several months, a vanishing bile duct syndrome has been described (20,49,53,54), and fatal cases have also been reported (9,53,55). Two population-based epidemiological studies using the U.K.-based General Practice Research Database estimated the risk of cholestatic liver disease within 45 days after first-time use of flucloxacillin at about 7 in 100,000 patients (56,57). Similar incidences were reported from Sweden (1:11,000–1:30,000 prescriptions) (29) and Australia (1:15,000–1:26,000) (58). Female sex, age, and high daily doses

Antibacterial and Antifungal Agents

529

seemed to be associated with higher risk of liver reactions from flucloxacillin (29). In response, the use of flucloxacillin was restricted in 1994 by the Australian Department of Human Services and Health. In the United Kingdom, only a warning was issued. In a recent cohort study, again using data from the U.K. General Practice Research Database, 283,097 and 131,189 patients were identified with a first-time prescription for flucloxacillin or oxytetracycline, respectively, from 1992 to 2002. The risk of cholestatic liver disease per 100,000 first-time users was 8.5 (95% CI 5.4, 12.6) during the 1 through 45 days and 1.8 (95% CI 0.6, 4.1) during the 46 through 90 days after starting flucloxacillin, and 0.8 (95% CI 0.02, 4.3) during the 1 through 45 days after starting oxytetracycline (59). In addition, despite published warnings, the frequency of first-time use of flucloxacillin was found to have remained stable between 1991 and 2000 (59). Like the other oxypenicillins, the mechanism for the apparent idiosyncratic immunoallergic injury associated with flucloxacillin is unknown. Hepatocytes, mainly via CYP3A4 activity, have been found to generate flucloxacillin metabolite(s), including 5 0 -hydroxymethylflucloxacillin, which may induce cytotoxicity in susceptible biliary epithelial cells (60). In addition, flucloxacillin treatment in rats has been shown to result in the formation of hepatic protein adducts that may be immunogenic (61). Liver injury related to amoxicillin–clavulanic acid, the orally (and parenterally, outside of the United States) administered combination of a semisynthetic penicillin and b-lactamase inhibitor, has been extensively described since the first case reports in 1989 (62–66). A cholestatic or mixed hepatocellular–cholestatic pattern of injury predominates, although a hepatocellular pattern of injury and granulomatous hepatitis have also been described (67). Use of amoxicillin–clavulanic acid is associated with a higher incidence of acute hepatic injury than the use of amoxicillin alone. In a retrospective cohort study in the United Kingdom, the incidences of reported acute liver injury associated with the combination of amoxicillin and clavulanic acid and amoxicillin alone were 1.7 (1.1–2.7) and 0.3 (0.02–0.5) per 10,000 prescriptions, respectively (19). Hepatic dysfunction is seen primarily in elderly patients, with a male preponderance, and after multiple courses of amoxicillin–clavulanic acid (19,68,69). The risk of developing acute liver injury was more than three times greater after a course of two or more consecutive prescriptions than after a single course of therapy; the combination of age above 65 years and repeated prescriptions resulted in a risk of developing acute liver injury greater than 1 per 1000 users of amoxicillin–clavulanic acid (19). The onset of signs and symptoms can be delayed from several days to several weeks and may occur during or several days after cessation of therapy. Features of hypersensitivity, including fever, rash, arthralgias, and eosinophilia, occur in up to 60% of patients. Complete resolution of physical and biochemical abnormalities generally occurs in a few weeks to few months (usually within six weeks). However, amoxicillin–clavulanic acid therapy has been associated with severe liver injury (70), and acute liver failure, vanishing bile duct syndrome (20,71,72), and deaths have been reported (73,74). Although the incidence of amoxicillin–clavulanic acid-related hepatotoxicity in the pediatric age group is rare, vanishing bile duct syndrome has also been reported in children (75,76). Ursodeoxycholate therapy may play a role in the management of patients with prolonged cholestasis from amoxicillin–clavulanic acid (76,77). Although the mechanism for amoxicillin–clavulanic acid-induced hepatitis is unknown, a significant association with the DRB1*1501–DRB5*0101–DQB1*0602 human leukocyte antigen (HLA) haplotype has been reported (78,79). These data support the view that an immunologic idiosyncrasy, mediated through HLA class II antigens, may play a role in the pathogenesis of amoxicillin–clavulanic acid-induced hepatitis. However, given the frequency of this HLA haplotype in normal controls (w12%) and the frequency of amoxicillin clavulanic acidinduced hepatitis (w0.1–0.01%), other factors must also be involved. The clavulanic acid component of the combination is thought to be responsible for the liver injury (80). There have been a few reports of liver injury from the ureidopenicillins (mezlocillin, azlocillin, and piperacillin) (81–84). Other Beta-Lactam Antibiotics The frequency of abnormal liver chemistries associated with the use of the carbapenems, meropenem, imipenem, and ertapenem, and the monobactam, aztreonam, is similar to

530

Moseley

the frequencies associated with other antibiotics in clinical trials. There is a single case report of cholangiopathy possibly associated with imipenem–cilastatin use (84). Cholestatic liver injury has rarely been reported with cephalosporin use (85–88). The mechanism appears to be a hypersensitivity reaction similar to liver injury from penicillin antibiotics (86). Biliary sludge formation has been associated with ceftriaxone, a semisynthetic third-generation cephalosporin. Analysis of the biliary concretions induced by this agent reveals a calcium salt of ceftriaxone (89). Macrolide Antibiotics Erythromycin has long been associated with liver injury. Of the several ester derivatives formulated to promote stability or enhance intestinal absorption of the erythromycin base, the erythromycin estolate was initially considered to be solely responsible for hepatotoxicity (90,91). However, other erythromycin derivatives, including erythromycin ethylsuccinate, erythromycin propionate, and erythromycin stearate, were all subsequently found to be associated with liver injury (92–96). Two studies support the conclusion that liver injury associated with erythromycin does not significantly differ among the various derivatives. In the first study, using prescription-event monitoring in 12,208 patients, three cases of jaundice were attributed to erythromycin stearate in contrast to no cases of jaundice in patients prescribed erythromycin estolate (97). In the second study, using data from a voluntary reporting system, the incidence of hepatotoxicity did not differ significantly between the various derivatives of erythromycin (98). Although there are no prospective trials, a retrospective cohort study, using data automatically recorded on general practitioners’ office computers in the United Kingdom, estimated the risk of liver injury associated with erythromycin to be in the range of 3.6 per 100,000 users (95% CI 1.9–6.1) (99). Subsequently, using Medicaid billing data from Michigan and Florida between 1980 and 1987, the number of patients developing acute symptomatic liver disease resulting in hospitalization from a 10-day course of erythromycin was determined to be 2.28 cases for each million patients treated (3). Liver injury related to erythromycin is predominantly cholestatic. The clinical picture may include abdominal pain, nausea, fever, which may resemble acute cholecystitis. Serum eosinophilia, occasional rash, rapid recurrence after rechallenge, and a lag period of 6 to 20 days between initiation of antibiotic and the onset of symptoms are consistent with an immunoallergic reaction; cross-reactivity between erythromycin estolate and ethylsuccinate has also been described (95). The onset of liver injury may occur one to two weeks after cessation of treatment, particularly when the duration of treatment is less than eight days (99). Features on liver biopsy include centrizonal cholestasis, in association with portal and lobular inflammation, presence of eosinophils, and mild hepatocellular necrosis (93). In most cases, recovery is complete after antibiotic withdrawal, although cases of ductopenia have been reported (100,101). In these rare cases, histopathology resembles primary biliary cirrhosis and positive antimitochondrial antibodies were reported in one case (102). One fatal case from parenteral erythromycin lactobionate has been described (103). In a recent analysis of all the reports of suspected hepatic adverse drug reactions received by the Swedish Adverse Drug Reactions Advisory Committee from 1970 to 2004, erythromycin was distinguished from other antibiotics by the absence of fatal cases (10). Serum bilirubin, aspartate and alanine aminotransferase (AST and ALT) levels were significantly less than other cases; age was significantly younger in the erythromycin cases as well (10). Cholestatic injury has also been reported with azithromycin, an erythromycin derivative that is a member of the subgroup of the macrolides known as azolides (104–106). Clarithromycin has the same macrolide, 14-membered lactone ring as erythromycin; the only difference is that, at position six, a methoxy group replaces the hydroxyl group (107). Cholestatic liver injury has been associated with clarithromycin, particularly in elderly patients on high doses of antibiotic (108,109). In one patient with a prior history of exposure to erythromycin, a progressive form of cholestatic injury developed leading to death (110). In addition, several cases of acute liver failure associated with clarithromycin have been reported (111–114).

Antibacterial and Antifungal Agents

531

Telithromycin, a member of the ketolide class of antimicrobials, has been associated with asymptomatic elevations in serum transaminases (115). Despite concerns voiced about its potential hepatotoxicity (116), significant liver injury associated with telithromycin had been reported, up until recently, only in one patient (6). However, severe hepatotoxicity in patients receiving telithromycin, resulting in death in one patient and liver transplantation in another, has now been described (117). Telithromycin is a substrate for P-glycoprotein and multidrug resistance–associated protein 2 (118), although the role, if any, of hepatobiliary transport in telithromycin-associated hepatotoxicity is not known. Several studies suggest a direct hepatotoxic effect of erythromycin, including erythromycin estolate inhibition of bile flow in the isolated perfused rat liver (119) and inhibition of canalicular bile acid efflux in regular and collagen-sandwich cultured human hepatocytes by erythromycin estolate and troleandomycin, but less so by roxithromycin, spiramycin, telithromycin, and erythromycin bases (120). These experimental findings have been linked with clinical findings consistent with an immunoallergic process that takes into account the induction by and metabolism of the macrolide antibiotics by cytochrome P450 isoforms, particularly CYP3A4 (121). Under this proposal, nitrosoalkanes formed from the demethylation and oxidation of macrolides not only react with glutathione, but also might covalently bind to the SH-groups of proteins. Direct hepatotoxicity from macrolides may produce enough hepatocyte necrosis to release hepatic proteins altered by the covalent binding of metabolites into the circulation. Such modified liver antigens may be recognized as foreign and trigger, in susceptible individuals, an immunoallergic type of clinical hepatitis (121,122). Sulfonamides Liver injury has been reported with several of the sulfonamides, either alone (sulfamethoxazole, sulfasalazine, sulfamethizole, and sulfamethoxypyridazine) or as part of a combination drug (TMP/SMX and pyrimethamine–sulfadoxine) (123). Hepatic injury associated with TMP/ SMX is particularly high in patients with AIDS (124). Cholestatic and mixed hepatocellular– cholestatic injury patterns predominate. Granulomatous hepatitis has rarely been described (125–127). Occasionally, severe liver injury has been associated with sulfonamide use, resulting in liver transplantation or death (9,128–131). Cases of prolonged cholestasis, with ductopenia and vanishing bile duct syndrome developing in several patients, have been associated with TMP/SMX use (132–136). Concentric arrangements of lamellar membranous structures in lysosomes in hepatocytes, resulting in the appearance of hepatic phospholipidosis, were a striking feature observed in one case of intrahepatic cholestasis associated with TMP/SMX use (137). The mechanism of liver injury from sulfonamides appears to be immunoallergic. Recurrent hepatitis, with a reduced latent period, has been reported after rechallenge (131,138). In addition, it has been proposed that susceptibility to a hypersensitivity reaction to sulfonamides might be enhanced by a slow acetylator phenotype, resulting in greater hepatic metabolism by CYP3A4 and 2C9 to reactive hydroxylamines (139). Given the low incidence of hepatic adverse events associated with sulfonamides, slow acetylation was never considered the sole factor in susceptibility, and recent studies in HIV-positive patients with sulfonamide hypersensitivity suggest a limited, if any, role for acetylator status (140,141). Tetracyclines Hepatotoxicity from tetracycline is largely of historical interest since the withdrawal of the intravenous form in the United States in 1991. High intravenous or oral doses greater than 2 g/day had been associated with clinical signs of nausea, vomiting, abdominal pain, and mild jaundice and biochemical abnormalities including elevations in serum AST (typically less than 500 IU/L) and serum amylase (142–144). Pregnancy and renal insufficiency increased the susceptibility to tetracycline-associated liver injury (145,146). Fatal outcomes were observed with high parenteral doses (147). Histopathologically, microvesicular steatosis with minimal necrosis is observed.

532

Moseley

The mechanism for liver injury appears to be the result of tetracycline inhibition of the mitochondrial b-oxidation of fatty acids (148). In addition, tetracycline, but not doxycycline, affects hepatic lipoprotein secretion by inhibiting the activity of microsomal triglyceride transfer protein, an endoplasmic reticulum protein that lipidates apolipoprotein B into triglyceride-rich very low-density lipoprotein particles (149). Liver injury from oral low-dose tetracycline appears to be extremely rare. An analysis of all hepatic adverse events reported to the Swedish Adverse Drug Reactions Advisory Committee from 1965 to 1995 estimated the incidence of liver injury from oral tetracycline to be approximately 1 in 18 million daily doses (150). In addition, the risk of developing acute symptomatic liver disease resulting in hospitalization from a 10-day course of tetracycline was estimated to be 1.56 cases per million (3). However, both cholestatic and hepatocellular patterns of injury have been reported (150). Tetracycline has also been implicated in two patients with prolonged cholestatic injury associated with paucity of bile ducts on liver biopsy (151). Doxycycline and minocycline, the semisynthetic derivatives of tetracycline widely used in the treatment of acne, have been associated with hepatic injury significantly different from the microvesicular steatosis observed with parenteral tetracycline. Cholestatic injury has been reported with doxycycline (152). Acute and chronic hepatitis with autoimmune features have been attributed to minocycline (153–162). Acute liver failure requiring liver transplantation has been described and several deaths have occurred in association with minocycline use (163–165). A systematic review of the literature conducted in 2000 identified 65 reported cases of hepatitis or liver damage in association with minocycline; 58% of cases occurred in females and 94% were under 40 years old (166). Minocycline-induced acute hepatitis presents with a hypersensitivity reaction in association with fever, eosinophilia, lymphadenopathy, and exfoliative dermatitis within 35 days of starting therapy (167,168). A cohort analysis performed in the United Kingdom demonstrated that the risk of acute liver injury associated with minocycline therapy for acne was very small, was not significantly greater than that associated with oxytetracycline/tetracycline therapy, and was greatest in the first month of use (169). Autoimmune hepatitis associated with lupus-like symptoms, including elevated serum gamma-globulin levels, the appearance of antinuclear antibodies, fever, and arthralgias, occurs after a median duration of exposure to minocycline of 365 days in females (nZ20) and 730 days in males (nZ9) (166). Features of autoimmune hepatitis in association with minocycline use are indistinguishable from type 1 autoimmune hepatitis; however, drug withdrawal should lead to rapid improvement and corticosteroid therapy is not required (170). Tigecycline is the first in a new generation of tetracyclines known as glycylcyclines. Because the Food and drug Administration (FDA) approved tigecycline only in June, 2005, there is limited postmarketing experience. In phase 3 studies of efficacy and safety, elevated serum AST and ALT were reported but at rates that were either the same or significantly less than imipenem–cilastatin- and vancomycin–aztreonam-treated controls, respectively (171,172). Quinolones Despite the widespread use of quinolones in the treatment of infections caused by variety of Gram-positive and Gram-negative organisms, particularly in patients with chronic liver disease, there are relatively few reports of significant hepatotoxicity related to the use of these antibiotics (173). Asymptomatic elevations in serum transaminases appear to be a class effect, but hepatocellular (174–176) and cholestatic forms of injury (177–181), including cases of ductopenia (182,183), acute liver failure and death (184–186) have all been reported. There may be a lag period of several weeks between the end of treatment and the onset of jaundice, which may hamper the diagnosis. In the cases of acute hepatitis associated with quinolones, liver biopsy reveals centrizonal necrosis and a mixed inflammatory infiltration containing abundant eosinophils (187), suggesting an immunoallergic mechanism. An exception to the safety profile of the quinolones, with respect to adverse liver effects, is trovafloxacin. Postmarketing surveillance and case reports (188–190) after this drug was approved in the United States in December 1997 revealed significant hepatotoxicity (out of 140 cases reported to the FDA, representing an incidence rate of adverse liver events of approximately 0.006%, 14 cases of acute liver failure, 4 requiring liver transplantation, and 5

Antibacterial and Antifungal Agents

533

liver-related deaths) that led to its withdrawal from European markets in June 1999 and significant restrictions on its use in the United States (191). Trovafloxacin hepatotoxicity may be the result of its difluorinated side chain that is not found in the other quinolones and that renders it highly lipophilic (188). In addition, Liguori et al. (192) have recently used microarray analysis on isolated human hepatocytes to understand the mechanisms underlying the idiosyncratic toxicity induced by trovafloxacin. Their results identified changes induced by trovafloxacin, but not other marketed quinolone agents, in genes that are involved in mitochondrial damage, RNA processing, transcription, and inflammation that may play a role in the hepatotoxicity associated with this antibiotic. Accumulating evidence from experimental models demonstrates that an episode of inflammation during drug treatment predisposes animals to liver injury (193–195), suggesting that inflammation could be a factor for idiosyncratic drug toxicity in humans. Waring et al. (196) tested this hypothesis in the case of trovafloxacin, demonstrating that the combination of bacterial lipopolysaccharide (LPS) and trovafloxacin resulted in hepatotoxicity in rats, which was not observed with either agent alone. In contrast, treatment with LPS and levofloxacin did not result in these changes. Microarray analysis identified unique changes induced by the combination of trovafloxacin and LPS, including enhanced expression of chemokines, suggestive of liver neutrophil polymorphonuclear leukocyte (PMN) accumulation and activation. The finding that depletion of PMNs with anti-PMN serum attenuated the liver injury supported a role for PMNs in the hepatotoxicity induced by the combination of LPS and trovafloxacin. It remains to be seen whether the gene expression changes identified by these and other studies can be used as biomarkers for idiosyncratic toxicity and improve our understanding of the mechanism(s) underlying hepatotoxicity associated not only with trovafloxacin but also with other drugs. Vancomycin Vancomycin, a glycopeptide antibiotic, is used in the oral treatment of Clostridium difficileassociated colitis and systemically for methicillin-resistant Staphylococcus aureus infections. Despite its increasing use, elevated serum AST and ALT levels have been reported in only one patient, to date (197). Nitrofurantoin The incidence of liver injury associated with nitrofurantoin, a synthetic nitrofuran most commonly used for the prophylaxis of urinary tract infections, has been estimated to be 0.0003% (198,199). Acute cholestatic and hepatocellular injury, granulomatous hepatitis, and chronic hepatitis with or without cirrhosis have all been reported with nitrofurantoin use (199–204). Acute liver injury typically manifests, along with fever, rash and eosinophilia, within one to six weeks of beginning nitrofurantoin (204). Chronic liver disease from nitrofurantoin use was typically observed in women taking this antibiotic for extended periods of time and, with increased recognition of this adverse effect, is rarely seen now. The clinical and histological features of chronic hepatitis associated with prolonged nitrofurantoin use resemble autoimmune hepatitis, including the presence of hypergammaglobulinemia, antinuclear and anti-smooth muscle antibodies, and interface hepatitis on liver biopsy (201,203–205). The clinical, biochemical, and histological improvement upon drug withdrawal in most cases distinguishes the two disorders (203,205–207). The risk of developing liver injury from nitrofurantoin increases with age and is higher in women than in men (208). The mechanism of liver injury associated with nitrofurantoin is presumed to be immunoallergic. In support of this are the presence of allergic manifestations in acute forms of liver injury and documentation of liver injury after rechallenge, even after a 17-year latent period (209). Cross-reactivity to furazolidine, a furan antibiotic closely related to nitrofurantoin, manifesting as cholestatic injury in a patient with jaundice during previous exposure to nitrofurantoin, has also been documented (210). Additional insight into the possible pathogenesis of nitrofurantoin-associated hepatotoxicity was provided by the recent finding that nitrofurantoin is excreted into bile by the breast cancer resistance protein (BRCP/ABCG2), a member of the ATP-binding cassette family of transporters (211). Analysis of hepatic expression

534

Moseley

of human BCRP also demonstrated a higher expression in men compared with women (212). Gender-specific and interindividual differences in the expression of BCRP in the liver may lead to variability in the pharmacokinetics of BCRP substrates, with potential impact on the susceptibility to hepatotoxicity. Rifampin The hepatic uptake of rifampin, a member of the rifamycin class of antibiotics, is mediated by members of the organic anion–transporting polypeptide (OATP) family of membrane transporters (213,214). As a result, rifampin inhibits the uptake of unconjugated bilirubin (215) and increases serum bile acid concentrations (216). Liver injury associated with rifampin is typically observed in patients receiving other antituberculous agents (see Chap. 26) and the hepatotoxic potential, if any, of this antibiotic is unclear. In fact, rifampin is used in the treatment of pruritus in patients with cholestatic liver disease, taking advantage of its ability to induce drugmetabolizing enzymes and transporters (217). However, three patients with primary biliary cirrhosis have been described, who developed hepatotoxicity when rifampin was given to treat their pruritus, necessitating liver transplantation in one patient (218). Clindamycin Hepatotoxicity associated with clindamycin has rarely been reported. A mixed hepatocellular– cholestatic injury pattern was observed in a patient with the severity of the underlying sepsis being a major confounding factor (219). Cholestatic injury with ductopenia in another report was most likely the result of concomitant TMP/SMX use (133). Daptomycin Daptomycin is a novel cyclic lipopeptide antibiotic recently approved for the treatment of complicated skin and skin structure infections caused by aerobic Gram-positive bacteria, including those caused by methicillin-resistant and -susceptible Staphylococcus aureus. Asymptomatic liver chemistry abnormalities have been described (220). However, the one reported case of possible daptomycin-related hepatotoxicity represented instead serum AST elevation due to a myopathy, a recognized adverse effect of this antibiotic (221). Oxazolidinones To date, there have been no reports of clinically significant hepatic injury associated with linezolid, a member of a new class of antibiotics, the oxazolidinones (222). Hepatic metabolism of this antibiotic with activity against a wide variety of Gram-positive bacteria, including those resistant to methicillin and vancomycin, occurs via morpholine ring oxidation, which is independent of the CYP450 enzyme system (222). Streptogramins In phase 3 trials of more than 2200 patients treated with dalfopristin–quinupristin, members of the streptogramin family of compounds isolated from Streptomyces pristinaespiralis, increases in total and conjugated bilirubin levels to greater than five times normal levels and elevations in liver enzyme levels were observed in 4% to 7%, respectively (223). However, based on an analysis of liver biopsy specimens, hyperbilirubinemia during quinupristin–dalfopristin therapy is likely multifactorial rather than antibiotic-related hepatotoxicity (224). ANTIFUNGAL AGENTS Amphotericin B Liver injury has rarely been associated with the use of amphotericin B. Of the three documented reports, that of a patient with cryptococcal meningitis treated with amphotericin B intermittently over a one-year period was, on autopsy, found to have marked centrilobular fatty

Antibacterial and Antifungal Agents

535

infiltration and congestion without inflammation that was confounded by the co-administration of chlorpropamide (225). Asymptomatic elevations in serum alkaline phosphatase, aminotransferase, and bilirubin levels after 18 days of amphotericin B treatment were described in a patient with acute myelogenous leukemia and fungal pneumonia that resolved with discontinuation of therapy and recurred with rechallenge (226). In the third case, a patient with pulmonary blastomycosis developed elevated liver chemistries in a hepatocellular injury pattern 10 days after amphotericin B was added to itraconazole therapy (227). Liver chemistry abnormalities resolved with discontinuation of the amphotericin B; liver biopsy demonstrated mild focal fatty change without acute or chronic inflammatory changes. In clinical trials, liposomal formulations of amphotericin resulted in mild elevations in serum AST and alkaline phosphatase levels (228). Although it was initially considered to be no greater for liposomal than for conventional amphotericin, severe liver injury was first reported in 2001 in a patient with acute myelogenous leukemia after repeat administration of liposomal amphotericin (229). A recent case-control analysis demonstrated that liposomal amphotericin B was associated with a substantial increase in the risk of hepatotoxicity in bone marrow transplant recipients, as defined by serum transaminase levels greater than three times ULN, that worsened with continued use; a smaller increase in risk was found for fluconazole and conventional amphotericin B carried no statistically significant risk (230). Although a strict definition of hepatotoxicity was chosen in this analysis, as noted in an accompanying editorial (231), elevations in serum transaminase levels greater than 10 times ULN were observed in 26% and 12% and serum bilirubin levels greater than 10 mg/dL were noted in 32% and 8% of patients on liposomal amphotericin B and fluconazole therapy, respectively. Caspofungin Caspofungin is the first of a new class of parenterally administered antifungal agents, the echinocandins, which act by inhibiting the synthesis of a component of the fungal cell wall. In a series of 239 patients with invasive candidiasis, although fewer hepatic adverse events were observed with caspofungin compared with conventional amphotericin B, elevated serum ALT, AST, alkaline phosphatase, and total bilirubin were still observed in 3.7%, 1.9%, 8.3%, and 2.8%, respectively, of caspofungin-treated patients (232). Similarly, in a series of 90 patients with invasive aspergillosis refractory to or intolerant to other antifungal agents, liver test abnormalities occurred, but in less than 2% of patients (233). Isolated cases of clinically significant liver injury have been reported, although it is unclear whether they were drug related (234). Oral Antifungal Agents Ketoconazole, fluconazole, itraconazole, and voriconazole are synthetic oral azole agents used in the treatment of systemic fungal infections. In a cohort of 69,830 users of antifungal agents, ketoconazole and itraconazole were associated with the greatest risk of developing acute liver injury; the incidence rates of acute liver injury were 134.1 per 100,000 person-months for ketoconazole and 10.4 per 100,000 person-months for itraconazole (235). In a total of 54,803 users of either fluconazole or itraconazole, the rate of serious adverse liver events was 1.4/ 100,000 prescriptions (95% CI 0.25–8.2) and 3.2/100,000 prescriptions (95% CI 0.6–17.9) for fluconazole and itraconazole, respectively (236). Ketoconazole The incidence of symptomatic hepatic injury associated with ketoconazole has been estimated at approximately 1:2000 (237) to 1:10,000 (238). The incidence, as well as the severity and course, of liver injury associated with ketoconazole were also determined in a cohort study of 211 patients with onychomycosis (239). Asymptomatic serum transaminase elevations and overt hepatitis, which resolved after drug withdrawal, occurred in 17.5% and 2.9% of patients, respectively (239). Female gender and older age are associated with a greater risk for ketoconazole-associated liver injury (237,239). Liver injury associated with ketoconazole is predominantly hepatocellular, although mixed hepatocellular–cholestatic and cholestatic patterns have been reported (240). Clinical

536

Moseley

features resemble those of viral hepatitis, and fever, rash, and eosinophilia are rare (240). Onset of illness typically occurs over four weeks after starting the drug (237,239,240). Spotty necrosis and mononuclear cell infiltrates in the portal tract are seen on liver biopsy, except in fatal cases (241,242) where massive necrosis has been reported. Cholestasis, however, was the predominant feature on liver biopsy in other cases (243). Complete recovery from nonfatal cases of ketoconazole-associated acute liver injury uniformly occurs, although there is a case report of cirrhosis developing as a sequela (244). Given the lack of hallmarks of hypersensitivity and the variable duration of ketoconazole administration before onset, the mechanism of the liver injury is thought to be the result of a metabolic idiosyncrasy. Ketoconazole is extensively metabolized by hepatic microsomal enzymes, but the route of formation and identity of toxic metabolites are largely unknown. It is recommended that liver chemistries should be conducted before ketoconazole, or, for that matter, any oral azole agent, is initiated and periodically thereafter, and patients must be informed of the symptoms of hepatitis. Since most cases of ketoconazole-associated liver injury occur during the first few months of treatment, monitoring is especially important during this time and ketoconazole therapy should be discontinued in patients who have symptomatic hepatitis or laboratory evidence of progressive or persistent hepatic dysfunction (245). Fluconazole Asymptomatic mild elevations in liver chemistries occur in less than 7% of patients during fluconazole therapy (245). The incidence may be greater in immunocompromised patients, including patients with AIDS (246) and bone marrow transplant recipients (230). Several case reports describe acute hepatocellular injury, including three deaths, associated with fluconazole use (247–253). Mixed hepatocellular–cholestatic and cholestatic injury patterns have been reported. Although liver biopsy revealed nonspecific findings, electron microscopy in one case of liver injury after prolonged (O3 months) fluconazole use revealed giant mitochondria with paracrystalline inclusions and enlarged smooth endoplasmic reticulum (254). Itraconazole Asymptomatic minor liver chemistry abnormalities were reported in 7% of patients on chronic itraconazole therapy for systemic mycoses (255). Subsequent studies have reported even lower rates of liver chemistry abnormalities with this azole agent (256). In contrast, only two cases of symptomatic liver injury have been reported (257,258). Affected patients exhibited a mixed cholestatic–hepatocellular injury pattern with resolution within six weeks after itraconazole was discontinued. However, cholestatic injury with ductopenia developed in three patients on itraconazole, suggesting that this agent should be added to the list of possible causes of the vanishing bile duct syndrome (259). Pulse versus continuous (O1 month) itraconazole appears to be safer in the treatment of onychomycosis (260). Voriconazole Abnormal liver chemistries have been reported in 5% to 15% of patients treated with this new triazole antifungal agent (261). The usual pattern described has been elevations in the serum levels of ALT and AST, although elevated serum alkaline phosphatase levels have also been noted. Although most patients have asymptomatic elevation of hepatic enzyme levels, several patients with severe life-threatening hepatitis have been described (262). The risk of developing hepatitis appears to increase with increased serum voriconazole levels and resolves following drug withdrawal (262). Of note, voriconazole therapy was well tolerated in a patient with coccidioidal meningitis that developed hepatotoxicity related to fluconazole (263). Terbinafine Terbinafine is an orally and topically active agent used in the treatment of onychomycosis. Clinically significant cholestatic liver injury was reported in two patients in a postmarketing surveillance study of 25,884 patients treated with terbinafine (264). In addition, there have been several case reports of terbinafine-associated hepatotoxicity (265–270). In general, liver injury associated with terbinafine exhibits a mixed cholestatic–hepatocellular pattern, although a predominant and prolonged cholestatic injury has been described. Acute liver failure requiring

Antibacterial and Antifungal Agents

537

liver transplantation has been reported (271). In two patients, reduced numbers of interlobular bile ducts were found on liver biopsy, suggesting that terbinafine-associated liver injury may lead to a vanishing bile duct syndrome (272,273). The latency period between the start of medication and the development of liver injury is reported to be approximately four to six weeks (267). Although the allylic aldehyde metabolite of terbinafine, 7,7-dimethylhept-2-ene-4-ynal (TBF-A), has been proposed to play a role in the pathogenesis of its hepatotoxicity (274), liver injury associated with this antifungal agent is rare (estimated at 1:45,000–1:54,000) and most likely idiosyncratic. Flucytosine Asymptomatic liver chemistry abnormalities have been reported in up to 41% of patients on flucytosine (275). Severe liver necrosis has been reported in two patients who received flucytosine for treatment of candidal endocarditis (276). The mechanism of flucytosinerelated hepatotoxicity is unknown, but it seems to be concentration dependent, predictable, possibly avoidable if peak concentrations of less than 100 mg/L are maintained, and reversible with temporary discontinuation of the drug or a reduction in dose. Griseofulvin Although griseofulvin causes a variety of acute and chronic effects in rodents, including induction of protoporphyria, where it serves as a useful experimental model, and hepatocellular carcinoma (277), hepatotoxicity is rare in humans. Cholestatic jaundice, reversible upon drug discontinuation, has been reported in a patient receiving griseofulvin (278). In a controlled cohort study of 211 patients with onychomycosis treated with ketoconazole or griseofulvin, no biochemical abnormalities were observed in patients during griseofulvin treatment (239). Griseofulvin alters porphyrin metabolism in humans and acute intermittent porphyria has been precipitated by the use of this agent (279). Investigational Antifungal Agents There are no data available on the potential for hepatic injury of two triazole antifungal agents, posaconazole and ravuconazole, and two echinocandins, anidulafungin and micafungin, that are currently under clinical investigation (280). CONCLUSION Liver injury from antibacterial and antifungal agents is typically asymptomatic and of little clinical significance. However, there are several notable exceptions, such as prolonged cholestasis from amoxicillin–clavulanic acid, severe acute hepatitis seen with trovafloxacin use, minocycline-associated autoimmune hepatitis, and chronic liver disease from the prolonged use of nitrofurantoin. Prompt recognition, discontinuation of the drug and substitution, if clinically indicated, with an alternative agent, and avoidance of rechallenge are the most effective means of managing suspected hepatotoxicity from antibacterial and antifungal agents. There are no specific measures to treat liver injury from antibacterial and antifungal agents other than supportive care. Liver transplantation may be required in the setting of acute liver failure. The judicious use of antibiotics, education of patients regarding the potential for, and the symptoms of, liver injury, and the reporting of suspected adverse hepatic effects to monitoring agencies are all important steps to be taken by the prescribing healthcare provider. REFERENCES 1. Trauner M, Boyer JL. Bile salt transporters: molecular characterization, function, and regulation. Physiol Rev 2003; 83:633–71. 2. Moseley RH. Sepsis and cholestasis. Clin Liver Dis 2004; 8:83–94. 3. Carson JL, Strom BL, Duff A, et al. Acute liver disease associated with erythromycins, sulfonamides, and tetracyclines. Ann Intern Med 1993; 119:576–83.

538

Moseley

4. Sgro C, Clinard F, Ouazir K, et al. Incidence of drug-induced hepatic injuries: a French populationbased study. Hepatology 2002; 36:451–5. 5. de Abajo FJ, Montero D, Madurga M, Garcia Rodriguez LA. Acute and clinically relevant druginduced liver injury: a population based case-control study. Br J Clin Pharmacol 2004; 58:71–80. 6. Galan MV, Potts JA, Silverman AL, Gordon SC. Hepatitis in a United States tertiary referral center. J Clin Gastroenterol 2005; 39:64–7. 7. Meier Y, Cavallaro M, Roos M, et al. Incidence of drug-induced liver injury in medical inpatients. Eur J Clin Pharmacol 2005; 61:135–43. 8. Andrade RJ, Lucena MI, Fernandez MC, et al. Drug-induced liver injury: an analysis of 461 incidences submitted to the Spanish registry over a 10-year period. Gastroenterology 2005; 129:512–21. 9. Bjornsson E, Jerlstad P, Bergqvist A, Olsson R. Fulminant drug-induced hepatic failure leading to death or liver transplantation in Sweden. Scand J Gastroenterol 2005; 40:1095–101. 10. Bjornsson E, Olsson R. Outcome and prognostic markers in severe drug-induced liver disease. Hepatology 2005; 42:481–9. 11. Parry MF. The pencillins. Med Clin North Am 1987; 71:1093–112. 12. Goldstein LI, Ishak KG. Hepatic injury associated with penicillin therapy. Arch Pathol 1974; 98:114–7. 13. Williams CN, Malatjalian DA. Severe penicillin-induced cholestasis in a 91-year-old woman. Dig Dis Sci 1981; 26:470–3. 14. Onate J, Montejo M, Aguirrebengoa K, Ruiz-Irastorza G, Gonzalez de Zarate P, Aguirre C. Hepatotoxicity associated with penicillin V therapy. Clin Infect Dis 1995; 20:474–5. 15. Andrade RJ, Guilarte J, Salmeron FJ, Lucena MI, Bellot V. Benzylpenicillin-induced prolonged cholestasis. Ann Pharmacother 2001; 35:783–4. 16. Cavanzo FJ, Garcia CF, Botero RC. Chronic cholestasis, paucity of bile ducts, red cell aplasia, and the Stevens–Johnson syndrome: an ampicillin-associated case. Gastroenterology 1990; 99:854–6. 17. Koklu S, Yuksel O, Filik L, Uskudar O, Altundag K, Altiparmak E. Recurrent cholestasis due to ampicillin. Ann Pharmacother 2003; 37:395–7. 18. Koklu S, Koksal AS, Asil M, Kiyici H, Coban S, Arhan M. Probable sulbactam/ampicillin-associated prolonged cholestasis. Ann Pharmacother 2004; 38:2055–8. 19. Garcia Rodriguez LA, Stricker BH, Zimmerman HJ. Risk of acute liver injury associated with the combination of amoxicillin and clavulanic acid. Arch Intern Med 1996; 156:1327–32. 20. Davies MH, Harrison RF, Elias E, Hubscher SG. Antibiotic-associated acute vanishing bile duct syndrome: a pattern associated with severe, prolonged, intrahepatic cholestasis. J Hepatol 1994; 20:112–6. 21. Miller WI, Souney PF, Chang JT. Hepatic dysfunction following nafcillin and cephalothin therapy in a patient with a history of oxacillin hepatitis. Clin Pharm 1983; 2:465–8. 22. Mazuryk H, Kastenberg D, Rubin R, Munoz SJ. Cholestatic hepatitis associated with the use of nafcillin. Am J Gastroenterol 1993; 88:1960–2. 23. Schuman R, Miskovitz P. Nafcillin-associated jaundice. Am J Gastroenterol 1994; 89:952. 24. Presti ME, Janney CG, Neuschwander-Tetri BA. Nafcillin-associated hepatotoxicity: report of a case and review of the literature. Dig Dis Sci 1996; 41:180–4. 25. Wilson FM, Belamaric J, Lauter CB, Lerner AM. Anicteric carbenicillin hepatitis. Eight episodes in four patients. JAMA 1975; 232:818–21. 26. Siegmund JB, Tarshis AM. Prolonged jaundice after dicloxacillin therapy. Am J Gastroenterol 1993; 88:1299–300. 27. Kleinman MS, Presberg JE. Cholestatic hepatitis after dicloxacillin-sodium therapy. J Clin Gastroenterol 1986; 8:77–8. 28. Saab S, Venkataramani A, Yao F. Possible granulomatous hepatitis after dicloxacillin therapy. J Clin Gastroenterol 1996; 22:163–4. 29. Olsson R, Wiholm BE, Sand C, Zettergren L, Hultcrantz R, Myrhed M. Liver damage from flucloxacillin, cloxacillin and dicloxacillin. J Hepatol 1992; 15:154–61. 30. Konikoff F, Alcalay J, Halevy J. Cloxacillin-induced cholestatic jaundice. Am J Gastroenterol 1986; 81:1082–3. 31. Pascual J, Orofino L, Marcen R, Quereda C, Ortuno J. Cloxacillin-induced cholestasis in a renal allograft patient with chronic hepatitis. Am J Gastroenterol 1990; 85:335–6. 32. Lotric S, Lejko-Zupanc T, Jereb M. Cloxacillin-induced cholestasis. Clin Infect Dis 1994; 19:981–2. 33. Goland S, Malnick SD, Gratz R, Feldberg E, Geltner D, Sthoeger ZM. Severe cholestatic hepatitis following cloxacillin treatment. Postgrad Med J 1998; 74:59–60. 34. Klein I, Tobias H. Oxacillin-associated hepatitis. Am J Gastroenterol 1976; 65:546. 35. Pollock AA, Berger SA, Simberkoff MS, Rahal JJ. Hepatitis associated with high-dose oxacillin therapy. Arch Intern Med 1978; 138:915–7. 36. Onorato IM, Axelrod JL. Hepatitis from intravenous high-dose oxacillin therapy: findings in an adult inpatient population. Ann Intern Med 1978; 89:497–500.

Antibacterial and Antifungal Agents

539

37. Bruckstein AH, Attia AA. Oxacillin hepatitis. Two patients with liver biopsy, and review of the literature. Am J Med 1978; 64:519–22. 38. Goldstein LI, Granoff M, Waisman J. Hepatic injury due to oxacillin administration. Am J Gastroenterol 1978; 70:171–4. 39. Halloran TJ, Clague MD. Hepatitis associated with high-dose oxacillin therapy. Arch Intern Med 1979; 139:376–7. 40. D’Angelo LJ. Oxacillin and hepatotoxicity. Ann Intern Med 1979; 90:442. 41. Tauris P, Jorgensen NF, Petersen CM, Albertsen K. Prolonged severe cholestasis induced by oxacillin derivatives. A report on two cases. Acta Med Scand 1985; 217:567–9. 42. Al-Homaidhi H, Abdel-Haq NM, El-Baba M, Asmar BI. Severe hepatitis associated with oxacillin therapy. South Med J 2002; 95:650–2. 43. Maraqa NF, Gomez MM, Rathore MH, Alvarez AM. Higher occurrence of hepatotoxicity and rash in patients treated with oxacillin, compared with those treated with nafcillin and other commonly used antimicrobials. Clin Infect Dis 2002; 34:50–4. 44. Saliba B, Herbert PN. Oxacillin hepatotoxicity in HIV-infected patients. Ann Intern Med 1994; 120:1048. 45. Aderka D, Livni E, Salamon F, Weinberger A, Pinkhas J. Use of macrophage inhibition factor and mast-cell degranulation tests for diagnosis of cloxacillin-induced cholestasis. Am J Gastroenterol 1986; 81:1084–6. 46. Horikawa M, Kato Y, Tyson CA, Sugiyama Y. Potential cholestatic activity of various therapeutic agents assessed by bile canalicular membrane vesicles isolated from rats and humans. Drug Metab Pharmacokinet 2003; 18:16–22. 47. Lobatto S, Dijkmans BA, Mattie H, Van Hooff JP. Flucloxacillin-associated liver damage. Neth J Med 1982; 25:47–8. 48. Bengtsson F, Floren CH, Hagerstrand I, Soderstrom C, Aberg T. Flucloxacillin-induced cholestatic liver damage. Scand J Infect Dis 1985; 17:125–8. 49. Turner IB, Eckstein RP, Riley JW, Lunzer MR. Prolonged hepatic cholestasis after flucloxacillin therapy. Med J Aust 1989; 151:701–5. 50. Fairley CK, McNeil JJ, Desmond P, et al. Risk factors for development of flucloxacillin associated jaundice. Br Med J 1993; 306:233–5. 51. Devereaux BM, Crawford DH, Purcell P, Powell LW, Roeser HP. Flucloxacillin associated cholestatic hepatitis. An Australian and Swedish epidemic? Eur J Clin Pharmacol 1995; 49:81–5. 52. Piotrowicz A, Polkey M, Wilkinson M. Ursodeoxycholic acid for the treatment of flucloxacillinassociated cholestasis. J Hepatol 1995; 22:119–20. 53. Eckstein RP, Dowsett JF, Lunzer MR. Flucloxacillin induced liver disease. Histopathological findings at biopsy and autopsy. Pathology 1993; 25:223–8. 54. Miros M, Kerlin P, Walker N, Harris O. Flucloxacillin induced delayed cholestatic hepatitis. Aust NZ J Med 1990; 20:251–3. 55. Koek GH, Stricker BH, Blok AP, Schalm SW, Desmet VJ. Flucloxacillin-associated hepatic injury. Liver 1994; 14:225–9. 56. Jick H, Derby LE, Dean AD, Henry DA. Flucloxacillin and cholestatic hepatitis. Med J Aust 1994; 160:525. 57. Derby LE, Jick H, Henry DA, Dean AD. Cholestatic hepatitis associated with flucloxacillin. Med J Aust 1993; 158:596–600. 58. George DK, Crawford DH. Antibacterial-induced hepatotoxicity. Incidence, prevention and management. Drug Saf 1996; 15:79–85. 59. Russmann S, Kaye JA, Jick SS, Jick H. Risk of cholestatic liver disease associated with flucloxacillin and flucloxacillin prescribing habits in the U.K.: cohort study using data from the U.K. General Practice Research Database. Br J Clin Pharmacol 2005; 60:76–82. 60. Lakehal F, Dansette PM, Becquemont L, et al. Indirect cytotoxicity of flucloxacillin toward human biliary epithelium via metabolite formation in hepatocytes. Chem Res Toxicol 2001; 14:694–701. 61. Carey MA, van Pelt FN. Immunochemical detection of flucloxacillin adduct formation in livers of treated rats. Toxicology 2005; 216:41–8. 62. Dowsett JF, Gillow T, Heagerty A, Radcliffe M, Toadi R, Isle I, Russell RC. Amoxycillin/clavulanic acid (Augmentin)-induced intrahepatic cholestasis. Dig Dis Sci 1989; 34:1290–3. 63. Stricker BH, Van den Broek JW, Keuning J, Eberhardt W, Houben HG, Johnson M, Blok AP. Cholestatic hepatitis due to antibacterial combination of amoxicillin and clavulanic acid (augmentin). Dig Dis Sci 1989; 34:1576–80. 64. Reddy KR, Brillant P, Schiff ER. Amoxicillin–clavulanate potassium-associated cholestasis. Gastroenterology 1989; 96:1135–41. 65. Schneider JE, Kleinman MS, Kupiec JW. Cholestatic hepatitis after therapy with amoxicillin/clavulanate potassium. NY State J Med 1989; 89:355–6. 66. Verhamme M, Ramboer C, Van de Bruaene P, Inderadjaja N. Cholestatic hepatitis due to an amoxycillin/clavulanic acid preparation. J Hepatol 1989; 9:260–4.

540

Moseley

67. Silvain C, Fort E, Levillian P, et al. Granulomatous hepatitis due to combination of amoxicillin and clavulanic acid. Dig Dis Sci 1992; 37:150–2. 68. Larrey D, Vial T, Micaleff A, Babany G, Morichau-Beauchant M, Michel H, Benhamou JP. Hepatitis associated with amoxycillin–clavulanic acid combination report of 15 cases. Gut 1992; 33:368–71. 69. Thomson JA, Fairley CK, Ugoni AM, et al. Risk factors for the development of amoxycillin– clavulanic acid associated jaundice. Med J Aust 1995; 162:638–40. 70. Gresser U. Amoxicillin–clavulanic acid therapy may be associated with severe side effects-review of the literature. Eur J Med Res 2001; 6(4):139–49. 71. Ryley NG, Fleming KA, Chapman RW. Focal destructive cholangiopathy associated with amoxycillin/clavulanic acid (Augmentin). J Hepatol 1995; 23:278–82. 72. Richardet JP, Mallat A, Zafrani ES, Blazquez M, Bognel JC, Campillo B. Prolonged cholestasis with ductopenia after administration of amoxicillin/clavulanic acid. Dig Dis Sci 1999; 44:1997–2000. 73. Hebbard GS, Smith KG, Gibson PR, Bhathal PS. Augmentin-induced jaundice with a fatal outcome. Med J Aust 1992; 156:285–6. 74. Fontana RJ, Shakil AO, Greenson JK, Boyd I, Lee WM. Acute liver failure due to amoxicillin and amoxicillin/clavulanate. Dig Dis Sci 2005; 50:1785–90. 75. Chawla A, Kahn E, Yunis EJ, et al. Rapidly progressive cholestasis. An unusual reaction to amoxicillin/clavulanic acid therapy in a child. J Pediatr 2000; 136:121–3. 76. Smith LA, Ignacio JR, Winesett MP, et al. Vanishing bile duct syndrome: amoxicillin–clavulanic acid associated intra-hepatic cholestasis responsive to ursodeoxycholic acid. J Pediatr Gastroenterol Nutr 2005; 41:469–73. 77. Katsinelos P, Vasiliadis T, Xiarchos P, Patakiouta F, Christodoulou K, Pilpilidis I, Eugenidis N. Ursodeoxycholic acid (UDCA) for the treatment of amoxycillin–clavulanate potassium (Augmentin)-induced intra-hepatic cholestasis: report of two cases. Eur J Gastroenterol Hepatol 2000; 12:365–8. 78. Hautekeete ML, Horsmans Y, Van Waeyenberge C, et al. HLA association of amoxicillin– clavulanate-induced hepatitis. Gastroenterology 1999; 117:1181–6. 79. O’Donohue J, Oien KA, Donaldson P, Underhill J, Clare M, MacSween RN, Mills PR. Co-amoxiclav jaundice: clinical and histological features and HLA class II association. Gut 2000; 47:717–20. 80. Berg P, Hahn EG. Hepatotoxic reactions induced by beta-lactamase inhibitors. Eur J Med Res 2001; 6:535–42. 81. Parry MF. The safety and tolerance of azlocillin. Arzneimittelforschung 1985; 35:1292–4. 82. Lang R, Lishner M, Ravid M. Adverse reactions to prolonged treatment with high doses of carbenicillin and ureidopenicillins. Rev Infect Dis 1991; 13:68–72. 83. Hargreaves JE, Herchline TE. Severe cholestatic jaundice caused by mezlocillin. Clin Infect Dis 1992; 15:179–80. 84. Quattropani C, Schneider M, Helbling A, Zimmermann A, Krahenbuhl S. Cholangiopathy after short-term administration of piperacillin and imipenem/cilastatin. Liver 2001; 21:213–6. 85. Ammann R, Neftel K, Hardmeier T, Reinhardt M. Cephalosporin-induced cholestatic jaundice. Lancet 1982; 2:336–7. 86. Eggleston SM, Belandres MM. Jaundice associated with cephalosporin therapy. Drug Intell Clin Pharm 1985; 19:553–5. 87. Bjornsson E, Olsson R. Acute liver injury due to loracarbef. J Hepatol 1997; 26:739–40. 88. Skoog SM, Smyrk TC, Talwalkar JA. Cephalexin-induced cholestatic hepatitis. J Clin Gastroenterol 2004; 38:833. 89. Shiffman ML, Keith FB, Moore EW. Pathogenesis of ceftriaxone-associated biliary sludge. In vitro studies of calcium-ceftriaxone binding and solubility. Gastroenterology 1990; 99:1772–8. 90. Johnson DF, Hall WH. Allergic hepatitis caused by propionyl erythromycin ester of lauryl sulfate. N Engl J Med 1961; 265:1200–2. 91. Reed C, Ritchie F. Toxic jaundice due to propionyl erythromycin ester lauryl sulphate (Ilosone). Med J Aust 1962; 49:810–2. 92. Tolman KG, Sannella JJ, Freston JW. Chemical structure of erythromycin and hepatotoxicity. Ann Intern Med 1974; 81:58–60. 93. Zafrani ES, Ishak KG, Rudzki C. Cholestatic and hepatocellular injury associated with erythromycin esters: report of nine cases. Dig Dis Sci 1979; 24:385–96. 94. Viteri AL, Greene JF, Jr., Dyck WP. Erythromycin ethylsuccinate-induced cholestasis. Gastroenterology 1979; 76:1007–8. 95. Keeffe EB, Reis TC, Berland JE. Hepatotoxicity to both erythromycin estolate and erythromycin ethylsuccinate. Dig Dis Sci 1982; 27:701–4. 96. Diehl AM, Latham P, Boitnott JK, Mann J, Maddrey WC. Cholestatic hepatitis from erythromycin ethylsuccinate. Report of two cases. Am J Med 1984; 76:931–4. 97. Inman WH, Rawson NS. Erythromycin estolate and jaundice. Br Med J 1983; 286:1954–5. 98. Avila P, Capella D, Laporte JR, Moreno V. Which salt of erythromycin is most hepatotoxic? Lancet 1988; 1:1104.

Antibacterial and Antifungal Agents

541

99. Derby LE, Jick H, Henry DA, Dean AD. Erythromycin-associated cholestatic hepatitis. Med J Aust 1993; 158:600–2. 100. Geubel AP, Nakad A, Rahier J, Dive C. Prolonged cholestasis and disappearance of interlobular bile ducts following chlorpropamide and erythromycin ethylsuccinate. Case of drug interaction? Liver 1988; 8:350–3. 101. Degott C, Feldmann G, Larrey D. Drug-induced prolonged cholestasis in adults: a histological semiquantitative study demonstrating progressive ductopenia. Hepatology 1992; 15:244–51. 102. Lazarczyk DA, Duffy MC. Erythromycin-induced primary biliary cirrhosis. Dig Dis Sci 2000; 45:1115–8. 103. Gholson CF, Warren GH. Fulminant hepatic failure associated with intravenous erythromycin lactobionate. Arch Intern Med 1990; 150:215–6. 104. Longo G, Valenti C, Gandini G, Ferrara L, Bertesi M, Emilia G. Azithromycin-induced intrahepatic cholestasis. Am J Med 1997; 102:217–8. 105. Cascaval RI, Lancaster DJ. Hypersensitivity syndrome associated with azithromycin. Am J Med 2001; 110:330–1. 106. Chandrupatla S, Demetris AJ, Rabinovitz M. Azithromycin-induced intrahepatic cholestasis. Dig Dis Sci 2002; 47:2186–8. 107. Sturgill MG, Rapp RP. Clarithromycin: review of a new macrolide antibiotic with improved microbiologic spectrum and favorable pharmacokinetic and adverse effect profiles. Ann Pharmacother 1992; 26:1099–108. 108. Yew WW, Chau CH, Lee J, Leung CW. Cholestatic hepatitis in a patient who received clarithromycin therapy for a Mycobacterium chelonae lung infection. Clin Infect Dis 1994; 18:1025–6. 109. Brown BA, Wallace RJ, Griffith DE, Girard W. Clarithromycin-induced hepatotoxicity. Clin Infect Dis 1995; 20:1073–4. 110. Fox JC, Szyjkowski RS, Sanderson SO, Levine RA. Progressive cholestatic liver disease associated with clarithromycin treatment. J Clin Pharmacol 2002; 42:676–80. 111. Shaheen N, Grimm IS. Fulminant hepatic failure associated with clarithromycin. Am J Gastroenterol 1996; 91:394–5. 112. Masia M, Gutierrez F, Jimeno A, Navarro A, Borras J, Matarredona J, Martin-Hidalgo A. Fulminant hepatitis and fatal toxic epidermal necrolysis (Lyell disease) coincident with clarithromycin administration in an alcoholic patient receiving disulfiram therapy. Arch Intern Med 2002; 162:474–6. 113. Christopher K, Hyatt PA, Horkan C, Yodice PC. Clarithromycin use preceding fulminant hepatic failure. Am J Gastroenterol 2002; 97:489–90. 114. Tietz A, Heim MH, Eriksson U, Marsch S, Terracciano L, Krahenbuhl S. Fulminant liver failure associated with clarithromycin. Ann Pharmacother 2003; 37:57–60. 115. Nguyen M, Chung EP. Telithromycin: the first ketolide antimicrobial. Clin Ther 2005; 27:1144–63. 116. Peters TS. Do preclinical testing strategies help predict human hepatotoxic potentials? Toxicol Pathol 2005; 33:146–54. 117. Clay KD, Hanson JS, Pope SD, Rissmiller RW, Purdum PP III, Banks PM. Brief communication: severe hepatotoxicity of telithromycin: three case reports and literature review. Ann Intern Med 2006; 144:415–20. 118. Yamaguchi S, Zhao YL, Nadai M. Involvement of the drug transporters P-glycoprotein and multidrug resistance-associated protein Mrp2 in telithromycin transport. Antimicrob Agents Chemother 2006; 50:80–7. 119. Gaeta GB, Utili R, Adinolfi LE, Abernathy CO, Giusti G. Characterization of the effects of erythromycin estolate and erythromycin base on the excretory function of the isolated rat liver. Toxicol Appl Pharmacol 1985; 80:185–92. 120. Kostrubsky VE, Strom SC, Hanson J, et al. Evaluation of hepatotoxic potential of drugs by inhibition of bile-acid transport in cultured primary human hepatocytes and intact rats. Toxicol Sci 2003; 76:220–8. 121. Pessayre D, Larrey D, Funck-Brentano C, Benhamou JP. Drug interactions and hepatitis produced by some macrolide antibiotics. J Antimicrob Chemother 1985; 16(Suppl. A):181–94. 122. Westphal JF, Brogard JM. Antibacterials and antifungal agents. In: Kaplowitz N, DeLeve LD, eds. Drug-Induced Liver Disesase. New York: Marcel Dekker, 2003. 123. Dujovne CA, Chan CH, Zimmerman HJ. Sulfonamide hepatic injury. Review of the literature and report of a case due to sulfamethoxazole. N Engl J Med 1967; 277:785–8. 124. Gordin FM, Simon GL, Wofsy CB, Mills J. Adverse reactions to trimethoprim–sulfamethoxazole in patients with the acquired immunodeficiency syndrome. Ann Intern Med 1984; 100:495–9. 125. Espiritu CR, Kim TS, Levine RA. Granulomatous hepatitis associated with sulfadimethoxine hypersensitivity. JAMA 1967; 202:985–8. 126. Callen JP, Soderstrom RM. Granulomatous hepatitis associated with salicylazosulfapyridine therapy. South Med J 1978; 71:1159–60. 127. Lazar HP, Murphy RL, Phair JP. Fansidar and hepatic granulomas. Ann Intern Med 1985; 102:722.

542

Moseley

128. Zaman F, Ye G, Abreo KD, Latif S, Zibari GB. Successful orthotopic liver transplantation after trimethoprim–sulfamethoxazole associated fulminant liver failure. Clin Transplant 2003; 17:461–4. 129. Alberti-Flor JJ, Hernandez ME, Ferrer JP, Howell S, Jeffers L. Fulminant liver failure and pancreatitis associated with the use of sulfamethoxazole–trimethoprim. Am J Gastroenterol 1989; 84:1577–9. 130. Zitelli BJ, Alexander J, Taylor S, et al. Fatal hepatic necrosis due to pyrimethamine–sulfadoxine (Fansidar). Ann Intern Med 1987; 106:393–5. 131. Ransohoff DF, Jacobs G. Terminal hepatic failure following a small dose of sulfamethoxazole– trimethoprim. Gastroenterology 1981; 80:816–9. 132. Yao F, Behling CA, Saab S, Li S, Hart M, Lyche KD. Trimethoprim–sulfamethoxazole-induced vanishing bile duct syndrome. Am J Gastroenterol 1997; 92:167–9. 133. Altraif I, Lilly L, Wanless IR, Heathcote J. Cholestatic liver disease with ductopenia (vanishing bile duct syndrome) after administration of clindamycin and trimethoprim–sulfamethoxazole. Am J Gastroenterol 1994; 89:1230–4. 134. Kowdley KV, Keeffe EB, Fawaz KA. Prolonged cholestasis due to trimethoprim sulfamethoxazole. Gastroenterology 1992; 102:2148–50. 135. Abi-Mansur P, Ardiaca MC, Allam C, Shamma’a M. Trimethoprim–sulfamethoxazole-induced cholestasis. Am J Gastroenterol 1981; 76:356–9. 136. Nair SS, Kaplan JM, Levine LH, Geraci K. Trimethoprim–sulfamethoxazole-induced intrahepatic cholestasis. Ann Intern Med 1980; 92:511–2. 137. Munoz SJ, Martinez-Hernandez A, Maddrey WC. Intrahepatic cholestasis and phospholipidosis associated with the use of trimethoprim–sulfamethoxazole. Hepatology 1990; 12:342–7. 138. Thies PW, Dull WL. Trimethoprim–sulfamethoxazole-induced cholestatic hepatitis: inadvertent rechallenge. Arch Intern Med 1984; 144:1691–2. 139. Wolkenstein P, Carriere V, Charue D, et al. A slow acetylator genotype is a risk factor for sulphonamide-induced toxic epidermal necrolysis and Stevens–Johnson syndrome. Pharmacogenetics 1995; 5:255–8. 140. O’Neil WM, MacArthur RD, Farrough MJ, et al. Acetylator phenotype and genotype in HIV-infected patients with and without sulfonamide hypersensitivity. J Clin Pharmacol 2002; 42:613–9. 141. Alfirevic A, Stalford AC, Vilar FJ, Wilkins EG, Park BK, Pirmohamed M. Slow acetylator phenotype and genotype in HIV-positive patients with sulphamethoxazole hypersensitivity. Br J Clin Pharmacol 2003; 55:158–65. 142. Dowling HF, Lepper MH. Hepatic reactions to tetracycline. JAMA 1964; 188:307–9. 143. Zimmerman HJ. Hepatotoxicity: The Adverse Effects of Drugs and Other Chemicals on the Liver. 2nd ed. Philadelphia: Lippincott Williams & Wilkins, 1999. 144. Combes B, Whalley PJ, Adams RH. Tetracycline and the liver. Prog Liver Dis 1972; 4:589–96. 145. Whalley PJ, Adams RH, Combes B. Tetracycline toxicity in pregnancy. JAMA 1964; 189:357–62. 146. Allen ES, Brown WE. Hepatic toxicity of tetracycline in pregnancy. Am J Obstet Gynecol 1966; 95:12–8. 147. Schultz JC, Adamson JS, Workman WW, Norman TD. Fatal liver disease after intravenous administration of tetracycline in high dosage. N Engl J Med 1963; 269:999–1004. 148. Freneaux E, Labbe G, Letteron P, et al. Inhibition of the mitochondrial oxidation of fatty acids by tetracycline in mice and in man: possible role in microvesicular steatosis induced by this antibiotic. Hepatology 1988; 8:1056–62. 149. Letteron P, Sutton A, Mansouri A, Fromenty B, Pessayre D. Inhibition of microsomal triglyceride transfer protein: another mechanism for drug-induced steatosis in mice. Hepatology 2003; 38:133–40. 150. Bjornsson E, Lindberg J, Olsson R. Liver reactions to oral low-dose tetracyclines. Scand J Gastroenterol 1997; 32:390–5. 151. Hunt CM, Washington K. Tetracycline-induced bile duct paucity and prolonged cholestasis. Gastroenterology 1994; 107:1844–7. 152. Westermann GW, Bohm M, Bonsmann G, Rahn KH, Kisters K. Chronic intoxication by doxycycline use for more than 12 years. J Intern Med 1999; 246(6):591–2. 153. Burette A, Finet C, Prigogine T, De Roy G, Deltenre M. Acute hepatic injury associated with minocycline. Arch Intern Med 1984; 144:1491–2. 154. Davies MG, Kersey PJ. Acute hepatitis and exfoliative dermatitis associated with minocycline. Br Med J 1989; 298:1523–4. 155. Malcolm A, Heap TR, Eckstein RP, Lunzer MR. Minocycline-induced liver injury. Am J Gastroenterol 1996; 91:1641–3. 156. Gough A, Chapman S, Wagstaff K, Emery P, Elias E. Minocycline induced autoimmune hepatitis and systemic lupus erythematosus-like syndrome. Br Med J 1996; 312:169–72. 157. Teitelbaum JE, Perez-Atayde AR, Cohen M, Bousvaros A, Jonas MM. Minocycline-related autoimmune hepatitis: case series and literature review. Arch Pediatr Adolesc Med 1998; 152:1132–6. 158. Crosson J, Stillman MT. Minocycline-related lupus erythematosus with associated liver disease. J Am Acad Dermatol 1997; 36:867–8.

Antibacterial and Antifungal Agents

543

159. Golstein PE, Deviere J, Cremer M. Acute hepatitis and drug-related lupus induced by minocycline treatment. Am J Gastroenterol 1997; 92:143–6. 160. Bhat G, Jordan J, Jr., Sokalski S, Bajaj V, Marshall R, Berkelhammer C. Minocycline-induced hepatitis with autoimmune features and neutropenia. J Clin Gastroenterol 1998; 27:74–5. 161. Goldstein NS, Bayati N, Silverman AL, Gordon SC. Minocycline as a cause of drug-induced autoimmune hepatitis. Report of four cases and comparison with autoimmune hepatitis. Am J Clin Pathol 2000; 114:591–8. 162. Nietsch HH, Libman BS, Pansze TW, Eicher JN, Reeves JR, Krawitt EL. Minocycline-induced hepatitis. Am J Gastroenterol 2000; 95:2993–5. 163. Min DI, Burke PA, Lewis WD, Jenkins RL. Acute hepatic failure associated with oral minocycline: a case report. Pharmacotherapy 1992; 12:68–71. 164. Boudreaux JP, Hayes DH, Mizrahi S, Hussey J, Regenstein F, Balart L. Fulminant hepatic failure, hepatorenal syndrome, and necrotizing pancreatitis after minocycline hepatotoxicity. Transplant Proc 1993; 25:1873. 165. Pohle T, Menzel J, Domschke W. Minocycline and fulminant hepatic failure necessitating liver transplantation. Am J Gastroenterol 2000; 95:560–1. 166. Lawrenson RA, Seaman HE, Sundstrom A, Williams TJ, Farmer RD. Liver damage associated with minocycline use in acne: a systematic review of the published literature and pharmacovigilance data. Drug Saf 2000; 23:333–49. 167. Maddrey WC. Drug-induced hepatotoxicity: 2005. J Clin Gastroenterol 2005; 39(Suppl. 2):S83–9. 168. Kaufmann D, Pichler W, Beer JH. Severe episode of high fever with rash, lymphadenopathy, neutropenia, and eosinophilia after minocycline therapy for acne. Arch Intern Med 1994; 154:1983–4. 169. Seaman HE, Lawrenson RA, Williams TJ, MacRae KD, Farmer RD. The risk of liver damage associated with minocycline: a comparative study. J Clin Pharmacol 2001; 41:852–60. 170. Liu ZX, Kaplowitz N. Immune-mediated drug-induced liver disease. Clin Liver Dis 2002; 6:755–74. 171. Babinchak T, Ellis-Grosse E, Dartois N, Rose GM, Loh E, Tigecycline 301 Study Group, Tigecycline 306 Study Group. The efficacy and safety of tigecycline for the treatment of complicated intraabdominal infections: analysis of pooled clinical trial data. Clin Infect Dis 2005; 41(Suppl. 5):S354–67. 172. Ellis-Grosse EJ, Babinchak T, Dartois N, Rose G, Loh E, Tigecycline 300 cSSSI Study Group, Tigecycline 305 cSSSI Study Group. The efficacy and safety of tigecycline in the treatment of skin and skin-structure infections: results of 2 double-blind phase 3 comparison studies with vancomycin–aztreonam. Clin Infect Dis 2005; 41(Suppl. 5):S341–53. 173. Owens RC, Jr., Ambrose PG. Antimicrobial safety: focus on fluoroquinolones. Clin Infect Dis 2005; 41(Suppl. 2):S144–57. 174. Lopez-Navidad A, Domingo P, Cadafalch J, et al. Norfloxacin-induced hepatotoxicity. J Hepatol 1990; 11:277–8. 175. Blum A. Oflxacin-induced acute severe hepatitis. South Med J 1991; 84:1158. 176. Villeneuve JP, Davies C, Cote J. Suspected ciprofloxacin-induced hepatotoxicity. Ann Pharmacother 1995; 29:257–9. 177. Sherman O, Beizer JL. Possible ciprofloxacin induced cholestatic jaundice. Ann Pharmacother 1994; 28:1162–4. 178. Hautekeete ML. Cholestatic hepatitis related to quinolones: a report of two cases. J Hepatol 1995; 23:759–60. 179. Labowitz JK, Silverman WB. Cholestatic jaundice induced by ciprofloxacin. Dig Dis Sci 1997; 42:192–4. 180. Lucena MI, Andrade RJ, Sanchez-Martinez H, Perez-Serrano JM, Gomez-Outes A. Norfloxacininduced cholestatic jaundice. Am J Gastroenterol 1998; 93:2309–11. 181. Soto S, Lopez-Roses L, Avila S, Lancho A, Gonzalez A, Santos E, Urraca B. Moxifloxacin-induced acute liver injury. Am J Gastroenterol 2002; 97:1853–4. 182. Bataille L, Rahier J, Geubel A. Delayed and prolonged cholestatic hepatitis with ductopenia after long-term ciprofloxacin therapy for Crohn’s disease. J Hepatol 2002; 37:696–9. 183. Cheung O, Chopra K, Yu T, Nalesnik MA, Amin S, Shakil AO. Gatifloxacin-induced hepatotoxicity and acute pancreatitis. Ann Intern Med 2004; 140:73–4. 184. Grassmick BK, Lehr VT, Sundareson AS. Fulminant hepatic failure possibly related to ciprofloxacin. Ann Pharmacother 1992; 26:636–9. 185. Fuchs S, Simon Z, Brezin M. Fatal hepatic failure associated with ciprofloxacin. Lancet 1994; 343:738–9. 186. Coleman CI, Spencer JV, Chung JO, Reddy P. Possible gatifloxacin-induced fulminant hepatic failure. Ann Pharmacother 2002; 36:1162–7. 187. Zimpfer A, Propst A, Mikuz G, Vogel W, Terracciano L, Stadlmann S. Ciprofloxacin-induced acute liver injury: case report and review of literature. Virchows Arch 2004; 444:87–9. 188. Lucena MI, Andrade RJ, Rodrigo L, et al. Trovafloxacin-induced acute hepatitis. Clin Infect Dis 2000; 30:400–1.

544

Moseley

189. Chen HJ, Bloch KJ, Maclean JA. Acute eosinophilic hepatitis from trovafloxacin. N Engl J Med 2000; 342:359–60. 190. Lazarczyk DA, Goldstein NS, Gordon SC. Trovafloxacin hepatotoxicity. Dig Dis Sci 2001; 46:925–6. 191. Bertino J, Jr., Fish D, Bertino J, Fish D. The safety profile of the fluoroquinolones. Clin Ther 2000; 22:798–817. 192. Liguori MJ, Anderson LM, Bukofzer S, et al. Microarray analysis in human hepatocytes suggests a mechanism for hepatotoxicity induced by trovafloxacin. Hepatology 2005; 41:177–86. 193. Buchweitz JP, Ganey PE, Bursian SJ, Roth RA. Underlying endotoxemia augments toxic responses to chlorpromazine: is there a relationship to drug idiosyncrasy? J Pharmacol Exp Ther 2002; 300:460–7. 194. Roth RA, Luyendyk JP, Maddox JF, Ganey PE. Inflammation and drug idiosyncrasy—is there a connection? J Pharmacol Exp Ther 2003; 307:1–8. 195. Ganey PE, Luyendyk JP, Maddox JF, Roth RA. Adverse hepatic drug reactions: inflammatory episodes as consequence and contributor. Chem Biol Interact 2004; 150:35–51. 196. Waring JF, Liguori MJ, Luyendyk JP, et al. Microarray analysis of LPS potentiation of trovafloxacininduced liver injury in rats suggests a role for proinflammatory chemokines and neutrophils. J Pharmacol Exp Ther 2006; 316:1080–7. 197. Cadle RM, Mansouri MD, Darouiche RO. Vancomycin-induced elevation of liver enzyme levels. Ann Pharmacother 2006; 40:1186–9. 198. D’Arcy PF. Nitrofurantoin. Drug Intell Clin Pharm 1985; 19:540–7. 199. Goldstein LI, Ishak KG, Burns W. Hepatic injury associated with nitrofurantoin therapy. Am J Dig Dis 1974; 19:987–98. 200. Sharp JR, Ishak KG, Zimmerman HJ. Chronic active hepatitis and severe hepatic necrosis associated with nitrofurantoin. Ann Intern Med 1980; 92:14–19. 201. Black M, Rabin L, Schatz N. Nitrofurantoin-induced chronic active hepatitis. Ann Intern Med 1980; 92:62–4. 202. Sippel PJ, Agger WA. Nitrofurantoin-induced granulomatous hepatitis. Urology 1981; 18:177–8. 203. Stricker BH, Blok AP, Claas FH, Van Parys GE, Desmet VJ. Hepatic injury associated with the use of nitrofurans: a clinicopathological study of 52 reported cases. Hepatology 1988; 8:599–606. 204. Klemola H, Penttila O, Runeberg L, Tallqvist G. Anicteric liver damage during nitrofurantoin medication. Scand J Gastroenterol 1975; 10:501–5. 205. Fagrell B, Strandberg I, Wengle B. A nitrofurantoin-induced disorder simulating chronic active hepatitis. A case report. Acta Med Scand 1976; 199:237–9. 206. Stromberg A, Wengle B. Chronic active hepatitis induced by nitrofurantoin. Br Med J 1976; 2:174–5. 207. Iwarson S, Lindberg J, Lundin P. Nitrofurantoin-induced chronic liver disease. Clinical course and outcome of five cases. Scand J Gastroenterol 1979; 14:497–502. 208. Holmberg L, Boman G, Bottiger LE, Eriksson B, Spross R, Wessling A. Adverse reactions to nitrofurantoin. Analysis of 921 reports. Am J Med 1980; 69:733–8. 209. Paiva LA, Wright PJ, Koff RS. Long-term hepatic memory for hypersensitivity to nitrofurantoin. Am J Gastroenterol 1992; 87:891–3. 210. Engel JJ, Vogt TR, Wilson DE. Cholestatic hepatitis after administration of furan derivatives. Arch Intern Med 1975; 135:733–5. 211. Merino G, Jonker JW, Wagenaar E, van Herwaarden AE, Schinkel AH. The breast cancer resistance protein (BCRP/ABCG2) affects pharmacokinetics, hepatobiliary excretion, and milk secretion of the antibiotic nitrofurantoin. Mol Pharmacol 2005; 67:1758–64. 212. Merino G, van Herwaarden AE, Wagenaar E, Jonker JW, Schinkel AH. Sex-dependent expression and activity of the ATP-binding cassette transporter breast cancer resistance protein (BCRP/ABCG2) in liver. Mol Pharmacol 2005; 67:1765–71. 213. Vavricka SR, Van Montfoort J, Ha HR, Meier PJ, Fattinger K. Interactions of rifamycin SV and rifampicin with organic anion uptake systems of human liver. Hepatology 2002; 36:164–72. 214. Tirona RG, Leake BF, Wolkoff AW, Kim RB. Human organic anion transporting polypeptide-C (SLC21A6) is a major determinant of rifampin-mediated pregnane X receptor activation. J Pharmacol Exp Ther 2003; 304:223–8. 215. Acocella G, Nicolis FB, Tenconi LT. The effect of an intravenous infusion of rifamycin SV on the excretion of bilirubin, bromsulphalein and indocyanine green in man. Gastroenterology 1965; 49:521–5. 216. Galeazzi R, Lorenzini I, Orlandi F. Rifampicin-induced elevation of serum bile acids in man. Dig Dis Sci 1980; 25:108–12. 217. Marschall HU, Wagner M, Zollner G, et al. Complementary stimulation of hepatobiliary transport and detoxification systems by rifampicin and ursodeoxycholic acid in humans. Gastroenterology 2005; 129:476–85. 218. Prince MI, Burt AD, Jones DE. Hepatitis and liver dysfunction with rifampicin therapy for pruritus in primary biliary cirrhosis. Gut 2002; 50:436–9. 219. Elmore M, Rissing JP, Rink L, Brooks GF. Clindamycin-associated hepatotoxicity. Am J Med 1974; 57:627–30.

Antibacterial and Antifungal Agents

545

220. Arbeit RD, Maki D, Tally FP, Campanaro E, Eisenstein BI, Daptomycin 98-01 and 99-01 Investigators. The safety and efficacy of daptomycin for the treatment of complicated skin and skin-structure infections. Clin Infect Dis 2004; 38:1673–81. 221. Echevarria K, Datta P, Cadena J, Lewis JS. Severe myopathy and possible hepatotoxicity related to daptomycin. J Antimicrob Chemother 2005; 55:599–600. 222. Stevens DL, Dotter B, Madaras-Kelly K. A review of linezolid: the first oxazolidinone antibiotic. Expert Rev Anti Infect Ther 2004; 2:51–9. 223. Rubinstein E, Prokocimer P, Talbot GH. Safety and tolerability of quinupristin/dalfopristin: administration guidelines. J Antimicrob Chemother 1999; 44:37–46. 224. Linden PK, Bompart F, Gray S, Talbot GH. Hyperbilirubinemia during quinupristin–dalfopristin therapy in liver transplant recipients: correlation with available liver biopsy results. Pharmacotherapy 2001; 21:661–8. 225. Carnecchia BM, Kurtzke JF. Fatal toxic reaction to amphotericin B in cryptococcal meningoencephalitis. Ann Intern Med 1960; 53:1027–36. 226. Miller MA. Reversible hepatotoxicity related to amphotericin B. Can Med Assoc J 1984; 131:1245–7. 227. Gill J, Sprenger HR, Ralph ED, Sharpe MD. Hepatotoxicity possibly caused by amphotericin B. Ann Pharmacother 1999; 33:683–5. 228. Meunier F, Prentice HG, Ringden O. Liposomal amphotericin B [AmBisome]: safety data from a phase II/III clinical trial. J Antimicrob Chemother 1991; 28(Suppl. B):83–91. 229. Ellis M, Shamoon A, Gorka W, Zwaan F, al-Ramadi B. Severe hepatic injury associated with lipid formulations of amphotericin B. Clin Infect Dis 2001; 32:E87–9. 230. Fischer MA, Winkelmayer WC, Rubin RH, Avorn J. The hepatotoxicity of antifungal medications in bone marrow transplant recipients. Clin Infect Dis 2005; 41:301–7. 231. Wingard JR, Leather H. Hepatotoxicity associated with antifungal therapy after bone marrow transplantation. Clin Infect Dis 2005; 41:308–10. 232. Mora-Duarte J, Betts R, Rotstein C, et al. Comparison of caspofungin and amphotericin B for invasive candidiasis. N Engl J Med 2002; 347:2020–9. 233. Maertens J, Raad I, Petrikkos G, et al. Efficacy and safety of caspofungin for treatment of invasive aspergillosis in patients refractory to or intolerant of conventional antifungal therapy. Clin Infect Dis 2004; 39:1563–71. 234. Bennett JE. Echinocandins for candidemia in adults with neutropenia. N Engl J Med 2006; 355:1154–9. 235. Garcia Rodriguez LA, Duque A, Castellsague J, Perez-Gutthann S, Stricker BH. A cohort study on the risk of acute liver injury among users of ketoconazole and other antifungal drugs. Br J Clin Pharmacol 1999; 48:847–52. 236. Bradbury BD, Jick SS. Itraconazole and fluconazole and certain rare, serious adverse events. Pharmacotherapy 2002; 22:697–700. 237. Stricker BH, Blok AP, Bronkhorst FB, Van Parys GE, Desmet VJ. Ketoconazole-associated hepatic injury. A clinicopathological study of 55 cases. J Hepatol 1986; 3:399–406. 238. Janssen PA, Symoens JE. Hepatic reactions during ketoconazole treatment. Am J Med 1983; 74:80–5. 239. Chien RN, Yang LJ, Lin PY, Liaw YF. Hepatic injury during ketoconazole therapy in patients with onychomycosis: a controlled cohort study. Hepatology 1997; 25:103–7. 240. Lewis JH, Zimmerman HJ, Benson GD, Ishak KG. Hepatic injury associated with ketoconazole therapy. Analysis of 33 cases. Gastroenterology 1984; 86:503–13. 241. Duarte PA, Chow CC, Simmons F, Ruskin J. Fatal hepatitis associated with ketoconazole therapy. Arch Intern Med 1984; 144:1069–70. 242. Bercoff E, Bernuau J, Degott C, et al. Ketoconazole-induced fulminant hepatitis. Gut 1985; 26:636–8. 243. Lake-Bakaar G, Scheuer PJ, Sherlock S. Hepatic reactions associated with ketoconazole in the United Kingdom. Br Med J 1987; 294:419–22. 244. Kim T-H, Kim B-H, Kim Y-W, et al. Liver cirrhosis developed after ketoconazole-induced acute hepatic injury. J Gastroenterol Hepatol 2003; 18:1426–9. 245. Como JA, Dismukes WE. Oral azole drugs as systemic antifungal therapy. N Engl J Med 1994; 330:263–72. 246. Munoz P, Moreno S, Berenguer J, Bernaldo de Quiros JCL, Bouza E. Fluconazole-related hepatotoxicity in patients with acquired immunodeficiency syndrome. Arch Intern Med 1991; 151:1020–1. 247. Franklin IM, Elias E, Hirsch C. Fluconazole-induced jaundice. Lancet 1990; 336:565. 248. Wells C, Lever AM. Dose-dependent fluconazole hepatotoxicity proven on biopsy and rechallenge. J Infect 1992; 24:111–2. 249. Jacobson MA, Hanks DK, Ferrel LD. Fatal acute hepatic necrosis due to fluconazole. Am J Med 1994; 96:188–90. 250. Trujillo MA, Galgiani JN, Sampliner RE. Evaluation of hepatic injury arising during fluconazole therapy. Arch Intern Med 1994; 154:102–4.

546

Moseley

251. Gearhart MO. Worsening of liver function with fluconazole and review of azole antifungal hepatotoxicity. Ann Pharmacother 1994; 28:1177–81. 252. Bronstein JA, Gros P, Hernandez E, Larroque P, Molinie C. Fatal acute hepatic necrosis due to dosedependent fluconazole hepatotoxicity. Clin Infect Dis 1997; 25:1266–7. 253. Schottker B, Dosch A, Kraemer DM. Severe hepatotoxicity after application of desloratadine and fluconazole. Acta Haematol 2003; 110:43–4. 254. Guillaume MP, De Prez C, Cogan E. Subacute mitochondrial liver disease in a patient with AIDS: possible relationship to prolonged fluconazole administration. Am J Gastroenterol 1996; 91:165–8. 255. Tucker RM, Haq Y, Denning DW, Stevens DA. Adverse events associated with itraconazole in 189 patients on chronic therapy. J Antimicrob Chemother 1990; 26:561–6. 256. Hay RJ. Risk/benefit ratio of modern antifungal therapy: focus on hepatic reactions. J Am Acad Dermatol 1993; 29:S50–4. 257. Lavrijsen AP, Balmus KJ, Nugteren-Huying WM, Roldaan AC, van’t Wout JW, Stricker BH. Hepatic injury associated with itraconazole. Lancet 1992; 340:251–2. 258. Gallardo-Quesada S, Luelmo-Aguilar J, Guanyabens-Calvet C. Hepatotoxicity associated with itraconazole. Int J Dermatol 1995; 34:589. 259. Adriaenssens B, Roskams T, Steger P, Van Steenbergen W. Hepatotoxicity related to itraconazole: report of three cases. Acta Clin Belg 2001; 56:364–9. 260. Gupta AK, Chwetzoff E, Del Rosso J, Baran R. Hepatic safety of itraconazole. J Cutan Med Surg 2002; 6:210–3. 261. Potoski BA, Brown J. The safety of voriconazole. Clin Infect Dis 2002; 35:1273–5. 262. Johnson LB, Kauffman CA. Voriconazole: a new triazole antifungal agent. Clin Infect Dis 2003; 36:630–7. 263. Spellberg B, Rieg G, Bayer A, Edwards JE, Jr. Lack of cross-hepatotoxicity between fluconazole and voriconazole. Clin Infect Dis 2003; 36:1091–3. 264. Hall M, Monka C, Krupp P, O’Sullivan D. Safety of oral terbinafine: results of a postmarketing surveillance study in 25,884 patients. Arch Dermatol 1997; 133:1213–9. 265. Zapata Garrido AJ, Romo AC, Padilla FB. Terbinafine hepatotoxicity. A case report and review of literature. Ann Hepatol 2003; 2:47–51. 266. Fernandes NF, Geller SA, Fong TL. Terbinafine hepatotoxicity: case report and review of the literature. Am J Gastroenterol 1998; 93:459–60. 267. Gupta AK, del Rosso JQ, Lynde CW, Brown GH, Shear NH. Hepatitis associated with terbinafine therapy: three case reports and a review of the literature. Clin Exp Dermatol 1998; 23:64–7. 268. Lazaros GA, Papatheodoridis GV, Delladetsima JK, Tassopoulos NC. Terbinafine-induced cholestatic liver disease. J Hepatol 1996; 24:753–6. 269. van’t Wout JW, Herrmann WA, de Vries RA, Stricker BH. Terbinafine-associated hepatic injury. J Hepatol 1994; 21:115–7. 270. Lowe G, Green C, Jennings P. Hepatitis associated with terbinafine treatment. Br Med J 1993; 306:248. 271. Agarwal K, Manas DM, Hudson M. Terbinafine and fulminant hepatic failure. N Engl J Med 1999; 340:1292–3. 272. Mallat A, Zafrani ES, Metreau JM, Dhumeaux D. Terbinafine-induced prolonged cholestasis with reduction of interlobular bile ducts. Dig Dis Sci 1997; 42:1486–8. 273. Anania FA, Rabin L. Terbinafine hepatotoxicity resulting in chronic biliary ductopenia and portal fibrosis. Am J Med 2002; 112:741–2. 274. Ajit C, Suvannasankha A, Zaeri N, Munoz SJ. Terbinafine-associated hepatotoxicity. Am J Med Sci 2003; 325:292–5. 275. Vermes A, Guchelaar HJ, Dankert J. Flucytosine: a review of its pharmacology, clinical indications, pharmacokinetics, toxicity and drug interactions. J Antimicrob Chemother 2000; 46:171–9. 276. Record CO, Skinner JM, Sleight P, Speller DC. Candida endocarditis treated with 5-fluorocytosine. Br Med J 1971; 1:262–4. 277. Libbrecht L, Meerman L, Kuipers F, Roskams T, Desmet V, Jansem P. Liver pathology and hepatocarcinogenesis in a long-term mouse model of erythropoietic protoporphyria. J Pathol 2003; 199:191–200. 278. Chiprut RO, Viteri A, Jamroz C, Dyck WP. Intrahepatic cholestasis after griseofulvin administration. Gastroenterology 1976; 70:1141–3. 279. Berman A, Franklin RL. Precipitation of acute intermittent porphyria by griseofulvin therapy. JAMA 1965; 192:1005–7. 280. Song JC, Deresinski S. Hepatotoxicity of antifungal agents. Curr Opin Investig Drugs 2005; 6:170–7.

26

Hepatotoxicity of Antituberculosis Drugs Sumita Verma and Neil Kaplowitz

Keck School of Medicine, University of Southern California, Los Angeles, California, U.S.A.

INTRODUCTION Tuberculosis (TB) is a major public health concern in developing countries. Nearly one-third of the world’s population is infected and this disease kills almost three million people per year, more than any other infectious agent in the world. In the mid-1980s, there was a resurgence of outbreaks in developed countries like the United States, largely due to the human immunodeficiency virus (HIV) epidemic, the development of drug-resistant strains, and the continued legal and illegal immigration from endemic areas. Since 1953, when Center for Disease Control (CDC) initiated public health surveillance for TB in the United States, the TB case rate has declined almost 10-fold from 53 cases per 100,000 to 5.1 per 100,000 in 2003. However, the 1.3% decline from 2002 to 2003 is the smallest yearly decline since 1992. Furthermore, for the first time since 1989, TB deaths increased by 5% in 2002. This data indicates that TB remains a major health problem even in developed countries like the United States (1). The first specific drug for TB became available in 1944 when the American microbiologist Selman Abraham Waksman discovered streptomycin (SM). This was followed by the development of p-aminosalisylate (PAS) (1949), isoniazid (INH) (1952), pyrazinamide (PZA) (1954), ethambutol (ETB) (1962), and rifampin (RMP) (1963). Of the four most commonly used first-line drugs for TB (INH, RMP, PZA, ETB), three are potentially hepatotoxic (INH, RMP, PZA). Table 1 lists the antitubercular drugs being used in the United States and Table 2 indicates the recommended treatment schedules (2). Though SM is as efficacious as ETB, because of increasing frequency of resistance to this drug, it is no longer considered first-line therapy. In those with latent TB infection (LTBI) (Table 2), RMP and PZA for two months had been considered for those where long-term compliance is an issue (2). However, this latter regimen is associated with an increased risk of severe hepatotoxicity [4.5% discontinued therapy because they either developed aspartate aminotransferase (AST) Ofive times the upper limit of normal (ULN) or symptoms of hepatitis] and is now not routinely recommended by either the CDC or the American Thoracic Society (ATS) (3). In addition, because of increased risk of hepatotoxicity from INH with older age, treatment for LTBI with INH should be restricted to those less than 35 years of age and only prescribed in older subjects if risks for reactivation of TB are significant (2,4). In the first few years after introduction of INH, only sporadic cases of hepatotoxicity were noted and were often attributed to concurrent viral infections or other drugs such as PAS (5–7). In 1959, Berte et al. reported an excellent hepatic safety profile of INH with no cases of hepatitis occurring in 513 treated patients (8). Subsequently, in 1963 the ATS recommended that one year of INH prophylaxis should be offered to all tuberculin-positive patients irrespective of age or duration of tuberculin positivity (9). Concerns regarding INH hepatotoxicity were first brought to attention in 1969 (17 years after INH was introduced) by Scharer et al. who observed liver test abnormalities in 10.3% of their cohort receiving INH (10). This failed to attract attention until 1971, when Garibaldi et al. retrospectively analyzed data on 2321 subjects receiving INH prophylaxis (following an outbreak on Capitol Hill), and reported clinical hepatitis in 19 (0.81%), of whom 13 developed overt jaundice with two deaths (0.08%), though mortality amongst those with hepatitis was 2 out of 19 (10%) (11). This unsettling data prompted the United States Public Health Service (USPHS) to initiate a large prospective multicenter

548

Verma and Kaplowitz

TABLE 1 Antitubercular Drugs Being Used in the United States First-line drugs

Second-line drugs Amikacina Capreomycin Cycloserine Ethionamide Gatifloxacina Levofloxacina Moxifloxacina p-Aminosalicylic acid Streptomycin

Ethambutol Isoniazid Pyrazinamide Rifabutina Rifampin Rifapentine

a

Not approved by FDA for treatment of tuberculosis in the United States. Source: Adapted from Ref. 2.

surveillance study to determine the incidence of INH-related hepatotoxicity. In this study probable INH related hepatitis was defined as AST O250 Karmen units, or AST !250 Karmen units but ALT OAST and negative HBV serology, with no other apparent cause of hepatitis. Possible hepatitis was defined as AST !250 Karmen units, or AST O250 Karmen units in the presence of other causes of liver disease, or AST O250 Karmen units, but lacking other biochemical tests. The overall rate of hepatitis in 13,838 subjects was 1.25% (total of 174 probable and possible cases) with eight deaths (0.06%) due to acute liver failure. Black females accounted for five of the eight deaths (12). The unexpectedly high mortality resulted in termination of the study. Both Garibaldi et al.’s study and the one commissioned by USPHS finally brought to attention the hepatotoxic potential of INH. INCIDENCE OF HEPATOTOXICITY It is difficult to be certain of the precise incidence of antituberculosis therapy (ATT)-induced hepatotoxicity as different authors have used varying criteria to diagnose drug-induced liver injury (DILI). In fact very few studies have followed the WHO guidelines in which DILI is classified as mild (transaminases !five times ULN), moderate (transaminases 5–10 times ULN) and severe (transaminases O10 times ULN) (13). Also in many of these studies, other causes of abnormal liver tests such as acute viral hepatitis, other hepatotoxins and biliary tract disease were not always diligently excluded (14). Finally, the occurrence of DILI with ATT appears to be higher in developing countries (8–10%) (14,15) compared to developed (4%) countries (16) probably due to higher prevalence of viral hepatitis and malnutrition (14,17,18), although genetic factors cannot be excluded. TABLE 2 Recommended ATT Schedule for TB Clinical situation Active TB Initial phase Continuation phaseb Latent tuberculosis infection TB in a patient with HIV

a

Drug therapy INHCRMPCPZACETBa for 2 mo INHCRMP for 4 mo for most patients INHCRMP for 7 mo if: cavitatory TB with positive sputum after 2 mo therapy Initial phase did not include PZA (if underlying liver disease) INH for 9 mo RMP with or without INH for 4 mo Managed in a similar fashion, though the continuation phase INH/RMP is recommended either daily as in a non-HIV patient or at least thrice weekly; remember drug interactions between RMP and antiretroviral agents

If and when drug susceptibility results are known and the organism in not INH resistant, then ETB need not be added to the initial phase. In children, as visual acuity cannot always be accurately monitored, ETB is not recommended unless there is strong clinical suspicion of infection being caused by INH resistant organisms or if the child has adult type TB (upper lobe cavitation). b The continuation phase may be given daily or twice /thrice weekly by directly observed therapy. Abbreviations: TB, Tuberculosis; INH, isoniazid; RMP, rifampin; PZA, pyrazinamide; ETB, ethambutol. Source: Adapted from Ref. 2.

549

Hepatotoxicity of Antituberculosis Drugs

TABLE 3 Incidence of Hepatotoxicity with the Various ATT Regimens Drug 1. 2. 3. 4.

INH alone Multidrug INH regimens without RMP Regimens containing RMP and not INH Regimens containing both INH and RMP

Number evaluated

Clinical hepatitis number (%)

38,257 2053 1264 6105

82 (0.6) 33 (1.6) 14 (1.1) 156 (2.5)

Abbreviations: ATT, antituberculosis therapy; INH, isoniazid; RMP, rifampin. Source: Adapted from Ref. 5.

Despite these limitations, the incidence of hepatotoxicity associated with multidrug regimens and INH monotherapy is best illustrated in an excellent meta-analysis performed by Steele et al. (5) which included 34 studies (22 in adults and 12 in children) published between 1966 and 1989. Only clinical trials or surveys of public health departments, where strict hepatitis criteria were stated (elevated bilirubin, clinical manifestations of hepatitis in conjunction with AST O100 units/dL), were included. Table 3 shows the incidence of clinical hepatitis with the various regimens, though the authors were unable to stratify incidence of hepatitis according to age, severity of underlying illness, and alcohol use due to lack of data. The overall incidence of hepatotoxicity in regimens containing INHCRMP was significantly higher than those that had INH without RMP (pZ0.04) or RMP without INH (pZ0.008). However, the incidence of clinical hepatitis was similar in regimens that contained INH without RMP or RMP without INH (pZ0.87) (Table 3). In children, the incidence of hepatitis in INHalone group was 0.2%, in multidrug INH regimens without RMP it was 1%, and in INHCRMP group it was 6.9%. Thus, in both adults and children, hepatitis with ATT was significantly higher with regimens that contained both INH and RMP, rather than when these drugs were used separately (3). A British study published a few years later involving 1317 subjects reported the incidence of hepatitis as follows: INH 0.3%, RMP 1.4%, and PZA 1.25% (19). However, since PZA was utilized for only two months and INH and RMP being used for six months or longer, the hepatitis rate per patient month was three times higher with PZA compared to RMP, and five times higher than for INH. This indicates that PZA probably has the most hepatotoxic potential of the drugs used for ATT (20). A recent study of INH monotherapy using AST monitoring (Ofive times ULN) revealed age-dependent risk of hepatotoxicity of 0.44% for ages 25 to 34, 0.85% for ages 35 to 49, and 2.08% for 50 years old or older. (21). MECHANISM OF HEPATOTOXICITY The pathogenesis of INH hepatotoxicity is not fully understood and is probably related to a nonallergic idiosyncratic reaction. Drug hypersensitivity is unlikely because of the variable and often prolonged time lag between exposure to the drug and the onset of symptoms and also because most patients can be rechallenged with INH without recurrence of the hepatotoxicity (22). Absence of fever, eosinophilia, and rash also argue against a hypersensitivity phenomenon. However, in rare cases, there is evidence of an allergic phenomenon in the form of prominent eosinophils on liver biopsy and development of hepatotoxicity on rechallenge (23,24). The other and more widely accepted possibility is a direct toxic effect of the drug or its metabolite. Both overt and subclinical hepatotoxicity is not related to serum INH or acetylhydrazine concentrations. However, the idiosyncratic nature of INH toxicity reflects the unique features of INH metabolism, which along with response to injury, probably dictates who will develop subclinical or overt toxicity. Thus, toxicity does not need to exhibit a relation to dose or blood level concentrations as the toxicity occurs at normal levels when a downstream toxic pathway is favored. Figure 1 illustrates the metabolism of INH. It is first acetylated into acetylisoniazid by the polymorphic N-acetyltransferase (NAT2), which is then hydrolyzed into acetylhydrazine and isonicotinic acid. Acetylhydrazine can then be further acetylated (again by NAT2) to a stable metabolite, diacetylhydrazine, or it can be hydrolyzed (by an amidase) to hydrazine. A small proportion of INH is directly hydrolyzed by the amidase to isonicotinic acid and hydrazine (Fig. 1) (utilization of this pathway is greater in slow acetylators) (25,26).

550

Verma and Kaplowitz

O Acetylisoniazid O NH-NH-C-CH3 C

Isoniazid O C NH-NH2 1 N 2

N

Isonicotic acid OH O C

N NH2-NH2 Hydrazine

1 2

1.N acetyl transferase (NAT) 2. Amidase 3. CYP reductase vs. CYP (not clear which)

2

O NH2-NH-C-CH3 Acetylhydrazine 1 O

O

CH3-C-NH-NH-C-CH3 Diacetylhydrazine 3 Toxic metabolite FIGURE 1 Metabolism of isoniazid. Source: From Ref. 25.

Hydrazine therefore can be generated directly by hydrolysis of INH or indirectly by hydrolysis of acetylhydrazine, both sources involving the activity of the amidase (26). Hydrazine is a known hepatotoxin in animal models (27,28) and the most plausible mechanistic model of INH toxicity involves the conversion to hydrazine. It is unclear if hydrazine involved in toxicity is derived directly from INH or N-acetylhydrazine. Hydrazine appears to undergo conversion by NADPH CYP450 reductase to a nitrogen-centered radical or by CYP to a carbon-centered radical (29–31). Irrespective of the uncertainty of the ultimate toxic species and the relative roles of CYP versus reductase in its formation, strong support exists for the hydrazine as the proximate toxin. Cytochrome P450 2E1 (CYP2E1) increases hydrazine-induced hepatotoxicity in rats (32). The most reliable animal model of INH toxicity is the rabbit, when given repeated doses over two days hepatic necrosis develops (26,33). Plasma hydrazine but not acetylhydrazine or INH levels correlated with necrosis. Phenobarbitone pretreatment increased toxicity, suggesting the possible potentiation of INH hepatotoxicity with induction of CYP450 enzymes (28,33). In the rabbit model, inhibition of the amidase by bis-p-nitrophenyl phosphate (BNPP), protected against necrosis and steatosis (Fig. 2) (26). Glutathione (GSH) depletion is modest in this model and does not correlate with the toxicity—therefore no evidence for a role for GSH-induced detoxification of the toxic metabolite (26). Using an up-to-date “omic” approach to assess the effects of hydrazine in rats has provided some insights—e.g., upregulation of mRNA of GRP78 and downregulation of GSH and superoxide dismutase and genes related to lipid transport. Most of these changes were confirmed at the level of protein expression. Metabolite profiling (NMR of serum) revealed alterations in lipid and glucose metabolism (34). However, it is not certain if these omic profiles reflect pathways leading to injury or response to injury. Recently, the combination of INH–RMP (but not either one) caused liver injury in mice associated with apoptosis and steatosis. Moderate GSH depletion accompanied the injury and was associated with lipid peroxidation. Predepleting GSH worsened the injury and BNPP protected (35). However, it remains uncertain if GSH involvement is through direct detoxification of the hydrazine metabolite or through modulation of oxidative stress. This work also implicates mitochondria as an important target of hydrazine-induced oxidative stress, GSH depletion, and permeability

551

Hepatotoxicity of Antituberculosis Drugs

300 250

240 ± 81

200 ALT 150 62.8 ± 24

100 50 0

24.9 ± 2.3

Control

INH

INH + BNPP

FIGURE 2 Effect of amidase inhibitor on serum ALT levels. BNPP, bis-p-nitrophenyl phosphate. Abbreviations: ALT, alanine aminotransferase; BNPP, bis-p-nitrophenyl phosphate. Source: From Ref. 25.

transition. The model however is not very robust showing rather mild injury and more prolonged drug administration may unmask greater effects (only three-day treatment studied). RMP is metabolized in the liver mainly by desacetylation followed by glucuronidation and is excreted in large concentrations in the bile as desacetyl rifampin (36). The mechanisms underlying RMP hepatotoxicity are also unclear. An allergic reaction is a possibility but is probably responsible for only 1% to 3% of the cases (37). RMP does transiently increase serum bilirubin (mostly unconjugated), but this is related to competitive inhibition of bilirubin uptake or excretion at the level of the hepatocyte membrane and is not indicative of hepatotoxicity (38,39). This effect is more prominent in children and this may contribute to the increased incidence of jaundice in children compared with adults (5). It is however well established that INHCRMP combination is associated with greater risk of hepatotoxicity than either drug alone. During INH metabolism, reactive toxic metabolites are produced by oxidation of hydrazine by the microsomal CYP enzymes. Since RMP induces the microsomal enzymes, it could theoretically result in increased production of this toxic metabolite. This would account for both the enhanced and the more rapid hepatotoxicity when RMP is coadministered with INH. Though this attractive and plausible hypothesis was first postulated more than two decades ago, it remains to be confirmed (40–42). In addition, RMP also induces the metabolism of INH by the amidase (especially in slow acetylators) resulting in increased direct formation of hydrazine (43–45). It is nonetheless uncertain if the hepatotoxicty of the INHCRMP combination is additive or synergistic. The mechanism responsible for PZA hepatotoxicity also remains to be elucidated. Lack of hypersensitivity signs and symptoms argue against an allergic phenomenon. It may be a direct toxic effect of the drug as prolonged duration of therapy and higher doses increase risk of hepatotoxicity (40,46). At present, there is no data to support a deleterious effect of RMP on PZA as PZA is not metabolized by the CYP system, but by both a microsomal deaminase and xanthine oxidase and RMP does not appear to enhance any of these enzymes (40). Nonetheless as already discussed, the combination of RMP–PZA is associated with severe hepatotoxicity in treatment of LTBI (3). HEPATOTOXICITY OF INDIVIDUAL DRUGS Isoniazid This first-line TB drug shows bactericidal activity against both extra and intracellular organisms (40). Data from surveillance programs suggest that approximately 10% of subjects who receive INH monotherapy develop abnormal transaminases (usually !three times ULN). In most of these cases, the subjects report no symptoms and discontinuation of therapy is not warranted. Only about 1% of patients on INH develop the more serious form of hepatotoxicity-overt hepatitis. This occurs mostly within the first three months and can be as

552

Verma and Kaplowitz

35 30 25

Incidence of hepatotoxicity (%)

20 15 10 5 0

1

2

3

4

5 6 7 8 9 Months of therapy

10

11

12

FIGURE 3 Incidence of probable INH hepatotoxicity depending on age. Abbreviation: INH, isoniazid. Source: From Ref. 12.

early as one week after drug initiation (Fig. 3). In contrast to those with asymptomatic transaminase elevations, those with hepatitis are symptomatic with anorexia, nausea, vomiting, abdominal pain, and jaundice. The significance of these symptoms in someone on ATT cannot be overemphasized, and it is mandatory that the treating physician informs and regularly questions patients on ATT about these symptoms. Besides symptoms, other characteristic features of severe hepatitis include significant elevations in transaminases (O10 times ULN). Jaundice is a poor prognostic sign as it is associated with a high mortality (10%). Of those with clinical hepatitis, about 5% to 10% (0.05–0.1% of overall treated patients) may develop a fulminant course characterized by coagulopathy and hepatic encephalopathy for which an urgent referral to a liver transplant unit may be necessary (41,47,48). Overall death rates with INH toxicity are 14 of every 100,000 people (0.014%) started on preventive therapy and between 23.2 and 57.9 of 100,000 people (0.023–0.057%) completing therapy (12,48). However, pooled results of published studies where patients received INH chemoprophylaxis and were monitored according to the ATS guidelines revealed a much lower mortality: overall 0.0009% (2 of 202,497) and in those older than 35 years 0.002% (1 of 43,334) (49). Factors associated with increased mortality after INH therapy include increasing age, female gender, delayed onset of hepatitis (two months or more after treatment initiation), continuing INH use after onset of symptoms, and serum bilirubin O350 mmol/L (O17 times ULN) (10,41,48,50). Is hepatotoxicity with INH dose dependent? In two large studies, it did not appear to be so (51,52). However in another series, 9 out of 18 patients who developed fulminant liver failure after multidrug ATT regimens had received INH in higher than the recommended dose (10 mg/kg) (46). Pessayre et al. also reported six patients with fulminant hepatitis who had received INH in doses ranging from 9.5 to 19 mg/kg (53). Therefore, it appears that mild variations from the recommended dose of INH (5 mg/kg) may not increase risk of hepatotoxicity, but increasing dosage to around 10 mg/kg would probably incur higher risk of hepatotoxcity (40). Rifampin This drug is also bactericidal against intra- and extracellular organisms. Almost all cases of hepatotoxicity that have been associated with RMP have occurred in patients who were also prescribed INH and it is therefore difficult to distinguish between the potentially hepatotoxic effects of the different drugs. Though RMP monotherapy is rarely associated with hepatotoxicity, unintentional overdose has resulted in liver failure (54). In the few cases where liver injury has been attributed to RMP, it was associated with a cholestatic liver panel (elevated bilirubin and alkaline phosphatase). This contrasts to the hepatitis-like liver injury (elevated transaminases) seen with INH (40,55). Dosing schedule may be important as use of daily versus twice weekly RMP therapy (in patients who received both INHCRMP) resulted in higher incidence of hepatotoxicity in the former (21% vs. 5%) (15).

Hepatotoxicity of Antituberculosis Drugs

553

Pyrazinamide PZA is only active against intracellular organisms. Liver injury is the most common and serious side effect of PZA. It generally results in a hepatitis-like liver injury similar to INH. Fever, arthalgia, skin rashes, and eosinophila are usually absent. After its introduction in 1954, PZA was used in high doses (40–50 mg/kg) and asymptomatic elevation of transaminases and symptomatic hepatitis were observed in 20% and 10% respectively. (40,56). Fatal fulminant hepatitis was also reported, resulting in the drug being abandoned as first-line therapy for TB. More recently, PZA was reintroduced as first-line therapy (as the incidence of TB increased) to overcome problems related to resistant strains. However, the current trend is to use a lower dose (30 mg/kg) for a shorter duration (two months) (40). There is very little data on incidence of hepatotoxicity with PZA monotherapy. In most cases when hepatotoxicity is reported, PZA has been used as part of a mutidrug regimen. There is some evidence to suggest that addition of PZA to INHCRMP increases risk of hepatotoxicity. In a case-control study with 60 patients who had evidence of ATT-induced liver injury, use of PZA was more frequent in the hepatitis group (70% vs. 42%) (57). Durand et al. studied 18 patients with fulminant/subfulminant liver failure due to ATT of whom nine had received PZA (30 mg/kg). Those in the non-PZA group developed liver failure usually within two weeks of drug initiation with a good overall prognosis (spontaneous survival 8 out of 9). In the PZA group, two patterns were seen: one with onset within 15 days (similar to non PZA group) with a good prognosis, and a second pattern where the liver failure occurred later (18–244 days after initiation) associated with a dismal prognosis (overall survival 2 out of 9) (46). Continuation of PZA after first manifestation of hepatitis increases risk of fatal outcome (2,46). However, Snider et al. did not find increased risk of hepatotoxicity with addition of PZA to the ATT regimen (58). HEPATOTOXICITY WITH COMBINATION DRUGS As Steele et al. demonstrated in their meta-analysis, the incidence of hepatotoxicity is significantly higher in patients who receive combination therapy than those who receive these drugs individually (5). The incidence of hepatotoxicity varies from 3% in the United States, 4% in the UnitedKingdom, to approximately 11% in countries like India (16,58,59). However, it is difficult to accurately assess the contribution of each individual drug or if this toxicity is additive or synergistic. Increased transaminases are seen in approximately 10% of those that receive INH monotherapy, but this is increased to 20% if RMP is added. In addition, the hepatitis occurs sooner (after two rather than four weeks) (5,41,42,53). Use of daily and twice weekly RMP–PZA combination for treatment of LTBI is also associated with an increased risk of severe hepatotoxicity (3,60). From January 2000 to June 2002, CDC monitored data on hepatotoxicity associated with RMP–PZA, which were reported to CDC through June 2003. For surveillance purposes, severe liver injury was defined as leading to hospitalization or death. Of the total 7737 patients who started RMP–PZA for LTBI, 5980 (77%) received daily doses and 1757 received twice weekly doses. A total of 204 (2.6%) patients discontinued therapy because of AST O5 ULN. An additional 146 (1.9%) patients stopped therapy because of symptoms of hepatitis. There were 48 (0.6%) cases of severe liver injury, of whom 11 (23%) died (3). Finally, use of PZA in combination with INHCRMP is also associated with a higher incidence of hepatic adverse effects as was mentioned above (46,57). It is uncertain whether the hepatotoxic potential of INH and PZA is affected by coadministration of ETB and SM (53). RISK FACTORS FOR ATT HEPATOTOXICITY Table 4 lists the factors associated with increased risk of hepatotoxicity with ATT. Acetylator Status NAT2 plays an important role in the metabolism of INH. About 60% of Caucasians and Blacks and 20% of Asians (Chinese and Japanese) are slow acetylators (61). Some regard slow

554

Verma and Kaplowitz

TABLE 4 Factors Associated with Antituberculosis Therapy-Induced Hepatotoxicity Acetylator status Age Alcohol use Enzyme inducers other than rifampin Gender (including pregnancy) and race Genetic factors Malnutrition Underlying chronic liver disease (includes coinfection with hepatitis B, C, and HIV)

acetylators to be at increased risk of INH hepatotoxicity especially if INH is used with RMP (15). Others have reported DILI more often in fast acetylators (23,51). But subsequently, three studies showed no impact of acetylator status on risk of developing hepatotoxicity (59,62,63). This included a study from South India by Gurumurty et al. with about 3000 patients who received various INH-containing regimens (62). A possible hypothesis for lack of association between fast acetylators and hepatotoxicity may be that though fast acetylators form acetylhydrazine more rapidly than slow acetylators, they also tend to inactivate this precursor of the reactive metabolite more quickly into the stable metabolite diacetylhydrazine (40). However, all these above studies phenotyped (rather than genotyped) acetylator status, and this may be influenced by many extrinsic factors. To determine more accurately if acetylator status impacted the risk of hepatotoxicity, Huang et al. genotyped N-acetyltransferase 2 (NAT2) in over 200 patients on ATT. They observed slow acetylators to be at increased risk of hepatotoxicity after multidrug ATT (26% vs. 11%). In addition, once the slow acetylators developed hepatotoxicity, they were prone to develop more serious hepatic injury than the rapid acetylators (64). Overall, genetic polymorphism of NAT2 (slow acetylators) appears to be a risk factor for INH toxicity but alone cannot account for the idiosyncratic nature of this problem. The high frequency of slow acetylator status precludes use of genotyping in decisions regarding the use of INH. Age It is well established that INH hepatotoxicity is related to age and is very uncommon below the age of 20 years (Fig. 4) (12). A more recent study from Belgium corroborated these findings and observed that patients more than 60 years of age were significantly more likely to have

2.5 2 1.5

Incidence of hepatotoxicity (%)

1 0.5 0

<20

20–34 35–49 Age group

50–64

FIGURE 4 Incidence of INH hepatotoxicity depending on duration of therapy. Abbreviation: INH, isoniazid. Source: From Ref. 12.

555

Hepatotoxicity of Antituberculosis Drugs

TABLE 5 Age-Dependent Mortality Associated with Isoniazid Age (years) !20 20 34 35 49 50 64 O65

Percentage of all deaths 9.2 12.4 17.6 37.9 22.9

Source: Adapted from Ref. 48.

abnormal transaminases after administration of INH and RMP compared to their younger counterparts (38% vs. 18%) (65). A Danish study and a recent report from the U.S. Public Health Service also observed older age to be a predictor of INH hepatotoxicity (21,66). Finally, mortality due to INH-induced liver injury is also age dependent as Snider et al. observed in their review of 177 INH-related deaths (Table 5) (48). It should be noted that the association of hepatotoxicity with age has also been observed when INH is used in combination with other hepatotoxic TB drugs (46). Alcohol Use It has been suggested that alcohol abusers are more likely to develop hepatotoxicity both with multidrug antitubercular regimens and INH monotherapy (67). Pande et al. observed high alcohol intake in 19.8% of those with ATT hepatotoxicity compared to 4.9% of the controls (18). The association between alcohol excess and ATT is not surprising as firstly, alcohol-dependent patients may be GSH depleted, [although there is no evidence that GSH detoxifies the reactive metabolite of INH, the suggestion that hydrazine depletes mitochondrial GSH is a point of convergence (35)], and secondly, chronic alcohol abuse induces the CYP2E1 enzyme, which could further enhance the formation of the toxic metabolite (38). However, some studies have shown that alcoholics are not at greater risk for hepatotoxicity from ATT or INH monotherapy (21,66–68). Enzyme Inducers Other Than RMP There are a number of other drugs that induce hepatic CYP enzymes increasing the risk of hepatotoxicity with INH. Fulminant hepatic failure has been reported in patients who initiated therapy with INH and RMP after receiving general anesthesia with halothane (53). It has been reported that induction of CYP2E1 by INH increases risk of acetaminophen toxicity and acute liver failure has also been observed in patients on ATT who were prescribed therapeutic doses of acetaminophen (!4 g/d) (69–71). In Nolan et al.’s series, all three patients were on multidrug ATT, including RMP, which is also a known inducer of the microsomal enzymes (40). Crippin et al.’s case report involved a 21-year old Asian female who had been on INH chemoprophylaxis for approximately six months (liver panel was normal at five months), and after ingesting only 3.25 g of acetaminophen the patient developed acute liver failure (AST O20,000 U/L, bilirubin 159 mmol/L, PT 26 seconds). Treatment with N-acetylcysteine was started and a referral initiated for a liver transplant evaluation, but she improved clinically and biochemically over the next week (70). Murphy et al. reported on acetaminophen-induced hepatotoxicity after ingestion of 11.5 g by a patient being treated with 300 mg INH daily (71). Thus, when hepatotoxicity develops in a patient receiving INH, careful assessment of acetaminophen use is critical. Accurate attribution of cause (INH vs. acetaminophen) may be very challenging. Towering transaminases (O3000 U/L) may be a clue pointing to acetaminophen. The employment of a test for serum acetaminophen–protein adducts shows promise as a means of verifying the role of acetaminophen (72). We recommend that patients receiving INH limit acetaminophen to 2 g/day or less, similar to the recommendation for alcoholics.

556

Verma and Kaplowitz

Gender Women, especially if pregnant or elderly, appear to be at increased risk of developing ATTrelated hepatotoxicity (14,42). This is a near-universal observation in DILI. Snider et al. in their review of 177 INH-associated deaths, reported a higher prevalence of women (69%) with 38% of the deaths occurring in the postpartum period (48). In a large series of pregnant women, incidence of INH-induced hepatitis and death were twofold and fourfold higher respectively compared to nonpregnant women (73). This apparently increased mortality in women could be related to the fact that a larger number of women than men were being treated, that women were more compliant, had better follow-up especially during pregnancy, and finally could be due to gender-related variations in drug metabolism. However, in a series of 20 INH-related deaths reported from California, 80% (of whom 25% were postpartum) were women and, even taking into account the fact that more women than men received INH, the difference in mortality was still striking (74). Thus, it appears that the increased INH-related mortality in women (especially if pregnant) is real and not spurious. It is therefore recommended that pregnant women who require INH should receive it based on their prepregnancy body weight, in order to minimize the risk of overdose (40), and be monitored closely. Race/Ethnicity A study done in 1975 with more than 14,000 subjects showed that ATT-induced hepatitis occurred in 55 out of 5190 whites (1.05%), 36 out of 6140 blacks (0.58%) and 23 out of 2608 (0.88%) Asians. Because a smaller percentage of Asians dropped out of the study (34%) compared to whites (52%) or blacks, the risk appeared to be the lowest in the Asians. Black females appeared to be at particularly high risk (12,41). A more recent study in patients with HIV suggested non-white race/ethnicity to be associated with increased risk of hepatotoxicity (75). Other Genetic Factors During metabolism of INH, CYP2E1 may play an important role in the oxidation of hydrazine to toxic metabolites. This enzyme exhibits genetic polymorphisms in intron 6 and its three genotypes are classified as c1/c1 (wild-type), c1/c2, and c2/c2 (76). Huang et al. performed a study in over 300 patients to assess whether genetic polymorphism of CYP2E1 influenced susceptibility to ATT hepatotoxicity. After adjustment for acetylator status and age, the CYP2E1 c1/c1 genotype was an independent risk factor for hepatotoxicity (77). The mutant c2/c2 genotype (25% of Asian population) was somewhat protective. Basal CYP2E1 is not affected by this polymorphism and its functional significance is unclear; the 2E1 activity (phenotype) was less inhibited by INH in c1/c1 patients. It is also possible that this gene is in linkage disequilibrium with other genes. Other genetic factors that have been studied include the major histocompatability complex class II alleles. Sharma et al. observed that absence of HLADQA1*0102 and presence of HLA-DQB1*0201 were independent risk factors for ATT-induced liver injury (78). Malnutrition In studies from India, a higher incidence of INHCRMP hepatotoxicity has been reported in malnourished patients (18). Singh et al. also reported lower BMI in those who developed hepatotoxicity (17.2 vs. 19.5) (57). Drug metabolizing processes in the liver including acetylation or GSH detoxification are altered in states of protein energy malnutrition. In fact, a significant decrease in INH metabolism has been observed in Kwashiorkor (79). Another reason poor nutritional status may be a risk factor for ATT hepatotoxicity is the tendency for such patients to receive higher doses of the drug on body weight basis (57). Underlying Chronic Liver Disease Including Coinfection with Hepatitis B and C Gronhagen-Riska et al. observed that approximately 50% of patients who developed moderate increases in transaminases (O150 U/L) after INH-RMP therapy had a history of either alcohol excess or underlying chronic liver disease (42) and this was corroborated by Kopanoff et al. (12).

Hepatotoxicity of Antituberculosis Drugs

557

There is also a case series of three patients with primary biliary cirrhosis who initiated RMP as an antipruritic agent. All developed significant hepatitis and impairment of synthetic function within two months of initiation of therapy, necessitating liver transplantation in one patient (80). In liver disease, metabolism of both INH and RMP is impaired resulting in increased bioavailability and serum elimination half-life (36). Furthermore, Acocelia et al. have also observed kinetic interactions between INH and RMP in patients with cirrhosis (81). These data indicate that these drugs should be avoided in patients with moderate to severe liver disease, or if that is not possible, then dose reduction should be considered (36,81). Presence of chronic hepatitis B infection (HBV) appears to increase the risk for ATT hepatotoxicity. Following INHCRMP therapy, Wu et al. observed higher peak ALT (1353 U/L vs. 885 U/L) and bilirubin (18.1 mg/dl vs. 4.4 mg/dl) levels, incidence of fulminant /subacute liver failure and mortality (47% vs. 4%) in patients who were HBV carriers than in controls. The carriers (all were HBcIgM negative) also developed the hepatotoxicity after significantly longer duration of therapy (110 days vs. 52 days) (Fig. 3) (82). A study in Chinese patients with chronic HBV (of whom 26% were HBeAg positive) found a higher prevalence (26.3% vs. 8.8%) of DILI (transaminases O1.5 ULN or O1.5 times baseline) and more severe histological injury in carriers compared to noncarriers, even after excluding those with elevated pretreatment transaminases and periportal interface hepatitis (a direct HBV effect) (83). Another interesting observation in this and Wu et al.’s study (82) was that episodes of hepatotoxicity were preceded by increase in HBV DNA levels. This raises the issue of using prophylactic antiviral agents prior to the introduction of ATT in patients with chronic HBV infection, but begs the question if the toxicity is really due to ATT or to reactivation of HBV due to the ATT (since untreated HBV controls had far less spontaneous hepatitis). In a more recent and the largest such study from Korea, Lee et al. report on 110 inactive HBsAg carriers who were treated with INH, RMP, ETB, and or PZA. All had normal pretreatment transaminases, were HBeAg negative, HBeAb positive, and had less than 105 copies/ml of HBV DNA. Thirty-five percent of the HBV carriers versus 20% controls developed increase in transaminases (pZ0.01). Moderate to severe hepatotoxicity (transaminases O5 ULN) also occurred more frequently in the carriers (8% vs. 2%, pZ0.05), though ATT could be safely reintroduced in over 50% of the carriers (17). Unfortunately, possible reactivation of HBV was not assessed in this study. Chemoprophylaxis with INH in young HBV carriers (!35 years of age) does not appear to be associated with increased risk of hepatotoxicity (84). Coinfection with hepatitis C (HCV) and HIV may also increase the risk of ATT hepatitis. Ungo et al. showed that the relative risk of ATT hepatotoxicity in patients being treated for TB was fivefold and fourfold respectively in presence of HCV or HIV. The risk was increased more than 14 times if both viruses were present. Four of these patients (who had had recurrence of DILI upon rechallenge with ATT) underwent a liver biopsy, which revealed active inflammation, attributed at least in part to HCV. They were treated with interferon and once the liver panel improved were able to undergo reintroduction of ATT without reoccurrence of liver injury (85). Therefore, as with coinfection with HBV, the interpretation of liver injury with chronic HCV remains a murky area; is it the ATT, reactivation of the virus or improved immune system due to control of the TB infection leading to attack on the virus? There is no certain answer at present and probably all are true. Such patients can therefore present a diagnostic dilemma (DILI vs. viral reactivation) and sometimes the only diagnostic test will be to treat the underlying virus before reinitiating ATT. However, as observed with chronic HBV infection, use of INH monotherapy as chemoprophylaxis is not associated with an increased risk of hepatotoxicty in presence of chronic HCV infection (86). CLINICAL, BIOCHEMICAL, AND HISTOLOGICAL FEATURES About 20% of patients develop abnormal transaminases with multidrug ATT though serious hepatotoxicity (AST, ALT O5!ULN) occurs in up to 4% (16). Common symptoms include fatigue, anorexia, nausea, jaundice, and right upper quadrant pain. Severe cases can develop fulminant or subacute liver failure characterized by hepatic encephalopathy and coagulopathy. Ascites is usually absent as are features of drug hypersensitivity (fever, drug rash, eosinophilia). Symptoms precede jaundice and liver failure by only a few days. Symptoms are a reliable

558

Verma and Kaplowitz

indicator of hepatotoxicity when present, but their absence does not negate the possibility of DILI. In a study of 72 consecutive patients with ATT-induced hepatitis from India, 61% presented with jaundice, 39% experienced nausea, vomiting, and abdominal discomfort with only 1 out of 77 (1.3%) reporting a skin rash or fever; 12(15%) developed fulminant/subacute liver failure of whom 75% died (59). It is of paramount importance that patients on ATT be regularly questioned about these symptoms, as continuing drugs in presence of existing liver damage can worsen the liver injury and ultimate prognosis. However, we recommend ALT monitoring in patients receiving multidrug regimens (see below). Hepatotoxicity with INH monotherapy usually develops within the first few months of therapy with the peak incidence being two months after initiation of treatment. With multiple drug ATT (INH, RMP, PZA) two patterns of hepatotoxicity can be recognized. The first is characterized by elevated transaminases occurring within the first two weeks and probably being related to RMP-induced hepatotoxicity of INH. This short interval between drug initiation and hepatotoxicity is associated with a relatively good prognosis even in the presence of fulminant hepatic failure. The second pattern consists of hepatotoxicity that occurs more than four weeks after starting ATT. As discussed above, this is associated with a poorer prognosis. Most cases of late hepatotoxicity appear to be associated with concomitant use of PZA. However, this classification of ATT-induced hepatotoxicity into an early and late type is not absolute, and it is imperative that all ATT be suspended in the presence of moderate to severe liver injury (40,46). Biochemically, ATT-induced hepatotoxicity is generally characterized by a “hepatitislike” picture with elevated transaminases (usually !1000 U/L). If the transaminase level is less than five times ULN, the hepatotoxicity is considered mild, 5 to 10 times ULN moderate, and greater than 10 times ULN severe (13). The clinical pattern however can be a “mixed” one, with evidence of increased bilirubin (as high as O10 times ULN) and alkaline phosphatase. However, if there is disproportionate elevation of bilirubin or alkaline phosphatase, that is indicative of RMP hepatotoxicity (13,20,55). Those with underlying liver disease can develop more marked abnormalities in the liver panel (42). In an earlier study from California reporting on 20 INH-related deaths (from 1973 to 1986), transaminases and bilirubin as high as 3685 U/L and 46 mg/dl respectively were observed (60% had ASTOALT). However, in 12 of the 19 cases where information was available, the patient had an underlying condition or was receiving a substance associated with fatty change in the liver; this included two or three who were consuming alcohol daily, two who were given anesthesia containing barbiturates while on INH, one or two who used therapeutic doses of acetaminophen, and one who was on long-term tetracycline (74). Clinically and biochemically, it can be difficult to distinguish ATT hepatotoxicity from acute viral hepatitis, and there have been reports of acute viral hepatitis being misdiagnosed as ATT hepatotoxicity (87). Therefore, a negative viral serology is essential before diagnosing DILI. It is also important to exclude other hepatotoxins (e.g., acetaminophen) and biliary tract disease by performing an ultrasound examination of the liver and gall bladder. Histologically, ATT-induced hepatotoxicity is characterized by focal necrosis and or confluent necrosis and portal inflammation (53,83). In the 20 fatal cases reported from California, the histological changes were more marked with 39% and 61% showing histological evidence of submassive and massive hepatic necrosis respectively, though as indicated above other hepatotoxins may have been involved (74). MANAGEMENT INCLUDING REFERRAL FOR LIVER TRANSPLANT Approximately 20% of those treated with standard four-drug regimen will develop asymptomatic elevations in transaminases (5). As long as the transaminases are less than 5 times ULN (in the absence of symptoms) there is no need to stop therapy, though the patient needs to be monitored frequently (see below). However, if transaminases are greater than five times ULN in absence of symptoms or greater than or equal to three times ULN in presence of symptoms, most hepatologists recommend that ATT should be stopped immediately and we strongly concur (88). Development of jaundice or elevations in alkaline phosphatase also is a cause for concern. Continuation of therapy in presence of these clinical and biochemical features has been associated with a poorer prognosis (41,46,59).

Hepatotoxicity of Antituberculosis Drugs

559

TABLE 6 Kings College Criteria to List Patients for Liver Transplant in Non-Acetaminophen-Induced Liver Failure 1. Prothrombin time greater than 100 sec (irrespective of grade of encephalopathy) or 2. Any three of the following criteria in presence of hepatic encephalopathy: Age less than 11 yr or greater than 40 yr Etiology of non-A/non-B hepatitis, halothane hepatitis, or idiosyncratic drug reactions (e.g., ATT) Duration of jaundice for more than seven days before onset of encephalopathy Prothrombin time greater than 50 sec Serum bilirubin level greater than 17 mg/dL (300 mmol/L) Abbreviation: ATT, antituberculosis therapy. Source: From Ref. 89.

After cessation of therapy, most patients will develop spontaneous resolution of the biochemical abnormalities after which ATT can be cautiously recommenced. However, a small proportion may develop worsening jaundice and confusion with abnormal INR. If this unfortunate course of events does occur, it is recommended that the patient be referred to a center where expertise for liver transplantation is available. The most widely used criteria to list patients for a liver transplant after ATT-induced liver injury are those that have been developed by Kings College, London (Table 6) (89). In a recent study, Russo et al. analyzed data from United Network for Organ Sharing, and observed that between 1990 and 2002, 15% (nZ370) of liver transplants performed for acute liver failure were due to drug-induced hepatotoxicity. Of the 270 subjects in whom complete data were available, a single drug was implicated in 258 (96%) with the remainder having multidrug associated DILI. Acetaminophen was the most common drug responsible (46%) followed by INH (17.5%), propylthiouracil (9.5%), phenytoin (7.3%), and valproate (7.3%). Of the 24 with INH-induced liver failure, 67% were women and 33% were African American (only 10% were African American in the acetaminophen group) (90). These data indicate that INH remains a very important and common cause of acute liver failure in the United States. ALTERNATIVE THERAPY FOR UNDERLYING TB AND REINTRODUCTION OF ATT Depending on the urgency of the need for ATT, it may be reasonable to wait for the liver tests to improve before a rechallenge. However, if that is not an option, then there will be a need to use regimens with no/minimal hepatotoxic potential. Such a combination might include SM, ETB, a fluoroquinolone, and another second-line oral drug. However, there are no data on choice, duration, or clinical effectiveness of such therapy. Expert opinion indicates that such a regimen needs to be continued for approximately 18 to 24 months, in case the original ATT regimen cannot be reintroduced. Once the transaminases return to less than 2 times ULN (or in those with underlying liver disease return to baseline levels), the original ATT can be slowly reinstituted. Because RMP is much less likely to cause hepatotoxicity compared to INH and PZA and is the most effective agent, it should be started first (5,2,91). It can be begin at doses of 75 mg/day, increased to 300 mg/day after two to three days and then 450 mg/day (!50 kg) or 600 mg/day (O50 kg) after a further two to three days. If the liver panel remains normal, INH can be reinstituted (starting with a low dose of 50 mg, increasing after every two–three days to 300 mg) and after one more week PZA can be restarted [starting at 250 mg, increasing to 1000 mg after two–three days and then to 1500 mg (!50 kg) or 2000 mg (O50 kg)]. If symptoms recur or the transaminases again become abnormal, the last drug should be stopped. If INH and RMP are well tolerated and the hepatotoxicity was severe, then PZA should be the presumed offending agent and not restarted. In such a situation, it may be necessary to extend the usual six-month therapy with INHCRMP to nine months (2). Though ATT can be reintroduced safely in the vast majority, there does remain a risk of recurrence of severe hepatitis. However, most experts agree that rechallenge should be attempted since non-INH/RMP ATT regimens are problematic due to uncertainty regarding duration of therapy and clinical efficacy, and the development of resistance (92). Singh et al. were able to safely reintroduce therapy with INH and RMP (in appropriate doses calculated according to body weight) in 35 out of 41 patients, with 6 (15%) redeveloping liver injury. Of these six, successful reintroduction was possible at the second attempt in three of them (59). It is unclear why hepatitis does not occur in the

560

Verma and Kaplowitz

majority after rechallenge. One reason may be the improved general condition as these patients had received ATT for some time with possible reduction in bacterial load. Another could be that the phased reintroduction may have reduced the hepatotoxic potential (92), promoting an adaptive protective mechanism. We would recommend however that the original ATT should not be reintroduced in the following clinical setting: (1) if the liver injury was a life threatening event (jaundice along with INR O1.5 times ULN and or encephalopathy), or (2) in patients with moderately severe underlying liver disease (Child B/C cirrhosis). ATT IN PATIENTS WITH UNDERLYING LIVER DISEASE Treatment of TB in patients with unstable or advanced liver disease is problematic for a number of reasons. Firstly, the likelihood of developing hepatotoxicity may be greater (especially in presence of HBV and HCV coinfection). Secondly, due to the poor hepatic reserve the liver injury may be life threatening. Thirdly, it may be impossible to differentiate symptoms of ATT hepatotoxicity from those of the underlying liver disease, and finally TB may itself cause abnormalities in the liver tests (2). Because of the effectiveness of the first-line ATT, these remain the drugs of choice in spite of the underlying liver disease. However, these patients need close monitoring (see the list given below). The following regimens can be used (2), though there have been no firm recommendations as to the choice: 1. Though the frequency of PZA hepatitis is less than INH, it may be more severe, so one could utilize INH, RMP, and ETM for two months followed by INH/RMP for seven more months. In our opinion, this is probably a reasonable approach when chronic liver disease is well compensated (Child Class A). 2. For those with more advanced disease (Child’s B or C) or even compensated cirrhosis, only one potentially hepatotoxic drug is retained, usually RMP, with the addition of EMB, fluoroquinolone, cycloserine, and injectable agents. The duration of such therapy should be 12 to 18 months. 3. As an alternative to (2) in those with severe liver disease, it may be best to avoid all hepatotoxic agents. The regimen recommended is SM, ETB, a fluoroquinolone and another second-line oral drug and total duration of therapy needs to be extended to 18 to 24 months. Overall, we recommend avoiding PZA in patients with underlying liver disease and avoiding INH if chronic liver disease is clinically severe, but would not argue with the judgement towards INH if chronic liver disease is well compensated. HEPATOTOXICITY OF ATT IN LIVER TRANSPLANT RECIPIENTS TB is a significant opportunistic infection in transplant recipients. The frequency of Mycobacterium tuberculosis infection after liver transplantation varies from 0.9% to 2.3%, but may be as high as 15% in areas of high endemicity (93). Approximately 50% of these patients develop disseminated TB (94). Mortality in liver transplant recipients from TB ranges from 21% to 33%, with death directly attributable to M. tuberculosis in 57% to 83% (93–95). It is recommended that liver transplant candidates, with a tuberculin reactivity of greater than or equal to 5 mm or recent conversion, receive chemoprophylaxis with INH (94). This is generally initiated in the immediate posttransplant period as immunosuppression is maximum at that time period. In such a clinical situation, it may be difficult to differentiate INH hepatotoxicity from other causes of abnormal liver tests such as rejection. However, liver biopsies have suggested the latter to be a more likely diagnosis (94,96). Overall hepatotoxicity requiring discontinuation of therapy occurs in 25% to 41% of liver transplant recipients, and this exceeds the rate of INH hepatotoxicity in renal (2.5%) and heart or lung (4.5%) transplant recipients (93,94,96). An alternative approach may be to use INH chemoprophylaxis in the pretransplant period, thereby eliminating diagnostic confusion with graft rejection and reducing

Hepatotoxicity of Antituberculosis Drugs

561

the likelihood of premature discontinuation. Obviously, this can be problematic in patients with decompensated cirrhosis. Nevertheless, this would also avoid drug interactions between ATT and immunosuppressive agents, which could potentially impact graft survival (96). In those liver transplant recipients with active TB, use of first-line multidrug antitubercular regimens can be problematic. Meyers et al. describe the outcome in nine liver transplant recipients with TB (of whom 78% had disseminated disease) who received ATT (INH, RMP, PZA, ETB). Overall, during induction therapy, histological features consistent with INH/druginduced hepatotoxicity developed in more than 80%, with 50% showing additional features of rejection. These patients were not rechallenged with the original ATT; instead continuation therapy was provided by ETM and ofloxacin (97). RECOMMENDATIONS FOR MONITORING PATIENTS WHILE ON ATT Monitoring patients on INH monotherapy or multidrug ATT has remained controversial (2,40,89,92), and we offer our opinion here. Monotherapy with INH for LTBI 1. If the patient is less than 35 years old, monthly interrogation for symptoms at time of drug refill; ALT monitoring is unnecessary as need for hospitalization for symptomatic hepatitis is less than 1:10,000. This is a widely accepted approach. 2. If the patient is more than 35 years old, we recommend adding monthly ALT monitoring with the usual stopping rules (ALT Ofive times ULN or ALT Othree times ULN with symptoms). If ALT is greater than three times ULN without symptoms, weekly ALT monitoring is needed until resolution or worsening (discontinue therapy). We acknowledge that this is more likely to be an acceptable strategy to hepatologists than to infectious disease specialists, but in view of increased risks with increasing age, maximizes patient safety. 3. If patients have underlying hepatitis B, C, and/or HIV, monthly monitoring of serum ALT. 4. For liver transplant candidates who need treatment for LTBI, it may be best to initiate therapy with INH in the pretransplant period with close (twice weekly monitoring) of the liver panel. For Patients on Multidrug Regimens 1. It is useful to do a baseline evaluation of the liver panel in all patients prior to initiation of ATT. At the onset, the patients and their care providers must be clearly educated about the hepatotoxic potential of ATT and be given clear instructions to watch for symptoms such as nausea, vomiting, abdominal pain, and jaundice. At the first sign of these symptoms, patients must be instructed to immediately discontinue therapy and seek urgent medical attention. 2. ATT should be prescribed in conjunction with either a respiratory or an infectious disease specialist. 3. Alcohol abuse should be actively discouraged while ATT is being continued. 4. In those with no underlying liver disease and no risk factors for ATT hepatotoxicity, it is reasonable to write an initial prescription for one month. We would recommend monitoring liver panel every two weeks for the first eight weeks and then every four weeks until the completion of therapy. If the transaminases are elevated to less than 2 times ULN, it would be reasonable to continue to monitor the liver panel as above, i.e., every four weeks. If the transaminases are more than two times ULN but less than 5 times ULN, the ATT can be continued as long as there are no symptoms. These patients need the liver panel monitoring every week and then bimonthly until the transaminases stabilize. However, it is essential that treatment be stopped if: a. transaminases increase greater than or equal to five times ULN in the absence of symptoms, b. transaminases greater than or equal to three times ULN in the presence of symptoms,

562

Verma and Kaplowitz

c. Serum bilirubin greater than or equal to 1.5 times ULN (along with ALT Othree times ULN) or prothrombin increased greater than or equal to 1.5 times ULN irrespective of presence or absence of symptoms. Isolated increase in unconjugated serum bilirubin is usually a benign and transient phenomenon related to interference with bilirubin excretion by RMP. Conjugated hyperbilirubinemia (irrespective of absence or presence of symptoms) however mandates cessation of all ATT. Conjugated bilirubin can defined as more than 35% direct (total O1.5 times ULN) or bilirubinuria. The ATT can be restarted once bilirubin returns to normal and transaminases return to less than 2 times ULN. Because of insufficient evidence, no firm recommendations can be made for those who develop isolated or predominant increase in serum alkaline phosphatase. 5. Those with underlying liver disease or risk factors for ATT hepatotoxicity should receive not more than two weeks supply of drugs initially and liver tests should be monitored at two weekly intervals until end of therapy. 6. In those with baseline liver test abnormalities, we would recommend continuing therapy until transaminases increase to more than twice baseline and or jaundice or symptoms develop. If this occurs, ATT should be stopped and the liver panel monitored at weekly and then two weekly intervals. The drugs can be cautiously reintroduced once the transaminases approach baseline. 7. Treatment of TB in a liver transplant recipient is a complex situation and needs to be individualized. Most of these patients may not tolerate full duration of first-line antitubercular drugs. It may be reasonable to initiate such a therapy for the induction phase only and then convert to safer second-line drugs for continuation phase. Liver panel needs to be monitored every one to two weeks.

CONCLUSIONS TB remains the leading cause of death worldwide from a single infectious agent. Three of the first-line drugs for TB, while being highly effective antitubercular agents, are potentially hepatotoxic. With INH monotherapy about 10% develop asymptomatic elevations in transaminases with less than 1% developing more serious liver injury. With multiple drug regimens up to 4% (higher in the developing countries) may develop overt hepatitis. After acetaminophen, INH is the second most common cause of drug-induced acute liver failure in the United States. Once patients develop jaundice, the mortality is high (10%). Therefore, the decision to initiate INH monotherapy or mutidrug ATT must be made judiciously and under close supervision, especially in those with risk factors for ATT hepatotoxicity. It is imperative that patients be made aware of the danger symptoms of hepatotoxicity and instructed to stop all medication if such signs and symptoms develop. Initial prescriptions for combination ATT should be written for only two to four weeks and close monitoring of the liver panel is indicated throughout the duration of the therapy. Due to the hepatotoxic potential of standard ATT and drug resistance, about 20% of patients with TB do not receive the best possible treatment. Therefore new, safer drugs are urgently needed. Note added: The American Thoracic Society has recently published its official recommendations on ATT (98). Generally, their recommendations cover a range of options from monitoring only patients with liver disease, to including older than age 35, to no monitoring. Our recommendations largely fall within their range with ours being on the more cautious side with respect to risk of mono- and multidrug therapies. REFERENCES 1. CDC. Trends in tuberculosis-United States 1998–2003. MMWR 2004; 53:209–14. 2. Blumberg HM, Burman WJ, Chaisson RE, Daley CL, Etkind SC, Friedman LN, Fujiwara P, Grzemska M, Hopewell PC, Iseman MD, et al. American Thoracic Society. Centers for Disease

Hepatotoxicity of Antituberculosis Drugs

3. 4. 5. 6. 7. 8.

