Lipid A in Cancer Therapy (Advances in Experimental Medicine and Biology Vol 667)

Lipid A in Cancer Therapy

ADVANCES IN EXPERIMENTAL MEDICINE AND BIOLOGY EditorialBoard: NATHAN BACK,State University ofNew Yorkat Buffalo IRUN R. COHEN, The Weizmann InstituteofScience ABEL LAJTHA, N.S. Kline Institutefor PsychiatricResearch JOHN D. LAMBRIS, University ofPennsylvania RODOLFOPAOLETTI, University ofMilan Recent Volumes in this Series Volume 660 PARAOXONASES IN INFLAMMATION, INFECTION, AND TOXICOLOGY Editedby Srinu Reddy Volume 661 MEMBRANE RECEPTORS, CHANNELS AND TRANSPORTERS IN PULMONARY CIRCULATION Editedby Jason X. -J. Yuan, and Jeremy P.T. Ward Volume 662 OXYGENTRANSPORT TO TISSUEXXXI Edited by Duane F. Bruley and Eiji Takahasi Volume 663 STRUCTURE AND FUNCTION OF THE NEURALCELLADHESION MOLECULE NCAM Edited by VladimirBerezin Volume 664 RETINAL DEGENERATIVE DISEASES Edited by Robert E. Anderson, Joe G. Hollyfield, and MatthewM. LaVail Volume 665 FORKHEAD TRANSCRIPTION FACTORS Edited by KennethMaiese Volume 666 PATHOGEN-DERIVED IMMUNOMODULATORY MOLECULES Editedby PadraicG. Fallon Volume 667 LIPIDA IN CANCERTHERAPY Editedby Jean-Francois Jcannin

AContinuation Order Planisavailable forthis series. Acontinuation order willbring delivery ofeachnewvolume immediately upon publication. Volumes are billed onlyupon aetual shipment. For further information please contact thepublisher.

Lipid A in Cancer Therapy Editedby Jean-Francois Jeannin

Tumor Immunology and Immunotherapy Laboratory Ecole Practique des Hautes Etudes Inserm U866, University ofBurgundy, Dijon , Fran ce

Springer Science+Business Media, LLC Landes Bioscience

Springer Science+Business Media, LLC LandesBioscience Copyright©2009LandesBioscience and SpringerScience+Business Media,LLC All rights reserved. No partof thisbookmaybe reproduced or transmitted inany formor by anymeans, electronic or mechanical, includingphotocopy, recording, or any information storageandretrievalsystem,withoutpermission in writingfromthe publisher, with the exceptionof any materialsuppliedspecifically for the purposeof being enteredand executedon a computersystem; for exclusive use by the Purchaserof the work. Printedin the USA. SpringerScience+Business Media,LLC, 233 SpringStreet, New York, New York10013, USA http://www.springer.com Pleaseaddress all inquiriesto the publishers: LandesBioscience, 1002WestAvenue, Austin, Texas7870I, USA Phone: 5121637 6050; FAX: 512/637 6079 http://www.landesbioscience.com The chaptersin this book are availablein the MadameCurie Bioscience Database. http://www.landesbioscience.comlcurie

LipidA in Cancer Therapy, editedbyJean-Francois Jeannin. Landes Bioscience1Springer Science+Business Media, LLC dual imprint1 Springerseries: Advances in Experimental Medicineand Biology. ISBN: 978-1-4419-1602-0 Whilethe authors,editorsandpublisherbelievethatdrugselectionand dosageand thespecifications and usage of equipment and devices, as set forth in this book, are in accordwith current recommendations and practice at the time of publication, they make no warranty, expressedor implied, with respect to material describedin this book. In view of the ongoingresearch, equipmentdevelopment, changes in governmental regulationsand the rapidaccumulation of information relating to the biomedical sciences, the reader is urgedto carefullyreviewand evaluatethe information providedherein.

Library of Congress Cataloging-in-Publication Data LipidA in cancertherapy1 editedby Jean-Francois Jeannin. p, ; em. -- (Advances in experimental medicineand biology; 667) Includes bibliographical references and index. ISBN 978-1-4419-1602-0 1. Cancer--Immunotherapy. 2. Microbial lipids-Therapeutic use. 3. Endotoxins--Therapeutic use. r. Jeannin, Jean-Francois, 1948- II. Series: Advances in experimental medicine and biology, v. 667. 0065-2598 ; [DNLM: 1. Lipid A--pharmacology. 2. Lipid A--therapeutic use. 3. Neoplasms-drug therapy. WI AD559v.667 20091 QU 85 L7605 2009] RC27I.I45L572009 616.99'4061--dc22 2009035583

FOREWORD Cancer remains a major challenge for modem society. Not only does cancer rank among the first three causes of mortality in most population groups but also the therapeutic options available for most tumor types are limited. The existing ones have limited efficacy, lack specificity and their administration carry major side effects. Hence the urgent need for novel cancer therapies. One of the most promising avenues in research is the use of specific immunotherapy. The notion that the immune system may have important anti-tumor effects has been around for more than a century now. Every major progress in microbiology and immunology has been immediately followed by attempts to apply the new knowledge to the treatment of cancer. Progress has reached a point where it is well established that most cancer patients mount specific T cell responses against their tumors . The molecular identity of the antigens recognized by anti-tumor T cells has been elucidated and several hundreds oftumor-derived antigenic peptides have been discovered. Upon recognition of such peptides presented by self MHC molecules, both CD8 and CD4 T cells are activated, expand to high numbers and differentiate into effective anti-tumor agents. CD8 T cells directly destroy tumor cells and can cause even large tumors to completely regress in experimental mouse models . These observations have spurred intense research activity aimed at designing and testing cancer vaccines . Over 100 years ago Coley successfully used intratumoral injection of killed bacteria to treat sarcomas. The important anti-tumor effects observed in a fraction of these patients fueled major research efforts. These led to major discoveries in the 80s and the 90s. It turns out that bacterial lipopolysaccharides stimulate the production of massive amounts of a cytokine still known today as tumor necrosis factor (TNF-a). They do so by engagement of a rather complex set of interactions culminating in the ligation ofa Toll-like receptor, TLR-4. Ensuing signaling through this receptor initiates potent innate immune responses. Unfortunately the clinical use of both TNF-a and LPS can not be generalized due to their very narrow therapeutic margin. Importantly, synthetic Lipid A analogs have been identified that retain useful bioactivity and yet possess only mild toxicity. v

vi

Foreword

The relatively large body of information accumulated thus far on the molecular and cellular interactions set in motion by administration of LPS as well as by the synthetic Lipid A analogs allow to place this family ofbacterially-derived molecules at the crossroads between innate and adaptive immunity. By virtue of this key position, the therapeutic applications being pursued aim at using these compounds either as direct anti-tumor agents or as vaccine adjuvants. The clinical experience acquired so far on these two avenues is asymmetric. Few clinical trials using Lipid A analogs as single anti-cancer agents involving less than 100 patients with advanced cancer have been reported. In contrast, Lipid A has been tested in over 300,000 individuals in various vaccines trials, including therapeutic cancer vaccines. Clearly most of the work needed to develop Lipid A as effective anti-cancer agents and/or as vaccine adjuvant lies ahead in the near future . This book is a timely contribution and provides a much needed up-to-date overview of the chemical, biological and physiological aspects of Lipid A. It should be a beacon to all those involved in this field of research.

Jean-Charles Cerottini, MD UniversityofLausanne, FormerDirector, LudwigInstitutefor Cancer Research Lausanne Branch PedroRomero, MD University ofLausanne, Member, Ludwig Institutefor Cancer Research Lausanne Branch

ABOUT THE EDITOR...

JEAN-FRAN<;OIS JEANNIN is Professor ofImmunology at Ecole Practique des Hautes Etudes (EPHE) and director of the EPHE Tumor Immunology and Immunotherapy Laboratory, an INSERM (National Institute of Health and Medical Research) team . His main research interests have included the effects oflipopolysaccharides in the tumor immune response and the immunotherapy ofcolon cancer with Lipid A. Now he is investigating mechanisms ofimmunotherapy with synthetic Lipid A analogs in cancer patients and animal cancer models. He is especially interested in the sensitization of tumor cell death by nitric oxide produced in tumors during Lipid A immunotherapy. Jean-Francois Jeannin has been Dean and President of the Life Sciences faculty of EPHE and a member of numerous scientific organizations.

vii

PARTICIPANTS Shizuo Akira Department of Host Defense Research Institute for Microbial Diseases Osaka University and ERATO Japan Science and Technology Corporation Osaka Japan

JorgAndra

Biophysics Division Research Center Borstel Leibniz-Center for Medicine and Biosciences Borstel Germany

Klaus Brandenburg Division of Biophysics Research Center Borstel Leibniz-Center for Medicine and Biosciences Borstel Germany Jean-Charles Cerottini University of Lausanne Ludwig Institute for Cancer Research Lausanne Switzerland Christopher W. Cluff GlaxoSmithKline Biologicals Hamilton, Montana USA

Marc Bardou Clinical Pharmacology Unit and Laboratory of Cardiovascular Experimental Physiology and Pharmacology Dijon France

Thomas Gutsmann Department of Immunochemistry and Biochemical Microbiology Research Center Borstel Leibniz-Center for Medicine and Biosciences Borstel Germany

Ali Bettaieb Tumor Immunology and Immunotherapy Laboratory Ecole Practique des Hautes Etudes Inserm U866 University of Burgundy Dijon France

Masahito Hashimoto Department of Nanostructure and Advanced Materials Graduate School of Science and Engineering Kagoshima Univers ity Korimoto , Kagoshima Japan ix

x

Jean-Francois Jeannin Tumor Immunology and Immunotherapy Laboratory Ecole Practique des Hautes Etudes Insenn U866 University of Burgundy Dijon France Kazuyoshi Kawahara Department ofApplied Material and Life Science College of Engineering Kanto Gakuin University Yokohama, Kanagawa Japan

Participants

Nilofer Qureshi Department of Basic Medical Science School of Medicine Shock Trauma Research Center University of Missouri Kansas City, Missouri USA Catherine Paul Tumor Immunology and Immunotherapy Laboratory Insenn U866 University of Burgundy Dijon France

Shoichi Kusumoto Suntory Institute for Bioorganic Research Shimamoto-cho Mish ima-gun, Osaka Japan

Daniele Reisser Tumor Immunology and Immunotherapy Laboratory Insenn U866 University of Burgundy Dijon France

Amandine Martin Tumor Immunology and Immunotherapy Laboratory Insenn U866 University of Burgundy Dijon France

Cheryl E. Rockwell Department of Basic Medical Science School of Medicine Shock Trauma Research Center University of Missouri Kansas City, Missouri USA

David C. Morrison Department of Basic Medical Science School of Medicine Shock Trauma Research Center University of Missouri Kansas City, Missouri USA

Pedro Romero University of Lausanne Ludwig Institute for Cancer Research Lausanne Switzerland

Mareike Muller Department of Immunochemistry and Biochemical Microbiology Research Center Borstel Leibniz-Center for Medicine and Biosciences Borstel Germany

Nejia Sassi Tumor Immunology and Immunotherapy Laboratory Insenn U866 University of Burgundy Dijon France

Participants

xi

Andra B. Schromm Department of Immunochemistry and Biochemical Microbiology Research Center Borstel Leibniz-Center for Medicine and Biosciences Borstel Germany

Kiyoshi Takeda Department of Molecular Genetics Medical Institute of Bioregulation Kyushu University Fukuoka Japan

Ulrich Seydel Division of Biophysics Research Center Borstel Leibniz-Center for Medicine and Biosciences Borstel Germany

Masahiro Yamamoto Department of Host Defense Research Institute for Microbial Diseases Osaka University Osaka Japan

CONTENTS FOREWORD

v

Jean-Charles Cerottini and Pedro Romero

1. INTRODUCTION: HISTORICAL BACKGROUND

1

Jean-Francois Jeannin

2. STRUCTURE AND SYNTHESIS OF LIPID A

5

Shoichi Kusumoto, Masahito Hashimoto and Kazuyoshi Kawahara Introduction General Architecture of Lipid A Structural Variations of Lipid A Chemical Synthesis of Lipid A Conclusion

3. CONFORMATION AND SUPRAMOLECULAR STRUCTURE OF LIPID A.

5 6 9 17 19

25

Klaus Brandenburg and Ulrich Seydel Abstract Introduction Aggregate Structure and Molecular Conformation Intramolecular Conformation Phase States and Transitions between Them Molecular Modelling Physicochemical Data in Relation to Biological Activity Conformational Concept of Lipid A Action: How Does Endotoxin Interact with Immune Cells?

25 25 27 28 29 31 32 33

xiii

xiv

Contents

4. INTERACTIONS BETWEEN LIPID A AND SERUM PROTEINS......... 39 Jorg Andra, Thomas Gutsmann, Mareike Miiller and Andra B. Schromm Abstract Proteins Involved in Lipid A/LP8-Mediated Immune Cell Activation Detection of Lipid A by Immunoglobulins Proteins Involved in Lipid A/LPS Transport Lipoproteins Proteins Neutralizing the Immune-Cell Activating Properties of Lipid A/LPS Conclusion

39 39 41 41 42 44 46

5. THE LIPID A RECEPTOR

53

Kiyoshi Takeda Abst ract Introduction Components of the Lipid A Receptor Conclusion

53 53 53 56

6. LIPID A RECEPTOR TLR4-MEDIATED SIGNALING PATHWAyS .... 59 Masahiro Yamamoto and Shizuo Akira Abstract Introduction The MyD88-Dependent and MyD88-Independent Pathways TRIF: The TIR Domain-Containing Signal Transducer for the MyD88-Independent Pathway TIRAP and TRAM: Anothe r Two TIR Domain-Containing Molecules Negative Regulation of LPS-Induced Signaling Pathways Two-Step Gene Induction Program in TLR4-Mediated Immune Responses Conclusion

59 59 60

7. LIPID A-INDUCED RESPONSES IN VIVO

69

60 62 62 64 66

Nejia Sassi, Catherine Paul, Amandine Martin, Ali Bettaieb and Jean-FrancoisJeannin Abstract Fate of Lipid A in the Bloodstream General Immune Responses to Lipid A Systemic Toxicity Lipid A Tolerance Tumor Immune Responses Vascular Response Conclus ion

69 69 70 71 72 73 75 76

Contents

xv

8. LIPID A-MEDIATED TOLERANCE AND CANCER THERAPy

81

Cheryl E. Rockwell , David. C. Morrison and Nilofer Qureshi Abstract Early, Late and Cross Tolerance The Isolation of Various Lipid A Structures and Synthesis of Analogs Relevance of Tolerance to the Use of LPS/Lipid A in Cancer Mechanisms of Early LPS/Lipid A-Mediated Tolerance Mechanisms of Tolerance of Other Lipid A Structures and LPS Antagonists Conclusion

81 81 82 83 86 91 91

9. LIPID A IN CANCER THERAPIES: PRECLINICAL RESULTS ......... 101 Daniele Reisser and Jean-Francois Jeannin Abstract Introduction LPS Treatments Lipids A Treatments DT-5461 ONO-4007 OM-174 Conclusion

101 101 101 102 103 104 106 107

10. MONOPHOSPHORYL LIPID A (MPL) AS AN ADJUVANT FOR ANTI-CANCER VACCINES: CLINICAL RESULTS

111

Christopher W. Cluff Abstract Section 1: Vaccines Targeting Specific Cancer Types Section 2: Vaccines Targeting Specific TAAs Expressed on Multiple Tumor Types Conclusion

11. ANTITUMORAL EFFECTS OF LIPIDS A, CLINICAL STUDIES

111 112 115 119

125

Marc Bardou and Daniele Reisser Abstract Introduction Immunological Background Underlying the Clinical Potential Interest.. Clinical Studies Conclusion

125 125 126 127 129

xvi

12. CONCLUSION

Contents

133

Jean-Francois Jeannin INDEX

135

CHAPTER!

Introduction: Historical Background Jean-Fran~oisJeannin

T

reatm ent of canc er patients with lipid A analogs is no w feasible. Thi s is the culmination ofa long story, beginning hundreds of years ago, of progress in different scientific fields bacteriology, chemistry, immunology, genetics, cell biology and experimental medicine. Knowing the hi story ofthis domain ofresearch is important in understanding why there is increa sing acceptance among the scientific comm un ity ofattempts to tr eat cancer with lipid A analogs. Immunotherapy has been used to treat cancer patients for a longtime even before the principles ofimmunology were known. Already several hundred years ago, the coin cidenc e between the spontaneous curing oftumors and the occurrence ofacute infections had been noted. At the beginning of the 18th century: Deidier described the regression of tumors that had been scarified with pus from boils. Coley, after first using live bacteria, used a preparation ofkilled Streptococcus pyogenes and Serratia marcescens , later known as Coley'stoxin, " to successfully treat cancer patients. In 1943, Shear and Turner" found that the antitumor effect of Coley's toxin was due to endotoxins. Pfeiffer,d working at the end ofthe 19th century, had introduced the term endotoxin, as opposed to toxins or exotoxins, to name toxic bacterial components that are not excreted from living bacteria but are released during bacteriolysis.The first indications as to th e biochem istr y ofendotoxins were given by Boivin ' between 1935 and 1945. He purified substances composed ofpol ysaccharide and lipid (with only small amounts ofprotein), named lipopolysaccharid es (LPS) by Shear and T urner,' Now we know that LPS are con stituents of the outer membrane of Gram-negative bacteria. The chemical and biological properties ofLPS were extensi vely studied by Westphal and Luderirz.l In the 1960s , the y proposed that all LPS had a common architecture made up of3 components: an O vspecific side cha in, a core and lipid A. Lipid A, the lipid component, is stru ct ur ally the most stable part of LPS. Lipid A consi sts of a diglucosamine backbone acylated by ester-linked and amide-linked long chain fatty acids. The nature of th e backbone glycosyl residues is the sam e for many LPS. The pre cise structures of lipid A from Escherichia coli and Salm onella ryph imuriurn were obtained at the beginning of the 1980s by Taka yama, Kusurnot o, Galanos and Riet schel.g.h·2 Lipid A molecules were first chem ically synthesized by Kusumoto and collaborators in 1985 .' Since then it has become po ssible to synthesize lipid A analogs. Lip id A , eith er from bacteria that cannot synthesize polysaccharides or from chemical synthesis, reproduces the toxic effect ofLPS in vivo so it is cons idered to be the biologically active part ofLPS. The di scovery in 1968 that a strain of mice failed to respond to LPS was a decisive step in understanding how LPS act.' The genetic locus responsible for this resistance was mapped and -De ld le r, A., Deux d issertations rnedic inales et chirurgicales, l'une sur la maladie vener ienne, l'autre sur la natu re et la curation des tumeurs, D'Houzy, CM . Paris: Impr imeur 1725. "Co ley, WB . The measurement of inoperable sarcoma with the mixed toxin of Erysipelas and Bacillus prodigiosus. J Am Med Assoc 1898; 3:589-595. "Shear MJ and Turner FC, Chemical Treatment of Tumors. V. Isolat ion of the Hemorrhage-Producing Fraction from Serratia marcescens (Bac illus prodigiosus) Culture Filtrate. J Nat Cancer Inst 1943; 4:81-9 7. ·Correspond ing Author: jean -Francois Jeannin- EPHE, Dijon, F-21000, France; Inserm U866, Dijon, F-21000, France. Email: jean-francois.jea [email protected]

Lipid A in Cancer Therapy , edited by jean-Francoisjeannin, ©2009 Landes Bioscien ce and Springer Science+Busine ss Media.

2

LipidA in Cancer Therapy

called the Ipsgene.' In 1989 Wright et al. discovered a serum protein they named LBP,LPS binding protein, that binds LPS and enhances macrophage responses to LPS.5 They also found that LBP-LPS complexes bind the leukocyte antigen CD146 thought to be the LPS receptor. However CD14 lacks transmembrane and intracellular domains and cannot transmit a signal. The Drosophila Toll receptors are essential for the production of antimicrobial pep tides in response to pathogens. In 1997 Medzhirov et al. identified and cloned a human Toll-like receptor (TLR) and showed that innate immunity is conserved from Drosophila to humans? A family of mammalian TLR were discovered which recognize microbial products and activate inflammatory responses. Poltorak et al. reported in 1998 that the mouse Ips gene encodes TLR4.8 TLR4 can be considered as the receptor oflipid A and LPS ; it is mainly expressed by immune cells but also by endothelial and epithelial cells which are potentially lipid A responsive. Since the experiments of Coley, huge progress has been made in immunology and while it is not the place here to summarize all these findings, it is useful to underline those principles which are important in antitumor immune activity. The innate immune system is a universal and ancient defense mechanism ofthe host, whereas the adaptive immune system has originated more recently and is only found in vertebrates. In the innate immune system a limited number of conserved molecules produced by pathogenic microbes are recognized by receptors encoded by invariable genes. On the contrary, the adaptive immune system recognizes many specific details (epiropes) of proteins or carbohydrates whatever their origin , via receptors encoded by gene segments that have undergone necessary rearrangements. In the innate immune system, the same receptors are expressed by all the cells ofthe same type (neutrophils, macrophages, etc.), while in the adaptive immune system, specific receptors are expressed only on cells coming from the clonal expansion ofone cell (T or B cell) activated by the specific recognition ofone epitope. There are important interactions between the innate and adaptive immune systems. It was proposed that B or T cell activation by antigen must be followed by a second signal from helper T cellsfor proliferation and differentiation to be complete,"To produce this second signal, helper T cells must be activated by innate immune cells (co-stimulation).]aneway suggested that innate immune cells, macrophages or antigen presenting cells (APC), are not constitutively active'? but must themselves be stimulated to produce the co-stimulatory signals. He proposed that APC are stimulated through pattern recognition receptors (PPR) that recogn ize conserved patterns of molecules found only on evolutionarily ancient organisms . II While the first model can explain the absence ofa tumor immune response by the absence ofco-stimulation, the second model does not explain the immune response to tumors. Matzinger has since developed the Danger Model proposing that APC are activated by a danger signal from injured cells, a signal which cannot be sent by healthy cells or by cells undergoing physiological death." According to this model, the tumor cell death (necrosis) induced by innate immune cells allows APC to capture, process and present tumor antigens and activates AP C to produce co-stimulatory signals. The triggering event is the induction of tumor cell death due to innate immune cells. This book summarizes the answers to important questions posed by the use oflipid A analogs in clinic. It is presented in two parts : the rationale behind clinical trials-the chemical and biochemical aspects and preclinical results ; and the clinical data itself-the use of lipid A as an adjuvant in vaccines against cancer and lipid A analogs in Phase I (toxicity) trials. What are the naturally occurring lipid A structures, is it possible to synthesize lipid A analogs and what are the difficulties encountered? Answers to these central questions are given in Chapter 2. What are the conforma-

dPfeiffer R. Untersuchungen uber das Choleragift. Z Hygiene 1892; 11 : 393-412. "Boivin, A., Travaux recents sur la constitution chimique et les proprietes biologiquesdes antlgenes bacteriens, Schweiz Z Pathol Bacteriol 1946; 9: 505 -541. fsee foot note 3. 81moto M, Kusumoto S, Shiba T et al. Chemical structure of E. coli lipid A: linkagesite of acyl groups in the disaccharide backbone. Tetrahedron Lett 1983; 24:4017-4020. hWestphal 0, l.iiderltz 0 , Galanos C et al. In: Chedid, Hadden et ai, eds. Adv Immunopharmacol. Oxford: Pergamon Press 1985:13-34. 'lrnoto M, Yoshimura H, Sakaguchi N et al. Total synthesis of Escherichia coli lipid A. Tetrahedron Lett 1985; 26:1545-1548.

Introduction

3

tions and the supramolecular structures oflipid A molecules and analogs able to activate immune cells? Answers are given in Chapter 3. Chapter 4 presents the latest reports on the fate of lipid A molecules in serum and their interactions with serum protein in vitro . Chapter 5 swnmarizes current knowledge ofthe lipid A receptors and Chapter 6 covers the signaling pathways induced when TLR4 binds lipid A. What responses are induced in vivo by lipid A and its analogs in humans, mice and rats? These are described in Chapter 7. The role oflipid A tolerance in cancer therapy is discussed in Chapter 8. Chapter 9 swnmarizes the preclinical results oflipid A cancer therapy in animal models . Chapter 10 presents clinical results ofthe use of lipid A as an adjuvant in cancer vaccination and the last chapter the results ofPhase I clinical trials with lipid A analogs.

References

1. Galanos C, Lehmann V, Luderitz 0 et al. Endotox ic properties of chemically synthesized lipid A part structures . Comparison of synthetic lipid A precursor and synthetic analogues with biosynthetic lipid A precursor and free lipid A. Eur J Biochern 1984; 140:221-227. 2. Takayama K, ~eshi N. and Mascagni P et al. Complete structure of lipid A obtained from the Iipopolysaccharides of the heproseless mutant of Salmonella typhimurium. J Bioi Chern 1983; 258: 12801-12803. 3. Sultzer B M. Genetic control of leucocyte responses to endotoxin . Nature 1968; 219:1253-1254. 4. Rosensrreich DL , Vogel SN, Jacques AR., Wahl LM et al. Macrophage sensitivity to endotoxin : genetic control by a single codominant gene. J Imrnunol 1978; I2 1:1664-1670 . 5. Wright SD. Tobias PS. Ulevirch RJ et al. Lipopolysaccharide(LPS) binding protein opsonizesLPS-bearing particles for recognition by a novel receptor on macrophages. J Exp Med 1989; 170:1231-1241. 6. Wright SD. Ramos RA, Tobias PS et al. CDI4. a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 1990; 249:1431-1433 . 7. Medzh itov R. Preston-Hurlburt P and Janeway CA. Jr. A human homologue of the Dro sophila Toll protein signals activation of adaptive immunit y. Nature 1997; 388:394-397. 8. Poltorak A. He X. Smirnova I. et al. Defective LPS signaling in C3H/HeJ and C57BL/I0ScCr mice: mutat ions in Tlr4 gene. Science 1998; 282:2085-2088. 9. Bretscher P, and Cohn M. A theory of self-nonself discriminat ion. Science 1970; 169:1042-1049. 10. Medzhitov R. and Janeway CA. Jr. Innate immunity : impact on the adaptive immune response. Curr Opin Immuno11997; 9:4-9. 11. Medzhitov R. and Janeway, CA .Jr. Innate immune recognition and control of adaptive immune responses. Semin Immunol1998; 10:351-353. 12. Matzinger P. The danger model: a renewed sense of sel£ Science 2002 ; 296:301305 .

CHAPTER 2

Structure and Synthesis ofLipid A Shoichi Kusumoto,* Masahito Hashimoto and Kazuyoshi Kawahara

Introduction

L

ipid A is the lipophilic partial structure of bacterial lipopolysaccharide (LPS), which is a characteristic and essential component of the cell surface architecture of Gram negative bacteria. LPS constitutes the outer leaflet of the lipid bilayer of outer membrane which covers the outermost surface ofbacterial cells. Structurally, LPS is composed ofcovalently bound three distinct parts, i.e., O -antigenic polysaccharide, core oligosaccharide and glycolipid called lipid A (Fig. 1). Westphal and Liideritz found that the linkage between the core oligosaccharide and glycolipid is selectively cleaved by mild acid hydrolysis of LPS. They also observed that the liberated glycolipid, which they named lipid A, is responsible for the endotoxic activity ofLPS. 1 The discovery oflipid A was a real epoch in the history ofthe human's efforts towards the iden tification ofthe endotoxic principle which causes serious clinical problems during Gram-negative Infections.' Even after that, however, further progress ofstructural study oflipid A was not straightforward because of the amphiphilic nature and intrinsic heterogeneity of this group of complex glycolipids. Lipid A has a strong tendency to aggregate, which inhibited isolation ofhomogeneous single molecular species to be subjected to the precise structural studies. The early structural studies were, therefore, carried out with heterogeneous mixtures ofcongeners ofien containing other cell components and even artifacts formed by partial degradations during isolation procedures. Nevertheless, the correct basic structure of enterobacterial lipid A was elucidated as follows.v' These lipid A share the common hydrophilic backbone consisting ofa ~(1-6)-linked disaccharide of2-amino-2-deoxy-D-glucose (D -glucosamine, GleN) 1,4'-bisphosphate which is acylated at two amino and several hydroxy groups . The major facey acids linked to the backbone are hydroxylated at their 3- (or ~- )positions in R-configurations, while certain proportion ofnonhydroxylated normal fatty acids are always present. Because of its strong aggregation, however, the correct molecular weight oflipid A was not available. So that the possibility was not excluded of a macromolecular structure formed by phosphodiesters linking the backbone disaccharide units. Further advance was made during 1980s when the major components of lipid A from Escherichia coli and Salmonella Re mutant were isolated as single molecular species atter chemical modifications and the complete structures of parent lipid A were elucidared.P" Chemically synthesized lipid A preparation according to the proposed structure showed endotoxic activity indistinguishable from that of the natural counterpart isolated from bacteria."!' This was the final evidence for the concept that lipid A is the active principle ofbacterial endotoxin. The sto ry ofendotoxin research including all the important historical progresses is summarized in a review

article.'

In the next section ofthis Chapter are first described the briefoutline ofthe historical structural studies on enterobacterial lipid A and the general structural feature oflipid A drawn as the result by subsequent analytical works . The third section summarizes the structural variations of lipid *Corresponding Author: Shoichi Kusumoto-Suntory Institute for Bioorganic Research, Wakayamadai 1-1-1 , Shimamoto-cho, Mishima-gun, Osaka 618-8503, Japan. Email: [email protected]

LipidA in Cancer Therapy, edited by jean-Francois jeannin, ©2009 Landes Bioscience and Springer Science+Business Media.

6

Lipid A in Cancer Therapy

n

1L-.

O-Polysaccharide

...J1

1..'

--'

Outer core Inner core Core oligosaccharide

Lipid A

Figure 1. Schematic structure of lipopolysaccharide (LPS).

A within and beyond these general structural features : through recent investigations of lipid A from various sources, many further novel structures have been found. The structural information accumulated is expected to help explain the endotoxic and other biological functions oflipid A on molecular basis. The fourth section deals with chemical synthesis oflipid A which played a decisive role in confirming the endotoxic activity oflipid A and is expected to contribute to elucidation of molecular mechanism ofthe action oflipid A.

General Architecture ofLipid A ChemicalStructure ofLipidA from Escherichia coli and Salmonella

On the basis of fundamental researches carried out for many years,2.12 the complete structure of lipid A was elucidated by independent two research groups during 1980s. They both extensively purified major components of lipid A and investigated their structures by chemical and spectroscopic analyses; the latter includes emerging new methodologies such as soft ion ization mass spectrometry (MS) and high field nuclear magnetic resonance spectroscopy (NMR) with superconductive magnets.'! Shortly before these works, the possibility ofpolymeric structures of lipid A was excluded by 31P_NMR analysis ofLPS from E. coli heptose-less mutant." The structure oflipid A from anE. coli Re mutant 1 was determined as shown in Figure 2.15-17 This bacteria was selected because it was known to produce less heterogeneous lipid A than the one ofSalmonella minnesota R595 most frequently studied until that time . The major component of E. coli lipid A was isolated by preparative thin-layer chromatography (TLC) after removal of the labile glycosyl phosphate and subsequent methyl esterification of the remaining phosphate. The previously proposed hydrophilic backbone structure, ~(1-6) GleN disaccharide 1,4'-bisphosphare, was reconfirmed by detailed NMR analysis. This lipid A contains total six fatty acids, l.e., four 3-hydroxytetradecanoic [14:0(3-0H)] and each one ofdodecanoic and tetradecanoic acids (12:0 and 14:0). With the aid of high field two dimensional NMR spectroscopic analysis, the positions of acylation on the backbone were determined: the location of the so-called primary acyl groups, which directly bind to the backbone, was unequivocally fixed to be the two amino and two hydroxy groups at the 2-, 2'- and 3-, 3'-positions ofthe disaccharide, respectively," This necessarily led to the conclusion that the structure 2 (Fig. 2) ofa disaccharide biosynthetic precursor to lipid A, which was isolated earlier and known to contain only four 3-hydroxytetradecanoic acids as its acyl components." With regard to the mature lipid A from E. coli, the location ofthe two additional acyl groups (12:0 and 14:0) had yet to be fixed. These are called secondary acyl groups which bind to the 3-hydroxy groups of the primary acyl groups to form 3-acyloxyacyl moieties . Their linkage positions were deduced by chemical and MS analyses to lead to the total structure ofE. coli lipid A as 1,16.17which was immediately confirmed by total chemical synthesis as described in the fourth section. Qureshi and Takayama investigated the structure oflipid A from Salmonella thyphimurium by chemical as well as MS and NMR analyses after extensive separation ofcomponents by TLC and

Structure and Synthesis ofLipid A

7

~ ~~O

(HOhP-OO~~ o

o

HO

14

14

1

0

0

0

HN O-P(OH)

0

HO

14

14

12

14

HO

0

HO

14

14

HN

2

HO

14 14

2

HO~ o

o

HO

0 HO

0 HN O-P(OHh 0

0 0

14 14

3

14

14

12

14

4 16

figure 2. Structure of Escherichia coli lip id A (1), disaccharide and monosaccharide biosynthetic precursor (2 and 3) and Salmonella minnesota lipid A (4). The dotted lines in 4 show nonstoichiometric presence of the substituents.

high performance liquid chromatography (HPLC) on reverse phase columns.P:" Among at least five components detected, they concentrated on the major one and elucidated its structure as 1 which was identical with that of E. coli described above. Shortly before reaching this conclusion, the same authors determined the structure 3 (Fig. 2) ofan acylated GleN l-phosphate which was isolated from a certain mutant of E. coli and named lipid X.21.22 Lipid X was assumed to be an early biosynthetic precursor to lipid A and LPS and , if this is the case, the structure of the disaccharide biosynthetic precursor must have the structure 2. This assumption was later proved true by elucidation of the biosynthetic scheme of lipid A by Raetz et al.23 The structures 4 (Fig. 2) of one of the lipid A components from S. minnesota was also elucidate ."

General Structure ofLipidA

After the chemical structures oftypical enterobacterial lipid A were elucidated as described in above the subsection, lipid A components from a large number of Gram-negative bacteria have also been investigated and their structures precisely determined. From the results ofthese works it can be concluded that the basic architecture is well conserved in most ofthe lipid A components investigated. They contain a I3(I-4)-linked disaccharide of GleN. In some very limited cases,

8

Lipid A in Cancer Therapy

one or both of GleN residues is replacedwith 2,3-diamino-2,3-dideoxy-D-glucose (GleN3N). Substitutions with other sugar moieties were also reported. They have been so far regarded as seldomexceptions, so that in this chapter their structuresaresummarized in a separatesubsection as unusuallipid A. In viewof recent findings, however, wemayencounter more examples of such "unusuallipid A" through investigation of wide varieties of bacteria,particularly, of those living in unusualenvironments. The two hydroxygroups of the GleN disaccharide, Le., one at the glycosidic position of the reducing terminal GleN [abbreviated as GleN(I)] and the other at the 4-position of the distal GleN [GleN(II)], are usually phosphorylatedto form the so-called hydrophilicbackboneofl ipid A (Fig.3). The latter phosphate is often calledthe 4'-phosphateaccordingto the numbering rule ofthe positions in the disaccharide. Thedisaccharide backboneisacylated withlongchainfattyacids to completethe basicstructure of lipid A. As in the caseof E. coliand S. minesota lipid A (1 and 3), the locationsof the primary acylgroups,which are (R)- 3-hydroxy fatty acylgroupsdirectlybound to the backbone, are at the two amino and two hydroxygroups at the 2-, 2'- and 3-, 3'-positions, respectively, of the disaccharide.Ofthese four,however, not all the positions are always acylatedin lipid A isolatedfrom variousbacteria.The 3-hydroxy groups of the primary acylgroups are often, though not always, esterified by secondaryacylgroupsto form 3-acyloxyacyl structureswhosepresenceis one of the typicalfeatures of bacteriallipid A. Thus,the total number,distributionsand the chainlengthsof acylgroupsgeneratequite widevariations of lipid A structural analogs. Though the term "lipidA" is routinelyusedwithout problems, it should be noted here that the term has,in a strict sense, two different meanings: one denotesthe glycolipidic partialstructureas presentin LPS,whiletheother represents the glycolipid liberatedbymildacidhydrolysis ofLPS.The latter is referred to as "free lipidA"when a strict expression is requiredto avoidpossible confusion. Therefore, lipid A (freelipid A in this case) neverexists in nature asa componentof bacterial cells but isan artificial molecule obtainedonlythroughhydrolytic cleavage ofLPS.The6-hydroxy group

c

Hydrophilic backbone

A : Position for hydrogen or primary acyl group 8 : Position for hydrogen or polar substituent C : Linkage position of core saccharide in LPS Figure 3. The general structure of lipid A.

Structure andSynthesis ofLipidA

9

of GlcN(II), i.e., the 6'-position of the disaccharide, always remains unsubstituted in free lipid A, because this particular hydroxy group is the linkage position ofthe core and outer saccharide chains in LPS via a kerosidic bond of the proximal3-deoxy-D-manno-octulosonic acid (Kdo) residue of the saccharide. One point should be, however, noted here that this traditional understanding may have to be slightly modified. The concept that usually two Kdo, in some limited case at least one," Kdo residues linked to lipid A are inevitable for survival ofbacceria" may not apply to all bacterial species. Recent investigations showed that some bacteria such as Y.pestis live without Kdo. 26-29 This could mean that "free lipid A:' does actually exist on the cell surface ofliving bacteria. Lipid A from some bacteria lacks one ofthe two phosphates: in most cases the 4'-phospates are lacking. The phosphate moieties of lipid A hence represent other source of structural variations. The occurrence of such dephospho lipid A and occasional polar substitutions on either one or both ofthe phosphate moieties create other diversities in the lipid A structures. In the next section are described the representative structures oflipid A so far elucidated with references to the types ofvariations. More comprehensive survey ofthe chemical structures oflipid A is available in the form of outstanding reviews written by Zahringer et al.6•13 Additional novel structures oflipid A reported after the publication of these reviews are also cited in the next section. Structural variations ofendotoxin, particularly those in lipid A and the inner core region of LPS, are reviewed by Trent et al from the standpoint ofbiosynthesis." These authors also discuss the relationship of the chemical structures and pathogenesis ofthe source bacteria.

