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).