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

563

Control and Prevention and the Infectious Diseases Society. American Thoracic Society/Centers for Disease Control and Prevention/Infectious Diseases Society of America: treatment of tuberculosis. Am J Respir Crit Care Med 2003; 167:603–62. Update: Adverse event data and revised American Thoracic Society/CDC recommendations against use of rifampin and pyrazinamide for treatment of latent tuberculosis infection-United States, 2003. August 8, 2003/52(31);735–9. American Thoracic Society. Preventive therapy of tuberculosis infection. Am Rev Respir Dis 1974; 110:371–4. Steele MA, Burk RF, DesPrez RM. Toxic hepatitis with isoniazid and rifampin. A meta-analysis. Chest 1991; 99(2):465–71. Randolph H, Joseph S. Toxic hepatitis with jaundice occurring in a patient treated with isoniazid. JAMA 1953; 152(1):38–40. Berte SJ, Dewlett HJ. Isoniazid and para-aminosalicylic acid toxicity in 513 cases: a study including high doses of INH and gastrointestinal intolerance to PAS. Dis Chest 1959; 36(2):146–51. Berte SJ, Dimase JD, Christianson CS. Isoniazid, para-aminosalycylic acid, and streptomycin tolerance in 1,744 patients. An analysis of reactions to single drug and drug groups plus data on multiple reactions, type and time of reactions and desensitization. Am Rev Respir Dis 1964; 90:598–606. American Thoracic Society. Chemoprophylaxis for the prevention of tuberculosis; a statement by an ad-hoc committee. Am Rev Respir Dis 1965; 96:558–60. Scharer L, Smith JP. Serum transaminase elevations and other hepatic abnormalities in patients receiving isoniazid. Ann Intern Med 1969; 71(6):1113–20. Garibaldi RA, Drusin RE, Ferebee SH, Gregg MB. Isoniazid-associated hepatitis. Report of an outbreak. Am Rev Respir Dis 1972; 106(3):357–65. Kopanoff DE, Snider DE, Jr., Caras GJ. Isoniazid-related hepatitis: a U.S. Public Health Service cooperative surveillance study. Am Rev Respir Dis 1978; 117(6):991–1001. World Health Organization Collaborating Center for International Drug monitoring. Adverse drug reaction terminology (ART), 1979. (Accessed January 2006. at http://www.WHO-UMC.org) Turktas H, Unsal M, Tulek N, Oruc O. Hepatotoxicity of antituberculosis therapy (rifampicin, isoniazid and pyrazinamide) or viral hepatitis. Tuberc Lung Dis 1994; 75(1):58–60. Parthasarathy R, Sarma GR, Janardhanam B, et al. Hepatic toxicity in South Indian patients during treatment of tuberculosis with short-course regimens containing isoniazid, rifampicin and pyrazinamide. Tubercle 1986; 67(2):99–108. British Thoracic Association. A controlled trial of six month chemotherapy in pulmonary tuberculosis. First report: results during chemotherapy. Br J Dis Chest 1981; 75:141–53. Lee BH, Koh WJ, Choi MS, et al. Inactive hepatitis B surface antigen carrier state and hepatotoxicity during antituberculosis chemotherapy. Chest 2005; 127(4):1304–11. Pande JN, Singh SP, Khilnani GC, Khilnani S, Tandon RK. Risk factors for hepatotoxicity from antituberculosis drugs: a case-control study. Thorax 1996; 51(2):132–6. Ormerod LP. Analysis of frequency of drug reactions to antituberculosis drugs. Thorax 1994; 45:403–8. Ormerod LP. Hepatotoxicity of antituberculosis drugs. Thorax 1996; 51:111–3. Fountain FF, Tolley E, Chrisman CR, Self TH. Isoniazid hepatotoxicity associated with treatment of latent tuberculosis infection: a 7-year evaluation from a public health tuberculosis clinic. Chest 2005; 128(1):116–23. Maddrey WC. Isoniazid-induced liver disease. Semin Liver Dis 1981; 1(2):129–33. Mitchell JR, Thorgeirsson UP, Black M, et al. Increased incidence of isoniazid hepatitis in rapid acetylators: possible relation to hydranize metabolites. Clin Pharmacol Ther 1975; 18(1):70–9. Maddrey WC, Boitnott JK. Isoniazid hepatitis. Ann Intern Med 1973; 79(1):1–12. Sarich TC, Adams SP, Petricca G, Wright JM. Inhibition of isoniazid-induced hepatotoxicity in rabbits by pretreatment with an amidase inhibitor. J Pharmacol Exp Ther 1999; 289(2):695–702. Timbrell JA, Mitchell JR, Snodgrass WR, Nelson SD. Isoniazid hepatoxicity: the relationship between covalent binding and metabolism in vivo. J Pharmacol Exp Ther 1980; 213(2):364–9. Patrick RL, Black KC. Pathology and toxicity of repeated doses of hydrazine and 1,1-dimethylhydrazine in monkeys and rats. Ind Med Surg 1965; 34:430–5. Sarich TC, Zhou T, Adams SP, Bain AI, Wall RA, Wright JM. A model of isoniazid-induced hepatotoxicity in rabbits. J Pharmacol Toxicol Methods 1995; 34(2):109–16. Sarich TC, Adams SP, Wright JM. The role of L-thyroxine and hepatic reductase activity in isoniazidinduced hepatotoxicity in rabbits. Pharmacol Res 1998; 38(3):199–207. Sinha BK. Activation of hydrazine derivatives to free radicals in the perfused rat liver: a pin trapping study. Biochim Biophys Acta 1987; 924:261–9. Sinha BK, Motten AG. Oxidative metabolism of hydrazine. Evidence for nitrogen centered radicals formations. Biochem Biophys Res Commun 1982; 105(3):1044–51. Yue J, Peng RX, Yang J, Kong R, Liu J. CYP2E1 mediated isoniazid-induced hepatotoxicity in rats. Acta Pharmacol Sin 2004; 25(5):699–704.

564

Verma and Kaplowitz

33. Sarich TC, Youssefi M, Zhou T, Adams SP, Walls RA, Wright JM. The role of hydrazine in the mechanism of isoniazid hepatotoxicity in rabbits. Arch Toxicol 1996; 70:835–40. 34. Kleno TG, Kiehr B, Baunsgaard D, Sidelmann UG. Combination of “omics” data to investigate the mechanism(s) of hydrazine-induced hepatotoxicity in rats and to identify potential biomarkers. Biomarkers 2004; 9(2):116–38. 35. Chowdhury A, Santra A, Bhattacharjee K, Ghatak S, Saha DR, Dhali GK. Mitochondrial oxidative stress and permeability transition in isoniazid and rifampicin induced liver injury in mice. J Hepatol 2006; 45(1):117–26. 36. McConnell JB, Powell-Jackson PR, Davis M, Williams R. Use of liver function tests as predictors of rifampicin metabolism in cirrhosis. Q J Med 1981; 50(197):77–82. 37. Mulder De Jong MT, Mulder RJ. Drugs used in the treatment of tuberculosis and leprosy. In: Dukes MUG, ed. Meyler’s Side Effects of Drugs, Vol. 1. Oxford: Excerpta Medica, 1977:676–89. 38. Cohn HD. Clinical studies with a new rifamycin derivative. J Clin Pharmacol J New Drugs 1969; 9(2):118–25. 39. Kenwright S, Levi AJ. Sites of competition in the selective hepatic uptake of rifamycin-SV, flavaspidic acid, bilirubin, and bromsulphthalein. Gut 1974; 15(3):220–6. 40. Durand F, Jebrak G, Pessayre D, Fournier M, Bernuau J. Hepatotoxicity of antitubercular treatments. Rationale for monitoring liver status. Drug Saf 1996; 15(6):394–405. 41. Black M, Mitchell JR, Zimmerman HJ, Ishak KG, Epler GR. Isoniazid-associated hepatitis in 114 patients. Gastroenterology 1975; 69(2):289–302. 42. Gronhagen-Riska C, Hellstrom PE, Froseth B. Predisposing factors in hepatitis induced by isoniazidrifampin treatment of tuberculosis. Am Rev Respir Dis 1978; 118(3):461–6. 43. Sarma GR, Immanuel C, Kailasam S, Narayana AS, Venkatesan P. Rifampin-induced release of hydrazine from isoniazid. A possible cause of hepatitis during treatment of tuberculosis with regimens containing isoniazid and rifampin. Am Rev Respir Dis 1986; 133(6):1072–5. 44. Dickinson DS, Bailey WC, Hirschowitz BI. The effect of acetylation status on isoniazid (INH) hepatitis. Am Rev Respir Dis 1977; 115:395. 45. Weber WW, Hein DW, Litwin A, Lower GM, Jr. Relationship of acetylator status to isoniazid toxicity, lupus erythematosus, and bladder cancer. Fed Proc 1983; 42(14):3086–97. 46. Durand F, Bernuau J, Pessayre D, et al. Deleterious influence of pyrazinamide on the outcome of patients with fulminant or subfulminant liver failure during antituberculous treatment including isoniazid. Hepatology 1995; 21(4):929–32. 47. Mitchell JR, Zimmerman HJ, Ishak KG, Thorgeirsson UP, Timbrell JA, Snodgrass WR, Nelson SD. Isoniazid liver injury: clinical spectrum, pathology, and probable pathogenesis. Ann Intern Med 1976; 84(2):181–92. 48. Snider DE, Jr., Caras GJ. Isoniazid-associated hepatitis deaths: a review of available information. Am Rev Respir Dis 1992; 145(2 Pt 1):494–7. 49. Salpeter SR. Fatal isoniazid-induced hepatitis. Its risk during chemoprophylaxis. West J Med 1993; 159(5):560–4. 50. Farrell GC. Drug-Induced Liver Disease. Edinburgh, Scotland: Churchill Livingstone, 1994. 51. Yamamoto T, Suou T, Hirayama C. Elevated serum aminotransferase induced by isoniazid in relation to isoniazid acetylator phenotype. Hepatology 1986; 6(2):295–8. 52. Mitchell JR, Long MW, Thorgeirsson UP, Jollow DJ. Acetylation rates and monthly liver function tests during one year of isoniazid preventive therapy. Chest 1975; 68(2):181–90. 53. Pessayre D, Bentata M, Degott C, et al. Isoniazid-rifampin fulminant hepatitis. A possible consequence of the enhancement of isoniazid hepatotoxicity by enzyme induction. Gastroenterology 1977; 72(2):284–9. 54. Rothwell DL, Richmond DE. Hepatorenal failure with self-initiated intermittent rifampicin therapy. Br Med J 1974; 2(917):481–2. 55. Alford RH. Antimycobacterial agents. In: Mandell GL, Douglas RJ, Benett JE, eds. Principles and Practices of Infectious Diseases. 3rd ed. New York: Churchill Livingstone, 1990:350–61. 56. Mcdermott W, Ormond L, Muschenheim C, Deuschle K, McCune RM, Jr., Tompsett R. Pyrazinamideisoniazid in tuberculosis. Am Rev Tuberc 1954; 69(3):319–33. 57. Singh J, Arora A, Garg PK, Thakur VS, Pande JN, Tandon RK. Antituberculosis treatment-induced hepatotoxicity: role of predictive factors. Postgrad Med J 1995; 71(836):359–62. 58. Snider DE, Jr., Long MW, Cross FS, Farer LS. Six-months isoniazid-rifampin therapy for pulmonary tuberculosis. Report of a United States Public Health Service Cooperative Trial. Am Rev Respir Dis 1984; 129(4):573–9. 59. Singh J, Garg PK, Tandon RK. Hepatotoxicity due to antituberculosis therapy. Clinical profile and reintroduction of therapy. J Clin Gastroenterol 1996; 22(3):211–4. 60. McNeill L, Allen M, Estrada C, Cook P. Pyrazinamide and rifampin vs. isoniazid for the treatment of latent tuberculosis: improved completion rates but more hepatotoxicity. Chest 2003; 123(1):102–6.

Hepatotoxicity of Antituberculosis Drugs

565

61. Ellard GA. Variations between individuals and populations in the acetylation of isoniazid and its significance for the treatment of pulmonary tuberculosis. Clin Pharmacol Ther 1976; 19(5 Pt 2):610–25. 62. Gurumurthy P, Krishnamurthy MS, Nazareth O, et al. Lack of relationship between hepatic toxicity and acetylator phenotype in three thousand South Indian patients during treatment with isoniazid for tuberculosis. Am Rev Respir Dis 1984; 129(1):58–61. 63. Martinez-Roig A, Cami J, Llorens-Terol J, de la Torre R, Perich F. Acetylation phenotype and hepatotoxicity in the treatment of tuberculosis in children. Pediatrics 1986; 77(6):912–5. 64. Huang YS, Chern HD, Su WJ, et al. Polymorphism of the N-acetyltransferase 2 gene as a susceptibility risk factor for antituberculosis drug-induced hepatitis. Hepatology 2002; 35(4):883–9. 65. van den Brande P, van Steenbergen W, Vervoort G, Demedts M. Aging and hepatotoxicity of isoniazid and rifampin in pulmonary tuberculosis. Am J Respir Crit Care Med 1995; 152(5 Pt 1):1705–8. 66. Dossing M, Wilcke JT, Askgaard DS, Nybo B. Liver injury during antituberculosis treatment: an 11-year study. Tuber Lung Dis 1996; 77(4):335–40. 67. Cross FS, Long MW, Banner AS, Snider DE, Jr. Rifampin-isoniazid therapy of alcoholic and nonalcoholic tuberculous patients in a U.S. Public Health Service Cooperative Therapy Trial. Am Rev Respir Dis 1980; 122(2):349–53. 68. Schaberg T, Rebhan K, Lode H. Risk factors for side-effects of isoniazid, rifampin and pyrazinamide in patients hospitalized for pulmonary tuberculosis. Eur Respir J 1996; 9(10):2026–30. 69. Nolan CM, Sandblom RE, Thummel KE, Slattery JT, Nelson SD. Hepatotoxicity associated with acetaminophen usage in patients receiving multiple drug therapy for tuberculosis. Chest 1994; 105(2):408–11. 70. Crippin JS. Acetaminophen hepatotoxicity: potentiation by isoniazid. Am J Gastroenterol 1993; 88(4):590–2. 71. Murphy R, Swartz R, Watkins PB. Severe acetaminophen toxicity in a patient receiving isoniazid. Ann Intern Med 1990; 111:799–800. 72. Muldrew KL, James LP, Coop L, et al. Determination of acetaminophen-protein adducts in mouse liver and serum and human serum after hepatotoxic doses of acetaminophen using high-performance liquid chromatography with electrochemical detection. Drug Metab Dispos 2002; 30:446–51. 73. Franks AL, Binkin NJ, Snider DE, Jr., Rokaw WM, Becker S. Isoniazid hepatitis among pregnant and postpartum Hispanic patients. Public Health Rep 1989; 104(2):151–5. 74. Moulding TS, Redeker AG, Kanel GC. Twenty isoniazid deaths in one state. Am Rev Respir Dis 1989; 140:700–5. 75. Gordin FM, Cohn DL, Matts JP, Chaisson RE, O’Brien RJ. Terry Beirn Community Programs for Clinical Research on AIDS, Adult AIDS Clinical Trials Group, Centers for Disease Control and Prevention. Hepatotoxicity of rifampin and pyrazinamide in the treatment of latent tuberculosis infection in HIV-infected persons: is it different than in HIV-uninfected persons? Clin Infect Dis 2004; 39(4):561–5. 76. Stephens EA, Taylor JA, Kaplan N, et al. Ethnic variations in the CYP2E1 gene:polymorphism analysis of 695 African–Americans, European–Americans and Taiwanese. Pharmacogemetics 1994; 4:185–92. 77. Huang YS, Chern HD, Su WJ, et al. Cytochrome P450 2E1 genotype and the susceptibility to antituberculosis drug-induced hepatitis. Hepatology 2003; 37(4):924–30. 78. Sharma SK, Balamurugan A, Saha PK, Pandey RM, Mehra NK. Evaluation of clinical and immunogenetic risk factors for the development of hepatotoxicity during antituberculosis treatment. Am J Respir Crit Care Med 2002; 166(7):916–9. 79. Buchanan N, Eyberg C, Davis MD. Isoniazid pharmacokinetics in kwashiorkor. S Afr Med J 1979; 56(8):299–300. 80. Prince MI, Burt AD, Jones DE. Hepatitis and liver dysfunction with rifampicin therapy for pruritus in primary biliary cirrhosis. Gut 2002; 50(3):436–9. 81. Acocella G, Bonollo L, Garimoldi M, Mainardi M, Tenconi LT, Nicolis FB. Kinetics of rifampicin and isoniazid administered alone and in combination to normal subjects and patients with liver disease. Gut 1972; 13(1):47–53. 82. Wu JC, Lee SD, Yeh PF, et al. Isoniazid-rifampin-induced hepatitis in hepatitis B carriers. Gastroenterology 1990; 98(2):502–4. 83. Wong WM, Wu PC, Yuen MF, et al. Antituberculosis drug-related liver dysfunction in chronic hepatitis B infection. Hepatology 2000; 31(1):201–6. 84. McGlynn KA, Lustbader ED, Sharrar RG, Murphy EC, London WT. Isoniazid prophylaxis in hepatitis B carriers. Am Rev Respir Dis 1986; 134(4):666–8. 85. Ungo JR, Jones D, Ashkin D, et al. Antituberculosis drug-induced hepatotoxicity. The role of hepatitis C virus and the human immunodeficiency virus. Am J Respir Crit Care Med 1998; 157:1871–6. 86. Fernandez-Villar A, Sopena B, Vazquez R, et al. Isoniazid hepatotoxicity among drug users: the role of hepatitis C. Clin Infect Dis 2003; 36(3):293–8.

566

Verma and Kaplowitz

87. Kumar A, Misra PK, Mehotra R, Govil YC, Rana GS. Hepatotoxicity of rifampin and isoniazid. Is it all drug-induced hepatitis? Am Rev Respir Dis 1991; 143(6):1350–2. 88. Adams PC, Arthur MJ, Boyer T D, et al. Screening in liver disease: report of an AASLD clinical workshop. Hepatology 2004; 39(5):1204–12. 89. O’Grady JG, Alexander GJ, Hayllar KM, Williams R. Early indicators of prognosis in fulminant hepatic failure. Gastroenterology 1989; 97(2):439–45. 90. Russo MW, Galanko JA, Shrestha R, Fried MW, Watkins P. Liver transplantation for acute liver failure from drug induced liver injury in the United States. Liver Transpl 2004; 10(8):1018–23. 91. Thompson NP, Caplin ME, Hamilton MI, Gillespie SH, Clarke SW, Burroughs AK, McIntyre N. Antituberculosis medication and the liver: dangers and recommendations in management. Eur Respir J 1995; 8(8):1384–8. 92. Garg PK, Tandon RK. Antituberculous agents-induced liver injury. In: Deleve L, Kaplowitz N, Deleve L, Kaplowitz N, eds. Drug Induced Liver Disease. New York: Marcel Dekker Inc., 2002:505–17. 93. Aguado JM, Herrero JA, Gavalda J, et al. Clinical presentation and outcome of tuberculosis in kidney, liver, and heart transplant recipients in Spain. Spanish Transplantation Infection Study Group, GESITRA. Transplantation 1997; 63(9):1278–86. 94. Singh N, Wagener MM, Gayowski T. Safety and efficacy of isoniazid chemoprophylaxis administered during liver transplant candidacy for the prevention of post transplant tuberculosis. Transplantation 2002; 74(6):892–5. 95. Verma A, Dhawan A, Wade JJ, et al. Mycobacterium tuberculosis infection in pediatric liver transplant recipients. Pediatr Infect Dis J 2000; 19(7):625–30. 96. Singh N, Paterson DL. Mycobacterium tuberculosis infection in solid-organ transplant recipients: impact and implications for management. Clin Infect Dis 1998; 27(5):1266–77. 97. Meyers BR, Papanicolaou GA, Sheiner P, Emre S, Miller C. Tuberculosis in orthotopic liver transplant patients: increased toxicity of recommended agents; cure of disseminated infection with nonconventional regimens. Transplantation 2000; 69(1):64–9. 98. Saukkonen J, Cohn D, Jasmer R, et al. An official ATS Statement: hepatotoxicity of antituberculosis therapy. Am J Respir Crit Care Med 2006; 174:935–52.

27

Hepatic Toxicity of Antiviral Agents Ulrich Spengler

€ Bonn, Department of General Internal Medicine, Rheinische Friedrich Wilhelms Universitat Bonn, Germany

INTRODUCTION Viral infections still pose a considerable threat to the human population, because they can spread rapidly and only a few drugs with antiviral activity are as yet available. The recent development of multiple effective antiretroviral drugs to treat HIV infection, however, has illustrated the impressive therapeutic potential antiviral drugs can display. Now, the study of liver function during antiviral therapy of HIV-infected patients, especially those with concurrent hepatitis B and C coinfection, has become an emerging issue. The availability of highly active antiretroviral therapy (HAART) has led to a marked reduction of opportunistic diseases including hepatobiliary complications associated with severe immunodeficiency. However, short- and long-term liver toxicities have been reported for virtually each single antiretroviral agent and may add to the liver damage caused by the frequently underlying chronic hepatitis B and/or hepatitis C (1–3). Novel and potent pharmaceutical regimens are also available to effectively treat chronic hepatitis B and C, e.g., polyethylene glycol-coupled recombinant interferons and antiviral drugs directly interfering with viral replication such as ribavirin, lamivudine, adefovir, tenofovir, and entecavir. Still more antiviral drugs with activity against HIV or hepatitis B and C are expected to be licensed in the near future. The extensive and prolonged use of these drugs has the potential for interactions and may further increase the risk of liver toxicity (3–5). Liver toxicity of antiviral therapy is closely linked to substances from three groups of agents: nucleoside reverse transcriptase inhibitors (NRTI), non-nucleoside reverse transcriptase inhibitors (NNRTI) and protease inhibitors (PI) used for combination therapy of HIV (Table 1). Randomized controlled trials (RCTs) and longitudinal surveys of cohorts on antiviral treatment have frequently reported on elevated aminotransferases as well as on an increased incidence of hepatic injury, life-threatening hepatotoxic events and end-stage liver disease in patients (6–9).

PROBLEMS ASSOCIATED WITH THE ASSESSMENT OF HEPATOTOXICITY CAUSED BY ANTIVIRAL DRUGS The potential of antiviral drugs to cause liver damage has been increasingly recognized since 1994 (10). Although most authors report that 5% to 15% of patients on antiretroviral therapy develop elevated liver enzymes, estimates on the true incidence of hepatotoxicity are still inaccurate. Analysis of liver toxicity is hampered by the fact that toxicity has not been reported uniformly along international consensus guidelines, which suggests that severe liver damage should be characterized with respect to jaundice, coagulopathy, and encephalopathy (11). Instead, most studies evaluate the risk of hepatotoxicity by assessing asymptomatic elevations The author of this chapter has relationships with the following corporations: he has been a consultant to BMS and Novartis, and has received grants from Roche and Essex.

568

Spengler

TABLE 1 Overview of Currently Available Antiviral Drugs (Exclusive Interferons) Substance

Mode of action

Aciclovir Famciclovir Ganciclovir Foscarnet Ribavirin

NuA NuA NuA Pyrophosphate NuA

Oseltamivir Zanamivir Cidofovir Valganciclovir Zidovudine Lamivudine Abacavir Stavudine Zalcitabine Didanosine Emtricitabine Tenofovir Adefovir Entecavir Delavirdine Efavirenz Nevirapine Amprenavir Fosamprenavir Atazanavir Indinavir Lopinavir Nelfinavir Ritonavir Saquinavir Tipranavir Enfuvirtide

Activity

Dose

Hepatoxic potential

Herpesviruses Herpesviruses, HBV CMV CMV RSV, HCV

5!200 800 mg 2!250 mg i.v. 7.5 15 mg/kg i.v. 200 mg/kg 800 1200 mg

Inhibitor of Neuramidinase Inhibitor of Neuramidinase NuA NuA NuA, NRTI NuA, NRTI NuA, NRTI NuA, NRTI NuA, NRTI NuA, NRTI NuA, NRTI NutA, NRTI NuA, (NRTI) NuA NNRTI NNRTI NNRTI PI

Influenzaviruses

2!75 mg

Low Low Low Low Low, but interacts with other NuAs Low

Influenzaviruses

2!10 mg inhaled

Low

Herpesviruses Herpesviruses HIV HIV, HBV HIV HIV HIV HIV HIV, HBV HIV, HBV HBV, (HIV) HBV HIV HIV HIV HIV

Low Low Considerable Low Low Considerable Considerable Considerable Low Low Low Low Possible Possible Considerable Possible

PI PI PI PI PI

HIV HIV HIV HIV HIV

i.v. 5 mg/kg § 2!900 mg 2!250 mg 1!300 mg 2!300 mg 1 2!15 20 mg 3!0.75 mg 1!400 mg 1!200 mg 1!300 mg 1!10 mg 1!0.5 1 mg 3!400 mg 1!600 mg 2!200 mg 2!1200 mg 2!1400 mg 1!300 mg (Crtv) 2!800 mg (Crtv) 2!400 mg (Crtv) 2!1250 mg Baby dose 100 mg

PI PI Fusion inhibitor

HIV HIV HIV

2!1000 mg (Crtv) 2!500 mg (Crtv) s.c. 2!90 mg

Jaundice Jaundice Possible Possible High in full dose, low in baby dose Possible Considerable Low

A low risk was attributed, if a drug was only rarely associated with hepatic toxicity. A possible risk signifies hepatotoxicity that is encountered in a small proportion of patients and that does not significantly exceed the risk seen with other drugs from the same class. A considerable risk indicates drugs, which have a particularly prominent profile for liver toxicity, usually observed in the 10% range of treated patients. Jaundice refers to drugs, which frequently inhibit bilirubin excretion without actual liver toxicity. Abbreviations: NuA, nucleoside analogue; Nut, nucleotide analogue; NRTI, nucleoside analogue reverse transcriptase inhibitor; NNRTI, non-nucleoside analogue reverse transcriptase inhibitor; PI, protease inhibitor; CMV, cytomegalovirus; HBV, hepatitis B virus; HCV, hepatitis C virus; RSV, respiratory syncytial virus; (Crtv), Dose, if given in combination with baby dose (100 mg) ritonavir; i.v., intravenously; s.c., subcutaneously; §, To be given in 3 wk intervals to avoid renal toxicity.

in aminotransferases levels as a surrogate marker for the risk of developing clinical liver disease. However, the frequency of abnormal liver function tests is not a reliable predictor of rare, severe clinical hepatotoxicity. This is nicely illustrated in the comprehensive safety analysis of nevirapine-associated liver toxicity, which found the incidence of elevated aminotransferases [Rfive times the upper limit of normal (ULN)] to be 10% among 2545 nevaripine-treated patients studied in controlled trials (9). However, approximately two-thirds of these patients had been free of symptoms associated with severe liver damage, and fortunately fatal outcomes of liver toxicity occurred in four patients only. A grading system based on elevated aminotransferase levels has been proposed by the AIDS Clinical Trial Group (ACTG) (12) for use in its trials. Under this definition, severe hepatotoxicity (grade 3 or 4) is defined by increases of alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) to 5!ULN. However, these criteria have been modified by many other authors, mostly in an attempt to better account for the effects of elevated baseline aminotransferases. For example,

Hepatic Toxicity of Antiviral Agents

569

Sulkowski et al. (2,13) and Nun˜ez et al. (14) considered ALT/AST above 3.5 times the baseline level to define hepatotoxicity in patients with elevated baseline levels; Palmon et al. (15) used a threshold of five times the baseline level. Two studies based all changes on baseline levels (16,17) rather than ULN, while two further studies applied the ACTG criteria in combination with an absolute increase in aminotransferase levels greater than 100 U/L or 200 U/L (1,7). The source of information on liver toxicity may give rise to further bias in assessing the potential of an antiviral drug for hepatotoxicity. The relationship between antiviral therapy and liver toxicity has been established in RCTs, observational data base studies and cohort analyses. RCTs offer the most convincing evidence for the efficacy of new treatment options (18). The availability of a control group means that it is possible to directly assess the proportion of toxicity events which can be attributed to the new type of treatment. A further benefit of RCTs is that the frequency and standard of monitoring for hepatotoxicity is defined by the study protocol. However, it has been well documented that subjects who take part in clinical trials may be very different in terms of demographic data and lifestyle as compared to those who choose not to participate (19,20). Moreover, those patients at the highest risk for hepatotoxicity because of prior complex treatment histories, predisposing hepatic abnormalities or concomitant medications are frequently excluded from clinical trials (21,22) resulting in a study population with lower risk of liver toxicity than the general population. Finally, data on long-term toxicity may not always be captured in RCTs. There is a clear potential for bias if a patient with mild signs of hepatotoxicity switches treatment during an early stage before severe hepatotoxicity may become clinically manifest. One major benefit of observational data is that all patients who consented to have their data recorded can be included into the analysis. Thus, the experience with the drug is more likely to reflect a real-life setting. Although many clinics have protocols to monitor adverse events, the lack of a strict study protocol may lead to ascertainment bias, if patients with some type of treatment or those considered to be particularly at risk are monitored more closely than others. Finally, data on possible confounding factors such as the concomitant use of nonprescription hepatotoxic drugs, herbs, and alcohol are often recorded less robustly, if at all, in observational cohorts compared with RCTs. Establishing causality between liver damage and the drug under consideration is a further difficult issue, because most patients are simultaneously treated with several potentially hepatotoxic drugs. Of note, in the study reported by Martinez et al. (17) univariate analysis revealed that receiving stavudine was associated with a relative hazard of 2.29 (95% confidence interval 1.25–4.2) for experiencing a threefold increase in aminotransferases while taking nevirapine. Likewise, the combined use of nevirapine and stavudine was implicated in the development of severe hepatotoxicity in a study reported in abstract form by Sanne et al. (23). These data suggest that mitochondrial toxicity associated with the use of stavudine might have been partially responsible for the observed elevations in liver enzymes. Moreover, drug interactions, active drug abuse, and poorly understood interactions of the drug with concomitant liver disease may hamper establishing a firm proof of causality. For instance, in the Amsterdam cohort lamivudine withdrawal was associated with flares of aminotransferases in patients with hepatitis B virus (HBV) coinfection (7). Such flares were correlated with a rise in HBV viraemia in most instances, suggesting that the observed aminotransferase elevations reflected loss of control over HBV replication rather than drug toxicity. Likewise, viral breakthrough during treatment of chronic hepatitis B with novel nucleoside analogues may lead to massive and potentially fatal hepatic decompensation which must not be mistaken for hepatic drug toxicity (24). MECHANISMS OF DRUG TOXICITY Mitochondrial Toxicity Nucleoside analogues are designed to be incorporated into viral nucleic acids and to act as chain terminators, thus inhibiting efficient viral replication. NRTIs are prodrugs, which have to be activated intracellularly via three phosphorylation steps before they can inhibit the HIV reverse transcriptase (25). NRTI triphosphates also inhibit human g-polymerase, which is

570

Spengler

responsible for the replication of mitochondrial DNA (mtDNA) (26,27). Thus, the inhibition of g-polymerase by NRTIs leads to depletion of mtDNA, a small circular molecule normally present in multiple copies in each mitochondrion and in hundreds of copies in most human cells. This type of acquired mitochondrial toxicity might be caused by either (i) direct inhibition of the polymerase without incorporation; (ii) chain termination by incorporation of the analogues into mtDNA; (iii) altered proofreading fidelity of g-polymerase; (iv) persistence of incorporated analogues in mtDNA due to inefficient excision; or any combination thereof. In vitro, NRTIs reduced mtDNA content and altered mitochondrial morphology in a dose-dependent fashion which preceded or coincided with a decline in the expression of cytochrome oxidase II, a decline in cell growth, increased production of lactate and increased intracellular lipids (28). In vitro, the relative hierarchy of mitochondrial toxicities due to g-polymerase inhibition is as follows: zalcitabineZstavudineZdidanosine[lamivudineO tenofovirOazidothymidineOabacavir (25,28–30). In liver biopsies obtained from patients who had been treated with the so-called D-drugs didanosine, stavudine or zalcitabine, content of mtDNA was reduced by 47% (31). In contrast, non-D drugs, e.g., lamivudine, abacavir, adefovir, tenofovir as well as other nucleoside analogues inhibiting HBV DNA polymerase seem to exert negligible mitochondrial toxicity, because these compounds are poorly incorporated into mtDNA (32–35). Mitochondrial toxicity is dependent on the cellular concentration of nucleoside analogues. High-NRTI concentrations cause a more pronounced mtDNA depletion compared to low concentrations, and for several NRTIs recommended dosing may be close to the limit of tolerance concerning mitochondrial toxicity (36). Mitochondrial toxicity requires prolonged drug exposure, because changes in mitochondrial metabolism are only observed if depletion of mtDNA exceeds a certain threshold. As a consequence, the onset of mitochondrial toxicity is rare in the first few months of HAART. On the other hand, long-term NRTI exposure may lead to mitochondrial toxicity despite relatively low NRTI concentrations. Concomitant liver diseases such as hepatitis C coinfection, which in itself can lower hepatocellular content of mtDNA (37), increase the vulnerability of patients. There may be additive or synergistic mitochondrial toxicities if NRTIs are used in combination. This effect was more pronounced in the combinations of zalcitabine–stavudine, didanoside–ribavirin and azidothymidine–lamivudine, while toxicity of the combination didanoside–stavudine did not exceed the toxicity of didanoside alone (28,38). Mitochondrial toxicity is tissue specific. Tissue specificity is explained by the fact that cellular uptake of the NRTI-prodrugs, activation by intracellular phosphorylation and translocation into the mitochondria may vary between different cell types (27). Zidovudine is an exception because its triphosphate is only a weak inhibitor of g-polymerase. Nevertheless, severe depletion of mtDNA is found when zidovudine is combined with lamivudine. Furthermore, zidovudine can exert mitochondrial toxicity by other mechanisms. This toxicity is characterized by a reduction of cell growth despite normal levels of mtDNA and cytochrome oxidase II expression, while enormously enlarged mitochondria with a disrupted mitochondrial reticulum together with increased lactate and intracellular lipids can be observed (28). Of note, zidovudine can be converted into stavudine by a nonenzymatic reaction (39,40); this may explain why depletion of mtDNA has occasionally been reported for treatment with zidovudine. Zidovudine may bind to adenylate kinase, an enzyme involved in ATP formation (41). In addition, zidovudine has been shown to inhibit the ADP/ATP translocator in rat liver and heart tissue in vitro, which may impair oxidative phosphorylation (42,43). However, inhibition of the ADP/ATP translocator has not yet been reported in humans. Zidovudine monophosphate has been shown to inhibit the activities of human DNA poly-g exonuclease (29) and rat nucleoside diphosphate kinase (44). Inhibition of protein glycosylation (45) and early effects on protein synthesis have been proposed as further mechanisms underlying the mitochondrial toxicity of zidovudine (46,47). Finally, increased levels of oxidatively damaged DNA residues have been observed in livers of zidovudine-treated mice (48,49). Mitochondria provide most of the hepatocellular energy and are involved in fatty acid oxidation, the tricarboxylic acid cycle and oxidative phosphorylation. The biological function of mtDNA is to encode for most enzyme subunits of the respiratory chain, which is located in the inner mitochondrial membrane. The respiratory chain oxidatively synthesizes ATP and

Hepatic Toxicity of Antiviral Agents

571

consumes NADH and FADH as end products of fatty acid oxidation. Therefore, mtDNA depletion due to NRTIs leads to a defect in respiratory chain function and eventually inhibits oxygen-dependent catabolic pathways such as pyruvate and fatty acid oxidation. As a result, anaerobic glycolysis and lactate synthesis are increased, and impaired mitochondrial fatty acid oxidation leads to an intracellular increase of triglycerides and nonesterified fatty acids, the cause of microvesicular steatosis. Beyond that, normal function of the respiratory chain is also essential for the synthesis of DNA, because de novo synthesis of pyrimidine nucleosides depends on dihydroorotat-ubiquinone oxidoreductase (DHODH), an enzyme located in the inner mitochondrial membrane (50). Mitochondria are also the main source of intracellular reactive oxygen species. Interruption of electron transfer in the mitochondrial respiratory chain results in increased cellular release of reactive oxygen species, which severely augments hepatic tissue damage by initiating lipid peroxidation of accumulated fatty acids, by releasing proinflammatory cytokines, by inducing fibrosis or by additionally damaging both mitochondrial and nuclear DNAs (51–55). Finally, loss of mitochondrial integrity results in release of cytochrome C into the cytosol which activates caspases and triggers apoptosis (56,57). Thus, fulminant hepatic failure has become a serious complication of mitochondrial hepatotoxicity which has increasingly been recognised in the recent past (58,59). Electron microscopy can help to identify tissue damage due to mitochondrial toxicity by demonstrating enlarged mitochondria with matriceal densities and occasional vacuoles (60). In more advanced cases, NRTI-induced hepatic injury is readily recognized by light microscopy (Fig. 1) (61). Histology typically shows a combination of micro- and macrovesicular hepatic steatosis together with features of cholestasis, while lobular microarchitecture is preserved and lymphocytic infiltrates or hepatocellular necrosis are rare. NRTI-related mitochondrial hepatotoxicity is frequently associated with a couple of additional abnormalities. A typical complication is elevated serum lactate. Such hyperlactataemia was more frequently described with prolonged stavudine treatment (62,63), especially when combined with didanosine. The significance of asymptomatic hyperlactataemia is unclear. When hyperlactataemia is associated with symptoms, these are often nonspecific such as nausea, right upper quadrant abdominal tenderness or myalgias. Mitochondrial myopathy was first described in HIV patients on antiretroviral therapy with high-dose zidovudine (64). Myopathy may cause skeletal muscle weakness under static or dynamic exercise but serum levels of creatine kinase are often normal or only minimally elevated. This complication can be diagnosed by muscle biopsy. Prolonged treatment with D-drugs may also frequently lead to symmetric distal sensory polyneuropathy (65). Asymptomatic elevations in serum lipase are frequent under HAART, but usually do not herald the onset of pancreatitis (66). Pancreatitis is associated with the use of didanosine in particular, and reexposure to didanosine should be avoided in affected patients. A peculiar combination of peripheral and facial lipoatrophy together with abnormal accumulation of adipose tissue at certain other sites, the so-called NRTI-associated lipodystrophy syndrome, occurs after several months of antiretroviral treatment. Hepatic mitochondrial toxicity is particularly associated with marked lipoatrophy of the face, legs, and arms (63). Thus, the presence of peripheral lipodystrophy in patients on NRTIs should alert the clinician to search for the possibility of additional hepatic mitochondrial toxicity. Taking into account possible drug interactions may help to reduce the risk of mitochondrial toxicity. For instance, mitochondrial toxicity of didanosine is augmented due to increased drug levels if didanosine is combined with ribavirin, hydroxyurea, or allopurinol (67). Thus, these combinations should be avoided. When didanosine is given together with tenofovir, the dose of didanosine must be reduced to 250 mg QD. Impaired mitochondrial function may also result from comedication with ibuprofen, valproic acid, and acetylsalicylic acid, because these agents inhibit mitochondrial utilization of fatty acids. In addition, acetylsalicylic acid may directly damage mitochondria. Life-threatening lactic acidosis was observed when valproic acid was given to HIV-infected patients on antiretroviral therapy and in patients with inherited mutations of mtDNA. Amiodarone and tamoxifen inhibit mitochondrial synthesis of ATP. Acetaminophen and alcohol abuse impair mito-chondrial antioxidative defense via reducing levels of glutathione, thus allowing for their enhanced

572

Spengler

(A)

(B) FIGURE 1 (A) Mitochondrial hepatotoxicity in a patient taking an antiretroviral regimen comprising azidothymidine and stavudine as nucleoside analogues. By light microscopy, there is mixed microvesicular [lower left of figure (A)] and macrovesicular [upper right of figure (A)] steatosis of hepatocytes. (B) Electron microscopy demonstrates intracellular fat droplets and multiple enlarged mitochondria. Source: Courtesy of Prof. Fischer, Institute of Pathology, University of Bonn.

damage by oxidative radicals. Aminoglycoside antibiotics and chloramphenicol not only inhibit protein synthesis in bacteria but can also impair mitochondrial peptide synthesis under certain circumstances. Adefovir and cidofovir are also inhibitors of g-polymerase. Hypersensitivity Reactions Certain drugs cause liver damage to occur rapidly within the first six weeks after onset of therapy (68). In addition to hepatocellular injury, there may be coexistent clinical features such as fever, skin rash, or eosinophilia that specify this type of liver injury as acute immune-

Hepatic Toxicity of Antiviral Agents

573

mediated hypersensitivity reactions. Immune-mediated adverse drug reactions are among the idiosyncratic type of drug reactions. They are infrequent and unpredictable and not related to the given dose of drugs (69). Among the antiretroviral drugs hypersensitivity reactions have been observed with all NNRTIs, the nucleoside analogue abacavir and the PI amprenavir. Hepatic damage due to hypersensitivity reactions is seen most frequently in nevirapine-based regimens but may occur with any of the other above-mentioned drugs (70). The risk of nevirapine-related hypersensitivity is greater in women with CD4 counts above 250 cells/ml and in men with CD4 counts above 400 cells/ml. Hypersensitivity reactions to antiviral drugs are linked to immunogenetic factors, because HLA-B*5701 and other associated alleles carried on the 57.1 ancestral haplotype have been identified as strong risk factors for abacavirassociated hypersensitivity in Caucasian patients (Fig. 2) (71–73). As a consequence, tissue

(A)

(B) FIGURE 2 Hypersensitivity reaction in a patient taking abacavir for 3 wk as part of the antiretroviral regimen. (A) The liver biopsy shows a mixed inflammatory infiltrate also comprising eosinophils. (B) In addition, there was a marked skin rash giving a hint that this liver injury was caused by hypersensitivity reaction. Source: Courtesy of Prof. Fischer, Institute of Pathology, University of Bonn.

574

Spengler

typing and skin patch testing have been proposed as measures to identify patients at particularly high risk for abacavir hypersensitivity (74,75). Likewise, nevaripine hypersensitivity is independently associated with both HLA-DRB1*0101 and high CD4 counts (76). HLADR molecules are needed to present antigens to CD4C T cells in a haplotypic fashion. Thus, the identification of CD4C counts and a distinct HLA-DR allele as independent risk factors of nevirapine hypersensitivity supports the concept that abnormal immune recogition is critically involved in the pathogenesis of nevirapine hypersensitivity. Tissue damage, on the other hand, appears to be related to CD8C T effector cells as well as release of cytotoxic cytokines (75,77–79). Cytokine release may also cause the constitutional symptoms such as fever or hypotension, which are frequently observed in hypersensitivity reactions. However, it remains unclear at present whether the antiretroviral drugs themselves or their metabolites act as triggering or target antigens. Hepatic histopathology can vary considerably but commonly shows a mixed hepatocellular pattern of inflammation (8,80). Typical features include hepatocellular necrosis, cholestasis, tissue eosinophilia, parenchymal and periportal infiltrates with a predominance of lymphocytes and plasma cells. The reaction tends to resolve when the offending agent is withdrawn and typically recurs upon rechallenge, usually after a shortened time interval (81). Continued therapy, however, carries a considerable risk of progression to fulminant hepatic failure (59). With nevaripine, liver failure sometimes progresses despite discontinuation of the drug (82). Hepatotoxicity Due to Other Mechanisms Hyperbilirubinaemia is seen in some patients treated with the PIs indinavir (83) and atazanavir (84,85), because these drugs inhibit UDP-glucuronosyltransferase activity. Unconjugated hyperbilirubinaemia can be observed in up to 25% of patients treated with these drugs, and in extreme cases, may be a reason to discontinue the drug (85,86). This abnormality, however, is not associated with clinical or other biochemical signs of liver damage, and thus reflects interaction of the drugs with intrahepatic bilirubin transport pathways rather than actual drugrelated hepatotoxicity. Bilirubin levels usually do not exceed 6 mg/dL, and hemolysis is the most important differential diagnosis, that must be considered. Saquinavir is also a competitive inhibitor of UDP-glucuronosyltransferase, but has a higher K(i) and is used at substantially lower levels than indinavir (83). Thus, saquinavir is not associated with clinically manifest hyperbilirubinaemia. About 10% of the population have Gilbert’s syndrome, an unconjugated form of hyperbilirubinaemia caused by a genetic polymorphism in the promoter region of the UDP-glucuronosyltransferase gene (87). Although indinavir also causes hyperbilirubinaemia in subjects who carry exclusively the UDP-glucuronosyltransferase wild-type gene, increases in serum bilirubin are significantly higher in hetero- and homozygous carriers of the Gilbert’s syndrome mutant allele (83,88). The hepatotoxic potential of PIs appears to be related to their molecular structure in a poorly defined way. Of note, PIs may aggravate the liver damage induced by NRTI and/or concomitant additional liver diseases via their metabolic effects. PIs may contribute to hepatic steatosis, particularly in patients concomitantly taking stavudine (89). Hepatic steatosis appears to accelerate progression to cirrhosis caused by other diseases such as chronic hepatitis C (90–92). PIs have been associated with inhibition of insulin-mediated uptake of glucose, decreased insulin sensitivity and development of overt diabetes mellitus in some patients (93). With the exception of atazanavir, PIs induce hepatic fatty acids and sterol biosynthesis by activating the sterol-regulatory-element-binding protein 1-c which regulates several lipogenic genes, thus leading to substantial increases in plasma lipids and the accumulation of triglycerides within hepatocytes (93,94). Genuine hepatotoxic mechanisms are also recognized with some NRTIs. In addition to their mitochondrial toxicities, stavudine can induce apoptosis by DNA intercalation in a similar way to the cytostatic drugs fludarabin and gemcitabine (95); and zidovudine localizes to human telomeric DNA, possibly accelerating cellular senescence by telomere shortening (96). Finally, in addition to being a potent cause of drug hypersensitivity reactions, intrinsic dosedependent hepatotoxicity has also been suggested for nevirapine (97).

Hepatic Toxicity of Antiviral Agents

575

Interferons and Interferon-Induced Hepatotoxicity Type I interferons were the first therapeutic agents for successful antiviral therapy in chronic hepatitis B, D, and C, which had acceptable side effects. Several recombinant a interferons (interferon a-2a, a-2b, and a-2c), a fully synthetic type I interferon (consensus interferon), mixtures containing several type I interferons from natural sources (lymphoblastoid interferon, human natural leukocyte interferon), and interferon b are available worldwide and can be used to treat chronic viral hepatitis. Recombinant interferons have also been coupled to polyethylene glycol molecules to modify their pharmacokinetic properties. This modification results in delayed elimination of the compound, altered drug distribution within the body compartments and prolonged half-life. PEGylated interferons a-2a and a-2b have intensified activity against hepatitis viruses and exhibit significantly improved therapeutic efficacy. PEGylated interferons have largely replaced conventional interferons in the treatment of hepatitis B, and combined with ribavirin have now become also the therapeutic standard to treat chronic hepatitis C. Interferons comprise a group of related proteins which exert antiviral activity, regulate the growth of cells, inhibit angiogenesis, regulate cell differentiation, upregulate expression of major histocompatibility complex antigens, and exhibit a wide range of other immunomodulatory activities. Most cells have interferon receptors and respond to interferons by upregulation of antiviral genes such 2 0 , 5 0 oligoadenylate synthetase, a 60 kDa protein kinase, and the Mx protein homologue (98). Interferons also act as potent immunostimulants. This fact is exemplified in the treatment of chronic hepatitis B, where successful control of the virus is closely linked to a competent immune system (99). Furthermore, in responders, elimination of HBV is heralded by a transient flare of aminotransferases followed by their normalization and loss of serum HBVDNA. This flare is attributed to destruction of virus-infected hepatocytes by cytolytic immune effector mechanisms (100). Unlike hepatitis B, a transient flare of aminotransferases is usually not observed in interferon therapy of patients with chronic hepatitis C. Moreover, PEGylated interferons are still quite effective in the treatment of HIV/HCV coinfected patients (101). Several different forms of hepatotoxicity have been observed in interferon therapy. An asymptomatic rise in serum aminotransferases has been noted in up to 25% of patients given interferon for the treatment of malignancies. Importantly, elevated aminotransferases were observed in as many as 80% of those patients who had received doses in excess of 100 million units per week. Abnormalities were usually seen at the onset of interferon therapy and elevated serum aminotransferases resolved with continued therapy or after dose reduction (102,103). Direct interferon-associated liver toxicity is rare during interferon treatment of viral hepatitis and is not considered to be immune mediated. Furthermore, hepatotoxicity has to be distinguished from the typical flare of aminotransferases during therapy of chronic hepatitis B preceding seroconversion. Nevertheless, fatal hepatic decompensation with ascites, jaundice, and hepatic encephalopathy can occur occasionally during such flares (104). Chronic interferon therapy has the potential to induce autoimmune hepatitis in some patients with chronic viral hepatitis B and C, although this is considered a rare complication. This complication is characterized by aggravation of clinical symptoms, an increase in serum aminotransferase levels and the presence of nonorgan-specific autoantibodies (105–113). In one report of 144 patients with hepatitis C, deterioration in liver function and autoantibody formation occurred in seven (5 percent) during interferon therapy, all of whom were women (107). Some of these patients had preexisting autoantibodies, suggesting predisposition to autoimmune hepatitis during interferon therapy (108). The amino acid sequences of the hepatitis C virus (HCV) share homologies with the human cytochromes and cellular structures, thus molecular mimicry may contribute to virustriggered autoimmunity. At the molecular level, different specificities of anti-LKM-1 antibodies, which are directed against cytochrome P450 2D6 (CYP2D6), were demonstrated when LKM-1 autoantibody positive patients with hepatitis C and patients with idiopathic autoimmune hepatitis were compared (114,115). This distinction between autoimmune hepatitis and chronic hepatitis C is particularly important with respect to therapy. Autoimmune hepatitis benefits from immunosuppression by corticosteroids and/or azathioprine, while hepatitis C infection is generally treated with interferon in combination with ribavirin. However, inappropriate

576

Spengler

interferon therapy of autoimmune hepatitis can exacerbate the liver disease (107,108), while immunosuppressive therapy in hepatitis C infection lowers aminotransferase levels but generally leads to a rise of viral loads. On the other hand, most patients with hepatitis C and preexisting nonorgan-specific autoantibodies appear to benefit from interferon treatment to the same extent as autoantibody negative patients. However, the need for meticulous monitoring is mandatory during interferon treatment of such patients, since flares of aminotransferases without subsequent clearance of HCV RNA have been observed (116,117). When exacerbation of liver disease occurs during interferon therapy, interferon must be withdrawn. The benefit of introducing immunosuppression at this point must be balanced against the associated risks of increased viral replication. CLINICAL PRESENTATION OF HEPATOTOXICITY INDUCED BY ANTIVIRAL DRUGS NRTIs Asymptomatic elevations of aminotransferases greater than five times ULN are noted in approximately 6% of patients (118), and clinically manifest symptoms of NRTI-related hepatotoxicity occur with an average incidence of 1.3 events per 1000 person years (119). Symptoms of liver damage may develop occasionally within a few days after the start of treatment but normally occur gradually over several months with prolonged intake of the drugs (120,121). Patients usually complain of non-specific symptoms such as abdominal pain, vomiting, anorexia, and right upper quadrant tenderness. Laboratory tests reflect hepatocellular damage by elevated serum aminotransferases and increased biochemical markers of cholestasis such as g-glutamyl transpeptidase or alkaline phosphatase. Additional biochemical features may comprise elevated amylase, lipase, or creatine kinase frequently combined with a disturbed acid-base balance, lactic acidosis, and loss of bicarbonate. Features of the lipodystrophy syndrome are frequently simultaneously present in these patients with hepatic dysfunction (63), and a liver biopsy confirms that liver damage (microvesicular steatosis) is caused by mitochondrial toxicity in the great majority of patients. Of note, this type of injury can occur in patients receiving exclusively NRTI-based regimens (122). However, the individual contribution of each single NRTI compound is less clear. Stavudine has been incriminated as a major causative hepatotoxic agent (21,123) and didanosine seems to be particularly dangerous when it is used in high doses (250–375 mg b.i.d.) (124,125), or in combination with stavudine (21). Importantly, ribavirin markedly increases didanosine levels which may cause fatal hepatic mitochondrial toxicity, pancreatitis, and lactic acidosis (126), and combining ribavirin with didanoside should be completely avoided. In contrast, mitochondrial toxicity is not a major clinical problem with abacavir, lamivudine, tenofovir, and adefovir (see mitochondrial toxicity above) (123), but abacavir hypersensitivity reactions can be a cause of liver damage in predisposed individuals. On the other hand, there also exist patient-related factors determining the risk to develop NRTI-related mitochondrial toxicity such as acquired nutritional deficiencies, preexisting obesity, viral hepatitis, Black African descent and female sex. A particularly high risk for severe mitochondrial liver injury appears to be associated with NRTI treatment during pregnancy (127). Duration of NRTI exposure as well as the concomitant use of additional drugs, alcohol, and liver toxins are further determinants modifying NRTI-related liver toxicity. Infrequently, progression of liver damage towards fulminant hepatic failure has been noted (10,58,59). At least initially, this complication may be mistaken for an acute surgical abdomen (128). The risk of hepatic failure is high in patients with severe pancreatic involvement and metabolic acidosis. Fortunately, most patients surviving an acute episode of NRTI-induced hepatotoxicity finally achieve full hepatic recovery (120,128). Nevertheless, occasionally gradual progression to chronic liver injury has been observed despite initial recovery from hepatic damage (58). The long-term outcome in patients with asymptomatic elevations of liver enzymes is unknown at present, but it is re-assuring that elevated aminotransferases have been reported to return to normal in as many as 87% of the patients (118).

Hepatic Toxicity of Antiviral Agents

577

In the case of severe liver damage (i.e., evidence for deteriorating liver function such as coagulopathy and excretory failure and/or aminotransferases exceeding five times ULN) drugs should be stopped temporarily. An alternative regimen with less mitochondrial toxicity can be initiated, once recovery of liver function tests has been achieved, and patients must be monitored closely (129). The severity of the acute toxicity may be blunted by administration of riboflavin (130) and/or thiamine (131). The supplementation of uridine is a novel promising strategy to prevent or treat clinically manifest mitochondrial toxicity of NRTIs. Impaired respiratory chain function leads to inhibition of DHODH, an essential enzyme of uridine synthesis and its pyrimidine derivatives. Decreased intracellular pyrimidine pools enhance the relative excess of NRTIs over natural pyrimidines available for mitochondrial g-polymerase, thus further aggravating mtDNA deficiency. By supplementing uridine, this vicious circle may be interrupted. In hepatocytes, uridine abolished all the effects of mtDNA depletion, and normalized lactate production, cell proliferation, the rate of cell death and intracellular steatosis (132). Uridine could also prevent the loss of mtDNA, accumulation of lipids and apoptosis in adipocytes exposed to stavudine (133). Oral substitution with uridine is well tolerated by humans, even at high doses (134,135). Moreover uridine does not interfere with the antiviral efficacy of NRTIs (136,137). Mitocnol, a food supplement from sugarcane with a high content of nucleosides, increased human serum concentrations of uridine to levels which are protective in vitro experiments (138) and reversed the clinical manifestations of mitochondrial toxicity in a patient with stavudine toxicity (139). NNRTIs There is a perception that serious clinical liver toxicity is more frequent with NNRTI-based regimens than NRTI or PI-based regimens. Nevirapine has most often been incriminated as the putative cause of NNRTI-related hepatotoxicity. Thus, the manufacturer issued a safety warning in 2000 (140). Nevirapine can cause several types of hepatotoxicity: (i) asymptomatic elevations of aminotransferases, (ii) early onset hypersensitivity reactions as well as (iii) a delayed-onset, dose-related hepatotoxicity. Elevated serum g-glutamyl transpeptidase activity was the most frequently reported abnormality; followed by increased aminotransferases, bilirubin and alkaline phosphatase, while symptomatic liver injury was initially estimated to occur at relative incidences ranging from 1% to 2% (23,141). Two comprehensive analyses of all randomized clinical trials on nevirapine demonstrated that 10% of nevirapine-treated patients developed elevated aminotransferases greater than five times ULN (122,142). However, in this analysis, only approximately 5% of patients had symptomatic liver disease. Although severe adverse hepatic events are most common in the first six weeks after the start of treatment, liver toxicity can occur at any time. In the analysis by Dieterich et al., 46% of patients with symptomatic hepatic events (2.2% of all patients treated with nevirapine) had a concomitant rash or other immune-related events (142). 2.7% of nevirapine-treated patients developed hepatotoxicity without a rash. Rates of rash-associated liver toxicity varied considerably between the different studies from less than 0.5% in the BI1090 study to 6% in the FTC-302 study. This variability in frequency was not seen for nevirapine-associated liver toxicity without a rash. The onset of nevirapine-associated hepatotoxicity was variable, most likely reflecting two alternative mechanism of hepatic damage. Constitutional symptoms such as fever or hypotension were usually present in those patients who had hepatic dysfunction and a rash. Key risk factors for this type of nevirapine-related liver injury were female sex, and high baseline CD4 counts of more than 250 cells/ml in women, and more than 400 cells/ml in men. (143). Combining nevirapine with a PI further increases the risk (144), while preexisting elevated baseline levels of aminotransferases and concomitant viral hepatitis B and C have been identified as risk factors in some but not all studies (9,13,15,142,144,145). Nevirapine should not be used for multidose postexposure prophylaxis, because severe and fatal hepatotoxic events have occurred in this setting (146). Risk factors for nevirapine-associated hepatotoxicity have recently been highlighted in a letter from the U.S. Food and Drug Administration (147) and by the manufacturer of the drug (148). Immune-mediated nevirapine hepatotoxicity takes a benign clinical course if the drug is withdrawn early. Nevertheless, two patients died after nevirapine exposure in the South African

578

Spengler

FTC-302 and the BI1090 trials (9), respectively. In both trials, a fourfold higher proportion of Black African patients had been enrolled than in the trials without fatal liver toxicity. Two additional nevirapine-related deaths occurred in the cohort reported by Sulkowski et al. (13), and one more patient—taking nevirapine as part of his postexposure prophylaxis regimen— required emergency liver transplantation (149). However, in the combined analysis of all randomized trials, fatal outcomes of nevirapine-related liver damage was infrequent occurring in only 0.1% of all nevirapine-treated patients (9,142). Late-onset nevirapine-associated hepatotoxicity seems to reflect cumulative direct hepatotoxicity rather than hypersensitivity (150). Typically, liver enzymes start to increase gradually after four to five months of nevirapine treatment (97,147). Liver damage seems to be related to the cumulative dose of the drug. Elevated baseline levels of aminotransferases, high drug serum concentrations, and coinfection with HBV but apparently not coinfection HCV were more frequent in patients with symptomatic late-onset nevirapine-related hepatotoxicity; sex and race were not risk factors (142). Unfortunately, the histopathological lesions underlying late-onset nevaripine-related hepatotoxicity are still poorly defined. A comprehensive analysis done by the Adult AIDS Clinical Trials Group, which included 10,611 patients from 21 studies, reported asymptomatic grade 3 and 4 increases of serum aminotransferases (Ofive times ULN) in 10.8% and 8.9% of the nevirapine- and efavirenztreated patients, respectively (122). In the same analysis, only 3.6% of delaviridine-treated patients developed severely increased liver enzymes. The author concluded that patients treated with delavirdine had a significantly lower risk of liver toxicity than patients treated with the other NNRTIs. A recent retrospective analysis, however, showed that there were no differences concerning the incidence of liver toxicity in patients treated with nevirapine, efavirenz, or delavirdine (15). In addition, this retrospective analysis showed that there was more apparent NNRTI-related toxicity in patients with chronic hepatitis B and C. PIs In clinical trials, both indinavir and atazanavir were associated with increases of unconjugated bilirubin in 6% to 40% of the treated patients depending on the concomitant boostering with ritonavir. However, this phenomenon does not reflect hepatocellular damage but indicates inhibition of the hepatic enzyme UDP-glucuronosyltransferase. Jaundice is more pronounced in patients with Gilbert’s syndrome, a harmless inherited abnormality of bilirubin glucuronidation. In randomized trials, the risk of severe grade 3 and 4 hepatotoxicity associated with PIs was estimated to be 1% to 10% of the treated patients. In these registration trials, however, symptoms, dose-limiting toxicity, and sustained liver damage were rare events. Shortly after licensure, several case reports provided details on severe liver injury in patients coinfected with HBV and HCV (151–153). Subsequent analysis of several cohorts confirmed a 10% to 20% risk for hepatotoxicity in PI-treated patients worldwide (1,2,7,14,16,21,25,118,154–159). The patients in these cohorts differed from those in the registration studies by a variable proportion of HBV and HCV coinfected subjects, and inadvertently increased exposure to PIs as a consequence of impaired liver function may have contributed importantly to the observed hepatic toxicity (160–162). The relative risk of hepatic toxicity of individual PIs is controversial. Ritonavir and indinavir were a leading cause of hepatic injury in the case reports (151,152,163), probably simply reflecting the widespread use of these agents as compared to other compounds. Highdose ritonavir (1200 m/day) was the single agent, which in several studies was associated with an increased risk of hepatotoxicity compared to other antiretroviral regimens (2,7,16,21,151, 156). However, high-dose ritonavir is no longer recommended as a first line therapy. Ritonavir is a potent inhibitor of the cytochrome P450 3A4 (CYP3A4) isoenzyme, and is coadministered in low doses with other PIs such as lopinavir, saquinavir, and indinavir to improve the bioavailability and half-life of these drugs. Of note, current guidelines for the treatment of HIV issued by the U.S. Department of Health and Human Sciences recommend ritonavir-boosted lopinavir in combination with two nucleoside or nucleotide analogues as the preferred regimen. A safety analysis of ritonavir-boosted regimens revealed that liver toxicity was low with ritonavir-boosted lopinavir and indinavir, but hepatotoxicity was more frequent in those

Hepatic Toxicity of Antiviral Agents

579

patients with HCV coinfection (157). Although saquinavir itself has a low incidence of severe liver toxicity, when boostered with high doses 26% incidence of grade 3 and 4 aminotransferase elevations was observed in HCV coinfected patients, probably because saquinavir was increased to toxic levels (164). In a retrospective study of 1168 HIV/HCV coinfected patients receiving PI, the rate of grade 3 and 4 liver enzyme elevations was 3% in nelfinavir-treated patients and 8% in the patients treated with other PIs (165). Moreover, liver toxicity related to amprenavir appeared to be quite rare in a review of data from 358 adults and 268 children (166). Tipranavir seems to exert considerable liver toxicity; the true incidence and underlying mechanisms are currently still a matter of debate (167). Aminotransferases start to rise after the first 100 days of therapy. On average, elevated aminotransferases return to normal within the next three to five months with continued treatment in most patients, but may remain persistently elevated in approximately 10% of the affected patients (1,118). PI-related hepatotoxicity of clinical relevance has been observed in less than 5% of patients (118,151). These patients frequently complained of abdominal pain, fever, jaundice, vomiting, and pruritus. In three large clinical trials involving approximately 1500 patients, mortality due to PI-associated hepatotoxicity was not observed (2,118,151). Nevertheless, fulminant hepatic failure has rarely been encountered in patients with extensive pretreatment (152), severe comorbidity, (153) or increased hepatic vulnerability due to other reasons (163). It currently remains quite unclear if histopathological findings play any role in the diagnosis of PI-induced liver injury. Histological changes are quite variable and heterogeneous. Injury of intrahepatic bile ducts, ductular proliferation, hepatocellular necrosis, and Mallory bodies have been reported repeatedly, and a recently conducted pathological survey identified ballooning of hepatocytes, activation of Kupffer cells and pericellular fibrosis in zone 3 as distinctive histological features in PI-related liver injury (168). However, all these histological alterations have poor diagnostic specificity. DRUG INTERACTIONS All PIs as well as the NNRTIs efavirenz, nevirapine, and delaviridine are metabolized via cytochrome P450 enzymes, mainly isoenzymes CYP3A4 and CYP2D6 or other distinct cytochrome P450 isoenzymes (164). Thus, antiretroviral drugs importantly alter the half lives and serum levels of other drugs that share the same routes of hepatic biotransformation. Both NNRTIs and PIs interact with the same cytochrome P450 isoenzymes and can mutually affect their serum half lives if given in combination. For example, delaviridine is a potent inhibitor of cytochrome P450 and increases plasma concentrations of various PIs up to fivefold (169). Similar effects on PI serum half-life can result from the simultaneous administration of efavirenz and a PI (170). Likewise, even small doses of the PI ritonavir (100–200 mg) can prolong and enhance the action of other PIs, which is exploited to boost bioavailability of PIs. Toxic drug levels and liver damage can occur when such interactions are not taken into account by adjusting the doses. Antimicrobial medication which is frequently required in HIV-infected patients is an important cause for potentially hepatotoxic drug interactions in patients receiving PIs or NNRTIs. For instance, simultaneous administration of indinavir with ketoconazole or itraconazole increases the area under the curve (AUC) of indinavir by roughly 70% (171). Therefore, indinavir doses must be reduced to 600 mg t.i.d. in this combination. Simultaneous administration of clarithromycin and a PI will result in increased concentration of both drugs. Azithromycin represents a safe alternative that should be preferred in this setting whenever possible (172). Treatment of tuberculosis constitutes a particularly difficult problem, because most antituberculous drugs potently interact with cytochrome P 450 enzymes and can induce toxic liver damage themselves. Since a recent analysis indicated a fourfold increased risk for drug-associated hepatitis in patients treated simultaneously against HIV and tuberculosis (173), the U.S. Centers for Disease Control strongly advised against combining rifampicin and any PI (174). Rifampicin should be exchanged for rifabutin which, however, should not be combined with saquinavir hard gel capsules, ritonavir, or delaviridine. Nevertheless, further dose adjustments of rifabutin and/or the PIs may still be necessary to avoid severe liver toxicity.

580

Spengler

TABLE 2 Recommendations for Antiretroviral Drug Dosages Drug Indinavir Lopinavir Saquinavir Nelfinavir Atazanavir Amprenavir Efavirenz Nevirapine

Cytochrome P450 isoenzymes relevant for drug metabolism CYP3A, CYP2D6 CYP3A4, CYP2C9 CYP3A, CYP2C9 CYP3A, CYP2D6 CYP3A4 CYP3A4, CYP2D6, CYP2C9 CYP3A, CYP2C9, CYP2C9 CYP3A

Target serum concentration 300 800 ng/mL 3500 6000 ng/mL 100 800 ng/mL 1000 2500 ng/mL 150 1000 ng/mL 750 2500 ng/mL 1000 4000 ng/mL 3000 5000 ng/mL

Note: Therapeutic target levels may vary depending on the laboratory methods. Abbreviations: CYP3A, cytochrome P450 3A; CYP3A4, cytochrome P450 3A4; CYP2D6, cytochrome P450 2D6; CYP2C9, cytochrome P450 2C9.

Monitoring of drug levels may provide some guidance as to how to adjust antiretroviral drugs in difficult cases (Table 2). However, therapeutic ranges and more importantly safety limits of antiviral drugs have not yet been agreed upon. Moreover, PIs and NNRTI not only change serum levels of other drugs metabolized via cytochrome P450 but may profoundly change their pharmacokinetic profiles indicating that further drug metabolising enzymes are affected as well. Thus, even apparently well-established therapeutic drug levels may be misleading if the drugs are given in combination with antiretroviral agents (175). A detailed discussion of this complex issue is beyond the scope of this review. However, the reader will find useful information at the project inform internet page on drug interactions (176) or in detailed expert reviews on this issue (177). At the population level, allelic variation can be detected in the genes encoding cytochrome P450 isoenzymes CYP3A4, CYP2C9, and CYP2D6. Interestingly, ten of 27 patients with elevated antiviral drug levels were carriers of the CYP2D6 allele defining the poor metabolizer phenotype, while antiviral drug levels were apparently not correlated to mutations in the genes encoding CYP3A4 and CYP2C19 (178). Homozygosity CC at codon 1145 (exon 26) of the P-glycoprotein gene was a further genetic polymorphism affecting antiviral drug levels in this study. Nevertheless, it remains unclear at present if such polymorphisms of drug metabolism actually contribute to liver toxicity, considering that most antiviral drugs can be metabolized via several enzymatic pathways. LIVER INJURY OF ANTIVIRAL DRUGS IN THE PRESENCE OF CONCOMITANT HEPATIC DISEASE Most studies analyzing liver injury unequivocally identified preexisting chronic hepatitis B and C as well as elevated baseline levels of aminotransferases as important predictors indicating an increased risk for hepatotoxic side effects of subsequent antiviral therapy. In a recent trial, the presence of concomitant chronic hepatitis C or hepatitis B was associated with a threefold increased risk for severe HAART-associated hepatotoxicity (1). This study also revealed that hepatotoxicity occurred earlier in the HCV or HBV-coinfected HIV patients than in mono-infected patients and was associated with significantly greater elevations of aminotransferases. Coinfection with hepatitis viruses is frequent in patients infected with HIV, because these viruses share the same routes of transmission (179). The prevalence of HCV and HIV varies from 10% to 90% depending on geographic location, and this coinfection is particularly frequent in hemophiliac patients and in intravenous drug users (180–182). Prevalence of HBV chronic infection ranges form 7% to 15% in HIV-infected patients. Hepatitis B Coinfection In several studies, hepatitis B surface antigen(HBsAg) seropositivity has clearly been identified as an independent predictor of HAART-related hepatotoxicity in HIV positive patients

Hepatic Toxicity of Antiviral Agents

581

(1,2,13,118). However, in these studies, hepatotoxicity has been defined as an arbitrary and heterogeneous increase in serum aminotransferase levels after the onset of HAART, without taking into account details of HBV serological markers and changes in HBV viral loads. Furthermore, analysis of any potentially confounding factors such as alcohol consumption had not been undertaken. Possible mechanisms of HBV-associated liver toxicity include direct toxic drug effects, inadvertent withdrawal of drugs with anti-HBV activity, HBV escape mutations and HBVrelated immune reconstitution syndromes. Intrinsic toxic drug effects are often dose related but the effects of HBV-seropositivity on this type of liver injury remains speculative. The serological and virological state of the patient should be identified at the start of therapy and monitored regularly during therapy with antiretroviral drugs. In particular, it should be taken into account whether drugs are given that affect HBV replication such as lamivudine and tenofovir. Of note, withdrawal of antiretroviral drugs with anti-HBV activity was a leading cause of elevated aminotransferases in the Amsterdam cohort (7), and a flare of hepatitis B following lamivudine withdrawal was even observed in previously anti-HBs positive patients who had lost their antiHBV immunity due to progressive HIV-related immunodeficiency (183). Finally, exacerbation of HBV replication due to the appearance of escape mutations may be a further mechanism that might cause elevated aminotransferases during antiviral therapy in HIV/HBV coinfected patients and in patients with hepatitis B alone who are treated with nucleoside or nucleotide analogues (184,185). It is also clear that recovery of immune function plays an important role in both reactivation of the HBV as well as in its clearance. It is therefore not surprising that flares of hepatitis have also been observed during immune reconstitution following successful antiretroviral therapy in some HBV/HIV coinfected patients. It is remarkable that such flares have been followed by successful resolution of chronic HBV infection (186–188). These findings underscore the need to properly interpret the abnormalities of liver function in patients with chronic hepatitis B and HIV coinfection, and illustrate the pivotal role of baseline hepatitis markers for the further management of such patients. Hepatitis C Coinfection Steatosis is a histological hallmark of chronic hepatitis C, and is an independent predictor of fibrosis progression and treatment failure (92,189). Several mechanisms are discussed as potential causes. Importantly, hepatitis C infection seems to interfere with the integrity of mitochondrial function leading to the depletion of mtDNA, inhibition of oxidative phosphorylation, and the intracellular release of reactive oxygen radicals (30,190–192). Thus, chronic hepatitis C seems to synergize with the mechanism of mitochondrial toxicity exerted by nucleoside analogues. In line with this concept, Sulkowski et al. reported steatosis in 40% of HIV/HCV coinfected patients with extensive antiretroviral therapy, which was associated with more severe liver disease (89). Excess weight, hyperglycemia and the use of stavudine were risk factors in this population, which could be modified. Furthermore, HCV core and envelope proteins may trigger mitochondrial pathways of apoptosis (193,194), and HCV core protein may importantly modify hepatocellular lipid metabolism (195). Both these mechanisms may further enhance the cellular damage resulting from the metabolic effects of antiretroviral drugs. A direct role of hepatitis C in aggravating drug-induced liver toxicity in HIV-coinfected patients on HAART is further supported by the clinical observation that pretreatment of hepatitis C markedly reduces the risk of subsequent antiretroviral drug hepatotoxicity in HIV/ HCV coinfected patients (196). Moreover, chronic liver toxicity seems to occur in HIV/HCV coinfected patients not responding to antiretroviral therapy (156). On the other hand, transient flares of hepatitis C have been observed in patients successfully treated by HAART (197–199). However, liver biopsies taken during the flares did not show typical features of drug toxicity but rather marked inflammatory infiltrates (200). Increased HCV replication due to a putative reciprocal relationship between HIV and HCV replication (200–202), or infection with more “toxic” HCV genotype 3 isolates have been proposed as potential causes for this increased drug-associated hepatic inflammatory activity (203). Of note, this type of liver injury was found predominantly in HIV/HCV coinfected patients who had achieved an extraordinary recovery of their CD4 counts on antiretroviral

582

Spengler

therapy and was seen infrequently with ritonavir and saquinavir, which interfere adversely with proteasome activity, antigen presentation and cytotoxic CD8C T cell responses (204,205). Finally, sustained disappearance of HCV viraemia in single patients has been reported following such HAART-associated exacerbations of hepatitis (206). Thus, these flares may be linked to restoration of HCV-specific immune responses possibly resulting in substantially increased immune recognition and destruction of HCV-infected hepatocytes. However, direct proof of this hypothesis is still lacking (207,208). Whatever their underlying cause, hepatitis-like inflammatory exacerbations in HAART-treated patients with hepatitis C should probably not be considered as hepatotoxic side effects. Liver Cirrhosis and Reduced Liver Function Both in hepatitis B and C, progression to fibrosis and cirrhosis is accelerated in HIV coinfected patients. Fibrotic livers with reduced hepatocellular function may be particularly vulnerable to the additional toxic effects of drug therapy, and consequently an increased frequency of hepatotoxic events has been reported during antiretroviral therapy of patients with advanced fibrosis (209). This problem is also illustrated in the APRICOT trial on treatment of hepatitis C with PEGylated interferon a-2a and ribavirin in HIV/HCV coinfected patients, which disclosed an increased treatment-associated risk of hepatic decompensation and death in patients with preexisting cirrhosis (210). Impaired liver function may also alter the pharmacokinetics of antiretroviral drugs; a factor contributing importantly to increased susceptibility in patients with preexisting liver disease. In particular, the drug metabolism of PIs and NNRTIs depends on the amount of functional cytochrome P450, which is reduced in compromised livers. Indeed, increased or even toxic drug concentrations have been reported in cirrhotic patients receiving standard doses of either ritonavir, indinavir, saquinavir, nelfinavir, lopinavir, amprenavir, nevirapine and efavirenz (157,161,171,211–214). Moreover, the AUC was increased by 40% in nevirapine-treated patients who had only moderate hepatic dysfunction (215). However, thus far dose reductions have been recommended for patients with cirrhosis only with respect to indinavir (reduction to 600 mg t.i.d.), amprenavir (reduction to 450 mg or 300 mg b.i.d.) and atazanavir (reduction to 400 mg q.d), in order to avoid toxic damage (160,171,216,217). CONSEQUENCES OF ANTIVIRAL DRUG TOXICITY FOR CLINICAL PRACTICE Circumstantial evidence suggests that it is generally safe to tolerate AST/ALT elevations up to a fivefold increase during HAART, unless these alterations are considered to be manifestations of a drug hypersensitivity reaction. In the setting of HAART, based exclusively on NRTI and PIs, elevations of liver transaminases or g-GT levels may be tolerable, provided that these alterations do not coincide with clinical symptoms and major indicators of hepatic organ dysfunction (e.g., severe reduction in clotting factors and serum pseudocholinesterase or increased blood ammonia levels). Indeed, several independent studies have indicated that a spontaneous recovery of these abnormalities occurs in the majority of patients despite unaltered, continued therapy (1,21,33,218). Continued observation combined with frequent monitoring at short-term intervals seems to be justified in those cases. The decision to continue unaltered treatment in the face of elevated liver enzymes should, however, be preceded by measuring serum drug levels and making appropriate dose modifications, if necessary. The identification of the causative agent is of particular relevance for the future treatment plan. NRTI-related liver toxicity is characterized by hepatic steatosis occurring typically in conjunction with lactic acidosis and elevated levels of the pancreatic enzymes, amylase or lipase. However, the diagnosis of NRTI-related liver damage is most reliably established by a liver biopsy. If severe mitochondrial damage is confirmed, all antiviral drugs should be temporarily stopped. An alternative regimen can be initiated, once recovery of liver function tests has been achieved. Patients with idiosyncratic drug reactions usually develop their hepatic symptoms shortly after the onset of HAART. Frequently, the cause is a hypersensitivity reaction, which is associated with characteristic extrahepatic features such as fever or skin

Hepatic Toxicity of Antiviral Agents

583

rashes. Immediate drug discontinuation is mandatory in these patients to avoid severe liver failure. A reexposure to the incriminated drug must be avoided under all circumstances. Corticosteroids have been proven effective in the acute management of nevirapine-associated liver damage (219). If liver injury occurs in patients with concomitant viral hepatitis, it can be difficult to determine whether the dysfunction is actually related to drug toxicity rather than to flares of the underlying hepatitis induced by exaggerated immune reconstitution. A liver biopsy may be required to make a conclusive decision. In the case of an immune reconstitution syndrome, a slow but steady long-term recovery of hepatocellular function can be expected in the vast majority of patients, and antiretroviral drugs must be discontinued in potentially lifethreatening situations only. REFERENCES 1. Den Brinker M, Wit FW, Wertheim-van Dillen PM, et al. Hepatitis B and C virus con-infection and the risk for hepatotoxicity of highly active antiretroviral therapy in HIV-1 infection. AIDS 2000; 14:2895–902. 2. Sulkowski MS, Thomas DL, Chaisson RE, Moore RDl. Hepatotoxicity associated with antiretroviral therapy in adults infected with human immunodeficiency virus and the role of hepatitis C or B virus infection. JAMA 2000; 283:74–80. 3. Ensoli F, Sirianni MC. HIV/HCV co-infection: clinical and therapeutic challenges. AIDS 2002; 16:1419–20. 4. Bruno R, Sacchi P, Puoti M, Soriano V, Filice G. HCV chroni hepatitis in patients with HIV: clinical management issues. Am J Gastroenterol 2002; 97:1598–606. 5. Benhamou Y. Antiretroviral therapy and HIV/hepatitis B virus coinfection. Clin Infect Dis 2004; 38(Suppl. 2):S98–103. 6. Bica I, McGovern B, Dhar R, Stone D, McGowan K, Scheib R, Snydman DR. Increasing mortality due to end-stage liver disease in patients with human immunodeficiency virus infection. Clin Infect Dis 2001; 32:492–7. 7. Wit FWNM, Weverling GJ, Weel J, Jurriaans S, Lange JMA. Incidence and risk factors for severe hepatotoxicity associated with antiretroviral combination therapy. J Infect Dis 2002; 186:25–31. 8. Puoti M, Torti C, Ripamonti D, Castelli F, Zaltron S, Zanini B, Spinetti A, Putzolu V, Casari S, Tomasoni L, et al. Severe hepatotoxicity during combination antiretroviral treatment: incidence, liver histology, and outcome. J Acquir Immune Defic Syndr 2003; 32:259–67. 9. Stern JO, Robinson PA, Love J, Lanes S, Imperiale MS, Mayers DL. A comprehensive hepatic safety analysis of nevirapine in different populations of HIV infected patients. J Acquir Immune Defic Syndr 2003; 34:S21–33. 10. Bissuel F, Bruneel F, Habersetzer F, et al. Fulminant hepatitis with severe lactate acidosis in HIVinfected patients on didanosine therapy. J Intern Med 1994; 235:367–71. 11. Anonymus. Report on an international consensus meeting. Criteria of drug-induced liver disorders. J Hepatol 1990; 11:272–6. 12. AIDS Clinical Trials Group. Table of grading severity of adult adverse experiences. Rockville MD: Division of AIDS, National Institute of Allergy and Infectious Diseases, 1996. 13. Sulkowski MS, Thomas DL, Mehta SH, Chaisson RE, Moore RD. Hepatotoxicity associated with nevaripine or efavirenz-containing antiretroviral therapy: role of hepatitis C and B infections. Hepatology 2002; 35:182–9. 14. Nun˜ez M, Lana R, Mendoza JL, Martin-Carbonero L, Soriano V. Risk factors for severe hepatic injury after introduction of highly active antiretroviral therapy. J Acquir Immune Defic Syndr 2001; 27:426–31. 15. Palmon K, Koo BCA, Shoultz DA, Dieterich DT. Lack of hepatotoxicity associated with nonnucleoside reverse transcriptase inhibitors. J Acquir Immune Defic Syndr 2002; 29:340–5. 16. Bonfanti P, Landonio S, Ricci E, et al. Risk factors for hepatotoxicity in HIV-infected patients treated with highly active antiretroviral therapy. J Acquir Immune Defic Syndr 2001; 167(316):8 (Letter). 17. Martinez E, Blanco JL, Arnaiz JA, et al. Hepatotoxicity in HIV-1-infected patients receiving nevirapine-containing antiretroviral therapy. AIDS 2001; 15:1261–8. 18. Dybul M, Fauci AS, Bartlett JG, Kaplan JE, Pau AK. Panel on Clinical Practices for the Treatment of HIV. Guidelines for using antiretroviral agents among HIV-infected adults and adolescents: recommendations of the Panel on Clinical Practices for Treatment of HIV. MMWR Morb Mort Wkly Rep 2002; 51(RR-07):1–64. 19. Madge S, Mocroft A, Wilson D, et al. Participation in clinical studies among patients infected with HIV-1 in a single treatment centre over 12 years. HIV Med 2000; 1:212–8.

584

Spengler

20. Gifford AL, Cunningham WE, Heslin KC, et al. Participation in research and access to experimental treatments by HIV-infected patients. N Engl J Med 2002; 346:1373–82. 21. Gisolf HE, Dreezen C, Danner SA, Weel JL, Weverling GJ. Prometheus Study Group. Risk factors for hepatotoxicity in HIV infected patients receiving ritonavir and saquinavir with or without stavudine. Clin Infect Dis 2000; 31:1234–9. 22. Stawzewski S, Keiser P, Montaner J, Raffi F, Gathe J, Brotas V, Hicks C, Hammer SM, Cooper D, Johnson M, et al. Abacavir-lamivudine-zidovudine vs. indinavir-lamivudine-zidovudine in antiretroviral-naive HIV-infected adults. JAMA 2001; 285:1155–63. 23. Sanne I. Severe liver toxicity in patients receiving two nucleoside analogues and a non-nucleoside reverse transcriptase inhibitor. AIDS 2000; 14(Suppl. 4):S12 (Abstract PL9.3). 24. Fung SK, Andreone P, Han SH, et al. Adefovir resistant hepatitis B can be associated with viral rebound and hepatic decompensation. J Hepatol 2005; 43:937–43. 25. Kakuda TN. Pharmacology of nucleoside and nucleotide reverse transcriptase inhibitor-induced mitochondrial toxicity. Clin Ther 2000; 22:685–708. 26. Lewis W, Dalakas M. Mitochondrial toxicity of antiviral drugs. Nat Med 1995; 1:417–22. 27. Lewis W, Day BJ, Copeland WC. Mitochondrial toxicity of NRTI antiviral drugs: an integrated cellular perspective. Nat Rev Drug Discov 2003; 2:812–22. 28. Walker UA, Setzer B, Venhoff N. Increased long-term mitochondrial toxicity in combinations of nucleoside analogue reverse transcriptase inhibitors. AIDS 2002; 16:2165–73. 29. Lim SE, Copeland WC. Differential incorporation and removal of antiviral deoxynucleotides by human polymerase-g. J Biol Chem 2001; 276:23616–23. 30. Johnson AA, Ray AS, Hanes J, Suo Z, Colacino JM, Andersen KS, Johnson KA. Toxicity of antiviral nucleoside analogs and the human mitochondrial DNA polymerase. J Biol Chem 2001; 276:40847–57. 31. Walker UA, B€auerle J, Laguno M, et al. Depletion of mitochondrial DNA in liver under antiretroviral therapy with didanosine, stavudine or zalcitabine. Hepatology 2004; 39:311–7. 32. Hart GJ, Orr DC, Penn CR, Figueiredo HT, Gray NM, Boehme RE, Cameron JM. Effects of (K)-2 0 deoxy-3 0 -thiacytidine (3TC) 5 0 -triphosphate on human immunodeficiency virus reverse transcriptase and mammalian DNA polymerases alpha, beta, and gamma. Antimicrob Agents Chemother 1992; 36:1688–94. 33. Birkus G, Hitchcock MJM, Cihlar T. Assessment of mitochondrial toxicity in human cells treated with tenofovir: comparison with other nucleoside reverse transcriptase inhibitors. Antimicrob Agents Chemother 2002; 46:716–23. 34. Birkus G, Gibbs CS, Cihlar T. Comparative effects of adefovir and selected nucleoside inhibitors of hepatitis B virus DNA polymerase on mitochondrial DNA in liver and skeletal muscle cells. J Viral Hepat 2003; 10:50–4. 35. Cihlar T, Birkus G, Greenwalt DE, Hitchcock MJ. Tenofovir exhibits low cytotoxicity in various human cell types: comparison with other nucleoside reverse transcriptase inhibitors. Antiviral Res 2002; 54:37–45. 36. Sanchez-Conde M, Mendoza C, Jimenez-Nacher I, Barreiro P, Gonzalez-Lahoz J, Soriano V. Reductions in stavudine dose might ameliorate mitochondrial-associated complications without compromising antiviral activity. HIV Clin Trials 2005; 6:197–202. 37. B€auerle J, Laguno M, Mauss S, et al. Mitochondrial DNA depletion in liver tissue of patients infected with hepatitis C virus: contributing effect of HIV infection? HIV Med 2005; 6:135–9. 38. Harvie P, Omar RF, Duserre N, et al. Ribavirin potentiates the efficacy and toxicity of 2 0 ,3 0 -dideoxyinosine in the murine acquired immunodeficiency syndrome model. J Pharmacol Exp Ther 1996; 279:1009–17. 39. Becher F, Pruvost AG, Schlemmer DD, et al. Significant levels of intracellular stavudine triphosphate are found in HIV-infected zidovudine-treated patients. AIDS 2003; 17:555–61. 40. Bonora S, Boffito M, D’Avolio A, et al. Detection of stavudine concentrations in plasma of HIVinfected patients taking zidovudine. AIDS 2004; 18:577–8. 41. Barile M, Valenti D, Hobbs GA, et al. Mechanism of toxicity of 3 0 -azido-3 0 deoxythymidine. Its interaction with adenylate kinase. Biochem Pharmacol 1994; 48:1405–12. 42. Barile M, Valenti D, Passarella S, Quagliariello E. 3 0 -azido-3 0 deoxythymidine uptake into isolated rat liver mitochondria and impairment of ADP/ATP translocator. Biochem Pharmacol 1997; 53:913–20. 43. Valenti D, Barile M, Passarella S. AZT inhibition of the ADP/ATP antiport in isolated rat heart mitochondria. Int J Mol Med 2000; 6:93–6. 44. Valenti D, Barile M, Quagliariello E, Passarella S. Inhibition of nucleoside diphosphate kinase in rat liver mitochondria by added 3 0 -azido-3 0 -deoxythymidine. FEBS Lett 1999; 444:291–5. 45. Hall ET, Yan JP, Melancon P, Kuchta RD. 3 0 -azido-3 0 -deoxythymidine potently inhibits protein glycosylation. A novel mechanism for AZT toxicity. J Biol Chem 1994; 269:14355–85. 46. Weidner DA, Sommadossi JP. 3 0 -azido-3 0 -deoxythymidine inhibits globin gene transcription in butyric acid-induced K562 human leukemia cells. Mol Pharmacol 1990; 38:797–804. 47. Hobbs GA, Keilbaugh SA, Rief PM, Simpson MV. Cellular targets of 3 0 -azido-3 0 -deoxythymidine: an early (non-delayed) effect on oxidative phosphorylation. Biochem Pharmacol 1995; 50:381–90.

Hepatic Toxicity of Antiviral Agents

585

48. Hayakawa M, Ogawa T, Sugiyama S, Tanaka M, Ozawa T. Massive conversion of guanosine to 8-hydroxy-guanosine in mouse liver mitochondrial DNA by administration of azidothymidine. Biochem Biophys Res Commun 1991; 176:87–93. 49. De la Asuncion JG, Del Olmo ML, Sastre J, Pallardo FV, Vina J. Zidovudine (AZT) causes an oxidation of mitochondrial DNA in mouse liver. Hepatology 1999; 25:985–7. 50. Lo¨ffler M, Jo¨ckel J, Schuster G, Becker C. Dihydroorotat-ubiquinone oxidoreductase links mitochondria in the biosynthesis of pyrimidine nucleotides. Mol Cell Biochem 1997; 174:125–9. 51. Fromenty B, Berson A, Pessayre D. Microvesicular steatosis and steatohepatitis: role of mitochondrial dysfunction and lipid peroxidation. J Hepatol 1997; 26(Suppl. 1):13–22. 52. Day CP, James OF. Hepatic steatosis: innocent bystander or guilty party? Hepatology 1998; 27:1463–6. 53. Pessayre D, Mansouri A, Haouzi D, Fromenty B. Hepatotoxicity due to mitochondrial dysfunction. Cell Biol Toxicol 1999; 15:367–73. 54. Pessayre D, Berson A, Fromenty B, Mansouri A. Mitochondria in steatohepatitis. Semin Liver Dis 2001; 21:57–69. 55. Tsukamoto H, Kim CW, Luo ZZ, Horn W, Su LC, Brittenham GM. Role of lipid peroxidation in in vivo and in vitro models of liver fibrogenesis. Gastroenterology 1993; 104:A101. 56. Susin SA, Zamzani N, Kroemer G. Mitochondria as regulators of apoptosis: doubt no more. Biochim Biophys Acta 1998; 1366:151–65. 57. Pessayre D, Haouzi D, Fau D, Robin MA, Moreau A, Neumann M. Hepatocyte apoptosis triggered by natural substances. In: Cameron RG, Feuer G, eds. Apoptosis and Its Modulation by Drugs. Handbook of Experimental Pharmacology, Vol. 142. Heidelberg: Springer, 2000:59–108. 58. Carr A, Morey A, Mallon P, Williams D, Thorburn DR. Fatal portal hypertension, liver failure, and mitochondrial dysfunction after HIV-1 nucleoside analogue-induced hepatitis and lactic acidaemia. Lancet 2001; 357:1412–4. 59. Clark SJ, Creigthon S, Portmann B, Taylor C, Wendon JA, Cramp ME. Acute liver failure associated with antiretroviral treatment for HIV: a report of six cases. J Hepatol 2002; 36:295–301. 60. Moyle G. Toxicity of antiretroviral nucleoside and nucleotide analogues. Drug Saf 2000; 23:467–81. 61. Hass S, Rockstroh JK, Spengler U, Fischer HP. Nucleoside induced hepatopathy in HIV patients. Diagnostic value of liver biopsy assessment. Der Pathologe 2004; 25:406–11. 62. Saint-Marc T, Touraine JL. The effects of discontinuing stavudine therapy on clinical and metabolic abnormalities in patients suffering from lipodystrophy. AIDS 1999; 13:2188–9. 63. Carr A, Miller J, Law M, Cooper DA. A syndrome of lipoatrophy, lactic acidaemia and liver dysfunction associated with HIV nucleoside analogue therapy: contribution to protease inhibitorrelated lipodystrophy syndrome. AIDS 2000; 14:F25–32. 64. Arnaudo E, Dalakas M, Shanske S, Moraes CT, DiMauro S, Schon EA. Depletion of muscle mitochondrial DNA in AIDS patients with zidovudine-induced myopathy. Lancet 1991; 337:508–10. 65. Simpson DM, Tagliati M. Nucleoside analogue-associated peripheral neuropathy in human immunodeficiency virus infection. J Acquir Immune Defic Syndr 1995; 9:153–61. 66. Maxson CJ, Greenfield SM, Turner JL. Acute pancreatitis as a common complication of 2 0 ,3 0 dideoxyinosine therapy in the acquired immunodeficiency syndrome. Am J Gastroenterol 1992; 87:708–13. 67. Ray AS, Olson L, Fridland A. Role of purine nucleoside phosphorylase in interactions between 2 0 ,3 0 dideoxyinosine and allopurinol, ganciclovir, or tenofovir. Antimicrob Agents Chemother 2004; 48:1089–95. 68. Pollard RB, Robinson P, Dransfield K. Safety profile of nevirapine, a nonnucleoside reverse transcriptase inhibitor for the treatment of human immunodeficiency virus infection. Clin Ther 1998; 20:1071–92. 69. Lee W. Drug-induced hepatotoxicity. N Engl J Med 2003; 349:474–85. 70. Carr A, Cooper DA. Pathogenesis and management of HIV-associated drug hypersensitivity. In: Volberding P, Jacobson MA, eds. AIDS Clinical Review 1995/96. New York: Marcel Dekker, 1996:65–97. 71. Mallal S, Nolan D, Witt C, et al. Association between presence of HLA-B*5701, HLA-DR7, and HLADQ3 and hypersensitivity to HIV-1 reverse transcriptase inhibitor abacavir. Lancet 2002; 359:727–32. 72. Heatherington S, Hughes AR, Mosteller M, et al. Genetic variations in HLA-B region and hypersensitivity reactions to abacavir. Lancet 2002; 359:1121–2. 73. Martin AM, Nolan D, Gaudieri S, et al. Predisposition to abacavir hypersensitivity conferred by HLA-B*5701 and a haplotypic Hsp 70-Hom variant. Proc Natl Acad Sci USA 2004; 101:4180–5. 74. Hughes DA, Vilar FJ, Ward CC, Alfirevic A, Park BK, Pirmohamed M. Cost-effectiveness analysis of HLA-B*5701 genotyping in preventing abacavir hypersensitivity. Pharmacogenetics 2004; 14:335–42. 75. Phillips EJ, Sullivan JR, Knowles SR, Shear NH. Utility of patch testing in patients with hypersensitivity syndromes associated with abacavir. AIDS 2002; 15:2223–5. 76. Martin AM, Nolan D, James I, et al. Predisposition to nevirapine hypersensitivity associated with HLA-DRB1*0101 and abrogated by low CD4-T-cell counts. AIDS 2005; 19:97–9.

586

Spengler

77. Easterbrook PJ, Waters A, Murad S, et al. Epidemiological risk factors for hypersensitivity reactions to abacavir. HIV Med 2003; 4:321–4. 78. King D, Tomkins S, Waters A, et al. Intracellular cytokines may model immunoregulation of abacavir hypersensitivity in HIV-infected subjects. J Allergy Clin Immunol 2005; 115:1081–7. 79. Phillips E, Wong GA, Kaul R, et al. Clinical and immunogenetic correlates of abacavir hypersensitivity. AIDS 2005; 19:979–87. 80. De Maat MMR, ter Hein R, van Gorp ECM, Mulder JW, Mairuhu ATA, Beijnen JH. Case series of acute hepatitis in a non-selected group of HIV-infected patients on vevirapine-containing antiretroviral treatment. AIDS 2003; 17:2209–14. 81. Spengler U, Lichterfeld M, Rockstroh H. Antiretroviral drug toxicity—a challenge for the hepatologist? J Hepatol 2002; 36:283–94. 82. Wooltorton E. HIV drug nevirapine (Viramune): risk of severe hepatotoxicity. CMAJ 2004; 170:1091. 83. Zucker SD, Quin X, Rouster SD, Yu F, Green RM, Keshavan P. Mechanism of indinavir-induced hyperbilirubinemia. Proc Natl Acad Sci USA 2001; 98:12671–6. 84. FDA notifications. Reyataz is approved for HIV treatment. AIDS Alert 2003; 18:119–20. 85. Busti AJ, Hall RG, Margolis DM. Atazanavir for the treatment of human immunodeficiency virus infection. Pharmacotherapy 2004; 24(12):1732–47. 86. Robinson BS, Riccardi KA, Gong YF, et al. BMS-232632, a highly potent human immunodeficiency virus protease inhibitor that can be used in combination with other available antiretroviral agents. Antimicrob Agents Chemother 2000; 44:2093–9. 87. Raijmakers MT, Jansen PL, Steegers EA, Peters WH. Association of human liver bilirubin UDPglucuronosyltransferase activity with a polymorphism in the promoter region of the UGT1A1 gene. J Hepatol 2000; 33:348–51. 88. Rotger M, Taffe P, Bleiber G, et al. Gilber syndrome and the development of antiretroviral therapyassociated hyperbilirubinemia. J Infect Dis 2005; 192:1381–6. 89. Sulkowski MS, Mehta SH, Tobenson M, Afdahl NH, Mirel L, Moore RD, Thomas DL. Hepatic steatosis and antiretroviral drug use among adults coninfected with HIV and hepatitis C virus. AIDS 2005; 19:585–92. 90. Lonardo A, Adinolfi LE, Loria P, Carulli N, Ruggiereo G, Day CP. Steatosis and hepatitis C virus: mechanisms and significance for hepatitis and extrahepatic disease. Gastroenterology 2004; 126:586–97. 91. Adinolfi LE, Gambardella M, Andreana A, Tripodi MF, Utili R, Ruggiero G. Steatosis accelerates the progression of liver damage of chronic hepatitis C patients and correlates with specific HCV genotype and visceral obesity. Hepatology 2001; 33:1358–64. 92. Patton HM, Patel K, Behling C, et al. The impact of steatosis on disease progression and early and sustained treatment response in chronic hepatitis C patients. J Hepatol 2004; 40:484–90. 93. Carr A, Samaras K, Thorisdottir A, Kaufmann GR, Chisholm DJ, Cooper DA. Diagnosis, prediction, and natural course of HIV-1 protease-inhibitor-associated lipodystrophy, hyperlipidaemia, and diabetes mellitus: a cohort study. Lancet 1999; 353:2093–9. 94. Riddle TM, Kuhel DG, Woollett LA, Fichtenbaum CJ, Hui DY. HIV protease inhibitor induces fatty acid and sterol biosynthesis in liver and adipose tissues due to the accumulation of activated sterol regulatory element-binding proteins in the nucleus. J Biol Chem 2001; 276:37514–9. 95. Hashimoto KI, Tsunoda R, Okamoto M, Shigeta S, Baba M. Stavudine selectively induces apoptosis in HIV type 1-infected cells. AIDS Res Hum Retroviruses 1997; 13:193–9. 96. Strahl C, Blackburn EH. Effects of reverse transcriptase inhibitors on telomere length and telomerase activity in two immortalized human cell lines. Mol Cell Biol 1996; 16:53–65. 97. Cattelan AM, Erne E, Salatino A, Trevenzoli M, Carretta G, Meneghetti F, Cadrobbi P. Severe hepatic failure related to nevirapine treatment. Clin Infect Dis 1999; 29:455–6. 98. Peters M. Mechanisms of action of interferons. Semin Liver Dis 1989; 9:235–9. 99. McDonald JA, Caruso L, Karayannis P, Scully LJ, Harris JR, Forster GE, Thomas HC. Diminished responsiveness of male homosexual chronic hepatitis B virus carriers with HTLV-III antibodies to recombinant alpha-interferon. Hepatology 1987; 4:719–23. 100. Scully LJ, Brown D, Lloyd C, Shein R, Thomas HC. Immunological studies before and during interferon therapy in chronic HBV infection. Identification of factors predicting response. Hepatology 1990; 12:1111–7. 101. Rockstroh JK, Spengler U. HIV and HCV coinfection. Lancet Infect Dis 2004; 4:437–44. 102. Quesada JR, Tapaz M, Rios A, Kurzrock R, Gutterman JU. Clinical toxicity of interferons in cancer patients: a review. J Clin Oncol 1986; 4:234–43. 103. Spiegel RJ. The alpha interferons: clinical overview. Semin Oncol 1987; 14(Suppl. 2):1–12. 104. Janssen HL, Brouwer JT, Nevens F, Sanchez-Tapias JM, Craxi A, Hadziyannis S. Fatal hepatic decompensation associated with interferon alfa. European Concerted Action on Viral Hepatitis (EUROHEP). BMJ 1993; 306:107–8. 105. Silva MO, Reddy KR, Jeffers LJ, Hill M, Schiff ER. Interferon-induced chronic active hepatitis? Gastroenterology 1991; 101:840–2.

Hepatic Toxicity of Antiviral Agents

587

’

106. Ruiz-Moreno M, Rua JM, Carren˜o V, Quiroga JA, Manns M, Meyer zum Bu¨schenfelde KH. Autoimmune chronic hepatitis type 2 manifested during interferon therapy in children. J Hepatol 1991; 12:265–6. 107. Papo T, Marcellin P, Bernuau J, Durand F, Poynard T, Benhamou JP. Autoimmune hepatitis exacerbated by alpha-interferon. Ann Intern Med 1992; 116:51–3. 108. Shindo M, DiBisceglie AM, Hoofnagle JH. Acute exacerbation of liver disease during interferon alfa therapy for chronic hepatitis C. Gastroenterology 1992; 102:1406–8. 109. Garcia-Buey L, Garcia-Monzon C, Rodriguez S. Latent autoimmune hepatitis triggered during interferon therapy in patients with chronic hepatitis. Gastroenterology 1995; 108:1770–7. 110. Cianciara J, Laskus T. Development of transient autoimmune hepatitis during interferon treatment of chronic hepatitis B. Dig Dis Sci 1995; 40:1842–4. 111. Bellary S, Schiano T, Hartman G, Black M. Chronic hepatitis with combined features of autoimmune chronic hepatitis and chronic hepatitis C: favorable response to prednisone and azathioprine. Ann Intern Med 1995; 123:32–4. 112. Lorke J, Erhardt A, H€aussinger D. Induction of autoimmune hepatitis by pegylated interferon alfa-2b in chronic hepatitis C. Clin Gastroenterol Hepatol 2004; 2:XX. 113. Sezaki H, Arase Y, Tsubota A, et al. Type C-chronic hepatitis patients who had autoimmune phenomenon and developed jaundice during interferon therapy. J Gastroenterol 2003; 38:493–500. 114. Yamamoto AM, Cresteil D, Homberg JC, Alvarez FI. Characterization of anti-liver kidney microsome antibody (anti-LKM1) from hepatitis C virus-positive and -negative sera. Gastroenterology 1993; 104:1762–7. 115. Muratori L, Lenzi M, Ma Y, Cataleta M, Mieli-Vergani G, Vergani D, Bianchi FB. Heterogeneity of liver/kidney microsomal antibody type 1 in autoimmune hepatitis and hepatitis C virus related liver disease. Gut 1995; 37:406–12. 116. Muratori L, Lenzi M, Cataleta M, et al. Interferon therapy in liver/kidney microsomal antibody type 1-positive patients with chronic hepatitis C. J Hepatol 1994; 21:199–203. 117. Gschwantler M, Schrutka-Kolbl C, Weiss W. Acute exacerbation of anti-liver cytosol antibodypositive autoimmune chronic hepatitis by alpha-interferon. Am J Gastroenterol 1995; 90:2239–40. 118. Saves M, Vandentorren S, Daucourt V, et al. Severe hepatic cytolysis: incidence and risk factors in patients treated by antiretroviral combinations. Aquitaine Cohort, France, 1996–1998. Groupe d Epidemiologie Clinique de Sida en Aquitaine (GECSA). AIDS 1999; 17:F115–21. 119. Fortgang IS, Belitsos PC, Chaisson RE, Moore RD. Hepatomegaly and steatosis in HIVinfected patients receiving nucleoside analog antiretroviral therapy. Am J Gastroenterol 1995; 90:1433–6. 120. Mokrzycki MH, Harris C, May H, Laut J, Palmisano J. Lactic acidosis associated with stavudine administration: a report of five cases. Clin Infect Dis 2000; 30:198–200. 121. Olano JP, Borucki MJ, Wen JW, Haque AK. Massive hepatic steatosis and lactic acidosis in a patient with AIDS who was receiving zidovudine. Clin Infect Dis 1995; 21:973–6. 122. Reisler R, Liou S, Servoss J, Robbins G, Theodore D, Murphy R, Chung R. Incidence of hepatotoxicity and mortality in 21 adult antiretroviral treatment trials (Abstract 43). 1st IAS Conference on HIV Pathogenesis and Treatment. July 2001. Argentina: Buenos Aires. 123. Ormseth EJ, Holtzmuller KC, Goodman ZD, Colonna JO, Batty DS, Sjogren MH. Hepatic decompensation associated with lamivudine: a case report and review of lamivudine-induced hepatotoxicity. Am J Gastroenterol 2001; 96:1619–22. 124. ter Hofstede HJ, de Marie S, Foudraine NA, Danner SA, Brinkman K. Clinical features and risk factors of lactic acidosis following long-term antiretroviral therapy: 4 fatal cases. Int J STD AIDS 2000; 11:611–6. 125. Lai KK, Gang DL, Zawacki JK, Cooley TP. Fulminant hepatic failure associated with 2 0 ,3 0 -dideoxyinosine (ddI). Ann Intern Med 1991; 115:283–4. 126. Lafeuillade A, Hittinger G, Chadapaud S. Increased mitochondrial toxicity with ribavirin in HIV/HCV coinfection. Lancet 2001; 357:280–1. 127. European Agency for the Evaluation of Medicinal Products (EMEA). Reports of lactic acidosis in pregnant women treated with Zerit and Videx. Press release January 31st, 2001. 128. Ellozy S, Massen R, Chao L, Divino CM. Antiretroviral-induced hepatic steatosis and lactic acidosis: case report and review of the literature. Am Surg 2001; 67:680–3. 129. Lonergan JT, Barber RE, Mathews WC. Safety and efficacy of switching to alternative nucleoside analogues following symptomatic hyperlactatemia and lactic acidosis. AIDS 2003; 17:2495–9. 130. Fouty B, Frerman F, Reves R. Riboflavin to treat nucleoside analogue-induced lactic acidosis. Lancet 1998; 352:291–2. 131. Schramm C, Wanitschke R, Galle PR. Thiamine for the treatment of nucleoside analogue-induced severe lactic acidosis. Eur J Anaesthesiol 1999; 16:733–5. 132. Walker UA, Venhoff N, Koch E, Olschweski M, Schneider J, Setzer B. Uridine abrogates mitochondrial toxicity related to nucleoside analogue reverse transcriptase inhibitors in HepG2 cells. Antivir Ther 2003; 8:463–70.

588

Spengler

133. Walker UA, Auclair M, Lebrecht D, Kornprobst M, Capeau J, Caron M. Uridine abrogates the adverse effects of stavudine and zalcitabine on adipose cell functions. Antivir Ther 2004; 9:L21–2. 134. van Groeningen CJ, Leyva A, Kraal I, Peters GJ, Pinedo HM. Clinical and pharmacokinetic studies of prolonged administration of high-dose uridine intended for rescue from 5-FU toxicity. Cancer Treat Rep 1986; 70:745–50. 135. Kelsen DP, Martin D, O’Neil J, et al. Phase I trial of PN401, an oral prodrug of uridine, to prevent toxicity from fluorouracil in patients with advanced cancer. J Clin Oncol 1997; 15:1511–7. 136. Sommadossi JP, Carlisle R, Schinazi RF, Zhou Z. Uridine reverses the toxicity of 3 0 -azido-3 0 deoxythymidine in normal human granulocyte-macrophage progenitor cells in vitro without impairment of antiretroviral activity. Antimicrob Agents Chemother 1988; 32:997–1001. 137. Koch EC, Schneider J, Weiss R, Penning B, Walker UA. Uridine excess does not interfere with the antiretroviral efficacy of nucleoside analogue reverse transcriptase inhibitors. Antivir Ther 2003; 8:485–7. 138. Venhoff N, Zilly M, Lebrecht D, et al. Uridine pharmacokinetics of Mitocnol, a sugarcane extract. AIDS 2005; 19:739–40. 139. Walker UA, Langmann P, Miehle N, Zilly M, Klinker H, Petschner F. Beneficial effects of oral uridine in mitochondrial toxicity. AIDS 2004; 18:1085–6. 140. Boehringer Ingelheim. International Viramune Product Monograph http://hiv.net/link. php?id=121. 141. Cahn P, Johnson M, Nusrat R, Hall D, Robinson P. Hepatic safety with nevirapine and two nucleosides in patients with advanced HIV infection, from a placebo controlled clinical endpoint trail (Abstract PL 8.6). 5th International Congress in HIV infection. October 2000. Glasgow, Scotland. 142. Dieterich DT, Robinson PA, Love J, Stern JO. Drug-induced liver injury associated with the use of nonnucleoside reverse-transcriptase inhibitors. Clin Infect Dis 2004; 38(Suppl. 12):S80–9. 143. Lebray P, Benhamou Y, Pol S, Myers R, Poynard T, Di Martino V. NNRTI-related or unrelated hepatotoxicity. Hepatology 2002; 36:513–4 (Letter). 144. Ena J, Amador C, Benito C, Fenoli V, Pasquau F. Risk and determinants of developing severe liver toxicity during therapy with nevirapine- and efavirenz-containing regimens in HIV-infected patients. Int J STD AIDS 2003; 14:776–81. 145. Bonnet F, Lawson-Ayayi S, Thiebaut R, et al. A cohort study of nevirapine tolerance in clinical practice: French Aquitaine Cohort, 1997–1999. Clin Infec Dis 2002; 3:1231–7. 146. Centers for Disease Control and Prevention. Serious adverse events attributed to nevirapine regimens for postexposure prophylaxis after HIV exposures-worldwide, 1997–2000. Morb and Mortal Wkly Rep 2001; 49:1153–6. 147. Viramune (nevirapine) important new safety information. February 2004. (www.fda.gov/ medwatch/SAFETY/2004/safety04.htm#viramune). 148. EMEA public statement on Viramune (nevirapine) - severe and life-threatening cutaneous and hepatic reactions. (http://hiv.net/link.php?id=120). 149. Johnson S, Chan J, Bennet CL. Hepatotoxicity after prophylaxis with a nevirapine-containing regimen. Ann Intern Med 2002; 347:146–7. 150. Clarke S, Harrington P, Condon C, Kelleher D, Smith OP, Mulcahy F. Late onset hepatitis and prolonged deterioration in hepatic function associated with nevirapine therapy. Int J STD AIDS 2000; 11:336–7. 151. Arribas JR, Ibanez C, Ruiz-Antoran B, et al. Acute hepatitis in HIV-infected patients during ritonavir treatment. AIDS 1998; 12:1722–4. 152. Matsuda J, Gohchi K. Severe hepatitis in patients with AIDS and haemophilia B treated with indinavir. Lancet 1997; 350:364. 153. Brau N, Leaf HL, Wieczorek RL, Margolis DM. Severe hepatitis in three AIDS patients treated with indinavir. Lancet 1997; 349:924–5. 154. Saves M, Raffi F, Clevenbergh P, et al. Hepatitis B or hepatitis C virus infection is a risk factor for severe hepatic cytolysis after initiation of a protease inhibitor-containing antiretroviral regimen in human immunodeficiency virus-infected patients. The APROCO Study Group. Antimicrob Agents Chemother 2000; 44:3451–5. 155. Cooper CL, Pabrhakar MA, Angel JB. Hepatoxicity associated with antiretroviral therapy containing dual versus single protease inhibitors in individuals coinfected with hepatitis C virus and human immunodeficiency virus. Clin Infect Dis 2002; 34:1259–63. 156. Aceti A, Pasquazzi C, Zechini B, De Bac C. Hepatoxicity development during antiretroviral therapy containing protease inhibitors in patients with HIV: the role of hepatitis B and C virus infection. J Acquir Immune Defic Syndr 2002; 29:41–8. 157. Sulkowski MS, Mehta SH, Chaisson RE, Thomas DL, Moore RD. Hepatotoxicity associated with protease inhibitor-based antiretroviral regimens with and without concurrent ritonavir. AIDS 2004; 18:2277–84.

Hepatic Toxicity of Antiviral Agents

589

158. Gonzalez-Requena D, Nun˜ez M, Jimenez-Nacher I, Gonzalez-Lahoz J, Soriano V. Liver toxicity of lopinavir containing regimens in HIV-infected patients with and without hepatitis C co-infection. AIDS Res Hum Retrovirus 2004; 20:698–700. 159. Bonfanti P, Ricci E, Penco G, et al. Low incidence of hepatotoxicity in a cohort of HIV patients treated with lopinavir/ritonavir. AIDS 2005; 19:1433–4. 160. Sulkowski MS. Drug-induced liver injury associated with antiretroviral therapy that includes HIV-1 protease inhibitors. Clin Infect Dis 2004; 38:S90–7. 161. Almond LM, Boffito M, Hoggard PG, et al. The relationship between nevirapine plasma concentrations and abnormal liver function tests. AIDS Res Hum Retrovirus 2004; 20:716–22. 162. Arribas J, Pulido F, Peng J, et al. Evaluation of multiple-dose pharmacokinetics of lopinavir/ ritonavir (LPV/r) in HIV and HCV co-infected patients with mild or moderate hepatic insufficiency. 9th European AIDS Conference Warsaw 2003; Abstract #F2/6. 163. Picard O, Rosmorduc O, Cabane J. Hepatotoxicity associated with ritonavir. Ann Intern Med 1998; 129:670–1. 164. Malaty LI, Kuper JJ. Drug interactions of HIV protease inhibitors. Drug Saf 1999; 20:147–69. 165. Dieterich D, Fischl M, Rimland D, Sepulveda G. The safety and efficacy of protease inhibitors in HCV co-infected patients. 9th CROI, Seattle 2002; Abstract 663-M. 166. Pedneault L, Brothers C, Pagano G, Tymkewycz P, Yeo J, Millard J, Fetter A. Safety profile of amprenavir in the treatment of adult and pediatric patients with HIV infection. Clin Ther 2000; 22:1378–94. 167. Pierone G, Drulak M, Arasteh K, et al. A long-term open-label rollover trial assessing the safety and tolerability of combination tipranavir and ritonavir (TPV/r) us in HIV-1-infected patients. 3rd IAS Conference Rio de Janeiro 2005. (WePe6.2C05). 168. Kemmer N, Molina C, Fuchs J, Miskovsky E, Asmuth D, Borucki S, Weinman S. A distinctive histologic pattern of liver injury in HIV-positive patients on HAART: a possible hepatotoxic effect of protease inhibitors (Abstract 431). Annual Meeting of the American Association for the Study of the Liver. November 2000. Dallas, U.S.A. 169. Cox SR, Ferrey JJ, Batts DH, et al. Delavirdine and marketed protease inhibitors: pharmakokinetic interaction studies in healthy volunteers (Abstract 372). 4th Conference on Retroviruses and Opportunistic Infections. January 1996. Washington, DC, U.S.A. 170. Fiske W, Benedek Ich, Joseph JL. Pharmacokinetics of efavirenz and ritonavir after multiple oral doses in healthy volunteers (Abstract 42269). 12th World AIDS Conference. June 1998. Geneva, Switzerland. 171. Indinavir package insert by Merck, Sharpe & Dohme Company. 172. Ouellet D, Hsu A, Granneman GR, Carlson G, Cavanaugh J, Guenther H, Leonard JM. Pharmacokinetic interaction between ritonavir and clarithromycin. Clin Pharmacol Ther 1998; 64:355–62. 173. Ungo JR, Jones D, Ashkin D, Hollender ES, Bernstein D, Albanese AP, Pitchenik AE. Antituberculosis drug-induced hepatotoxicity. The role of hepatitis C virus and the human immunodeficiency virus. Am J Respir Crit Care Med 1998; 157:1871–6. 174. Centers for Disease Control and Prevention. Prevention and treatment of tuberculosis among patients with human immunodeficiency virus: principles of therapy and revised recommendations. MMWR Morb Mortal Wkly Rep 1998; 47:1–58. 175. Vogel M, Voigt E, Michaelis HC, et al. Management of drug-to-drug interactions between cyclosporine A and the protease-inhibitor lopinavir/ritonavir in liver-transplanted HIV-infected patients. Liver Transpl 2004; 10:939–44. 176. www.projinf.org/fs/drugin.html 177. Robertson SM, Penzak SR, Pau AK. Drug interactions in the management of HIV infection. Expert Opin Pharmacother 2005; 6:233–53. 178. Fellay J, Marzolini C, Greub G, Buclin T, Eap CB, Telenti, A. Predictive power of p-glycoprotein and CYP2D6 polymorphisms for plasma levels of antiretroviral agents. 8th CROI Chicago 2001; Abstracts 260. 179. Kim AY, Chung RT, Polsky B. Human immunodeficiency virus and hepatitis B and C coinfection: pathogenic interactions, natural history and therapy. AIDS Clin Rev 2000–2001:263–306. 180. Ockenga J, Tillmann HL, Trautwein C, Stoll M, Manns MP, Schmidt RE. Hepatitis B and C in HIVinfected patients. Prevalence and prognostic value. J Hepatol 1997; 27:18–24. 181. Dieterich DT. Hepatitis C virus and human immunodeficiency virus: clinical issues in coinfection. Am J Med 1999; 107:S79–84. 182. Diamondstone LS, Blakley SA, Rice JC, Clark RA, Goedert JJ. Prognostic factors for all cause mortality among hemophiliacs infected with human immunodeficiency virus. Am J Epidemiol 1995; 142:304–13. 183. Altfeld M, Rockstroh JK, Addo M, Kupfer B, Pult I, Will H, Spengler U. Reactivation of hepatitis B in a long-term anti-HBs-positive patient with AIDS following lamivudine withdrawal. J Hepatol 1998; 29:306–9.

590

Spengler

184. Cooley L, Ayres A, Bartholomeusz A, et al. Prevalence and characterization of lamivudine-resistant hepatitis B virus mutations in HIV-HBV co-infected individuals. AIDS 2003; 17:1649–57. 185. Locarnini S, McMillan J, Bartholomeusz A. The hepatitis B virus and common mutants. Semin Liver Dis 2003; 23:5–20. 186. Carr A, Cooper DA. Restoration of immunity to chronic hepatitis B infection in an HIV-infected patient on protease inhibitor. Lancet 1997; 349:996–7. 187. Mastroianni CM, Trinchieri V, Santopadre P, et al. Acute clinical hepatitis in an HIV-positive hepatitis B carrier receiving protease inhibitor therapy. AIDS 1998; 112:1939–40. 188. Velasco M, Moran A, Tellez MJ. Resolution of chronic hepatitis B after ritonavir treatment in an HIVinfected patient. N Engl J Med 1999; 340:1765–6. 189. Castera L, Chouteau P, Hezode C, Zafrani ES, Dhumeaux D, Pawlotsky JM. Hepatitis C virusinduced hepatocellular steatosis. Am J Gastroenterol 2005; 100:711–5. 190. Barbaro G, Di Lorenzo G, Asti A, et al. Hepatocellular mitochondrial alterations in patients with chronic hepatitis C: ultrastructural and biochemical findings. Am J Gastroenterol 1999; 94:2198–205. 191. Okuda M, Li K, Beard MR, Showalter LA, Scholle F, Lemon SM, Winman SA. Mitochondrial injury, oxidative stress, and antioxidant gene expression are induced by hepatitis C virus core protein. Gastroenterology 2002; 122:366–75. 192. Korenaga M, Wang T, Li Y, Showalter LA, Chan T, Sun J, Winman SA. Hepatitis C virus core protein inhibits mitochondrial electron transport and increases ROS production. J Biol Chem 2005; 280: 37481–8. 193. Chou AH, Tsai HF, Wu YY, Hu CY, Hwang LH, Hsu PI, Hsu PN. Hepatitis C virus core protein modulates TRAIL-mediated apoptosis by enhancing Bid cleavage and activation of mitochondria apoptosis signaling pathway. J Immunol 2005; 174:2160–6. 194. Balasubramian A, Koziel M, Groopman JE, Ganju RK. Molecular mechanism of hepatic injury in coinfection with hepatitis C virus and HIV. Clin Infect Dis 2005; 41(Suppl. 1):S32–7. 195. Yamaguchi A, Tazuma S, Nishioka T, Ohishi W, Hyogo H, Momura S, Chayama K. Hepatitis C virus core protein modulates fatty acid metabolism and thereby causes lipid accumulation in the liver. Dig Dis Sci 2005; 50:1361–71. 196. Uberti-Foppa C, De Bona A, Morsica G, Galli L, Gallotta G, Boeri E, Lazzarin A. Pretreatment of chronic active hepatitis C in patients coinfected with HIV and hepatitis C virus reduces the hepatotoxicity associated with subsequent antiretroviral therapy. J Acquir Imm Defic Syndr 2003; 33:146–52. 197. John M, Flexman J, French MAH. Hepatitis C virus-associated hepatitis following treatment of HIV-infected patients with HIV protease inhibitors: an immune restoration disease? AIDS 1998; 12:2289–93. 198. Pe´rez-Olmeda M, Garcı´a-Samanı´ego J, Soriano V. Hepatitis C viraemia in HIV-HCV co-infected patients having immune restoration with highly active antiretroviral therapy. AIDS 2000; 14:212. 199. Torti C, Lapadula G, Casari S, Puoti M, Nelson M, Quiros-Roldan E, Bella D, Pastore G, Ladisa N, Minoli L, et al. BMC Infect Dis 2005; 5:58. 200. Rutschmann OT, Negro F, Hirschel B, Hadengue A, Anwar D, Perrin LH. Impact of treatment with human immunodeficiency virus (HIV) protease inhibitors on hepatitis C viremia in patients coinfected with HIV. J Infect Dis 1998; 177:783–5. 201. Vento S, Garofano T, Renzini C. Enhancement of hepatitis C replication and liver damage in HIVconinfected patients on antiretroviral combination therapy. AIDS 1998; 12:116–7. 202. Chung RT, Evans SR, Yang Y, et al. Immune recovery is associated with persistent rise in hepatitis C virus RNA, infrequent liver test flares, and is not impaired by hepatitis C virus in co-infected subjects. AIDS 2002; 16:1915–23. 203. Nu´n˜ez M, Rı´os P, Martı´n-Carbonero L, Pe´rez-Olmeda M, Gonza´lez-Lahoz J, Soriano V. Role of hepatitis C virus genotype in the development of severe transaminase elevation after the introduction of antiretroviral therapy. J Acquir Imm Defic 2002; 30:65–8. 204. Andre´ P, Goettrup M, Klenerman P, et al. An inhibitor of HIV protease modulates proteasome activity, antigen presentation, and T cell responses. Proc Natl Acad Sci USA 1998; 95:12120–4. 205. Michelet C, Chapplain JM, Petsaris O, Arvieux C, Ruffault A, Lotteau V, Andre´ P. Differential effect of ritonavir and indinavir on immune responses to hepatitis C in HIV-1 infected patients. AIDS 1999; 13:1995–6. 206. Fialaire P, Payan C, Vitour D, Chennebault JM, Loison J, Pichard E, Lunel F. Sustained disappearance of hepatitis C viremia in patients receiving protease inhibitor treatment for HIV infection. J Infect Dis 1999; 180:574–5. 207. Zylberberg H, Chaix ML, Rabian C, et al. Tritherapy for human immunodeficiency virus infection does not modify replication of hepatitis C virus in coinfected subjects. Clin Infect Dis 1998; 26:1104–6.

Hepatic Toxicity of Antiviral Agents

591

208. Martı´n-Carbonero L, Nu´n˜ez M, Rı´os P, Pe´rez-Olmeda M, Gonza´lez-Lahoz J, Soriano V. Liver injury after beginning antiretroviral therapy in HIV/hepatitis C virus co-infected patients is not related to immune reconstitution. AIDS 2002; 16:1423–5. 209. Aranzabal L, Casado JL, Moya J, et al. Influence of liver fibrosis on highly active antiretroviral therapy-associated hepatotoxcity in patients with HIV and hepatitis C virus coinfection. Clin Infect Dis 2005; 40:588–93. 210. Mauss S, Valenti W, De Pamphilis J, et al. Risk factors for hepatic decompensation in patients with HIV/HCV coinfection and liver cirrhosis during interferon-based therapy. AIDS 2004; 18:F21–5. 211. Cameron DW, Valdes J, Garber C. Ritonavir pharmacokinetics in HIV-infected patients with underlying hepatic disease (Abstract 359). 5th Conference on Retroviruses and Opportunistic Infections. February 1998. Chicago, U.S.A. 212. Tachikawa N, Yoshizawa S, Kikuchi Y, Yasuoka A, Oka S. Saquinavir therapy in patients with advanced HIV infection and liver cirrhosis. Jpn J Infect Dis 1999; 52:177. 213. Maserati R, Villani P, Seminari E, Pan A, Lo Caputo S, Regazzi MB. High plasma levels of nelfinavir and efavirenz in two HIV-positive patients with hepatic disease. AIDS 1999; 13:870–1. 214. Gonza´lez de Requena D, Nu´n˜ez M, Jime´nez-Na´cher I, Soriano V. Liver toxicity caused by nevirapine. AIDS 2002; 16:290–1. 215. Lamson M, Maldonado S, Hutman H. The effects of underlying renal or hepatic dysfunction on the pharmacokinetics of nevirapine (Abstract TuPeB 3301). XIII International Conference on AIDS. July 2000. Durban, South Africa. 216. Agenerase (amprenavir) capsules and oral solution. Package insert, Research Triangle Park, NC, Glaxo Wellcome 1999. 217. Reyataz (atazanavir sulfate). Prescribing Information (U.S.A.). Princeton, NJ: Bristol Myers Squibb, 2003. 218. French AL, Benning L, Anastos K, Augenbraun M, Nowicki M, Sathasivam K, Terrault NA. Longitudinal effect of antiretroviral therapy on markers of hepatic toxicity: impact of hepatitis C coinfection. Clin Infect Dis 2004; 39:402–10. 219. Leitze Z, Nadeem A, Choudhary A, Saul Z, Roberts I, Manthous CA. Nevirapine-induced hepatitis treated with corticosteroids? AIDS 1998; 12:1115–7.

28

Hepatotoxicity of Cardiovascular and Antidiabetic Medications Sidharth S. Bhardwaj and Naga P. Chalasani

Division of Gastroenterology and Hepatology, Indiana University School of Medicine, Indianapolis, Indiana, U.S.A.

INTRODUCTION Drugs used in the treatment of cardiovascular disorders and diabetes are among the most commonly prescribed medications in the United States. Figures 1 and 2 show a classification of cardiovascular and antidiabetic medications. Use of these agents has continuously increased due to the improved recognition of these disorders and their complications. Additionally, the ever-growing problem of obesity and the metabolic syndrome has caused an increase in the incidence of diabetes, dyslipidemia, hypertension, and subsequent coronary artery disease and dysrhythmias. There is high prevalence of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH) in patients with obesity, diabetes, and dyslipidemia, and distinguishing medication-induced liver disease from underlying liver disease can be difficult. This leads to difficulty in distinguishing whether the liver disease is due to cardiovascular or diabetic drugs themselves or is a manifestation of underlying liver disease resulting from metabolic disorders and diabetes. Evolving data suggest that patients with diabetes may have higher risk of drug-induced liver disease than nondiabetic individuals. Diabetic patients who take oral hypoglycemics have a higher risk of developing druginduced liver disease than those not taking these medications (1,2). An excellent review was recently published by Wang and colleagues that provided a detailed review of existing human and animal studies supporting the notion that patients with diabetes are generally at higher risk for developing drug hepatotoxicity (3).

CARDIOVASCULAR DRUGS Antiarrhythmic Drugs A broad range of medications are now used for the control of cardiac dysrhythmias. As a class, they generally function by prolonging the action potential of myocardial conductive cells. This discussion will focus on potassium channel blockers (class III antiarrhythmics), sodium channel blockers (class I antiarrhythmics), and miscellaneous antiarrhythmics such as adenosine and digoxin. b-blockers (class II antiarrhythmics) and calcium channel blockers (class IV antiarrhythmics) are discussed in the section on antihypertensives. When considering drugs mainly used for the treatment of arrhythmias, the available literature suggests that most clinically evident hepatic injury is produced by three medications: amiodarone, quinidine, and procainamide. Although each class of antiarrhythmics contains medications that share a mechanism of action, the drugs are diverse in regard to their biochemical Naga Chalasani has been a consultant to Merck, Ortho-McNeil, Abbott, Metabasis, Atherogenics, and Advanced Life Sciences, on issues generally related to hepatotoxicity of prescription medications.

594

1.

2.

Bhardwaj and Chalasani

Anti-anginal drugs a. Nitrates—nitroglycerin, isosorbide dinitrate, amyl nitrate b. Calcium-channel blockers—bepridil; others as below c. Beta blockers—as below Antihypertensive agents a. Diuretics—amiloride, bumetanide, furosemide, chlorothiazide, chlorthalidone, b. c. d. e. f.

3.

5. 6. 7. 8.

enalapril, fosinopril, lisinopril, moexipril, perindopril, quinapril, ramipril, trandolapril Angiotensin-2 Receptor Antagonists—candesartan, eprosartan, irbesartan, losartan, telmisartan, valsartan Beta Blockers—acebutolol, atenolol, betaxolol, bisoprolol, carteolol, metoprolol, nadolol, propranolol, sotalol, timolol Combined alpha/beta blockers—labetalol Calcium Channel Blockers—amlodipine, diltiazem, felodipine, nifedipine, nimodipine, nisoldipine, verapamil CNS-acting antihypertensives—clonidine, methyldopa, guanabenz Direct vasodilators—hydralazine, minoxidil, nitroprusside, diazoxide Vascular smooth muscle relaxer—nesiritide Alpha-1-adronergic antagonist—prazosin, terazosin Adrenergic neuron blocking agents—guanethedine, reserpine Endothelin-receptor antagonists—bosentan

g. h. i. j. k. l. Antiarrhythmic drugs a. Class I (sodium channel blockers)—disopyramide, flecainide, lidocaine, b. c.

4.

hydrochlorothiazide, indapamide, spironolactone, triamterene

Angiotensin-Converting Enzyme (ACE) Inhibitors—benazepril, captopril,

mexiletine, moricizine, procainamide, procanabid, propafenone, quinidine, tocainide, encainide, aprindine Class II (beta blockers)—as above Class III (potassium channel blockers)—amiodarone, azimilide, bepridil, dofetilide, ibutalide, sotalol, tedisamil Class IV (calcium channel blockers)—as above Miscellaneous—adenosine, digoxin

d. e. Cholesterol-lowering drugs a. Fibrates—clofibrate, gemfibrozil b. Nicotinic Acid c. Resins—cholestyramine, colestipol, colesevelam d. Statins—atorvastatin, fluvastatin, lovastatin, pravastatin, rosuvastatin, simvastatin e. Ezetimibe Anticoagulant drugs—dalteparin, danaparoid, enoxaparin, heparin, tinzaparin, warfarin,

lepirudin, bivalirudin, argatroban, melagatran, ximelagatran

Antiplatelet agents—dipyridamole, aspirin, cilostazol, clopidogrel, ticlodipine Glycoprotein IIb/IIIa receptor antagonists—abciximab inj, eptifibatide inj, tirofiban inj Thrombolytics—tissue plasminogen activator, urokinase, streptokinase, anistreplase

FIGURE 1 Classification of cardiovascular drugs.

properties. Therefore, most of these medications will be discussed individually. Please refer to the outline for a classification of these medications. Amiodarone There has been a dramatic rise in the popularity of amiodarone since its first clinical use as an antianginal. Although its use in the recent past was limited to the treatment of refractory 1. 2.

Insulins Sulfonylureas—acetohexamide, chlorpropamide, tolbutamide, tolazamide, glipizide, glyburide,

3. 4. 5. 6.

Biguanides—metformin Thiazolidinediones—rosiglitazone, pioglitozone; [historical—troglitazone] Alpha-glucosidase inhibitors—acarbose, miglitol Meglitinides—repaglenide, nateglinide

glimepiride

FIGURE 2 Classification of antidiabetic drugs.