Structural Variations ofLipid A Variation ofAcyl Groups: Locations, Typesand Chain Lengths

Lipid A contains ester- and amide -linked fatty acids as its major constituents. Among them saturated 3-hydroxy fatty acids of(R)-configuration are the standard and inevitable building blocks oflipid A. Saturated nonhydroxylated fatty acids are also present in various molar ratios depending on the source oflipid A. Maximum four ofthe former hydroxy acids are directly bound to the backbone disaccharide as the primary acyl groups , whereas the latter nonhydroxylated acids are linked to any of the 3-hydroxy groups ofthe former . Lipid A actually isolated from bacterial cells contains from four to seven total acyl moieties, though theoretically eight , i.e., each four of primary and secondary acyl groups, could find their positions per one disaccharide. Typical examples of lipid A structures with variou s numbers of acyl groups are summarized in Table 1, where the location and chain lengths of acyl groups as well as the presence /absence of phosphates (see below) are also indicated. More comprehensive information is available in the review articles cited above.6.13 The heptaacylated lipid A 4 (Fig. 2 and Table 1) of S. minnesota represents the one with the highest acylation among lipid A so far described. The secondary hexadecanoyl group at the 2-position is known to be nonstoichiometric, so that lipid A ofthi s bacteria contains variable proportion of 4 and the hexaacyl congener 1. Another example of similarly heptaacylated lipid A was also reported (Table 1) .31.32 Hexaacyllipid A are quite frequently observed in nature. They are structurally divided in two groups according to the distribution patterns oftheir acyl moieties on the two GleN residues. One has an asymmetrical acylation pattern of the (4 + 2)-type , where GleN (II) has four while GleN(!) has two acyl groups . E. coli lipid A 1 is a typical example belonging to thi s group. The other group has a symmetrical or even (3 + 3)-type acylation pattern as in the lipid A from Chromobacterium uiolacerum (Table 1). Most of naturally occuring pentaacylated lipid A are of the (3 + 2)-type and all tetraacyllipid A are of the (2 + 2)-type. Besides the diversity in the numbers ofacyl groups and their distributions, the carbon numbers ofindividual fatty acids linked to the particular positions can differ depending on the source bacteria. Such large heterogeneity in acyl groups alone represents one ofthe origins ofthe enormous diversity seen in lipid A structures, a part ofwhich are listed in Table 1.

Erwinia carorovora Proteus mirabilis Salmonella minessota (4 + 2)-type hexaacyl Campylobacter jejunf Escherichia coli HaemophiJus influenzae Leptospira interrogens" Plesiomonas shigelloides Salmonellatyphimunum Yersinia petls" (3 + 3)-type hexaacyl Chromobacterium violaceum Comamonas testeroni Neisseria meningitidis Pseudomonas aeruginosa Pseudomonas reactans Xanthomonas campestris (3 + 2)-type pentaacyl Bacteroides fragilis Burkholderia cepacis

(4 + 3)-type heptaacyl

Bacteria

17:0[3-(15:0)]i 16:0 [3-0(14:0)]

16:0[3-0H] 14:0[3-0H]

16:0[3-0H] 16:0[3-0H]

12:0[3-0H]i

15:0[3-0H] 14:0[3-0H]

12:0[3-0(12:0)] 10:0[3-0(14:0)] 14:0[3-0(12:0)] 12:0[3-0(12:0)C] 12:0[3-0(12:0)C] 12:0[3-0H]i

10:0[3-0H] 10:0[3-0H] 12:0[3-0H] 10:0[3-0H] 10:0[3-0H] 10:0[3-0(10:0)]i

12:0[3-0(12:0)] 10:0[3-0(12:0)C] 14:0[3-0(12:0)] 12:0[3-0(12:0)C]

10:0[3-0H] 10:0[3-0H] 12:0[3-0H] 10:0[3-0H] 10:0[3-0H] 10:0[3-0(10:0)]i 12:0[3-0(12:0Y]

14:0[3-0H] 14:0[3-0H] 14:0[3-0H]C 16:0[3-0H] 14:0[3-0H] 14:0[3-0H] 14:0[3-0H]

14:0[3-0H] 14:0[3-0H] 14:0[3-0H] 12:0[3-0H] 14:0[3-0H] 14:0[3-0H] 14:0[3-0H]

14:0[3-0(16:0)] 14:0[3-0(12:0)C] 14:0[3-0(14:0)] 16:0[3-0(14:1 )e] 14:0[3-0(16:1 )8] 14:0[3-0(12:0)] 14:0[3-0(16:1 )g]

14:0[3-0(16:0)] 14:0[3-0(14:0)] 14:0[3-0(14:0)] 12:0[3-0(12:1)e] 12:0[3-0(12:0)] 14:0[3-0(14:0)] 14:0[3-0(12:0)]

14:0[3-0(16:0)] 14:0[3-0(16:0)b] 14:0 [3-O(16:0)b]

2-N

14:0[3-0H] 14:0[3-0H] 14:0[3-0H]

3-0

14:0[3-0(12:0)] 14:0[3-0(14:0)] 14:0[3-0(12:0)]

2'-N

14:0[3-0(12:0)] 14:0[3-0(14:0)] 14:0[3-0(14:0)C]

3'-0

Carbon Numbers and Types of Acyl or Acyloxyacyl Groups Linked to the Positions GlcN (I) GlcN(II)

P

H p

p p

p

p

p p

p p

p

P H H

p

P P

P P

P

P P

P

P

P P

pf

P

P

P

P

P

109 47

104 105 54 106 107 108

46 19,20 49,50

71 16,17,20 25 73

31 32 24

References

continued on next page

1-0

GlcN(I)

p p

p

4'-0

GlcN (II)

Presence of Phosphates"

Table 1. Selected examples of structural variations of lipid A from various bacteria: distribution of acyl groups and phosphates on the disaccharides

~

~

~

"'t

~

S· ~ ;:s

~

~

~ ~.

c;:::)

......

3'-0

1O:0[3 -0H] 14:0[3-0H] 14:0[3-0H] 16:0[ 3-0H] 16:0[3-0H)]

10:0[3-0(10:0)Q) 14:0[3-0H] 14:0[3 -0(12:0)] 18:0[3 -0(14:0)] 18:0[ 3-0(18:0)]

14:0[3-0H)] 14:0[3-0H)] 16:0[3-0H)] 18:0[3-0H)]

H

10:0[3-0(10:0Y.Q]

H H H H

P

(l-GaIUN P

P

Hm

H (l-GaIUN P

.4'-0

17:0[3-0H)]i 16:0[3-0H)] 14:0[3-0H) 17:0[3-0H)1 18:0[3-0H) 14:0[3-0xo) 14:0[3-0xo) 18:0[3-0H]

2-N

GIcN(I)

72 75 47 35,36 38,76 39,41,43 43 77

P ~-GaIUA'

18 48 111 64-66

P P P

H

110

P

P P H P P P

References

1-0

GIcN(1I)

Presence of Phosphates'

groups are partiall y 2-hydroxylat ed. dEither on e or both of the backbone GleN are repl aced with GIeN3N . eThe secondary acy l groups (ci s-b.5-12 :1 and ci s-b.7- 14:1) may be interconverted. fTh e phosph ate is meth yl esterifie d. gThe secondary acyl group is cis-b.9-16:1. hat 25·C. Another (2 + 2)-type tetraacyl lipid A is produ ced at 37 ·C (see text). iThe primary acyl groups have non stoichiometric methyl substitutions at co-l and w-2 pos itions. jAcyl groups with odd carbon numbers are of iso-types . ka-D -galactopyranosiduronic acid . I~-d-galactopyranosiduronic acid . mThe corresponding l,4-bisphosphate was also detected . nThis long chain fatty acid is hydroxylated at the 27 pos itions. oThe secondary acyl group is cis-b.7-14:1. pThe secondary acyl group is cis-b.5-12 :1. qA pr art of th e secondary acyl is 12:0.

' P denotes th e position is phosphorylated, wh ile H means th e hydroxy group is free. bThe secondary acyl groups are nonstoichiometric. 'Th e secondary acyl

15:0[3-0H)1 14:0[3 -0H) 12:0[3-0H) 16:0[3-0H) 14:0[3-0H) 1O:0[3-0H) 1O:0[3-0H) 14:0[3-0H)

3-0

17:0[3 -0(15:0)]i 16:0[3-0H] 14:0[3-0(12:0)] 17:0[3-(16:0)]1 14:0[3-0(28:0)"] 14:0[3-0(14:1)0] 14:0[3-0xo] 18:0[3 -0(28:0)"]

2'-N

Carbon Numbers and Types of Acyl or Acyloxyacyl Groups linked to the Positions GIcN(II) GleN (I)

16:0[3-0H) Flavobacterium meningosepticuttt' 14:0[3-0(18:0)] Leptospira interrogens" 12:0[3-0H) Neisseria gonorrhoeae Porphyromonas (Bacteroides) gingiva/is 15:0[3 -0H)i 14:0[3-0H) Rhizobium etli 10:0[ 3-0H) Rhodobacter sphaeroides 14:0[3-0 (12:1 )P) Rhodobacter capsu/atus 14:0[3-0H) Shinorhizhobium sp. (2 + 3)-type pentaacyl Marinomona s vaga (2 + 2)-type tetraacyl Escherichia coli mutant 14:0[3-0H] Escherichia coli mutant --Francisse/a tu/arensis --Helicobacter pylori ---

Bacteria

Table 1. Continued

......

......

::...

it

r:-.; .....

~

~ ::; .

So

~ ;:s

~ l:. ;:s I:>..

5::' i1 l}

12

Lipid A in Cancer 1herapy

Majority of fatty acids present in lipid A are around the range of C IO to C 14, but longer and shorter chains are also present, though quite seldom. Some bacteria like Rhizobium etliare reported to produce lipid A which contains an extraordinary long carbon chain of C 28.33,34 Almost all fatty acid components oflipid A are ofsaturated structures with normal, i.e., nonbranched chains ofeven carbon numbers. Some lipid A, however, contain branched fatty acids with odd number carbons. Most ofthe branched acids are ofso-called iso-type structures which have terminal isopropyl structures having methyl branches at the (00-1) position ofthe main chains as seen in lipid A fromPorphyromonas (Bacteroides)gingivalis 5.35.36Some bacteria grown at ambient temperature produce lipid A with unsaturated acyl groups (Table 1). Several other bacteria, whose lipid A usually contain only saturated acids, were shown to incorporate unsaturated fatty acids when grown at low temperatures.5.37.38 This seems to be a rational and could thus be general, strategy ofbacteria, since the presence ofunsaturated acids in lipid A molecule is expected to increase the fluidity of the outer membrane, which certainly favors the growth at lower temperatures. Lipid A 6 from Rhodobacter (Rhodopseudomonas)sphaeroides39' 41 and Rhodobacter capslatus42.43 have a common unique structural feature having 3-keto acids as their N-bound acyl moieties in place ofordinary 3-hydroxy acids in other lipid A.In addition, these lipid A contain unsaturated acids as their components. These unique structures may be responsible for their strong antagonistic activity to suppress the endotoxic action ofLPS simultaneously given to experimental animals or cultured cells. The presence of2-hydroxy fatty acidswas also observed since very earlystageoflipid A research." These acids occur as one of the secondary acyl groups of lipid A usually in nonstoichiometric molar ratio. In other word, some of the secondary acyl groups are incompletely substituted with the corresponding S-configurated 2-hydroxy acids of the same chain length. Such incomplete substitution is another source of the heterogeneity typical of bacterial LPS. According to recent investigation on the biosynthesis ofLPS, 2-hydroxy acids are formed by the action ofa dioxygenase which directly oxidizes certain proportions of the specific carbon atom in an already completed lipid A structure with atmospheric oxygen.t' There are many bacteria, even a single strain ofwhich produces a mixture ofdifferently acylated lipid A.37.45In such cases, too, the structure ofindividual molecular species of lipid A is usually constant with regard to the chain lengths and location of the individual acyl groups, which are rather strictly controlled by the specificity ofthe acyl transferase participating in the introduction of each acyl groUp.23.30 The heterogeneity in the degree of acylation is the result of incomplete action of some of the transferases. Yet, in some lipid A were observed acyl groups with different chain lengths at the same positions.46•47 The total number and location of acyl groups in lipid A influence the biological activity as typically represented by the completely different activity between the hexaacyllipid A of E. coli 1 and the tetraacyl disaccharide precursor 2 mentioned in the section following: the former is fully endotoxic, whereas the latter suppress the action of the former as an antagonist. Another hexaacyllipid A recently isolated from E. coli FS1S was also found to act as an antagonist," A quite unique and interesting strategic ability of Yersinia pestis was unveiled in relation to its virulence to cause plaque. This bacteria contains hexaacylated lipid A 7 of (4 + 2)-type like E. coli when grown at around 2S"C, whereas it switches to produce (2 + 2)-type tetraacyllipid A 8 at the mammalian body temperature of3TC.49.50 The latter does not activate the mammalian innate immune system, likewise the tetraacyl precursor 2 does not, so that Y. pestis readily evades the defense system and is allowed to grow rapidly to cause serious damage to the host. The direct relation ofthe structural change oflipid A to the virulence ofthe bacteria was recently proved by molecular genetic experiments ." Biological activities ofvarious structural variations oflipid A were sununarized in a reviewardele." Seydel et al discussed the endotoxic and antagonistic activities oflipid A analogs in relation to the acylation patterns and molecular conformations determined by physical methods."

Structure and Synthe sis ofL ipidA

13

O~OH II

(HOl2P-O0

0

OJ HO

~

HN HO 0 0 00

o

0

OJ HO

0

HN 0 -P(OH)2

0

10 10 14 14

16

14

5

6

Figure 4. Structures of Porphyromonas gingiva/is lipid A (5) and Rhodobacter sphaeroides lipid A (6).

Modifications ofthe Phosphate Moiety

The two phosphate moieties of lipid A represent other points of structural modifications (Fig. 3) . The presence ofso-called polar substiruents on the phosphates has been often observed in many bacterial species. Typical ones are amino alcohols such as 2-aminoethanol (Em) and 4-amino-4-deoxy-L-arabinose (4-AraN), which are linked to one or both ofthe two phosphates of lipid A formingphosphodiester structures. Lipid A of Y. pestis (Figs. 5,7 and 8)50 mentioned above have 4-AraN substituents on both phosphates. In Table 2 some further selected examples of such polar subsriruents are listed. The amino alcohols, particularly 2-aminoethanol, are sometimes not directly linked to the phosphate of'lipid A but via an additional phosphate forming a diester of a pyrophosphate as observed in lipid A 9 from Neisseria meningitidis. 54 These modifications ofphosphates are usually not stoichiometric: the respective phosphate moieties are only partially modified. The possibility ofsuch substitutions on phosphates provides with an additional source of heterogeneity in lipid A of bacterial origins. In earlier structural investigations, much efforts have been paid to determine the position of substituents. According to recent investigations, however, the degree and position of the polar substiruenrs can vary depending even on culture conditions.P'"

o __\~n o-~-oo~o

0

NH,

~

o-~-o~o

14

7

NH,

14

8

Figure 5. Structures of two Yersinis pestis lipidA produced at differenttemperatures: hexaacyl component 7 at 25'C and tetraacyl component 8 at 37'C. The dotted lines indicate that the polar substituents are nonstoichiometric.

Lipid A in Cancer Therapy

14

Table 2. Selected examples of polar substituents on the backbone phosphate moieties oflipidN Substituents on the Phosphates at Bacteria

4'-Position

l-Position

Reference

Campylobacter jejunlb Escherichia coli Escherichia coli mutant Helicobacter pylori Neisseria meningitidis Porphyromonas gingivalis Proteus mirabilis Pseudomonas aerginosa Salmonella minessota Salmonella typhimurium Yershinia pestis

Etn-P

-- c

Etn P P-Etn Etn PEtn Etn

Ara4N Ara4N Ara4N Etn-P Ara4N

Ara4N P-Etn Ara4N Ara4N

71 63 14 64,65 54 36 32 57 112 113 50

Etn-P -- c

EtnP

' Po lar substituents are usually not stoichiometric. "The backbone disaccharide is a mixture of GIeN3N-GlcN, GIeN3N-GlcN3N and GleN-GleN.
Unusual LipidA

As mentioned already, the basic structural feature of lipid A is well conserved. Most of Gram-negative bacteria producelipidAcomposedofN,O-polyacylatedand I,4-0-bisphosphorylated ~(I-6) disaccharide of GleN. Even in very limited cases, however, there have been found lipid A analogs to which this structural principle does not apply.

Structure and Synthesis ofLipidA

H2N

o 0 """'vO-~---O-~-O OH

k

OH 0

o

15

OH

HO~:':O

OH

0

o

0

HNHO

0

HO~~

0

o

0

o

o

HO

.. 9

0 NH HN" /'V' 2 O-P-O OH

1.

"

2

0 HO

f.

10

Figure 6. Structures of Neisseria meningitidis lipid A (9) and Helieobaeter pylori lipid A (10). The dotted lines indicate that the polar substituents are not always present.

One of such examples is substitution of the GleN residues, where one or both of GleN are substituted with a closely related diamino sugar, GleN3N. The hydrophilic backbone of lipid A of Campylobacterjejuni is composed of a mixture ofbisphosphates of three ~(1-6) disaccharides, GleN3N-GleN, GlcN3N-GleN3N and GlcN3N-GleN3N in an approximate molar ratio of 6:1.2:1. The complete structure 11 ofthe hexaacylated major component is shown in Figure 7 as a typical example of this type of lipid A.71 Lipid A from Flavobacterium meningosepticum has a structure ofpentaacylated backbone , which is a mixture of GleN-GleN and GleN3N-GleN (in a ratio of 1:0.35) l-monophosphates." Through genetic and biosynthetlc study on a lipid A-synthesizing spirochete, Leptospira interrogans, the biosynthetic pathway of GlcN3N-type lipid A became understood. 2-Acetamido -2-deoxy-D-glucose (N-acetylglucosamine, GleNAc) is enzymatically converted into 2-acetamido-3-amino-2,3-dideoxy-D-glucose (GlcNAc3N), which then comes into a route similarto the normal biosynthesis ofGleN-type lipid A in E. coli,73 Since both GleN and GleN3N share the same stereochemistry at alltheir chiral centers, substitution ofthe backbone disaccharide

HO Co,H

HO~~OH

o .....s0~o

(HO);-~N~~ o

o

0

0

HNHO

0

o

0 HO

0

0 0

0

n

HN O-P(OH

HO

o o HN

0

0

0

o HO

HN

~ ~HOOH

HO HN 0 HO

°HO

0 HO

HN O

0

C0 2H

14

,. ,.

14 14

11

18

,.

14

12

,.

Figure 7. Structures of GlcN3N-containing lipid A: a major component of Campy/obaeter

jejuni lipid A (11) and that from hyperthermophilic Aquiflexpyrophi/us (12).

Lipid A in Cancer Therapy

16

with the latter is assumed not to bring a big change in the overall molecular shape oflipid A. The complete structure of L. interrogans lipid A was elucidated by combination of NMR and MS analysis to be hexaacylated GleN3N disaccharide I-monophosphaee (Table 1). It shares the same distribution ofsixacylgroups as that ofE. coli though the chain lengths ofacylgroups are different, where both ofthe two secondary acyl groups are monounsaturated (cis-I4:1 and -12:1). Another unique feature of this lipid A is that the I -O-phosphate is a methyl ester substituent which have never been observed in lipid A from other sourses." Another unique lipid A with novel backbone structure was isolated from obligate predatory bacteria, Bdellovibrio bacteriovorus74 The hydrophilic backbone composed of~( 1-6) disaccharide ofGleN3N lacks both phosphates. Instead, D-mannopyranosyl residues are a-glycosidically linked to the 1- and 4'-positions of the disaccharide. Though their linkage positions in the (4 + 2)-type distribution have not been completely assigned, the nature of the unique fatty acid components were determined. The all but one primary and secondary acids are iso-type branched ones. The backbone oflipid A 12 (Fig. 7) from hyperthermophilic bacteria Aquiflexpyrophilus is also composed of GleN3N disaccharide and rwo sugar units , though galacturonic acids (GalVA) in this case, in place ofconventional phosphares." The chemical stability ofthis lipid A as compared to ones with conventional structures may account for, at least in part, the ability ofthis bacteria to live at temperature as high as 95· C. After some confusion, the quite novel structures were determined of lipid A from Gram-negative Rhizobium and related bacteria which fix nitrogen during simbiosis in the root of leguminous plants.38.76.77 Lipid A components of R. etli were separated and their structures elucidated by modern high resolution NMR and MS analyses. The lipid A of this bacterium are structurally divided in two groups . The one group is composed ofan acylated normal GleN disaccharide which, however, lacks both phosphates as represented by the structure 13 in Figure 8. The l-phosphare is lacking, while an a-glycosidically linked D-galacturonic acid replaces the 4'-phosphate in the conventional backbone. In the other group oflipid A, the GleN (I) residue

14

14

14

13

"y"',r,0

28

HO 0

Figure 8. Structures of two types of lipid A (13 and 14) from endosymbiotic Rhysobium etli.

Structure and Synthesis ofLipidA

17

in the former is converted to a 2-amino-2-deoxy-gluconic acid as seen in structure 14. Another characteristic feature of these lipid A is that they contains an extremely long-chain C 28 fatty acid as a sole secondary acyl group. The possible relation of these unique structural features of lipid A is discussed to the ability ofthe particular bacteria to facilitate symbiosis by evading the innate immune system of the host plants."

Chemical Synthesis ofLipid A

Systematic synthetic studies toward lipid A had already started78.79 before its correct precise structure was drawn out as described above. On the basis of the knowledge accumulated during the precedingworks,78.80chemical synthesis ofthe newly proposed definite structures was straightforward. Immediately after the exact location ofthe primary acyl groups on the disaccharide backbone had been determined, chemical synthesis ofthe disaccharide precursor (named precursor Ia or lipid IVa) 2 was successfully achieved. 81.82The first chemically synthesized preparation of the lipid A analog 2, which was absolutely free from any contaminants originating from bacterial cell components, showed expected endotoxic activity both in vivo and in vitro tests.7-9Thisgave the first definite evidence to prove that the glycolipid is the active principle ofbacterial endotoxin. The first chemical synthesis of the structure 1 corresponding to mature E. coli lipid A followed soon after the above synthesis of2.16.83 One ofthe synthetic intermediate to 1 was thereby converted to the dimethyl ester of a disaccharide 4'-phosphate having all six acyl groups at the correct locations. The synthetic compound was identified with the natural counterpart derived from bacterial lipid A; both showed ind istinguishable NMR spectra." This was the first direct comparison ofsynth etic and natural lipid A derivatives. The synthetic hexaacylated bisphosphare 1 showed full endotoxic activity as the active principle of endoroxin.'?" In Scheme 1 is illustrated an improved efficient synthetic route to 1.85 Synthetic homogeneous preparations of 1 and 2 are now commercially available as standard compounds for various biological studies .

Bno~ o1 , a a ocos

BnO

It

'~-o

a

+

ssorr

__ 84% ""

OH

I) Zn- Cu

-

BnO

0

3CO,H AcOH 2) R DCC -

C(~ ~-w~~~,~

2) R'C02H DCC, DMAP

97%

-&

I) Ir complex

2) 120 H,o

0

a

0 OCOR

cf

a

OCOR2

NHCOR2

NHTroc

a

a

OCOR

0,.

OH

a

Bno~o~ 0 a 1 0 a acOR acOR2

C(=

0

~~-O

BnO NHCOR3

OAllyl NHCOR2

I) (IMS}zNLi

0 0

Bn0E(o~2)(BnO}z~O'hOBn}z Bn0Er:0~

92%C('0a ocoa' "1p-o I "" cf -&

0

NHT~"" ~~~ C(j-~.kt:

CG I ""

HO

,CCI3

BnO NHCOR3

ocos

H

NHCOR2

75% .. a a OCO~ C("I I p-o cf -&

Troc : 2,2,2-trichloroethoxycarbonyl

Scheme 1. An improved synthetic route to E.coli lipid A (1).

BnO NHCOR3

H . Pd

OCO~ a~

I

O-P(OBn}z NHCOR2

1

18

Lipid A in Cancer Therapy

Quite unexpected fact was then observed in the experiments with the synthetic specimen that the tetraacyl derivative 2 acts as an antagonist to suppress the endotoxicity of lipid A and LPS when tested in human cell systems." Availability ofpure synthetic preparations ofboth endotoxic and antagonistic activities is being utilized in recent investigations on the molecular mechanism ofrecognition by the receptors.87.88 Natural lipid A from certain bacterial species were also found to exhibit very low or negligible endotoxic activity or even show antagonistic activity. A typical example is lipid A from of Rhodobacter (Rhodopseudomonas) sphaeroides described above.39.89-91 Like that ofR. sphaeroides described above, there are also several natural lipid A which contain double bonds in their acyl groups. The above synthetic routes to 1 and 2 can not be applied to such unsaturated targets, because the hydrogenolytic deprotection at the final step would lead to saturation ofthe double bonds in the substrates. A modified route was reported for the synthesis of R. sphaeroides lipid A by employing allyl type protecting groups which can be removed by transition metal catalyzed reactions," the use of catalytic hydrogenolysis being avoided. The outline of the synthesis ofa structure 6 proposed for R. sphaeroides lipid A is illustrated in Scheme 2. In view of the strong antagnositic activity of R. spbaeroides lipid A and related compounds to suppress the endotoxic activities of bacterial LPS, various structural analogs were designed and synthesized," Among them, one ofartificial analogs is expected to be resistant against biological degradation in human body and suitable for clinical applications." Until today synthesis has become theoretically possible of any structural analogs oflipid A .95 A number ofstructural analogs of lipid A , including both natural and unnatural ones, were synthesized and their biological activities investigated in relation to their structures. 96-99 Even more complex structure of E. coli Re LPS containing Kdo disaccharide as a part ofthe core saccharide was already chemically synthesized and lipid A with a hydrophilic substituent on the phosphate (see above) is now also accessible.loo.101 Chemical syntheses have also been reported oftritium-labeled

HO~

AlIocOE>, 011 OCOR 0'C,CCI3 + (Allyl012P-O

N3

NH

1) 5n(1I)reagent 2) R3C02H, DCC 56%

AllceO o I'

o

OCOR'

(AlIyI012P-O

AllocO

OTBS NHCOR2

AllocO

°a 0

OCORt

AllocO NHCOR3

OCOR'

AllocO AgOTf ----

OCORI

Po

(AllylO);zP-O

o~ o

OCOR'

AllocO NHCOR3

o~o

OTaS NHCOR2

AllocO

OTBS NHCOR2

l)HF 2) (AlIyIOhPN(Pt)2 3)mCPBA

80%

0

II

O-P(OAllyl12 NHCOR2

o CH3(CH2)SCH=CH(CHvsAO

R4co :

~

),o..J.........C.....

CH3(CH 2

Alice : a1lyloxycarbonyl

OH~ CH3(CHV8.AvC .....

TBS : t-butyldirnethylsilyl

Scheme 2. Synthetic route to the proposed structure of Rhodobacter sphaeroides lipid A (6) w ith an unsaturated acyl group.

Structure and Synthesis ofLipid A

19

lipid A derivatives with high specific radioactivities as well as a fluorescence- and biotin-labeled derivative.102.103

Conclusion

As shown in this article, isolation and structure elucidation have been achieved with lipid A components from various sources. Recent great advances in the isolation techniques and in MS or MS/MS and NMR spectroscopy enabled the facile analysis ofthe target molecules with regard to backbone structures, compositions and locations of fatty acyl components as well as the presence/location ofpolar substituents, Analysis can be achieved with small amounts ofcompounds. Individual components in a mixture can sometimes be analyzed by virtue of effective separation ofions in MS techniques. In this way many novel lipid A structures are being unveiled from various bacterial origins. In parallel to the structural and synthetic studies of lipid A, our understanding on the biosynthesis ofLPS, in particular oflipid A and the inner core region, was greatly improved during the last two decades. Almost all genes and enzymes participating in the biosynthesis have thereby been characrerized." The structural variations observed in lipid A from various bacteria and even their influence on the pathogenesis can be discussed from biosynrhetic and thus genetic standpoints.'? Manipulation ofbiosynthetic pathways by molecular genetic techniques has now become possible. On the basis ofthese rapidly growing knowledge on the molecular structure oflipid A and its recognition by host systems, the communication between bacteria and hosts via LPS will be discussed in more details in terms ofmolecular interactions oflipid A and receptor proteins. Recent understanding of the innate immune defense system of higher animals and its utilization will be described in depth in other chapters ofthis book. From an inverse point ofview, the sophisticated strategies ofbacteria may provide with interest issues, too: they modify the chemical structures of their vital lipid A components to evade the defense mechanism ofhost or to survive under harsh or extraordinary living conditions. Discussions have already appeared on the possible relation of the chemical structure oflipid A and symbiotic/parasitic behavior of the bacteria. Further contribution ofchemistry, in collaboration with biochemistry and molecular biology, is expected to help our deeper understanding ofthe conversation between hosts and microbes on molecular basis.

References

1. Wesrphal 0, Liideritz O. Chernische Erforschung von Lipopolysacchariden Gram-negariver Bacterien. Angew Chern 1954; 66 :407-401. 2. Rierschel ET. Westphal O. Endotoxin: Historical perspectives. In : Endotoxin in health and disease. Brade H , Opal SM. Vogd SN, Morrison DC. eds, New York: Marcd Dekker, 1999:1-31. 3. Gmeiner L, Liideritz O. Wesrphal O. Biochemical studies on lipopolysaccharide of Salmonella R mutant 6. Investigation on the structure of the lipid A component. Eur J Biochern 1969; 7:270-379. 4. Hase S. Rierschel ET. Isolation and analysis of the lipid A backbone. Lipid A structure of lipopolysaccharide from various bacterial group. Eur J Biochem 1976; 63:101-107. 5. Takayama K. Qureshi N. Chemical structure of lipid A. In: Morrison DC. Ryan JL, eds. Bacterial endoroxic lipopolysaccharides Vol: I Molecular biochemi stry and cellular biology. Boca Raron : CRC Press. 1992:43-65. 6. Ziihringer U, Lindner B. Rietschel ET. Molecular structure of lipid A. the endoroxic center of bacterial lipopolysaccharides. Adv Carbohydr Chern Biochem 1994; 50:211-276. 7. Galanos C. Lehman V, Liideritz L. Endoroxic properties of chemically synthesized lipid A part structures-comparison of synthetic lipid A precursor and synthetic analogues with biosynrhet ic lipid A precursor and free lipid A. Eur J Biochem 1984; 140:221-227 . 8. Korani S, Takada H . Tsujimoto M et al. Immunobiologically active lipid A analogs synthesized according to a revised structural modd of natural lipid A. Infect Immun 1984; 45:293·296. 9. Takada H, Korani S, Tsujimoto M et al. Immunopharmacological activities of a synthetic counterpart of a biosyntheric lipid A precursor molecule and of its analogs. Infect Immun 1985; 48:219-227 . 10. Galanos C. Liiderirz O. Rietschel ET et al. Synthetic and natural Escherichia coli free lipid A express identical endotoxic activities. Eur J Biochem 1985; 148:1-5.

20

Lipid A in Cancer Therapy

11. Kotani S. Takada H. Tsujimoto M et a1. Synthetic lipid A with endotoxic and related biological activities comparable to those of a natu ral lipid A from an Escherichia coli Re-rnutant. Infect Immun 1985: 49:225-237. 12. Galanos C. Liideritz O. Rietschel ET er al. Newer aspect of the chemistry and biology of bacterial lipopolysaccharides with special reference to their lipid A component. In: Goodwin T. ed. International Review of Biochemistry Vol 14: Biochemistry of lipid II. Baltimore : University Park Press. 1977:239-335. 13. Zahringer U. Lindner B. Rierschel ET. Chemical structure of lipid A: Recent advances in structural analysisof biologicallyactive molecule. In: Endotoxin in health and disease. Brade H. Opal SM. VogelSN, Morrison DC. eds, New York: Marcel Dekker, 1999:1-31. 14. Rosner MR , Khorana HG. Satterhwait AC. The structure of the lipopolysaccharide from a heptose-less mutant Escherichia coli K-12. J Bioi Chern 1979: 254:5918-5925 . 15. Imoto M. Kusumoto S, Shiba T ec al. Chemical structure ofE. coli lipid A: Linkage site of acyl groups in the disaccharide backbone. Tetrahedron Lett 1983: 24:4017-4020. 16. Imoto M. Yoshimura H . Kusumoto S et al. Total synthesis of lipid A, active principle of bacterial endotoxin. Proc Japan Acad Ser B 1984: 60:285-288. 17. Imoto M, Kusumoto S. Shiba T et al. Chemical structure of Escherichia coli lipid A. Tetrahed ron Lett 1985: 26:907-908. 18. Lehman V. Isolation. purification and properties of an intermediate in 3-deoxy-D-mannooctulosonic acid-lipid A biosynthesis. Eur J Biochern 1977; 75:257-266. 19. Qureshi N, Takayama K. Heller D et a1. Position of ester groups in the lipid A backbone of lipopolysaccharides obtained from Salmonella typhimurium. J Bioi Chem 1983; 258:12947-12951. 20. Takayama K. Qureshi N. Mascagni P. Complete structure of lipid A obtained from Salmonella typhimurium. J Bioi Chern 1983: 258:12801-12803 . 21. Nishijima M. Raetz CRH. Characterization of two membrane-associated glycolipids from an Escherichia coli mutant deficient in phosphatidylglycerol. J Bioi Chern 1981: 256:10690-10896. 22. Takayama K, Qureshi N. Mascagni P et al. Fatty acyl derivatives of glucosamine l-phosphate in Escherichia coli and their relation to lipid A. J Bioi Chern 1983; 258:7379-7385. 23. Raetz CR. Whitfield C. Lipopolysaccharide endoroxins, Annu Rev Biochem 2002 ; 71:635-700. 24. Qureshi N. Mascagni P. Ribi E et al. Monophosphoryl lipid A obtained from lipopolysaccharides of Salmonella minnesota R595. J Bioi Chern 1985; 260:5271-5278 . 25. Helander 1M. Lindner B. Brade H et al. Chemical stru cture of the lipopolysaccharide of Haemophylus inlluenzae strain 1-69 RD -/b+. Eur J Biochem 1988: 177:483-492. 26. Tyeng ZL. Datta A. Kolli VK et al. Endotoxin of Neisseria meningitid is composed only of intact lipid A: Inactivation of the meningococcal 2-deoxy-D-manno-octulosonic acid transferase. J Barcriol 2002; 2379-2388. 27. Mohan S. Raetz CRH. Endotoxin biosynthesis in Pseudomonas aeruginosa: Enzymatic incorporation of laurate before 3-deoxy-D-manno-octulosonate. J BacterioI1994:6944-6951. 28. Tan J, Darby C. Yersinia pestis is viable with endotoxin composed of only lipid A. J Bacreriol 2005: 187:6599-6600 . 29. Tzeng YL. Datta A, Kolli VK et al. Endotoxin of Neisseria meningitid is composed only of lipid A: inact ivation of the menigococcal 3-deoxy-D-manno-octuloson ic acid transferase. J Bacreriol 2002 ; 184:2379-2388 . 30. Trent MS. Stead CM. Tran AX ct al. Diversity of endotoxin and its impact on pathogenesis.J Endotoxin Res 2006 ; 12:205-223. 31. Fukuoka S. Kamishima H . Nagawa Y et al. Structural characterization of lipid A component of Erwinia carorovora lipopolysaccharide. Arch Microbiol 1992; 157:311-318. 32. Sidorczyk Z. Zahringer U. Rierschel ET. Chemical structure of the lipid A component in the lipopolysaccharide from Proteus mirabilis Re-mutant . Eur J Biochem 1983; 137:15-22. 33. Hollingsworth RI. Carlson RW. 27-Hydroxyoctacosanoic acid is major structural fatty acyl component of the lipopolysaccharide of Rhizobium trifolii ANU 843. J Bioi Chern 1989; 264:9300-9303. 34. Bhat UR. Forsberg LS. Carlson RW. Structure of lipid A component of Rhizobium leguminosarum bv, Phaseoli lipopolysaccharide. J Bioi Chem 1994; 269:14402-1441 o. 35. Ogawa T. Chemical structure of lipid A from Porphyromonas (Bacteroides) gingivalislipopolysaccharide. FEBS Letters 1993: 332:197-201. 36. Kumada H , Haish ima Y, Umemoto T et al. Structural study on the free lipid A isolated from lipopolysaccharide of Porphyromonas gingivalis. J Bacteriol1995; 177:2098-2106. 37. Seydel U. Lindner B. Wollenweber HW et al. Structural studies on the lipid A component of enterobacteriallipopolysaccharides by laser desorption mass spectrometry. Eur J Biochem 1984; 145:505-509. 38. Que NLS. Lin S. Cotter RJ et al. Purification and mass spectrometry of six lipid A species from the bacterial endosymbionr Rhysobium erli. J Bioi Chern 2000; 275:28006-28016.