Cardiovascular and Antidiabetic Medications

595

ventricular tachyarrhythmias, it is now becoming a common drug for the prevention and treatment of atrial fibrillation (4). Various organ toxicities have led to discontinuation of the drug in a significant proportion of patients, and are mainly due to adverse effects on the lung, liver, cornea, thyroid, and skin (5–7). The drug is withdrawn from patients up to 25% of the time at its highest estimate but generally is quoted to be around 12% to 15% of the time in most centers (5,8–12). Amiodarone hepatotoxicity is mostly seen with long-term administration but may also occur acutely. Acute liver injury is rare but has been demonstrated in several case reports and case series of patients undergoing IV amiodarone loading (13–21). A recent case series and review of literature revealed that roughly two-thirds of the acute cases of hepatic laboratory abnormalities occurred within 24 hours of initiation of the drug with the majority of cases occurring within three days (13,15). A greater than tenfold increase in the transaminases occurred in most cases and an elevation of bilirubin was seen in about half of the patients (5,22,23). Histologic findings predominantly consist of centrilobular necrosis. If the outcome is not fatal, acute amiodarone hepatotoxicity is generally reversible (14,15,24). Rechallenge with oral amiodarone after discontinuation of IV amiodarone has been well tolerated according to the currently available published reports (14,15,23,25–27). A cosolvent of IV amiodarone, polysorbate 80, has been postulated to either impair hepatocyte cell integrity or impair liver perfusion as a result of polysorbate-induced hypotension (13,18,22,28). Additionally, IV amiodarone hepatotoxicity may be related to higher plasma concentrations of the drug and a rapid onset of the toxic effects generally seen with long-term administration (14,18). As opposed to IV amiodarone, oral amiodarone hepatotoxicity with long-term administration is a well-studied entity. Hepatotoxicity of amiodarone and its primary metabolite, desethylamiodarone, frequently takes the form of an asymptomatic elevation in transaminases in fully one-quarter of patients on the drug (10,29). Once symptomatic disease has begun, the mortality rate can be high (24,29–33). Two proposed pathophysiologic mechanisms for amiodarone hepatotoxicity have become generally well-accepted. The first mechanism involves phospholipidosis due to inhibition of the intralysosomal enzyme phospholipase A (34,35). This is a result of the binding of the drug to phospholipids and leads to the inability of hepatic lysosomes to release phospholipids (Fig. 3). This, in turn, leads to oxidative damage and becomes a cornerstone of the histologic findings of pseudoalcoholic liver disease (discussed below) (29,30,34,38). The second pathophysiologic mechanism is a mitochondrial dysfunction that leads to the induction of apoptosis and hepatocyte necrosis. Inhibition of mitochondrial enzymes of the electron transport chain and, subsequently, of b-oxidation leads to an uncoupling of oxidative phosphorylation (39,40). This second hypothesis is supported by the demonstration that apoptosis is reduced by the addition of the antioxidant ascorbate (38). Clinically, hepatic disease due to amiodarone toxicity can range from minor asymptomatic biochemical abnormalities to full-blown cirrhosis (5,24,41). The usually modest elevation in transaminases does not often exceed four to five times the upper limit of normal (ULN) although much higher elevations have been reported (7,42). Transaminase levels usually return to normal after dose reduction or discontinuation of the drug (7,43). On the other hand, amiodarone, and its metabolites may still be found in the body several months after the discontinuation of the drug due to its long half-life and lipophilic properties. In this case, hepatic abnormalities may persist despite withdrawal of the drug (44). The presence of lab abnormalities usually predicts the physical exam finding of hepatomegaly in these patients (45). Additional clinical features include those of other chronic liver diseases and/or cirrhosis, e.g., fatigue, anorexia with weight loss, splenomegaly, and findings of portal hypertension (ascites, caput medusae, etc.) (45). Rarely, disease similar to Reye’s syndrome with the development of coma or alcoholic liver disease (termed “pseudoalcoholic liver disease”) can develop due to steatohepatitis (5,7,29,30,34,46). Physical exam findings in this syndrome may include spider angiomata (5). Histologic findings are generally nonspecific but may show several characteristic features (Fig. 3A–C). These include microvesicular steatosis often confused with Reye’s syndrome or alcoholic hepatitis (7,29,30,47). There also may rarely be associated mallory bodies and necrosis yielding even further similarities to alcoholic liver disease or NASH (7,29,47,48). Fibrosis is not

596

Bhardwaj and Chalasani

(A)

(B)

(C)

FIGURE 3 (A) Liver pathology of a patient with long-term amiodarone ingestion. Nearly all of the hepatocytes in the field exhibit ballooning and mallory bodies (600!). (B) Liver pathology of a patient with long-term amiodarone ingestion. Ubiquitin immunostain. Some nodules showed extensive mallory body formation (200!). This degree of mallory body formation is very rare outside the setting of amiodarone toxicity. (C) Electron microscopy showing characteristic lysosomal inclusions in a patient with amiodarone hepatotoxicity. Source: From Refs. 36, 37.

an uncommon feature (47,49). Electron microscopy yields the most pathognomic feature of amiodarone hepatotoxicity: large lysosomes engorged with phospholipids with a laminated appearance (29,47,50). The collective literature reveals that toxicity is dose-dependent with hepatotoxicity being relatively infrequent when amiodarone is prescribed in lower doses (7,50). However, lab monitoring has typically been difficult for prescribers due to a seemingly arbitrary monitoring interval and limited availability of evidence-based guidelines for dosage adjustments. It is

Cardiovascular and Antidiabetic Medications

597

conventionally recommended that liver enzymes be checked before initiating therapy and every six months thereafter (44,50). A recent five-year prospective study evaluated the correlation of amiodarone concentrations with the risk of toxicity and suggested that alanine aminotransferase (ALT) monitoring should be done at baseline, one month, three months, then semiannually (51). Further studies will be necessary to confirm if this is an appropriate strategy to detect hepatotoxicity. Quinidine Quinidine-induced hepatotoxicity has been shown to occur at a frequency of about two percent of patients on the drug and occurs by a mechanism believed to be a hypersensitivity reaction (52,53). Elevated aminotransferases and the acute onset of fever characterize the clinical syndrome of quinidine hepatotoxicity and generally occur one to two weeks after initiation of the medication (54,55). This characteristic pattern of toxicity, in addition to a more prompt recurrence of these findings upon rechallenge with the medication, support the theory that hypersensitivity leads to the pathogenesis of quinidine hepatotoxicity (52). Discontinuation of the medication leads to a rapid return to normal hepatic function (53,55,56). Hepatosplenomegaly with fever often accompanies the biochemical markers of liver injury. Jaundice is relatively rare (57). Biopsy findings often include periportal inflammation with centrilobular necrosis and may also consist of nonspecific focal necrosis but more often reveal granulomatous hepatitis with eosinophilia (58–60). The recognition of quinidine hepatotoxicity may be aided by several clinical indicators. The hepatic injury is generally preceded by fever, which may be accompanied by gastrointestinal symptoms, rash, and thrombocytopenia (56,61,62). As noted, these symptoms reverse after discontinuation of the medication and reappear with reinitiation of the medication (53,56). Procainamide Hepatotoxicity with procainamide is extremely rare (63). Liver injury may be seen anywhere from one day to six weeks after the initiation of the medication (63). The mechanism for liver dysfunction is likely delayed drug hypersensitivity and, thus, may include symptoms of fevers, chills, rash, abdominal pain; arthralgias, and nausea with vomiting (64–66). The acetylated metabolite of the drug is generally considered to be less toxic but may have an additive hepatotoxic effect (67,68). Thus, patients who are genetically “slow acetylators” have a theoretic increased likelihood of developing procainamide-induced liver injury (67). Generally, a mixed pattern of liver disease is seen on laboratory assessment with a more prominent elevation in the transaminases (63). There is, however, another pattern of liver injury that may occur consisting of a largely cholestatic pattern of liver injury that is mainly seen after intravenous loading of procainamide (64). Procainamide may cause intrahepatic cholestasis or granulomatous hepatitis upon histologic examination (65). Discontinuation of the drug leads to the resolution of symptoms within several days although resolution of the biochemical and histologic markers of liver disease may take up to several weeks (63,66). Rechallenge with the medication causes a rapid recurrence of the liver disease (63). Aprindine Aprindine is not available for clinical use in the United States. Cholestatic jaundice is a side effect of aprindine which can be minimized by the careful titration of the dose (69). Elevations in the levels of the transaminases as well as an increase in alkaline phosphatase have been noted in conjunction with a lobular pattern of inflammation (70,71). Hepatotoxicity is usually seen within three to six weeks of initiating therapy and improved quickly upon drug withdrawal (70). Granulomatous hepatitis can also be seen within six weeks of initiating therapy and also resolves rapidly with discontinuation of drug use (72). Propafenone Propafenone toxicity is a highly uncommon occurrence; it has been shown to cause a mixed pattern of hepatocellular and cholestatic liver injury with a prolonged recovery of alkaline phosphatase upon drug withdrawal (73–76). Liver enzyme elevation is usually minor and asymptomatic and can sometimes improve with dose reduction (75). Most patients present

598

Bhardwaj and Chalasani

with jaundice usually about two to six weeks after initiation of the drug (75,76). Liver biopsy generally reveals portal inflammation with granulocytes and lymphocytes and perhaps, rare ballooning of hepatocytes and bile ductular proliferation (73). The mechanism of injury is not well understood but is likely due to either direct toxic injury or from drug hypersensitivity (73). A direct toxic reaction is supported by the pattern of injury and the short time course between ingestion and toxicity in some cases, whereas a hypersensitivity reaction is suggested by the presence of mild eosinophilia in several cases (74). Other Antiarrhythmics Flecainide has been shown to cause hepatocellular injury that reverses rapidly after withdrawal of the drug (77,78). A case of cholestasis has been reported when flecainide was combined with enalapril (78). Disopyramide ingestion has been associated with severe hepatocellular damage or intrahepatic cholestasis that resolved within weeks after cessation of therapy (79–81). Encainide may cause asymptomatic minimal elevations in the aminotransferases and, very rarely, a more significant transaminase elevation (up to 30 times the ULN) (82,83). There have been reports of reversible cholestatic hepatitis with mexiletine and lidocaine (84–87). There is a case report of granulomatous hepatitis with tocainide use (88,89). Lipid Lowering Agents Although the hydroxymethylglutaryl-CoA (HMG-CoA) reductase inhibitors (or “statins”) have received the most attention for their proclivity to cause hepatotoxicity, any of the lipid lowering agents may cause drug-induced liver injury, with the exception of cholestyramine which actually may be hepato-protective (90,91). Sustained-release (SR) niacin is the only drug in this category that causes significant and symptomatic hepatotoxicity (91,92). The other agents and classes (including the statins) rarely cause clinically significant liver injury and may, unfortunately, be prematurely withdrawn from patients who may benefit from them (92). HMG-CoA Reductase Inhibitors (Statins) Statins are prescribed extensively for primary and secondary prevention of cardiovascular disease (93). A number of clinical trials have shown that statins significantly decrease mortality from coronary artery disease and decrease the incidence of myocardial infarction, revascularization procedures, stroke, and peripheral vascular disease (94). Currently, there are six statins available for clinical use including lovastatin which may become over the counter in the United States. These medications are widely prescribed and more than 100 million prescriptions were written in the USA in year 2004 alone (Table 1) (96). Numerous clinical trials and extensive post-marketing experience have shown that currently available statins are extremely safe. Early animal toxicology studies suggested that statins may cause hepatotoxicity. At usual doses, lovastatin did not cause significant liver injury but given at very high doses it caused hepatocellular necrosis in rabbits (97). Similarly, high doses of simvastatin caused hepatocellular necrosis in guinea pigs (98). Such liver injury could be prevented or reversed by supplementing animals with mevalonate, suggesting that depletion of mevalonate or its downstream metabolite might be responsible for liver injury (97,98). However, there is no evidence that statins cause hepatocellular necrosis in humans in a predictable or dose-dependent fashion. Asymptomatic elevations in aminotransferases are common in patients receiving statins but clinically TABLE 1 Number of Prescriptions Written for Various Statins in the United States for Years 2003 and 2004 Atorvastatin Simvastatin Pravastatin Fluvastatin Lovastatin Rosuvastatin Source: From Ref. 95.

2003 (in millions)

2004 (in millions)

53.4 24 13.4 5 1.7 N/A

62.5 23.8 12.0 1.9 7.4 6.3

599

Cardiovascular and Antidiabetic Medications

TABLE 2 Relationship Between the Dose of Statin and the Incidence of Persistent Elevation of ALT Greater than Three Times ULN Placebo (%) Lovastatin Simvastatin Pravastatin Fluvastatin Atorvastatin Rosuvastatin

Statin 10 mg (%)

0.1 1.3 0.28

0.2 0

Statin 20 mg (%)

Statin 40 mg (%)

Statin 80 mg (%)

0.1 0.7

0.9 0.9 1.4 1.5 0.6 0.1

2.3 2.1

0.2 0.2 0

2.7 2.3

Abbreviations: ALT, alanine aminotransferase; ULN, upper limit of normal. Source: Reproduced from Ref. 101.

significant hepatotoxicity (e.g., direct hyperbilirubinemia or symptoms) is very uncommon. Asymptomatically elevated aminotransferases is a class effect of statins (94,99,100). The incidence of ALT greater than or equal to three times the ULN in association with different statins is less than 3% and there is a minor dose-related increase in its incidence (Table 2) (94,100). Very rarely, after initiating statins, the liver enzymes can rise to significantly higher levels with no evidence of liver dysfunction (102). In general, the incidence of liver enzyme elevations in the statin treated patients has not been consistently different than in the placebo-treated patients (102,103). Clinically significant liver injury is very uncommon with statins and there is no signature pattern for liver injury associated with statins. All common patterns of liver injury such as hepatocellular, cholestatic or mixed types have been reported (104–107). Isolated case reports suggest that statins may very rarely unmask autoimmune liver disease in genetically susceptible individuals (108–110). Irreversible liver damage leading to death or liver transplantation appears to be extremely uncommon with statins (105,106,111). Based on 232 cases of acute liver injury, potentially associated with lovastatin reported to Merck’s World-wide Adverse Event Database (WAES), it was estimated that risk of fulminant liver failure attributable to lovastatin was two in one million patients (105,106). The mechanism by which statins cause asymptomatic elevation in aminotransferases is unknown. It has been speculated that increased aminotransferases may actually represent a pharmacodynamic effect of lipid lowering, rather than a direct effect of the statins (106). Idiosyncratic and immunoallergic mechanisms have been implicated in the pathogenesis of rare cases of clinically significant liver injury caused by statins (99,104–107). Although statins have been shown to cause various tumors, including hepatic adenomas and carcinomas in rodents (97,112), it does not appear that statins are carcinogenic in humans. A meta-analysis of five large clinical trials consisting of 30,817 patients (15,420 receiving statins) demonstrated no association between statin use over a five-year period and the risk of fatal and nonfatal cancers (113). More recently, the Scandinavian Simvastatin Survival Study (4S) reported that there was no difference in the cancer incidence or mortality between the original simvastatin and placebo groups over a ten-year follow-up (114). There has been recent interest in assessing the safety of statins in patients with underlying liver disease such as NAFLD. This is an important issue because fatty liver is common in hyperlipidemic patients and type 2 diabetics (primary targets for long-term statins) (115–118) and the presence of fatty liver disease may indicate higher cardiovascular risk (119–121). Recent studies have shown that hyperlipidemic patients with unexplained elevation baseline liver enzymes are at not higher risk for statin hepatotoxicity (defined biochemically) those with normal transaminases (102,122,123). A retrospective cohort study conducted on adult members of the Northern California Kaiser Permanente Medical Care Program also suggested that lovastatin is not hepatotoxic to patients with preexisting liver disease (124). In this study, 13,492 patients with lovastatin exposure had no evidence of increased hepatotoxicity when compared to 79,628 patients without lovastatin exposure (124). This study is available only in an abstract form and thus the results should be viewed with caution. It is currently recommended that liver biochemistries be checked before and periodically after starting statin therapy. However, this recommendation is not evidence-based and is controversial (97). Recently, some authors have called for reexamination of these recommendations (101,125). The National Lipid Association has recently conducted a Statin Safety Task

600

Bhardwaj and Chalasani

TABLE 3 Recommendations of the Liver Expert Panel: National Lipid Association Safety Task Force Meeting (Summary) 1. 2. 3. 4. 5.

Asymptomatic elevations in aminotransferases are a class effect of statin and they do not indicate liver dysfunction Liver failure causing death or hospitalization or requiring liver transplantation is very rare from statins Current evidence does not support routine monitoring of liver enzymes and liver biochemistries in patients receiving statins Presence of chronic liver disease and Child’s A cirrhosis should not be considered a contraindication for statin use Current evidence supports use of statins to treat hyperlipidemia in patients with nonalcoholic fatty liver disease and NASH

Abbreviation: NASH, nonalcoholic steatohepatitis. Source: From Ref. 125 for full report of the recommendations.

Force Meeting (Washington DC, 2005) and the recommendations of its Liver Expert Panel listed in Table 3 (125). Niacin Niacin, otherwise known as nicotinic acid, results in modest reductions in total cholesterol, lowdensity lipoprotein cholesterol, triglycerides, and lipoprotein (90). It is more effective in causing an elevation of high-density lipoprotein cholesterol (90). Since certain preparations have become easily available over-the-counter, many patients may potentially be using toxic doses unbeknownst to their physicians. It is generally inadvisable to ingest greater than three grams per day of any formulation and unsupervised use of the over-the-counter SR formulation should be discouraged as niacin may exhibit dose-related toxicity (90,91,126,127). Hepatotoxic effects may appear anywhere between one week to 48 months after the initiation of therapy and usually subside with discontinuation of the drug (92,128–132). Recovery is usually seen in one to two months (126,128,131). There is, however, a potential for fulminant hepatic failure (129,133,134). Combination with statin products does not seem to increase the incidence of adverse effects of either medication (135). Niacin may cause a mixed pattern of hepatocellular and cholestatic injury but is usually associated with an elevation of transaminases (126,136,137). Histologic findings generally include varying patterns of necrosis but may also include centrilobular cholestasis or even steatosis which may simulate the radiographic appearance of hepatobiliary neoplasia (138–140). Clinical clues of nicotinic acid hepatotoxicity include nausea and vomiting in addition to signs and symptoms of acute liver failure (137,141,142). A sharp reduction of serum lipid levels may precede the hepatotoxicity (141,143). When fatty infiltration of the liver occurs, it is common to see a presentation similar to an acute abdomen in addition to a cholestatic pattern of liver disease (143,144). Niacin products are well known to cause hepatotoxicity, especially in doses that exceed 2 to 3 g/day; however, there is considerable difference in the hepatotoxic potential of each formulation (145–148). Niacin is available in three formulations: immediate-release (IR), SR, and extended-release (ER) (149,150). The SR niacin, which is available over-the-counter, is not Food and Drug Administration (FDA) approved for the treatment of dyslipidemia; about half of treated patients develop a symptomatic elevation in transaminases with its use (127,147,149, 151–153). The hepatotoxic profile of niacin is related to the proportion of the drug that undergoes one of two metabolic pathways (Fig. 4) (149,150,154,155). Cutaneous flushing is caused by the conjugation pathway via a prostaglandin-mediated vasodilation (149,150,155). The amidation pathway leads to hepatotoxicity due to nicotinamide and pyrimidine metabolites (149,150,155). The amidation pathway is one of higher affinity than the conjugation pathway but has a lower capacity. Thus, the quickly-released formulation (IR) rapidly overwhelms the amidation pathway and the majority is metabolized using the conjugation pathway. Due to slow release of the SR formulation, most of the drug is able to be processed by amidation (149,154). ER niacin has an intermediate rate of dissolution and can be associated with both flushing and hepatotoxicity (149). Fibrates Much has been published regarding the hepatotoxic and carcinogenic potential of fibrate drugs in rodents. Both clofibrate and gemfibrozil have been found to initiate hepatocarcinogenesis

601

Cardiovascular and Antidiabetic Medications

Nicotinomide and pyrimidine metabolites

Hepatotoxicity

Amidation (high affinity, low capacity)

Nicotinic acid

IR

ER

SR

Conjugation (low affinity, high capacity) Prostaglandin -mediated vasodilation

Cutaneous flushing

FIGURE 4 Metabolic pathways leading to niacin toxicity. Abbreviations: IR, immediate-release niacin; ER, extendedrelease niacin; SR, sustained-release niacin.

in the rat (156–158). However, this does not seem to translate into human cancer risk (159). Clofibrate has also been noted to cause mitochondrial damage and programmed cell death in mouse models. There has been no evidence that the effects seen in rodents are predictive of human toxicity, namely hepatomegaly, hepatocyte apoptosis, and hepatocarcinogenicity. Gemfibrozil has been noted to cause hepatocellular injury and there is a case report of cholestatic hepatitis originating in Spain (159–161). Fenofibrate has been implicated as a cause of autoimmune hepatitis with ductopenia, chronic hepatits, and fibrosis, especially in combination with statin medications (159,162). Ezetimibe Much of what is known about the hepatotoxicity of ezetimibe comes from the manufacturer’s prerelease clinical trials. When evaluated for safety, ezetimibe was generally well tolerated with an incidence of any adverse event similar to that reported with placebo (90). Ezetimibe selectively inhibits the intestinal absorption of cholesterol; it is rapidly absorbed and glucuronidated, yielding an active metabolite (163). Inactivation is primarily by excretion into the bile and there is significant enterohepatic recirculation (163). There does not seem to be any significant hepatotoxicity when used alone, however the frequency of increased transaminases was slightly higher in patients receiving ezetimibe administered with HMG-CoA reductase inhibitors than in patients treated with HMG-CoA reductase inhibitors alone (1,164–168). As our experience with this medication increases, more information may be discovered regarding hepatotoxicity with this medication. Antihypertensive Drugs Hepatotoxicity due to medications used to treat hypertension is relatively rare with the exception of toxicity related to hydralazine or a-methyldopa. Additionally, toxicity may be seen with the use of angiotensin-converting enzyme (ACE) inhibitors and certain b-blockers. Table 4 outlines specific biochemical and histologic patterns of injury as well as the currently understood hepatotoxic potential. Antihypertensives generally exert a class effect in regards to hepatotoxicity with the exceptions noted in the table. Centrally Acting (Central Nervous System) Agents Drugs in this class exert their antihypertensive effect by direct stimulation of a-2 adrenergic receptors, resulting in reduced adrenergic outflow from the central nervous system (CNS) and

Cholestatic injury Hyperbilirubinemia (jaundice) Elevated ATs without clinical LD Hepatocellular injury Hepatocellular injury common but may also cause elevated ATs without clinical LD; possible cholestatic injury; rarely FHF Hepatocellular injury possibly with cholestatic injury; possible FHF

Histologic pattern of hepatotoxicity

Granulomatous hepatitis; necrosis; autoimmune pattern; possible steatohepatitis Granulomatous hepatitis with or without cholestasis; bridging necrosis

Possible cholestasis with or without ductopenia Cholestasis

Possible necrosis Cholestasis Massive necrosis Possible steatohepatitis or mallory bodies Cholestasis and granulomatous hepatitis, steatohepatitis Cholestasis, possible ductopenia

a Specific drugs are listed in cases of unique or additional injury compared with the drug class. Abbreviations: CNS, central nervous system; FHF, fulminant hepatic failure; ACE, angiotensin-converting enzyme; ATs, aminotransferases; LD, liver disease.

Direct vasodilator hydralazine

Diuretics Furosemide Spironolactone a-blockers CNS-acting drug a-methyldopa

Captopril Angiotensin receptor blockers

Cholestatic injury; rarely FHF; also possible elevated ATs without clinical LD Mixed cholestatic and hepatocellular injury Cholestatic and/or hepatocellular injury

ACE inhibitors

Biochemical pattern of hepatic injury Hepatocellular injury Cholestatic injury Hepatocellular injury Hepatocellular injury Cholestatic injury

b-blockers Atenolol Combined a/b-blockers (labetalol) Calcium-channel blockers Diltiazem

Antihypertensive medication or classa

TABLE 4 Hepatotoxicity of Various Antihypertensive Medications Hepatotoxicity potential

High

Probably low, but may cause chronic LD or cirrhosis High Unknown but thought to be low Low Low Low Low High; may lead to chronic LD or cirrhosis

Low Low High Low Low

602 Bhardwaj and Chalasani

Cardiovascular and Antidiabetic Medications

603

an overall reduced total peripheral resistance. a-Methyldopa is a prodrug that is metabolized into a-methyl-norepinephrine. This active metabolite is stored in neurosecretory vesicle in place of norepinephrine and, when released, acts as a potent central adrenergic receptor agonist. Methyldopa

Hepatotoxicity related to a-methyldopa use is common (almost 5% of patients) (169). a-Methyldopa may cause hepatotoxicity with a striking variability in latency period which can be several weeks (which is the majority of cases) to several years but usually occurs within the first six months (170–174). Hepatotoxicity is generally asymptomatic but, when symptomatic, almost invariably causes upper abdominal pain with nausea and vomiting, anorexia, malaise, and often, fever (175–177). Thus, the symptoms can be very difficult to distinguish from acute viral hepatitis (169). Clinical signs of methyldopa toxicity follow these symptoms after several days to weeks, are more uncommon than the biochemical changes, and include jaundice and tender hepatomegaly (171,175,178). Patients who develop hepatotoxicity after long-term use of the drug often have milder disease but with a more prolonged course (179). Laboratory assessment most often reveals a hepatocellular pattern of liver injury with most cases showing a moderate or severe acute hepatitis but patients may also present with cholestatic disease in the form of a chronic active hepatitis (180–182). This latter form of hepatotoxicity is more commonly seen in patients after long-term use of the drug or in those with underlying liver disease (179,180). The mechanism of toxicity appears to be autoimmune in nature as evidenced by the presence of autoimmune markers and the often present plasma cells and eosinophils on histologic examination (183,184). Other histopathologic findings on examination of the biopsy specimen include a continuum of changes ranging from fatty change to focal hepatocellular necrosis to massive hepatic necrosis to steatosis and, rarely, hepatic granulomas (185–187). Fatty change and granulomatous disease tends to be more common in patients who develop methyldopa-related hepatitis after long-term use of the drug (179,186). The inflammation, necrosis, and steatosis most commonly occur in a lobular pattern although hepatits after shortterm use of the drug has been associated with massive periportal necrosis (188). Post-necrotic cirrhosis has been demonstrated in case reports and case series (189). a-Methyldopa is often used in the pregnant patient as it is one of the few agents felt to be safe for the fetus (190). Diagnosis of drug-induced liver injury in this setting may be difficult owing to the multitude of disorders that may cause hepatitis in the pregnant patient. The pattern of injury in the few available published reports differs from drug toxicity in the nonpregnant patient (190,191). Such patients were documented to have rash and eosinophilia, which are conspicuously absent in the nonpregnant patient (192,193). Unless the patient presents in fulminant hepatic failure (which is rare but possible), clinical improvement is usually rapid after withdrawal of the drug, i.e., seen within 48 hours, although resolution may take months in patients with toxicity after long-term use of the drug (185,189,194). Less commonly, a-methyldopa has been shown to cause a chronic autoimmunetype hepatitis despite discontinuation of the drug (184,188,195). Conversely, in some rare cases, the transaminase elevations resolve despite continued use of the medication (196). Reexposure to the medication may lead to a more rapid and severe recurrence of the hepatitis (189,197). A small cohort of patients had histologic follow-up after discontinuation of the medication (196). Patients with prior histologic evidence of fatty infiltration showed a decrease in the severity of disease whereas patients with submassive necrosis progressed to cirrhosis (189). Clonidine

Clonidine does not seem to cause hepatotoxicity to any significant degree according to our present understanding of this medication (198,199). Direct Vasodilators Hydralazine

Hydralazine leads to decreased blood pressure by its effect on the peripheral vasculature; it causes relaxation within the smooth muscle of arterioles, thereby causing reduced systemic

604

Bhardwaj and Chalasani

vascular resistance. Hepatic injury may be seen within a few months or after several years of therapy with hydralazine (200,201). Fever and eosinophilia may be seen prior to the development of hepatic injury (202). Severe hepatitis or even acute liver failure with a prolonged prothrombin time and hepatic encephalopathy are not uncommon (203). A mainly hepatocellular insult is seen with marked elevation in the transaminases usually coupled with an elevation in bilirubin levels (201,204). Hydralazine has been shown to cause centrilobular necrosis to varying degrees upon histologic examination; bridging necrosis is almost pathognomonic of the diagnosis in the setting of drug exposure (205,206). There are also some case reports documenting cholestasis or granulomatous disease on biopsy (207–209). There is one case report of acute cholangitis attributed to hydrazaline reexposure with clinical findings of fever and right upper quadrant pain (210). The mechanism of injury is likely immunologic although it is still debated that the mechanism may be from direct injury of a hepatotoxic metabolite (207,211). Peripheral eosinophilia with rash and fever suggests allergic-type drug hypersensitivity as the mechanism of injury (202). Discontinuation of the drug has led to clinical, biochemical, and histologic recovery, however, some cases of fibrotic sequelae have been reported (212,213). Recovery, when it occurs, is usually complete within six months of removal of the agent from the patient’s medication regimen (214,215). When hepatotoxicity occurs after days to a few months of therapy, the hepatitis tends to be more severe but often resolves without sequelae, whereas toxicity occurring several months to years after initiation of therapy tends to cause a milder disease with a rare possibility of fibrosis (213,216–218). ACE Inhibitors ACE inhibitors are competitive inhibitors of angiotensin-converting enzyme, mimicking the structure of its substrate. ACE inhibitors directly block the formation of angiotensin II and also increase bradykinin level which results in reduced vasoconstriction, reduced sodium and water retention, and increased vasodilation (through the action of bradykinin). Captopril and lisinopril are active molecules but the others are prodrugs that need to be converted to active metabolites (219,220). ACE inhibitors exert a class effect in regard to drug-induced liver injury, which is usually cholestatic in nature but hepatocellular injury has been described (219). Rarely, acute liver failure is seen (221). Toxicity is usually seen in three to four months but can vary from several days to almost a year after the initiation of therapy (221). Drug-induced liver injury has been noted even after several years and can be triggered by a dose escalation (219). Jaundice is very common and may be associated with pruritis, nausea, and right upper quadrant pain (222). There seems to be cross-reactivity of the various agents with respect to hepatotoxicity. Several reports have documented adverse reaction to two or more different ACE inhibitors (223). Other reports have shown the safe use of another agent (namely enalapril) after the resolution of liver injury from another agent (223,224). This absence of cross-reactivity may be attributable to the lack of proline in the molecular structure of the ACE inhibitors that do not exhibit hepatotoxicity when used after ACE inhibitors that do (224). There is no evidence of crossreactivity with angiotensin receptor blockers (ARBs). The toxicity is believed to be produced by a common mechanism, perhaps from an eicosanoid-mediated, inflammatory mechanism (224). There also may be an allergic type of hepatotoxicity in patients who develop fever and rash (219). These two separate mechanisms of toxicity may also explain cross-reactivity in some case reports and the lack of cross-reactivity in others (219). The liver injury usually resolves with the discontinuation of the ACE inhibitors but can progress to acute liver failure or chronic liver disease with cirrhosis should the offending agent be continued (219). It is not uncommon for the alkaline phosphatase to remain elevated for up to 18 months (219). Representative medications with associated patterns of toxicity are described below. Captopril

Captopril produces mainly cholestatic liver injury usually with jaundice but may sometimes produce a mixed pattern of liver injury with a (less significant and usually transient) elevation in transaminases as well (225–228). This hepatic injury is rarely seen and, when it occurs,

Cardiovascular and Antidiabetic Medications

605

is generally not severe and also quickly reversible (225). The jaundice, however, may last several weeks to months before resolution (227). There is no report of definitive chronic disease due to captopril as opposed to the other ACE inhibitors. Most reported cases have occurred within one to two months of use, although a longer duration of treatment prior to the development of elevated liver enzymes does not eliminate the possibility of drug-induced liver injury (225,229). Captopril-related hepatotoxicity may be confused with obstructive biliary disease or even cholangitis as symptoms of acholic stools, pruritis, right upper quadrant pain, or fever may occur (229–231). Liver biopsy reveals extensive cholestasis, often with portal inflammation and rarely with hepatic necrosis (231,232). It has been suggested that a gap between the bilirubin and lactate dehydrogenase lactate dehydrogenase (LDH) (with a high bilirubin and almost normal LDH) is another diagnostic clue that a patient’s liver dysfunction is caused by captopril (225). In addition, imaging studies should not show any biliary abnormalities (230). Enalapril

The development of jaundice is common in the setting of enalapril-induced hepatotoxicity, which can occur anytime between several days to several years after initiation of the medication (233–236). Imaging studies do not reveal any evidence of extrahepatic cholestasis (234). Progressive liver failure can be seen despite withdrawal of the drug if hepatotoxicity appears after long-term use (i.e., years) (233,237,238). Ductopenia has been reported after longterm use as well (239). Other ACE Inhibitors

Both acute hepatic failure and chronic hepatitis have been reported with the use of lisinopril (240). As with the other ACE inhibitors, continuation of lisinopril after the occurrence of jaundice may lead to fulminant disease with the development of hepatic encephalopathy and coagulopathy (241). Fosinopril-related hepatotoxicity is extremely rare and gives a clinical picture similar to enalapril, with possible biliary ductular damage (242). Ramipril-associated liver injury is similar to the patterns of injury described above (243). There is a case report of duct necrosis and biliary fibrosis in a patient with prolonged cholestasis (243). ARBs ARBs competitively inhibit angiotensin II and lead to reduced vasoconstriction and reduced sodium and water retention. Hepatotoxicity related to ARBs is very uncommon and has been associated with tender hepatomegaly and jaundice (244,245). The agents currently in use have an extremely rare incidence of liver toxicity; tasosartan was believed to have a higher incidence of hepatotoxicity but was withdrawn by the pharmaceutical company prior to release onto the market (246). Losartan, irbesartan, and candesartan are the most frequently mentioned ARBs with hepatotoxicity in the current literature. Centrilobular (perivenular) cholestasis is the hallmark of ARB-related hepatotoxicity upon histologic examination (246). Losartan has been noted to cause jaundice and tender hepatomegaly with a mixed pattern of liver injury (transaminases in the several hundreds to thousands) (244,245,247). All known cases have shown resolution of the hepatic injury after withdrawal of the drug (244,245). Candesartan use has been associated with several patterns of liver toxicity: transient elevations in transaminases, hepatocellular injury, acute cholestatic hepatitis, and bile ductopenia have all been seen (248–250). Jaundice and hepatomegaly are noted upon physical examination (249,250). Nausea, vomiting, and right upper quadrant pain have been described (249). Irbesartan has been noted to cause cholestatic hepatitis with a protracted course that resolves slowly after drug withdrawal, taking up to a year for the resolution of biochemical evidence of liver injury (246,251). As opposed to candesartan, noted above, there has not been any evidence of ductopenia or bile duct damage in cases of irbesartan-related liver injury despite the prolonged nature of recovery from toxicity (246). The reported cases generally presented with right upper quadrant pain, jaundice, and anorexia (252). Many of the published reports describe a mixed pattern of injury upon biochemical assessment and hepatocellular disease upon examination of the biopsy (246). There has been one recent case of an autoimmune reaction requiring steroid therapy presumably provoked by irbesartan

606

Bhardwaj and Chalasani

(253). Valsartan hepatotoxicity has been noted in two cases (254,255). Cholestatic liver injury was seen in addition to right upper quadrant pain, nausea, and jaundice (254). Combined a1/b Antagonist Labetalol had been known to cause asymptomatic elevation of the transaminases before its commercial release in about 8% of cases (256). However, it has since been shown to cause clinically significant hepatotoxic damage with hepatocellular necrosis as well (256–258). The duration of treatment prior to the development of liver disease is generally several weeks to several months but may be as long as four years, as noted in one case report (256,257). This time course may point to a metabolic idiosyncracy, i.e., the development of toxic metabolites as the pathogenesis of labetalol hepatotoxicity (256,259). A hypersensitivity-type reaction is unlikely given the absence of fever, rash, or peripheral eosinophila (although there may be some eosinophilia upon biopsy analysis) (260,261). Histologic examination reveals submassive or massive necrosis but zone 1 necrosis may also be seen, especially with an eosinophilic infiltrate (260,262). Clinical and biochemical recovery is seen quickly after discontinuation of the drug with the exception of cases that had progressed to fulminant failure prior to drug withdrawal, in which case progression to death was not uncommon (218,262). b-Blockers b-blockers have been shown to cause rare instances of hepatocellular damage with possible necrosis (263). The incidence of drug-induced liver injury in these cases is extremely low and generally reversible with discontinuation of the drug (263). The exception to this low incidence and specific pattern of disease is atenolol, which may cause cholestatic disease in a higher than anticipated number of patients (but still very rare) compared to other medications in the drug class (264). Liver disease with b-blockers is usually asymptomatic but may include jaundice with pruritis, anorexia, right upper quadrant pain, or a nonspecific flu-like syndrome (263). There has been evident cross-reactivity with carvedilol followed by metoprolol (265). Atenolol may, infrequently, lead to cholestatic injury with or without a mixed hepatocellular picture (264). Carvedilol has rarely caused hepatocellular injury; most instances consist of an asymptomatic elevation of transaminases (265). The liver injury is reversible following withdrawal of this medication (265). Acute hepatocellular injury has been seen with propranolol, metoprolol, and acebutolol (212,263,266–268). Calcium-Channel Blockers Calcium-channel blockers, especially the dihydropyridines, have been implicated as the cause of drug-related liver damage in a small number of patients (269,270). A class or subclass effect may exist as reversible liver injury has been documented after the use of amlodipine in a patient with known previous nifedipine hepatotoxicity (271). Recovery is seen within several weeks of drug withdrawal in most cases (269). Examination may reveal jaundice and hepatomegaly (272). A mixed pattern of hepatocellular and cholestatic liver injury may be seen, although hepatocellular injury alone is more common (270,272). Lesions resembling mallory bodies, in the background of steatosis, have been noted with this subclass of drugs (273). Cholestasis has been documented with the dihydropyridines and the nondihydropyridine calcium-channel antagonists (270,274). Diltiazem has been associated with granulomatous hepatitis (275,276). Steatohepatitis with mallory bodies has been seen on biopsy in many of these agents as well, mimicking disease seen with alcohol or fatty infiltration of the liver (273). However, NASH may coincidentally occur in these patients so the association with fatty liver disease may be overestimated (270). Cholestasis with jaundice has been described as has hepatocellular disease (277,278). Several reports seem to indicate drug hypersensitivity as the mechanism of liver toxicity due to the presence of peripheral eosinophilia, fever, nausea, and liver tenderness in some patients (279). Hepatic injury with verapamil has been demonstrated in very few patients so the general pattern and mechanism of injury remain speculative, although a hypersensivity reaction has been proposed (280,281). The cases that have been documented thus far reveal a mainly hepatocellular pattern of injury (280,282–284).

Cardiovascular and Antidiabetic Medications

607

Diuretics Tienilic Acid

Tienilic acid is no longer available commercially as it was withdrawn from the market due to its hepatotoxicity. The hepatotoxic effects are believed to be due to the actions of the metabolites on liver proteins: it produces an autoimmune-type reaction with the development of liver-kidney microsomal antibodies (anti-LKM) (285–291). Furosemide

Hepatotoxicity from furosemide is extremely rare and when it occurs, it appears to be hepatocellular in presentation. Hepatotoxicity from furosemide is induced by reactive intermediates (292,293). Several studies in rodent models have attempted to document the mechanism of injury (294–296). However, an adequate explanation has not yet been described. The theory that a depletion of cellular reduced glutathione or mitochondrial total glutathione may contribute to a problem with the respiratory chain has been debunked by these studies (297). Centrilobular congestion with necrosis has been seen on pathologic examination (296,298,299). Sloughing of sinusoidal cells with ribosomal disaggregation and vesiculation of the endoplasmic reticulum has been noted upon electron microscopy (299,300). Hydrochlorothiazide

Hydrochlorothiazide has been noted to cause mixed hepatocellular and cholestatic injury that reverses upon drug withdrawal and there have been rare instances of cholestasis with the other thiazide diurectics (301). There is a report of fulminant hepatic failure with a combination of hydrochlorothiazide and benazepril (302). Spironolactone

Spironolactone-related hepatotoxicity is an extremely rare cause of hepatocellular injury (303). It is associated with a histologic examination showing foci of lobular necrosis and pervenular cholestasis and inflammation (303,304). This may be caused by interference with biliary secretion as is the case with certain steroids (303,305). The lab abnormalities are both hepatocellular and cholestatic in nature and resolve in about one to two months with discontinuation of the drug (305). Endothelin Receptor Antagonists Bosentan blocks the action of endothelin, an endogenous substance that has been linked to pulmonary hypertension. It is hepatically metabolized to three compounds, one of which is also an active metabolite (306). Bosentan-induced liver injury is reported to occur in 11% of patients per the package insert; this assertion has been confirmed by a review of the clinical safety database (307,308). Most of these cases (O90%) occurred within the first six months of initiation of the drug and, in all cases, the elevated hepatic transaminase levels resolved spontaneously after treatment discontinuation or dose reduction with no liver failure or ongoing liver injury reported in any patient (307). The transaminase elevations usually occurred over the course of several weeks (306–308). In almost all cases, the elevation in liver tests did not correlate with clinically significant disease, however less than 2% of patients had symptoms of fever, nausea, vomiting, and abdominal pain, and there has been one reported case of jaundice (307). A review of eight placebo-controlled studies found that the administration of bosentan was associated with a dose-related elevation in hepatic transaminase levels of greater than three times the ULN in 74 (11.2%) of 658 patients versus five (1.8%) of 280 patients receiving placebo (307). Higher doses of bosentan were associated with a higher rate of abnormally elevated hepatic enzyme levels and did not result in a greater number of patients with increased clinical symptoms (307–309). In vitro studies have shown that bosentan competitively inhibits the bile salt export pump (BSEP) but it is not clear if this mechanism has any significant mechanistic relationship with bosentan-related liver injury (310). Federal Drug Administration guidelines mandate a baseline and then monthly monitoring of hepatic enzymes upon initiation of the drug (Table 5) (307).

608

Bhardwaj and Chalasani

TABLE 5 Dosage Adjustment and Monitoring of Patients Developing Aminotransferase Abnormalities on Bosentan Therapy ALT/AST Levels O3 and %5 times the ULN

O5 and %8 times the ULN

O8 times the ULN

Recommendations Confirm by another aminotransferase test. If confirmed, reduce the daily dose or interrupt treatment and monitor aminotransferase levels at least every two weeks. If the aminotransferase levels return to pretreatment values, continue or reintroduce the treatment as appropriated.a Confirm by another aminotransferase test. If confirmed, reduce the daily dose or interrupt treatment and monitor aminotransferase levels at least every two weeks. If the aminotransferase levels return to pretreatment values, consider reintroduction of the treatment as appropriate.a Treatment should be stopped and reintroduction of bosentan should not be considered. There is no experience with reintroduction of bosentan in these circumstances.

a

If reintroduced, initiate bosentan therapy at the starting dose. Check aminotransferase levels within three days and at least every two weeks therafter. If liver aminotransferase elevations are accompanied by clinical symptoms of liver injury (such as nausea, vomiting, fever, abdominal pain, jaundice, or unusual lethargy or fatigue) or increases in bilirubin of greater than or equal to two times the ULN, treatment should be stopped. There is no experience with reintroduction of bosentan in these circumstances. Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; ULN, upper limit of normal. Source: From Ref. 307.

Antianginal Medications The organic nitrates (e.g., nitroglycerin, isosorbide dinitrate, and amyl nitrate) have been used for many years without definitive evidence of hepatotoxicity related to their ingestion. There has been evidence that rats can develop liver tumors after exposure to nitrates, however there is no evidence that this occurs in human subjects (311,312). Calcium-channel blockers and b-blockers are discussed in detail in the antihypertensive medication section. Anticoagulants Warfarin and Coumarin-Derived Agents Warfarin and the other coumarin derivatives prevent and treat thrombotic disorders by inhibiting the synthesis of vitamin K-dependent clotting factors via inhibition of vitamin K epoxide reductase. Liver injury is rare and is felt to be due to an epoxide intermediate, leading to an allergic-type hypersensitivity reaction (313–317). Hepatic injury, when it occurs, is predominantly of the cholestatic type and, when severe, may lead to fulminant hepatic failure (318–320). Hepatocellular disease has also been shown to occur (321). Histologically, coumarin-induced liver disease may vary from mild nonspecific changes to severe disease that consists of portal and lobular hepatitis with possible bridging necrosis (which may be confluent and massive) (322,323). Zone 3 necrosis appears to be particularly common (323). Almost every patient with coumarin-related liver injury exhibits portal and periportal hepatitis and most also have evidence of eosinophils in the areas of inflammation (324). Usually these changes correlate with the severity of biochemical injury seen in each individual patient (325). A latency period of three months at the minimum has been seen but clinical and biochemical evidence of disease can also take greater than five years to manifest (319,324). Reexposure to warfarin generally results in a much shorter latency period (usually one week to two months) but there have been reports of successful reintroduction of the same or different coumarin agent after very mild cases of disease (324). Heparins Unfractionated Heparin

Heparin exerts its anticoagulant action by binding to antithrombin III and inducing a conformational change that yields a more active compound. The heparin-antithrombin complex inactivates a number of coagulation enzymes, including thrombin (factor IIa), as well as factors Xa, IXa, XIa, and XIIa. Clinically significant liver injury from heparin is extremely rare (only a few cases in the literature) but laboratory abnormalities are frequently

Cardiovascular and Antidiabetic Medications

609

reported. Hepatic laboratory abnormalities greater than three times the ULN have been reported in 5% of patients taking the drug (326). Heparin may cause mild to moderate increases in transaminase levels and alkaline phosphatase; bilirubin elevation and subsequent jaundice is not seen (327–329). Enzyme abnormalities develop after 10 days to one month and return to normal with discontinuation of the drug within four months (327). Low Molecular Weight Heparins

Low molecular weight heparins, such as enoxaparin, ardeparin, dalteparin, and tinzaparin, have been associated with elevated aminotransferases at a rate of approximately 5% (330,331). Enoxaparin has been the most frequently studied of these medications, with a reported incidence of hepatotoxicity between 2% and 9% (308,326,331–333). The similar rates of hepatotoxicity between unfractionated and low-molecular weight heparin suggest a class effect; therefore, it is recommended to refrain from all heparin products when an adverse reaction is seen in one heparin product (326,332). The pattern of injury seen is mostly hepatocellular, although an elevation in alkaline phosphatase may sometimes be seen, producing a mixed pattern of liver injury (330). The degree of liver injury is usually, but not universally, related to the dose of the medication (331). In most studies, liver injury occurred with the first week to 10 days after the initiation of therapy (331). Recovery after withdrawal of the medication is prolonged; return of the transaminases to baseline may take several months (326,330). Return of the transaminases to normal levels is almost universal upon discontinuation of enoxaparin (331,334). Hepatic function studies generally do not return to normal unless the drug has been discontinued (331). Liver histologic changes are rare, but when abnormal, typically show ballooning degeneration, minimal hepatic necrosis and a scattered mononuclear cell infiltrate (326). The pathogenesis of the liver injury remains unknown although various mechanisms of toxicity have been hypothesized. Studies have been published supporting theories of altered hepatocyte membranes, direct toxic injury, and immune-mediated mechanisms (308,335). Antithrombin Agents The antithrombin agents act by directly inhibiting thrombin, including thrombin which is bound to clot (336). These drugs are analogs of hirudin, derived from the medicinal leech, and include medications such as lepirudin, bivalirudin, and argatroban (the only three medications in the drug class approved for use in the United States) (99). These medications, in addition to melagatran (another hirudin analogue which has a slightly different mechanism of action, binding reversibly only to the active site of thrombin) have been developed only for short-term clinical use and must be administered parenterally (336,337). Ximelagatran is an oral medication and a prodrug of melagatran that is absorbed in the small intestine (99). There have been several reports of probable hepatotoxic injury due to ximelagatran and, therefore, the US FDA withheld its approval in the United States (and the manufacturer has removed the drug from the world market). Ximelagatran was associated with an asymptomatic but significant increase of the ALT (O3 times the ULN) in 6.0% to 9.6% of patients in the initial clinical trials using various indications, mostly comparing efficacy with heparin, enoxaparin, or warfarin (338–340). The increase usually occurred within six months after the initiation of therapy (with a mean of 2.5 months). Although most patients had the medication discontinued in the setting of each particular trial, the ALT levels decreased whether or not the medication was stopped (336,338–340). A randomized trial investigating stroke prevention in patients with atrial fibrillation had one patient who died after the development of biopsy-proven hepatic necrosis and another with marked elevation of the ALT to greater than 11 times the ULN (340). According to FDA documentation, there is no evidence that short-term use of ximelagatran (e.g., orthopedic procedures) is associated with hepatotoxicity; however, the length of follow up for these patients has only been up to six weeks (341). The concern during the clinical trials has involved a high incidence of liver enzyme elevation observed in patients undergoing prolonged therapy. However, in data presented to the FDA on all patients receiving long-term ximelagatran (defined as greater than 35 days of therapy), an increase in ALT greater than three times normal occurred in 7.9% of patients compared

610

Bhardwaj and Chalasani

FIGURE 5 Cumulative risk of ALT greater than three times ULN (%) versus time after randomization into ximelagatran and comparator arms. Abbreviations: ALT, alanine aminotransferase; ULN, upper limit of normal. Source: From Ref. 341.

with 1.2% of patients receiving comparator therapy (Fig. 5) (341). Using the definition of severe hepatotoxicity as total bilirubin greater than two times the ULN, this level of liver injury was observed in 0.53% versus 0.08% taking the comparator therapy (341). Of the nine deaths noted in these clinical trials, three were deemed possibly related to ximelagatran (341,342). The mechanism of hepatotoxicity has not yet been established. On the basis of these data, the advisory committee to the FDA recommended against the approval of ximelagatran. In fact, this medication has been taken off the market worldwide. Antiplatelet Agents Drugs used in the treatment of thrombosis and disorders related to thrombosis include those with antiplatelet actions. These medications inhibit ADP-induced platelet activation. Both medication described below are metabolized in the liver by the cytochrome P450 system to active and inactive metabolites. Ticlopidine Ticlopidine has been associated with asymptomatic abnormal liver function studies in up to 5% of patients (343–346). The pattern of liver disease appears to involve mostly hepatocellular damage early in the course of injury with a mixed pattern of damage later in the course of injury (343,347). The severity and pattern of liver injury may vary. The alkaline phosphatase level can increase 5 to 10 times while the transaminases may be normal or increased up to the several thousands (343,348–350). Jaundice with or without abdominal pain is common upon physical assessment of the patient (343,351,352). Histologically, ticlopidine may cause a mostly nonspecific pattern of cholestatic drug-induced liver injury consisting of mostly macrovesicular steatosis and bile ductular damage (345,352,353). However, a portal infiltrate of mononuclear and neutrophilic cells may also be seen (presumably early in the course of disease) (352). There has been evidence of several mechanisms of liver injury, including drug hypersensitivity, immune-mediated injury, and toxicity related to one of the metabolites, although the exact mechanism remains elusive (345,354). The latency period prior to demonstrable injury has been shown to range from one week to one and a half months (345,355). Evidence of drug toxicity usually resolves within one to three months after discontinuation of the drug (345). As in other cases of cholestatic liver injury, the cholestasis may show a prolonged recovery phase lasting upwards of several months to a year (345,356). Importantly, there have been no cases of fatal outcome or irreversible liver injury

Cardiovascular and Antidiabetic Medications

611

when the drug was discontinued after recognition of hepatic damage (352). Ursodeoxycholic acid may have a beneficial effect on ticlopidine-induced liver damage, but, as recovery is almost universal, the utility of adding this agent is questionable (355). Clopidogrel Clopidogrel may cause a pattern of liver injury that is similar to that of ticlopidine described above (357). Both hepatocellular and mixed liver injury have been seen just as in ticlopidinerelated disease (358). The incidence is much rarer, occurring in one in several thousand patients on the drug (357). The latency and recovery also tend to be similar in both drugs (359). There does not seem to be any cross-sensitivity between ticlopidine and clopidogrel and, in fact, there has been a case of successful clopidogrel use after an incidence of ticlopidine-induced liver injury (360). Glycoprotein IIb/IIIa Receptor Blocker Antagonists of the glycoprotein IIb/IIIa receptor include abciximab, eptifibatide, and tirofiban. There is no current definitive evidence that these agents lead to drug-induced liver injury. If there is any hepatotoxicity with these agents it is likely a rare occurrence as evidenced by the lack of documentation of liver injury despite several years of use in virtually all patients undergoing cardiac catheterization. Thombolytics Thrombolytics are a rare cause of significant liver disease. There are many reports detailing subcapsular hemorrhage but few reports detailing actual hepatic parenchymal disease. Tissue plasminogen activator (TPA) and urokinase have not been noted to cause liver disease. Streptokinase, however, has been reported to cause an asymptomatic elevation in transaminases (361). ANTIDIABETIC DRUGS Insulins Insulin preparations carry a low incidence of hepatotoxicity (362). Insulin, as an endogenous compound, is generally not considered to be hepatotoxic. It has been proposed that minor contaminants in insulin preparations are responsible for the hepatic injury seen with the administration of this drug; there are, however, conflicting reports of improvement with changes in insulin preparation (362,363). Usually improvement, when seen, occurs when human insulin is replaced with porcine insulin (362). Some studies do not show this improvement with change of insulin preparation (362,364). Therefore, exogenous human insulin itself has been theorized to cause the liver injury (362). The biochemical pattern of liver toxicity is dominated by an elevation in g-glutamyl transpeptidase, often greater than 30 times the ULN (1,164,362). This is usually accompanied by an elevation of the transaminases (up to 45 times the ULN). This effect has been shown to be reversible upon drug withdrawal and recurs after rechallenge (362). Biopsies, when performed, have generally been nonspecific (362). Although fatty liver has been seen upon imaging of the liver and subcapsular steatosis has been reported in peritoneal dialysis patients, this may be coincidental as diabetic patients have a tendency to develop fatty liver disease (365). It has been suggested that the fatty liver can cause or augment the noted liver response to insulins (1,366,367). Sulfonylureas These agents stimulate the pancreatic b cells to secrete insulin. With the development of newer second- and third-generation agents, the sulfonylurea class of oral hypoglycemics has become safer and more efficacious. Since the prototype agents that may cause severe liver injury (e.g., carbutamide and metahexamide) are no longer in use, the risk of hepatotoxicity has fallen significantly (366–368). However, there remains a rare possibility of serious liver injury with

612

Bhardwaj and Chalasani

the existing agents, especially chlorpropamide, owing, perhaps, to similarities in structure with carbutamide and metahexamide (369). The first-generation sulfonylureas (acetohexamide, tolbutamide, tolazamide, and chlorpropamide) have a greater tendency to cause hypoglycemia and, therefore, their use has decreased (1,366–368,370). All sulfonylureas are capable of causing mixed liver injury with prominent cholestasis probably due to drug hypersensitivity (perhaps to the sulfa moiety) although a primarily hepatocellular reaction is seen with the older sulfonylurea compounds (1,371). Hepatic injury usually occurs within the first month of therapy but has been known to occur several months later as well (1,372,373). These findings tend to be reversible and subside with discontinuation of the offending agent, although irreversible cholestasis has been reported when chlorpropamide is ingested in combination with erythromycin (366,374). Additionally, tolazamide has been noted to cause chronic liver injury (1,375). Fulminant hepatic failure is very rare but has been documented (376). Cross-reactivity between agents has been shown in some case studies, but sulfonylurea therapy with glyburide has been successfully initiated in patients with idiosyncratic reactions to the first-generation agents, chlorpropamide and tolbutamide (377–383). However, insufficient evidence exists whether second- and third-generation agents are a safe alternative when first-generation agents lead to hepatotoxicity. Chlorpropamide A mixed pattern of hepatocellular and cholestatic injury is seen with chlorpropamide toxicity (379,381–383). This usually resolves spontaneously after discontinuation of the medication (379, 383–385). Granulomatous hepatitis with associated fever and exfoliative dermatitis (indicating a hypersensitivity reaction) is seen in some cases (372,380). When used with erythromycin, a profound cholestasis with disappearance of the interlobular bile ducts can be noted which has a tendency to be irreversible (386–388). Glyburide (Glibenclamide) Glyburide has been noted to cause a mixed pattern of liver injury likely due to drug hypersensitivity, although modest hepatotocellular injury with increased aminotransferases has been reported within a couple of months of initiation of the drug (387,389). It has also been theorized that the cholestasis generally seen with glyburide toxicity may come from inhibition of the BSEP (390,391). Biopsy usually reveals hepatic necrosis and mild centrilobular cholestasis, although granulomatous hepatitis has been seen (386,388). Normalization of lab parameters is generally seen with discontinuation of the drug (374,392). Tolazamide Liver injury with tolazamide causes a different form of hepatic injury compared with most other sulfonylureas. While most cases resolve with a normalization of liver function tests after discontinuation of the drug, chronic liver injury is quite common and can occur in up to onethird of patients with a hepatotoxic reaction to tolazamide (392). A cholestatic pattern of injury is seen, with a characteristic elevation in alkaline phosphatase which may persist several months after discontinuation of the medication (374). Liver biopsy reveals portal injury with bile duct proliferation and fibrosis (369,393,394). Tolbutamide Tolbutamide hepatotoxicity is evidenced by the development of cholestatic hepatitis with a possible eosinophilic cutaneous eruption, similar to chlorpropamide (394). Discontinuation of therapy is associated with complete reversal of injury (395–397). Glimepiride Patients on glimepiride have developed cholestatic liver injury within several weeks of initiation of the drug (395). Like most sulfonylureas, zone 3 necrosis and cholestatic changes are seen on liver biopsy (398,399).

Cardiovascular and Antidiabetic Medications

613

Gliclazide Gliclazide may cause hepatotoxicity within several weeks of initiation of therapy (398,400). Like other sulfonylureas, a centrilobular necroinflammatory reaction is seen upon histologic examination (399,400). Immediate discontinuation results in complete recovery in a matter of weeks (373). Acetohexamide Acetohexamide can cause cholestatic injury as well as eosinophilic hepatitis (373). An acute hepatitis with fever and jaundice may develop even after prolonged use of the medication. Hepatic function studies return to normal quickly after withdrawal of the drug (1,366,401–404). Thiazoladinediones Thiazolidinediones act primarily through a PPARg receptor-mediated signal enhancing insulin responsiveness in muscle or adipose tissues, or decreasing insulin resistance by increasing insulin sensitivity of insulin-dependent tissues and inhibiting hepatic gluconeogenesis (405–407). These drugs are widely used to treat type 2 diabetes, but they are not effective in type 1 diabetics who are insulinopenic. When administered, experimental animals, including mice and rats but not nonhuman primates, develop cardiac hypertrophy, interscapular brown fat hyperplasia, and endothelial cell proliferation (408–411). The cardiac enlargement is prevented by the coadministration of ACE inhibitors, indicating a role for the renin-angiotensin system (412). Extensive animal studies did not reveal abnormalities in safety evaluation models (408–410,413–415). The adipose tissue of rodents undergoes significant metabolic or morphological changes with glitazones (416,417), while exerting significant effects on adipocyte differentiation (418,419). The significance of these findings is not clear over the long-term in diabetics (420). Troglitazone (TGZ) TGZ was withdrawn from the market by the US FDA in March 2000 because it was associated with significant hepatotoxicity and acute liver failure. This is an oral antihyperglycemic agent that normalizes blood glucose by increasing target cell responses to insulin. This compound was discovered by Japanese researchers in the 1980s and is the same group that developed the first HMG-CoA inhibitor (421). TGZ reduces hepatic glucose output and enhances insulindependent glucose disposition in skeletal muscle. Its mechanism of action is related to binding to peroxisome-proliferator-activated receptor in the nucleus that regulates the transcription of a number of insulin-responsive genes required for the control of glucose and lipid metabolism. The TGZ molecule contains two chiral centers and each of the four stereoisomers shows similar pharmacological effects; the tocopherol group appears to confer antioxidant activity (422). In clinical trials consisting 2510 patients, 1.9% of patients receiving TGZ had ALT elevation three times the ULN compared to 0.6% of patients who received placebo (423). In this cohort of 2510 patients, TGZ induced overt liver injury and jaundice occurred in three patients (424,425). Significant liver injury associated with TGZ is unpredictable and is neither time nor dose dependent, and cannot be reproduced in animals (metabolic idiosyncracy). Although rare instances of pure cholestatic injury or mixed hepatocellular/cholestatic injury have reported, the signature pattern of significant liver injury from TGZ was hepatocellular and acute in nature with high peak serum aminotransferases occurring weeks or months after its initiation (426). Elevation of serum bilirubin was delayed relative to the aminotransferase elevations, with jaundice typically appearing weeks after the onset of liver injury (426). A striking feature of this “signature” was that after stopping TGZ, the serum aminotransferases generally continued to rise for many days or weeks (426). Resolution of liver injury was prolonged, often taking weeks or months. When available in the United States, 1.92 million patients received TGZ and 94 of them were reported to have liver failure (89 acute and 5 chronic) (427). In a detailed case study, Graham and colleagues have described the characteristics and outcome of 89 patients with TGZ associated acute liver failure (Table 6) (427). In a subsequent pharmacoepidemiological study of data from a large multistate health care

614

Bhardwaj and Chalasani

TABLE 6 Characteristics of the 89 Patients in the United States with Troglitazone-Induced Acute Liver Failure Characteristic Age (years) Female sex Troglitazone use Daily dose (mg) Days of therapy Monotherapy Clinical features of hepatotoxicity Hepatocellular (%) Jaundice as first sign (%) Days from jaundice to failure Outcome Died (%) Liver transplant (%)a Spontaneous recovery (%) Unknown (%)b

Number in whom data were available

Value (% or averageGs.d.)

85 87

63G13 67

69 76 89

387G128 184G142 27

71 70 77

85 54 24G19

84 84 84 5

77 12 13

a

Two patients who received liver died shortly afterwards. At last follow-up, one patient was awaiting liver transplant and one had been discharged from the hospital with encephalopathy (not a candidate for liver transplant). Source: Modified from Ref. 426.

b

organization (United Health Group) Graham and colleagues estimated that, per million personyears of TGZ exposure, the incidence rates of liver injury requiring hospitalization was 1244 (95% CI: 404–2900), hospitalized jaundice was 995 (95% CI: 271–2546) and fulminant liver failure was 240 (95% CI: 6.3–1385) (424). The pathogenesis of TGZ-induced significant liver injury is unknown. A variety of mechanisms including accumulation of TGZ in liver cells, oxidative stress and mitochondrial damage, effects on bile salt export protein and liver cell apoptosis have been suggested (428). However, data are not convincing that any of these mechanisms are significant factors in the pathogenesis of TGZ hepatotoxicity in humans (429). In an unpublished toxicogenomic study consisting of 4079 diabetic patients receiving TGZ, combined genetic polymorphisms involving specific genes were associated with increased susceptibility to TGZ-induced liver injury (430). These genetic polymorphisms included heterozygous single point mutations for CYP1A1 (increased drug interactions), NQO1 (reduced metabolic activity), GLUT-1 (increased risk for diabetes type 2) and PPARg-1431 (increased leptin levels) (430). In another small study consisting of diabetic patients, combined glutathione-S-transferase GSTT1–GSTM1 null genotype was significantly associated with TGZ-induced liver injury (422). Second Generation Thiazolidinediones Currently, two second generation TZDs are available in the United States and these are rosiglitazone and pioglitazone. They have been prescribed extensively since their approval. Except for isolated case reports of significant liver injury, these two agents are very safe from the hepatotoxicity standpoint. In the following sections, we review pertinent aspects of reported hepatotoxicity from these two agents. Rosiglitazone

Rosiglitazone increases insulin sensitivity similar to other TZDs but it improves glycemic control and reduces circulating insulin levels at lower doses than TGZ. It is a highly selective and potent agonist of the PPARg receptor that is found in key target tissues for insulin action such as liver, skeletal muscle, and adipose tissue. Rosiglitazone has clinical activity when used as monotherapy or in combination with other agents, such as metformin (431,432). Available clinical trial data and post-marketing experience showed that this agent is remarkably safe except for rare reports of significant liver injury. In placebo-controlled clinical trials, 0.2% of patients had reversible ALT elevations greater than three times the ULN and this frequency was not different than controls (433). Sixty-eight cases of “hepatitis” or “acute liver failure” due to rosiglitazone have been reported to the FDA; however, in many cases

615

Cardiovascular and Antidiabetic Medications

TABLE 7

Hepatotoxicity from Thiazolidinediones: Summary from NDA Safety Databases and AERS Database (US FDA) NDA clinical trials data

Troglitazone (TGZ) TGZ Placebo Rosiglitazone Rosiglitazone Placebo Pioglitazone Pioglitazone Placebo

AERS database Fulminant liver failure (estimated per 106 pt-yrs)

N

ALTO3!ULN

ALTO10!ULN

2510 475

1.9% 0.6%

0.68% 0

3503 574

0.2 0.2

0 0

3

1526 793

0.3 0.3

0 0

4

63

Fulminant liver failure: fatal cases of liver failureCliver transplants. Abbreviations: NDA, new drug application; AERS, adverse event reports and surveillance; ALT, alanine aminotransferase; ULN, upper limit of normal. Source: From Ref. 434.

there were competing etiologies for liver injury (e.g., heart failure or concomitant medications) and thus establishing causal relationship was very difficult (434). Isolated cases of severe cholestatic hepatitis, hepatocellular damage and acute liver failure from rosiglitazone have been reported in the literature (434). Based on AERS database (US FDA), it is estimated that the risk of fulminant liver failure (either fatal or requiring liver transplantation) is 3 per 106 patient-years of rosiglitazone usage (Table 7) (432). The package insert for rosiglitazone (and pioglitazone) recommends monitoring ALT levels in patients receiving these agents and they be not used in those with increased liver enzymes (ALT O2.5 ULN) (435). These recommendations are not evidence-based and are problematic because a significant proportion of diabetics likely to benefit from these agents have underlying fatty liver disease. Two recent studies have shown that the risk of hepatotoxicity from rosiglitazone in patients with elevated baseline aminotransferases is not significantly different than those with normal aminotransferases receiving rosiglitazone (Table 8) (436,437). It is noteworthy that several reports suggest that rosiglitazone and other TZDs may improve liver status (by imaging or biochemically or histologically) in patients with NAFLD (438–440). Pioglitazone

Pioglitazone monohydrochloride is approved to treat type 2 diabetes mellitus (DM), and chemically it is a racemic mixture with both enantiomers of equal pharmacological activity. Pioglitazone is extensively metabolized by hydroxylation and oxidation; the metabolites are partly converted to glucuronide or sulfate conjugates. Metabolites M-II and M-IV (hydroxyl derivatives of pioglitazone) and M-III (keto derivatives of pioglitazone) are pharmacologically TABLE 8 The Course of Liver Biochemistries Following Rosiglitazone Prescription in Patients with Elevated Baseline Aminotransferases (Cohort 1) and Normal Baseline Aminotransferases (Cohort 2) AST during f/u, meanGs.d. (IU/L) ALT during f/u, meanGs.d (IU/L) Bilirubin during f/u meanGs.d (mg/dl) Discontinuation of rosiglitazonea Mild-moderate elevations in liver biochemistriesa Severe elevations in liver biochemistriesa Hy’s rulea a

Cohort 1 (nZ210)

Cohort 2 (nZ628)

38G35 36G30 0.4G0.4 8.6% 10% 0.9% 0

29G28 26G23 0.5G0.6 8.1%b 6.6%b 0.6%b 2 (0.3%)b

See Reference 436 for definitions. pZns versus variable in the next column. Abbreviations: AST, aspartate aminotransferase; ALT, alanine aminotransferase. Source: Modified from Ref. 436.

b

616

Bhardwaj and Chalasani

active in animal models of type 2 DM. At steady-state, pioglitazone comprises approximately 30% to 50% of the total peak serum concerntrations and 20% to 25% of the total serum area under the curve. In vitro data demonstrate that multiple CYP isoforms are involved in the metabolism of pioglitazone. The CYPs involved are cytochrome P450 2C8 and, to a lesser degree, cytochrome P450 3A4 with additional contribution from a variety of other CYPs including extrahepatic CYP1A1. During preapproval placebo-controlled trials in the United States, 4 out of 1526 patient receiving pioglitazone (0.26%) and 2 out of 793 placebo-treated patients (0.25%) had ALT values greater than or equal to three times the ULN (441). It is reported that 37 cases of “hepatitis” or “acute liver failure” were reported to FDA or published in the literature (434). Based on AERS database (US FDA), it is estimated that the risk of fulminant liver failure (either fatal or requiring liver transplantation) is 4 per 106 patient-years of pioglitazone usage (Table 7) (432). It is currently not recommended that pioglitazone or rosiglitazone be used in individuals who had documented hepatotoxicity to TGZ. Biguanides Metformin increases the effects of circulating insulin by augmenting the number and effectiveness of peripheral insulin receptors. It rarely causes a cholestatic hepatitis with a mixed laboratory pattern of hepatocellular and cholestatic enzyme elevations (403). Jaundice is an uncommon but a possible clinical finding (366,404,442). The enzyme elevations usually occur within one to two months of institution of the drug, often with a more prominent elevation of transaminases compared to alkaline phosphatase (403,442). Transaminases fall quickly upon discontinuation of the drug, however, alkaline phosphatase may remain persistently elevated (366,403). Liver biopsy has shown a mild portal inflammation with severe cholestasis in the currently available literature (which is sparse) (403,443,444). Metformin exhibits first-pass metabolism in the liver, but there is no evidence of dose-related toxicity in the cases reported thus far (401,443). Metformin has also been associated with severe acute hepatitis with lethargy, fatigue, and diarrhea (401,443). In this case, centrilobular necroinflammatory injury is seen in addition to portal venular endothelialitis (402,404,443). The hepatotoxicity in either form of drug-induced liver injury has been shown to resolve spontaneously upon removal of the drug from the patient’s medication regimen (445,446). a-Glucosidase Inhibitors Acarbose Acarbose is an oligosaccharide that causes reversible inhibition of enzymes (collectively called a-glucosidase) on the brush border of the small intestine and, consequently, causes an inhibition of the absorption of carbohydrates. Due to the nature of this medication as a minimally absorbable compound in the gut (!1% systemic absorption), the drug was generally regarded as safe in terms of any major organ toxicity (446–448). However a number of cases exist, documenting adverse hepatic effects with its use. The latency prior to evidence of injury may be several months as the period of recovery can be which usually occurs after discontinuation of the drug. Evidence of hypersensitivity such as eosinophilia with fever and rash has not been reported (445). This would seem to suggest an idiosyncratic type of injury rather than an intrinsic dose-dependent type of toxicity or an allergic-type injury, although these have not been ruled out. Dose-related toxicity may yet be a possibility as most cases of hepatic injury occur at doses of greater than 100 mg daily (445,447–451). A hepatocellular pattern of injury with a potential transaminase elevation up to the several thousands has been documented (446–448,451). The transaminase elevations are, however, rarely seen and are usually asymptomatic. In those cases with moderate to severe elevations in transaminases, hepatomegaly has been noted upon examination in several of the cases (445–448). Liver biopsies have been performed on patients with acarbose-related liver injury and have revealed a mild, pan-lobular inflammatory infiltrate consisting of mainly lymphocytes (445,448,449,452). Mild periportal fibrosis has also been seen in some specimens. Recovery with discontinuation of acarbose is usually seen within two to four months (445).

Cardiovascular and Antidiabetic Medications

617

It is intriguing that the majority of cases have been reported in Spain. It has been suggested that the toxicity may be due to genetic differences in metabolism and, subsequently, higher levels of 4-methylpyrogallol derivatives of acarbose (447,452). The theory of metaboliterelated hepatotoxic injury is supported by the relatively long duration from the start of therapy until the onset of liver injury. There seems to be no other explanation for the mechanism of injury as a minute amount of acarbose actually reaches the liver (445). Other a-Glucosidase Inhibitors Very few reports have been published on the hepatotoxic potential of voglibose, the use of which has been associated with intrahepatic cholestasis acutely and submassive hepatic necrosis in one report after long-term use (453). Treatment with miglitol has not been associated with clear-cut evidence of hepatocellular injury although experience with this medication is much less than that of acarbose (454,455). Metiglinide Analogues Metiglinide and its analogues stimulate the secretion of insulin from pancreatic b cells (456,457). Thus far, two case reports of repaglinide-induced liver toxicity have been published (456). Both cases describe an idiosyncratic hepatic reaction with a cholestatic pattern of liver injury and modest elevations in the transaminases. Both cases have shown that biochemical evidence of liver injury reversed upon withdrawal of the drug. One biopsy revealed mild bile duct proliferation and a mainly lymphocytic portal infiltrate (458,459). To date, there are no reported cases of hepatotoxicity with nateglinide. ACKNOWLEDGMENT

This work is supported in part by K24 DK 069290 (NC).

REFERENCES 1. Chan KA, et al. A cohort study of the incidence of serious acute liver injury in diabetic patients treated with hypoglycemic agents. Arch Intern Med 2003; 163(6):728–34. 2. Huerta C, Zhao SZ, Garcia Rodriguez LA. Risk of acute liver injury in patients with diabetes. Pharmacotherapy 2002; 22(9):1091–6. 3. Avins AL, et al. Lovastatin is not hepatotoxic to patients with pre-existing liver disease. Abstract presented at DDW 2006. 4. Waxman HL, et al. Amiodarone for control of sustained ventricular tachyarrhythmia: clinical and electrophysiologic effects in 51 patients. Am J Cardiol 1982; 50(5):1066–74. 5. Breuer HW, et al. Amiodarone-induced severe hepatitis mediated by immunological mechanisms. Int J Clin Pharmacol Ther 1998; 36(6):350–2. 6. Coste T, et al. Chronic hepatitis associated with the intake of amiodarone. Gastroenterol Clin Biol 1986; 10(10):689–90. 7. Gay G, et al. A case of amiodarone-induced hyperthyroidism and hepatitis: results of clinical laboratory and histological tests. Biomed Pharmacother 1986; 40(2):54–7. 8. Aldershvile J, Hansen PF. Amiodarone and liver damage. Ugeskr Laeger 1987; 149(22):1470–1. 9. Anastasiou-Nana MI, et al. High incidence of clinical and subclinical toxicity associated with amiodarone treatment of refractory tachyarrhythmias. Can J Cardiol 1986; 2(3):138–45. 10. Jeyamalar R, et al. Hepatotoxicity of amiodarone. Ann Acad Med Singapore 1992; 21(6):838–40. 11. Kalantzis N, et al. Acute amiodarone-induced hepatitis. Hepatogastroenterology 1991; 38(1):71–4. 12. Lewis JH, et al. Amiodarone hepatotoxicity: prevalence and clinicopathologic correlations among 104 patients. Hepatology 1989; 9(5):679–85. 13. Agozzino F, Picca M, Pelosi G. Acute hepatitis complicating intravenous amiodarone treatment. Ital Heart J 2002; 3(11):686–8. 14. Benbassat C, Shahar A. Acute fulminant hepatic failure after short-term amiodarone. Harefuah 1991; 121(10):378–80. 15. Bravo AE, et al. Hepatotoxicity during rapid intravenous loading with amiodarone: description of three cases and review of the literature. Crit Care Med 2005; 33(1):128–34 (discussion 245–246). 16. Capron-Chivrac D, et al. Acute hepatopathy caused by amiodarone. Study of a case and review of the literature. Gastroenterol Clin Biol 1985; 9(6–7):535–9. 17. de Korwin JD, et al. Acute hepatitis attributed to amiodarone. Gastroenterol Clin Biol 1986; 10(10):688–9.

618

Bhardwaj and Chalasani

18. Giannattasio F, et al. Three cases of severe acute hepatitis after parenteral administration of amiodarone: the active ingredient is not the only agent responsible for hepatotoxicity. Ann Ital Med Int 2002; 17(3):180–4. 19. Gonzalez Galilea A, et al. Early-onset acute toxic hepatitis induced by intravenous amiodarone administration. Gastroenterol Hepatol 2002; 25(6):392–4. 20. Gregory SA, Webster JB, Chapman GD. Acute hepatitis induced by parenteral amiodarone. Am J Med 2002; 113(3):254–5. 21. Lupon-Roses J, et al. Probable early acute hepatitis with parenteral amiodarone. Clin Cardiol 1986; 9(5):223–5. 22. Guglin M. Intravenous amiodarone: offender or bystander? Crit Care Med 2005; 33(1):245–6 (comment). 23. James PR, Barrera Groba C. Intravenous amiodarone in intensive care: time for a reappraisal? Intensive Care Med 2001; 27(8):1433. 24. Gilinsky NH, Briscoe GW, Kuo CS. Fatal amiodarone hepatoxicity. Am J Gastroenterol 1988; 83(2):161–3. 25. Iliopoulou A, et al. Reversible fulminant hepatitis following intravenous amiodarone loading. Amiodarone hepatotoxicity. Int J Clin Pharmacol Ther 1999; 37(6):312–3. 26. James PR, Hardman SM. Acute hepatitis complicating parenteral amiodarone does not preclude subsequent oral therapy. Heart 1997; 77(6):583–4. 27. Lopez-Gomez D, et al. The use of oral amiodarone as a chronic treatment in a patient with prior fulminant hepatitis due to intravenous amiodarone. Rev Esp Cardiol 1999; 52(3):201–3. 28. Rhodes A, Eastwood JB, Smith SA. Early acute hepatitis with parenteral amiodarone: a toxic effect of the vehicle? Gut 1993; 34(4):565–6. 29. Bach N, et al. Amiodarone hepatotoxicity: progression from steatosis to cirrhosis. Mt Sinai J Med 1989; 56(4):293–6. 30. Elving LD, et al. Amiodarone hepatotoxicity. Neth J Med 1986; 29(10):303–7. 31. Lim PK, et al. Neuropathy and fatal hepatitis in a patient receiving amiodarone. Br Med J Clin Res Ed 1984; 288(6431):1638–9. 32. MacFadyen RJ, Palmer TJ, Hisamuddin K. Rapidly fatal acute amiodarone hepatitis occurring in the context of multiple organ failure. Int J Cardiol 2003; 91(2–3):245–7. 33. Richer M, Robert S. Fatal hepatotoxicity following oral administration of amiodarone. Ann Pharmacother 1995; 29(6):582–6. 34. Lamproye A, et al. Pseudo-alcoholic hepatitis and cirrhosis caused by amiodarone (Cordarone). Rev Med Liege 1998; 53(3):121–4. 35. Rigas B. The evolving spectrum of amiodarone hepatotoxicity. Hepatology 1989; 10(1):116–7. 36. Courtesy of David Kleiner, MD, National Cancer Institute. 37. Courtesy of Dr. Oscar W. Cummings, Dept. of Pathology and Laboratory Medicine, Indiana University Medical Center, Indianapolis, Indiana. 38. Flaharty KK, et al. Hepatotoxicity associated with amiodarone therapy. Pharmacotherapy 1989; 9(1):39–44. 39. Fornaciari G, et al. Amiodarone-induced acute hepatitis: case report. J Clin Gastroenterol 1992; 15(3):271–3. 40. Rinder HM, Love JC, Wexler R. Amiodarone hepatotoxicity. N Engl J Med 1986; 314(5):318–9. 41. Maker AV, Orgill DP. Rapid acute amiodarone-induced hepatotoxicity in a burn patient. J Burn Care Rehabil 2005; 26(4):341–3. 42. Morse RM, et al. Amiodarone-induced liver toxicity. Ann Intern Med 1988; 109(10):838–40. 43. Olsson R. Acute hepatitis after parenteral amiodarone. Ital Heart J 2003; 4(5):355–6 (comment). 44. Chang CC, et al. Severe intrahepatic cholestasis caused by amiodarone toxicity after withdrawal of the drug: a case report and review of the literature. Arch Pathol Lab Med 1999; 123(3):251–6. 45. Rigas B, et al. Amiodarone hepatotoxicity. A clinicopathologic study of five patients. Ann Intern Med 1986; 104(3):348–51. 46. Simon JB, et al. Amiodarone hepatotoxicity simulating alcoholic liver disease. N Engl J Med 1984; 311(3):167–72. 47. Lewis JH, et al. Histopathologic analysis of suspected amiodarone hepatotoxicity. Hum Pathol 1990; 21(1):59–67. 48. Somani P, et al. Amiodarone- and desethylamiodarone-induced myelinoid inclusion bodies and toxicity in cultured rat hepatocytes. Hepatology 1990; 11(1):81–92. 49. Puli SR, et al. Hepatic cirrhosis caused by low-dose oral amiodarone therapy. Am J Med Sci 2005; 330(5):257–61. 50. Josephson SA, Kessel ER. Amiodarone hepatotoxicity. Dig Dis 1997; 15(4–5):312. 51. Pollak PT, Shafer SL. Use of population modeling to define rational monitoring of amiodarone hepatic effects. Clin Pharmacol Ther 2004; 75(4):342–51. 52. Geltner D, et al. Quinidine hypersensitivity and liver involvement. A survey of 32 patients. Gastroenterology 1976; 70(5 Pt 1):650–2.

Cardiovascular and Antidiabetic Medications

619

53. Knobler H, et al. Quinidine-induced hepatitis. A common and reversible hypersensitivity reaction. Arch Intern Med 1986; 146(3):526–8. 54. Bourliere M, et al. Quinidine phenylethylbarbiturate-induced fulminant hepatitis in a pregnant woman. A case report. J Hepatol 1988; 6(2):214–6. 55. Dzur JR. Letter: quinidine hepatotoxicity. JAMA 1976; 235(9):908. 56. Guharoy SR, Shahin J, Levin S. Quinidine-induced hepatotoxicity revisited. Vet Hum Toxicol 1991; 33(6):613–4. 57. Alix M, et al. Cholestatic hepatitis caused by quinidine phenylethylbarbiturate, and sarcoidosis. Apropos of a case. Therapie 1989; 44(1):69. 58. Bramlet DA, Posalaky Z, Olson R. Granulomatous hepatitis as a manifestation of quinidine hypersensitivity. Arch Intern Med 1980; 140(3):395–7. 59. Chajek T. Letter: quinidine and granulomatous hepatitis. Ann Intern Med 1975; 82(2):282. 60. Chajek T, et al. Quinidine-induced granulomatous hepatitis. Ann Intern Med 1974; 81(6):774–6. 61. Handler SD, et al. Quinidine hepatitis. Arch Intern Med 1975; 135(6):871–2. 62. Koch MJ, et al. Quinidine hepatotoxicity. A report of a case and review of the literature. Gastroenterology 1976; 70(6):1136–40. 63. Chuang LC, et al. Possible case of procainamide-induced intrahepatic cholestatic jaundice. Ann Pharmacother 1993; 27(4):434–7. 64. Worman HJ, et al. Hypersensitivity reaction associated with acute hepatic dysfunction following a single intravenous dose of procainamide. J Intern Med 1992; 232(4):361–3. 65. Rotmensch HH, et al. Granulomatous hepatitis: a hypersensitivity response to procainamide. Ann Intern Med 1978; 89(5 Pt 1):646–7. 66. Farber HI. Fever, vomiting, and liver dysfunction with procainamide therapy. Postgrad Med 1974; 56(1):155–6. 67. Drayer DE, Reidenberg MM. Clinical consequences of polymorphic acetylation of basic drugs. Clin Pharmacol Ther 1977; 22(3):251–8. 68. du Souich P, Erill S. Metabolism of procainamide and p-aminobenzoic acid in patients with chronic liver disease. Clin Pharmacol Ther 1977; 22(5 Pt 1):588–95. 69. Brandes JW, et al. Jaundice after aprindin (author’s transl). Dtsch Med Wochenschr 1976; 101(4):111–3. 70. Herlong HF, et al. Aprindine hepatitis. Ann Intern Med 1978; 89(3):359–61. 71. Stoel I, Hagemeijer F. Aprindine: a review. Eur Heart J 1980; 1(2):147–56. 72. Elisaf M, et al. Aprindine-induced hepatic granulomata. J Hepatol 1992; 14(2–3):276–9. 73. Mondardini A, et al. Propafenone-induced liver injury: report of a case and review of the literature. Gastroenterology 1993; 104(5):1524–6. 74. Schuff-Werner P, et al. Propafenon-induced cholestatic liver injury—a further example for allergic drug hepatitis (author’s transl). Z Gastroenterol 1981; 19(10):673–9. 75. Cocozzella D, et al. Propafenone hepatotoxicity: report of two new cases. Dig Dis Sci 2003; 48(2):354–7. 76. Grieco A, et al. Propafenone-related cholestatic hepatitis in an elderly patient. Ital Heart J 2002; 3(7):431–4. 77. Kuhlkamp V, Haasis R, Seipel L. Flecainide-induced hepatitis. Z Kardiol 1988; 77(10):678–80. 78. Mikloweit P, Bienmuller H. Drug-induced intrahepatic cholestasis caused by flecainide acetate and enalapril. Internist 1987; 28(3):193–5 (erratum appears in Internist (Berl) 1987 August 28(8):568). 79. Antonelli D, Koltun B, Barzilay J. Acute hepatotoxic effect of disopyramide. Chest 1984; 86(2):274. 80. Bakris GL, Cross PD, Hammarsten JE. Disopyramide-associated liver dysfunction. Mayo Clin Proc 1983; 58(4):265–7. 81. Doody PT. Disopyramide hepatotoxicity and disseminated intravascular coagulation. South Med J 1982; 75(4):496–8. 82. Quart BD, et al. Drug interaction studies and encainide use in renal and hepatic impairment. Am J Cardiol 1986; 58(5):104–13. 83. Dorne JL, et al. Human variability in polymorphic CYP2D6 metabolism: is the kinetic default uncertainty factor adequate? Food Chem Toxicol 2002; 40(11):1633–56. 84. Higa K, Hirata K, Dan K. Mexiletine-induced severe skin eruption, fever, eosinophilia, atypical lymphocytosis, and liver dysfunction. Pain 1997; 73(1):97–9. 85. McCarthy JD, Picazo JG. Lidocaine in the common bile duct. Effects and possible toxicity. JAMA 1969; 209(12):1904–5. 86. Pernot C, et al. Reversible hepatic anomalies during treatment with mexiletine. Presse Med 1983; 12(31):1938. 87. Uematsu H, Hiei K, Kawasaki H. Unknown fever and abnormal liver functions after repeated epidural blocks with lidocaine for management of herpes zoster pain. Masui 1994; 43(3):405–8. 88. Farquhar DL, Davidson NM. Possible hepatoxicity of tocainide. Scott Med J 1984; 29(4):238. 89. Tucker LE. Tocainide-induced granulomatous hepatitis. JAMA 1986; 255(24):3362.

620

Bhardwaj and Chalasani

90. Ducobu J. Innovative therapies in metabolic diseases: ezetimibe (Ezetrol), nicotinic acid (Niaspan), acids omega-3 (Omacor), rimonabant (Acomplia). Rev Med Brux 2005; 26(4):S393–405. 91. Parra JL, Reddy KR. Hepatotoxicity of hypolipidemic drugs. Clin Liver Dis 2003; 7(2):415–33. 92. Tato F, Vega GL, Grundy SM. Effects of crystalline nicotinic acid-induced hepatic dysfunction on serum low-density lipoprotein cholesterol and lecithin cholesteryl acyl transferase. Am J Cardiol 1998; 81(6):805–7. 93. Eidelman RS, Lamas GA, Hennekens CH. The new National Cholesterol Education Program guidelines: clinical challenges for more widespread therapy of lipids to treat and prevent coronary heart disease. Arch Intern Med 2002; 162(18):2033–6. 94. Pasternal RC, et al. ACC/AHA/NIHLBI clinical advisory on the use and safety of statins. Circulation 2002; 106:1024–8. 95. Accessed July 2, 2006 at www.drugtopics.com 96. Pharmacy facts and figures, (Accessed January 5, 2005 at www.drugtopics.com). 97. MacDonald JS, et al. Preclinical evaluation of lovastatin. Am J Cardiol 1988; 62(15):16–27J. 98. Horsmans Y, Desager JP, Harvengt C. Biochemical changes and morphological alterations of the liver in guinea-pigs after administration of simvastatin (HMG CoA reductase-inhibitor). Pharmacol Toxicol 1990; 67(4):336–9. 99. Mohapatra R, et al. A review of the oral direct thrombin inhibitor ximelagatran: not yet the end of the warfarin era. Am Heart J 2005; 150(1):19–26. 100. Farmer JA, Torre-Amione G. Comparative tolerability of the HMG-CoA reductase inhibitors. Drug Saf 2000; 23(3):197–213. 101. Chalasani N. Statins and hepatotoxicity: focus on patients with fatty liver. Hepatology 2005; 41(4):690–5. 102. Pfeffer MA, et al. Safety and tolerability of pravastatin in long-term clinical trials. Prospective pravastatin pooling (PPP) project. Circulation 2002; 105:2341–6. 103. de Denus S, et al. Statins and liver toxicity: a meta-analysis. Pharmacotherapy 2004; 24:584–91. 104. Chitturi S, George J. Hepatotoxicity of commonly used drugs: nonsteroidal anti-inflammatory drugs, antihypertensives, antidiabetic agents, anticonvulsants, lipid lowering agents, psychotropic drugs. Semin Liver Dis 2002; 22(2):169–83. 105. Tolman KG. The liver and lovastatin. Am J Cardiol 2002; 89:1374–80. 106. Tolman KG. Defining patient risks from expanded preventive therapies. Am J Cardiol 2000; 85:15–9E. 107. Perger L, et al. Fatal liver failure with atorvastatin. J Hepatol 2003; 39:1096–107. 108. Graziadei IW, et al. Drug-induced lupus-like syndrome associated severe autoimmune hepatitis. Lupus 2003; 12:409–12. 109. Pelli N, et al. Autoimmune hepatitis revealed by atorvastatin. Eur J Gastroenterol Hepatol 2003; 15:921–4. 110. Alla V, Abraham J, Siddiqui J, et al. Autoimmune hepatitis triggered by statins. J Clin Gastroenterol 2006; 40(8):757–61. 111. Russo MW, et al. Liver transplantation from drug induced liver injury in the United States. Liver Transpl 2004; 10:1018–23. 112. Newman TB, Hulley SB. Carcinogenecity of lipid-lowering agents. JAMA 1996; 275:55–60. 113. Guallar E, Goodman SN. Statins and cancer: a case of meta-uncertainty. Am J Med 2001; 110:738–40. 114. Strandberg TE, et al. Mortality and incidence of cancer during 10-year follow-up of the Scandinavian Simvastatin Survival Study (4S). Lancet 2004; 364:771–7. 115. Caldwell SH, Zaidman JS, Hespenheidec EE. The liver and statin therapy: uncertain navigation in the sea of risk-benefit. Pharmacoepidemiol Drug Saf 2003; 12:303–6. 116. Assay N, et al. Fatty infiltration of liver in hyperlipidemic patients. Dig Dis Sci 2000; 45:1229–34. 117. Kemmer NM, et al. High prevalence of NASH among Mexican American females with type II diabetes mellitus. Gastroenterology 2001; 120:A117. 118. Foster KJ, et al. Liver disease in patients with diabetes mellitus. Postgrad Med J 1980; 56:767–72. 119. Browning JD, et al. Prevalence of hepatic steatosis in an urban population in the United States: impact of ethnicity. Hepatology 2004; 40:1387–95. 120. Matteoni CA, et al. Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity. Gastroenterology 1999; 116:1413–9. 121. Chalasani N, Deeg MA, Crabb DW. Systemic levels of lipid peroxidation and its metabolic and dietary correlates in non-alcoholic steatohepatitis. Am J Gastroenterol 2004; 99:1497–502. 122. Chalasani N, et al. Patients with elevated liver enzymes are not at higher risk for statin hepatotoxicity. Gastroenterology 2004; 126(5):1287–92. 123. Vuppalanchi R, Teal E, Chalasani N. Patients with elevated baseline liver enzymes do not have higher frequency of hepatotoxicity from lovastatin than those with normal baseline liver enzymes. Am J Med Sci 2005; 329(2):62–5. 124. Avins AL, et al. Lovastatin is not hepatotoxic to patients with pre-existing liver disease (submitted).

Cardiovascular and Antidiabetic Medications

621

125. Cohen D, Anania F, Chalasani N. Report of the liver expert panel. Statin Safety Task Force, National Lipid Association. Am Journal of Cardiol (in press). 126. Dalton TA, Berry RS. Hepatotoxicity associated with sustained-release niacin. Am J Med 1992; 93(1):102–4. 127. Etchason JA, et al. Niacin-induced hepatitis: a potential side effect with low-dose time-release niacin. Mayo Clin Proc 1991; 66(1):23–8. 128. Henkin Y, Johnson KC, Segrest JP. Rechallenge with crystalline niacin after drug-induced hepatitis from sustained-release niacin. JAMA 1990; 264(2):241–3. 129. Hodis HN. Acute hepatic failure associated with the use of low-dose sustained-release niacin. JAMA 1990; 264(2):181. 130. Reimund E, Ramos A. Niacin-induced hepatitis and thrombocytopenia after 10 years of niacin use. J Clin Gastroenterol 1994; 18(3):270–1. 131. Rizakallah GS, et al. Clinical inquiries. Should liver enzymes be checked in a patient taking niacin? J Fam Pract 2005; 54(3):265–8. 132. Rourk RM, Rehman NU. Niacin. Help for your cholesterol—harm for your liver. N C Med J 1998; 59(2):87–8. 133. Mullin GE, Greenson JK, Mitchell MC. Fulminant hepatic failure after ingestion of sustained-release nicotinic acid. Ann Intern Med 1989; 111(3):253–5. 134. Clementz GL, Holmes AW. Nicotinic acid-induced fulminant hepatic failure. J Clin Gastroenterol 1987; 9(5):582–4. 135. Moon YS, Kashyap ML. Niacin extended-release/lovastatin: combination therapy for lipid disorders. Expert Opin Pharmacother 2002; 3(12):1763–71. 136. Anonymous. ASHP therapeutic position statement on the safe use of niacin in the management of dyslipidemias. American Society of Health-System Pharmacists. Am J Health Syst Pharm 1997; 54(24):2815–9. 137. Jungnickel PW, Maloley PA. Comment: adverse-effect profile of sustained-release niacin. DICP 1991; 25(9):1014–5. 138. Scheer MS, et al. Ultrasonographic findings in niacin-induced hepatitis. J Ultrasound Med 1999; 18(4):321–3. 139. Ferenchick G, Rovner D. Hepatitis and hematemesis complicating nicotinic acid use. Am J Med Sci 1989; 298(3):191–3. 140. Kristensen T, Olcott EW. Effects of niacin therapy that simulate neoplasia: hepatic steatosis with concurrent hepatic dysfunction. J Comput Assist Tomogr 1999; 23(2):314–7. 141. Coppola A, Brady PG, Nord HJ. Niacin-induced hepatotoxicity: unusual presentations. South Med J 1994; 87(1):30–2. 142. Lavie CJ, Milani RV. Safety and side effects of sustained-release niacin. JAMA 1994; 272(7):513–4 (author reply 514–5). 143. Knopp RH. Niacin and hepatic failure. Ann Intern Med 1989; 111(9):769. 144. Gray DR, et al. Efficacy and safety of controlled-release niacin in dyslipoproteinemic veterans. Ann Intern Med 1994; 121(4):252–8. 145. McCarty MF. Co-administration of equimolar doses of betaine may alleviate the hepatotoxic risk associated with niacin therapy. Med Hypotheses 2000; 55(3):189–94. 146. Capuzzi DM, et al. Efficacy and safety of an extended-release niacin (Niaspan): a long-term study. Am J Cardiol 1998; 82(12A):74–81U (discussion 85–86U). 147. Malloy MJ, Frost PH, Kane JP. Niacin—the long and the short of it. West J Med 1991; 155(4):424–6. 148. Rader JI, Calvert RJ, Hathcock JN. Hepatic toxicity of unmodified and time-release preparations of niacin. Am J Med 1992; 92(1):77–81. 149. Pieper JA. Overview of niacin formulations: differences in pharmacokinetics, efficacy, and safety. Am J Health Syst Pharm 2003; 60(13 Suppl. 2):S9–14 (quiz S25). 150. McKenney J. Niacin for dyslipidemia: considerations in product selection. Am J Health Syst Pharm 2003; 60(10):995–1005. 151. Schwenk TL, Fisher M. Hepatitis caused by low-dose sustained-release niacin. J Am Board Fam Pract 1994; 7(3):242–4. 152. Lawrence SP. Transient focal hepatic defects related to sustained-release niacin. J Clin Gastroenterol 1993; 16(3):234–6. 153. Patterson DJ, et al. Niacin hepatitis. South Med J 1983; 76(2):239–41. 154. McKenney JM, et al. A comparison of the efficacy and toxic effects of sustained vs. immediate-release niacin in hypercholesterolemic patients. JAMA 1994; 271(9):672–7. 155. Knopp RH. Clinical profiles of plain versus sustained-release niacin (Niaspan) and the physiologic rationale for nighttime dosing. Am J Cardiol 1998; 82(12A):8–24U (discussion 39–41U). 156. Chen H, et al. Effect of the peroxisome proliferators ciprofibrate and perfluorodecanoic acid on hepatic cell proliferation and toxicity in Sprague–Dawley rats. Carcinogenesis 1994; 15(12):2847–50. 157. Cornwell PD, De Souza AT, Ulrich RG. Profiling of hepatic gene expression in rats treated with fibric acid analogs. Mutat Res 2004; 549(1–2):131–45.

622

Bhardwaj and Chalasani

158. Dix KJ, Coleman DP, Jeffcoat AR. Comparative metabolism and disposition of gemfibrozil in male and female Sprague–Dawley rats and Syrian golden hamsters. Drug Metab Dispos 1999; 27(1):138–46. 159. Alsheikh-Ali AA, Kuvin JT, Karas RH. Risk of adverse events with fibrates. Am J Cardiol 2004; 94(7):935–8. 160. Bustamante Balen M, et al. Acute hepatitis caused by gemfibrozil. Gastroenterol Hepatol 1998; 21(8):419–20. 161. de Diego Lorenzo A, et al. Cholestatic hepatitis caused by gemfibrozil. Rev Esp Enferm Dig 2001; 93(9):610–1. 162. Athyros VG, et al. Safety and efficacy of long-term statin–fibrate combinations in patients with refractory familial combined hyperlipidemia. Am J Cardiol 1997; 80(5):608–13. 163. Landmesser U, et al. Simvastatin versus ezetimibe: pleiotropic and lipid-lowering effects on endothelial function in humans. Circulation 2005; 111(18):2356–63. 164. Scheen AJ. Combined thiazolidinedione-insulin therapy: should we be concerned about safety? Drug Saf 2004; 27(12):841–56. 165. Cruz-Fernandez JM, et al. Efficacy and safety of ezetimibe co-administered with ongoing atorvastatin therapy in achieving low-density lipoprotein goal in patients with hypercholesterolemia and coronary heart disease. Int J Clin Pract 2005; 59(6):619–27. 166. Goldman-Levine JD, Bohlman LG. Ezetimibe/Simvastatin (vytorin) for hypercholesterolemia. Am Fam Physician 2005; 72(10):2081–2. 167. Reyderman L, et al. The effect of fluvastatin on the pharmacokinetics and pharmacodynamics of ezetimibe. Curr Med Res Opin 2005; 21(8):1171–9. 168. van Heyningen C. Drug-induced acute autoimmune hepatitis during combination therapy with atorvastatin and ezetimibe. Ann Clin Biochem 2005; 42(Pt 5):402–4. 169. Olsson R. Liver damage due to antihypertensive drugs. Acta Med Scand Suppl 1979; 628:53–6. 170. Bonkowsky HL, Brisbane J. Colitis and hepatitis caused by methyldopa. JAMA 1976; 236(14):1602–3. 171. Breland BD, Hicks GS, Jr. Hepatitis and hemolytic anemia associated with methyldopa therapy. Drug Intell Clin Pharm 1982; 16(6):489–92. 172. Dybing E. Alpha-methyldopa hepatitis. Tidsskr Nor Laegeforen 1977; 97(31):1639. 173. Granlien M, Ostergaard Kristensen HP. Methyldopa hepatitis in 2 sisters. Ugeskr Laeger 1978; 140(17):978–80. 174. Seggie J, et al. Patterns of hepatic injury induced by methyldopa. S Afr Med J 1979; 55(3):75–83. 175. Daghfous R, et al. Methyldopa-induced hepatitis: 3 case reports. Tunis Med 1994; 72(1):47–50. 176. Shalev O, et al. Methyldopa-induced immune hemolytic anemia and chronic active hepatitis. Arch Intern Med 1983; 143(3):592–3. 177. Stanley P, Mijch A. Methyldopa: an often overlooked cause of fever and transient hepatocellular dysfunction. Med J Aust 1986; 144(11):603–5. 178. Rodman JS, Deutsch DJ, Gutman SI. Methyldopa Hepatitis. A report of six cases and review of the literature. Am J Med 1976; 60(7):941–8. 179. Arranto AJ, Sotaniemi EA. Morphologic alterations in patients with alpha-methyldopa-induced liver damage after short- and long-term exposure. Scand J Gastroenterol 1981; 16(7):853–63. 180. Haider Z, Bano KA, Uddin F. Liver function profile in patients taking methyldopa. J Pak Med Assoc 1978; 28(8):98–100. 181. Moses A, Zahger D, Amir G. Cholestatic liver injury after prolonged exposure to methyldopa. Digestion 1989; 42(1):57–60. 182. Sotaniemi EA, et al. Hepatic injury and drug metabolism in patients with alpha-methyldopainduced liver damage. Eur J Clin Pharmacol 1977; 12(6):429–35. 183. Delpre G, et al. Case report. Immunological studies in a case of hepatitis following methyldopa administration. Am J Med Sci 1979; 277(2):207–13. 184. Neuberger J, et al. Antibody mediated hepatocyte injury in methyl dopa induced hepatotoxicity. Gut 1985; 26(11):1233–9. 185. Bezahler GH. Fatal methyldopa-associated granulomatous hepatitis and myocarditis. Am J Med Sci 1982; 283(1):41–5. 186. Miller AC, Jr., Reid WM. Methyldopa-induced granulomatous hepatitis. JAMA 1976; 235(18):2001–2. 187. Mirada Canals A, et al. Methyldopa-induced granulomatous hepatitis. DICP 1991; 25(11):1269–70. 188. Balazs M, Kovach G. Chronic aggressive hepatitis after methyldopa treatment. Case report with electron-microscopic study. Hepatogastroenterology 1981; 28(4):199–202. 189. Lee WM, Denton WT. Chronic hepatitis and indolent cirrhosis due to methyldopa: the bottom of the iceberg? J S C Med Assoc 1989; 85(2):75–9. 190. Minakami H, et al. Hepatic injury caused by methyldopa in a pre-eclamptic patient. Nippon Sanka Fujinka Gakkai Zasshi—Acta Obstetrica et Gynaecologica Japonica 1986; 38(1):139–42. 191. Smith GN, Piercy WN. Methyldopa hepatotoxicity in pregnancy: a case report. Am J Obste Gynecol 1995; 172(1 Pt 1):222–4.

Cardiovascular and Antidiabetic Medications

623

192. Picaud A, et al. Fatal toxic hepatitis in pregnancy. A discussion of the role of methyldopa. J Gynecol Obstet Biol Reprod 1990; 19(2):192–6. 193. Thomas LA, Cardwell MS. Acute reactive hepatitis in pregnancy induced by alpha-methyldopa. Obstet Gynecol 1997; 90(4 Pt 2):658–9. 194. Puppala AR, Steinheber FU. Fulminant hepatic failure associated with methyldopa. Am J Gastroenterol 1977; 68(6):578–81. 195. Balazs M, Kovach G. Aggressive chronic hepatitis during methyldopa therapy. Orv Hetil 1980; 121(12):705–7. 196. Thomas E. Methyldopa liver injury. J Assoc Physicians India 1976; 24(12):851–3. 197. Thomas E, et al. Spectrum of methyldopa liver injury. Am J Gastroenterol 1977; 68(2):125–33. 198. de Graaf EM, et al. A case of tizanidine-induced hepatic injury. J Hepatol 1996; 25(5):772–3. 199. Bhadoria DP, et al. Safe use of clonidine, nifedipine and vitamin K in hepatic porphyria. J Assoc Physicians India 1988; 36(11):675–6. 200. Anonymous. Dihydralazine-induced hepatitis. Schweizerische Medizinische Wochenschrift Journal Suisse de Medecine 1977; 107(19):649–50. 201. Boccaccio F, et al. Acute hepatitis caused by dihydralazine. Gastroenterol Clin Biol 1987; 11(8–9):614. 202. Pariente EA, et al. Dihydralazine hepatitis: report of a case and review of the literature. Digestion 1983; 27(1):47–52. 203. Barnett DB, Hudson SA, Golightly PW. Hydrallazine-induced hepatitis? Br Med J 1980; 280(6224):1165–6. 204. Roschlau G, Baumgarten R, Fengler JD. Dihydralazine hepatitis. Morphologic and clinical criteria for diagnosis. Zentralblatt fur Allgemeine Pathologie und Pathologische Anatomie 1990; 136(1–2):127–34. 205. Bartoli E, et al. Acute hepatitis with bridging necrosis due to hydralazine intake. Report of a case. Arch Intern Med 1979; 139(6):698–9. 206. Roschlau G. Dihydralazine-induced hepatitis with bridging necrosis. A clinico-pathologic survey of 20 cases. Zentralblatt fur Allgemeine Pathologie und Pathologische Anatomie 1983; 127(5–6):385–93. 207. Itoh S, et al. Hydralazine-induced hepatitis. Hepatogastroenterology 1981; 28(1):13–6. 208. Jori GP, Peschile C. Hydralazine disease associated with transient granulomas in the liver. A case report. Gastroenterology 1973; 64(6):1163–7. 209. Rice D, Burdick CO. Granulomatous hepatitis from hydralazine therapy. Arch Intern Med 1983; 143(5):1077. 210. Myers JL, Augur NA, Jr. Hydralazine-induced cholangitis. Gastroenterology 1984; 87(5):1185–8. 211. Itoh S, et al. Hydralazine-induced liver injury. Dig Dis Sci 1980; 25(11):884–7. 212. Kunze KD, et al. Liver lesions induced by dihydralazine and propranolol. Zentralblatt fur Allgemeine Pathologie und Pathologische Anatomie 1985; 130(1):19–29. 213. Roschlau G, Hass S, Schmehl V. Chronic drug-induced hepatitis caused by dihydralazine. Deutsche Zeitschrift fur Verdauungs- und Stoffwechselkrankheiten 1988; 48(1):41–6. 214. Colombel JF, et al. Dihydralazine hepatitis. Presse Med 1985; 14(9):552. 215. Hod M, et al. Hydralazine-induced hepatitis in pregnancy. Int J Fertility 1986; 31(5):352–5. 216. Forster HS. Hepatitis from hydralazine. N Engl J Med 1980; 302(24):1362. 217. Machnik G, et al. Drug-induced hepatitis after dihydralazine treatment with fatal consequences. Zentralblatt fur Allgemeine Pathologie und Pathologische Anatomie 1988; 134(2):167–77. 218. Stumpf JL. Fatal hepatotoxicity induced by hydralazine or labetalol. Pharmacotherapy 1991; 11(5):415–8. 219. Alderman CP. Adverse effects of the angiotensin-converting enzyme inhibitors. Ann Pharmacother 1996; 30(1):55–61. 220. Jurima-Romet M, Huang HS. Comparative cytotoxicity of angiotensin-converting enzyme inhibitors in cultured rat hepatocytes. Biochem Pharmacol 1993; 46(12):2163–70. 221. Hilburn RB, Bookstaver D, Whitlock WL. Comment: angiotension-converting enzyme inhibitor hepatotoxicity: further insights. Ann Pharmacother 1993; 27(9):1142–3. 222. Hagley MT, Hulisz DT, Burns CM. Hepatotoxicity associated with angiotensin-converting enzyme inhibitors. Ann Pharmacother 1993; 27(2):228–31. 223. Hagley MT, Benak RL, Hulisz DT. Suspected cross-reactivity of enalapril- and captopril-induced hepatotoxicity. Ann Pharmacother 1992; 26(6):780–1. 224. Zannad F, et al. Cytolytic hepatitis with captopril but not with enalapril. Cardiovasc Drugs Ther 1988; 1(6):669–70. 225. Schattner A, Kozak N, Friedman J. Captopril-induced jaundice: report of 2 cases and a review of 13 additional reports in the literature. Am J Med Sci 2001; 322(4):236–40. 226. Martinez Gutierrez A, et al. Hepatotoxicity caused by captopril. Aten Primaria 1998; 21(8):586–7. 227. Perez Roldan F, et al. Cholestatic hepatitis caused by captopril. Rev Esp Enferm Dig 1998; 90(5):378–9. 228. Pedersen JH, et al. Captopril-induced toxic hepatitis. Ugeskr Laeger 1992; 154(42):2911–2. 229. Bellary SV, Isaacs PE, Scott AW. Captopril and the liver. Lancet 1989; 2(8661):514.

624

Bhardwaj and Chalasani

230. Deira JL, et al. Captopril hepatotoxicity in a case of renal crisis due to systemic sclerosis. Nephrol Dial Transplant 1997; 12(8):1717–8. 231. Nissan A, et al. Captopril-associated “pseudocholangitis”. A case report and review of the literature. Arch Surg 1996; 131(6):670–1. 232. Vandenburg M, et al. Hepatitis associated with captopril treatment. Br J Clin Pharmacol 1981; 11(1):105–6. 233. Bas V, et al. Toxic hepatitis due to enalapril in childhood. Pediatr Int 2003; 45(6):755–7. 234. Muela Molinero A, et al. Enalapril induced cholestasic hepatitis. An Med Interna 2002; 19(9):492–3. 235. Jeserich M, et al. Acute liver failure due to enalapril. Herz 2000; 25(7):689–93. 236. Valle R, et al. Enalapril-induced hepatotoxicity. Ann Pharmacother 1993; 27(11):1405. 237. Gonzalez de la Puente MA, et al. Fatal hepatotoxicity associated with enalapril. Ann Pharmacother 2001; 35(11):1492. 238. Rosellini SR, et al. Hepatic injury related to enalapril. Gastroenterology 1989; 97(3):810. 239. Macias FM, et al. Ductopenic hepatitis related to Enalapril. J Hepatol 2003; 39(6):1091–2. 240. Lindgren A, Olsson R. Liver damage following antihypertensive therapy. A case report of hepatitis induced by lisinopril and a review. Lakartidningen 1993; 90(16):1557–8. 241. Larrey D, et al. Fulminant hepatitis after lisinopril administration. Gastroenterology 1990; 99(6):1832–3. 242. Romero-Gomez M, et al. Hepatotoxicity induced by fosinopril. J Hepatol 2001; 35(2):309–10. 243. Yeung E, et al. Ramipril-associated hepatotoxicity. Arch Pathol Lab Med 2003; 127(11):1493–7. 244. Bosch X. Losartan-induced hepatotoxicity. JAMA 1997; 278(19):1572. 245. Nygaard B, Strandgaard S. Marked hepatotoxicity associated with losartan treatment. Blood Press 1996; 5(3):190–1. 246. Andrade RJ, et al. Cholestatic hepatitis related to use of irbesartan: a case report and a literature review of angiotensin II antagonist-associated hepatotoxicity. Eur J Gastroenterol Hepatol 2002; 14(8):887–90. 247. Tabak F, et al. Losartan-induced hepatic injury. J Clin Gastroenterol 2002; 34(5):585–6. 248. Basile G, et al. Candesartan cilexetil-induced severe hepatotoxicity. J Clin Gastroenterol 2003; 36(3):273–5. 249. Gonzalez-Jimenez D, et al. Candesartan and acute liver injury. Eur J Clin Pharmacol 2000; 56(9–10):769–70. 250. Vallejo I, Garcia Morillo S, Pamies E. Acute hepatitis induced by candesartan. Med Clin 2000; 115(18):719. 251. Foruny Olcina JR, et al. Irbesartan-induced acute hepatitis. Gastroenterol Hepatol 2004; 27(2):71–2. 252. Peron JM, et al. Irbesartan induced acute hepatitis: one case. Gastroenterol Clin Biol 2005; 29(6–7):747–8. 253. Annicchiarico BE, Siciliano M. Could irbesartan trigger autoimmune cholestatic hepatitis? Eur J Gastroenterol Hepatol 2005; 17(2):247–8. 254. Kiykim A, et al. Valsartan-induced hepatotoxicity in a HBs-Ag-Positive patient. Am J Gastroenterol 2003; 98(2):507. 255. Rene JM, et al. Acute hepatitis induced by valsartan. Med Clin 2001; 117(16):637–8. 256. Thiele DL. Labetalol hepatotoxicity. Am J Med 1989; 87(3):361. 257. Stronkhorst A, Bosma A, van Leeuwen DJ. A case of labetalol-induced hepatitis. Neth J Med 1992; 40(3–4):200–2. 258. Anonymous. Labetalol hepatotoxicity. Ann Int Med 1991; 114(4):341. 259. Chon EM, Middleton RK. Labetalol hepatotoxicity. Ann Pharmacother 1992; 26(3):344–5. 260. Marinella MA. Labetalol-induced hepatitis in a patient with chronic hepatitis B infection. J Clin Hypertens 2002; 4(2):120–1. 261. Maria VA, Victorino RM. Hypersensitivity immune reaction as a mechanism for dilevalol-associated hepatitis. Ann Pharmacother 1992; 26(7–8):924–6. 262. Clark JA, Zimmerman HJ, Tanner LA. Labetalol hepatotoxicity. Ann Int Med 1990; 113(3):210–3 (erratum appears in Ann Intern Med 1990 September 15; 113(6):485). 263. Lahoti S, Lee WM. Hepatotoxicity of anticholesterol, cardiovascular, and endocrine drugs and hormonal agents. Gastroenterol Clin North Am 1995; 24(4):907–22. 264. Yusuf SW, Mishra RM. Hepatic dysfunction associated with atenolol. Lancet 1995; 346(8968):192. 265. Hagmeyer KO, Stein J. Hepatotoxicity associated with carvedilol. Ann Pharmacother 2001; 35(11):1364–6. 266. Epeirier JM, et al. Fulminant hepatitis after carbimazole and propranolol administration. Eur J Gastroenterol Hepatol 1996; 8(3):287–8. 267. Larrey D, et al. Metoprolol-induced hepatitis: rechallenge and drug oxidation phenotyping. Ann Int Med 1988; 108(1):67–8. 268. Lennard MS. Metoprolol-induced hepatitis: is the rate of oxidation related to drug-induced hepatotoxicity? Hepatology 1989; 9(1):163–4.

Cardiovascular and Antidiabetic Medications

625

269. Basile C, Mascia E. Dihydropyridine calcium channel blockers: a rare and reversible cause of hepatotoxicity with cholestasis in a CAPD patient. Nephrol Dial Transplant 1999; 14(11):2776–7. 270. Lafuente NG. Calcium channel blockers and hepatotoxicity. Am J Gastroenterol 2000; 95(8):2145. 271. Khemissa-Akouz F, et al. Amlodipine-induced acute hepatitis. Gastroenterol Clin Biol 2002; 26(6–7):637–8. 272. de Luis DA, Aller R, Moreira V. Hepatic toxicity and amlodipine. Med Clin 1998; 110(16):638–9. 273. Stravitz RT, Sanyal AJ. Drug-induced steatohepatitis. Clin Liver Dis 2003; 7(2):435–51. 274. Lopez Vivancos J, Bara Olivan B, Muniz Garcia R. Toxic hepatitis caused by amlodipine. Gastroenterol Hepatol 2002; 25(2):112. 275. Sarachek NS, London RL, Matulewicz TJ. Diltiazem and granulomatous hepatitis. Gastroenterology 1985; 88(5 Pt 1):1260–2. 276. Toft E, Vyberg M, Therkelsen K. Diltiazem-induced granulomatous hepatitis. Histopathology 1991; 18(5):474–5. 277. Kanathur N, et al. Simvastatin–diltiazem drug interaction resulting in rhabdomyolysis and hepatitis. Tenn Med 2001; 94(9):339–41. 278. Lewin JJ, III, Nappi JM, Taylor MH. Rhabdomyolysis with concurrent atorvastatin and diltiazem. Ann Pharmacother 2002; 36(10):1546–9. 279. Traverse JH, Swenson LJ, McBride JW. Acute hepatic injury after treatment with diltiazem. Am Heart J 1994; 127(6):1636–9. 280. Brodsky SJ, et al. Hepatotoxicity due to treatment with verapamil. Ann Int Med 1981; 94(4 Pt 1):490–1. 281. Hare DL, Horowitz JD. Verapamil hepatotoxicity: a hypersensitivity reaction. Am Heart J 1986; 111(3):610–1. 282. Burgunder JM, Abernethy DR, Lauterburg BH. Liver injury due to verapamil. Hepatogastroenterology 1988; 35(4):169–70. 283. de Arriba G, et al. Hepatotoxicity due to verapamil hydrochloride. Eur J Med 1993; 2(3):179–81. 284. Guarascio P, et al. Liver damage from verapamil. BMJ (Clinical Research Ed) 1984; 288(6414):362–3. 285. Anonymous. Multicentre controlled trial of tienilic acid in hyperuricaemic, hypertensive subjects, with intensive monitoring of liver function. Br J Clin Pharmacol 1983; 15(5):513–8. 286. Biour M, et al. Fulminant hepatitis due to tienilic acid. Lancet 1991; 338(8771):891. 287. Homberg JC, Andre C, Abuaf N. A new anti-liver-kidney microsome antibody (anti-LKM2) in tienilic acid-induced hepatitis. Clin Exp Immunol 1984; 55(3):561–70. 288. Lechevalier L, et al. Fulminating hepatitis caused by tienilic acid. Gastroenterol Clin Biol 1987; 11(3):262–3. 289. Lecoeur S, Andre C, Beaune PH. Tienilic acid-induced autoimmune hepatitis: anti-liver and -kidney microsomal type 2 autoantibodies recognize a three-site conformational epitope on cytochrome P4502C9. Mol Pharmacol 1996; 50(2):326–33. 290. Lecoeur S, et al. Specificity of in vitro covalent binding of tienilic acid metabolites to human liver microsomes in relationship to the type of hepatotoxicity: comparison with two directly hepatotoxic drugs. Chem Res Toxicol 1994; 7(3):434–42. 291. Neuberger J, Williams R. Immune mechanisms in tienilic acid associated hepatotoxicity. Gut 1989; 30(4):515–9. 292. Mehrotra R, Tandon P, Nath P. An experimental study of furosemide induced hepatotoxicity. Indian J Pathol Microbiol 1986; 29(3):271–6. 293. Mitchell JR, et al. Metabolic activation of furosemide to a chemically reactive, hepatotoxic metabolite. J Pharmacol Exp Ther 1976; 199(1):41–52. 294. Grewal KK, Rafeiro E, Racz WJ. Bromobenzene and furosemide hepatotoxicity: alterations in glutathione, protein thiols, and calcium. Can J Physiol Pharmacol 1996; 74(3):257–64. 295. Massey TE, et al. Furosemide toxicity in isolated mouse hepatocyte suspensions. Toxicology 1987; 43(2):149–60. 296. Spitznagle LA, et al. The role of biliary excretion in the hepatotoxicity of furosemide in the mouse. Toxicol Appl Pharmacol 1977; 39(2):283–94. 297. Wong SG, Card JW, Racz WJ. The role of mitochondrial injury in bromobenzene and furosemide induced hepatotoxicity. Toxicol Lett 2000; 116(3):171–81. 298. Mousson C, et al. Acute granulomatous interstitial nephritis and hepatitis caused by drugs. Possible role of an allopurinol-furosemide combination. Nephrologie 1986; 7(5):199–203. 299. Walker RM, McElligott TF. Furosemide induced hepatotoxicity. J Pathol 1981; 135(4):301–14. 300. Klouchek E, Popov A. Experimental liver damage by furosemide for studying drugs with hepatoprotective activity. Eksperimentalna Meditsina i Morfologiia 1989; 28(3):47–50. 301. Arinzon Z, Alexander P, Berner Y. Hydrochlorothiazide induced hepato-cholestatic liver injury. Age Ageing 2004; 33(5):509–10. 302. Hourmand-Ollivier I, et al. Fatal subfulminant hepatitis probably due to the combination benazepril–hydrochlorothiazide (Briazide). Gastroenterol Clin Biol 2000; 24(4):464.

626

Bhardwaj and Chalasani

303. Thai KE, Sinclair RD. Spironolactone-induced hepatitis. Australas J Dermatol 2001; 42(3):180–2. 304. Shuck J, et al. Spironolactone hepatitis in primary hyperaldosteronism. Ann Int Med 1981; 95(6):708–10. 305. Renkes P, Gaucher P, Trechot P. Spironolactone and hepatic toxicity. JAMA 1995; 273(5):376–7. 306. Rubin LJ, et al. Bosentan therapy for pulmonary arterial hypertension. N Engl J Med 2002; 346(12):896–903 (erratum appears in N Engl J Med 2002 April 18;346(16):1258). 307. Cohen H, et al. Bosentan therapy for pulmonary arterial hypertension. Am J Health Syst Pharm 2004; 61(11):1107–19. 308. Guevara A, Labarca J, Gonzalez-Martin G. Heparin-induced transaminase elevations: a prospective study. Int J Clin Pharmacol Ther Toxicol 1993; 31:137–41. 309. Sitbon O, et al. Effects of the dual endothelin receptor antagonist bosentan in patients with pulmonary arterial hypertension: a 1-year follow-up study. Chest 2003; 124(1):247–54. 310. Fattinger K, et al. The endothelin antagonist bosentan inhibits the canalicular bile salt export pump: a potential mechanism for hepatic adverse reactions. Clin Pharmacol Ther 2001; 69(4):223–31. 311. Bastide J, Bastide P. Phosphoglucomutase activity (renal, hepatic and blood) in rats during experimental poisoning with uranyl nitrate. Revue Francaise d Etudes Cliniques et Biologiques 1968; 13(9):893–6. 312. Garcia Roche MO, et al. Effect of ascorbic acid on the hepatotoxicity due to the daily intake of nitrate, nitrite and dimethylamine. Nahrung 1987; 31(2):99–104. 313. Casley-Smith JR, Casley-Smith JR. Frequency of coumarin hepatotoxicity. Med J Aust 1995; 162(7):391. 314. Cohen AJ. Critical review of the toxicology of coumarin with special reference to interspecies differences in metabolism and hepatotoxic response and their significance to man. Food Cosmetics Toxicol 1979; 17(3):277–89. 315. Cox D, O’Kennedy R, Thornes RD. The rarity of liver toxicity in patients treated with coumarin (1,2benzopyrone). Hum Toxicol 1989; 8(6):501–6. 316. Lake BG. Investigations into the mechanism of coumarin-induced hepatotoxicity in the rat. Arch Toxicol Suppl 1984; 7:16–29. 317. Lake BG, et al. Studies on the hepatotoxicity and metabolism of coumarin in the rat. Biochem Soc Trans 1980; 8(1):96–7 (proceedings). 318. Adler E, Benjamin SB, Zimmerman HJ. Cholestatic hepatic injury related to warfarin exposure. Arch Intern Med 1986; 146(9):1837–9. 319. Ehrenforth S, Schenk JF, Scharrer I. Liver damage induced by coumarin anticoagulants. Semin Thromb Hemost 1999; 25(1):79–83. 320. Rehnqvist N. Intrahepatic jaundice due to warfarin therapy. Acta Medica Scandinavica 1978; 204(4):335–6. 321. Bamanikar A, Hiremath S. Hepatotoxic reaction to warfarin in a recovering hepatitis patient with hypoalbuminenia. J Assoc Physicians India 2002; 50:1456. 322. Beinssen AP. Possible coumarin hepatotoxicity. Med J Aust 1994; 161(11–12):725 (erratum appears in Med J Aust 1995 March 20;162(6):336). 323. Ghosh P, Markin RS, Sorrell MF. Coumarin-induced hepatic necrosis. Am J Gastroenterol 1997; 92(2):348–9. 324. Loprinzi CL, Sloan J, Kugler J. Coumarin-induced hepatotoxicity. J Clin Oncol 1997; 15(9):3167–8. 325. Jack KR, et al. Further studies on the lobar heterogeneity in response to coumarin-mediated hepatotoxicity. Int J Exp Pathol 1996; 77(2):79–82. 326. Carlson MK, Gleason PP, Sen S. Elevation of hepatic transaminases after enoxaparin use: case report and review of unfractionated and low-molecular-weight heparin-induced hepatotoxicity. Pharmacotherapy 2001; 21(1):108–13. 327. KA AL-M, Sherker AH. Heparin-induced hepatotoxicity. Can J Gastroenterol 2001; 15(8):527–30. 328. Manfredini R, et al. Cholestatic liver reaction and heparin therapy. Arch Intern Med 2000; 160(20):3166. 329. Barreiro Lopez B, Canet JJ, Ochoa de Echaguen Aguilar A. Liver toxicity from heparin. An Med Interna 2000; 17(3):164. 330. Christiansen HM, Lassen MR, Borris LC. Biologic tolerance of two different low molecular weight heparins. Semin Thromb Hemost 1991; 17:450–4. 331. Hui CK, et al. Low molecular weight heparin-induced liver toxicity. J Clin Pharmacol 2001; 41(6):691–4. 332. Colwell CW, Spiro TE, Trowbridge AA. Use of enoxaparin, a low-molecular weight heparin, and unfractionated heparin for the prevention of deep venous thrombosis after elective hip replacement. J Bone Joint Surg 1994; 76:3–14. 333. Harenberg J, Giese C, Demple CE. Biological activity and safety of low molecular weight heparin for 8 days in human volunteers. Thromb Haemost 1989; 61:357–62. 334. Nielsen HK, Husted SE, Koopman HD. Heparin-induced increase in serum levels of aminotransferases. Acta Medica Scandinavia 1986; 215:231–3.

Cardiovascular and Antidiabetic Medications

627

335. Lambert M, et al. Modifications of liver enzymes during heparin therapy. Acta Clin Belg 1986; 41:307–10. 336. Gurewich V. Ximelagatran—promises and concerns. JAMA 2005; 293(6):736–9. 337. Hirsh J, O’Donnell M, Weitz JI. New anticoagulants. Blood 2005; 105(2):453–63. 338. Wallentin L, et al. Oral ximelagatran for secondary prophylaxis after myocardial infarction: the ESTEEM randomised controlled trial. Lancet 2003; 362(9386):789–97. 339. Schulman S, et al. Secondary prevention of venous thromboembolism with the oral direct thrombin inhibitor ximelagatran. N Engl J Med 2003; 349(18):1713–21. 340. Albers GW, et al. Ximelagatran vs. warfarin for stroke prevention in patients with nonvalvular atrial fibrillation: a randomized trial. JAMA 2005; 293(6):690–8. 341. Cardiovascular and Renal Drugs Advisory Committee September 10, 2004: EXANTA (xim elagatran) tablets NDA 21–686. Food and Drug Administration Advisory Committee web site. Available at: http://www.fda.gov/ohrms/dockets/ac/04/briefing/2004-4069B1_01_AstraZenecaBackgrounder.pdf. Accessed April 26, 2006. 342. O’Brien CL, Gage BF. Costs and effectiveness of ximelagatran for stroke prophylaxis in chronic atrial fibrillation. JAMA 2005; 293(6):699–706. 343. Adams L, et al. Ticlopidine-associated cholestatic hepatitis. Intern Med J 2002; 32(7):359–60. 344. Cassidy LJ, Schuster BG, Halparin LS. Probable ticlopidine-induced cholestatic hepatitis. Ann Pharmacother 1995; 29(1):30–2. 345. Chen LK, et al. Ticlopidine-induced hepatitis. Chung Hua i Hsueh Tsa Chih—Chin Med J 2001; 64(1):59–63. 346. Skurnik YD, et al. Ticlopidine-induced cholestatic hepatitis. Ann Pharmacother 2003; 37(3):371–5. 347. Blanco JR, et al. Early hepatopathy induced by ticlopidine. An Med Interna 2001; 18(1):48–9. 348. Berent R, et al. Cholestatic hepatitis as a rare side effect of therapy with ticlopidine. Z Gastroenterol 2000; 38(7):587–91. 349. Tsai MH, et al. Ticlopidine-induced cholestatic hepatitis with anti-nuclear antibody in serum. J Formos Med Assoc 2000; 99(11):866–9. 350. Wu MS, et al. Ticlopidine-induced severe cholestatic hepatitis. Chung Hua i Hsueh Tsa Chih—Chin Med J 2000; 63(8):663–6. 351. Gandolfi A, et al. Ticlopidine-induced acute cholestatic hepatitis. A case report. Recenti Prog Med 2004; 95(2):96–9. 352. Grieco A, et al. Cholestatic hepatitis due to ticlopidine: clinical and histological recovery after drug withdrawal. Case report and review of the literature. Eur J Gastroenterol Hepatol 1998; 10(8):713–5. 353. Ruiz-Valverde P, et al. Ticlopidine-induced granulomatous hepatitis. Ann Pharmacother 1995; 29(6):633–4. 354. Lopez ML, Lopez AM, Feliu RM. Ticlopidine, cholestasis and hepatic nodules. Med Clin 2001; 117(12):477. 355. Martinez Perez-Balsa A, et al. Hepatotoxicity due to ticlopidine. Ann Pharmacother 1998; 32(11):1250–1. 356. Iqbal M, et al. Ticlopidine-induced cholestatic hepatitis: report of three cases and review of the literature. Dig Dis Sci 1998; 43(10):2223–6. 357. Beltran-Robles M, et al. Hepatotoxicity induced by clopidogrel. J Hepatol 2004; 40(3):560–2. 358. Chau TN, et al. Clopidogrel-induced hepatotoxicity after percutaneous coronary stenting. Hong Kong Med J 2005; 11(5):414–6. 359. Ramos JCR, et al. Clopidogrel-induced hepatotoxicity. Med Clin 2003; 120(4):156–7. 360. Zeolla MM, Carson JJ. Successful use of clopidogrel for cerebrovascular accident in a patient with suspected ticlopidine-induced hepatotoxicity. Ann Pharmacother 1999; 33(9):939–41. 361. Polkey MI, Oliver RM, Walker JM. Hepatic dysfunction induced by streptokinase. Am J Gastroenterol 1992; 87(8):1062. 362. Tawata M, et al. A type 2 diabetic patient with liver dysfunction due to human insulin. Diabetes Res Clin Pract 2000; 49(1):17–21. 363. Huang A, Raskin P. Thiazolidinediones and insulin: rationale for use and role of combination therapy in type 2 diabetes mellitus. Treat Endocrinol 2005; 4(4):205–20. 364. Masaki N, et al. Enhancement of carbon tetrachloride-induced liver injury by glucagon and insulin treatment. Res Exp Med 1988; 188(1):27–33. 365. Olsson R, et al. Elevated aminotransferases and alkaline phosphatases in unstable diabetes mellitus without ketoacidosis or hypoglycemia. J Clin Gastroenterol 1989; 11(5):541–5. 366. Ghali P, Lindor KD. Hepatotoxicity of drugs used for treatment of obesity and its comorbidities. Semin Liver Dis 2004; 24(4):389–97. 367. Paice BJ, Paterson KR, Lawson DH. Undesired effects of the sulphonylurea drugs. Adverse Drug React Acute Poisoning Rev 1985; 4(1):23–36. 368. Bolkent S, et al. The morphological and biochemical effects of glibornuride on rat liver in experimental diabetes. Hum Exp Toxicol 2004; 23(5):257–64.

628

Bhardwaj and Chalasani

369. Boyer JF, Fornaris M, Donnet V. Action of tolbutamide on experimental cirrhosis development in the dog. Comptes Rendus des Seances de la Societe de Biologie et de Ses Filiales 1972; 166(2):414–7. 370. Clementsen P, Hansen CL, Hoegholm A. Glipizide induced toxic hepatitis. Ugeskr Laeger 1986; 148(13):771–2. 371. Watanabe S, et al. Two cases of drug induced hepatic injury probably due to glycopyramide. Nippon Shokakibyo Gakkai Zasshi 1978; 75(2):239–45 (author’s transl). 372. Geubel AP, et al. Prolonged cholestasis and disappearance of interlobular bile ducts following chlorpropamide and erythromycin ethylsuccinate. Case of drug interaction? Liver 1988; 8(6):350–3. 373. Rank JM, Olson RC. Reversible cholestatic hepatitis caused by acetohexamide. Gastroenterology 1989; 96(6):1607–8. 374. Nakao NL, et al. A case of chronic liver disease due to tolazamide. Gastroenterology 1985; 89(1):192–5. 375. Zhu XX, et al. Addition of rosiglitazone to existing sulfonylurea treatment in chinese patients with type 2 diabetes and exposure to hepatitis B or C. Diabetes Technol Ther 2003; 5(1):33–42. 376. Rumboldt Z, Bota B. Favorable effects of glibenclamide in a patient exhibiting idiosyncratic hepatotoxic reactions to both chlorpropamide and tolbutamide. Acta Diabetol Latina 1984; 21(4):387–91. 377. Abdullaev N. The effect of chlorpropamide on pathochemical changes during experimental heliotrin hepatitis. Patol Fiziol Eksp Ter 1966; 10(6):49–52. 378. Abdullaev N. The effect of chlorpropamide on liver tissue respiration in experimental chronic toxic hepatitis. Biulleten Eksperimentalnoi Biologii i Meditsiny 1967; 63(1):50–2. 379. Bell RL. Chlorpropamide hepatitis. Ala Med 1984; 53(9):20. 380. Crespo Valades E, et al. Hepatotoxic reaction associated with metformin and chlorpropamide treatment. Rev Clin Esp 1999; 199(2):118–9. 381. Leites SM, Abdulaev N. The influence of chloropropamide on certain biochemical indices of the liver under normal conditions and in experimental toxic hepatitis. Biulleten Eksperimentalnoi Biologii i Meditsiny 1966; 61(3):55–9. 382. Meadow P, Tullio CJ. Glyburide-induced hepatitis. Clin Pharm 1989; 8(7):470. 383. Schneider HL, et al. Chlorpropamide hepatotoxicity: report of a case and review of the literature. Am J Gastroenterol 1984; 79(9):721–4. 384. Grottum KA. Liver damage due to phenylbutazone and chlorpropamide. Tidsskr Nor Laegeforen 1966; 86(16):1096–8. 385. Rigberg LA, Robinson MJ, Espiritu CR. Chlorpropamide-induced granulomas. A probable hypersensitivity reaction in liver and bone marrow. JAMA 1976; 235(4):409–10. 386. Goodman RC, et al. Glyburide-induced hepatitis. Ann Intern Med 1987; 106(6):837–9. 387. Perez-Roldan F, et al. Cholestatic hepatitis caused by glybenclamide in a patient with hepatitis C virus. Rev Esp Enferm Dig 1995; 87(2):174–6. 388. van Basten JP, et al. Glyburide-induced cholestatic hepatitis and liver failure. Case-report and review of the literature. Neth J Med 1992; 40(5–6):305–7. 389. Saw D, et al. Granulomatous hepatitis associated with glyburide. Dig Dis Sci 1996; 41(2):322–5. 390. Byrne JA, et al. The human bile salt export pump: characteristics of substrate specificity and identification of inhibitors. Gastroenterology 2002; 123(5):1649–58. 391. Kemp DC, Brouwer KL. Viability assessment in sandwich-cultured rat hepatocytes after xenobiotic exposure. Toxicol In Vitro 2004; 18(6):869–77. 392. Van Thiel DH, et al. Tolazamide hepatotoxicity. A case report. Gastroenterology 1974; 67(3):506–10. 393. Ananth JV, Ban TA, Lehmann HE. Tolbutamide jundice associated with multiple drug therapy. Can Med Assoc J 1970; 103(11):1194. 394. Wilson RB, Wilson WD. Hepatotoxicity of tolbutamide in dogs. J Am Vet Med Assoc 1970; 156(11):1557–66. 395. Chounta A, et al. Cholestatic liver injury after glimepiride therapy. J Hepatol 2005; 42(6):944–6. 396. Dusoleil A, et al. Glimepiride-induced acute hepatitis. Gastroenterol Clin Biol 1999; 23(10):1096–7. 397. Sitruk V, et al. Acute cholestatic hepatitis induced by glimepiride. Gastroenterol Clin Biol 2000; 24(12):1233–4. 398. Caksen H, et al. Gliclazide-induced hepatitis, hemiplegia and dysphasia in a suicide attempt. J Pediatr Endocrinol 2001; 14(8):1157–9. 399. Dourakis SP, et al. Gliclazide-induced acute hepatitis. Eur J Gastroenterol Hepatol 2000; 12(1):119–21. 400. Chitturi S, et al. Gliclazide-induced acute hepatitis with hypersensitivity features. Dig Dis Sci 2002; 47(5):1107–10. 401. Babich MM, Pike I, Shiffman ML. Metformin-induced acute hepatitis. Am J Med 1998; 104(5):490–2. 402. Barquero Romero J, Perez Miranda M. Metformin-induced cholestatic hepatitis. Gastroenterol Hepatol 2005; 28(4):257–8. 403. Desilets DJ, et al. Cholestatic jaundice associated with the use of metformin. Am J Gastroenterol 2001; 96(7):2257–8.

Cardiovascular and Antidiabetic Medications

629

404. Paterson KR, Paice BJ, Lawson DH. Undesired effects of biguanide therapy. Adverse Drug React Acute Poisoning Rev 1984; 3(3):173–82. 405. Adams MD, Raman P, Judd RL. Comparative effects of englitazone and glyburide on gluconeogenesis and glycolysis in the isolated perfused rat liver. Biochem Pharmacol 1998; 55:1915–20. 406. Chen C. Troglitazone: an antidiabetic agent. Am J Health Syst Pharmacol 1998; 55:905–25. 407. Wang M, et al. Troglitazone, an antidiabetic agent, inhibits cholesterol biosynthesis through a mechanism independent of peroxisome proliferator-activated receptor-gamma. Diabetes 1999; 48:254–60. 408. de la Iglesia FA, et al. Chronic toxicity study of the antidiabetic troglitazone in Wistar rats. Toxicol Sci 1998; 42:50 (abstr. 246). 409. Herman JR, et al. Carcinogenicity study of the antidiabetic troglitazone in Wistar rats. Toxicol Sci 1998; 42:71 (abstr. 351). 410. Herman JR, et al. Subchronic toxicity of the antidiabetic troglitazone in Wistar rats. Fundam Appl Toxicol 1997; 36:273 (abstr. 1387). 411. Breider MA, et al. Troglitazone-induced heart and adipose tissue cell proliferation in mice. Toxicol Pathol 1999; 27:545–52. 412. Nakano N, et al. Role of angiotensin II in the regulation of a novel vascular modulator, hepatocyte growth factor (HGF), in experimental hypertensive rats. Hypertension 1997; 30:1448–54. 413. McGuire EJ, et al. Carcinogenicity study of the antidiabetic troglitazone in B6C3F1 mice. Toxicol Sci 1998; 42:50 (abstr. 247). 414. McGuire EJ, et al. Subchronic toxicity of the antidiabetic troglitazone in B6C3F1 mice. Fundam Appl Toxicol 1997; 36:273 (abstr. 1388). 415. Rothwell CE, et al. 52 week oral toxicity study of troglitazone in cynomolgus monkeys. Fundam Appl Toxicol 1997; 36:273 (abstr. 1386). 416. Okuno A, et al. Troglitazone increases the number of small adipocytes without the change of white adipose tissue mass in obese Zucker rats. J Clin Invest 1998; 101:1354–61. 417. Rothwell NJ, Stock MJ, Tedstone AE. Effects of ciglitazone on energy balance, thermogenesis and brown fat activity in the rat. Mol Cell Endocrinol 1987; 51:253–7. 418. Tafuri SR. Troglitazone enhances differentiation, basal glucose uptake, and Glut1 protein levels in 3T3L1 adipocytes. Endocrinology 1996; 137:1406–12. 419. Tai TAC, et al. Activation of the nuclear receptor peroxisome proliferator-activated receptor gamma promotes brown adipocyte differentiation. J Biol Chem 1996; 271:29909–17. 420. Horikoshi H, Yoshioka T. Troglitazone—a novel antidiabetic drug for treating insulin resistance. Drug Discov Today 1998; 3:79–88. 421. Inoue I, et al. Troglitazone has a scavenging effect on reactive oxygen species. Biochem Biophys Res Commun 1997; 235:113–6. 422. Physicians’ Desk Reference. Montvale, NJ: Medical Economics Company, 1998:2218–21. 423. Graham DJ, Drinkard CR, Shatin D. Incidence of idiopathic acute liver failure and hospitalized liver injury in patients treated with troglitazone. Am J Gastroenterol 2003; 98:175–9. 424. Willman D. Fears grow over delay in removing Rezulin: company’s conduct in development of the drug also comes under fire. Los Angeles Times. March 10, 2000; A18. 425. Watkins PB. Insight into hepatotoxicity: the troglitazone experience. Hepatology 2005; 41:229–30. 426. Graham DJ, et al. Troglitazone-induced liver failure: a case study. Am J Med 2003; 11:299–306. 427. Smith MT. Mechanisms of troglitazone hepatotoxicity. Chem Res Toxicol 2003; 16:679–87. 428. Chojkier M. Troglitazone and liver injury: in search of answers. Hepatology 2005; 41:237–46. 429. de la Iglesia FA, Haskins JR, Feuer G. Hepatotoxicity of cardiovascular and antidiabetic drugs. In: Kaplowitz N, DeLeve LD, eds. Drug Induced Liver Disease. New York: Marcel Dekker, 2003:549–92. 430. Watanabe I, et al. A study to survey susceptible genetic factors responsible for troglitazoneassociated hepatotoxicity in Japanese patients with type 2 diabetes mellitus. Clin Pharmacol Ther 2003; 73:435–55. 431. Brown MN. The thiazolidinediones or “glitazones” a treatment option for type 2 diabetes mellitus. Med Health RI 2000; 83:118–20. 432. Avandia (rosiglitazone maleate) prescribing information. Research Triangle Park, NC: GlaxosmithKline; 2005. 433. Tolman KG, et al. Narrative review: hepatobiliary disease in type 2 diabetes mellitus. Ann Intern Med 2004; 141:946–56. 434. Avigan, M. Characterization of DILI risk from post marketing reports. in FDA/CDER-AASLDPhRMA HepTox Steering Group Meeting January 2006. 435. Lebovitz HE, Kreider M, Freed MI. Evaluation of liver function in type 2 diabetic patients during clinical trials. Evidence that rosiglitazone does not cause hepatic dysfunction. Diabetes Care 2002; 25:815–21. 436. Chalasani N, Teal E, Hall SD. Effect of rosiglitazone on serum liver biochemistries in diabetic patients with normal and elevated baseline liver enzymes. Am J Gastroenterol 2005; 100:1317–21.