Structure and Synthesis ofLipid A

21

39. Salimath PV, WeckesserJ. Strittmarer W et al. Structural studies on the nontoxic lipid A from Rhodopseudomonas sphaeroides ATCC 17023. Eur J Biochem 1983; 136:195-200. 40. Qureshi N. Honovich JP. Hara H et al. Location of fatty acids in lipid A obtained from lipopol ysaccharide of Rhodoseudomonas sphaeroides ATCC 17023. J BioI Chern 1988; 263:5502-5504. 41. Qureshi N. Takayama K. Meyers KC et al. Chemical reduction of3-oxo and unsaturated groups in fatty acids of diphosphoryllipid A from lipopolysaccharide of Rhodoseudomonas sphaeroides. J BioI chem 1991; 266:6532-6538 . 42. Krauss JH, Seydel U. Weckesser J et al. Structural analysis of the nontoxic lipid A of Rhodobacrer capsulatus 37b4. Eur J Biochem 1989; 180:519-526. 43. Meyer H, Merkofer T. Warth C et al. Position and configuration of double bonds inlipid A-associated monounsaturated fatty acids ofProteobaceteria and Rhodobacter casulatus 37b4. J Endotoxin Res 1996; 3:345-352. 44. Gibbons HS. Lin S, Comer R] et al. Oxygen requirement for the biosynthesis of the S-2-hydroxymyristate moiety in Salmonella typhimurium lipid A. J BioI Chem 2000; 275:32940 -32949. 45. Balzer LH. Mamby-Balzer I. Heterogeneity of lipid A: Structure determination by IJC and 31 p NMR of lipid A fraction from lipopolysaccharide of Escherichia coli 0111. Biochemistry 1986; 25:3570-3575. 46. Lukasiewicz J. Dzieciatkowska M. Niedziela T et al. Complete lipopolysaccharide of Plesiomonas shigelloides 074:H5 (strain CNCTC 144/92). 2. Lipid A. its structural variability. the linkage to the core oligosaccharide and the biological activity of the lipopolysaccharide. Biochemistry 2006 ; 45: 10434-10447. 47. Takayama K. Qureshi N. Hyver K et al. Characterization of a structural series of lipid a obtained from the lipopolysaccharide of Neisseria gonorrhoeae . J Bioi Chern 1986; 261:10624-10631. 48. Zahringer U. Salvetzki R. Wanger F et al. Structure and biological characteristics of a novel tetra -acyl lipid A from Escherichia coli F515 lipopolysaccharide acting endotoxin antagonist in human monocytes. J Endotoxin Res 2001; 7:133-146 . 49. Kawahara K. Tsukano H. Watanabe H et al. Modification of the structure and activity of lipid A in Yersinia pestis lipopolysaccharide by growth temperature . Infect Immun 2002 ; 70:4092-4098 . 50. Knirel YA. Dentovskaya SV. Senchenkova SN et al. Structural feature and structural variability of the lipopolysaccharide of Yersinia pestis. the cause of plaque. J Endotoxin Res 2006; 12:3-9. 51. Montminy SW; Khan N. McGrath S et al. Virulence factor of Yersinia pestis are overcome by a strong lipopolysaccharide response. Nature Immunology 2006 ; 7:1066-1073. 52. Takada H. Kotani S. Structure-function relationships of lipid A. In : Morrison DC. RyanJL. eds. Bacterial endotoxic lipopolysaccharides Vol: I Molecular biochemistry and cellular biology. Boca Raton : CRC Press. 1992:43-65. 53. Seydel U. Weise A. Schrom AB et al. A Biophysical view on the function and activity of endotoxin. In: Endotoxin in health and disease. Brade H , Opal SM. Vogel SN. Morrison DC. eds. New York: Marcel Dekker. 1999:195-219. 54. Kulshin VA. Zahringer U. Lindner B et al. Structural characterizat ion of the lipid A component of pathogenic Neisseria meningitidis. J Bacteriol 1992; 174:1793-1800. 55. Zhou Z. Ribeiro AA. Raetz CHR. H igh-resolution NMR spectroscopy of lipid A molecules containing 4-amino-4-deoxy-L-arabinose and pho sphoethanolamine susbstirucnts. J Bioi Chem 2000 ; 275:13542-13551. 56. Vaara M. Vaara T. Jensen A et al. Characterization of the lipopolysacchride from the polymyxin-resistant pmrA mutants of Salmonella typhimurium. FEBS letters 1981; 129:145-149. 57. Bhat R. Marx A, Galanos C et al. Structural studies of lipid A from Psudomonas aeruginosa PA01 : Occuence of 4-anino -4-deoxyarabinose. J Baceriol 1990; 172:6631-6636. 58. Qureshi N. Kaltashov I. Walker K et al. Structure of the monophosphoryllipid A moiety obtained from the lipopolysaccharide of Chlamydia trachomaris, J Bioi Chem 1997; 272:10594-10600. 59. Trent MS. Riberio AA. Doerrler WT et al. Accumulation of a pol yisoprene-linked amino sugar in polymyxin -resistant Salmonella typhimurium and Esche richia coli. J Bioi Chem 200 I; 276:43132-43144. 60. Lee H , Hsu FE Turk ] et al. The PmrA-regulated pmrC gene mediates phospho ethanolamine modification of lipid A and plymixin resistance in Salmonella enterica, J Bacteriol 2004 ; 186:4122-4133. 61. Tran AX. Lester ME. Stead CM ct al. Resistance to the ant imicrobial peptide polymixin requires myristoylation of Escherichia coli and Salmonella t yphimurium lipid A. J Bioi Chem 2005; 280:28286-28194. 62. Helander 1M. Kilpelainen 1, Vaara M. Increased substitution of phosphate groups in lipopolysaccharide of the polymixin-resistanr pmrA mutants of Salmonella typhimurium: a 3I P_NMR study. Mol Microbiol 1994:11:481-487.

22

Lipid A in Cancer Therapy

63. Nummila K, Kilpelainen I. Zahringer U ec al. Lipopolysaccharides of polymyxin Bvresitanr mutants of Escherichia coli are extensively substituted by 2-aminoethanol pyrophosphate and contain aminoarabinose in lipid A. Molecular Microbiol1995; 16:271-278. 64. Moran AP. Lindner B. Walsh E]. Structural characterization of the lipid A component of Helicobacter pylori rough- and smooth-form lipopolysaccharides, J Bacteriol1997; 179:6453-7463. 65. Suda Y, Ogawa T. Kashihara W ec al. Chemical structure of lipid A from Helicobacter pylori Strain 206-1 lipopolysaccharide. J Biochem 1997; 121:1129-1133 . 66. Suda Y, Kim YM. Ogawa T et al. Chemical structure and biological activity of a lipid A component from Helicobacrer pylori main 206. J Endotoxin Res 2001 ; 7:95-104. 67. Phillips N. Apicella MA. Griffis M et al. Structural characterization of the cell surface lipopolysaccharides from a Nontypable strain of Haemophilus influenzae. Blochemstry 1992; 31:4515-4526. 68. Karbarz MJ. Kalb SR, Cotter RL er al. Expression cloning and biochemical characterization of a Rhizobium leguminosarum lipid A l -phosphatase. J Bioi Chern 2003; 278:39269-39279. 69. Wang X. McGrath SC. Cotter RL et al. Expression cloning and periplasmic orientation of the Francisella novicida lipid A 4'-phosphatase LpxF. J Bioi Chern 2006; 281:9321-9330. 70. Holst O. Borowiak D, WeckesserJ et al. Structural studies on the phosphate-free lipid A of Rhodomicrobium vannielii ATCC 17100. Eur J Biochern 1983; 137:325-332. 71. Moran AP, Zahringer U. Seydel U et al. Structural analysis of the lipid A component of Campylobacrer jejuni CCUG 10936 (serotype 0 :2) lipopolysaccharide. Eur J Biochern 1991; 198:459-469. 72. Kaco H , Haishima Y,Iida T et al. Chemical structure of lipid A from Flavobacterium meningosepricum lipopolysaccharide. J Bacteriol1998; 180:3891-3899 . 73. Que-Gewirrh NLS, Ribeiro AA. Kalb SR et al. A methylated phosphate group and four amide-linked acyl chains in Leptospira interrogans lipid A. J Bioi Chern 2004; 279:25420-25429. 74. Schuwdke D, Linsheid M. Strauch E ec al. The obligate predatory Bddlovbibro bacteriovorus posseses a neutral lipid a containing u-Dvmannoses that replace phosphate residues. ]. Bioi Chern 2003 ; 278:27502-27512. 75. Plotz BM. Lindner B. Stetter KO er al. Characterization of A novel lipid A containing D-galacturonic acid that replaces phosphate residues. J Bioi Chern 2000; 275: 11222-11228. 76. Que NLS. Ribeiro AA. Raetz CRH. Two dimensional NMR spectroscopy and structures of six lipid A species from Rhysobium etli CE3 . J Bioi Chern 2000; 275:28017-28027. 77. Gudlavalleti SK. Forsberg LS. Structural characterization of the lipid a component of Shinorizobium sp. NGR234 rough and smooth form lipopplysaccharide. J Bioi Chern 2003; 278:3957-3968. 78. Kusumoto S. Inage M. Chaki H er al. Chemical synthesisof lipid A for the elucidanon ofstructure-activity relashionship, In: Bacterial lipopolysaccharide. Anderson L. Unger FM. eds, ACS Symposium Series 231. Washington DC : American Chemical Society. 1983:237-254. 79. Kiso M. Hasegawa M. Synthetic studies on the lipid A component of bacterial lipopolysaccharide. In: Bacterial lipopolysaccharide. Anderson L. Unger FM. cds. ACS Symposium Series 231. Washington DC: American Chemical Society. 1983:277-300. 80. Inage M. Chaki H , Imoto M er al. Synthetic approach to lipid A: Preparation of phosphorylated disaccharides containing (R)-3-hydroxyacyl and (R)-3-acyloxyacyl groups. Tetrahedron Lett 1983; 24: 2011-2014. 81. Imoto M. Yoshimura H . Yamamoto M cr al. Chemical synthesis of phosphorylated tetraacyl disaccharide corresponding to a biosynrhetic precursor of lipid A. Tetrahedron Lett 1984; 25:2667-2670. 82. Irnoto M. Yoshimura H. Yamamoto M et al. Chemical synthesis of a biosynrhetic precursor of lipid A with a phosphorylated reeraacyl disaccharide structure. Bull Chern Soc Jpn 1987; 60:2197-2204. 83. Imoto M. Yoshimura H , Shimamoto T et al. Total synthesis of Escherichia coli lipid A. the endotoxically active principle of cell-surface lipopolysaccharide. Bull Chern Soc Jpn 1987; 60:2205-2214 . 84. Kusumoto S, Yoshimura H. Imoto M er al. Chemical synthesis of l-dephospho derivative of Escherichia coli lipid A. Tetrahedron Lett 1985; 26:909-912 . 85. Liu WC. Oikawa M. Fukase K et al. A divergent synthesis of lipid A and its chemically stable unnatural analogues. Bull Chern Soc Jpn 1999; 72:1377-1385 . 86. Loppnow H, Brade L. Brade H et al. Induction of human inrerleukin 1 by bacterial and synthetic lipid A. Eur J Immuno11986; 16:1263-1267. 87. Kobayashi M. Saitoh S. Tanimura N ct al. Regulatory roles for MD-2 and TLR-4 in ligand-induced receptor clustering. J Immunol 2006; 176:6211-6218. 88. Saito S. Miyake K. Mechanism regulating cell surface expression and activation of Toll-like receptor 4. Chern Recd 2006; 6:311-319. 89. Takayama K. Qureshi N. Beutler B et al. Diphosphoryl lipid A from Rhodopseudomonas sphaeroides ATCC 1702 blocks induction of cachectin in macrophage by lipopolysaccharide. Infec Immun 1989; 57:1336-1338.

Structure and Synthesis ofLipidA

23

90. Kirkland TN, Qureshi N. Takayama K. Diphposphoryllipid A derivative from lipopolysaccharide (LPS) of Rhodopseudomonas sphaeroides inhibits activation of 70Z/3 cells by LPS. Infect Immun 1991; 59:131-136. 91. Golenbock D. Hampton RY, Qureshi N ct al. Lipid A-like molecules that antagonize the effects of endotoxins on human monocytes. ] BioI Chern 1991; 266 :19490-19498. 92. Christ Wj, McGuiness PD, Asano 0 et al. Total synthesis of the proposed structu re of Rhodobacter sphaeroides lipid A resulting in the synthesis of new potent lipopolysaccharide anragonisrs.] Am Chern Soc 1994; 116:3637-36 38. 93. Rossignol DP, Christ W], Hawkins LD ec al. Synthetic endotoxin antagonist. In: Endotoxin in health and disease. Brade H. Opal SM. Vogel SN, Morrison DC. eds, New York: Marcel Dekker. 1999:699-717 . 94. Christ W]. Asano 0 , Robidoux AL et al. E5531. a pure endotoxin antagonist of high potenc y. Science 1995; 268 :80-83. 95. Kusumoto S. Fukase K. Oikawa M. The chemical synthesis of lipid A. In : Endotoxin in health and disease. Brade H. Opal SM, Vogel SN. Morrison DC. eds, New York: Marcel Dekker. 1999:243-256. 96. Liu WC. Oikawa M. Fukase K er al. A divergent synthesis oflipid A and its chemically stable unnatural analogues. Bull Chern Soc jpn 1999; 72:1377 -1385. 97. Fukase K, O ikawa M. Suda Y er al. New synthesis and conformational analysis of lipid A: biological activity and supramolecular assembly.] Endotoxin Res 1999; 5:46-51. 98. Sakai Y, Oikawa M, Yoshizaki H er al. Synthesis of Helicobacrer pylori lipid A and its analogue using p-(trifluoromethyl)benzyl protecting group. Tetrahedron Lett 2000; 41 :6843-6847. 99. Fukase K. Ueno A. Fukase Y er al. Synthesis and biological activities of lipid A analogs possessing ~-glycos idic linkage at l-pos ition . Bull Chern Soc Jpn 2003 ; 76 :485-500. 100. Yoshizaki H . Fukuda N, Sato K et al. First total synt hesis of the Re-type lipopolysaccharide. Angew Chern Int Ed 2001; 40:1475-1480 . 101. Sakai Y. Oikawa M. Yoshizaki H et al. Synthesis of Helicobacter pylori lipid A and its analogue using p-Itrifluoromerhyljbenzyl protecting group . Tetrahedron Lett 2000; 41 :6843-6847. 102. Fukase K, Kirikae T. Kirikae F er al. Synthesis of [3H]-labeled bioactive lipid A analogs and their use for detection of lipid A-binding proteins on murine macrophages . Bull Chern Soc ]pn 200 I; 74:2189-2197. 103. Fujimoto Y, Kimura E, Murata S et al. Synthesis and bioactivity of Huorescence- and biotin-l abeled lipid A analogues for investigation of recognition mechanism in innate immunity. Tetrahedron Lett 2006; 47:539-543 . 104. Wollenweber HW; Seydel U, Lindner Bet al. Nature and location of (R)-3-acyloxyacyl groups in lipid A of lipopolysaccharides from various gram-negative bacteria. Eur] Biochern 1984; 135:265-272. 105. lida T. Haishima Y, Tanaka A er al. Chemical structure of lipid A isolated from Comamonas testeron i lipopolysaccharide. Eur] Biochem 1996; 237 :468-475 . 106. Kulshin VA, Zahringcr U, Lindner B et al. Structural characte rization of the lipid A component of Pseudomonas aeruginosa wild-type and rough mutant lipopolysaccharide. Eur J Biochern 1991; 198:697-704 . 107. Silipo A, Lanzetta R. Garozzo D er aI. Structural determination of lipid A of the lipopolysaccharide from Pseudomonas reacrans, A pathogen of cultivated mushrooms . Eur] Biochem 2002 ; 269 :2498-2505 . 108. Silipo A. Molinaro A. Sturiale L et al. The elicitation of plant innate immunity by lipopolysaccharide of Xanthomonas campestris.] Bioi Chern 2005; 280 :33660 -33668 . 109. Weintraub A, Zahringer U. Wollenweber HW ct al. Structural characterization of the lipid A component of Bacteroides fragilis strain NCTC9343 lipopolysaccharide. Eur J Biochem 1989; 183:425-431. 110. Krasikova IN . Kapustina NY, Isakov VV et al. Detailed structure of lipid A isolated from lipopolysaccharide from marine proteobacterium Marinomonas vaga ATCC 27119T• Eur J Biochern 2004 ; 271 :2895-2904. Ill. Vinogradov E, Perry MB. Conlan ]W. Structural analysis of Francissela tularensis lipopol ysaccharide. Eur J Biochern 2002; 269 :6112-6118. 112. Batley N, Packer NH. Redmond JW; Configurations of glycosidic phosphates of lipopolysaccharide from Salmonella minnesota R595. Biochemistry 1982; 21:6580-6586. 113. Strain SM Armitage 1M. Anderson L et al. Location of polar substiruenrs and fatt y acyl chains on lipid A precursors from a 3-deoxy-D-manno-octulosonic acid-deficient mutant of Salmonella typhimurium. Studies by 'H. 13C and 31p nuclear magnetic resonance.} BioI Chern 1985; 260:16089-16098 .

CHAPTER 3

Conformation and Supramolecular Structure ofLipid A Klaus Brandenburg and Ulrich Seydel"

Abstract

I

n recent years, lipid A as 'endotoxic principle' of bacterial lipopolysaccharide (LPS) and derivatives thereofhave become increasinglyimportant in the fieldofbiomedical application such as for vaccination or as therapeutical, e.g., anti-tumor agent. For an understanding of these biological processes, however,a basic physicochemicalcharacterization oflipid A and lipid A-like structures is a necessaryprerequisite. This includes the determination of parameters like criticalmicellarconcentration, the type of aggregatestructure, the molecularconformation, the gel to liquid crystalline phase behaviour, and theoretical approacheslike molecular modelling. In this chapter, data from literature are summarizedshowing that the unusual chemicalstructure of lipid A-type moleculesis connected with a very complexstructural polymorphism, which issensitively dependent on the particular chemical primary structure, particular on the acylation pattern and on the number ofphosphate groups at the diglucosamine backbone.

Introduction

Early attempts to correlate biological activities of LPS to specificchemical structures of the macromoleculeshaveprovided evidencethat no changesin endotoxin activitywasprovokedwhen LPS wasmodified chemicallywithout altering the lipid A moiety.' From these observationsit was concluded that the lipid A part represents the 'endotoxic princlple;' and that the polysaccharide portion of lipopolysaccharides should function merely as 'a solubilizing carrier' for the biologicallyactivepart. Thus, the variabilityofthe core oligosaccharide, and evenmore of the a-antigen, between different bacterial strains and genera should be indicative of the main role of the sugar part as epitop e for antibod y binding and as barrier for hydrophobic drugs, thus contributing to the overallpermeability barrier ofthe bacterial organism.' Alterations ofthe chemicalstructure of the lipid A moiety werefound to influencethe bioactivitydramatically.t There are lipid A variants, which exhibit-despite of an identical backbone to that ofenterobacterial lipid A-significantly reduced or evenno biologicalactivities.This fact isconnected to variationsin the acylchain moiety, mainly to a decreaseofthe number of fatty acid residuesand/or to their shortening (as examples for such structures see Fig. lA). The change in the primary chemical structure may even lead to antagonistic molecules,antagonizing biologicallyactive LPS or lipid A.5.6 For a better understanding ofthe followingand to avoidconfusion, a fewdefinitions should be put at the beginning: The term 'structure' has a double meaning, the chemical or primary structure of a molecular speciesand the supramolecular or tertiary structure ofaggregates or assemblies of a large number of identical molecules. The latter definition is the most relevant in this chapter. The term 'conformation' stands for the secondary structure of an individual molecule, i.e., its *Correspond ing Author : Ulrich Seydel-Forschungszentrum Borstel, Leibniz-Zentrum fUr Med izin und Biowissenschaften, LG Biophysik, D-23845 Borstel, Germany. Email: useydelefz-borstel

LipidA in Cancer Therapy, edited byjean-FrancoisJeannin. ©2009 Landes Bioscience and Springer Science+BusinessMedia.

26

Lipid A in Cancer Therapy

A

tbon_.

N_.....

LIpIcIAS-.

I

m

Hoa_,t

U

-,t

n

0

1.

1.

1.

u

1.

1.

1.-

It

1.

1.

1.

It

12

II1lodoap/rtlJum fuIrwn

1.-

1.

~-gala-

1.

12

1.

14

~

U

14

14

1.

~coll

~-~ c.""",....jaJunI

-

ClV'IlIlJ'J tEC .......

B

..

riot. . . .

Y,O

I' I'

It'

,. ,. ,. ,. ,. Hap ,. ,.

1.

1.-

1.12

1.-

U

R'

..........

,.

Gai'·12:1

,.

,. ,. ,.

,.

R'

R'

1.:0

12,.

H

H

12,.

H

01ID

0lt0

H

11:0

1.-,.

H

12:1

12:1

12:0

11:0

H

. ..1.'.12:1

H

H

12:.

12,.

1"0

,.

GaIA

It'

~o 0 HO~

R"""O

NH R3....

R

2....

0

NH .... X R1.... 0

SUbstituents Sample

516 506 505 504 406 606 OM·174 E5564

R1

R2

R3

R4

X

Y

C....Q-C" C....QH C....QH C....QH C....()H C....()H

C••..QH C••..QH C....QH C••..QH C....QH H H C..

C....Q-C•• C••..Q-C.. C....Q-C.. C,•..Q-C.. C,•..QH C....QH C....Q-C..

C••..Q-C.. C....Q-C.. C....Q-C.. C....Q-C.. C,•..QH H H C,•..QH

P P

P P H P P P P P

C,..OH C,..OH

CiSd'·12:1

p

H p

P P

p

Figure 1. Chemical structures of various (A) natural lipid A variants, (B) synthetic lipid A analogues and part structures. three-dimensional geometry. The conformation of a given molecule defines the structure of the aggregate formed by a large number ofidentical molecules. Lipid A as arnphiphilic molecules form aggregates in aqueous environments above a critical concentration (critical micellar concentration, CMC, sometimes also termed critical aggregate concentration, CAe). First rough estimations of the CMC for the lipid A precursor lipid IVa (tetraacyl lipid A corresponding to synthetic compounds '406') by Hofer et aF gave a value

Conformation and Supramolecular Structure ofLipid A

27

of < 10-7 M . Takayama et al8 have determined the CMC of deep rough mutant LPS (LPS Re), which carries in addition to the lipid A moiety only two negatively charged monosaccharides (Kdo) . They applied equilibrium dialysis and arrived at values of the solubility of 3.3 .10- 8 M at 22°C and 2.8 .10- 8 M at 37°C. Own still unpublished experiments utilizing a dialysis technique and radio-labelled lipid A compounds yielded values ofabout 10-7 M in pure water and of 10- 8 M in NaCV thus roughly confirming the data ofTakayama et al.In a paper published in 1998. Aurell and Wistrom 1o claimed to have determined the 'critical aggregation concentration' oflipid A and LPS. However. their results for CACs, lying in the range of4 to 10 fLM for different endotoxins, reflect the sensitivity limit of the applied measuring techniques, fluorescence spectroscopy with N-phenyl naphryl amine and 90°-light scattering, rather than provide reliable data for the CAC or CMC. Using fluorescence spectroscopy (with a variety of dyes) and 90°-light scattering. we have found that for various diacyl phospholipids as well as endotoxins (lipid A and various rough mutant and wild type LPS from Salmonella minnesota with differently long sugar chains) the signals all disappear in the micromolar range. although the values ofthe CMC should vary by orders of magnitude for the different lipids under investigation (unpublished results). The structure of lipid A aggregates depends on the conformation of the aggregate-forming molecules. which is again determined by their primary chemical structure. and is influenced by ambient conditions like temperature, pH. water content, and concentration ofmono- and divalent cations. Additionally, the molecular conformation ofa given lipid A within a supramolecular aggregate will depend. at least in part, on the fluidity (inversely correlated to the state oforder) ofits acyl chains. which can assume two main phase states, the well-known gel (~) and liquid-crystalline (a) phases. At a characteristic phase transition temperature To, a reversible transition between these two phases takes place. To depends on the length and the degree ofsaturation ofthe acyl chains as well as on the conformation and the charge density ofand its distribution within the headgroup region . From the variability in the acylation patterns, a complex phase behaviour and structural polymorphism are to be expected for free lipid A and part structures thereof. A more detailed discussion ofthe phase behaviour oflipid A will follow in a later section . The structural polymorphism oflipid A is insofar ofhigh biological importance. as it was shown that for the activation ofhuman mononuclear cells lipid A aggregates and not monomers are the endotoxically active unit." :" However, there is also evidence that monomers may be active. for example for the activation ofendothelial cells." In this context, a comprehensive review ofthe variety of biological activities expressed by different synthetic lipids A analogues and part structures is given by Takada and Kotani,!"

Aggregate Structure and Molecular Conformation

The structure ofan aggregateformed by a large number ofidentical amphiphilic molecules above the CMC can be predicted on the basis ofa simple geometric model, which relates the aggregate structure to the ratio ofthe effective cross-sectional areas a, ofthe hydrophilic polar and ahofthe hydrophobic apolar regions . respectively. of the composing individual molecules. Israelachvili et all 5 have introduced a dimensionless shape parameter S, which is defined as S =v/ (aoA}J =ah/a, (v: volume per molecule ofthe hydrophobic moiety; 1,: length ofthe fully extended hydrophobic portion). From the absolute values ofthe shape parameter. which may be estimated from energy minimization calculations (see later section). a predicted structure of the supramolecular aggregates can be deduced. According to this simple model, lipids with Y2 < S S 1 should assume lamellar (L) structures. Lipid A with S > 1. which also have a prominent axis of the headgroup, should adopt inverted hexagonal H ll structures. These structures are characterized by water-filled. hexagonally arranged circular rods. which are lined with the hydrated lipid headgroups, and the remaining volume is filled by the fluid hydrocarbon chains. Stable intermediate structures between the lamellar structure with S s 1 and the inverted hexagonal structure H ll with S > 1 are cubic structures (Q).16The L-, H ll - , and Qvstrucrures have a 1-, 2-, or 3-dimensional geometry. respectively. The

28

Lipid A in Cancer Therapy

Qjseructures exist in various space groups, from which the bicontinous cubic structures'? are the most important for lipid A. For the determination of the long-range order (supramolecular structure) and that of the shore-range order (arrangement ofthe acyl chains) small-angle and wide-angle scattering, respectively,with X-rays or neutrons can be used. A complete phase diagram was established for lipid A from S. minnesota and Escherichia coli over a wide range ofwater content (20 to 9S%) and Mg2+-concentrations (molar ratio [lipid A] : [Mg2+] from 1:0 to 1:1) and dependence on temperature.P''? The phase diagram included the determination of the aggregate structures with synchrotron radiation X-ray scattering (SAXS) and that of the phase behaviour with Fourier-transform infrared (FTIR) spectroscopy. Briefly, the results can be summarized as follows: In pure lipid-water systems, free lipid A forms lamellar structures at water contents below appro60% and non-lamellar cubic structures at higher water concentrations already at T < T, (see later section). In other words, a lyotropic structural transition takes place around 60% water content. In the presence of divalent cations [e.g., Mg2+) at a molar ratio of [lipid A] : [cation] = 1, non-lamellar structures are suppressed below T, Concomitantly, '( is shifted to higher values. Once the acyl chain melting process begins, free lipid A assumes non-lamellar cubic structures over the whole range of water content. With the completion of chain melting, the cubic structures change into inverted hexagonal H n structures.P Details ofthe structural polymorphism at high water contents are shown in Figure 2. Measurements ofthe aggregate structure were extended to other lipid A samples, i.e.,enterobacteriallipid A in different salt forms, monophosphoryllipid A, and lipid A from nonenterobacterial sources like those ofRhodobacter capsulatus, Rhodopseudomonas uiridis, Rubrivivax (formerly Rhodocyclus) gelatinosus. Rhodospirillum fulvum. Campylobacter jejuni and Chromobacterium violaceum (chemical structures Fig. lA).21.22 The measurements were carried out exclusivelyunder near physiological conditions with the purpose of directly correlating the results to data from biological test systems. It was found that different nonenterobacrerial lipid A samples showed a variety of aggregate structures ranging from H n (lipid A from Ru.gelatinosus) over mixed cubic/ lamellar (monophosphoryllipid A from S. min. and lipid A from C. jejuni) to pure lamellar structures (lipid A from C. oiolaceum, Rb.capsulatus,Rp. oiridis and Rs.fulvum). Also some synthetic lipid A analogues and part structures were investigated. Aggregates formed by eM-derivatives of lipid A-one or two phosphate groups are substituted by carboxymethyl groups-exhibited a structural variability similar to that of the phosphate-containing compounds:" thus, tetraacyl lipid A compound 406 as well as Bis-CM-406 adopt a multilamellar structure, whereas hexaacyl lipid A S06 as well as Bis-CM-S06 adopt a cubic inverted structure at 40·C and a hexagonal H n structure at T > 60·C. Very interestingly. the triacyllipid A part structure OM-I?4 (structure Fig. 1B) with a normal bisphosphorylared disaccharide backbone exhibits a non-lamellar structure." This is. however, a direct micellar rather than an inverted micellar (HI) structure, connected with lower but not zero activity in the cytokine assay. which is in contrast to the tetraacyllipid A 406 exhibiting no agonistic activity (Fig. 3). Moreover. the latter is antagonistically effective by blocking the LPS-induced cytokine production when administered to the cells before LPS addirion." Also synthetic triacyl monophosphoryl lipid A part structures were investigated and showed a very specific dependence ofthe aggregate structure and the intramolecular conformation (see next section) on the binding sites to the sugar backbone." It will be shown later that these different structural preferences are ofimportance for the expression of biological activity. An overview ofaggregate structures and other physico-chemical parameters oflipid A ofdifferent origin and their relative biological activities is given in a previous review.26

Intramolecular Conformation

Byapplying FTIR spectroscopy using an attenuated total reflectance unit with polarized IR light. the orientation ofmolecular groups within lipid A could be determined." For enterobacterial lipid A. an inclination of the diglucosamine backbone (normal to the sugar plane) with respect to the

Conformation and SupramolecularStructure ofLipid A

HII

70

-.a ~ !

l!!

[

E

29

coexistence region

x

Q212

60

50

coexistence region

~

Q224

Q229

40 UQ229

Q230

30 70

75

80

UQ212

85

Water content I %

90

95

Figure 2. Phasediagramof lipidAfrom LPS ofSalmonellaminnesota R595 inthe water concentration range 70 to 95% at low Mg2+ concentrations ([lipid A] : [M g2+] » 1). Llamellar phase, Q212, Q224, Q229, Q230 cubic phases of space groups 212, 224, 229 and 230, respectively.":" HII inverted hexagonal phase. X = unassigned cubic phase. Reproduced from: Brandenburg Ket al. Chem Phys Li pids 91 :53-69; ©1998 with permission from Elsevier." direction ofthe hydrocarbon chains ofmore than 45° was observed, lower acylated lipid A carrying a reduced number ofacyl chains (penta and tetra, Fig. IA,B) as well as some nonenterobacterial lipid A exhibit a significantly smaller angle (<20 °), while monophosphoryllipid A (4'-phosphoryl lipid A corresponding to synthetic compound 504) has an intermediate inclination (36 to 38°) of the backbone. Accordingly, backbones oflipid A having a well expressed conical conformation are strongly inclined, those with a conical conformation that deviates only slightly from a cylindrical conformation, have a medium inclination, and those with a cylindrical conformation have-ifat all-only a small inclination. It should be emphasized that the two parameters, molecular conformation and inclination, are interdependent and apparently express the packing constraints of the hydrocarbon chains with respect to the available cross sectional areas, and determine, most importantly, also the ability of lipid A to act agonistically or antagonistically (see later section).

Phase States and Transitions between Them

As already mentioned above, lipid A as do other lipids exists in different phase states of the hydrocarbon chains, the gel (~) and the liquid-crystalline (~) states, between which a first order trans ition takes place at a phase transition temperature 1;;, which depends critically on the chemical structure ofthe given lipid A. The acyl chain melting is connected with a transition from an all-trans configuration into a less ordered phase.P The transition is accompanied by an increase in the cross-sectional area ofthe apolar region of the involved molecules (decrease of the effective length of the acyl chains) and may, therefore, be also connected with a change of the aggregate structure of the supramolecular assemblies.

30

Lipid A in Cancer 1herapy

-... Lipid A from E. coli Re-LPS

100 -A- Tetraacyl fipid A (406)

-

-~-OM-174

:€

- 10 ~ c

o

U :::s

ea.

1

W ~

0.1 1

10

100

1000

10000

Lipid A concentration (ng/ml) Figure 3. Stimulation of interleukin-6 production in human mononuclear cells after incubation w ith different concentrat ions of OM-174, lipid A from E. coli-LPS, or synthetic tetraacyllipid A '406'. Reproduced from: Brandenburg K et al. Eur J Biochem 2000 ; 267:3370-3377, with perm ission from Blackwell Publishing."

The systematic investigationof the phasetransition behaviourof freelipid A providedTc-values in the range40-43°C for freelipidA from E. coliand ofaround 45°C for that from S. minnesota.29.30 Other groups havearrivedat similarresults," Moreover, for synthetic lipid A of E. coli and S. minnesota (compounds 506 and 516, respectively, seeFig.lB) Tc-values in closeagreementto thoseof the natural compounds were found.32.33 Also variousnon-enterobacterial lipid A wereinvestigated.Thus, for hexa- and heptaacylated lipid A from Erwinia carotovera (structure see Fig. lA), which is chemicallyvery similar to lipid A from Salmonella, Tc-values were found to be similar to those of the latter. Usually, however, significantly lowerT.,-values than those ofEnterobacteriaceae were observed.P '" This observation correlateswith the fact that the former often contain shorter acyl chains with a higher degree of unsaturation. In an infrared spectroscopicstudy, alsothe Tc-values and the statesof order ofthe acyl chainsin the singlephasesofvarious lipidA analogues and part structureswerederermined.P A characteristic dependenceofTcon the number ofacylchainswasobserved.Thus, Tc issigni6cantlybelowO°C for the diacylanaloguecompound 606, between 15 and 20°C for the tetraacylcompound 406, which corresponds to the lipid A precursors Ia or IVa,and at 43°C for the hexaacyl compound 506. The addition ofdivalent cations aswell as a loweringofthe pH -value causea significantrigidification of the acyl chains of free lipid A preparations and lead to an increaseof the Tc-value. In a recent study,also the synthetic carboxymethylated lipid A analogues [phosphate groups are substituted bycarboxymethyl(CM) groups]wereinvestigated.23 Interestingly, for tetraacylcompound 406 the substitution of the phosphate groups causes an increaseofthe T.,-value from ",,20 to 45°C, while the valuesofTc are more comparable for synthetic compounds 506, CM-sa6, and Bis-CM-sa6, indicating a possibleinfluenceofthe aggregatestructure (seeabove). The lipid A analogue E5531, a pentaacyllipid A based on natural lipid A from R. capsulatus (structure in Fig. lA), blocksin analogyto its mother compound the cytokineinduction in human mononuclear cells,35but alsothe endotoxin response in humans with experimental endoroxernia-" Thiswasfound to be true similarly for the tetraacylatedcompound E5564 (structurein Fig. IB).37.38

Conformation and Supramolecular Structure ofLipid A

31

Asai et al have measured the influence ofcompound E5531 on phospholipid membrane properties with optical and calorimetric techniques .P'v They found that th is compound decreased the Tc-valueofDPPC concomitant with an increase in fluidity, while "the effect of complete lipid A was different from that ofE5531" without discussing this difference . The authors speculated that the biological effect ofantagonism is mainly a result ofa change in membrane fluidity ofthe phospholipids, probably by unspecific intercalation ofthe lipid A antagonist. However, it has been shown that membrane intercalation happens readily only in the presence of transport proteins such as the LPS-binding protein LBPY

Molecular Modelling

On the basis ofthe known primary chemical structure, the application ofmolecular modelling techniques has allowed to calculate the potential conformations accessible to a given molecule . In an early paper by Labischinski et al,43 results from theoretical calculations on the conformation of the heptaacyllipid A component of S. m innesota have been published. Most strikingly, these calculations suggested that the fatty acid chains are arranged in a hexagonal lattice and that the terminal methyl groups do not form a plane parallel to the membrane surface. The authors calculate a length of2.6 nm ofthe lipid A molecule and a cross-section of0.6 to 0.8 nm for the smaller and 1.2 to 1.6 nm for the longer side of the rectangular cross-section of the acyl chain moiety. In this model, the tilt angle ofthe diglucosamine backbone was determined to 53.7° with respect to the membrane surface with the reducing side ofthe diglucosamine, i.e., the l-phosphate emerged in the hydrophobic moiety of the neighbouring molecule and the 4'-phosphate sticking out into the aqueous phase. As pointed out above, we could , however, verify that the l-phosphate is surrounded by water and the 4'-phosphate is buried in the hydrophobic region , probably facing the 3-hydroxyl groups ofthe neighbouring molecules.27Although this packing ofthe acyl chains is not optimal, evidence for it was prov ided by the following observations: (I) several lipid A analogues and partial structures do not assume a hexagonal dense packing and (ii) lipid A exhibits a strongly reduced enthalpy change (~Hc) ofthe gel to liquid-crystalline (~-) phase transition as compared to saturated phospholipids with identical acyl chain lengths. From this, a comparison on the basis of the number ofmethylene groups indicates a reduced packing density ofthe former. In a molecular dynamics study, Frecer et al44 have investigated different lipid A (bis- or monophosphorylated compounds 506, 504, and 505 and unphosphorylated compound 503, Fig. IB) with respect to the conformation ofthe diglucosamine headgroup, order and packing ofthe acyl chains, solvation of phosphate groups , and other parameters and deduced from these results also possible mechanisms of action on target cells. The authors did not find significant geometrical differenc es in the molecular conformation, mutual position, and orientation of the GIcN rings, orientation ofthe phosphate groups, or compact alignment ofthe fatty acid chains . Also, the angle between the backbone axis ofthe diglucosamine headgroup and the vector aligned along the acyl chains was estimated. The authors found values of84°, 69°, 75° and 87° for compounds 503, 505 , 504 and 506 , respectively, which would correspond to an inclination angle as described above of 6°,21 °, 15° and 3°, respectively. Obviously, these values are in sharp contrast to those determined experimentally (at least for compounds 504 and 506, see above). Furthermore, the authors found an increase in th e tendency to form non-lamellar aggregates with decreasing number ofphosphate groups (l.e., compound 503 has the highest tendency to adopt non-lamellar structures and compound 506 the lowest). This is again the reversed sequence as found in experiments,22.27.33and can be explained by the fact that the authors did not consider the bridging function ofdivalent cations: The non-lamellar structure ofcompound 506 is essentially due to the fact that cation bridging of neighbouring phosphate groups leads to a smaller effective backbone cross-section . The influence of cation bridging is smaller for the monophosphoryl and absent for the dephosphorylated lipid A, thus explaining their preference to form lamellar structures. Kato and coworkers" have chosen a different approach for the determination ofthe molecular geometry by crystallizing synthetic lipid A (crystallization of natural lipid A failed) . These crystals were analyzed by electron diffraction, from which the anisotropic cross-section ofthe acyl