630

Bhardwaj and Chalasani

437. Sanyal AJ, et al. A randomized controlled pilot study of pioglitazone and vitamin E versus vitamin E for nonalcoholic steatohepatitis. Clin Gastroenterol Hepatol 2004; 2:1107–15. 438. Neuschwander-Tetri BA, et al. Improved nonalcoholic steatohepatitis after 48 weeks of treatment with the PPAR-gamma ligand rosiglitazone. Hepatology 2003; 38:1008–17. 439. Promrat K, et al. A pilot study of pioglitazone treatment for nonalcoholic steatohepatitis. Hepatology 2004; 39:188–96. 440. Bajaj M, et al. Decreased plasma adiponectin concentrations are closely related to hepatic fat content and hepatic insulin resistance in pioglitazone treated type 2 diabetics. J Clin Endocrinol Metab 2004; 89:200–6. 441. Actos (pioglitazone hydrochloride) prescribing information. Deerfield, IL: Takeda Pharmaceutical; 2005. 442. Deutsch M, Kountouras D, Dourakis SP. Metformin hepatotoxicity. Ann Intern Med 2004; 140(5):W25. 443. Nammour FE, Fayad NF, Peikin SR. Metformin-induced cholestatic hepatitis. Endocr Pract 2003; 9(4):307–9. 444. Swislocki AL, Noth R. Case report. Pseudohepatotoxicity of metformin. Diabetes Care 1998; 21(4):677–8. 445. Andrade RJ, et al. Acarbose-associated hepatotoxicity. Diabetes Care 1998; 21(11):2029–30. 446. Benavente Fernandez A, et al. Acute hepatitis induced by acarbose. Med Clin 2001; 117(8):317–8. 447. Carrascosa M, Pascual F, Aresti S. Acarbose-induced acute severe hepatotoxicity. Lancet 1997; 349(9053):698–9. 448. de la Vega J, et al. Acarbose-induced acute hepatitis. Report of two events in the same patient. Gastroenterol Hepatol 2000; 23(6):282–4. 449. Mennecier D, et al. Acarbose-induced acute hepatitis. Gastroenterol Clin Biol 1999; 23(12):1398–9. 450. Krahenbuhl S. Acarbose and acetaminophen—a dangerous combination? Hepatology 1999; 29(1):285–7. 451. Madonia S, Pietrosi G, Pagliaro L. Acarbose-induced liver injury in an anti-hepatitis C virus positive patient. Dig Liver Dis 2001; 33(7):615–6. 452. Diaz-Gutierrez FL, Ladero JM, Diaz-Rubio M. Acarbose-induced acute hepatitis. Am J Gastroenterol 1998; 93(3):481. 453. Carlson RF. Miglitol and hepatotoxicity in type 2 diabetes mellitus. Am Fam Physician 2000; 62(2):315. 454. Malaisse WJ. Pharmacology of the meglitinide analogs: new treatment options for type 2 diabetes mellitus. Treat Endocrinol 2003; 2(6):401–14. 455. Rosenstock J, et al. Repaglinide versus nateglinide monotherapy: a randomized, multicenter study. Diabetes Care 2004; 27(6):1265–70. 456. Lopez-Garcia F, et al. Cholestatic hepatitis associated with repaglinide. Diabetes Care 2005; 28(3):752–3. 457. Nan DN, et al. Acute hepatotoxicity caused by repaglinide. Ann Intern Med 2004; 141(10):823. 458. Gerich J, et al. PRESERVE-beta: two-year efficacy and safety of initial combination therapy with nateglinide or glyburide plus metformin. Diabetes Care 2005; 28(9):2093–9. 459. Miwa S, et al. Efficacy and safety of once daily gliclazide (20 mg/day) compared with nateglinide. Endocr J 2004; 51(4):393–8.

29

Cancer Chemotherapy Laurie D. DeLeve

Keck School of Medicine, University of Southern California, Los Angeles, California, U.S.A.

INTRODUCTION This chapter is comprised of three major sections. The first section examines the factors that alter liver metabolism and how this can affect efficacy and toxicity from anticancer drugs. The second section will review liver toxicity in hematopoietic stem cell transplantation; this topic warrants its own section, since several of the conditioning regimens for stem cell transplantation have a high incidence of liver toxicity and a high case-fatality rate. The final section is the more traditional review of hepatotoxicity listed for individual anticancer drugs. HEPATIC METABOLISM OF ANTICANCER DRUGS Most anticancer drugs are potentially toxic compounds with a narrow therapeutic index. Thus changes in pharmacokinetics or drug metabolism that effect the disposition of these drugs may easily alter efficacy or cause toxicity. This section will discuss how variability in hepatic metabolism will effect disposition of anticancer drugs and how this may predispose to liver toxicity. Drug Interactions Drug Interactions and Drug Disposition In general, the incidence of adverse drug reactions increases exponentially with the number of drugs prescribed. This is partially because individuals on multiple drugs tend to have more severe underlying illness and partially due to falsely attributing disease symptoms to drug toxicity. However drug interactions contribute to the disproportionate rise in adverse reactions due to polypharmacy. Cancer patients are at particular risk because of the narrow therapeutic index of anticancer drugs, use of anticancer drugs in combination regimens, and many supportive drugs used in this population. Supportive drugs used in patients on antineoplastic regimens include antiemetics, antipyretics, analgesics, antihistamines, antifungals, antivirals, and antibacterial antibiotics, many of which are potential hepatotoxins. Drug interactions that affect drug metabolism by the liver may inhibit detoxification, induce metabolic activation, or inhibit excretion and thereby increase efficacy or toxicity. Conversely, inhibition of metabolic activation or induction of detoxification pathways may diminish efficacy of anticancer drugs. These interactions are more important when a drug is more extensively metabolized and when the affected pathway contributes a large share to overall metabolism. Increased efficacy and/or toxification. Inhibition of detoxification by oxidative metabolism and/or conjugation reactions by concomitant medication may lead to drug toxicity. CYP3A4 is the most abundant P450 in the liver (around 30% of total hepatic P450), it is inducible and it is responsible for the metabolism of over 50% of xenobiotics. Inhibition of CYP3A4 is responsible for many clinically significant interactions. Vinca alkaloids (vincristine, vinblastine, vindesine, and vinorelbine) and docetaxel are mainly detoxified by CYP3A4 catalyzed metabolism and are predominantly excreted into bile. Ketoconazole, itraconazole, and fluconazole—antifungals frequently used in this population— are P450 3A4 inhibitors that should be avoided during Dr. DeLeve is a consultant for Johnson & Johnson, Ono USA, and Wyeth-Ayerst.

632

DeLeve

therapy with these vinca alkaloids or docetaxel, since the unmetabolized parent compound may reach toxic levels. Indeed, there are several case reports of neurotoxicity from the combination of vincristine with itraconazole that have been attributed to inhibition of CYP3A4 by itraconazole. 6-Mercaptopurine, 6-thioguanine, and azathioprine are prodrugs (Fig. 1). When orally administered these thiopurines are subjected to first-pass metabolism by xanthine oxidase in the liver and intestine. When 6-mercaptopurine is given orally, e.g., in maintenance therapy for acute lymphoblastic leukemia, inhibition of xanthine oxidase by allopurinol, inhibits first-pass metabolism in the intestine and liver, increases bioavailability and can precipitate toxicity (1). Allopurinol has little effect on the area under the curve (AUC) when these thiopurines are given intravenously (1). The mechanism of hepatotoxicity of these drugs will be discussed more in-depth in the last section of this chapter. The liver extensively metabolizes paclitaxel: around 90% of the dose is converted to metabolites. Total fecal excretion is approximately 70% of the dose and the largest component by far is 6a-hydroxypaclitaxel, a detoxification product. Given the extensive metabolism, paclitaxel might be expected to be prone to toxicity through interaction with inhibitors of metabolism. In humans, CYP2C8 is responsible for the formation of the major metabolite, 6ahydroxypaclitaxel, whereas CYP3A4 catalyses formation of two hydroxylation products that are minor metabolites in most individuals. Several compounds inhibit paclitaxel metabolism in vitro in human mic rosome studies, but major effects have not been reported clinically. This is likely because of the lack of clinically used drugs that inhibit CYP2C8. Irinotecan is now a component in the chemotherapy of advanced colorectal cancer. It is metabolized by carboxylesterase-2 (2) to SN-38 (7-ethyl-10-hydroxy camptothecin), the putative active metabolite, which is conjugated by uridine diphosphate glucuronosyl transferase 1A1 (3) (UGT-1A1) to the glucuronide, SN-38-glucuronide (Fig. 2). Irinotecan and its metabolites are actively excreted in bile (4) and fecal excretion accounts for elimination of two-thirds of the dose (5). Increased concentrations of SN-38 in the intestine are thought to be the cause of the delayed-onset diarrhea associated with irinotecan, whereas glucuronidation of SN-38 protects against this toxicity. Valproate (6) and ketoconazole (7) inhibit UGT and

6-Thiouric acid

6-Methylmercaptopurine TPMT Azathioprine

GST

Xanthine oxidase 6-Mercaptopurine

Allopurinol –

Thioguanine HPRT

HPRT

Thionosine monophosphate Thioguanosine monophosphate TPMT

Adenosine kinase Thioguanine nucleotides

Methylthionosine monophosphate Methylmercaptopurine nucleotide

TPMT Methylthioguanosine monophosphate

DNA

? DNA

FIGURE 1 Thiopurine metabolism. A simplified schematic of 6-mercaptopurine, 6-thioguanine, and azathioprine metabolism. Thioguanine nucleotides are the presumptive toxic metabolites. Abbreviations: GST, glutathione Stransferase; HPRT, hypoxanthine phosphoribosyltransferase; TPMT, thiopurine S-methyltransferase.

633

Cancer Chemotherapy

Irinotecan Carboxylesterase-2 CYP3A4

SN-38

UGT-1A1

SN-38-glucuronide b-glucuronidase MRP2

Aminopentane carboxylic acid and primary amine metabolites

Biliary excretion

FIGURE 2 Irinotecan metabolism. Irinotecan is metabolized by carboxylesterase-2 to the toxic metabolite, SN-38. SN-38-glucuronide is a detoxified metabolite that can be excreted into bile via ABCC2/cMOAT. Intestinal toxicity may be due to bacterial deconjugation of the glucuronide with regeneration of SN-38.

reduce formation of SN-38-glucuronide experimentally (8). Thus treatment with valproate or ketoconazaole may increase the risk of intestinal toxicity from irinotecan. Induction of metabolic activation by P450 will enhance efficacy and could potentially cause toxicity. Cyclophosphamide requires metabolic activation by CYP3A4/5 at the higher range of therapeutic concentrations and by CYP2C9 at low concentrations to form 4-hydroxycyclophosphamide (9). 4-Hydroxycyclophosphamide is an intermediate metabolite in the formation of phosphoramide mustard, the active metabolite, and acrolein, the metabolite responsible for much of its toxicity. Phenobarbital induces P450 activation of cyclophosphamide, which will increase both efficacy and toxicity. A similar mechanism would apply to ifosfamide. Drug-induced inhibition of biliary excretion may occur at the level of the active transporters on the canalicular pole of the hepatocyte (see transporters in Table 1; mechanical obstruction of the biliary tree is discussed below). P-glycoprotein, the multidrug resistance drug efflux pump that is overexpressed in some cancer cells, is comprised of three transporters at the canalicular pole of the hepatocyte, MDR1, BSEP, and MDR3 (Table 1). These transmembrane-spanning proteins are responsible for biliary excretion of hydrophobic cationic antineoplastic drugs. TABLE 1 Transporters of Chemotherapeutic Agents Transporter a

MDR1/ABCB1

BSEP/ABCB11a MDR3/ABCB4a MRP1/ABCC1b

MRP2/ABCC2/cMOAT

MRP3/ABCC3b MRP4/ABCC4

BCRP/MXR/ABCG2

Chemotherapeutic agents

Comments

Doxorubicine, Paclitaxel, Tamoxifen, Vinblastine Vinblastine Paclitaxel, Vinblastine Anthracyclines (10,11), Camptothecins (10 13) (Irinotecan, SN-38), Epipodophyllotoxins (10,11), Methotrexate, Vinca alkaloids (10,11) Antracyclines (10,11), Epipodopllyotoxins (10,11), Camptothecins (10 14) (Irinotecan, SN-38), Methotrexate, Taxanes, Vinca alkaloids (10,11), Cyclophosphamideglutathione conjugate Vinca alkaloids (15,16), Epipodopllyotoxins (15,16), Methotrexate Camptothecins (17) (Irinotecan, SN-38), cyclic nucleotide transporters (6-thioguanine), (cyclophosphamide and ifosfamide?), methotrexate (10,11) Anthracyclines (18), mitoxantrone (18), camptothecins (18,19), imatinib mesylate (20)

Widely expressed, including canalicular membrane Liver specific, canalicular membrane Liver specific, canalicular Present on basolateral membrane of hepatocyte. GSH depletion decreases efflux Present on canalicular membrane of hepatocyte. GSH depletion decreases efflux

Present on basolateral membrane of hepatocyte GSH depletion decreases efflux and abolishes resistance; not important in liver Canalicular membrane of hepatocytes

Transporters listed in bold are the most significant ones for liver transport. Ref. (21), chapter 6 was the source of information for drugs listed above in column 2 that are not referenced in the table. a MDR1, BSEP, and MDR3 are P-glycoproteins in the liver. b MRP1 and MRP3 are only expressed on the basolateral membrane during special conditions, such as cholestasis (MRP1 and MRP3) or liver regeneration (MRP1) (21). Abbreviations: ABC, ATP-binding cassette family of transporters; BCRP, breast cancer resistance protein; GSH, glutathione; MRP, multidrug resistance associate protein; PgP, P-glycoprotein; SN-38, 7-ethyl-10-hydroxycamptothecin, the active metabolite of irinotecan.

634

DeLeve

Inhibition of P-glycoprotein at the canalicular pole of hepatocytes will block excretion of substrates into bile and increase drug levels. Verapamil and cyclosporin are both inhibitors of P-glycoprotein, but through different mechanisms. Verapamil is a substrate for P-glycoprotein and is a competitive inhibitor of this pump, whereas cyclosporin inhibits transport function by interfering with substrate recognition and ATP hydrolysis (22). Experimental studies have demonstrated that decreased clearance of drugs through inhibition of P-glycoprotein translates clinically into increased AUC (23,24) and an increase in toxicity (24). Examples of this type of drug interaction are the decrease in vincristine clearance in the presence of verapamil (23), of paclitaxel, or etoposide clearance in the presence of Cremophor (a drug solubiliser used in the formulation of paclitaxel) (25,26), and of etoposide (24) or doxorubicin clearance in the presence of cyclosporin [see review (23)]. The effect on biliary excretion is only part of the pharmacokinetic change that occurs with inhibition of P-glycoprotein. Inhibition of P-glycoprotein on renal tubular epithelium will inhibit urinary excretion, analogous to the effect on the hepatocyte. Also, P-glycoprotein pumps drugs from the enterocyte back into the intestinal lumen, so that inhibition of P-glycoprotein may increase intestinal uptake of orally administered drugs. The toxicity due to decreased excretion of the drug needs to be taken into account when pharmacologic inhibition of P-glycoprotein is used to overcome drug resistance in tumor cells that express P-glycoprotein: expression of P-glycoprotein allows the tumor cell to efflux the chemotherapeutic agent, whereas pharmacological inhibition of P-glycoprotein-mediated transport allows intracellular accumulation of drug in cells. Inhibition of P450 or P-glycoprotein in the small intestine may increase bioavailability of orally administered drugs through decreased first-pass metabolism (see previous and below). Decreased efficacy and/or toxicity. Fluconazole causes a significant reduction in plasma clearance of cyclophosphamide due to inhibition of P450 activation that is presumably accompanied by reduced formation of 4-hydroxycyclophosphamide. The effect of this interaction on clinical efficacy has not been established (27). A well-documented case report describes prevention of hepatotoxicity with loss of chemotherapeutic efficacy of fotemustine by N-acetylcysteine therapy given with the intention of preventing recurrent hepatotoxicity (28). This case report is an excellent example that when the mechanism is the same for the wanted chemotherapeutic effect and the unwanted adverse reaction, potential antidotes for the toxicity can backfire. Antioxidants widely used by cancer patients (29,30), are described in the literature as a safe complementary medicine (31–33), and are being incorporated into clinical trials (31,32). Interference with the chemotherapeutic effect should be a consideration when including antioxidants in a chemotherapeutic regimen. Induction of detoxification by P450 may decrease efficacy. Vincristine is partially metabolized by CYP3A4. Concomitant use of carbamezepine or phenytoin increases vincristine clearance (34). Docetaxel is detoxified by CYP3A4. In liver microsomes prepared from patients who had been treated with phenobarbital, docetaxel hydroxylation is induced (35). Paclitaxel induces CYP3A4 and P-glycoprotein through activation of the nuclear receptor PXR (36,37). This will induce its own metabolism by CYP3A4 to 6a-hydroxypaclitaxel, but will alter the CYP3A4 metabolism of other chemotherapeutic agents used in combination regimens and of ancillary medications used in this population. Cyclophosphamide induces CYP3A4 and CYP2B and activates PXR. However the dose response for PXR activation did not seem to be sufficient to account for the extent of induction of CYP3A4 and CYP2B (38). CYP3A4 and CYP2B6 catalyze metabolism of cyclophosphamide to 4-hydroxycyclophosphamide, the alkylating metabolite. Thus analogous to paclitaxel (see previous paragraph), cyclophosphamide will induce its own metabolism, but also that of other drugs. Experimentally, phenobarbital pretreatment reduces the AUC of irinotecan and the active metabolite, SN-38, with an increase in the SN-38-glucuronide (Fig. 2) (8). This is most likely due to induction of CYP3A4 and UGT. The relevance of this is demonstrated by a comparison of irinotecan pharmacokinetics in patients with colorectal cancer versus those with malignant glioma (39). Ninety-one percent of the patients with malignant glioma were on phenytoin, carbamazepine, or phenobarbital. Enhanced irinotecan clearance and reduced concentrations of SN-38 and SN-38-glucuronide are consistent with induction of CYP3A4 metabolism of irinotecan. The low incidence of severe toxicity in the malignant glioma patients in conjunction

Cancer Chemotherapy

635

with the low plasma concentrations of irinotecan suggests that treatment with anticonvulsants has a significant effect on irinotecan disposition. Drug Interactions and Liver Toxicity Imatinib (GleevecTM) is predominantly metabolized by CYP3A4 with a minor contribution to the metabolism by CYP1A2, CYP2D6, CYP2C9, and CYP2C19. Imatinib also inhibits CYP3A4 and CYP2D6 (40,41). CYP3A4 inducers such as rifampin and St. Johns’ wort reduce imatinib AUC by 30% to 50% (42,43), whereas the CYP3A4 inhibitor ketoconazole increases the AUC of imatinib by 40% (44). Grade III to IV (Table 4) elevations of serum transaminases or bilirubin occur in 2.6% of patients on imatinib alone (40), but in 53% of patients on imatinib in combination chemotherapy regimens (45). There have been five reported cases of liver injury (46–49) and one report of liver failure, the latter in a patient on chronic imatinib who received acetaminophen (40). Liver injury at doses similar to the clinical dose in the dog in preclinical testing and reversible bilirubin and serum transaminases in a patient given an accidental overdosage (40) suggest that the therapeutic index of imatinib mesylate may be narrow. The significant impact of CYP3A4 inhibitors on imatinib AUC and the marked increase in liver test abnormalities when imatinib is combined with conventional chemotherapy suggest that drug interactions may play a significant role in imatinib hepatotoxicity. Patients treated with combination chemotherapy commonly develop mild to moderate transient elevations of liver tests, although the individual drugs have little or no risk of liver injury. Interactions of drugs may lead to such abnormalities. For example, a large retrospective study of patients with breast cancer treated with cyclophosphamide, doxorubicin, and 5-FU reported liver test abnormalities in around 85% of patients without known liver metastases (50). The abnormal liver tests did not require discontinuation of the treatment regimen and 90% of patients had normal liver tests one year after discontinuing the therapy. Each of these drugs individually has a low incidence of liver test abnormalities at conventional doses, suggesting an interaction between the drugs in this regimen. The underlying mechanism for these interactions is unknown. Similarly, although the individual drugs rarely cause toxicity, some conditioning regimens for stem cell transplantation have a high incidence of hepatic venoocclusive disease. This is likely due to additive or synergistic toxicity of the drugs, as will be discussed below. Multidrug resistance associated proteins (MRP) MRP1 and MRP2 cotransport many of their substrates with glutathione. Depletion of glutathione blocks transport of MRP1/ABCC1substrates doxorubicin, daunorubicin, vincristine, and VP-16, thereby abolishing resistance in cancer cells (51) and increasing unwanted toxicity in non-cancerous cells. Clinical examples of increased toxicity of doxorubicin, daunorubicine, vincristine, and VP-16 due to glutathione depletion have not been reported. However glutathione depletion is not an uncommon consequence of metabolism of chemotherapeutic agents, so that a combination of a glutathionedepleting chemotherapeutic drug with one of the anthracyclines or vinca alkaloids might indeed have greater toxicity and/or chemotherapeutic efficacy. Underlying Liver Disease Underlying Liver Disease and Hepatic Metabolism In general, underlying liver disease may affect hepatic metabolism. A special consideration in this setting is the involvement of the liver by the underlying neoplasm. Infiltrative liver disease by primary or metastatic tumor may lead to liver dysfunction through several mechanisms. Extensive replacement of normal liver tissue can lead to a decrease in metabolic capacity. Impairment of biliary drug excretion may occur due to compression of intrahepatic bile ductules or the extrahepatic biliary tree. Tumor invasion may compromise normal hepatic blood flow and impair clearance of drugs by the liver. Given the narrow therapeutic range of most anticancer drugs, there is a heightened awareness of the need to alter dosing when drug clearance may be impaired. Attempts to modify dosing of anticancer drugs based on changes in a single liver test, regardless of the type of underlying liver disease, have lead to empiric decisions about dose reduction of anticancer drugs that have not always been confirmed by experimental data. A better approach, when

636

DeLeve

possible, is to correlate abnormal disposition of a drug to the severity of a specific type of liver disease, rather than to changes of a single liver test independent of the underlying pathology. Some clinical research has correlated clearance of anticancer drugs with tests of liver function. Both antipyrine and lidocaine are metabolized by P450 and are measures of liver function, with the distinction that the former is flow-independent and the latter flow-dependent. Similarly, the erythromycin breath test is a measure of hepatic CYP3A4 activity that has been used in clinical research to predict toxicity or to correlate liver function with drug disposition. A limitation in the oncology literature is that patients with liver test abnormalities are often excluded from the early clinical trials of new anticancer drugs. Thus the disposition of a drug in patients with liver disease may not be known until phase III clinical trials are completed and the drug has been made more widely available. The tubulin acting drugs-vincristine, vinblastine, vindesine, vinorelbine, paclitaxel, and docetaxel are predominantly detoxified in the liver and mainly excreted into bile. The vinca alkaloids and docetaxel are mainly metabolized by CYP3A4 (52–54), whereas paclitaxel is predominantly metabolized by CYP2C8 and, to a lesser degree, by CYP3A4 (55). Cancer patients with underlying liver disease may have decreased clearance of these drugs, either due to decreased hepatic function or to decreased biliary excretion. For example, patients with extensive liver metastases have reduced vinorelbine clearance and clearance of the drug correlates with lidocaine clearance (56). Ondansetron, a 5-hydroxytryptamine receptor 3 antagonist used for nausea, is extensively metabolized by P450. The major metabolites are formed by CYP1A2, CYP1A1, and CYP2D6. First-pass metabolism of orally administered ondansetron is impaired in cirrhotic patients, with a significant increase in bioavailability (57). AUC after intravenous administration was also significantly higher than in healthy controls. Ondansetron clearance was closely correlated to antipyrine clearance. There are conflicting reports as to whether systemic concentrations of 5-hydroxytryptamine receptor 3 antagonists correlate with effect, so that it is unclear whether changes in clearance will alter the therapeutic response. Underlying Liver Disease and Liver Toxicity Pre-existent liver disease is the main risk factor for progressive liver disease due to methotrexate. Portal fibrosis in the pretreatment liver biopsy or risk factors for steatohepatitis, such as chronic alcohol abuse, obesity, and diabetes mellitus, predispose to progressive liver disease with chronic low-dose methotrexate. In children treated with methotrexate and/or 6-mercaptopurine for acute lymphoblastic leukemia, viral hepatitis is a risk factor for elevated serum transaminases during therapy, but there appears to be little if any increased risk of persistent elevation of serum aminotransferase (58–60). It is unclear whether treatment with these drugs adds to the risk of hepatitis C-related fibrosis and/or cirrhosis in children with acute lymphoblastic leukemia. In a large cohort of children with childhood leukemia treated with unspecified regimens of chemotherapy, none of the patients had developed evidence of severe chronic liver disease 13 to 27 years after the estimated time of infection (58). Hepatitis C is a risk factor for hepatotoxicity for some drugs (61–65). Prevalence of hepatitis C exposure is 1.8% or 3.9 million individuals, nationwide. Sixty-five percent of these individuals are between 30 and 49 years of age, with a prevalence in this age group of 3% to 3.9% (66). As this cohort of individuals’ ages, the number of cancer patients with concomitant hepatitis C will rise. Thus liver injury due to the interaction between anticancer drugs and hepatitis C is likely to increase in the coming decades. Withdrawal of chemotherapy can lead to reactivation of hepatitis B in carriers and exacerbation of chronic active hepatitis B. The presumptive mechanism is increased viral synthesis during immunosuppression with consequent hepatocyte infection. Withdrawal of chemotherapeutic or immunosuppressive drugs is accompanied by restoration of immune function with rapid destruction of infected hepatocytes. The course of hepatitis in such patients can be fulminant and there is a high incidence of liver failure and death. This has been most commonly described for patients with hepatitis B and hematological malignancies, but has also

Cancer Chemotherapy

637

been described for hepatitis C and for solid tumors. Reactivation of hepatitis is much less common in patients who receive steroid-free chemotherapy (67). Age Age and Hepatic Metabolism With the aging of the population, the median age for cancer is now 70 years. This has stimulated an interest in the effect of aging on disposition of and response to anticancer drugs in the elderly. Chapter 14 reviews ageing and its effect on hepatic metabolism and drug-induced liver disease. In univariate analysis, 5-fluorouracil clearance is significantly reduced with age (68) and age is an independent risk factor for 5-fluorouracil toxicity (69). 5-Fluorouracil is metabolized by hepatic dihydropyrimidine dehydrogenase (DPD), but hepatic DPD activity does not change with age (70). Mitomycin is primarily metabolized in the liver and excreted in the bile (71). AUC of mitomycin increases with age in patients with normal hepatic, renal, cardiac, and bone marrow function (72). Epirubicin is a stereoisomer of doxorubicin that is predominantly cleared by the liver. Epirubicin clearance decreases with age in females, but the study that demonstrated this did not have sufficient males to determine whether this applied to males as well (73). Daunorubicin clearance decreases and AUC and cardiotoxicity increase in rats with age (74,75). The mechanism for the age-related changes in pharmacokinetics of 5-fluorouracil, mitomycin, epirubicin, and daunorubicin has not been defined, but may relate to changes in the liver described in Chapter 14. Age and Liver Toxicity There is an increased incidence of sinusoidal obstruction syndrome (SOS) in patients under 36 months of age treated with actinomycin-D, cyclophosphamide, and vincristine for rhabdomyosarcoma (76). The drugs in this combination that have been linked to SOS are actinomycin-D and cyclophosphamide. It is sometimes stated that age is a risk factor for methotrexate toxicity. However age has not been found to be an independent risk factor for methotrexate-induced aminotransferase elevation (77). It remains to be established whether age per se predisposes to fibrosis in chronic methotrexate therapy. However the age-related decline in renal function is a risk factor for methotrexate hepatotoxicity. Gender Anthracycline metabolism occurs mainly in the liver. Females have significantly lower clearance of doxorubicin and epirubicin than males (73,78). Females also have an incidence of cardiotoxicity that is twice as high as males (79) and lower clearance may play a role in this. This increased risk of cardiac dysfunction in females is already apparent in childhood (80,81). Females have increased risk of toxicity from 5-fluorouracil (69,82,83). This is consistent with the finding that 5-fluoruracil clearance is significantly lower in females than in males (84). However current data have not determined why clearance is lower in females. 5-Fluorouracil is metabolized by hepatic—DPD and low levels of DPD or DPD deficiency (see Genetic Predisposition) predispose to decreased 5-fluorouracil clearance and toxicity. Interestingly, females with 5-fluorouracil toxicity are more likely to have low DPD levels than males with toxicity (83). Thus one would speculate that the increased incidence of toxicity and the decreased clearance in females is due to lower DPD levels in the female population. However two prospective studies, i.e., not in patients with 5-fluorouracil toxicity, did not find an effect of gender on DPD activity in peripheral blood mononuclear cells and liver (70,85). Hormonal status does not appear to affect DPD activity and the gene for DPD is on chromosome 1, an autosomal chromosome. It has been observed that patients with breast cancer have lower DPD activity than healthy controls (86), so that underlying disease or nutritional status may be confounders in the studies done in patients with 5-fluorouracil toxicity. Clearly more studies are needed to reconcile these findings.

638

DeLeve

Nutrition Nutrition and Hepatic Metabolism Undernourishment is common in patients with advanced cancer. Protein-calorie malnutrition or a low daily intake of protein can decrease oxidative metabolism by 20% to 40% (87). Protein depletion also causes a significant decrease in hepatic DPD activity in rats that is associated with significantly decreased hepatic metabolism and clearance of 5-fluorouracil and increased morbidity and mortality (88). Doxorubicin clearance is decreased and AUC is increased in rabbits fed a low-protein diet (89). Nutrition and Liver Toxicity Protein malnourishment is a risk factor in the human epidemics of SOS (venoocclusive disease) due to pyrrolizidine alkaloids. However it has not been identified as a risk factor for SOS due to high dose chemotherapy for hematopoietic stem cell transplantation. However protein deprivation depletes hepatic glutathione levels and experimental data suggest that decreased hepatic glutathione levels may predispose to SOS (90,91). A low-protein diet also increases liver toxicity due to irradiation in rats (92). Genetic Polymorphisms The reader is referred to a review on the impact of metabolic polymorphisms on the efficacy and toxicity of various anticancer drugs (93). The therapeutic effect of 5-fluorouracil is through formation of nucleotides that block normal nucleic acid formation. This is balanced by catabolism by DPD in the liver. More than 85% of 5-fluorouracil is broken down by DPD and DPD activity is therefore a major determinant of 5-fluorouracil activity and toxicity. DPD enzyme activity follows a Gaussian distribution with up to a sixfold inter individual variation. In addition to the normal variation of DPD activity, there are also mutations in the DPD gene that can lead to DPD deficiency, which occurs in less than 3% of the population. Low DPD levels and DPD deficiency reduce 5-fluorouracil clearance and can lead to severe or life-threatening toxicity. There is a relatively weak correlation between DPD activity in peripheral blood mononuclear cells and in the liver (85), yet low DPD activity in peripheral blood mononuclear cells is a strong predictor of reduced 5-fluorouracil clearance (68). As described earlier and in Figure 2, irinotecan is metabolized by carboxyesterase-2 to the active metabolite, SN-38. SN-38 is detoxified to SN-38-glucuronide by UGT-1A1 and UGT-1A9. Both SN-38 and the detoxified conjugate, SN-38-glucuronide, are actively excreted into bile (4). The limiting side effects of irinotecan, intestinal toxicity, and neutropenia, are due to unconjugated SN-38. Genetic variants in UGT1A that alter glucuronidation kinetics are determinants of efficacy and toxicity of SN-38 (3,94–97). The polymorphism associated with Gilbert’s syndrome, UGT1A1*28, has a longer TATAA element in the promoter region, which is associated with decreased transcription and consequent decreased glucuronidation (98). The UGT1A1*28 polymorphism was the first one linked to irinotecan toxicity (99). Genetic variants linked to altered irinotecan efficacy and toxicity include polymorphisms in hepatic UGT1A1 and UGT1A9 and extrahepatic UGT1A7 (95). For the Caucasian population functional variants in UGT1A9 are rare and therefore not likely to be clinically significant in determining risk of irinotecan, whereas the UGT1A1 variants seem to be the most important ones (100). In the Japanese and Korean populations, polymorphisms in both UGT1A1 and UGT1A9 may be relevant to SN-38 toxicity (101,102). LIVER TOXICITY AND HEMATOPOIETIC STEM CELL TRANSPLANTATION Hematopoietic stem cell transplantation, the new phrase for what used to be referred to as bone marrow transplantation, is commonly complicated by drug-induced and non-drug related forms of liver disease. The drug-induced complications include hepatic SOS, the new name for what used to be referred to as hepatic venoocclusive disease, and nodular regenerative hyperplasia. These need to be differentiated from other liver diseases frequently seen in this

Cancer Chemotherapy

639

setting, notably graft-versus-host disease, viral hepatitis, fungal liver disease, tumor infiltration of the liver, cholestasis of sepsis, and liver injury due to total parenteral nutrition [see recent review (103)]. This section will provide an extensive discussion of hepatic SOS, since it has the highest mortality of any chemotherapy-induced liver disease. Hepatic Sinusoidal Obstruction Syndrome (Venoocclusive Disease) SOS or “bush tea disease” was first recognized in humans in the early 20th century as a complication of pyrrolizidine alkaloids ingested as herbal teas or contaminating the food supply (104). Sporadic cases have been described in patients treated with a wide variety of antineoplastic drugs at conventional doses. At present the most common cause of SOS in North America and Western Europe is the preparative regimen used for hematopoietic stem cell transplantation. In hematopoietic stem cell transplantation, SOS is caused by synergistic toxicity between the drugs used in high-dose combination chemotherapy or high-dose chemotherapy plus total body irradiation, the so-called conditioning regimen in transplantation. In studies that included more than 100 patients transplanted to treat malignancies and non-malignant conditions, the incidence of SOS has varied between 1% and 54%. The incidence is generally around 20% to 40% for the more toxic myelosuppressive regimens. The wide range in risk of SOS from myelosuppressive regimens between transplant units is largely due to differences in patient selection criteria, choice of chemotherapy/ irradiation regimen, and criteria to diagnose SOS. In contrast to the myelosuppressive regimens, non-myeloablative regimens have little or no risk for SOS. The incidence of SOS is dropping due to the increasing use of nonmyeloablative regimens, avoidance of cyclophosphamidecontaining regimens, lower doses of total body irradiation, exclusion of patients with hepatitis C, and avoidance of norethisterone, which increases the risk of SOS. Diagnosis of SOS SOS presents clinically with tender hepatomegaly, weight gain, and jaundice. The diagnosis of SOS in the setting of stem cell transplantation with a myelosuppressive regimen is based on these clinical features and the exclusion of other likely causes. Confounding diagnoses include (hyper) acute graft-versus-host disease, cholestatic jaundice due to sepsis, fluid overload secondary to cardiac or renal failure, drug-induced cholestasis, viral or fungal infections of the liver, and tumor infiltration. The diagnosis may be more difficult in a patient with more than one of these causes of liver injury and fluid retention and it may be difficult to determine which is the main cause of illness. Ten to twenty percent of patients with liver injury in the first 20 days after the stem cell transplantation may not be diagnosed with certainty based on signs and symptoms (105,106). Investigators in Seattle and Baltimore have published diagnostic criteria for research purposes (69,70). The Seattle criteria require two of three findings occurring within 20 days of transplantation: bilirubin O2 mg/dl, hepatomegaly or right upper quadrant pain of liver origin, and greater than 2% weight gain due to fluid accumulation. The temporal criterion of signs and symptoms occurring within the first 20 days was developed for cyclophosphamidecontaining regimens and this may not apply for regimens without cyclophosphamide. The Baltimore criteria require hyperbilirubinemia plus two of three other findings: bilirubin O2 mg/dl (usually painful), hepatomegaly, greater than 5% weight gain, and ascites. Retrospective comparison of these criteria found that more patients fulfilled the Seattle criteria and that the Baltimore criteria identified a sicker population (71). Although the Baltimore criteria identify a more clinically relevant population, the patients identified by these criteria may be further along in their course by the time they fulfill the criteria. SOS may be classified as mild, moderate, or severe. Mild disease is clinically apparent but resolves without therapy; moderate disease requires diuretics or pain medication, but resolves completely; and severe disease requires treatment but does not resolve before death or day 100. Published graphs derived from retrospective data allow prediction of severity of disease for patients treated with regimens that contain cyclophosphamide (72).

640

DeLeve

Diagnosis is usually based on the clinical diagnostic criteria above. Ultrasound or other imaging can confirm hepatomegaly and ascites, may exclude tumor infiltration of the hepatic parenchyma and vasculature, and will detect biliary tract disease from benign or malignant causes. However there is no single diagnostic feature of SOS (107–109). Gray scale or color Doppler ultrasonography features suggestive of SOS include reversal of portal venous flow, attenuation of hepatic venous flow, gallbladder wall edema, and perhaps increased resistive indices of hepatic artery flow. The most useful additional diagnostic tool is transvenous liver biopsy and this may be utilized for moderately or severely ill patients with an uncertain diagnosis. Given the thrombocytopenia, coagulopathy, and ascites that may be present in this population, percutaneous liver biopsy is often contraindicated. Transvenous liver biopsy also offers the advantage that a hepatic venous pressure gradient can be measured. In stem cell transplantation patients, hepatic venous pressure gradient higher than 10 mmHg has a 90% specificity for SOS and a positive predictive value of greater than 85% (110). The complication rate of transvenous liver biopsy after stem cell transplantation is 7% to 18% with 0% to 3% fatalities (110,111). Drugs that Cause SOS Anticancer drugs may cause SOS at conventional doses, but the risk is substantially higher at the high doses used for stem cell transplantation. Even at high doses, the components of the conditioning regimen by themselves are not particular hepatotoxic, but exhibit toxicity in combination regimens. Thus high-dose cyclophosphamide by itself rarely causes SOS (112), high-dose busulfan did not cause SOS in the small number of patients treated with it as a single agent (113), melphalan has little or no effect on liver tests (114–116) and total body irradiation used alone within the dose range used in stem cell transplantation is not hepatotoxic (117). The highest incidence of SOS occurs with combination regimens such as cyclophosphamide-total body irradiation, busulfan-cyclophosphamide, BCNU-cyclophosphamideetoposide, and carboplatin-cyclophosphamide-BCNU. Cyclophosphamide, a common element in these regimens, is particularly toxic to sinusoidal endothelial cells in vitro (90). In contrast, patients treated with regimens combining busulfan and melphalan have a much lower incidence of SOS compared to the cyclophosphamide containing regimens listed above (118,119). It has been shown in several studies that busulfan toxicity and efficacy benefit from dosage adjustment through therapeutic drug monitoring (120). Similarly, SOS associated with high-dose cyclophosphamide may be more common in individuals with higher concentrations of cyclophosphamide metabolites (121,122), so that therapeutic monitoring may prove to be of value. Lower doses of total body irradiation decrease the risk of SOS, but increase the risk of leukemic relapse. Gemtuzumab ozogamicin (Mylotargw) is used for acute myeloid leukemia and may cause SOS. In patients who relapse after gemtuzumab ozogamicin and go on to stem cell transplantation or in patients who relapse after stem cell transplantation and then receive treatment with gemtuzumab ozogamicin, the risk of SOS can be substantial (123,124). In a small series of patients who received gemtuzumab ozogamicin followed by stem cell transplantation, the risk of SOS was 90% when the interval between treatments was less than 3.5 months (124). Thus the major risk for gemtuzumab ozogamicin seems to be in the population who have received two myelosupressive insults within a short time frame. Mechanism of Disease Two clinical features provide clues to the mechanisms involved in SOS. First, in other intrinsic liver diseases parenchymal disease precedes the development of portal hypertension. However in SOS, the signs and symptoms of portal hypertension precede evidence of parenchymal damage. In SOS, disruption of the liver circulation is the cause and not the consequence of the parenchymal disease. Second, venoocclusive lesions of the hepatic veins are not essential to the development of the clinical picture: 45% of patients with mild or moderate disease and 25% of patients with severe SOS did not have occluded hepatic venules at autopsy (125). In-depth examination of the changes in the experimental model (see below) has established that the

Cancer Chemotherapy

641

essential change occurs at the level of the sinusoid. Occlusion of central veins is associated with more severe disease and the development of ascites (125,126), which suggests that venoocclusive lesions may add to the impairment of the circulation that occurs at the level of the sinusoid. The original name accorded to this disease, hepatic venoocclusive disease, reflects the fact that the venoocclusive lesion is easily recognized on light microscopy. The more recent recognition that involvement of the vein is not essential has lead to the name change from hepatic venoocclusive disease to hepatic SOS. Experimental studies have confirmed that changes in the hepatic sinusoid are the earliest changes in SOS. In vitro studies have shown that sinusoidal endothelial cells are more susceptible than hepatocytes to drugs that cause SOS (90,127,128). This is consistent with studies done in the rat using monocrotaline, a pyrrolizidine alkaloid that is one of the beststudied toxins involved in SOS. In this model the first morphological change noted by electron microscopy is loss of sinusoidal endothelial cell fenestration and the appearance of gaps in the sinusoidal endothelial cell barrier (129). The gaps increase in size over time and at the same time there is a decrease in the number of Kupffer cells and loss of venous endothelium. Studies with in vivo microscopy and confirmation by electron microscopy have shown that blood cells begin to track under sinusoidal endothelial cells that round up. The accumulation of blood in the space of Disse dissects off the sinusoidal lining, which creates an embolus of sinusoidal lining cells downstream that obstructs flow (130). By the time hepatocyte necrosis is observed, there is extensive sloughing of the sinusoidal lining, i.e., Kupffer cells, sinusoidal endothelial cells, and stellate cells. At this time point there is also a significant influx of monocytes within the sinusoids, which exacerbates the obstruction of sinusoidal flow by the embolized sinusoidal lining cells. These studies demonstrate that rounding up or swelling of sinusoidal endothelial cells is the initiating event in this experimental model of SOS and that this leads to dissection and embolization of sinusoidal lining cells that block the microcirculation. The SOS-inducing drugs and toxins examined to-date all profoundly deplete sinusoidal endothelial cell glutathione prior to cell death and support of sinusoidal endothelial cell glutathione will prevent cell death (90,127,128). A continuous infusion of glutathione or N-acetylcysteine into the portal vein prevents the rounding up of the sinusoidal endothelial cells and the subsequent events that lead to the development of SOS in the monocrotaline model (91). If the glutathione infusion is discontinued several days after monocrotaline has been eliminated, full-blown SOS develops rapidly. Although glutathione protection may be (partially) by preventing profound glutathione depletion, another mechanism for protection must be invoked to explain glutathione protection several days after monocrotaline has been eliminated. Since full-blown SOS in this model normally takes 72 hours to develop, the accelerated development of SOS within 24 hours after discontinuation of glutathione indicates that glutathione is suppressing a persistent change in the sinusoid (see below). One possible explanation for the rounding up of the sinusoidal endothelial cells may be increased activity of matrix metalloproteinases (MMPs) that allows the cells to let loose from the extracellular matrix in the space of Disse. In the experimental model, de novo synthesis of MMP-9 (gelatinase B) and increased MMP-9 activity occur 12 hours after monocrotaline, which coincides with rounding up of the sinusoidal endothelial cells (SEC) (131). The increased synthesis of MMP-9 is due to toxin-induced depolymerization of F-actin in the sinusoidal endothelial cells. Inhibition of MMP activity completely prevents SOS. MMP expression and activity are regulated by redox status and can be suppressed by glutathione and N-acetylcysteine (132–135). Thus the protective effect of glutathione and N-acetylcysteine, particularly the protection that occurs after the toxic drug is eliminated (see previous paragraph), may be (partially) due to inhibition of MMP activity. An additional biochemical change that has been observed experimentally in SOS relates to nitric oxide. In the in vivo model hepatic vein, nitric oxide decreases in parallel with the loss of Kupffer cells and sinusoidal endothelial cells, the major cellular sources of nitric oxide in the liver (136). Manipulations of nitric oxide production experimentally also suggest that decreased nitric oxide contributes to the development of SOS (136). Analogous to findings with glutathione, continuous infusion of a nitric oxide pro-drug with normalization of hepatic nitric oxide levels prevents upregulation of MMP-9, the rounding up of the sinusoidal endothelial cells,

642

DeLeve

the stripping of the sinusoidal lining cells, and the development of SOS (136), further supporting the importance of the sinusoidal endothelial cell in the pathogenesis of this disease. Although the disease is defined as a nonthrombotic obstruction of flow, the issue of clotting has been a recurring topic of research interest. The reader is referred to an in-depth review of this topic (137). Many of the observed changes interpreted to indicate a procoagulant state may have been sequelae of functional impairment and damage to the liver in SOS: hepatic dysfunction with decreased synthesis of anticoagulants (protein C, FVII, and ATIII), endothelial damage with release of membrane and intercellular proteins (vWF, FVIII, PAI-1, and tPA) and increased synthesis of acute phase reactants (fibrinogen). Immunohistochemical studies in livers of patients with SOS have detected Factor VIII and fibrinogen in the wall of the central veins, but not in the sinusoids or vascular lumen (138). Platelet glycoproteins could not be detected by immunostaining in the livers of patients with SOS (138). Electron microscopy of pyrrolizidine alkaloid-induced SOS in humans did not detect clotting (139). Sequential observations during the development of SOS in the experimental model by in vivo microscopy and electron microscopy have not demonstrated any evidence of clotting (129). Thus a role for coagulation has not been ruled out, but current studies do not provide evidence to support a role for activation of coagulation and local excess fibrin production as essential elements of SOS. Nodular Regenerative Hyperplasia In two case series, liver biopsy or autopsy material from patients within 100 days of stem cell transplantation has shown nodular regenerative hyperplasia in 8% and 23% of cases, respectively (125,140). Although the symptoms of nodular regenerative hyperplasia, i.e., hepatomegaly, ascites, and mild elevations of serum bilirubin, are the same symptoms used to diagnose SOS, the time of onset of symptoms is usually much later (103). The lesions may or may not be detectable with diagnostic imaging and the features are non-diagnostic. Doublespiral computerized tomography may aid in differentiating nodular regenerative hyperplasia from hepatocellular carcinoma. Based on studies of liver histology it has been postulated that nodular regenerative hyperplasia is due to local variation in blood flow within the liver: impairment of sinusoidal perfusion leading to atrophy with compensatory regeneration in adjacent areas (141–144). If this is true, heterogeneity of flow in the microcirculation could be due to impaired circulation at either the venular or sinusoidal level. Semiquantitative evaluation of the histology in a large autopsy study of patients with nodular regenerative hyperplasia showed that the most common vascular abnormality was obliteration of portal veins (145). Changes at the level of the sinusoids were demonstrated by ultrastructural studies of biopsies from three renal transplant patients who developed nodular regenerative hyperplasia, SOS, sinusoidal fibrosis, and/or peliosis hepatis due to azathioprine (146). This study demonstrated damage to and loss of sinusoidal endothelial cells, consistent with the selective toxicity of azathioprine for sinusoidal endothelial cells noted by in vitro studies (128). Since damage to sinusoidal endothelial cells is also the postulated mechanism of injury for SOS, it is not surprising that chemotherapy that causes SOS would also cause nodular regenerative hyperplasia. Cyclosporine-Induced Cholestasis Clinically significant liver toxicity from cyclosporine is uncommon. The most common form is bland cholestasis. Cyclosporine impairs canalicular function by inhibiting several of the canalicular transporters, with impairment of both bile acid-dependent and bile acid-independent bile flow (147–156). HEPATOTOXICITY BY ANTICANCER THERAPY Diagnosis Mild or moderate transient elevations of liver tests without clinical toxicity are common in patients treated with anticancer therapy and these may often be ignored. However when clinical toxicity occurs, early diagnosis and discontinuation of the offending drug from the

Cancer Chemotherapy

643

TABLE 2 Causes of Abnormal Liver Tests in Cancer Patients Toxicity of anticancer drugs Toxicity of supportive medications Drug interactions Radiation-induced liver disease Tumor infiltration of the liver Budd-Chiari syndrome (hypercoagulable state, tumor obstruction) Paraneoplastic syndrome (Stauffer’s syndrome) Graft-versus-host disease (stem cell transplantation) Total parenteral nutrition Viral hepatitis Fungal liver disease Sepsis Hemolysis Congestive hepatopathy

anticancer regimen is imperative. The diagnosis of drug-induced liver toxicity is often difficult in any clinical setting. The difficulty of diagnosis in cancer patients is compounded by a variety of other causes of liver disease related to the underlying cancer or cancer therapy (Table 2). There are systematic approaches to the diagnosis of drug-induced liver disease (157,158), but these do not lend themselves to drugs given cyclically in chemotherapy regimens. As in any setting, diagnosis is based on knowledge of the type of liver injury previously observed with a particular drug, the temporal relationship between drug exposure and evidence of liver toxicity, recurrent or exacerbated response upon re-exposure, histological appearance if a biopsy is warranted and exclusion of competing causes of liver disease. Hepatotoxicity Due to Specific Anticancer Drugs The majority of anticancer drugs in use today were first tested fairly long ago, many in the 1950 to 1970s. Much of our information on the type and frequency of liver injury is derived from an era prior to hepatitis C testing and in many cases prior to testing for hepatitis B or even hepatitis A. A good example of this is methotrexate. In a frequently cited report from 1960, there was a very significant increase in the percentage of children with fibrosis noted on autopsy after aminopterin and methotrexate were introduced in 1948 (see details later in this section). The extremely high incidence of liver injury seen in that era is not observed in the high-dose regimens today. Given the liberal use of transfusions in a period prior to hepatitis testing, viral hepatitis was likely a major confounder. During much of the period when many of the currently used anticancer drugs were first tested, liver imaging modalities were often inadequate to rule out occult involvement of the liver by tumor. Even current imaging modalities may not always detect tumor infiltration (159). Thus the bulk of the case reports, which are the major source for determining patterns of liver injury in humans, occurred in an era when jaundice due to liver metastases could easily have been attributed to chemotherapy. The most clear-cut evidence of hepatotoxicity comes from studies using a drug as a single agent, since attribution of causality in a multidrug regimen may be difficult and since toxicity may be due to an interaction of two or more drugs. Most anticancer drugs will be incorporated into a multidrug regimen once initial safety testing is complete. Thus our information on hepatotoxicity due to single drug exposure will often be 20 or more years old for many drugs. Thus much of the information of these drugs as single agents was reported prior to availability of diagnostic studies for viral hepatitis and of current imaging modalities. A final complication in cancer patients is the multitude of factors that may affect the liver (Table 2). Although a high frequency of case reports may reliably link toxicity to a given drug, the bulk of our knowledge base for anticancer drugs comes from sporadic reports in the older literature, which is fraught with the difficulties described in the preceding paragraphs. This is particularly problematic for host-dependent, dose-independent toxicity (idiosyncratic reactions): these reactions usually occur with low frequency and the only reported cases may

644

DeLeve

TABLE 3 Conventional Dose Chemotherapy: Effect on Liver Tests and Liver Injury Drug

Transient liver test abnormalities 0

2-Chloro-3 deoxyadenosine 6-Mercaptopurine 6-Thioguanine

Actinomycin D Arsenic trioxide Azathioprine

24% of patients with inflammatory bowel disease developed serum transaminase elevations; median elevation was grade I (160)

Liver test abnormalities reported in 18% (163) Liver test abnormalities reported in 13% of patients with inflammatory bowel disease (164)

Bortezomib Busulfan Carmustine (BCNU) Chlorozotocin Cisplatin

Up to 25% (165,166) Up to 25% elevations of amino-transferases (167)

Cyclophosphamide

Uncommon; see text for high dose

Cyproterone acetate

10% with alkaline phosphatase and 3% with aminotransferase elevations (168)

Cytarabine

Frequent in one series but confounders obscure true causality; high dose may give transient ab normalities. Mild, transient elevation of amino-transferases in up to 50% (169) Low incidence of moderate, transient liver test abnormalities (170)

Dacarbazine Etoposide Fluorodeoxyuridine (into hepatic artery) Flutamide Gemcitabine Imatinib (GleevecTM)

L-asparaginase

Lomustine (CCNU) Megestrol acetate

Aminotransferase elevations Aminotransferase elevations WHO grade I IIa in 60% of patients, grade III/IV in 5 10% (106,172 174) grade III IV elevations of serum transaminase in 1 3% and of bilirubin in 0.4 3.5% when treated with imatinib alone (40); 53% grade III IV liver test abnormalities in imatinib combined with conventional chemotherapy (45) Abnormalities in O50% of patients in older literature; considerably less in recent years Uncommon (175)

Liver injury 1 case report peliosis hepatis in hairy cell leukemia Numerous reports of either hepatocellular or cholestatic liver disease Case reports of SOS; 1 case of peliosis hepatis when given with cytarabine; Nodular regenerative hyperplasia reported in 18 53% of patients treated for inflammatory bowel disease (161,162) SOS, particularly for right-sided Wilm’s tumors Hepatotoxicity reported in 10% (nZ16) of patients with inflammatory bowel disease: 1 patient with serum transaminase elevations O30 times ULN, 13 patients with grade I II serum transaminase elevations (i.e., !5 times ULN), 2 patients with mixed injury Case report of hepatocellular injury 2 case reports of cholestatic liver injury; SOS in high-dose combination regimens Case reports of liver injury, some fatalities SOS 3 case reports of severe cholestatic liver injury in 1 series Rare cases of steatosis and cholestasis; case reports of hepatocellular injury at high doses Rare case reports at conventional doses; SOS with high dose 13 case reports of hepatocellular in jury with 8 fatalities; a series of 96 cases with 33 fatalities and 5 additional case reports with 2 fatalities not available to author for review Case reports of cholestatic jaundice; case report of peliosis when given with 6-thioguanine O15 case reports of SOS Case reports of hepatocellular injury Sclerosing cholangitis By 1996, 46 reported cases of severe cholestatic hepatitis (3/10,000 users) with 20 fatalities (171) 1 case report of fatal, fulminant liver failure 5 reported cases of liver injury; case of fatal liver failure in patient on imatinib plus acetaminophen

Steatosis (40 90% incidence on au topsy); occasional hepatocellular necrosis Rare case reports of liver injury, some fatalities. SOS with high dose Case report of bland cholestasis (Continued )

645

Cancer Chemotherapy

TABLE 3 Conventional Dose Chemotherapy: Effect on Liver Tests and Liver Injury (Continued) Drug

Transient liver test abnormalities

Methotrexate

Aminotransferase elevation is common at high doses (139 141)

Mithramycin

Uncommon with dosing for hypercalcemia; frequent with daily dosing Transient abnormalities in bilirubin or aminotransferases (176,177) Bilirubin elevations, 1.5 3 times ULN in 23% (88)

Mitoxantrone Gemtuzumab ozagamicin Paclitaxel Raltitrexed Streptozotocin Tamoxifen

Up to 67% in 1973 study, but most patients had liver metastases (181)

Liver injury Steatosis, fibrosis, and cirrhosis with maintenance therapy; case reports of hepatocellular carcinoma following fibrosis/cirrhosis Frequent liver injury with daily dosing

Incidence of SOS ranges from 12% to 48% with a case fatality rate of 60% (88 90) One published case of fatal hepatic coma in patient with multiple liver metastases (178) 1 case report of severe liver dysfunction (179); 2 fatal liver failure cases (180) Steatosis; case reports of nonalcoholic steatohepatitis, some with cirrhosis; case report of peliosis; sinusoidal dilatation (also seen in experimental animals) (182)

Unless stated, these findings pertain to conventional-dose rather than high-dose chemotherapy. Toxicity of high-dose treatment is discussed in the text. The WHO grades apply to aminotransferases, alkaline phosphatase, and bilirubin. a See Table 4. Abbreviations: SOS, sinusoidal obstruction syndrome; ULN, upper limit of normal.

have occurred several decades ago. Thus before one assumes that an anticancer drug does indeed cause hepatocellular injury, jaundice or fibrosis, it is imperative to note the year of publication of the case reports that describe the injury. Chemotherapeutic agents commonly cause transient increases in liver test abnormalities without evidence of clinically significant liver injury. Table 3 lists causes of transient liver test abnormalities at conventional doses of chemotherapy (as opposed to high-dose regimens used in stem cell transplantation and some experimental regimens). The right-hand column of Table 3 provides descriptions of the type of clinically apparent liver injury, which is mostly derived from case reports. The relative infrequency of significant liver injury by chemotherapeutic agents is somewhat unexpected. The liver may be relatively protected since many of these compounds target rapidly proliferating cells, whereas liver cells have a slow turnover. As with other categories of drugs, the liver is also protected because of the relative strength of the detoxification pathways within the hepatocyte that protect it from electrophilic metabolites and drug-induced oxidative stress. SELECTED ANTICANCER DRUGS AND MODALITIES Alkylating Agents Cyclophosphamide Cyclophosphamide at standard doses is an uncommon cause of liver toxicity, with few reported cases of liver test abnormalities and rare case reports of clinically significant hepatocellular necrosis (183). SOS is seen almost exclusively at high doses of cyclophosphamide and in conjunction with synergistic agents, such as busulfan, total body irradiation or BCNU. The incidence of SOS in high-dose regimens that contain cyclophosphamide is often in the range of 20% to 40%. Cyclophosphamide metabolism (Fig. 3) by CYP2C9 and 3A4 (9) in hepatocytes yields 4-hydroxycyclophosphamide, which appears in the circulation. 4-Hydroxycyclophosphamide equilibrates with aldophosphamide, which follows two pathways: spontaneous decomposition

646

DeLeve

Cyclophosphamide P-450 4-Hydroxycyclophosphamide GSH

Aldophosphamide Spontaneous

4-Glutathionylcyclophosphamide

Phosphoramide mustard + Acrolein GSH

Hydroxypropyl mercapturic acid

FIGURE 3 Simplified scheme of cyclophosphamide metabolism.

to phosphoramide mustard and acrolein and metabolism by aldehyde dehydrogenase 1 (ALD1) to carboxyethyl phosphoramide mustard. Phosphoramide mustard is the putative antineoplastic moiety and acrolein is the proximate toxic metabolite. There is wide interindividual variation in metabolism of intravenously administered cyclophosphamide and this variability may contribute to the risk of SOS (121,122). The presumptive mechanism of toxicity in SOS is through hepatocyte metabolism of cyclophosphamide and formation of acrolein, the proximate toxic metabolite (90). Acrolein is toxic to endothelial cells in general (184,185), but toxicity is greatest in the sinusoidal endothelial cells due to their proximity to hepatocytes. Busulfan As a single agent, liver toxicity due to high-dose busulfan is described as cholestatic (113), but only two case reports have described cholestatic liver disease at standard chemotherapeutic doses. A large case series found that patients treated with the combination of busulfan plus 6-thioguanine had a significant incidence of non-cirrhotic portal hypertension and/or of nodular regenerative hyperplasia (186). The literature on high-dose busulfan as a single agent is too limited to comment on whether it can induce SOS at all, but the risk seems to be low (113) and it is certainly less, than with dimethylbusulfan alone (187,188). This may reflect differences in cellular toxicity. Busulfan is equally toxic to hepatocytes and sinusoidal endothelial cells, whereas dimethylbusulfan toxicity is selective for sinusoidal endothelial cells as is seen with other drugs closely linked to SOS (L.D. DeLeve, unpublished observation). Busulfan is a weak alkylating agent. Experimental in vitro studies of busulfan toxicity and genotoxicity have been difficult to interpret as the doses needed to induce either interstrand crosslinking or toxicity have been significantly higher than therapeutic plasma concentrations in most of the studies. Busulfan toxicity requires glutathione S-transferase (GSH) mediated conjugation to glutathione, which leads to oxidative stress (189). There are two mechanisms of oxidative stress. In an environment high in glutathione (e.g., within the hepatocyte) the busulfan-glutathione conjugate itself causes oxidative stress. In a low glutathione environment, busulfan conjugation to glutathione further depletes intracellular glutathione and the glutathione depletion causes oxidative stress. Dacarbazine Dacarbazine-induced SOS has been reported in at least 15 case reports. The disease varies from the usual presentation of SOS in that it is associated with peripheral eosinophilia and thrombosis of the central venules and veins. The peripheral eosinophilia has been seen after the first exposure to dacarbazine (190), which suggests that this is related to the toxicity rather than implicating a hypersensitivity reaction. The only site of microsomal activation of dacarbazine is in the liver. The initial microsomal metabolite transforms spontaneously through several steps to the proximate

Cancer Chemotherapy

647

methyl-donating metabolite, either a methylcarbonium or a methyldiazonium ion (191). This drug is selectively toxic to sinusoidal endothelial cells, which can metabolically activate it, and it is detoxified by glutathione (127). Melphalan At standard chemotherapeutic doses, melphalan is not hepatotoxic. As a single agent, high-dose melphalan (140 mg/m2) has been associated with either mild, transient elevations of serum aminotransferase and bilirubin (114,192) or no abnormalities at all (115,116). In high-dose multidrug conditioning regimens, melphalan is associated with SOS. However, it has been reported that the incidence and severity of SOS following the busulfan–melphalan regimen was lower than that seen historically due to the busulfan–cyclophosphamide regimen (118). Melphalan, or L-phenylalanine mustard is a bifunctional alkylating agent. Conjugation of melphalan to glutathione requires GSH and the glutathionyl conjugate is a non-competitive inhibitor of GSH (193). Efflux of the GSH conjugate by MRP1 should determine ongoing conjugation of the parent compound, since efflux from the cell reduces the concentration of intracellular conjugate available to inhibit (GSH) (193,194). The evidence that GSH conjugation detoxifies melphalan is not entirely straightforward: depletion of GSH exacerbates toxicity, but the addition of GSH or the GSH precursor N-acetylcysteine does not attenuate toxicity (195). Antimetabolites Methotrexate Concern about methotrexate-induced fibrosis was raised by a much-cited 1960 study of 273 children treated for acute leukemia (196). The study described a 31% incidence of fibrosis in autopsy cases prior to the use of chemotherapy between 1940 and 1947 and an increase to 78% incidence of fibrosis between 1948 and 1951 when the folic acid antagonists, aminopterin, and methotrexate, were used. No information was provided about the frequency of transfusion before and after the introduction of chemotherapy and testing for viral hepatitis was of course not available. More recent studies have not found a significant incidence of liver injury (see below). In high-dose methotrexate therapy for osteosarcoma, childhood acute lymphocytic leukemia, and adult non-Hodgkins lymphomas, methotrexate may be infused in doses of 3–15 g/m2. During maintenance therapy, weekly oral doses of low-dose methotrexate may be interspersed with additional infusions of high-dose methotrexate. Thus very large cumulative doses can be achieved. Although high-dose methotrexate has a high incidence of transient elevation of aminotransferase, the current literature suggests that this does not result in chronic liver disease (59,197–199). However the incidence of fibrosis in this population may be underestimated. Liver tests correlate poorly with histological abnormalities and clinical studies using high-dose methotrexate have not routinely performed liver biopsies after treatment with significant cumulative doses of methotrexate. Thus clinically asymptomatic fibrosis likely goes undiagnosed. Given the significant numbers of long-term disease-free survivors, long-term toxicity is a concern. Two case reports of hepatocellular carcinoma following methotrexate-induced fibrosis were reported in 1977 and 1987 (200,201); it should be noted that these reports preceded hepatitis C testing in leukemia patients, a population with a high incidence of hepatitis C. In a large meta-analysis, patients with psoriasis and rheumatoid arthritis have a 7% chance of progression of histological abnormalities on liver biopsy for every gram of methotrexate (202). Even if one assumes that asymptomatic fibrosis goes undiagnosed if systematic biopsies are not done, given the very high cumulative doses of methotrexate administered in cancer therapy, the incidence of significant liver injury would seem to be much lower than that suggested by the literature for rheumatoid arthritis and psoriasis. Thiopurines In a prospective study of 161 patients treated with azathioprine or 6-mercaptopurine for inflammatory bowel disease, 13% developed liver test abnormalities (164). In this same study, 10% (nZ16) reportedly had liver injury, but it was unclear how liver injury was defined. One patient with a hepatocellular pattern of injury had serum transaminases 30 times upper limit of

648

DeLeve

normal (ULN). However most patients with a hepatocellular pattern (12 of 14) had grade I or II serum transaminases, i.e., !5 times ULN. Two of the patients had a mixed pattern with grade II liver test abnormalities. All patients were on azathioprine at the time of hepatotoxicity. This study demonstrates that azathioprine can cause significant liver injury, but that it also has a significant incidence of mild liver test abnormalities. Liver injury from 6-mercaptopurine usually presents with jaundice that may be accompanied by pruritus. The injury is most commonly cholestatic, but hepatocellular necrosis may also be present. Although liver injury due to 6-mercaptopurine has been frequently reported, the actual incidence of toxicity is difficult to estimate but may be lower than suggested by early studies. In 1952, 6-mercaptopurine was first incorporated into the treatment for leukemia. Between 1954 and 1959, 6-mercaptopurine (typically 2.5 mg/kg/day, but as high as 5 mg/kg/day) was reported to induce jaundice in 6% to 14% of leukemic patients (203–205). The incidence in a 1964 study of 38 patients was reported to be as high as 42% (206). In current regimens for acute lymphoblastic leukemia, 6-mercaptopurine is used within multi-drug regimens, usually with other potential hepatotoxins such at L-asparaginase, cytarabine, and methotrexate. Liver injury is not singled out as a major toxicity in most of the larger studies and when it does occur cannot be attributed to any particular drug in the regimen. Only one clearcut case of drug-induced cholestatic hepatitis was found in an 18 year follow-up study of 396 patients with inflammatory bowel disease treated with 6-mercaptopurine (50 mg/day or 1.5 mg/kg) (207). Metabolism of 6-mercaptopurine, 6-thioguanine, and azathioprine are shown in Figure 1. Oral administration of thiopurines leads to extensive first-pass metabolism in the intestine and liver by xanthine oxidase. Cytotoxicity is due to formation of thioguanine nucleotides. In target tissues the effect of thiopurines is determined by the balance between detoxification by thiopurine S-methyltransferase (TPMT) and toxification by hypoxanthine phosporibosyltransferase (HPRT): HPRT initiates formation of thioguanine nucleotides, whereas methylation reactions by TPMT shunt the drug away from thioguanine nucleotide formation. There is a genetic polymorphism of TPMT activity with a trimodal distribution: 90% of persons have wildtype which confers high activity, 10% are heterozygotes with intermediate activity, and 0.3% are homozygous mutants with little or no detectable TPMT activity. Since TPMT shunts metabolites away from the activation pathway, low TPMT activity confers increased risk for hematopoietic toxicity [see review (208)] and perhaps also for secondary malignancies, including brain tumors and acute myelogenous leukemia (209). However, hepatotoxicity from these thiopurines may be due to a different mechanism, namely high first-pass metabolism in the liver by TPMT to 6-methylmercapopurine. What evidence is there to suggest this? TPMT mRNA is highly expressed in the liver (208) and hepatotoxicity is more frequent among individuals with higher TPMT activity (210). Studies that compared patients with and without 6-mercapopurine hepatotoxicity demonstrated pharmacokinetics consistent with increased first-pass metabolism in the patients with toxicity, notably delayed time to peak concentrations, lower peak levels and lower AUCs (211). High erythrocyte 6-methymercaptopurine levels have been reported to correlate with hepatotoxicity (212). Alternatively or additionally, TPMT may contribute to toxicity through formation of methylmercaptopurine nucleotides derived from methylthioinosine monophosphate (Fig. 3) (213). In 123 patients treated with 6-thioguanine for inflammatory bowel disease, 24% developed serum transaminase elevations, but the median elevation was only grade I (1.4 times ULN) (160). Risk factors for transaminase elevations were male gender and preferential 6-methylmercaptopurine production. 6-Thioguanine has been reported to cause nodular regenerative hyperplasia in 18% to 53% of patients treated for inflammatory bowel disease (161,162). An injury described as SOS has been linked to 6-thioguanine, often in combination with cytarabine (214–217). These cases have most often resembled radiation induced liver disease (RILD) in some of the clinical and histological features. Patients present with hepatomegaly that is sometimes described as painful and ascitic, but without jaundice. Bilirubin is often normal or marginally elevated, as is seen in RILD. There are venoocclusive lesions and centrilobular congestion, but centrilobular hepatocyte atrophy is described rather than necrosis. Several of the cases also had underlying cirrhosis.

Cancer Chemotherapy

649

It is not surprising that this group of liver injuries is linked to one drug, such as 6-thioguanine. SOS, nodular regenerative hyperplasia, peliosis hepatis, and sinusoidal dilatation often share similar causes and in some cases up to all four have been described within the same liver. Azathioprine (all four lesions), urethane (peliosis and SOS), heroin (sinusoidal dilatation, sinusoidal, and perivenular fibrosis), thorotrast (peliosis, venoocclusive lesions with hepatocyte atrophy), and oral contraceptives (sinusoidal dilatation and peliosis hepatis) cause an overlap of these injuries. Damage to sinusoidal endothelial cells and sometimes to hepatic venular endothelial cells seems to be the common link in these four types of liver injury (128,146,218–221). Azathioprine, another thiopurine, has been shown to be selectively toxic to sinusoidal endothelial cells (128), but this has not been examined specifically for 6-thioguanine. 3-Fluorodeoxyuridine or floxuridine (FUDR) may be infused into the hepatic artery to treat hepatic metastases. This may cause liver test abnormalities and, in more serious cases, sclerosing cholangitis. High-grade obstruction of the common hepatic duct may extend into the left and right hepatic ducts and there are usually also multiple strictures of the intrahepatic ducts. The incidence varies widely in the literature; in a recent study of 32 patients who received an average of 7.3 cycles of chemotherapy, the incidence of liver test abnormalities was 15.6% and of severe biliary sclerosis was 9.3% (222). In another recent study in which 38 patients received two cycles of FUDR plus dexamethasone, 22% had elevations of aspartate aminotransferase (AST and/or alkaline phosphatase and 7% had bilirubin elevations greater than three times the ULN (223). Concomitant treatment with dexamethasone has been suggested to reduce liver injury from intraarterial FUDR, but this has not been well established (223–225). Progressive elevation of bilirubin levels, pruritus, or sepsis in patients with sclerosing cholangitis may be successfully palliated with percutaneous transhepatic biliary drainage (226). There are histological changes in the hilar vessels that suggest organization of occlusive thrombi (227). The biliary duct strictures of sclerosing cholangitis may therefore be due to circulatory impairment secondary to drug-induced damage to the peribiliary vascular plexus. Monoclonal Antibodies MylotargTM (gemtuzumab ozogamicin) is used to treat acute myeloid leukemia and seems to have a significant incidence of SOS (88,89,188). The incidence of SOS is described in the earlier section on drugs that cause SOS. To summarize, gemtuzumab ozogamicin has a relatively low risk by itself of causing SOS. However there may be a very high incidence of SOS in patients who have received stem cell transplantation followed within a short time frame by gemtuzumab ozogamicin, or in the opposite order, i.e., gemtuzumab ozogamicin followed by stem cell transplantation. Gemtuzumab ozogamicin is a conjugate of calicheamicin linked to the “humanized” monoclonal anti-CD33 antibody. CD33 is a myeloid surface antigen that is expressed on more than 90% of blast cells in acute myeloid leukemia, but not on hematopoietic stem cells or lymphoid cells. The presumptive mechanism of action of gemtuzumab ozogamicin is preferential binding to cells with the CD33 antigen, internalization of the conjugate, and release of the calicheamicin moiety by acid hydrolysis within lysosomes (228,229). Calicheamicin has a methyltrisulfide group that is reduced by glutathione. The resulting diradical species binds to the minor groove of double-stranded DNA and causes sequence-selective oxidation of deoxyribose leading to DNA strandbreaks (230,231). The mechanism leading to gemtuzumab ozogamicin-induced SOS is undefined. Since both sinusoidal endothelial cells and Kupffer cells are of bone marrow origin (232,233), it is possible that one or both may have CD33 surface antigen that binds gemtuzumab ozogamicin. Miscellaneous Anticancer Drugs L-asparaginase L-asparaginase

is used in the treatment of acute lymphoblastic leukemia. It is a microbial product derived from Escherichia coli or from Erwinia chrysanthemi. Interference with liver function is manifested by decreased serum albumin, coagulation factors, and lipoproteins. Liver toxicity is thought to be due to inhibition of protein synthesis by asparaginase and

650

DeLeve

TABLE 4 WHO Grades of Liver Test Abnormalities Grade 0 Grade 1 Grade 2 Grade 3 Grade 4

/4 1.25 times ULN 1.26 2.5 times ULN 2.6 5 times ULN 5.1 10 times ULN O10 times ULN

3

Abbreviation: ULN, upper limit of normal. Source: From Ref. 238.

glutaminase activity (234,235). In the past a particularly high incidence of liver toxicity was described with elevated alkaline phosphatase and serum aminotransferase and depression of liver synthetic function. 40% to 87% of patients were found to have hepatic steatosis on autopsy and this could be detected up to nine months after the last dose of L-asparaginase (236,237). The incidence of liver test abnormalities and of frank liver disease is considerably lower in the recent literature, although severe and fatal cases are still reported. In adults the incidence of transient WHO grade I or II liver test abnormalities (WHO grades listed in Table 4) is around 50% (239) with few cases of significant liver test abnormalities (239,168). A recent study of 245 patients with acute lymphoblastic leukemia randomized to conventional dose or high dose L-asparaginase as part of a multi-drug regimen found antithrombin III levels, less than 50% in 0 and 2.5%, antithrombin III levels of 50% to 70% in 1.7% and 10.3%, and fibrinogen less than 100 mg/dl in 8.4% and 10.3% of patients, respectively, in the conventional and high dose L-asparaginase treatment groups (240). Liver test abnormalities and liver injury were not listed among the major toxicity reactions. Another recent study of 377 patients with acute lymphoblastic leukemia treated with a regimen that included L-asparaginase, liver test abnormalities and liver injury were not listed among the more frequent toxicities; among a subgroup of 43 patients in this study who did not complete the full 30 week treatment regimen, 2% developed hepatitis (241). Actinomycin D Actinomycin D (or dactinomycin) has been associated with SOS. There may be some synergistic toxicity with abdominal irradiation or vincrisitine. The risk may correlate with the dose of radiation and perhaps also the dose of actinomycin D (242). Most of the reported cases of actinomycin D have been in patients with right-sided Wilms’ tumors (243). The risk of SOS may be greater when the right-sided tumors are large (244). The incidence of actinomycin D-induced SOS is low for left-side Wilms’ tumors and for rhabdomyosarcoma (243,171). One potential explanation is that external vascular compression from these large right-sided nephroblastomas is impeding hepatic venous outflow, i.e., causing a Budd-Chiari syndrome. Another contributing factor may be intravascular extension of the nephroblastoma, causing Budd-Chiari syndrome, which may be under diagnosed in the older literature (245). In one study, intravascular extension of tumor was diagnosed by ultrasonography in only 40% of cases, perhaps due to the lesser imaging quality of older ultrasonography equipment in the early cases of the series (245). Budd-Chiari syndrome and SOS share many of the same clinical and histological features. The unanswered question is whether the increased incidence of SOS with right-sided Wilms’ tumors is due to hepatic venous outflow obstruction acting in concert with actinomycin D-induced SOS, due to undiagnosed Budd-Chiari syndrome or due to some other unidentified risk factor. Radiation RILD or radiation hepatitis occurs after hepatic irradiation. With conventional fractionation of irradiation, RILD occurs at doses in excess of 30–35 Gy in adults. Children or adults who have recently undergone partial hepatectomy may develop RILD at lower doses. The signs and symptoms of RILD resemble Budd-Chiari syndrome or SOS with hepatomegaly, weight gain, and varying amount of ascites. Common histological features are sinusoidal congestion, sinusoidal fibrosis, and subendothelial and adventitial fibrosis of the

651

Cancer Chemotherapy

TABLE 5 Differences Between Sinusoidal Obstruction Syndrome (SOS) and Radiation-Induced Liver Disease (RILD) SOS in stem cell transplantation Time of onset Resolution of signs/symptoms RUQ pain Bilirubin Histology

Day 0 30 30 60 days Marked O2 mg/dl, often markedly elevated Centrilobular necrosis

RILD 2 weeks 4 months (usually 1 2 months) Months Mild Normal or minimal elevation Centrilobular atrophy

Abbreviation: RUQ, right upper quadrant.

central veins. However RILD differs from SOS due to stem cell transplantation conditioning therapy in several ways (246) (Table 5): (1) The diagnostic criteria for SOS in stem cell transplantation include elevations of bilirubin O2 mg/dl and tenderness of the liver, whereas in RILD bilirubin elevations are usually minimal and right upper quadrant (RUQ) pain is much less pronounced (246,165). (2) A characteristic histologic feature of SOS is centrilobular necrosis, whereas in RILD there is atropy of the centrilobular cords and coagulative necrosis is uncommon (246,165,166). The presence of atrophy rather than necrosis may reflect the fact that RILD becomes clinically apparent at a much later time-point. (3) In RILD fibrin has been identified within the central vein by electron microscopy, but fibrin has not been demonstrated by electron microscopic examination of SOS. (4) In SOS onset of the first clinical signs can occur as early as day 0, the day of stem cell infusion, or as late as 30 days after exposure to the conditioning regimen. Onset of RILD typically occurs one to two months after irradiation, although it may occur as early as two weeks or as late as seven months afterwards (246). (5) Signs of SOS resolve within 30 to 60 days of onset in patients who survive the disease, whereas in RILD evidence of liver injury can persist for months after the insult has been discontinued (165,167). Thus clinical presentation, histology, and time course of disease distinguish RILD from SOS. Historically, RILD has limited use of hepatic irradiation in the treatment of intrahepatic cancers. However with the advent of three dimensional radiation therapy treatment planning, much higher doses of radiation can be delivered to the liver with a low incidence of RILD (169). Hepatic irradiation can also act synergistically with chemotherapy to cause liver toxicity. Total body irradiation contributes to the risk of SOS in hematopoietic stem cell transplantation, although the doses of irradiation used (10–16 Gy) are well below the hepatotoxic dose for radiation alone. In combination with cyclophosphamide, the incidence of SOS is higher in single-dose than in hyperfractionated total-body irradiation (170). Higher doses of total body irradiation, i.e., O12 Gy, may also increase the risk (106). Doses of irradiation O20 Gy may also contribute to the incidence of SOS when used in conjunction with L-asparaginase for Wilms’ tumor. No features have been identified that distinguish SOS due to high-dose combination chemotherapy (i.e., chemotherapeutic drugs without irradiation) from that resulting from hepatic irradiation plus high-dose chemotherapy. Hormones Tamoxifen. Tamoxifen a nonsteroidal drug with antiestrogenic and estrogenic properties is widely used in the chemoprevention of breast cancer. Reported liver injury from tamoxifen includes nonalcoholic fatty liver disease (106,172–177), peliosis hepatis (178), acute hepatitis (181), and hepatocellular cancer (182,238). Nonalcoholic fatty liver disease (NAFLD) is the most common form of liver injury due to tamoxifen. A Japanese study that screened 105 women on tamoxifen by annual abdominal CT examination found a 38% incidence of fatty liver, which developed during the first two years of therapy in 35 of the 40 cases (85%) (177). 40% of the patients with fatty liver (16 of 40 patients) had sustained elevations of aminotransferases. In addition, there have been three reported cases of cirrhosis in the presence of steatohepatitis by liver biopsy (173,175). In contrast, a prospective study of 5408 women treated with tamoxifen and followed for a median of seven years found that tamoxifen-associated non-alcoholic steatohepatitis was largely an indolent disease and more common in women with the metabolic syndrome (247).

652

DeLeve

Rats treated with tamoxifen develop nodular regenerative hyperplasia, hepatic adenomas and hepatocellular carcinomas (248). Tamoxifen is metabolized by CYP3A4. There are substantial differences between rats and humans in the rate of tamoxifen metabolism and in the metabolites formed. To achieve clinically relevant serum concentrations, rats must be given high doses, which result in very high liver concentrations of tamoxifen and its metabolites (249,250). Rats are also more susceptible to liver DNA damage from tamoxifen, albeit at liver concentrations that are much higher than those seen in humans (250,251). The propensity for hepatocellular cancer in rats may be very species specific, since liver tumors are not found in mice or hamsters. There is currently no epidemiological evidence in humans of a significant increase in the incidence of liver cancer [see review (248)], but there have been case reports of hepatocellular cancer after long-term use of tamoxifen (182,238). Given the expanded indications for use of tamoxifen, future clinical studies will undoubtedly carefully monitor the risk of hepatocellular carcinoma. Cyproterone acetate is a synthetic progesterone derivative with antiandrogenic and progesterone-like activity, used in the treatment of advanced prostate cancer. It is marketed in the United Kingdom and Germany, but is not approved by the food and drug administration (FDA). In a large surveillance study of 1685 patients receiving cyproterone acetate for indications other than prostate cancer, elevated liver tests were noted in 10% of patients treated with 50 mg/day and 20% of those receiving O100 mg/day (252). A retrospective analysis of 78 patients receiving 50 mg/day of cyproterone acetate for advanced prostate cancer reported elevation of alkaline phosphatase in 14% of patients without known liver involvement and elevated aminotransferases in 2.5% (253). There have been 18 case reports of cyproteroneassociated hepatitis with six fatalities. A review article (252) cited a report of 96 hepatotoxic events with 33 fatalities attributed to cyproterone acetate (254). There has also been one report of cirrhosis in a pediatric patient treated for precocious puberty (255). Cyproterone acetate is mitogenic, tumorigenic, and induces DNA-adducts and DNArepair synthesis in rat liver [see reviews (256,257)]. High levels of DNA-adducts are also formed in human hepatocytes. It has been suggested that formation of the reactive metabolite is catalyzed by hydroxysteroid-sulfotransferases. Although long-term exposure to cyproterone acetate at high doses may potentially induce hepatocellular carcinomas, this is unlikely to be clinically relevant given the current life expectancy for advanced prostate cancer. Flutamide. Liver injury from flutamide presents with marked elevations of bilirubin and a wide range in elevation of serum aminotransferases. The predominant histological feature determined on autopsy is marked to massive hepatic necrosis. In a multicenter study of 905 patients treated with flutamide, liver tests with elevations greater than four times the ULN occurred in 0.8% of patients (258). According to postmarketing surveillance, severe liver disease due to flutamide has occurred in 46 patients with 20 fatalities (259). The rate of serious liver injury is estimated to be 3 per 10,000 flutamide users, based on number of prescriptions written. Flutamide is a synthetic non-steroid that is a competitive antagonist of the androgen receptor. After oral administration, flutamide undergoes extensive first-pass metabolism with formation of several oxidized metabolites. Formation of electrophilic metabolites is catalyzed via CYP3A and CYP1A (260). Experimental studies in rat hepatocytes (261) suggest that toxicity from the electrophilic metabolites occurs through depletion of hepatocyte glutathione, which is accompanied by oxidative stress. Toxicity to mitochondria is manifested by depression of mitochondrial respiration and ATP formation, but the study did not report whether mitochondrial toxicity was due to depletion of mitochondrial glutathione.

CONCLUSIONS Transient liver test abnormalities are a common occurrence in cancer chemotherapy. Our knowledge of chemotherapy-induced liver injury is confounded by the injurious effects of a variety of insults to the liver in cancer patients as well as by the limitations in the older literature. Nevertheless it is clear that liver injury is a relatively frequent complication in some of the current treatments for hematologic malignancies, but that clinically apparent liver injury occurs much less frequently with conventional dose chemotherapy.

Cancer Chemotherapy

653

REFERENCES 1. Zimm S, Colins JM, O’Neill D, Chabner BA, Poplack DG. Inhibition of first-pass metabolism in cancer chemotherapy: interaction of 6-mercaptopurine and allopurinol. Clin Pharmacol Ther 1983; 34:810–7. 2. Humerickhouse R, Lohrbach K, Li L, Bosron WF, Dolan ME. Characterization of CPT-11 hydrolysis by human liver carboxylesterase isoforms hCE-1 and hCE-2. Cancer Res 2000; 60:1189–92. 3. Iyer L, King CD, Whitington PF, et al. Genetic predisposition to the metabolism of irinotecan (CPT-11). Role of uridine diphosphate glucuronosyltransferase isoform 1A1 in the glucuronidation of its active metabolite (SN-38) in human liver microsomes. J Clin Invest 1998; 101:847–54. 4. Sugiyama Y, Kato Y, Chu X. Multiplicity of biliary excretion mechanisms for the camptothecin derivative irinotecan (CPT-11), its metabolite SN-38, and its glucuronide: role of canalicular multispecific organic anion transporter and P-glycoprotein. Cancer Chemother Pharmacol 1998; 42(Suppl.):S44–9. 5. Slatter JG, Schaaf LJ, Sams JP, et al. Pharmacokinetics, metabolism, and excretion of irinotecan (CPT-11) following I.V. infusion of [(14)C]CPT-11 in cancer patients. Drug Metab Dispos 2000; 28:423–33. 6. Anderson GD. A mechanistic approach to antiepileptic drug interactions. Ann Pharmacother 1998; 32:554–63. 7. Yong WP, Ramirez J, Innocenti F, Ratain MJ. Effects of ketoconazole on glucuronidation by UDPglucuronosyltransferase enzymes. Clin Cancer Res 2005; 11:6699–704. 8. Gupta E, Wang X, Ramirez J, Ratain MJ. Modulation of glucuronidation of SN-38, the active metabolite of irinotecan, by valproic acid and phenobarbital. Cancer Chemother Pharmacol 1997; 39:440–4. 9. Ren S, Yang JS, Kalhorn TF, Slattery JT. Oxidation of cyclophosphamide to 4-hydroxycyclophosphamide and deschloroethylcyclophosphamide in human liver microsomes. Cancer Res 1997; 57:4229–35. 10. Tian Q, Zhang J, Chan E, Duan W, Zhou SF. Multidrug resistance protein (MRPs) and implication in drug development. Drug Dev Res 2005; 51:1–18. 11. Borst P, Oude Elferink RP. Mammalian ABC transporters in health and disease. Annu Rev Biochem 2002; 71:537–92. 12. Luo FR, Paranjpe PV, Guo A, Rubin E, Sinko P. Intestinal transport of irinotecan in Caco-2 cells and MDCK II cells overexpressing efflux transporters Pgp, cMOAT, and MRP1. Drug Metab Dispos 2002; 30:763–70. 13. Chu XY, Kato Y, Sugiyama Y. Multiplicity of biliary excretion mechanisms for irinotecan, CPT-11, and its metabolites in rats. Cancer Res 1997; 57:1934–8. 14. Gupta E, Safa AR, Wang X, Ratain MJ. Pharmacokinetic modulation of irinotecan and metabolites by cyclosporin A. Cancer Res 1996; 56:1309–14. 15. Kool M, van der Linden M, de Haas M, et al. MRP3, an organic anion transporter able to transport anti-cancer drugs. Proc Natl Acad Sci USA 1999; 96:6914–9. 16. Zeng H, Bain LJ, Belinsky MG, Kruh GD. Expression of multidrug resistance protein-3 (multispecific organic anion transporter-D) in human embryonic kidney 293 cells confers resistance to anticancer agents. Cancer Res 1999; 59:5964–7. 17. Norris MD, Smith J, Tanabe K, et al. Expression of multidrug transporter MRP4/ABCC4 is a marker of poor prognosis in neuroblastoma and confers resistance to irinotecan in vitro. Mol Cancer Ther 2005; 4:547–53. 18. Allen JD, Schinkel AH. Multidrug resistance and pharmacological protection mediated by the breast cancer resistance protein (BCRP/ABCG2). Mol Cancer Ther 2002; 1:427–34. 19. Yang CJ, Horton JK, Cowan KH, Schneider E. Cross-resistance to camptothecin analogues in a mitoxantrone-resistant human breast carcinoma cell line is not due to DNA topoisomerase I alterations. Cancer Res 1995; 55:4004–9. 20. Burger H, van Tol H, Boersma AWM, Brok M, Wiemer EAC, Stoter G, Nooter K. Imatinib mesylate (TTI571) is a substrate for the breast cancer resistance pump (BCRP)/ABCG2 drug pump. Blood 2004; 104:2940–2. 21. Kaplowitz N, DeLeve LD. Drug-Induced Liver Disease. New York: Marcel Dekker, Inc., 2003. 22. Ambudkar SV, Dey S, Hrycyna CA, Ramachandra M, Pastan I, Gottesman MM. Biochemical, cellular, and pharmacological aspects of the multidrug transporter. Ann Rev Pharmacol Toxicol 1999; 39:361–98. 23. Kivisto¨ KT, Kroemer HK, Eichelbaum M. The role of human cytochrome P450 enzymes in the metabolism of anticancer agents: implications for drug interactions. Br J Clin Pharmacol 1995; 40:523–30. 24. Lum BL, Kaubisch S, Yahanda AM, et al. Alteration of etoposide pharmacokinetics and pharmacodynamics by cyclosporine in a phase I trial to modulate multidrug resistance. J Clin Oncol 1992; 10:1635–42.

654

DeLeve

25. Ellis AG, Webster LK. Inhibition of paclitaxel elimination in the isolated perfused rat liver by Cremophor EL. Cancer Chemother Pharmacol 1999; 43:13–8. 26. Ellis AG, Crinis NA, Webster LK. Inhibition of etoposide elimination in the isolated perfused rat liver by Cremophor EL and Tween 80. Cancer Chemother Pharmacol 1996; 38:81–7. 27. Yule SM, Walker D, Cole M, et al. The effect of fluconazole on cyclophosphamide metabolism in children. Drug Metab Dispos 1999; 27:417–21. 28. Webster GJM, Kurtovic J, Singh-Grewal I, Friedlander M, Riordan SM. Prevention of hepatotoxicity but loss of antimelanoma effect with combined fotemustine and anti-oxidant treatment. Intern Med J 2004; 34:298–9. 29. Davies AA, Davey SG, Harbord R, Bekkering GE, Sterne JAC, Beynon R, Thomas S. Nutritional interventions and outcome in patients with cancer or preinvasive lesions: systematic review. J Natl Cancer Inst 2006; 98:961–73. 30. Vickers AJ, Cassileth BR. Unconventional therapies for cancer and cancer-related symptoms. Lancet Oncol 2001; 2:226–32. 31. Hoh C, Boocock D, Marczylo T, et al. Pilot study of oral silibinin, a putative chemopreventive agent, in colorectal cancer patients: silibinin levels in plasma, colorectum, and liver and their pharmacodynamic consequences. Clin Cancer Res 2006; 12:2944–50. 32. Drisko JA, Chapman J, Hunter VJ. The use of antioxidants with first-line chemotherapy in two cases of ovarian cancer. J Am Coll Nutr 2003; 22:118–23. 33. Eli R, Fasciano JA. An adjunctive preventive treatment for cancer: ultraviolet light and ginkgo biloba, together with other antioxidants, are a safe and powerful, but largely ignored, treatment option for the prevention of cancer. Med Hypotheses 2006; 66:1152–6. 34. Villikka K, Kivisto KT, Maenpaa H, Joensuu H, Neuvonen PJ. Cytochrome P450-inducing antiepileptics increase the clearance of vincristine in patients with brain tumors. Clin Pharmacol Ther 1999; 66:589–93. 35. Royer I, Monsarrat B, Sonnier M, Wright M, Cresteil T. Metabolism of docetaxel by human cytochromes P450: interactions with paclitaxel and other antineoplastic drugs. Cancer Res 1996; 56:58–65. 36. Synold TW, Dussault I, Forman BM. The orphan nuclear receptor SXR coordinately regulates drug metabolism and efflux. Nat Med 2001; 7:584–90. 37. Mani S, Huang H, Sundarababu S, Liu W, Kalpana G, Smith AB, Horwitz SB. Activation of the steroid and xenobiotic receptor (human pregnane X receptor) by nontaxane microtubule-stabilizing agents. Clin Cancer Res 2005; 11:6359–69. 38. Lindley C, Hamilton G, McCune JS, et al. The effect of cyclophosphamide with and without dexamethasone on cytochrome P450 3A4 and 2B6 in human hepatocytes. Drug Metab Dispos 2002; 30:814–22. 39. Friedman HS, Petros WP, Friedman AH, Schaaf LJ, Kerby T, Lawyer J, Parry M, Houghton PJ, Lovell S, Rasheed K, et al. Irinotecan therapy in adults with recurrent or progressive malignant glioma. J Clin Oncol 1999; 17:1516–25. 40. Cohen MH, Williams G, Johnson JR, et al. Approval summary for imatinib mesylate capsules in the treatment of chronic myelogenous leukemia. Clin Cancer Res 2002; 8:935–42. 41. Lyseng-Williamson K, Jarvis B. Imatinib Drugs 2001; 61:1765–74 (discussion 1775–1766). 42. Bolton AE, Peng B, Hubert M, Krebs-Brown A, Capdeville R, Keller U, Seiberling M. Effect of rifampicin on the pharmacokinetics of imatinib mesylate (Gleevec, STI571) in healthy subjects. Cancer Chemother Pharmacol 2004; 53:102–6. 43. Smith P, Bullock JM, Booker BM, Haas CE, Berenson CS, Jusko WJ. The influence of St. John’s wort on the pharmacokinetics and protein binding of imatinib mesylate. Pharmacotherapy 2004; 24:1508–14 (erratum appears in Pharmacotherapy 2004; 24(12):1837). 44. Dutreix C, Peng B, Mehring G, Hayes M, Capdeville R, Pokorny R, Seiberling M. Pharmacokinetic interaction between ketoconazole and imatinib mesylate (Glivec) in healthy subjects. Cancer Chemother Pharmacol 2004; 54:290–4. 45. Wassmann B, Pfeifer H, Goekbuget N, Beelen DW, Beck J, Stelljes M, Bornhauser M, Reichle A, Perz J, Haas R, et al. Alternating versus concurrent schedules of imatinib and chemotherapy as frontline therapy for Philadelphia-positive acute lymphoblastic leukemia (PhCALL). Blood 2006; 108:1469–77. 46. Ohyashiki K, Kuriyama Y, Nakajima A, et al. Imatinib mesylate-induced hepato-toxicity in chronic myeloid leukemia demonstrated focal necrosis resembling acute viral hepatitis. Leukemia 2002; 16:2160–1. 47. James C, Trouette H, Marit G, Cony-Makhoul P, Mahon FX. Histological features of acute hepatitis after imatinib mesylate treatment Leukemia. 17 2003:978–9 (comment). 48. Rocca P, El Jastimi S, Troncy J, et al. Imatinib mesylate-induced acute cytolytic hepatitis. Gastroenterol Clin Biol 2004; 28:918–9. 49. Ayoub WS, Geller SA, Tran T, Martin P, Vierling JM, Poordad FF. Imatinib (Gleevec)-induced hepatotoxicity. J Clin Gastroenterol 2005; 39:75–7.

Cancer Chemotherapy

655

50. Larroquette CA, Hortobagyi GN, Buzdar AU, Holmes FA. Subclinical hepatic toxicity during combination chemotherapy for breast cancer. JAMA 1986; 256:2988–90. 51. Zaman GJ, Lankelma J, van Tellingen O, et al. Role of glutathione in the export of compounds from cells by the multidrug-resistance-associated protein. Proc Natl Acad Sci USA 1995; 92:7690–4. 52. Zhou-Pan XR, Seree E, Zhou XJ, Placidi M, Maurel P, Barra Y, Rahmani R. Involvement of human liver cytochrome P450 3A in vinblastine metabolism: drug interactions. Cancer Res 1993; 53:5121–6. 53. Zhou XJ, Zhou-Pan XR, Gauthier T, Placidi M, Maurel P, Rahmani R. Human liver microsomal cytochrome P450 3A isozymes mediated vindesine biotransformation. Metabolic drug interactions. Biochem Pharmacol 1993; 45:853–61. 54. Kajita J, Kuwabara T, Kobayashi H, Kobayashi S. CYP3A4 is mainly responsibile for the metabolism of a new vinca alkaloid, vinorelbine, in human liver microsomes. Drug Metab Dispos 2000; 28:1121–7. 55. Rahman A, Korzekwa KR, Grogan J, Gonzalez FJ, Harris JW. Selective biotransformation of taxol to 6 alpha-hydroxytaxol by human cytochrome P450 2C8. Cancer Res 1994; 54:5543–6. 56. Robieux I, Sorio R, Borsatti E, et al. Pharmacokinetics of vinorelbine in patients with liver metastasis. Clin Pharmacol Ther 1996; 59:32–40. 57. Figg WD, Dukes GE, Pritchard JF, et al. Pharmacokinetics of ondansetron in patients with hepatic insufficiency. J Clin Pharmacol 1996; 36:206–15. 58. Locasciulli A, Testa M, Pontisso P, et al. Prevalence and natural history of hepatitis C infection in patients cured of childhood leukemia. Blood 1997; 90:4628–33. 59. Farrow AC, Buchanan GR, Zwiener RJ, Bowman WP, Winick NJ. Serum aminotransferase elevation during and following treatment of childhood acute lymphoblastic leukemia. J Clin Oncol 1997; 15:1560–6. 60. Meir H, Balawi I, Nayel H, El Karaksy H, El Haddad A. Hepatic dysfunction in children with acute lymphoblastic leukemia in remission: relation to hepatitis infection. Med Pediatr Oncol 2001; 36:469–73. 61. Horina JH, Wirnsberger GH, Kenner L, Holzer H, Krejs GJ. Increased susceptibility for CsA-induced hepatotoxicity in kidney graft recipients with chronic viral hepatitis C. Transplantation 1993; 56:1091–4. 62. Zylberberg H, Carnot F, Mamzer M-F, Blancho G, Legendre C, Pol S. Hepatitis C virus-related fibrosing cholestatic hepatitis after renal transplantation. Transplantation 1997; 63:158–60. 63. Ungo JR, Jones D, Ashkin D, Hollender ES, Bernstein D, Albanese AP, Pitchenik AE. Antituberculosis drug-induced hepatotoxicity. The role of hepatitis C virus and the human immunodeficiency virus. Am J Respir Crit Care Med 1998; 157:1871–6. 64. Strasser SI, Myerson D, Spurgeon CL, et al. Hepatitis C virus infection and bone marrow transplantation: a cohort study with 10-year follow-up. Hepatology 1999; 29:1893–9. 65. Sulkowski MS, Thomas DL, Chaisson RE, Moore RD. Hepatotoxicity associated with antiretroviral therapy in adults infected with human immunodeficiency virus and the role of hepatitis C or B virus infection. JAMA 2000; 283:74–80. 66. Alter MJ, Kruszon-Moran D, Nainan OV, et al. The prevalence of hepatitis C virus infection in the United States, 1988 through 1994. N Engl J Med 1999; 341:556–62. 67. Cheng AL. Steroid-free chemotherapy decreases the risk of hepatitis flare-up in hepatitis B virus carriers with non-Hodgkin’s lymphoma. Blood 1996; 87:1202 (letter). 68. Etienne MC, Chatelut E, Pivot X, Lavit M, Pujol A, Canal P, Milano G. Co-variables influencing 5-fluorouracil clearance during continuous venous infusion. A NONMEM analysis. Eur J Cancer 1998; 34:92–7. 69. Stein BN, Petrelli NJ, Douglass HO, Driscoll DL, Arcangeli G, Meropol NJ. Age and sex are independent predictors of 5-fluorouracil toxicity. Cancer 1994; 75:11–7. 70. Lu Z, Zhang R, Diasio RB. Population characteristics of hepatic dihydropyrimidine dehydrogenase activity, a key metabolic enzyme in 5-fluorouracil chemotherapy. Clin Pharmacol Ther 1995; 58:512–22. 71. den Hartigh J, McVie JG, van Oort WJ, Pinedo HM. Pharmacokinetics of mitomycin C in humans. Cancer Res 1983; 43:5017–21. 72. Miya T, Sasaki Y, Karato A, Saijo N. Pharmacokinetic study of mitomycin C with emphasis on the influence of aging. Jpn J Cancer Res 1992; 83:1382–5. 73. Wade JR, Kelman AW, Kerr DJ, Robert J, Whiting B. Variability in the pharmacokinetics of epirubicin: a population analysis. Cancer Chemother Pharmacol 1992; 29:391–5. 74. Cusack BJ, Mushlin PS, Johnson CJ, Vestal RE, Olson RD. Aging increases the cardiotoxicity of daunorubicin and daunorubicinol in the rat. J Gerontol A Biol Sci Med Sci 1996; 51:B376–84. 75. Cusack BJ, Young SP, Vestal RE, Olson RD. Age-related pharmacokinetics of daunorubicin and daunorubicinol following intravenous bolus daunorubicin administration in the rat. Cancer Chemother Pharmacol 1997; 39:505–12.

656

DeLeve

76. Arndt C, Hawkins D, Anderson JR, Breitfeld P, Womer R, Meyer W. Age is a risk factor for chemotherapy-induced hepatopathy with vincristine, dactinomycin, and cyclophosphamide. J Clin Oncol 2004; 22:1894–901. 77. Rheumatoid Arthritis Clinical Trial Archive Group. The effect of age and renal function on the efficacy and toxicity of methotrexate in rheumatoid arthritis. J Rheumatol 1995; 22:218–23. 78. Dobbs NA, Twelves CJ, Gillies H, James CA, Harper PG, Rubens RD. Gender affects doxorubicin pharmacokinetics in patients with normal liver biochemistry. Cancer Chemother Pharmacol 1995; 36:473–6. 79. Grenier MA, Lipshultz SE. Epidemiology of anthracycline cardiotoxicity in children and adults. Semin Oncol 1998; 25:72–85. 80. Lipshultz SE, Lipsitz SR, Mone SM, et al. Female sex and higher drug dose as risk factors for late cardiotoxic effects of doxorubicin therapy for childhood cancer. N Engl J Med 1995; 332:1738–43. 81. Silber JH, Jakacki RI, Larsen RL, Goldwein JW, Barber G. Increased risk of cardiac dysfunction after anthracyclines in girls. Med Ped Oncol 1993; 21:477–9 (see comments). 82. Zalcberg J, Kerr D, Seymour L, Palmer M. Haematological and non-haematological toxicity after 5-fluorouracil and leucovorin in patients with advanced colorectal cancer is significantly associated with gender, increasing age and cycle number. Tomudex International Study Group. Eur J Cancer 1998; 34:1871–5. 83. Milano G, Etienne MC, Pierrefite V, Barberi-Heyob M, Deporte-Fety R, Renee N. Dihydropyrimidine dehydrogenase deficiency and fluorouracil-related toxicity. Br J Cancer 1999; 79:627–30. 84. Bressolle F, Joulia JM, Pinguet F, Ychou M, Astre C, Duffour J, Gomeni R. Circadian rhythm of 5-fluorouracil population pharmacokinetics in patients with metastatic colorectal cancer. Cancer Chemother Pharmacol 1999; 44:295–302. 85. Chazal M, Etienne MC, Rene´e N, Bourgeon A, Richelme H, Milano G. Link between dihydropyrimidine dehydrogenase activity in peripheral blood mononuclear cells and liver. Clin Cancer Res 1996; 2:507–10. 86. Lu Z, Zhang R, Carpenter JT, Diasio RB. Decreased dihydropyrimidine dehydrogenase activaty in a population of patients with breast cancer: implication for 5-fluorouracil-based chemotherapy. Clin Cancer Res 1998; 4:325–9. 87. Walter-Sack I, Klotz U. Influence of diet and nutritional status on drug metabolism. Clin Pharmacokinet 1996; 31:47–64. 88. Davis LE, Lenkinski RE, Shinkwin MA, Kressel HY, Daly JM. The effect of dietary protein depletion on hepatic 5-fluorouracil metabolism. Cancer 1993; 72:3715–22. 89. Cusack BJ, Young SP, Loseke VL, Hurty MR, Beals L, Olson RD. Effect of a low-protein diet on doxorubicin pharmacokinetics in the rabbit. Cancer Chemother Pharmacol 1992; 30:145–8. 90. DeLeve LD. Cellular target of cyclophosphamide toxicity in the murine liver: role of glutathione and site of metabolic activation. Hepatology 1996; 24:830–7. 91. Wang X, Kanel GC, DeLeve LD. Support of sinusoidal endothelial cell glutathione prevents hepatic veno-occlusive disease in the rat. Hepatology 2000; 31:428–34. 92. Geraci JP, Mariano MS, Jackson KL. Radiation hepatology of the rat: microvascular fibrosis and enhancement of liver dysfunction by diet and drugs. Radiat Res 1992; 129:322–32. 93. Iyer L, Ratain MJ. Pharmacogenetics and cancer chemotherapy. Eur J Cancer 1998; 34:1493–9. 94. Carlini LE, Meropol NJ, Bever J, et al. UGT1A7 and UGT1A9 polymorphisms predict response and toxicity in colorectal cancer patients treated with capecitabine/irinotecan. Clin Cancer Res 2005; 11:1226–36. 95. Gagne J-F, Montminy V, Belanger P, Journault K, Gaucher G, Guillemette C. Common human UGT1A polymorphisms and the altered metabolism of irinotecan active metabolite 7-ethyl10-hydroxycamptothecin (SN-38). Mol Pharmacol 2002; 62:608–17. 96. Hanioka N, Ozawa S, Jinno H, Ando M, Saito Y, Sawada J. Human liver UDP-glucuronosyltransferase isoforms involved in the glucuronidation of 7-ethyl-10-hydroxycamptothecin. Xenobiotica 2001; 31:687–99. 97. Massacesi C, Terrazzino S, Marcucci F, et al. Uridine diphosphate glucuronosyl transferase 1A1 promoter polymorphism predicts the risk of gastrointestinal toxicity and fatigue induced by irinotecan-based chemotherapy. Cancer 2006; 106:1007–16. 98. Bosma PJ, Roy Chowdhury J, Bakker C, et al. The genetic basis of the reduced expression of bilirubin UDP-glucuronosyltransferase I in Gilbert’s syndrome. N Engl J Med 1998; 333:1171–5. 99. Wasserman E, Myara A, Lokiec F, et al. Severe CPT-11 toxicity in patients with Gilbert’s syndrome: two case reports. Ann Oncol 1997; 8:1049–51. 100. Paoluzzi L, Singh AS, Price DK, et al. Influence of genetic variants in UGT1A1 and UGT1A9 on the in vivo glucuronidation of SN-38. J Clin Pharmacol 2004; 44:854–60. 101. Sai K, Saeki M, Saito Y, Ozawa S, Katori N, Jinno H, Hasegawa R, Kaniwa N, Sawada J-I, Komamura K, et al. UGT1A1 haplotypes associated with reduced glucuronidation and increased serum bilirubin in irinotecan-administered Japanese patients with cancer. Clin Pharmacol Ther 2004; 75:501–15.

Cancer Chemotherapy

657

102. Han J-Y, Lim H-S, Shin ES, et al. Comprehensive analysis of UGT1A polymorphisms predictive for pharmacokinetics and treatment outcome in patients with non-small-cell lung cancer treated with irinotecan and cisplatin. J Clin Oncol 2006; 24:2237–44. 103. Strasser SI, McDonald GB. Gastrointestinal and hepatic complications. In: Thomas ED, Blume KG, Forman SJ, eds. Hematopoietic Cell Transplantation. Cambridge, MA: Blackwell Scientific Publications, 1999:627–58. 104. Willmot FC, Robertson GW. Senecio disease, or cirrhosis of the liver due to senecio poisoning. Lancet 1920; 2:848–9. 105. McDonald GB. Review article: management of hepatic disease following haematopoietic cell transplant. Aliment Pharmacol Ther 2006; 24:441–52. 106. McDonald GB, Hinds MS, Fisher LD, et al. Veno-occlusive disease of the liver and multiorgan failure after bone marrow transplantation: a cohort study of 355 patients. Ann Intern Med 1993; 118:255–67. 107. Hommeyer SC, Teefey SA, Jacobson AF, Higano CS, Bianco JA, Colacurcio CJ, McDonald GB. Venocclusive disease of the liver: prospective study of US evaluation. Radiology 1992; 184:683–6. 108. McCarville MB, Hoffer FA, Howard SC, Goloubeva O, Kauffman WM. Hepatic veno-occlusive disease in children undergoing bone-marrow transplantation: usefulness of sonographic findings. Pediatr Radiol 2001; 31:102–5. 109. Teefey SA, Brink JA, Borson RA, Middleton WD. Diagnosis of venoocclusive disease of the liver after bone marrow transplantation: value of duplex sonography. AJR 1995; 164:1397–401. 110. Shulman HM, Gooley T, Dudley MD, Kofler T, Feldman R, Dwyer D, McDonald GB. Utility of transvenous liver biopsies and wedged hepatic venous pressure measurements in sixty marrow transplant recipients. Transplantation 1995; 59:1015–22. 111. Carreras E, Granena A, Navasa M, et al. Transjugular liver biopsy in BMT. Bone Marrow Transplant 1993; 11:21–6. 112. Deeg HJ, Shulman HM, Schmidt E, Yee GC, Thomas ED, Storb R. Marrow graft rejection and venoocclusive disease of the liver in patients with aplastic anemia conditioned with cyclophosphamide and cyclosporine. Transplantation 1986; 42:497–501. 113. Peters WP, Henner WD, Grochow LB, et al. Clinical and pharmacological effects of high dose single agent busulfan with autologous bone marrow support in the treatment of solid tumors. Cancer Res 1987; 47:6402–6. 114. Ayash LJ, Elias A, Wheeler C, et al. Double dose-intensive chemotherapy with autologous marrow and peripheral-blood progenitor-cell support for metastatic breast cancer: a feasibility study. J Clin Oncol 1994; 12:37–44. 115. Singhal S, Powles R, Treleaven J, Horton C, Swansbury GJ, Mehta J. Melphalan alone prior to allogeneic bone marrow transplantation from HLA-identical sibling donors for hematologic malignancies: alloengraftment with potential preservation of fertility in women. Bone Marrow Transplant 1996; 18:1049–55. 116. Moreau P, Fiere D, Bezwoda WR, et al. Prospective randomized placebo-controlled study of granulocyte-macrophage colony-stimulating factor without stem-cell transplantation after highdose melphalan in patients with multiple myeloma. J Clin Oncol 1997; 15:660–6. 117. Fajardo LF, Berthrong M. Radiation injury in surgical pathology. Am J Surg 1978; 2:159–99. 118. Lee JL, Gooley T, Bensinger W, Schiffman K, McDonald GB. Veno-occlusive disease of the liver after busulfan, melphalan, and thiotepa conditioning therapy: incidence, risk factors, and outcome. Biol Blood Marrow Transplant 1999; 5:306–15. 119. Srivastava A, Bradstock KF, Szer J, de Bortoli L, Gottlieb DJ. Busulphan and melphalan prior to autologous bone marrow transplantation. Bone Marrow Transplant 1993; 12:323–9. 120. McCune JS, Gibbs JP, Slattery JT. Plasma concentration monitoring of busulfan. Does it improve clinical outcome? Clin Pharmacokinet 2000; 39:155–65. 121. McDonald GB, Slattery JT, Bouvier ME, et al. Cyclophosphamide metabolism, liver toxicity, and mortality following hematopoietic stem cell transplantation. Blood 2003; 101:2043–8. 122. McDonald GB, McCune JS, Batchelder A, et al. Metabolism-based cyclophosphamide dosing for hematopoietic cell transplant. Clin Pharmacol Ther 2005; 298–308. 123. Rajvanshi P, Shulman HM, Sievers EL, McDonald GB. Hepatic sinusoidal obstruction following gemtuzumab ozogamicin (Mylotargw). Blood 2002; 99:2310–4. 124. Wadleigh M, Richardson PG, Zahrieh D, et al. Prior gemtuzumab ozogamicin exposure significantly increases the risk of veno-occlusive disease in patients who undergo myeloablative allogeneic stem cell transplantation. Blood 2003; 102:1578–82. 125. Shulman HM, Fisher LB, Schoch HG, Henne KW, McDonald GB. Venoocclusive disease of the liver after marrow transplantation: histological correlates of clinical signs and symptoms. Hepatology 1994; 19:1171–80. 126. Rollins BJ. Hepatic veno-occlusive disease. Am J Med 1986; 81:297–306. 127. DeLeve LD. Dacarbazine toxicity in murine liver cells: a novel model of hepatic endothelial injury and glutathione defense. J Pharmacol Exp Ther 1994; 268:1261–70.

658

DeLeve

128. DeLeve LD, Wang X, Kuhlenkamp JF, Kaplowitz N. Toxicity of azathioprine and monocrotaline in murine sinusoidal endothelial cells and hepatocytes: the role of glutathione and relevance to hepatic venooclusive disease. Hepatology 1996; 23:589–99. 129. DeLeve LD, McCuskey RS, Wang X, Hu L, McCuskey MK, Epstein RB, Kanel G. Characterization of a reproducible rat model of hepatic veno-occlusive disease. Hepatology 1999; 29:1779–91. 130. DeLeve LD, Ito Y, Machen NW, McCuskey MK, Wang X, McCuskey RS. Embolization by sinusoidal lining cells causes the congestion of hepatic venoocclusive disease. Gastroenterology 2000; 118:A2345. 131. DeLeve LD, Wang X, Tsai J, Kanel G, Tokes Z. Prevention of hepatic venooclusive disease in the rat by inhibition of matrix metalloproteinases. Gastroenterology 2001; 120:A54. 132. Cai T, Fassina GF, Morini M, et al. N-acetylcysteine inhibits endothelial cell invasion and angiogenesis. Lab Invest 1999; 79:1151–9. 133. Tyagi SC, Ratajska A, Weber KT. Myocardial matrix metalloproteinase(s): localization and activation. Mol Cell Biochem 1993; 126:49–59. 134. Tyagi SC, Kumar S, Borders S. Reduction–oxidation (redox) state regulation of extracellular matrix metalloproteinases and tissue inhibitors in cardiac normal and transformed fibroblast cells. J Cell Biochem 1996; 61:139–51. 135. Upadhya GA, Strasberg SM. Glutathione, lactobionate, and histidine: cryptic inhibitors of matrix metalloproteinases contained in University of Wisconsin and histidine/tryptophan/ketoglutarate liver preservation solutions. Hepatology 2000; 31:1115–22. 136. DeLeve LD, Wang X, Kanel GC, et al. Decreased hepatic nitric oxide production contributes to the development of rat sinusoidal obstruction syndrome. Hepatology 2003; 38:900–8. 137. Korte W. Veno-occlusive disease of the liver after bone marrow transplantation: is hypercoagulability really part of the problem? Blood Coagul Fibrinolysis 1997; 8:367–81. 138. Shulman HM, Gown AM, Nugent DJ. Hepatic veno-occlusive disease after bone marrow transplantation—immunohistochemical identification of the material within occluded central venules. Am J Pathol 1987; 127:549–58. 139. Brooks SEH, Miller CG, McKenzie K, Audretsch JJ, Bras G. Acute veno-occlusive disease of the liver. Arch Path 1970; 89:507–20. 140. Snover DC, Weisdorf S, Bloomer J, McGlave P, Weisdorf D. Nodular regenerative hyperplasia of the liver following bone marrow transplantation. Hepatology 1989; 9:443–8. 141. Wanless IR, Godwin TA, Allen F, Feder A. Nodular regenerative hyperplasia of the liver in hematologic disorders: a possible response to obliterative portal venopathy. A morphometric study of nine cases with an hypothesis on the pathogenesis. Medicine 1980; 59:367–79. 142. Thomas PA, McCusker JJ, Merrigan EH, Conte NF. Lobar cirrhosis with nodular hyperplasia (hamartoma) of the liver treated by left hepatic lobectomy. Am J Surg 1966; 112:831–4. 143. Whelan TJ, Baugh JH, Chandor S. Focal nodular hyperplasia of the liver. Ann Surg 1973; 177:150–8. 144. Travers H, D’Amato NA. Vascular alterations in focal nodular hyperplasia of the liver. Mil Med 1978; 143:96–101. 145. Wanless IR. Micronodular transformation (nodular regenerative hyperplasia) of the liver: a report of 64 cases among 2,500 autopsies and a new classification of benign hepatocellular nodules. Hepatology 1990; 11:787–97. 146. Haboubi NY, Ali HH, Whitwell HL, Ackrill P. Role of endothelial cell injury in the spectrum of azathioprine-induced liver disease after renal transplant: light microscopy and ultrastructural observations. Am J Gastroenterol 1988; 83:256–61. 147. Bo¨hme M, Mu¨ller M, Leier I, Jedlitschky G, Keppler D. Cholestasis caused by inhibition of the adenosine triphosphate-dependent bile salt transport in rat liver. Gastroenterology 1994; 107:255–65. 148. Kadmon M, Klu¨nemann C, Bo¨hme M, et al. Inhibition by cyclosporin A of adenosine triphosphatedependent transport from the hepatocyte into bile. Gastroenterology 1993; 104:1507–14. 149. Bo¨hme M, Bu¨chler M, Mu¨ller M, Keppler D. Differential inhibition by cyclosporins of primary-active ATP-dependent transporters in the hepatocyte canalicular membrane. FEBS Lett 1993; 333:193–6. 150. Moran D, De Buitrago JM, Fernandez E, Galan AI, Munoz ME, Jimenez R. Inhibition of biliary glutathione secretion by cyclosporine A in the rat: possible mechanisms and role in the cholestasis induced by the drug. J Hepatol 1998; 29:68–77. 151. Yasumiba S, Tazuma S, Ochi H, Chayama K, Kajiyama G. Cyclosporin A reduces canalicular membrane fluidity and regulates transporter function in rats. Biochem J 2001; 354:591–6. 152. Azer SA, Stacey NH. Differential effects of cyclosporin A on the transport of bile acids by human hepatocytes. Biochem Pharmacol 1993; 46:813–9. 153. Mosely RH, Johnson TR, Morrissette JM. Inhibition of bile acid transport by cyclosporine A in rat liver plasma membrane vesicles. J Pharmacol Exp Ther 1990; 253:974–80. 154. Tamai I, Safa AR. Competitive inhibition of cyclosporins with the vinca alkaloid-binding site of P-glycoprotein in multidrug-resistant cells. J Biol Chem 1990; 265:16509–13.

Cancer Chemotherapy

659

155. Roma´n ID, Monte MJ, Gonzalez-Buitrago JM, Esteller A, Jimenez R. Inhibition of hepatocytary vesicular transport by cyclosporin A in the rat: relationship with cholestasis and hyperbilirubinemia. Hepatology 1990; 12:83–91. 156. Roma´n ID, Coleman R. Disruption of canalicular function in isolated rat hepatocyte couplets caused by cyclosporin A. Biochem Pharmacol 1994; 48:2181–8. 157. Maria VA, Victorino RM. Development and validation of a clinical scale for the diagnosis of druginduced hepatitis. Hepatology 1997; 26:664–9. 158. Danan G, Benichou C. Causality assessment of adverse reactions to drugs—I. A novel method based on the conclusions of international consensus meetings: application to drug-induced liver injuries. J Clin Epidemiol 1993; 46:1323–30. 159. Mori T, Sugita K, Suzuki T, Ishikawa T, Kurosawa H, Matsui A. Histopathologic features of the biopsied liver at onset of childhood B-precursor acute lymphoblastic leukemia presenting as severe jaundice. J Pediatr Gastroenterol Nutr 1997; 25:354–7. 160. Dubinsky MC, Vasiliauskas EA, Singh H, et al. 6-Thioguanine can cause serious liver injury in inflammatory bowel disease patients. Gastroenterology 2003; 125:298–303. 161. Geller SA, Dubinsky MC, Poordad FF, et al. Early hepatic nodular hyperplasia and submicroscopic fibrosis associated with 6-thioguanine therapy in inflammatory bowel disease. Am J Surg Pathol 2004; 28:1204–11. 162. Seiderer J, Zech CJ, Reinisch W, et al. A multicenter assessment of liver toxicity by MRI and biopsy in IBD patients on 6-thioguanine. J Hepatol 2005; 43:303–9. 163. Mathews V, Desire S, George B, et al. Hepatotoxicity profile of single agent arsenic trioxide in the treatment of newly diagnoses acute promyelocytic leukemia, its impact on clinical outcome and the effect of genetic polymorphisms on the incidence of hepatotoxicity. Leukemia 2006; 20:881–3. 164. Bastida G, Nos P, Aguas M, Beltran B, Rubin A, Dasi F, Ponce J. Incidence, risk factors and clinical course of thiopurine-induced liver injury in patients with inflammatory bowel disease. Aliment Pharmacol Ther 2005; 22:775–82. 165. Ingold JA, Reed GB, Jr., Kaplan HS, Bagshaw MA. Radiation hepatitis. AJR 1965; 93:200–8. 166. Reed GB, Jr., Cox AJ, Jr. The human liver after radiation injury: a form of veno-occlusive disease. Am J Pathol 1966; 48:597–611. 167. Selzer G, Parker RGF. Senecio poisoning exhibiting as Chiari’s syndrome: a report on twelve cases. Am J Pathol 1950; 27:885–907. 168. Wiernik PH, Dutcher JP, Paietta E, et al. Long-term follow-up of treatment and potential cure of adult acute lymphocytic leukemia with MOAD: a non-anthracycline containing regimen. Leukemia 1993; 7:1236–41. 169. McGinn CJ, Ten Haken RK, Ensminger WD, Walker S, Wang S, Lawrence TS. Treatment of intrahepatic cancers with radiation doses based on a normal tissue complication probability model. J Clin Oncol 1998; 16:2246–52. 170. Grinsky T, Benhamou E, Bourhis JH, Dhermain F, Guillot-Valls D, Ganansia V, Luboinski M, Perez A, Cosset J-M, Socie G, et al. Prospective randomized comparison of single-dose versus hyperfractionated total-body irradiation in patients with hematologic malignancies. J Clin Oncol 2000; 18:981–6. 171. Ortega JA, Donaldson SS, Ivy SP, Pappo A, Maurer HM. Venoocclusive disease of the liver after chemotherapy with vincristine, actinomycin D, and cyclophosphamide for the treatment of rhabdomyosarcoma. Cancer 1997; 79:2435–9. 172. Saibara T, Onishi S, Ogawa Y, Yoshida S, Enzan H. Non-alcoholic steatohepatitis. Lancet 1999; 354:1299–300. 173. Van Hoof M, Rahier J, Horsmans Y. Tamoxifen-induced steatohepatitis. Ann Intern Med 1996; 124:855–6 (letter). 174. Pratt DS, Knox TA, Erban J. Tamoxifen-induced steatohepatitis. Ann Intern Med 1995; 123:236 (letter). 175. Oien KA, Moffat D, Curry GW, Dickson J, Habeshaw T, Mills PR, MacSween RN. Cirrhosis with steatohepatitis after adjuvant tamoxifen. Lancet 1999; 353:36–7 (letter). 176. Cortez Pinto H, Baptista A, Camilo ME, Bruno de Costa E, Valente A, Carneiro de Moura M. Tamoxifen-associated steatohepatitis—report of three cases. J Hepatol 1995; 23:95–7. 177. Murata Y, Ogawa Y, Saibara T, et al. Unrecognized hepatic steatosis and non-alcoholic steatohepatitis in adjuvant tamoxifen for breast cancer patients. Oncol Rep 2000; 7:1299–304. 178. Loomus GN, Aneja P, Bota RA. A case of peliosis hepatis in association with tamoxifen therapy. Am J Clin Pathol 1983; 80:881–3. 179. Santini D, Picardi A, Vincenzi B, Binetti P, Massacesi C, La Cesa A, Tonini G. Severe liver dysfunction after raltitrexed administration in an HCV-positive colorectal cancer patient. Cancer Invest 2003; 21:162–3. 180. Raderer M, Fiebiger W, Wrba F, Scheithauer W. Fatal liver failure after the administration of raltitrexed for cancer chemotherapy: a report of two cases. Cancer 2000; 89:890–2.

660

DeLeve

181. Storen EC, Hay JE, Kaur J, Zahasky K, Hartmann L. Tamoxifen-induced submassive hepatic necrosis. Cancer J 2000; 6:58–60. 182. Moffat DF, Oien KA, Dickson J, Habeshaw T, McLellan DR. Hepatocellular carcinoma after longterm tamoxifen therapy. Ann Oncol 2000; 11:1195–6. 183. Modzelewski JR, Jr., Daeschner C, Joshi VV, Mullick FG, Ishak KG. Veno-occlusive disease of the liver induced by low-dose cyclophosphamide. Mod Pathol 1994; 7:967–72. 184. Patel JM, Block ER. Acrolein-induced injury to cultured pulmonary artery endothelial cells. Toxicol Appl Pharmacol 1993; 122:46–53. 185. Kachel DL, Martin WJ, II. Cyclophosphamide-induced lung toxicity: mechanism of endothelial cell injury. J Pharmacol Exp Ther 1994; 268:42–6. 186. Shepherd PC, Fooks J, Gray R, Allan NC. Thioguanine used in maintenance therapy of chronic myeloid leukaemia causes non-cirrhotic portal hypertension. Results from MRC CML. II. Trial comparing busulphan with busulphan and thioguanine. Br J Haematol 1991; 79:185–92. 187. Shulman HM, McDonald GB, Matthews D, Doney KC, Kopecky KJ, Gauvreau JM, Thomas ED. An analysis of hepatic venocclusive disease and centrilobular hepatic degeneration following bone marrow transplantation. Gastroenterology 1980; 79:1178–91. 188. Kanfer EJ, Petersen FB, Buckner CD, et al. Phase I study of high-dose dimethylbusulfan followed by autologous bone marrow transplantation in patients with advanced malignancy. Cancer Treat Rep 1987; 71:101–2. 189. DeLeve LD, Wang X. Role of oxidative stress and glutathione in busulfan toxicity in cultured murine hepatocytes. Pharmacology 2000; 60:143–54. 190. Paschke R, Heine M. Pathophysiological aspects of dacarbazine-induced human liver damage. Hepatogastroenterology 1985; 32:273–5. 191. Hill DL. Microsomal metabolism of triazenylimidazoles. Cancer Res 1975; 35:3106–10. 192. Lazarus HM, Herzig RH, Graham-Pole J, et al. Intensive melphalan chemotherapy and cryopreserve autologous bone marrow transplantation for the treatment of refractory cancer. J Clin Oncol 1983; 1:359–67. 193. Awasthi S, Bajpai KK, Piper JT, et al. Interactions of melphalan with glutathione and the role of glutathione S-transferase. Drug Metab Dispos 1996; 24:371–4. 194. Paumi CM, Ledford BG, Smitherman PK, Townsend AJ, Morrow CS. Role of multidrug resistance protein 1 (MRP1) and glutathione S-transferase A1-1 in alkylating agent resistance. Kinetics of glutathione conjugate formation and efflux govern differential cellular sensitivity to chlorambucil versus melphalan toxicity. J Biol Chem 2001; 276:7952–6. 195. Vahrmeijer AL, van Dierendonck JH, Schutrups J, van de Velde CJ, Mulder GJ. Effect of glutathione depletion on inhibition of cell cycle progression and induction of apoptosis by melphalan (Lphenylalanine mustard) in human colorectal cancer cells. Biochem Pharmacol 1999; 58:655–64. 196. Hutter RVP, Shipkey FH, Tan CTC, Murphy ML, Chowdhury M. Hepatic fibrosis in children with acute leukemia. A complication of therapy. Cancer 1960; 13:288–307. 197. Weber BL, Tanyer G, Poplack DG, Reaman GH, Feusner JH, Miser JS, Bleyer WA. Transient acute hepatotoxicity of high-dose methotrexate therapy during childhood. NCI Monogr 1987; 207–12. 198. Schmiegelow K, Pulczynska M. Prognostic significance of hepatotoxicity during maintenance chemotherapy for childhood acute lymphoblastic leukaemia. Br J Cancer 1990; 61:767–72. 199. Rask C, Albertioni F, Bentzen SM, Schroeder H, Peterson C. Clinical and pharmacokinetic risk factors for high-dose methotrexate-induced toxicity in children with acute lymphoblastic leukemia—a logistic regression analysis. Acta Oncol 1998; 37:277–84. 200. Ruymann FB, Mosjczuk AD, Sayers RJ. Hepatoma in a child with methotrexate-induced hepatic fibrosis. JAMA 1977; 238:2631–3. 201. Fried M, Kalra J, Ilardi CF, Sawitsky A. Hepatocellular carcinoma in a long-term survivor of acute lymphocytic leukemia. Cancer 1987; 60:2548–52. 202. Whiting-O’Keefe QE, Fye KH, Sack KD. Methotrexate and histologic hepatic abnormalities: a meta analysis. Am J Med 1991; 90:711–6. 203. Frei E, III, Holland JF, Schneiderman MA, et al. A comparative study of two regimens of combination chemotherapy in acute leukemia. Blood 1958; 13:1126–48. 204. Farber S. Summary of experience with 6-mercaptopurine. Ann NY Acad Sci 1954; 60:412–4. 205. McIlvanie SK, MacCarthy JD. Hepatitis in association with prolonged 6-mercaptopurine therapy. Blood 1959; 14:80–90. 206. Einhorn M, Davidsohn I. Hepatotoxicity of mercaptopurine. JAMA 1964; 188:802–6. 207. Present DH, Meltzer SJ, Krumholz MP, Wolke A, Korelitz BI. 6-Mercaptopurine in the management of inflammatory bowel disease: short- and long-term toxicity. Ann Int Med 1989; 111:641–9. 208. Krynetski EY, Evans WE. Pharmacogenetics as a molecular basis for individualized drug therapy: the thiopurine S-methyltransferase paradigm. Pharm Res 1999; 16:342–9. 209. McLeod HL, Krynetski EY, Relling MV, Evans WE. Genetic polymorphism of thiopurine methyltransferase and its clinical relevance for childhood acute lymphoblastic leukemia. Leukemia 2000; 14:567–72.

Cancer Chemotherapy

661

210. Relling MV, Hancock ML, Rivera GK, et al. Mercaptopurine therapy intolerance and heterozygosity at the thiopurine S-methyltransferase gene locus. J Natl Cancer Inst 1999; 91:2001–8 (see comments). 211. Berkovitch M, Matsui D, Zipursky A, et al. Hepatotoxicity of 6-mercaptopurine in childhood acute lymphocytic leukemia: pharmacokinetic characteristics. Med Ped Oncol 1996; 26:85–9. 212. Dubinsky MC, Lamothe S, Yang HY, Targan SR, Sinnett D, Theoret Y, Seidman EG. Pharmacogenomics and metabolite measurement for 6-mercaptopurine therapy in inflammatory bowel disease. Gastroenterology 2000; 118:705–13. 213. Dervieux T, Blanco JG, Krynetski EY, Vanin EF, Roussel MF, Relling MV. Differing contribution of thiopurine methyltransferase to mercaptopurine versus thioguanine effects in human leukemic cells. Cancer Res 2001; 61:5810–6. 214. Merino JM, Casanova F, Sa´ez-Royuela F, Velasco A, Gonzalez JB. Veno-occlusive disease of the liver associated with thiopurines in a child with acute lymphoblastic leukemia. Ped Hematol Oncol 2000; 17:429–31. 215. Griner PF, Elbadawi A, Packman CH. Veno-occlusive disease of the liver after chemotherapy of acute leukemia. Report of two cases. Ann Intern Med 1976; 85:578–82. 216. D’Cruz CA, Wimmer RS, Harcke HT, Huff DS, Naiman JL. Veno-occlusive disease of the liver in children following chemotherapy for acute myelocytic leukemia. Cancer 1983; 52:1803–7. 217. Satti MB, Weinbren K, Gordon-Smith EC. 6-Thioguanine as a cause of toxic veno-occlusive disease of the liver. J Clin Pathol 1982; 35:1086–91. 218. de Araujo MS, Gerard F, Chossegros P, Porto LC, Barlet P, Grimaud JA. Vascular hepatotoxicity related to heroin addiction. Virchows Arch A Pathol Anat Histopathol 1990; 417:497–503. 219. Trigueiro de Araujo MS, Gerard F, Chossegros P, Guerret S, Grimaud JA. Lack of hepatocyte involvement in the genesis of the sinusoidal dilatation related to heroin addiction: a morphometric study. Virchows Arch A Pathol Anat Histopathol 1992; 420:149–53. 220. Zafrani ES, Cazier A, Baudelot AM, Feldmann G. Ultrastructural lesions of the liver in human peliosis. A report of 12 cases. Am J Pathol 1984; 114:349–59. 221. DeLeve LD. Glutathione defense in non-parenchymal cells. Sem Liver Dis 1998; 18:403–13. 222. Porta C, Danova M, Accurso S, Tinelli C, Girino M, Riccardi A, Palmeri S. Sequential intrahepatic and systemic fluoropyrimidine-based chemotherapy for metastatic colorectal cancer confined to the liver. A phase II study. Cancer Chemother Pharmacol 2001; 47:423–8. 223. Kemeny N, Gonen M, Sullivan D, et al. Phase I study of hepatic arterial infusion of floxuridine and dexamethasone with systemic irinotecan for unresectable hepatic metastases from colorectal cancer. J Clin Oncol 2001; 19:2687–95. 224. Campos LT. A randomized trial of intrahepatic infusion of fluorodeoxyuridine with dexamethasone versus fluorodeoxyuridine alone in the treatment of metastatic colorectal cancer. Cancer 1993; 71:875–6 (letter comment). 225. Kemeny N, Seiter K, Niedzwiecki D, et al. A randomized trial of intrahepatic infusion of fluorodeoxyuridine with dexamethasone versus fluorodeoxyuridine alone in the treatment of metastatic colorectal cancer. Cancer 1992; 69:327–34. 226. Brown KT, Kemeny N, Berger MF, et al. Obstructive jaundice in patients receiving hepatic artery infusional chemotherapy: etiology, treatment implications, and complications after transhepatic biliary drainage. J Vasc Interv Radiol 1997; 8:229–34. 227. Ludwig J, Kim CH, Wiesner RH, Krom RA. Floxuridine-induced sclerosing cholangitis: an ischemic cholangiopathy? Hepatology 1989; 9:215–8. 228. Press OW, Shan D, Howell-Clark J, et al. Comparative metabolism and retention of iodine-125, yttrium-90, and indium-111 radioimmunoconjugates by cancer cells. Cancer Res 1996; 56:2123–9. 229. van Der Velden VH, te Marvelde JG, Hoogeveen PG, Bernstein ID, Houtsmuller AB, Berger MS, van Dongen JJ. Targeting of the CD33-calicheamicin immunoconjugate Mylotarg (CMA-676) in acute myeloid leukemia: in vivo and in vitro saturation and internalization by leukemic and normal myeloid cells. Blood 2001; 97:3197–204. 230. Nicolaou K, Smith A, Yue E. Chemistry and biology of natural and designed enediynes. Proc Natl Acad Sci USA 1993; 90:5881–8. 231. Dedon PC, Goldberg IH. Free-radical mechanisms involved in the formation of sequence-dependent bistranded DNA lesions by the antitumor antibiotics bleomycin, neocarzinostatin, and calicheamicin. Chem Res Toxicol 1992; 5:311–32. 232. Gao Z, McAlister VC, Williams GM. Repopulation of liver endothelium by bone-marrow-derived cells. Lancet 2001; 357:932–3. 233. Takezawa RI, Watanabe Y, Akaike T. Direct evidence of macrophage differentiation from bone marrow cells in the liver: a possible origin of Kupffer cells. J Biochem 1995; 118:1175–83. 234. Durden DL, Salazar AM, Distasio JA. Kinetic analysis of hepatotoxicity associated with antineoplastic asparaginases. Cancer Res 1983; 43:1602–5. 235. Villa P, Corada M, Bartosek Il-. L-asparaginase effects on inhibition of protein synthesis and lowering of the glutamine content in cultured rat hepatocytes. Toxicol Lett 1986; 32:235–41.

662

DeLeve

236. Land VJ, Sutow WW, Fernbach DJ, Lane DM, Williams TE. Toxicity of L-asparginase in children with advanced leukemia. Cancer 1972; 30:339–47. 237. Pratt CB, Johnson WW. Duration and severity of fatty metamorphosis of liver following L-asparaginase therapy. Cancer 1971; 28:361–4. 238. Law CH, Tandan VR. The association between tamoxifen and the development of hepatocellular carcinoma: case report and literature review. Can J Surg 1999; 42:211–4. 239. Chim CS, Kwong YL, Chu YC, Chan CH, Chan YT, Liang R. Improved treatment outcome in adult acute lymphoblastic leukemia using the intensive German protocol, a preliminary report. Hematol Oncol 1997; 15:19–26. 240. Rizzari C, Valsecchi MG, Arico M, et al. Effect of protracted high-dose L-asparaginase given as a second exposure in a Berlin-Frankfurt-Mu¨nster-based treatment: results of the randomized 9102 intermediate-risk childhood acute lymphoblastic leukemia study—a report from the Associazione Italiana Ematologia Oncologia Pediatrica. J Clin Oncol 2001; 19:1297–303. 241. Silverman LB, Gelber RD, Dalton VK, et al. Improved outcome for children with acute lymphoblastic leukemia: results of Dana-Farber Consortium Protocol 91-01. Blood 2001; 97:1211–8. 242. Flentje M, Weirich A, Potter R, Ludwig R. Hepatotoxicity in irradiated nephroblastoma patients during postoperative treatment according to SIOP9/GPOH. Radiother Oncol 1994; 31:222–8. 243. Tornesello A, Piciacchia D, Mastrangelo S, Lasorella A, Mastrangelo R. Veno-occlusive disease of the liver in right-sided Wilms’ tumours. Eur J Cancer 1998; 34:1220–3. 244. Ludwig R, Weirich A, Hofmann WJ, Waldherr R. Veno-occlusive disease as hepatotoxic side effect of the nephroblastoma SIOP-9 treatment protocol: preliminary results of the German group. Med Pediatr Oncol 1992; 20:434. 245. Mushtaq I, Carachi R, Roy G, Azmy A. Childhood renal tumours with intravascular extension. Brit J Urol 1996; 78:772–6. 246. Lawrence TS, Robertson JM, Anscher MS, Jirtle RL, Ensminger WD, Fajardo LF. Hepatic toxicity resulting from cancer treatment. Int J Radiat Oncol Biol Phys 1995; 31:1237–48. 247. Bruno S, Maisonneuve P, Castellana P, et al. Incidence and risk factors for non-alcoholic steatohepatitis: prospective study of 5408 women enrolled in Italian tamoxifen chemoprevention trial. Br Med J 2005; 330:932. 248. Wilking N, Isaksson E, von Schoultz E. Tamoxifen and secondary tumours. An update. Drug Saf 1997; 16:104–17. 249. Dragan YP, Fahey S, Street K, Vaughan J, Jordan VC, Pitot HC. Studies of tamoxifen as a promoter of hepatocarcinogenesis in female Fischer F344 rats. Breast Cancer Res Treat 1994; 31:11–25. 250. Carthew P, Martin EA, White IN, et al. Tamoxifen induces short-term cumulative DNA damage and liver tumors in rats: promotion by phenobarbital. Cancer Res 1995; 55:544–7. 251. Martin EA, Rich KJ, White IN, Woods KL, Powles TJ, Smith LL. 32P-postlabelled DNA adducts in liver obtained from women treated with tamoxifen. Carcinogenesis 1995; 16:1651–4. 252. Rabe T, Feldmann K, Heinemann L, Runnebaum B. Cyproterone acetate: is it hepato- or genotoxic? Drug Saf 1996; 14:25–38. 253. Hinkel A, Berges RR, Pannek J, Schulze H, Senge T. Cyproterone acetate in the treatment of advanced prostatic cancer: retrospective analysis of liver toxicity in the long-term follow-up of 89 patients. Eur Urol 1996; 30:464–70. 254. Anonymous. Hepatic reactions with cyproterone acetate (Cyprostat, Androcur). Curr Probl Pharmacovigilance 1995; 21:1. 255. Garty BZ, Dinari G, Gellvan A, Kauli R. Cirrhosis in a child with hypothalamic syndrome and central precocious puberty treated with cyproterone acetate. Eur J Ped 1999; 158:367–70. 256. Werner S, Topinka J, Kunz S, Beckurts T, Heidecke CD, Schwarz LR, Wolff T. Studies on the formation of hepatic DNA adducts by the antiandrogenic and gestagenic drug, cyproterone acetate: 1. adduct levels in various species including man and 2. persistence and accumulation in the rat. Adv Exp Med Biol 1996; 387:253–7. 257. Schwarz LR, Werner S, Topinka J, Andrae U, Neumann I, Wolff T. The liver as origin and target of reactive intermediates exemplified by the progesterone derivative, cyproterone acetate. Adv Exp Med Biol 1996; 387:243–51. 258. Oosterlinck W, Casselman J, Mattelaer J, Van Velthoven R, Kurjatkin O, Schulman C. Tolerability and safety of flutamide in monotherapy, with orchiectomy or with LHRH-a in advanced prostate cancer patients. A Belgian multicenter study of 905 patients. Eur Urol 1996; 30:458–63. 259. Wysowski DK, Fourcroy JL. Flutamide hepatotoxicity. J Urol 1996; 155:209–12. 260. Berson A, Wolf C, Chachaty C, et al. Metabolic activation of the nitroaromatic antiandrogen flutamide by rat and human cytochrome P-450, including froms belonging to the 3A and 1A subfamilies. J Pharmacol Exp Ther 1983; 265:366–72. 261. Fau D, Eugene D, Berson A, Letteron P, Fromenty B, Fisch C, Pessayre D. Toxicity of the antiandrogen flutamide in isolated rat hepatocytes. J Pharmacol Exp Ther 1994; 269:954–62.