32

Lipid A in Cancer Therapy

Table 1. Bilayer periodicities for various lipidA preparations under different ambientconditions LipidA Samples Lipid A from LPS LipidA from LPS LipidA synthetic Lipid A from LPS

R595 R595 E.coli-type R595

LipidA from rough mutant LPS from S. minnesota

Bilayer Periodicity (nm)

Authors

Remarks

5.20 4.48 4.93 4.90 at 10% 5.10 at 50% 5.50 at 90% 4.80 5.40

4'(1985) 19(1993) 44(1996) 18(1990)

Dry and hydrated samples Dried samples Crystallized samples At 40°C and varying water content

19(1993)

Dry samples At 70% water content and [endotoxin): [Mg' +ll :l M

chains was determined to be 1.67 nm x 0.924 nm and the length of the lipid A part to 2.96 nm . When looking perpendicular to the membrane surface, the authors found that the acyl chains of synthetic E. coli-type lipid A form a hexagonal lattice with a lattice constant of 0.462 nm. The longitudinal axis, corresponding to the bilayer thickness, was determined to 4.93 nm . Finally, a monolayer thickness of2.47 nm was derived from this value. Thus, the data of these authors are in close agreement to those published by Labischinski er al.43 When comparing the data from various investigations on the molecular geometry for lipid A, very similar values for the bilayer thickness are found: in the absence ofwater or at low water content, values ranging from 4.5 to 5.2 nm are found, which increase to 5.5 nm at high water content.'! It should be noted that the values at high water content were obtained in the presence of'Mg", which was necessary because of the tendency of these compounds to adopt nonlamellar structures in the absence of divalent cations (see above). An overview ofthe bilayer periodicities ofvarious lipid A preparations under different ambient conditions is given in Table 1.46

Physicochemical Data in Relation to Biological Activity

A basic parameter describing endoroxicity is the aggregate structure of lipid A as described above, which is a measure ofits molecular conformation ('conformational concept'22.27). Striking correlations were found between the biological activity of lipid A from different species and their ability to adopt particular supramolecular structures: lipid A samples assuming lamellar structures (having a cylindrical molecular conformation), such as those from Rb. capsulatus and C. oiolaceum as well as tetra- and pentaacyllipid A from enterobacterial origin, are completely inactive. Those assuming mixed lamellar/cubic structures (partly conical molecular conformation, such as monophosphoryllipid A and lipid A from C.jejuni) have intermediate activity, and those samples preferring pure nonlamellar structures (Q. H ll , conical molecular conformation), such as enterobacterial hexaacyllipid A, are highly active. The inactive lipid A samples, which adopt lamellar structures, are found to act as efficient antagonists against biologically active LPS as long as they carry a sufficiently high effective negative charge." Uncharged lipid A with lamellar structures, such as lipid A from Rhodomierobium oannielii and dephospho hexaacyllipid A, are inactive agonistically as well as antagonistically. Similar data were obtained for the monosaccharide monophosphoryl triaycllipid A structures mentioned above: Tho se compounds with an acyloxyacyl chain in position 3 adopt a nonlamellar aggregate structure and exhibit bioactivity, which is, however, lower than E. coli lipid A, whereas those compounds with the acyloxyacyl chain linked to position 2 are inactive." The importance of the aggregate structure becomes evident also from the data for the triacyl lipid A OM-174 (see above) . Its preference for the micellar HI structure is connected with a moderate ability to induce cytokines. Concomitantly, as data from the group ofJeannin show, it

Conformation and Supramolecular StructureofLipid A

33

can successfully be utilized as effective antitumor drug in a rat model, but also in preclinical and clinical studies .48'5l As an example ofa possible importance ofthe value ofT., the differences in the bioactivities of lipid A and rough mutant LPS should be mentioned. It is known since long that lipid A, although representing the 'endotoxic principle' ofLPS, can induce TNFa or interleukins in mononuclear cells only at two orders of magnitude higher concentrations than LPS.22.47This can be correlated with the values of T, lying at 4S"C for lipid A and at 30 to 3S"C for LPS, i.e., LPS has a higher fluidity of the acyl chains at 37"C than lipid A. Of course, the higher fluidity of LPS at 37"C is also directly correlated with a better 'solubility' of these compounds as compared to lipid A (in this context, the term solubility is not defined, however, and is used only as expression for different aggregate sizes/shapes ofthe compounds). Strong evidence for the influence ofacyl chain fluidity on bioactivity was provided by Toman et al,34 who found even a higher TNFa production induced by lipid A than by the parent smooth LPS from Coxiella burnetii-although overall activity ofboth samples was significantly lower than that ofenterobacterial endotoxins, a fact which was assumed to be due to a lessbulky hydrophobic lipid A moiety (lower tendency for inverted structures of the lipid A moiety). As explanation for the relatively high TNFa-inducing capacity of this lipid A is its relatively high fluidity at 37"C. Its phase transition temperature T. is significantly below 37"C (the phase transition is in the range 10 to 30"C). The higher mobility ofthe acyl chains in the case of the lipid A moiety ofLPS may facilitate the interaction with target structures. Since it has been found that the aggregate structure oflipid A and for example LPS Re are essentially the same, an influence ofthe latter quantity may be excluded (for a more complete presentation ofthe influence ofthe acyl chain fluidity the reader is referred to an earlier review"), It must be emphasized, however, that a high acyl chain fluidity is not per se a determinant of bioactivity. Thus, the acyl chain moiety of the LPS as well as its lipid A part of Rb. capsulatus are highly fluid, however these molecules are both biologically inactive."

Conformational Concept ofLipid A Action: How Does Endotoxin Interact with Immune Cells?

The obvious correlation between the supramolecular aggregate structure and with that of the conformation of individual lipid A molecules on the one side and their bioactivity on the other side provokes the question to the early steps ofinteraction ofendotoxins with the host cell finally leading to its activation. This question leads automatically to the further que stion concerning the active unit of endotoxins: Is it the individual molecule or rather the supramolecular aggregate of a number of more or less identical molecules? Further questions are concerning the transport of the endotoxin units to their receptors/signalling complex and the molecules involved in this transport. The question of how bacterial endotoxins activate immune cells is not yet conclusively answered. There are many proteins-soluble as well as membrane-bound-known to be involved in endotoxin-induced cell activation. One important membrane-bound protein is mCD14, a SS kDa protein anchored to the membrane by a glycosylphosphatidylinositol anchor," Due to its lack of a transmembrane domain, mCD 14 cannot transfer a signal across the cell membrane. Instead, the probably most important function described for mCD 14 is the binding and transfer ofLPS or lipid A to other proteins, thus facilitating the delivery ofLPS or lipid A to the host cell and enhancing cell activation. Toll-like receptor (TLR) 4 was identified as a signal transducing coreceptor.Y? For functional signal transduction in response to endotoxin, TLR4 requires the presence of the secreted protein MD-2, which binds to the extracellular portion ofTLR4. 56.57 There is increasing evidence that, in addition to these proteins several other proteins, such as a potassium channel,58.59 heat shock proteins 70 and 90,60.61 and others are engaged in building a complex receptor cluster. Another important protein in this context is the LPS-bindingprotein LBP,described asa 6S kDa soluble acute-phase serum-protein produced by hepatocytes, Due to its positive charges, LBP binds

34

Lipid A in Cancer Therapy

to negatively charged molecules like endotoxin, but also to anionic lipids like phosphatidylserine or

cardlollpln.f It enhances LPS-binding to the cell membrane, thus enhancing cell activation.63.64 It

could be shown previously that LBP not only mediates a transport ofendotoxin and phospholipids to cell membranes, but also intercalates itself into cell membranes. This membrane-bound LBP (mLBP) plays a role in cell activation most likely by binding LPS similar to the function ofCD 14. In addition to its activating function, LBP also plays a role as aLPS-neutralizing protein.65.66 In order to understand the initial molecular steps in cell activation by LPS, the question has to be answered, whether endotoxin activates immune cells in the aggregated or in the monomerized form . The information on the nature ofthe biologically active unit ofLPS is ofoutmost importance and extremely value for an understanding of the molecular mechanisms ofcell activation. In this context it is interesting to note that preincubation of sLBP with LPS leads to a trans ition from cubic inverted to lamellar structures ofthe LPS-aggregates.These lamellarized LBP-LPS complexes are neither able to intercalate into liposomal membranes nor do they induce cytokine production in immune cells67 (and own unpublished results). The commonly accepted model of cell activation by endotoxin proposes that LBP removes monomers from endotoxin aggregates, transports them to mCDI4 in a kind of shuttle mecha nism , from where it is passed onto TLR4 which then, together with MD-2, is respons ible for transmembrane signaling and triggering of the intracellular signaling cascade, finally leading to cytokine production and release.64.68A recent paper by the group oO.P. Weiss suggests a vectorial transport alongsoluble proteins, with LBP binding to the LPS aggregate, opsonizing it for sCD 14, which then monomerizes the aggregate. This sCD 14-LPS monomer complex binds to sMD2, and the resulting sMD-2-LPS monomer complex binds then to TLR4, leading to cell activation/" This model proposes that it is the monomer that first interacts with the host cell. This preposi tion is backed by papers of Takayama et al who found that LPS is an order of2.5 more active in stimulating a pro-B-cellline in a disaggregated state and was more active in the LAL-assay than aggregated LPS.8.69 Inspired by the observation that natural isolates oflipid A from E. coli express a higher endotoxic activity than identical amounts ofcompound 506, Mueller et allooked closer at the chemical composition ofnatural Isolates" and found in these isolates that the largest fraction was hexaacylated, but that also significant amounts of penta- and rerraacylated molecules were present that, when administered to human mononuclear cells, may antagonise the induction of cytokines by biologicallyactive hexaacylated endotoxins.'?They prepared separate aggregatesofeither compound 506 or 406 and mixed these at different molar ratios as well as prepared mixed aggregates (by first mixing the compounds in an organic solvent) containing both compounds in the same ratios. Surprisingly, the latter mixtures showed higher endotoxic activity than that ofthe pure compound 506 up to an admixture of20% ofcompound 406 (Fig. 4) . Similar results were obtained when using various phospholipids instead ofcompound 406. Thus, when adding 406 or cardiolipin prepared as separated aggregates to compound 506, the antagonistic action of these compounds are well observable (Fig. 4, white bars) leading to inhibition ofcytokine induction. In contrast, 406 or CL molecularly mixed with 506 induce an increase ofthe cytokine activity of506 (FigA, black bars), probably by reducing the binding energy ofthe molecules within the aggregate. These findings can only be understood by assuming that the active unit of endotoxins is the aggregate. This interpretation wasverified by the authors by comparing monomeric (prepared from lipid A suspension by a dialysis process) and aggregated lipid A. They found that-at same con centrations-only aggregates were biologically active, whereas monomers showed no activiry.'! Based on these findings , the group of U. Seydel stated that all proteins involved are membrane-bound or at least membrane-associated and that it is the aggregate that first interacts with the immune cells. This is supported by a paper of Shnyra et aFO who showed that aggregated endotoxin activated the LAL cascade to a higher degree and was lethal in galactosamine-sensitized mice in lower doses than monomerized endotoxin. Seydel et al further propose that mLBP binds endotoxin aggregates and probably intercalates them into the cell membrane-as they could show for model membranes like phospholipid liposomes .64.65A strong support ofthis interpretation was

Conformation and Supramolecular Structure ofLipid A

35

A ";"

E

-

C)

c c o

5 4

-g

:-=-3

~

C

(J.)

oo

2

~

-.!-

.c

O ~'-iL-....Iii;;L--I_L.-..IiiiiiiiiiL-"""iiiiI_"'''''''''''-'_-----.A

1:0

E

5:1

2:1

1:1

0.5:1

0.2:1 0.1:1

0:1

[406]:[506]/molar ratio

B "7

10:1

1.2

-

C)

c c

-

~ 0.8 ~

c o c

(J.)

8

0.4

~

-.!-

.-c O~_"'_ "

1:0

__ 1:1

5:1

""'''''_'-''''_''''_'''.....l_'''~

0.2:1

0.1:1

0:1

[CL]:[506]/molar ratio Figure 4. TNFa production of human mononuclear cells after stimulation under serum-free conditions with (A) aggregate mixtures from hexaacyl lipid A compound 506 and tetraracyl lipid A 406 (light boxes) and mixed aggregates 406 (dark boxes) and (B) aggregates mixtures from 506 and cardiolipin (Cl) (light boxes), and mixed aggregates from 506 and Cl (black boxes). In the case of aggregate mixtures, the respective lip ids were prepared separately in buffer and then mixed, in the case of mixed aggregates, the lipids were dissolved together in organic solvent, and after the removal of the solvent buffer was added. Reproduced from : Mueller Met al. J Bioi Chem 2004; 279:26307-26313, with permission from ASBMB.12

36

Lipid A in Cancer Therapy

presented by Roes et al66 who found in liposomes single LBP molecules at low and LBP clusters at higher concentrations, and after external lipid A addition lipid A domains within the membranes. The formation oflipid A domains could be inhibited by anti-LBP-antibodies. Thus, intercalation of endotoxin aggregates into the membrane might bring it into close proximity of the proteins of the signaling complex. The conical conformation of lipid A is then responsible for the disturbance ofthe host cell membrane thus triggering conformational changes ofneighbouring membrane proteins. In this model, the aggregate is the active unit ofendotoxin, which interacts with the cell membrane and membrane-bound proteins.

Acknowledgements

The authors would like to thank Dr. M. Miiller, Dr. M .HJ. Koch and G. von Bussefor support in the biological, X-ray diffraction and FTIR experiments, respectively.

References

1. Rietschel ETh, Galanos C, Tanaka A et al. Biological activities of chemically modified endotoxins. Eur ] Biochem 1971; 22:218-224. 2. Rietschel ETh, Kirikae T, Schade FU ec al. Bacterial endotoxin: Molecular relationships of structure to activity and function. FASEB] 1994; 8:217-225. 3. Nikaido H, Vaara M. Molecular basis of bacterial outer membrane permeability. Microbial Rev 1985; 49:1-32. 4. Golenbock DT, Hampton RY, ~eshi N et al. Lipid A-like molecules that antagonize the effects of endotoxins on human monocytes.] Bioi Chern 1991; 266:19490-19498. 5. Mayer H , Krauss ]H, Yokota A ec al. Natural variants of lipid A. In: Friedman H, Klein TW; Nakano M and Nowotny A, eds Endotoxin, New York and London : Plenum press, 1990:45-70. 6. Loppnow H. Rierschel ETh. Brade H et al. Lipid A precursor Ia (compound 406) and Rhodobacrer capsulatus lipopolysaccharide: Potent endotoxin antagonists in the human sysrem in vitro. Levin.], Alving, CR. Munford. RS et al. P.337-348. 1993. Amsterdam. Elsevier. Bacterial endotoxin: Recognition and effector mechanisms. 7. Hofer M. Hampton RY. Raetz CRH er al. Aggregation behavior of lipid IVA in aqueous solutions at physiological pH . 1: Simple buffer solutions. Ch ern Phys Lipids 1991; 59:167-181. 8. Takayama K. Din ZZ , Mukerjee P et al. Physicochemical properties of the lipopolysaccharide unit that activates B lymphocytes.] Bioi Chern 1990; 265:14023-14029. 9. Buschner S. Bestimmung der kritischen Aggregatkonzentrationen von Lipiden: Implikationen fur die biologische Wirksamkeit von Endoroxinen, Universitat Kiel 1999; PhD Thesis. 10. Aurell CA. W istrom AO. Critical aggregation concentrations of gram-negative bacteriallipopolysaccharides (LPS). Biochem Biophys Res Commun 1998; 253:119-123. 11. Mueller M. Lindner B. Dedrick Ret al. Endotoxin : Physicalrequirements for cell activation.] Endotoxin Res 2005; 11:299-303. 12. Mueller M. Lindner B. Kusumoto S et al. Aggregates are the biologically active units of endotoxin . ] Bioi Chern 2004; 279:26307-26313. 13. Gioann ini TL. Teghanemt A. Zhang D et al. Monomeric endotoxin : Protein complexes are essential for TLR4-dependent cell activation.} Endotoxin Res 2005 ; 11:117-123. 14. Takada H . Kotani S. Structu ral requirements of lipid A for endoroxiciry and other biological activities. CRC Crit Rev Microbiol1989; 16:477-523. 15. Israelachvili ]N. Electrostatic Forces Between Surfaces in Lipids. Intermolecular and Surface Forces. London: Academic Press, 1991: 213-259. 16. Luzzati V. Vargas R. Mariani P et al. Cubic phases of lipid-contain ing systems. Elements of a theory and biological connotations . ] Mol Bioi 1993; 229:540-551. 17. Mariani P. Luzzati V. Delacroix H . Cubic phases of lipid-containing systems. Structure analysis and biological implications. ] Mol Bioi 1993; 204:165-189. 18. Brandenburg K. Koch MH]. Seydel U. Phase diagram of lipid A from Salmonella minnesota and Escherichia coli rough mutant lipopolysaccharide.] Struct Bioi 1990; 105:11-21. 19. Seydel U. Koch MH], Brandenburg K. Structural polymorphism of rough mutant lipopolysaccharides Rd to Ra from Salmonella minnesota.] Strucr Bioi 1993; 110:232-243. 20. Brandenburg K. Richter W; Koch MH] et al. Characterization of the nonlamellar cubic and H ll structures of lipid A from Salmonella enterica serovar Minnesota by X-ray diffraction and freeze-fracture electron microscopy. Chern Phys Lipids 1998; 91:53-69. 21. Brandenburg K. Mayer H . Koch MH] et al. Influence of the supramolecular structure of free lipid A on its biological activity. Eur] Biochem 1993; 218:555-563.

Conformation and Supramolecular Structure ofLipid A

37

22. Schromm AB. Brandenburg K. Loppnow H et al. Biological activities of lipopolysaccharides are determined by the shape of their lipid A portion. Eur J Biochern 2000; 267:2008-2013. 23. Seydel U. Schromm AB. Brade L ct al. Physicochemical characterization of carboxymethyl lipid A derivatives in rdation to biological activity. FEBS J 2005; 272:327-340 . 24. Brandenburg K. Lindner B. Schromm AB et al. Physico-chemical characteristics of triacyllipid A partial structure OM-174 in relation ro biological activity. Eur J Biochem 2000; 267:3370 -3377. 25. Brandenburg K. Matsuura M, Heine H ct al. Biophysical characterization of triacyl monosaccharide lipid A partial structures in relation to bioactivity. Biophys J 2002; 83:322-333. 26. Brandenburg K, Andra J, Muller M et al. Physicochemical properties of bacterial glycopolymers in relation to bioactivity. Carbohydr Res 2003; 338:2477 -2489. 27. Seydel U, Oikawa M, Fukase K ct al. Intrinsic conformation of lipid A is responsible for agonistic and antagonistic activity. Eur J Biochern 2000; 267:3032-3039. 28. Jain MK . Nonrandom lateral organization in bilayers and biomernbranes. In: Aloia RC Membrane Fluidity in Biology, Vol. I: Concepts of Membrane Structure , New York: Academic Press, 1983: 1-37. 29. Brandenburg K, Seydel U. Physical aspects of structure and function of membranes made from lipopolysaccharides and free lipid A. Biochim Biophys Acta 1984; 775:225 -238. 30. Brandenburg K, Seydel U. Investigation into the fluidity of lipopolysaccharide and free lipid A membrane systems by Fourier-transform infrared spectroscopy and differential scanninig calorimetry. Eur J Biochem 1990; 191:229-236. 31. Naumann D. Schultz C, Sabisch A et al. New insights into the phase behaviour of a complex anionic amphiphile: Architecture and dynamics of bacterial deep rough mutant lipopolysaccharide membranes as seen by FTIR, X-ray, and molecular moddling techniques . J Molec Strucr 1989; 214:213-246. 32. Naumann D, Schultz C. Born Jet al.Investigations into the polymorphism of lipid A from lipopolysaccharides of Escherichia coli and Salmonella minnesota by Fourier-transform infrared spectroscopy. Eur J Biochern 1987; 164:159-169. 33. Brandenburg K. Kusumoro S, Seydel U. Conformational studies of synthetic lipid A analogues and partial structures by infrared spectroscopy. Biochim Biophys Acta 1997; 1329:193-201. 34. Toman R, Garidd P, Andrs J er al. Physicochemical characterization of the endotoxins from Coxiella burnetii strain Priscilla in relation to their bioactivities. BMC Biochem 2004 ; 5:1. 35. Christ WJ. Asano 0 , Robidoux LC et al. E5531, a pure endotoxin antagonist of high potenc y. Science 1995; 268:80-83. 36. Bunnell E, Lynn M, Habet K et al. A lipid A analog, E5531, blocks the endotox in response in human volunteers with experimental endoroxemia. Crit Care Med 2000; 28:2713-2720 . 37. Mullarkey M, Rose JR . Brisrol JU et al. Inhibition of endotoxin response by E5564, a novel Toll-like receptor 4-directed endotoxin antagonist. J Pharmacol Exp Therap 2003; 304:1093-1102. 38. Lynn M, Rossignol DP, Wheder JL et al. Blocking of responses to endoroxin by E5564 in healthy volunteers with experimental endotoxernia, J Infect Diseases 2003; 187:631-639. 39. Asai Y, Iwamoro K, Watanabe S. The effect of the lipid A analog E5531 on phospholipid membrane properties . FEBS Lett 1998; 438:15-20. 40. Asai Y, Iwamoto K, Watanabe S. Characterization of the physicochemical properties of aggregates of the lipid A analog, E5531, prepared by a 'pH-jump method'. Chern Phys Lipids 1999; 97:93-104. 41. Asai Y, Watanabe S. Effect of divalent cations on the membrane properties of the lipid. A analog E5531. Drug Dev Ind Pharm 1999; 25:1107-1113. 42. Gutsmann T, Schromm AB, Koch MHJ et a1. Lipopolysaccharide-binding protein-mediated interaction of lipid A from different origin with phospholipid membranes. Phys Chern Chern Phys 2000; 2:4521-4528. 43. Labischinski H, Barnickd G, Bradaczek H er al. High state of order of isolated bacterial lipopolysaccharide and its possible contribution ro the permeation barrier property of the outer membrane. J Bacreriol 1985; 162:9-20. 44. Frecer V. Ho B. Ding JL. Molecular dynamics study on lipid A from Escherichia coli: Insights into its mechanism of biological action . Biochim Biophys Acta 2000; 1466:87-104. 45. Karo N, Naito S, Arakawa Y er al. Crystallization of synthetic Escherichia coli-type lipid A. Microb iol Immunol 1996; 40:33-38. 46. Brandenburg K, Wiese A. Endoroxins: Relationships between structure , function and activity. Curr Top Med Chern 2004; 4:1127-11 46. 47. Schromm AB, Brandenburg K, Loppnow H er al. The charge of endotoxin molecules influences their conformation and interleukin-6 inducing capacity. J Immunol 1998; 161:5464-5471. 48. Onier N, Hilpert S, Arnould L et al. Cure of colon cancer metastasis in rats with the new lipid A OM 174. Apoprosis of tumor cells and immunization of rats. Clin Exp Metastasis 1999; 17:299-306. 49. Reisser D, Pance A, Jeannin JE Mechanisms of the antitumoral effect of lipid A. BioEssays 2002; 24:284-289.

38

Lipid A in Cancer Therapy

50. Pance A. Reisser D. ]eannin ]E Antitumoral effects of lipid A: Preclinical and clinical srudies.} Invest Med 2002; 50:173 -178. 51. Larmonier CB. Arnould L. Larmonier N et al. Kinetics of tumor cell apoptosis and immune cell activation during the regression of tumors induced by lipid A in a rat model of colon cancer. Int ] Mol Med 2004 ; 13:355-361. 52. Seydel U. Labischinski H. Kastowsky M er al. Phase behaviour. supramole cular structure and molecular conformation of lipopolysaccharide. lmmunobiol 1993; 187:191-211. 53. Wright SD. Ramos RA. Tobias PS et al. CDI4. a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 1990; 249:1431 -1433. 54. Hoshino K. Takeuchi O. Kawai T et al. Cutting edge: Toll-like receptor 4 (TLR4)-deficient mice arc hyporesponsive to lipopolysaccharide: Evidence for TLR4 as the Lps gene product. ] lmmunol 1999; 162:3749-3752. 55. Poltorak A. He X. Smirnova I et al. Defective LPS signaling in C3H/He] and C5 7BLlI0ScCr mice: Mutations in Tlr4 gene. Science 1998; 282:2085-2088. 56. Shimazu R. Akashi S. Ogata H et al. MD-2, a molecule that confers lipopolysaccharide responsiveness on Toll-like receptor 4.] Exp Med 1999; 189:1777-1782. 57. Schromm AB. Lien E. Henneke P et al. Molecular genetic analysis of an endotoxin nonresponder mutant cell Line: A point mutation in a conserved region of MD -2 abolishes endoroxin-induced signaling. ] Exp Med 2001 ; 194:79-88. 58. Blunck R. Scheel O. Muller M et al. New insights inro endotoxin-induced activation of macrophages: Involvement of a K+ channel in transmembrane signaling.] lmmunol2001; 166:1009-1015. 59. Seydel U. Scheel O. Muller Met al. A K+ channel is involved in LPS signaling.] Endotoxin Res 2001; 7:243-247. 60. Triantafilou M. Miyake K. Golenbock DT et al. Mediators of innate immune recognition of bacteria concentrate in lipid rafts and facilitate lipopolysaccharide-induced cell activation. ] Cell Sci 2002 ; 1152:2603-2611. 61. Triantafilou K. Triantafilou M. Ladha S er al. Fluorescence recovery after photobleaching reveals that LPS rapidly transfers from CD14 to hsp70 and hsp90 on the cell membrane.] Cell Sci 2001; 114:2535 -2545. 62. Mueller M. Brandenburg K, Dedrick R et al. Phospholipids inhib it lipopolysaccharide (LPS)-induced cell activation: A role for LPS-binding protein.] lmmuno12005; 174:1091-1096. 63. Schumann RR. Leong SR. FlaggsGW et al. Structure and function of lipopolysaccharide binding protein . Science 1990; 249 :1429 -1431. 64. Hailman E. Lichenstein HS . Wurfel MM er al. Lipopol ysaccharide (LPS)-binding protein accelerates the bind ing of LPS to CDI 4. J Exp Med 1994; 179:269-277. 65. Gutsmann T. Haberer N. Carroll SF et al. Interaction between lipopolysaccharide (LPS). LPS-binding prote in (LBP) and planar membranes. Bioi Chern 2001 ; 382:425-434. 66. Gutsrnann T. Mueller M. Carroll SF er al. Dual role of lipopolysaccharide (LPS)-binding protein in neutralization of LPS and enhancement of LPS-induced activation of mononuclear cells. Infect lmmun 2001; 69:6942-6950. 67. Roes S. Mumm F. Seydel U et al. Localization of the lipopolysaccharide-binding protein in phospholipid membranes by atomic force microscopy.] Bioi Chern 2006; 281:2757-2763 . 68. Gioannini TL. Teghanemt A. Zhang D et al. Isolation of an endotoxin-MD-2 complex that produces Tell-like receptor 4-dependent cell activation at picomolar concentrations. Proc Nat! Acad Sci USA. 2004 ; 101:4186-4191. 69. Takayama K. Mitchell DH. Din ZZ et al. Monomeric Re lipopolysaccharide from Escherichia coli is more active than the aggregated form in the Limulus amebocyte lysate assay and in inducing Egr-I mRNA in murine peritoneal macrophages.] Bioi Chern 1994; 269;2241-2244. 70. Shnyra A. Hultenby K. Lindberg AA. Role of the physical State of Salmonella lipopolysaccharide in expression of biological and endoroxic properties. Infect lmmun 1993; 61:5351-5360.

CHAPTER 4

Interactions between Lipid A and Serum Proteins Jorg Andra,* Thomas Gutsmann, Mareike Miiller and Andra B. Schromm Abstract

E

nt ry ofendotoxin (lipopolysaccharide (LPS) or lipid A) into the blood stream is causative for the emergence ofsepsis and septic shock with all its pathophysiological consequences. I Serum contains a whole variety of proteins that interact with endotoxin. As large as the number of different proteins interacting with endotoxin, as broad are the consequences of the se interactions. Serum proteins can either enhance cell activation by endotoxin or attenuate the cellular response, they can detoxify and eliminate endotoxin from the blood stream. In this chapter we summarize work on the investigation of the interaction ofendotoxins with serum proteins. In four paragraphs we focus on proteins involved in the endotoxin-induced immune cell activation, detection by immunoglobulins, the transport of endotoxins and on proteins and peptides with the capability to neutralize the biological effects ofendotoxin (Fig. 1). There is a multitude ofstudies analyzing the interactions between serum proteins and endotoxin s, howe ver, with great differences in the sour ce and quality ofthe endotoxins used. The number ofstudies dealing with chemically well defined endotoxin structures are quite limited. In addition, though lipid A is the biologically active entity, the "endotoxic pr inciple", ofLPS, the majority of studies was performed with LPS. Therefore, to be comprehensive, we included also stu dies dealing with LPS and not with lipid A if fundamental scientific problems were add ressed. In that cases, we have to be aware that the re may be differences in the protein interactions oflipid A and LPS, and we tried to emphasize thi s point in the respective paragraphs.

Proteins Involved in Lipid A/LPS-Mediated Immune Cell Activation

Isolated lipid A is the 'endotoxic principle' ofLPS, the major component ofthe outer layer of the outer membrane ofGram-negative bacteri a. Like LPS, lipid A induces th e activation ofimmune cells leading to the production of a variety of pro-inflammatory mediators such as turnor-necrosis-factor-(TNF) c, interleukins, and tissue-factors (reviewed in ref. 2). The in itiation ofsignaling is preceeded by a transport oflipid A to cellular receptor proteins, mediated by the serum proteins soluble CD14 (sCD14)3and the lipopolysaccharide-binding protein (LBP).4Membrane bound CD14 (mCD14), anchored to the membrane by a glycosylphosphatidylinositol anchor, is the first cell surface structure involved in cell activation.' Due to its lack ofa transmembrane domain, mCD 14 cannot transfer a signal across the cell membrane. Instead, the probably most important function described for mCD14 is the binding and transfer ofLPS or lipid A to other proteins, facilitating deliver y of LPS or lipid A to the host cell and enhancing cell activation. Th e presence oflipid A and LPS is sensed by a complex receptor cluster on the surface ofimmune cells. Toll-like *Correspond ing Author: Jorg Andra-s-B lophyslcs Division, Research Center Borstel, Leibniz-Center for Medicine and Biosciences, Parkallee 10, 23845 Borstel, Germany. Email: jandraeofz-borstel.de

Lipid A in Cancer Therapy, edited byjean-Francois ]eannin. ©2009 Landes Bioscience and Springer Scien ce+ Business Media.

40

LipidA in Cancer Therapy

Transport

Neutralization

Figure 1. Overview over the lipid A interactingprotein groups described in this review. LBP, lipopolysaccharide-binding protein; TLR4, toll-like receptor 4; LOL, low-density lipoprotein; HOL, high-density lipoprotein; CAP18, cathelicidin cationic antimicrobial protein 18 kOa; BPI, bactericidal/permeability-increasing protein. receptor (TLR) 4 together with MD-2, a soluble protein bound to the extracellular domain of TLR4. has been shown to be the central receptor protein for the recognition ofLPS and lipid A, initiating the generation ofa signal across the cellular membrane upon binding oflipid A.6-10 LBP is a 55 kDa soluble acute-phase serum protein produced by hepatocytes.'! Its concentration in normal serum is in the range of 1-5 fLg/ml. At these concentrations, LBP accelerates LPS-binding to the cell membrane, thus enhancing cell activation." Targeted deletion ofthe gene for LBP in mice has underlined the role ofLBP in cell activation by LPS 12 as well as its importance in combating infections with Gram-negative bacteria.F'Ihe neutralization ofLBP in serum by an LBP-specific antibody protects mice against the lethal effects of endotoxic shock after challenge with LPS and with lipid A. 14 Due to its positive charges, LBP binds specifically and with high affinity to the negatively charged lipid A.IS Gazzano-Santoro er al 16 determined the binding affinity ofrecombinant LBP protein to lipid A with an apparent K o of58 nM. The lipid A binding site of LBP has been attributed to its N -terrninal domain (amino acid residues 1-197), whereas the binding ofLBP to CD 14 and the biological effects are mediated by the C-terminal domain ofLBP.17.18 Besides the transport oflipid A and LPS to CD 14. LBP also mediates a transport oflipid A into phospholipid membranes.P'" a process that has been discussed to intercalate lipid A molecules into the cytoplasmic membrane thereby delivering lipid A to the receptor molecules. CD 14 is a 53 kDa serum protein released from myelomonocytic cells by proteolytic cleavage from the phosphatidylinositol membrane-anchor. The binding sites for rough and smooth LPS in CD 14 have been extensively studied by mutagenesis and by epitope mapping with antibodies that inhibit LPS binding to CD 14.23-29 Four regions have been identified at the N-terminal domain of sCD14 that are important for the binding of LPS.30 The recently published crystal structure of CD 14 revealed that the N -terrnlnal domain of the protein forms a hydrophobic pocket which is the main binding site for LPS.31

Interactions between Lipid A and Serum Proteins

41

The function of CD 14 and LBP is not restricted to their accessory role enhancing cell activation by low concentrations of endotoxin. Recent investigations have shown that both proteins have dual roles in the process of cell activation, depending on their concentration in serum. At high concentrations, as they appear in acute phase serum, CDI4 and LBP have been shown to attenuate cell activation by lipid A and LPS.32.37 Recently, it is recognized that besides the soluble form ofCD 14. other proteins ofthe cellular receptor cluster exist in a soluble form in serum. Thus, it has been suggested that MD-2, a 20-25 kDa extracellular glycoprotein that associates with the extracellular domain ofTLR4, may also be active as soluble protein in serum. We and others have shown that the soluble form ofMD-2 enables cells that express only TLR4 but no MD-2 on the cell surface to respond to LPS.38.39 Kennedy and coworkers" found. that soluble MD-2-once it is secreted from the cells-losesits bioactivity in serum over a period of 24 hours at physiological temperature. They found, that this loss of activity can be prevented by bindingofLPS, but not oflipid A, to form stable MD-2 complexes. Others have shown that MD-2 in serum can attenuate activation ofcells that express the TLR4/MD-2 receptor complex, probably bya competition ofbindingofLPS to the cellular TLR4/MD-2 complex." Re and Strorninger? found that multimeric MD-2 formed by intermolecular disulfide bridging inactivates the protein, whereas the monomeric MD-2 harbors bioactivity to LPS. Pugin and coworkers" investigated the effect ofsoluble MD-2 in plasma from patients with severe sepsis and found that septic plasma imparts LPS-responsiveness to epithelial cells only expressing TLR4, indicating an activity of soluble MD-2. Visintin et al44 recently determined the amount ofsoluble monomeric MD-2 in human healthy serum by depletion ofnative MD-2 and subsequent reconstitution with recombinant MD-2 protein to be about 50 nM. An exceptionally attribute ofthe TLR4/MD-2 receptor complex is its unique species-specific reaction to the tetraacyllipid A precursor lipid IVa and its synthetic analog "compound 406" which activate mouse macrophages but are inactive and act as LPS antagonist in human macrophages.This species-specific pharmacology has been attributed to be conferred by TLR445.46 and by MD_2. 47,48 Mutational analysis identified amino acids 57-79 and 108-135 critical for the species specific recognition oflipid IVa and "compound 406" by cell associated MD_2. 48Viriyakosol et al49 performed mutational analysis ofmembrane bound and soluble MD-2 and defined the LPS binding region ofmonomeric soluble MD-2 as a cluster ofthe basic residues 125-131. The authors suggest that in both membrane and soluble MD-2 the MD-2 domains for bindingTLR4 and LPS are separate as well as overlapping.

Detection of Lipid A by Immunoglobulins

Anti-lipid A antibodies are present in the serum ofhealthy individuals and more frequently in patients with Gram-negative infections. These natural antibodies react only with free lipid A but not with the lipid A part within the intact LPS molecule." Immunization ofrabbits with lipid A and lipid A partial structures led to the discovery ofpolyclonal antibodies with specifines directed against the diglucosamine backbone, the l-phosphate and the 4' -phosphare group of lipid A,5I The immunogenicity oflipid A was strongly dependent on its physicochemical state, with the best results obtained for lipid A embedded in liposomes or coated on erythrocytes." A direct interaction with the sugar backbone oflipid A without the need ofthe lipid A acyl chains was also confirmed with monoclonal antibodies and synthetic diglucosamine linked to albumin." The discrimation between lipid A and LPS was attributed to a hydroxyl neoepitope at the C6' position of lipid A generated by the acid hydrolysis of the attached Kdo sugarY

Proteins Involved in Lipid A/LPS Transport Serum Albumin

Albumin is the major protein in serum. It is a single chain 65 kDa protein which acts as a carrier for small hydrophobic molecules like bilirubin, fatty acids and certain hormones. Due to its high concentration in serum (65% of all serum proteins) it is very likely that albumin is among the direct interacting partners for LPS and lipid A in vivo.

42

Lipid A in Cancer Therapy

An interaction of serum albumins from different sources (human HSA, bovine BSA, rabbit RSA) with LPS and lipid A have been reported already in the early 70th.54•55 Lipid A has been shown to form complexes with serum albumins. These complexes, when injected into a rabbit, elicit a significant highe r pyrogenicity than pure lipid A. This finding is in sharp contrast to the effects of other serum proteins like bactericidal/permeability-increasing protein (BPI) and lacroferrin which completely inhibit or reduce the endotoxic activity oflipid A.56.57 A biphasic binding reaction of lipid A to HSA was suggested by David and colleagues." By spectroscopic analysis of a sole tryptophane and fluorescently labelled sole cystein in albumin, they found a molar ratio of lipid A to albumin of2:1 with two distinct binding sites in domain III (referred to the crystal structure59) of albumin with affinities for both sites in the micromolar concentration range. Notably, these binding sites are different from that of the antibiotic polymyxin B which has been shown to bind to the lipid A headgroup phosphates/? The most comprehensive and recent investigation of the lipid A : albumin interaction was performed by Jiirgens et aL61 The authors used recombinant human serum albumin (rHSA) which is free of activity in the limulus-amoebocyte-Iysate (LAL) assay, i.e., free of contaminating endotoxins. They studied the effects of albumin on various physicochemical parameters of lipid A and LPS, i.e.,acyl chain fluidity, Zeta potential, conformation oflipid A and supramolecular organization. Interactions of albumin with lipid A seem to be driven by electrostatic as well as by hydrophobic forces. In the presence of serum albumin, the phase transition ofthe acyl chains oflipid A is shifted to slightly higher temperatures and the fluidity of the acyl chains is rigidified in the liquid crystalline phase . No change was observed for the backbone orientation of lipid A organized in lamellar multilayers. The inverted cubic aggregate structure oflipid A, a prerequisite for endotoxic activity (see also Chapter 3, Physical Properties ofLipid A), is converted into another not clearlydefined but also inverted structure, which might explain the enhanced endotoxic activity of albumin: lipid A complexes observed by others." Effects of rHSA on lipid A are small albeit definite. The Zeta potential, a measure for the accessiblesurface charge ofaggregates isonly partially compensated by the addition ofalbumin, reaching a saturation at a lipid A to albumin molar ratio ofapproximately 10: 1. This ratio is in contrast to the study ofDavid et aL It should be mentioned however, that Jurgens et al used highly purified and chemically defined lipid A , whereas the lipid A source of David et al is commercial and that the "structural heterogeneity of lipid A is impedant to a detailed analysis."! In conclusion,Jiirgens et al. suggest that, in contrast to many cationic compounds, albumin is not tightly bound to lipid A phosphates but along its perpendicular axis,which again is in consistence with David et al. In such a scenario, serum albumin acts a carrier for lipid A to the immune cells. Concomitantly, an essential role of albumin for endothelial cell activation by LPS was described by Gioannini et aL 62

Lipoproteins

Lipoproteins are spherical macromolecular particles composed ofa hydrophobic core containing cholesterol esters and triglycerides surrounded by a layer of phospholipids, uneseerified free cholesterol and one or more proteins called apolipoproteins. Cholesterol and apolipoproteins stabilize the particles. Apolipoproteins are also involved in lipid recruitment, modulation ofenzyme activity and modulation of receptor-mediated binding and endocytosis. In human blood, four major lipoprotein classescan be distinguished according to their density, i.e.,their protein content: chylomicrons, very low density lipoproteins (VLDL), low density lipoproteins (LDL), and high density lipoproteins (HDL). These lipoprotein classes differ with respect to size, electrophoretic mobility and lipid and apolipoprotein composition.63•64 In addition to their primary role in the transport oflipids, cholesterol and cholesteryl esters in blood and the lymphatic system,lipoproteins are known to have a great impact on the LPS-induced cell activation by controlling responses to endotoxin.f Severalstudies show the effect oflipoproteins on the biological activity ofLPS and lipid A. Interaction ofLPS with HDL leads to the reduction of cytokine and adhesion molecule productions in human monocytes, HUVECs, endothelial cells, and in whole blood assays.66-70 Binding of LDL to LPS reduced cytokine production in

Interactions between Lipid A and SerumProteins

43

human monocytes. T " HDL and LDL or chylomicrons reduced LPS-induced toxicity ofbovine endothelial cells," biological activity ofLPS in the LAL assay74 and reduced septic shock in rodent endotoxemia models .69.75 Corresponding to their phospholipid content, LPS binds strongest to HDL, then LDL, then VLDL,76 In an experimental setup utilizing LPS with different distribution of the fatty acids of the lipid A part and varying sugar chain length, Sprong et al showed that not the saccharide tail of the LPS causes differences in lipoprotein-dependent inhibition ofcytokine but the lipid A part of the LPS molecule." LPS ofthe same species with longer and shorter sugar chain length was used and made no difference in lipoprotein binding. The lipid A part, however, with its asymmetrical hexaacylated conformation in E. coli LPS and its symmetrical distribution of the six acyl chains in N. meningitidis LPS, was shown to account for the divergent effect. The bioactivity of an LPS molecule is strongly determined by the three-dimensional conformation of the lipid A part.?8.79 For LPS-lipoprotein interactions, a "leaflet insertion" model is proposed,66.80 in which the fatty acid acyl chains ofthe lipid A portion ofLPS are inserted into the lipid monolayer ofthe lipoproteins, thereby inactivating the "toxic moiety" ofLPS. Sprong et al presume that the lower degree of neutralization of meningococcal LPS by lipoproteins is caused by the symmetrical lipid A structure interacting more slowly with the lipoprotein lipid monolayers." Another mechanisms is proposed by Brandenburg and coworkers who could show that HDL leads to a rigidification ofthe fatty acids oflipid A. In the presen ce ofHDL, the aggregate structure of lipid A changes from cubic inverted to a mulrilamellar one which corresponds with a decrease ofbiological activity," In addition to affecting the biological activity ofagonistic LPS and lipid A, lipoproteins are able to reduce the activity ofantagonists as well. Nonpathogenic bacteria such as Rbodobaaer capsulatus synthesize LPS and lipid A that lack stimulatory activities and that can antagonize activation of cells by more toxic LPS.82 £5531 is a synthetic stabilized analogue ofR. capsulatus lipid A and has been described to be a potent in vitro and in vivo antagonist ofLPS.83The antagonistic effect of £5531 is reduced when £5531 is preincubated in blood, serum or serum components prior to the addition ofLPS. £5531 binds to HDL and the £5531 induced inhibition ofLPS induced TNFa production is dependent on the concentration ofHDL,84 Another lipid A analogue which antagonizes LPS-induced cell activation is £5564. 83 Several studies showed that £5564, like £5531 , binds to HDL and its inhibitory activity is decreased in its presence.i':" Not only whole lipoproteins like HDL are able to inhibit the biological activity of bacterial LPS. The apolipoprotein ApoA -I was able to reduce the biological activity ofLPS in the LAL assay to the same extend as HDL and LDL, 74Ma and coworkers showed that ApoA-I is sufficient and responsible for the inhibitory effect ofHDL on the LPS-induced L-929 cell mortality," Several stu dies with the apolipoprotein ApoA-I mimetic L-4F supported this finding. A mixture ofL4-F diminished VCAM-l expression, an adhesion molecule thought to playa crucial role in monocyte infiltration into the arterial wall ofhuman monocytes in a dose-dependent manner.89Additionally, Gupta et al found that L-4F prevented the binding ofLPS to LBP in a concentration-dependent manner, suggesting this to be attributable to an interaction between L-4F and LPS. Amphipathic molecules like LPS and lipid A form aggregates in aqueous solutions. It has been suggested that LPS is biologically active in the aggregated state and that the shape of the aggregate influences its binding to LBP ICD 14.90 The data of Gupta et al suggest that L-4F binds LPS and alters its aggregation state . Because the lipid A domain ofLPS is the binding site for LBP, they hypothesized that the binding ofL-4F to LPS induces a conformational change in LPS by altering its aggregation state, which would render the lipid A region inaccessible to LBP, thus preventing assembly ofan LPS-LBP complex and exposure to cell surface receptors." This idea is supported by the findings ofBrandenburget al that an admixture ofHDL to lipid A leads to a change in th e supramolecular structure." When Gram-negative bacteria or LPS enter the bloodstream the most important known mechanism for preventing excessive cellular responses is the transfer ofLPS to circulating plasma

44

Lipid A in Cancer Therapy

lipoproteins. When purified LPS aggregates are injected intravenously into animals, LPS binds rapidly to circulating lipoproteins." This lipoprotein-bound LPS is cleared from circulation by transport to the liver9z and is finally excreted to the bile.93•94 The rapid binding ofLPS to lipoproteins, a crucial step to LPS clearance , is facilitated by LBP and the phospholipids transfer protein (PLTP), thus preventing septic shock.

Proteins Neutralizing the Immune-Cell Activating Properties ofLipid A/LPS Lactoferrin

Lactoferrin (LF) is a 80 kDa iron binding, cationic host defense protein found in mucosal secretions (milk, tears, saliva) and in the secondary granules of polymorphonuclear neutrophils (PMN). It can be released into the bloodstream following infection in concentrations greater than 1 flgIml. LF exhibits antibacterial activities via two distinct mechanisms, by complexation of iron and by destabilization of the bacterial outer membrane. The latter mechanism is located in a natural N-terminal cleavage product ofLF called lactoferricin and is strongly connected to the ability oflactoferricin to bind and neutralize lipid A and LPS and prevent the activation of immune cells in vitro and in vivo. 57.95.96 Binding of LF purified from human milk to immobilized lipid A was demonstrated first by Appelmelk et al.97 They used synthetic lipid A (compound 506) and mono- and bisphosphoryllipid A from E. coli, Pseudomonasaeruginosa and other bacteria, and clearly demonstrated that the headgroup phosphates are crucial for the binding ofLF to lipid A and that LF binding competes with that of polymyxin B. The binding reaction appears saturable with an apparent affinity constant in the nanomolar range and a molar lipid A:LF binding stoichiometry of2: 1. Furthermore, the inhibition of LF binding to LPS by lipid A and not by Kdo suggests that LF binds to LPS also via the lipid A part ofthe molecule. Brandenburg et al95 reported that binding ofLF strongly but not exclusively involves the lipid A phosphates. They estimated a molar lipid A:LF binding stochiornetry of3:1 to 4: 1 by titration calorimetry. Binding ofLF to lipid A was accompanied by a dramatic change of the lipid A aggregate structure from an inverted cubic structure to a mul tilamellar one. Again, the binding was competetive to polymyxin B binding to lipid A with an affinity lower than polymyxin B. LF has an impact on the gel to liquid crystalline phase transition oflipid A: the enthalpy is lowered and the phase transition temperature is enhanced," The LPS binding region within human lactoferricin was attributed firstly to residues 28-34 (for peptide sequence, see Table 1).98 A second binding epitope was described by van Berkel et a199 who isolated N-terminally truncated LF from human milk and analyzed its interactions with immobilized ligands includ ing lipid A. Deletion offive amino acid residues , including four positively charged arginines, completely abolished binding ofLF to lipid A , heparin, and DNA. Noteworthy, Majerle et al 100 determined a medium affinity constant of 20.7 fLM for the binding of a 12-mer peptide fragment (LF 12, Table 1) oflactoferricin (res. 21-31) spanning the respective LPS-binding site to lipid A. This apparent discrepancy to the high affinity binding ofLF to lipid A 95·97 may result from a structural reorientation of the LF fragment from a-helical to fl-sheet or stress the significance ofthe second binding site for LPS/lipid A within LF. A very similar lactoferricin-derived peptide

Table 1. Amino acidsequencesof human lactoferrin and derivedneutralizing peptides Designation

Sequence

Reference

human lactoferrin (21-36) LF12 LFll

FQWQRNMRKVRGPPVS FQWQRNIRKVR-(HL) FQWQRNIRKVR-CONH z

98

100

96

CONH z and (HL) indicate terminal amide and homoserine lactone groups, respectively.

Interactions between Lipid A and SerumProteins

45

(LF II, Table 1) was effective in neutralizing the LPS-induced activation ofhuman mononuclear cells and, like the whole protein LF, converted the aggregate structure oflipid A into a multilamellar stack ," In concordance to findings with other LPS/lipid A neutralizing peptides and proteins (see also Chapter 3, Conformation and supramolecular structure oflipid A andlOl.I03), LPS/lipid A in such a conformation no longer is capable to stimulate cytokine production in immune cells (conformational concept of endoroxicity'P']. Thus , LF may act as a high-affinity scavenger that prevents the transport oflipid A to the receptor complex on the immune cell membrane.

Bactericidal/Permeability-IncreasingProtein (BPI) andLPS-BindingProtein (LBP)

BPI I05 and LBP both bind to LPS, and a sequence comparison for human LBP and BPI revealed 44% amino acid residue Identity," In contrast to BPI, LBP has no effect on the viability of Gram-negative bacteria at concentrations at which BPI isvery effective,l06 and the effects ofLBP and BPI on LPS-induced cytokine release from mononuclear phagocytic cells are counteractive.l'" The subsequent addition ofLBP and LPS, or vice versa, to mononuclear cells or macrophages induced an increase in TNFa production with increasing LBP concentration,I07.108 however, complexation ofLPS with LBP prior to application led to a decrease of TNFa production with increasing LBP concentration. These results are indicative of a direct LPS-neutralizing effect of LBP,independent ofother components like HDLor LD L. Variousbiophysical experiments showed that LPS/LBP complexes do not bind or intercalate into lipid matrices resembling the membrane of macrophages and even not to LBP already intercalated into the membranes." Thus , it can be concluded that complexation ofLPS with LBP prior to binding ofLPS to membrane-associated LBpl09results in LPS neutralization and , thus , to inhibition ofmononuclear cells activation. Most probably the neutralization oflipid A or LPS occurs due to (i) an increase of the binding energy of the individual lipid A molecules within the aggregate and (ii) the formation of multi-lamellar structures oflipid A .110 In in vitro (in the presence of serum) and, in particular in in vivo experiments, a neutralization ofLPS by LBP has only been found at high acute phase LBP concentration as described above. Thus, the clear distinction between the two roles ofLBP cannot easily be elaborated, because the effects are concentration-dependent, and a variety offurther proteins and other serum constituents such as soluble CD 14,III BPI ,107CAP 18 112 and lipoproteins'P enhance or suppress LPS-induced activation ofmononuclear cells. Molecular modeling suggests that the three-dimensional structures of BPII 14 and LBP are closely similar,I15 however, BPI does not increase the LPS -induced activation ofimmune cells but is a very potent inhibitor ofCD 14-dependent cell activation. 116 BPI is stored in primary granules of polymorphonuclear leucocytes. One recombinant fragment rBPI21 (containing residues 1-191) has all antibacterial and neutralizing properties ofholo-BPI. 117.118Th is fragment, which is also called NEUPREX, is non-toxic and non-immunogenic and has been tested in Phase II/III clinical trails with apparent therapeutic benefit.

Cathelicidins

Many antibacterial peptides and proteins which are an integral part of the host's defense barrier against invading bacteria are antimicrobial and they also neutralize lipid A and LPS released from Gram-negative bacteria. One group ofthese antibacterial proteins are the cathelicidins which have been identified in various mammals including hurnans.!" They contain a highly conserved N-terminal domain called cathelin and a C -terminal domain that comprises an antimicrobial pep tide . Human and rabbit cathelicidins are termed 18 kDa cationic antibacterial protein hCAP 18 and rCAP 18, respectively.CAP18 is sto red in the intracellular granules ofneutrophilic granulocytes and is liberated into the phagocytic vacuoles during phagocytosis po Interestingly, the human plasma concentration of CAP 18 (1.2 ~ ml) is several fold higher than that of other neutrophil specific granule proteins.'!' The bactericidal C -terminal fragment of hCAP 18 FALL-39 (hCAP 18 102.140) and its precursor(s) have also been found in wound fluid,122 and LL-37 (hCAP 18104.140) in alveolar macrophages, bronchial epithelial cellsand bronchial glands!" Antibacterial and cytotoxic activity of the human LL-37 (hCAPI8 104.14O) is inhibited in plasma by binding to apolipoprotein A-I.124·126

46

Lipid A in Cancer Therapy

The C-terminal domain ofCAP 18 exhibits (LPS)-binding, LPS-neutralizing, antibacterial and anticoagulant acrivities.P? A major difference in the interaction ofhuman and rabbit CAP 18 with cell membranes is observed in their effect on human red blood cells: the hCAP 18-fragment FALL- 39 (hCAP 18102.140) is haemolytic whereas rCAP 18 106-142 is not.128 CAP18 belongs to the group of antimicrobial peptides having a net positive charge , an a-helical conformation-most pronounced in the presence of negatively charged lipids, and a high amphipathicity.P"!" It has been shown that rabbit as well as human CAP18 fragments can inhibit the LPS-induced release of cytokines (TNFa, IL-l, IL-6), and nitric oxide from macrophages in in vitro experiments.!" Furthermore, rCAP 18106.142 could protect mice l32and rCAP 18 106. 137pig l33from lethal effects ofLPS. Some positive effects on the lethality ofmice could be shown for human CAP 18 fragmenes .P' The LPS-induced activation of macrophages can be inhibited by cathelicidins by blocking the binding ofLPS to CDI4. 135 In recent publications it has been shown that it is possible to augment the LPS-neutralizing activities of CAP 18-derived peptides by modifying their hydrophobicity and cationicity.136'138 These physical parameters ofthe CAP 18 fragments can be decisive for the interaction with lipid A and LPS.139 Peptides showing a high biological activity can bind to the phosphates ofthe lipid A, however, less active fragments do only interact with the Kdo ' s or other groups of the core region of the LPS.I40

Conclusion

Lipid A interacts with a variety ofserum proteins. The consequences ofthese interactions are extremely divergent as outlined in this chapter (see also Fig. 1). A comprehensive understanding ofthe interplay between endotoxins and host proteins may serve as basis for the development of specific drugs for example for the therapy of cancer. Notably, there are many crosslinks between the four groups ofproteins we have discussed in this chapter, such as the LBP-mediated transport oflipid A to LDL and HDL, which are likely to account for substancial differences in in vitro and in vivo assays. For the therapeutic application oflipid A or structural mimetics the complex and competing interplay between endotoxins and the diversity of proteins with different functions have to be considered.

Acknowledgements

This work was financiallysupported by the Deutsche Forschungsgemeinschafi:(SFB617 project A17 , SFB367 project B8 and Emmy-Noether grant SCHR 621/2-2).

References

1. Riecschel ET, Mamar U, Hamann L ct al. Bacterial endotoxins as inducers of septic shock. Novo Acta Leopoldina 1999; 307:93 -122. 2. Alexander C, Rietschel ET. Bactcrial lipopolysaccharides and innate immunity.] Endotoxin Res 2001 ; 7:167-202. 3. Frey EA, Miller DS, ]ahr TG er al, Soluble CD14 participates in the response of cells to lipopolysaccharide.] Exp Med 1992; 176:1665 -1671. 4. Schumann RR, Leong SR, Flaggs GW et aI. Structure and function of lipopolysaccharide binding protein . Science 1990; 249:1429-1431. 5. Wright SD, Ramos RA, Tobias PS, Ulevitch R], Math ison]C. CDI4, a recepto r for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 1990; 249:1431-1433. 6. Polrorak A, He X, Smirnova 1 et al. Defective LPS signaling in C3H/He] and C57BL/I0ScCr mice: mutations in Tlr4 gene. Science 1998; 282 :2085-2088. 7. Q!:reshi ST, Lariviere L, Leveque G et at Endotoxin-tolerant mice have mutations in Toll-like receptor 4 (Tlr4).] Exp Med 1999; 189:615-625. 8. Hoshino K, Takeuchi 0, Kawai T et al. Cutting edge: Toll-like receptor 4 (TLR4)-de6cient mice are hyporesponsive to lipopolysaccharide : evidence for TLR4 as the Lps gene product. ] Immunol 1999; 162:3749-3752. 9. Shimazu R, Akashi S, Ogata H ec al. MD-2, a molecule that confers lipopolysaccharide responsiveness on Toll-like receptor 4.] Exp Med 1999; 189:1777-1782. 10. Nagai Y, Akashi S, Nagafuku M ct al. Essential role ofMD-2 in LPS responsiveness and TLR4 distribution . Nat Irnmuno12002; 3:667-672 .

Interactions between LipidA and Serum Proteins

47

11. Ramadori G. Meyer zum Buschenfelde KH. Tobias PS er al. Biosynthesis of lipopolysaccharide-binding protein in rabbit hepatocytes, Pathobiology 1990; 58:89-94. 12. Wurfel MM. Monks BG, Ingalls RR er al. Targeted deletion of the lipopolysaccharide (LPS)-binding protein gene leads to profound suppression of LPS responses ex vivo. whereas in vivo responses remain intact . J Exp Med 1997; 186:2051-2056. 13. Jack RS, Fan X, Bemheiden M et al. Lipopolysaccharide-binding protein is required to combat a murine gram-negative bacterial infection. Nature 1997; 389:742-745. 14. Gallay P. Heumann D. Le Roy D et al. Lipopolysaccharide-binding protein as a major plasma protein responsible for endoroxemic shock. Proc Nat! Acad Sci USA. 1993; 90:9935-9938. 15. Tobias PS. Soldau K. Ulevitch R]. Identification of a lipid A binding site in the acute phase reactant lipopolysaccharide binding protein. J Bioi Chern 1989; 264: 10867-10871. 16. Gazzano-Santoro H. Meszaros K, Birr C er al. Competition between rBPI23. a recombinant fragment ofbactericidallpermeability-increasing protein. and lipopolysaccharide (LPS)-binding protein for binding to LPS and gram-negative bacteria. Infect Immun 1994; 62:1185-1191. 17. Theofan G. Horwitz AH , Williams RE et al. An amino-terminal fragment of human lipopolysaccharide-binding protein retains lipid A binding but not CDl4-stimulatory activity. J Immunol 1994; 152:3623-3629 . 18. Han J. Mathison JC. Ulevirch RJ et al. Lipopolysaccharide (LPS) binding protein , truncated at Ile-197, binds LPS but does not transfer LPS to CDI4. J Bioi Chern 1994; 269:8172-8175. 19. Schromm AB. Brandenburg K. Rierschel ET et al. Lipopolysaccharide-binding protein mediates CDl4-independent intercalation oflipopolysaccharide into phospholipid membranes. FEBS Lett 1996; 399:267-271. 20. Schromm AB. Brandenburg K, Loppnow H et al. The charge of endotoxin molecules influences their conformation and IL-6-inducing capacity. J Immunol1998; 161:5464-5471. 21. Schromm AB. Brandenburg K, Loppnow H et al. Biological activities oflipopolysaccharides are determined by the shape of their lipid A portion. Eur J Biochem 2000; 267:2008-2013. 22. Gutsmann T. Schromm AB, Koch MHJ et al. Lipopolysaccharide-binding protein-mediated interaction of lipid A from different origin with phospholipid membranes. Phys Chern Chern Phys 2000; 2:4521-4528 . 23. Viriyakosol S, Kirkland TN. A region of human CD14 required for lipopolysaccharide binding. J Bioi Chern 1995; 270:361-368. 24. McGinley MD , Narhi La. Kelley MJ et al. CDI4: physical properties and identification of an exposed site that is protected by lipopolysaccharide. J Bioi Chern 1995; 270:5213-5218. 25. Juan TS. Kelley MJ. Johnson DA et al. Soluble CD14 truncated at amino acid 152 binds lipopolysaccharide (LPS) and enables cellular response to LPS. J Bioi Chern 1995; 270:1382-1387. 26. Juan TS. Hailman E, Kelley MJ et al. Identification of a lipopolysaccharide binding domain in CD 14 between amino acids 57 and 64.J Bioi Chern 1995; 270:5219-5224. 27. Juan TS, Hailman E. Kelley MJ et al. Identification of a domain in soluble CD14 essential for lipopolysaccharide (LPS) signaling but not LPS binding . J Bioi Chern 1995; 270:17237-17242 . 28. Viriyakosol S. Kirkland TN. The N-terminal half of membrane CD14 is a functional cellular lipopolysaccharide receptor. Infect Immun. 1996; 64:653-656. 29. Stelter F. Bernheiden M. Menzel Ret al. Mutation of amino acids 39-44 of human CD14 abrogates binding of lipopolysaccharide and Escherichia coli. Eur J Biochem 1997; 243:100-109. 30. Cunningham MD , Shapiro RA. Seachord C et al. CD14 employs hydrophilic regions to 'capture" lipopolysaccharides. J Immunol 2000; 164:3255-3263. 31. Kim JI, Lee CJ. Jin MS er al. Crystal structure of CD14 and its implications for lipopolysaccharide signaling. J Bioi Chern 2005; 280:11347-11351. 32. Grunwald U, Kruger C. Schutt C. Endotoxin-neutralizing capacity of soluble CD 14 is a highly conserved specific function . Circ Shock 1993; 39:220-225. 33. Haziot A. Rong GW; Bazil V et al. Recombinant soluble CD14 inhibits LPS-induced tumor necrosis factor-alpha production by cells in whole blood. J Immunol 1994; 152:5868-5876 . 34. Kitchens RL. Thompson PA. Viriyakosol S et al. Plasma CD14 decreases monocyte responses to LPS by transferring cell-bound LPS to plasma lipoproteins. J Clin Invest 2001; 108:485-493. 35. Gutsmann T. Miiller M, Carroll SF et al. Dual role of lipopolysaccharide (LPS)-binding protein in neutralization of LPS and enhancement of LPS-induced activation of mononuclear cells. Infect Immun 2001 ; 69:6942-6950 . 36. Zweigner J, Gramm HJ. Singer OC et al. High concentrations of lipopol ysaccharide-binding protein in serum of patients with severe sepsis or septic shock inhibit the lipopolysaccharide response in human monocytes. Blood 2001; 98:3800-3808.

48

Lipid A in Cancer Therapy

37. Hamann L, Alexander C, Stamme C et al, Acute-phase concentrations of lipopolysaccharide (LPS)-binding protein inhibit innate immune cell aerivation by different LPS chemorypes via different mechanisms. Infect Immun 2005; 73:193-200. 38. Schromm AB. Lien E. Henneke P er aJ.Molecular genetic analysisof an endotoxin nonresponder mutant cell line: a point mutation in a conserved region of MD-2 abolishes endotoxin-induced signaling. J Exp Med 2001; 194:79-88. 39. Visintin A, Mazzoni A. Spitzer JA et al. Secreted MD-2 is a large polymeric protein that efficiently confers lipopolysaccharide sensitivity to Toll-like receptor 4. Proc Natl Acad Sci USA. 2001; 98:12156-12161. 40. Kennedy MN, Mullen GE. Leifer CA er al, A complex of soluble MD-2 and lipopolysaccharide serves as an activating ligand for Toll-like receptor 4. J Bioi Chern 2004; 279:34698-34704. 41. Viriyakosol S. Tobias PS. Kitchens RL et al. MD-2 binds to bacterial lipopolysaccharide. J Bioi Chern 2001; 276:38044-38051. 42. Re F, Strominger JL. Monomeric recombinant MD-2 binds toll-like receptor 4 tightly and confers lipopolysaccharide responsiveness. J Bioi Chern 2002; 277:23427-23432. 43. Pugin J. Stern-Voeffray S. Daubeuf B et al. Soluble MD -2 activity in plasma from patients with severe sepsis and septic shock. Blood 2004; 104:4071-4079. 44. Visintin A. Halmen KA, Latz E et al. Pharmacological inhibition of endotoxin responses is achieved by targeting the TLR4 coreceptor. MD-2 . J Immunol 2005 ; 175:6465 -6472. 45. Poltorak A. Ricciardi-Castagnoli P. Citterio S et al. Physical contact between lipopolysaccharide and toll-like receptor 4 revealed by genetic complementation. Proc Natl Acad Sci USA. 2000; 97:2163-2167 . 46. Lien E. Means TK. Heine H er al. Toll-like recepror 4 imparts ligand-specific recognition of bacterial lipopolysaccharide. J Clin Invest 2000; 105:497-504. 47. Akashi S. Nagai Y, Ogara H et al, Human MD -2 confers on mouse Toll-like recepror 4 species-specific lipopolysaccharide recognition . Int Immunol 2001; 13:1595-1599. 48. Muro i M. Tanamoto K. Structural regions of MD-2 that determine the agonist-antagon ist activity of lipid IVa. J Bioi Chern 2006 ; 281:5484-5491. 49. Viriyakosol S. Tobias PS, Kirkland TN. Mutational analysis of membrane and soluble forms of human MD -2. J Bioi Chern 2006 ; 281:11955-11964. 50. Galanos C, Freudenberg MA. Jay F et aJ. Immunogenic properties of lipid A. Rev Infect Dis 1984; 6:546 -552. 51. Brade L. Brandenburg K. Kuhn HM et al. The immunogenicity and antigenicity of lipid A are influenced by its physicochemical state and environment . Infect Immun 1987; 55:2636-2644. 52. Brade L, Holst O. Brade H . An artificial glycoconjugate containing the bisphosphorylated glucosamine disaccharide backbone of lipid A binds lipid A monoclonal antibodies. Infect Immun 1993; 61:4514-4517 . 53. Brade L, Engcl R, Chr ist WJ et aJ. A nonsubstituted primary hydroxyl group in position 6' of free lipid A is required for bind ing of lipid A monoclonal ant ibodies. Infect Imrnun 1997; 65:3961-3965 . 54. Galanos C, Rietschel ET, Liideritz 0 er al. Biological activities of lipid A cornplexed with bovine-serum albumin . Eur J Biochem 1972; 31:230-233. 55. Rierschel ET. Kim YB, Watson DW et al. Pyrogenicity and immunogen icity of lipid A complexed with bovine serum albumin or human serum albumin. Infect Immun Aug 1973; 8:173-177. 56. Elsbach P, Weiss J. Doerfler M et al. The bactericidal/permeability increasing protein of neutrophils is a potent antibacterial and anti-endotoxin agent in vitro and in vivo. Prog Clin Bioi Res 1994; 388:41-51. 57. Zhang G. Mann D. Tsai C. Neutralization of endotoxin in vitro and in vivo by a human laccoferrin-derived peptide. Infect Immun 1999; 67:1353-1358. 58. David SA. Balaram P. Machan VI. Characterization of the interaction of lipid A and lipopolysaccharide with human serum albumin : implications for an endotoxin carrier function for albumin. J Endotoxin Res 1995; 2:99-106. 59. He XM , Carter DC. Atomic structure and chemistry of human serum albumin . Nature 1992 ; 358:209-215. 60. Morrison DC, Jacobs DM. Binding of polymyxin B to the lipid A portion of bacterial Iipopolysaccharides. Immunochem istry 1976; 13:813-818. 61. Jurgens G. Miiller M. Garide! P et aI. Investigation into the interaction of recombinant human serum albumin with Re!ipopolysaccharide and lipid A. J Endotox in Res 2002 ; 8:115-126. 62. Gioannini TL. Zhang D. Teghanemt A er aJ. An essential role for albumin in the interaction of endo toxin with lipopolysaccharide-binding protein and sCD 14 and resultant cell activation. J Bioi Chern 2002 ; 277:47818-47825. 63. Ginsberg HN. Lipoprotein physiology. Endocrinol Merab Clin North Am 1998; 27:503-519.

Interactions between Lipid A and Serum Proteins

49

64. Van Amersfoort ES. Van Berkel TJ. Kuiper J. Receptors. mediators. and mechanisms involved in bacterial sepsis and septic shock . Clin Microbiol Rev 2003; 16:379-414. 65. Feingold KR . Funk JL . Moser AH et al. Role for circulating lipoproteins in protection from endotoxin toxicity. Infect Immun 1995; 63:2041-2046. 66. Parker TS. Levine DM . ChangJC et al. Reconstituted high-density lipoprotein neutralizes gram-negative bacteriallipopolysaccharides in human whole blood . Infect Immun 1995; 63:253 -258. 67. Baumberger C. Ulevitch RJ. Dayer JM . Modulation of endotoxic activity of lipopolysaccharide by high-density lipoprotein. Parhobiology 1991; 59:378-383. 68. Calabresi L. Franceschini G. Sirtori CR er al. Inhibition of VCAM-l expression in endothelial cells by reconstituted high density lipoproteins. Biochem Biophys Res Commun 1997; 238:61-65 . 69. Casas AT. Hubsch AP. Rogers BC ct al. Reconstituted high-density lipoprotein reduces LPS-stimulated TNFu. J Surg Res 1995; 59:544-552. 70. Munford RS. Diets chy JM . Effects of specific ant ibodies. hormones. and lipoproteins on bacterial lipopolysaccharides injected into the rat. J Infect Dis Ju11985; 152:177-184. 71. Flegel WA. Wolpl A. Mannel DN et al. Inhibition of endotoxin-induced activation of human monocytes by human lipoproteins. Infect Immun 1989; 57:2237-2245 . 72. Weinstock C , Ullrich H, Hohe Ret al. Low density lipoproteins inh ibir endotoxin activation of monocytes. Arterioscler Thromb. 1992; 12:341-347 . 73. Morel DW; DiCorieto PE, Chisolm GM . Modulation of endotoxin-induced endothelial cell toxicity by low density lipoprotein. Lab Invest 1986; 55:419-426. 74. Emancipator K, Csako G. Elin RJ. In vitro inactivation of bacterial endotoxin by human lipoproteins and apolipoproteins. Infect Immun 1992; 60 :596-601. 75. Harris HW; Rockey DC, Chau P. Chylomicrons alter the hepatic distribution and cellular response to endotoxin in rats. Hepatology 1998; 27:1341-1348. 76. Kitchens RL. Thompson PA. Munford RS et al. Acute inflammation and infection mainta in circulating phospholipid levels and enhance lipopolysaccharide binding to plasma lipoproteins. J Lipid Res 2003 ; 44 :2339-2348. 77. Sprong T. Netea MG . van der Ley P et al. Human lipoproteins have divergent neutralizing effects on E. coli LPS, N. meningitidis LPS. and complete Gram-negative bacter ia. J Lipid Res 2004 ; 45 :742-749. 78. Netea MG. van Deuren M . Kullberg BJ et al. Does the shape of lipid A determ ine the inte raction of LPS with Toll-like receptors ? Trends Immunol 2002 ; 23 :135-139. 79. Seydel U. Hawkins L. Schromm A er al. The generalized endotoxic pr inciple. Eur J Immunol 2003 ; 33:1586 -1592 . 80. Levine DM , Parker TS . Donnelly TM er al. In vivo protection against endotoxin by plasma high density lipoprotein. Proc Nat! Acad Sci U.S.A. 1993; 90:12040-12044. 81. Brandenburg K. Jurgens G. Andra J et al. Biophysical characterization of the interaction of high-density lipoprotein (HDL) with endoroxins, Eur J Biochem 2002 ; 269 :5972-5981. 82. Loppnow H . Libby P. Freudenberg M et al. Cyrokine induction by lipopolysaccharide (LPS) corresponds to lethal toxicity and is inhibited by nontoxic Rhodobacrer capsularus LPS. Infect Immun 1990; 58:3743-3750. 83. Christ W). Asano O. Robidoux AL et al. E5531. a pure endotoxin antagonist of high potency. Science 1995; 268 :80-83 . 84. Rose JR. Mullarkey MA, Christ WJ ct al. Consequences of int eraction of a lipophilic endotoxin antagon ist with plasma lipoproteins. Antimicrob Agents Chemother 2000 ; 44 :504-510. 85. Wasan KM. Strobel FW, Parrott SC ct al. Lipoprotein distribution of a novel endotoxin antagonist. E5531. in plasma from human subjects with various lipid levels. Antimicrob Agents Chemother 1999; 43 :2562 -2564 . 86. Wasan KM, Sivak O . Cote RA er al. Association of the endotoxin antagonist E5564 with high-d ensity lipoproteins in vitro : dependence on low-density and triglyceride-rich lipoprotein concentrations. Antimicrob Agents Chernother 2003 ; 47 :2796-2803. 87. Rossignol DP, Wasan KM, Choo E et al. Safety, pharmacokinetics. pharmacodynamics . and plasma lipo protein distr ibution of eritoran (E5564) during continuous int ravenous infusion into healthy volunteers . Antimicrob Agents Chernother 2004 ; 48 :3233 -3240. 88. Ma J. Liao XL, Lou B et al. Role of apolipoprotein A-I in protecting against endotoxin toxicity. Acta Bioch im Bioph ys Sin (Shanghai) 2004 ; 36:419-424. 89. Gupta H. Dai L. Datta G et al. Inhibition of lipopol ysaccharide-induced inflammatory responses by an apolipoprotein AI mimetic peptide. Circ Res 2005 ; 97:236 -243. 90. Mueller M. Lindner B. Kusumoto S er al. Aggregates are the biologically active un its of endotoxin. J BioI Chem 2004; 279 :26307-26313. 91. Freudenberg MA, Bog-Hansen TC, Back U et al. Interaction of lipopolysaccharides with plasma high-density lipoprotein in rats. Infect Immun 1980; 28:373-380.