30

Hepatotoxicity of Immunomodulating Agents and the Transplant Situation Timothy J. Davern

Gastroenterology Division and Liver Transplant Program, University of California, San Francisco, California, U.S.A.

INTRODUCTION An expanding array of agents are being developed for the treatment of autoimmune, inflammatory, and dermatological conditions and to prevent rejection following solid organ transplantation. Indeed, the search for novel immunosuppressant and immunomodulatory agents that are effective and well tolerated is currently amongt the most active areas of pharmaceutical research and development. Although the currently available immunosuppressant and immunomodulatory drugs have generally well-defined toxicities, hepatotoxicity is being increasingly recognized with some of these agents. Nonetheless, severe liver injury from these drugs is generally rare, appearing infrequently in several large surveys of drug-induced liver injury (1–4). Immunosuppressant-related liver injury represents an apparent paradox because it is generally well accepted that the innate and adaptive immune systems participate in the genesis and evolution of most if not all forms of liver injury (5,6). Indeed, most of the immunosuppressants described in this chapter have been used at least in the context of clinical trials for the treatment of liver disease. For example, uncontrolled studies report a potentially beneficial effect from corticosteroid therapy for some causes of drug-induced liver injury, particularly if there are autoimmune or hypersensitivity features (7). Causality assessment of liver injury that occurs in the setting of immunosuppressive therapy is often challenging for several reasons. First, immunosuppressant medications are often used in combinations, and defining the role of an individual drug can be difficult if not impossible. The situation after a liver transplant is further complicated by multiple other potential etiologies of abnormal liver biochemistries, including acute and chronic rejection, preservation injury, bile duct obstruction, vascular insufficiency, and recurrent disease. It is therefore very likely that low-level drug-induced liver injury often goes unrecognized in this setting. Second, in patients with an autoimmune diathesis, the possibility of previously unrecognized autoimmune liver disease must always be considered, and the differentiation of autoimmune hepatitis from drug-induced liver injury may be particularly difficult. Third, immunosuppression by definition compromises the host immune system and thereby predisposes to infection. As a consequence, it is quite likely that more episodes of acute liver injury occurring in the setting of immunosuppression reflect sepsis-induced liver dysfunction rather than direct hepatotoxicity per se. Immunosuppression can also result in reactivation of chronic viral hepatitis, particularly hepatitis B, which may be misattributed to drug-induced liver injury in the absence of appropriate serological and other testing. The possibility of hepatitis B reactivation following immunosuppression is an issue that deserves special attention and is discussed further below. Finally, the development of some forms of drug-induced liver injury due to immunosuppressants, such as nodular regenerative hyperplasia (NRH), may be quite insidious, as a result of which it is difficult to recognize. Early consideration of liver biopsy may be appropriate in some situations in order to facilitate the timely recognition of significant liver injury and withdrawal of the offending agent before irreversible liver injury has occurred. This chapter outlines the clinicopathological features and mechanisms of liver injury from commonly used immunosuppressant agents. For descriptions of liver injury associated with conventional nonsteroidal anti-inflammatory drugs (chaps. 20 and 21), chemotherapeutic

664

Davern

agents (chap. 29), antiviral agents including interferon-alpha and -beta (chap. 27), and methotrexate (chap. 31), the reader is referred to appropriate chapters in the text. Reactivation of Hepatitis B Hepatitis B virus (HBV) may surreptitiously persist in hepatocytes of a previously infected individual even after apparent serological recovery and may become clinically overt with immunosuppression. Although HBV deoxynucleic acid (DNA) is almost always detectable in patients with HBV reactivation, the level of viral DNA may be declining when the patient presents with overt liver injury, occasionally causing diagnostic uncertainty. Flares may also be associated with the appearance of HBV core IgM antibody and a significant rise in serum alphafetoprotein levels, and these may be misinterpreted as indicative of acute HBV infection and hepatocellular carcinoma, respectively. Reactivation appears to be particularly common in the setting of cancer chemotherapy (e.g., for lymphoma), especially upon withdrawal of such therapy. Such HBV flares may be quite severe and result in life-threatening acute liver failure (ALF) (8). Male gender, hepatitis B surface antigen positivity, and treatment for lymphoma appear to be risk factors for clinically overt flares (9,10). Corticosteroid-based immunosuppression can also be associated with HBV flares, typically presenting after immunosuppression is withdrawn. In addition to causing general immunosuppression, corticosteroids may enhance HBV replication directly through a steroid-responsive element in the HBV genome (11). Severe hepatitis from HBV has also been reported in the setting of immunosuppression from antagonists of tumor necrosis factor (TNF) (e.g., infliximab), rituximab, and is well described following solid organ transplantation (12,13). In most cases, liver injury from hepatitis B reactivation is preventable by prophylactic use of a nucleoside analog with anti-HBV activity, such as lamivudine (100 mg p.o. q.d.), begun a few weeks before and for at least six months following the period of immunosuppression (14). Specific testing for HBV infection in patients at risk (e.g., born in an endemic area) and appropriate antiviral therapy are thus recommended before virtually any iatrogenic immunosuppression. Compared with HBV infection, the influence of immunosuppression on chronic hepatitis C infection is more variable and generally less severe, although cases of severe hepatitis C flares in the setting of immunosuppression have been reported (15–18). Active chronic hepatitis does appear to significantly increase the risk of venoocclusive disease (VOD) in patients subjected to high-dose chemotherapy and hematopoietic stem cell transplantation (19).

CORTICOSTEROIDS Corticosteroids are frequently used for a wide variety of inflammatory and autoimmune conditions and still play a fundamental role in most immunosuppressive regimens following solid organ transplantation and for the treatment of rejection. Unlike estrogens and androgenic steroids, which are frequently linked to cholestatic liver injury, glucocorticoids are rarely classified as hepatotoxic. High-dose glucocorticoids, as used in the early posttransplant setting, may be associated with macrovesicular steatosis and hepatomegaly on this basis, but overt hepatic dysfunction is quite rare. Lower doses of glucocorticoids (10–15 mg/day) when administered for prolonged periods may also result in steatosis. Glucocorticoids may also precipitate nonalcoholic steatohepatitis (NASH) in predisposed individuals by exacerbating insulin resistance, central obesity, diabetes, and hypertriglyceridemia (20). Given that glucocorticoids are used frequently and the incidence of NASH appears to be growing to epidemic proportions in developed countries, this interaction may become increasingly relevant. Exactly how glucocorticoids, which enhance fat redistribution, increase plasma free fatty acids, and inhibit hepatic fatty acid estertification, influence the pathogenic factors involved in the development and progression of NASH—including lipotoxicity, oxidative stress, various cytokines, and other proinflammatory mediators—is currently not clear (21).

Hepatotoxicity of Immunomodulating Agents

665

CALCINEURIN INHIBITORS Cyclosporin A (CsA) and tacrolimus (FK-506) are calcineurin inhibitors that have enjoyed widespread clinical use, particularly for preventing rejection of solid organ transplants, and more recently for the treatment of severe autoimmune and inflammatory diseases. Indeed, after its introduction in the late 1970s, CsA revolutionized the field of solid organ transplantation. After binding their respective intracellular receptors, cyclophilin and FK-binding protein, CsA and FK-506 inhibit calcineurin, a calcium-dependent phosphatase that, in response to activating signaling through the T-cell receptor, normally initiates a signaling cascade culminating in transcription of cytokine genes, including interleukin-2 (IL-2), that are critical to T-cell proliferation and function. CsA and FK thereby effectively inhibit T-cell responses to both class I and II antigens. CsA-associated liver injury has been observed following liver, kidney, cardiac, and bone marrow transplantation, and it has also rarely been reported in nontransplant settings (22). Liver injury attributed to CsA was first reported in the early 1980s, when it was associated with relatively frequent but generally mild, dose-dependent cholestatic injury (22,23). CsA has also been hypothesized to contribute to the development of biliary tract sludge and stones (24). Experimentally, CsA has been shown to cause a dose-dependent decrease in both bile flow and secretion, possibly via inhibition of vesicular transport (25–27). Clinically, liver injury from CsA is characterized by a modestly increased serum alkaline phosphatase and variable but typically low-level elevations of serum aminotransferases and conjugated bilirubin. Histologically, liver injury attributed to CsA has been associated with hypertrophy of the bile ductal epithelium, with cytoplasmic vacuoles and the presence of “foamy” material within the hepatic sinusoids (28). CsA hepatotoxicity may be dose related and its incidence appears to be decreasing since the 1980s when reliable assays for CsA were not routinely available and higher doses of CsA were often used (29). In addition, during that era the causative agent for “non-A, non-B” hepatitis (i.e., hepatitis C) had not been defined and, as a result, some cases of hepatitis C flares or recurrence were probably misattributed to CsA hepatotoxicity. Liver injury attributed to FK-506 appears to be relatively rare, and the drug has been used safely in patients who developed mild liver injury attributed to CsA (30). Interestingly however, liver transplant recipients who have developed aminotransferase elevations in the setting of tacrolimus-based immunosuppression have also been successfully converted to CsA with resolution of liver injury (31,32). For example, one group described five transplant patients who developed elevation of serum aminotransferases and alkaline phosphatase attributed to tacrilimus 6 to 24 weeks after liver transplantation (33). Dose reduction or conversion to CsA appeared to rapidly improve but not necessarily normalize the liver test abnormalities. Liver histology showed perivenular heptatocyte dropout with sinusoidal dilation and congestion, features that may also be found with rejection (34). Nonetheless, the authors of this report believed that cellular rejection, viral hepatitis, and vascular compromise were less likely explanations than hepatotoxicity for the observed liver injury in these patients (33). FK-506 has also been associated with hepatic VOD in a patient following lung transplantation (35). Both CsA and FK-506 are metabolized by cytochrome P450 3A (CYP3A) and drugs that inhibit this pathway (e.g., erythromycin) may increase serum levels of these drugs and potentially increase the risk of liver injury (36). However, compared with other more frequent undesirable side effects of calcineurin inhibitors, including nephrotoxicity, neurotoxicity, hypertension, hyperlipidemia, and glucose intolerance, hepatotoxicity is a relatively infrequent and generally minor issue. In fact, perhaps the major importance of calcineurin inhibitor-related cholestasis, at least following liver transplantation, is that it may trigger an extensive, expensive, and potentially invasive evaluation with, for example, a liver biopsy and an endoscopic retrograde cholangiogram to rule out rejection and bile duct obstruction, respectively. SIROLIMUS Sirolimus (rapamycin; Rapamune) is a macrocyclic lactone antibiotic with immunosuppressive, antiproliferative, antifibrotic, antifungal, and antineoplastic properties. Both sirolimus and the

666

Davern

related compound everolimus bind the same intracellular receptor as FK-506, FKBP-12, but unlike FK-506 they do not inhibit calcineurin. Instead, the sirolimus–FKB-12 complex binds to the “mammalian target of rapamycin,” which blocks its kinase activity and interrupts downstream signaling pathways to a variety of growth factor receptors, including the IL-2 receptor (37). Sirolimus is a potent immosuppressant and because it is generally considered to lack nephrotoxicity, it has been used routinely to prevent rejection in transplant patients with renal insufficiency. Although sirolimus has been associated with a variety of serious side effects, including, leucopenia, thrombocytopenia, hypertriglyceridemia, pneumonitis, aphthous ulcers, edema, arthralgias, and wound dehiscence, clinically significant hepatotoxicity appears to be a rare complication of therapy. Reversible serum aminotransferase elevations have been reported in association with sirolimus use following both kidney and liver transplantation (38,39). More importantly, sirolimus has also been associated with an increased risk of hepatic artery thrombosis following liver transplantation in some but not all studies (40,41). This and other potential toxicities of sirolimus have curtailed its use as an immunosuppressant, especially early after liver transplantation. The mechanism of liver injury from sirolimus is not known. MYCOPHENOLATE MOFETIL (CELLCEPT) Mycophenolate mofetil (MMF) (CellCept) is an antimetabolite immunosuppressant that blocks de novo purine synthesis via the noncompetitive and reversible inhibition of inosine monophosphate dehydrogenase. Lymphocytes lack salvage pathway enzymes and thus MMF has a relatively selective lymphocyte antiproliferative effect that involves both T and B cells. MMF is commonly used in combination with other immunosuppressants, including a calcineurin inhibitor (e.g., FK-506) and corticosteroids, for solid organ transplantation, and has largely replaced azathioprine (AZA) in this setting because of its greater potency and better side effect profile. More recently, MMF has also been used as a steroid-sparing agent for a variety of autoimmune diseases. MMF is generally well tolerated, and does not have the undesirable side effects associated with other common immunosuppressants, including nephrotoxicity, hypertension, neurotoxicity, electrolyte abnormalities, hyperglycemia, hyperlipidemia, and structural bone loss. The most common side effects of MMF are reversible hematological and gastrointestinal disturbances, and the latter appear to be less problematic with a new enteric-coated form of the drug (42). Significant liver injury has rarely been reported secondary to MMF. Reversible liver injury has been reported in a small number of patients on MMF for the treatment of recalcitrant inflammatory scleritis, but whether MMF or another agent used in the multidrug immunosuppressive regimen was responsible is not clear (43). Recently, MMF-related hepatotoxicity was reported in 11 out of 79 (13.9%) Turkish renal transplant recipients (44). Most of the cases were young (median age 29 years; range 19–54 years) males (72.7%), and all had normal serum aminotransferases prior to MMF exposure. Elevated aminotransferases (median 222 IU; range 51–508 IU) regressed after the withdrawal (nZ6) or dosage reduction (nZ5) of MMF. The latency between exposure and ALT elevation was less than a month (median 28 days; range 4–70 days) and the values returned to the normal range in 16 (4–210) days after a 50% reduction or withdrawal of MMF. None of the cases developed jaundice or hepatic synthetic dysfunction. The mechanism of liver injury associated with MMF is not known. THIOPURINES: AZA, 6-MP, AND 6-THIOGUANINE The thiopurines, AZA, 6-mercaptopurine (6-MP), and 6-thioguanine (6-TG) have been used for decades for the treatment of a wide variety of autoimmune and inflammatory conditions, often as “steroid-sparing agents.” All the three drugs, which act as thiopurine analogs of the natural purine bases, are potentially hepatotoxic. Since they share common metabolic pathways (Fig. 1), it is not surprising that they also share some common clinicopathological patterns of liver injury. An impressive array of liver pathology has been associated with these drugs, particularly AZA,

667

Hepatotoxicity of Immunomodulating Agents C

N C

N

C

N

C

N

C

Purine nucleus S

H N

N N

N–CH3 O 2N

N

N Azathioprine(AZA)

H

SH

SH

N

N

H N

N

N N H2N 6-Thioguanine (6-TG)

N N 6-Mercaptopurine (6-MP)

6-MMP ribonucleotides

6-MMP

HPRT

6-MTIMP

TPMT

6-TIDP

TPMT

HPRT

AZA

Hepatotoxicity?

6-TIMP

6-MP

IMPDH GMPS

XO 6-TU

ITPase

6-TGNs

6-TITP

Pancreatitis?

Myelosuppression?

HPRT 6-TU

XO

AO

6-TG

TPMT

6-MTG

FIGURE 1 Diagram of thiopurine metabolism. Note the structural similarity of the thiopurines to the canonical purine base. Azathioprine is converted into 6-mercaptopurine (6-MP) by a rapid non-enzymatic reaction. 6-MP is metabolized by competing catabolic and anabolic enzymatic pathways: xanthine oxidase (XO) converts 6-MP into 6-thiouric acid (TU), which is an inactive metabolite; thiopurine (S)-methyltransferase (TPMT) methylates 6-MP to 6-methylmercaptopurine (6-MMP); hypoxanthine phosphoribosyl transferase (HPRT) converts 6-MP into 6-thioinosine 5’-monophosphate (6-TIMP) which can be metabolized in a multi-enzyme pathway into active 6-thioguanine nucleotides (6-TGNs). Abbreviations: Enzymes TPMT, thiopurine (S)-methyltransferase; XO, xanthine oxidase; HPRT, hypoxanthine phosphoribosyl transferase; IMPDH, inosine monophosphate dehydrogenase; GMPS, guanosine monophosphate synthetase; ITPase, inosine triphosphate pyrophosphatase; AO, aldehyde oxidase. Metabolites 6-MMP, 6-methylmercaptopurine; 6-TU, 6-thiouric acid; 6-TIMP, 6-thioinosine 5’-monophosphate; 6-MTIMP, 6-methylthioinosine 5’-monophosphate; 6-TIDP, 6-thioinosine 5’-diphosphate; 6-TITP, 6-thioinosine 5’-triphosphate; 6-MTG, 6-methyl thioguanine; 6-TGNs, 6-thioguanine nucleotides.

including asymptomatic liver biochemistry abnormalities, hepatocellular necrosis, cholestasis, and mixed hepatocellular/cholestatic liver injury. In addition, injury to the hepatic vascular endothelium manifested as VOD, NRH, and peliosis hepatitis is characteristic of the thiopurines. In particular, 6-TG, which has traditionally been used in the treatment of childhood leukemia and more recently for controlling inflammatory bowel disease (IBD) in patients intolerant or refractory to AZA or 6-MP, has been associated with the development of NRH, which is discussed further below. The mechanisms by which thiopurines preferentially injure the endothelial cells lining the sinusoids and the terminal hepatic venules are not fully defined. Using cultured murine sinusoidal endothelial cells (SECs) and hepatocytes, DeLeve et al. demonstrated that AZA causes profound depletion of cellular glutathione stores, and the depletion is greater in SECs than in hepatocytes (45). However, the drug and host factors which dictate the histological lesion that develops, be it VOD (more recently termed the “hepatic sinusoidal obstruction syndrome”), NRH (discussed further below), or peliosis hepatitis (characterized by blood-filled hepatic cysts), are currently undefined (46). Cholestatic or mixed cholestatic–hepatocellular liver injury are the most common forms of 6-MP hepatotoxicity that were seen in up to 40% of recipients in some older studies,

668

Davern

particularly when high doses of 6-MP were used (47). Adults appear to be more susceptible than children. The latency is highly variable, ranging from as little as one month to over one year after starting the drug. Cases of fatal ALF attributed to 6-MP have also been described, as have rare cases of recurrent liver injury upon rechallenge with the drug (47). AZA has been associated with a wider spectrum of liver injury than 6-MP, with cholestatic, hepatocellular, and hypersensitivity reactions all having been described in addition to the vascular lesions noted previously (48,49). Although the potential hepatotoxicity of AZA is well established, with at least occasional reports of rechallenge supporting causality, the incidence of significant liver injury from AZA is unclear (50,51). Most cases of AZA-related severe liver injury have occurred after six months of drug exposure, and male gender has been suggested as a risk factor for liver injury from AZA (51–53). In one study of patients with IBD, concomitant cortiscosteroid therapy also appeared to increase the risk of serum aminotransferase elevations attributed to AZA and 6-MP (54). 6-TG has been reported to cause acute hepatotoxicity with significant serum aminotransferase elevations noted only after two weeks of therapy (55). More common and concerning, the drug has been linked to both VOD and NRH in patients in treatment trials for leukemia and IBD (56–63). VOD has long been regarded as the characteristic histological lesion caused by 6-TG, but NRH has been increasingly recognized more recently (64). For example, Dubinsky et al. found histological evidence of NRH in over 75% of IBD patients treated with 6-TG with laboratory abnormalities (61). Importantly, they also noted NRH in 33% of IBD patients treated with 6-TG who had normal liver chemistries. The risk of NRH did not appear to be related to 6-TG cumulative dose or duration of therapy, but male gender increased the risk significantly (odds ratio, OR 2.9; 95% CI, 1.1–7.3; p!0.03) (61). The thiopurines may cause both allergic reactions with rash, fever, arthralgias, and possibly pancreatitis, and nonallergic (not dose-dependent) reactions including myelosuppression and hepatotoxicity. In addition, although it is usually nonallergic, liver injury can also rarely occur in the setting of a hypersensitivity reaction to these drugs. Impaired activity of the dephosphorylating enzyme inosine triphosphate pyrophosphatase may result in accumulation of a metabolite (6-thioinosine triphosphate) that has been hypothesized to contribute to these hypersensitivity symptoms (Fig. 1). In 62 patients treated with AZA for IBD, the ITPA 94COA deficiency–associated allele, present in approximately 5% of American-Caucasians and African-Americans in the general population, was significantly associated with adverse drug reactions (OR 4.2, 95% CI 1.6–11.5, pZ0.0034) (65). However, other studies have not confirmed the association of this ITPA mutation with hypersensitivity symptoms from AZA and the association remains controversial (66–69). Nonallergic reactions to AZA and 6-MP, including myelosuppression and hepatotoxicity, appear to be related to the concentration of thiopurine metabolites, which in turn is controlled by genetically polymorphic enzymes, in particular thiopurine (S)-methyltransferase (TPMT). All three thiopurines are prodrugs and must be metabolized before they have activity (Fig. 1) (70). After absorption, AZA is rapidly converted to 6-MP in the liver. 6-MP is metabolized by three critical enzymes: TPMT, xanthine oxidase (XO), and hypoxanthine phosphoribosyl transferase (HPRT). As shown in Figure 1, XO converts 6-MP into 6-thiouric acid, TPMT coverts 6-MP into 6-methylmercaptopurine (6-MMP), and HPRT metabolizes 6-MP into 6-thioinosine-5 0 -monophosphate (6-TIMP). 6-TIMP is transformed in a series of enzymatic reactions into 6-thioguanine nucleotides (6-TGNs) that are structurally similar to endogenous purine bases, and are incorporated into the DNA of dividing cells, including T and B lymphocytes, inhibiting their replication. The metabolism of 6-TG to 6-TGNs is less complicated than for AZA and 6-MP (Fig. 1). TMPT, which catalyzes thiopurine S-methylation, is a genetically polymorphic enzyme; over a dozen alleles have been described, with varying allele frequencies in different ethnic groups. Approximately 1/300 Caucasians inherit the TPMT deficiency phenotype as an autosomal recessive trait, most often associated with the TMPT*2, -*3A, and -*3C alleles (71). Patients homozygous for these variant alleles have poor methylator phenotypes and are at risk for severe, even life-threatening bone marrow toxicity when exposed to conventional therapeutic doses of AZA or 6-MP. The reason for this is that by converting 6-MP to 6-MMP, TMPT effectively shunts metabolism away from 6-TGNs. Poor TMPT activity thus results in higher

Hepatotoxicity of Immunomodulating Agents

669

concentrations of substrates for subsequent production of 6-TGNs that are associated with both clinical response and myelotoxicity. Conversely, high TMPT activity is predicted to result in lower 6-TGN levels, poor clinical response to treatment, and in higher concentrations of 6-MMP ribonucleotides, elevated serum concentrations of which have been associated with liver injury in some studies (61,72). TPMT activity may be assessed by commercially available genotype and phenotype assays, but the clinical benefits of assessing TPMT status remain controversial. In addition to genetic determinants of toxicity, other drugs may influence the metabolism of thiopurines. For example, allopurinol is a XO inhibitor commonly used for the treatment of gout that may have dramatic effects on thiopurine metabolism and result in profound myelosuppression with its attendant risk of sepsis. Nodular Regenerative Hyperplasia NRH of the liver is an increasingly recognized pathological condition characterized histologically by the presence of small (typically less than 3 mm) liver nodules that diffusely infiltrate the hepatic parenchyma, but in contradistinction to cirrhosis, are not surrounded by fibrosis. NRH has long been recognized in association with a variety of rheumatological, vascular, myeloproliferative, and lymphoproliferative disorders as well as certain drugs, including the thiopurines and some chemotherapeutic agents (73,74). NRH is often clinically silent but may be associated with significant portal hypertension manifested as splenomegaly and bleeding gastroesophageal varices. However, unlike cirrhosis, NRH rarely results in hepatic synthetic dysfunction, hepatic encephalopathy, or severe ascites. Because NRH is often asymptomatic and typically associated with only mild elevations of alkaline phosphatase and aminotransferases, its prevalence is probably underappreciated. Indeed, the diagnosis of NRH may be elusive and requires a high index of suspicion. Imaging by magnetic resonance imaging may be useful in the evaluation of NRH (57). However, a definitive diagnosis of NRH usually requires a careful histological examination by an experienced hepatopathologist, as the findings may be subtle and are easily overlooked on routine stains. For example, hematoxylin–eosin staining often shows a vague nodularity (Fig. 2A), while staining for collagen with a Masson trichrome characteristically shows no significant fibrosis (75). The histological diagnosis is best made on reticulin stain (Fig. 2B,C), which illustrates nodules with expanded liver cell plates (more than one cell thick) encircled by zones of reticulin compression composed of hepatocytes that are small, atrophic, and pressed together (75). The hepatocytes contained within the confines of a nodule are typically large and may even appear dysplastic. Inflammation, cholestasis, and fibrosis are notably absent, although mild periportal or perisinusoidal fibrosis may be present (74). The etiology of NRH is unclear but, based primarily on histological data, it is thought to be the result of alterations in portal blood flow—a so-called “obstructive portal venopathy” (76–78). The nodular areas in NRH appear to represent a hypertrophic response to adjacent hepatocyte atrophy resulting from obliteration of the portal flow feeding that portion of the lobule. Supporting this notion, NRH has been reported in complicating cases of congenital portal vein abnormalities, and some investigators have demonstrated histological obstruction of portal venules by platelet aggregates and thrombi generated in the spleen or portal vein (76). The natural history of NRH appears to be relatively indolent in most cases, although severe complications of portal hypertension, including massive variceal bleeding, may occur. Whether drug-induced NRH is reversible if the associated offending drug is withdrawn is unclear and probably dependent on the stage of the liver injury (79). Certainly, if a patient on a thiopurine develops abnormal liver biochemistries, consideration should be given to liver biopsy or discontinuation of the drug if possible (80). Given that all three thiopurines share a propensity to cause NRH, simply switching from one to another is probably not appropriate. LEFLUNOMIDE Leflunomide (Arava) is an inhibitor of pyrimidine synthesis and as such has a mechanism of action that differs from other existing disease-modifying antirheumatic drugs (DMARDs). It is rapidly converted by first-pass metabolism in the gut and liver to its active malononitrilamide

670

Davern

(A)

(B)

(C)

FIGURE 2 Photomicrographs of liver histology from a patient treated with 6-thioguanine for ulcerative colitis demonstrating nodular regenerative hyperplasia. See text for details. Source: Courtesy of Dr. Steve Geller.

Hepatotoxicity of Immunomodulating Agents

671

metabolite, A771726, which reversibly inhibits dihydroorotate dehydrogenase, a rate-limiting mitochondrial enzyme involved in de novo pyrimidine ribonucleotide biosynthesis (81,82). Activated T lymphocytes require a large pyrimidine pool to proliferate rapidly and thus appear to be particularly sensitive to inhibition of the de novo pathway (83). Leflunomide also has multiple other effects, including inhibition of tyrosine kinases and other signaling pathways, that may also influence its immunomodulatory properties (84–86). Leflunomide improves both clinical signs and symptoms and delays radiological progression of rheumatoid arthritis (RA), and appears to have efficacy generally comparable to other DMARDs such as methotrexate (87). Leflunomide was approved for the treatment of RA in 1998, and also appears to be effective for the treatment of plaque psoriasis and psoriatic arthritis (PA) (88,89). Preapproval clinical trial data demonstrated that leflunomide treatment was associated with elevated aminotransferases in a significant number of treated individuals, but these were generally mild (!2!ULN) and reversed even with continued leflunomide treatment. However, following a 2001 report by the European Agency for the Evaluation of Medicinal Products (EMEA) describing 129 serious hepatic reactions linked to leflunomide therapy, including 15 cases of ALF with 9 fatalities, a relatively intense monitoring regimen was adopted. Current recommendations suggest that patients treated with leflunomide have an ALT at baseline and at monthly intervals for six months and, if stable, every six to eight weeks thereafter. In patients who develop mild ALT elevations (2–3!ULN), a dose reduction of leflunomide is recommended (to 10 mg/day) (81). If the ALT remains elevated 2–3!ULN despite dose reduction, a liver biopsy has been recommended in patients that continue treatment, despite the known risks of this invasive procedure (90). If ALTO3!ULN persists despite dose reduction, leflunomide should be discontinued and “washout” treatment with cholestyramine (4–8 g p.o. tid for two to five days) is recommended (91). Leflunomide has a long serum half-life (approximately 15 days) and undergoes extensive enterohepatic circulation, which provides the rationale for cholestyramine therapy, although this has not been unequivocally demonstrated to be effective in the setting of leflunomide-induced liver injury. Of note, leflunomide treatment is not recommended in individuals with preexisting liver disease or those exposed to other hepatotoxic drugs. Within four years of leflunomide’s approval, and following the EMEA report, a public citizen’s group called on the Food and Drug Administration (FDA) to withdraw approval for the drug, citing at least 130 cases of severe liver injury and 12 deaths related to the use of the drug (92). Most of these postmarketing cases reportedly occurred within the first six months of leflunomide therapy, and there was no obvious age or gender predilection. Many of the affected patients had significant comorbidities, including liver disease, and most patients were exposed to multiple potential hepatotoxins in addition to leflunomide (93). Apparently, in response to this petition, the FDA undertook an extensive rereview of data regarding liver injury attributed to leflunomide from prospective clinical trials, retrospective cohort studies, and postmarketing reports, both in the literature and from the FDA adverse event reporting system (94). While mild serum ALT elevations appeared to be more common (2–4%) in patients exposed to leflunomide than placebo, no cases of jaundice associated with hepatocellular injury, associated with an estimated mortality of at least 10% per “Hy’s rule,” were found in a review of clinical trials including over 1700 patients on leflunomide monotherapy, or in a combined database search of over 13,700 patients. The conclusions of this FDA rereview were that, although the exact incidence could not be accurately calculated with the available data, leflunomideassociated ALF appeared to be very rare and liver injury from the drug did not have a consistent clinical or biochemical pattern or “signature” (94). A risk/benefit analysis by the FDA ultimately favored continued approval of leflunomide. There are only scattered reports of leflunomide-associated liver injury in the published literature, and most have confounding issues. For example, one case report describes a 75-yearold woman with RA treated with leflunomide for 14 months before undergoing coronary artery bypass grafting (95). The patient was taking multiple other medications but had normal serum aminotransferases prior to surgery. Postoperatively, she developed gastrointestinal bleeding, shock, renal, and hepatic failure, and ultimately succumbed to multiorgan failure. Abdominal computer tomography suggested hepatic ischemia, and autopsy ultimately revealed findings consistent with severe hepatic congestion. Nonetheless, the liver injury was inexplicably

672

Davern

attributed to leflunomide rather than hepatic ischemia (95). Another published case describes a middle-aged man with RA who developed signs of portal hypertension after 3.5 years of combined therapy with methotrexate and leflunomide (96). Viral hepatitis and autoimmune serologies were negative and there was no biochemical evidence of iron overload. Liver histology showed micronodular cirrhosis with mild steatosis but little or no inflammation, and the relative roles of leflunomide and methotrexate in the development of the cirrhosis were unclear (96). Another published case report describes a 67-year-old woman with RA who developed elevated serum aminotransferases (ALT 550 U/L), alkaline phosphatase (777 U/L), and bilirubin (3.5 mg/dL) 15 days after starting therapy with leflunomide and gold therapy (97). Other than a low titer (1:80) antinuclear antibody, her serological evaluation was unrevealing. A liver biopsy was performed approximately one month after presentation with liver injury and showed patchy hepatocyte dropout and focal lymphocytic infiltrates, interpreted as consistent with acute hepatitis. The patient was treated with cholestyramine and made a complete recovery with normalization of serum aminotransferases by three months after presentation. Of note, the patient was found to be homozygous for the uncommon CYP2C9*3 allele that is associated with low activity (98). This is of potential interest, as at least some other patients with apparent leflunomide-associated liver injury have been exposed concomitantly to CYP2C9 substrates, such as the nonsteroidal anti-inflammatory drug celecoxib and diclofenac, and the antifungal itraconazole (98). Paradoxically, leflunomide has also been shown to ameliorate liver injury in a variety of experimental settings. For example, it appears to attenuate injury to rat hepatocytes cocultured with lipopolysaccharide (LPS)-activated Kupffer cells by inhibiting release of proinflammatory cytokines, including tumor necrosis factor-alpha (TNF-a), from Kupffer cells (99). In vivo, leflunomide has also recently been reported to protect mice from T-cell-mediated liver injury induced by concanavalin A by the inhibition of the transcription factor NF-kB, thereby reducing TNF-a production and caspase-dependent hepatocyte death (100). Leflunomide has also been demonstrated to significantly decrease hepatic fibrogenesis and ameliorate ischemia– reperfusion injury in rat models of liver injury (101,102). The mechanism of hepatotoxicity from leflunomide is unknown.

TNF ANTAGONISTS TNF has pleiotropic effects on immune responses and appears to play a central role in the immunopathogenesis of many autoimmune and inflammatory disorders. Currently, three agents that directly bind and neutralize TNF-a are approved by the FDA: etanercept (Enbrel) for RA, juvenile idiopathic arthritis, psoriasis, PA, and ankylosing spondylitis (AS); infliximab (Remicade) for RA, Crohn’s disease, and AS; and Adalimumab (Humira) for RA and PA. Although structurally distinct, all three agents directly bind soluble and membrane-bound TNF and thereby antagonize TNF-a signaling. Etanercept comprises two extracellular domains of human recombinant soluble TNF receptor type 2 (p75) fused with the Fc domain of human IgG1 antibody, while infliximab is a chimeric IgG1 monoclonal antibody (25% murine, 75% human) and adalimumab is a humanized monoclonal antibody that has reduced immunogenity compared with infliximab. Other TNF antagonists have also been developed and are in clinical trials. Most cases of clinically significant liver injury associated with TNF antagonists have been linked to infliximab. In a phase III RA study, mild to moderate aminotransferase elevations (! 3!ULN) were observed more frequently in patient treated with infliximab (127/342; 37%) than placebo (25/85; 29%); all patients also received methotrexate in this study (103). Similarly, in a trial of maintenance therapy for Crohn’s disease, aminotransferase elevations were more common in the infliximab (160/385; 42%) than the placebo (68/187; 36%) group (104). In neither study did patients experience jaundice or hepatic synthetic dysfunction. However, by early 2003, the FDA had received 134 spontaneous reports of ALF attributed to infliximab and etanercept treatment, of which 50 (31 infliximab and 19 etanercept) appeared relatively well documented. However, in most (43 out of 50) of these cases, there appeared to be strong competing causes for liver injury, such as sepsis, other hepatotoxins, and viral hepatitis.

Hepatotoxicity of Immunomodulating Agents

673

Based on these data, including rare cases of severe, even fatal, liver injury in postmarketing studies, the FDA recently issued warnings regarding the potential risk of liver injury associated with infliximab use (105). Features of autoimmune hepatitis, including high titer antinuclear antibodies, have been observed in at least some of these cases. However, TNF antagonists as a group have been associated with the development of such antibodies even in the absence of liver injury, so their significance is unclear. Severe hepatic reactions have occurred between two weeks and more than one year after the initiation of infliximab treatment. The onset of liver injury associated with infliximab may apparently be quite sudden; elevations in hepatic aminotransferases were not observed prior to the onset of severe and progressive liver injury in some of the reported cases that were ultimately fatal or necessitated liver transplantation. Based on at least one case report, etanercept also appears to be associated with rare episodes of otherwise unexplained liver injury with autoimmune features (106). If confirmed, this association may have mechanistic implications; infliximab and etanercept are structurally quite different, and if both are potentially hepatotoxic, this suggests that in susceptible individuals inhibition of TNF signaling may itself initiate liver injury, and that liver injury may represent a “class effect” of these agents. Several other adverse effects associated with infliximab are shared with other TNF antagonists, including activation of latent tuberculosis and other infections, congestive heart failure, lymphoma, demyelinating disease, seizures, aplastic anemia, intestinal perforation, and cutaneous rash. The development of autoimmunity markers, including antinuclear antibodies and anti-double-stranded DNA antibodies, also appears to be a class effect of TNF antagonists; however, such antibodies are of unclear clinical significance (107). TNF has seemingly paradoxical effects on liver function. For example, TNF-a signaling appears to be critically involved in the development and progression of acute and chronic liver injury in experimental models (108). Conversely, experiments in knockout mice that lack TNF signaling demonstrate that this cytokine is also critical for normal hepatic regeneration in response to liver injury (109). Although the mechanisms have not been well elucidated, liver injury in the setting of TNF antagonist treatment is clearly idiosyncratic and characteristically has autoimmune features, suggesting that it results from an aberrant immune response induced by TNF blockade in genetically susceptible individuals. Supporting this possibility is a report describing the development of a cutaneous vasculitis in a patient with Crohn’s disease treated with etanercept that resolved when etanercept was stopped but subsequently reappeared when the patient was exposed to infliximab (110). The authors of this report suggest that TNF-a signaling is important for self-tolerance. Treatment with TNF antagonists may also increase the risk of lymphoma and possibly solid tumors (111,112). Recently, a black box warning was issued for infliximab based on six cases of hepatosplenic T-cell lymphoma (HSTCL), a rare and often fatal form of non-Hodgkin’s lymphoma that typically affects children and young adults. All cases to date have been young patients on infliximab for Crohn’s disease, and all had received prior treatment with a thiopurine (AZA or 6-MP) as well. Presenting symptoms of HSTCL include hepatosplenomegaly with elevated liver tests, fever, cytopenias, weight loss, and malaise, with the conspicuous absence of peripheral adenopathy (113). HYDROXYCHLOROQUINE (PLAQUENIL) The antimalarial, hydrochloroquine, an a-hydroxylated derivative of chloroquine, has been used for the treatment of cutaneous and joint manifestations of rheumatic disease. The drug is generally well tolerated with few serious side effects (114). Mildly elevated serum aminotransferases were frequently noted in Asian patients with both RA and coexisting chronic viral hepatitis while on hydroxychloroquine (115). Only a few cases of significant liver injury have been reported in association with hydroxychloroquine. The group from King’s College described two young women, one with systemic lupus erythematosis and the other with juvenile RA (Still’s disease), who developed life-threatening ALF within two weeks of being exposed to hydrochloroquine; one died and the other required liver transplantation (116). Neither patient had a history of antecedent liver disease. More recently, a 26-year-old woman

674

Davern

was described who developed fever and elevated serum aminotransferases (ALT 285 UI/L, AST 399 UI/L) within 10 hours of initial exposure to hydroxychloroquine; she had previously normal aminotransferases and no other obvious explanation for her liver injury (117). Interestingly, the patient was subsequently treated with chloroquine and did not develop evidence of recurrent liver injury. The mechanism of liver injury from hydroxychloroquine is unknown. GOLD SALTS Gold salts have been used for over 75 years for the treatment of RA and other autoimmune diseases, although the mechanism of immunosuppression caused by gold has been relatively mysterious. Recently, using a high-throughput screening assay, gold was found to strip peptides from human class II MHC proteins by an allosteric mechanism, thus blocking the ability of antigen-presenting cells to activate T cells (118). At one time, gold shots were a standard treatment for RA, but both parenteral and oral gold therapy has largely been replaced by newer, more effective, and safer DMARDs. Treatment with gold salts, the so-called chrysotherapy, may be associated with severe side effects, including dermatitis, kidney damage, and myelotoxicity, which have largely overshadowed any problems with potential hepatotoxicity. Liver injury from gold, which is cholestasic more commonly than hepatocellular, typically occurs within a month of initiating treatment and is associated with rash, fever, and eosinophilia in approximately 40% of cases (47). Massive hepatocellular necrosis has rarely been associated with parenteral gold therapy (119). The mechanisms underlying liver injury from gold are not well defined (120). Many cases have hypersensitivity features; some studies suggest that the risk of side effects from gold therapy may be higher in patients with the HLADR3 antigen, although liver injury does not appear to be more common in this group (121,122). D-PENICILLAMINE

Penicillamine was introduced 50 years ago as a copper chelator for Wilson’s disease, but is now rarely used as first-line therapy because effective and less toxic alternatives are available (123). However, penicillamine continues to be used occasionally for the treatment of RA and other conditions (124). Although significant liver injury is much less common than other serious side effects of the drug, including rash and other hypersensitivity symptoms, bone marrow depression, and kidney damage, jaundice has been reported in association with the use of penicillamine. Essentially all reported cases describe the relatively rapid onset (two to four weeks after the first exposure) of hypersensitivity symptoms, including rash, eosinophilia, fever, and cholestatic liver injury (47). Potential hepatotoxicity, in the form of moderate reversible aminotransferase elevations, has also been observed in a small number of Wilson’s disease patients, although causality assessment in this setting is obviously challenging (125,126). Although the mechanism of liver injury associated with penicillamine is not well defined, hypersensitivity appears to play a major role. SULFASALAZINE AND OTHER AMINOSALICYLATES Sulfasalazine (Azulfidine), which contains both mesalamine (5-aminosalicylate) and sulfapyridine moieties, has been used for decades as a treatment of IBD, particularly ulcerative colitis. Its use for IBD has decreased recently with the introduction of newer mesalamine agents lacking the sulfonamide component that is associated with severe hypersensitivity reactions. Sulfasalazine is still occasionally used for the treatment of RA and other rheumatological conditions (127). The mechanisms underlying liver injury from hypersensitivity related to drugs, including sulfonamides, are reviewed elsewhere in the text (chap. 7). Like other sulfonamides, sulfasalazine characteristically causes hepatocellular liver injury, but granulomatous hepatitis has also been reported. Fatalities due to ALF associated with the drug has been described (128,129). Clinical features include fever, rash, lymphadenopathy, hepatomegaly, leukocytosis (often with “atypical lymphocytes”), and eosinophilia (130,131). Onset of

675

Hepatotoxicity of Immunomodulating Agents

hypersensitivity symptoms is usually within two months of exposure. The sulfapyridine portion of sulfasalazine is responsible for most hypersensitivity reactions. Unlike sulfasalazine, other aminosalicylates lacking sulfapyridine are rarely associated with severe liver injury. One case report described a patient who developed rash to sulfasalazine subsequently developed severe liver injury, with hypersensitivity upon exposure to mesalazine (132). Another report attributed a cholestatic reaction to mesalazine in a patient with Crohn’s disease (133). THALIDOMIDE Despite its notorious history as the sedative and antiemetic for morning sickness associated with profound teratogenicity, thalidomide has recently enjoyed renewed use for a variety of indications based on its anti-inflammatory, immunomodulatory, and antiangiogenic properties. Thalidomide was FDA approved in 1998 for the treatment of erythema nodosum leprosum lesions, but it is also frequently used for advanced multiple myeloma, Kaposi’s sarcoma, HIVassociated aphthous ulcers, mucocutaneous lesions of Behcet’s disease, refractory cutaneous lesions associated with lupus, scleroderma, and sarcoidosis, and has been studied for the treatment of a variety of other cancers and rheumatological disorders (134). The mechanisms by which thalidomide modulates immune responses may be through its inhibition of TNF-a and other inflammatory cytokines. Thalidomide has been shown to inhibit TNF-a production from lipopolysaccharide-stimulated monocytes and macrophages by stimulating the degradation of its mRNA (135,136). Thalidomide may also be oxidized by cytochome 2C19 (CYP2C19) and its metabolites appear to inhibit inflammatory cascades initiated by both NF-kB and COX2 (137,138). Significant liver injury has rarely been associated with thalidomide use (139). Interestingly, all of the recent case reports that have implicated thalidomide as a cause of significant liver injury have described elderly women treated for refractory multiple myeloma, some of whom had had underlying liver disease, including chronic viral hepatitis (140). In all cases, thalidomide was associated with hepatocellular liver injury, often with marked aminotransferase elevations and variable degrees of jaundice, and the onset of liver injury after exposure varied between 1 and 28 weeks (140–142). The mechanism of liver injury associated with thalidomide is unknown. NON-HEPATOTOXIC IMMUNOSUPPRESSANTS Table 1 includes a partial listing of other immunosuppressants that have not yet been associated with a well-established track record of clinically significant liver injury. This apparent lack of TABLE 1 Non-Hepatotoxic Immunosuppressants Immunosuppressant Antithymocyte globulin (Thymoglobulin) Muronmonoab-CD3 (Orthoclone OKT3) Daclizumab (Zenapaxw) Basiliximab (Simulectw) Anakinra (Kineretw) Rituximab (Rituxanw) CTLA4IG (Abatacept/Orenciaw) Certolizumab pegol (CTZ-PEG) Natalizumab (Tysabriw) Efalizumab (Raptivaw) Tocilizumab (Actemraw)

Mechanism of action Polyclonal antiserum directed against human thymocytes that selectively destroys T cells Monoclonal antibody directed against CD3 on T cells Murine human chimeric monoclonal antibody to the IL-2Ra T-cell receptor Murine human chimeric monoclonal antibody to the IL-2Ra T-cell receptor Recombinant human interleukin-1 receptor antagonist Monoclonal antibody against CD20 receptor on B cells Recombinant fusion protein that blocks T-cell costimulation Humanized Fab fragment that acts as a TNF antagonist Monoclonal antibody directed against alpha-4-integrin preventing immune cell adhesion and extravascular migration Recombinant humanized monoclonal that binds to the a-subunit (CD11a) of LFA-1 and inhibits T-cell activation Humanized anti-interleukin-6 receptor monoclonal antibody

Abbreviations: CTLA4IG, cytotoxic T-lymphocyte antigen-4IG; IL-2Ra, interleukin-2Ra; LFA-1, lymphocyte function-associated antigen-1.

676

Davern

intrinsic hepatotoxicity may actually reflect the relatively limited and often short-term use of these drugs to date. Based on fundamental advances in the understanding of immunity and inflammation over the last decade, novel immunosuppressants are being developed at an unprecedented pace (143). Some of these novel immunomodulatory agents will almost certainly have clinically significant liver injury as an important side effect. ACKNOWLEDGMENTS The author gratefully acknowledges Sadie O. MacFarlane for the artwork in Figure 1 and Dr. Stephen Geller, Cedars-Sinai Medical Center, Los Angeles, CA for the generous contribution of the liver histology photomicrographs in Figure 2. REFERENCES 1. Andrade RJ, Lucena MI, Fernandez MC, et al. Drug-induced liver injury: an analysis of 461 incidences submitted to the Spanish registry over a 10-year period. Gastroenterology 2005; 129(2):512–21. 2. Bjornsson E, Olsson R. Outcome and prognostic markers in severe drug-induced liver disease. Hepatology 2005; 42(2):481–9. 3. Bjornsson E, Olsson R. Suspected drug-induced liver fatalities reported to the WHO database. Dig Liver Dis 2006; 38(1):33–8. 4. Russo MW, Galanko JA, Shrestha R, Fried MW, Watkins P. Liver transplantation for acute liver failure from drug induced liver injury in the United States. Liver Transpl 2004; 10(8):1018–23. 5. Ganey PE, Luyendyk JP, Maddox JF, Roth RA. Adverse hepatic drug reactions: inflammatory episodes as consequence and contributor. Chem Biol Interact 2004; 150(1):35–51. 6. Kaplowitz N. Idiosyncratic drug hepatotoxicity. Nat Rev Drug Discov 2005; 4(6):489–99. 7. Ferrero D, Pogliani EM, Rege-Cambrin G, et al. Corticosteroids can reverse severe imatinib-induced hepatotoxicity. Haematologica 2006; 91(Suppl. 6):ECR27. 8. Lau JY, Lai CL, Lin HJ, et al. Fatal reactivation of chronic hepatitis B virus infection following withdrawal of chemotherapy in lymphoma patients. Q J Med 1989; 73(270):911–7. 9. Yeo W, Chan PK, Zhong S, et al. Frequency of hepatitis B virus reactivation in cancer patients undergoing cytotoxic chemotherapy: a prospective study of 626 patients with identification of risk factors. J Med Virol 2000; 62(3):299–307. 10. Lok AS, Liang RH, Chiu EK, Wong KL, Chan TK, Todd D. Reactivation of hepatitis B virus replication in patients receiving cytotoxic therapy. Report of a prospective study. Gastroenterology 1991; 100(1):182–8. 11. Chou CK, Wang LH, Lin HM, Chi CW. Glucocorticoid stimulates hepatitis B viral gene expression in cultured human hepatoma cells. Hepatology 1992; 16(1):13–8. 12. Davern TJ, Lake JR. Recurrent disease after liver transplantation. Semin Gastrointest Dis 1998; 9(3):86–109. 13. Esteve M, Saro C, Gonzalez-Huix F, Suarez F, Forne M, Viver JM. Chronic hepatitis B reactivation following infliximab therapy in Crohn’s disease patients: need for primary prophylaxis. Gut 2004; 53(9):1363–5. 14. Lau GK, Yiu HH, Fong DY, et al. Early is superior to deferred preemptive lamivudine therapy for hepatitis B patients undergoing chemotherapy. Gastroenterology 2003; 125(6):1742–9. 15. Vento S, Cainelli F, Mirandola F, et al. Fulminant hepatitis on withdrawal of chemotherapy in carriers of hepatitis C virus. Lancet 1996; 347(8994):92–3. 16. Markovic S, Drozina G, Vovk M, Fidler-Jenko M. Reactivation of hepatitis B but not hepatitis C in patients with malignant lymphoma and immunosuppressive therapy. A prospective study in 305 patients. Hepatogastroenterology 1999; 46(29):2925–30. 17. Kawatani T, Suou T, Tajima F, et al. Incidence of hepatitis virus infection and severe liver dysfunction in patients receiving chemotherapy for hematologic malignancies. Eur J Haematol 2001; 67(1):45–50. 18. Zuckerman E, Zuckerman T, Douer D, Qian D, Levine AM. Liver dysfunction in patients infected with hepatitis C virus undergoing chemotherapy for hematologic malignancies. Cancer 1998; 83(6):1224–30. 19. Strasser SI, Myerson D, Spurgeon CL, et al. Hepatitis C virus infection and bone marrow transplantation: a cohort study with 10-year follow-up. Hepatology 1999; 29(6):1893–9. 20. Farrell GC. Drugs and steatohepatitis. Semin Liver Dis 2002; 22(2):185–94. 21. Farrell GC, Larter CZ. Nonalcoholic fatty liver disease: from steatosis to cirrhosis. Hepatology 2006; 43(2 Suppl. 1):S99–112. 22. Kassianides C, Nussenblatt R, Palestine AG, Mellow SD, Hoofnagle JH. Liver injury from cyclosporine A. Dig Dis Sci 1990; 35(6):693–7.

Hepatotoxicity of Immunomodulating Agents

677

23. Lorber MI, Van Buren CT, Flechner SM, Williams C, Kahan BD. Hepatobiliary and pancreatic complications of cyclosporine therapy in 466 renal transplant recipients. Transplantation 1987; 43(1):35–40. 24. Kahan BD, Flechner SM, Lorber MI, Golden D, Conley S, Van Buren CT. Complications of cyclosporine–prednisone immunosuppression in 402 renal allograft recipients exclusively followed at a single center for from one to five years. Transplantation 1987; 43(2):197–204. 25. Roman ID, Monte MJ, Gonzalez-Buitrago JM, Esteller A, Jimenez R. Inhibition of hepatocytary vesicular transport by cyclosporin A in the rat: relationship with cholestasis and hyperbilirubinemia. Hepatology 1990; 12(1):83–91. 26. Roman ID, Fernandez-Moreno MD, Fueyo JA, Roma MG, Coleman R. Cyclosporin A induced internalization of the bile salt export pump in isolated rat hepatocyte couplets. Toxicol Sci 2003; 71(2):276–81. 27. Stone BG, Udani M, Sanghvi A, et al. Cyclosporin A-induced cholestasis. The mechanism in a rat model. Gastroenterology 1987; 93(2):344–51. 28. Wisecarver JL, Earl RA, Haven MC, et al. Histologic changes in liver allograft biopsies associated with elevated whole blood and tissue cyclosporine concentrations. Mod Pathol 1992; 5(6):611–6. 29. Kowdley KV, Keeffe EB. Hepatotoxicity of transplant immunosuppressive agents. Gastroenterol Clin North Am 1995; 24(4):991–1001. 30. Yuan QS, Zheng FL, Sun Y, Yu Y, Li Y. Rescue therapy with tacrolimus in renal graft patients with cyclosporine A-induced hepatotoxicity: a preliminary study. Transplant Proc 2000; 32(7):1694–5. 31. Emre S, Genyk Y, Schluger LK, et al. Treatment of tacrolimus-related adverse effects by conversion to cyclosporine in liver transplant recipients. Transpl Int 2000; 13(1):73–8. 32. Mor E, Schwersenz A, Sheiner PA, Emre S, Schwartz ME, Miller CM. Reversal of gastrointestinal toxicity associated with long-term FK506 immunosuppression by conversion to cyclosporine in liver transplant recipients. Transplantation 1994; 57(7):1130–2. 33. Fisher A, Mor E, Hytiroglou P, et al. FK506 hepatotoxicity in liver allograft recipients. Transplantation 1995; 59(11):1631–2. 34. Tsamandas AC, Jain AB, Felekouras ES, Fung JJ, Demetris AJ, Lee RG. Central venulitis in the allograft liver: a clinicopathologic study. Transplantation 1997; 64(2):252–7. 35. Shah S, Budev M, Blazey H, Fairbanks K, Mehta A. Hepatic veno-occlusive disease due to tacrolimus in a single-lung transplant patient. Eur Respir J 2006; 27(5):1066–8. 36. Jensen CW, Flechner SM, Van Buren CT, et al. Exacerbation of cyclosporine toxicity by concomitant administration of erythromycin. Transplantation 1987; 43(2):263–70. 37. Kahan BD. Sirolimus: a comprehensive review. Expert Opin Pharmacother 2001; 2(11):1903–17. 38. Neff GW, Ruiz P, Madariaga JR, et al. Sirolimus-associated hepatotoxicity in liver transplantation. Ann Pharmacother 2004; 38(10):1593–6. 39. Niemczyk M, Wyzgal J, Perkowska A, Porowski D, Paczek L. Sirolimus-associated hepatotoxicity in the kidney graft recipient. Transpl Int 2005; 18(11):1302–3. 40. Montalbano M, Neff GW, Yamashiki N, et al. A retrospective review of liver transplant patients treated with sirolimus from a single center: an analysis of sirolimus-related complications. Transplantation 2004; 78(2):264–8. 41. Dunkelberg JC, Trotter JF, Wachs M, et al. Sirolimus as primary immunosuppression in liver transplantation is not associated with hepatic artery or wound complications. Liver Transpl 2003; 9(5):463–8. 42. Simmons WD, Rayhill SC, Sollinger HW. Preliminary risk-benefit assessment of mycophenolate mofetil in transplant rejection. Drug Saf 1997; 17(2):75–92. 43. Sen HN, Suhler EB, Al-Khatib SQ, Djalilian AR, Nussenblatt RB, Buggage RR. Mycophenolate mofetil for the treatment of scleritis. Ophthalmology 2003; 110(9):1750–5. 44. Balal M, Demir E, Paydas S, Sertdemir Y, Erken U. Uncommon side effect of MMF in renal transplant recipients. Ren Fail 2005; 27(5):591–4. 45. DeLeve LD, Wang X, Kuhlenkamp JF, Kaplowitz N. Toxicity of azathioprine and monocrotaline in murine sinusoidal endothelial cells and hepatocytes: the role of glutathione and relevance to hepatic venoocclusive disease. Hepatology 1996; 23(3):589–99. 46. DeLeve LD, Shulman HM, McDonald GB. Toxic injury to hepatic sinusoids: sinusoidal obstruction syndrome (veno-occlusive disease). Semin Liver Dis 2002; 22(1):27–42. 47. Zimmerman HJ. Hepatotoxicity—The Adverse Effects of Drugs and Other Chemicals on the Liver. 2nd ed. Philadelphia, PA: Linppincott Williams & Wilkins, 1999. 48. Knowles SR, Gupta AK, Shear NH, Sauder D. Azathioprine hypersensitivity-like reactions—a case report and a review of the literature. Clin Exp Dermatol 1995; 20(4):353–6. 49. Horsmans Y, Rahier J, Geubel AP. Reversible cholestasis with bile duct injury following azathioprine therapy. A case report. Liver 1991; 11(2):89–93. 50. Sterneck M, Wiesner R, Ascher N, et al. Azathioprine hepatotoxicity after liver transplantation. Hepatology 1991; 14(5):806–10.

678

Davern

51. Katzka DA, Saul SH, Jorkasky D, Sigal H, Reynolds JC, Soloway RD. Azathioprine and hepatic venocclusive disease in renal transplant patients. Gastroenterology 1986; 90(2):446–54. 52. Daniel F, Cadranel JF, Seksik P, et al. Azathioprine induced nodular regenerative hyperplasia in IBD patients. Gastroenterol Clin Biol 2005; 29(5):600–3. 53. Romagnuolo J, Sadowski DC, Lalor E, Jewell L, Thomson AB. Cholestatic hepatocellular injury with azathioprine: a case report and review of the mechanisms of hepatotoxicity. Can J Gastroenterol 1998; 12(7):479–83. 54. Bastida G, Nos P, Aguas M, et al. Incidence, risk factors and clinical course of thiopurine-induced liver injury in patients with inflammatory bowel disease. Aliment Pharmacol Ther 2005; 22(9):775–82. 55. Rulyak SJ, Saunders MD, Lee SD. Hepatotoxicity associated with 6-thioguanine therapy for Crohn’s disease. J Clin Gastroenterol 2003; 36(3):234–7. 56. de Boer NK, Reinisch W, Teml A, et al. 6-Thioguanine treatment in inflammatory bowel disease: a critical appraisal by a European 6-TG working party. Digestion 2006; 73(1):25–31. 57. Seiderer J, Zech CJ, Reinisch W, et al. A multicenter assessment of liver toxicity by MRI and biopsy in IBD patients on 6-thioguanine. J Hepatol 2005; 43(2):303–9. 58. Shepherd PC, Fooks J, Gray R, Allan NC. Thioguanine used in maintenance therapy of chronic myeloid leukaemia causes non-cirrhotic portal hypertension. Results from MRC CML. II. Trial comparing busulphan with busulphan and thioguanine. Br J Haematol 1991; 79(2):185–92. 59. Piel B, Vaidya S, Lancaster D, Taj M, Pritchard-Jones K. Chronic hepatotoxicity following 6-thioguanine therapy for childhood acute lymphoblastic leukaemia. Br J Haematol 2004; 125(3):410–1 (author reply 412). 60. Stoneham S, Lennard L, Coen P, Lilleyman J, Saha V. Veno-occlusive disease in patients receiving thiopurines during maintenance therapy for childhood acute lymphoblastic leukaemia. Br J Haematol 2003; 123(1):100–2. 61. Dubinsky MC, Vasiliauskas EA, Singh H, et al. 6-Thioguanine can cause serious liver injury in inflammatory bowel disease patients. Gastroenterology 2003; 125(2):298–303. 62. De Bruyne R, Portmann B, Samyn M, et al. Chronic liver disease related to 6-thioguanine in children with acute lymphoblastic leukaemia. J Hepatol 2006; 44(2):407–10. 63. Ravikumara M, Hill FG, Wilson DC, et al. 6-Thioguanine-related chronic hepatotoxicity and variceal haemorrhage in children treated for acute lymphoblastic leukaemia-A dual-centre experience. J Pediatr Gastroenterol Nutr 2006; 42(5):535–8. 64. Satti MB, Weinbren K, Gordon-Smith EC. 6-Thioguanine as a cause of toxic veno-occlusive disease of the liver. J Clin Pathol 1982; 35(10):1086–91. 65. Marinaki AM, Ansari A, Duley JA, et al. Adverse drug reactions to azathioprine therapy are associated with polymorphism in the gene encoding inosine triphosphate pyrophosphatase (ITPase). Pharmacogenetics 2004; 14(3):181–7. 66. van Dieren JM, van Vuuren AJ, Kusters JG, Nieuwenhuis EE, Kuipers EJ, van der Woude CJ. ITPA genotyping is not predictive for the development of side effects in AZA treated inflammatory bowel disease patients. Gut 2005; 54(11):1664. 67. Allorge D, Hamdan R, Broly F, Libersa C, Colombel JF. ITPA genotyping test does not improve detection of Crohn’s disease patients at risk of azathioprine/6-mercaptopurine induced myelosuppression. Gut 2005; 54(4):565. 68. Duley JA, Marinaki AM, Arenas M, Florin TH. Do ITPA and TPMT genotypes predict the development of side effects to AZA? Gut 2006; 55(7):1048 (author reply 1048-9). 69. De Ridder L, Van Dieren JM, Van Deventer HJ, et al. Pharmacogenetics of thiopurine therapy in paediatric IBD patients. Aliment Pharmacol Ther 2006; 23(8):1137–41. 70. Al Hadithy AF, de Boer NK, Derijks LJ, Escher JC, Mulder CJ, Brouwers JR. Thiopurines in inflammatory bowel disease: pharmacogenetics, therapeutic drug monitoring and clinical recommendations. Dig Liver Dis 2005; 37(4):282–97. 71. Yates CR, Krynetski EY, Loennechen T, et al. Molecular diagnosis of thiopurine S-methyltransferase deficiency: genetic basis for azathioprine and mercaptopurine intolerance. Ann Intern Med 1997; 126(8):608–14. 72. Dubinsky MC, Yang H, Hassard PV, et al. 6-MP metabolite profiles provide a biochemical explanation for 6-MP resistance in patients with inflammatory bowel disease. Gastroenterology 2002; 122(4):904–15. 73. Al-Mukhaizeem KA, Rosenberg A, Sherker AH. Nodular regenerative hyperplasia of the liver: an under-recognized cause of portal hypertension in hematological disorders. Am J Hematol 2004; 75(4):225–30. 74. Reshamwala PA, Kleiner DE, Heller T. Nodular regenerative hyperplasia: not all nodules are created equal. Hepatology 2006; 44(1):7–14. 75. Geller SA, Dubinsky MC, Poordad FF, et al. Early hepatic nodular hyperplasia and submicroscopic fibrosis associated with 6-thioguanine therapy in inflammatory bowel disease. Am J Surg Pathol 2004; 28(9):1204–11.

Hepatotoxicity of Immunomodulating Agents

679

76. Wanless IR, Godwin TA, Allen F, Feder A. Nodular regenerative hyperplasia of the liver in hematologic disorders: a possible response to obliterative portal venopathy. A morphometric study of nine cases with an hypothesis on the pathogenesis. Medicine (Baltimore) 1980; 59(5):367–79. 77. Wanless IR. Micronodular transformation (nodular regenerative hyperplasia) of the liver: a report of 64 cases among 2500 autopsies and a new classification of benign hepatocellular nodules. Hepatology 1990; 11(5):787–97. 78. Shimamatsu K, Wanless IR. Role of ischemia in causing apoptosis, atrophy, and nodular hyperplasia in human liver. Hepatology 1997; 26(2):343–50. 79. Seiderer J, Zech CJ, Diebold J, et al. Nodular regenerative hyperplasia: a reversible entity associated with azathioprine therapy. Eur J Gastroenterol Hepatol 2006; 18(5):553–5. 80. Gane E, Portmann B, Saxena R, Wong P, Ramage J, Williams R. Nodular regenerative hyperplasia of the liver graft after liver transplantation. Hepatology 1994; 20(1 Pt 1):88–94. 81. Li EK, Tam LS, Tomlinson B. Leflunomide in the treatment of rheumatoid arthritis. Clin Ther 2004; 26(4):447–59. 82. Panek JJ, Jezierska A, Mierzwicki K, Latajka Z, Koll A. Molecular modeling study of leflunomide and its active metabolite analogues. J Chem Inf Model 2005; 45(1):39–48. 83. Cherwinski HM, Cohn RG, Cheung P, et al. The immunosuppressant leflunomide inhibits lymphocyte proliferation by inhibiting pyrimidine biosynthesis. J Pharmacol Exp Ther 1995; 275(2):1043–9. 84. Jankovic V, Samardzic T, Stosic-Grujicic S, Popadic D, Trajkovic V. Cell-specific inhibition of inducible nitric oxide synthase activation by leflunomide. Cell Immunol 2000; 199(2):73–80. 85. Vergne-Salle P, Leger DY, Bertin P, Treves R, Beneytout JL, Liagre B. Effects of the active metabolite of leflunomide, A77 1726, on cytokine release and the MAPK signalling pathway in human rheumatoid arthritis synoviocytes. Cytokine 2005; 31(5):335–48. 86. Siemasko K, Chong AS, Jack HM, Gong H, Williams JW, Finnegan A. Inhibition of JAK3 and STAT6 tyrosine phosphorylation by the immunosuppressive drug leflunomide leads to a block in IgG1 production. J Immunol 1998; 160(4):1581–8. 87. Strand V, Cohen S, Schiff M, et al. Treatment of active rheumatoid arthritis with leflunomide compared with placebo and methotrexate. Leflunomide Rheumatoid Arthritis Investigators Group. Arch Intern Med 1999; 159(21):2542–50. 88. Nash P, Thaci D, Behrens F, Falk F, Kaltwasser JP. Leflunomide improves psoriasis in patients with psoriatic arthritis: an in-depth analysis of data from the TOPAS study. Dermatology 2006; 212(3):238–49. 89. Kaltwasser JP, Nash P, Gladman D, et al. Efficacy and safety of leflunomide in the treatment of psoriatic arthritis and psoriasis: a multinational, double-blind, randomized, placebo-controlled clinical trial. Arthritis Rheum 2004; 50(6):1939–50. 90. Friedman LS. Controversies in liver biopsy: who, where, when, how, why? Curr Gastroenterol Rep 2004; 6(1):30–6. 91. van Roon EN, Jansen TL, Houtman NM, Spoelstra P, Brouwers JR. Leflunomide for the treatment of rheumatoid arthritis in clinical practice: incidence and severity of hepatotoxicity. Drug Saf 2004; 27(5):345–52. 92. Charatan F. Arthritis drug should be removed from market, says consumer group. Br Med J 2002; 324(7342):869. 93. Suissa S, Ernst P, Hudson M, Bitton A, Kezouh A. Newer disease-modifying antirheumatic drugs and the risk of serious hepatic adverse events in patients with rheumatoid arthritis. Am J Med 2004; 117(2):87–92. 94. Goldkind L, Simon L. FDA Meeting March 2003: update on the safety of new drugs for rheumatoid arthritis. Part III: safety and efficacy update on leuflunomide (Arava). 2003. (Accessed August 6, 2006 at http://www.rheumatology.org/publications/hotline/0503leffda.asp?audZmem). 95. Thomasset SC, Ong SL, Large SR. Post-coronary artery bypass graft liver failure: a possible association with leflunomide. Ann Thorac Surg 2005; 79(2):698–9. 96. Weinblatt ME, Dixon JA, Falchuk KR. Serious liver disease in a patient receiving methotrexate and leflunomide. Arthritis Rheum 2000; 43(11):2609–11. 97. Sevilla-Mantilla C, Ortega L, Agundez JA, Fernandez-Gutierrez B, Ladero JM, Diaz-Rubio M. Leflunomide-induced acute hepatitis. Dig Liver Dis 2004; 36(1):82–4. 98. Legras A, Bergemer-Fouquet AM, Jonville-Bera AP. Fatal hepatitis with leflunomide and itraconazole. Am J Med 2002; 113(4):352–3. 99. Yao HW, Li J, Chen JQ, Xu SY. Leflunomide attenuates hepatocyte injury by inhibiting Kupffer cells. World J Gastroenterol 2004; 10(11):1608–11. 100. Imose M, Nagaki M, Kimura K, et al. Leflunomide protects from T-cell-mediated liver injury in mice through inhibition of nuclear factor kappaB. Hepatology 2004; 40(5):1160–9. 101. Karaman A, Fadillioglu E, Turkmen E, Tas E, Yilmaz Z. Protective effects of leflunomide against ischemia-reperfusion injury of the rat liver. Pediatr Surg Int 2006; 22(5):428–34.

680

Davern

102. Yao HW, Li J, Chen JQ, Xu SY. Inhibitory effect of leflunomide on hepatic fibrosis induced by CCl4 in rats. Acta Pharmacol Sin 2004; 25(7):915–20. 103. Maini R, St Clair EW, Breedveld F, et al. Infliximab (chimeric anti-tumour necrosis factor alpha monoclonal antibody) versus placebo in rheumatoid arthritis patients receiving concomitant methotrexate: a randomised phase III trial. ATTRACT Study Group. Lancet 1999; 354(9194):1932–9. 104. Hanauer SB, Feagan BG, Lichtenstein GR, et al. Maintenance infliximab for Crohn’s disease: the ACCENT I randomised trial. Lancet 2002; 359(9317):1541–9. 105. United States Food and Drug Administration. 2004 Safety alerts for drugs b, medical devices, and dietary supplements. (Accessed September 15, 2006 at http://www.fda.gov/medwatch/SAFETY/ 2004/safety04.htm#Remicade2). 106. Davern TJ, Moawad D, Ferrell L. Liver injury from TNF antagonists—a class effect? AASLD Single Topic Conference on Drug Induced Liver Injury, 2005. 107. Sarzi-Puttini P, Atzeni F, Capsoni F, Lubrano E, Doria A. Drug-induced lupus erythematosus. Autoimmunity 2005; 38(7):507–18. 108. Bradham CA, Plumpe J, Manns MP, Brenner DA, Trautwein C. Mechanisms of hepatic toxicity. I. TNF-induced liver injury. Am J Physiol 1998; 275(3 Pt 1):G387–92. 109. Yamada Y, Kirillova I, Peschon JJ, Fausto N. Initiation of liver growth by tumor necrosis factor: deficient liver regeneration in mice lacking type I tumor necrosis factor receptor. Proc Natl Acad Sci USA 1997; 94(4):1441–6. 110. Srivastava MD, Alexander F, Tuthill RJ. Immunology of cutaneous vasculitis associated with both etanercept and infliximab. Scand J Immunol 2005; 61(4):329–36. 111. Baecklund E, Askling J, Rosenquist R, Ekbom A, Klareskog L. Rheumatoid arthritis and malignant lymphomas. Curr Opin Rheumatol 2004; 16(3):254–61. 112. Bongartz T, Sutton AJ, Sweeting MJ, Buchan I, Matteson EL, Montori V. Anti-TNF antibody therapy in rheumatoid arthritis and the risk of serious infections and malignancies: systematic review and meta-analysis of rare harmful effects in randomized controlled trials. JAMA 2006; 295(19):2275–85. 113. Thayu M, Markowitz JE, Mamula P, Russo PA, Muinos WI, Baldassano RN. Hepatosplenic T-cell lymphoma in an adolescent patient after immunomodulator and biologic therapy for Crohn disease. J Pediatr Gastroenterol Nutr 2005; 40(2):220–2. 114. Singh G, Fries JF, Williams CA, Zatarain E, Spitz P, Bloch DA. Toxicity profiles of disease modifying antirheumatic drugs in rheumatoid arthritis. J Rheumatol 1991; 18(2):188–94. 115. Mok MY, Ng WL, Yuen MF, Wong RW, Lau CS. Safety of disease modifying anti-rheumatic agents in rheumatoid arthritis patients with chronic viral hepatitis. Clin Exp Rheumatol 2000; 18(3):363–8. 116. Makin AJ, Wendon J, Fitt S, Portmann BC, Williams R. Fulminant hepatic failure secondary to hydroxychloroquine. Gut 1994; 35(4):569–70. 117. Galvan VG, Oltra MR, Rueda D, Esteban MJ, Redon J. Severe acute hepatitis related to hydroxychloroquine in a woman with mixed connective tissue disease. Clin Rheumatol 2007; 26(6):971–2. 118. De Wall SL, Painter C, Stone JD, et al. Noble metals strip peptides from class II MHC proteins. Nat Chem Biol 2006; 2(4):197–201. 119. Watkins PB, Schade R, Mills AS, Carithers RL, Jr., Van Thiel DH. Fatal hepatic necrosis associated with parenteral gold therapy. Dig Dis Sci 1988; 33(8):1025–9. 120. te Boekhorst PA, Barrera P, Laan RF, van de Putte LB. Hepatotoxicity of parenteral gold therapy in rheumatoid arthritis: a case report and review of the literature. Clin Exp Rheumatol 1999; 17(3):359–62. 121. Shah P, Griffith SM, Shadforth MF, et al. Can gold therapy be used more safely in rheumatoid arthritis? Adverse drug reactions are more likely in patients with nodular disease, independent of HLA-DR3 status J Rheumatol 2004; 31(10):1903–5. 122. Gran JT, Husby G, Thorsby E. HLA DR antigens and gold toxicity. Ann Rheum Dis 1983; 42(1):63–6. 123. Brewer GJ, Askari FK. Wilson’s disease: clinical management and therapy. J Hepatol 2005; 42(Suppl. 1):S13–21. 124. Suarez-Almazor ME, Spooner C, Belseck E. Penicillamine for treating rheumatoid arthritis. Cochrane Database Syst Rev 2000; (4):CD001460. 125. Menara M, Zancan L, Sturniolo GC. Penicillamine hepatotoxicity in the treatment of Wilson’s disease. J Pediatr Gastroenterol Nutr 1992; 14(3):353–4. 126. Deutscher J, Kiess W, Scheerschmidt G, Willgerodt H. Potential hepatotoxicity of penicillamine treatment in three patients with Wilson’s disease. J Pediatr Gastroenterol Nutr 1999; 29(5):628. 127. Plosker GL, Croom KF. Sulfasalazine: a review of its use in the management of rheumatoid arthritis. Drugs 2005; 65(13):1825–49. 128. Ribe J, Benkov KJ, Thung SN, Shen SC, LeLeiko NS. Fatal massive hepatic necrosis: a probable hypersensitivity reaction to sulfasalazine. Am J Gastroenterol 1986; 81(3):205–8. 129. Marinos G, Riley J, Painter DM, McCaughan GW. Sulfasalazine-induced fulminant hepatic failure. J Clin Gastroenterol 1992; 14(2):132–5.

Hepatotoxicity of Immunomodulating Agents

681

130. Halmos B, Anastopoulos HT, Schnipper LE, Ballesteros E. Extreme lymphoplasmacytosis and hepatic failure associated with sulfasalazine hypersensitivity reaction and a concurrent EBV infection—case report and review of the literature. Ann Hematol 2004; 83(4):242–6. 131. Poland GA, Love KR. Marked a typical lymphocytosis, hepatitis, and skin rash in sulfasalazine drug allergy. Am J Med 1986; 81(4):707–8. 132. Hautekeete ML, Bourgeois N, Potvin P, et al. Hypersensitivity with hepatotoxicity to mesalazine after hypersensitivity to sulfasalazine. Gastroenterology 1992; 103(6):1925–7. 133. Stoschus B, Meybehm M, Spengler U, Scheurlen C, Sauerbruch T. Cholestasis associated with mesalazine therapy in a patient with Crohn’s disease. J Hepatol 1997; 26(2):425–8. 134. Franks ME, Macpherson GR, Figg WD. Thalidomide. Lancet 2004; 363(9423):1802–11. 135. Moreira AL, Sampaio EP, Zmuidzinas A, Frindt P, Smith KA, Kaplan G. Thalidomide exerts its inhibitory action on tumor necrosis factor alpha by enhancing mRNA degradation. J Exp Med 1993; 177(6):1675–80. 136. Sampaio EP, Sarno EN, Galilly R, Cohn ZA, Kaplan G. Thalidomide selectively inhibits tumor necrosis factor alpha production by stimulated human monocytes. J Exp Med 1991; 173(3):699–703. 137. Keifer JA, Guttridge DC, Ashburner BP, Baldwin AS, Jr. Inhibition of NF-kappaB activity by thalidomide through suppression of IkappaB kinase activity. J Biol Chem 2001; 276(25):22382–7. 138. Fujita J, Mestre JR, Zeldis JB, Subbaramaiah K, Dannenberg AJ. Thalidomide and its analogues inhibit lipopolysaccharide-mediated induction of cyclooxygenase-2. Clin Cancer Res 2001; 7(11):3349–55. 139. Clark TE, Edom N, Larson J, Lindsey LJ. Thalomid (Thalidomide) capsules: a review of the first 18 months of spontaneous postmarketing adverse event surveillance, including off-label prescribing. Drug Saf 2001; 24(2):87–117. 140. Fowler R, Imrie K. Thalidomide-associated hepatitis: a case report. Am J Hematol 2001; 66(4):300–2. 141. Hanje AJ, Shamp JL, Thomas FB, Meis GM. Thalidomide-induced severe hepatotoxicity. Pharmacotherapy 2006; 26(7):1018–22. 142. Trojan A, Chasse E, Gay B, Pichert G, Taverna C. Severe hepatic toxicity due to thalidomide in relapsed multiple myeloma. Ann Oncol 2003; 14(3):501–2. 143. Henson PM. Dampening inflammation. Nat Immunol 2005; 6(12):1179–81.

31

Methotrexate Controversies Adrian Reuben

Division of Gastroenterology and Hepatology, Medical University of South Carolina, Charleston, South Carolina, U.S.A.

INTRODUCTION Methotrexate (MTX), a classic antimetabolite that inhibits folic acid metabolism, has been licensed in the United States since 1953 and commercially available since 1955. An analog of both folic acid and the antecedent folic acid antagonist aminopterin (1) that was withdrawn because of toxicity, MTX competes with 5-methyltetrahydrofolate (found in food and the major folate in serum) and with folinic acid (5-formyltetrahydrofolate or leucovorin) for uptake into cells via the reduced folate carrier. Once inside the cell, MTX is polyglutamylated by folylpolyglutamate synthetase (2), and it is likely that this conversion, which impairs MTX egress from the cell (2), determines MTX’s activity and cytotoxicity (3,4). Intracellularly, MTX inhibits dihydrofolate reductase leading to a reduced supply of tetrahydrofolates (especially methylene- and methyl-tetrahydrofolate), and it also interferes with thymidylate synthase, which together impair synthesis of thymidylate (a pyrimidine precursor). MTX inhibition of glycinamide and aminoimidazoleoxamide ribonucleotide transformylases depletes purines. DNA biosynthesis is stopped and cell death ensues (5,6). Folic acid antagonists were originally used for childhood leukemia (7) but soon aminopterin, and later MTX (amethopterin), were shown to improve arthritic symptoms in patients with psoriasis and rheumatoid arthritis (8–10). Over successive decades, the oncological indications for MTX grew to cover the treatment of a variety of solid tumors including head and neck cancers, breast cancer, gestational and trophoblastic diseases, nonHodgkin lymphoma, pediatric tumors, and sarcoma—especially osteosarcoma. MTX has been used in both conventional and high-dose regimens, often in combination with leucovorin “rescue” and other chemotherapeutic agents. Similarly, MTX treatment of inflammatory disorders has expanded beyond psoriasis and rheumatoid arthritis. MTX is used now for connective tissue disorders, such as the conjunctivitis–sacroiliitis–urethritis (formerly called Reiter’s (11)) syndrome, polymyositis/dermatomyositis, and Wegener’s granulomatosis, as well as for uveitis, asthma, and sarcoidosis. MTX has been known to be hepatotoxic almost since its introduction into clinical practice (12). Thus, because of the seriousness of its hepatotoxic and other adverse effects (bone marrow toxicity, alopecia, mucositis, erythematous skin rashes, nephrotoxicity, interstitial pneumonitis, osteopenia, and neurotoxicity), MTX use was originally restricted to situations in which less hazardous remedies fail. But now there is a growing trend to use MTX early in the course of diseases such as rheumatoid arthritis, as rheumatologists turn their therapeutic pyramid on its head (13,14). In this context, the perennial controversy concerning the stringency of monitoring for hepatotoxicity persists, while the second controversy, whether MTX is an effective and safe treatment for certain liver diseases per se, notably primary biliary cirrhosis (PBC), seems to have faded. The controversy over the need for pre-MTX liver biopsies and the appropriate surveillance of patients on long-term MTX therapy for various inflammatory disorders had largely been resolved, especially for the treatment of rheumatoid arthritis and to a certain extent for the treatment of psoriasis. But recently the debate has been resurrected. The author of this chapter has relationships with the following corporations: clinical trials for Roche, Schering, Debiovision, Coley, and Valeant. He is on the speaker bureaus of Roche and Schering.

684

Reuben

Guidelines that have been published for hepatotoxicity monitoring in the treatment of rheumatoid arthritis (15) and psoriasis (16,17) are now being challenged, and since adherence to these published guidelines was not universally consistent (18,19), it is to be hoped that the less demanding recommendations being proposed now will be easier to follow. Guidelines are also emerging for monitoring MTX therapy for more recent indications such as Crohn’s disease, based on careful clinical, biochemical, and liver histological follow-up (20). In accordance with a growing consensus, the lively duel that once took place between protagonists and antagonists of MTX therapy for PBC and other liver disorders (21–28) has lost much of its passion, as the antagonists appear to have all but vanquished the protagonists (29). ACUTE LIVER INJURY CAUSED BY MTX As with many other drugs, reversible elevations of aminotransferases are quite common following initiation of MTX therapy, with a prevalence of approximately 14% in one report (30). In patients treated with cyclical therapy, the enzymes rise with each course— usually higher the more frequent the dosing—but subside within a month. With high-dose therapy (with or without leucovorin rescue), aminotransferases may rise up to 40-fold above normal, sometimes accompanied by hyperbilirubinemia (31,32) and with an occasional clinical presentation of “acute hepatitis” (33,34), by which is meant acute necrotic liver injury. Yet, even when there is deep jaundice and dramatic increases in alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels (35), these abnormalities resolve within a few weeks of stopping MTX almost invariably without long-term sequelae (36). Liver biopsy may show steatosis but not fibrosis or cirrhosis (36). Indeed, the severity of acute hepatotoxicity induced by MTX is thought to predict for a good oncological therapeutic response (37). Rare cases of reversible liver failure have been described (38) but incrimination of MTX as the sole culprit in these cases was not unequivocal. A case of rapidly progressive sub-fulminant liver failure requiring transplantation has also been reported (39). CHRONIC LIVER INJURY CAUSED BY MTX High-Dose MTX The most significant long-term side effect of chronic MTX therapy is hepatic fibrosis leading to cirrhosis. High-dose daily MTX, whether given for leukemia in children (12,40,41) or for psoriasis in adults (42–45), was associated with the development of hepatic fibrosis or cirrhosis. One impediment to assessing the true role of MTX in causing cirrhosis in adults with psoriasis had been the lack of baseline pre-MTX liver biopsies in many of the older studies. This was an especially important omission since it is known that liver histological abnormalities are present in psoriatic patients, even before MTX therapy (46). Preexisting liver abnormalities in patients with psoriasis, mostly steatosis (47), may be related to the systemic dermatological condition itself and/or comorbid conditions such as alcohol excess, obesity, diabetes, exposure to other hepatotoxins, hepatitis C, and so forth. Portal inflammation, steatosis, and portal fibrosis are also found in patients with rheumatoid arthritis (48,49). In some series, fibrosis has been reported in up to 22% and cirrhosis in 0% to 2% of psoriatic patients without MTX therapy (50). The prevalence of cirrhosis in psoriatics was reported in another series as being 0.6% without MTX therapy and as high as 25% after five years of 50 mg weekly bolus MTX treatment (51). In those patients with more than mild fibrosis, alcohol abuse and other causes of liver injury could be identified in most cases. Notwithstanding the difficulties in estimating the absolute fibrogenic potential of MTX (because of uncertainty about underlying preexisting liver pathology and other variables causing liver injury), there is little doubt that high-dose MTX, whether daily or weekly, does cause hepatic fibrosis and cirrhosis that may even deteriorate to warrant liver transplantation (52). Histological Changes Due to MTX Mindful of the pitfalls in interpretation alluded to above, there is agreement about the histopathological changes caused by MTX. The earliest changes are ultrastructural and include lysosomal and mitochondrial injury, endoplasmic reticulum hypertrophy, autophagic

685

Methotrexate Controversies

vacuole formation, and desmosomal injury (53,54). There is bile duct damage (55) and stellate cell (Ito cell, lipocyte, or fat-storing cell) hyperplasia (53,56). It has been presumed that the latter leads to deposition of collagen in the space of Disse (57), causing fibrosis and ultimately cirrhosis (58). A propensity for periportal inflammation to proceed to periportal, sinusoidal, and bridging fibrosis has been suggested (59), but it is dubious that MTX causes bona fide chronic hepatitis (38). An intriguing alternative or additional explanation for MTX-induced fibrosis has been put forward recently, namely that it originates with scarring of the canals of Hering that have undergone fibrous obliteration (60). Other lesions attributed to MTX include marked macrovesicular steatosis, zone 3 focal hepatocyte degeneration, hepatocyte nuclear pleomorphism and multinucleation, and Kupffer cell proliferation and ductular reaction (60,61). Roenigk et al. classified these light microscopic changes into five levels of severity (16), to permit ease of comparison both within and between patients (Table 1). Grades I and II of the Roenigk scheme refer to steatotic, nuclear, and necroinflammatory changes only, whereas grades III (a and b) and IV include degree of fibrosis. Although this grading system has been widely used to grade liver histopathology in psoriatic patients on MTX treatment (17) especially when deciding to continue or discontinue treatment, it is grossly subjective, insensitive to small changes, has features of unclear significance (like nuclear pleomorphism) and is only semiquantitative (especially for fibrosis). The Roenigk score has not been validated and is not used in any other liver disease assessment. Moreover, when the Roenigk scale was compared to well-accepted grading protocols, like those of Ishak and Scheuer (63), the Roenigk system performed poorly and, especially, it consistently overestimated the degree of fibrosis (64). Other grading systems have been proposed to supplant (64,65) or modify (66) the Roenigk scheme, including quantitative estimates of collagen content by image analysis (67). The fibrosis caused by MTX is typically periportal with extensions into the parenchyma in a sinusoidal or “chicken-wire” pattern, reminiscent of the fibrosis of alcoholic and nonalcoholic fatty liver disease. MTX cirrhosis is usually micronodular (59). Mechanism of MTX-Induced Liver Injury The mechanism of MTX-induced liver injury, especially chronic injury, is unknown but has been attributed to the intracellular buildup of polyglutamylated MTX and the toxic effects of 7-hydroxymethotrexate, the major MTX metabolite (68). However, it is unclear whether the fibrosis is initiated by hepatocyte or bile duct injury, by independent activation of hepatic stellate cells or by damage to progenitor cells in the canals of Hering (60). The earliest changes seen in the liver of MTX-treated patients before light microscopic abnormalities occur, is the increased appearance of matrix proteins, several collagens, and transforming growth factor (69), which suggests a primary role for activated stellate cells. TABLE 1 Roenigk Classification of Liver Histology in Chronic MTX Therapy Histology Classification

Fatty change

Nuclear pleornorphism

Grade I Grade II

Mild or none Moderate to severe

Mild or none Moderate to severe

None None

Grade IIIa

0/C

0/C

Grade IIIb Grade IV

0/C 0/C

0/C 0/C

Mild (fibrotic septa extending into the lobule) Moderate to severe Cirrhosis

0/C denotes absent/present. Source: Modified from Refs. 15, 61.

Fibrosis

Necroinflammatory change Portal tract inflammation mild Portal tract inflammation moderate to severe Hepatocellular necrosis moderate to severe 0/C 0/C 0/C

686

Reuben

Unfortunately, there is no animal model that mimics human MTX liver injury well. Acute MTX toxicity in the rat is cholestatic and appears to be caused by 7-hydroxymethotrexate precipitation in bile (68). Perfusion of isolated rat livers with high-dose MTX gives ultrastructural disruption that is similar to that produced in vivo, namely disruption of the hepatic plates, swelling and vacuolation of the hepatocytes with striking changes in the plasma membrane, microvilli, intercellular junctions, and sinusoidal endothelium (70). There is also disorganization of the endoplasmic reticulum and mitochondria, and redistribution of glycogen. Chronic daily oral administration of MTX to rats causes zone 3 necrosis and Kupffer cell enlargement and, in some animals, fibrosis (71). MTX-induced steatosis appears to result from interference with methionine metabolism and transmethylation reactions, as the striking steatogenic effect of MTX in rats can be mitigated by choline administration (72). Low-Dose MTX Therapy Psoriasis The enthusiasm for MTX as effective therapy for proliferative and inflammatory conditions like psoriasis and rheumatoid arthritis, respectively, was soon tempered by its toxicity in patients dosed daily, notably the advent of advanced hepatic fibrosis. Many years later, low-dose weekly regimens of MTX were introduced and found to be effective but there was still disagreement over safety and, in particular, the risk of advanced fibrosis and cirrhosis. MTX generally has been reserved for patients with moderate to severe psoriasis, meaning psoriatic erythroderma, moderate to severe psoriatic arthritis, more than 20% body surface involvement, localized pustular psoriasis, lack of response to phototherapy, PUVA, and retinoids, or psoriasis that affects certain areas of the body so that normal function and employment are prevented (73). In short, for MTX use, the psoriasis should be life-ruining physically, emotionally, or economically. Nonetheless, and despite the advent of new biological therapies too (73), MTX is the most widely used drug in Europe for psoriasis (19), and likely elsewhere too (73). The results of numerous biopsy studies, in both psoriasis (25 studies) and rheumatoid arthritis patients (17 studies) taking long-term weekly MTX, were reviewed in detail recently by West (50). In psoriatics on MTX, the prevalence of fibrosis ranged from 14% to 34% and of cirrhosis from 0% to 21%, but conclusions from these studies were often compromised by the lack of baseline biopsies. Unfortunately, even when studies were done of paired pretreatment and on-treatment histology, there was poor agreement over the risk of MTX causing fibrosis. In some series, clinically significant fibrosis that dictated cessation of therapy was rare (74), even when cumulative doses of 5.1 g were used (75). At the other extreme (51), 13% of patients who ingested a 2.2-g cumulative dose and 26% of those who ingested 4-g developed cirrhosis. Excluding extremes, the likely cirrhosis rate for psoriatics treated with MTX was estimated at approximately 7% to 10% (74–79), including a prediction of 6.7% increases in risk of progression of fibrosis for every additional gram of drug ingested (80). Although the reasons for those discrepancies are not proven, it seems likely that comorbid clinical variables influence the risk of developing MTX-associated fibrosis and cirrhosis. The most important of these appears to be heavy (but not light) (81) alcohol use. For example, previous or ongoing heavy alcohol use increases the risk of MTX-induced fibrosis 2.5- to 5-fold (80). It seems that weekly alcohol ingestion of 100 g is sufficient to influence the risk of progression to cirrhosis (80). In an enlightening retrospective analysis of his study published 25 years earlier, Zachariae recently drew attention to the discovery that several of his psoriatic patients who developed cirrhosis while taking MTX later admitted their heavy alcohol consumption (82). Obesity and diabetes together enhance the fibrogenic potential of MTX (83,84), but it is unclear whether either does so alone (78,85,86). Other potentiating factors are preexisting liver disease (46,77), excessive vitamin A ingestion (51,82), and renal failure (51,85), presumably because the latter raises MTX blood levels. Historically, prior arsenical therapy potentiated MTX hepatotoxicity (51,82), but this is no longer a consideration. Whether advancing years aggravate MTX-induced liver damage is unclear (46,77,86–88) but neither severity of psoriasis (86), gender (83), HLA phenotype (89), nor corticosteroid treatment (86) appears to influence MTX liver damage. Aside from comorbid conditions that must now include nonalcoholic steatohepatitis (90) and hemochromatosis (91), the single most important factor in MTX-induced liver fibrosis in psoriasis is the dosing regimen,

Methotrexate Controversies

687

and arguably the cumulative dose. Some years ago we performed a liver transplant in a patient who had received daily MTX doses for three years; there were no other risk factors for liver disease. Daily dosing has not been the standard of care in psoriasis and rheumatoid arthritis since the 1970s, when the change to weekly dosing was adopted. Even low doses, however, can result in up to a fourfold increase in prevalence of advanced fibrosis if MTX is given on a daily or alternate daily regimen (44,45,86). Daily dosing should be avoided. The vexing question whether MTX hepatotoxicity in psoriasis is related strictly to cumulative dose or to duration of therapy is important, as the answer will dictate the intensity of monitoring and, in particular, the frequency of liver biopsy. Even with weekly treatment, high dosing, i.e., in excess of 20 mg (51,81,92) is associated with cirrhosis more often than when lower doses are used (77,93). Liver Biopsy

In patients with risk factors for liver disease, it is still recommended that a baseline biopsy be done (62). However, since a small percentage of psoriasis patients may not continue to take MTX after the initial two to four months of therapy (because of adverse effects, lack of efficacy, etc.), the first liver biopsy can be postponed for this period until it is certain that long-term treatment is needed. In patients with risk factors for liver disease, namely those with a prior or current history of alcohol excess, abnormal liver test results, chronic hepatitis B or C infection, obesity or diabetes, other hepatotoxin exposure, or a family history of an inheritable liver disease, the baseline biopsy should be done early (i.e., before MTX is started or within the first two to four months of therapy). In psoriatics without these risk factors, the most recent official guidelines permit postponement of the first biopsy until the patient has consumed 1 to 1.5 g of drug (62). However, in a small but well-conducted recent study of serial liver biopsies in psoriatic patients receiving less than 20 mg MTX weekly (93), no significant hepatic fibrosis was seen at a mean cumulative dose of 1.5 g, whereas the cumulative probabilities of developing advanced hepatic fibrosis were 2.6% at 3 to 4.5 g and 8.2% at 5.6 g. In carefully supervised lowrisk patients, it is reasonable to postpone the first biopsy until a cumulative MTX dose of 4 g has been consumed (94). Some studies have shown a correlation between the degree of liver injury and cumulative dose (44,51,76,86,93,95,96) while others have not (45,46,75,76,78,79,82,88,97). It has been suggested that continuation of MTX following demonstration offibrosis or cirrhosis may actually not always lead to disease progression (45,51,76,93) and, contrary to expectation, only rarely to liver decompensation (93,98). These latter experiences have prompted some authorities to question the need for frequent liver biopsy in MTX-treated psoriasis (82,94,99), except in highrisk patients. Rather they suggest reducing the frequency of biopsy and substituting monitoring by serial measurement of serum aminoterminal peptide of type III procollagen (PIIIP), dynamic hepatic scintography nuclear medicine scanning or other noninvasive testing (99,100). Although the latest consensus opinion (62) still recommends follow-up biopsy every 1 to 1.5 g of incremental cumulative dose of MTX as long as the baseline liver biopsy was normal, it would appear that 2 to 4 g intervals are safe if there is little or no fibrosis until 10 g MTX consumption is reached (93,94). If liver chemistries (aminotransferases, alkaline phosphatase, bilirubin, and albumin) are abnormal (62,100), then repeat biopsy should be done after the next cumulative dose of 0.5 to 1 g (or approximately after 6 to 12 months of further therapy). Patients showing Roenigk grade I or II changes, i.e., without fibrosis (Table 1) may continue on therapy. Those with grade IIIa changes (corresponding to Ishak stage 2–3, Scheuer stage 2) (63) should undergo repeat liver biopsy six to 12 months later (or change to alternative therapy). The counsel of perfection for those with grade IIIb and IV changes (Ishak stage 4–6, Scheuer stage 3–4) is to stop MTX therapy if feasible. For some patients, discontinuing MTX is unacceptable because of the medically, emotionally, or economically disabling nature of their uncontrolled psoriasis. The decision not to interrupt MTX in those circumstances can only be taken if the patient is made fully aware of the risks of decompensated liver disease and signs an informed consent to continue with therapy. Some argue that follow-up liver biopsy should only be done if discontinuation of MTX is an option (94). Fortunately, new therapeutic regimens now allow MTX dose interruption (101), reduction or even discontinuation under cover of biological therapy (73,102).

688

Reuben

Toxicity Monitoring

Guidelines for monitoring MTX therapy in psoriasis have been established and revised repeatedly since 1972 (16,17,62,103,104). The latest recommended monitoring protocol consists of baseline and interval blood testing, baseline and interval urinalysis, and liver biopsy, as shown in modified form in Table 2. Once the mainstay of monitoring and clinical management, liver biopsy is likely to have less impact in the future, except perhaps to indicate that modification or change in therapy is needed, to trigger surveillance for complications of cirrhosis, and to check for and treat comorbid risk factors for liver disease progression, especially alcohol abuse. The time has also come to abandon the Roenigk Classification and to use one of the fibrosis scores that are in vogue for other liver disorders (63). Several noninvasive tests have been proposed for MTX monitoring but none have yet proved reliable enough to replace liver biopsy as a means of detecting significant fibrosis. Liver enzymes are poor predictors of liver injury in psoriatics, as 30% to 50% of patients will have normal aminotransferases despite significant histological abnormalities (16,86). Certainly, elevated bilirubin and/or enzymes or decreased serum albumin (107) are causes for concern. The diagnostic benefit of frequent hepatic panel monitoring, say every four to six weeks, has not been tested although it would certainly add to the cost of care. In patients with rheumatoid arthritis, four to eight weekly hepatic panel monitoring is still advocated (see below) as it is thought to be useful in early detection of significant liver injury. Some authors recommend four to eight weekly hepatic panel testing in psoriatics, too (100). Fasting serum bile salt concentrations (108), aminopyrine breath tests (109), galactose tolerance tests (110), and antipyrine clearance (111) have all failed as screening tests for MTX hepatotoxicity. More recently, measurement of serum levels of extracellular matrix derivatives and indicators of cytokine activation—used alone (112,113) or in combination with other markers, clinical features, imaging and other biochemical or hematological variables (114–120)—have been explored as surrogate tests for fibrosis in several liver diseases. But, whereas these models and TABLE 2 Monitoring of MTX Therapy in Psoriasis A. Baseline or prior to MTX therapy (i) Blood tests CBC BMP (i.e., BUN, creatinine, electrolytes, calcium, glucose) Hepatic panel (total bilirubin, AST, ALT, alkaline phosphatase, albumin) Iron, total iron-binding capacity, ferritin HBsAg, anti-HCV HIV antibodies (in high-risk patients) (ii) Urinalysis (iii) Liver biopsy (a) High-risk patients before or within 2 4 mo of initiation of MTX therapy (b) Low-risk patients after 4 g of MTX therapy (iv) Serum PIIIP measurement B. During MTX therapy (i) Blood tests CBC weekly for 2 wk, biweekly for next month, then approximately monthly BMP at 3 4 monthly intervals Hepatic panel every 3 4 mo (or more frequently until first liver biopsy is done see A (iii)) (ii) Urinalysis 3 4 monthly intervals (iii) Serum PIIIP levels every 2 3 mo (iv) Liver biopsy (a) High-risk patients after initial biopsy, at MTX cumulative doses of 1.5 and 3 g and in additional 1 to 1.5 g increments (b) Low-risk patients at MTX cumulative dose of 4 g, then at 2 to 4 g cumulative increments, depending upon histological findings and especially O10 g (c) Abnormal hepatic panel follow-up biopsy after incremental 0.5 to 1 g or after a further 6 12 mo of MTX (d) New or higher elevation in serum PIIIP level Abbreviations: MTX, methotrexate; CBC, complete blood count; BMP, basic metabolic panel; AST, aspartate aminotransferase; ALT, alanine aminotransferase; PIIIP, peptide of type III procollagen. Source: Modified from Refs. 16, 17, 62, 94, 95, 105, 106.

Methotrexate Controversies

689

indices are usually considered to show “promise,” most reviewers agree that they are not yet fully reliable substitutes for liver biopsy (121–124). In some diseases, the derived indices have been thought to be particularly reliable in predicting either the absence of fibrosis (125) or the presence of cirrhosis (126,127). Although results for individual markers often show overlap among patients with normal and fibrotic liver, especially at early stages of fibrogenesis before cirrhosis is established (112,113), PIIIP is thought to be one of the most promising markers thus far for monitoring patients with psoriasis on MTX. PIIIP does indeed rise during MTX therapy (128,129), but this occurs when liver histology is normal, only steatotic, or distinctly abnormal (130,131). In patients with normal liver biopsies by light microscopy, fibrosis can be seen in the space of Disse by electron microscopy (132). This could account partly for the elevation of PIIIP levels seen in these individuals, possibly reflecting subtle MTX-induced liver injury (132). Of course, inflammation anywhere in the body can also raise PIIIP levels. Advocates of PIIIP monitoring claim that persistent normality over repeated testing excludes significant liver injury and fibrosis (129,130). Justifiably, it appears that reliance on persistently normal PIIIP levels can avoid unnecessary liver biopsies (131), in almost half the patients on MTX therapy (105), resulting possibly in seven times fewer biopsies without missing liver toxicity (106). Older studies showed that colloid-isotope liver–spleen scans (133), computed tomography (134), magnetic resonance imaging (135), and isotope hepatobiliary scans (136) do not predict for significant early liver injury, and none are recommended for patient monitoring. Whereas liver sonography detects fat and fibrosis fairly reliably (136), it cannot distinguish between them. More advanced imaging techniques are fairly reliable in detecting hepatic steatosis, but cannot tell whether steatohepatitis is present too (137). Advanced Doppler sonography maybe helpful, although not foolproof, in detecting advanced fibrosis or cirrhosis (138,139). An early vote for dynamic hepatic scintigraphy (140) has been seconded to some extent for excluding liver histological abnormalities. Normality is predicted when there is greater than 52% portal contribution to liver blood flow (141) although this technique cannot grade severity of liver damage, when the latter is present. Neither patients (142) nor their physicians (18) are enthusiastic about liver biopsy, which explains the enthusiasm for the development of noninvasive screening tests for early hepatic fibrosis. Meanwhile, adherence to the so-called Manchester U.K. protocol (105,106), appears to be a safe and reliable way to follow MTX therapy in patients for cirrhosis while reducing the number of liver biopsies performed (143). Hepatotoxic MTX Drug Interactions

Nowadays physicians must actively look for potential interactions between therapeutic drugs (either prescribed or over-the-counter), herbal remedies, complementary (“alternative”) medicines, and foodstuffs that can interfere with the absorption, pharmacokinetics, metabolism, and disposal of other drugs they are prescribing. Such interactions can blunt or enhance the therapeutic action of prescribed drugs, and also cause toxicity. Additive and synergistic drug toxicity can also occur at the target organ level. The primary route of elimination of MTX is via the kidneys. Drugs that decrease renal clearance of MTX, and thereby enhance its toxicity, include recognized nephrotoxins (e.g., aminoglycosides, cyclosporin, and tacrolimus), some antibiotics (e.g., penicillins, cephalosporins, and sulfonamides), and many agents used for arthritis [salicylates, other nonsteroidal anti-inflammatory drugs (NSAIDs), and colchicine]. Ethanol has already been discussed as a synergistic hepatotoxin. Many drugs increase free blood levels of MTX by displacing it from protein binding in the serum (salicylates, probenecid, phenytoin, retinoids, sulfonylureas, and tetracycline) whereas others, such as dipyridamole, potentiate the intracellular accumulation of MTX and enhance its cytotoxicity. Other folate antagonists may act synergistically with MTX but this is more likely to cause bone marrow depression than hepatotoxicity. A full list of such interactions may be found in a review by Evans and Christensen (144), which clearly needs updating, especially with respect to interference with MTX metabolism. Kremer’s recent exhaustive review of MTX metabolism (6) gives the molecular and cellular basis for interference by insulin, corticosteroids, estrogens, and many other compounds that are weak organic acids and/or have aromatic and carboxylic side-chains.

690

Reuben

Rheumatoid Arthritis and Other Rheumatic Diseases Rheumatoid Arthritis

Many of the conclusions that were drawn from the extensive yet often disparate data on MTX therapy of psoriasis were originally applied to MTX therapy of rheumatoid arthritis. Rheumatologists initially endorsed and adopted the monitoring protocol used in MTX treatment of psoriasis (145). The data in patients with rheumatoid disorders seemed consistent with that in psoriatics. First, it was found that many patients with rheumatoid arthritis have some degree of underlying liver pathology that must be taken into account when judging lesions attributed to MTX. Second, some rheumatoid arthritis patients appear to be at greater risk than others of suffering MTX liver damage. As in psoriasis, conditions that predispose to hepatic steatosis and fibrosis, especially heavy alcohol use and diabetes combined with obesity, contribute to and even potentiate MTX hepatotoxicity. Similarly, underlying liver diseases, such as chronic hepatitis B and C, are also risk factors that may lead to fibrosis. Third, liver injury in patients with rheumatoid arthritis is arguably related to the extent of MTX exposure in terms of cumulative dose and possibly duration of therapy. However, an important and widely accepted difference between MTX therapy of psoriatic and rheumatoid patients is that hepatotoxicity appears to be less frequent in rheumatoid arthritis—possibly 2.5- to 5-fold lower—than in psoriasis. Thus, less aggressive liver biopsy monitoring is recommended for uncomplicated MTX-treated rheumatoid arthritis compared to psoriasis (13,146) (Table 3). In these guidelines, as with psoriasis monitoring, recommendations are based on liver histological abnormalities according to the Roenigk grading system, which the results of at least one study have shown to TABLE 3 Monitoring of MTX Therapy in Rheumatoid Arthritis A. Baseline or prior to MTX therapy (i) Blood tests CBC BMP Hepatic panel HBsAg, anti-HCV (ii) Liver biopsy High-risk patients only Prior excessive alcohol consumption Persistently abnormal baseline AST values Chronic hepatitis B or C infection Obesity and/or insulin-dependent diabetes B. During MTX therapy (i) Blood tests CBC, creatinine, AST, ALT, and albumin monthly for the first 6 mo then 4 8 weekly intervals thereafter For: (a) 1-fold !AST ! 2-fold ULN repeat in 2 4 wk (b) 2-fold !AST ! 3-fold ULN monitor every 2 4 wk with dose reduction (c) AST O 3-fold ULN discontinue MTX, do liver biopsy (ii) Liver biopsy Either when 5 of 9 AST values are abnormal in a given 12-month interval or when 6 of 12 AST values are abnormal with monthly testing, or when serum albumin is decreased despite good rheumatoid arthritis control (iii) For the following liver biopsy results (a) Roenigk grades I, II, or IIIa resume MTX and monitor as in B(i) and (ii) (b) Roenigk grade IIIb or IV discontinue MTX (iv) Discontinue MTX if AST or albumin abnormalities persist and the patient refuses liver biopsy Abbreviations: MTX, methotrexate; CBC, complete blood count; BMP, basic metabolic panel; AST, aspartate aminotransferase; ALT, alanine aminotransferase; ULN, upper limit of normal. Source: Modified from Refs. 13, 146.

Methotrexate Controversies

691

be unsatisfactory (66), and for which an infrequently used alternative system was suggested (147). Again, it would be preferable to use one of the popular accepted histopathology grading schemes for chronic hepatitis (63), or the recently devised classification for nonalcoholic liver disease (148), which resembles MTX injury in many ways (149). Although liver histological abnormalities are common in rheumatoid arthritis, fibrosis is now considered to be either uncommon or absent, unless there are comorbid causes of liver injury. Mild fibrosis (Roenigk grade IIIa) occurred in a maximum of 15% of rheumatoid arthritis patients in one series (49), in which two-thirds of all the patients had only Roenigk grade I changes and 17% had grade II. Most series that examined liver histology in untreated rheumatoid arthritis patients are devoid of cirrhosis, with rare exceptions (146,149). Similarly, the frequency of severe liver disease is low in rheumatoid arthritis patients on long-term MTX treatment (150). Admittedly, mild hepatic fibrosis can be seen with MTX therapy (146) and progression can occur (151), but MTX-induced cirrhosis is distinctly unusual. Stability of liver histology is the rule with MTX therapy in rheumatoid patients and even histological improvement has been reported (48,50,150,152,153). The worst estimate reported for cirrhosis in MTX-treated rheumatoid arthritis was a point prevalence of 2.6% (150), whereas, out of 719 MTX-treated patients whom were reported in 15 other studies reviewed by Kremer et al. (146), only 2 (0.3%) developed cirrhosis. In those instances in which cirrhosis occurs in MTX-treated rheumatoid arthritis, a potentiating cause such as obesity with diabetes (152–154) or alcohol abuse (81,153) can usually be found (155). In this setting, hepatic decompensation is a late event (156). On the whole, MTX is well tolerated in rheumatoid arthritis patients and its toxicity profile compares favorably with that of other disease-modifying antirheumatic drugs (13,157–159), and may even provide a substantial survival benefit by reducing cardiovascular mortality (160). Although liver enzyme elevation occurs commonly with MTX use, it is rare for this to necessitate cessation of therapy (161–163). Concomitant therapy with folate supplements ameliorates aminotransferase elevations (164) and this is now recommended adjuvant therapy (165,166) that may even offer a survival benefit too (167). Hydroxychloroquine may reduce MTX hepatotoxicity (168) by reducing MTX bioavailability (169), whereas aspirin (168) and NSAIDs (170) may exacerbate it. Results of more recent studies have emphasized the importance of obesity, untreated hyperlipidemia, the use of NSAIDs, lack of folate supplements, and renal impairment as risk factors for MTX toxicity, which also is more common in women than in men, and in those with longer duration of rheumatoid disease evolution (171–173). The accumulation of MTX in the liver does not appear to influence either its clinical efficacy or its toxicity (174). Instead, MTX toxicity may be better predicted in the future by assessing such molecular factors as the genotypes of methylenetetrahydrofolate reductase (175) and other folate-dependent enzymes (176), and the effects of MTX on enzymes of purine metabolism, especially the activity of serum 5 0 -nucleotidase, which correlates inversely with MTX hepatotoxicity (177). Studies of the pharmacogenetics of the intracellular targets of MTX (6) may allow better prediction in the future of MTX toxicity and efficacy (178). Kremer has analyzed the reasons why MTX-treated rheumatoid arthritis patients seem to fare better than their psoriasis counterparts (179). He proposes that lessons learned from the treatment of psoriasis were used to the advantage of rheumatoid patients, specifically by enforcing a strict ban on alcohol use (151,180,181) and by reducing the use of potential hepatotoxins (like steroids and NSAIDs), which was feasible because of a favorable response to MTX therapy. Kremer further stresses the benefit of frequent hepatic panel testing that was practiced by rheumatologists, who would reduce MTX doses whenever AST values rose or albumin levels fell. In earlier studies in MTX-treated psoriatic patients, doses of MTX were not lowered unless aminotransferases were two to three times elevated above normal. This difference in MTX dose regulation probably accounts partly for the low prevalence of MTX hepatotoxicity experienced by rheumatoid arthritis patients. The hypothesis that frequent blood testing is useful in managing MTX therapy was tested in a prospective study of 94 MTX-treated rheumatoid arthritis patients (in three cohorts), in whom very frequent AST and ALT testing was done at intervals as often as two weekly preceding liver biopsy (182). A total of 354 follow-up liver biopsies were done in these patients. The investigators found that the prevalence of abnormal ASTs and ALTs in the interval before

692

Reuben

liver biopsy correlated with the degree of histological injury, judged by the Roenigk grade. MTX doses were reduced when AST and albumin results were abnormal. There were no instances of bridging fibrosis (Roenigk grade IIIb) or cirrhosis (Roenigk grade IV) in these three cohorts of “teetotal” patients. Whereas the sensitivity of an elevated AST value to detect mild early fibrosis (Roenigk grade IIIa) was only 12%, if ASTs were elevated on less than half the occasions tested there was 97% certainty of having normal liver histology. The authors attributed the absence of severe lesions to their insistence on strict alcohol abstinence, and frequent adjustments of weekly MTX dose when they observed abnormalities of aminotransferases and serum albumin (179). Alkaline phosphatase elevations occurred in about one-third of the treated patients but were thought to be too sensitive to be of any value in MTX management. Of interest, the authors found that the probability of missing an elevated AST by infrequent sampling ranged from 9% for 30-day sampling to as high as 68% with 3-month sampling. From these and other data are derived the latest (revised) guidelines for monitoring MTX treatment of rheumatoid arthritis (13,15,146), as shown in Table 3. Presumably, these data have persuaded some dermatologists to suggest that four to eight weekly blood testing intervals be applied to MTX monitoring in psoriasis too (100). As in psoriasis, there is disagreement over whether the magnitude of the cumulative MTX dose has any impact on hepatotoxicity. Although the results of some studies have shown that AST elevations are more frequent in rheumatoid patients taking up to 25 mg MTX weekly compared with those on 10 to 15 mg weekly (183,184), this is not reproducible (185,186). There are, nonetheless, many published examples of extremely high cumulative MTX doses and prolonged durations of therapy in rheumatoid arthritis that did not lead to cirrhosis. In one study of 23 patients, up to 10 g of MTX was administered for over 10 years (152), yet only two patients progressed to Roenigk grade IIIb (bridging fibrosis) when therapy was continued after Roenigk grade IIIa changes (mild fibrosis) were found. In two other prospective studies, only 1 of 10 patients treated for six years (187) and 7 of 18 patients receiving 2 to 7 g MTX (161) developed mild fibrosis at worst, and none had cirrhosis. Whether it can really be proved that rheumatoid patients are inherently less susceptible to MTX toxicity then psoriatics, as some authors claim (188), is doubtful. However, it is clear that alcohol abstinence and minimizing MTX doses (whether guided by frequent AST monitoring or not) are safe strategies in treating both psoriatics and patients with rheumatoid arthritis. Liver Biopsy. In contrast to patients with psoriasis, baseline liver biopsies are recommended by the American College of Rheumatology (ACR) only for rheumatoid patients embarking on MTX therapy who are at high risk for underlying liver disease (13,15,146). This includes those with a history of extensive alcohol consumption, with hepatitis B or C infection, or with persistently elevated AST (and/or ALT). Similarly, follow-up liver biopsy on treatment is recommended only if AST elevation of any magnitude persists (i.e., 5 of 9 abnormal ASTs in a given 12-month interval or 6 of 12 when AST is tested monthly) or if a subnormal serum albumin cannot be explained on the basis of uncontrolled rheumatoid disease or, by implication, by proteinuria, or another nonhepatic cause (Table 3). Guidelines from the Health and Public Policy Committee of the American College of Physicians (ACP) (189) support the ACR ruling that restricts baseline liver biopsies to high-risk patients, whereas the American College of Gastroenterology (ACG) recommended baseline biopsies in patients with rheumatoid arthritis and psoriasis alike, before starting MTX (17). The ACG recommendation, which is no longer valid, was based on data available at the time, which showed that significant yet unsuspected fibrotic liver disease was common in rheumatoid patients (49,79,87,149). If follow-up liver biopsy shows Roenigk grade IIIa changes (non-bridging fibrosis) or less, MTX therapy can continue and standard monitoring is resumed. This contrasts with the recommended care of MTX-treated psoriatics, in whom the finding of grade IIIa change prompts a repeat biopsy after 6 to 12 months of further therapy (Table 3). If the biopsy in rheumatoid patients shows Roenigk grade IIIb (bridging fibrosis) or IV (cirrhosis) on MTX treatment, the drug is discontinued.

Methotrexate Controversies

693

Toxicity Monitoring. The ACR protocol for laboratory testing (13,15,146) in rheumatoid arthritis is more intense and stringent than that of the ACP (189) and the ACG (17). Current guidelines state that aminotransferases and albumin should be measured at monthly intervals for the first six months and then at four to eight weekly intervals, and any elevation of AST or ALT, or reduction in albumin is considered significant. ACP recommended monthly laboratory testing but considers enzyme rises significant only if they are threefold elevated above normal; ACG recommended one to two monthly testing. Although the utility of serial AST, ALT, and albumin testing of rheumatoid patients taking MTX has been challenged (190–192), this protocol is well defended by its principal author (193,194) as being derived from the largest data set ever published in which the effect of a potential hepatotoxin (MTX) on liver histology and simultaneously measured liver enzymes were examined (181,182), even though there were numerical discrepancies in the original data sets, and the patients waited on average 10 years before MTX therapy was initiated. Moreover, in a study designed to test the usefulness and cost savings of the ACR guidelines for monitoring MTX hepatotoxicity (195), 112 MTX-treated rheumatoid patients were followed prospectively using the strict guidelines recommended for psoriatics on MTX (as shown in Table 2). The results of applying ACR guidelines retrospectively were then examined using the same data. Under ACR guidelines (which advocate more frequent hepatic panel testing and less frequent liver biopsy), 15 instead of 66 patients would have undergone biopsy, on 18 instead of 110 occasions, respectively. Biopsy complications occurred in two patients, neither of whom would have been biopsied under ACR guidelines. Of the five patients who had Roenigk grade IIIb or IV liver histology, four would have been biopsied under ACR guidelines. This included two patients whose pre-MTX laboratory results did not indicate a biopsy, but who developed frequent AST elevations on MTX and would therefore have been biopsied on therapy. One patient with obesity and poorly controlled insulin-dependent diabetes had persistently normal AST and albumin values, yet the baseline biopsy showed grade IIIa changes that progressed to cirrhosis during therapy. Neither of these two biopsies was mandated by the ACR protocol. The authors estimated that application of ACR instead of psoriasis guidelines avoided 92 biopsies in 51 patients (including two complications) and saved almost $100,000. They further reasoned that adding poorly controlled diabetes to the indications for pretreatment liver biopsy would have given the ACR guidelines 100% sensitivity. A study that has so far appeared only in abstract form (196) does not endorse the ACR guidelines. On the other hand, a decision-analysis study (197) that compared the monitoring strategy of “no biopsy versus biopsy,” after 5 and 10 years of MTX therapy in rheumatoid arthritis, concluded that biopsy was not cost-effective after 5 years and even 10 years of therapy, since the costeffectiveness ratios were $1.9 million and $52,000 per year of life saved, respectively. Aside from hepatic panel monitoring, and here ALT is as effective as AST (146), noninvasive monitoring appears to be no better than in MTX-treated psoriasis. Studies on fibrogenesis markers in MTX-treated rheumatoid arthritis patients are limited. PIIIP is elevated in untreated rheumatoid arthritis and normalizes with MTX therapy (198), but it is not known whether PIIIP rises again in this setting when hepatic fibrosis starts. In a study (156) whose objectives were to determine quantitative liver function prospectively and to assess the relationship between such testing and liver histology, neither galactose elimination, aminopyrine breath tests, liver enzymes, g-glutamyl transpeptidase (transferase), serum bile acids, bilirubin, nor albumin was of any practical use. In particular, no relationship was found between changes in results of galactose elimination, aminopyrine breath tests, and MTX dose, age, enzyme elevation, alcohol intake, or liver histology. Ideal monitoring guidelines cannot be fully evidence-based until a large prospective study is done in which patients are stratified according to the indication for MTX therapy (e.g., psoriasis, psoriatic arthritis, rheumatoid arthritis, etc.), and clinical variables, such as obesity, diabetes, alcohol intake, and age; laboratory variables, such as liver enzymes, albumin, serological tests for fibrogenesis, fibrogenic and inflammatory cytokine levels in the serum; and modern liver imaging, such as transient elastography (199), are correlated with liver histology. As yet, liver biopsy is still the most reliable means of diagnosing fibrosis in MTX-treated patients and will be the standard until better noninvasive testing [such as automated assays of multiple serum markers of liver fibrosis (117)] are available and shown to be helpful. It remains

694

Reuben

to be seen whether more sensitive histological measures of fibrosis, such as the semiquantitative scoring system that was recently tested in MTX-treated rheumatoid arthritis patients (66), will enhance understanding of the natural history of MTX hepatotoxicity and its management, as has been the case with the grading and staging systems now used regularly for the assessment of liver histology in chronic hepatitis. Other Rheumatic and Inflammatory Conditions. Data are limited on hepatotoxicity in other conditions for which MTX is prescribed, mostly because the number of patients being reported is small and large prospective series are not available for analysis. Of these conditions, juvenile rheumatoid arthritis is a common indication for MTX. In two series of patients who had liver biopsy monitoring, only 2 of 63 discontinued therapy because of fibrosis (200,201), whereas in two other series combined none of 21 patients developed fibrosis or cirrhosis with cumulative doses of up to 3 g (202,203). In a later study (204), only modest liver histological changes (Roenigk grade I or II) were seen in 13 of 14 patients who also had modest enzyme abnormalities (which exceeded threefold elevation in 5 patients). Irrespective, none of the patients developed any significant fibrosis and no significant clinical consequences were apparent, despite doses that were either greater than 3 or 4 g/1.73 m2 body surface area. The addition of folinic acid (2.5–7.5 mg daily) to the regimen for children on low-dose weekly MTX (10–20 mg/m2) who had already experienced aminotransferase elevations and gastrointestinal symptoms, dramatically ameliorated both hepatotoxicity and gastrointestinal toxicity without affecting the clinical efficacy of MTX (205). In a recent study of the relationship between hepatotoxic risk factors and liver histology in 25 patients with juvenile rheumatoid arthritis (206), only two patients (6%) had grade IIIa liver changes at worst, and the only risk factors that correlated with liver histology were the frequency of serum biochemical abnormalities and the degree of obesity (body mass index). Neither age, gender, disease duration, arthritis subtype, course, duration of MTX treatment, cumulative dose, route of administration, concurrent use of other medications (including folic acid), nor other potential hepatotoxicity played any role in liver injury. The safety of MTX was further substantiated in 34 patients from Helsinki (207), as even those who received relatively high doses (in excess of 20 mg/m2) did not develop fibrosis. Thus, the hepatotoxic effect of MTX in juvenile rheumatoid arthritis appears to be even less severe than that seen in adult rheumatoid arthritis. It turns out that when the 1994 ACR guidelines are used to follow these children on MTX therapy, a frequency of blood testing every four to eight weeks is unnecessarily frequent for the yield of abnormality and three monthly testing is more suitable (208). Studies on the hepatotoxicity of MTX therapy of other disorders such as dermatomyositis (209), sarcoidosis, and asthma are either largely anecdotal or retrospective (210). When hepatic fibrosis occurs it is usually mild and may often be ascribed to comorbid states, such as diabetes (209). Wider experience is awaited. In sarcoidosis, liver involvement with the primary disease process is common, and when MTX has been used in both children and adults, liver toxicity had previously rarely been reported. In one study, however, in which liver biopsies were done in 33 of 50 patients, six had to discontinue therapy for hepatic reactions that were considered to be due to MTX (211). Since hepatic involvement in sarcoidosis is common and gives rise to elevated aminotransferases and alkaline phosphatase, liver biopsy is required to diagnose MTX hepatotoxicity that may be as high as 14% (212). MTX appears to benefit idiopathic granulomatous hepatitis, including loss of granulomas (213). The same is true for inflammatory bowel disease (20), especially Crohn’s disease (214). Yet, MTX is used by a minority of gastroenterologists in some countries, and when they do, liver biopsies are done frequently and probably often unnecessarily (215). Overall, however, MTX appears to be well tolerated in many inflammatory (20) and connective tissue disorders, in a manner comparable to that seen in rheumatoid arthritis. Hepatotoxic Drug Reactions. These reactions are the same in MTX-treated rheumatic disorders as they are in MTX-treated psoriasis, except that there is a higher likelihood that patients with rheumatic disorders use analgesics, especially NSAIDs. Thus, if there is any additive or synergistic effect between MTX and NSAIDs, it should be seen in this group of patients. Some authors have reported such interactions (170) but, in general, it has been difficult to

Methotrexate Controversies

695

implicate concurrent NSAID use as a significant cofactor in MTX hepatotoxicity. Concurrent use of a wide variety of NSAIDs does not appear to affect MTX pharmacokinetics (including the area under the curve following ingestion, total systemic clearance, distribution volume, or the half-life of MTX) but does lead to an increased interpatient variability of MTX blood levels, which may not be clinically important (216). Concomitant use of sulfasalazine and cyclosporin does not appear to enhance MTX hepatotoxicity (217,218) although there may be bone marrow, skin rash, and renal interactions. Notwithstanding, given the growing trend of using disease-modifying agents in combination, especially in rheumatoid arthritis (219,220), we must be on the lookout for more drug hepatotoxic interactions. For example, there is a high incidence of hepatotoxicity from isoniazid when used in rheumatoid arthritis patients who are taking MTX (221). The demonstration that insulin augments MTX polyglutamate synthesis in human tumor cell lines (222) has not been correlated with any clinical interaction between insulin and MTX. It is intriguing to speculate that with insulin-resistant states (such as obesity and type II diabetes), in which insulin levels are high, or during insulin administration, there may be another mechanism for enhancing MTX toxicity. Insulin apparently also suppresses g-glutamyl hydrolase, the enzyme that degrades polyglutamates, and this would enhance intracellular MTX levels (223). OVERVIEW OF MTX HEPATOTOXICITY AND MONITORING MTX use, which is so effective in treating many difficult inflammatory disorders, does have the propensity to cause liver fibrosis and even cirrhosis. The risk of MTX-induced cirrhosis has been assuaged greatly over the past two decades by dosing patients weekly rather than daily and monitoring them carefully with repeated liver biopsy or frequent simple laboratory testing. Until proved otherwise, it appears that rheumatology patients may be less susceptible to MTXinduced hepatic fibrosis than psoriasis patients. The guidelines for MTX monitoring of rheumatoid arthritis patients should be applied prospectively, in a study, to MTX-treated psoriatics, controlling for dose and risk factors such as alcohol. Such a study could show that psoriatics react no differently than rheumatoid patients to MTX. Until truly reliable markers of fibrogenesis in MTX-treated patients are available, a liver biopsy should be done when there is any doubt about the safety of MTX therapy. Guidelines are just that, guidelines, and are not infallible. Physicians must exercise common “clinical” sense in the care of their patients. Thus, patients on MTX therapy should be followed carefully, because when this is neglected, avoidable decompensated cirrhosis can develop leading to death or the need for liver transplantation (52). Physicians should be vigilant about restricting alcohol use in MTX-treated patients and should also look for other causes of liver disease (such as hepatitis C and, the increasingly prevalent, nonalcoholic steatohepatitis) as well as hidden causes of liver injury such as occupational and environmental exposure to nondrug hepatotoxins, at work and at home (224). In contrast to the monitoring of psoriasis patients, MTX therapy of uncommon conditions such as juvenile rheumatoid arthritis, other inflammatory skin conditions, Crohn’s disease, and sarcoidosis can probably be monitored safely using the currently recommended guidelines for rheumatoid arthritis, unless the patient is known or suspected to have underlying liver disease. In the United Kingdom, monitoring of Crohn’s disease treated with MTX is recommended to be closer to that used for psoriatics than for patients with rheumatoid arthritis (225). This is especially important now that MTX is recommended more frequently for Crohn’s disease (226). Although the data are not yet conclusive, it should be feasible to adopt a monitoring scheme that can be customized to the patient’s needs, until better evidence-based guidelines become available. MTX THERAPY OF LIVER DISEASE The controversy over the use of MTX to treat PBC is a classic in the genre of debates in clinical science. Acknowledged experts in the field, seasoned investigators, take diametrically opposed positions that they champion with fervor bordering on passion, using plausible data that somehow do not jibe with the evidence presented by the opposition. In this, as with other

696

Reuben

similar debates in hepatology, it seemed likely for a while that the truth lay between the two views (227). Primary Sclerosing Cholangitis The stimulus for using MTX to treat cholestatic liver disease was a serendipitous observation. When MTX was used to treat a patient with a life-threatening skin disorder, the coexistent primary sclerosing cholangitis (PSC) appeared to improve and not deteriorate as feared. This unexpected, but encouraging experience was repeated in several other similar patients (228) and in an open-labeled study (229) of at least one year of MTX therapy in 10 patients with earlystage PSC. All symptomatic patients became asymptomatic, liver enzymes improved (but not bilirubin), and six of nine patients who had repeat biopsies after one year of MTX had improvement in necroinflammatory liver histology. Even repeat cholangiograms in six patients either improved (2) or stabilized (4). Unfortunately, a follow-up double-blind, placebocontrolled randomized trial of MTX in 24 patients with PSC (of whom 50% already had cirrhosis) did not show efficacy (230); neither did a pilot study in five patients in which MTX was combined with ursodeoxycholic acid (UDCA) treatment (231). In the latter study, there were frequent extrahepatic complications of MTX, and a case report also documented lifethreatening Pneumocystis carinii pneumonia in a PSC patient on MTX therapy (232). MTX is not a treatment option for PSC. Primary Biliary Cirrhosis MTX was evaluated initially in nine symptomatic patients with PBC (233), some but not all of whom showed slow improvement in symptoms, liver enzymes, and liver histology. Transient aminotransferase elevations occurred that seemed to predict a favorable response to MTX. The positive responses to MTX were seen primarily in those patients who had precirrhotic PBC, and this was confirmed later in five more patients who experienced remission of symptoms, biochemical amelioration, and histological improvement (234). A pilot study conducted at the National Institutes of Health in nine PBC patients (235) also showed improvement in symptoms, alkaline phosphatase elevation, and histological inflammatory activity, but patients with more advanced disease did not benefit and fibrosis progressed. Fibrosis progression, despite improvement in inflammation, was also reported by Bach et al. (236) in a primarily histopathological study of MTX-treated PBC patients. The early positive results of MTX therapy in PBC have been criticized for the relatively short-term and uncontrolled nature of the studies in a disease that is notoriously slow to progress and sometimes shows periods of stability (22). An interim (24-month) analysis of a randomized double-blind trial comparing MTX with colchicine in 83 PBC patients showed greater symptomatic, biochemical, and possible histological improvement in the MTX-treated patients than those on colchicine (237). In contrast, the long-term results of a completed placebo-controlled trial did not show any efficacy for MTX in PBC (238). In that study (238), there was even a trend toward a threefold increase in the rate of death or liver transplantation as a result of liver disease during or after the trial in MTX-treated patients compared to placebotreated controls (in a Cox multivariate regression analysis). In contrast to an earlier report of severe interstitial pneumonitis complicating MTX therapy in 14% of patients (239), the MTXtreated patients in the trial by Hendrickse et al. (238) had few side effects (which were readily reversible). The results of a smaller trial in precirrhotic PBC patients in Argentina showed MTX to be ineffective in preventing progression to cirrhosis despite symptomatic improvement and a biochemical response (240). The lack of benefit of adding MTX to UDCA treatment of PBC in Chile (241) and in large study in the United States (242) contrasts with the observation in Boston that MTX improves liver biochemical tests, and some histology too, in PBC patients who respond incompletely to UDCA (243). How can we explain the differences in outcomes among these many studies and others (244,245)? While arguments go back and forth over which dose of MTX is best, how likely are intolerable side effects to occur, how long is a reasonable followup period, and what constitutes a “favorable response” (21–28), it is possible that MTX may be beneficial in some precirrhotic PBC patients (but probably not in cirrhotics), that side effects

Methotrexate Controversies

697

cannot be ignored (242) and may be risky, and that some patients who do not respond to other, more benign therapy, such as UDCA and colchicine, may experience slowing of liver disease progression with MTX. A favorable answer did not come in the largest controlled trial to date (245) of MTX for those patients who do not respond to UDCA, when the endpoints were not merely symptomatic relief (albeit an important benefit for the patient), biochemical results, and liver histology, but rates of death and/or transplantation. In fact, this large trial was stopped early because of futility. Most authorities in the field do not recommend MTX therapy (29,246). Whether MTX should be used to treat refractory autoimmune hepatitis, as recently reported (247,248), is yet another can of worms waiting to be opened. REFERENCES 1. Gubner R. Effect of aminopterin on epithelial tissues. Arch Dermatol Syphilol 1951; 64:688–99. 2. Nair MG, Baugh CM. Synthesis and biological evaluation of poly-gamma-glutamyl derivatives of methotrexate. Biochemistry 1973; 12:3923–7. 3. Chabner BA, Allegra CJ, Curt GA, Clendeninn NJ, Baram J. Polyglutamination of methotrexate: is methotrexate a pro-drug? J Clin Invest 1985; 76:907–12. 4. Barak AJ, Tuma DJ, Beckenhauer HC. Methotrexate toxicity. J Am Coll Nutr 1984; 3:93–6. 5. Bertino JR. The general pharmacology of methotrexate. In: Wilke WS, ed. Methotrexate Therapy in Rheumatic Disease. New York: Marcel Dekker, 1989:11–23. 6. Kremer JM. Toward a better understanding of methotrexate. Arthritis Rheum 2004; 50:1370–2. 7. Farber S, Diamond LK, Mercer RD, Sylvester RF, Wolff JA. Temporary remission in acute leukemia in children, produced by folic antagonist 4-amethopteroylglutamic acid (aminopterin). N Engl J Med 1948; 238:789–93. 8. Gubner R, August S, Ginsberg V. Therapeutic suppression of tissue reactivity. II. Effect of aminopterin in rheumatoid arthritis and psoriasis. Am J Med Sci 1951; 221:176–82. 9. O’Brien WM, Van Scott EJ, Black RL, Auerbach R, Eisen AZ, Bunim JJ. Clinical trial of amethopterin (methotrexate) in psoriatic and rheumatoid arthritis (preliminary report). Arthritis Rheum 1962; 5:312. 10. Black RL, O’Brien WM, Van Scott EJ, Auerbach R, Eisen AZ, Bunim JJ. Methotrexate therapy in psoriatic arthritis: double-blind study on 21 patients. JAMA 1964; 189:743–7. 11. Panush RS, Wallace DJ, Dorff EN, Engleman EP. Retraction of the suggestion to use the term “Reiter’s syndrome” sixty-five years later: The legacy of Reiter; a war criminal, should not be eponymic honor but rather condemnation. Arthritis Rheum 2007; 56:693–4. 12. Colsky J, Greenspan EM, Warren TN. Hepatic fibrosis in children with acute leukemia after therapy with folic acid antagonists. Arch Pathol 1955; 59:198–206. 13. Kwoh CK, Anderson LG, Greene JM, et al. Guidelines for the management of rheumatoid arthritis: 2002 update. Arthritis Rheum 2002; 46:328–46. 14. O’D ell JR. Treating rheumatoid arthritis early: a window of opportunity? Arthritis Rheum 2002; 46:283–5. 15. American College of Rheumatology Ad Hoc Committee on Clinical Guidelines. Guidelines for monitoring drug therapy in rheumatoid arthritis. Arthritis Rheum 1996; 39:723–31. 16. Roenigk HH, Jr., Auerbach R, Maibach HI, Weinstein GD. Methotrexate in psoriasis: revised guidelines. J Am Acad Dermatol 1988; 19:145–6. 17. Lewis J, Schiff E. Methotrexate-induced chronic liver injury: guidelines for detection and prevention. Am J Gastroenterol 1988; 88:1337–45. 18. Petrazzuoli M, Rothe MJ, Grin-Jorgensen C, Ramsey WH, Grant-Kels JM. Monitoring patients taking methotrexate for hepatotoxicity. J Am Acad Dermatol 1994; 31:969–77. 19. Boffa MJ. Methotrexate for psoriasis: current European practice. J Eur Acad Dermatol Venereol 2004; 19:196–202. 20. Te HS, Schiano TD, Kuan SF, Hanauer SB, Conjeevaram HS, Baker AL. Hepatic toxic effects of longterm methotrexate use in the treatment of inflammatory bowel disease. Am J Gastroenterol 2000; 95:3150–6. 21. Lindor KD. Primary biliary cirrhosis: questions and promises. Ann Intern Med 1997; 120:733–5 (editorial). 22. LaRusso N. Search for medical treatment for primary biliary cirrhosis. Lancet 1997; 351:1046 (letter). 23. Kaplan MM. Primary biliary cirrhosis. Lancet 1998; 351:216 (letter). 24. Dufour J-FJ, Kaplan MM. Methotrexate and liver disease. N Engl J Med 1996; 335:898–9 (letter). 25. Bonis PAL, Kaplan MM. Low-dose methotrexate in primary biliary cirrhosis. Gastroenterology 1999; 117:1510–1 (letter). 26. Gleeson D, Underwood JCE, Hendrickse MT, Giaffer MH. Low-dose methotrexate in primary biliary cirrhosis. Gastroenterology 1999; 118:1512 (letter).