50

LipidA in Cancer Therapy

92. Munford RS. Andersen JM. Diet schy JM. Sites of tissue binding and uptake in vivo of bacterial lipopolysaccharide-high density lipoprotein complexes: studies in the rat and squirrel monkey. J Clin Invest 1981; 68:1503-1513. 93. Freudenberg MA. Galanos C. Bacrerial lipopolysaccharides: structure . metabolism and mechanisms of action . Int Rev lmmunol1990; 6:207-221. 94. Read TE. Harris HW; Grunfdd C et al. Ch ylomicrons enhance endotoxin excretion in bile. Infect lmmun 1993; 61:3496-3502 . 95. Brandenburg K. Jiirgens G. Muller M et al. Biophysical characterization of lipopolysaccharide and lipid A inactivation by lacroferrin. BioI Chern 2001 ; 382:15-25. 96. Andr a J. Lohner K. Blondelle SE et al. Enhancement of endotoxin neutral ization by coupling of a C12-alkyl chain to a lacroferricin-derivcd peptide. Biochern J 2005 ; 385:135-143. 97. Appelmelk BJ. An YQ, Geerts M et al. Lactoferrin is a lipid A-binding protein . Infect lmmun Jun 1994; 62:2628-2632. 98. Elass-Rochard E. Roseanu A. Legrand D et al, Lacroferrin-Iipopolysaccharide int eraction: involvement of the 28-34 loop region of human lactoferrin in the high-affinity bind ing to Escherichia coli 055B5 lipopol ysaccharide. Biochem J 1995; 312:839 -845. 99. van Berkel PH . Geerts ME. van Veen HA er al. N-terminal stretch Arg2. Atg3. Atg4 and AtgS of human lacroferrin is essential for bind ing to heparin . bacterial lipopolysaccharide. human lysozyme and DNA. Biochem J 1997; 328:145-151. 100. Majerle A. Kidric J. jerala R. Enhancement of antibacterial and lipopolysaccharide binding activities of a human lacroferrin peptide fragment by the addition of acyl chain. J Ant imicrob Chernother. 2003; 51:1159-1165. 101. Brandenburg K. Koch MH. Seydel U. Biophysical characterization of lysozyme binding to LPS Re and lipid A. Eur J Biochern 1998; 258:686 -695. 102. Andra ] , Koch MHJ. Bartels Ret al. Biophysicalcharacterization of the endotoxin inactivation by NK-2. an antimicrobial peptide derived from mammalian NK-Lysin. Antimicrob Agents Chemother 2004; 48 :1593-1599. 103. Andra J. Garidd P. Majerle A et al. Biophysicalcharacterization of the interaction of Limulus polyphemus endotoxin neutralizing protein (ENP) with lipopolysaccharide. Eur J Biochem 2004; 271:2037-2046. 104. Brandenburg K. Andra J. Muller M et al. Physicochemical properties of bacterial glycopolymers. Carbohydr Res 2003 ; 338:2477-2489. 105. Elsbach P. The bactericidal/permeability-increasing protein (BPI) in ant ibacterial host defense.J Leukoc BioI 1998; 64: 14-18. 106. Tobias PS. Mathison JC . Ulevitch RJ. A family of lipopolysaccharide binding proteins involved in responses to gram-negative sepsis. ] Bioi Chern 1988: 263:134 79-13481. 107. Dentener MA . Von Asmurh EJ. Francot G] et al. Antagon istic effects of lipopolysaccharide binding protein and bactericidal/permeability-increasing protein on lipopolysaccharide-induced cytokine release by mononuclear phagocyres, Competition for binding to lipopol ysaccharide. J lmmunol 1993; 151:4258-4265. 108. Heumann D. Gallay P. Barras C er al. Control of lipopolysaccharide (LPS) bind ing and LPS-induced tumor necrosis factor secretion in human peripheral blood monoc ytes. J Immunol 1992; 148:3505-3512. 109. Gutsmann T. Haberer N. Carroll SF er al. Interaction between lipopolysaccharide (LPS). LPS-binding protein (LBP). and planar membranes. Bioi Chern 2001 ; 382:425-434 . 110. Roes S. Mumm F, Seydel U er al, Localization of the lipopolysaccharide-binding protein in phospholip id membranes by atomic force microscopy.J BioI Chern 2006 ; 281:2757-2763. Ill. Bazil V. Horejsi V, Baudys M et al. Biochemical characterization of a soluble form of the 53-kDa monocyte surface antigen. Eur. J. lmmunol. 1986; 16:1583-1589. 112. Larrick JW; Hirata M. Zheng H er al. A novel granulocyte-derived peptide with lipopolysaccharideneutraliz ing activity. J lmmunol 1994; 152:231-240. 113. Wurfd MM. Kunitake ST. Lichenstein H er al. Lipopolysaccharide (LPS)-binding protein is carried on lipoproteins and acts as a cofactor in the neutralization of LPS. J Exp Med 1994; 180:1025-1035. 114. Beamer LJ. Carroll SF, Eisenberg D. Crystal structure of human BPI and two bound phospholip ids at 2.4 angstrom resolution. Science. 1997; 276:1861 -1864. 115. Beamer LJ. Carroll SF. Eisenberg D. The BPI/LBP family of proteins : a structural analysis of conserved regions. Protein Sci 1998; 7:906-914. 116. Tobias PS. Soldau K. Iovine NM et al, Lipopolysaccharide (LPS)-binding proteins BPI and LBP form different types of complexes with LPS. J Bioi Chern 1997; 272:18682-18685. 117. Ooi CEoWeissJ. Elsbach P et al. A 25-kDa NH2-terminal fragment carries all the ant ibacterial activities of the human neutrophil 60-kD a bactericidal/permeability-increasing protein. J BioI Chern 1987; 262 :14891-14894.

Interactions between LipidA and Serum Proteins

51

118. Weiss]. Elsbach P, Shu C ct al. Human bactericidallpermeability-increasing protein and a recombinant NH2-terminal fragment cause killing of serum-resistant gram-negative bacteria in whole blood and inhibit tumor necrosis factor release induced by the bacteria. ] Clin Invest 1992; 90:1122-11230. 119. Zanetti M. Gennaro R. Romeo D. Cathelicidins: a novel protein family with a common pro region and a variable C-terminal antimicrobial domain . FEBS Leu 1995: 374:1-5. 120. Cowland ]B. Johnsen AH. Borregaard N. hCAP-18. a carhelin/pro-bactenecin-like protein of human neutrophil specific granules. FEBS Lett 1995; 368: 173-176. 121. Maim]. Sorensen O. Persson T et al. The human cationic antimicrobial protein (hCAP-18) is expressed in the epithelium of human epididymis. is present in seminal plasma at high concentrations, and is attached to spermatozoa. Infect Immun 2000; 68:4297-4302. 122. Frohm M. Gunne H. Bergman AC er al. Biochemical and antibacterial analysis of human wound and blister fluid. Eur] Biochem 1996; 237:86-92. 123. Agerberrh B. Grunewald]. Casranos-Velez E et al. Antibacterial components in bronchoalveolar lavagefluid from healthy individuals and sarcoidosis patients. Am] Respir Crit Care Med 1999; 160:283-290. 124. ]ohansson]. Gudmundsson GH . Rottenberg ME et al. Conformation-dependent antibacterial activity of the naturally occuring peptide LL-37.] BioI Chern 1998; 273:3718-3724. 125. Sorensen O. Bratt T. Johnsen AH et al. The human antibacterial cathelicidin, hCAP -18, is bound to lipoproteins in plasma.] Bioi Chern 1999; 274:22445-22451. 126. Wang Y. Agerberrh B. Lothgren A et al. Apolipoprotein A-I binds and inhibits the human antibacterial/cytotoxic peptide LL-37.] Bioi Chern 1998; 273:33115-33118. 127. Hirata M. Zhong]. Wright SC. Larrick ]W Structure and functions of endotoxin -binding pepcides derived from CAPI8. Prog Clin BioI Res 1995; 392:317-326. 128. Travis SM, Anderson NN. Forsyth WR er al. Bactericidal activity of mammalian cathelicidin-derived peptides, Infect lmmun 2000; 68:2748 -2755. 129. Chen C . Brock R. Luh F et al. The solution structure of the active domain of CAPI8-a lipopolysaccharide binding protein from rabbit leukocytes. FEBS Lett 1995; 370:46-52. 130. Turner ], Cho Y. Dinh NN er al. Activities of LL-37. a cathelin-associated antimicrobial peptide of human neurrophils, Antimicrob Agents Chemorher 1998; 42:2206-2214 . 131. Oren Z. Lerman ]C. Gudmundsson GH ct al. Structure and organization of the human antimicrobial peptide LL-37 in phospholipid membranes: relevance to the molecular basis for its non cell-selective activity. Biochem] 1999: 341:501 -513. 132. Hirata M. Shimomura Y. Yoshida M et al. Endotoxin-binding synthet ic peprides with endotoxin-neutralizing . antibacterial and anticoagulant activities. Prog Clin Bioi Res 1994; 388:147-159. 133. VanderMeer T]. Menconi M]. Zhuang] er al. Protective effects of a novel 32-amino acid C -terminal fragment of CAP18 in endotoxemic pigs. Surgery. 1995: 117:656-662. 134. Kirikae T. Hirata M. Yamasu H et al. Protective effects of a human 18-kilodalron cationic antimicrobial protein (CAP I 8)-derived peptide against murine endotoxernla. Infect Immun 1998; 66:1861-1868 . 135. Nagaoka I, Hirota S. Niyonsaba F er al. Cathelicidin family of antibacterial peptides CAP18 and CAPll inhibit the expression ofTNF-alpha by blocking the binding ofLPS to CDI4(+) cells.] Immunol2001; 167:3329-3338. 136. Nagaoka I, Hirota S, Niyonsaba F et al. Augmentation of the lipopolysaccharide-neutralizing activities of human cathelicidin CAP 18/LL -37-derived antimicrobial peprides by replacement with hydrophobic and cationic amino acid residues. Clin Diagn Lab Immunol 2002: 9:972-982. 137. Ciornei CD, Sigurdardonir T, Schrnidrchen A et al. Antimicrobial and chernoarrracrant activity. lipopolysaccharide neutralization. cyrotoxiciry, and inhibition by serum of analogs of human cathelicidin LL-37. Antimicrob Agents Chernorher 2005; 49:2845-2850. 138. Nell M], Tjabringa GS. Wafelman AR et al. Development of novel LL-37 derived antimicrobial peptides with LPS and LTA neutralizing and antimicrobial activities for therapeutic application. Peptides 2006; 27:649-660. 139. Gursmann T, Hagge SO. Larrick]W et al. Interaction of CAP IS-derived peprides with membranes made from endotoxins or phospholipids. Biophys] 2001; 80:2935-2945. 140. Gursmann T. Fix M. Larrick JW et al. Mechanisms of action of rabbit CAPI8 on monolayers and liposornes made from endotoxins or phospholipids.] Membr Bioi 2000 ; 176:223-236.

CHAPTERS

The Lipid A Receptor Kiyoshi Takeda"

T

Abstract

he lipid A receptor consists ofseveral subunits. Lipopolysaccharide binding protein (LBP) is a serum protein facilitating association oflipid A with CD 14. The Lipid A-LBP-CD 14 complex is further delivered to Toll-like receptor 4 (TLR4), which is essential for lipid A-mediated cellular activation. TLR4 associates with MD-2, which is required for surface expression and recognition oflipid A.In addition, the RP 10S-MD-1 complex mediates TLR4-mediated response in B cells and negatively regulates TLR4-mediated response in dendritic cells.Thus , lipid A receptor's function varies among cell populations.

Introduction

Lipopolysaccharide (LPS) is a major component of the outer membrane of Gram-negative bacteria and is a potent activator ofinnate immune cells including macrophages and dendritic cells. LPS consists of hydrophilic polysaccharides and hydrophobic lipid A. I Polysaccharide is divided into a core oligosaccharide and a polysaccharide called an O-antigen. Lipid A is a glucosamine-based phospholipid and is the core structure inducing activation of innate immunity. Excessive lipid A-mediated activation ofinnate immunity induces a harmful systemic disorder known as endotoxin shock. Therefore, identification ofthe lipid A receptor has long been requested. Recent advances in our understanding ofToll-like receptor (TLR)-mediated recognition ofmicrobial components helped to reveal the lipid A receptor complex.

Components ofthe Lipid A Receptor

LBPandCD14

LPS binding protein (LBP) was first identified as a molecule that associated with lipid A. LBP is a member ofa family oflipid-binding proteins that act as lipid-transport proteins in some cases.2•3The generation ofLBP-deficient mice revealed a nonredundant role for LBP in response to LPS.4 Formation of LPS/LBP complexes triggers association of these complexes with another LPS binding molecule, CD14. 5•6 CD14 is a glycosylphosphatidyl-inositol (GPI) anchored protein, which is preferentially expressed on the surface of mature myeloid cells. Soluble forms of CD 14 are also produced through escape from the GPI anchoring and proteolytic cleavage of the membrane-bound CD 14. Importance of CD 14 in LPS-mediated response has been demonstrated in CD 14-deficient mice , showing a reduced response to LPS.?Thus, LPS first binds LBP and then the LPS/LBP complex is transferred to CD14. Although CD14 is a membrane-anchored protein, it has no cytoplasmic region that would be required for cellular activation. Therefore, the LPS-LBP-CD 14 complex requires an additional receptor to transduce the signal from the membrane into the cytoplasm. However, this molecule had long been unrevealed.

*KiyoshiTakeda-Department of Molecular Genetics, Medical Institute of Bioregulation, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, 812-8582, Japan. Email: [email protected]

LipidA in Cancer Therapy, edited by jean-Francois ]eannin. ©2009 Landes Bioscience and Springer Science+Business Media.

54

Lipid A in Cancer Therapy

TLR4

A due to identify the signal transducing receptor ofLPS (lipid A) was given in a genetic study of

Drosophila. A mutant flylacking the transmembrane receptor Toll was shown to be highly vulner-

able to fungal infection.t One year later, a mammalian Toll receptor (which was later identified as TLR4) was shown to induce expression ofgenes involved in inflammatory responses." Subsequent studies revealed that there were several Toll receptors in mammals and these were designated Toll-like receptors (TLRs). TLRs in mammals were shown to recognize microbial components that are not present in mammals but are conserved between microorganisms, thereby detecting invasion ofpathogens such as bacteria, fungi, protozoa and viruses.to TLRs are Type I transmembrane proteins bearing leucine-rich repeats (LRRs) in the extracellular portion and contain the Toll/IL-l receptor (TIR) domain in the cytoplasmic portion. The TIR domain ofTLRs shows a high similarity with the cytoplasmic region ofthe IL-l receptor family and is essential for activation ofsignaling pathways leading to induction ofgene expression. I I Two mouse strains, C3H/He] and CS7BLlO/ScCr, are known to be hyporesponsive to lipid A and to be sensitive to Gram-negative bacterial infection. In 1998, the gene responsible for hyporesponsiveness to LPS was identified and C3H/He] and CS7BLl O/ScCr strains were shown to have mutations in the Tlr4 gene." C3H/He] mice showed a point mutation in the cytoplasmic region of the Tlr4 gene, resulting in an amino acid change from proline to histidine. This mutation resulted in defective TLR4-mediated signaling and had a dominant negative effect on LPSmediated responses," On the other hand, CS7BLlO/ScCr mice had a null mutation in the Tlr4 gene." TLR4 knockout mice further revealed the essential role of TLR4 in lipid A-mediated response'! i.e., TLR4 was an essential component in lipid A-mediated activation of signaling pathways. Lipid A ofLeptospira interrogam show some variability compared with enterobacterial lipid A, which is recognized by TLR4.14 Differential proinflammatory capacity of leptospiral lipid A is attributable to the degree ofacylation or the length ofacyl chains. In addition, leptospiral lipid A has been shown to be recognized by TLR2 and TLRl/TLR2 in human cells.15,16 TLR4 is implicated in the recognition of several ligands in addition to lipid A. Taxol, a diterpene pu rified from the bark of the Western yew (Taxus brevifllia), has been shown to activate mouse TLR4. but not human TLR4. 17 TLR4 is also involved in the recognition of viral products such as the fusion protein of respiratory syncytial virus (RSV) and the envelope glycoprotein ofmouse mammary tumor virus. 18,19 Furthermore. TLR4 has been shown to playa role in the recognition ofendogenous ligands such as heat shock proteins (HSP60 and HSP70). the extra domain A of fibronecrins, oligo saccharides of hyaluronic acid, heparan sulfate and fibrinogen." However. all of these endogenous ligands have to exist at high concentrations to activate TLR4. In addition. it has been shown that contaminating LPS in HSP70 preparations conferred the ability to activate TLR4. 21 LPS is a very potent immune-stimulator, so TLR4 can be activated by a very small amount ofLPS contaminating these endogenous ligand preparations. Therefore. more careful experiments are required before we can conclude that TLR4 recognizes these endogenous ligands.

MD-2

MD-2 was identified as a molecule that associated with the extracellular portion ofTLR4. 22 Expression ofboth MD-2 and TLR4. but not ofTLR4 alone resulted in lipid A-induced NF-KB activation in lipid A nonresponsive Ba/F3 cells. indicating that MD-2 functionally associated with TLR4. Physical association ofMD-2 and TLR4 on mouse peritoneal macrophages was also shown using a monoclonal antibody against TLR4/MD-2 complexes." Importance ofMD-2 in the response to lipid A was further demonstrated in genetic studies . The Chinese hamster ovary (CHO) cell line that showed an impaired response to lipid A. was shown to contain a mutation in the MD-2 gene." MD-2-deficient mice displayed severely impaired responses to lipid A and their phenotype was very similar to TLR4-deficient mice." Analysis of MD-2-deficient mice

1ht Lipid A Receptor

t

55

s

TLR4

RP105

MD-2

[ LBP 1 CD14

Figure 1. The lipid A receptor components.

further demonstrated that TLR4 was not expressed on the cell surface, although it was found in the Golgi apparatus . These findings indicated that MD -2 was required for surface expression of TLR4. 25 MD-2 associateswith TLR4 in the endoplasmic reticulum, then TLR4-MD-2 complexes might be delivered to the cell surface through the Golgi appararus.F'Ihc eukaryotic endoplasmic reticulum (ER) chaperone gp96, which is thought to be indispensable for cell survival, has been shown to be required for surface expression ofTLR4, since TLR4 is retained in the intracellular portion ofgp96-deficient cells," MD-2 also seems to mediate recognition of lipid A. Physical proximity between CD14 and TLR4-MD-2 complexes was induced in response to lipid A srimularion .P-" Introduction of mutations in N-glycosylat ion sites of MD-2 did not induce changes in surface expression of TLR4-MD-2 complexes, but led to arrest of interaction with LPS.30 In addition, a report indicated that MD-2 directl y regulated species-specific recognition ofthe lipid moiet y by TLR4.J1 On the other hand, another report indicated a direct interaction oflipid A with the TLR4-MD-2 complex rather than with MD-2 alone.F'Iherefore, the TLR4-MD-2 complex is the most stable and important receptor for the recognition oflipid A (Fig. 1).

56

Lipid A in Cancer Therapy

TLR4 0.2

RP105 MD-1

TLR4 MD-2

RP105 MD-1

Dendritic cells

B cells Figure 2. Role of RP105-MD-l in B cells and dendritic cells.

RPI05 andMD-I

In addition to the molecules described so far, RP lOS and MD-1 are involved in lipid A-mediated responses, especially in B cells. RP 1 contains an extracellular LRR domain that is structurally related to domains found in TLRs. However, unlike TLRs, RP10S only has a short cytoplasmic tail and is preferentially expressed on B cells.33 RP lOS-deficient mice showed a severely impaired response to LPS in B cells, indicating that RP10S is an essential component in the recognition ofLPS in B cells.t' Like TLR4. RPIOS was shown to associate with MD-!, which is structurally related to MD-2.35 Similarly to RP lOS-deficient mice, MD-1-deficient mice showed impairment in LPS -induced B-cell proliferation. antibody production and CD86 up-regulation." Furthermore, surface expression of RP lOS was abolished in MD-1-deficient B cells, indicating that MD-1 was essential in LPS-mediated response and in surface expression ofRP10S in B cells. Thus. the RP10S-MD-1 complex is indispensable to LPS-mediated responses in B cells. In contrast to its function in B cells. the RP 10S-MD-1 complex has a quite opposite function in dendritic cells." RP 1OS and TLR4 are expressed in similar cell populations. RP 1OS-MD-1 was shown to associate with the TLR4-MD-2 complex and to inhibit TLR4-mediated response in HEK293 cells. Furthermore, dendritic cellsderived from RP lOS-deficient mice showed enhanced response to lipid A in terms of cytokine production. Thus, in dendritic cells, the RPlOS -MD-1 complex has an inhibitory function on lipid A-mediated responses (Fig. 2).

as

Conclusion

Identification ofTLRs allowed elucidation ofthe lipid A receptor. TLR4 has now been established as a key component ofthe lipid A receptor and several TLR4-associating molecules mediating lipid A-induced responses have been identified. Targeting these lipid A receptor components would help developing therapeutic agents for endotoxin shock caused by lipid A.

Acknowledgments

I would like to thank M. Kurata for her secretarial assistance and M. Matsumoto for gathering the references.

The LipidA Receptor

References

57

1. Raetz CR , Whitfield C. Lipopolysaccharide endotoxins. Annu Rev Biochem 2002; 71:635-700 . 2. Tobias PS, Soldan K, Ulevitch RJ. Isolation of a lipopolysaccharide-binding acute phase reactant from rabbit serum. J Exp Med 1986; 164(3):777-793. 3. Schumann RR, Leong SR, Flaggs GW et al. Structure and function of lipopolysaccharide binding protein. Science 1990; 249(4975):1429-1431. 4. Wurfel MM , Monks BG, Ingalls RR et al, Targeted deletion of the lipopolysaccharide (LPS)-binding protein gene leads to profound suppression of LPS responses ex vivo, whereas in vivo responses remain intact. J Exp Med 1997; 186(12) :2051-2056 . 5. Wright SD. Ramos RA. Tobias PS et al, CDI4. a receptor for complexes of Iipopolysaccharide (LPS) and LPS binding protein. Science 1990; 249(4975):1431-1433. 6. Tobias PS. Soldan K. Gegner JA er al, Lipopolysaccharide binding protein-mediated complexation of lipopolysaccharide with soluble CDI4. J Bioi Chern 1995; 270(18) :10482-10488 . 7. Moore KJ andersson LP. Ingalls RR et al, Divergent response to LPS and bacteria in CDl4-deficient murine macrophages, J Immunol 2000; 165(8):4272-4280. 8. Lernaitre B. Nicolas E. Michaut L er al, The dorsoventral regulatory gene cassette spatzle/Toll/cactus controls the potent antifungal response in Drosophila adults. Cell 1996; 86(6) :973-983. 9. Medzhirov R. Preston-Hurlburt P, Janeway CA , Jr. A human homologue of the Drosophila Toll protein signals activation of adaptive immunity. Nature 1997; 388(6640) :394-397. 10. Takeda K. Kaisho T. Akira S. Toll-like receptors. Annu Rev Immunol 2003; 21:335-376. 11. Akira S, Takeda K. Toll-like receptor signalling. Nat Rev Immunol 2004 ; 4(7) :499-511. 12. Poltorak A. He X. Smirnova I et al, Defective LPS signaling in C3H/HeJ and C57BLIlOScCr mice: mutations in Tlr4 gene. Science 1998; 282(5396) :2085-2088 . 13. Hoshino K, Takeuchi 0, Kawai T er al. Cutting edge: Toll-like receptor 4 (TLR4)-deficient mice are hyporesponsive to lipopolysaccharide: evidence for TLR4 as the Lps gene product. J Immunol 1999; 162(7) :3749-3752. 14. Netea MG . van Deuren M. Kullberg BJ ec al. Does the shape of lipid A determine the interaction of LPS with Toll-like receptors? Trends Immunol 2002; 23(3) :135-139. 15. Werts C, Tapping RI. Mathison JC er al. Leptospiral lipopolysaccharide activates cells through a TLR2-dependent mechanism. Nat Immunol 2001 ; 2(4):346-352. 16. Nahori MA, Foumle-Amazouz E. ~e-Gewirth NS et al, Differential TLR recognition of leptospiral lipid A and lipopolysaccharide in murine and human cells. J Immunol2005; 175(9):6022-6031. 17. Byrd-Leifer CA. Block EF. Takeda K er al. The role of MyD88 and TLR4 in the LPS-mimetic activity ofTaxol. Eur J Immunol 2001; 31(8) :2448-2457 . 18. Kurt-Jones EA, Popova L. Kwinn L et al. Pattern recognition receptors TLR4 and CD14 mediate response to respiratory syncytial virus. Nat Immunol 2000 ; 1(5):398-401. 19. Rassa lC. Meyers JL. Zhang Y er al. Murine retroviruses activate B cells via interaction with roll-like receptor 4. Proc Nat! Acad Sci USA 2002;99(4) :2281-2286 . 20. Takeda K, Akira S. Toll-like receprors in innate immunity. Inc Immunol 2005 ; 17(1):1-14. 21. Gao B. Tsan ME Endotoxin contamination in recombinant human heat shock protein 70 (Hsp70) preparation is responsible for the induction of tumor necrosis factor alpha releaseby murine macrophages, J Bioi Chern 2003 ; 278(1) :174-179. 22. Shimazu R, Akashi S. Ogata H et al. MD-2, a molecule that confers lipopolysaccharide responsiveness on Toll-like recepror 4. J Exp Med 1999; 189(11):1777-1782 . 23. Akashi S. Shimazu R, Ogata H et al, Cutting edge: cell surface expression and lipopolysaccharide signaling via the toll-like receptor 4-MD -2 complex on mouse peritoneal macrophages, J Immunol 2000 ; 164(7):3471-3475. 24. Schromm AB. Lien E. Henneke P et al, Molecular genetic analysisof an endotoxin nonresponder mutant cell line: a point mutation in a conserved region of MD-2 abolishes endotoxin-induced signaling. J Exp Med 2001; 194(1):79-88. 25. Nagai Y.Akashi S. Nagafuku M et al, Essential role ofMD-2 in LPS responsiveness and TLR4 distribution. Nat Immunol 2002; 3(7) :667-672. 26. Visintin A. Mazzoni A. Spitzer JA et al. Secreted MD-2 is a large polym eric protein that efficiently confers lipopolysaccharide sensitivity ro Toll-like receptor 4. Proc Nat! Acad Sci USA 2001; 98(21 ):12156-12161. 27. Randow F, Seed B. Endoplasmic reticulum chaperone gp96 is required for innate immunity but not cell viability. Nat Cell Bioi 2001; 3(10):891-896. 28. Jiang Q. Akashi S, Miyake K er al, Lipopolysaccharide induces physical proximity between CDI4 and toll-like receptor 4 (TLR4) prior to nuclear translocation of NF-kappa B. J Immunol 2000 ; 165(7) :3541-3544.

58

Lipid A in Cancer Therapy

29. da Silva Correia J. Soldau K. Christen U et al. Lipopolysaccharide is in close proximity to each of the proteins in its membrane receptor complex transfer from CD14 to TLR4 and MD -2. ] Bioi Chern 2001; 276(24):21129-21135. 30. da Silva Correia J. Ulevicch RJ. MD-2 and TLR4 N-linked glycosylations are important for a functional lipopolysaccharide receptor. ] Bioi Chern 2002; 277(3) :1845-1854. 31. Akashi S. Nagai Y, Ogata H et al. Human MD-2 confers on mouse Toll-like receptor 4 species-specific lipopolysaccharide recognition . Int Immunol 2001; 13(12) :1595 -1599. 32. Akashi S, Saitoh S. Wakabayashi Y et al. Lipopolysaccharide interaction with cell surface Toll-like receptor 4-MD-2: higher affinity than that with MD-2 or CDI4. ] Exp Med 2003 ; 198(7):1035-1042 . 33. Miyake K, Yamashita Y. Ogata M et al. RPI05. a novel B cell surface molecule implicated in B cell activation. is a member of the leucine-rich repeat protein family.] Immuno11995; 154(7):3333-3340. 34. Ogata H, Su 1. Miyake K er al. The toll-like receptor protein RP105 regulates lipopolysaccharide signaling in B cells. ] Exp Med 2000 ; 192(1) :23-29. 35. Miyake K. Shimazu R. Kondo] et al. Mouse MD -l. a molecule that is physically associated with RPI05 and positively regulates its expression. ] Immuno11998; 161(3):1348-1353. 36. Nagai Y, Shimazu R, Ogata H er al. Requirement for MD -l in cell surface expression ofRPI05 /CD180 and B-cell responsiveness to lipopolysaccharide. Blood 2002; 99(5):1699-1705. 37. Divanovic S, Trompette A, Atabani SF et al. Negative regulation of Toll-like receptor 4 signaling by the Toll-like receptor homolog RPI05. Nat Immunol 2005; 6(6) :571-578.

CHAPTER 6

Lipid A Receptor TLR4-Mediated Signaling Pathways Masahiro Yamamoto and Shizuo Akira" Abstract

L

ipid A is a strongactivatorofmonoeytesto rdeaseimmunestimulators suchasproinRammatory eytokines. Overproduction of inflammatory cytokines such as TNF and 1L-6 is known to causesepticshock that frequentlyleadsto multipleorgan failureand finallyto death. In recent years,Lipid A hasalsobeen recognizedbyaToll-likereceptor,TLR4. Activation ofTLR4 byLPSor Lipid A triggerssignaltransduction viathe cytoplasmicdomain calledthe Toll/1L-1 Receptor (TIR) domain. IntracellularT1Rdomain-containingadaptor moleculesareinvolvedin the TLR4-mediated signalingpathways.Moreover,a subsetofLPS-inducible genesisregulated in two stepsby the induciblenuclearprotein .Additionally, the TLR4-mediated activation of signalingcascades iselaborately down-regulated by a number ofnegativeregulators. In this chapter, we discuss the mechanisms of the activation or de-activationprogram mediated by the Lipid A receptor TLR4 .

Introduction

Lipid A is the active component ofLPS and has properties similar to those ofLPS regarding the facilitation of cellular immune responses.' LPS has long been known to be a strong inducer of proinflammatory cytokines including TNF, which is mainly produced from macrophages and plays a leading role in endotoxin septic shock.' CD14 is the glycosylphosphoinositol-tethered leucine-rich protein on the surfaceofmonocytes and was the first identified cellular LPS receptor in association with a plasma LPS binding protein LBP.3In addition, in terms of signal transduction, LPS stimulates activation ofNF-KB and MAP kinases in cultured cells." Considering that both CD 14 and LBP lack a cytoplasmic domain, LPS is expected to tran smit the signal through a cell surface receptor(s) that possiblyinteracts with CD14 and LBP. Drosophila deficient in Toll,originally reported to be involvedin dorso-ventral formation and whose cytoplasmic domain shows striking homology with that of interleukin-I (1L-1) receptor, displayssusceptibilityto fugalinfecnons.' Subsequently, mammalianToll,known asToll-like receptor (TLR) , has been shown to be involved in NF-KB activation in cultured cells," Independently, positional cloning of the responsible gene for an LPS-resistant mouse strain, C3H/He], led to identification of a mutation in the tlr4 gene? Moreover, TLR4-deficient mice were completely defectivein LPS response.f These results indicate that TLR4 is the receptor for LPS that mediates cellular immune responses. Since the cytoplasmic domain ofTLR displayshomology with that ofIL-1R, this portion is known as the Toll/1L-1R (T1R) domain . The T1R-mediated signaling molecules for TLRs are mostly shared with those for 1L-1R.~ Among them, the roles ofMyD88, 1RAK family members and TRAF6 in the TLR4-mediated signaling pathways and their responses will be discussed in *CorrespondingAuthor: 5hizuoAkira-Department of Host Defense, Research Institute for Microbial Diseases, Osaka University, Osaka 565-0871 , Japan. Email: saklraeblken.osaka-u.ac.jp

LipidA in Cancer Therapy, edited byjean- Francois]eannin. ©2009 Landes Bioscience and Springer Science+Business Media .

60

Cell-Pre« Expression

this chapter. As well, we will discuss the immune responses specificallyfound in TLR4-, but not in IL-I R-, mediated pathways.

The MyD88-Dependent and MyD88-Independent Pathways

MyD88 is a signal transducer of the IL-IR-mediated signaling pathway." Indeed, MyD88-deficient mice show complete defects in not only IL-IR, but also in the family member IL-18R-mediated immune responses such as stimulus-dependent induction ofacute phase proteins in the liver and the activation ofNF-KB and MAP kinases.!' Since proinflammatory cytokine production is stimulated by LPS as well as IL-I, the response to LPS in MyD88-deficient cells has been examined; revealing that MyD88 is essential for LPS-induced pro inflammatory cytokine production in macrophages." In sharp contrast to the almost complete defects in the LPS-induced proinflammatory cytokine production, LPS-mediated activation ofNF-KB and MAP kinases is observed in MyD88-deficient cells. Further detailed analysis has revealed that, relative to wild-type cells, the activation ofNF-KB and MAP kinases in MyD88 -deficient cells is through delayed kinetics.P'Ihese results suggest that, although the IL-I R-mediated signaling is very MyD88-dependent, TLR4 stimulates not only the MyD88-dependent signaling, which culminates in proinflammatory cytokine production and splenocyte activation, but also the MyD88-independent pathway(s) that activates NF-KB and MAP kinases. Further, subtrac tion analyses demonstrated that interferon (IFN)-f3 and IFN-inducible genes such as IP-lO, RANTES and ISG-S4 are induced in response to LPS in MyD88-deficient cells." A family of transcription factors called IFN regulatory factor (IRF) has been shown to be implicated in type I interferon." In the TLR4-mediated MyD88-independent pathway, IRF3 is shown to participate in the expression ofIFN and IFN-inducible genes." Among the TLR family members, TLR3 and TLR4 stimulate MyD88-independent pathways, resulting in the expressions of IFN-f3 and IFN-inducible genes." Considering that a TIR domain-containing adaptor molecule MyD88 mediates the MyD88-dependent pathways in IL-I Rand TLR signaling, another TIR domain-containingadaptor(s) other than MyD88 may participate in the MyD88 -independent pathway in the TLR4 and TLR3-mediated signaling pathways (Fig. I) .

TRIF: The TIR Domain-Containing Signal Transducer for the MyD88-Independent Pathway

Another TIR domain-containing adaptor molecule has been identified through data base analysis and a yeast two-hybrid system using the cytoplasmic TIR domain ofTLR3 as bait. 17•18 The NF -KB-dependent promoter is activated by ectopic expression of the molecule as well as by MyD88. Moreover, in sharp contrast to MyD88, overexpression of the molecule strongly up-regulates the IFN-f3promoter; therefore, this molecule wasnamed the TIRdomain-containing adaptor molecule inducing IFN-f3 (TRIF) (alsoknown asTICAM-I). In vitro biochemicalanalysis has demonstrated that TRIF is associated with the full length or the TIR domain ofTLR3, but not TLR4. In addition, TRIF is bound with IRF3 .The deletion mutant ofTRIF only possessing the TIR domain, but not that of MyD88, inhibits TLR3-mediated activation of the IFN-f3 promoter, indicating that TRIF is involved in the TLR3-induced MyD88-independent signal transduction. TRIF-deficient and TRIF mutant mice (Lps2 mice) were generated and examined to ascertain whether TRIF is involvedin TLR4- as well as TLR3-mediated MyD88-independent pathways.19.20 TLR3-mediated expression ofIFN-f3 and IFN-inducible genes is completely abrogated in TRIF-deficient mice. Moreover, TLR4-mediated MyD88-independent IFN induction and activation ofIRF3 as well as NF-KB and MAP kinases are not observed in TRIF-deficient or-mutant mice, suggestingthat TRIF is the molecule responsible for TLR3- and TLR4-mediated MyD88-independent pathways. In addition to defects in the MyD88-independent expression of IFN -f3 and IFN-inducible genes, TRIF-deficient mice also show impairments in LPS-induced production of proinflammatory cytokines and proliferation of splenocytes, both of which have been considered to be the MyD88-dependent immune responses.W" Considering that, except for TLR3 and TLR4,

LipidA Receptor TLR4-Mediaud Signaling Pathways

61

Figure 1. MyD88-dependent and MyD88-independent pathways in TLR4 signaling. The MyD88-dependent pathway activates an early phase of NF-KB and MAP kinases via TIRAP, MyD88, IRAK-1/-4 and TRAF6. TheMyD88-independent pathwayleads to IRF3 and latephase activation of NF-KB and MAP kinases via TRAM and TRIF and apoptosis by PKR. the immune responses by the rest of the TLR family members are MyD88-dependent but TRIF-independent, TLR4 alone requires both adaptors for mediating full levelsof the immune responses. However, the detailed mechanisms remain unknown. In vitro analysis using a seriesofdeletion mutants ofTRIF revealed that both the N-terminal and C-terminal portion ofTRIF potentiate NF-KB activation.'? In contrast, overexpression of the N-terminal, but not the C vterminal, portion ofTRIF only activates the IFN-13-promoter, suggesting that TRIF may interact with distinct sets of signaling molecules in the N- or C-terminal portions. Subsequent studies have demonstrated that the N-terminal portion of TRIF is associated with TRAF6 through its TRAF binding sites." Moreover, RIPl , which plays an essential role in TNFR-mediated NF-KB activation, is associated with the C-terminal portion ofTRIF through its RHIM domain.22 These results indicate that RIP 1 and TRAF6 are signal transducers for TRIF-mediated NF-KB activation . Another TRIF-interacting molecule, a TANK-homologous protein called NAPl, is also reportedly involved in the TLR3-mediated MyD88-independent parhway." However, its involvement in TLR4 signaling is unknown. In addition to cellular activation by TRIF, overexpression ofTRIF is known to induce apoptosis in cultured cells.24•25 The pro-apopeotic signals from TRIF are mediated by PKR, which was originally identified as a double -stranded RNA -induced inhibition of protein synthesis and subsequent cell death." Considering that macrophage apoptosis by bacterial pathogens is partially TLR4-dependent, TRIF-mediated PKR activation may be required for bacteria -induced apoptosis (Fig. 1).

62

Cdl-Free Expression

TIRAP and TRAM: Another Two TIR Domain-Containing Molecules

To date. the TIR domain-containing adaptors comprise 5 family members in which TlRAP (also known as Mal) and TRAM (also known as TIRP or TICAM-2) are shown to be involved in TLR4-mediated signaling pathways in addition to MyD88 and TRIP. TlRAP was the second TIR domain-containing molecule to be identified and was originally considered to participate in the MyD88-independent pathways.27,28 However. LPS-induced expression ofIFN-~ and IFN-inducible genesis normally observed in TlRAP-deficient macrophages.P" Moreover. LPS-mediated up-regulation of surface molecules on dendritic cells is comparable between wild-type and TIRAP-deficient mice. On the other hand. TlRAP-deficient macrophages show defective proinflammatory cytokine production with either TLR2 or TLR4 stimulation. Moreover. both TLR2- and TLR4- mediated proliferation ofsplenocytes is severelyimpaired in TlRAP-deficient mice. demonstrating that TIRAP is not specific to TLR4 signaling and does not participate in the MyD88-independent pathway. but rather TlRAP plays a crucial role in the MyD88-dependent signaling pathway shared by TLR2 and TLR4. TRAM was the fourth adaptor possessingTIR domain to be identifiedY-33 Ectopic expression ofTRAM leads to activation ofNF-KB and the IFN-~ promoter. In addition. TRAM isassociated with TRIF and the TIR domain ofTLR4. indicating a possible function in the TLR4-mediated MyD88-independent pathway. TRAM-deficient macrophages display defective proinflammatory cytokine production in response to LPS. but not to other TLR ligands. In addition. LPS-induced expression of IFN-inducible genes and activation of IRF3 are abolished in TRAM-deficient mice. In sharp contrast. TLR3-mediated MyD88-independent immune responses are intact in TRAM-deficient mice. Taken together. TRAM can be seen to be specifically involved in the TLR4-mediated MyD88-independent pathway (Fig. 1). Among TLR family members. TLR4 alone utilizes four TIR domain-containing molecules to trigger the activation signal downstream to the nucleus. In TIR domain-containing adaptor molecules. MyD88 possessesa death domain in the N-terminal portion. through which MyD88 trans duces the signal to downstream molecules such as lRAK family members and TRAF6. TRIF also interacts with TRAF6. PKR. TBK1. IRF-3 and RIP 1 in either the N- or C-terminal portion. suggesting that MyD88 and TRIF are signal transducers rather than adaptor molecules." On the other hand. TlRAP and TRAM are relatively short length proteins and their motifs other than their TIR domain have not been well characterized; therefore. the significance oftheir existence in terms ofTLR4-mediated signal transduction remains enigmatic. Recently.the possiblesignificanceofTlRAP and TRAM hasbeen demonstrated in cellbiological studies. TRAM has been shown to localize in the plasma membrane and the Golgi apparatus. where it co-localizeswith TLR4. Membrane localization ofTRAM is the result ofmyrisroylarion and TRAM contains a putative myristoylation sequence in the Nvrerminal portion." A mutant form of TRAM possessing a defective myristoyiation sire does not activate IRF-3 and NF -KB. Moreover. a member of the PKC family. PKC-E. promotes LPS-induced phosphorylation of TRAM on serine-16.36These results clearlyindicate that membrane localization ofTRAM through the myristoylation sequence plays an important role in the TLR4-mediated MyD88-independent signal transduction. On the other hand. TlRAP is reported to reside on membranes that are shuttled between the plasma membrane and endosomes by an ADP ribosylation factor 6 (ARF6) dependent process. Moreover. TlRAP contains a phosphatidylinosiroI4,s-bisphosphate (PIP2) binding domain that mediates TlRAP recruitment to membranes and is required for TLR4 signaling. leading to the recruitment ofMyD88 to the TIR domain ofTLR4.37 The result demonstrates that TlRAP acts as an adaptor rather than a signal transducer of the TLR4-mediated MyD88-dependent pathway.

Negative Regulation ofLPS-Induced Signaling Pathways

Macrophages exposed to LPS show reduced responses to a second stimulation with LPS;38 a response that is termed LPS tolerance and is considered an important negative regulatory biological response for the limitation of systemic inflammation. As discussed above. TLR4

Lipid A Receptor TLR4-MediatedSignalingPathways

63

signals two distinct pathways. Downstream ofTLR4, TIRAP and MyD88 play essential roles in the subsequent IRAK and TRAF6 activation; thus, allowing activation ofNF-KB and MAP kinases in the MyD88-dependent pathways. On the other hand, the MyD88-independent pathway is activated via TRAM, TRIF, TBK1 and IRF3 , leading to mainly IFN-~ induction. Accumulating lines ofevidencedemonstrate that there are negative regulators in the LPS-induced MyD88-dependent and MyD88 -independent pathways." For instance, at the level of TLR4, a soluble form ofTLR4 called sTLR4 that lacks the intracellular TIR domain down-regulates the LPS response in a dominant-negative fashion." The TIR domain ofTLR4 is also a target of ubiquitin-proteasome dependent protein destruction pathways by an E3 ubiquitln-protein ligase member, TRIAD3A.41 Overexpression or knockdown ofTRIAD3A results in the reduction or enhancement of TLR4 expression, indicating that TRIAD3A may act a negative regulator in certain circumstances . Activation ofTIRAP and MyD88 are also negatively regulated. TIRAP is a target ofSOCS1 , which is known to be important in the suppression of cytokine signaling.f SOCS1 -ddicient mice are highly susceptible to LPS-induced septic shock, while SOCS1-deficient cells display augmented LPS response compared with wild-type ones.43.44 In vitro analysis demonstrated that SOCS 1 promotes tagging ubiquitin chains with TIRAP, leading to the degradation. Two MyD88 cDNA species, long and short forms, have been identified and encode full-length MyD88 and MyD88s that lacks the interdomain between the death and the TIR domains." Although MyD88s mRNA is detected in spleen and , to a lesser degree, in brain in unstimulated conditions, LPS stimulation up-regulates the expression in human macrophages. Overexpression ofMyD88s inhibits LPS- as well as IL-1-induced NF -KBactivation by facilitating the formation of the heterodimer with full-length MyD88 and negatively regulates the activation ofIRAKs. As well, recent studies demonstrate that IRF-4 is associated with MyD88 for negative regulation and that it is essential for LPS-tolerance. 46 The IRAKs consist of four family members, IRAK-1 , IRAK-2 , IRAK-4 and IRAK-MY IRAK-M-deficient mice show enhanced proinflammatory cytokine production in response to various TLR ligands including LPS.48 Moreover, LPS tolerance is significantly reduced in IRAK-M-deficient mice, indicating the negatively regulatory role in the MyD88-dependent pathway. IRAK-2 possesses 4 different splicing variants termed IRAK-2a, -Zb, -2c and _2d.49

Overexpressions of IRAK-2c and -2d inhibit LPS-induced NF-KB activation, indicating a possible negative effect on LPS-induced immune responses . IRAK-1 is inhibited through degradation by TOLLIP interaction, which is reported to down -regulate the level ofIRAK1 autophosphorylation in vitro." Thus, IRAK family members negatively regulate themselves, or are so by their interaction partners. TRAF6 is an E3 ubiquitin ligase that plays essential role in TLR-mediated signaling pathways." Downstream ofTRAF6, ubiquitination is positively required for the activation ofNF-KB and MAP kinases via a MAP3 kinase, TAK1.52.53 The polyubiquitin chain on TRAF6 is removed by A20, which isinitially identified as a TNF-induced zinc-finger protein that suppressesNF -KB activation in the TNNR-mediated signaling pathway." In addition to increased sensitivity to TNF, A20-deficient mice are extremely susceptible to LPS-induced septic shock , indicating the important function of LPS tolerance. Moreover, A20-deficient cells demonstrate augmented polyubiquitin chains of TRAF6. 55 A20 also negatively regulates the MyD88 -independent expression of IFN-~ by interaction with IRF -3,56 indicating that A20 is implicated in the negative regulation of both the MyD88-dependent and the MyD88-independent pathways in TLR4-mediated signaling. Although the precise mechanism remains uncertain, PI3K is a negative regulator of TLR4-mediated immune responses. Mice deficient in a p8S regulatory subunit ofPI3K show enhanced TLR signaling and IL-12 synthesis from dend ritic cells, indicating that PI3K might suppress MAP kinases or NF-KB in the pathwaysy ·58 Thus, TLR4-mediated signals are negatively regulated at many levelsof the various pathways

(Fig. 2).

64

Cdl-Pre« Expression

Figure 2. Negative Regulators of TLR4-Mediated Signaling Pathways. Negative regulators such as sTLR4, Triad3A, MyD88s, IRAK-2c, -2d, -M, Tollip, socs-i and A20 inhibit TLR4 signaling at multiple levels.

Two-Step Gene Induction Program in TLR4-Mediated Immune Responses

After LPS stimulation, TLR4 activates MyD88-dependent and TRIF-dependent signaling pathways at early and late time points. respectively. Regardless of the fact that both lead to activation of NF-KB and MAP kinases, a subset of TLR4-dependent genes is only inducedin the MyD88-dependent pathway." The MyD88-dependent, but not the MyD88-independent, pathwayfacilitates mRNA stability through the 3'-UTR portion.60,61 IKBt (also known as INAP or MAIL) is a genethat is specifically induced by the MyD88-dependent pathway62-64 and is a member of the IKB family of proteins, amongwhich Bcl-3 shows the highest homologywith IKBt. Unlike other IKB family members, IKBt is specifically induced in response to TLR ligands, but not TNF, which activates NF-KB and MAP kinases similarly to the TLR-mediated pathways. In vitrostudies showthat overexpression ofIKBt inhibits NF-KB activation. On the otherhand,ectopic expression ofIKBtpromotes LPS-induced IL-6production ina cell line.IKBt-deficient mice have been generated to examine itsphysiological roleinTLR-mediated immune responses.v IKBt-deficient mice show defective IL-6production in response to allTLRligands andIL-I,butnot to TNF.Regardingthe IL-6promoter, the NF-KB siteisresponsible forthe positive effect OfIKB!;, which specifically interacts withthepSO subunitofNF-KB.Additionally, microarrayanalysis comparingwild-rype andIKBt-deficient cells inresponse toLPSdemonstrated thatasubset ofLPS-inducible genes,suchasIL-12 p40, GM-CSF andG-CSF,areseverelyaffected byIKBt-deficiency. Given that theLPS-induced transcription OfIKBt occurs earlier than thetranscription ofthese genes,some TLR-mediated responses maybe regulated by a gene expression process involving atleast twosteps that also require inducible IKBl;. Moreover, mice deficient in an IKBt homologue nuclearprotein, IKBNS, display enhanced expressionofLPS-inducible secondary genes suchas IL-6 and IL-12 p40, suggesting that IKBNS co-operatively regulates the MyD88-dependent immune response with IKBt66.67 (Fig. 3).

Lipid A Receptor TLR4-Mediated Signaling Pathways

65

Figure 3. The two-step gene induction program in TLR4-Mediated Immune Response. The TLR4-mediated MyD88-dependent pathway stimulates not only NF-KB and MAP kinases but also mRNA stabilization, and leads to the first-step of gene expression including IKB~. The induced IKB~ subsequently activates second-step gene induction, such as of IL-6 and IL-12 p40 . The immune responses are negatively regulated by IKBNS.

Conclusion

AmongallTLR family members, the LipidA receptorTLR4 seems the onemostelaborately and heavily regulatedbya number of signaling molecules. Thisispresumably because TLR4 mediates activationsignals through not only the MyD88-dependent pathway, which is commonlyutilized by most other TLRs, but also the MyD88-independent pathwayspecifically found in TLR3 and TLR4 signaling. However, the aberrant regulationofTLR4-mediated signaling potentiallyleads to extremetoxicitysuchassepticshockand auto-immunity. Moredetailedanalysis willreveal novel target molecules and molecularmechanisms in TLR4-mediated signaltransduction that should help facilitatethe developmentof more usefuldrugsfor infectiousand inflammatorydiseases.

Acknowledgements

Supported by grants from Special Coordination Funds, the Ministry of Education, Culture, Sports, Science and Technology, Research Fellowships of the Japan Society for the Promotion of Science for Young Scientists, The Junior Research Associate from RIKEN and Exploratory Research for AdvancedTechnology, Japan Science and Technology Agency.

References

1. Freudenberg MA, Galanos e. Bacreriallipopolysaccharides: structure , metabolism and mechanisms of action . Int Rev Immuno11990; 6:207-21. 2. Lin W], Yeh We. Implication of Toll-like receptor and tumor necrosis factor a signaling in septic shock. Shock 2005 ; 24:206-9. 3. Wright SD, Ramos RA, Tobias PS er al. CD 14, a receptor for complexes of lipopolysaccharide (LPS) and LPS bind ing protein. Science 249:1431-3. 4. Miiller ]M, Ziegler-Heitbrock HW; Baeuerle P. Nuclear facror-xls, a mediator of lipopolysaccharide effects Immunobiology 1993; 187:233-56.

66

Cell-Free Expression

5. Lernaitre B. Nicolas E. Michaut L er aI. The dorsoventral regulatory gene cassette sparzle/Toll/cactus controls the potent antifungal response in Drosophila adults. Cell 1996; 86:973-83. 6. Medzhitov R. Preston-Hurlburt P. Janeway CA Jr. A human homologue of the Drosophila Toll protein signals acrivation of adaptive immunity. Nature 1997; 388:394-7 . 7. Poltorak A. He X. Smirnova I et al. Defective LPS signaling in C3H/HeJ and C57BL/I0ScCr mice: mutations in Tlr4 gene. Science 1998; 282:2085-8. 8. Hoshino K. Takeuchi O. Kawai T er al. Cutting edge: Toll-like receptor 4 (TLR4)-ddicient mice are hyporesponsive to lipopolysaccharide: evidence for TLR4 as the Lps gene product. J Immunol 1999; 162:3749-52. 9. Yamamoto M. Takeda K. Akira S. TIR domain-containing adaptors define the specificity ofTLR signaling. Mol Immunol 2004; 40:861-8. 10. Muzio M. Ni J. Feng P er aI. IRAK (Pelle) family member IRAK-2 and MyD88 as proximal mediators of IL-I signaling. Science 1997; 278:1612-5. 11. Adachi O. Kawai T. Takeda K et al. Targeted disruption of the MyD88 gene results in loss ofIL-l- and IL-18-mediated function . Immunity 1998; 9:143-50. 12. Kawai T. Adachi O. Ogawa T et aI. Unresponsiveness ofMyD88-deficient mice to endotoxin. Immunity 1999; 11:115-22. 13. Kawai T. Takeuchi O. Fujita T et al. Lipopolysaccharide stimulates the MyD88-independent pathway and results in activation of IFN-regulatory factor 3 and the expression of a subset of lipopolysaccharide-inducible genes. J Immunol 2001; 167:5887-94. 14. Taniguchi T. Ogasawara K. Takaoka A et al. IRF family of transcription factors as regulators of host defense. Annu Rev Immunol. 2001; 19:623-55. 15. Sakaguchi S. Negishi H. Asagiri M et al. Essential role of IRF -3 in lipopolysaccharide-induced interferon-beta gene expression and endotoxin shock. Biochem Biophys Res Commun 2003; 306:860-6. 16. Alexopoulou L. Holt AC . Medzhitov R er al. Recognition of double-stranded RNA and activation of NF-KB by Toll-like receptor 3. Nature 2001; 413 :732-8. 17. Yamamoto M. Sato S. Mori K et al. Cutting edge: a novel Toll/IL-l receptor domain-containing adapter that preferentially activates the IFN-~ promoter in the Toll-like receptor signaling. J Immunol 2002; 169:6668-72. 18. Oshiumi H. Matsumoto M. Funami K et aI. TICAM-l. an adaptor molecule that participates in Toll-like receptor 3-mediated interferon-beta induction. Nat Immunol 2003 ; 4:161-7. 19. Yamamoto M. Sato S. Hemrni H et aI. Role of adaptor TRIF in the MyD88-independent toll-like receptor signaling pathway. Science 2003; 301:640-3. 20. Hoebe K. Du X. George! P er aI. Identification of Lps2 as a key transducer of MyD88-independent TIR signalling. Nature 2003; 424:743-8. 21. Sato S. Sugiyama M. Yamamoto M er al. Toll/IL-l receptor domain-containing adaptor inducing IFN-~ (TRIF) associateswith TNF receptor-associated factor 6 and TANK-binding kinase 1 and activates rwo distinct transcription factors. NF-KB and IFN-regulatory factor-S, in the Toll-like receptor signaling. J Immuno12003; 171:4304-10. 22. Meylan E. Burns K. Hofmann K er al. RIPI is an essential mediator of Toll-like receptor 3-induced NF-KB activation. Nat Immunol 2004; 5:503-7. 23. Sasai M. Oshium i H. Matsumoto M ct al. Cutting Edge: NF-KB-activating kinase-associated protein 1 participates in TLR3/Toll-IL-l homology domain-containing adapter molecule-l-rnediated IFN regulatory factor 3 activation. J Immunol 2005 ; 174:27-30. 24. Han KJ, Su X. Xu LG et al. Mechanisms of the TRIF-induced interferon-stimulated response dement and NF-KB activation and apoptosis pathways. J BioI Chern 2004; 279:15652-61. 25. Kaiser WJ. Offermann MK. Apoptosis induced by the toll-like receptor adaptor TRIF is dependent on its receptor interacting protein homotypic interaction morif J Immunol 2005; 174:4942-52. 26. Hsu LC. Park JM. Zhang K er al. The protein kinase PKR is required for macrophage apoptosis alter activation of Toll-like receptor 4. Nature 2004 ; 428 :341-5. 27. Horng T. Barton GM. Medzhirov R. TIRAP: an adapter molecule in the Toll signaling pathway. Nat Immunol 2001; 2:835-41. 28. Fitzgerald KA. Palsson-McDermott EM. Bowie AG er al. Mal (MyD88-adapter-like) is required for Toll-like receptor-s signal transduction. Nature 2001 ; 413 :78-83. 29. Yamamoto M. Sato S. Hernrni H et aI. Essential role for TIRAP in activation of the signalling cascade shared by TLR2 and TLR4 . Nature 2002; 420:324-9. 30. Horng T. Barton GM. Flavell RA et al. The adaptor molecule TIRAP provides signalling specificity for Toll-like receptors. Nature 2002; 420:329-33. 31. Yamamoto M. Sato S. Hernrni H et al. TRAM is specifically involved in the Toll-like receptor 4-mediated MyD88-independent signaling pathway. Nat Immunol 2003 ; 4:1144-50.

Lipid A Receptor Tl.Re-Mediated' SignalingPathways

67

32. Fitzgerald KA. Rowe DC. Barnes BJ et al. LPS-TLR4 signaling to IRF-3/7 and NF-KB involves the roll adapters TRAM and TRIEJ Exp Med 2003 ; 198:1043 -55. 33. Oshiumi H, Sasai M. Shida K et al. TIR-containing adapter molecule (T IC AM)- 2. a bridging adapter recruiting to roll-like receptor 4 TICAM-I that induce s interferon-beta. J Bioi Chern 2003; 278 :49751 -62. 34. Akira S. Uernatsu S. Takeuchi O. Pathogen recognition and innate immun ity. Cell 2006; 124:783-801. 35. Rowe DC. McGe ttrick AF, Latz E et aI. The myristoylation ofTRIF-related adapror molecule is essential for Toll-like receptor 4 signal transduction. Proc Natl Acad Sci USA 2006 ; 103:6299-304. 36. McGmrick AF, Brint EK. Palsson-McDermorr EM er aI. Trif-relared adapter molecule is phosphorylated by PKC-E during Toll-like receptor 4 signaling. Proc Natl Acad Sci USA 2006 ;103: 9196 -201. 37. Kagan JC. Medzhitov R. Phosphoinosit ide-mediated adapror recruitment contr ols Toll-like recepror signaling . Cell 2006 ; 125:943-55 . 38. Medvedev AE. Sabroe I. Hasday JD et al. Tolerance to microbial TLR ligands : molecular mechanism s and relevance to disease. J Endotoxin Res 2006 ; 12:133-50. 39. Liew FY, Xu D. Brint EK et al. Negative regulation of roll-like receptor-mediated immune responses. Nat Rev Immunol 2005 ; 5:446-58. 40. Iwami KI. Matsuguchi T. Masuda A et al. Cutting edge: naturally occurring soluble form of mouse Toll-like recepror 4 inhibits lipopol ysaccharide signaling. J Immunol 2000; 165:6682-6 . 41. Chuang TH. Ulevirch RJ. Triad3A . an E3 ubiquirin-protein ligase regulating Toll-like receptors . Nat Immunol 2004 ; 5:495-502. 42. Mansell A. Smith R, Doyle SL et al. Suppressor of cyrokine signaling 1 negatively regulates Toll-like recepror signaling by mediating Mal degradation. Nat Immunol2006; 7:148-55. 43. Nakagawa R. Naka T. Tsutsui H et al. SOCS-l participates in negative regulation of LPS responses. Immunity 2002; 17:677-87. 44. Kinjyo I. Hanada T, Inagaki-Ohara K et al. SOCS l/JAB is a negative regulator of LPS-induced macrophage activation . Immunity 2002 ; 17:583-91. 45. Janssens S. Burns K. Tschopp J et al. Regulation of interleukin-l - and lipopol ysaccharide-induced NF-KB activation by alternative splicing of MyD88. Cure BioI 2002 ; 12:467 -71. 46 . Negishi H . Ohba Y, Yanai H er al. Negative regulation of Toll-like-receptor signaling by IRF-4 . Proc Nad Acad Sci USA 2005 ; 102:15989-94. 47. Janssens S. Beyaert R. Funct ional diversity and regulation of different interleukin- I recepror-associared kinase (IRAK) family member s. Mol Cell 2003; 11:293 -302. 48. Kobayashi K. Hernandez LD. Galan JE er al. lRAK-M is a negative regularor of Toll-like recept or signaling . Cell 2002; 110:191-202. 49. Hardy MP. O'Neill LA. The mur ine lRAK2 gene encodes four alternatively spliced isoforms, two of which are inhibitory. J BioI Chern 2004 ; 279:27699- 708 . 50. Burns K. Clarworrhy J. Mart in L et al. Tollip, a new component of the IL-IRI pathway. links IRAK to the IL-l receptor. Nat Cell Bioi 2000 ; 2:346-51. 51. Kobayashi T. Walsh MC . Choi Y. The role ofTRAF6 in signal transduction and the immune response. Microbes Infect 2004; 6:1333-8. 52. Saro S. Sanjo H. Takeda K ct al. Essential function for the kinase TAKI in innat e and adaptive immune respon ses. Nat Immunol 2005; 6:1087-95. 53. Shim JH. Xiao C, Paschal AE er al. TAKI . but not TABI or TAB2, plays an essential role in multiple signaling pathways in vivo. Gene s Dev 2005 ; 19:2668-81. 54. Lee EG. Boone DL . Chai S et al. Failure to regulate TNF-induced NF-KB and cell death responses in A20 -deficient mice. Science 2000; 289:2350-4. 55. Boone DL . Turer EE. Lee EG er al. The ubiquitin-modifying enzyme A20 is required for termination of Toll-like receptor responses. Nat Immunol2004; 5:1052-60. 56. Sairoh T, Yamamoro M. Miyagishi M et al. AlO is a negative regularor of IFN regularory factor 3 signaling.J Immuno12005; 174:1507 -12. 57. Fukao T. Tanabe M. Terauchi Yet al. PI3K-mediated negative feedback regulation ofIL-12 production in DCs. Nat Immunol 2002 ; 3:875-81. 58. Fukao T. Koyasu S. PI3K and negative regulation ofTLR signaling. Trends Immunol 2003; 24:358-63. 59. Hirotani T. Yamamoro M. Kumagai Y et aI. Regulation of lipopol ysaccharide-indu cible genes by MyD88 and Toll/IL-I doma in containing adaptor inducing IFN-~ . Biochem Biophys Res Commun 2005 ; 328 :383-92. 60. Dana S. Novotn y M. Li X et al. Toll IL-I receprors differ in the ir ability to promote the stabil ization of adenosine and undine-rich element s conta ining mRNA . J Immunol 2004 ; 173:2 755 -61. 61. Yamazaki S. Mura T. Matsuo S ct al. Stimulus-specific indu ction of a novel nuclear facror-KB regulator, IKB-1;. via Toll/Inrerleukin-I receptor is med iated by mRNA stabilization . J BioI Chern 200 5; 280 :1678-87.

68

Cell-Free Expression

62. Yamazaki S, Muta T, Takeshige K. A novel hcB protein, b:B-~, induced by proinflammatory stimuli. negatively regulates nuclear factor-KB in the nuclei. J Bioi Chern 2001; 276:27657-62. 63. Kitamura H, Kanehira K, Okira K et aI. MAIL, a novel nuclear lKBprote in that potentiates LPS-induced lL -6 production. FEBS Lett 2000; 485 :53-6. 64. Harura H , Kato A, Todokoro K. Isolation of a novel inrerleukin-Linducible nuclear protein bearing ankyrin-repeat motifs. J Bioi Chern 2001 ; 276:12485-8. 65. Yamamoto M. YamazakiS. Uernatsu Set aI. Regulation of Toll/Ik-l -recepror-medlated gene expression by the inducible nuclear protein lKB!;. Nature 2004; 430 :218-22. 66. Hirotani T, Lee PY, Kuwata H er al. The nuclear lKB protein lKBNS selectively inhib its lipopolysaccharide-induced IL-6 production in macrophages of the colonic lamina propria. J Immunol 2005; 174:3650-7. 67. Kuwata H, Matsumoto M. Atarashi K et al. IKBNS inhibits induction of a subset of Toll-like receptor-dependent genes and limits inflammation. Immunity 2006; 24:41-51.

CHAPTER 7

Lipid A-Induced Responses In Vivo Nejia Sassi, Catherine Paul, Amandine Martin, Ali Bettaieb and jean-Francois Jeannin*

Abstract

T

he lipid A analogs used in preclinical studies and clinical trial s are not naturally-occurring forms oflipid A; the y are synthetic molecules produced to be less toxic than lipid A itself and they do not reproduce the effects of natural lipid A molecules especially in vivo. The responses induced by lipid A analogs are summarized in this chapter: their fate in the blood stream and their toxicit y as well as the lipid A tolerance and the tumor immune responses they induce. Lipid A is not found in the mammalian organ ism under normal circumstances so its use in cancer therapy raises important questions as to its different effects in vivo and its toxicity, particularly in cancer patients. Lipid A has to be injected intravenously (Lv.) to study its effects. Injections ofchemi cally synthesized lipid A in humans and in animals produce sepsis symptoms, such as tachycardia , tachypnea, hyper or hypothermia and leukocytosis or leukopenia. Similar manifestations are observed mer injection of purified lipopolysaccharide (LPS) , whi ch is why lipid A is usually thought of as the active part ofLPS. While lipid A injection is therefore expected to induce reactions similar to septic shock, the lipid A molecules used to treat cancer are not natural forms but analogs. produced by chemical synthesi s or genetic engineering, specifically selected for the ir low toxicit y.The in vivo effects ofsuch low-toxicity lipid A analogs are summarized in this chapter.

Fate of Lipid A in the Bloodstream

When lipid A enters the bloodstream, it may bind to plasma proteins or to circulating cells (monocytes or granulocytes) or endothelial cells. Serum albumin is the most abundant plasma protein and it binds to lipid A analogs in vitro! and in vivo (C Paul and T Gautier, pers. comm.). When the lipid A analog Eritoran (an antagonist ofLPS) is injected i.v, into healthy human volunteers. the majority ofthe lipoprotein -bound lip id A is recovere d in the high-density lipoprotein (HDL) fraction, whereas the remainder is found in the low-density lipoprotein (LDL) fraction with some in the triglyceride plus chylom icron fraction. ' Eritoran is inactivated by HDL but not by LDL, very low density lipoproteins (VLDL) or albumin.' Therefore. plasm a lipoproteins in vivo can prevent lipid A inte racting with its circulating target cells. But plasma lipoproteins can act as lipid A carriers and transport it to target organs where it can be transferred to its receptors. Lipoprotein-bound lipid A is either taken up by the scavenger receptors ofKupffer cells as shown in mice,' or transferred to its receptor Toll-like receptor 4 (TLR4). A few minutes after being injected, a large proportion ofcirculating lipid A is bound to lipoproteins and cleared by Kupffer cells and hepaeocyres.' Clearance from th e circulation is principally carried out by the liver,3.4 then lipid A is detoxified and excreted into the bile.' Lipid A is detoxified by dea cylation by acyloxyacyl hydrolase " or by dephosphorylation by alkaline phosphatase in lysosome s to give the less toxic monophosphoryl lipid A (MPL),7A few hours after injection. lipid A is found mainly in the liver, · Corresponding Author: lean-Francois Jeannin -EPHE, D ijon, F-2 1000, France; Inserm UB66, Dijon , F-21000, France. Email : jean-francoi s.jeannin @u-bourgogneJr

Lipid A in Cancer Therapy, edited by jean-FrancoisJeannin . ©2009 Landes Bioscience and Springer Science+Busines s M edia.

70

LipidA in Cancer Therapy

but also in the lungs, spleen, adrenals and kidneys.! Other plasma proteins can bind lipid A in vivo,like lipopolysaccharide-binding protein (LBP),9 bactericidalpermeabilityincreasing (BPI) protein9.IOand lactoferrin,II whichregulatethe interactionsbetweenlipidA and its receptor. LBP transfers lipidA to CD 14,aglycosylphosphatidylinositol (GPI) anchoredprotein,whichfacilitates lipid A loading onto the TLR4-MD2 receptor'! (seeChapter 5). LBP enhancesthe response to lipid N2,13 while BPI inhibits the response'P" asdoes lactoferrin."

General Immune Responses to Lipid A Inflammatory Response

Allcellsexpressing TLR4 areableto respondto lipidA, primarilycirculatingrnonocytes" and neutrophils," but alsoendothelialcells." Tumor necrosis factor (TNF)-a and interleukin (IL)-6 (but not IL-l~) are found in human, murine and rat plasmaafter one i.v, injection of lipid A analogs (authors' unpublished results). In vitro, peripheral blood mononuclearcells (PBMC) or macrophages from the three species are activated by lipid A analogsto produce TNF-a and IL_6.19·23IL-I~ secretionby human PBMC activatedin vitro has been reported.24,25 Lipid A analogsactivate mousemacrophages to produce IL-I~ in vitr026,27although in some circumstances the IL-I~ is synthesizedbut not secreted,28.29 while lipid A analogsactivate rat macrophages to produce IL-I~ in vitro." Human PBMC and mouse macrophages are also activatedby lipid A analogsin vitro to produce IL-12 p4030,31 and prostaglandin (PG) E2Y In vitro, lipid A analogs induce the release of high levels of superoxide (0 2-) and hydrogen peroxide (H 20 2) in human monocyres" and granulocytes." In vivo MPL activates mouse macrophages to produceO2 and H 20 !> detectedexvivo.35,36 LipidA analogs alsoinducethe expression ofinduciblenitric oxide(NO) synthase(NOS II) and NO production in human monocytesand in mouseand rat macrophages in vitro.37'39 Inflammatory responses arenot onlydueto the actionofimmunecells but alsoto the reactionof cells in the microenvironment, particularly ofendothelialcells. In this respect,lipidA increases the expression of intercellularadhesion molecule(ICAM)-l in human endothelialcells, the adhesion ofPBMC to the endorhelium'" and the adhesionof weakly attached human granulocytes' v" and monocytes." Increased adhesionofPBMC to the endorheliumenhances the activityofPBMC and endothelial cells, i.e., cytokineproduction. TNF-a and interferon (IFN)-y then induceICAM-l expression in vivo and neutrophil infiltration into the lung," thus creating a positive feedback loop amplifyingthe inflammatoryresponse.

Other Responses

The inflammatoryresponse is in step with other responses like the activationof complement, platelets,cellmigration and phagocytosis. The classic human complementpathwayisactivatedbysomelipidA analogs 23.43 dependingon their structureY.44 Lipid A stimulates rabbit platelets" and human platelets as measuredby the secretionof serotonin and aggregation.46 The migrationof human monocytesand polymorphonuclearleukocytes (PMN)22,43,47and that of mousePMN40 and rnacrophages"is increasedby lipid A analogs, which seemto influencethe kineticsof migration but not chemotaxis. Murine macrophage phagocytosis is induced by lipid A analogsin vitro."

Adaptive Immune Response B-Cells and Antibody Production

The effects of lipid A on the B-cellimmune response are mentioned here,although the role of antibodiesin tumor regression has not yet been documented. Lipid A analogs havea mitogeniceffecton B-Iymphocytes of mice38.43 and rats."

Lipid A-InducedResponses In Vivo

71

Tested as adjuvants in vivo, lipid A analogs increase antibody production against bovine serum albumin (BSA)24.32 and sheep red blood cell (SRBC) antigen in miceY·47

Dendritic Cells

Antigen-presenting cells and particularly dendritic cells (DC) capture and process protein antigen to present them to CD4 and CD8 T-cells (see below). MPL in vitro induces the maturation ofhuman monocyte-derived DC, inducing IL-I2 production and increasing the expression ofHLA-DR, costimulatory molecules like CD40, CD80, CD86 and maturation marker CD83. 50DC matured with MPL and IFN-y express the functional chemokine receptor CCR-7 and produce more IL_1251than with MPL alone. In human DC, the CD40 ligand (CD40L) and MPL trigger persistent stimulation; DC matured in this way possessa high and prolonged migratory capacity and are able to produce IL-l2Y Systemic administration of lipid A analogs in mice results in the redistribution of fully mature DC in the T-cell area of the spleen,52.53 while subcutaneous (s,c.) injection induces the colocalization of mature DC with T-cells in the draining lymph nodes."

T-Cells

In the mixed lymphocyte reaction (MLR), both Th 1 and Th2 cyrokines are induced by human DC treated with MPL,50while human DC treated with MPL plus CD40L or IFN-y induce CD4 T-cells to produce Thl cytokines only, indicating a clonal expansion ofThI cellsY·54 In vivo in mice, MPL and the lipid A analog RC529 enhance short-term clonal expansion ofThI cellswith production ofIFN-y.55 MPL acts directly on human T-cells increasing the intracellular free calcium concentration in CD4 cells, but has no apparent effect in the absence ofT-cell receptor (TCR) triggering. MPL alone has no effect on CD40L expression in resting T-cells ; TCR engagement is requ ired for the enhancement ofCD40L expression at the cell surface by MPUo MPL also enhances the expansion ofantigen-specific CD8 cytotoxic T-cells (CTL).%·57

Systemic Toxicity

Among the lipid A analogs synthesized by chemists , some are oflow toxicity but are effective at inducing tumor regression or inhibiting tumor growth in animal models .21.58.60 Some of these lipid A analogs have been tested in Phase I oncology trials 61.62 (see Chapters 10 and 11) and are only weakly toxic to patients. Our own observations are consistent with these reports; the lipid A analog OM-I74 is biologically active in cancer patients at doses that are not toxic to them. Fever and rigors are reported to be the most common adverse events (grad e 1 or 2)62as we have observed with OM-I74 (N Isambert, unpublished results) . These events are dose-related and subside within halfan hour. The other adverse events (nausea, flushing, etc.) are not dose-related and do not occur systematically after each injection. The most serious event was one grade 3 fever and rigor with cyanosis, which required oxygen treatment; the patient recovered within 30 min. 62 Similar results were obtained with MPL,61 The toxicity observed in some patients is due to an inflammatory reaction. In animals, the severe toxicity observed with lipid A administered at very high doses to determine the lethal dose is due to huge inflammatory reactions. The massive reaction ofthe innate immune system consists in the activation of pro-inflammatory cascades (e.g., the complement system) and the appearance ofvarious pro -inflammatory mediators (TNF-a, IL-I13, IL-6, C5a, etc.) in an amplification loop . The symptoms of toxicity are ruffied fur, conjunctival discharge, diarrhea and prostration. When the inflammatory reaction gets out ofcontrol, animals die ofshock with disseminated intrava scular coagulation and organ or multi-organ failure (acute respiratory distress syndrome, kidney failure and liver failure). Little is known about the mechanisms of lipid A toxicity. In an attempt to find the lethal dose of DT-546 1 in rats, no effect on the number of circulating platelets and no disseminated intravascular coagulation were found, whil e the natural lipid A counterpart (compound 506) induced thrombocytopenia and disseminated int ravascular coagulation." In mice, high doses of

72

Lipid A in Cancer Therapy

ON0-4007 are the cause ofperipheral vascular permeability mediated by TNF-a and/or NO. 63 In rabbits, which are higWy susceptible to LPS, lipid A analogs are weakly or non pyrogenic ,44.64 but high doses ofMPL always reduce the cardiac index and the mean arterial pressure/" Preclinical experiments show that systemic toxicity is not linked to antitumor efficacy. Lipid A molecules are less toxic than the parent LPS and are more effective.21 Lipid A analogs are less toxic, sometimes a hundred or a thousand times so, than natural lipid A forms and can be more effective. However, studies to determine the systemic toxicity and anti-tumor efficacy of these compounds have been done in different animal models under different experimental conditions by different groups, so it is impossible to deduce general rules about structure, function and interactions oflipid A. Effective doses oflipid A analogs are probably far below the toxic doses, as shown in Phase I clinical trials and preclinical experiments. Cancer patients in Phase I trials bear advanced tumors and undergo numerous treatments and as such are probably imm unosuppressed. It is therefore possible that inflammatory reactions in cancer patients are less severe than in healthy volunteers. Furthermore, tolerance to lipid A develops with repeated administration, decreasing the intensity of the inflammatory reaction (see below and Chapter 8). Whether or not lipid A tolerance is a contributing factor to treatment efficacyis not known.

LipidA Tolerance

Lipid A and their analogs can induce a state of self-hyporesponsiveness. This effect is called lipid A tolerance by analogy to the well known LPS tolerance . LPS tolerance can be divided into early tolerance and late tolerance, the latter being mediated by anti-LPS antibodies. However, there is no late tolerance to lipid A molecules because they are not immunogenic. After one lipid A analog injection there is a transient period ofhypo responsiveness in vivo, during which a second injection oflipid A induces a lesserinflammatory reaction . Repeated injections oflipid A increase this tolerance/" Lipid A-induced tolerance also protects against LPS toxicityY.67.68 After repeated i.v, injections ofthe SDZ MRL 953 lipid A analog in cancer patients, a decrease in the serum concentration ofgranulocyte colony-stimulating factor (G -CSF) is observed and the initial decrease in circulating leukocytes lessens.66 This effect is dependent on the rate ofinjection, being observed with short injection intervals (2-5 days)66 but not with long intervals .62.66 In rats, repeated injections ofthe lipid A analog OM-174 reduce the concentration ofTNF-a, IL-6 and IFN-y in serum (authors' unpublished results). Regarding challenge with LPS, repeated injections ofSDZ MRL 953 inhibit the inflammatory response to subsequent injection ofLPS in humans; circulating TNF-a and IL-6 concentrations are reduced, while the number of circulating neutrophil granulocytes increases.f In mice, a first injection oflipid A analogs reduces the inflammatory response : levelsofTNF-a in serum 69.71 and ofIFN -a1~, CSF67.71 and IL-611 and ofsplenic IL-l~ mRNA are all reduced/" Lipid A tolerance is due to TLR4 positive cells, mainly monocytes and macrophages. When rats are injected with different lipid A molecules, we have shown that peritoneal macrophages produce high amounts ofTNF-a, IL-l~, IL-6 and NO in vitro after the first injection, but are no longer active after 2 to 5 injections. Furthermore, these cells do not respond in vitro to lipid A or LPS stimulation for at least 2 weeks after the in vivo injection, while normal macrophages do respond (authors' unpublished results) . The induction of lipid A tolerance is concomitant with corticosterone production and does not occur in adrenalectomized mice/" Glucocorticoids down-regulate many genes including those encoding TNF-a, IL-l~ and IL-6,72.73 because ofthe inhibition ofNFKB nuclear translocation and activity. In vitro, mouse and human monocytes/macrophages can be tolerized to lipid N 4-76 as can rat macrophages (authors' unpublished results). The in vitro induced tolerance lasts one or two days during which the lipid A receptor is not down-regulated." In some Phase I clinical oncology trials, to avoid potential toxic effects each patient is first injected with a low dose ofa lipid A analog and then with higher doses. Whether lipid A tolerance is beneficial or not to anti -tumor treatment is not known. Macrophages are the main cells shown

Lipid A-Induced Responses In Vivo

73

to be tolerized by lipid A analogs in vivo and the experiments on peritoneal or spleen macrophages suggest that tumor associated macrophages (TAM) may be tolerized too. Ifit is true, are TAM-I (anti-tumor macrophages) or TAM-2 (suppressor macrophages) tolerized?