698

Reuben

27. Angulo P, Dickson ER. Methotrexate in the treatment of primary biliary cirrhosis: the hype and the hope. Gastroenterology 1999; 117:492–5. 28. Angulo P, Dickson ER. Low-dose methotrexate in primary biliary cirrhosis. Gastroenterology 1999; 117:1512–3 (letter). 29. Gong Y, Gluud C. Methotrexate for primary biliary cirrhosis (Cochrane Review). Cochrane Database Syst Rev 2005. Issue 3. Art No.: CD004385; DOI: 10.1002/14651858.CD004385. 30. Berkowitz RS, Goldstein DP, Bernstein MR. Ten years’ experience with methotrexate and folinic acid as primary therapy for gestational trophoblastic disease. Gynecol Oncol 1986; 23:111–8. 31. Hersh EM, Wong VG, Henderson ES, Freireich EJ. Hepatotoxic effects of methotrexate. Cancer 1996; 19:600–6. 32. Jaffe N, Traggis D. Toxicity of high-dose methotrexate (NSC-740) and citrovorum factor (NSC-3590) in osteogenic sarcoma. Cancer Chemother Rep 1975; 6:31–6. 33. Taft LI. Methotrexate-induced hepatitis in childhood leukemia. Isr J Med Sci 1965; 1:823. 34. Banerjee AK, Lakhani S, Vincent M, Selby P. Dose-dependent acute hepatitis associated with administration of high dose methotrexate. Hum Toxicol 1988; 7:561–2. 35. Chan H, Evans WE, Pratt CB. Recovery from toxicity with high-dose methotrexate: prognostic factors. Cancer Treat Rep 1977; 61:797–804. 36. Weber BL, Tanyer G, Poplack DG, et al. Transient acute hepatotoxicity of high-dose methotrexate therapy during childhood. NCI Monogr 1987; 5:207–12. 37. Schmiegelow K, Pulczynska M. Prognostic significance of hepatotoxicity during maintenance chemotherapy of childhood acute leukemia. Br J Cancer 1990; 61:767–72. 38. Clegg DO, Furst DE, Tolman KG, Payne R. Acute, reversible hepatic failure associated with methotrexate treatment of rheumatoid arthritis. J Rheumatol 1989; 16:1123–6. 39. Hakim NS, Kobienia B, Benedetti E, Bloomer J, Payne WD. Methotrexate-induced hepatic necrosis requiring liver transplantation in a patient with rheumatoid arthritis. Int Surg 1998; 83:224–5. 40. Hutter RVP, Shipkey FH, Tan CTC, Murphy ML, Chowdhury M. Hepatic fibrosis in children with acute leukaemia: a complication of therapy. Cancer 1960; 13:288–307. 41. Mclntosh S, Davidson DL, O’Brien RT, Pearson HA. Methotrexate hepatotoxicity in children with leukemia. J Pediatr 1977; 90:1019–21. 42. O’Rourk RA, Eckert GE. Methotrexate-induced hepatic injury in an adult: a case report. Arch Intern Med 1964; 113:191–4. 43. Coe RO, Bull FE. Cirrhosis associated with methotrexate treatment of psoriasis. JAMA 1968; 206:1515–20. 44. Dahl MGC, Gregory MM, Scheuer PJ. Methotrexate in psoriasis—comparison of different dose regimens. Br Med J 1972; 1:654–6. 45. Podurgiel BJ, McGill DB, Ludwig J, Taylor WF, Muller SA. Liver injury associated with methotrexate therapy for psoriasis. Mayo Clin Proc 1973; 48:787–92. 46. Nyfors A, Poulsen H. Liver biopsy from psoriatics related to methotrexate therapy. II. Findings before and after methotrexate therapy in 88 patients: a blind study. Acta Pathol Microbiol Scand Sect A Pathol 1976; 84:262–70. 47. Nyfors A, Poulsen H. Liver biopsy from psoriatics related to methotrexate therapy. T. Findings in 123 consecutive non-methotrexate-treated patients. Acta Pathol Microbial Scand 1976; 84:253–61. 48. Lanse SB, Arnold GL, Goweens JD, Kaplan MM. Low incidence of hepatotoxicity associated with long-term, low-dose oral methotrexate in treatment of refractory psoriasis, psoriatic arthritis and rheumatoid arthritis: an acceptable risk/benefit-ratio. Dig Dis Sci 1985; 30:104–9. 49. Rau R, Karger T, Herbom G, Frenzel H. Liver biopsy findings in patients with rheumatoid anthritis undergoing longterm treatment with methotrexate. J Rheumatol 1989; 16:489–93. 50. West SG. Methotrexate hepatotoxicity. Rheum Dis Clin North Am 1997; 23:883–915. 51. Zachariae H, Kragbelle JT, Sogaard H. Methotrexate induced liver cirrhosis. Studies including serial liver biopsies during continued treatment. Br J Dermatol 1980; 102:407–12. 52. Gilbert SE, Klintmalm G, Mentor A, Silverman A. Methotrexate-induced cirrhosis requiring liver transplantation in three patients with psoriasis; a word of caution in light of the expanding use of this “steroid” sparing agent. Arch Intern Med 1990; 150:889–91. 53. Nyfors A, Hopwood D. Liver ultrastructure in psoriatics related to methotrexate therapy. I. A prospective study of findings in hepatocytes from 24 patients before and after methotrexate treatment. Acta Pathol Microbiol Scand Sect A Pathol 1977; 85:787–800. 54. Horvath E, Kovacs K, Ross RC, Saibil F, Kerenyi NA. Desmosomal abnormalities in the liver of methotrexate-treated psoriatics. Experientia 1977; 33:1202–4. 55. Hopwood D, Nyfors A. Liver ultrastructure in psoriatics related to methotrexate therapy. II. Findings in bile ducts from 11 methotrexate treated psoriatics and 2 controls. Acta Pathol Microbiol Scand Sect A Pathol 1977; 85:801–11. 56. Horvath E, Saibil FG, Kovacs K, Kerenyi NA, Ross RC. Fine ultrastructural changes in the liver of methotrexate-treated psoriatics. Digestion 1978; 17:488–502.

Methotrexate Controversies

699

57. Bjorkman DJ, Hammon EH, Lee RG, Clegg DO, Tolman KG. Hepatic ultrastructure after methotrexate therapy for rheumatoid arthritis. Arthritis Rheum 1988; 31:1465–72. 58. Albanis E, Friedman SL. Hepatic fibrosis. Clin Liver Dis 2001; 5:315–34. 59. Nyfors A, Poulsen H. Morphogenesis of fibrosis and cirrhosis in methotrexate-treated patients with cirrhosis. Am J Surg Pathol 1977; 1:235–43. 60. Hytiroglou P, Tobias H, Saxena R, Abramidou M, Papadimitrian CS, Theise ND. The canals of Hering might represent a target of methotrexate hepatic toxicity. Am J Clin Pathol 2004; 121:324–9. 61. Kevat S, Ahern M, Hall P. Hepatolotoxicity of methotrexate in rheumatic diseases. Med Toxicol 1988; 3:197–208. 62. Roenigk HH, Jr., Auerbach R, Maibach H, Weinstein G, Lebwohl M. Methotrexate in psoriasis: consensus conference. J Am Acad Dermatol 1998; 38:478–85. 63. Brunt EM. Grading and staging the histopathology of chronic hepatitis: the Knodell histology activity index and beyond. Hepatology 2000; 31:241–6. 64. Hall PD, Ahern MJ, Jarvis LR, Stoll P, Jenner MA, Harley H. Two methods of assessment of methotrexate hepatotoxicity in patients with rheumatoid arthritis. Ann Rheum Dis 1991; 50:471–6. 65. Fathi NH, Mitros F, Hoffman J, et al. Longitudinal measurement of methotrexate liver concentrations does not correlate with liver damage, clinical efficacy, or toxicity during a 3.5 year double blind study in rheumatoid arthritis. J Rheumatol 2002; 29:2092–8. 66. Richard S, Guerret S, Gerard F, Tebib JG, Vignon E. Hepatic fibrosis in rheumatoid arthritis patients treated with methotrexate: application of a new semi-quantitative scoring system. Rheumatology 2000; 39:50–4. 67. Nohlgard C, Rubio CA, Kocky Hammar H. Liver fibrosis quantified by image analysis in methotrexate-treated patients with psoriasis. J Am Acad Dermatol 1993; 28:40–5. 68. Bremnes RM, Smeland E, Huseby NE, Eide JJ, Aarbakke J. Acute hepatotoxicity after high dose methotrexate administration to rats. Pharmacol Toxicol 1991; 69:132–9. 69. Jaskewicz K, Voigt H, Blakolmer K. Increased matrix proteins, collagen and transforming growth factor are early markers of hepatotoxicity in patients on long-term methotrexate therapy. J Toxicol Clin Toxicol 1996; 34:301–5. 70. Al-Ali SY, Hassan IM, Sadek S. Ultrastructural changes in rat livers perfused in vitro and in vivo with a high dose of methotrexate. Histol Histopathol 2005; 20:1131–45. 71. Hall PD, Jenner MA, Ahem MJ. Hepatotoxicity in a rat model caused by orally administrated methotrexate. Hepatology 1991; 14:906–10. 72. Freeman-Narrod M. Choline antagonism of methotrexate liver toxicity in the rat. Med Pediatr Oncol 1997; 3:9–14. 73. Yamauchi PS, Rizk D, Kormeili T, Patnaik R, Lowe NJ. Current systemic therapies for psoriasis: Where are we now? J Am Acad Dermatol 2003; 49:S66–77. 74. Warin AP, Landells JW, Levene GM, Baker H. A prospective study of the effects of weekly oral methotrexate on liver biopsy: findings in severe psoriasis. Br J Dermatol 1975; 93:321–7. 75. Boffa MJ, Chalmers RJ, Haboubi NY, Shomaf M, Mitchell DM. Sequential liver biopsies during longterm methotrexate treatment for psoriasis: a reappraisal. Br J Dermatol 1995; 133:774–8. 76. Ashton RE, Millward-Sadler GH, White JE. Complications in methrotrexate treatment of psoriasis with particular reference to liver fibrosis. J Invest Dermatol 1982; 79:229–32. 77. Robinson JIT, Baughman RD, Auerbach R, Cimis RJ. Methotrexate hepatotoxicity in psoriasis: consideration of liver biopsies at regular intervals. Arch Dermatol 1980; 116:413–5. 78. Newman M, Auerbach R, Feiner H, et al. The role of liver biopsies in psoriatic patients receiving long-term methotrexate treatment improvement in the abnormalities after cessation of treatment. Arch Dermatol 1989; 125:1218–24. 79. Themido R, Loureiro M, Pecegueiro M, Brandao M, Campos MC. Methotrexate hepatotoxicity in psoriatic patients submitted to long-term therapy. Acta Derm Venereol 1992; 72:361–4. 80. Whiting-O’Keefe QE, Fye KM, Sack KD. Methotrexate and histologic hepatic abnormalities: a metaanalysis. Am J Med 1991; 90:711–6. 81. Malatjalian DA, Ross JB, Williams CN, Colwell SJ, Eastwood RJ. Methotrexate hepatotoxicity in psoriasis: report of 104 patients from Nova Scotia with analysis of risks from obesity, diabetes and alcohol consumption during long term follow-up. Can J Gastroenterol 1996; 10:369–95. 82. Zachariae H. Have methotrexate-induced liver fibrosis and cirrhosis become rare? Dermatology 2005; 211:307–8. 83. Roenigk HH, Jr., Bergfeld WF, St Jaques R, Owens FJ, Hawk WA. Hepatotoxicity of methotrexate. Arch Dermatol 1971; 103:250–61. 84. Weinstein GD, Cox JW, Suringa DWR, Millard MM, Kalser MF. Evolution of possible chronic hepatotoxicity from methotrexate for psoriasis. Arch Dermatol 1970; 102:613–8. 85. Reese LT, Grisham TW, Aach RD, Elsen AZ. Effects of methotrexate on the liver in psoriasis. J Invest Dermatol 1974; 62:597–602. 86. Weinstein G, Roenigk H, Maibach H, et al. Psoriasis-liver-methotrexate interactions. Arch Dermatol 1973; 108:36-42.

700

Reuben

87. O’Connor GT, Olmstead EM, Zug K, et al. Detection of hepatotoxicity associated with methotrexate therapy for psoriasis. Arch Dermatol 1989; 125:1209–17. 88. Van De Kerkhof PC, Hoefnagels WH, Van Haelst UJ, Mali JW. Methotrexate maintenance therapy and liver damage in psoriasis. Clin Exp Dermatol 1955; 10:194–200. 89. Pestana A, Halprin KM, Taylor JR, et al. Predictive value of HLA antigens for methotrexate-induced liver damage in patients with psoriasis. J Am Acad Dermatol 1985; 12:26–9. 90. Langman G, Hall PM, Todd G. Role of non-alcoholic steatohepatitis in methotrexate-induced liver injury. J Gastroenterol Hepatol 2001; 16:1395–401. 91. Mathew J, Leong MY, Morely N, Burt AD. A liver fibrosis cocktail? psoriasis, methotrexate and genetic hemochromatosis BMC Dermatol 2005; 5:12. 92. Nyfors N. Liver biopsies from psoriatics related to methotrexate therapy: 3. Findings in postmethotrexate liver biopsies from psoriatics. Acta Pathol Microbiol Scand 1977; 85:511–8. 93. Aithal GP, Haugk B, Das S, Card T, Burt AD, Record CO. Monitoring methotrexate-induced hepatic fibrosis in patients with psoriasis: are serial liver biopsies justified? Aliment Pharmacol Ther 2004; 19:391–9. 94. Thomas JA, Aithal GP. Monitoring liver function during methotrexate therapy for psoriasis. Are routine biopsies really necessary? Am J Clin Dermatol 2005; 6:357–63. 95. Tobias H, Auerbach R. Hepatotoxicity of long-term methotrexate therapy for psoriasis. Arch Intern Med 1973; 132:391–6. 96. Almeyda J, Barnado D, Baker H, Levene GM, Landells JW. Structural and functional abnormalities of the liver in psoriasis before and during methotrexate therapy. Br J Dermatol 1972; 87:625–31. 97. Haufstein U-F, Rytter M. Methotrexate in psoriasis: 26 years’ experience with low-dose long-term treatment. J Eur Acad Derm Venereol 2000; 14:382–8. 98. Zachariae H, Sogaard H, Heickendorff L. Methotrexate-induced liver cirrhosis: clinical, histological and serological studies—a further 10-year follow-up. Dermatology 1996; 192:343–6. 99. Zachariae H. Liver biopsies and methorexate: a time for reconsideration? J Am Acad Dermatol 2000; 42:531–4. 100. Cuellar ML, Espinoza LR. Methotrexate use in psoriasis and psoriatic arthritis. Rheum Dis Clin North Am 1997; 23:797–809. 101. van Doreen-Greebe RJ, Kuijpers AL, Termorshuizen F, van der Kerkhof PC. Interruption of longterm methotrexate treatment in psoriasis. Evaluation of clinical course and laboratory parameters after discontinuation and reintroduction of weekly oral methotrexate. Acta Derm Venereol 1995; 75:393–6. 102. Saporito FC, Menter MA. Methotrexate and psoriasis in the era of new biologic agents. J Am Acad Dermatol 2004; 50:301–9. 103. Roenigk H, Maibach H, Weinstein G. Guidelines on methotrexate therapy for psoriasis. Arch Dermatol 1972; 105:363–5. 104. British Association of Dermatologists. Guidelines for management of patients with psoriasis. Br Med J 1991; 303:829–35. 105. Maurice PD, Maddox AJ, Green CA, Tatnall F, Schofield JK, Stott DJ. Monitoring patients on methotrexate: hepatic fibrosis not seen in patients with normal serum assays of aminoterminal peptide of type III procollagen. Br J Dermatol 2005; 152:451–8. 106. Chalmers RJ, Kirby B, Smith A, et al. Replacement of routine liver biopsies by procollagen III aminopeptide for monitoring patients with psoriasis receiving long-term methotrexate: a multicentre audit and health economic analysis. Br J Dermatol 2005; 152:444–50. 107. Reynolds FS, Lee WM. Hepatotoxicity after long-term methotrexate therapy. South Med J 1986; 79:536–9. 108. Lawrence CM, Srange RC, Summerly RA, et al. Assessment of liver function using fasting bile salt concentrations in psoriasis prior to and during methotrexate therapy. Clin Chim Acta 1983; 129:341–51. 109. Williams CN, McCauley D, Malatjalian DO, Turnbull GK, Ross JB. The aminopyrine breath test, an inadequate early indicator of methotrexate-induced liver disease in patients with psoriasis. Clin Invest Med 1987; 10:54–8. 110. Lenler-Petersen P, Sogard H, Thestrup-Pederson H, Zarchariae H. Galactose tolerance test and methorexate-induced liver fibrosis and cirrhosis in patients with psoriasis. Acta Derm Venereol (Stockh) 1982; 62:448–550. 111. Palmer HM. Hepatic toxicity of methotrexate treatment of psoriasis. Practitioner 1973; 211:324–8. 112. Wu J, Danielsson A. Detection of hepatic fibrogenesis: a review of available techniques. Scand J Gastroenterol 1995; 30:817–25. 113. Hayasaka A, Saisho H. Serum markers as tools to monitor liver fibrosis. Digestion 1998; 59:381–4. 114. Lu L-G, Zeng M-D, Mao Y-M, Li J-Q, Qiu D-K, Fang J-Y, Cao A-P, Wan M-B, Li C-Z, Ye J, et al. Relationship between clinical and pathologic findings in patients with chornic liver diseases. World J Gastroenterol 2003; 9:2796–800.

Methotrexate Controversies

701

115. Wai CT, Greenson JK, Fontana RJ, et al. A simple noninvasive index can predict both significant fibrosis and cirrhosis in patients with chronic hepatitis C. Hepatology 2003; 38:518–26. 116. Afdhal NH, Nunes D. Evaluation of liver fibrosis; a concise review. Am J Gastroenterol 2004; 99:1160–72. 117. Rosenberg WMC, Voelker M, Thiel R, et al. Serum markers detect the presence of liver fibrosis: a cohort study. Gastroenterology 2004; 127:1704–13. 118. Naveau S, Raynard B, Ratziu V, et al. Biomarkers for the prediction of liver fibrosis in patients with chronic alcoholic liver disease. Clin Gastroenterol Hepatol 2005; 3:167–74. 119. Kelleher TB, Afdhal N. Noninvasive assessment of liver fibrosis. Clin Liver Dis 2005; 9:667–83. 120. Rockey DC, Bissell M. Noninvasive measures of liver fibrosis. Hepatology 2006; 43:S113–20. 121. Bissell DM. Assessing fibrosis without a liver biopsy: are we there yet? Gastroenterology 2004; 127:1847–9. 122. Afdhal N. Can serum markers be used to reliably detect liver fibrosis? Nat Clin Pract Gastroenterol Hepatol 2005; 2:132–3. 123. Mardini H, Record C. Detection assessment and monitoring of hepatic fibrosis: biochemistry or biopsy? Ann Clin Biochem 2005; 42:441–7. 124. Sebastiani G, Alberti A. Non invasive fibrosis biomarkers reduce but not substitute the need for liver biopsy. World J Gastroenterol 2006; 12:3682–94. 125. Forns X, Ampurdanes S, Llovet JM, et al. Indentification of chronic hepatitis C patients without hepatic fibrosis by a simple predictive model. Hepatology 2002; 36:982–6. 126. Kaul V, Friedenberg FK, Braitman LE, et al. Development and validation of a model to diagnose cirrhosis in patients with hepatitis C. Am J Gastroenterol 2002; 97:2623–8. 127. Islam S, Antonsson L, Westin J, Lagging M. Cirrhosis in hepatitis C virus-infected patients can be excluded using an index of standard biochemical serum markers. Scand J Gastroenterol 2005; 40:867–72. 128. Zachariae H, Aslam HM, Bjerring P, Sogaard H, Zachariae E, Heickendorff L. Serum aminoerminal propeptide of type III precollagen in psoriasis and psoriatic arthritis: relation to liver fibrosis and arthritis. J Am Acad Dermatol 1991; 25:50–3. 129. Boffa MJ, Smith A, Chalmers RJG, et al. Serum type III procollagen aminopeptide for assessing liver damage in methotrexate-treated psoriasis patients. Br J Dermatol 1996; 135:538–44. 130. Mitchell D, Smith A, Rowan B, et al. Serum type III precollagen peptide, dynamic liver function tests and hepatic fibrosis in psoriatic patients receiving methotrexate. Br J Dermatol 1990; 122:1–7. 131. Zachariae H, Heickendorff L, Sogaard H. The value of amino-terminal propeptide of type III collagen in routine screening for methotrexate-induced liver fibrosis: a 10-year follow-up. Br J Dermatol 2001; 144:100–3. 132. Oogarah PK, Rowland PC, Mitchell DM, et al. Abnormalities of serum type III procollagen aminoterminal peptide in methotrexate-treated psoriatic patients with normal liver histology do not correlate with hepatic ultrastructural changes. Br J Dermatol 1995; 133:512–8. 133. Geronemus RG, Auerbach R, Tobias H. Liver biopsies versus liver scans in methotrexate-treated patients with psoriasis. Arch Dermatol 1982; 118:649–51. 134. The Clinical NMR Group. Magnetic resonance imaging of parenchymal liver disease: a comparison with ultrasound, radionuclide scintigraphy and x-ray computer tomography. Clin Radiol 1987; 38:495-502. 135. Rademaker M, Webb JA, Lowe DG, Meynck-Thomas RH, Kirby JD, Munro DD. Magnetic resonance imaging as a screening procedure for methorexate-induced liver damage. Br J Dermatol 1987; 117:311–6. 136. Arias JM, Morton KA, Albro TE, et al. Comparison of methods for identifying early methotrexateinduced hepatotoxicity in patients with rheumatoid arthrisis. J Nucl Med 1993; 34:1905–9. 137. Lupsor M, Badea R. Imaging diagnosis and quantification of hepatic steatosis: is it an accepted alternative to biopsy? Rom J Gastroenterol 2005; 14:419–25. 138. Aube C, Winkfield B, Oberti F, et al. New Doppler ultrasound signs improve the non-invasive diagnosis of cirrhosis or severe liver fibrosis. Eur J Gastroenterol Hepatol 2004; 16:743–51. 139. Haktanir A, Cihan BS, Celenk C, Cihan S. Value of Doppler sonography in assessing the progression of chronic viral hepatitis and in the diagnosis and grading of cirrhosis. J Ultrasound Med 2005; 24:311–21. 140. McHerney PM, Bingham EA, Callender ME, et al. Dynamic hepatic scintigraphy in the screening of patients for methotrexate-induced hepatotoxicity. Br J Dermatol 1992; 127:122–5. 141. Van Dooren-Greebe R, Kuijpers A, Buijs W, et al. The value of dynamic hepatic scintigraphy and serum aminoterminal propeptide of type III procollagen for early detection of methorexate-induced hepatic damage in psoriasis patients. Br J Dermatol 1996; 134:481–7. 142. Ferraz MB, Ciconelli RM, Vilar MJ. Patient’s preference regarding the option of performing unselective liver biopsy following methotrexate treatment in rheumatoid arthritis. Clin Exp Rheumatol 1994; 12:621–5.

702

Reuben

143. Khan S, Subedi D, Chowdhury MMU. Use of amino terminal type III procollagen peptide (P3MP) assay in methotrexate therapy for psoriasis. Postgrad Med J 2006; 82:353–4. 144. Evans WE, Christensen ML. Drug interactions with methotrexate. J Rheumatol 1985; 12(Suppl. 12):15–20. 145. Furst DE, Kremer JM. Methotrexate in rheumatoid arthritis. Arthritis Rheum 1988; 81:305–14. 146. Kremer JM, Alarcon GS, Lightfott RW, Jr., et al. Methotrexate for rheumatoid arthritis: suggested guidelines for monitoring liver toxicity. Arthritis Rheum 1994; 37:316–28. 147. Chevalier M, Guerret S, Chossegros P, Gerard F, Grimaud JA. A histological semiquantitative scoring system for evaluation of hepatic fibrosis in needle biopsy specimens: comparison with morphometric studies. Hepatology 1994; 20:349–55. 148. Brunt EM, Janney CG, Di Bisceglie AM. Neuschwander-Tetri BA, Bacon R. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 1999; 94:2467–74. 149. Brick JE, Morland LW, Al-Kawas F, Chang WWL, Layne RD, DiBartolomeo AG. Prospective analysis of liver biopsies before and after methotrexate therapy in rheumatoid arthritis patients. Semin Arthritis Rheum 1989; 19:31–44. 150. Chandran G, Ahern MJ, Pall PD, et al. Cirrhosis in patients with rheumatoid arthritis receiving lowdose methotrexate. Br J Rheumatol 1994; 33:981–4. 151. Kremer JM, Lee RG, Tolman KG. Liver histology in rheumatoid arthritis patients receiving long-term methotrexate therapy: a prospective study with baseline and sequential biopsy samples. Arthritis Rheum 1989; 31:305–14. 152. Aponte J, Petrelli M. Histopathologic findings in the liver of rheumatoid arthritis patients treated with long-term bolus methotrexate. Arthritis Rheum 1988; 12:1457–64. 153. Shergy WJ, Polisson RP, Caldwell DS, Rice JR, Pisetsky DS, Allen NB. Methotrexate-associated hepatotoxicity: retrospective analysis of 210 patients with rheumatoid arthritis. Am J Med 1988; 85:771–4. 154. Minocha A, Dean HA, Pittsley RA. Liver cirrhosis in rheumatoid arthritis patients with long-term methotrexate. Vet Hum Toxicol 1993; 35:45–8. 155. Walker AM, Funch D, Dreyer NA, et al. Determinants of serious liver disease among patients receiving low-dose methotrexate for rheumatoid arthritis. Arthritis Rheum 1993; 36:329–35. 156. Beyeler C, Richen J, Thomann SR, Lautcrberg BM, Gerber NJ. Quantitative liver function in patients with rheumatoid arthritis treated with low-dose methotrexate: a longitudinal study. Br J Rheumatol 1997; 36:338–44. 157. Singh G, Fries JF, Williams CA, Zatarain E, Spitz P, Bloch DA. Toxicity profiles of disease modifying antirheumatic drugs in rheumatoid arthritis. J Rheumatol 1991; 18:188–94. 158. Borchers AT, Keen CL, Cheema GS, Gershwin ME. The use of methotrexate in rheumatoid arthritis. Semin Arthritis Rheum 2004; 34:465–83. 159. Kinder AJ, Hassell AB, Brand J, Brownfield A, Grove M, Shadforth MF. The treatment of inflammatory arthritis with methotrexate in clinical practice: treatment duration and incidence of adverse drug reactions. Rheumatology 2005; 44:61–6. 160. Choi HK, Herna¯n MA, Seeger JD, Robins JM, Wolfe F. Methotrexate and mortality in patients with rheumatoid arthritis: a prospective study. Lancet 2002; 359:1173–7. 161. Leonard PA, Clegg DO, Carson CC, Canon GW, Egger MJ, Ward JR. Low-dose pulse methotrexate in rheumatoid arthritis: an 8-year experience with hepatotoxicity. Clin Rheumatol 1987; 6:575–82. 162. Yazici Y, Erkan D, Harrison MJ, Nikolov NP, Paget SA. Methotrexate use in rheumatoid arthritis is associated with few clinically significant liver function test abnormalities. Clin Exp Rheumatol 2005; 23:517–20. 163. Yazici Y, Sokka T, Kautiainen H, Swearingen C, Kulman I, Pincus T. Long term safety of methotrexate in routine clinical care: discontinuation is unusual and rarely the result of laboratory abnormalities. Ann Rheum Dis 2005; 64:207–11. 164. Morgan SL, Baggott JE, Vaughn WH, et al. Supplementation with folic acid during methotrexate therapy for rheumatoid arthritis: a double-blind placebo-controlled trial. Ann Intern Med 1994; 121:833–41. 165. Ortiz Z, Shea B, Suarez Almazor M, Moher D, Wells G, Tugwell P. Folic acid and folinic acid for reducing side effects in patients receiving methotrexate for rheumatoid arthritis. Cochrane Database Syst Rev 1999, DOI:10.1002/14651858, CD000951. 166. Whittle SL, Hughes RA. Folate supplementation and methotrexate treatment in rheumatoid arthritis: a review. Rheumatology 2004; 43:267–71. 167. Hoekstra M, van der Laar MA, Bernelot Moens HJ, Kruijsen MW, Haagsma CJ. Longterm observational study of methotrexate use in a Dutch cohort of 1022 patients with rheumatoid arthritis. J Rheumatol 2003; 30:2325–9. 168. Fries JF, Singh A, Lenert L, Furst DE. Aspirin, hydroxychloroquine, and hepatic enzyme abnormalities with methotrexate in rheumatoid arthritis. Arthritis Rheum 1990; 33:1611–9.

Methotrexate Controversies

703

169. Seideman P, Albertioni F, Beck O, Esksborg S, Peterson C. Chloroquine reduces the bioavailability of methotrexate in patients with rheumatoid arthritis. A possible mechanism of reduced hepatotoxicity. Arthritis Rheum 1994; 37:830–3. 170. Rabinovitz M, Van Thiel DH. Hepatotoxicity of nonsteroidal anti-inflammatory drugs. Am J Gastroenterol 1992; 87:1696–704. 171. Ros S, Juanola X, Condom E, et al. Light and electron microscopic analysis of liver biopsy samples from rheumatoid arthritis patients receiving long-term methotrexate therapy. Scand J Rheumatol 2002; 31:330–6. 172. Hoekstra M, van Ede AE, Haagsma CJ, et al. Factors associated with toxicity, final dose, and efficacy of methotrexate in patients with rheumatoid arthritis. Ann Rheum Dis 2003; 62:423–6. 173. Kent PD, Luthratis Michet CJ, Jr. Risk factors for methotrexate-induced abnormal laboratory monitoring results in patients with rheumatoid arthritis. J Rheumatol 2004; 31:1727–31. 174. Fathi NJ, Mitros F, Hoffman J, et al. Longitudinal measurement of methotrexate liver concentrations does not correlate with liver damage, clinical efficacy, or toxicity during a 3.5 year double blind study in rheumatoid arthritis. J Rheumatol 2002; 29:2092–8. 175. Berkun Y, Levartovsky D, Rubinow A, et al. Methotrexate related adverse effects in patients with rheumatoid arthritis are associated with the A1298C polymorphism of the MTHFR gene. Ann Rhematol Dis 2004; 63:1227–31. 176. Weisman MH, Furst DE, Park GS, et al. Risk genotypes in folate-dependent enzymes and their association with methotrexate-related side effects in rheumatoid arthritis. Arthritis Rheum 2006; 54:607–12. 177. van Ede AE, Laan RFJM, De Abreu RA, Stegeman ABJ, van de Putte LBA. Purine enzymes in patients with rhematoid arthritis treated with methotrexate. Ann Rheum Dis 2002; 61:1060–4. 178. Ranganathan P, Eisen S, Yokoyama WM, McLeod HL. Will pharmacogenetics allow better prediction of methotrexate toxicity and efficacy in patients with rheumatoid arthritis? Ann Rheum Dis 2003; 62:4–9. 179. Kremer JM. Liver toxicity does not have to follow methotrexate therapy of patients with rheumatoid arthritis. Am J Gastroenterol 1997; 92:184–96. 180. Kremer JM, Kaye GI. Electron microscopic analysis of sequential liver biopsies from patients with rheumatoid arthritis: correlation with light microscopic findings. Arthritis Rheum 1992; 32:1202–13. 181. Kremer JM, Kaye GI, Kaye NW, Ishak KG, Axiotis CA. Light and electron microscopic analysis of sequential liver biopsy samples from patients with rheumatoid arthritis receiving long-term therapy: follow-up over prolonged treatment intervals and correlation with clinical and laboratory variables. Arthritis Rheum 1995; 38:1194–203. 182. Kremer JM, Furst DE, Weinblatt ME, Blotner SD. Significant changes in AST across hepatic histological biopsy grades: prospective analysis of 3 cohorts receiving methotrexate therapy for rheumatoid arthritis. J Rheumatol 1996; 23:459–61. 183. Gabriel S, Cregsen E, O’Fallen WM, Jaquith J, Bunch T. Treatment of rheumatoid arthritis with higher dose intravenous methotrexate. J Rheumatol 1990; 17:460–5. 184. Schnabel A, Reinhold-Keller E, Willmann V, Gross WL. Tolerability of methotrexate starting with 15 or 25 mg/week for rheumatoid arthritis. Rheumatol Int 1994; 14:33–8. 185. Furst DE, Koehnke R, Burmeister LF, Kohler J, Cargill I. Increasing methotrexate effect with increasing dose in the treatment of resistant rheumatoid arthritis. J Rheumatol 1989; 16:313–20. 186. Thompson RN, Watts C, Edelman J, Esdaile J, Russell AS. A controlled two-center trial of parenteral methotrexate therapy for refractory rheumatoid arthritis. J Rheumatol 1984; 11:760–3. 187. Weinblatt ME, Weissman BN, Holdsworth DE, et al. Long-term prospective study of methotrexate in the treatment of rheumatoid arthritis. Arthritis Rheum 1992; 35:129–37. 188. Weinstein A, Marlowe S, Korn J, Farouhar F. Low-dose methotrexate treatment of rheumatoid arthritis: long-term observations. Am J Med 1985; 79:331–7. 189. Health and Public Policy Committee, American College of Physicians. Methotrexate in rheumatoid arthritis. Ann Intern Med 1987; 107:418–9. 190. Levin DM. Re: JM Kremer—editorial. Am J Gastroenterol 1997; 92:1402–3. 191. Yazici Y, Erkan D, Paget SA. Monitoring methotrexate hepatic toxicity in rheumatoid arthritis: is it time to update the guidelines? J Rheumatol 2002; 29:1586–9. 192. Yazici Y, Erkan D, Paget SA. Monitoring by rheumatologists for methotrexate-, etamercep-, infliximab-, and anakinra-associated adverse events. Arthritis Rheum 2003; 48:2769–72. 193. Kremer JM. Response to Dr. Levin. Am J Gastroenterol 1997; 92:1402–3 (letter). 194. Kremer JM. Not yet time to change the guidelines for monitoring methotrexate liver toxicity: they have served us well. J Rheumatol 2002; 29:1590–2. 195. Erickson AR, Reddy V, Vogelgesang SA, West SG. Usefulness of the American College of Rheumatology recommendations for liver biopsy in methotrexate-treated rheumatoid arthritis. Arthritis Rheum 1995; 38:1115–9. 196. Jackson CG, Sawitzke A, Clegg DO. Monitoring methotrexate toxicity in rheumatoid arthritis: failure of ACR guidelines to predict hepatic fibrosis. Arthritis Rheum 1994; 37:S361 (abstract).

704

Reuben

197. Berquist SR, Felson DT, Prashker MJ, Freedberg KA. The cost-effectiveness of liver biopsy in rheumatoid arthritis patients treated with methotrexate. Arthritis Rheum 1995; 38:326–33. 198. Biasi D, Bambara LM, Carletto A, et al. Effects on fibrogenesis markers of rheumatoid arthritis therapy with methoirexate. J Rheumatol 1996; 23:453–4. 199. Gomez-Dominquez E, Mendoza J, Rubio S, Moreno-Monteagudo JA, Garcia-Buey L, MorenoOtero R. Transient elastography: a valid alternative to biopsy in patients with chronic liver disease. Aliment Pharmacol Ther 2006; 24:513–8. 200. Danao T, Steinbrunner J, Medendorp SV. Methotrexate in juvenile rheumatoid arthritis. Arthritis Rheum 1989; 32(Suppl. 9):28. 201. Keim D, Ragsdale C, Heidelberger K, Sullivan D. Hepatic fibrosis with the use of methotrexate for juvenile rheumatoid arthritis. J Rheumatol 1990; 17:846–8. 202. Graham LD, Myones BL, Rivas-Chacon RF, Pachman LM. Morbidity associated with long-term methotrexate therapy in juvenile rheumatoid arthritis. J Pediatr 1992; 120:468–73. 203. Kugathasan S, Newman AJ. Liver biopsy findings in patients with juvenile rheumatoid arthritis receiving long-term weekly methotrexate therapy. J Pediatr 1996; 128:149–51. 204. Hashkes PJ, Balistreri WF, Bove KE, Ballard ET, Passo MH. The long-term effect of methotrexate therapy on the liver in patients with juvenile rheumatoid arthritis. Arthritis Rheum 1997; 40:2226–30. 205. Ravelli A, Migliavacca D, Viola S, Ruperto N, Pistorio A, Martini A. Efficacy of folinic acid in reducing methotrexate toxicity in juvenile idiopathic arthritis. Clin Exp Rheumatol 1999; 17:625–7. 206. Hashkes PJ, Balistreri WF, Bove KE, Ballard ET, Passo MH. The relationship of hepatotoxic risk factors and liver histology in methotrexate therapy for juvenile rheumatoid arthritis. J Pediatr 1999; 134:47–52. 207. Lahdenne P, Rapola J, Ylijoki H, Haapasaari J. Hepatotoxicity in patients with juvenile rheumatoid arthritis receiving long-term methotrexate therapy. J Rheumatol 2002; 29:2442–5. 208. Ortiz-Alvarez O, Morishita K, Avery G, et al. Guidelines for blood test monitoring of methotrexate toxicity in juvenile idiopathic arthritis. J Rheumatol 2004; 31:2501–6. 209. Zieglschmid-Adams ME, Pandya AG, Cohen SB, Sontheimer RD. Treatment of dermatomyositis with methotrexate. J Am Acad Dermatol 1995; 32:754–7. 210. Davies H, Olson L, Gibson P. Methotrexate as a steroid-sparing agent for asthma in adults. Cochrane Database Syst Rev 1998, DOI: 10.1002/14651858, CD000391. 211. Lower EE, Baughman RP. Prolonged use of methotrexate for sarcoidosis. Arch Intern Med 1995; 155:846–51. 212. Baughman RP, Koehler A, Bejarano PA, Lower EE, Weber FL, Jr. Role of liver function tests in detecting methotrexate-induced liver damage in sarcoidosis. Arch Intern Med 2003; 163:615–20. 213. Knox TA, Kaplan MM, Gelfand JA, Wolff SM. Methotrexate treatment of granulomatous hepatitis. Ann Intern Med 1995; 122:592–5. 214. Schro¨der O, Stein J. Low dose methotrexate in inflammatory bowel disease: current status and future directions. Am J Gastroenterol 2003; 98:530–7. 215. Chande N, Ponich T, Gregor J. A survey of Canadian gastroenterologists about the use of methotrexate in patients with Crohn’s disease. Can J Gastroenterol 2005; 19:553–8. 216. Iqbal MP, Baig JA, Ali AA, Niazi SK, Mehboobali N, Hussain MA. The effects of non-steroidal antiinflammatory drugs on the disposition of methotrexate in patients with rheumatoid arthritis. Biopharm Drug Dispos 1998; 19:163–7. 217. Nisar M, Carlisle L, Amos RS. Mcthotrexate and sulphasalzine as combination therapy in rheumatoid arthritis. Br J Rheumatol 1994; 33:651–4. 218. Stein CM, Pincus T, Yocum D, et al. Combination treatment of severe rheumatoid arthritis with cyclosporine and methotrexate for forty-eight weeks: an open-label extension study: the Methotrexale-Cyclosporine Combination Study Group. Arthritis Rheum 1997; 40:1843–51. 219. Kremer JM. Rational use of new and existing disease-modifying agents in rheumatoid arthritis. Ann Intern Med 2001; 134:695–706. 220. Silva MA, Ishii-Iwamoto EL, Bracht A, et al. Efficacy of combined methotrexate chloroquine therapy in adjuvant-induced arthritis. Fundam Clin Pharmacol 2005; 19:479–89. 221. Vanhoof J, Landewe S, van Wijngaerden E, Geusens P. High incidence of hepatotoxicity of isoniazide treatment for tuberculosis chemoprophyalxis in patients with rheumatoid arthritis treated with methotrexate or sulphasalazine and anti-tumor necrosis factor inhibitors. Ann Rheum 2003; 62:1241–2. 222. Schilsky RL, Ordway FS. Insulin effects on methotrexate polyglutamale synthesis and enzyme binding in cultured human breast cancer cells. Cancer Chemother Pharmacol 1985; 15:272–7. 223. Galivan J, Rees MS. Insulin-dependent suppression in glutamyl hydrolase activity and elevated cellular methotrexate polyglutamates. Biochem Pharmacol 1995; 50:1659–63. 224. Redlich CA, Cullen MR. Nonalcoholic steatohepatitis. Ann Intern Med 1997; 127:410 (letter). 225. Rampton DS. Therapy update: methotrexate in Crohn’s disease. Gut 2001; 48:790–1.

Methotrexate Controversies

705

226. Dome`nech E. Inflammatory bowel disease: current therapeutic options. Digestion 2006; 73(Suppl. 1):67–76. 227. Tang H, Neuberger J. Methotrexate in gastroenterology—dangerous villain or simply misunderstood? Aliment Pharmacol Ther 1996; 10:851–8. 228. Kaplan MM, Arora S, Pincus SH. Primary sclerosing cholangitis and low dose oral pulse methotrexate. Ann Intern Med 1987; 106:231–5. 229. Knox TA, Kaplan MM. Treatment of primary sclerosing cholangitis with oral methotrexate. Am J Gastroenterol 1991; 86:546–52. 230. Knox TA, Kaplan MM. A double-blind controlled trial of oral-pulse methotrexate in the treatment of primary sclerosing cholangitis. Gastroenterology 1994; 106:494–9. 231. Lindor KD, Jorgensen RA, Anderson ML, Gores GJ, Hofmann AF, La Russo NF. Ursodeoxycholic acid and methotrexate for primary sclerosing cholangitis: a pilot study. Am J Gastroenterol 1996; 91:511–5. 232. Duerkson DR, Blondell-Hill E, Bailey RJ. Pneumocystis carinii pneumonia complicating methotrexate treatment of primary sclerosing cholangitis. Am J Gastroenterol 1995; 90:1886–7. 233. Kaplan MM, Knox TA. Treatment of primary biliary cirrhosis with low-dose weekly methotrexate. Gastroenterology 1991; 101:1332–8. 234. Kaplan MM, De Lellis RA, Wolfe HJ. Sustained biochemical and histologic remission of primary biliary cirrhosis in response to medical treatment. Ann Intern Med 1997; 126:682–8. 235. Bergasa NV, Jones EA, Kleiner DE, et al. Pilot study of low-dose oral methotrexate treatment for primary biliary cirrhosis. Am J Gastroenterol 1996; 91:295–9. 236. Bach N, Thung SN, Schafner F. The histologic effect of low-dose methotrexate therapy for primary biliary cirrhosis. Arch Pathol Lab Med 1998; 122:342–5. 237. Kaplan MM, Schmid C, Provenzale D, Sharma A, Dickstein G, McKusick A. A prospective trial of colchicine and methotrexate in the treatment of primary biliary cirrhosis. Gastroenterology 1999; 117:1173–80. 238. Hendrickse MT, Rigney E, Giaffer MH, et al. Low-dose methotrexate is ineffective in primary biliary cirrhosis: long-term results of a placebo-controlled trial. Gastroenterology 1999; 117:400–7. 239. Sharma A, Provenzale D, McKusick A, Kaplan MM. Interstitial pneumonitis after low-dose methotrexate therapy in primary biliary cirrhosis. Gastroenterology 1994; 107:266–70. 240. Sorda J, Findor J. Methotrexate therapy in primary biliary cirrhosis. Acta Gastroenterol Latinoam 2000; 30:221–5. 241. Gonzalez-Koch A, Brahm J, Antezana C, Smok G, Cumsille MA. The combination of ursodeoxycholic acid and methotrexate for primary biliary cirrhosis is not better then ursodeoxycholic acid alone. J Hepatol 1997; 27:143–9. 242. Bach N, Bodian C, Bodenheimer H, et al. Methotrexate for primary biliary cirrhosis. Am J Gastroenterol 2003; 98:187–93. 243. Bonis PAL, Kaplan MM. Methotrexate improves biochemical tests in patients with primary biliary cirrhosis who respond incompletely to ursodiol. Gastroenterology 1999; 117:395–9. 244. Lindor KD, Dickson ER, Jorgensen RA. The combination of ursodeoxycholic acid and methotrexate for patients with primary biliary cirrhosis: the results of a pilot study. Hepatology 1995; 22:1158–62. 245. Combes B, Emerson SS, Flye NL, Munoz SJ, Luketic VA, Mayo MJ, McCashland TM, Zetterman RK, Peters MG, Di Bisceglie AM, et al. Methotrexate (MTX) plus ursodeoxycholic acid (UDCA) in the treatment of primary biliary cirrhosis. Hepatology 2005; 42:1184–93. 246. Oh YH, Neuberger J. Options for treatment of primary biliary cirrhosis. Drugs 2004; 64:2261–71. 247. Burak KW, Urbanski SJ, Swain MG. Successful treatment of refractory type 1 autoimmune hepatitis with methotrexate. J Hepatol 1998; 29:990–3. 248. Venkataramani A, Jones MB, Sorrell MF. Methotrexate therapy for refractory chronic active autoimmune hepatitis. Am J Gastroenterol 2001; 96:3432–4.

32

Adverse Effects of Hormones and Hormone Antagonists on the Liver Shivakumar Chitturi and Geoffrey C. Farrell

Department of Gastroenterology and Hepatology, Australian National University Medical School at the Canberra Hospital, Australian Capital Territory, Australia

INTRODUCTION Hepatotoxicity is a recognized adverse effect of treatment with several hormonal agents and also with their corresponding antagonists. In this chapter, we first discuss the liver toxicity associated with oral contraceptive steroids (OCS) (Table 1) and anabolic androgenic steroids (AAS). Liver disorders with the OCS and AAS encompass a range of histologic appearances including acute hepatitis, cholestasis, hepatic vascular toxicity and liver tumors, both benign and malignant. Although there are broad similarities between the OCS and AAS with regard to the liver disorders they can induce, there are important differences between the groups. For example, cholestasis is a feature common to both groups, but hepatic and portal vein thrombosis have only been associated with OCS. Conversely, peliosis hepatis, and hepatocellular carcinoma (HCC) have stronger associations with AAS than with OCS. The second section of this chapter addresses the growing recognition that tamoxifen hepatotoxicity can be associated with fatty liver disease, and the role of metabolic factors in this association. While the medium-term outcome of tamoxifen-associated hepatotoxicity seems benign and allows for interrupted treatment, long-term safety data are critical because tamoxifen remains a valuable drug in breast cancer. Likewise, androgen blockade, a potent tool in prostate cancer management, brings with it the risk of infrequent but significant liver injury. Deaths from acute liver failure have been ascribed to both flutamide and cyproterone. The final section of this chapter explores the hepatic adverse effects of currently used antithyroid drugs. The liver toxicity associated with the use of antidiabetic agents is a large important topic that is discussed separately in chapter 28.

ORAL CONTRACEPTIVE STEROIDS Cholestasis OCS are characteristically associated with intrahepatic cholestasis. However, by influencing gallbladder motility and by altering the lithogenicity of bile, they promote gallstone formation and potentially, extrahepatic biliary obstruction (1,2). Thus, in all cases of presumed OCS induced cholestasis, extrahepatic cholestasis, should be excluded by appropriate hepatobiliary imaging. The frequency of developing intrahepatic cholestasis with OCS varies with ethnic group, from 3 to 4 per 100,000 women (3) in Western Europe, up to 1 per 4000 in Chile (4) and Scandinavia (5). Individuals with a personal or family history of intrahepatic cholestasis of pregnancy or OCS-induced cholestasis are also at high risk (up to 50%) (6,7). The estrogenic component of OCS has been implicated in inducing cholestasis; estrogen impairs the function of pumps involved in bile secretion, and such effects comprise the prevailing view on the pathogenesis of this reaction (8). (See also Chap. 6 for mechanisms of drug-induced cholestasis). However, progestin-only contraceptives or high-dose progestin regimes used in breast cancer may also rarely produce cholestasis (9–11).

708

Chitturi and Farrell

TABLE 1 The Spectrum of Oral Contraceptive Steroid-Induced Liver Injury Liver disorder Cholestasis Vascular lesions Peliosis hepatis Portal vein thrombosis Hepatic vein thrombosis Liver tumors Cavernous hemangioma Hepatic adenoma Hepatocellular carcinoma Other liver disorders Focal nodular hyperplasia

Predisposing factors Frequent in certain ethnic groups and in families with a history of OCS-associated cholestasis and/or intrahepatic cholestasis of pregnancy Not identified. Other causes of peliosis should be ruled out Presence of hereditary or other acquired thrombogenic risk factors (see text) Presence of hereditary or other acquired thrombogenic risk factors (see text) Probably do not influence development or natural history (see text) Duration of use (O5 years), dose (low risk with low-dose estrogens) Increased relative risk with duration of use and dose (O5 years, high dose) Probably do not influence their development or natural history (see text)

Abbreviation: OCS, oral contraceptive steroids.

In clinical practice, liver biopsy is not usually performed for suspected OCS-induced cholestasis as the disorder settles following discontinuation of OCS. The histological appearances are of a bland cholestasis, with insignificant hepatic lobular injury or inflammation (12). The clinical features of OCS-induced cholestasis often start with mild prodromal symptoms such as anorexia and nausea, two to three months after starting OCS (rarely as late as nine months) (13,14). This is followed by pruritus. As expected for cholestasis, serum alkaline phosphatase is moderately elevated but serum aminotransferases (AT) can show transient elevation, occasionally over 10-fold above the upper limit of normal in the early stages. Jaundice is usually mild, with serum bilirubin levels usually below 10 mg/dL. An unusual biochemical feature is that serum gamma-glutamyl transpeptidase levels are often normal, thereby resembling benign recurrent intrahepatic cholestasis and differing from most other causes of jaundice. Cholestasis generally settles within days to a few weeks after drug discontinuation. Prolonged cholestasis is infrequent (13). Vascular Disorders of the Liver Peliosis Hepatis The hallmark of peliosis hepatis is multiple blood-filled cystic cavities in the hepatic parenchyma. Oral contraceptives are among a long list of drugs and disorders associated with this vascular lesion (15,16); other causes include tuberculosis, human immunodeficiency virus infection, lymphoma, malnutrition, Bartonella infection, and drugs including anabolic– androgenic steroids, 6-thioguanine, and azathioprine (17). Toxic injury to sinusoidal endothelial cells may be central to the development of peliosis. Peliosis hepatis can be detected incidentally at hepatic imaging or laparoscopy in asymptomatic individuals but intraperitoneal hemorrhage can occur. Cases with hepatomegaly, portal hypertension, and liver failure have also been reported (18). Treatment options include withdrawal of OCS and/or surgical resection in patients with large lesions or in those presenting with intraperitoneal hemorrhage. Portal Vein and Hepatic Vein Thrombosis OCS users are more than twice as likely to develop hepatic vein thrombosis as compared with nonusers [RR, 2.37 (95% CI 1.05–5.34)] (19). Further, the risk of hepatic vein thrombosis approaches that of stroke and myocardial infarction in the OCS group. Others have also reproducibly documented the thrombogenic potential of OCS in relation to venous thromboses at various extrahepatic sites (calf veins, inferior vena cava, cerebral venous sinuses) (20) and within the hepatobiliary system (portal and hepatic veins) (21,22). However, these reports predate the emergence of hereditary thrombophilia as a major factor in the development of systemic venous thromboses. As an illustrative example, a young woman using OCS was

Effects of Hormones and Hormone Antagonists

709

admitted to a Dutch hospital with features of hepatic vein thrombosis (23). Tests for myeloproliferative disorders and paroxysmal nocturnal hemoglobinuria were negative. She was found to have a double heterozygous defect in the thrombophilia screen (Factor V Leiden and G20210A), a pattern also seen in her father who had never experienced thromboembolism. The authors suggest that OCS use may have unmasked a hitherto unknown thrombogenic tendency. This interaction of thrombogenic risk factors is also brought out well in a study of, 43 patients with Budd-Chiari syndrome (BCS) and 92 with portal vein thrombosis (PVT) (24). The coexistence of acquired risk factors (such as OCS, surgery) and inherited thrombophilia mutations were observed in 26% of patients with BCS and 37% of those with PVT (24). More than three thrombotic risk factors (acquired or congenital) were identified in 7% to 10% of these cases. Therefore, even in cases where OCS appear to be strongly incriminated as the only obvious thrombogenic risk factor, it is critical to exclude other factors including hereditary thrombophilia and latent or overt myeloproliferative disease. Benign Neoplasms of the Liver Hemangioma Unlike hepatic adenomas, the involvement of OCS in the development of liver hemangioma is not supported by clinical studies. Among 40 women with hepatic hemangiomas and 109 agematched controls with normal liver imaging (25) there were no differences found, and the frequencies of OCS usage between these groups (30% and 27%, respectively) were similar. However, there have been occasional case reports that use of OCS increases the size of preexisting liver hemangiomas, and similar findings have been reported during pregnancy (26). Further, hemangiomas have recurred in OCS users after previous resection (27). This suggests that an estrogen, induced trophic effect in some cases (28). However, the overall risk appears low and current contraceptive prescribing practice does not preclude use of OCS in patients with liver hemangiomas. Focal Nodular Hyperplasia This lesion is not neoplastic but is discussed here because it presents as a focal liver nodule that needs to be differentiated from a hepatic adenoma. Focal nodular hyperplasia (FNH) most likely represents a hyperplastic response to a local vascular abnormality. Unlike hepatic adenoma, most patients with FNH are asymptomatic. The threat of intralesional or intraperitoneal hemorrhage is minimal and there is no risk of malignant transformation. The demographic profile of individuals with FNH is similar to those presenting with hepatic adenoma. Most (up to 86%) occur in young women, during their childbearing years (29). It has been suggested that FNH may also be an estrogen-derived and estrogen-dependent liver tumor. However, this association appears tenuous on several grounds. First, the incidence of FNH has not increased in parallel with OCS use. Second, FNH can occur in persons not using OCS (30). The results of case-control studies exploring the relationship between OCS and FNH have been conflicting (30,31). In the first study, Scalori and colleagues compared the frequencies of oral contraceptive use between 23 patients with FNH and 94 hospital controls (31). Among the FNH group, 89% reported ever using OCS, including 22% who had used these agents for longer than three years. The corresponding figures for the control group were 53% and 9%, respectively. The odds ratios for the risk of FNH among OCS users and those using it for more than three years were 2.8; 95% CI (0.8–9.4) and 4.5; 95% CI (1.2–16.9), respectively. However, in a larger French study of over 200 patients with FNH, Mathieu and colleagues could not demonstrate any relationship between the pattern of OCS use and the size or number of FNH lesions (30). Discontinuation of OCS failed to influence the size of FNH, and estrogen-dependent tumors were rare. The authors suggest that previously reported “estrogen-driven” FNH lesions may have been misclassified under hepatic adenomas. In general, the absence of a tumor capsule, presence of bile duct proliferation, and a central scar favor the histologic diagnosis of FNH rather than hepatic adenoma. Thus, on the basis of current evidence, OCS can be used in patients with known

710

Chitturi and Farrell

FNH lesions and need not be discontinued if the diagnosis is well established by histologic or by hepatic magnetic resonance imaging. Hepatic Adenoma The reported association between OCS and hepatic adenomas, first by Baum and colleagues in 1973 (32), was initially met with skepticism. However, numerous publications have validated their conclusions. Observational studies confirm an increase in the frequency of hepatic adenoma since the introduction of OCS in 1960. Among 50,000 autopsies performed over a 36-year period before 1954, there were only two cases of hepatic adenoma (33). Epidemiologic studies in the 1970s found that annual risk of hepatic adenoma with OCS was approximately three to four per 100,000 exposed persons (33,34). The risk with low (estrogen) dose preparations appears to be much lower. There is also a relationship with duration of OCS use. Thus, the risk of hepatic adenoma (compared to nonusers) for long-term OCS users (O10 years) is increased 100-fold. Occasional cases of liver adenoma have been described with short duration OCS use (!24 months) (35). Liver adenomas may be identified as an incidental finding during hepatobiliary imaging for other indications (36). However, patients can present with right upper quadrant pain or occasionally with hemoperitoneum secondary to hepatic rupture. Except in the face of such complications, the initial approach is to stop OCS, and periodic liver imaging is then conducted to assess the desired regression in size (37). Surgical resection is reserved for cases when the diagnosis is uncertain and also in patients with larger lesions (O5 cm) to avert possible rupture (29) and to prevent the small but definite risk of malignant transformation. Routine surveillance with liver imaging is not indicated for women receiving OCS. While current trends are to use OCS with lower estrogenic potency (30–35 mg of estradiol), long uninterrupted periods of OCS use should still be avoided. Hepatocellular Carcinoma Several case-control studies have demonstrated an association between OCS use and HCC as recently summarized by Yu (38,39–46). These case series were documented from the U.S.A and Europe and therefore involved mainly white subjects, most of whom did not have liver cirrhosis. In a summary of eight such studies, the relative risk for HCC was 2.5 (95% CI 1.7– 3.5) among women who had ever taken OCS, and 5.8 (95% CI 3–11) in long-term users (O8 years) compared with age-matched controls (38). Even so, it must be emphasized that estrogen-related HCC is rare, accounting for less than 2% of primary liver cancer in western countries (47). Further, in Asia and Africa, where there is a high prevalence of HCC, OCS use did not appear to be an independent risk factor (48,49). However, these studies may have been inadequately powered to detect such an association because the sample size would have to be large, to take into account the contributions of chronic viral hepatitis and aflatoxin to HCC in these regions (50). It may be important to note that hormone replacement therapy with estrogen-containing products has not been shown to have an association with HCC (43,51,52). OCS-associated HCC usually appears about five years after commencement. The median age at presentation is around 30 years. These tumors are well-differentiated HCCs. In the shortto medium- term, the clinical outcome is better than with HCC due to other causes, but most cases are ultimately fatal. Other Liver Tumors Rare instances of other liver tumors such as epithelioid hemangioendothelioma (EHE) and angiosarcoma have been reported in persons using OCS (53–56). Because EHE occurs in women during their reproductive years and 17-b estradiol receptors have been observed in tumors at other sites (e.g. lung), it has been suggested that hormonal factors may be relevant to its pathogenesis. Some authors have speculated that histologic changes seen with long-term OCS use (hepatocyte proliferation, sinusoidal dilatation and peliosis hepatis) could represent the precursor lesions of hepatic angiosarcoma (56). However, the evidence linking OCS to

Effects of Hormones and Hormone Antagonists

711

hepatic angiosarcoma is less robust when placed in context with the effects of known carcinogens, such as arsenic, thorotrast, and vinyl chloride. ANABOLIC ANDROGENIC STEROIDS Cholestasis and Hepatitis Bland cholestasis is reported with high doses of AAS. It is usually a reversible reaction that occurs usually within one to six months of starting treatment. Cholestasis is best characterized with C17-alkylated AAS, but has also been described with non-C17 alkylated congeners (57). Other presentations include prolonged jaundice with ductopenia, (58–60), acute hepatitis with very high serum AT (over 10,000 IU/L) (61) and intraabdominal hemorrhage from rupture of a large intracapsular hepatic hematoma; in the latter case, the underlying hepatic parenchyma showed extensive necrosis (62). A rare presentation is biliary ductopenia with an inflammatory pseudotumor of the liver, mimicking liver malignancy (63). Vascular Lesions Peliosis hepatis is well recognized with AAS. The clinical presentation is similar to that described with OCS (see above) (64,65). Neoplasms of the Liver Both benign (hepatic adenoma) and malignant liver tumors HCC have been reported with AAS (14,66,67). A high proportion of these patients (over 35 cases) had Fanconi’s anemia, a disorder with a high background incidence of neoplasms (68,69); the cumulative probability of liver tumors in Fanconi’s anemia has been estimated to be close to 50% by the age of 50. In light of this predisposition, the risk of liver tumor development in androgen treated patients was considered to be disease-specific. However, there are now over 100 instances of liver tumors (hepatic adenoma and HCC) in disorders other than Fanconi’s anemia, providing more convincing evidence of probable etiologic association (70,71). Most cases are in men. However, there is ample evidence that AAS-associated HCC is not gender-specific; thus, HCC has been observed among female body builders, women with systemic lupus erythematosus, idiopathic thrombocytopenic purpura and other medical disorders treated with danazol and similar agents (72). One review of androgen-associated hepatic neoplasms concluded that HCC was more frequently observed in recipients of oxymetholone or methyltestosterone (69), whereas hepatic adenomas were often observed with danazol (72). However, these apparent differences are more likely to represent general trends only, and both benign and malignant liver tumors have been described with all androgens. Other than for oxymetholone (no association found), there are insufficient data to conclude whether there exists a dose relationship to liver tumor development. Likewise, the general perception that liver tumors occur only with oral androgens has been challenged by reports in patients receiving parenteral androgens (73). Likewise, with the exception of patients with Fanconi’s anemia, there appears to be no clear association of liver tumor development with patient age or duration of androgen use. Patients with Fanconi’s anemia develop hypoplastic anemia at an early age and are therefore likely to have commenced AAS at a much earlier age. The natural history of these androgen-related liver tumors is different from typical HCC cases. Some metastasize very late, others show regression (67,74) and serum alfa fetoprotein levels are usually within the normal range. Liver histology often fails to resolve the true nature of these tumors, because seemingly well-differentiated HCCs are often indistinguishable from hepatic adenomas (75). Hepatic angiosarcoma has been associated with AAS use; evidence of causality is lacking (76). Liver adenomas can regress after discontinuation of androgens. However, surgical resection may be necessary in cases where there is incomplete resolution (77). Persons receiving long-term androgen therapy should undergo periodic hepatic imaging (66). Former users of anabolic androgen steroids should probably also undergo surveillance liver imaging because liver tumors have been described up to 24 years after discontinuation (78).

712

Chitturi and Farrell

ESTROGEN-RECEPTOR ANTAGONISTS Tamoxifen Tamoxifen, an antiestrogenic compound, is widely used as adjuvant therapy in breast cancer. There is a growing body of literature about its effects on the liver. Up to 43% of long-term users of tamoxifen exhibit hepatotoxicity. Non-alcoholic fatty liver disease (NAFLD) is the predominant form of liver injury associated with tamoxifen (Table 2) (79–89). The frequency of fatty liver disease exceeds other well-recognized adverse effects, such as endometrial hyperplasia (16%) and deep vein thrombosis (1–2%) (81). Rare instances of submassive hepatic necrosis have also been reported with tamoxifen (90,91). Many case series have relied heavily on abdominal ultrasound or computerized tomographic (CT) imaging for diagnosis of NAFLD. However, liver biopsies obtained from patients with radiologic features of severe hepatic steatosis, have shown mild to moderate steatohepatitis. Advanced hepatic fibrosis and even micronodular cirrhosis have also been described (79,88). Pathogenesis of Tamoxifen-Induced Steatosis/Steatohepatitis Insulin Resistance

In a prospective, randomized controlled study of tamoxifen versus placebo in over 5400 patients with breast cancer, fatty liver was observed predominantly in overweight and obese women (79); after five years, the cumulative incidence of NAFLD among obese, overweight, and normal weight women (defined by body mass index R30, 25–30 and %25 kg/m2, respectively), was 3.8%, 2.4%, and 0.7%, respectively (79). Similar conclusions were reached in an earlier study from Greece (80). An additional observation from this report was the higher frequency of impaired glucose intolerance (69%) among those with steatosis as compared to BMI matched controls without steatosis (24%) (80). Thus, the metabolic profile of patients with tamoxifen-associated NAFLD seems indistinguishable from that observed in individuals with “primary” NAFLD (92–94). The presence of general or visceral obesity and features of the insulin resistance syndrome (metabolic syndrome) are common themes in both groups of patients. It can therefore be concluded that hepatic steatosis/steatohepatitis is circumstantial (women taking tamoxifen are often overweight), or results from the synergistic relationship TABLE 2 Tamoxifen and Non-Alcoholic Fatty Liver Disease Method of diagnosis

Review of the Literature

Author

Year

Cases

Liver histology

Bruno

2005

52

Elfesiniotis

2004

60

Nishino

2003

67

Ultrasound; liver biopsy NASH (12), simple (13 cases) steatosis (1) Ultrasound; liver biopsy Steatosis (1), NASH (9 of 26 cases with steatosis) (6), cirrhosis (2) CT; steatosis in 29 cases (43%)

Nguyen Cai Dray X Oien

2001 2000 2000 1999

32 1 1 2

Abdominal CT Abdominal CT Liver biopsy Liver biopsy

Ogawa

1998

66

Pinto HC Van Hoof Pratt

1996 1996 1995

3 1 1

CT; Liver biopsy in 7 cases (lowest liver spleen attenuation) Liver biopsy Liver biopsy Liver biopsy

NASH with cirrhosis NASH with cirrhosis NASH in (6) of (7), simple steatosis (1) NASH (3) NASH/cirrhosis (1) NASH (1)

Abbreviations: ALT, serum alanine aminotransferase; NASH, non-alcoholic steatohepatitis.

Outcome after drug discontinuation Non progressive course ALT improved in 18 of 26 cases ALT improved; no clinical or radiological signs of progression Not reported ALT improved Not reported ALT improved in 1 case; fluctuating ALT in the 2nd patient ALT improved ALT improved Not reported Normal ALT

Effects of Hormones and Hormone Antagonists

713

of tamoxifen with other NAFLD risk factors. However, it has also been proposed that tamoxifen could induce hepatic steatosis by effects on mitochondrial structure and function. Mitochondrial Effects

In common with other drugs causing steatohepatitis such as perhexiline maleate, tamoxifen induces steatosis in mice by a mechanism that appears linked to impaired mitochondrial fatty acid beta-oxidation (b-oxidation) and respiratory chain function (95). Tamoxifen has also been shown to deplete mitochondrial DNA by inhibiting topoisomerases and mitochondrial DNA synthesis (96). In human studies, tamoxifen was shown to inhibit b-oxidation of BMIPP, an I123labelled fatty acid analoge, thereby reducing clearance of this agent (97). A mechanism by which tamoxifen decreases mitochondrial fatty acid beta oxidation appears to involve reduction of hepatic fatty acid synthase activity (98). The ensuing accumulation of malonylCoA impairs activity of carnitine palmitoyl transferase 1 (CPT-1), the step involved with fatty acid uptake into the mitochondrion, a prerequisite for fatty acid b-oxidation (98). However, others have been unable to demonstrate significant disruption of mitochondrial respiratory chain in mitchondria extracted from livers of rats treated with tamoxifen (98). Further studies are required to clarify whether tamoxifen, at the doses used to prevent breast cancer recurrence, can cause clinically significant mitochondrial toxicity in humans. Genetic Factors

Ohnishi et al., studied the frequencies of cytochrome P450 17 (CYP17) gene polymorphisms in 180 recipients of tamoxifen, including 57 with and 123 without hepatic steatosis (99). CYP17 gene encodes cytochrome CYP17a, a key enzyme in estrogen synthesis. Of the two alleles studied (A1, A2), these workers observed an increased frequency of A2 in the steatosis group [57% vs. 41%, respectively; odds ratio 1.9, 95% CI 1.9(21–2.99)]. The A2 allele is associated with increased transcriptional activity of CYP17a and consequently high estrogen levels. However, in animal studies, estrogen deficiency, not excess, is associated with hepatic steatosis (100). This is illustrated by studies with aromatase-deficient mice, which lack estrogen and develop hepatic steatosis as a consequence of impaired mitochondrial b-oxidation (100); estrogen repletion ameliorates hepatic steatosis in these animals. Ohnishi et al. also found no association between tamoxifen-induced steatosis and polymorphisms of several genes, including those involved in obesity (b-2 adrenergic receptor), lipid metabolism (microsomal triglyceride transfer protein), inflammation, and insulin resistance (tumor necrosis factor-a, peroxisome proliferator activated receptor-g, insulin receptor substrate-1) (99). The role of genetic factors in tamoxifen induced steatosis remains therefore unclear. Treatment and Outcome

The natural history of tamoxifen-related fatty liver disease has not been studied, but seven year follow up has not shown any clinical, biochemical, and radiologic deterioration in the liver disease (79). Discontinuation of tamoxifen normalizes the serum AT. Improvement has been documented in the abdominal CT liver–spleen attenuation coefficient, an indirect marker of hepatic steatosis (81). However, the outcome of patients with cirrhosis is uncertain, and they should be observed for signs of hepatic decompensation. The treatment of tamoxifen-related steatohepatitis has not been systematically evaluated. Nonpharmacologic strategies to improve insulin resistance, such as diet and exercise, should be considered, as they are first-line treatment in ‘primary’ non-alcoholic steatohepatitis (NASH). Preliminary experience with bezafibrate appears promising and may allow for continued use of tamoxifen (101). Bezafibrate is a peroxisome proliferator-activated receptor-a (PPAR-a) agonist and is presumed to work by PPAR-a driven pathways of fatty acid b-oxidation and triglyceride secretion from the liver. Improvement in liver imaging characteristics and AT was observed in two patients with tamoxifen-induced steatohepatitis (101). Interestingly, commensurate with reduction of hepatic steatosis (by abdominal CT), BMIPP clearance rates, a marker of peroxisomal b-oxidation, also improved after treatment with bezafibrate, (97). An alternative approach to use of bezafibrate is replacement of tamoxifen with toremifene, an antiestrogen with a lower frequency of hepatic steatosis.

714

Chitturi and Farrell

Surveillance

Patients receiving tamoxifen should be monitored for the development of hepatic steatosis by periodic physical examination for hepatomegaly and other signs of chronic liver disease, liver biochemistry, and hepatic imaging (ultrasonography or computed tomography). Particularly if liver test abnormalities fail to resolve after stopping tamoxifen, liver biopsy is required to establish the severity of the disorder as well as to exclude metastatic breast cancer in difficult cases. Toremifene Steatosis and steatohepatitis have also been reported with Toremifene, a tamoxifen analoge (102). However, hepatic effects with this agent appeared to be less frequent (4 of 52 cases, 7.7%) as compared with historical controls receiving tamoxifen (30%) (102). In this study, the diagnosis of hepatic steatosis was principally based on CT characteristics; liver histology, available for one case, showed steatohepatitis. Cyclofenil This nonsteroidal selective estrogen receptor modulator is used in Europe for suppressing lactation and treating anovulatory infertility. Acute hepatitis has been recorded with cyclofenil (103). Positive rechallenge was documented in one case. Most cases presented within three months. All affected individuals have recovered completely. ANTIANDROGENS Antiandrogens are mainly used in treating prostate cancer but are also being increasingly prescribed for acne, hirsutism, and hypersexual behavior disorders. They comprise two classes of drugs: steroidal and the nonsteroidal androgens. Cyproterone acetate is the principal steroidal antiandrogen in use while flutamide and its long acting congeners, nilutamide and bicalutamide, belong to the latter group. Liver injury is a well-recognized adverse effect of all antiandrogens (Table 3). Thus, among 79 cases of all published cases identified between 1986 and 2003, flutamide (46), cyproterone (21), nilutamide (4), and bicalutamide (1) were implicated (104,105). The frequency of liver injury with flutamide ranges from 0.18% to 5% and with cyproterone is less than 1% (104–106). Hepatic reactions occur in up to 30% of recipients of cyproterone (107). While hepatotoxicity due to co-prescribed drugs could not be excluded in several reports, there are TABLE 3 The Spectrum of Antiandrogen Hepatotoxicity Clinical presentation

Liver histology

Causality

Outcome

Cholestatic hepatitis; massive, submassive or zonal hepatic necrosis

Positive rechallenge. Only drug used in 15 of 46 reported cases

Most recover. Liver-related deaths in 22% (10 of 46 cases)

Nilutamide

Cholestatic hepatitis; acute hepatitis; acute liver failure Acute hepatitis

Submassive necrosis

Bicalutamide

Acute hepatitis

Not studied

2 deaths (of 4 cases) Recovered

Cyproterone acetate

Acute hepatitis

Used as monotherapy in 1 of 4 reported cases Causality disputed; previous flutamide not excluded Likely; no other hepatotoxic drugs in 14 of 21 reported cases Association less robust (see text)

Flutamide

Hepatocellular carcinoma Cirrhosis

Mortality 52% (11 of 21 cases)

Effects of Hormones and Hormone Antagonists

715

well-documented instances where antiandrogens were used alone and were probably implicated in causing liver injury. Bicalutamide, which has been incriminated in one report of acute hepatitis (108), is an exception; the causality association between bicalutamide and liver injury has been questioned (109). The patient had taken two doses of bicalutamide but had also received other antiandrogen therapy in the recent past, including three months of flutamide and one month of cyproterone acetate. There appears to be no clear relation between antiandrogen dose and liver injury, duration of use or gender (liver toxicity is also seen in women). However, low-dose cyproterone acetate (1–2 mg) regimes used in acne appear to have a low risk of liver injury when compared with conventional cyproterone protocols (50–100 mg) used in prostate cancer (107). Liver histologic appearances include acute hepatitis or cholestatic hepatitis. In severe cases, bridging hepatic necrosis and massive hepatic necrosis have been reported (110). Flutamide liver toxicity manifests earlier than cyproterone (mean, 14 weeks vs. 30 weeks after drug exposure). Cyproterone has been implicated in causing HCC and cirrhosis (111,112). Among six cases of HCC reported in cyproterone recipients, two had Turner’s syndrome and had also received human growth hormone and buserelin. The duration of cyproterone use ranged from 4 to 728 weeks (median, 280 weeks). In rat hepatocytes, cyproterone increases the proportion of cells in the replicative (G2 phase), which could be mechanistically relevant to hepatocarcinogenesis (113). However, the association with HCC has been documented in only a small number of patients, and hepatitis C virus infection was not excluded in some of these cases. Considering the widespread use of cyproterone (59,000 prescriptions dispensed in Australia for the year 2003 alone), this appears to be a rare association. Likewise, the role of cyproterone in causing cirrhosis is not clear and it has been implicated in one case only; a 10-year-old boy with hypothalamic syndrome and central precocious puberty who had received cyproterone for longer than four years (112). Mechanism of Liver Injury The lack of hypersensitivity features with cases of antiandrogen-associated liver injury is consistent with direct drug and/or drug metabolite-induced toxicity. Flutamide exerts cytotoxic effects to isolated rat hepatocytes via CYP3A and CYP1A-mediated formation of electrophilic metabolites (114). In addition, this drug (and also nilutamide) impairs mitochondrial respiration and decreases ATP levels in vitro (114). Nilutamide also induces oxidative stress in isolated rat liver cells (115,116), but the relevance of studies in short-term suspensions of hepatocytes to human toxicology is unclear. Outcome and Treatment Although most patients usually recover after antiandrogens are discontinued (117), there are instances of severe hepatitis where the outcome is less favorable; at least 10 liver-related deaths have been recorded with both flutamide and cyproterone (118,119). Ursodeoxycholic acid (UDCA) has been used with success in a case with flutamide-associated cholestasis (120). In a retrospective Japanese study, patients receiving UDCA concomitantly with flutamide were less likely to have an increase in serum transaminases than those on flutamide alone (11% vs. 32%), respectively (121). However, it is not clear from this report whether any of the patients with deranged transaminases had jaundice or had symptoms of an acute hepatitis. The applicability of UDCA for patients developing acute liver failure with flutamide is therefore unclear. Transfer to a tertiary center for liver transplant assessment is critical and remains the standard of care. CORTICOSTEROIDS Hepatic steatosis has been described with the prolonged use of corticosteroids. Corticosteroidinduced steatosis is likely a consequence of exacerbation of metabolic disorders linked to non-alcoholic fatty liver disease, such as obesity, hypertriglyceridemia, and type 2 diabetes (95). Interestingly, even in the absence of corticosteroid therapy recent data indicate cortisol

716

Chitturi and Farrell

hypersecretion and overactivation of the hypothalamo–pituitary–adrenaline axis in persons with NAFLD (122). It is therefore tempting to speculate that supraphysiologic doses of corticosteroids could aggravate these hormonal changes, and this is worthy of future study. ANTITHYROID DRUGS Liver toxicity resulting from antithyroid drugs has been recognized for over 50 years. Even so, when a patient with hyperthyroidism is referred with altered liver tests, non-drug related causes should first be carefully excluded. These causes include autoimmune hepatitis, which can accompany (123) or present for the first time after commencing antithyroid therapy (124). Hyperthyroidism itself can induce liver enzyme changes, and this should be taken into consideration (125). In one study of 95 propylthiouracil (PTU) recipients, abnormal liver tests were observed in 72 (76%) cases at baseline (126). Of these, an increase in total serum alkaline phosphatase was most frequent (64%), and was largely attributable to a rise in the bone isoenzyme fraction. Serum AT, gamma glutamyl transpeptidase, and bilirubin levels were elevated in 27% to 37%, 17%, and 5%, respectively (126). Of the patients with raised serum alanine aminotransferase (ALT), 62% normalized their serum transaminases while continuing PTU. The remainder continued to show transient but asymptomatic elevations in ALT (126). Jaundice is infrequent in patients with hyperthyroidism, and when present, tends to be mild. It is usually attributable to hyperthyroidism-associated heart failure, but can occur in its absence (127). Severe intrahepatic cholestasis (serum bilirubin O35 mg/dL) is an extremely rare complication of inadequately treated hyperthyroidism, and is reversed upon treatment with antithyroid drugs (128). Methimazole and Carbimazole Methimazole and carbimazole have been implicated in more than 20 cases of cholestasis or cholestatic hepatitis (129–132). The onset is within 2 and12 weeks after beginning treatment. Uneventful recovery is generally the rule, but rare instances of liver-related deaths or acute liver failure necessitating liver transplantation have been documented (130,133,134). In one fatal case, the patient had hepatitis B-related cirrhosis, which may have contributed to the outcome (134). Methimazole and carbimazole-induced liver injury are probably examples of immunemediated drug reactions. Thus, recurrence of cholestasis within five hours of rechallenge has been documented with methimazole (135). Likewise, positive rechallenge within 24 hours has been documented with carbimazole (136). Further, immune phenomena such as erythema nodosum may accompany liver injury in some cases (136). It is also possible that metabolic idiosyncrasy may underlie some cases of methimazole hepatotoxicity. Like acetaminophen, methimazole is metabolized by cytochrome P450 to generate a reactive metabolite, N-methylthiourea (137). Glutathione-depleted mice developed severe liver injury when treated with methimazole, an observation, which could indicate that inadequate detoxification of this reactive metabolite, is central to methimazole-induced liver injury (137,138). Propylthiouracil Unlike methimazole and carbimazole, PTU is associated with a predominantly hepatocellular pattern of liver injury (139–142). While ALT elevations are frequently observed during PTU treatment, clinical evidence of drug hepatitis is less common. Thus, Kim and colleagues observed six (1.2%) cases of symptomatic liver injury in over 900 patients treated with PTU (143). The case definition of symptomatic liver injury was jaundice or symptoms of hepatitis, together with a threefold increase in serum ALT above the upper limit of normal. A collective review of 50 years of experience with PTU identified 36 cases of acute hepatitis, including seven deaths from acute liver failure (139). Contemporary data show that PTU liver injury remains an important cause of severe hepatitis. The United Network for Organ Sharing (UNOS) liver transplant database (1990–2002) was examined for the leading causes of drug-induced liver failure (144). Besides acetaminophen (133 cases, 49%), other drugs

Effects of Hormones and Hormone Antagonists

717

implicated were isoniazid (24 cases, 17.5%), PTU (13 cases, 9.5%), phenytoin, and valproate (10 cases each, 7.3%). Clinical Presentation The clinical profile is indistinguishable from other causes of acute hepatitis. Most cases present within the first few months of PTU treatment (range, 16–49 days); rarely, acute hepatitis may develop beyond as late as 14 months. There are no clear predisposing factors for PTU-induced liver injury; there is no association with age, thyroid status, or PTU dose (139,141). In keeping with the typical gender profile of patients with hyperthyroidism, women predominate in most case series. Neonatal hepatitis associated with maternal PTU use has been described (145). Pathology and Pathogenesis Liver histology shows portal inflammation, cholestasis, and zone 3 necrosis in mild cases, to submassive or massive hepatic necrosis in patients developing severe liver toxicity (146). The frequent presence of accompanying hypersensitivity manifestations and a positive lymphocyte stimulation test in some cases is consistent with immune-mediated liver injury. Treatment and Outcome Most patients recover following PTU discontinuation. However, in severe cases, there can be a progressive decline to acute or subacute liver failure. Early discussions with the regional liver transplant center should therefore be initiated. Anecdotally, corticosteroids, fresh frozen plasma, and plasmapheresis (147) have been used with success in some cases, but their use remains investigational (141). In view of the severe, potentially fatal hepatic reaction, rechallenge with propylthiouracil is not recommended. REFERENCES 1. Bennion LJ, Ginsberg RL, Gemick MB, et al. Effects of oral contraceptives on the gallbladder bile of normal women. N Engl J Med 1976; 294(4):189–92. 2. Kern F, Jr., Everson GT, DeMark B, et al. Biliary lipids, bile acids and gall bladder function in the human female: effects of pregnancy and the ovulatory cycle. J Clin Invest 1981; 68(5):1229–42. 3. Rooks JB, Ory HW, Ishak KG, et al. The cooperative liver tumor study group. Epidemiology of hepatocellular adenoma: the role of oral contraceptive use. JAMA 1979; 242(7):644–8. 4. Haemmerli UP, Wyss HI. Recurrent intrahepatic cholestasis of pregnancy. Report of six cases, and review of the literature. Medicine (Baltimore) 1967; 46(4):299–321. 5. Eisalo A, Jarvinen PA, Luukkainen T. Hepatic impairment during the intake of contraceptive pills: clinical trial with postmenopausal women. Br Med J 1964; 5406:426–7. 6. Holzbach RT, Sivak DA, Braun WE. Familial recurrent intrahepatic cholestasis of pregnancy: a genetic study providing evidence for transmission of sex-limited, dominant trait. Gastroenterology 1983; 85(1):175–9. 7. Dalen E, Westerholm B. Occurrence of hepatic impairment in women jaundiced by oral contraceptives and their mothers and sisters. Acta Med Scand 1974; 195(6):459–63. 8. Stieger B, Fattinger K, Madon J, et al. Drug- and estrogen-induced cholestasis through inhibition of the hepatocellular bile salt export pump (Bsep) of rat liver. Gastroenterology 2000; 118(2):422–30. 9. Anand V, Gorard DA. Norethisterone-induced cholestasis. QJM 2005; 98(3):232–4. 10. Foitl DR, Hyman G, Lefkowitch JH. Jaundice and intrahepatic cholestasis following high-dose megestrol acetate for breast cancer. Cancer 1988; 63(3):438–9. 11. Langlands AO, Martin WM. Jaundice associated with norethisterone-acetate treatment of breast cancer. Lancet 1975; 1(7906):584–5. 12. Schaffner F. The effect of oral contraceptives on the liver. JAMA 1966; 198:1019–21. 13. Chitturi S, Farrell GC. Drug-induced cholestasis. Semin Gastrointest Dis 2001; 12(2):113–24. 14. Farrell GC. Drug-induced cholestasis. In: Farrell GC, ed. Drug-Induced Liver Disease. London: Churchill Livingstone, 1994:331 (chap. 14). 15. Schonberg LA. Peliosis hepatis and oral contraceptives. J Reprod Med 1982; 27(12):753–6. 16. Staub PG, Leibowitz CB. Peliosis hepatis associated with oral contraceptive use. Australas Radiol 1996; 40(2):172–4. 17. Corpa MV, Bacchi MM, Bacchi CE, et al. Peliosis hepatis associated with lymphoplasmacytic lymphoma: an autopsy case report. Arch Pathol Lab Med 2004; 128(11):1283–5. 18. Gushiken FC. Peliosis hepatis after treatment with 2-chloro-3 0 -deoxyadenosine. South Med J 2000; 93(6):625–6.

718

Chitturi and Farrell

19. Valla D, Le MG, Poynard T, et al. Risk of hepatic vein thrombosis in relation to recent use of oral contraceptives. A case-control study. Gastroenterology 1986; 90(4):807–11. 20. Gomes MP, Deitcher SR. Risk of venous thromboembolic disease associated with hormonal contraceptives and hormone replacement therapy: a clinical review. Arch Intern Med 2004; 164(18):1965–76. 21. Chu G, Farrell GC. Portal vein thrombosis associated with prolonged ingestion of oral contraceptive steroids. J Gastroenterol Hepatol 1993; 8(4):390–3. 22. Maddrey WC. Hepatic vein thrombosis (Budd-Chiari syndrome): possible association with the use of oral contraceptives (review). Semin Liver Dis 1987; 7(1):32–9. 23. Minnema MC, Janssen HL, Niermeijer P, et al. Budd-Chiari syndrome: combination of genetic defects and the use of oral contraceptives leading to hypercoagulability. J Hepatol 2000; 33(3):509–12. 24. Janssen HL, Meinardi JR, Vleggaar FP, et al. Factor V Leiden mutation, prothrombin gene mutation, and deficiencies in coagulation inhibitors associated with Budd-Chiari syndrome and portal vein thrombosis: results of a case-control study. Blood 2000; 96(7):2364–8. 25. Gemer O, Moscovici O, Ben-Horin CL, et al. Oral contraceptives and liver hemangioma: a casecontrol study. Acta Obstet Gynecol Scand 2004; 83(12):1199–201. 26. Morley JE, Myers JB, Sack FS, et al. Enlargement of cavernous haemangioma associated with exogenous administration of oestrogens. S Afr Med J 1974; 48(16):695–7. 27. Conter RL, Longmire WP, Jr. Recurrent hepatic hemangiomas: possible association with estrogen therapy. Ann Surg 1988; 207(2):115–9. 28. Xiao X, Hong L, Sheng M. Promoting effect of estrogen on the proliferation of hemangioma vascular endothelial cells in vitro. J Pediatr Surg 1999; 34(11):1603–5. 29. Reddy KR, Kligerman S, Levi J, et al. Benign and solid tumors of the liver: relationship to sex, age, size of tumors, and outcome. Am Surg 2001; 67(2):173–8. 30. Mathieu D, Kobeiter H, Maison P, et al. Oral contraceptive use and focal nodular hyperplasia of the liver. Gastroenterology 2000; 118(3):560–4. 31. Scalori A, Tavani A, Gallus S, et al. Oral contraceptives and the risk of focal nodular hyperplasia of the liver: a case-control study. Am J Obstet Gynecol 2002; 186(2):195–7. 32. Baum JK, Bookstein JJ, Holtz F, Klein EW. Possible association between benign hepatomas and oral contraceptives. Lancet 1973; 2(7835):926–9. 33. Edmondson HA, Henderson B, Benton B. Liver-cell adenomas associated with use of oral contraceptives. N Engl J Med 1976; 294(9):470–2. 34. Edmondson HA. Tumors of the liver and intrahepatic bile ducts. Atlas of Tumor Pathology. Washington, DC: AFIP, 1958:193–206 (Section 7, Fascicle 25). 35. Blayney AW, O’Callaghan T, MacErlean DP, et al. Fatal outcome of an hepatic adenoma following short term oral contraceptive use. Ir Med J 1977; 70(15):455–6. 36. Kerlin P, Davis GL, McGill DB, et al. Hepatic adenoma and focal nodular hyperplasia: clinical, pathologic, and radiologic features. Gastroenterology 1983; 84(5 Pt 1):994–1002. 37. Aseni P, Sansalone CV, Sammartino C, et al. Rapid disappearance of hepatic adenoma after contraceptive withdrawal. J Clin Gastroenterol 2001; 33(3):234–6. 38. Yu MC, Yuan JM. Environmental factors and risk for hepatocellular carcinoma. Gastroenterology 2004; 127(5 Suppl. 1):S72–8. 39. Henderson BE, Preston-Martin S, Edmondson HA, et al. Br J Cancer 1983; 48(3):437–40. 40. Neuberger J, Forman D, Doll R, et al. Oral contraceptives and hepatocellular carcinoma. Br Med J 1986; 292(6532):1355–7. 41. Forman D, Vincent TJ, Doll R. Cancer of the liver and the use of oral contraceptives. Br Med J 1986; 292(6532):1357–61. 42. Palmer JR, Rosenberg L, Kaufman DW, et al. Oral contraceptive use and liver cancer. Am J Epidemiol 1989; 130(5):878–82. 43. Yu MC, Tong MJ, Govindarajan S, et al. Nonviral risk factors for hepatocellular carcinoma in a lowrisk population, the non-Asians of Los Angeles County, California. J Natl Cancer Inst 1991; 83(24):1820–6. 44. Hsing AW, Hoover RN, McLaughlin JK, et al. Oral contraceptives and primary liver cancer among young women. Cancer Causes Control 1992; 3(1):43–8. 45. Tavani A, Negri E, Parazzini F, et al. Female hormone utilisation and risk of hepatocellular carcinoma. Br J Cancer 1993; 67(3):635–7. 46. The Collaborative MILTS Project Team. Oral contraceptives and liver cancer. Results of the Multicentre International Liver Tumor Study (MILTS). Contraception 1997; 56(5):275–84. 47. Ishak KG, Zimmerman HJ. Hepatotoxic effects of the anabolic/androgenic steroids. Semin Liver Dis 1987; 7(3):230–6. 48. The WHO Collaborative Study of Neoplasia and Steroid Contraceptives. Combined oral contraceptives and liver cancer. Int J Cancer 1989; 43(2):254–9.

Effects of Hormones and Hormone Antagonists

719

49. Kew MC, Song E, Mohammed A, Hodkinson J. Contraceptive steroids as a risk factor for hepatocellular carcinoma: a case/control study in South African black women. Hepatology 1990; 11(2):298–302. 50. Prentice RL. Epidemiologic data on exogenous hormones and hepatocellular carcinoma and selected other cancers. Prev Med 1991; 20(1):38–46. 51. Dourakis SP, Tolis G. Sex hormonal preparations and the liver. Eur J Contracept Reprod Health Care 1998; 3(1):7–16. 52. Gelfand MM, Wiita B. Androgen and estrogen–androgen hormone replacement therapy: a review of the safety literature, 1941 to 1996. Clin Ther 1997; 19(3):383–404. 53. Makhlouf HR, Ishak KG, Goodman ZD. Epithelioid hemangioendothelioma of the liver: a clinicopathologic study of 137 cases. Cancer 1999; 85(3):562–82. 54. Dean PJ, Haggitt RC, O’Hara CJ. Malignant epithelioid hemangioendothelioma of the liver in young women. Relationship to oral contraceptive use. Am J Surg Pathol 1985; 9(10):695–704. 55. Shi EC, Fischer A, Crouch R, Ham JM. Possible association of angiosarcoma with oral contraceptive agents. Med J Aust 1981; 1(9):473–4. 56. Thung SN, Gerber MA. Precursor stage of hepatocellular neoplasm following long exposure to orally administered contraceptives. Hum Pathol 1981; 12(5):472–4. 57. Yoshida EM, Erb SR, Scudamore CH, et al. Severe cholestasis and jaundice secondary to an esterified testosterone, a non-C17 alkylated anabolic steroid. J Clin Gastroenterol 1994; 18(3):268–70. 58. Glober GA, Wilkerson JA. Biliary cirrhosis following the administration of methyltestosterone. JAMA 1968; 204(3):170–3. 59. Farrell GC. Drug-induced liver disease. In: Gitnick G, ed. Current Hepatology. Vol. 17. St. Louis, MO: Mosby-Year Book, 1997:131–68. 60. Gurakar A, Caraceni P, Fagiuoli S, et al. Androgenic/anabolic steroid-induced intrahepatic cholestasis: a review with four additional case reports. J Okla State Med Assoc 1994; 87(9):399–404. 61. Stimac D, Milic S, Dintinjana RD, et al. Androgenic/anabolic steroid-induced toxic hepatitis. J Clin Gastroenterol 2002; 35(4):350–2. 62. Schumacher J, Muller G, Klotz K-F. Large hepatic hematoma and intraabdominal hemorrhage associated with abuse of anabolic steroids. N Engl J Med 1999; 340(14):1123–4. 63. Capra F, Nicolini N, Morana G, et al. Vanishing bile duct syndrome and inflammatory pseudotumor associated with a case of anabolic steroid abuse. Dig Dis Sci 2005; 50(8):1535–7. 64. Zafrani ES, von Pinaudeau Y, Dhumeaux D. Drug-induced vascular lesions of the liver. Arch Intern Med 1983; 143(3):495–502. 65. Soe KL, Soe M, Gluud C. Liver pathology associated with the use of anabolic–androgenic steroids. Liver 1992; 12:73–9. 66. Johnson FL, Lerner KG, Siegel M, et al. Association of androgenic–anabolic steroid therapy with development of hepatocellular carcinoma. Lancet 1972; 2(7790):1273–6. 67. Farrell GC, Joshua DE, Uren RF, et al. Androgen-induced hepatoma. Lancet 1975; 1(7904):430–2. 68. Touraine RL, Bertrand Y, Foray P, et al. Hepatic tumors during androgen therapy in Fanconi anaemia. Eur J Pediatr 1993; 152(8):691–3. 69. Velazquez I, Alter BP. Androgens and liver tumors: Fanconi’s anemia and non-Fanconi’s conditions. Am J Hematol 2004; 77(3):257–67. 70. Westaby D, Portmann B, Williams R. Androgen related primary hepatic tumors in non-Fanconi patients. Cancer 1983; 51(10):1947–52. 71. Middleton C, McCaughan GW, Painter DM, et al. Danazol and hepatic neoplasia: a case report. Aust NZ J Med 1989; 19(6):733–5. 72. Bork K, Pitton M, Harten P, et al. Hepatocellular adenomas in patients taking danazol for hereditary angio-oedema. Lancet 1999; 353(9158):1066–7. 73. Shahidi NT. A review of the chemistry, biological action, and clinical applications of anabolic– androgenic steroids. Clin Ther 2001; 23(9):1355–90. 74. McCaughan GW, Bilous MJ, Gallagher ND. Long-term survival with tumor regression in androgeninduced liver tumors. Cancer 1985; 56(11):2622–6. 75. Anthony PP. Hepatoma associated with androgenic steroids? Lancet 1975; 1(7908):685–6. 76. Falk H, Popper H, Thomas LB, et al. Hepatic angiosarcoma associated with androgenic–anabolic steroids. Lancet 1979; 2(8152):1120–2. 77. Bagia S, Hewitt PM, Morris DL. Anabolic steroid-induced hepatic adenomas with spontaneous haemorrhage in a bodybuilder. Aust NZ J Surg 2000; 70(9):686–7. 78. Guy JT, Auslander MO. Androgenic steroids and hepatocellular carcinoma. Lancet 1973; 1(7795):148. 79. Bruno S, Maisonneuve P, Castellana P, et al. Incidence and risk factors for non-alcoholic steatohepatitis: prospective study of 5408 women enrolled in Italian tamoxifen chemoprevention trial. Br Med J 2005; 330(7497):932. 80. Elefsiniotis IS, Pantazis KD, Ilias A, et al. Tamoxifen induced hepatotoxicity in breast cancer patients with pre-existing liver steatosis: the role of glucose intolerance. Eur J Gastroenterol Hepatol 2004; 16(6):593–8.

720

Chitturi and Farrell

81. Nishino M, Hayakawa K, Nakamura Y, et al. Effects of tamoxifen on hepatic fat content and the development of hepatic steatosis in patients with breast cancer: high frequency of involvement and rapid reversal after completion of tamoxifen therapy. Am J Roentgenol 2003; 180(1):129–34. 82. Nguyen MC, Stewart RB, Banerji MA, et al. Relationships between tamoxifen use, liver fat and body fat distribution in women with breast cancer. Int J Obes Relat Metab Disord 2001; 25(2):296–8. 83. Cai Q, Bensen M, Greene R, et al. Tamoxifen-induced transient multifocal hepatic fatty infiltration. Am J Gastroenterol 2000; 95(1):277–9. 84. Dray X, Tainturier MH, De La Lande P, et al. Cirrhosis with non alcoholic steatohepatitis: role of tamoxifen. Gastroenterol Clin Biol 2000; 24(11):1122–3. 85. Oien KA, Moffat D, Curry GW, et al. Cirrhosis with steatohepatitis after adjuvant tamoxifen. Lancet 1999; 353(9146):36–7. 86. Ogawa Y, Murata Y, Nishioka A, et al. Tamoxifen-induced fatty liver in patients with breast cancer (letter). Lancet 1998; 351(904):725. 87. Cortez-Pinto H, Baptista A, Camilo ME, et al. Tamoxifen-associated steatohepatitis-report of three cases. J Hepatol 1995; 23(1):95–7. 88. Van Hoof M, Rahier J, Horsmans Y. Tamoxifen-induced steatohepatitis (letter). Ann Intern Med 1996; 124(9):855–6. 89. Pratt DS, Knox TA, Erban J. Tamoxifen-induced steatohepatitis (letter). Ann Intern Med 1995; 123(3):236. 90. Storen EC, Hay JE, Kaur J, et al. Tamoxifen-induced submassive hepatic necrosis. Cancer J 2000; 6(2):58–60. 91. Ching CK, Smith PG, Long RG. Tamoxifen-associated hepatocellular damage and agranulocytosis. Lancet 1992; 339(8798):940. 92. Chitturi S, Abeygunasekera S, Farrell GC, et al. NASH and insulin resistance: insulin hypersecretion and specific association with the insulin resistance syndrome. Hepatology 2002; 35(2):373–9. 93. Marchesini G, Forlani G. NASH: from liver diseases to metabolic disorders and back to clinical hepatology. Hepatology 2002; 35(2):497–9. 94. Pagano G, Pacini G, Musso G, et al. Nonalcoholic steatohepatitis, insulin resistance, and metabolic syndrome: further evidence for an etiologic association. Hepatology 2002; 35(2):367–72. 95. Farrell GC. Drugs and steatohepatitis. Semin Liver Dis 2002; 22(2):185–94. 96. Larosche I, Letteron PH, Mansouri A, et al. Tamoxifen inhibits mitochondrial function, topoisomerases and mitochondrial DNA synthesis, and causes progressive mitochondrial DNA depletion and steatosis in mouse liver. J Hepatol 2005; 42(Suppl. 2):23–4. 97. Fukumoto M, Masuda K, Ogawa Y, et al. In vivo hepatic fatty acid metabolism in patients with nonalcoholic steatohepatitis using semiquantitative 123 I-BMIPP liver scan. Hepatol Res 2005; 33(2):105–9. 98. Lelliott CJ, Lopez M, Curtis RK, et al. Transcript and metabolite analysis of the effects of tamoxifen in rat liver reveals inhibition of fatty acid synthesis in the presence of hepatic steatosis. FASEB J 2005; 19(9):1108–19. 99. Ohnishi T, Ogawa Y, Saibara T, et al. CYP17 polymorphism as a risk factor of tamoxifen-induced hepatic steatosis in breast cancer patients. Oncol Rep 2005; 13(3):485–9. 100. Nemoto Y, Toda K, Ono M, et al. Altered expression of fatty acid-metabolizing enzymes in aromatase-deficient mice. J Clin Invest 2000; 105(12):1819–25. 101. Saibara T, Onishi S, Ogawa Y, et al. Bezafibrate for tamoxifen-induced non-alcoholic steatohepatitis. Lancet 1999; 353(9166):1802. 102. Hamada N, Ogawa Y, Saibara T, et al. Toremifene-induced fatty liver and NASH in breast cancer patients with breast-conservation treatment. Int J Oncol 2000; 17(6):1119–23. 103. Olsson R, Tyllstrom J, Zettergren L. Hepatic reactions to cyclofenil. Gut 1983; 24(3):260–3. 104. Manso G, Thole Z, Salgueiro E, et al. Spontaneous reporting of hepatotoxicity associated with antiandrogens: data from the Spanish pharmacovigilance system. Pharmacoepidemiol Drug Saf 2006; 15(4):253–9. 105. Thole Z, Manso G, Salgueiro E, et al. Hepatotoxicity induced by antiandrogens: a review of the literature. Urol Int 2004; 73(4):289–95. 106. Wysowski DK, Fourcroy JL. Flutamide hepatotoxicity. J Urol 1996; 155(1):209–12. 107. Anon. High dose cypropterone and hepatotoxicity. Aust Adv Drug React Bull 2004; 23(1):3. 108. Dawson LA, Chow E, Morton G. Fulminant hepatic failure associated with bicalutamide. Urology 1997; 49(2):283–4. 109. Chodak GW. Bicalutamide-associated fulminant hepatic failure. Urology 1997; 50(6):1027. 110. Wysowski DK, Freiman JP, Tourtelot JB, et al. Fatal and nonfatal hepatotoxicity associated with flutamide. Ann Intern Med 1993; 118(11):860–4. 111. Watanabe S, Yamasaki S, Tanae A, et al. Three cases of hepatocellular carcinoma among cyproterone users. Ad hoc Committee on Androcur Users. Lancet 1994; 344(8936):1567–8. 112. Garty BZ, Dinari G, Gellvan A, et al. Cirrhosis in a child with hypothalamic syndrome and central precocious puberty treated with cyproterone acetate. Eur J Pediatr 1999; 158(5):367–70.

Effects of Hormones and Hormone Antagonists

721

113. Grasl-Kraupp B, Luebeck G, Wagner A. Quantitative analysis of tumor initiation in rat liver: role of cell replication and cell death (apoptosis). Carcinogenesis 2000; 21(7):1411–21. 114. Fau D, Eugene D, Berson A, et al. Toxicity of the antiandrogen flutamide in isolated rat hepatocytes. J Pharmacol Exp Ther 1994; 269(3):954–62. 115. Fau D, Berson A, Eugene D, et al. Mechanism for the hepatotoxicity of the antiandrogen, nilutamide. Evidence suggesting that redox cycling of this nitroaromatic drug leads to oxidative stress in isolated hepatocytes. J Pharmacol Exp Ther 1992; 263(1):69–77. 116. Berson A, Schmets L, Fisch C, et al. Inhibition by nilutamide of the mitochondrial respiratory chain and ATP formation. Possible contribution to the adverse effects of this antiandrogen. J Pharmacol Exp Ther 1994; 270(1):167–76. 117. Gomez JL, Dupont A, Cusan L, et al. Incidence of liver toxicity associated with the use of flutamide in prostate cancer patients. Am J Med 1992; 92(5):465–70. 118. Friedman G, Lamoureux E, Sherker AH. Fatal fulminant hepatic failure due to cyproterone acetate. Dig Dis Sci 1999; 44(7):1362–3. 119. Famularo G, De Simone C, Minisola G, et al. Flutamide-associated acute liver failure. Ann Ital Med Int 2003; 18(4):250–3. 120. Cicognani C, Malavolti M, Morselli-Labate AM, et al. Flutamide-induced toxic hepatitis. Potential utility of ursodeoxycholic acid administration in toxic hepatitis. Dig Dis Sci 1996; 41(11):2219–21. 121. Kojima M, Kamoi K, Ukimura O, et al. Clinical utility of ursodeoxycholic acid in preventing flutamide-induced hepatopathy in patients with prostate cancer: a preliminary study. Int J Urol 2002; 9(1):42–6. 122. Targher G, Bertolini L, Rodella S, et al. Associations between liver histology and cortisol secretion in subjects with nonalcoholic fatty liver disease. Clin Endocrinol (Oxf) 2006; 64(3):337–41. 123. Nagai T, Imamura M, Kamiya Y, et al. Graves’ disease accompanied by anti-myeloperoxidase antibody-related nephropathy and autoimmune hepatitis. Intern Med 2004; 43(6):516–20. 124. Marino M, Morabito E, Altea MA, et al. Autoimmune hepatitis during intravenous glucocorticoid pulse therapy for Graves’ ophthalmopathy treated successfully with glucocorticoids themselves. J Endocrinol Invest 2005; 28(3):280–4. 125. Malik R, Hodgson H. The relationship between the thyroid gland and the liver. QJM 2002; 95(9):559–69. 126. Huang MJ, Li KL, Wei JS, et al. Sequential liver and bone biochemical changes in hyperthyroidism: prospective controlled follow-up study. Am J Gastroenterol 1994; 89(7):1071–6. 127. Fong TL, McHutchison JG, Reynolds TB. Hyperthyroidism and hepatic dysfunction. A case series analysis. J Clin Gastroenterol 1992; 14(3):240–4. 128. Hasan MK, Tierney WM, Baker MZ. Severe cholestatic jaundice in hyperthyroidism after treatment with 131-iodine. Am J Med Sci 2004; 328(6):348–50. 129. Vitug AC, Goldman JM. Hepatotoxicity from antithyroid drugs. Horm Res 1985; 21(4):229–34. 130. Epeirier JM, Pageaux GP, Coste V, et al. Fulminant hepatitis after carbimazole and propranolol administration. Eur J Gastroenterol Hepatol 1996; 8(3):287–8. 131. Schwab GP, Wetscher GJ, Vogl W, et al. Methimazole-induced cholestatic liver injury, mimicking sclerosing cholangitis. Langenbecks Arch Chir 1996; 381(4):225–7. 132. Lunzer M, Huang SN, Ginsburg J. Jaundice due to carbimazole. Gut 1975; 16(11):913–7. 133. Binder C, Lang W. Necrotizing hepatitis with a fatal outcome after carbimazole therapy. Dtsch Med Wochenschr 1993; 118(42):1515–9. 134. Kang H, Choi JD, Jung IG, et al. A case of methimazole-induced acute hepatic failure in a patient with chronic hepatitis B carrier. Korean J Intern Med 1990; 5(1):69–73. 135. Schmidt G, Borsch G, Muller KM, et al. Methimazole-associated cholestatic liver injury: case report and brief literature review. Hepatogastroenterology 1986; 33(6):244–6. 136. Marazuela M, Sanchez de Paco G, Jimenez I, et al. Acute pancreatitis, hepatic cholestasis, and erythema nodosum induced by carbimazole treatment for Graves’ disease. Endocr J 2002; 49(3):315–8. 137. Mizutani T, Murakami M, Shirai M, et al. Metabolism-dependent hepatotoxicity of methimazole in mice depleted of glutathione. J Appl Toxicol 1999; 19(3):193–8. 138. Mizutani T, Yoshida K, Murakami M, et al. Evidence for the involvement of N-methylthiourea, a ring cleavage metabolite, in the hepatotoxicity of methimazole in glutathione-depleted mice: structuretoxicity and metabolic studies. Chem Res Toxicol 2000; 13(3):170–6. 139. Williams KV, Nayak S, Becker D, et al. Fifty years of experience with propylthiouracil-associated hepatotoxicity: what have we learned? J Clin Endocrinol Metab 1997; 82(6):1727–33. 140. Jonas MM, Eidson MS. Propylthiouracil hepatotoxicity: two pediatric cases and review of the literature. J Pediatr Gastroenterol Nutr 1988; 7(5):776–9. 141. Ichiki Y, Akahoshi M, Yamashita N, et al. Propylthiouracil-induced severe hepatitis: a case report and review of the literature. J Gastroenterol 1998; 33(5):747–50. 142. Ruiz JK, Rossi GV, Vallejos HA, et al. Fulminant hepatic failure associated with propylthiouracil. Ann Pharmacother 2003; 37(2):224–8.

722

Chitturi and Farrell

143. Kim HJ, Kim BH, Yan YS, et al. The incidence and clinical characteristics of symptomatic propylthiouracil-induced hepatic injury in patients with hyperthyroidism: a single-center retrospective study. Am J Gastroenterol 2001; 96(1):165–9. 144. Russo MW, Galanko JA, Shrestha R, et al. Liver transplantation for acute liver failure from drug induced liver injury in the United States. Liver Transpl 2004; 10(8):1018–23. 145. Hayashida CY, Duarte AJ, Sato AE, et al. Neonatal hepatitis and lymphocyte sensitization by placental transfer of propylthiouracil. J Endocrinol Invest 1990; 13(11):937–41. 146. Testa G, Trevino J, Bogetti D, et al. Liver transplantation for propylthiouracil-induced acute hepatic failure. Dig Dis Sci 2003; 48(1):190–1. 147. Aydemir S, Ustundag Y, Bayraktaroglu T, et al. Fulminant hepatic failure associated with propylthiouracil: a case report with treatment emphasis on the use of plasmapheresis. J Clin Apheresis 2005; 20(4):235–8.

33

Mushroom Poisoning: A Clinical Model of Toxin-Induced Centrilobular Necrosis FranÅois Durand and Dominique Valla

Service d’HØpatologie, Hospital Beaujon, Clichy, France

INTRODUCTION A few non-toxic mushrooms are highly appreciated by gourmets because of their delicate and inimitable flavor. However, other mushrooms yielding in the same geographical areas, Amanita in particular, are well known to be highly toxic to humans as well as animals, even when ingested in small amounts. Since Amanita can be easily distinguished from non-toxic mushrooms on the basis of their general aspect and, according to the testimony of survivors from intoxication the taste of this toxic mushroom is rather disgusting, it is a paradox that cases of fatal intoxication with acute liver failure continue to be reported worldwide. Unfortunately, greediness and heedless curiosity frequently prevail on cautiousness. In this chapter, we will essentially discuss about the mechanisms and consequences of Amanita mushroom poisoning, because this species, the most dangerous, represents a model of liver toxicity. We will also discuss about the management of Amanita mushroom poisoning which still represents a challenging issue. Other species of wild, uncultivated, mushrooms which can also be a source of intoxication (1) with gastrointestinal disorders, central nervous system toxicity, myopathy, and acute renal failure will not be detailed in this chapter.

AMANITA: THE MUSHROOMS Of more than 5000 species of mushrooms, only about 50 are poisonous for humans. Again Amanita represents the most poisonous genus. Not all Amanita species are toxic for humans. Those containing amatoxins are A. phalloides, A. verna, A. virosa, A. ocreata, A. bisporigera, A. suballiaceae, A. tenuifolia, and A. hygroscopica (2). Other species belonging to Lepiota and Galerina genus, although less frequently involved, also contain amatoxins. Amanita phalloides mushrooms, frequently termed as “death cap” or “destroying angel,” are white capped with a yellowish to greenish top at the center of the cap. They are characterized by white close gills, a white stalk enlarging to a basal bulb and a membranous volva. Amanita verna only differs from A. phalloides by a uniformly white cap. These mushrooms are found on the ground below the woods or in the grass near trees. They yield from late summer through fall to early winter in temperate climate geographical areas. They can be found in most parts of Europe, Russia, North America (especially along northern pacific coast), Central and South America, Australia (3), New Zealand, Malaysia, Thailand (4), India and South Africa. The incidence of amanita mushroom poisoning seems more frequent in Eastern Europe than in other parts of the world. However, intoxications are likely to be underreported in some developing geographical areas. A. phalloides predominates in Europe while A. verna is more frequently involved in the United States. Amanita mushroom poisoning often results from confusion with non-toxic species such as Gyromitra esculenta and Agaricus campestris. It can also be confused with edible members of the amanita species, Amanita cesarea in particular. The risk of confusion is especially high for young, non-mature, mushrooms.

724

Durand and Valla

THE TOXINS Toxic amanita mushrooms contain three main groups of toxins: amatoxins, phallotoxins and virotoxins. Amatoxins are the most dangerous. Amatoxins The amatoxins are bicyclic octapeptides (Fig. 1). There are different amatoxins; alpha, beta, and gamma amatoxins being the most dangerous. Their molecular weight is about 900. Amatoxins are heat-stable and thus not destroyed by cooking. Neither are they destroyed by freezing and freeze drying. Amatoxins remain toxic after prolonged periods of storage. Amatoxin concentration varies in the different parts of mushroom. The highest concentration is found in the ring (about 6 mg/g) while the bulb has the lowest toxin content (about 0.2 mg/g) (5). Overall, the toxin content is higher in mature mushrooms than in young specimens. Additionally, it seems that amatoxin content varies according to different geographical areas. It is slightly higher in A. phalloides specimens from Europe than in those from North America (5). The absorption of amatoxins varies according to different species. Rodents are not affected while guinea pigs, dogs, and human beings, among other species, readily absorb the toxin. Amatoxins are powerful toxins. Ingested amounts as low as 0.1 mg/kg are sufficient to be lethal (6). A single full-grown specimen of A. phalloides, weighing 20 g, contains about 5 to 8 mg of amatoxin and therefore, is potentially lethal. Due to their lower body weight, children are especially at risk. Amatoxins are rapidly absorbed after ingestion. The distribution of the toxin to different compartments and tissues is also rapid. As a result, amatoxin is unlikely to be detected in the serum more than 36 to 48 hours after ingestion. Amatoxins are not bound to plasma proteins. The main route of elimination is the kidney. Interestingly, toxins can still be found in urine after they have become undetectable in serum. Early after ingestion, urine concentration of amatoxins is 10 to 100 times higher than blood concentration (7). There is also biliary excretion of amatoxins which are found at a high concentration in gastroduodenal fluid. Importantly, there is an enterohepatic recycling of this toxin, contributing to a vicious circle. Experimental studies have shown that amatoxin transport into human hepatocytes is mediated by the nonspecific organic anion transporting polypeptide (OATP) 1B3, on the basolateral membrane (8). The other transporters OATP1B1 (which is the main transporter of phallotoxin) and OATP2B1 seem to have little or no transport activity for amatoxin. Under experimental conditions, membrane transport of amatoxin and hepatocyte damage are inhibited by penicillin G, silibinin, rifampicin, and cyclosporine A (8). Cholecystokinin octapeptide also inhibits amanitin uptake into hepatocytes.

O

H HO CH3 H H C 2 O N C HOH2C C C C H H H C HN H2C O C HO

N O C H C H2C R

H N

CH2 O C NH C

N H

S H2C

N H

C

OH

H C O

NH

C O

C H

N H

C

FIGURE 1 Chemical structure of the amatoxin.

CH2 O

CH3 CH CH2CH3

Mushroom Poisoning

725

The mechanism of amatoxin toxicity is related to the propensity of this compound to form tight complexes with RNA polymerase II (2,9). By complexing with the intracellular RNA polymerase II enzyme, amatoxins inhibit the formation of mRNA and, in turn, protein synthesis, rapidly resulting in cell necrosis. Cells providing transport device are mainly affected. Experimental studies in mice have shown that, besides cell necrosis, apoptosis is also involved in amatoxin-induced toxicity (10). Induction of apoptosis may result from a synergistic action of amatoxin with endogenous tumor necrosis factor (TNF). Indeed, a 50% reduction in transcription by amatoxin is sufficient for sensitizing liver cells towards toxicity of endogenous TNF, therefore promoting TNF-induced apoptosis (10). In chronological order, and in parallel with the absorption and subsequent distribution of amatoxins, the principal targets are the intestinal mucosa, liver cells, and proximal tubules of the kidney. In addition to binding with RNA polymerase II, amatoxins may interact with cytokines (tumor necrosis factor in particular), causing further cell damage via apoptosis (10). Amatoxin concentration can be analyzed in urine and blood using radioimmunoassay or high performance liquid chromatography (11,12). None of these techniques are routinely used. Neither direct identification of the toxin is absolutely necessary for establishing the diagnosis of A. phalloides poisoning, nor measurements of toxin concentration are helpful for the prognosis. Phallotoxins and Virotoxins In contrast to amatoxins, phallotoxins are poorly or not absorbed after ingestion (2). They only play a role in the initial gastrointestinal disorders after ingestion (13). However, they proved to exert severe toxicity after intraperitoneal injection in animal models (9). Similarly, it seems that the monocyclic heptapeptides virotoxins do not exert significant toxicity during A. phalloides intoxication. AMANITA POISONING: MANIFESTATIONS Amanita poisoning is mainly observed during falls and early winter, the period during which amanita species yield. Experience shows that expert mushroom hunters represent a substantial source of accidental intoxications. Otherwise, intoxications are observed in non-experts seeking for organic food products, immigrants who get confused with edible mushrooms from their own country, children, and finally young adults seeking hallucinogenic mushrooms. In this latter group, amanita poisoning can result from confusion with Psilocybe (Mellow Mushroom), a species known to have psychogenic effects. It is quite common that several family members who shared a meal containing amanita mushrooms are simultaneously intoxicated. Finally, voluntary intoxication in an attempt to suicide can occasionally occur, including after eating stored mushrooms (14) (personal observation). After an asymptomatic period of 6 to 12 hours following ingestion, initial manifestations are characterized by the abrupt onset of violent abdominal pain, vomiting, and profuse choleralike diarrhea. The intensity of watery diarrhea frequently results in significant dehydration with hypotension within a few hours. It is not uncommon that patients present with a first peak of serum transaminases [over ten times the upper limit of normal with aspartate amino transferase (AST) level superior to alanine amino transferase (ALT)] and some degree of renal insufficiency with oliguria at this early stage. These early manifestations are likely to result from hypovolemia and liver hypoxia rather than the toxic effects of amatoxins. Serum transaminases as well as serum creatinine tend to decrease after fluid resuscitation. Diarrhea persists for at least one to two days but is often more prolonged. The second phase, corresponding to direct liver toxicity, occurs 24 to 48 hours after ingestion. This phase is characterized by a marked increase (or re-elevation) in serum transaminases. Transaminases may rapidly peak over 100 times the upper limit of normal, AST being superior to ALT. Initially, serum bilirubin level is normal or only moderately elevated, remaining below 50 mmol/L. Jaundice only occurs at the latest stages. A massive decrease in coagulation factors can be observed as early as 48 hours after ingestion. There is a decrease in prothrombin index and factor V level, and a corresponding increase in international

726

Durand and Valla

normalized ratio (INR) and prothrombin time. These coagulation changes reflect the profound alteration of liver functions. However, the decrease in coagulation factors may also be related to a lesser extent to intravascular coagulation. Besides coagulopathy, a number of disorders including hyper-ammonemia, metabolic acidosis, hypophosphoremia, and high lactate level are frequently observed. Hypoglycemia is less common. The liver is normal in size. Imaging techniques are unhelpful for the diagnosis. Renal dysfunction is especially frequent in patients with a prothrombin index below 50% of normal (zINR over 1.7). Renal dysfunction may initially result from diarrhea-related hypovolemia and, later on, from the alterations of renal perfusion accompanying acute liver failure as well as direct amatoxin toxicity. Manifestations resemble those of prerenal failure with oliguria and low urinary sodium concentration. Some patients progress to acute renal failure with anuria, which precipitates metabolic acidosis. However, in some patients with established acute liver failure, aggressive fluid resuscitation enables to preserve adequate kidney function. Therefore not all patients with a fatal outcome have evidence of renal insufficiency, even at the latest stages. Following the initial vascular collapse which may occur a few hours after the onset of diarrhea, patients with acute liver failure frequently develop hemodynamic instability. Hemodynamic instability is multifactorial in origin, hypovolemia and vasoplegia being the main factors. Massive diarrhea usually persists throughout the second phase of amanita mushroom poisoning and may be responsible for protracted fluid losses, compromising renal function. Additionally, some patients experience massive bloody diarrhea related to diffuse lesions mimicking ulcerative colitis. In such patients, blood loss frequently justifies repeated transfusions of red blood cell units. Encephalopathy and brain edema occur at the latest stages and are associated with an especially poor outcome unless emergency transplantation is performed. However, there are some particularities with respect to encephalopathy in the setting of mushroom poisoning. Firstly, a minority of patients have an unfavorable outcome because of irreversible shock with an extremely rapid course, despite the fact that they never developed overt encephalopathy. This profile is rarely observed in patients with other causes of acute liver failure, except in some patients with paracetamol overdose. Secondly, in most cases, once encephalopathy occurs, the progression of multiorgan failure is especially rapid, giving little time for bridging patients to emergency transplantation. Thirdly, some patients may occasionally recover after a transient period of encephalopathy (15). Overall, an unfavorable outcome is characterized by multiorgan failure with massive metabolic acidosis and refractory shock. Conversely, in patients with a favorable outcome, improvement is characterized by a rapid increase in prothrombin index (a decrease in INR). There is no evidence that patients who recover following massive liver cell necrosis will have persistent liver lesions with longstanding inflammatory process and/or significant fibrosis, unless a distinct cause of chronic liver disease is associated.

LIVER LESIONS A. phalloides poisoning may induce massive necrosis of the liver parenchyma (Fig. 2). Liver necrosis may appear as ballooning degeneration or more often as coagulative necrosis involving nearly all the lobule. Such severe injury can leave only a few spared periportal hepatocytes, which often display microvesicular or macrovesicular steatosis. In portal tracts, a poorly developed inflammation made of mononuclear cells is usually seen. In the collapsed areas, Kupffer cells and myofibroblasts are present. Some degree of endothelialitis of terminal hepatic veins can be observed.

PROGNOSIS The proportion of patients who die or need emergency liver transplantation after A. phalloides poisoning ranges from 10% to 30% according to different series (6,16–18). The most recently reported series suggest that a 10% rate of fatalities is likely to be underestimated.

Mushroom Poisoning

727

(A)

(B)

FIGURE 2 Acute liver failure related to Amanita phalloides poisoning. (A) Massive necrosis sparing few periportal hepatocytes (PT/portal tract). (B) Reticulin staining showing drop out hepatocytes and collapse of reticulin network in centro- and mediolobular zones.

Establishing the prognosis of individuals with acute liver failure following Amanita mushroom poisoning still represents a challenging issue. Basically, the issue is to identify individuals who will not recover spontaneously and, therefore, who need emergency transplantation provided time interval until death is long enough for an efficient bridge to transplantation. Indeed, some patients with massive liver cell necrosis may have complete recovery while in others, encephalopathy and multiorgan failure progress within a few hours or days, providing little time for deciding and performing rescue transplantation. The most widely used criteria for deciding emergency transplantation in the setting of acute liver failure, namely, Clichy’s (19) and King’s College criteria (20), have obvious limitations in the context of Amanita mushroom poisoning. Firstly, as indicated above, encephalopathy generally occurs in patients with established mutliorgan failure. Thus, encephalopathy appears to be a too late criterion for making a decision of transplantation. Secondly, a substantial proportion of the patients who have a major decrease in coagulation factors (and a major increase in INR) within the first three days after ingestion will recover spontaneously, with a drop in serum transaminases and a re-elevation of coagulation factors. Thirdly, serum bilirubin, one of the variables used in King’s College criteria is only modestly elevated in case of Amanita mushroom poisoning due to the hyperacute course of the disease. An original set of prognostic criteria has been proposed recently by Ganzert M et al. (16). According to this retrospective analysis of a large series of patients, the combination of a prothrombin index less than 25% of normal with an increase in serum creatinine over 106 mmol/L from day three after ingestion predicted a fatal outcome with a 100% and 98% sensitivity. The authors proposed that these criteria could be used for deciding emergency transplantation since the median time period between the first occurrence of the criteria in

728

Durand and Valla

non-survivors and death was about 60 hours. These time dependant prognostic criteria are obviously attractive. However, the 25% limit for prothrombin index (INR of z3) can be viewed as somewhat high. For instance, an INR of 6.5 represents the threshold limit for making a decision of transplantation according to King’s College criteria (20). Additionally, in our experience, not all patients who died or had to be transplanted had an increase in serum creatinine as a result of an aggressive fluid resuscitation policy (18). Finally, potentially important variables including blood lactates could not be analyzed in this retrospective series. In a multicenter study, we found that an interval between ingestion of the mushrooms and the onset of diarrhea shorter than eight hours was an early marker of poor prognosis (18). Later on, independent of any other complication (encephalopathy and renal insufficiency in particular), a decrease in prothrombin index below 10% of normal (INR of about six) four days or more after ingestion had an excellent accuracy for predicting a poor outcome. Female sex was significantly associated with a poor outcome. MANAGEMENT General Management Whatever the initial intensity of the manifestations, patients with A. phalloides intoxication should be promptly referred to a specialized unit, in the vicinity of a liver transplant center. The initial phase of the management consists in fluid resuscitation. Once patients experience vomiting and profuse diarrhea, the aim is to prevent vascular collapse, superimposed ischemic liver cell necrosis and renal insufficiency. The administration of any potentially nephrotoxic or hepatotoxic agents as well as sedative drugs should be avoided. As long as patients experience profuse diarrhea, arterial blood pressure must be carefully monitored and adequate blood volume has to be maintained through appropriate fluid infusion. Systematic gastric lavage or induction of emesis by ipeca syrup is useless unless performed within the first hour after ingestion (17). Similarly, there is no evidence that oral administration of activated charcoal helps limit toxicity (17,21). During the second phase of the syndrome, as indicated above, a profound decrease in coagulation factors can be observed. Since, even in such circumstances, spontaneous bleeding complications are highly uncommon, systematic administration of fresh frozen plasma is not justified. In addition, the administration of fresh frozen plasma would hamper the interpretation of the key prognostic marker represented by the changes in coagulation profile over time. In patients with massive liver cell necrosis, the occurrence of acute renal failure and/or severe metabolic acidosis is frequent. Renal replacement therapy (intermittent dialysis or continuous hemodiafiltration) is the only efficient means for controlling severe acidosis. “Specific” Treatments For more than 30 years, several agents considered to represent “specific” treatments or “antidotes” have been used in patients with Amanita toxin poisoning. These so called “specific” treatments include beta-lactam antibiotics (penicillin G in particular), silymarin, thioacetic acid, L-ascorbic acid (vitamin C) and cimetidine. In a large review, it was shown that the proportion of reported patients receiving one or several of these agents in combination was as high as 80.9% (17). A rationale, relying on protective effects in experimental models, exists for beta-lactam antibiotics (22) and silymarin (23,24) while the use of other agents is empirical (Table 1). Among TABLE 1 Rationale for Use and Efficacy of the Specific Therapies for Amanita phalloides Poisoning Treatment Beta-lactam antibiotics Silymarin Thioacetic acid L-ascorbic acid Cimetidine

Rationale for use Hepatoprotective Hepatoprotective Antioxydant Antioxydant Cytochrome P450 inhibitor

Efficacy in experimental models

Evidence for efficacy in clinical practice

Yes (22) Yes (23,24) No No No

No No No No No

Mushroom Poisoning

729

these agents, silymarin has been associated with the best outcomes. This observation, however, is considerably weakened by the impossibility to adjust the outcome according to baseline prognostic markers as well as by the absence of control groups. Nonetheless, silymarin has an excellent safety profile. Overall, it is doubtful that any significant benefit can be expected from these agents in case of acute liver failure. As indicated above, TNF-induced apoptosis may play an important role in liver cell damage after amatoxin intoxication (10). Whether TNF antagonists could help prevent liver cell damage in this context is unknown. N-acetyl cysteine is widely used in patients with acute liver failure, including those with non-paracetamol causes. The rationale for using N-acetyl cysteine is that it helps restore glutathione stores. Glutathione is a free radical scavenger which limits the toxicity of paracetamol metabolites. N-acetyl cysteine has been empirically used in about 5% of previously reported patients with A. phalloides poisoning (17). Whether it improved the outcome is uncertain. Extracorporeal “Detoxification” Theoretically, extracorporeal detoxification systems based on hemodialysis or hemofiltration could be helpful since amatoxins are small, easily dialyzable, molecules. However, the benefits expected from blood detoxification is limited since amatoxins are only detectable in the serum in the very early period following ingestion (24–28 hours) and serum concentration is low compared to urine concentration (7,25,26). Hemoperfusion across charcoal cartridges has been used on the grounds that amatoxins have a high affinity for activated charcoal, without evidence for any superiority compared to conventional hemodialysis (27). More recently, cases of patients successfully treated with the molecular adsorbent recirculating system (MARSw) have been reported (28–30). MARS system includes anion exchanger and charcoal cartridges. It also offers the advantage of a dialysate enriched in albumin. Accordingly, MARS could be more effective for removing some circulating proteinbound toxins compared to conventional systems. However, circulating amatoxins are not bound to proteins. Additionally, as indicated above, the concentration of circulating amatoxin is very low and no massive detoxification can be expected from any of these systems, including MARS. Finally, the fact that some patients treated with MARS recovered after A. phalloides intoxication is not evidence that this technique had a significant impact on survival. Past experience shows that a majority of patients survived without any extracorporeal support. Overall, detoxification systems aimed at removing amatoxins (namely, hemodialysis, hemofiltration, plasmapoheresis, and MARS) can only be useful if initiated very early (within the first 24 hours) after ingestion. Afterwards, extracorporeal systems should rather be viewed as means for correcting the consequences of A. phalloides intoxication (acute renal failure, metabolic acidosis, hyperkalemia) rather than limiting the extent of liver damage by removing toxins. Liver Transplantation Emergency liver transplantation is the only option in patients who develop life threatening complications in association with profound and persistent liver insufficiency, encephalopathy, brain edema, and multiorgan failure. A number of case reports or small series have been reported, most of them being successful (31–39). However, it can be suspected that unsuccessful transplantations have been underreported. In the absence of large series, it remains difficult to establish clear cut criteria for deciding emergency liver transplantation. Establishing criteria for emergency transplantation is even more difficult since improvement leading to complete recovery as well as deterioration can occur within especially short time periods. Noteworthy, once patients develop encephalopathy, a key prognostic variable according to Clichy’s and King’s College criteria, there is very little time for bridging them to transplantation (19,20). Ganzert’s criteria, namely, prothrombin index below 25% of normal and serum creatinine over 106 mmol/L from day three after ingestion, represent an interesting alternative to more general criteria for transplantation (16). However, as indicated in the prognosis section, they may have pitfalls. In a retrospective study, we found that King’s college criteria were more efficacious than Clichy’s and Ganzert’s criteria for

730

Durand and Valla

TABLE 2 Proposed Guidelines for Deciding Transplantation in Patients with Amanita phalloides Poisoning Assessment at admission Strongly consider that liver transplantation may be necessary if the interval between ingestion and the onset of diarrhea is below 8 hoursa The need for transplantation is unlikely if the interval between ingestion and the onset of diarrhea exceeds 8 hours Post-admission criteria Listing for emergency transplantation if King’s College criteria are met and/or prothrombin index is lower than 10% of normal (INRz6) 4 days or more after ingestion a

Especially in female patient. Source: From Ref. 18.

deciding transplantation (18). Practical guidelines for deciding emergency transplantation are proposed in Table 2. Multiorgan failure does not represent a contraindication for transplantation, even though it is a source of substantial postoperative morbidity. Most patients are transplanted with a significant deterioration of renal function and most of them need transient extracorporeal renal support in the early post transplantation course. There is no clear evidence that toxicity recurs post operatively on the transplanted liver. Even if emergency transplantation can be a life saving option, it leads to lifelong immunosuppression and potentially exposes to severe side effects in the long term. An alternative option is to perform auxiliary transplantation. This procedure consists in performing partial hepatectomy in the recipient and transplanting a partial liver graft. While the transplanted graft rapidly corrects liver failure, the remnant part of the native liver can undergo delayed regeneration, offering the opportunity to stop immunosuppression definitely. A. phalloides poisoning belongs to the list of causes of acute liver failure with a potential for complete regeneration of the native liver after auxiliary transplantation. Observations of successful auxiliary transplantation with discontinuation of immunosuppression have been reported (18,40–42). Interestingly, complete regeneration may even occur in patients found to have near total liver cell necrosis on the explanted native liver. Even if attractive, auxiliary transplantation remains technically complex and should only be applied in selected patients without multiorgan failure. CONCLUSIONS Amanita mushroom poisoning is a model of toxin-induced centrilobular necrosis. It often results in massive hepatocyte necrosis leading to hyperacute liver failure. Patients with acute liver failure following A. phalloides intoxication have a high mortality rate. There is no clear evidence that any pharmacological agent or any extracorporeal detoxification system helps prevent the extent of liver cell necrosis. Therefore, the management essentially relies on symptomatic care and the prevention of external aggravating factors. Emergency liver transplantation is the only option in patients who do not experience rapid regeneration and develop encephalopathy and/or multiorgan failure. In selected patients, auxiliary transplantation can be an option, with a potential for delayed regeneration and immunosuppression withdrawal. Although A. phalloides is a rare cause of acute liver failure, education about mushrooms characteristics and toxicity is still needed to prevent accidental ingestion. ACKNOWLEDGMENT The authors are indebted to Dr Vale´rie Paradis for her contribution to the preparation of pathological illustrations. REFERENCES 1. Diaz JH. Evolving global epidemiology, syndromic classification, general management, and prevention of unknown mushroom poisonings. Crit Care Med 2005; 33:419–26. 2. Karlson-Stiber C, Persson H. Cytotoxic fungi—an overview. Toxicon 2003; 42:339–49.

Mushroom Poisoning

731

3. Cole FM. Amanita phalloides in Victoria. Med J Aust 1993; 158:849–50. 4. Chaiear K, Limpaiboon R, Meechai C, et al. Fatal mushroom poisoning caused by Amanita virosa in Thailand. Southeast Asian J Trop Med Public Health 1999; 30:157–60. 5. Vetter J. Toxins of Amanita phalloides. Toxicon 1998; 36:13–24. 6. Faulstich H. Mushroom poisoning. Lancet 1980; 2:794–5. 7. Jaeger A, Jehl F, Flesch F, et al. Kinetics of amatoxins in human poisoning: therapeutic implications. J Toxicol Clin Toxicol 1993; 31:63–80. 8. Letschert K, Faulstich H, Keller D, et al. Molecular characterization and inhibition of amanitin uptake into human hepatocytes. Toxicol Sci 2006; 91:140–9. 9. Wieland T. The toxic peptides from Amanita mushrooms. Int J Pept Protein Res 1983; 22:257–76. 10. Leist M, Gantner F, Naumann H, et al. Tumor necrosis factor-induced apoptosis during the poisoning of mice with hepatotoxins. Gastroenterology 1997; 112:923–34. 11. Faulstich H, Zobeley S, Trischmann H. A rapid radioimmunoassay, using a nylon support, for amatoxins from Amanita mushrooms. Toxicon 1982; 20:913–24. 12. Pastorello L, Tolentino D, D’Alterio M, et al. Determination of alpha-amanitin by high-performance liquid chromatography. J Chromatogr 1982; 233:398–403. 13. Mas A. Mushrooms, amatoxins and the liver. J Hepatol 2005; 42:166–9. 14. Koppel C. Clinical symptomatology and management of mushroom poisoning. Toxicon 1993; 31:1513–40. 15. Rengstorff DS, Osorio RW, Bonacini M. Recovery from severe hepatitis caused by mushroom poisoning without liver transplantation. Clin Gastroenterol Hepatol 2003; 1:392–6. 16. Ganzert M, Felgenhauer N, Zilker T. Indication of liver transplantation following amatoxin intoxication. J Hepatol 2005; 42:202–9. 17. Enjalbert F, Rapior S, Nouguier-Soule J, et al. Treatment of amatoxin poisoning: 20-year retrospective analysis. J Toxicol Clin Toxicol 2002; 40:715–57. 18. Escudie´ L, Francoz C, Vinel JP, et al. Amanita phalloides poisoning: reassessment of prognostic factors and indications for emergency liver transplantation. J Hepatol 2007; 46:466–73. 19. Bernuau J, Goudeau A, Poynard T, et al. Multivariate analysis of prognostic factors in fulminant hepatitis B. Hepatology 1986; 6:648–51. 20. O’Grady JG, Alexander GJ, Hayllar KM, et al. Early indicators of prognosis in fulminant hepatic failure. Gastroenterology 1989; 97:439–45. 21. Toxicologists. AAoCTEAoPCaC. Position statement: single dose activated charcoal. Clin Toxicol 1997; 35:721–41. 22. Tamai I, Terasaki T, Tsuji A. Evidence for the existence of a common transport system of beta-lactam antibiotics in isolated rat hepatocytes. J Antibiot (Tokyo) 1985; 38:1774–80. 23. Floersheim GL. Treatment of experimental poisoning produced by extracts of Amanita phalloides. Toxicol Appl Pharmacol 1975; 34:499–508. 24. Vogel G, Tuchweber B, Trost W, et al. Protection by silibinin against Amanita phalloides intoxication in beagles. Toxicol Appl Pharmacol 1984; 73:355–62. 25. Mydlik M, Derzsiova K, Klan J, et al. Hemoperfusion with alpha-amanitin: an in vitro study. Int J Artif Organs 1997; 20:105–7. 26. Mercuriali F, Sirchia G. Plasma exchange for mushroom poisoning. Transfusion 1977; 17:644–6. 27. Splendiani G, Zazzaro D, Di Pietrantonio P, et al. Continuous renal replacement therapy and charcoal plasmaperfusion in treatment of amanita mushroom poisoning. Artif Organs 2000; 24:305–8. 28. Covic A, Goldsmith DJ, Gusbeth-Tatomir P, et al. Successful use of Molecular Absorbent Regenerating System (MARS) dialysis for the treatment of fulminant hepatic failure in children accidentally poisoned by toxic mushroom ingestion. Liver Int 2003; 23(Suppl. 3):21–7. 29. Rubik J, Pietraszek-Jezierska E, Kaminski A, et al. Successful treatment of a child with fulminant liver failure and coma caused by Amanita phalloides intoxication with albumin dialysis without liver transplantation. Pediatr Transplant 2004; 8:295–300. 30. Faybik P, Hetz H, Baker A, et al. Extracorporeal albumin dialysis in patients with Amanita phalloides poisoning. Liver Int 2003; 23(Suppl. 3):28–33. 31. Montalti R, Nardo B, Beltempo P, et al. Liver transplantation in fulminant hepatic failure: experience with 40 adult patients over a 17-year period. Transplant Proc 2005; 37:1085–7. 32. Kucuk HF, Karasu Z, Kilic M, et al. Liver failure in transplanted liver due to Amanita falloides. Transplant Proc 2005; 37:2224–6. 33. Burton JR, Jr., Ryan C, Shaw-Stiffel TA. Liver transplantation in mushroom poisoning. J Clin Gastroenterol 2002; 35:276–80. 34. Woodle ES, Moody RR, Cox KL, et al. Orthotopic liver transplantation in a patient with Amanita poisoning. JAMA 1985; 253:69–70. 35. Meunier BC, Camus CM, Houssin DP, et al. Liver transplantation after severe poisoning due to amatoxin-containing Lepiota—report of three cases. J Toxicol Clin Toxicol 1995; 33:165–71. 36. Klein AS, Hart J, Brems JJ, et al. Amanita poisoning: treatment and the role of liver transplantation. Am J Med 1989; 86:187–93.

732

Durand and Valla

37. Pawlowska J, Pawlak J, Kaminski A, et al. Liver transplantation in three family members after Amanita phalloides mushroom poisoning. Transplant Proc 2002; 34:3313–4. 38. Broussard CN, Aggarwal A, Lacey SR, et al. Mushroom poisoning—from diarrhea to liver transplantation. Am J Gastroenterol 2001; 96:3195–8. 39. Galler GW, Weisenberg E, Brasitus TA. Mushroom poisoning: the role of orthotopic liver transplantation. J Clin Gastroenterol 1992; 15:229–32. 40. Rosenthal P. Auxiliary liver transplantation for toxic mushroom poisoning. J Pediatr 2001; 138:449–50. 41. Chenard-Neu MP, Boudjema K, Bernuau J, et al. Auxiliary liver transplantation: regeneration of the native liver and outcome in 30 patients with fulminant hepatic failure—a multicenter European study. Hepatology 1996; 23:1119–27. 42. Jaeck D, Boudjema K, Audet M, et al. Auxiliary partial orthotopic liver transplantation (APOLT) in the treatment of acute liver failure. J Gastroenterol 2002; 37(Suppl. 13):88–91.

34

Hepatotoxicity of Herbal Medicines, Vitamins, and Natural Hepatotoxins Lawrence U. Liu and Thomas D. Schiano Division of Liver Diseases and Recanati/Miller Transplantation Institute, The Mount Sinai Medical Center, New York, New York, U.S.A.

INTRODUCTION This chapter discusses the increasing use of complementary and alternative medicine (CAM), vitamins, and natural hepatotoxins and why hepatotoxicity has become a major problem associated with these treatments. Diagnosis, treatment, and the pathophysiology of hepatotoxicity from these agents are outlined. Examples of specific hepatotoxic products are cited, and drug–product interactions pertinent to patients having chronic liver disease are reviewed.

BACKGROUND There has been a growing trend in Western countries to use alternative medicine, particularly herbal preparations, to treat a range of medical maladies. It has been estimated that 36% of all adults in the United States use some form of herbal products (1). Commonly used herbs in the United States not having direct hepatotoxicity and reasons for their use are shown in Table 1 (2). Because herbal products are not marketed as drugs, they are not subjected to rigorous safety and efficacy testing. Herbal supplements are regulated by the Dietary Supplement Health and Education Act, which was passed by the U.S. Congress in 1994. With the exception of new dietary ingredients, dietary supplements need neither Food and Drug Administration (FDA) approval before marketing nor the registration of their products or manufacturer with the FDA. This regulatory approach is in contrast to the drug approval process in which both proven efficacy and safety must be established in humans (3). Their purchase of these products by mail order and in health food and vitamin stores carries an implicit assumption of safety, and because they are considered natural and are available without a prescription, many users ignore the potential for toxicity (4,5). Many herbal remedies have not been submitted to rigorous scientific testing and are largely justified by prescribers via trial-and-error experience. Within individual disciplines, CAM treatment regimens may also vary considerably between practitioners (6). In addition to the lack of demonstrable proof of efficacy, the ingredients of these preparations are not rigorously substantiated. Thus, toxicity may or may not be caused by the products of the specific plants or herbs listed as ingredients. For example, milk thistle is a common folk medicine for treatment of liver disease and devoid of hepatotoxicity, but it may be combined with other potentially hepatotoxic substances. Risks of toxicity are particularly high when therapeutic preparations contain numerous different plants or herbs, which is more often than not the case (4). One ingredient may potentiate another through the induction of certain hepatic enzymes or the formation of secondary toxic metabolites because of alterations in pharmacokinetics or pharmacodynamics. Other factors potentially contributing to the hepatotoxicity of herbal preparations include adulteration of the product by various chemicals and prescription medications, improper storage, misidentification of the ingredients, and mislabeling of the final product. There may be batch-to-batch variation between

734

Liu and Schiano

TABLE 1 Non-Hepatotoxic Medicinal Plants Currently Popular in the United States and Their Indications Aloe Chamomile Cranberry (Vaccinium oxycoccus) Dandelion root Feverfew Garlic (Allium sativum) Ginger Ginkgo biloba (Maidenhair tree) Ginseng Goldenseal (Hydrastis canadensis) Peppermint (Mentha piperita) St. John’s wort Saw palmetto (Serenoa repens)

Wound healing Insomnia, antiseptic Urinary tract infections Diuretic Arthritis, migraines Cholesterol-lowering, antihypertensive Concentration, increase stamina Memory improvement, claudication Stimulant Digestive disorders Analgesic Antidepressant Diuretic

Source: Data from Ref. 2.

manufacturers in their preparation of certain herbal medicinals (7). Plants can also vary based on the geographic location of their growth and the time of the harvest (8). The initial processing and storage condition of plant material also affects its purity. Patients should be aware that an aqueous extract would not contain the same amount of a herb as an alcohol extract, or as the ground leaves of the herb without extraction. Identification of all the herbs present in a combination product may be made difficult because many plants are marketed under various different names. For example, Echinacea is also known as cone flower, black sampson, black susan, comb flower, scurvyroot, hedgehog, Missouri snakeroot, and Indian head (9). It is often difficult to demonstrate the hepatotoxicity of herbal remedies. The patient usually omits mentioning the use of herbal products during medical evaluation because of their perceived safety of use. Eisenberg et al. (10) reported that 72% of patients using unconventional therapies failed to inform their physicians that they were doing so. Use of CAM seems more common among women, patients with a college education, and those with a higher income (7). Furthermore, only about 10% of adverse events associated with herbal remedies are reported (11). For all these reasons, herbal remedy-associated liver injury commonly can go unnoticed. For example, a Swedish health control laboratory showed that of 15 patients investigated daily, one or more patients regularly had abnormal liver chemistry tests that normalized at the time they stopped taking herbal preparations (12). Use of CAM is common and popular in persons with chronic liver disease, especially in patients with hepatitis C (7). Its use has also been reported among liver transplant recipients; a case of allograft rejection secondary to colostrum ingestion has recently been reported (13). This is because herbal medicines receive wide coverage in the media and their use raises patient expectations at the time medical treatment is not always effective or is difficult to tolerate, such as with interferon-based therapies. A more candid discourse between physician and patient about herbal medicines will prevent individuals from using potentially hepatotoxic herbs or using these approaches instead of prescription medications having excellent response rates and life-saving potential (e.g., corticosteroids in the treatment of autoimmune hepatitis). Patients should not be dissuaded from being treated by CAM practitioners, but should avoid “doctor shopping.” They should be warned that advice given in health food stores can at times be dangerously misleading (14). Direct communication with the CAM practitioner to better understand what herbal preparation the patient is receiving is recommended. Overall, it is important for the hepatologist to keep an open mind with regard to CAM, and to inquire about them in as nonthreatening and non-pejorative a manner as possible, because it will help consolidate the patient–physician relationship (7). Ongoing research regarding the use of CAM has been carried out largely by the National Center for Complementary and Alternative Medicine and the Cochrane Database. The current practice of CAM often means that it is difficult to recruit adequate numbers of patients to investigate the effects of individual agents (15).

Herbal Medicines, Vitamins, and Natural Hepatotoxins

735

DIAGNOSIS AND TREATMENT As with any drug-induced liver injury, early suspicion of herbal or botanical hepatotoxicity is paramount because morbidity is greatly increased if the product is continued after liver enzyme test abnormalities appear or clinical symptoms develop (16,17). The use of these products should be a part of any clinical history taking when considering obscure cases of acute or chronic liver disease. For example, Echinacea, if used continuously for more than eight weeks, may cause hepatotoxicity (18). The first clinical manifestations of hepatotoxicity often can be nonspecific with constitutional symptoms occurring much earlier than jaundice. A detailed history of all preparations used, duration of therapy, and use of concomitant drugs is essential (16,17). Many herbal products have a good safety record, with hepatotoxicity occurring at the time recommended dose thresholds are exceeded. At times, it can be helpful to examine the packaging of the herbal products that have been ingested, review the often long list of ingredients of some of these preparations, and to contact a poison control center. Other causes of liver disease, however, must be ruled out by careful assessment of clinical, radiologic, biochemical, and serologic testing. The possibility of botanical or herbal hepatotoxicity superimposed on preexisting liver disease must always be considered, especially because many herbal remedies are used as liver “remedies” in patients with hepatitis C and other chronic liver conditions. Patients with chronic liver disease use many of the herbs listed in Table 1 in addition to frequently using Astragali radix (Astragalus), milk thistle (Silybum marianum), Curcuma longa (turmeric), Schizandra root, melatonin, Phyllanthus amarus, and SAMe (s-adenosyl-methionine), none of which have documented hepatotoxicity. Herbal- and botanical-induced liver injury can clinically and histologically mimic many different types of primary liver disease (Table 2) (8). Some uncommon histological patterns of liver injury, however, should arouse the suspicion of herbal hepatotoxicity, including zonal necrosis, necrotic lesions with steatosis or bile duct injury, and vascular injury, especially veno-occlusive disease, now known as sinusoidal obstruction syndrome (SOS) (8,19). In a series of 20 patients with fulminant hepatic failure (FHF), Estes et al. (20) found a high prevalence of potentially hepatotoxic dietary supplement usage. Seven (35%) of these patients had no other identifiable risk factor for developing liver failure. More frequently though, the histological appearance is nonspecific. The main treatment of herbal hepatotoxicity is withdrawal of the offending agent (16,17). Once hepatotoxicity has occurred, continued use of the herbal product may result in cirrhosis or acute liver failure. Physicians should caution their patients not to take any herbal remedy without first consulting them, to be cognizant of the signs of liver disease, and to know that herbal products are not always as innocuous as perceived. Patients who are known to use multiple vitamin supplements, herbal preparations, and actively seek out alternative medicine approaches should have any products that they regularly use periodically reviewed and the ingredients recorded.

HEPATOTOXICITY OF SPECIFIC HERBAL, COMPLEMENTARY, AND ALTERNATIVE MEDICINES Pyrrolizidine Alkaloids The hepatotoxicity of these alkaloids, which are found in more than 350 plant species worldwide, has been well recognized for over 70 years. These alkaloids are widely distributed, botanically and geographically. Pyrrolizidines of the Senecio, Heliotropium, Crotalaria, and Symphytum (comfrey) genera have been noted to be especially toxic. Pyrrolizidine poisoning is endemic in areas such as Jamaica and many parts of Africa, where the alkaloids are most often ingested as herbal tea (21). Sporadic cases are still being observed in Western countries (5,8). Epidemic intoxications from accidentally contaminated flour have been noted in India and Afghanistan (22). Of note, certain species of Echinacea contain pyrrolizidine alkaloids (23). Comfrey (Symphytum officinale) has been used in the treatment of inflammatory conditions, e.g., arthritis, gout, thrombophlebitis, and diarrhea. Its toxic effects on the liver appear to be due to various pyrrolizidine alkaloids, such as lasiocarpine and symphytine,

736

Liu and Schiano

TABLE 2 Histopathological Patterns of Complementary and Alternative Medicine Hepatotoxicity Histopathological pattern Acute hepatitis Autoimmune hepatitis Chronic hepatitis, C/K fibrosis

Cirrhosis

Cholestatic hepatitis

Fulminant hepatic failure

Giant cell hepatitis Massive hepatic necrosis

Microvesicular steatosis Sinusoidal obstruction syndrome Zonal necrosis

Additional data provided by the authors. Source: Data adapted from Ref. 8.

Hepatotoxin Carp gall bladder Cycad Lipokinetix Black cohosh Ma Huang Syo-saiko-to Germander Greater celandine Jin Bu Huan Prostata Syo-saiko-to Chaparral Cycad Germander Greater celandine Iron Jin Bu Huan Vitamin A (Retinol) Ackee Ba Jiao Lian Cascara sagrada Chaparral Greater celandine Green tea Jin Bu Huan Kava Morinda Shou Wu Pian Syo-saiko-to Teucrium capitatum Atractylis gummifera Bacillus cereus Black cohosh Callilepsis laureola (Impila) Chaparral Chinese herbal medicines Cocaine Germander Green tea Kava Lipokinetix Pennyroyal Teucrium polium Usnic acid Isabgol Aflatoxin Chaparral Cocaine Cyanobacteria Germander Greater celandine Iron Kava Ma Huang Pennyroyal Bacillus cereus Margosa oil Syo-saiko-to Pyrrolizidine alkaloids Skullcap Atractylis gummifera Callilepsis laureola Cocaine Germander Jin Bu Huan Pennyroyal oil Syo-saiko-to Teucrium polium

Herbal Medicines, Vitamins, and Natural Hepatotoxins

737

and their related N-oxides. The mechanisms of hepatotoxicity are still not completely understood, and may be related to the biotransformation of alkaloids by hepatic microsomal enzymes, thereby producing highly reactive pyrroles that act as powerful alkylating agents (24). The main liver injury induced by pyrrolizidine alkaloids is SOS. It was first publicized in the 1950s as a disease of Jamaican children who drank bush teas made from plants of the Senecio and Crotalaria species as a folk remedy for acute illness (25). SOS may take the form of acute, subacute, or chronic liver injury, all characterized by abdominal distention, hepatomegaly, ascites, and edema. In the acute form, abdominal pain occurs suddenly, sometimes with markedly elevated aminotransferases and conjugated hyperbilirubinemia (26). Histopathologic findings include nonthrombotic occlusion of the lumen of small terminal hepatic venules, leading to hepatic congestion and, ultimately, hemorrhagic centrilobular necrosis. There also may be perivenular congestion with sinusoidal distention and dilatation (5,26). In the subacute and chronic forms, there is central fibrosis and bridging extending from the perivenular areas, insidiously culminating in cirrhosis and portal hypertension (25). The prognosis of the acute form of SOS is characterized by a rapidly fatal course in 15% to 20% of patients, with the prognosis worse in adults as compared with children. There seems to be complete recovery in one-half of patients. Approximately 15% of patients with acute disease develop subacute or chronic injury and die from decompensated liver disease within several years. Many of the remaining patients ultimately develop cirrhosis and portal hypertension (17,25,27). The hepatotoxicity of pyrrolizidine alkaloids is reproducible and dose dependent in laboratory animals. Hepatotoxicity seems to be related to the biotransformation by cytochrome P450 of unsaturated alkaloids into unstable toxic metabolites that may serve as alkylating agents. Liver injury seems to depend on what type of pyrrolizidine alkaloid and the total dose that is ingested along with the susceptibility of the person to the alkaloid. Acute SOS may result from a short exposure to high doses of alkaloids, whereas chronic lesions appear to be caused by prolonged exposure to small doses of alkaloids (28–30). Chaparral Chaparral leaf (Larrea tridentata) is prepared by grinding the leaves of the evergreen desert shrub known as the creosote bush or greasewood. The leaves can be brewed for tea or made into capsules or tablets. It grows in the southwestern United States and northern Mexico and was traditionally used by Native Americans for the treatment of a variety of ailments (5,31). It has been espoused as a free radical scavenger to retard aging and is also used in burn salves, weight loss teas, liver tonics, and as treatment for chronic skin conditions. In addition, it is perceived to have antifungal, antibacterial, and antiviral properties by way of its active ingredient nordihydroguaiaretic acid, a potent and selective inhibitor of cycloxygenase and lipoxygenase pathways. Chaparral leaf was often used as one of several natural products in alternative medicine approaches by patients with HIV infection (32). There have been multiple reports of chaparral-induced hepatotoxicity. Sheikh et al. (33) reviewed 18 case reports of adverse events associated with the ingestion of chaparral reported to the FDA between 1992 and 1994, and they found that in 13 of the cases there was evidence of hepatotoxicity. This was manifested by jaundice with a generally marked increase in total serum bilirubin and aminotransferases. Hepatitis occurred 3–52 weeks after chaparral ingestion, resolving 1–17 weeks after most individuals discontinued its use. In four patients, there was progression to cirrhosis, and in two patients, FHF developed, necessitating liver transplantation. Histologic findings on liver biopsy ranged from cholangitic changes and cholestasis to massive hepatocellular necrosis (33). Hepatic fibrosis after eight weeks of daily ingestion of chaparral tablets has been reported (34). The mechanism of chaparral’s hepatotoxicity is unknown but may be related to nordihydroguaiaretic acid, which itself was removed by the FDA in 1968 from the “generally recognized as safe” list for use in foods, because of the results of animal toxicity testing. Besides hepatotoxicity, renal and dermatologic toxicity have also been associated with the ingestion of chaparral leaf (33).

738

Liu and Schiano

Kava Kava is a popular supplement derived from the pepper plant (Piper methysticum) (18). Other frequently used names for kava include ava, awa, kava root, kava kava, P. methysticum, kava kew, sakau, tonga, rauschpfeffer, wurzelstock, or yangora. Kava has gained popularity as a herbal preparation due to its anxiolytic and sedative properties (35). The psychoactive activity in kava is contained in a series of compounds called kavalactones. Kava has a high potential for causing herb–drug interactions through inhibition of CYP450 enzymes responsible for the majority of the metabolism of pharmaceutical agents. Genetic polymorphism of many CYP enzymes, leading to interindividual variation in drug metabolism, may be another important factor in the marked variability in hepatotoxic response (36). Recent reports of possible hepatotoxicity from commercially prepared kava raised concerns in the medical community which forced several countries, including the United States, to either ban these products or to issue health advisories. Despite these actions, the traditional kava beverage continues to be consumed in the Pacific Islands without reports of liver injury. The International Kava Conference, held in Suva, Fiji, in 2004, was conducted to identify the origins of the reported hepatotoxicity associated with kava use. Various hypotheses proposed for the etiology of hepatotoxicity include the method of extraction (i.e., commercial preparations use organic solvents whereas the traditional beverage is a water infusion), the use of bark and other parts of the plant that may contain toxic substances (e.g., pipermethystine), preexisting liver and kidney dysfunction, overdosage, interactions between kava and other drugs, and genetic differences that may impact the rate of kavalactone metabolism (36). Chinese and Other Asian Herbal Medicines and Teas Traditional Chinese herbal medicines are widely used by Asian communities throughout the world. Over 7000 species of medicinal plants are used in China, and these are finding everincreasing popularity in the United States, especially in the treatment of eczema and atopic dermatitis (8,37,38). Some Chinese medicinal plants have well-known toxicities. For example, Aconitum kusnezoffii (caowu root) and Aconitum carmichaeli (chuanwa root), both used for rheumatic complaints, can cause severe neurologic and gastrointestinal symptoms along with fatal ventricular arrhythmias (39). Although most Chinese medicinal preparations appear safe, Melchart et al. (40) noted that 1% of patients in a German population using traditional Chinese medicines had elevated liver chemistry tests; there was a greater risk for this with the use of formulas containing Glycyrrhizae radix and Atractylodis macrophalae. Numerous reports of hepatitis (41,42) and FHF (43) have been ascribed to the use of Chinese herbal medicinals. Most of these medicinal preparations have multiple herbal ingredients; the actual hepatotoxic constituent remains unidentified. The identification of singular hepatotoxic herbal ingredients is difficult because these products and other Chinese proprietary medicines often are adulterated with substituted herbs and Western medicines (i.e., nonsteroidal anti-inflammatory agents, phenylbutazone, corticosteroids, and acetaminophen) (44). Huang et al. (45) showed that 24% of samples of traditional herbal medicines in Taiwan contained at least one conventional pharmacologic compound. Common plant species have been found in several of the hepatotoxic preparations, including Paeonia (red peony root), Gentianae (gentian root), Bupleurum (hare’s ear root), and most commonly, Dictamnus dasycarpus; no reproducible liver injury, however, has been found with any of these plants when used alone (37,41,42,46). Ma Huang (Ephedra) Ma Huang, a herbal product derived from plants of the Ephedra species, has been implicated in two cases of severe necroinflammatory hepatitis (47,48). As its active constituent is ephedrine, Ma Huang seems to increase metabolic activity and is used as an aid to weight loss (47,49). Recently, Neff et al. (50) described 12 cases of hepatotoxicity that were thought to be related to the ingestion of herbal weight loss compounds from various ingredients, including Ma Huang and usnic acid. Among these 12 subjects, eight recovered upon withdrawal of the herbal product. However, three required liver transplantation and one subject died.

Herbal Medicines, Vitamins, and Natural Hepatotoxins

739

Shou Wu Pian There have been two reports of acute cholestatic hepatitis induced by Shou Wu Pian, a herbal product derived from Polygonum multiflorum, which contains potentially hepatotoxic anthraquinones (51,52). The processed roots and vines of this plant are used as a tonic for a range of conditions that include dizziness, the premature graying of hair, and for liver problems. This again highlights the fact that some herbal products espoused to alleviate liver problems may cause liver dysfunction instead. Ba Jiao Lian Ba Jiao Lian (Dysosma pleianthum), one species in the May apple family, has been widely used in China for thousands of years as a general remedy and for the treatment of snakebite, weakness, condyloma acuminata, lymphadenopathy, and tumors. Textbooks of traditional Chinese medicine mention little about the toxicity of Ba Jiao Lian. Kao et al. (53) reported five patients who exhibited nausea, vomiting, diarrhea, abdominal pain, thrombocytopenia, leukopenia, abnormal liver chemistries, sensory ataxia, and altered consciousness after drinking infusions made with Ba Jiao Lian. These clinical manifestations were deemed to be consistent with podophyllum intoxication, podophyllotoxin being one of the main ingredients of the Ba Jiao Lian root. Jin Bu Huan Jin Bu Huan tablets (Lycopodium serratum), available as Jin Bu Huan Anodyne tablets, is a traditional herbal remedy used for more than 1000 years in China as a sleeping aid and analgesic. The hepatotoxicity of Jin Bu Huan has now been well characterized. Woolf et al. (54) described seven patients who developed acute hepatitis a mean of 20 weeks after taking appropriate doses of Jin Bu Huan tablets. Clinical hepatitis was associated with constitutional symptoms, pruritus and abdominal pain, with evidence of hepatomegaly and jaundice. Histologic examination demonstrated periportal focal necrosis, cholestasis, eosinophilic inflammation in one patient (there was peripheral eosinophilia that resolved with cessation of Jin Bu Huan), mild-to-moderate periportal fibrosis, lymphocytic inflammation with focal hepatocellular necrosis, and microvesicular steatosis in another. Aminotransferase elevation was marked in some patients with only minor abnormalities in serum alkaline phosphatase. The hepatitis resolved in most patients within a mean of eight weeks. Decreasing the dose of Jin Bu Huan improved liver chemistry test abnormalities in one patient, and readministration of Jin Bu Huan led to abrupt recrudescence of hepatitis in two patients (54). The mechanism of Jin Bu Huan’s hepatotoxicity is unknown. The package insert of the incriminated Jin Bu Huan indicated that the tablets contained 70% starch and 30% levo-alkaloid from the plant Polygala chinensis (54). Analysis subsequently revealed the presence of levotetrahydropalmatine, an alkaloid present in the plant genera Stephania and Cordydalis, but not in Polygala. Levo-tetrahydropalmatine exhibits some structural similarity with hepatotoxic pyrrolizidine alkaloids. Direct hepatotoxicity of levo-tetrahydropalmatine is unproven, however, and the clinical presentation of the affected patients more closely resembled a hypersensitivity reaction and not the SOS typically seen with pyrrolizidine alkaloid liver injury. Levo-tetrahydropalmatine is structurally related to the berberine alkaloids (55) that have been associated with the development of kernicterus and death in preterm infants given Huang Lian (Coptis chinensis), a traditional Chinese medicine containing 7% to 9% berberine (56). Horowitz et al. (57,58) outlined similar hepatotoxicity in children. In the Jin Bu Huan used by their patients, ulcer pain was listed in the product insert as an indicated use. Although the FDA has limited regulatory authority over the manufacture and packaging of products sold as dietary supplements, the unsubstantiated medical claims advertised in these cases made the product subject to FDA regulation as a drug and enabled the FDA to ban the legal importation of Jin Bu Huan Anodyne into the United States. Breynia Breynia officinalis has the Chinese proprietary name, Chi R Yun. It has been used, in combination with other traditional Chinese medicines, to treat contusions, heart failure, venereal diseases, growth retardation, and conjunctivitis. Poisoning in humans involves the neurologic,

740

Liu and Schiano

gastrointestinal, hepatic, urinary, and respiratory systems. Hepatic injury is primarily hepatocellular rather than cholestatic. Two cases of poisoning were reported to the Poison Control Center in Taiwan; in both cases, the stems and roots of the plant were ingested (59). Dose-related toxic effects were seen in 14 of 19 patients who ingested a soup that was accidentally contaminated with B. officinalis stems. Clinical manifestations were mostly gastrointestinal (e.g., diarrhea, nausea, vomiting, abdominal fullness) and elevated liver aminotransferases; marked jaundice did not develop. Peak liver chemistries were reached between three and 12 days post-ingestion and the majority resolved within 6 months (60). Syo-saiko-to Syo-saiko-to (Xiao Chai Hu Tang, Da Chai Hu Tang, TJ-9) is an Asian herbal medicine containing Bupleuri root, Pinelliae tuber, Scutellaria root, ginseng root, ginger rhizome, glycyrrhiza root, and jujube fruit (61,62). It has been used mainly in China and Japan for hundreds of years to treat common cold symptoms and fever, and has recently become popular as a treatment for chronic liver disease, especially hepatitis C. Unlike milk thistle, which has no reported associated hepatotoxicity, there have been several reports of liver injury occurring in patients taking Syo-saiko-to. Itoh et al. (61) reported four patients who had worsening of abnormal serum aminotransferases while taking Syo-saikoto. These abnormalities resolved with cessation of the drug and recurred at the time the herbal product was re-administered. The investigators postulated that the Scutellaria (skullcap) fraction might be the likely hepatotoxic ingredient. Kamiyama et al. (63) reported a patient who had significant worsening of her autoimmune hepatitis after the administration of Da Chai Hu Tang. Germander The blossoms of wall germander, a plant of the Labiatae (Teucrium chamaedrys) family, have long been used for relieving fever and abdominal disorders and for supposed choleretic and antiseptic properties. Preparations of germander include herbal teas, capsules containing germander alone or combined with other herbs, and some medicinal liquors. As a traditional folk medicine, the plant generally was considered safe. Starting in 1986, germander was marketed and widely used in France for the treatment of obesity. Large-scale use soon led to the recognition of germander-related hepatotoxicity. Following reports of over 30 cases of liver injury, including several cases of fatal FHF, all germander capsules and teas were barred from sale in France in May 1992 (5,64). Most cases of hepatotoxicity involved middle-aged women who were ingesting germander at recommended doses (600–1600 mg/day). Generally, liver injury occurred after two months of treatment; there was marked elevation of the serum bilirubin and aminotransferases that almost always resolved within two to six months of discontinuing the germander product. In a few patients using germander for longer periods of time or in large doses, liver damage occurred insidiously with the initial insult manifesting as chronic hepatitis or cirrhosis (5,64). In affected patients accidentally reexposed to germander, the liver injury relapsed after a shorter latent period, suggesting an immunoallergic component to the toxicity. Histologic examination demonstrated centribolular hepatocellular necrosis and inflammation. The mechanisms of germander hepatotoxicity have now been well characterized and serve as an example of how complex the pathophysiology of herbal hepatotoxicity can be, and how various inherent patient characteristics may play a role in inducing the liver injury. Germander is a complex composition of glycosides, flavonoids, saponins, and furan-containing neoclerodane diterpenoids (65,66). These constituents are oxidized by cytochrome P450 into reactive metabolites, similar to those produced by other diterpenoid-like compounds (i.e., aflatoxin). Furano neoclerodane diterpenoids appear to covalently bind to proteins, deplete cellular glutathione and protein thiols, and cause plasma membrane blebbing and cell disruption in isolated rat hepatocytes (4,66). These toxic metabolites are partially inactivated by glutathione and can be enhanced by inducers of cytochrome P450. Crude fractionation has indicated that the furano neoclerodane diterpenoids are responsible for the liver injury. Since toxic doses of germander in mice are higher than those used in patients with germander-

Herbal Medicines, Vitamins, and Natural Hepatotoxins

741

induced hepatitis, the use of lower doses in humans may help explain why most patients tolerated germander for so long. Fau et al. (67) have demonstrated that the incriminated furano neoclerodane diterpenoids can induce apoptosis in isolated rat hepatocytes. Several biochemical changes, such as increased cytosol calcium and ATP concentration and glutathione depletion resulting in nucleosomal DNA fragmentation, are associated with this apoptotic change. Apoptotic cell death is markedly increased in rats fed a sulfur amino acid-deficient diet, which reduces hepatic glutathione levels (4). As germander hepatotoxicity was frequently noted in patients on weight control diets, it has been suggested that food restriction and thus glutathione reduction may have played a role in its clinic toxicity. DeBerardinis et al. (68) recently demonstrated that teucrin A (germander’s major neoclerodane diterpenoid) can trigger antibodies against human microsomal epoxide hydrolase on the plasma membrane of hepatocytes. This may help explain the immune component of germander hepatotoxicity, and is a unique example of a plant substance inducing human autoantibodies. Genetic predisposition (i.e., cytochrome P450 polymorphisms) may also need to be present to develop this immune-related toxicity. Herbs sharing the same genus as germander include Teucrium polium and Teucrium capitatum. T. polium is a traditional herbal medicine in the Mediterranean region used as a cicatrizant and as an anti-inflammatory, cholesterol-lowering, or antimicrobial agent. A case of acute liver failure secondary to T. polium use resulting in massive hepatic necrosis has been reported (69). T. capitatum is a medicinal plant that has been used to lower blood glucose and cholesterol levels. Acute hepatitis and jaundice were described in a 62-year-old man four months after continued intake of a tea containing this herb (70). Liver biopsy revealed changes consistent with acute hepatitis and bridging necrosis. After intake of this tea ceased, the patient recovered clinically and liver chemistry abnormalities returned to normal levels within nine weeks. Pennyroyal Pennyroyal (squawmint oil), a herb consisting of the leaves of either Hedeoma pulegoides or Mentha pulegium, contains pulegone and smaller amounts of other monoterpenes that are encountered in mint species. It is usually in the form of oil but is also ingested in the form of tablets, leaves, teas, and as essence of pennyroyal. Pennyroyal has been espoused for hundreds of years as an abortifacient and an agent to induce menses, despite a lack of proven efficacy and multiple instances of centrilobular hepatic necrosis and death in connection with its use. Pennyroyal also is used as a pesticide for controlling fleas on pets (8,71). Pulegone is well-known to cause hepatic, pulmonary, and central nervous system toxicity in rodents. Studies have shown that cytochrome P450-induced oxidative metabolites of pulegone (i.e., menthofuran) are responsible for pennyroyal’s toxicity (72,73). Pulegone and menthofuran may be further oxidized by cytochrome P450 to other hepatotoxins. These reactive oxidative metabolites further bind to target cell proteins, and the extent of this binding parallels the extent of cellular damage (71–73). In addition, pulegone forms reactive metabolites that deplete hepatic glutathione levels (55). Replacement of glutathione with N-acetylcysteine may theoretically provide protection from pennyroyal hepatotoxicity. Anderson et al. (71) reviewed 18 cases of toxicity occurring in patients ingesting at least 10 mL of pennyroyal oil. They also reported four additional patients, one of whom died from multiorgan failure and severe hepatocellular necrosis, and another who was successfully treated with N-acetylcysteine. In pennyroyal intoxication, patients typically develop severe gastrointestinal upset and central nervous system effects within one to two hours of ingestion of 10 mL or less of the oil. Fatal ingestions and severe hepatotoxicity appear to occur with ingestion of 15 mL or more of pennyroyal oil (71). Atractylis Gummifera and Callilepsis Laureola Intoxications of Atractylis gummifera (Mediterranean glue thistle) typically occur in Mediterranean countries, where 26 species have been identified. A. gummifera traditionally has been used as a purgative, emetic, diuretic, and antipyretic. A whitish-yellowish, glue-like substance is

742

Liu and Schiano

secreted by the plant and often used by children as chewing gum. Intoxication commonly occurs in the spring with ingestion of the root extracts in which the toxins are more concentrated. Intoxication also may result from confusion between the plant and a variety of wild artichoke (5,8). Initial symptoms include headache, vomiting, and abdominal pain within 24 hours of ingestion. Necroinflammatory hepatitis, severe hypoglycemia, and renal failure ensue, culminating in acute liver failure. Similar hepatotoxicity has been reproduced in rats and may be related to potassium atractylate and gummiferin, two compounds able to inhibit mitochondrial oxidative phosphorylation and the Krebs cycle (74,75). Metabolites may also trigger apoptosis by altering mitochondrial membrane permeability and releasing caspaseactivating proteases (76). Several cases of fatal FHF associated with renal tubular necrosis have been ascribed to Callilepsis laureola (Impila), a traditional Zulu plant medicine (8,77). The clinical features include an abrupt onset of abdominal pain, gastrointestinal disturbance, encephalopathy, profound hypoglycemia, and liver/renal failure. There is 90% mortality within five days. This plant contains compounds chemically related to potassium atractylate; the cause of its hepatotoxicity may be similar to that of A. gummifera. Greater Celandine Extracts of Chelidonium majus (greater celandine) are used in Europe to treat dyspepsia and gallstones. C. majus has also been used as part of a herbal mixture for treatment of irritable bowel syndrome. Benninger has described 10 women who developed acute hepatitis associated with greater celandine use (78). The course of the hepatitis ranged from mild to severe with ALT peaking at 1000 IU/L with marked cholestasis (alkaline phosphatase up to 450 IU/L) in half of these patients. Liver biopsy was consistent with drug hepatotoxicity in seven patients. After discontinuation of the herbal product, rapid recovery was observed in all patients and liver chemistry tests normalized within two to six months. Unintentional rechallenge led to a second flare of hepatic inflammation in one patient. The mechanism of the liver injury is unknown but an idiosyncratic reaction seems likely because there was no dose dependency and the latent period was variable. Presence of low-titer autoantibodies in most patients and eosinophilic infiltrates on liver biopsy suggests an immune mechanism, although these features can be a nonspecific response to any form of liver injury. Lipokinetix Lipokinetix is marketed as an over-the-counter dietary supplement for weight loss; it contains norephedrine hydrochloride, sodium usniate, caffeine, yohimbine hydrochloride, and 3,5-diiodo-thyronine. Favreau et al. (79) described seven patients who between July and December 2000, after the ingestion of Lipokinetix at doses according to the manufacturer’s recommendation, developed acute hepatitis, including one patient who developed fulminant liver failure complicated by cerebral edema. Jaundice occurred in five patients; five patients had used the supplement for one month or less. Liver chemistry tests normalized within four months of stopping Lipokinetix. Three different lot numbers of Lipokinetix capsules were analyzed by the FDA and found not to contain contaminants. The etiology of the hepatotoxicity is unknown; the FDA has issued warning letters regarding Lipokinetix to the manufacturer, the physicians, and the public. As a key ingredient in Lipokinetix, usnic acid is also a component of some crude drugs and supplements sold throughout the world, e.g., Cladonia and Kambala teas. Han et al. (80) demonstrated the effects of usnic acid in mouse hepatocytes; it was found to inhibit mitochondrial function directly, causing an increase in reactive oxygen species production and leading to subsequent cell death. Besides the report by Neff (50), Durazo et al. (81) reported a case of FHF in a 28-year-old female taking usnic acid for weight loss who eventually underwent liver transplantation. Review of the patient’s history and examination of her native liver did not reveal other causes for the organ failure.

Herbal Medicines, Vitamins, and Natural Hepatotoxins

743

Black Cohosh Black cohosh (Cimicifuga racemosa or Actaea racemosa) is a herb used primarily for the treatment of menopausal symptoms; Remifemin Menopause (GlaxoSmithKline Consumer) is a common brand containing this herb. Reported cases of liver disease stemming from its use include acute hepatitis (82), autoimmune hepatitis (83), and FHF (84); treatment modalities have included prednisone and liver transplantation. The direct mechanism of this herb’s hepatotoxicity remains unknown. Morinda There have been recent reports of potential liver toxicity with the use of Morinda citrofolia, available commercially as Noni juice or Noni preparation. Over the past decade, various preparations of Noni have been sold as products to promote health, including cancer prevention (85). Millonig et al. (86) first described a case of hepatotoxicity after a few weeks’ ingestion of Noni juice (Morinda citrifolia). Peak ALT level was 1995 IU/mL; other tests did not reveal an etiology for the acute hepatitis. Liver biopsy revealed acute hepatitis with mixed inflammatory infiltrates and numerous eosinophils in the portal tracts, along with zone 3 hepatocellular cholestasis, and Kupffer cells and PAS-positive histiocytes in the sinusoids. Liver dysfunction gradually returned to normal upon discontinuation of this herbal product. Two more cases were reported by Stadlbauer et al. (87), with one patient requiring liver transplantation. The most hepatotoxic components of the Noni plant are anthraquinones such as nordamnacanthal, moridone, and rubiadin, and may be the cause of liver injury.

OTHER HEPATOTOXIC HERBS AND BOTANICALS Cycasin, a glycoside found in the fruit of the cycad tree (Cycas circinalis and Cycas revoluta), is a potent hepatotoxin and hepatocarcinogen in experimental animals. Cycads are indigenous in the tropics and subtropics (17). One report from Miyako Island, Okinawa, described a small epidemic of acute hepatitis attributable to ingestion of cycad nuts during a famine. The indigenous population was forced to subsist on cycad nuts because all other agriculture was destroyed by serious weather conditions. Long-term follow-up of islanders showed a high death rate from cirrhosis and chronic liver disease. The mechanism of the purported hepatotoxicity is unknown. Populations that ingest small amounts of the cycad nut in their diets do not appear to suffer any adverse effects (88). Green tea (Camellia sinensis) is popularly taken for health benefits, including the lowering of cholesterol levels and the prevention of cancer. Gloro et al. (89) described fulminant hepatitis occurring in a 48-year-old woman three weeks after taking Exolise (a predominantly ethanolic dry extract of green tea) for weight reduction. The patient eventually underwent successful liver transplantation. Although rat studies have shown green tea (C. sinensis) to have hepatoprotective properties, the biological and medical properties of green tea are difficult to study since many of its components remain unknown. Recently, Bonkovsky (90) reviewed published reports on 10 subjects who developed hepatotoxicity in relation to C. sinensis use. Liver dysfunction was usually manifested by a mixed hepatocellular–cholestatic pattern, with all patients’ liver test results normalizing within four months after cessation of product use. The author recommended the avoidance of taking large amounts or concentrated preparations of C. sinensis due to possible adverse effects. Ackee Fruit Ackee (Blighia sapida), the national fruit of Jamaica, is a food staple in many Jamaican diets. The fruit splits open, while still on the tree, to reveal three glassy black seeds surrounded by a thick, oily, yellow aril. It is consumed either in its fresh state or in its canned form. The ackee tree is indigenous to West Africa where it is called ankye or ishin. It also grows in other West Indian islands, in Central America, and in Southern Florida. Illegal in the United States, a case of jaundice and the development of abnormal liver chemistries in a 27-year-old male chronically

744

Liu and Schiano

ingesting this fruit was reported; signs and symptoms resolved shortly after intake ceased (91). Currently, only two cases of ackee poisoning have been reported in the United States. Ackee is associated with Jamaican vomiting sickness and toxic hypoglycemic syndrome. Signs and symptoms include profound vomiting and convulsions that may progress to coma and death; profound hypoglycemia is a frequent laboratory abnormality. Unripe ackee contains hypoglycin A, a water-soluble liver toxin that induces hypoglycemia by inhibiting gluconeogenesis due to its limitation of cofactors (CoA and carnitine) that are essential for oxidation of long-chain fatty acids. Histopathology findings are depleted hepatic glycogen stores and fatty metamorphosis of the liver, kidney, and other organs. Potential risk factors for ackee poisoning include selection and cooking of unripe ackee, purchase of tampered and forcibly opened ackee fruit, and the reuse of water in which unripe ackee has been cooked. In Jamaica, the epidemiology of ackee poisoning has not been well characterized; the true incidence and mortality rates may be underreported (92). Margosa Oil Margosa oil is a fatty-rich extract of the seeds of the neem tree (Azadirachta indica), which is found in tropical parts of the world. The oil is used in treating asthma, intestinal parasitism, arthritis, and leprosy. It has been shown to cause a Reye’s syndrome-like illness among children in Malaysia and India. Demonstration of this substance’s effects in mouse livers showed rapid action within 30 minutes, with mitochondrial injury, loss of liver glycogen, and hepatocyte fat globule ingestion via endocytosis (93). Sinniah and Baskaran (94) reported vomiting, drowsiness, metabolic acidosis, polymorphonuclear leukocytosis, and encephalopathy developing in 13 infants within hours of ingesting margosa oil. Liver biopsy of one infant demonstrated pronounced fatty infiltration; electron microscopy revealed mitochondrial damage. Case reports of other hepatotoxic herbs and CAM are shown in Table 3 (8,19,95).

VITAMINS Iron Iron is essential for cellular functions; in excessive amounts, it is toxic to cells. Hepatotoxicity can occur with acute iron overdose and chronic ingestion. Hepatic necrosis in iron-induced hepatotoxicity occurs generally in zone 1. Progressive hepatic iron deposition leads to fibrosis and cirrhosis. Heavy iron overload, as seen in primary and secondary hemochromatosis, can cause fibrosis of the liver, heart, and pancreas. Cellular injury is induced by free radicals generated by iron that leads to peroxidation of lipids, proteins, and nucleic acids. The free radical mechanism of iron-mediated tissue damage depends on the availability of ferrous iron to promote free radical reactions (96). In the liver, lipid peroxidation leads to damage to hepatocellular components such as mitochondria and lysosomes (97). Hepatotoxicity caused by acute iron poisoning is associated with a high mortality rate. Clinical hepatic injury, however, is uncommon and most symptoms relate to the intestinal tract instead, e.g., nausea, vomiting, diarrhea. Acute iron toxicity most commonly occurs in children less than five years of age. A case of FHF from iron overdose in an adult has been described. Full recovery was achieved with prompt deferoxamine and empiric N-acetylcysteine treatment. Peak serum iron level (340 mcg/dL) was significantly lower than that previously reported to cause hepatotoxicity (O1700 mcg/dL) (98). Vitamin A (Retinol) Sources of vitamin A include dietary products from animals, fruits, vegetables, and oils. Only preformed retinol and other retinoids can cause acute toxicity; b-carotene and other provitamin A carotenoids from food are not toxic. After intake of a physiologic dose of vitamin A, serum concentrations reach higher peaks in old people than in young people (99).

Herbal Medicines, Vitamins, and Natural Hepatotoxins

745

TABLE 3 Medicinal Plants and Complementary and Alternative Medicines with Hepatotoxic Potential Ackee fruit Atractylis gummifera (Mediterranean glue thistle) Azadirachta indica (neem tree) Ba Jiao Lian Berberis vulgaris Callilepsis laureola (Impila) Camphor Cascara sagrada Carp gall bladder Cassia angustifolia (senna) Chelidonium majus (greater celandine) Chinese herbal remedies and teas Lycopodium serratum (Jin Bu Huan) Ma Huang (Ephedra) Shou-Wu-Pian (Polygonum multiflorum) Syo-saiko-to (Xiao-chai-hu-tang) Cocaine (Erythroxylon coca) Copaltra Coptis chinensis (Huang Lian) Cycasin Germander (Teucrium chamaedrys) Kava Kombucha mushroom Isabgol Larrea tridentata (chaparal, creosote bush, greasewood) Lipokinetix Pennyroyal (Hedeoma pulegoides, Mentha pulegium) Pyrrolizidine alkaloids Crotalaria Heliotropium MatØ (Paraquay) tea Senecio Symphytum officinale (comfrey) Tussilago farfara (coltsfoot) T’u-san-chi’ (Compositae) Sassafras albidum (sassafras) Scutellaria (skullcap) Shark cartilage Teucrium polium Valeriana officinalis (valerian) Venencapsan (horse chestnut leaf) Viscum album (mistletoe) Source: Adapted from Refs. 8, 19, 95.

Acute hypervitaminosis A occurs after a single or a limited number of large doses is taken over a short period of time. Signs and symptoms include nausea and vomiting, increased cerebrospinal fluid pressure, headaches, blurred vision, lack of muscular coordination, and constitutional symptoms such as fever, anorexia, and fatigue. In infants, an excessive dose can cause transient bulging of the fontanelles. Similar signs and symptoms can also be seen in chronic hypervitaminosis A, in addition to edema, arthralgias and muscle stiffness, hepatosplenomegaly, and skin and hair changes. Chronic hypervitaminosis A is caused by moderately high doses taken frequently, usually daily, over a span of months to years. Chronic hypervitaminosis A in humans has been reported after recurrent intake of retinol in amounts R10 times the recommended dietary allowance, i.e., R10 mg for adults (typically in food faddists) and R3.75 mg in infants. Chronic vitamin A toxicity produces liver abnormalities that include perisinusoidal fibrosis, stellate cell hypertrophy, and a spontaneous green fluorescence of sinusoidal cells (Fig. 1). Intrinsic liver disease and/or intake of hepatotoxic substances hasten this liver damage, and can be fatal (100). A meta-analysis performed by Myrhe et al. (101)

746

Liu and Schiano

demonstrated that the chances of developing chronic hypervitaminosis A was higher using the water-miscible or emulsified form (w0.2 mg retinol/kg per day for 1–4 weeks) rather than the oil-based preparations (O2 mg retinol/kg for weeks or months).

FOOD CONTAMINANTS Aflatoxin Aflatoxins are potent hepatocarcinogens produced by Aspergillus flavus and Aspergillus parasiticus fungi during production, harvest, storage, and food processing of susceptible food items (e.g., rice, corn, cassava, peanuts, soy beans, chilies, spices). Factors promoting fungal growth include production, storage, and processing of these food items. Aflatoxins are considered by the U.S. FDA to be an unavoidable contaminant of food items. Highly exposed populations are primarily those residing in sub-Saharan Africa, and east and Southeast Asia, i.e., in areas located at latitudes between 408N and 408S of the equator. Synergistic interactions between hepatitis B virus (HBV) and aflatoxins in increasing the risk of developing hepatocellular carcinoma (HCC) have been observed in Shanghai (102) and in Taiwan (103). Susceptibility to aflatoxin is greatest in the young, and there are very significant differences between gender and individuals (differing abilities to detoxify aflatoxin by biochemical processes). HBV positivity also reduces the person’s ability to detoxify aflatoxin (104). Aflatoxicosis is the poisoning that results from ingesting aflatoxins. Two forms of aflatoxicosis have been identified: acute severe intoxication, which results in direct liver damage and subsequent illness or death, and chronic subsymptomatic exposure. All doses have a cumulative effect on the risk of developing cancer. The increased risk of developing HCC is caused by deletion mutations in the p53 tumorsuppressing gene and by activation of dominant oncogenes. The suggested mechanism for the synergy between HBV and aflatoxin in liver cancer development is that aflatoxin suppresses DNA repair mechanisms that help limit the development of cancer from HBV (105). There is little information on whether aflatoxins are related to cases of HCC in the United States. Singapore and Shanghai, both historically high-risk areas for HCC, have been experiencing steadily decreasing incidence rates over the past two decades. The decrease in aflatoxin contamination in the food supply is likely a result of economic development and their residents’ increase in the standard of living (106). Argemone Oil Epidemic dropsy results from the consumption of edible oils adulterated with Argemone mexicana oil or from contamination of seeds (e.g., mustard) with argemone seeds. Several outbreaks of epidemic dropsy have occurred in India, Mauritius, Fiji, and South Africa. Damage is caused mostly by free radical (singlet oxygen and hydroxyl radical) generation. The illness begins as a gastroenteric illness that is followed by oliguria and pedal edema. Other clinical manifestations include nausea, vomiting, diarrhea, erythema, and breathlessness; in extreme cases, glaucoma and death have been reported. The toxicity of argemone has been attributed to two of its alkaloids, sanguinarine and dihydrosanguinarine (107). Sanguinarine given to animals has been shown to cause immediate stimulation of intestinal tone and gut peristalsis. Sanguinarine also causes a decrease in hepatic glycogen levels, thereby leading to pyruvate accumulation in the blood resulting in breathlessness due to the uncoupling of oxidative phosphorylation. Sanguinarine has also been shown to inhibit Na(C), K(C)ATPase activity of different organs, e.g., brain, heart, liver, intestine, and skeletal muscle, thereby causing decreased active transport of glucose. Argemone oil enhances hepatic microsomal and mitochondrial lipid peroxidation. Studies have suggested that singlet oxygen and hydroxyl radical formation are involved in argemone oil toxicity. The oil’s toxic alkaloids induce widespread capillary dilatation and permeability causing leakage of proteinrich plasma into the interstitial tissues of various organs (108). Interstitial fluid accumulation in

Herbal Medicines, Vitamins, and Natural Hepatotoxins

747

FIGURE 1 Chronic vitamin A hepatotoxicity. Liver needle biopsy specimen showing vacuolated Ito cells (arrow) with accompanying perisinusoidal fibrosis (arrowheads). Trichrome stain, original magnification 200!. Source: Courtesy of Dr. M. Isabel Fiel, Department of Pathology, The Mount Sinai Medical Center, New York.

pulmonary alveoli causes restrictive ventilatory dysfunction resulting in hypertension and right-sided failure with well-preserved LV function. This, in turn, causes hepatic venous congestion. Therapy, at present, is mainly symptomatic relief. Cyanobacteria Cyanobacteria (blue-green algae) blooms have been identified in nutrient-enriched fresh water bodies all over the world. Three groups of toxic compounds are associated with cyanobacteria: lipopolysaccharide endotoxins, neurotoxins, and hepatotoxins. Microcystins and nodularins are the most extensively studied group of cyanobacterial hepatotoxins. Exposure to microcystins most likely occurs through consumption of contaminated drinking water or through recreational activities such as swimming; it may also occur through consumption of contaminated food items such as freshwater mussels, fish, edible plants (109). In 1996, 100 patients developed acute liver failure at a hemodialysis clinic in Caruaru, Brazil; the major contributor to their demise was intravenous exposure to microcystins (110). Reported cases of microcystins causing animal and human fatalities have led the World Health Organization to set a recommended limit of 1 mg of microcystin-LR per liter of drinking water (111). The toxicity of microcystins is associated with the inhibition of serine/threonine protein phosphatases 1 and 2A, which can lead to acute hepatocyte necrosis and hemorrhage (112). Microcystins also have potent tumor-promoting activity. Increased rates of primary liver cancer in some areas of China have been attributed to chronic microcystin exposure from ingestion of contaminated drinking water (113). Bacillus cereus Bacillus cereus is a gram-positive or facultatively anaerobic spore-forming rod that can cause a self-limiting gastroenteritis due to two toxins: enterotoxin and cereulide. Immunosuppressed

748

Liu and Schiano

patients can manifest with non-gastrointestinal tract infections, e.g., ocular infections (114) and sepsis (115). Cooked rice is the most commonly reported food item that serves as the growth medium; food poisoning caused by this microorganism is commonly associated with Chinese restaurants. In case reports of B. cereus-associated hepatotoxicity, pasta (116,117) has also been implicated as the food source. Cereulide can cause hepatocyte mitochondrial injury, resulting in reduced fatty-acid metabolism and the formation of microvesicular steatosis. Significant mitochondrial damage results in hepatocyte death and possible liver failure. FHF (116,117) and ruptured hepatic abscess (118) have both been ascribed to this microorganism. OTHER NATURAL HEPATOTOXINS Carp Gall Bladder Raw bile of the grass carp (Ctenopharyngodon idellus) is sometimes ingested in parts of Asia, in the form of pills, for the improvement of visual acuity, rheumatism, and general health. Although ciguatera and scombroid poisoning are the most commonly recognized forms of fish poisoning, ingestion of the raw gall bladder of the grass carp has been reported to cause acute renal failure and acute hepatitis. After ingestion, most patients initially present with gastrointestinal symptoms, i.e., abdominal pain, nausea, vomiting, watery diarrhea after several hours, followed by central nervous system abnormalities and oliguria. Acute hepatitis generally precedes the renal dysfunction, and spontaneous resolution of hepatic dysfunction within a few days is the usual pattern. Liver histology generally demonstrates focal hepatitis; this finding is attributed to bile toxins. A bile component, cyprinol, was found to cause multiple focal necrosis in the liver when administered to rats. Nephrotoxicity can result in oliguric or non-oliguric acute renal failure, occurring usually within 48 to 72 hours after ingestion (while the hepatic dysfunction is resolving) (119). INTERACTIONS BETWEEN PRESCRIPTION DRUGS AND HERBALS/NATURAL PRODUCTS Eisenger et al. (10) have noted that close to 20% of U.S. adults regularly taking prescription medication concurrently use at least one herbal product or high-dose vitamin. Many of these herbs and CAM may have interactions with certain prescription medications used commonly by individuals having liver disease or who have undergone liver transplantation (Table 4) (8,120). Herbs that induce hepatic cytochrome P450 enzymes [i.e., St. John’s wort (121,122)] can potentiate the injury of intrinsic hepatotoxins such as germander and acetaminophen, by way of the increased production of toxic metabolites. Similarly, prescription medications such as anticonvulsants that induce cytochrome P450 may potentiate these toxic metabolites. Grapefruit juice in large amounts has been shown to depress cytochrome P450 function (123). Garlic, ginkgo, ginger, feverfew (Tanacetum parthenium), bilberry (Vaccinium myrtillus), dong quai (Angelica sinensis), turmeric, meadowsweet (Filipendula ulmaria), and willow may all inhibit qualitative platelet function and aggregation; ginseng, St. John’s wort, motherworth (Leonurus cardiaca), chamomile, horse chestnut (Aesculus hipposcastanum), red clover (Trifolium pretense), fenugreek, and cod liver oil may increase the International Normalized Ratio (INR) (23). Garlic has antiplatelet aggregation properties occurring in a dose-dependent manner (124). Serious hematologic side effects have been reported when garlic has been taken in conjunction with blood-thinning prescription drugs such as warfarin (125). Ginkgo contains compounds that act as anticoagulants, inhibiting platelet aggregation. Its long-term use has been associated with increased bleeding time and spontaneous hemorrhage, and thus is contraindicated in patients taking prescription anticoagulants such as warfarin (126,127). Ginger has been shown to act as a potent inhibitor of thromboxane synthetase, raising levels of prostacyclin without a concomitant rise in prostaglandin E2 or prostaglandin F2 alpha, and having implications in prolonging bleeding time (128). Animal studies have demonstrated that ginsenocides contained in ginseng prolong coagulation times of thrombin and activated partial thromboplastin (129). Patients should be discouraged from taking these preparations, especially in combination, before liver biopsy or in the setting of preexisting thrombocytopenia

749

Herbal Medicines, Vitamins, and Natural Hepatotoxins

TABLE 4 Some HerbDrug Interactions Pertinent to Patients with Liver Disease Herb Danshen (Salvia miltiorrhiza) Devil’s Claw (Harpagophy cumbens) Dong quai (Angelica sinensis) Grapefruit juice Glycyrrhizin (licorice root) St. John’s wort (Hypericum perforatum, goatweed) Garlic Ginger Papaya extract Tamarind (Tamarindus indica)

Interacting medication Corticosteroids Anticoagulants, aspirin Corticosteroids Cyclosporine, calcium channel blockers Prednisolone, spironolactone Cyclosporine, indinavir, and other protease inhibitors Anticoagulants, aspirin Anticoagulants, aspirin Anticoagulants, aspirin Anticoagulants, aspirin

Source: Data from Refs. 8, 120.

(in cirrhosis or with the use of interferon), just as they are similarly told to avoid nonsteroidal anti-inflammatory drugs and aspirin-containing compounds. Because of potential effects on cytochrome P450 enzyme systems and the altered metabolism of immunosuppression medications such as cyclosporine and tacrolimus, the use of most herbal medications is discouraged after liver transplantation. Echinacea, Astragalus, glycyrrhizin, alfalfa sprouts, vitamin E, and zinc all have purported immunostimulating effects; theoretically, this effect can interfere with immunosuppression post-transplantation although it remains unproven (130). Ginseng may augment the central nervous system toxic effects of corticosteroids in predisposed patients. Glycyrrhizin may offset the effects of spironolactone. It has no direct hepatotoxicity but can cause fluid retention that may exacerbate ascites and peripheral edema in decompensated cirrhosis. Wormwood, sage, evening primrose oil, and borage may lower seizure threshold, pertinent to patients taking other neurotoxic medications such as cyclosporine and tacrolimus. Horseradish and kelp may suppress thyroid function (130). Tannin-containing herbs (e.g., chamomile, feverfew, and St. John’s wort) may interact with iron, inhibiting its absorption (130). Goldenseal, celery seed, dandelion root, juniper, pokeroot, shepherd’s purse, and yarrow are all aquaretics. Aquaretics differ from diuretics in that sodium is not excreted with water so these herbs are not well suited to the treatment of edema/ascites related to portal hypertension (130). Glycyrrhizin may increase the plasma concentration of prednisolone. Shosaiko-to and Saiboku-to have been shown to alter the pharmacokinetics of prednisolone (131). St. John’s wort (Hypericum perforatum L.) has been reported to decrease serum cyclosporine levels, with resulting liver allograft rejection (132). It may cause reduced bioavailability of indinavir and other protease inhibitors in HIV patients.

SUMMARY Though some herbal medicines have been shown to protect against or treat experimental liver injury in vitro, and many may possess one or a combination of antioxidant, antifibrotic, immunomodulatory, or antiviral activities, they have not been shown effective in human trials. It has been extremely difficult to construct randomized, controlled trials using CAM because of an incomplete understanding of their modes of action, the lack of standardization in their manufacture, and the complexity of ingredients in any herbal extract (7,133). This may become easier once more standardized and broad-based regulatory oversight of marketing and manufacture of these products is achieved. Despite this, the use of CAM is ever increasing, especially in patients having chronic liver disease. With this growing popularity, it is becoming more apparent that many of these treatments possess the potential for appreciable hepatotoxicity, in some instances resulting in significant morbidity and mortality. Until these products are more closely regulated and their advertising better scrutinized, all physicians and patients should become more familiar with the natural and alternative products that are commonly used, and recognize which can be harmful.

750

Liu and Schiano

REFERENCES 1. http://nccam.nih.gov/news/camsurvey_fs1.htm (Accessed October 2005). 2. Ernst E. Harmless herbs? A review of the recent literature. Am J Med 1998; 104(2):170–8. 3. Willett KL, Roth RA, Walker L. Workshop overview: hepatotoxicity assessment for botanical dietary supplements. Toxicol Sci 2004; 79(1):4–9. 4. Kaplowitz N. Hepatotoxicity of herbal remedies: insights into the intricacies of plant–animal welfare and cell death. Gastroenterology 1997; 113(4):1408–12. 5. Larrey D. Hepatotoxicity of herbal remedies. J Hepatol 1997; 26(Suppl. 1):47–51. 6. Fogden E, Neuberger J. Alternative medicines and the liver. Liver 2003; 23(4):213–20. 7. Seeff LB, Lindsay KL, Bacon BR, et al. Complementary and alternative medicine in chronic liver disease. Hepatology 2001; 34(3):595–603. 8. Stedman C. Herbal hepatotoxicity. Semin Liver Dis 2002; 22(2):195–206. 9. Ernst E. The risk-benefit profile of commonly used herbal therapies: ginkgo, St. John’s wort, ginseng, echinacea, saw palmetto, and kava. Ann Intern Med 2002; 136(1):42–53. 10. Eisenberg DM, Davis RB, Ettner SL, et al. Trends in alternative medicine use in the United States, 1990–1997: results of a follow-up national survey. JAMA 2000; 280(18):1569–75. 11. Marcus DM, Grollman AP. Botanical medicines—the need for new regulations. N Engl J Med 2002; 347(25):2073–6. 12. Carlsson C. Herbs and hepatitis. Lancet 1990; 2(8722):1068. 13. Neff GW, O’Brien C, Montalbano M, et al. Consumption of dietary supplements in a liver transplant population. Liver Transpl 2004; 10(7):881–5. 14. Phillips LG, Nichols MH, King WD. Herbs and HIV. South Med J 1995; 88(9):911–3. 15. Nahin RL, Straus SE. Research into complementary and alternative medicine: problems and potential. Br Med J 2001; 322(7279):161–4. 16. Zimmerman HJ, Ishak KG. General aspects of drug-induced liver disease. Gastroenterol Clin North Am 1995; 24(4):739–57. 17. Zimmerman HJ, Lewis JH. Chemical- and toxin-induced hepatotoxicity. Gastroenterol Clin North Am 1995; 24(4):1027–45. 18. Kumar NB, Allen K, Bell H. Perioperative herbal supplement use in cancer patients: potential implications and recommendations for presurgical screening. Cancer Control 2005; 12(3):149–57. 19. Chitturi S, Farrell GC. Herbal hepatotoxicity: an expanding but poorly defined problem. J Gastroenterol Hepatol 2000; 15(10):1093–9. 20. Estes JD, Stolpman D, Olyaei A, et al. High prevalence of potentially hepatotoxic herbal supplement use in patients with fulminant hepatic failure. Arch Surg 2003; 138(8):852–8. 21. McGee J, Patrick RS, Wood CB, et al. A case of veno-occlusive disease of the liver in Britain associated with herbal tea consumption. J Clin Pathol 1976; 29(9):788–94. 22. Tandon HD, Tandon BN, Ramalingaswami V. Epidemic of toxic hepatitis in India of possible mycotoxic origin. Arch Pathol Lab Med 1978; 102(7):372–6. 23. Abebe W. Herbal medication: potential for adverse interactions with analgesic drugs. J Clin Pharm Ther 2002; 27(6):391–401. 24. Stickel F, Seitz HK. The efficacy and safety of comfrey. Public Health Nutr 2000; 3(4A):501–8. 25. Bras G, Jelliffe DB, Stuart KL. Veno-occlusive disease of liver with nonportal type of cirrhosis occurring in Jamaica. Arch Path 1954; 57(4):285–300. 26. Valla D, Benhamou JP. Drug-induced vascular and sinusoidal lesions of the liver. Clin Gastroenterol 1988; 2(2):481–500. 27. Ridker PM, Ohkuma S, McDermott WV, et al. Hepatic venoocclusive disease associated with the consumption of pyrrolizidine-containing dietary supplements. Gastroenterology 1985; 88(4):1050–4. 28. DeLeve L, Wang X, Kuhlenkamp J, et al. Toxicity of azathioprine and monocrotaline in murine sinusoidal endothelial cells and hepatocytes: the role of glutathione and relevance to hepatic venoocclusive disease. Hepatology 1996; 23(3):589–99. 29. Huxtable RJ. New aspects of the toxicology and pharmacology of pyrrolizidine alkaloids. Mol Pharmacol 1979; 10(3):159–67. 30. Mattocks AR. Toxicity of pyrrolizidine alkaloids. Nature 1968; 217(5130):723–8. 31. Gordon DW, Rosenthal G, Hart J, et al. Chaparral ingestion. The broadening spectrum of liver injury caused by herbal mediations. JAMA 1995; 273(6):489–90. 32. Kassler WJ, Blanc P, Greenblatt R. The use of medicinal herbs by human immunodeficiency virusinfected patients. Arch Intern Med 1991; 151(11):2281–8. 33. Sheikh NM, Philen RM, Love LA. Chaparral-associated hepatotoxicity. Arch Intern Med 1997; 157(8):913–9. 34. Kauma H, Koskela R, Ma¨kisalo H, et al. Toxic acute hepatitis and hepatic fibrosis after consumption of chaparral tablets. Scand J Gastroenterol 2004; 39(11):1168–71. 35. Singh YN, Singh NN. Therapeutic potential of kava in the treatment of anxiety disorders. CNS Drugs 2002; 16(11):731–43.

Herbal Medicines, Vitamins, and Natural Hepatotoxins

751

36. Singh YN. Potential for interaction of kava and St. John’s wort with drugs. J Ethnopharmacol 2005; 100(1–2):108–13. 37. Chan TY, Chan JC, Tomlinson B, et al. Chinese herbal medicines revisited: a Hong Kong perspective. Lancet 1993; 342(8886–8887):1532–4. 38. Kane JA, Kane SP, Jain S. Hepatitis induced by traditional Chinese herbs; possible toxic components. Gut 1995; 36(1):146–7. 39. Tomlinson B, Chan TY, Chan JC, et al. Herb-induced aconitine poisoning. Lancet 1993; 341(8841):370–1. 40. Melchart D, Linde K, Weidenhammer W, et al. Liver enzyme elevations in patients treated with traditional Chinese medicine. JAMA 1999; 282(1):28–9. 41. Graham-Brown R. Toxicity of Chinese herbal remedies. Lancet 1992; 340(8820):673–4. 42. Perharic-Walton L, Murray V. Toxicity of Chinese herbal remedies. Lancet 1992; 340(8820):674. 43. Yoshida EM, McLean CA, Cheng ES, et al. Chinese herbal medicine, fulminant hepatitis, and liver transplantation. Am J Gastroenterol 1996; 91(12):2647–8. 44. Ernst E. Adulteration of Chinese herbal medicines with synthetic drugs: a systematic review. J Intern Med 2002; 252(2):107–13. 45. Huang WF, Wen KC, Hsiao ML. Adulteration by synthetic therapeutic substances of traditional Chinese medicines in Taiwan. J Clin Pharmacol 1997; 37(4):344–50. 46. McRae CA, Agarwal K, Mutimer D, et al. Hepatitis associated with Chinese herbs. Eur J Gastroenterol Hepatol 2002; 14(5):559–62. 47. Nadir A, Agrawal S, King PD, et al. Acute hepatitis associated with the use of a Chinese herbal product, Ma-Huang. Am J Gastroenterol 1996; 91(7):1436–8. 48. Borum ML. Fulminant exacerbation of autoimmune hepatitis after the use of Ma Huang. Am J Gastroenterol 2001; 96(5):1654–5. 49. Nelson R. FDA issues alert on Ephedra supplements in the U.S.A. Lancet 2004; 363(9403):135. 50. Neff GW, Reddy R, Durazo FA, et al. Severe hepatotoxicity associated with the use of weight loss diet supplements containing Ma Huang or usnic acid. J Hepatol 2004; 41(6):1062–4. 51. But PP, Tomlinson B, Lee KL. Hepatitis related to the Chinese medicine Shou-Wu-Pian manufactured from Polygonum multiflorum. Vet Hum Toxicol 1996; 38(4):280–2. 52. Park GJ, Mann SP, Ngu MC. Acute hepatitis induced by Shou-Wu-Pian, a herbal product derived from Polygonum multiflorum. J Gastroenterol Hepatol 2001; 16(1):115–7. 53. Kao WF, Hung DZ, Tsai WJ, et al. Podophyllotoxin intoxication: toxic effect of Bajiaolian in herbal therapeutics. Hum Exp Toxicol 1992; 11(6):480–7. 54. Woolf GM, Petrovic LM, Rotjer SE, et al. Acute hepatitis associated with the Chinese herbal product Jin Bu Huan. Ann Intern Med 1994; 121(10):729–35. 55. Thomassen D, Slattery JT, Nelson SD. Menthofuran-dependent and independent aspects of pulegone hepatotoxicity: roles of glutathione. J Pharmacol Exp Ther 1990; 253(2):567–72. 56. Yeung CY. Neonatal hyperbilirubinemia in Chinese. Trop Geogr Med 1973; 25(2):151–7. 57. Horowitz RS, Dart RC, Gomez H, et al. Jin Bu Huan toxicity in children—Colorado, 1993. MMWR Morb Mortal Wkly Rep 1993; 42(33):633–5. 58. Horowitz RS, Feldhaus K, Dart RC, et al. The clinical spectrum of Jin Bu Huan toxicity. Arch Intern Med 1996; 156(8):899–903. 59. Lin TJ, Tsai MS, Chiou NM, et al. Hepatotoxicity caused by Breynia officinalis. Vet Hum Toxicol 2002; 44(2):87–8. 60. Lin TJ, Su CC, Lan CK, et al. Acute poisonings with Breynia officinalis-an outbreak of hepatotoxicity. J Toxicol Clin Toxicol 2003; 41(5):591–4. 61. Itoh S, Marutani K, Nishijima T, et al. Liver injuries induced by herbal medicine, syo-saiko-to (xiaochai-hu-tang). Dig Dis Sci 1995; 40(8):1845–8. 62. Yamashiki M, Nishimura A, Suzuki H, et al. Effects of the Japanese herbal medicine “Sho-saiko-to” (TJ-9) on in vitro interleukin-10 production by peripheral blood mononuclear cells of patients with chronic hepatitis C. Hepatology 1997; 25(6):1390–7. 63. Kamiyama T, Nouchi T, Kojima S, et al. Autoimmune hepatitis triggered by administration of an herbal medicine. Am J Gastroenterol 1997; 92(4):703–4. 64. Larrey D, Vial T, Pauwels A, et al. Hepatitis after germander (Teucrium chaemaedrys) administration: another instance of herbal medicine hepatotoxicity. Ann Intern Med 1992; 117(2):129–32. 65. Lekehal M, Pessayre D, Lereau JM, et al. Hepatotoxicity of the herbal medicine, germander. Metabolic activation of its furano dierpenoids by cytochrome P4503A depletes cytoskeletonassociated protein thiols and forms plasma membrane blebs in rat hepatocytes. Hepatology 1996; 24(1):212–8. 66. Loeper J, Descatoire V, Letteron P, et al. Hepatotoxicity of germander in mice. Gastroenterology 1994; 106(2):464–72. 67. Fau D, Lekehal M, Farrell G, et al. Diterpenoids from germander, an herbal medicine, induce apoptosis in isolated rat hepatocytes. Gastroenterology 1997; 113(4):1334–46.

752

Liu and Schiano

68. De Berardinis V, Moulis C, Maurice M, et al. Human microsomal epoxide hydrolase is the target of germander-induced autoantibodies on the surface of human hepatocytes. Mol Pharmacol 2000; 58(3):542–51. 69. Mattei A, Rucay P, Samuel D, et al. Liver transplantation for severe acute liver failure after herbal medicine (Teucrium polium) administration. J Hepatol 1995; 22(5):597. 70. Dourakis SP, Papanikolaou IS, Tzemanakis EN, et al. Acute hepatitis associated with herb (Teucrium capitatum L.) administration. Eur J Gastroenterol Hepatol 2002; 14(6):693–5. 71. Anderson IB, Mullen WH, Mecker JE, et al. Pennyroyal toxicity: measurement of toxic metabolite levels in two cases and review of the literature. Ann Intern Med 1996; 124(8):726–34. 72. Madyastha KM, Raj CP. Biotransformations of R-(C)pulegone and menthofuran in vitro: chemical basis for toxicity. Biochem Biophys Res Commun 1990; 173(3):1086–92. 73. McClanahan RH, Thomassen D, Slattery JT, et al. Metabolic activation of (R)-(C)-pulegone to a reactive enonal that covalently binds to mouse liver proteins. Chem Res Toxicol 1989; 2(5):349–55. 74. Georgiou M, Sianidou L, Hatzis T, et al. Hepatotoxicity due to Atractylis gummifera-L. J Toxicol Clin Toxicol 1988; 26(7):487–93. 75. Hedili A, Warnet JM, Thevenin M, et al. Biochemical investigation of Atractylis gummifera-L hepatotoxicity in the rat. Arch Toxicol Suppl 1989; 13:312–5. 76. Stewart MJ, Steenkamp V. The biochemistry and toxicity of atractyloside: a review. Ther Drug Monit 2000; 22(6):641–9. 77. Mokhobo KP. Herb use and necrodegenerative hepatitis. S Afr Med J 1976; 50(28):1096–9. 78. Benninger J, Schneider HT, Schuppan D, et al. Acute hepatitis induced by greater celandine (Chelidonium majus). Gastroenterology 1999; 117(5):1234–7. 79. Favreau JT, Ryu ML, Braunstein G, et al. Severe hepatotoxicity associated with the dietary supplement LipoKinetix. Ann Intern Med 2002; 136(8):590–5. 80. Han D, Matsumaru K, Rettori D, et al. Usnic acid-induced necrosis of cultured mouse hepatocytes: inhibition of mitochondrial function and oxidative stress. Biochem Pharmacol 2004; 67(3):439–51. 81. Durazo FA, Lassman C, Han SH, et al. Fulminant liver failure due to usnic acid for weight loss. Am J Gastroenterol 2004; 99(5):950–2. 82. Whiting PW, Clouston A, Kerlin P. Black cohosh and other herbal remedies associated with acute hepatitis. Med J Aust 2002; 177(8):432–5. 83. Cohen SM, O’Connor AM, Hart J, et al. Autoimmune hepatitis associated with the use of black cohosh: a case study. Menopause 2004; 11(5):575–7. 84. Levitsky J, Alli TA, Wisecarver J, et al. Fulminant liver failure associated with the use of black cohosh. Dig Dis Sci 2005; 50(3):538–9. 85. Wang MY, West BJ, Jensen CJ, et al. Morinda citrifolia (Noni): a literature review and recent advances in Noni research. Acta Pharmacol Sin 2002; 23(12):1127–41. 86. Millonig G, Stadlmann S, Vogel W. Herbal hepatotoxicity: acute hepatitis caused by a Noni preparation (Morinda citrofolia). Eur J Gastroenterol Hepatol 2005; 17(4):445–7. 87. Stadlbauer V, Fickert P, Lackner C, et al. Hepatotoxicity of NONI juice: report of two cases. World J Gastroenterol 2005; 11(30):4758–60. 88. Hirono I, Kachi H, Kato T, et al. A survey of acute toxicity of cycads and mortality rate from cancer in the Miyako Islands Okinawa. Acta Pathol Jap 1970; 20(3):327–37. 89. Gloro R, Hourmand-Ollivier I, Mosq B, et al. Fulminant hepatitis during self-medication with hydroalcoholic extract of green tea. Gastroenterol Hepatol 2005; 17(10):1135–7. 90. Bonkovsky HL. Hepatotoxicity associated with supplements containing Chinese green tea (Camellia sinensis). Ann Intern Med 2006; 144(1):68–71. 91. Larson J, Vender R, Camuto P. Cholestatic jaundice due to ackee fruit poisoning. Am J Gastroenterol 1994; 89(9):1577–8. 92. Centers for Disease Control (CDC). Toxic hypoglycemic syndrome—Jamaica, 1989–1991. MMWR Morb Mortal Wkly Rep 1992; 41(4):53–5. 93. Sinniah R, Sinniah D, Chia LS, et al. Animal model of margosa oil ingestion with Reye-like syndrome. Pathogenesis of microvesicular fatty liver. J Pathol 1989; 159(3):255–64. 94. Sinniah D, Baskaran G. Margosa oil poisoning as a cause of Reye’s syndrome. Lancet 1981; 1(8218):487–9. 95. Schiano TD. Hepatotoxicity and complementary and alternative medicines. Clin Liver Dis 2003; 7(2):453–73. 96. Fraga CG, Oteiza PI. Iron toxicity and antioxidant nutrients. Toxicology 2002; 180(1):23–32. 97. Ramm GA, Ruddell RG. Hepatotoxicity of iron overload: mechanisms of iron-induced hepatic fibrogenesis. Semin Liver Dis 2005; 24(4):433–49. 98. Daram SR, Hayashi PH. Acute liver failure due to iron overdose in an adult. South Med J 2005; 98(2):241–4. 99. Krasinski SD, Cohn JS, Schaefer EJ, et al. Postprandial plasma retinyl ester response is greater in older subjects compared with younger subjects. J Clin Invest 1990; 85(3):883–92.

Herbal Medicines, Vitamins, and Natural Hepatotoxins

753

100. Allen LH, Haskell M. Estimating the potential for vitamin A toxicity in women and young children. J Nutr 2002; 132(Suppl. 9):2907S–19. 101. Myhre AM, Carlsen MH, Bøhn SK, et al. Water-miscible, emulsified, and solid forms of retinol supplements are more toxic than oil-based preparations. Am J Clin Nutr 2003; 78(6):1152–9. 102. Qian GS, Ross RK, Yu MC, et al. A follow-up study of urinary markers of aflatoxin exposure and liver cancer risk in Shanghai, People’s Republic of China. Cancer Epidemiol Biomarkers Prev 1994; 3(1):3–10. 103. Wang LY, Hatch M, Chen CJ, et al. Aflatoxin exposure and risk of hepatocellular carcinoma in Taiwan. Int J Cancer 1996; 67(5):620–5. 104. Henry SH, Bosch FX, Bowers JC. Aflatoxin, hepatitis and worldwide liver cancer risks. Adv Exp Med Biol 2002; 504:229–33. 105. Groisman IJ, Koshy R, Henkler F, et al. Downregulation of DNA excision repair by the hepatitis B virus-x protein occurs in p53-proficient and p53-deficient cells. Carcinogenesis 1999; 20(3):479–83. 106. Yu MC, Yuan JM. Environmental factors and risk for hepatocellular carcinoma. Gastroenterology 2004; 127(5 Suppl. 1):S72–8. 107. Das M, Khanna SK. Clinicoepimediological, toxicological, and safety evaluation studies on argemone oil. Crit Rev Toxicol 1997; 27(3):273–97. 108. Sharma BD, Bhatia V, Rahtee M, et al. Epidemic dropsy: observations on pathophysiology and clinical features during the Delhi epidemic of 1998. Trop Doct 2002; 32(2):70–5. 109. McElhiney J, Lawton LA. Detection of the cyanobacterial hepatotoxins microcystins. Toxicol Appl Pharmacol 2005; 203(3):219–30. 110. Azevedo SM, Carmichael WW, Jochimsen EM, et al. Human intoxication by microcystins during renal dialysis treatment in Caruaru—Brazil. Toxicology 2002; 181–182:441–6. 111. WHO. Guidelines for drinking-water quality. Addendum to Volume 2, Health Criteria and Other Supporting Information. 2nd ed. Geneva: World Health Organisation, 1998:93. 112. Bhattacharya R, Sugendran K, Dangi RS, et al. Toxicity evaluation of freshwater cyanobacterium Microcystis aeruginosa PCC 7806: II. Nephrotoxicity in rats. Biomed Environ Sci 1997; 10(1):93–101. 113. Yu SZ. Primary prevention of hepatocellular carcinoma. J Gastroenterol Hepatol 1995; 10(6):674–82. 114. David DB, Kirkby GR, Noble BA. Bacillus cereus endophthalmitis. Br J Ophthalmol 1994; 78(7):577–80. 115. Gaur AH, Patrick CC, McCullers JA, et al. Bacillus cereus bacteremia and meningitis in immunocompromised children. Clin Infect Dis 2001; 32(10):1456–62. 116. Dierick K, Van Coillie E, Swiecicka I, et al. Fatal family outbreak of Bacillus cereus-associated food poisoning. J Clin Microbiol 2005; 43(8):4277–9. 117. Mahler H, Pasi A, Kramer JM, et al. Fulminant liver failure in association with the emetic toxin of Bacillus cereus. N Engl J Med 1997; 336(16):1142–8. 118. Latsios G, Petrogiannopoulos C, Hartzoulakis G, et al. Liver abscess due to Bacillus cereus: a case report. Clin Microbiol Infect 2003; 9(12):1234–7. 119. Lin YF, Lin SH. Simultaneous acute renal and hepatic failure after ingesting raw carp gall bladder. Nephrol Dial Transplant 1999; 14(8):2011–2. 120. Fugh-Berman A. Herb–drug interaction. Lancet 2000; 355(9198):134–8. 121. Barone GW, Gurley BJ, Ketel BL, et al. Herbal supplements: a potential for drug interactions in transplant recipients. Transplantation 2001; 71(2):239–41. 122. Moore LB, Goodwin B, Jones SA, et al. St. John’s wort induces hepatic drug metabolism through activation of the pregnane X receptor. Proc Natl Acad Sci USA 2000; 97(13):7500–2. 123. Bailey DG, Arnold JMD, Spence JD. Grapefruit juice and drugs: how significant is the interaction? Clin Pharmacol 1994; 26(2):91–8. 124. Apitz-Castro R, Badimon JJ, Badimon L. Effect of ajoene, the major antiplatelet compound from garlic, on platelet thrombus formation. Thromb Res 1992; 68(2):145–55. 125. Izzo AA, Ernst E. Interactions between herbal medicines and prescribed drugs: a systematic review. Drugs 2001; 61(15):2163–75. 126. Le Poncin Lafitte M, Rapin J, Rapin JR. Effects of Ginkgo Biloba on changes induced by quantitative cerebral microembolization in rats. Arch Int Pharmacodyn Ther 1980; 243(2):236–44. 127. Smith PF, Maclennan K, Darlington CL. The neuroprotective properties of the Ginkgo biloba leaf: a review of the possible relationship to platelet-activating factor (PAF). J Ethnopharmacol 1996; 50(3):131–9. 128. Backon J. Ginger as an antiemetic: possible side effects due to its thromboxane synthetase activity. Anaesthesia 1991; 46(8):705–6. 129. Park HJ, Lee JH, Song YB, et al. Effects of dietary supplementation of lipophilic fraction from Panax ginseng on cGMP and cAMP in rat platelets and on blood coagulation. Biol Pharm Bull 1996; 19(11):1434–9. 130. Miller LG. Herbal medicinals. Selected clinical considerations focusing on known or potential drug– herb interactions. Arch Intern Med 1998; 158(20):2200–11.

754

Liu and Schiano

131. Homma M, Oka K, Ikeshima K, et al. Different effects of traditional Chinese medicines containing similar herbal constituents on prednisolone pharmacokinetics. J Pharm Pharmacol 1995; 47(8):687–92. 132. Karliova M, Treichel U, Malago` M, et al. Interaction of Hypericum perforatum (St. John’s wort) with cyclosporin a metabolism in a patient after liver transplantation. J Hepatol 2000; 33(5):853–5. 133. Schuppan D, Jia JD, Brinkhaus B, et al. Herbal products for liver diseases: a therapeutic challenge for the new millennium. Hepatology 1999; 30(4):1099–104.

35

Occupational and Environmental Hepatotoxicity Keith G. Tolman and Anthony S. Dalpiaz

Division of Gastroenterology, University of Utah School of Medicine, Salt Lake City, Utah, U.S.A.

INTRODUCTION From the time of the industrial revolution through much of the twentieth century, man has ignored the contamination of his environment and workplace. Thousands of chemicals, many of them untested for toxicity, exist in the workplace and environment. President Theodore Roosevelt was the first political figure to recognize the importance of this contamination, stating in 1907: “Conservation of our natural resources and their proper use constitute the fundamental problem which underlies almost every other problem of our national life.” Alice Hamilton, the first American physician to devote a career to industrial medicine, would comment in 1943: “American medical authorities had never taken industrial diseases seriously. workers accepted the risk with fatalistic submissiveness as part of the price one must pay for being poor” (1). To this day, our solid waste tends to end up in the poor areas of our society because poor people have little political influence. Rachel Carson, in her enlightened book Silent Spring, first drew public attention to this problem (2). Forty years later, roughly 3000 chemicals are produced annually in quantities exceeding one million pounds. The National Research Council has concluded that 78% of these compounds lack minimal toxicology information (3). The Environmental Defense Fund in 1997 reported that such data were lacking for 71% of chemicals produced in large quantity (4). As Robert F. Kennedy, Jr. commented during the William J. Taylor Executive Lecture Series (on the environment) at Westminster College in March 2000, “The most devastating impact of the free market is the suspension of laws that protect us.” In November 2005, nearly 100 years after President Roosevelt’s admonition, the problem continues with the Environmental Protection Agency frustratingly attempting to make the Clean Air Act rules for coal-fired plants more industry-friendly. One of the possible implications for hepatotoxicity is an increased release of mercury which becomes bioconcentrated in the food chain after being absorbed by fish (5). Mr Kennedy would later say, “Investment in the environment is an investment in our infrastructure.” Yet, as the late Dr Hyman Zimmerman noted, “The issues have been clouded, however, by the incompleteness of the database, and the judgments are compromised by the efforts to balance the potential and proposed adverse effects of many pollutants against the important sociologic, economic and medical benefits. Containment of the risks posed by environmental contamination requires systematic and coordinated epidemiologic, toxicologic and clinical studies to set the stage for the proper control measures” (6). A decade later, fewer than 30% of potentially toxic chemicals have been adequately tested and there is continuing exposure in the environment and workplace to known hepatotoxins, such as vinyl chloride (7) and yet-to-be-identified hepatotoxic chemicals, as recently reported in petrochemical workers in Brazil (8).

TYPES OF INJURY Virtually, all types of liver disease may be mimicked by toxic exposure. Many chemicals in both the workplace and environment are capable of injuring the liver but rarely do so because the lungs and skin are at the front end of exposure making them more vulnerable targets.

756

Tolman and Dalpiaz

TABLE 1 Types of Hepatic Injury Type of injury Hepatocellular

Steatosis

Examples Carbon tetrachloride Tetrachloroethane Tetrachloroethylene Dimethylacetamide Hydrochlorofluorocarbons Vinyl chloride Yellow phosphorus Poisonous mushrooms Herbal preparations 2-Nitropropane Dose-dependent toxicity Hypoglycin (Jamaican vomiting illness) Toxic oil (rapeseed oil aniline) Trinitrotoluene Tetrachloroethane

Cholestasis

Toxic oil Methylene dianiline (Epping jaundice) Paraquat

Subacute necrosis/cirrhosis

Trinitrotoluene Tetrachloroethane

Veno-occlusive disease Hepatoportal Sclerosis Cirrhosis

Hepatocellular carcinoma Angiosarcoma

Polychlorinated biphenyls Pyrrolizidine alkaloids Arsenic Vinyl chloride Trinitrotoluene Chlorinated hydrocarbons Arsenic Trichloroethane Aflatoxin B1 Arsenic Vinyl chloride

Exposure Household inhalation Industrial solvent Dry-cleaning industry Plastics and rubber industry Refrigerants and solvents Plastics industry Rat poison and firecrackers Environmental Household Paint coating and varnish remover Residual insecticide Accidental, unripe Akee fruit Contaminated cooking oil Munitions industry Industrial solvent, chemical manufacturing Contaminated bread Residual herbicide Accidental/suicide Munitions industry Industrial solvent, chemical manufacturing Residual, electrical soldering Accidental plant ingestion Vintners Plastics industry Munitions industry Printing industry Vintners Solvent Stored food contamination Vintners, contaminated water Plastics industry

The liver is typically a bystander organ or, more likely, acts to detoxify the potentially toxic chemical. Occasionally, however, the detoxification process goes awry leading to activation of chemicals to toxic metabolites. This problem may be potentiated by simultaneous consumption of substances that enhance the toxicity of hepatotoxins, e.g., carbon tetrachloride and alcohol. Virtually, the entire spectrum of liver disease including steatosis, hepatitis, chronic hepatitis, cirrhosis, and neoplasms can occur as the result of toxic exposure. The types of injury, examples of the substances involved, and their potential sources are identified in Table 1. TYPES OF EXPOSURE A chemical must cross cell membrane barriers in order for exposure to occur. The structure and relative lipid solubility of the compound are the major determinants of transport across cellular membranes. The major routes of exposure are via the skin (dermal), gastrointestinal tract (ingestion), or lungs (inhalation). A parenteral route is also possible, although this is rare. Industrial exposure occurs primarily through inhalation and dermal contact, while environmental exposure occurs primarily through inhalation and ingestion. The route of exposure may affect the toxicity of the chemical. For example, if the compound is detoxified by the liver, exposure by inhalation may be more toxic than by ingestion since the chemical will bypass the liver en route to the target organ. If, however, the liver activates a chemical to a toxic

Occupational and Environmental Hepatotoxicity

757

metabolite, exposure by ingestion may be more toxic. Many inhaled compounds are toxic to the skin, lung, kidney, and bone marrow, but hepatotoxic injury due to environmental pollutants is rare (9). Inhalation exposure involves two distinctly different barriers: the alveolar cells and the epithelium (10). Exposure of the solute across the alveolar barrier leads to rapid transport into the systemic circulation with subsequent activation or deactivation in the liver. In this way, the liver has high exposure to potential toxins. Epithelial exposure, on the other hand, results in slow passage into the systemic circulation and localized concentration in the bronchial epithelium predisposing to lung toxicity. The importance of this is that measuring inhalation exposure in animals may underestimate the toxicity to the lung while overestimating the toxicity to the liver or vice versa.

SUSCEPTIBILITY TO INJURY Individual susceptibility to hepatic injury is affected by age, gender, nutritional status, genetic predisposition, concurrent xenobiotic exposure and, possibly, the presence of underlying disease. Most of the information about susceptibility relates to acute or subacute injury from exposure to drugs rather than chronic exposure to chemicals. With regard to age, adults appear to be more susceptible to isoniazid (11), while children are more susceptible to valproate (12) and aspirin (13). Genetic susceptibility appears to be related to polymorphisms in drug-metabolizing enzymes. Examples include acetylator status for isoniazid where the data are conflicting (14,15), aryl epoxide-converting capacity with phenytoin (16) and sulfoxidation capacity with chlorpromazine (17). The importance of genetic susceptibility is underscored by the occurrence of halothane hepatotoxicity in more than one member of a family (18), the correlation with specific human leukocyte antigen haplotypes (19), and the identification of constitutional susceptibility as described by Farrell et al. (20). Female gender may enhance susceptibility to drug-induced chronic hepatitis (21). Females also appear to be more susceptible to a number of drugs, including halothane (22), diclofenac (23), and chlorpromazine (24). The role of nutritional status remains largely theoretical based on the fact that some foods as well as increased protein intake can induce the cytochrome P-450 system while protein deprivation can deplete the system (25). There is some evidence that fasting can enhance the toxicity of acetaminophen in animals (26) and, perhaps, man. Obesity induces CYP2E1 and in so doing enhances susceptibility to halothane (22). Concurrent medication or environmental/occupational exposure may predispose to toxic drug interactions. While there are many examples of drug–drug interactions, it is unclear if there are interactions with occupational or environmental exposure other than CCl4. There is concern, however, that dioxins, potent inducers of CYP1A2, contaminate some of the rivers in the United States (27). This could enhance other xenobiotics metabolized by CYP1A2. It is not known if the levels of dioxin found in the environment are capable of enhancing the toxicity of CYP1A2 metabolized drugs, such as estradiol, haloperidol, naproxen, warfarin, etc. It is unclear whether underlying systemic disease or liver disease enhances susceptibility to environmental or occupational toxins. It does appear that patients with chronic inflammatory disease such as systemic lupus erythematosis and rheumatoid arthritis have enhanced susceptibility to aspirin. In a recent court case, a worker with hepatitis C was denied work at a petrochemical factory (28). In this case, Echazabel v Chevron, it was argued that Mr Echazabel had been denied his rights under the American Disabilities Act (ADA). Chevron believed that Echezabel’s exposure to hepatotoxic chemicals in its refinery would pose a risk to his health. Advocates of the ADA argued that Chevron’s claim was paternalistic and that employees had the right to determine, through their own consent, the risks that they would take. The Supreme Court sided with Chevron ruling that risks to a disabled worker, if established by medical opinion (NB: the first author was an expert witness for the defendant), can disqualify the worker from protection offered by the ADA. This decision poses ethical questions but also highlights industry’s responsibility to protect workers. Interestingly, the medical opinion in

758

Tolman and Dalpiaz

that case was largely based on a parallel circumstance, i.e., the increased susceptibility of patients with hepatitis C to alcohol.

TESTING FOR OCCUPATIONAL HEPATOTOXICITY Methodology for acute toxic exposure and maximum tolerated exposure levels has advanced greatly while methodology for chronic low-dose exposure, the more typical exposure in the workplace, is lacking. To date, this methodology consists mostly of screening exposed populations for increased serum alanine aminotransferase (ALT) (29) and exposed individuals for enzyme induction (27). In a recent study of workers in a petrochemical plant in Venezuela, air exposure to various chemicals (benzene, toluene, and xylene) was correlated with liver enzyme changes. However, activities of serum ALT, aspartate aminotransferase (AST), g-glutamyl transpeptidase (GGT), and alkaline phosphatase were not influenced by exposure to the solvents; rather, elevations of serum GGT were associated with obesity and alcohol consumption. In another study in Brazil, ALT, AST, and GGT were measured in workers in a petrochemical plant. Using logistic regression analysis for age, weight, years working, use of alcohol, smoking, occupational exposure, and medical history of hepatitis, jaundice, and obesity, the prevalence of elevated enzymes in the refinery workers was 3.56 times greater than the reference population. Thus, it remains unclear whether or not measurement of liver enzymes is a sensitive marker of hepatotoxicity in exposed workers. Since enzyme induction (or inhibition) is important in the mediation of hepatotoxicity, measurement of enzyme induction may reflect the functional characteristics of exposure. Such testing has been done using [N15]methacetin to assess disturbances in CYP450-oxidation (30) which do reflect multicomponent exposure. Human-derived cell lines (HepG2) and rat hepatocytes have also been used to test for genotoxicity (31–33), but application in vivo is awaiting further study. Methods are underway using multiple laboratory tests (hepascore) to detect hepatotoxicity in the work place (e.g., anesthetic exposure), but have not been fully assessed for uniformity and specificity (34). Thus, it is still not possible to predict hepatotoxicity for chronic exposure to chemicals and the surveillance testing remains unsatisfactory. Past experience underlies the concern, perhaps best illustrated by the example of angiosarcoma secondary to subtle exposure to vinyl chloride (35).

MECHANISMS OF INJURY Most chemicals exert their effect through selective targeting of specific tissue sites. The toxicity is a function of the ionization state, the specific interaction with the target organ, and the organ’s ability to metabolize the potentially toxic chemical. Most environmental and industrial toxins are highly lipid soluble and undergo metabolic transformation rather slowly and thus tend to become bioconcentrated in body fat stores. This allows them to exert their effects for prolonged periods. The polychlorinated biphenyls (PCBs) and vinyl chloride are good examples of such toxins that stay in the environment and the food chain for many years. Cell membranes are lipid bilayers and only chemicals in a nonionized and thus lipidsoluble state can enter the cell. The degree of ionization in solution is dependent upon the pH of the solution and the acidic dissociation constant (pKa) of the chemical. The pKa is the pH at which one-half of the compound is in the ionized state and another one-half in the nonionized state. By convention, this is expressed as the acidic pKa. For an acid, a low pKa indicates a strong acid and for a base, a low pKa indicates a weak base. At a pH below the pKa, acids exist in the nonionized form while bases exist in the ionized form. Thus aspirin, a strong acid, exists in the nonionized state at gastric pH, as do nonsteroidal drugs, which for the most part are weak bases. Whenever there is a pH gradient across a membrane, a concentration gradient for the nonionized compound will exist. This is especially true in the stomach and kidneys. Thus,

Occupational and Environmental Hepatotoxicity

ALT elevation

759

Adaptation

Drug

Non-toxic metabolite

CYP450 bioactivation

Detoxification e.g. GSH conjugation

Reactive metabolite Covalent binding

Oxidative stress

Neoantigen formation

Dysfunction of Formation of reactive essential cell proteins oxygen species

Immune-mediated cell injury

DNA and protein damage ?Carcinogenicity, Apoptosis

FIGURE 1 Potential mechanism of toxic cell injury.

aspirin at the acidic pH of the stomach is nonionized and tends to bioconcentrate across the cell membrane of the stomach but is excreted at high concentrations in alkaline urine. Hepatic biotransformation usually converts lipid-soluble compounds to less toxic watersoluble compounds that can be eliminated in urine or bile. This process is usually protective to the host. However, it can go awry leading to the formation of toxic-reactive metabolites. The biotransformation is mostly carried out by cytochrome P450-mediated oxidation in the microsomal enzyme system, a system that may be induced or inhibited by various compounds. Chemicals that either inhibit or induce this system may enhance or reduce the toxicity dependent upon whether the parent or a metabolite is the toxic compound. This is the case with CCl4 toxicity, which is worsened by alcohol which accelerates the conversion of CCl4 to a toxic-reactive metabolite (36,37). Environmental interactions may also occur. For example, 2,3,7,8-tetrachlorodibenzo-p-dioxin present in many rivers, is a potent inducer of aryl hydrocarbon hydroxylase, which metabolizes ubiquitous polycyclic aromatic hydrocarbons to potentially carcinogenic metabolites (38). The mechanisms of toxicity remain speculative for most environmental and occupational hepatotoxins. It is clear that formation of toxic metabolites leads to the cytotoxicity and carcinogenicity of many compounds. Such is the case with the cytotoxicity of CCl4 (39) and the carcinogenicity of vinyl chloride (40). A general model is outlined in Figure 1. The potential mechanisms include covalent binding of metabolites to critical cellular proteins necessary for cell survival. Another mechanism involves the formation of reactive oxygen species resulting in oxidative stress. Cell injury or damage to DNA may then occur resulting in carcinogenicity or mutagenicity. A third potential mechanism, albeit not known to occur with industrial toxins, involves formation of neoantigens that stimulate an immune attack. While these mechanisms are well documented as a cause of acute injury, the mechanisms for subtle chronic injury remain unclear.

HEPATOTOXIC CHEMICALS The National Institute for Occupational Safety and Health has published a pocket guide to hazardous chemicals (41). A total of 667 industrial chemicals are listed of which 228 have reference to liver toxicity through either animal experimentation or clinical observation. The incidence and severity of liver toxicity, however, appear to be low and many of the chemicals are only of historical interest at this time. The most widely known are listed in Table 2.

760

Tolman and Dalpiaz

TABLE 2 Most Widely Known Hepatotoxins Chemical class Halogenated aromatic hydrocarbons Polychlorinated biphenyls Chloronaphthalenes Chlorinated benzenes Halogenated aliphatic hydrocarbons Carbon tetrachloride

Use

Teratogenic

Carcinogenic

Environmental contaminant (fish) HCN steatosis Environmental contaminant (fish) HCN HCN

? ? ?

? ? ?

Experimental hepatotoxin

Steatosis, HCN, acute liver failure HCN HCN, steatosis HCN, steatosis

?

Yes

? ? ?

Yes Yes Yes

HCN HCN

? ?

Yes Yes

Hepatic sclerosis

Yes

Yes

Steatosis Steatosis Necrosis Steatosis, cirrhosis

? ? ? ? ?

Yes

HCN, J, steatosis HCN, steatosis

? ?

? ?

HCN HCN, jaundice HCN HCN, steatosis HCN, jaundice

? ? ? ? ?

? Yes ? ? ?

? ? ?

Yes Yes ?

?

?

1,1,2,2-Tetrachloroethane Chemical manufacture 1,1,1-Trichloroethane Varnish, solvent Chloroform Pharmaceutical manufacture, sniffing Halothane Anesthetic Hydrochlorofluorocarbons Refrigerants, cleaning agents, solvents (HCFC 123 and HCFC 124) Chlorinated ethylenes Vinyl chloride Manufacture of PVC, plastics, food wrapping, ground water Vinylidine chloride Manufacture of plastics trans-Dichloroethylenes Solvent cis-Dichloroethylene Solvent Trichloroethylene Solvent, degreaser Perchloroethylene Solvent, drycleaning, paint, pesticides, fluorocarbons N-substituted amides Dimethylacetamide Solvent, resins, polymers Dimethylformamide Solvent, resins, polymers Nitroaromatic compounds Dinitrobenzene 2,6-Dinitrotoluene Picric acid Tetryl Trinitrotoluene Miscellaneous organics Diphenyl oxide Dimethylnitrosamine Experimental Methylene dianiline Plastic Pyridine Solvent, chemical manufacturing Paraquat Herbicide Diquat Herbicide Miscellaneous inorganics Arsenic Insecticide, miners, vineyard workers, homicide, experimental Beryllium Experimental Copper Hydrazine Aliphatic hydrocarbons Benzene Toluene Selenium

Type of injury

Fungicide Rocket fuel Glue sniffing Semiconductors, photoconductors

HCN HCN HCN, J HCN HCN, cholestasis HCN Steatosis, HCN, AS Granulomas, zone 2 necrosis, ALT, elevations, steatosis Granulomas, AS Steatosis, ALT, elevations Steatosis Steatosis, HCN

Yes Yes

Yes ?

Yes

?

?

Abbreviations: AS, angiosarcoma; J, jaundice; HCN, hepatocellular necrosis; ALT, alanine aminotransferase.

Occupational and Environmental Hepatotoxicity

761

HALOGENATED AROMATIC HYDROCARBONS Polychlorinated Biphenyls The dielectric properties, chemical stability, and non-inflammability of PCBs make them highly desirable for manufacturing high-voltage electrical apparatus and insulating material. However, the exposure of workers has resulted in serious hepatic disease (42). Manufacturing of PCBs was discontinued in 1977. However, their resistance to biodegradation and insolubility in water has allowed them to remain in the environment (43–45) and contaminate wildlife and edible fish. They have thus entered the food chain and are present in breast milk (46) in contaminated areas of the United States as well as in Third World countries where polluting industries have emerged (47). The hepatotoxicity and carcinogenicity is a function of the number of chlorines per molecule with those containing five or more chlorines being the most toxic (45,48). PCBs are potent inducers of CYP450 isoenzymes, which in turn are potential inducers of carcinogenic compounds, but there is no convincing evidence that PCBs are hepatocarcinogenic in humans (49). However, they also induce d-aminolevulinic synthase, which probably accounts for the cases of porphyria cutanea tarda among those with occupational exposure. Hepatic injury in humans has occurred as the result of industrial exposure or by accidental exposure to contaminated cooking oil (Yusho disease in Japan) (50). The current level of exposure that occurs from residual amounts of PCBs in the environment does not appear to cause liver disease in humans (reviewed in 43–45). However, because these agents are such potent inducers of the CYP450 system, there are continuing concerns about their potential enhancement of the hepatotoxic effects of other chemicals and drugs (51,52). In addition, PCBs have been shown to be carcinogenic in rodents (45). Studies of the environmental effects of PCBs, have been compromised by cross-contamination with dioxin, a well-known hepatotoxin in animals (38,53).

HALOGENATED ALIPHATIC HYDROCARBONS Carbon Tetrachloride Carbon tetrachloride is one of the three classic hepatotoxins (the others are yellow phosphorus and toxic mushrooms) that cause acute hepatic and renal failure with hepatocellular necrosis and steatosis. In the early 1900s, CCl4 was used as a vermifuge. Subsequently, it was used as a solvent and in fire extinguishers leading to household and industrial exposure (54). Exposure typically occurred by inhalation of fumes during the cleaning of vats or by accidental ingestion in the household (54–56). CCl4 is no longer available in the United States other than as an experimental hepatotoxin. It is now largely of historical interest with no cases having been reported in the United States since 1985. It should not be forgotten, however, that in some parts of the world, CCl4 is still used as an ingredient in fire extinguishers and refrigerants. The toxicity of CCl4 is mediated by a trichloromethyl radical generated by cytochrome P2E1. Cell membrane injury from lipid peroxidation inhibits triglyceride secretion with subsequent steatosis and necrosis (39). Fulminant hepatic failure ensues (57). The injury is augmented by alcohol, which induces CYP2E1 (58,59). Experimental studies suggest that nutritional status—particularly starvation and high fat diets enhance toxicity (56,60,61). Tetrachloroethane 1,1,2,2-Tetrachloroethane is a chemical intermediate used in the manufacture of the solvents trichloroethylene and tetrachloroethane. It is similar to CCl4 in terms of hepatotoxicity. Hepatotoxicity is characterized by hepatitis (and/or steatosis with low dose exposure) and has been reported in 25 patients (62). There has been one fatal case in a person with cirrhosis and superimposed hepatitis (63). There is also a report of exposure in a penicillin-manufacturing facility, where 50% of the workers developed abnormal liver chemistry tests over a threeyear period (64). Fatty degeneration is seen experimentally in rodents (65).

762

Tolman and Dalpiaz

Hydrochlorofluorocarbons Hydrochlorofluorocarbons (HCFCs), specifically 1,1-dichloro-2,2,2-tetrafluoroethane (HCFC 124), are being increasingly substituted for ozone-depleting chlorofluorocarbons (CFCs) (66). This was the result of a convention of international experts in June 1990, in Montreal, Canada where it was determined that depletion of the ozone layer was occurring as a result of release of active chlorine from CFCs. The potential health consequences of the ozone layer depletion created an urgent need for development of partially halogenated HCFCs, which are now used as refrigerants, cleaning agents, industrial solvents, and foam blowing agents. In 1997, Hoet et al. reported an “epidemic” of nine industrial workers who had reported accidental exposure to a mixture of HCFC 123/124 (67). One of the workers developed a picture of “acute mixed hepatitis” with ALT, 1298 U/L; alkaline phosphatase, 303 U/L; prothrombin time, 51%; and bilirubin, 289 mmol/L. He recovered completely over the next two months and then relapsed when he returned to work. A liver biopsy showed hepatocellular coagulative necrosis and canalicular bile plugs in zone 3 with bridging necrosis. There was a moderate lymphoid infiltrate. Immunohistochemical staining demonstrated trifluoroacetyl protein adducts similar to those seen with halothane toxicity. The chemical structures of HCFC 123 and HCFC 124 are similar to that of halothane, an inhalation anesthetic known to cause hepatitis in susceptible individuals after repeated exposure. HCFCs 123/124 are metabolized through the same oxidative pathway as halothane to a reactive trifluoroacetyl halide that forms a hapten (68). Animal studies, primarily in rats, have demonstrated some hepatotoxicity for HCFC 123, but not HCFC 124 (69–71). Acute exposure to HCFC 123 in guinea pigs has been demonstrated to be hepatotoxic (72). The toxicity is enhanced by glutathione depletion (73). Hepatic adenomas have been seen in rats in chronic studies at subacute doses. Dekant has emphasized the need for more research into the chronic effects and the mechanisms of injury (74). Five out of six HCFC 123/124-affected workers analyzed had antibodies to protein disulfide isomerose isoform p58 and CYP2E1, again similar to that seen with halothane hepatotoxicity. The current production of HCFCs is measured in kilotons per year but is expected to greatly increase owing to the ban on CFCs. The hepatotoxicity and possible carcinogenicity of HCFCs in humans raise questions about the net risk–benefit ratio of the ban on CFCs and introduction of HCFCs. CHLORINATED ETHYLENES Tetrachloroethylene (Perchloroethylene) Tetrachloroethylene is used primarily in the dry-cleaning industry and for textile processing. It is also used as an insulating fluid in the production of fluorocarbons and, to a lesser extent, in the production of adhesives, aerosols, and paint. Most exposure occurs through inhalation and dermal contact as an industrial contaminant. Liver injury, including cirrhosis, has been reported in workers with low-dose exposure over a two- to six-year period (75). Cases of accidental high-level exposure resulting in hepatotoxicity have also been reported (76–79). Tetrachloroethylene is hepatocarcinogenic in rodents (80). Vinyl Chloride The recognition of angiosarcoma of the liver in factory workers exposed to vinyl chloride (35) after four decades of apparent safe use led to widespread interest in the toxicity of this agent which is used in the manufacture of polyvinyl chloride. Since that time, several hepatic lesions including acute hepatocellular injury, hepatoportal sclerosis, peliosis, cirrhosis, and hepatocellular carcinoma have been described (56). The histologic sequence of events from preneoplastic changes (hepatocyte and sinusoidal cell hyperplasia, fibrosis, sinusoidal dilation, and peliosis hepatis) to angiosarcoma has been described (81). The mechanistic sequence of events appears to involve metabolic activation to chloroethylene oxide followed by DNA binding of the reactive metabolite exocyclic ethanol adducts, oxidative stress, and interaction with or mutations of tumor suppressor genes (40). A significant gene–environment interaction

Occupational and Environmental Hepatotoxicity

763

between vinyl chloride exposure and polymorphisms in the DNA repair protein XRCC1 in the occurrence of mutant p53 biomarkers has been demonstrated. This may be related to polymorphisms in glutathione-S-transferases which may be involved with Phase II metabolism of the reactive metabolites of vinyl chloride (82). N-SUBSTITUTED AMIDES Dimethylacetamide N,N-dimethylacetamide is a solvent used in the synthesis of resins and polymers. Jaundice has been seen in people repeatedly exposed to small amounts of the solvent (83). Workers monitored over a 2- to 10-year period have shown abnormal liver chemistries (84), while hepatitis after dermal exposure has been experienced by those employed in an acrylic manufacturing plant (85). As a result, biological monitoring of urinary metabolites is now recommended in exposed workers. Fatty infiltration in rats (86) and focal necrosis in dogs (87) have been observed after highdose experimental exposure. Dimethylformamide N,N-dimethylformamide is a widely used solvent in the resin and polymer industries. It is also used in the manufacture of paint, film, and adhesives. Currently more than 100,000 workers per year are exposed to the solvent. Inhalation, ingestion, and dermal exposure have all been shown experimentally to lead to injury (88–93). Acute hepatitis is seen in rats (92), but there have been few occupational exposures resulting in tissue injury in humans (94,95). There have been scattered reports of mild hepatotoxicity (96–98). A relatively recent report, however, described cases of hepatic injury in a group of fabric workers with more than one year of exposure (99). Liver biopsies showed steatosis. CHLORPHENOXY COMPOUNDS More than 15 isomers of chlorophenol are available for use as insecticides and pesticides. These include paraquat, chlordane, heptachlor, aldrin, dieldrin, lindane, and chlordecone (kepone) (100,101). While zonal necrosis and steatosis can occur with these compounds (102), there is little evidence of hepatotoxicity in humans, even when significant amounts are ingested. Ingestion of large quantities of dichloro-diphenyl-trichloroethane (DDT) produces hepatic necrosis (103) but hepatic abnormalities are not usually seen among workers in DDT factories (104). Experimental hepatocarcinogenesis has been shown for aldrin, amitrole, aramite, captan, chlorobenzilate, chlordane, chlordecone, DDT, dieldrin, heptachlor, lindane, mirex, and other pesticides (105–108). There is a continuing concern about Agent Orange, a defoliant used in Vietnam. It is a mixture of the chlorophenoxy herbicides 2,4-dichlorophenoxyacetic acid, 2,4,5-trichlorophenoxyacetic acid, and dioxin (109–111). Dioxin is a potent hepatocarcinogen. The chlorophenoxy compounds have been associated with porphyria cutanea tarda (112,113) but there is little evidence of hepatic injury. Nevertheless, the latency period from exposure to development of tumors is measured in years in humans and this situation deserves continued monitoring. Chlordecone (Kepone) Chlordecone is a pesticide that attracted considerable attention in the lay press in the 1970s when workers at a small plant developed neurological and systemic symptoms (114,115). Contamination of an abutting structure and far-off waters was subsequently demonstrated (116–118). A year later chlordecone was still evident in the food chain (119). Chlordecone is a relatively mild hepatotoxin with only slight biochemical and histological changes noted despite high-tissue levels in liver and other organs (114,120). Steatosis and mild necrosis have been seen in humans and animals. It is hepatocarcinogenic in mice and rats (114).

764

Tolman and Dalpiaz

It is a potent inducer of the CYP450 system and has the ability to enhance the toxicity of CCl4 and CHCl3. Guzelian has demonstrated that treatment with cholestyramine enhances excretion of this agent (114). The large body stores of chlordecone and its hepatocarcinogenicity in rats and mice raise concern about long-term safety but no evidence of hepatotoxicity in humans has appeared to date. Paraquat Paraquat is a toxic herbicide used as a crop defoliant and weed killer. Two fatal cases of hepatic injury were reported in 1966 (121). A number of cases have since been reported as the result of attempted suicide (121,122) or accidental ingestion (115). The case fatality rate is 50% to 70% but the cause of death is usually pulmonary (123). Patients exposed to paraquat present with severe vomiting, diarrhea, abdominal pain, and irritation of the oropharyngeal area. Hepatic manifestations, including jaundice, appear on the second or third day. Histological changes include hepatocellular necrosis followed by severe cholangitis. Biochemical changes are mixed hepatocellular-cholestatic (124). Treatment consists of vigorous diuresis and intragastric administration of activated charcoal. There are also case reports crediting dexamethasone and cyclophosphamide with improved recovery. MISCELLANEOUS ENVIRONMENTAL HEPATOTOXINS Aflatoxins The aflatoxins are groups of closely related furanocoumarins that are potent hepatotoxins and hepatocarcinogens (62,125). They were initially recognized as the cause of an epidemic of acute necrosis in turkey poults, duckling, and pheasants in England in 1960 (126,127). The most potent aflatoxin is AFB1 derived from Aspergillus flavus found to contaminate peanuts, cashews, wheat, corn, oats, soybeans, rice, and cottonseed primarily in warm, moist conditions. Animal aflatoxicosis occurs worldwide, albeit primarily in third world countries. There have not been any cases in the United States. Acute ingestion of large doses of aflatoxins results in acute hepatic necrosis with variable zonal distribution depending on the species (reviewed in 128). Susceptibility is species specific with rainbow trout being highly susceptible while coho salmon and channel catfish are relatively resistant (128,129) as are chickens (130). Cirrhosis has been produced in screened animal species, including cattle, pigs, chickens, and rabbits (131). The lesion in monkeys resembles the acute fatty liver of Reye’s syndrome (132). Similar cases have been described in humans in third world countries (128). The clinical syndrome is characterized by vomiting, abdominal pain, seizures, hepatitis, pulmonary edema, and death. Aflatoxins are highly carcinogenic. AFB1 is the most potent known hepatocarcinogen in rats (128,133). There is epidemiologic evidence that aflatoxins are carcinogenic in humans (132,134–136). The mechanism of injury appears to be impairment of nucleic acid synthesis due to DNAdependent RNA polymerase activity (125,137). Both the acute injury and carcinogenesis result from active metabolites of aflatoxins. The metabolism of AFB1 to 2,3-epoxide is mediated primarily by CYP3A4 and, to some extent, by CYP1A2, CYP2E1, CYP2B7, and CYP3A3 (138). The 2,3-expoxide forms adducts with DNA resulting in a mutation in p53 (substitution of a guanine for thymine at codon 249) (138). This mutation has been found in the P53 tumor suppressor gene in human hepatocellular carcinoma (137,139). Phosphorus Poisoning by yellow phosphorus is relatively rare in this country since being outlawed in 1942 prior to which it was used in matches and firecrackers (140). Since that time, cases have occurred as the result of accidental or suicidal ingestion of rodent poison (141–143). The clinical syndrome progresses from acute gastrointestinal symptoms consisting of nausea, vomiting, abdominal pain, and diarrhea with a garlic odor lasting up to 24 hours,

Occupational and Environmental Hepatotoxicity

765

to a latent phase of one to three days, followed by acute hepatic and renal failure. Jaundice and delirium follow. The fatality rate is high (141). Liver histology is characterized by zone 1 steatosis and necrosis (141,144). The only effective management is early gastric lavage. SUMMARY Contamination of the environment and workplace by toxic chemicals continues to pose a health hazard, the magnitude of which remains uncertain. The risk has been greatly reduced in terms of acute overt toxicity. The long-term effects, however, of low-level exposure in causing chronic liver disease and hepatic cancer remains a concern. The latency period from exposure to chronic liver disease and carcinoma is measured in years, yet the monitoring that is in place is confined to acute exposure without continuing monitoring of people who have moved from a potentially toxic environment. The discovery of latent injury in vinyl chloride workers (81) and of nonalcoholic steatohepatitis in workers and nearby residents of a petrochemical plant in Brazil (145,146) underscores the importance of monitoring and enhanced testing of potentially toxic chemicals. The science has been compromised by the paucity of biological models for chronic exposure and the persistence of often-conflicting medical, economic, social, and political priorities. Whether we will learn from the lessons of the past remains uncertain. It seems self-evident that a person’s health is, in large part, a function of the health of his or her environment (147). As Kennedy said, “In a true free market, you clean up after yourself.” REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

Hamilton A. Exploring the Dangerous Trades. Boston, NY: Little Brown, 1943:1. Carson R. Silent Spring. Cambridge, MA: Riverside Press, 1962. National Research Council. Toxicity Testing. Washington, DC: National Academic Press, 1984. Roe D, Pearse W, Horini K, et al. Toxic Ignorance: The Continuing Absence of Basic Health Testing for Top-Selling Chemicals in the United States. New York: Environmental Defense Fund, 1997. http://www.whitehouse.gov/news/releases/2002/02/20020214.html, Accessed October 24, 2005. Zimmerman H. Hepatotoxicity. Adverse Effects of Drugs and Other Chemicals on the Liver. New York: Appleton-Crofts, 1978:333. Pirastu R, Comba P, Reggiani A, Foa V, Masina A, Maltoni C. Mortality from liver disease among Italian vinyl chloride monomer/polyvinyl chloride manufacturers. Am J Ind Med 1990; 17(2):155–61. Barberino JL, Carvalho FM, Silvany-Neto AM, et al. Liver changes in workers at an oil refinery and in a reference population in the state of Bahia Brazil. Rev Panam Salud Publica 2005; 17(1):30–7. Guzelian PS. Hepatic injury due to environmental agents. Clin Lab Med 1984; 4(3):483–8. Gerde P. Animal models and their limitations: on the problem of high-to-low dose extrapolations following inhalation exposures. Exp Toxicol Pathol 2005; 57(Suppl. 1):143–6. Comstock GW, Edwards PQ. The competing risks of tuberculosis and hepatitis for adult tuberculin reactors. Am Rev Respir Dis 1975; 111(5):573–7. Dreiff FE, Santilli N. Valproic acid hepatic fatalities: analysis of United States cases. Neurology 1986; 36(Suppl. 1):133. Zimmerman HJ. Effects of aspirin and acetaminophen on the liver. Arch Intern Med 1981; 141(3):333–42. Mitchell JR, Long MW, Thorgeirsson UP, Jollow DJ. Acetylation rates and monthly liver function tests during one year of isoniazid preventive therapy. Chest 1975; 68(2):181–90. Timbrell JA. Isoniazid metabolism in relation to hepatotoxicity. In: Davis M, Tredger JM, Williams R, eds. Drug Reactions and The Liver. Bath, UK: Pittman, 1981:190–6. Spielberg SP. In vitro assessment of pharmacogenetic susceptibility to toxic drug metabolites in humans. Fed Proc 1984; 43(8):2308–13. Watson RG, Olomu A, Clements D, Waring RH, Mitchell S, Elias E. A proposed mechanism for chlorpromazine jaundice—defective hepatic sulphoxidation combined with rapid hydroxylation. J Hepatol 1988; 7(1):72–8. Hoft RH, Bunker JP, Goodman HI, Gregory PB. Halothane hepatitis in three pairs of closely related women. N Engl J Med 1981; 304(17):1023–4. Berson A, Freneaux E, Larrey D, et al. Possible role of HLA in hepatotoxicity. An exploratory study in 71 patients with drug-induced idiosyncratic hepatitis. J Hepatol 1994; 20(3):336–42. Farrell G, Prendergast D, Murray M. Halothane hepatitis. Detection of a constitutional susceptibility factor. N Engl J Med 1985; 313(21):1310–4.

766

Tolman and Dalpiaz

21. Seeff LB. Drug-induced chronic liver disease, with emphasis on chronic active hepatitis. Semin Liver Dis 1981; 1(2):104–15. 22. Stock JG, Strunin L. Unexplained hepatitis following halothane. Anesthesiology 1985; 63(4):424–39. 23. Banks AT, Zimmerman HJ, Ishak KG, Harter JG. Diclofenac-associated hepatotoxicity: analysis of 180 cases reported to the Food and Drug Administration as adverse reactions. Hepatology 1995; 22(3):820–7. 24. Ayd FJ, Jr. Chlorpromazine: ten years’ experience. JAMA 1963; 184:51–4. 25. McLean AEM. Host factors in hepatotoxicity. Isr J Med Sci 1974; 10:431. 26. Whitcomb DC, Block GD. Association of acetaminophen hepatotoxicity with fasting and ethanol use. JAMA 1994; 272(23):1845–50. 27. Abraham K, Geusau A, Tosun Y, Helge H, Bauer S, Brockmoller J. Severe 2,3,7,8-tetrachlorodibenzop-dioxin (TCDD) intoxication: insights into the measurement of hepatic cytochrome P450 1A2 induction. Clin Pharmacol Ther 2002; 72(2):163–74. 28. Daniels N. Chevron v Echazabal: protection, opportunity, and paternalism. Am J Public Health 2003; 93(4):545–8. 29. Fernandez-D’Pool J, Orono-Osorio A. Liver function of workers occupationally exposed to mixed organic solvents in a petrochemical industry. Invest Clin 2001; 42(2):87–106. 30. Rehwagen M, Krumbiegel P, Koschny I, Rolle-Kampczyk U, Richter M, Herbarth O. The [15N]methacetin liver function test characterizes multicomponent exposure of children in industrially polluted regions. Isotopes Environ Health Stud 2001; 37(2):167–74. 31. Knasmuller S, Cavin C, Chakraborty A, et al. Structurally related mycotoxins ochratoxin A, ochratoxin B, and citrinin differ in their genotoxic activities and in their mode of action in human-derived liver (HepG2) cells: implications for risk assessment. Nutr Cancer 2004; 50(2):190–7. 32. Knasmuller S, Mersch-Sundermann V, Kevekordes S, et al. Use of human-derived liver cell lines for the detection of environmental and dietary genotoxicants; current state of knowledge. Toxicology 2004; 198(1–3):315–28. 33. Fastner J, Heinze R, Humpage AR, Mischke U, Eaglesham GK, Chorus I. Cylindrospermopsin occurrence in two German lakes and preliminary assessment of toxicity and toxin production of Cylindrospermopsis raciborskii (Cyanobacteria) isolates. Toxicon 2003; 42(3):313–21. 34. Pollet C, Bar F, Battista S, et al. Identification and evaluation of liver damage in subjects exposed to occupational risk from general anesthetics. Med Lav 2003; 94(5):421–31. 35. Creech JL, Jr., Johnson MN. Angiosarcoma of liver in the manufacture of polyvinyl chloride. J Occup Med 1974; 16(3):150–1. 36. Recknagel RO, Glende EA, Jr., Dolak JA, Waller RL. Mechanisms of carbon tetrachloride toxicity. Pharmacol Ther 1989; 43(1):139–54. 37. Zimmerman HJ. Effects of alcohol on other hepatotoxins. Alcohol Clin Exp Res 1986; 10(1):3–15. 38. Poland A, Kende A. 2,3,7,8-Tetrachlorodibenzo-p-dioxin: environmental contaminant and molecular probe. Fed Proc 1976; 35(12):2404–11. 39. Dianzani M. Biochemical aspects of fatty liver. In: Meeks RG, Harrison SD, Bull RJ, eds. Hepatotoxicity. Boca Raton, FL: CRC Press, 1991:327. 40. Bolt HM. Vinyl chloride-a classical industrial toxicant of new interest. Crit Rev Toxicol 2005; 35(4):307–23. 41. NIOSH. Pocket Guide to Chemical Hazards. U.S. Government Printing Office, ed. U.S. Department of Health and Human Services Publications 94-116. U.S. Government Printing Office, 1994. 42. Strauss N. Hepatotoxic effects following occupational exposure to halowax (chlorinated hydrocarbons). Rev Gastroenterol 1994; 11:381. 43. D’Itri FM, Kamrin MA, eds. PCBs: Human and Environmental Hazards. Woburn, MA: Butterworths, 1983. 44. Peakall DB, Lincer JL. Poly chlorinated biphenyls: another long-life, widespread chemical in the environment. Bioscience 1970; 20:958. 45. Kimbrough RD. Human health effects of polychlorinated biphenyls (PCBs) and polybrominated biphenyls (PBBs). Annu Rev Pharmacol Toxicol 1987; 27:87–111. 46. Fries GF. The PBB episode in Michigan: an overall appraisal. Crit Rev Toxicol 1985; 16(2):105–56. 47. LaDou J. The export of environmental responsibility. Arch Environ Health 1994; 49(1):6–8. 48. Kimbrough RD. The toxicity of polychlorinated polycyclic compounds and related chemicals. CRC Crit Rev Toxicol 1974; 2(4):445–98. 49. Goldstein JA, Friesen M, Scotti TM, Hickman P, Hass JR, Bergman H. Assessment of the contribution of chlorinated dibenzo-p-dioxins and dibenzofurans to hexachlorobenzene-induced toxicity, porphyria, changes in mixed function oxygenases, and histopathological changes. Toxicol Appl Pharmacol 1978; 46(3):633–49. 50. Kuratsune M, Yoshimura T, Matzuzaka J, Yamaguchi A. Epidemiologic study on Yusho, a poisoning caused by ingestion of rice oil contaminated with a commercial brand of polychlorinated biphenyls. Environ Health Perspect 1972; 1:119–28.

Occupational and Environmental Hepatotoxicity

767

51. Fouts JR. Interactions of chemicals and drugs to produce effects on organ function. In: Lee DHK, Koten P, eds. Multiple Factors in the Causation of Environmentally Induced Disease. New York: Academic Press, 1972:109–18. 52. Mitchell JR, Gillette JR. Drug-chemical interactions as a factor in environmentally induced disease. In: Lee DHK, Koten P, eds. Multiple Factors in the Causation of Environmentally Induced Disease. New York: Academic Press, 1972:119–31. 53. Poland A, Knutson JC. 2,3,7,8-Tetrachlorodibenzo-p-dioxin and related halogenated aromatic hydrocarbons: examination of the mechanism of toxicity. Annu Rev Pharmacol Toxicol 1982; 22:517–54. 54. Hardin BL, Jr. Carbon tetrachloride poisoning; a review. Ind Med Surg 1954; 23(3):93–105. 55. Jennings RB. Fatal fulminant acute carbon tetrachloride poisoning. AMA Arch Pathol 1955; 59(3):269–84. 56. Von Oettingen WF. The Halogenated Hydrocarbons of Industrial and Toxicological Importance. Amsterdam: Elsevier, 1964. 57. Recknagel RO, Glende EA, Jr. Carbon tetrachloride hepatotoxicity: an example of lethal cleavage. CRC Crit Rev Toxicol 1973; 2:263–97. 58. Reichert D. Biological actions and interactions of tetrachloroethylene. Mutat Res 1983; 123(3):411–29. 59. Reuber MD, Glover EL. Cirrhosis and carcinoma of the liver in male rats given subcutaneous carbon tetrachloride. J Natl Cancer Inst 1970; 44(2):419–27. 60. Drill VA. Hepatotoxic agents; mechanism of action and dietary interrelationship. Pharmacol Rev 1952; 4(1):1–42. 61. Rouiller CH. Experimental toxic injury of the liver. In: Rouiller CH, ed. The Liver, Vol. II. New York: Academic Press, 1964:335–476. 62. Schoental R. Liver disease and “natural” hepatotoxins. Bull World Health Organ 1963; 29:823–33. 63. Gurney R. Tetrachloroethane intoxication. Early recognition of liver damage and means of prevention. Gastroenterology 1943; 1:1112. 64. Jeney E, Bartha F, Kondor L, Szendrei S. Prevention of industrial tetrachloroethane intoxication. Part 3. Egeszegtudmany 1967; 1:155–64. 65. Horiuchi K, Moriguchi S, Monimoto K, et al. Studies on industrial tetrachloroethane poisoning. Osaka City Med J 1962; 8:29–38. 66. WHO. Environmental Health Criteria 139. Partially Halogenated Chlorofluorocarbons (Ethane Derivatives). Geneva: WHO, 1992. 67. Hoet P, Graf ML, Bourdi M, et al. Epidemic of liver disease caused by hydrochlorofluorocarbons used as ozone-sparing substitutes of chlorofluorocarbons. Lancet 1997; 350(9077):556–9. 68. Harris JW, Jones JP, Martin JL, et al. Pentahaloethane-based chlorofluorocarbon substitutes and halothane: correlation of in vivo hepatic protein trifluoroacetylation and urinary trifluoroacetic acid excretion with calculated enthalpies of activation. Chem Res Toxicol 1992; 5(5):720–5. 69. Malley LA, Carakostas M, Hansen JF, Rusch GM, Kelly DP, Trochimowicz HJ. Two-year inhalation toxicity study in rats with hydrochlorofluorocarbon 123. Fundam Appl Toxicol 1995; 25(1):101–14. 70. Urban G, Speerschneider P, Dekant W. Metabolism of the chlorofluorocarbon substitute 1,1-dichloro2,2,2-trifluoroethane by rat and human liver microsomes: the role of cytochrome P450 2E1. Chem Res Toxicol 1994; 7(2):170–6. 71. Malley LA, Carakostas M, Elliott GS, et al. Subchronic toxicity and teratogenicity of 2-chloro-1,1,1,2tetrafluoroethane (HCFC-124). Fundam Appl Toxicol 1996; 32(1):11–22. 72. Marit GB, Dodd DE, George ME, Vinegar A. Hepatotoxicity in guinea pigs following acute inhalation exposure to 1,1-dichloro-2,2,2-trifluoroethane. Toxicol Pathol 1994; 22(4):404–14. 73. Lind RC, Gandolfi AJ, Hall PD. Biotransformation and hepatotoxicity of HCFC-123 in the guinea pig: potentiation of hepatic injury by prior glutathione depletion. Toxicol Appl Pharmacol 1995; 134(1):175–81. 74. Dekant W. Toxicology of chlorofluorocarbon replacements. Environ Health Perspect 1996; 104(Suppl. 1):75–83. 75. Lukaszewski T. Acute tetrachloroethylene fatality. Clin Toxicol 1979; 15(4):411–5. 76. Hake CL, Stewart RD. Human exposure to tetrachloroethylene: inhalation and skin contact. Environ Health Perspect 1977; 21:231–8. 77. Meckler LC, Phelps DK. Liver disease secondary to tetrachloroethylene exposure. A case report. JAMA 1966; 197(8):662–3. 78. Stewart RD. Actue tetrachloroethylene intoxication. JAMA 1969; 208(8):1490–2. 79. Stewart RD, Gay HH, Erley DS, et al. Human exposure to tetrachloroethylene vapor. Arch Environ Health 1961; 2:516–22. 80. National Cancer Institute. Bioassay of Tetrachloroethylene for Possible Carcinogenicity. NCI TR 13. DHEW (NIH) Pub No. 77-813: NTIS Pub No PB-272-940. Springfield, VA: National Technical Service. 81. Thomas LB, Popper H, Berk PD, Selikoff I, Falk H. Vinyl-chloride-induced liver disease. From idiopathic portal hypertension (Banti’s syndrome) to Angiosarcomas. N Engl J Med 1975; 292(1):17–22.

768

Tolman and Dalpiaz

82. Li Y, Zhou M, Marion MJ, Lee S, Brandt-Rauf PW. Polymorphisms in glutathione S-transferases in French vinyl chloride workers. Biomarkers 2005; 10(1):72–9. 83. Johnson MD. Letter from Medical Director of Chemstrand Corportation to the TLV Committee, March 1961. 84. Corsi GC. On occupational pathology caused by dimethylacetmide (with special reference to hepatic function). Med Lav 1971; 62(1):28–42. 85. Spies GJ, Rhyne RH, Jr., Evans RA, et al. Monitoring acrylic fiber workers for liver toxicity and exposure to dimethylacetamide. 1. Assessing exposure to dimethylacetamide by air and biological monitoring. J Occup Environ Med 1995; 37(9):1093–101. 86. Kelly DP. Subchronic inhalation toxicity of dimethyl acetamide in rats. Toxicologist 1984; 4:65. 87. Horn HJ. Toxicology of dimethylacetamide. Toxicol Appl Pharmacol 1961; 3:12–24. 88. Clayton JW, Jr., Barnes JR, Hood DB, Schepers GW. The inhalation toxicity of dimethylformamide (DMF). Am Ind Hyg Assoc J 1963; 24:144–54. 89. Craig DK, Weir RJ, Wagner W, Groth D. Subchronic inhalation toxicity of dimethylformamide in rats and mice. Drug Chem Toxicol 1984; 7(6):551–71. 90. Lundberg I, Lundberg S, Kronevi T. Some observations on dimethylformamide hepatotoxicity. Toxicology 1981; 22(1):1–7. 91. Wiles JS, Narcisse JK, Jr. The acute toxicity of dimethylamides in several animal species. Am Ind Hyg Assoc J. 1971; 32(8):539–45. 92. Kennedy GL, Jr., Sherman H. Acute and subchronic toxicity of dimethylformamide and dimethylacetamide following various routes of administration. Drug Chem Toxicol 1986; 9(2):147–70. 93. Kennedy GL, Jr. Biological effects of acetamide, formamide, and their monomethyl and dimethyl derivatives. Crit Rev Toxicol 1986; 17(2):129–82. 94. Massman W. Toxicological investigations on dimethylformamide. Br J Ind Med 1956; 13:51–4. 95. Wang JD, Lai MY, Chen JS, et al. Dimethylformamide-induced liver damage among synthetic leather workers. Arch Environ Health 1991; 46(3):161–6. 96. Potter HP. Dimethylformamide-induced abdominal pain and liver injury. Arch Environ Health 1973; 27(5):340–1. 97. Buylaert W, Calle P, De Paepe P, et al. Hepatotoxicity of N,N-dimethylformamide (DMF) in acute poisoning with the veterinary euthanasia drug T-61. Hum Exp Toxicol 1996; 15(8):607–11. 98. Poelmans L, van Besien B, Van Ganse W, Deprez A. Toxic hepatitis due to dimethylformamide: case reports and literature review. Acta Clin Belg 1996; 51(5):360–6. 99. Redlich CA, West AB, Fleming L, True LD, Cullen MR, Riely CA. Clinical and pathological characteristics of hepatotoxicity associated with occupational exposure to dimethylformamide. Gastroenterology 1990; 99(3):748–57. 100. Vettorazzi G. Toxicological decisions and recommendations resulting from the safety assessment of pesticide residues in food. CRC Crit Rev Toxicol 1975; 4(2):125–83. 101. Von Oettingen WF. The Halogenated Aliphatic, Olephinic, Cyclic, Aromatic and Aliphatic–Aromatic Hydrocarbons Including the Halogenated Insecticides: Their Toxicity and Potential Dangers U.S. Department H.E.W. Washington, DC: U.S. Government Printing Office, 1995. 102. Beloskurshaya GI, Korolchuk EI. Time course of the liver aspect enzymogram in workers engaged in phosphorus production and in patients with chronic phosphorus poisoning after hypobaric oxygenation. Gigiens Trunda I Professionalye Zabolevani. July 11, 1984. 103. Smith NJ. Death following accidental ingestion of DDT. JAMA 1948; 136:469–71. 104. Morgan DP, Roan CC. Liver function in workers having high tissue stores of chlorinated hydrocarbon pesticides. Arch Environ Health 1974; 29(1):14–7. 105. Tomatis L. The IARC program on the evaluation of the carcinogenic risk of chemicals to man. Ann NY Acad Sci 1976; 271:396–409. 106. Timbrell JA, Scales MD, Streeter AJ. Studies on hydrazine hepatotoxicity. 2. Biochemical findings. J Toxicol Environ Health 1982; 10(6):955–68. 107. Weisberger EK. Halogenated substances: environmental and industrial materials. In: Khan MAQ, Stanton RH, eds. Toxicology of Halogenated Hydrocarbons: Health and Ecological Effects, Vol. 3. New York: Pergamon Press, 1981:22. 108. Ecobichon DJ. Toxic effects of pesticides. In: Klaassen CD, ed. Cassarettand Doull’s Toxicology: The Basic Science of Poisons. 5th ed. New York: McGraw-Hill, 1996:643–90. 109. Wade N. Viets and vets fear herbicide health effects. Science 1979; 204(4395):817. 110. Sterling TD, Arundel A. Review of recent Vietnamese studies on the carcinogenic and teratogenic effects of phenoxy herbicide exposure. Int J Health Serv 1986; 16(2):265–78. 111. Council on Scientific Affiars of the AMA. The Health Effects of “Agent Orange” and Poly-chlorinated Dioxin Contaminants. Chicago, IL: American Medical Association, 1981. 112. Jirasek L, Kalensky J, Kubec K, Pazderova J, Lukas E. Chloracne, porphyria cutanea tarda, and other poisonings due to the herbicides. Hautarzt 1976; 27(7):328–33. 113. Pazderova-Vejlupkova J, Lukas E, Nemcova M, Pickova J, Jirasek L. The development and prognosis of chronic intoxication by tetrachlordibenzo-p-dioxin in men. Arch Environ Health 1981; 36(1):5–11.

Occupational and Environmental Hepatotoxicity

769

114. Guzelian PS. Comparative toxicology of chlordecone (Kepone) in humans and experimental animals. Annu Rev Pharmacol Toxicol 1982; 22:89–113. 115. Blanke RV, Fariss MW, Griffith FD, Jr., Guzelain PS. Analysis of chlordecone (Kepone) in biological specimens. J Anal Toxicol 1977; 1:57. 116. Huggett RJ, Bender ME. Kepone in the James River. Environ Sci Technol 1980; 14:918–23. 117. Huggett RJ, Nichols MM, Bender ME. Kepone contamination of the James River Estuary. Contam Sediments 1980; 1:33. 118. Sterrett FS, Boss CA. Careless Kepone. Environment 1977; 19:30. 119. The Kepone catastrophe. Washington Post, December 21, 1975. 120. Larson PS, Egle JL, Jr., Hennigar GR, Lane RW, Borzelleca JF. Acute, subchronic, and chronic toxicity of chlordecone. Toxicol Appl Pharmacol 1979; 48(1 Pt 1):29–41. 121. Bullivant CM. Accidental poisoning by paraquat: report of two cases in man. Br Med J 1966; 5498:1272–3. 122. Carson DJ, Carson ED. The increasing use of paraquat as a suicidal agent. Forensic Sci 1976; 7(2):151–60. 123. Bus JS, Aust SD, Gibson JE. Superoxide- and singlet oxygen-catalyzed lipid peroxidation as a possible mechanism for paraquat (methyl viologen) toxicity. Biochem Biophys Res Commun 1974; 58(3):749–55. 124. Mullick FG, Ishak KG, Mahabir R, Stromeyer FW. Hepatic injury associated with paraquat toxicity in humans. Liver 1981; 1(3):209–21. 125. Wogan GN. Mycotoxins and liver injury. In: Gall EA, Mostofi FK, eds. The Liver. Baltimore, MD: Williams & Wilkins, 1973:161–81. 126. Swarbock O. Disease of turkey poults. Vet Rec 1960; 72:671. 127. Smith KM. Disease of turkey poults. Vet Rec 1960; 72:652. 128. Goldblatt LA, ed. Aflatoxin: Scientific Background: Control and Implications. New York: Academic Press, 1969. 129. Newberne PM. Chemical carcinogenesis: mycotoxins and other chemicals to which humans are exposed. Semin Liver Dis 1984; 4(2):122–35. 130. Butler WH. Liver injury induced by aflatoxin. Prog Liver Dis 1970; 3:408–18. 131. Newberne PM, Butler WH. Acute and chronic effects of aflatoxin on the liver of domestic and laboratory animals: a review. Cancer Res 1969; 29(1):236–50. 132. Alpert E, Serck-Hanssen A, Rajagopolan B. Aflatoxin-induced hepatic injury in the African monkey. Arch Environ Health 1970; 20(6):723–8. 133. Anonymous. IRAC Monograph on the Evaluation of Carcinogenic Risks of Humans. Vol. 56. Lyon, France: International Agency for Research on Cancer, 1993. 134. Wogan GN. Aflatoxins as risk factors for hepatocellular carcinoma in humans. Cancer Res 1992; 52(Suppl. 7):2114s–8s. 135. Montesaano R, Kirby GM. Chemical carcinogenesis in human liver cancer. In: Brechot C, ed. Primary Liver Cancer: Etiological and Progression Factor. Boca Raton, FL: CRC Press, 1994:57–77. 136. Wogan GN, Edwards GS, Newberne PM. Structure-activity relationships in toxicity and carcinogenicity of aflatoxins and analogs. Cancer Res 1971; 31(12):1936–42. 137. Pong RS, Wogan GN. Time course and dose-response characteristics of aflatoxin B1 effects on rat liver RNA polymerase and ultrastructure. Cancer Res 1970; 30(2):294–304. 138. Eaton DL, Gallagher EP. Mechanisms of aflatoxin carcinogenesis. Annu Rev Pharmacol Toxicol 1994; 34:135–72. 139. Hsu IC, Metcalf RA, Sun T, Welsh JA, Wang NJ, Harris CC. Mutational hotspot in the p53 gene in human hepatocellular carcinomas. Nature 1991; 350(6317):427–8. 140. Fletcher GF, Galambos JT. Phosphorus poisoning in humans. Arch Intern Med 1963; 112:846–52. 141. Diaz-Rivera RS, Collazo PJ, Pons ER, Torregrosa MV. Acute phosphorus poisoning in man: a study of 56 cases. Medicine (Baltimore) 1950; 29(4):269–98. 142. Rodriguez-Iturbe B. Acute yellow-phosphorus poisoning. N Engl J Med 1971; 284(3):157. 143. McCarron MM, Gaddis GP, Trotter AT. Acute yellow phosphorus poisoning from pesticide pastes. Clin Toxicol 1981; 18(6):693–711. 144. LaDue JS, Schenken JR, Kuker LH. Phosphorus poisoning. A report of sixteen cases with repeated liver biopsies in a recovered case. Am J Med Sci 1944; 208:223. 145. Cotrim HP, Freitas LA, Parana R, et al. Non alcoholic steatohepatitis (NASH) and environmental toxins: a liver disease in workers from an industrial area. Hepatology 1996; 24:337 (abstract). 146. Cotrim HP, Parana R, Portugal M, et al. Non-alcoholic steatohepatitis (NASH) and industrial toxins: follow up of patients removed from one industrial area. Hepatology 1997; 26:149 (abstract). 147. Tolman KG, Sirrine R. Industrial hepatotoxicity. Clin Liver Dis 1998; 2(3):563–89.

PART IV

REGULATORY PERSPECTIVES

36

Regulatory Perspectives John R. Senior

Office of Surveillance and Epidemiology, Center for Drug Evaluation and Research, Food and Drug Administration, Silver Spring, Maryland, U.S.A.

INTRODUCTION Public Concerns About Drug Safety Much has happened in the field of drug-induced liver injury (DILI) during the several years since the previous chapter (1) on Regulatory Perspectives was drafted for the first edition of this book. Both within the Food and Drug Administration (FDA, also called the Agency) and in the broader communities of medicine and society, perspectives have changed and are continuing to do so. Growing concerns are being discussed about the safety of drug products, the balance of benefits and risks of treatment, and the adequacy of current systems for protecting the people used by the FDA, as reported in March 2006 by the Government Accountability Office (2). A committee of the Institute of Medicine has been holding hearings since the summer of 2005, and is preparing to report to Congress in the latter half of 2006 on what it recommends for legislative action (3). Drug-Induced Liver Disease as a Model This volume focuses on drug-induced liver disease, an increasing problem as our aging world population is exposed to more and more drugs, both prescribed and self-purchased as over-the-counter medications, in addition to environmental chemicals and self-selected and consumed so-called dietary supplements that may have considerable drug-like effects, special diets, alcohol, and illicit agents. We appreciate that drugs may injure other organs and tissues than the liver (such as brain, heart, kidney, muscle, skin, and others), but the liver is a frequent target for drug-induced injury and is a useful model for the more general problem. We focus here on what regulatory processes are being used to protect the people, what problems and issues must be addressed, and what authorization, funding, and research are needed. There are several groups of issues, each of which will be addressed in the sections that follow: detection and evaluation of DILI; reporting it to the Agency; attribution of causality as really drug induced or not; attempts to understand and elucidate the mechanisms by which the injuries were produced and the host responses to the injuries, and how best to minimize risk and Comments made do not reflect official Agency policies or positions, but are personal opinions based on experiences of the writer at the FDA, and upon other experiences derived for many years before from the practice of medicine, academic research, and teaching (20 years), executive positions in pharmaceutical companies (5 years), and corporate consulting (11 years). Since then the writer has served the FDA for 11 years: 5 years as medical reviewer of gastrointestinal drug product applications and 6 years as a consulting hepatologist to the Agency in assessing possible drug-induced liver injury cases reported from the 15 to 17 review divisions engaged in evaluation of new drugs and biologics for their potential approval for clinical use.

772

Senior

balance drug efficacy and safety. This revised chapter will review briefly the regulatory processes set forth previously, but will focus on recent changes and reasons for them. Background of Special FDA Interest in DILI An initial impetus arose at the FDA when we recognized in mid-1998 a need to devote attention to the problem of drugs as both substrates for metabolism by the liver and conversely upon their metabolites as causes for possible injury to the liver cells. This gave rise to a teaching course at the Center for Drug Evaluation and Research (CDER) on “Drugs and the Liver: What They Do To Each Other,” attended by 325 reviewers from various disciplines (medicine, chemistry, toxicology, pharmacology, and statistics) when offered in April 1999, and again in November of that year by another 75 reviewers (4). This occurred at a time when public and FDA concern was rising about the need for risk assessment, communication, and management (5) of the problems of drug-induced injuries. Convergence of the CDER course and Agency interest in risk assessment led to broadening involvement to additional constituencies, including the pharmaceutical industry, represented by the Pharmaceutical Research and Manufacturers of America, and the academic community, represented by the American Association for the Study of Liver Diseases. We made plans in the summer and fall of 2000 for a national conference that resulted in a two-day meeting in February 2001 on “DrugInduced Liver Injury: A National and Global Problem” (6). Interested persons from each of those constituencies have formed themselves into a Hepatotoxicity Steering Committee that has continued to talk together by telephone two or three times a year and to meet annually for presentations of papers on issues and ideas pertinent to the subject, most recently in late January 2007 (7). A parallel interest in DILI has been developed by the Liver Disease Research Branch of the National Institute of Diabetes and Digestive and Kidney Diseases, which hosted in October 2000 a conference (8) on mechanisms of and test systems for drug-induced liver injury, and then funded a consortium of five university medical centers (9) as a drug-induced liver injury network (DILIN) for 2003–2007 (Table 1). CURRENT FDA PROCESSES FOR EVALUATION OF NEW DRUGS The work of the FDA CDER, and parallel centers for biologics, devices, veterinary products, foods, and toxicology research, is not well understood by most practicing or academic physicians, and even less by patients and the general public. Yet that work is of great importance and interest to all of them, as evidenced by the daily plethora of articles in the press, expressing strong opinions on perceptions of it. The FDA is charged with regulation of products that constitute almost a quarter of all consumer expenditures (the gross domestic product), well over a trillion dollars worth of goods annually. It attempts to do this on a budget of $1.95 billion for the current fiscal year 2007, and to deal with many very difficult issues that TABLE 1 A Chronology of Special Conferences on Drug-Induced Liver Injury Drugs and the liver: what they do to each other Drug-induced liver injury: mechanisms and test systems Drug-induced liver injury: a national and global problem Drug-induced liver injury network Hepatotoxicity steering group: annual meetings Drug-induced liver injury

FDA/CDER course for NDA reviewers NIH/NIDDK course FDA/CDER, AASLD, and PhRMA NIH/NIDDK funded program 2003 2007 See FDA public website (10) AASLD single topic conference

April and November 1999, Gaithersburg and Rockville, MD October 2000, Lister Hill Center, Bethesda, MD February 2001, Westfields Conference Center, Chantilly, VA Universities of CT, NC, IN, MI, CA (UCSF), and Duke Rockville and Bethesda, MD, 2003, 2004, 2005, 2006 September 2005, Emory University, Atlanta, GA

Abbreviations: AASLD, American Association for the Study of Liver Diseases; CDER, Center for Drug Evaluation and Research; FDA, Food and Drug Administration; NDA, new drug application; NIH, National Institutes of Health; NIDDK, National Institute of Diabetes and Digestive and Kidney Diseases; PhRMA, Pharmaceutical Research and Manufacturers of America.

Regulatory Perspectives

773

involve science but transcend it into societal, ethical, and moral nuances. The human drug portion of the 2007 budget is $535 million, of which $225 million was appropriated and $224 million was derived from user fees (11) to support 1105 of the 2382 employees at CDER. Contrary to some popular beliefs, the FDA does relatively little research itself; the great bulk of the expenses are borne by the pharmaceutical industry that discovers, develops, and markets new drugs. The process often fails, as drug candidates are found to be ineffective or toxic, and total costs per approved new drug are estimated now to range from $0.5 to $2 billion (12). The major activities at CDER are oriented toward careful evaluation of data gathered, analyzed, and submitted by the pharmaceutical sponsors as new drug applications (NDAs). The clinical data are gathered by the sponsoring companies under rigorous rules applied to study investigational new drugs (INDs), providing exemptions from the laws of 1938 and 1962 that require drugs to be safe and effective before they are allowed to be administered to people. After chemical and animal preclinical work, candidate drugs that continue to look promising may be considered for study in humans under IND regulations. Such human studies are usually carried out in three phases: (i) careful dose ranging and tolerance or safety studies in modest numbers of healthy volunteers or stable patients; (ii) initial treatment studies in patients with the target disease to find efficacious yet safe doses; and (iii) much larger clinical trials in the specific type of patients intended to be treated, using the formulations and regimens intended to be recommended for clinical use after approval and marketing. The conduct of these clinical studies grows increasingly expensive with advancing phases, and the amount of data accumulated may be enormous. Review and evaluation of the submitted data by chemists, toxicologists, clinical pharmacologists, statisticians, and physicians is a complex process that takes many months, although the increased supplemental funding and additional reviewers provided by the user fees authorized under the 1992 Prescription Drug User Fee Act (PDUFA) did succeed in reducing the total review time and eliminating the backlog of NDAs awaiting review. This law enabled the FDA to charge fees to pharmaceutical companies for reviewing and evaluating their NDAs for possible approval. Although the 1992 Act was intended to provide supplementary funding to the FDA to recruit additional reviewers, it expired after five years. Subsequent Congresses in 1997 and 2002 used the revised laws to substitute the pharmaceutical corporate fees for Congressional appropriations, so that in fiscal 2007 money from PDUFA fees provided almost 50% of total CDER funding (11). Preclinical and Nonclinical Research and Development For decades, pharmaceutical firms have screened large numbers of organic compounds for desired pharmacologic properties and absence of toxicity in their search for new drugs, but in recent years they have turned to development of agents targeted to specific receptors and selected by pharmacogenomic studies. New methods also for proteomic and metabonomic assays toward more personalized drug development and less toward mass marketing have promised much but as yet have not delivered, resulting in an actual decline in new molecular entities submitted for NDA evaluation since 1997, despite steadily increasing costs. Concepts arising out of basic research and new understanding lead to drug prototype design and discovery of new molecular entities with desired biological properties. Most of the compounds screened in the initial research and development are discarded because of insufficient efficacy or excessive toxicity seen in animals, and are never administered to humans. Drug candidates that appear promising are then considered for clinical development, under regulated conditions. Studies of INDs in Humans Sponsoring companies or investigators who intend to conduct clinical studies of experimental, unapproved new drugs must apply for exemption from the laws requiring that drugs be approved as safe and effective before they may be placed into interstate commerce, and must conduct the studies under procedures specified in the Code of Federal Regulations (13) for INDs. All investigators of new drugs are well advised to become familiar with these regulations, especially for knowing their responsibilities (13: §312.60–69) and when to submit

774

Senior

safety or adverse event reports (13: §312.32). Review of IND applications by the FDA is carried out by an appropriate medical specialty division of CDER, and must be completed within 30 days of their receipt. The main emphasis of the IND review is on safety to proceed with the initial protocol for first human exposure to the new agent, or to notify the sponsor of imposed clinical hold if there are safety questions. A sponsoring company or individual investigator incurs responsibilities under IND regulations, including: & & & & & & & &

& &

preparation and submission of a protocol for the initial human study of the new drug, review and approval of the initial plan for investigation and continuing studies by an institutional review board, careful record keeping and documentation (maintained for at least two years following approval for marketing), secure control of and accounting for the investigational drug product, selection of qualified coinvestigators, and keeping them well informed, operating only under written protocols, with informed consent of participants, careful monitoring of the progress of the investigations, prompt or immediate reporting of serious adverse events (disabling, life threatening or fatal, prolonging or causing hospitalization, congenital anomaly/birth defect, including important medical events that require intervention to prevent one of the listed events), annual (within two months) reporting of progress, findings, and participants studied, and providing easy access to all study records to FDA inspectors.

Information about studies carried out under IND regulations, even the existence of the IND itself, is confidential, not to be disclosed by the FDA unless publicly acknowledged by the sponsor. However, individual study participants may request from the FDA a copy of their IND safety reports, if any. During the conduct of studies under IND regulations, there is close interaction between the sponsoring company or investigator and the members of the reviewing team of the Division assigned responsibility for the review, based on the nature of the drug and the disease, and the expertise of the Division. Controlled Clinical Trials New drugs are first given to people in carefully increasing doses as “first in man” Phase 1 (13: §312.21) studies to determine the safety and tolerability of the drug in relatively small groups of healthy volunteers or sometimes in selected stable patients. Studies of absorption, distribution, metabolism, and excretion rates are done to establish pharmacokinetic parameters for the drug, its metabolic products or metabolites are determined, cytochrome P450 or other enzymes are identified that catalyze their production, and studies may be done on effects of other drugs that might be used along with the drug on induction or suppression of the various enzyme systems. Additional nonclinical studies may be done concurrently. Somewhat larger studies of patients with the disease to be treated are then carried out as Phase 2 controlled clinical trials in an attempt to find a suitable dose and regimen that is effective in treating the disease but does not cause dose-related adverse effects. The design of these studies generally employs selecting suitable patients according to standardized inclusion and exclusion criteria, randomizing those selected into groups to receive varying doses of the drug or placebo or a control agent, and then comparing effects observed in the several groups. Such randomized, double-blinded, prospective, multiarmed clinical trials in patients have evolved to become the “gold standard” for reducing bias and confounding (14). The number of patients selected to participate in each group is based upon assumptions and estimates of differences in effectiveness outcome measures to achieve statistically significant results. The numbers of study participants are usually insufficient to have any reasonable chance of detecting relatively rare adverse effects. Phase 3 controlled clinical trials are usually the largest in number of patients studied, typically using the final formulation, dose, and regimen intended for clinical use in patients with the specified disease or disorder if the drug is approved. Again the randomized, prospective, blinded, placebo- or alternative treatment-controlled design is used. In order to

775

Regulatory Perspectives

enroll sufficient numbers of patients with the disease or disorder to be treated, such studies are often carried out using multiple centers in geographically separate locations, under a common protocol. Such studies ideally aim to simulate the target patient population to be treated, using the drug formulation and regimen to be advised if the drug is approved for clinical use and marketing, but for practical reasons of avoiding confounding, bias, and too many adverse effects, extensive inclusion and exclusion criteria are still used to select study participants. The Phase 3 studies are the most expensive to carry out, and may involve many hundreds of patients. Even so, almost three times as many patients as the reciprocal of the true incidence rate must be observed to have at least 95% chance of finding at least one with a relatively rare problem. This so-called “rule of three” requires almost 1500 patients per study arm if the true incidence is 1 per 500 and all problems are detected, a simple consequence of the binomial possibility that a person will show the problem or will not show it (Table 2). Thus, rare but serious adverse events may be missed, simply because not enough patients are studied. In some cases, additional studies may be done after approval of the drug for clinical use and marketing. Before market approval, FDA may discuss and agree with a sponsoring company that they obtain supplemental information about different doses or schedules of dosing or duration of treatment, use in other populations or other stages of the disease, or investigate special safety problems. These Phase 4 studies unfortunately are not always carried out promptly or at all, because of company emphasis on marketing, and the FDA is currently not authorized to require the studies to be done. It may not be the best strategy to force companies to do safety studies that may have consequences of restricting or limiting use of the drug and sales thereof. The FDA is currently drafting guidance for industry sponsors of controlled clinical trials for premarketing evaluation of DILI. When completed, the draft guidance document will be announced in the Federal Register as available for comment on the Agency’s guidance page. NDA Review and Approval Within the CDER, there are currently 17 Divisions oriented by medical specialty (cardio-renal, pulmonary, endocrine-metabolic, etc.), to which NDAs may be assigned when the sponsoring company or investigator has accumulated, often in consultation with the Division, sufficient information about a new drug product to merit review and evaluation by the Division’s chemists, pharmacotoxicologists, clinical pharmacologists, microbiologists, statisticians, and medical reviewers. The NDA submitted undergoes preliminary assessment within 60 days to determine whether the information is sufficiently complete and clear to permit definitive review, after which the application is filed by the Agency. The date of filing the NDA is then 60 days after receipt of the application, and begins a 180-day “clock” specified in law (section 505(c) of the 1938 act) for reaching a decision. Details concerning the content of NDAs are specified in the Code of Federal Regulations (15), parts A, B, and C, and details concerning timing of review, dispute resolution, exceptions, hearing procedures, and miscellaneous provisions in parts D, E, and G. TABLE 2 Calculations for the Rule of Three IZ1/i

i

uZ1Li

uN!0.05

N

N/I

10 50 100 500 1000 5000 10,000

0.1 0.02 0.01 0.002 0.001 0.0002 0.0001

0.9 0.98 0.99 0.998 0.999 0.9998 0.9999

0.047101 0.049282 0.049536 0.049937 0.049962 0.049992 0.049999

29 149 299 1497 2995 14978 29956

2.90000 2.98000 2.99000 2.99400 2.99500 2.99560 2.99560

Note: If an event occurs i, or not, u, then uZ1Ki, a binomial possibility. For this to be evaluated for N people, the binomial expression may be given as (iCu)N, which expands to NC1 terms. The first term is iN, the last uN, and the next to last Niu(NK1). The power of i gives the number of occurrences of i, so the chance that all N show the occurrence is iN, and that none show it uN. All N terms that have an i at least one occurrence, and only the last term uN has none. When uN is less than 0.05, the chance that there will be at least one occurrence will be O0.95, or Ni, which approaches 3 as I increases.

776

Senior

Before the PDUFA of 1992, the legally specified time of 180 days was often not sufficient for review teams to complete their work, and extensions were negotiated between the FDA and the applicant. As a result of obtaining and training additional personnel recruited with PDUFA funds, review times were shortened greatly, from medians of two to three years before 1992 (16) to about a year for routine drugs (17) and to six months or less for urgently needed drugs, and the large backlog of NDAs awaiting review was eliminated. In recent years, with increasing concerns for safety, total review times have risen a bit for new molecular entities and new biologic agents never before approved (18), but the six-month time has held for the high priority drug products. NDAs submitted for review usually contain vast amounts of information, formerly in hundreds of volumes (each about two inches thick) of printed information, but recently often as hundreds of megabytes or even many gigabytes of electronic information. Sections of data submissions are assigned to reviewers by discipline (chemistry, pharmacotoxicology, microbiology, clinical pharmacology, statistics, medicine), who work together also as a team. Completed reviews are discussed in multidisciplinary meetings, exchanged, and sometimes combined medical–statistical reviews are prepared. In some cases, separate medical reviews are done for efficacy and for safety. Each review usually concludes with a recommendation for (i) approval, (ii) as approvable with further information obtained, or (iii) not approvable, with reasons and justifications. These recommendations are not binding, but carry considerable weight on the final regulatory decisions made by the Division or Office supervising the overall review. The science and art of NDA review is a learned medical subspecialty, and like a residency or fellowship in a special branch of medicine takes about two or three years to attain competence and skill. Before reaching a final decision, the reviewing Division may seek external advice from academic or practicing physicians or other consultants who serve appointed four-year terms as members of Advisory Committees (ACs), and some who serve ad hoc for single meetings. Such external experts may also have served as consultants to industrial sponsors in the past, and considerable effort is made to avoid conflicts of interest. Service on an AC is only modestly compensated by the FDA, and the principal reasons for doing so are mainly prestige and public service. Meetings of ACs are open to the public and press, and allow time for brief comments from patients, physicians, special interest groups, and others, as well as detailed presentations by the sponsoring company and its consultants, the FDA review Division and its consultants. The members of the AC hear and discuss the presentations, and may also appoint members to prepare presentations as well. The public meetings of ACs are often intense and dramatic, newsworthy and well covered by the press, and are influential on predicted valuation of the company and its share prices. Specific questions for the AC are usually prepared in advance by the FDA review Division, and are voted upon, but although the consensus may carry great weight, AC opinions and recommendations are not binding on the Agency. Before final approval is granted, there are discussions between the Agency and the applicant about the labeling language that describes exactly to whom the drug is to be given, by what dose and route and regimen, for how long, warnings about adverse effects to be expected, and summaries of selected pharmacological and clinical information. The final approved labeling becomes the basis for allowable advertising and promotion of the drug product by the company, and for instructions to physicians on how the drug should be used. Special patient-oriented medication guides may be prepared as well, and patients increasingly are interested in the content of the drug labeling (the full set of information, as distinct from the label on the package or bottle describing the drug name and amount). It should be understood that the labeling is not just a marketing or advertising piece written by the company that manufactures and sells the product, but every word in it must be negotiated, supported by data, justified and agreed to by the FDA on behalf of patients who will take the drug and the physicians who will prescribe it, or recommend it if it is allowed to be sold without prescription. This labeling may be printed in the Physicians Desk Reference (PDR), now available on the Internet. In January 2006, the FDA announced that the content and format of drug labeling for newly approved drugs would be changed to make the information provided in labeling more consistent and clear (19), to begin June 30, 2006.

777

Regulatory Perspectives

The process that began as a confidential interaction between the sponsoring company and FDA in the IND development, even the existence of which could not be disclosed, becomes public and widely reported at AC meetings, which involves external consultants to both parties, as well as some patients and practicing physicians at large. After approval, the CDER reviews of NDA submissions become publicly available on request (20), under the 1966 Freedom of Information Act and its amendment of 1996 to include electronic material. However, the original data from the studies and sponsors’ reports about data submitted for review generally remain confidential property of the sponsor. Only after approval can the sponsoring company market the drug and start its advertising and promotion program. Within the company, it is typical that the primary oversight of the newly approved drug product shifts from the scientific–medical staff to the marketing and sales staff who have a strong drive to maximize sales, gain market share competitively, and return profits to offset development costs. A natural dynamic tension exists between backgrounds of training, experience, and viewpoints of the conservative medical and scientific staff and the enthusiastic, aggressive marketing people. If new safety problems are discovered after marketing the new drug to large numbers of patients, there may be differences of opinion between the two groups that is both understandable and predictable, but the Agency must deal with them. Reluctance by doctors to prescribe the drug, or by patients to accept it, that may result from findings of special studies of drug safety, may generate marketing staff concerns about meeting sales targets. There are many serious questions as to whether they should be asked to conduct postmarketing safety studies that are not perceived to be in their interest. The problem of failure to do postmarketing studies promptly or at all led to inclusion of a provision in the Food and Drug Modernization Act of 1997 that sponsors of approved drugs and biologics must report annually to the FDA on the progress of postmarketing study commitments, to be published by the FDA in the Federal Register (21). The most recent report, for 2005, shows that about 67% of the commitments for postmarketing drug studies have not been started or are delayed, as was reported for 54% of the biologic products. The Critical Path for Drug Development Because of a significant decline over several years in submissions of applications for FDA review of new drugs, biologic agents, and devices, the FDA proposed in March 2004 a process for working with the larger scientific community, called the Critical Path Initiative (22). The critical path for drug development begins with candidate molecule selection for preclinical and controlled clinical studies, and proceeds through progressively more rigorous evaluation steps aimed at eventual approval for clinical use and marketing (Fig. 1).

Basic research

Prototype design, discovery

Preclinical development

Clinical development

Market application

NDA review, approval

Launch, marketing, clinical use

Approval

Critical path

FIGURE 1 The critical path for medical product development. This figure shows an idealized critical path that encompasses the process of development for a drug or biological product or device. Ideas coming out of basic research are used to discover and design new products, which are then tested in animal or in vitro systems and then in humans. Only a small fraction of candidate products survive the more and more rigorous evaluation steps to reach the approval and marketing stage for clinical use. Source: Adapted from Ref. 22.

778

Senior

This initiative addressed problems of how to harness new knowledge, create new tools to get better answers to safety and effectiveness of new products, faster and at lower cost. The FDA is directed by Congress to promote and protect the public health, one responsibility of which is to ensure that safe and effective medications are available to patients, as stated in the second sentence of the FDA mission statement (23): “The FDA is responsible for protecting the public health by assuring the safety, efficacy, and security of human and veterinary drugs, biological products, medical devices, our nation’s food supply, cosmetics, and products that emit radiation. The FDA is also responsible for advancing the public health by helping to speed innovations that make medicines and foods more effective, safer, and more affordable; and helping the public get the accurate, science-based information they need to use medicines and foods to improve their health.” Sequencing the human genome in 2000 raised general hope that a new era in targeted drug development was about to begin, but instead a further slowdown occurred from the 1997 peak of total applications received for review of new molecular entities. Much discussion has taken place of what should be done, how, and by whom. FDA recently published (24) a followup report two years later in March 2006 summarizing programs and proposing a list of 76 critical path opportunities for research and improvements for consideration and funding, several of which are especially pertinent to issues of DILI. Here are several of these opportunities, numbered as in the March 2006 list (24): 1. 9. 20. 30. 31. 41. 44. 46. 47. 52. 73.

Biomarker qualification Markers of disease progression in hepatitis C Modernizing predictive toxicology Improving extrapolation from animal data to human experience Better animal disease and tissue injury models Measuring patient-centered endpoints Development of data standards Identification and qualification of safety biomarkers Virtual control groups in clinical trials Failure analysis Drug metabolism and therapeutic response

Investigators may wish to consult the more detailed descriptions of these listed opportunities to guide their work, or to modify them appropriately. Surveillance of New Drugs After Approval and Marketing The principal method for detecting adverse effects of new drugs that were not discovered during the controlled clinical trials has been the voluntary, spontaneous process of reporting using the MedWatch system (25). More than 3.5 million such reports have been received by the Agency and more pour in at over 1300 per day. They are organized as the adverse events reporting system (AERS), a database maintained by the CDER Office of Surveillance and Epidemiology (OSE, formerly the Office of Drug Safety). Reports may be submitted directly to the FDA via MedWatch online electronically, or by telephone, fax, or mail, by physicians, other health care personnel, patients, family members, and pharmacists. Reports are most frequently submitted by the manufacturers after their sales representatives or others in their employ learn of the event when physicians or others notify them, and such reports are mandatory. Over 90% of the reports entered into AERS are submitted by pharmaceutical companies. It is recognized that voluntary reporting represents a considerable underestimate of the true number of serious drug-induced events (26), but the precise extent of underreporting is very difficult to determine. In addition, the quality of information contained in the reports is highly variable (27), and often is insufficient to allow accurate determination of whether the adverse event was indeed drug induced or not, and if so by which drug or other agent if more than one is possible.

Regulatory Perspectives

779

DILI AND DYSFUNCTION DILI may be mild, moderate, or very severe, and it may be transient and reversible, whether or not administration of the causative drug is stopped or continued, or it may be progressive to serious liver dysfunction, acute liver failure, and either death or need for liver transplantation. Currently, DILI is the leading cause for acute liver failure in patients who are referred to liver transplant centers in the United States (28), exceeding all other causes combined. Therefore, the FDA is particularly concerned about serious DILI, meaning that which is disabling, requires or prolongs hospitalization, or is life threatening. We make a distinction between liver injury and liver dysfunction, and note that increases in serum enzyme activities (including ALT, AST, ALP, GGT, or others) may reflect injury to liver cells but are not measures of overall liver functions. Other commonly done tests, such as serum bilirubin concentration and plasma prothrombin time (or its international normalized ratio, INR), are better measures of certain true liver functions such as clearance of bilirubin from the plasma and protein synthesis. Despite this, the serum enzyme activities are still erroneously referred to as “liver function tests,” but do not measure any known function of the liver. The distinction is not just semantic and trivial, as was pointed out three decades ago by the late Dr. Hyman Zimmerman, who observed in 1978 that “drug-induced hepatocellular jaundice is a serious lesion, with mortality of 10% to 50%.” Hy, as he was known to his many friends, never did specify exactly how he determined whether or not the liver abnormalities were drug induced, what exact levels of serum transaminase rise was needed to determine hepatocellular injury, nor what level of serum TBL elevation was enough to cause jaundice. Nevertheless, he adhered firmly to the central ideas of this observation even to the time of his death in 1999, and he repeated the statement in the second edition of his book (29) on hepatotoxicty from drugs and chemicals published that year. The essence of his observation, based upon his many consultations and encyclopedic recall of the literature, was that injury to enough hepatocytes sufficient to impair overall liver function of clearing bilirubin was clear clinical indication of serious liver injury and dysfunction that was potentially life threatening. Dr. Robert Temple, currently Director for Medical Policy in FDA/CDER, had participated in organizing the Fogarty conference of 1978 on “Hepatotoxicity of Drugs” held at the National Institutes of Health in Bethesda, MD (30) that established by consensus that a 3!ULN elevation of ALT indicated significant liver injury. He was a great admirer of Dr. Zimmerman, found his observation to be true repeatedly over more than two decades, and dubbed it “Hy’s Law.” Over the years, this observation had been used informally at the FDA to assess severity of putative cases of DILI, and to focus attention on them for special scrutiny. We began discussing it more publicly at an FDA internal course (4) for CDER reviewers on “Drugs and the Liver: What They Do To Each Other” in April 1999, with Dr. Zimmerman present. He did not wish to define it more precisely, although he still believed it in principle and expressed the hope that we could validate it using the large amount of data from clinical trials submitted to the FDA for review and evaluation. Over subsequent years we have come to use combined peak observed values of ALTR3!ULN and TBLR2!ULN, not necessarily at the same time (bilirubin tends to rise somewhat later than the ALT in hepatocellular injury), as a compound biomarker to suggest a possible case of serious DILI. It has not, however, been used as an automatic reason to call an investigational drug hepatotoxic, or reject it, in absence of additional clinical information (31–34). The best time to detect and investigate DILI is before drug approval, during controlled clinical trials, which are conducted under protocols and under IND regulations, and study participants are well observed. Outlined below are some suggestions as to how this might be done. Detection of DILI It is not clear whether vague and often nonspecific symptoms of malaise, fatigue, anorexia, nausea, vomiting, and right upper quadrant abdominal discomfort, or of dark urine and light stools in cholestatic cases, precede or follow the initial elevations of serum enzyme activities, and it may vary from person-to-person. The onset of any evidence of DILI may be abrupt or

780

Senior

insidiously slow, and may show a latent period of days to weeks or months after the initial exposure to the drug causing the injury. After a potentially lethal dose of oral acetaminophen ingestion, there is often a period of no symptoms for 24 to 48 hours that may follow some immediate gastric discomfort. But the towering levels of serum transaminase activities that appear then suggest that indeed something very damaging to hepatocytes has been going on during that apparent latent period. For most drugs, we know very little about exact chronology of the onset of DILI, and it seems that more attention should be paid to even such vague symptoms, rather than simply waiting for serum transaminase elevations to be discovered at some preset monitoring interval. Patients or study participants can monitor themselves daily, at no cost, and can be taught to report such symptoms immediately, when they have been exposed to a new drug or combination. This approach was used quite successfully by Nolan and colleagues (35), who carefully instructed patients starting on INH prophylaxis to watch out for symptoms, reminding them at every visit to do so daily and report immediately, with no mortality and only 1 per 1000 patients having to stop INH permanently. For asymptomatic onset, we rely upon the preset monitoring of ALT or AST to detect hepatocellular injury, and upon ALP and GGT to detect cholestatic problems, looking for a doubling of the enzyme activity above the ULN. But there is a practical limit on how often such serum enzyme activity can be measured, how much blood can be drawn, how much the study participant will tolerate, and the costs of the measurements, most of which will be negative or within the normal range. It may be advisable to use both symptoms and enzyme monitoring, and to act upon whichever first may show abnormalities. Confirmation of Liver Injury When either symptoms or elevated serum enzyme activities are reported, it is advisable to confirm immediately, within 48 hours, whether the findings are valid, and whether increasing or disappearing. In many studies, participants are outpatients and not resident at the study site, so confirmation will have to be carried out by their local physicians and laboratories. At this time of prompt confirmation, the person should be questioned for symptoms, examined for physical findings, and a full set of serum tests done, including ALT, AST, ALP, TBL, INR, adjusted for any differences in normal ranges of test results from those of the study central laboratory, if there is one. Communication of findings to the principal investigator should be immediate. If impending liver injury is confirmed by this prompt recheck, it is important to note whether the problem is evolving slowly or rapidly in either direction, but of special concern is rapid worsening. Study participants, investigators, and personal physicians should be briefed ahead of time about what to do. Close Observation of Evolving Injury When apparent liver injury is detected and confirmed, either after symptoms or abnormal test results, the critical concern is whether the injury will be self-limited and will disappear, or will progress to serious liver injury with dysfunction and possible liver failure. The liver is an amazingly robust organ, with many defense and healing mechanisms, is accustomed to dealing with both internal and external chemical substances by metabolizing them, transporting them out of hepatocytes, excreting them into the bile, making more water-soluble derivatives for renal system excretion, and in a few instances volatile derivatives for exhalation. Further, the liver can heal itself after injury, and may initiate vigorous regenerative processes. We do not know at present how to evaluate if a given person will respond to injury from a given drug, i.e., how to predict the outcome. It has become appreciated increasingly that mild liver injury is usually overcome by hepatic defense and healing mechanisms, that transient rises in serum enzyme activities are not reliably predictive of more serious liver injury and dysfunction, even though the offending drug continues to be given. The principal variable may be more the person receiving the drug than either the drug, dose, or duration administered. Mild or moderate injury will resolve by these adaptive processes in most people, but in a few cases the person is so susceptible or so unable to adapt that progressive injury follows. A great and still unresolved problem is when

781

Regulatory Perspectives

the drug administration should be stopped, for there usually is no available treatment for serious DILI other than discontinuation of the drug that caused the injury and general supportive care. Despite earlier opinions, stopping administration of the offending drug, also called “dechallenge,” often does not lead to immediate improvement, but serum enzymes may continue to rise for days or weeks, and serum TBL or plasma INR may begin to rise to signal functional impairment (36). Because drug-induced hepatocellular injury with jaundice is a serious disorder, possibly life threatening in 10% or more of patients who show it, we have taken rises of either serum transaminase to 3!ULN or more plus rise in TBL to R2!ULN as signals that the drug should be stopped immediately (37). Despite this, there have been cases of such injury caused by isoniazid in which the drug administration was not stopped, and the patients adapted successfully even to such serious injury and became tolerant of the drug (38), as shown in Figure 2. Close observation means just that, and in confirmed cases of liver injury should include laboratory retesting every two or three days, careful interview for details of medical history and observation of physical findings, search for possible other causes of the clinical findings than the suspected drug, and consultation by a gastroenterologist or hepatologist. It is essential to determine whether the liver injuries are getting better or worse, and the pace of change. It should be realized that falling serum enzyme levels are not always indications of recovery, but may occur as fewer hepatocytes are left to leak enzymes into the blood, especially if function measures such as bilirubin and prothrombin become perturbed and findings of encephalopathy, renal failure, and metabolic derangements appear. In such cases, falling transaminase levels are an ominous sign, not an encouraging one. In most clinical trials, such cases that require this careful but costly close observation either will not be seen at all or will be quite rare. If not, the investigational drug may be of dubious value indeed, and it would be better to discover this sooner rather than later. Whether hospitalization will be advisable, or liver biopsy attempted, should be determined on clinical grounds. Liver biopsy is not conclusively 100.0

Mitchell, et al., 1975 30.7x

22.1x AST, TBL: log10xUL

14.6x

M49w M61b

M49w M61b

M39b

M39b

10.0 4.3x

3.3x 2.8x

1.0

60

56

52

48

44

40

36

32

28

24

20

16

12

8

4

0

–4

–8

On INH 0.1

Weeks

FIGURE 2 Data for three patients on isoniazid prophylaxis for a year. This figure shows monthly levels of AST activity, in triangles, and TBL concentration, in squares, for three male patients treated with continuous isoniazid (INH) prophylaxis for up to a year in 1974, without stopping the INH despite rises in AST and TBL to levels now considered to represent serious liver injury, fulfilling requirements for Hy’s Law. All three patients adapted to the drug and became tolerant, without progressing to serious liver disease or liver failure. Values for AST and TBL are shown as multiples of the upper limit of the normal range, x. Abbreviations: AST, aspartate aminotransferase; b, Black; w, White; M, male; age, in years; TBL, total bilirubin. Source: From Ref. 38.

782

Senior

diagnostic of DILI, which may mimic any known liver disease, but it may be useful in some cases. In general, most DILI is acute, and hepatocellular injury is more of immediate concern than cholestatic liver injury. The latter tends to develop somewhat more slowly and to be less life threatening, so it is useful to note whether the initial onset of the process is predominantly hepatocellular or cholestatic. A useful measure is the ratio, R, of the relative elevations of ALT/ALP, both expressed in multiples of the upper limit of their normal ranges (!ULN), and both measured at the first detection of a significant rise in either to 2!ULN or more (39). If R isR5 or ALTO2!ULN only, then the initial indication is of hepatocellular injury; if R%2 and ALPO2!ULN, cholestatic; and if 2!R!5 and ALTO2!ULN and ALPOULN, mixed hepatocellular and cholestatic injury. Later in the course of DILI, mixed patterns are often seen. In addition to these two broad categories of liver injury caused by drugs, there are some other patterns of liver injury and response that are notably different, including allergic– immunologic and mitochondrial liver injuries. The anesthetic agent halothane, introduced in the late 1950s as a noninflammable alternative to diethyl ether, was found to induce in some people a sensitization process so that re-administration caused a very severe liver injury, when the first exposure had not, which set off a 25-year controversy and hundreds of papers before it became accepted (40). Other drugs have since shown a similar tendency to produce more severe and more rapid injury on repeated exposure, rather than the more common process of adaptation and development of tolerance. Fialuridine, a new treatment for chronic hepatitis B infection, when administered for more than a few weeks was found to cause injury to liver and other mitochondria, with loss of intracellular energy from oxidative phosphorylation (41). Aspirin administered to children with viral infections also appeared to cause mitochondrial injury, with severe liver injury, metabolic acidosis, and encephalopathy (42). How many more patterns of liver injury we shall identify cannot yet be stated, but it is likely we have not found them all. Beyond the different patterns, or “signatures” of drugs that may injure the liver, the responses of certain individual patients may differ, so that drugs such as the combination of amoxicillin and clavulanic acid (Augmentin) that usually cause cholestatic reactions may in some people cause severe hepatocellular injury (43). Exclusion of Alternative Causes The diagnosis of DILI may be very difficult, and there are no truly pathognomonic findings to reach a definite conclusion; hence the diagnosis is made by exclusion of other possible causes. Acute viral hepatitis A and B should be ruled out by history and serological markers, hepatitis C by history and presence of serum HCAb (but it is less often acute in onset, more likely to be insidious), alcoholic and autoimmune hepatitis by history, physical findings and AST/ALT ratio and serological antibodies, biliary disease by history and ultrasound examination, and ischemic hepatic by history of hypotension or cardiac failure. Beyond these relatively more common liver diseases, it may be unreasonable or prohibitively time-consuming and expensive to search in all cases for hemochromatosis, Wilson’s disease, a1 antitrypsin deficiency, viral hepatides D, E, EBV, CMV, herpes, etc. Unless unusually complete information is available, in many cases, it cannot be concluded that a drug definitely caused the liver injury, and the best that can be done is to say that the drug probably or possibly caused it. This is even more difficult when multiple drugs are taken, more than one of which possibly could have been the cause. However, it should not be concluded that because some alternative drug was taken or some disease was present that might have caused the injury, therefore the drug did not. The problem of possible drug injury added to or aggravating effects of another drug or disease has not been resolved, and differential diagnosis of DILI remains one of the most difficult and challenging problems in medicine (44). Some attempts have been made to clarify the attribution problem, as scoring systems in which points are awarded for various findings. Among the first of these directed specifically toward DILI was the Roussel Uclaf causality assessment method (RUCAM) developed in France and still in use there (45). In the RUCAM, points are awarded for onset of enzyme rises within 5 to 90 days after drug administration is started, for decreasing serum enzymes promptly (by R50% within 8 days after drug administration is stopped, for absence of other

Regulatory Perspectives

783

possible disease causes, for previously published and labeled reports of DILI attributed to the drug, and especially for prompt doubling of ALT on readministration of the drug alone (rechallenge). Points are also subtracted for presence of any other possible cause or alternative causes not excluded entirely, lack of serum enzyme reduction by at least 50% within 30 days, coadministration of a known hepatotoxic agent with compatible time relationship, and negative rechallenge response. In use, the RUCAM system has proved confusing and subject to varying interpretations, and some of the basic assumptions of the RUCAM appear to be flawed (46). Absence of adequate information for the differential diagnosis results in a RUCAM score indicating that DILI is unlikely. Lee (36) has not found that increased age is associated with worse outcome in the Acute Liver Failure Study, and the blanket statement about any use of alcohol in the RUCAM was also unconfirmed. Special Investigations into Susceptibility, Mechanisms A key question is what may be different about people who react adversely to a drug, when most people do not. Is there some difference in susceptibility in them that causes a severe reaction to a drug given in the same dose for the same duration to another person of the same age and gender who shows no ill effect? Or if a mild drug-induced injury to the liver occurs, in which some liver cells are killed and release enzymes into the blood, why do most people recover from this injury by healing responses, but a few people do not? The focus of standard drug development is on differences in the drug and dose administered, and not on differences in the recipients thereof. Such differences may be genetic, or equally likely may be acquired. Animal models do not seem able to solve this problem, and it seems likely that more investigation should be directed toward looking for people who respond differently, using all the new tools we have: genomic, proteomic, and metabonomic. The best time to make such observations is while the adverse process is occurring, for time is a critical variable. Samples of tissue fluids, such as urine and blood, can be obtained prospectively during controlled clinical trials, frozen for later analyses, in the uncommon or rare individuals who show confirmed liver injury possibly induced by a drug. It is reasonable that some matched control persons (for age, gender, dose, and duration of drug exposure) who do not show any liver test abnormalities should also be studied. Duration of drug administration may be a critical variable, as has been observed with bromfenac, trovofloxacin, fialuridine, and ximelagatran. The drug may be tolerated for a few days or weeks, but after longer administration serious DILI may emerge. What is going on, metabolically and immunologically during this period is not yet known. These and many other questions about mechanisms of drug-induced injury, and hepatic responses, need to be investigated while the process is evolving, and cannot be done as well retrospectively. Follow-Up of DILI Cases to Resolution It is important to follow the course of the liver injury and dysfunction to resolution, which may mean to the baseline state existing before the drug was administered, to normality, or to some other endpoint. Nothing is learned by simply stopping the drug when the ALT becomes elevated to R3!ULN and removing the participant from the study, and the practice may be dangerous. We have learned that stopping administration of an offending drug does not always or even usually lead to immediate reversal of the injury process, which may continue to worsen for days or weeks after the drug is withdrawn. It has been observed that most people who show elevations of serum transaminase activities only will resolve, whether or not administration of drug is stopped, as has been learned from long experience with isoniazid and more recently with tacrine and other drugs (47). The liver is a very robust organ, quite capable of coping with xenobiotic substances, as discussed above in section “Close Observation of Evolving Injury.” It is not well understood, however, whether preexisting liver diseases or disorders may interfere with the healing, recovery, and regeneration processes that ordinarily lead to adaptation and later tolerance of the drug (48), even if underlying disease does not increase the chance of sustaining some liver injury. These adaptive processes may lead to an apparent “negative rechallenge” test, and falsely to a conclusion that the drug did not cause the injury.

784

Senior

POST-APPROVAL ISSUES The problems of detecting and managing DILI after approval of a new drug for prescription and marketing are even greater than those in controlled clinical trials. The emphasis of the sponsoring manufacturer is upon promotion, sales, increasing market share, as quickly as possible, and the public assumes that an approved new drug is safe and effective for everyone. Practicing doctors are besieged to prescribe the new drug to as many patients as possible, and prospective patients are recruited by direct-to-consumer advertising. Yet, the new drug may not be safe for everyone who takes it, but quite dangerous for some despite being well tolerated by most people. Detection of Uncommon Adverse Events Uncommon or rare adverse reactions to drugs may not be discoverable during controlled clinical trials, which are limited by cost to reasonable numbers of exposed study participants, and also by reasonable duration of exposure and dose. Marketing zeal may lead to broadening the type of patient well beyond those studied during controlled studies, and observation of patients may in some cases be less intense. Patients may be treated with higher doses or much longer durations than recommended. Detection of new rare but serious adverse effects is less systematic after approval for clinical use, and depends on astute observation and spontaneous, voluntary reporting. It is estimated that only a small fraction of actual events are reported to the MedWatch program (26), and that the quality of the reports is inadequate for making the difficult diagnosis of whether the cases represent DILI or not (27). Problems of Reporting Why is there underreporting of adverse events? There are many reasons, some of which are: & & & & & & & & & & & & & &

Inadequate follow-up or monitoring of the patient (problem not detected), Failure of the patient, or family member, to recognize or report it to the physician, Problem not believed to be severe enough to be “worth reporting,” Sense of guilt, for having caused harm to the patient by prescribing the drug, Fear of lawsuit liability, Erroneous attribution of the events to some other cause than the drug, Lack of reimbursement for the time and effort to report, Procrastination, inertia, unwilling to take the time to report, Reluctance to “get involved” in reporting, Knowledge that others have reported such effects, and that it is not “news,” Denial that the new drug could have caused the effect, Intent to publish the finding, and therefore keep it confidential, Dislike or distrust of the “government,” and Other reasons peculiar to individual potential reporters.

Attribution of Causality as DILI or Not As mentioned in section “Exclusion of Alternative Causes,” the attribution of causality to DILI or disease, or to one drug or another given concomitantly, is among the most difficult of differential diagnoses, even if complete information is available. For most reports of putative DILI postmarketing, decisions on causality are opinion based on inadequate information. Even the DILIN group of leading experts in liver disease, after careful and exhaustive work on the topic of causality assessment, still relies on consensus to make decisions as to the probability that the drug really caused the liver injury. It is the natural tendency of drug manufacturers trying to market new drugs to seek other causes than their drug to explain liver injury and dysfunction that might restrict use of the drug or lower sales, which may bias their opinions and those of their hired consultants. There is great need for a better and more objective method of causality assessment, based on data and clinical evidence rather than upon opinions.

Regulatory Perspectives

785

Epidemiologic, Retrospective, Observational Studies The FDA is responsible for the safety of drugs used by the people of the United States, so that the problems of rare but serious DILI in a few individuals have to be considered in the context of benefits and risks for the population exposed to the drug. It is very difficult indeed to balance modest and perhaps delayed future benefits to many people against severe and immediate risks to a few persons, as experienced in the troglitazone controversy (49) that ranged from 1997 to 2000 and was only resolved when two other drugs (rosiglitazone and pioglitazone) of the same class, with similar efficacy but much less tendency to cause DILI, were approved and used widely. After detection of “signals” of possible DILI, either by reports to MedWatch or by published case reports, the FDA through its OSE (the successor to the previous Office of Drug Safety and Office of Postmarketing Drug Risk Assessment) may mount epidemiologic investigations to seek confirmation and elucidation of the problems by analyses of large databases of clinical information. These databases are often constructed to obtain financial information on the costs of care, laboratory testing, drug use, and other purposes, but can be modified (50,51) to provide better information on the risk factors and incidence rate (observed or diagnosed cases in a given population of people exposed to or prescribed the drug in a given time interval). Reporting Rates Versus True Incidence: Hazard Rates It must be understood that “reporting rates” are not true incidence rates, and may be very far from the truth. As derived from analyses of the AERS database, reported cases are far fewer than the actual number of cases that occur, or the many reasons discussed above. It is also very difficult to know exactly how many people were exposed to how much drug for how long, if only data on prescriptions filled are available. Both the numerator and denominator of that rate ratio are badly flawed and unreliable. But true incidence rates are essential to evaluation of risk management programs to determine whether or not the programs are effective in reducing the incidence of the problem. Only by implementation of prospective safety studies, with known numerators of cases that are detected by adequate methods and known denominators of how many people are exposed to what doses for how long, can true incidence rates be established. After implementation of risk management programs, repeat prospective studies can then establish how effective were the programs intended to reduce the risks. The hazard rate for the occurrence of DILI over time cannot be assumed to be constant, and will probably vary from drug-to-drug, and among people exposed, depending upon the mechanisms of drug-induced injury and counterresponses. As has been noted above, the hazard rate (new cases becoming detectable in a given interval of time per number exposed) appears to increase with more prolonged exposure, for drugs such as bromfenac, trovofloxacin, fialuridine, and ximelagatran. On the other hand, the hazard rate may diminish for many drugs that have been taken for long periods of time without adverse effects, provided that some secondary disease or other drug does not trigger a late reaction. Very little is known about effects of combinations of drugs or drugs–foods, drugs–disease, drugs–other chemicals, in leading to onset of DILI. Risk Management: Processes Versus Results Most of the risk management programs to date have relied on processes that are intended to reduce risks, and measures of adherence to those processes rather than to measures of actual reduction of the incidence. Reporting rates are not sufficiently reliable to evaluate effectiveness of risk management programs, and reliance on processes may be wishful thinking. It seems likely that only by developing ways to determine the true incidence of risks such as serious DILI can this dilemma be resolved. Prospective Safety Studies Prospective safety studies need to be large enough and long enough and intensive enough for sufficient opportunity for rare adverse events to occur, be detected and evaluated, which

786

Senior

means they will be costly and difficult. They are not likely to be just “large, simple studies,” and will require careful design and execution, and special funding. It may be argued that they should not be delegated to a sponsoring company that is at the same time trying to develop and increase its market share for a new product, or even to retain a market for an older product. It would be far preferable to have them done by an agency whose primary interest is in the welfare and safety of the people taking the drug, who should and who should not be exposed to the drug, and the best way to use it, accounting for both the risks and benefits. Ideally, agencies such as the National Institutes of Health (NIH) or the Agency for Healthcare Research and Quality (AHRQ) should carry out such studies, funded by public money rather than by the pharmaceutical industry, and the findings evaluated by the FDA (36). New sources of public money are needed to support this, since appropriations are already inadequate.

WHAT NEEDS TO BE DONE? The liver disease research branch of the National Institute for Diabetes and Digestive and Kidney Diseases (NIDDK) developed and published (52) in December 2004 an action plan for needed research in the field of liver disease over the period from 2005 to 2014, including a chapter on drug- and toxicant-induced liver disease. Short-term (one to three years), mediumterm (four to six years), and long-term (seven to ten years) goals were developed by 16 working groups of national and international experts, reviewed by separate teams of experts. Many of those goals are included in the topics discussed in this chapter, but certain research questions seem especially pertinent to the issues faced and debated in the context of regulatory perspectives for DILI. Among them are: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

Is ALT monitoring effective in preventing serious DILI? Should liver disease symptoms be monitored instead or in addition? How can we know if the DILI will become serious? What should be the rule for stopping drug administration? When should rechallenge be attempted, and how? When is liver biopsy helpful in diagnosis of DILI? How can we be more certain that the problem is DILI and not disease? How can we measure the hazard rate, as constant or increasing or decreasing? Does preexisting or concurrent liver disease affect outcome of DILI? How can we measure the true incidence of DILI? Why are certain people idiosyncratically susceptible to DILI? Can a test be devised to identify the susceptible people before the drug is given? What is happening during the latent period from exposure to drug to onset of DILI? What is the mechanism by which the drug causes the injury? Can the liver be protected against DILI?

Many of these points need to be discussed and debated, by experts in the field and by both patients and physicians. The Institute of Medicine of the National Academy of Sciences has been holding hearings (3) for over the year 2005–2006 and is scheduled to report its recommendations to Congress in late summer 2006. Some of the ideas that Congress possibly might consider include: (i) conditional approval of new drugs for five years; (ii) authorization of prospective safety studies for cause; (iii) new sources of public funding, such as a small tax on sales of regulated products; (iv) third party (NIH, AHRQ) design and execution of prospective studies; and (v) stronger FDA safety review policies and funding support. The debate is beginning, as may be indicated by the Grassley–Dodd bill (53) introduced in April 2005, and the Enzi–Kennedy bill in July 2006 (54). Readers of this volume will be especially interested in following closely the progress of these debates, as well as scientific progress in the field.

Regulatory Perspectives

787

REFERENCES 1. Senior JR. Regulatory perspectives. In: Kaplowitz N, DeLeve L, eds. Drug-Induced Liver Disease. 1st ed. New York: Marcel Dekker Inc., 2003 (Chap. 30). 2. Government Accountability Office. Drug safety: improvement needed in FDA’s postmarket decisionmaking and oversight process. GAO Report 06–402 to Congessional Requestors, March 2006; 1–62. (Accessed 28 July 2006 at http://www.gao.gov/new.items/d06402.pdf). 3. Institute of Medicine. Committee on Assessment of the U.S. Drug Safety System. (Accessed 28 July 2006 at http://www.iom.edu/drugsafety). 4. Drugs and the Liver: What They Do to Each Other. FDA/CDER teaching course, April 1999, University of Maryland Conference Center at Shady Grove, Gaithersburg, MD. (Accessed 28 July at http://www.fda.gov/cder/livertox/courses.pdf). 5. Task Force on Risk Management. Managing the risks from medical product use: creating a risk management framework. Report to the FDA Commissioner, May 1999. (Accessed 28 July at http:// www.fda.gov/oc/tfrm/riskmanagement.pdf). 6. Drug-Induced Liver Injury: A National and Global Problem, 12–13 February 2001, Chantilly, VA. (Accessed 28 July at http://www.fda.gov/cder/livertox/default2001.htm). 7. HepTox Steering Group. 2007 Meeting, 24–5 January. (Accessed 25 April 2007 at http://www.fda. gov/cder/livertox). 8. Bissell DM, Gores GJ, Laskin DL, Hoofnagle JH. Drug-induced liver injury: mechanisms and test systems. Hepatology 2001; 33(4):1009–13. 9. Hoofnagle JH. Drug-induced liver injury network (DILIN). Hepatology 2004; 40(4):773. 10. www.fda.gov/cder/livertox 11. FDA. FY 2007 Budget Summary. All Purpose Tables 79–81. (Accessed 25 April 2007 at http://www. fda.gov/oc/oms/ofm/budget/2007/TOC.htm). 12. Adams CP, Brantner VV. Estimating the cost of new drug development: is it really 802 million dollars? Health Aff 2006; 25(2):420–8. 13. Office of the Federal Register. Part 312—Investigational New Drug Application. 21 Code of Federal Regulations. (Accessed 28 July 2006 at http://ecfr.gpoaccess.gov/; select Title 21, 300–499, 312). 14. Friedman LM, Furberg CD, DeMets DL. Fundamentals of Clinical Trials. 3rd ed. New York: Springer, 1998. 15. Office of the Federal Register. Part 314—Applications for FDA Approval to Market a New Drug. 21 Code of Federal Regulations. (Accessed 28 July 2006 at http://ecfr.gpoaccess.gov/; select Title 21, 300–499, 314). 16. Kaitin KI, Manoccchia M, Seibring M, Lasagna L. The new drug approvals of 1990, 1991, and 1992: trends in drug development. J Clin Pharmacol 1994; 34(2):120–7. 17. Kros JF, Qian J. Analysis of U.S. Food and Drug Administration review intervals for drugs approved during the period 1997–2002. Am J Ther 2004; 11(5):337–43. 18. CDER Report to the Nation: 2004. (Accessed 28 July 2006 at http://www.fda.gov/cder/reports/rtn/ 2004/rtn2004-1.htm). 19. FDA. New Requirements for Prescribing Information. (Accessed 28 July 2006 at http://www.fda. gov/cder/regulatory/physLabel/default.htm). 20. Handbook for Requesting Information and Records from FDA. (Accessed 28 July 2006 at http:// www.fda.gov/opacom/backgrounders/foiahand.html). 21. FDA. Postmarketing Study Commitments. (Accessed 28 July 2006 at http://www.fda.gov/cder/ pmc/default.htm). 22. FDA. Innovation or Stagnation: Challenge and Opportunity on the Critical Path to New Medical Products, 16 March 2004. (Accessed 28 July 2006 at http://www.fda.gov/oc/initiatives/criticalpath/ whitepaper.html). 23. FDA. Mission Statement. (Accessed 28 July 2006 at http://www.fda.gov/opacom/morechoices/ mission.html). 24. FDA. Critical Path Opportunities Report and List, March 2006. (Accessed 28 July 2006 at http:// www.fda.gov/oc/initiatives/criticalpath/; Opportunities Report and List). 25. Ahmad SR. Adverse drug event monitoring at the Food and Drug Administration. J Gen Intern Med 2003; 18(1):57–60. 26. Begaud B, Martin K, Haramburu F, Moore N. Rates of spontaneous reporting of adverse drug reactions in France. JAMA 2002; 288(13):1588. 27. Gary MM, Harrison DJ. Analysis of severe adverse event related to use of mifepristone as an abortifacient. Ann Pharmacother 2006; 40(2):191–7. 28. Larson AM, Polson J, Fontana RJ, et al. Acetaminophen-induced acute liver failure: results of a United States multicenter, prospective study. Hepatology 2005; 42(6):1364–72. 29. Zimmerman HJ. Hepatotoxicity: The Adverse Effects of Drugs and Other Chemicals on the Liver. 2nd ed. Philadelphia: Lippincott Williams & Wilkins, 1999.

788

Senior

30. Davidson CS, Leevy CM, Chamberlayne EC, eds. Guidelines for detection of hepatotoxicity due to drugs and chemicals. Fogarty Conference, 1978, NIH Publication No. 79–313. Washington, DC: U.S. Government Printing Office, 1979. 31. Lewis JH. “Hy’s Law,” the “Rezulin Rule,” and other predictors of severe drug-induced hepatotoxicity: putting risk-benefit into perspective. Pharmacoepidemiol Drug Saf 2006; 15(4):221–9. 32. Kaplowitz N. Rules and laws of drug hepatotoxicty. Pharmacoepidemiol Drug Saf 2006; 15(4):231–3. 33. Senior JR. How can ‘Hy’ s law’ help the clinician? Pharmacoepidemiol Drug Saf 2006; 15(4):235–9. 34. Temple R. Predicting serious hepatotoxicity. Pharmacoepidemiol Drug Saf 2006; 15(4):241–3. 35. Nolan C, Goldberg SV, Buskin SE. Hepatotoxicity associated with isoniazid preventative therapy: a 7-year survey from a public health tuberculosis clinic. JAMA 1999; 281(11):1014–8. 36. Lee WM, Senior JR. Recognizing drug-induced liver injury: current problems, possible solutions. Toxicol Pathol 2005; 33(1):1155–64. 37. Temple R, Senior J. Clinical white paper. Drug-Induced Liver Injury: A National and Global Problem, 12–13 February 2001, Chantilly, VA. (Accessed 28 July at http://www.fda.gov/cder/livertox/ clinical.pdf). 38. Mitchell JR, Long MW, Thorgeirsson UP, Jollow DJ. Acetylationrates and monthly liver function tests during one year of isoniazid preventative therapy. Chest 1975; 68(2):181–90. 39. Danan G. International consensus meeting. Criteria of drug-induced liver disorders. J Hepatol 1990; 11(2):272–6. 40. Neuberger J, Williams R. Halothane hepatitis. Dig Dis 1988; 6(1):52–64. 41. McKenzie R, Fried MW, Sallie R, et al. Hepatic failure and lactic acidosis due to fialuridine (FIAU), an investigational nucleoside analogue for chronic hepatitis B. N Engl J Med 1995; 333(17):1099–105. 42. Linnemann CJ, Jr., Shea L, Partin JC, Schubert WK, Schiff GM. Reye’s syndrome: epidemiologic and viral studies, 1963–1974. Am J Epidemiol 1975; 101(6):517–26. 43. Fontana RJ, Shakil AO, Greenson JK, Boyd I, Lee WM. Acute liver failure due to amoxicillin and amoxicillin/clavulanate. Dig Dis Sci 2005; 50(10):1785–90. 44. Seeff L. The problem of establishing causality. HepTox Steering Group. 2006 Meeting, North Bethesda, 25–26 January. (Accessed 28 July 2006 at http://www.fda.gov/cder/livertox/presentations2006/ seeff.htm). 45. Danan G, Be´nichou C. Causality assessment of adverse reactions to drugs—I. A novel method based on the conclusions of international consensus meetings: application to drug-induced liver injuries. J Clin Epidemiol 1993; 46(11):1323–30. 46. Freston J. Use and limitations of the RUCAM. HepTox Steering Group. 2006 Meeting, North Bethesda, 25–26 January. (Accessed 28 July 2006 at http://www.fda.gov/cder/livertox/presentations2006/ freston.htm). 47. Senior JR. Adaptation to liver injury: tacrine, isoniazid, ethanol, experimental drugs. HepTox Steering Group. 2006 Meeting, North Bethesda, 25–26 January. (Accessed 28 July 2006 at http://www.fda. gov/cder/livertox/presentations2006/senior.htm). 48. Mehendale H. Liver tissue repair, survival factors, and adaptation to injury. HepTox Steering Group. 2006 Meeting, North Bethesda, 25–26 January. (Accessed 28 July 2006 at http://www.fda.gov/cder/ livertox/presentations2006/mehendale.htm). 49. Watkins PB. Insight into hepatotoxicity: the troglitazone experience. Hepatology 2005; 41(2):229–30. 50. Bilker WB, Berlin JA, Gail MH, Strom BL. An efficient design for verifying disease outcome status in large cohorts with rare exposures and low disease rates. Stat Med 1999; 18(22):3021–36. 51. Chan KA, Truman A, Gurwitz JH, et al. A cohort study o the incidence of serious acute liver injury in diabetic patients treated with hypoglycemic agents. Arch Int Med 2003; 163(6):728–34. 52. Drug- and toxicant-induced liver disease. In: Action Plan for Liver Disease Research (chap. 8), December 2004. (Accessed 28 July 2006 at http://www.niddk.nih.gov/fund/divisions/ddn/ldrb/ ldrb_action_plan.htm). 53. Grassley C, Dodd C. Senators Introduce Bill to Establish New FDA Prescription Drug Safety Office, U.S.A. (Accessed 28 July 2006 at http://www.medicalnewstoday.com/medicalnews.php?newsidZ 23584). 54. Enzi M, Kennedy E. The Enhancing Drug Safety and Innovation Act of 2006. (Accessed 28 July 2006 at http://help.senate.gov/S_.pdf). Note added in proof: The Committee on Assessment of the U.S. Drug Safety System of the Institute of Medicine issued its draft report in September 2006, and the final published report in January 2007 (The Future of Drug Safety: Promoting and Protecting the Health of the Public). Accessed 25 April 2007 at; http://books. nap.edu/catalog.php?record_id=11750.

Index

Abacavir, 57–59, 570, 573 Acarbose, 3, 194 ACE inhibitors, 3 Acetaminophen (paracetamol), 6, 22, 55–56, 89–91, 161, 188, 226–227, 239–241, 254, 265, 347 hepatotoxicity acetaminophen-induced, risk factors for, 390–391 pathology and clinical presentation of, 389–405 liver disease, acetaminophen-induced, mechanisms, 351–387 covalent binding to hepatocellular proteins, 359–361 CYP 1A2, 355 CYP 2A6, 356 CYP 2D6, 356 CYP 2E1, 356 CYP 3A4, 356 CYP oxidation pathways, 354–356 early events, 352–367–368 factors influencing, 356–357 age, 356–357 drug-drug interactions, 357–358 late events in, 367–368–374. See also Late events in acetaminophen-induced liver disease lipid peroxidation and related oxidative events, 364–368 major non-P450 pathways, 352–354 metabolic pathways, 353 ‘noncovalent’ interactions with cellular proteins, 364 pathways of acetaminophen metabolism, 352–358 reactive metabolite disposition, 359 reactive metabolite formation, mechanisms of, 358–359 liver injury, acetaminophen-induced, 350 poisoning epidemiology of, 393–394 acetaminophen-induced acute liver failure, 396–397 biochemical and other laboratory features, 394–396 clinical features, 394 diagnosis, 397–399 outcome, 400–401 pathological features, 396 treatment, 399–400

[Acetaminophen (paracetamol)] suicidal, 391 suicidal versus unintentional APAP cases, 392 unintentional, 391–393 toxic doses, 390 toxicity chronic, 393 metabolism and, 19 Acetohexamide, 265, 613 Ackee fruit, 744 Acquired immune system, 215–216 ACS1. See Acyl-coenzyme A synthetase Actinomycins, 56, 91 Acute hepatitis, 224, 238 Acute liver failure (ALF), 664 management, patient, 349–350 Acute myeloid leukemia, 640 Acyl glucuronides binding, to proteins, reversible, 136 biochemical aspects of, 128–129 biosynthesis of, 129–130 chemical reactivity of, 126–127 glycation, 127 transacylation, 126–127 covalent binding of acidic drugs, predictability of, 148–149 to proteins, 137–146 in vitro covalent binding to plasma and tissue proteins, 138–139 in vivo covalent binding, 141 in vivo tissue protein binding, 141–144 mechanism of, 139–141 procedures to assay, 137–138 protein adducts, stability of, 144 tissue proteins, 145–146 degradation of, influencing factors, 134–136 hydrolysis of, 132 intramolecular acyl migration of, 133 mechanistic role of, 125–157 reactive, potential toxicological significance of, 150–151 separation and quantification of, 130–131 HPLC conditions for, 131 stability of, 132–136 hydrolysis rates, predictability, 133 stereochemical aspects of, 146–148 synthesis, isolation, and characterization of, 129–132 structural characterization, 131–132

790 Acyl-coenzyme A synthetase (ACS1), 410 Adalimumab, 672 Adefovir, 570 Adenine nucleotide translocase (ANT), 87 Adverse events reporting system (AERS), 778 Advilw, 448–449 AERS. See Adverse events reporting system Affymetrixe microarrays, 309 Aflatoxins, 22, 254, 746, 764 Age acetaminophen-induced liver disease and, 356–357 as risk factor, 292 Alanine aminotransferase (ALT), 1, 56, 207, 441–457, 493, 536, 568, 597–599, 611–613, 684, 725 Alcohol abuse, 296–297 alcoholic hepatitis, 35 mitochondrial injury due to, 55 ALF. See Acute liver failure Algorithms or clinical scales, 315, 337 Naranjo adverse drug reaction probability scale, 315 probabilistic approaches, 315 Aliphatic hydrocarbons, halogenated, 761–762 carbon tetrachloride, 761 hydrochlorofluorocarbons, 762 tetrachloroethane, 761 Alkaline phosphatase (ALP), 485 Allopurinol, 3, 225, 248, 265, 632 Allyl alcohol, 161, 188 ALP. See Alkaline phosphatase a-Methyldopa, 243, 248, 254, 265, 603 a-Tocopherol (vitamin E), 40 ALT. See Alanine aminotransferase Amanita, 723–730 liver lesions, 726 management, 728–730 extracorporeal detoxification, 729 general management, 728 liver transplantation, 729–730 specific treatments, 728–729 poisoning manifestations, 725–726 prognosis, 726–728 toxins, 724–725 amatoxins, 724–725 chemical structure, 724 phallotoxins, 725 virotoxins, 725 Amanitin, 254 a-amanitin, 91 Amatoxins, 724–725 chemical structure, 724 Amineptine, 62, 209, 214, 254, 265, 511 Amino acid sequence, 311 Aminoglutethimide, 265 Aminosalicylates, 674 Aminosalicylic acid, 265 Aminotransferases, 579 dosage adjustment, 608

Index Amiodarone, 64, 230–231, 254, 594–597 Amitriptyline, 254, 265, 511 Amoxicillin-clavulanate, 217, 225, 265, 327 Amphotericin B, 534–535 Ampicillin, 528 Amprenavir, 579 Amsacrine, 56 ANA. See Antinuclear antibodies Anabolic androgenic steroids, 711 cholestasis and hepatitis, 711 neoplasms of the liver, 711 vascular Lesions, 711 Anaproxw, 449 Anesthetic agents, 465–479 anesthetic-induced liver injury, 470–477 animal models, 470–471 halothane, 470–471 immune responses, 476–477 other volatile agents, 471 immune hypothesis, 472–477 enflurane, isoflurane, and desflurane, 475–476 halothane, 472–475 hydrochlorofluorocarbon refrigerants, 476 sevoflurane, 476 individual susceptibility factors, 477–478 immunological factors, 478 metabolic factors, 477 volatile anesthetic-induced hepatotoxicity clinical features of, 465–467 halothane, 465–466 other volatile agents, 466–467 mechanism of immune-mediated hepatotoxicity, 475 metabolism, 467–470 desflurane, 467–469 enflurane, 467–469 halothane, 467–468 isoflurane, 467–469 sevoflurane, 469–470 “Angel dust,” 520 Angiosarcoma, 224, 710 Angiotensin receptor blockers (ARBs), 604 Angiotensin-converting enzyme, 604–605 Animal models of anesthetic-induced liver injury, 470–471 Ankylosing spondylitis (AS), 672 Anorexia, 465 Ansaidw, 450 ANT. See Adenine nucleotide translocase Anthracyclines, 56 Anthranilic acids, 452 Antiarrhythmic drugs, 593–598 amiodarone, 594–597 pathophysiologic mechanisms, 595 other antiarrhythmics, 598 aprindine, 597 procainamide, 597 propafenone, 597–598 quinidine, 597

Index

Antibacterial agents, 527–534 beta-lactam antibiotics, 528–530 penicillins, 528–529 clindamycin, 534 daptomycin, 534 macrolide antibiotics, 530–531 nitrofurantoin, 533–534 oxazolidinones, 534 quinolones, 532–533 rifampin, 534 streptogramins, 534 sulfonamides, 531 tetracyclines, 531–532 vancomycin, 533 Anticancer drugs, 631–638 hepatic metabolism, 631–638 age, 637 and hepatic metabolism, 637 and liver toxicity, 637 drug interactions, 631–635. See also Drug interactions gender, 637 genetic polymorphisms, 638 nutrition, 637–638 and hepatic metabolism, 637–638 and liver toxicity, 638 underlying liver disease, 635–636 and hepatic metabolism, 635–636 and liver toxicity, 636 and modalities, 645–652 alkylating agents, 645–647 busulfan, 646 cyclophosphamide, 645–646 dacarbazine, 646 melphalan, 646–647 antimetabolites, 647–649 methotrexate, 647 thiopurines, 647–649 hormones, 651–652 cyproterone acetate, 652 flutamide, 652 tamoxifen, 651–652 miscellaneous anticancer drugs, 649–650 actinomycin d, 650 l-asparaginase, 649–650 monoclonal antibodies, 649 radiation, 650–651 Anticancer therapy, 642–645 Anticonvulsant agents, 6, 485–499. See also Carbamazepine; Felbamate; Phenytoin; Valproic acid aromatic anticonvulsants, cross-sensitivity between, 492 lamotrigine, 499 management, 499 oxcarbazepine, 489 phenobarbital, 492 Anticonvulsants, 6 Antidepressants, 510–514 liver injury, characteristics of, 512 monoamine oxidase inhibitor, 513

791 [Antidepressants] iproniazid, 513 specific serotonin reuptake inhibitor, 513 mianserine, 513–514 nefazodone, 514 trazodone, 514 tricyclic antidepressants, 510–513 amineptine, 511 amitriptyline, 511 cross-sensitization, 511–513 imipramine, 511 other tricyclic antidepressants, 511 Antidiabetic drugs, 611–617 biguanides, 616 a-glucosidase inhibitors, 616–617 acarbose, 616–617 other a-glucosidase inhibitors, 617 insulins, 611 metiglinide analogues, 617 sulfonylureas, 611–612 acetohexamide, 613 chlorpropamide, 613 gliclazide, 613 glimepiride, 612 glyburide, 612 tolazamide, 612 tolbutamide, 612 thiazoladinediones, 613–616 second generation thiazolidinediones, 614–616 pioglitazone, 615–616 rosiglitazone, 614–615 troglitazone, 613–614 Antidotes and steroids for hepatitis-like reactions, 348 Antiemetics, 254 Antifungal agents, 534–537 amphotericin B, 534–535 caspofungin, 535 oral antifungal agents, 535–537 fluconazole, 536 flucytosine, 537 griseofulvin, 537 itraconazole, 536 ketoconazole, 535–536 terbinafine, 536–537 voriconazole, 536 Antihistamine, 18 Antihypertensive drugs, 601–608 ace inhibitors, 604–605 captopril, 604–605 enalapril, 605 other ACE inhibitors, 605 ARBs, 605–606 b-blockers, 606 calcium-channel blockers, 606 centrally acting agents, 601–603 clonidine, 603 methyldopa, 603 combined a 1/b antagonist, 606 direct vasodilators, 603–604 hydralazine, 603–604

792 [Antihypertensive drugs] diuretics, 607 furosemide, 607 hydrochlorothiazide, 607 spironolactone, 607 tienilic acid, 607 endothelin receptor antagonists, 607–608 Anti-inflammatory cytokines, 167 Antimalarial drugs, 65 Antinuclear antibodies (ANA), 232 Antioxidant defense, in liver injury, 33–48. See also Antioxidant defense systems; Oxidant stress antioxidant defense in nonparenchymal cells, 42 in vascular space, 42–43 antioxidant defense systems, 38–41 enzymatic defense mechanisms, 38–39 intracellular and vascular sources of oxidants, 33–36 cytosol, 35 microsomes, 34 mitochondria, 33–34 peroxisomes, 34 reactive oxygen species generation, 34 vascular oxidant stress, 35–36 xanthine dehydrogenase, 35 low molecular weight antioxidants, 40–41 metal-binding proteins, 41 in nonparenchymal cells, 42 oxidant stress, pathophysiological consequences of, 36–38 reactive oxygen and nitrogen intermediates, 33 hydrogen peroxide, 33 nitric oxide, 33 peroxynitrite, 33 peroxynitrous acid, 33 superoxides, 33 Trxs and Prxs, 39–40 in the vascular space, 42–43 Antioxidant response element (ARE), 37 Antiplatelet agents, 610–611 Antipyrine, 636 Antithyroid drugs, 716–717 Antituberculosis drugs, 298, 547–562 hepatotoxicity, 548–562 alternative therapy, 559–560 clinical, biochemical, and histological features, 557–558 combination drugs, 553 incidence, 548–549 of individual drugs, 551–553 isoniazid, 551–552 pyrazinamide, 552–553 rifampin, 552 in liver transplant recipients, 560–561 management, 558–559 mechanism, 549–551 risk factors, 553–557 acetylator status, 553–554 age, 554

Index [Antituberculosis drugs] alcohol use, 554–555 enzyme inducers, 555 gender, 555 hepatitis B and C, coinfection with, 556–557 malnutrition, 556 other genetic factors, 556; race/ethnicity, 555–556 liver disease, in patients with, 560 monitoring, recommendations for, 561 LTBI, 561 monotherapy with INH, 561 multidrug regimens, 561–562 Antituberculosis therapy (ATT), 549–562 Antiviral agents, 567–583 assessment, problems with, 567–569 clinical presentation, 576–579 NNRTIs, 577–578 NRTIs, 576–577 PIs, 578–579 drug interactions, 579–580 dosages recommendations, 580 drug toxicity, mechanisms of, 569–576 hypersensitivity reactions, 572–574 interferons, 574–576 other mechanisms, 574 mitochondrial toxicity, 569–572 cause, 569–570 liver injury, 580–582 hepatitis B coinfection, 580–581 hepatitis C coinfection, 581–582 liver cirrhosis and reduced liver function, 582 toxicity, consequences of, 582–583 Anxiolytic agents, 517 benzodiazepines, 517 buspirone, 517 Apical (canalicular) transporters, characteristics, 102 properties, function and regulation of, 101–103 Aplastic anemia, 498 Apoptosis, 85 apoptosis cascade, 86 apoptotic cell death, 85–88 machinery of, 86 mitochondria in, 86 triggering of, 87 Aprindine, 265, 597 ARBs. See Angiotensin receptor blockers Ardeparin, 609 ARE. See Antioxidant response element Argatroban, 609 Argemone oil, 746–748 Aromatic hydrocarbons, halogenated, 761 polychlorinated biphenyls, 761 Array-based genotyping, 311 Arteritis, 264 Aryl hydrocarbon receptor (AhR), 17 2-Arylpropionate nonsteroidal anti-inflammatory drugs, 62 AS. See Ankylosing spondylitis Ascorbate (vitamin C), 40

793

Index

Aspartate aminotransferase (AST), 441, 443, 447–450, 454, 493, 536, 568, 684, 725 Aspirin, 60, 243, 254, 443–444 Astragali radix, 735 Atazanavir, 574 Atenolol, 265 ATP synthase, inhibition, 65–66 ATT. See Antituberculosis therapy Augmentinw, 3 Autoantibodies, in drug-induced liver disease, 7 Autoimmunity, 329 Azapropazone, 452 Azathioprine, 211, 265, 632, 647, 666–669 Azidothymidine (AZT), 254, 572 Bacillus cereus, 748 Bacterial serology, 329 Barbiturate therapy (P450 2B inducer), 16 Basolateral (sinusoidal) transporters, 98–101, 103 characteristics, 100 BCS. See Budd-Chiari syndrome Beclobric acid, 129, 135, 138, 141, 149 Benorilate, 445 Benoxaprofen, 3, 135, 138, 265, 293, 450–451 Benzamides, 515 Benzarone, 65, 243 Benzbromarone, 65 Betulinic acid, 68 Bezafibrate, 713 Biguanides, 616 Bile duct injury, 263–264, 269 chlorpromazine, 269 chlorpropamide, 269 Bile ductular transport systems, 103 Bile formation, 98–103 Bile salt export pump (BSEP), 101, 216, 607 Bile salt-exporting protein (cBsep), 414 Biliary cirrhosis, primary, 696–697 Biliary excretion, 412–413 Bilirubin, 493, 534 Bioactivation, by P450 s and other enzymes, 19–24 basic aspects, 19–22 P450-dependent bioactivation, 22–24 acetaminophen, 22 aflatoxin B1, 22 trichloroethylene (TCE), 22–23 troglitazone, 23–24 to reactive intermediates toxicity, 18–19 Bioactive lipids, 169–170 Bivalirudin, 609 Bleomycin, 254 Borates, 254 Boric acid, 254 Bosentan, 3, 8 Bovine serum albumin (BSA), 136 Breast cancer resistance protein (BRCP), 533 Breynia, 740 Bromfenac, 2–3, 9, 228, 248, 446–447 Bromperidol, 515 BSEP. See Bile salt export pump

Budd-Chiari syndrome (BCS), 650, 709 Buprenorphine, 65, 520 Bupropion, 243 “Bush tea disease,” 639 Busulfan, 646 Butyrophenones, 515 Cadmium, 161, 254 Caffeine, 742 Calcineurin inhibitors, 665 Calcium hopantenate, 63 Calcium-channel blockers, 606 Calicheamicin, 649 Calpain, 198–199 Camellia sinensis, 743 Camphor, 254 Cancer chemotherapy, 631–652. See also Anticancer drugs; Hematopoietic stem cell transplantation chemotherapeutic agents, transporters of, 633 conventional dose chemotherapy, 644–645 Cannabis, 519 Capillarization, 293 Captopril, 248, 265, 604–605 Carbamazepine (CBZ), 225, 248, 265, 485–489 clinical manifestations, 486 diagnosis, 486–487 pathology, 486 postulated mechanisms, 487–489 bioactivation, pathways of, 488 susceptibility factors, 487 Carbarsone, 265 Carbimazole, 265, 716 Carbon tetrachloride (CCl4), 89, 161, 188, 254, 761 Carbutamide, 243 Cardiovascular drugs, 593–611. See also Antiarrhythmic drugs; Heparins; Antihypertensive drugs antianginal medications, 608 anticoagulants, 608–610 antithrombin agents, 609–610 heparins, 608–609 warfarin and coumarin-derived agents, 608 antihypertensive drugs, 601–608 antiplatelet agents, 610–611 clopidogrel, 611 ticlopidine, 610–611 classification, 594 glycoprotein iib/iiia receptor blocker, 611 lipid lowering agents, 598–601 ezetimibe, 601 fibrates, 600–601 HMG-CoA reductase inhibitors, 598–600 niacin, 600 metabolic pathways, 601 thombolytics, 611 Carisoprodol, 265 Carnitine and/or direct inhibition of b-oxidation, 60–63 Carprofen, 129, 135, 138, 141 Carvedilol, 606

794 Caspofungin, 535 Catalase, 39 Causality assessment, 326–343 diagnosis in the clinical setting, 326–332 drug exposure data and chronology, 327 hepatotoxic potential, 327–328 exclusion of alternative causes, 328–330 incriminating criteria, 330–332 drug-allergy features, 330 course after drug withdrawal (de-challenge), 331 response to re-administration (re-challenge), 2331 algorithms or clinical scales, 335–337 CIOMS/RUCAM scale in, 336 clinical diagnostic scale (CDS), 336 decision tree in, 336 “gold standard,” 332 hepatotoxicity assessment methods, comparison, 337–340 liver biopsy and the drug “signature,” 331–332 standardized causality assessment methods, advantages and limitations, 333 CDS. See Clinical diagnostic scale Cefadroxil, 265 Cefazolin, 265 Celecoxib, 453 Cell death and drug hepatotoxicity, 85–95. See also Apoptosis; Necrosis in drug hepatotoxicity, 88–92 acetaminophen, 89–91 carbon tetrachloride (CCl4), 89 GSH role in susceptibility to FasL and TNFa, 91 NF-kB role, 91–92 reactive oxygen and, 37 transcriptional arrest in hepatocytes, 91 CellCeptw, 666 Cellular glutathione peroxidase (cGPx-1), 39 Centrilobular hepatic necrosis, 466 Centrilobular necrosis, 604 Cereulide, 748 Chaparral, 243, 248 Chemokines, 167 Chinese herbal medicines, 248 Chlorambucil, 265 Chloramphenicol, 60, 254, 571 Chlordecone, 763–764 Chlordiazepoxide, 248, 265 Chlorinated ethylenes, 762–763 tetrachloroethylene, 762 vinyl chloride, 762–763 Chloroform, 188, 254 Chloroquine, 65, 254 40-Chlorodiazepam, 68 Chlorothiazide, 265 Chlorpheniramine, 254 Chlorphenoxy compounds, 763–764 chlordecone, 763–764 paraquat, 764

Index Chlorpromazine, 213, 243, 254, 265, 295, 514–515, 757 Chlorpropamide, 265, 535 Chlortetracycline, 254, 265 Chlorthalidone, 265 Cholangiocarcinoma, 224 Cholecystokinin octapeptide, 724 Cholehepatic circulation, 413–414 Cholestasis management, ursodeoxycholic acid for, 348–349 Cholestasis, 224, 314, 490 Cholestatic hepatitis, 528 Cholestatic injury, 259–263 cholestasis, simple, 259–262 with inflammation, 262–263, 265 allopurinol, 266 chlorpromazine, 266 clarithromycin, 267 ketoconazole, 268 niacin, 268 Choline salicylate, 444 Chromate, 254 Chronic hepatitis, drug-induced, 224, 232–233 Chrysotherapy, 674 Cimetidine, 248, 265 Cinchophen, 2 CIOMS. See Council for International Organizations of Medical Sciences CIOMS/RUCAM scale, 336 Ciprofloxacin, 56 Cirrhosis, 686 Cisplatin, 254, 265 Clarithromycin, 243, 248, 265, 530 Clindamycin, 534 Clinical diagnostic scale (CDS), 336 Clinicopathological patterns of drug-induced liver disease, 223–235 assessment, 230–231 amiodarone, 230–231 clinical features of, 227–228 bromfenac, 228 diclofenac, 228 isoniazid, 227–228 NSAIDs, 228 sulindac, 228 drug-induced chronic hepatitis, 232–233 dihydralazine, 233 minocycline, 232 nitrofurantoin, 232–233 oxyphenisatin, 233 tienilic acid, 233 identification and diagnosis, 225–227 mechanisms of injury, 233–234 risk-benefit considerations, 231–232 signals of hepatotoxicity, 229–230 intermediate, 229 major, 229 minor, 229 Clinoril, 447–448 Clofibrate, 601 Clofibric acid, 135, 136, 141

Index

Clometacin, 243, 254, 448 Clonidine, 603 Clopidogrel, 611 Clorazepate, 265 Clozapine, 515–517 CoA, sequestration of, 60–63 Coagulopathy, 557 Cocaine, 241, 254, 254, 519 Coccidioidal meningitis, 536 Colchicine, 188 colchicine antimitosis, 189 Comfrey, 735 Constitutive androgen receptor (CAR), 17, 293, 357 Copper chelator, 674 Corticosteroids, 254, 664, 715–716 Coumarin, 608 Council for International Organizations of Medical Sciences (CIOMS), 336–337 Covalent binding data, use in drug safety, 28 COX inhibition, 415 Cremophor, 634 Creosote bush, 737 C-ring rotor, 66 Crohn’s disease, 672–674 Cross-reactivity, as risk factor, 296 Curcuma longa, 735 Cyamemazine, 515 Cyanamide, 254 Cyanobacteria, 748 Cyclofenil, 714 Cyclophosphamide, 634, 639, 645–646 Cyclosporin, 265, 633–634 cholestasis, induced, 642 Cyproterone, 652, 714 Cytochrome (CYP), 313 CYP 1A2, 355 CYP oxidation pathways, 354–356 CYP2C19, 317 CYP2C9, 317 CYP2D6, 316–317 CYP2E1, 295295 Cytochrome p450 (P450) activation of toxins and hepatotoxicity, 13–31. See also P450 enzymes; Safety prediction; Toxicities contexts of toxicity, 17–19 metabolism contribution to drug toxicity, 24–26 P450 enzymes, 13–17 P450-mediated bioactivation, 409 safety prediction of new drugs, 26–28 Cytokines, 165 anti-inflammatory cytokines, 167 chemokines, 167 proinflammatory cytokines, 165–166 role in toxin-induced liver disease, 4–5 TGFb, 166–167 Cytosol, 35 Dacarbazine, 248, 265, 646 Dalteparin, 609 “Danger” hypothesis, 18 Dantrolene, 2–3, 243, 254, 265

795 Dapsone, 243 Daptomycin, 534 Daypro, 450 “D-drugs,” 58 Demeclocycline, 254 Dequalinium, 57 Desferrioxamine, 254 Desflurane, 466–467 immune hypothesis, 475–476 Diabetes, liver tissue repair in, 194–196 Diacetylhydrazine, 554 Diazepam, 265 Diazepines, 515–517 1,2-Dichlorobenzene, 161, 188 Dichloroethylene, 254 Diclofenac, 2–3, 6, 68–69, 130, 135, 136, 138, 141, 213, 228, 243, 265, 299, 327, 441, 445–446, 757 Didanosine, 254,570 Dideoxyinosine, 243, 248, 254 2’,3’-Dideoxynucleosides and abacavir, 57–59 mtDNA replication by, 58 4 0 ,4 0 -Diethyaminoethoxyhexestrol, 64 Diflunisal, 129, 135, 136, 138, 141, 146, 149, 445 Digoxin, 327 Dihydralazine, 3, 19, 233, 243 dihydralazine-induced hepatitis, 20 Dihydroorotat-ubiquinone oxidoreductase (DHODH), 571, 577 Dihydropyrimidine dehydrogenase (DPD), 637–638 DILI. See Drug-induced liver injury Diltiazem, 606 Dimethylformamide, 254 Diphenylhydantoin, 234, 347 Disease modifying anti-rheumatic drugs (DMARDs), 453–454, 669 Disopyramide, 265 Disulfiram, 3, 243 Ditercalinium, 57 Diuretics, 607 DMARDs. See Disease modifying anti-rheumatic drugs Docetaxel, 631, 636 Doppler ultrasonography, 639 Doxorubicin, 634, 637 Doxycycline, 532 D-penicillamine, 225, 674 DRESS syndrome, 486 Droxicam, 451 Drug development strategies used in, 27 Drug hepatotoxicity, cell death and, 85–95. See also Cell death Drug interactions, 631–635 and drug disposition, 631–635 toxification, 631–633 biliary excretion, inhibition of, 633–634 metabolic activation, induction of, 633 and liver toxicity, 635 as risk factor, 295–296 Drug-allergy features, 330

796 Drug-drug interactions, acetaminophen-induced liver disease and, 357–358 Drug-induced chronic hepatitis, 232–233 Drug-induced hepatotoxitity, risk factors for, 6 genetic determinants, 6 Drug-induced liver injury (DILI), 1–11, 548, 553–559 autoantibodies in, 7 cases and controls, appropriate source for, 207–208 diagnosis, concepts, 330 identification of, 208 not related to reactive metabolites, 216–218 causality assessment, 7 clinical overview, 1–3 diagnosis, 6–7 algorithm, 326 specific targets for, 326 drug development, 7–8 hepatic manifestations of, 2 idiosyncratic hepatotoxins in clinical trials, 8 pathogenesis, 3–5 cytokine balance role, 4–5 genetic polymorphisms of enzymes, 4 metabolism, 3 postmarketing monitoring, 8–9 issue in, 9 predictable reactions of, 2 profiling technologies, 307–312 risk factors, 5–6 environmental factors, 5 genetic factors, 5 toxic potential of drug, 5 unpredictable reactions, 2 Drugs of abuse, 519–520 illegal and recreational compounds, 519–520 amphetamines and derivatives, 519–520 buprenorphine, 520 cocaine, 519 methylphenidate, 520 phencyclidine, 520 Dubin-Johnson syndrome, 412 Echinacea, 734 Eicosanoids, 169 Elevated aminotransferase levels, 224 ELISA. See Enzyme-linked immunosorbent assay Ellipticines, 56 Enalapril, 265, 605 Endoplasmic reticulum (ER), 52 Endothelial cells, 163–164 Endotoxemia, 35 Endotoxin, 161 Enflurane, 2, 466–469 immune hypothesis, 475–476 Enoxaparin, 609 Enterohepatic circulation, 413 Enterotoxin, 748 Environmental hepatotoxicity, 764–765 aflatoxins, 764 phosphorus, 764–765

Index Enzymatic defense mechanisms, 38–39 Enzyme-linked immunosorbent assay (ELISA), 472 Eosinophilia, 445 Epilepsy, 347 Epirubicin, 637 Epithelioid hemangioendothelioma (EHE), 710 Erwinia chrysanthemi, 649 Erythromycin estolate, 531 Erythromycins, 3, 60, 248, 265, 530, 613 Escherichia coli, 649 Estrogen-receptor antagonists, 712–714. See also Tamoxifen cyclofenil, 714 toremifene, 714 Etanercept, 672 Ethambutol, 212, 217 Ethanol, 161, 243, 254 Ethchlorvynol, 265 Ethidium bromide, 57 Ethionamide, 248, 265 Ethionine, 254 Ethyl bromide, 254 Ethyl chloride, 254 Etodolac, 129, 130, 135, 136, 138, 149, 446 Etoposide, 56, 634 Etretinate, 243, 254 European Agency for the Evaluation of Medicinal Products (EMEA), 671 Ezetimibe, 601 FADD. See Fas-associated death domain Familial cholestasis type 1 (FIC1), 101 Fanconi’s anemia, 711 Fas-associated death domain (FADD), 85 Fat oxidation and energy production, in mitochondria, 50–51 primary impairment of, 52–53 Fatty change inflammation (steatohepatitis), 254 FDA. See Food and Drug Administration Felbamate, 3, 496–498 clinical manifestations, 496–497 pathology, 496–497 postulated mechanisms, 497–498 bioactivation, mechanism of, 498 Felty’s syndrome, 441 Female sex hormones, 63 Fenamates, 452 Fenclozic acid, 447 Fenestration, 293 Fenofibrate, 243 Fenofibric acid, 138, 149 Fenoprofen, 130, 135, 136, 138, 141, 449–450 Feprazone, 452 Fexofenadine, 18 Fialuridine, 59, 254 Fibrates, 600–601 Fibrous obliteration, 264 FIC1. See Familial cholestasis type 1 Fipexide, 517 Floxuridine, 254 Flucloxacillin, 265, 528–529

Index

Fluconazole, 536 Flucytosine, 537 Flufenamic acid, 135, 138 Flunoxaprofen, 134, 135, 451 Fluorescence detection technique, 145 5-Fluorouracil, 637 Fluoromethyl-2,2,-difluoro-1-(trifluoromethyl)vinyl ether (FDVE), 469–471, 476–477 Fluoxymesterone, 265 Fluphenazine, 265 Flurazepam, 254, 265 Flurbiprofen, 450 Flutamide, 3, 129, 134–136, 265, 652 Focal nodular hyperplasia, 709 Food and Drug Administration (FDA), 445, 600 DILI, interest in, 772 new drugs evaluation, 772–773 Food contaminants, 746–748 aflatoxin, 746 argemone oil, 746–748 Bacillus cereus, 748 cyanobacteria, 748 other natural hepatotoxins, 748–749 carp gall bladder, 748–749 Fulminant hepatic failure (FHF), 443–449, 453, 486 Furosemide, 607 Galactosamine, 91, 161 Gamma glutamyl transferase (g-GT),485 g-Glutamyl transpeptidase, 611 Ganciclovir, 59 Gemfibrozil, 130, 134–136, 138, 141, 601 Gemtuzumab ozogamicin, 649 Gene array studies, 310 Gene transcription, reactive oxygen and, 37–38 Genetic susceptibility to drug-induced liver disease, 207–222, See also DILI cases acquired immune system, role of, 215–216 association studies targeting drug metabolizing enzymes, 209–215 genetic variation in events not involving metabolism, 215 lymphocyte toxicity assays, 208–209 Genomics, of drug-induced liver disease, 307–324 global assessment of protein expression, 311–312 global gene expression profiling, 309–311 global profiling technologies role in, 312–316 human variability role in toxicogenomics, 316–321, See also Human variability role personal genomes, 307–309 Gentamicin, 254 Germander, 740–741 Gilbert’s syndrome, 574 Glibenclamide, 265, 612 Gliclazide, 613 Glimepiride, 612 Global profiling technologies expression profiling, for hepatotoxicity, 313–316 future application of, 321–322 role in drug-induced liver disease, 312–316 Glucocorticoids, 63, 254, 664

797 Glutathione (GSH), 3, 40 Glutathione disulfide (GSSG), 90 Glutathione peroxidase (GPx), 35 Glutathione transferases, 210, 528, 550, 635, 641 glutathione-S-transferase (GST), 39, 473 Glyburide, 612 Glycation, 127 Gold salts, 453–454, 674 Gold sodium thiomalate, 254, 265 “Gold standard,” in causality assessment, 332 Granulomas, 224, 253–254 Greasewood, 737 Green tea, 743 Griseofulvin, 265, 537 HAART. See Highly active antiretroviral therapy Halogenated hydrocarbons, 244, 254, 265 Haloperidol, 265, 515 Halothane, 2–3, 209, 214, 217, 347, 465–468, 471–478, 757 animal models, 470–471 immune hypothesis, 472–475 HapMap, 308 “Hapten” hypothesis, 18 HCC. See Hepatocellular carcinoma Heat shock protein (HSP), 87 Hedeoma pulegoides, 741 Hemangioma, 709 Hematopoietic stem cell transplantation, 638–642. See also Hepatic sinusoidal obstruction syndrome and liver toxicity, 638–642 Heparins, 608–609 low molecular weight heparins, 609 unfractionated heparin, 608–609 Hepatic adenoma, 224, 710 Hepatic encephalopathy, 557 Hepatic granulomas, from therapeutic drugs, 225 Hepatic injury DMARDS, due to, 453–454 NSAID-induced, 440–453 Hepatic neutrophil recruitment, 35 Hepatic nonparenchymal cells, and tissue injury, 159–165 agents whose toxicity is associated with, 161 endothelial cells, 163–164 Kupffer cells and inflammatory macrophages, 159–163 NK cells, 165 stellate cells, 164–165 Hepatic sinusoidal obstruction syndrome, 639–642 diagnosis, 639–640 drugs causing, 640 mechanism, 640–642 vs. radiation-induced liver disease, 651 Hepatic vein obstruction, 224 Hepatic vein thrombosis, 708–709 Hepatitis B, 580–581, 636 reactivation of, 664 Hepatitis C virus (HCV), 575, 581–582, 636

798 Hepatobiliary transporters. See also Hepatocellular transport systems for drug-induced liver disease, 97–114 hepatic transport systems, 98–103 apical (canalicular) transporters, 101–103 basolateral (sinusoidal) transporters, properties, function and regulation, 98–101 basolateral and apical cholangiocyte transporters, 103 bile ductular transport systems, 103 physiology and bile formation, 98–103 Hepatocellular carcinoma (HCC), 710, 746 Hepatocellular injury, 237–259 See also Lobular necrosis fatty change, 253 cocaine, 256 ethanol, 255 methotrexate, 256–257 mushrooms, 255 sulfasalazine, 258 tetracycline, 257 vitamin A, 258 granulomas, 253–254 epithelioid type, 253 ethanol, 260 inflammatory type, 253 mineral oil, 261 sulfasalazine, 260–261 hepatocellular jaundice drugs which cause, 2 mallory bodies, 254–259 morphologic variants, 238 bile duct injury, 238 cholestatic injury, 238 hepatocellular injury, 238 inclusions, 238 neoplasia, 238 pigments, 238 portal fibrosis, 238 vascular injury, 238 Hepato, covalent binding to, 361–362 Hepatocellular transport systems age and gender, 105–107 drug-transporter interactions, 104–105 environmental risk factors, 104 genetic risk factors, 107–109 impaired hepatocellular transporter expression, 106 preexisting liver disease, 107 role in drug-induced liver injury, 103–109 Hepatocyte role in immune-mediated liver injury, 119–120 Hepatomegaly, 674 Hepatosplenic T-cell lymphoma (HSTCL), 673 Hepatotoxicity. See also individual drug types assessment methods, comparison, 337–340 by anticancer therapy, 642–645 diagnosis, 642–643 specific anticancer drugs, due to, 643–645 due to mitochondrial injury, 49–84 expression profiling for, 313–316

Index [Hepatotoxicity] hepatotoxic potential, 327–328 risk of, 5 signals of, 229–230 Herbal medicines, 518–519, 733–744 background, 733–734 cannabis, 519 complementary and alternative medicine (CAM), 733 histopathological patterns, 736–737 diagnosis and treatment, 735 hepatotoxicity, 735–743 atractylis gummifera and callilepsis laureola, 742 black cohosh, 743 chaparral, 737–738 Chinese and other Asian herbal medicines and teas, 738–740 ba jiao lian, 739 breynia, 740 jin bu huan, 739–740 ma huang, 739 shou wu pian, 739 syo-saiko-to, 740 germander, 740–741 greater celandine, 742 kava, 738 lipokinetix, 742–743 morinda, 743 pennyroyal, 741–742 pyrrolizidine alkaloids, 735–737 interactions with prescription drugs, 749–750 kava-kava, 518–519 other hepatotoxic herbs and botanicals, 743–744 ackee fruit, 744 margosa oil, 744 valerian, 519 Highly active antiretroviral therapy (HAART), 570–571, 580–582 Histopathology of drug-induced liver disease, 237–287 bile duct injury, 263–264 cholestatic injury, 259–263 hepatocellular injury, 237–259 inclusions, 272–276 neoplasia, 271–272 pigments, 277–284 portal fibrosis, 265–271 vascular injury, 264–265 HIV/AIDS, 298 HLA genotypes, 318–320 Hopantenate, 254 Hormones, adverse effects of, 707–717. See also Anabolic androgenic steroids; Oral contraceptive steroids antithyroid drugs, 716–717 methimazole and carbimazole, 716 propylthiouracil, 716–717 clinical presentation, 717 pathology and pathogenesis, 717 treatment and outcome, 717

Index

[Hormones, adverse effects of] corticosteroids, 715–716 estrogen-receptor antagonists antiandrogens, 714–715 mechanism of liver injury, 715 outcome and treatment, 715 Human leukocyte antigen (HLA), 216, 489, 529 Human variability role in toxicogenomics, 316–321 CYP2C19, 317 CYP2C9, 317 CYP2D6, 316–317 genotyping, 320–321 non-P450 genes, 317–320 Hy’s law, 1, 328 Hydralazine, 225 Hydrazine, 254, 550 Hydrochlorofluorocarbons, 476, 762 Hydrochloroquine, 68 Hydrochlorothiazide, 607 Hydrolysis and systemic cycling, 412 Hydrolytic enzymes, 170 3-Hydroxyacetaminophen, 354 Hydroxychloroquine, 673 7-Hydroxymethotrexate, 685 6a-Hydroxypaclitaxel, 632 Hyperbilirubinemia, 534 Hyperechoic liver, 329 Hypersensitivity, 2, 9, 18–19 Hypoglycin A, 254 Hypoxanthine phosporibosyltransferase (HPRT), 648 Hypoxia, 470–471 IBD. See Inflammatory bowel disease Ibufenac, 135, 138 Ibuprofen, 129, 135, 138, 254, 265, 448–449 ICP. See Intrahepatic cholestasis of pregnancy Idiosyncrasy, 72–73 drug-induced liver injury, 321 hepatitis, 2 drugs associated with, 3 hepatotoxicity, 291 direct evidence for, 326 liver toxicity caused by NSAIDs, 427–428 toxicity, 18–19 metabolic and comorbidity factors involved in, 73 Imatinib, 635 Imipramine, 265, 511 Immune reactions, 18–19, 71–72 mechanism of immunization, 71 mechanisms of cell death, 71–72 Immune-mediated hepatic injury, 422–425 Immunoblotting, 472 Immunological mechanisms in drug-induced liver injury, 115–123 hepatocyte role in immune-mediated liver injury, 119–120 lymphocyte-mediated cytotoxicity, 118–119 stimuli that trigger, 115–117

799 Immunomodulating agents, 663–676 calcineurin inhibitors, 665 corticosteroids, 664 D-penicillamine, 674 gold salts, 674 hepatitis B, reactivation of, 664 hydroxychloroquine, 673 leflunomide, 669–672 mycophenolate mofetil, 666 nodular regenerative hyperplasia, 669 non-hepatotoxic immunosuppressants, 675–676 sirolimus, 665–666 sulfasalazine and other aminosalicylates, 674 thalidomide, 675 thiopurines, 666–669 AZA, 666–669 6-mercaptopurine, 666–669 6-thioguanine, 666–669 TNF antagonists, 672–673 Impaired mitochondrial function, consequences of, 52–54 In vitro lymphocyte-stimulation test, 326 Inclusions, 272–276 procainamide, 277 Indinavir, 574, 578 Indocin, 447 Indomethacin, 135, 138, 254, 265, 447 Inflammatory bowel disease (IBD), 298, 667 Inflammatory cells, 42 Inflammatory macrophages, 159–184 Inflammatory mediators implicated in hepatotoxicity, 165–170. See also Cytokines bioactive lipids, 169–170 reactive nitrogen intermediates, 168–169 reactive oxygen intermediates, 167–168 Infliximab, 672 INH. See Isoniazid INR. See International normalized ratio Insulin, 327, 611 Intercellular adhesion molecule-1 (ICAM-1), 35 Interferon-a, mtDNA transcripts treated with, 59 International normalized ratio (INR), 725–726 Intracellular and vascular sources of oxidants, in antioxidant defense, 33–36 Intrahepatic cholestasis of pregnancy (ICP), 107 Intrinsic cell death pathway, 368–372 Investigational new drugs (INDs), 773 humans, studies in, 773–774 Iodipamide meglumine, 265 Iproniazid, 2, 513 Irinotecan, 632 metabolism, 633 Iron, 744–745 Isocarboxazid, 265 Isoflurane, 466–469 immune hypothesis, 475–476 Isoniazid (INH), 2–3, 6, 225, 227–228, 254, 265, 327, 548–549, 551–552 age-dependent mortality, 555 metabolism of, 550

800 Isotretinoin, 3 Isoxepac, 135 Itraconazole, 536 Jamaican vomiting sickness, 494 JRA. See Juvenile rheumatoid arthritis Juvenile idiopathic arthritis, 672 Juvenile rheumatoid arthritis (JRA), 298, 443–444, 447 Kaplan-Meyer survival plot, 200 Kaposi’s sarcoma, 675 Kava-kava, 518–519 Kepone, 763–764 Ketoconazole, 2–3, 265, 535–536, 635 Ketoprofen, 135, 136, 138, 141, 254, 450 Ketorolac, 446 Keyhole limpet hemocyanin (KLH), 150, 424 KLH. See Keyhole limpet hemocyanin Killer immunoglobulin-like receptors (KIR), 117 Kupffer cell, 579, 641, 672 activation, 35 and inflammatory macrophages, 159–163 Labetalol, 2–3, 606 Lamivudine, 570 Lamotrigine, 499 L-asparaginase, 254 Late events in acetaminophen-induced liver disease, 368–374 cellular, tissue, and whole body factors, 371–372 extrinsic cell death pathway, 372–373 intrinsic cell death pathway, 368–372 macromolecular and organellar events, 369–371 molecular and biochemical events, 368–369 Latent TB infection (LTBI), 547, 551–553, 561 Leflunomide, 3, 453, 669–672 Lepirudin, 609 Leukocytosis, 674 Leukotriene receptor antagonists hepatotoxicity of, 454–455, 457 montelukast, 454–455 zafirlukast, 454 zileuton, 455 Lidocaine, 636 Lipid peroxidation (LPO), 36 and related oxidative events, 367–368 Lipophilicity, 409 Lipopolysaccharide (LPS), 160, 533 release and inflammatory changes, 426 Liver biopsy and drug “signature,” 294 psoriasis, 687 rheumatoid arthritis, 692 Liver cirrhosis, 582 Liver injury, tissue repair role in, 185–206. See also Tissue repair Liver test abnormalities WHO grades, 650 Liver tissue repair in diabetes, 194–196

Index Liver transplantation, patient with, management, 350–351 Lobular inflammatory infiltrate, 262 Lobular necrosis with inflammation, 238–248 bupropion, 243 halogenated hydrocarbons, 244 methotrexate, 247 nitrofurantoin, 244–245 phenytoin, 245 rifampin, 246 sertraline, 246 lobular confluent necrosis with inflammation, 248 halogenated hydrocarbons, 248 isoniazid, 250 niacin, 251 sertraline, 252 troglitazone, 252 with minimal to absent inflammation, 237–238 acetaminophen, 239–241 cocaine, 241–242 mushrooms, 242 Lodine, 446 Long-chain 3-hydroxyacyl-CoA dehydrogenase (LCHAD), 63 Lonidamide, 68 Lopinavir, 578 Lovastatin, 598–599 Low molecular weight antioxidants, 40–41 Loxapine, 517 LPO. See Lipid peroxidation Lyell syndrome, 330 Lymphadenopathy, 490, 674 Lymphocyte toxicity assays, 209 Lymphocyte-mediated cytotoxicity, 118–119 Macrolide antibiotics, 530–531 Macrovesicular fatty change, 253 Malaise, 465 Mallory bodies, 254–259, 596 amiodarone, 263 ethanol, 262 Mammalian catalase, 39 Management of patient with drug-induced liver disease, 345–352 acute liver failure, 349–350 adverse hepatic drug reactions and their complications, treatment, 348–349 criteria for, 346 antidotes and steroids for hepatitis-like reactions, 348 class sensitization and rechallenge, 346–348 drug-induced liver injury diagnosis and liver tests monitoring, 345–346 liver transplantation, 350–351 ursodeoxycholic acid for cholestasis management, 348–349 Margosa oil, 254, 744 Mass spectrometry approaches, 27 Masson trichrome, 669

Index

Matrix metalloproteinases (MMP), 164, 641 Matthew-Rumack nomogram, 398 MDR. See Multidrug resistance Medium chain acyl-CoA dehydrogenase (MCAD) mutation, 63 MedWatch system, 9 Mefenamic acid, 129, 135, 136, 138, 452 Melatonin, 735 Meloxicam, 451 Melphalan, 646–647 Membrane permeability transition in mitochondria, 417–418 Membrane transporters, 318 Mentha pulegium, 741 Meprobamate, 265 6-Mercaptopurine, 265, 632, 648, 666–669 Mesalazine, 674 Metabolic factors involved in idiosyncrasy, 73 “Metabolic idiosyncrasy,” 327 Metabolism and drug toxicity, 24–26 metabolite-mediated hepatocyte apoptosis, 69–71. See also Immune reactions metabolomics in diagnosis drug-induced liver disease, 307–324 Metabolomics, 28, 312, 315 Metal-binding proteins, 41 Metformin, 194, 616 Methimazole, 254, 716 Methotrexate controversies, 683–697 acute liver injury, 684 chronic liver injury, 684–695 high-dose methotrexate, 684 histological changes, 684–685 low-dose methotrexate, 686–695. See also Psoriasis; Rheumatic diseases mechanism, 685–686 Roenigk classification, 685 MTX therapy, monitoring 695–696 primary biliary cirrhosis, 696–697 primary sclerosing cholangitis, 696 7-Methoxyresorufin, 14 Methyl bromide, 254 Methyl chloride, 254 1-Methyl-4-phenylpyridinium ion (MPPC), 57 Methyl salicylate, 254 Methyldichloride, 254 Methyldopa, 3, 225 Methylphenidate, 520 Metiglinide, 617 Mianserine, 513–514 Microcycline, 254 Microsomal triglyceride transfer protein (MTP), 52 Microsomes, 34 Microvesicular fatty change, 253–254 Midecamycin, 327 Milk thistle, 735 Minocycline, 3, 232, 254, 532 Mitochondria, 33–34 in apoptosis, 86 mitochondrial b-oxidation,

801 [Mitochondria] inhibition of, 418–419 mitochondrial effects of valproate, 61 origin and structure of, 49 roles of, 50–52 cell survival and cell death, 51–52 fat oxidation and energy production, 50–51 mitochondrial ROS formation, 51 Mitochondrial injury, hepatotoxicity, 49–84. See also mtDNA; Mitochondrial permeability transition amphiphilic cationic drugs effects on, 64 diverse mechanisms, 54–68 degradation of mtDNA, 55–59 impairment of mtDNA replication by drugs, 56–57 mtDNA transcripts in cells, 59 by nucleoside analogs incorporated into mtDNA, 57–59 fatty acid oxidation, primary impairment of, 52–53 idiosyncrasy, 72–73 impaired mitochondrial function, consequences of, 52–54 inhibition of b-oxidation and/or respiration, 64–65 inhibition of ATP synthase or the ANT, 65–66 mitochondrial respiration uncoupling, by protonophoric drugs, 66–67 sequestration of CoA or carnitine and/or direct inhibition of b-oxidation, 60–63 uncoupling protein 2, induction of, 68 Mitochondrial outer membrane permeabilization (MOMP), 87 Mitochondrial permeability transition (MPT), 36, 49, 68–72, 87. See also Metabolite-mediated hepatocyte apoptosis direct toxicity of reactive metabolites, 69–71 parent drugs, 68 betulinic acid, 68 diclofenac, 68–69 40-chlorodiazepam, 68 hydrochloroquine, 68 lonidamide, 68 nimesulide, 68–69 peripheral benzodiazepine receptor (PBR), 68 salicylic acid, 68–69 troglitazone, 69 valproic acid, 68–69 Mitochondrial respiration primary impairment of, 53–54 Mitochondrial toxicity, 569 Mitomycin C, 254 Mixed lipids, formation, 426–427 MOMP. See Mitochondrial outer membrane permeabilization Montelukast, 454–455 Motrinw, 448–449 MPT. See Mitochondrial permeability transition

802 mRNA expression profiles, 311 mtDNA. See also Mitochondrial injury degradation of, 55–59 acetaminophen (paracetamol), 55–56 alcohol, 55 mtDNA transcripts, decreased translation of, 60–67 chloramphenicol, 60 erythromycins, 60 thiamphenicol, 60 nucleoside analogs incorporated into mtDNA, 57–59 2’, 3’-dideoxynucleosides and abacavir, 57–59 replication by dideoxynucleosides, 58 fialuridine, 59 ganciclovir, 59 replication impairment by drugs, 56–57 dequalinium, 57 ditercalinium, 57 ethidium bromide, 57 1-methyl-4-phenylpyridinium ion (MPPC), 57 pentamidine, 57 polyamine analogs and methylglyoxal bis(guanine hydrazone), 57 quinolone antibiotics, 56 tacrine and tamoxifen, 56–57 replication, impairment by drugs, 56 treated with interferon-a, 59 MMP. See Matrix metalloproteinases Multidrug resistance (MDR), 101 multidrug resistance-associated protein (Mrp2), 41, 357 Multidrug resistance associated proteins (MRP), 635 Mushroom poisoning, 242, 254, 723–730. See also Amanita toxin-induced centrilobular necrosis, 723–730 Mycophenolate mofetil, 666 Myopathy, 571 N-(3060-demethoxy cinnamoyl)-antranilic acid (tranilast), 211 Nabumetone, 448 N-acetylcysteine (NAC), 348, 399, 641 N-acetyl-para-benzoquinoneimine (NAPQI), 89, 295 N-acetyltransferase, 317–318 NADRPS. See Naranjo adverse drug reaction probability scale NAFLD. See Nonalcoholic fatty liver disease Nalfonw, 449–450 Naprosyn, 449 Naproxen, 130, 135–136, 138, 254, 265, 449 Naranjo adverse drug reaction probability scale (NADRPS), 333, 340 NASH. See Nonalcoholic steatohepatitis Nateglinide, 617 NDA. See New drug application Necrosis, 85 Nefazodone, 2–3, 514

Index Nelfinavir, 578 Neoplasia/mass lesions, 224, 271–272 benign, 274 malignant, 274 oxymethalone, 275 Thorotras, 275–276 Neuroleptic agents, 514–517 benzamides, 515 sulpiride, 515 butyrophenones, 515 bromperidol, 515 haloperidol, 515 diazepines and oxazepines, 515–517 clozapine, 515–517 loxapine, 517 olanzapine, 517 liver injury, characteristics of, 516 molindone, 517 phenothiazines, 514–515 chlorpromazine, 514–515 cross-sensitization, 515 cyamemazine, 515 thioridazine, 515 trifluoperazine, 515 risperidone, 517 thioxanthenes, 515 Nevirapine, 3, 577 New drug application (NDA), 772 review and approval, 775–777 surveillance, 778 NF-kB role in cell death, 91–92 Niacin, 251, 265, 600 Nicotinic acid, 600 Nifedipine, 254, 265 Nimesulide, 68–69, 452–453 Nitric oxide synthase (NOS), 34 Nitrofurantoin, 2–3, 217, 232–233, 244–245, 254, 265, 533–534 NK cells, 165 N-methylthiourea, 716 NNRTI. See Non-nucleoside reverse transcriptase inhibitors Nodular regenerative hyperplasia, 642, 669 Nomifensine, 265 Nonalcoholic fatty liver disease (NAFLD), 593, 599, 615, 651, 712 Nonalcoholic steatohepatitis (NASH), 593, 606, 664 Non-hepatotoxic immunosuppressants, 675–676 Non-hepatotoxic medicinal plants, 734 Non-Hodgkin’s lymphoma, 673 Non-nucleoside reverse transcriptase inhibitors (NNRTI), 567–568, 577–579 Non-P450 genes, 317–320 HLA genotypes, 318–320 membrane transporters, 318 N-acetyltransferase, 317–318 pathways, 352–354 Nonparenchymal cells, inflammatory macrophages and hepatotoxicity, 159–184 hydrolytic enzymes, 170

803

Index

[Nonparenchymal cells, inflammatory macrophages and hepatotoxicity] inflammatory mediators implicated in hepatotoxicity, 165–170 and tissue injury, 159–165 Nonsteroidal anti-inflammatory drugs (NSAIDs), 228, 439–457 bioactivation to reactive metabolites, 409–412 coenzyme A to acyl-CoA thioesters, 410–411 cytochrome P450-mediated bioactivation, 409 peroxidases to pro-oxidant radicals, 410 UDP-glucuronosyltransferase to reactive acyl glucuronides and iso-glucuronides, 411–412 carboxylic acid-containing NSAIDs, 411 cellular and molecular mechanisms of NSAIDinduced hepatotoxicity, 414–427 disposition and metabolism of, 408–414 lipophilicity, 409 plasma protein binding, 408–409 hepatotoxicity, mechanisms, 407–437 biliary excretion, 412–413 cholehepatic circulation, 413–414 enterohepatic circulation, 413 hydrolysis and systemic cycling, 412 of NSAIDs, 407–412 renal excretion, 414 NSAID-induced hepatic injury, acetic acid derivatives, 445–448 bromfenac, 446–447 clometacin, 448 diclofenac, 445–446 etodolac, 446 fenclozic acid, 447 indomethacin, 447 ketorolac, 446 nabumetone, 448 sulindac, 447–448 tolmetin, 448 clinical and biochemical spectrum, 441–443 clincopathological features, 442 cyclooxygenase-2-selective inhibitors, 452–453 nimesulide, 452–453 fenamates, 452 mefenamic acid, 452 incidence of, 440 NSAID monitoring, 455–457 monitoring recommendations, 456 oxicams, 451 propionic acid derivatives, 448–451 benoxaprofen, 450–451 fenoprofen, 449–450 flurbiprofen, 450 ibuprofen, 448–449 ketoprofen, 450 naproxen, 449 oxaprozin, 450 pyrazolone derivatives, 451–452 azapropazone, 452 feprazone, 452 oxyphenbutazone, 452

[Nonsteroidal anti-inflammatory drugs (NSAIDs)] rheumatic diseases effect on the liver, 440–441 salicylates, 443–445 aspirin, 443–444 other salicylates, 444–445 Norephedrine hydrochloride, 742 Novobiocine, 56 NSAID-induced hepatotoxicity, cellular and molecular mechanisms of, 414–427. See also Protein adduct formation COX inhibition, 415 disruption of mitochondrial function, 415–416 hepatobiliary transport of endogenous or exogenous cholephilic compounds, 425–426 idiosyncratic liver toxicity caused by NSAIDs, 427–428 lipopolysaccharide release and inflammatory changes, 426 membrane permeability transition in mitochondria, 417–418 mitochondrial b-oxidation, inhibition of, 418–419 mixed lipids, formation, 426–427 oxidant stress and induction of apoptosis, 419 peroxisome proliferator-activated receptors, 415 pharmacological targets, 414 protein adduct formation and immune-mediated toxicity, 420–425 uncoupling of oxidative phosphorylation, 416–417 NSAIDs. See Nonsteroidal anti-inflammatory drugs N-substituted amides, 763 dimethylacetamide, 763 dimethylformamide, 763 Nucleoside reverse transcriptase inhibitors (NRTI), 567–571, 576–577 Nutritional status, as risk factor, 297 Obesity, 73 Obstructive cholestasis, 35 Occupational hepatotoxicity, 755–760 exposure, types of, 756–757 hepatotoxic chemicals, 759–760 injury, susceptibility to, 757–758 injury, types of, 755–756 mechanisms of injury, 758–759 testing, 758 Off-target toxicity, 18–19 Olanzapine, 517 Ondansetron, 636 On-target toxicity, 18–19 Oraflexe, 450–451 Oral contraceptive steroids, 707–710 cholestasis, 707–708 liver, benign neoplasms of, 709–710 focal nodular hyperplasia, 709 hemangioma, 709 hepatic adenoma, 710 hepatocellular carcinoma, 710

804 [Oral contraceptive steroids] other liver tumors, 710 vascular disorders of the liver, 708–709 peliosis hepatis, 708 portal vein and hepatic vein thrombosis, 708–709 Organic anion-transporting polypeptide (OATP), 41, 99, 534, 724 Organic solvents, 254 Organotin compounds, 66 Orotic acid, 254 Orudis, 450 Oxacillin, 265, 528 Oxaprozin, 135, 136, 138, 149, 265, 450 Oxazepines, 515–517 Oxazolidinones, 534 Oxicams, 451 Oxidant stress, 33–48. See also Antioxidant defense and induction of apoptosis, 419 pathophysiological consequences of, 36–38 lipid peroxidation (LPO), 36 peroxynitrite formation, 36 reactive oxygen and cell death, 37 reactive oxygen and gene transcription, 37–38 Oxygen and nitrogen intermediates, in antioxidant defense, 33 Oxyphenbutazone, 452 Oxyphenisatin, 233, 265 Oxytetracycline, 254, 529 P450 enzymes, 13–17 bioactivation by, 19–24 classification, 14 contribution in drug metabolism, 14 induction, 16–17 mechanisms, 16 inhibition, 15–16 regulation of, 15 variability effect on pharmacokinetics, 15 Paclitaxel, 634 Panadiplon, 61 Papaverine, 265 Para-aminosalicylic acid, 265 Paraquat, 764 Pathogen-associated molecular patterns (PAMP), 117 Pathology anesthetic agents, 465–479 leukotriene receptor antagonists, 454–457 nonsteroidal anti-inflammatory drugs, 439–457 PEGylated interferons, 575 Peliosis hepatis, 224, 264, 648, 708 Pemoline, 2–3 Penicillamine, 265 Penicillins, 265, 528–529 Pennyroyal oil, 254, 741–742 Pentamidine, 57 Pentenoic acid, 254 Perhexiline maleate, 64, 73, 254 Peripheral benzodiazepine receptor (PBR), 68

Index Peripheral blood mononuclear cells (PBMC), 117 Perisinusoidal fibrosis, 224 Peroxiredoxin (Prx), 39–40 Peroxisome proliferator-activated receptors (PPARg), 231 Peroxisome proliferators (P450 4A inducer), 16 Peroxisomes, 34 Peroxynitrite formation, 36 Perphenazine, 265 Personal genomes, 307–309 P-glycoprotein, 634 Phalloidin, 254 Phallotoxins, 725 Phencyclidine, 520 Phenobarbital, 211, 265, 492 Phenobarbitone, 470 Phenothiazines, 3, 514–515 Phenylbutazone, 2, 225, 254, 265 Phenytoin, 2–3, 209, 225, 245, 265, 490–492 clinical manifestations, 490 diagnosis, 490 pathology, 490 postulated mechanisms, 491–492 susceptibility factors, 490–491 Phosphoinositide-3-kinase (PI3K), 99 Phospholipidosis, 224, 254 Phosphorus, 254, 764–765 Phyllanthus amarus, 735 “P-i concept,” 18 Pigments, 277–284 Pioglitazone, 615–616 Piper methysticum rhizoma, 518–519 Piperazine, 265 Piroxicam, 254, 265, 451 Pirprofen, 254, 450 Pivampicillin, 62 Pizotyline, 265 Plaquenil, 673 Plasma antioxidants, 42 Plasma protein binding, 408–409 Plastic anemia, 496 Polyamine analogs and methylglyoxal bis(guanine hydrazone), 57 Polychlorinated biphenyls, 761 Polymyalgia rheumatica, 441 Polythiazide, 265 Portal fibrosis, 265–271, 272 bile ducts, 269 ethanol, 273 methotrexate, 272–273 Portal vein thrombosis (PVT), 708–709 Prajmalium, 265 Pregnane X receptor (PXR), 17, 293, 357 Primaquine, 65 Primary impairment of mitochondrial respiration, 53–54 Probenecid, 135, 141 Procainamide, 225, 265, 597 Prochlorperazine, 265 Proinflammatory cytokines, 165–166 Proliferator receptor g coactivator 1 (PGC1), 59

Index

Proliferator-activated receptor a (PPARa), 17 Promethazine, 254 Propafenone, 597–598 Propoxyphene, 265 Propylthiouracil, 3, 265, 716–717 Protease inhibitors (PI), 567–568, 577–579 Protein adduct formation and immune-mediated toxicity, 420–425 hapten formation and immune-mediated hepatic injury, 422–425 inactivation of critical protein function, 422 NSAID acyl glucuronides and iso-glucuronides, mechanisms and molecular targets of, 420–422 Protein expression, global assessment of, 311–312 Proteome analyses categories, 311 global expression profile, 311 modular proteomics, 311 organelle-based proteomics, 311 “Proteome,” 311–312 Proteomics, 27 of drug-induced liver disease, 307–324 Protonophoric drugs, mitochondrial respiration uncoupling by, 66–67 Psilocybe, 725 Psoriasis, 672, 686–689 hepatotoxic MTX drug interactions, 689 liver biopsy, 687 toxicity monitoring, 688–689 Psychotropic drugs, 507–519. See also Anxiolytic agents; Antidepressants; Neuroleptic agents epidemiological aspects, 507–508 hepatotoxicity contributing factors, 509 acquired factors, 509–510 genetic factors, 510 herbal medicines, 518–519 cannabis, 519 kava-kava, 518–519 valerian, 519 liver injury, identification of, 508–509 diagnosis, difficulties in, 508 other drugs, 517–518 fipexide, 517 riluzole, 518 tacrine, 518 tolcapone, 518 Pulegone, 741 Pulmonary blastomycosis, 535 Pyrazinamide, 3, 217, 552–553 Pyrexia, 466 Pyrrolizidine alkaloids, 254 Pyruvate dehydrogenase, 478 Quinethazone, 265 Quinidine, 225, 597 Quinine, 65 Quinolones, 56, 532–533 4-quinolones, 56

805 Radiation induced liver disease (RILD), 648, 650–651 vs. hepatic sinusoidal obstruction syndrome, 651 Reactive metabolite disposition, 359 “Reactive metabolite hypothesis,” 209 Reactive nitrogen intermediates, 168–169 nitric oxide synthase (NOSII), 168 nitric oxide, 168–169 peroxynitrite, 168 Reactive oxygen and cell death, 37 and gene transcription, 37–38 reactive oxygen intermediates, 167–168 reactive oxygen species (ROS), 33 Receptor interacting protein (RIP), 86 Regulatory perspectives, 771–786 controlled clinical trials, 774 drug development, critical path for, 777–778 drug safety, public concerns about, 771 drug-induced liver disease, 771–783 alternative causes exclusion, 782–783 close observation, 780–782 confirmation, 780 detection of DILI, 779–780 DILI and dysfunction, 779 FDA interest, 772 follow-up to resolution, 783 as a model, 771–772 susceptibility, mechanisms, 783 post-approval issues, 784 causality attribution, 785 epidemiologic, retrospective, observational studies, 785 hazard rates, 785 reporting rates vs. true incidence prospective safety studies, 785–786 reporting, problems of, 784 risk management, 785 processes vs. results, 785 uncommon adverse events detection, 784 preclinical and nonclinical research and development, 773 Reiter’s syndrome, 441 Relafenw, 448 Respiration, impairment of, 53–54 Respiratory chain polypeptides, 49 Retinoid X receptor (RXR), 17, 357 Retinol, 745–746 Reye’s syndrome (RS), 60–61, 441, 444, 494, 595 Rheumatic disease, 690–695 effect on the liver, 440–441 other rheumatic and inflammatory conditions, 694–695 rheumatoid arthritis, 690–694 liver biopsy, 692 MTX therapy, monitoring of, 690 toxicity monitoring, 693–694 Rheumatoid arthritis, 440 disease modifying anti-rheumatic drugs, 453–454 gold salts, 453–454 leflunomide, 453

806 Rifabutin, 579 Rifampicin, 579 Rifampin, 212, 213, 217, 246, 254, 265, 534, 552, 635 RILD. See Radiation induced liver disease Riluzole, 518 RIP. See Receptor interacting protein Risk factors for acetaminophen-induced hepatotoxicity, 390–391 for drug-induced liver disease, 291–305 age, 292 alcohol abuse, 296–297 cross-reactivity, 296 drug interactions, 295–296 genetic and familial predisposition, 294–295 nutritional status, 297 sex, 294 genetic and familial predisposition, 294–295 Ritonavir, 578–579 Rofecoxib, 453 Rolitetracycline, 254 Rosiglitazone, 348, 614–615 Roussel-Uclaf casuality assessment method scoring system, 7 Roxithromycin, 531 S-adenosyl-methionine, 735 Safety prediction, of new drugs, 26–28 use of covalent binding data, 28 mass spectrometry approaches, 27 metabonomics, 28 proteomics, 27 silico prediction, 27 transcriptonomics or mRNA profiling, 27 Salicylates, 298, 443–445 Salicylic acid, 68–69, 135, 136, 138, 141 Salsalate, 445 Saquinavir, 574, 579 Schizandra root, 735 Sclerosing cholangitis, primary, 696 SDS-polyacrylamide gel electrophoresis (SDS-PAGE), 472–473 Selenoprotein P, 42–43 Sertraline, 246, 252 Serum lipase, 571 Sevoflurane, 466–467, 469–470, 476 formation and biotransformation fluoromethyl-2,2,-difluoro-1-(trifluoromethyl)vinyl ether, 470 immune hypothesis, 476 Sex, as risk factor, 294 Signal transducers and activator of transcription 3 (STAT3), 163 Silico prediction, 27 Silybum marianum, 735 Simulate alcoholic hepatitis, 224 Simulate primary biliary cirrhosis, 224 Simulate primary sclerosing cholangitis, 224 Simvastatin, 598–599 Sinusoidal dilatation, 224, 264 Sinusoidal endothelial cells (SEC), 641, 667

Index Sirolimus, 665–666 Sjogren’s syndrome, 441 SLE. See Systemic lupus erythematosus Sodium usniate, 742 Sodium-taurocholate-cotransporting polypeptide (NTCP), 99 Specific serotonin reuptake inhibitor (SSRI), 513 Spectrum of drug-induced liver disorders, 223–225 Spiramycin, 531 Spironolactone, 254, 607 Splenomegaly, 490 St. John’s wort, 635 Statins, 598–600 Stavudine, 570, 572 Steatohepatitis, 64–65, 224, 253, 636 Steatosis, 313 Stellate cells, 38, 164–165 Steroids, for hepatitis-like reactions, 348 Stevens-Johnson syndrome/toxic epidermal necrolysis, 319, 489 Streptogramins, 534 Streptomycin, 212, 213, 217, 327 Sudoxicam, 451 Sulfamethoxazole trimethoprim, 265 Sulfapyridine, 674 Sulfasalazine, 254, 265, 674 Sulfonamides, 3, 6, 209, 212, 225, 265, 531 Sulfonylureas, 611–613 Sulindac, 2–3, 228, 254, 265, 447–448 Sulpiride, 515 Superoxide dismutases (SOD), 38, 90 Suprofen, 129, 135, 138 Systemic lupus erythematosus (SLE), 298, 441–444, 448 Tachyarrhythmias, 594 Tacrine, 56–57, 213, 231, 518 Tamoxifen, 56–57, 65, 254, 265, 651–652, 712–714 steatosis/steatohepatitis, pathogenesis of, 712–714 genetic factors, 713 insulin resistance, 712–713 mitochondrial effects, 713 surveillance, 714 treatment and outcome, 713 Telithromycin, 531 Telmisartan, 135 Tenofovir, 570 Tenoxicam, 451 Terbinafine, 3, 536–537 Terfenadine, 18–19 Tetrachloroethane, 254, 761–762 Tetracyclines, 62, 73, 254, 531–532 TGFb, 166–167 Thalidomide, 675 Thallium compounds, 254 Thiabendazole, 265 Thiamphenicol, 60 Thiazoladinediones, 348, 613–616 Thioacetamide, 161, 188 6-Thioguanine, 265, 632, 666–669

Index

Thiopental, 265 Thiopurine S-methyltransferase (TPMT), 648 Thiopurines, 632, 647–649, 666–669 metabolism, 632 Thioredoxin-1 (Trx-1), 39–40 Thioridazine, 265, 515 Thioxanthenes, 515 6-Thiouric acid, 668 Thombolytics, 611 Thrombocytopenia, 597 Tianeptine, 62–63 Ticlopidine, 265, 610–611 Ticrynafen, 2, 254 Tienilic acid, 233, 607 Tigecycline, 532 Tinzaparin, 609 Tipranavir, 579 Tissue repair role, in liver injury, 185–206 dose response for, 188 drugs and toxicants that stimulate, 188 mixtures and/or combinations, 188 single chemicals, 188 factors affecting, 191–196 age, 192–193 caloric restriction, 193–194 liver tissue repair in diabetes, 194–196 nutrition, 193 species and strain differences, 191–192 following a dose response, 186–190 higher initiation of liver injury and tissue repair, relationship between, 190–191 progression of liver injury, 196–199 death proteins role in, 198 mechanisms, 197 regression of injury, 199–201 significance, 201 two-stage model of toxicity, 186 TNF receptor associated factor 2 (TRAF2), 86 TNFR associated death domain (TRADD), 85 Tocainide, 265 Tolazamide, 265, 612 Tolbutamide, 265, 612 Tolcapone, 3, 211, 518 Tolectin, 448 Toll-like receptor-4 (TLR-4), 160 Tolmetin, 129, 135, 138, 141, 149, 254, 448 Topoisomerases, 56 Toradolw, 446 Toremifene, 714 Total parenteral nutrition, 254, 265 Toxic oil (rapeseed), 265 Toxicities acetaminophen metabolism and, 19 bioactivation to reactive intermediates, 18 categories, 17 contexts of, 17–19 hypersensitivity and immune reactions, 18 idiosyncratic, 18 off-target, 18 on-target, 18 Transacylation, 126–127

807 Transcriptional arrest in hepatocytes, 91–92 “Transcriptome,” 27, 309–311 Transferrin saturation, 329 Tranylcypromine, 265 Trazodone, 514 Triazolam, 265 Trichloroethylene (TCE), 22–23, 188, 254 Tricyclicsa, 3 Trifluoperazine, 265, 515 Trifluoroacetyl hapten, 477 Trifluoroacetylated liver neoantigens, 473 mechanism of generation, 474 Triggering receptor on myeloid cells (TREM), 163 Triglycerides (TG), 52 Trimethobenzamide, 265 Tripelennamine, 265 Troglitazone, 2–3, 8–9, 23–24, 69, 194, 213, 231–232, 252, 613–614 activation by P450 3A4, 25 Troleandomycin, 265, 531 Trovafloxacin, 2–3, 327, 533 Tuberculosis recommended ATT schedule, 549 Tumor necrosis factor (TNF), 664, 672–673, 725 Tumor necrosis factorreceptor 1 (TNFR1), 85 Turmeric, 735 UDP-glucuronosyltransferase (UGT), 411, 574 UDP-glycoprotein glucosyltransferase (UGGT), 474 UGTs (UDP-glucuronosyltransferases), 352 Uncoupling, of oxidative phosphorylation, 416–417 Uncoupling protein 2 (UCP2), 68 Upper limit of normal (ULN), 1, 441, 445–447, 453–457, 548–553, 556–562, 577, 595 Uranium compounds, 254 Uridine 50-diphosphate (UDP)-glucuronic acid (UDPGA), 128 Ursodeoxycholic acid, 715 for cholestasis management, 348–349 Valdecoxib, 453 Valerian, 519 Valproate, 6, 73, mitochondrial effects of, 61 valproate (VPA)-induced microvesicular steatosis, 293 Valproic acid (VPA), 2–3, 60–61, 68–69, 134, 135, 141, 254, 265, 492–496 clinical manifestations, 492–493 pathology, 493 postulated mechanisms, 494–492 biotransformation, pathways for, 495 susceptibility factors, 493–494 Vancomycin, 533 Vascular injury, 264–265 Vascular lesions, 224 Vascular oxidant stress, 35–36 VCM, 211 Veno-occlusive disease, 224, 264. See also Hepatic sinusoidal obstruction syndrome Verapamil, 265, 606, 633–634

808 Very low-density lipoprotein (VLDL), 52 Vinca alkaloids, 631, 636 Vincrisitine, 650 Vinyl chloride, 762–763 Viral serology, 329 Virotoxins, 725 Vitamins, 744–746 iron, 744–745 vitamin A, 254, 745–747 chronic vitamin A hepatotoxicity, 747 Voriconazole, 536 VPA. See Valproic acid Warfarin, 254, 608 Wilson’s disease, 674

Index Wy-18251, 135 Wy-41770, 135 Xanthine dehydrogenase, 35 Ximelagatran, 8, 609 Yohimbine hydrochloride, 742 Zafirlukast, 3, 454 Zalcitabine, 570 Zenarestat, 135 Zidovudine, 570 Zileuton, 3, 455 Zimelidine, 265 Zomepirac, 129, 135–136, 138, 141, 146, 149

FIGURE 13.10 Lobular necrosis with inflammation: halogenated hydrocarbons. The parenchyma shows striking diffuse necroinflammatory change with prominent Kupffer cell hyperplasia. Mild macrovesicular fatty change is also present. (See p. 244 )

FIGURE 13.31 Fatty change: ethanol. Almost all of the hepatocytes contain microvesicular fat, defined as fat globules smaller than the size of the liver cell nucleus. This type of alcoholic liver disease is termed acute foamy degeneration. (See p. 257 )

FIGURE 13.33 Fatty change: Tetracycline (Oil Red O). This stain for neutral lipids performed on frozen section of fresh or formalin-fixed tissue shows that all of the liver cells contain abundant fat. Source: From Ref. 35. (See p. 258 )

FIGURE 13.34 Fatty change: Tetracycline (autofluorescence). The abundant golden-brown strongly fluorescent pigment within the liver cell cytoplasm represents tetracycline. Source: From Ref. 35. (See p. 258 )

FIGURE 13.53 Bile duct injury: chlorpropamide. This high power photomicrograph of a portal tract shows considerable irregularity of the size and shape of the duct epithelium and nuclei. Lymphocytes can be seen surrounding and focally infiltrating into the duct wall and lumen itself (non-suppurative cholangitis). (See p. 272 )

FIGURE 13.55 Vascular, veno-occlusive disease: Busulfan (conditioning regimen for bone marrow transplantion) (trichrome). This terminal hepatic venule shows considerable intraluminal collagen deposition with almost total occlusion of the vascular lumen. Collagen is also present within the adjacent sinusoids. Source: Courtesy of Drs. Howard Shulman and Fred Hutchison; From Ref. 35. (See p. 274 )

FIGURE 13.64 Inclusions: particulate injectant (IV drug users) (polarized light). The injectant admixed with the drug usually appears within portal tracts as a crystalline-like material that is strongly birefringent under polarized light. Source: From Ref. 35. (See p. 280 )

FIGURE 13.66 Pigment: thorotrast. The coarsely granular pigment representing thorotrast is seen in this cluster of Kupffer cells and histiocytes within the lobule in a patient with non-cirrhotic portal fibrosis ("idiopathic portal hypertension"), a disorder that can be associated with thorotrast exposure. (See p. 282 )