Tumor Immune Responses Molecular Immune Response

The main cytokine, whose production is induced by lipid A and lipid A analogs, is TNF-a. This was the first mechanism described to explain the efficacy of LP S on tumor regression in experimental models. In BALBIc mice bearing Meth A sarcomas treated with LPS, the hemorrhagic necrosis observed within tumors was attributed to a circulating factor," Early studies linked the in vivo efficacyofLPS to the in vitro sensitivity oftumor cell lines to TNF-a. Lipid A analogs induce tumor necrosis 58.79 concomitant with TNF-a production.F" However the phenomenon is not completely explained as injections ofantibodies directed against TNF-a do not prevenr" or only reduce 83.84 the tumor regression. In the same vein, we have shown that lipid A analogs induce the regression oftumors provoked by TNF-a-insensitive cells.21.85For complete regression to occur, an adaptive immune response is necessary as tumor regression is not seen in nude rats. 86 So TNF-a is not sufficient to induce tumor regression. It is now accepted that the antitumor effect ofTNF-a is indirect and dependent on the development ofa specific immune response. IFN-y , whose production can be induced by lipid A in vivo, is a cytokine produced mainly by natural killer (NK) cells (see below) and T-cells.This effect may be indirect as macrophages activated by lipid A produce IL-I2 andlor IL-18 which are IFN-y-inducing cytokines, The requirement of IFN-y for tumor regression has been evaluated in vivo." It was shown that IFN-y-unresponsive Meth A fibrosarcomas are not rejected by BALB/c mice treated with LPS. IFN-y mRNA levels are higher in KDH-8 hepatomas and the protein levelsare higher in Prob colon carcinomas when rats are treated with lipid A analogs. 60.88Moreover, in tumor-bearing mice, the lipid A analog DT-S46I causes an increase in the concentration of circulating IFN-y and anti-IFN-y antibodies reduce the activity and TNF-a production, showing the involvement ofIFN-y in lipid A efficacy.84Other IFNs may be involved since the lipid A analog DT-S46 I could act through IFN-a/13 as well as IFN_y.84 IL-I13 is another cytokine whose secretion can be induced by lipid A in vivo. IL-I13 is principally secreted by macrophages, but also by various immune cells such as DC and by endothelial cells during lipid A treatment. IL-I13 mRNA levels are higher in KDH-8 hepatomas and the protein levels are higher in Prob colon carcinoma when rats are treated with lipid A analogs. 6O·88 IL-I13 may be directly cytotoxic to tumor cells' or in vivo with IFN-a against BI6 melanoma in C57BL/6 rnice." A positive correlation was found between the abilit y oflipid A analogs to activate macrophages to produce IL -I13 in vitro and their anti-tumor activity in vivo." Other cytokines could be involved in the lipid A anti-tumor effect. Lipid A analogs induce the production ofIL-6, which promotes lymphokine-activated killer (LAK) celldifferentiation, as does IL-12, the production ofwhich is also induced in vivo by lipid A in KDH-8 hepatornas.P'" NO is a mediator ofmacrophage cytotoxicity in vitro," but its role in controlling tumor growth is not yet well defined.?? In a model of colon cancer (Prob) in BDIX rats, spleen macrophages are shown to inhibit T-cell proliferation via NO, preventing an adaptive immune response." In this model, the lipid A analog OM-174 induces the regression of tumors that is concomitant with: 1. the induction oflipid A tolerance in spleen and peritoneal macrophages, causing them to stop producing NO (authors' unpublished results); 2. the synthesis ofIL-I13 and IFN-y in tumors, cytokines shown to activate NOS II and NO production in Prob tumor cells;6O 3. the activation of NOS II and the production of NO in tumors, particularly in tumor cells;6O·92

a

Cavaillon JM. Interleuk in-1. In: Cava illon JM, ed. Les cytokines. Paris: Masson ,

1996:93-117.

74

Lipid A in Cancer Therapy

4. the decrease in transforming growth factor (TGF)-~ I production in tumor cells shown to inhibit NOS II expression in Prob tumor cells;93 S. the apoptosis oftumor cells.85.92 Thus it was concluded that 0 M-174 induces the production ofN0 that sensitizes tumor cells to apoptosis. In the model ofEMT-6 mammary tumors in BALBIc mice, using knockout (KO) mice for the NOS II gene and a NOS II inhibitor, we have shown that the NOS II oftumor cellsinhibits tumor growth via NO while the NOS II ofstromal cells promotes tumor growth." In the same model, M-17 4 inhibits tumor growth by activating N OS II in tumor cells (authors' unpublished results) which is consistent with the above results. Furthermore. lipid A via NO induces the expression and the aggregation ofFas at the plasma membrane oftumor cells. In contrast with other models. in rats bearing KD H -8 hepatomas, NO production by peritoneal macrophages is suppressed and restored by ON0-4007.88

o

Cellular Immune Responses Cells ofthe Innate Immune System

Historically, the first cells considered to be involved in the anti-tumor immune response were macrophages. In the 70s, they were first shown to nonspecifically kill tumor cells in vitro in the presence ofLPS .95 Spleen and peritoneal macrophages are activated in vivo by lipid A analogs to produce TNF-a, ILl-~. IL-6, IL-12. IL-18 and N0,21 ,60.81.88 just as when they are activated in vitro. 21,81 Kupffer cellsand TAM are also activated in vivo to produce TNF-a. 96However, rolerization to ON0-4007 affects Kuppfer cells as found in the serum of tumor-bearing mice (MM46 mammary tumor in C3He/He mice), while TAM are not tolerized. ON0-4007 has an anti-tumor effect 96 and stimulates constant TNF-a production in MM46 tumors. Are TAM modified by tumor cells or their microenvironment such that they are not tolerized? Or is this a property of ON0-4007 since the parent lipid A (LA-IS PP) induces tolerance in TAM ofMM46 tumorsr" Two types ofmacrophage have now been described; Type I stimulates Type I helper T-cells, while Type 2 develops pro -tumor activity and down-regulates Th I cell activit y.97 It is possible that TAM. which could differentiate in tumors under the control ofthe microenvironment, differentiate into TAM-I in the presence oflipid A . IFN -y or TNF-a. orinto TAM -2 in the presence ofIL-lO. IL-6, PGE2 or TGF-~. TAM-I could develop anti-tumor activit y by producing IL-12 and stimulating the Th-I immune response while TAM-2 could develop pro -tumor activity by inhibiting the Th I immune response. It is not known whether lipid A tolerization affects TAM- lor TAM-2 or both typ es ofmacrophage. In any case. the importance ofmacrophages should be emphasized because they are the main producer ofTNF-a in response to lipid A analogs. NK cellsare mononuclear cellsthat participate in the innate immune response. NK cellsare also able to lysetumor cellswithout prior sensitization to specific antigen. When NK cellsare activated in vitro, they differentiate into LAK cells which are cytotoxic to tumor cells. When injected one day before BI6 melanoma cells,GLA-27 reduced the number ofmetastases via NK cell activation." Furthermore, ONO-4007 pretreatment increased NK activity in the liver ofCS7BL/6 mice and decreased the number ofEL4 hepatic metastases via secretion ofthe NK-activating cytokine IL-12 by Kuppfer cells." While these lipid A analogs activate NK cells when injected into mice before the tumor cells, OM-174 activates NK cells and reduces melanoma growth when administered after BI6 melanoma cells are administered.I'" However in none of these reports are NK cells demonstrated to be involved in the anti-tumor effect oflipid A. In humans as in animals , the innate immune response induced in vivo by lipid A analogs, precedes the adaptive immune response. All the results discussed here show that the inflammatory response induced by lipid A analogs is certainly involved in their anti-tumor activity. Although inflammation contributes to tumor progression as shown by the preventive effect ofnonsteroidal anti-inflammatory drugs , it is likely that a local inflammatory response induced in the tumor microenvironment has an anti-tumor effect. One mechanism could be by the activation ofNOS II by inflammatory cytokines.Ieading to NO production in the tumor microenvironment. This could

Lipid A-InducedResponses In Vivo

75

occur because NO sensitizes tumor cells to cell death induced by small amounts ofcytokines like IFN-y, FasLor TRAIL, so there would be selective death oftumor cells that are more sensitive to these eytokines than normal cells.An adaptive immune response can be activated as a consequence, either by way ofcell death or by way ofnitrosylation ofproteins that become immunogenic.

Cells ofthe Adaptive Immune Response

Antigen presenting cells (APe) are necessary for the development of an effective adap tive immune response . Among APC, DC playa key role because they are capable of antigen cross-presentation. Apoptotic tumor cells or tumor celllysates are sources of antigens. After the uptake ofantigens DC migrate to and mature in draining lymph nodes . Migration is induced by local production ofcytokines like TNF-a or IL-1~. Maturation occurs after DC encounter bacterial products (LPS), cytokines, Tvcell-derived signals (CD40L) and/or cell products (tumor cell lysates or heat shock proteins (HSP)). Maturation is a process in which major histocompatibility complex (MHC) and costimulatory molecules are upregulated (e.g., CD40 ) .101 Therefore lipid A analogs could induce DC maturation in tumor-bearing animals as they do in normal animals (see AdaptiveImmuneResponse). The treatment of rats bearing colon peritoneal carcinomatosis with OM-174 induces the infiltration ofactivated DC into tumor nodules ," suggesting a possible role for DC in the efficacy ofthe treatment. Antigens acquired exogenously are processed and presented as peptides on MHC class II molecules for activation of CD4 T-cells. Following recognition of antigen on MHC class II molecules, CD4 T-cells can differentiate into Th1 cells and Th2 cells, which produce IFN-y and IL-4, respectively. The production ofIL-12 by mature DC is critical for the differentiation ofThl cells.DC also cross-present numerous tumor antigens , including HSP. on MHC classI molecules to CD8 Tvcells.'?' CD8 T-cells will different iate into cytotoxic T-cells able to recognize and kill tumor cells presenting the same peptides on MHC class I molecules. In 1973, the importance ofT-lymphocytes in the respon se oftumor-bearing animals to treatment with LPS or lipid A was evoked by Parr et al.102 Lipid A analogs are not active in nude rats or nude mice, which is evidence that mature T-cells have a role in the treatment efficacyl03 (authors' unpublished results). Furthermore, rats and mice cured from their tumors by lipid A analogs are protected against a second challenge of tumor cells which fails to give rise to a tumor8S•104 (authors' unpublished results). This immune memory is transferred to naive rats by non adherent splenocytes, taken from rats cured from their tumors by the lipid A analogs OM-174 (authors' unpublished results) or ON 0 4007 .104 However, immunohistological examination oftumor slices shows that T-cells are not in close contact with tumor cells during tumor regression, so tumor cells are not killed directly by T_cells.86•92 The involvement ofT-cells shown in nude animals is probably regulatory. Lipid A analogs have a mitogenic effect on B-lymphocytes of mice (see Adaptive Immune

Response).

The lipid A analog ONO-4007 is mitogenic to B-cells from BALBIc mice. lOS However the relevance of the antibody response to the anti-tumoral effect oflipid A has not been documented.

Vascular Response

Blood flow and angiogenesis are ofgreat relevance in cancer biology since they modulate the supply ofoxygen and nutrients to tumors and the clearance oftoxic metabolites, all ofwhich are needed for tumor growth. On the contrary, for tumor regression, blood flow and angiogenesis determine the accessoftoxic cytokines or immune cellsto tumor cells.The lipid A analog DT-5461 decreases the blood flow in subcutaneous Meth A fibrosarcoma s in BALB/c mice 106 and in liver nodules ofVX2 carcinomas in rabbits .107This decrease involves the action ofTNF-a, IFN-a/~ and IFN- yl06 but could not be reproduced with Lv. injections ofTNF-a alone.I'" Another effect of lipid A analogs is the inhibition oftumor angiogenesis as evidenced in C57BL/6 mice bearing B16 melanomas treated with IFN-y and then the lipid A analog GL-60. 108This effect was confirmed in the same model using DT-5461 alone.P The effect is likely mediated by TNF-a, which is toxic to the endothelium and inhibits the motility and proliferation ofendothelial cells.

76

LipidA in Cancer Therapy

Furthermore, the lipid A analog OM-174 inhibits ascite formation in a model ofrat colon peritoneal carcinomatosis, owing to a decrease in plasma leakage, probably mediated by an inhibition ofvessel permeability added to a decrease in tumor volume (authors' unpublished results) .

Conclusion

The in vivo reponses oftumor-bearing rats or mice to lipid A analogs are first an innate immune response, then an adaptive response leading to a specific anti-tumor effect and, in some models, a concomitant vascular response. The innate immune response is a local inflammatory response, toxic to tumor cells and to endothelial cells in some models, with attraction and activation of PMN, macrophages and DC and production of cytokines, chemokines and inflammatory mediators. When the balance ofthese numerous and sometimes antagonist effects is favorable to an anti-tumor response , an adaptive immune response completes the tumor elimination and develops an immune memory. The local inflammatory response has no general toxicity and especially no auto-immune toxicity. Why this ubiquitous treatment is highly selective oftumors is not known. In animal models it has been demonstrated that the toxic effect oflipid A can be separated from the beneficial effect, but the molecular basis of the difference is not yet understood. Transfer to human cancer therapy may well be successful as the first clinical assaysindicate immune responses occur at lipid A analog concentrations that are not toxic to patients.

References

1. Rossignol DP, Wasan KM, Choo E et aI. Safety. pharmacokinetics. pharmacodynamics and plasma lipoprotein distribution of eritoran (E5564) during continuous intravenous infusion into healrhy volunteers. Antimicrob Agents Chemother 2004 , 48(9) :3233-3240 . 2. Hampton RY, Golenbock DT. Penman M et al. Recognition and plasma clearance of endotoxin by scavenger receptors. Nature 1991,352(6333):342-344. 3. Hopf U. Ramadori G. Moller B et al. Hepatocellular clearance function of bacterial Iipopolysaccharides and free lipid A in mice with endoroxic shock. Am] Emerg Med 1984,2(1):13-19. 4. Munford RS, Andersen JM. Dietschy ]M. Sites of tissue bind ing and uptake in vivo of bacterial lipopolysaccharide-high density lipoprotein complexes: studies in the rat and squirrel monkey. ] Clin Invest 1981; 68(6) :1503-1513. 5. Kleine B, Freudenberg MA, Galanos C. Excretion of radioactivity in faeces and urine of rats injected with 3H.14C-lipopolysaccharide. Br J Exp Patho11985; 66(3) :303-308. 6. Erwin AL. Munford RS. Deacylation of structurally diverse lipopolysaccharides by human acyloxyacyl hydrolase. J Bioi Chern 1990,265(27):16444-16449. 7. Bentala H , VerweijWR. Huizinga-Van der Vlag A et al. Removal of phosphate from lipid A as a strategy to detoxify lipopolysaccharide. Shock 2002 , 18(6):561-566. 8. Reddy TS, Kishore V. Distribution and localization of monophosphoryl Iipid A in selected tissues of the rat. Immunopharmacol Immunoroxicol Ivve: 18(1):145-159. 9. Tobias PS. Soldau K. Ulevitch R]. Isolation of a lipopolysaccharide-binding acute phase reactant from rabbit serum.] Exp Med 1986; 164(3) :777-793. 10. Jiang Z, Hong Z. Guo W et al. A synthetic peptide derived from bactericidal!permeabilit y-increasing protein neutralizes endotoxin in vitro and in vivo. Int Immunopharmacol 2004, 4(4) :527-537. 11. Appelmelk B]. An YQ, Geerts M er al. Lactoferrin is a lipid A-binding protein . Infect Immun 1994, 62(6) :2628 -2632. 12. Miyake K. Innate immune sensing of pathogens and danger signals by cell surface Toll-like receptors. Semin Immunol 2007 , 19(1):3-10. 13. Azuma M, Matsuo A, Fujimoto Y et al. Inhibition of lipid A-mediated type I interferon induction by bacter icidallpermeability-increasing protein (BPI). Biochern Biophys Res Commun 2007, 354(2) :574-578. 14. Gazzano-Sanroro H. Meszaros K. Birr C et al. Competition between rBPI23. a recombinant fragment ofbactericidal!permeability-increasing protein and lipopolysaccharide (LPS)-binding protein for binding to LPS and gram-negative bacteria. Infect Immun 1994,62(4):1185-1191. 15. Chodaczek G. Zimecki M. LukasiewiczJ er al. A complex of lacroferrin with monophosphoryllipid A is an efficient adjuvant of the humoral and cellular immune response in mice. Med Microbiol Immunol 2006, 195(4):207-216. 16. Lien E. Chow]C. Hawkins LD ct al. A novel synthetic acyclic lipid A-like agonist activates cells via the lipopolysaccharide/toll-like receptor 4 signaling pathway, ] Bioi Chern 2001, 276(3):1873-1880.

Lipid A-Induced Responses In Vivo

77

17. Rasool O. Freer E. Moreno E ct aI. Effect of Brucella abortus lipopolysaccharide on oxidative metabolism and lysozyme release by human neutrophils . Infect Immun 1992; 60(4) :1699-1702 . 18. Dunzendorfer S. Lee HK, Soldau K er aI. Toll-like receptor 4 functions intracellularly in human coronary artery endothelial cells: roles of LBP and sCD14 in mediating LPS responses. FASEB J 2004; 18(10) :1117-1119. 19. Feist W; Ulmer AJ. Musehold J et al. Induction of tumor necrosis factor-alpha release by lipopolysaccharide and defined lipopolysaccharide partial structures. Immunobiology 1989; 179(4-5):293-307. 20. Ogawa T. Suda Y, Kashihara W et al. Immunobiological activities of chemically defined lipid A from Helicobacter pylori LPS in comparison with Porphyromonas gingivalis lipid A and Escherichia coli-type synthetic lipid A (compound 506). Vaccine 1997; 15(15) :1598-1605. 21. Jeannin JF, Onier N. Lagadec P et al. Antitumor effect of synthetic derivatives of lipid A in an experimental model of colon cancer in the rat. Gastroenterology 1991; 101(3):726-733. 22. Korani S. Takada H . Tsujimoto M et al. Synthetic lipid A with endoroxic and related biological activities comparable to those of a natural lipid A from an Escherichia coli rernutant. Infect Immun 1985; 49(1) :225-237. 23. Takahashi I. Kotani S. Takada H er al. Requirement of a properly acylated beta(I-6)-D-glucosamine disaccharide bisphosphate structure for efficient manifestation of full endotoxic and associated bioactivities of lipid A. Infect Immun 1987; 55(1) :57-68. 24. Caroff M. Cavaillon JM. Fitting C et al. Inability of pyrogenic. purified Borderella pertussis lipid A to induce interleukin -I release by human monocytes. Infect Immun 1986; 54(2) :465-471. 25. Hurme M. Serkkola E. Comparison of interleukin 1 release and interleukin 1 mRNA expression of human monocytes activated by bacteriaIlipopolysaccharide or synt hetic lipid A. Scand J Immunol 1989; 30(2) :259-263. 26. Kotani S. Takada H . Takahashi I et al. Low endotoxic activities of synthetic Salmonella-type lipid A with an additional acyloxyacyl group on the 2-amino group of beta (1·6) glucosamine disaccharide 1,4'-bisphosphate. Infect Immun 1986; 52(3) :872-884. 27. Arata S. Kasai N. Klein TW et aI. Legionella pneumophila growth restriction and cytokine production by murine macrophages activated by a novel Pseudomonas lipid A. Infect Immun 1994; 62(2) :729-732. 28. Cavaillon JM . Fitting C. Caroff M et al. Dissociation of cell-associated interleukin -l (IL-I) and IL-l release induced by lipopolysaccharide and lipid A. Infect Immun 1989; 57(3):791-797. 29. Okernoro K. Kawasaki K. H anada K er al. A potent adjuvant monophosphoryl lipid A triggers various immune responses. but not secretion of IL-I beta or activation of caspase-I. J Immunol 2006; 176(2) :1203-1208. 30. Martin M. Michalek SM. Katz J. Role of innate immune factors in the adjuvant activity of monophosphoryllipid A. Infect Immun 2003 ; 71(5) :2498 -2507. 31. Salkowski CA. Detore GR. Vogel SN. Lipopolysaccharide and monophosphoryl lipid A differentially regulate interleukin-LZ, gamma interferon and interleukin-IO mRNA production in murine macrophages. Infect Immun 1997; 65(8) :3239-3247. 32. Galanos C. Lehmann V, Luderitz 0 et al. Endoroxic properties of chemically synthesized lipid A part structures . Comparison of synrhetic lipid A precursor and synthetic analogues with biosynrhetic lipid A precursor and free lipid A. Eur J Biochem 1984; 140(2):221-227. 33. Saha DC. Barua RS. Astiz ME et al. Monophosphoryl lipid A stimulated up-regulation of reactive oxygen intermediates in human monocytes in vitro. J Leukoc BioI 2001; 70(3):381-385. 34. Dah inden CA . Fehr J. Hugli TE. Role of cell surface contact in the kinetics of superoxide production by granulocytes. J Clin Invest 1983; 72(1) :113-121. 35. Chen YF. Solem L. Johnson AG. Activation of macrophages from aging mice by detoxified lipid A. J Leukoc BioI 1991; 49(4) :416-422. 36. Dijksrra J. Mellors JW; Ryan JL. Altered in vivo activity of liposome-incorporated lipopolysaccharide and lipid A. Infect Immun 1989; 57(11) :3357-3363 . 37. Saha DC. Astiz ME. Lin RY er al. Monophosphoryl lipid A stimulated up-regulation of nitric oxide synthase and nitric oxide release by human monocytes in vitro. Immunopharmacology 1997; 37(2-3):175-184. 38. Tanarnoto K. Azumi S. Haishima Y et al. Endotoxic properties of free lipid A from Porphyromonas gingivalis. Microbiology 1997; 143(Ptl):63-71. 39. Lopez-Urrutia L. Alonso A. Nieto ML et al. Lipopolysaccharides of Brucella abortus and Brucella melitensis induce nitric oxide synthesis in rat peritoneal macrophage s. Infect Immun 2000; 68(3) :1740-1745 . 40. Renzi PM. Lee CH. Anti-lip id A monoclon al antibodies alter human endothelial cell ICAM-l expression in vitro. Shock 1995; 3(5) :329-336.

78

LipidA in Cancer Therapy

41. Dahinden C. Galanos C. Fehr J. Granulocyte activation by endotoxin . 1. Correlation between adherence and other granulocyte functions and role of endotoxin structure on biologic activity. J Immunol 1983; 130(2):857-862. 42. Kang BH. Manderschied BD. Huang YC et al. Contrasting response oflung parenchymal cells to instilled TNF alpha and IFN gamma: the inducibility of specific cell ICAM-l in vivo. Am J Respir Cell Mol BioI 1996; 15(4):540-550. 43. Kotani S. Takada H. Tsujimoto M et al. Immunobiological activities of synthetic lipid A analogs and related compounds as compared with those of bacterial lipopolysaccharide. reglycollpld, lipid A and muramyl dipeptide. Infect Immun 1983; 41(2):758-773. 44. Tanamoto K. Zahringer U. McKenzie GR er al. Biological activities of synthetic lipid A analogs: pyrogenicicy,lethal toxicity. anticomplement activity and induction of gelation of Limulus amoebocyte lysate. Infect Immun 1984; 44(2):421-426. 45. Ramsey RB. Hamner MB. Alving BM et al. Effects of lipid A and liposomes containing lipid A on platelet and fibrinogen production in rabbits. Blood 1980; 56(2):307-310. 46. Grabarek], T immons S. Hawiger J. Modulation of human platelet protein kinase C by endotoxic lipid A. J Clin Invest 1988; 82(3) :964-971. 47. Takada H . Korani S, Tsujimoto M et al. Immunopharmacological activities of a synthetic counterpart of a biosynrheric lipid A precursor molecule and of its analogs. Infect Immun 1985; 48(1) :219-227. 48. Verghese MW; Snyderman R. Endotoxin (LPS) stimulates in vitro migration of macrophages from LPS-resistant mice but not from LPS-sensitive mice. J Immuno11982; 128(2) :608-613. 49. Sagara-Ishijima N, Furuhama K. Toxic characteristics of the synthetic lipid A derivative DT-5461 in tats and monkeys. Toxicol Sci 1999; 49(2) :324-331. 50. Ismaili J, Rennesson J. Aksoy E er al. Monophosphoryllipid A activates both human dendritic cells and T-cells. J Immunol 2002 ; 168(2):926-932. 51. Lapteva N, Seeehammagari MR, Hanks BA et al. Enhanced activation of human dendritic cells by inducible CD40 and Toll-like receptor-4ligation. Cancer Res 2007 ; 67(21) :10528-10537 . 52. De Becker G. Moulin V. Pajak B et al. The adjuvant monophosphoryllipid A increases the function of antigen-presenting cells. Inr Immunol 2000; 12(6):807-815. 53. Pajak B, Garze V. Davies G et al. The adjuvant OM -174 induces both the migration and maturat ion of murine dendritic cells in vivo. Vaccine 2003 ; 21(9-10):836-842. 54. Ten Brinke A, Karsten ML, Dicker MC et al. The clinical grade maturation cocktail monophosphoryl lipid A plus IFNgamma generates monocyte -derived dendritic cells with the capacity to migrate and induce Thl polarization . Vaccine 2007; 25(41) :7145-7152. 55. Thompson BS. Chilton PM, Ward JR et al. The low-toxicity versions of LPS, MPL adjuvant and RC529. are efficient adjuvants for CD4+ T-ceUs. J Leukoc Bioi 2005 ; 78(6) :1273-1280. 56. EvansJT, Cluff CW; Johnson DA er al. Enhancement of antigen-specificimmunity via the TLR4 ligands MPL adjuvant and Ribi.529. Expert Rev Vaccines 2003; 2(2) :219-229. 57. Baldridge JR . McGowan P, EvansJT et al. Taking a Toll on human disease: Toll-like receptor 4 agonists as vaccine adjuvants and monorhcrapcutic agents. Expert Opin BiolTher 2004; 4(7):1129-1138. 58. Kumazawa E, Tohgo A. Soga T et al. Significant antitumor effect of a synthetic lipid A analogue. DT-5461. on murine syngeneic tumor models. Cancer Immunol Immunother 1992; 35(5) :307-314. 59. Yang D. Satoh M, Ueda H et al. Activation of rumor-infiltrating macrophages by a synthetic lipid A analog (ON0-4007) and its implication in antitumor effects. Cancer Immunol Immunother 1994; 38(5) :287-293. 60. On ier N, Hilpert S, Reveneau S er al. Expression of inducible nitric oxide synthase in tumors in relation with their regression induced by lipid A in rats. Inr J Cancer 1999; 81(5):755 -760. 61. Vosika GJ, Barr C, Gilbertson D. Phase-I study of intravenous modified lipid A. Cancer Immunol Irnmunorher 1984; 18(2):107-112. 62. de Bono JS, Dalgleish AG, Carmichael J et al. Phase I study of ON0-4007, a synthetic analogue of the lipid A moiety of bacterial lipopolysaccharide. Clin Cancer Res 2000; 6(2) :397-405. 63. Ishida H, Fujii E, Irie K et al. Role of inflammatory mediators in lipid A analogue (ON0-4007)-induced vascular permeability change in mouse skin. Br J Pharmacol2000; 130(6):1235-1240. 64. Lam C. Schutze E, Hildebrandt J et al. SDZ MRL 953. a novel immunostimulato ry monosaccharidic lipid A analog with an improved therapeutic window in experimental sepsis. Antimicrob Agents Chemorher 1991; 35(3) :500-505. 65. Ali KH, Feeley TW; Bieber M ct al. Cardiovascular effect of int ravenous lipid A in rabbits. Circ Shock 1987; 23(4) :285-293. 66. Kiani A, Tschiersch A. Gaboriau E et al. Downregulation of the proinflammatory cyrokine response to endotoxin by pretreatment with the nontoxic lipid A analog SDZ MRL 953 in cancer patients. Blood 1997; 90(4) :1673-1683 .

Lipid A-InducedR(sp0n5(S In Vivo

79

67. Madonna GS. Peterson JE. Ribi EE et al. Early-phase endotoxin tolerance : induction by a detoxified lipid A derivative. monophosphoryllipid A. Infect Immun 1986; 52(1):6-11. 68. Astiz ME, Rackow EC, Kim YB et al. Monophosphoryllipid A induces tolerance to the lethal hemodynamic effects of endotoxernia. Circ Shock 1991; 33(2):92-97. 69. Zuckerman SH. ~eshi N. In vivo inhibition of lipopolysaccharide-induced lethality and tumor necrosis factor synthesis by Rhodobacter sphaeroides diphosphoryllipid A is dependent on corticosterone induction. Infect Immun 1992; 60(7):2581-2587. 70. Matsuura M. Kiso M. Hasegawa A et al. Multistep regulation mechan isms for tolerance induction to lipopolysaccharide lethality in the tumor-necrosis-factor-alpha-mediated parhway. Application of nontoxic monosaccharide lipid A analogues for elucidat ion of mechanisms. Eur J Biochem 1994; 221(1) :335-341. 71. Henricson BE. Benjamin WR. VogelSN. Differential cytokine induction by doses oflipopolysaccharide and monophosphoryllipid A that result in equivalent early endotoxin tolerance. Infect Immun 1990; 58(8) :2429-2437. 72. Beutler B, Krochin N. Milsark lW et al, Control of cachectin (tumor necrosis factor ) synthesis: mechanisms of endotoxin resistance. Science 1986; 232(4753):977-980. 73. Amano Y. Lee SW; Allison AC. Inhibition by glucocorticoids of the formation of interleukin-I alpha. inrerleukin-I beta and interleukin-6: mediation by decreased mRNA stability. Mol Pharmacol 1993; 43(2):176-182. 74. Henricson BE, Perera PY. ~eshi N et al. Rhodopseudomonas sphaeroides lipid A derivatives block in vitro induction of tumor necrosis factor and endotoxin tolerance by smooth lipopolysaccharide and monophosphoryllipid A. Infect Immun 1992; 60(10):4285-4290. 75. Galanos C, Luderitz 0, Rietschel ET et al. Synthetic and natural Escherichia coli free lipid A express identical endotoxic activities. Eur J Biochem 1985; 148(1) :1-5. 76. Knopf HP, Otto F, Engelhardt R er al. Discordant adaptation of human peritoneal macrophages to stimulation by lipopolysaccharide and the synthetic lipid A analogue SDZ MRL 953. Down-regulation ofTNF-alpha and IL-6 is paralleled by an up-regulation ofIL-l beta and granulocyte colony-stimulating factor expression.] Immunol1994; 153(1):287-299. 77. Kitchens RL, Ulevirch R], Munford RS. Lipopol ysaccharide (LPS) partial structures inhib it responses to LPS in a human macrophage cell line without inhibiting LPS uptake by a CD 14-mediated pathway. J Exp Med 1992; 176(2):485-494. 78. Carswell EA. Old LJ, Kassel RL et al. An endotoxin-induced serum facror that causes necrosis of rumors. Proc Natl Acad Sci USA 1975; 72(9) :3666-3670. 79. van de W iel PA, van der Pijl A, Bloksma N. Role of tumour necrosis factor in the tumour-necrotizing activity of agents with diverging toxicity. Cancer Immunollmmunother 1991; 33(2) :115-120. 80. Kuramitsu Y, Nishibe M. Ohiro Y ct al. A new synthetic lipid A analog, ON0-4007. stimulates the production of tumor necrosis factor-alpha in tumor tissues. resulting in the rejection of transplanted rat hepatoma cells. Anticancer Drugs 1997; 8(5) :500-508. 81. Akimoto T, Kumazawa E.]imbo T et al. Antitumor effect ofDT-546Ia, a synthetic low-toxicity lipid A analog, involves endogenous tumor necrosis factor induction subsequent to macrophage activation. Int J Immunopharmacol 1994; 16(11):887-893. 82. Sato K, Yoo YC. Mochizuki M et al. Inhibition of tumor-induced angiogenesis by a synthetic lipid A analogue with low endotoxicity, DT-5461. ]pn] Cancer Res 1995; 86(4 ):374-382. 83. Matsumoto N. Oida H, Aze Y et al. Intrarumoral tumor necrosis factor induction and tumor growth suppression by ON0-4007, a low-toxicity lipid A analog. Anticancer Res 1998; 18(6A):4283-4289. 84. Kumazawa E. Akimoto T. Kira Y et al. Intratumoral production of tumor necrosis factor augmented by endogenous interferons results in potent antitumor effects of DT-5461, a synthetic lipid A analog. J Irnmunother Emphasis Tumor Immunol1995 ; 17(3):141-150 . 85. Onier N. Hilpert S. Arnould L et al. Cure of colon cancer metastasis in rats with the new lipid A OM 174. Apoptosis of tumor cells and immunization of rats. Clin Exp Metastasis 1999; 17(4):299-306. 86. Onier N, Lejeune P, Mart in M et al. Involvement of T-lymphocytes in curative effect of a new immunomodulator OM 163 on rat colon cancer metastases. Eur J Cancer 1993; 29A(14) :2003-2009. 87. Dighe AS. Richards E. Old L] et al. Enhanced in vivo growth and resistance to rejection of tumor cells expressing dominant negative IFN gamma receptors. Immunity 1994; 1(6):447-456. 88. Matsushita K. Kuramitsu Y, Ohi ro Yet al. ON0-4007, a synt hetic lipid A analog, induces Th l-t ype immune response in tumor eradication and restores nitric oxide production by peritoneal macrophages. Int J Oncol 2003; 23(2 ):489-493. 89. Hibbs JB Jr. Taintor RR, VavrinZ er al. Nitric oxide: a cytotoxic activated macrophage effector molecule. Biochem Biophys Res Commun 1988; 157(1):87-94. 90. Jeannin ]F, Leon L. Cortier M er al. Nitric oxide-induced resistance or sensitization to death in tumor cells. Nitric Oxide 2008; 19(2):158-163.

80

Lipid A in Cancer Therapy

91. Lejeune P, Lagadec P, Onier N et at. Nitric oxide involvement in tumor-induced immunosuppression. ] Immuno11994; 152(10):5077-5083 . 92. Larmonier CB. Arnould L. Larmonier N et at. Kinetics of tumor cell apoptosis and immune cell activation during the regression of tumors induced by lipid A in a rat model of colon cancer. Inr ] Mol Med 2004; 13(3):355-361. 93. Lagadec P. Raynal S. Lieubeau B et at. Evidence for control of nitric oxide synthesis by intracellular transforming growth factor-beta1 in tumor cells. Implications for tumor development. Am ] Parhol 1999; 154(6):1867-1876. 94. Gauthier N. Lohm S. Touzery C et at. Tumour-derived and host-derived nitric oxide differentially regulate breast carcinoma metastasis to the lungs. Carcinogenesis 2004; 25(9):1559-1565. 95. Alexander P. Evans R. Endotoxin and double stranded RNA render macrophages cytotoxic. Nat New Bioi 1971; 232(29) :76-78. 96. Satoh M. Ando S. Shinoda T et at. Clearance of bacteriallipopolysaccharides and lipid A by the liver and the role of argininosuccinate synthase. Innate Immun 2008; 14(1):51-60. 97. Verreck FA, de Boer T. Langenberg DM et at. Human lL -23-producing type 1 macrophages promote but lL-10-producing type 2 macrophages subvert immunity to (myco)bacteria. Proc Nat! Acad Sci USA 2004 ; 101(13):4560-4565. 98. Nakatsuka M. Kumazawa Y, Homma ]Y et at. Inhibition in mice of experimental metastasis of B16 melanoma by the synthetic lipid A-subunit analogue GLA-60. Int ] Immunopharmacol 1991; 13(1):11-19. 99. Takahashi M. Ogasawara K. Takeda K et at. LPS induces NKl.l+ alpha beta T-cells with potent cytotoxicity in the liver of mice via production of IL-12 from Kupffer cells. ] Immunol 1996; 156(7):2436-2442. 100. D'Agostini C. Pica F. Febbraro G et at. Antitumour effect ofOM-174 and cyclophosphamide on murine B16 melanoma in different experimental conditions . Inr Immunopharmacol 2005; 5(7-8):1205-1212. 101. Bhardwaj N. Processing and presentation of antigens by dendritic cells: implications for vaccines. Trends Mol Med 2001; 7(9) :388-394. 102. Parr 1, Wheeler E. Alexander P. Similarities of the anti-tumour actions of endotoxin, lipid A and double-stranded RNA. Br] Cancer 1973; 27(5):370 -389. 103. Andreani V. Gatti G. Simonella L et at. Activation of Toll-like receptor 4 on tumor cells in vitro inhibits subsequent tumor growth in vivo. Cancer Res 2007; 67(21):10519-10527. 104. Matsushita K. Kobayashi M. Totsuka Yet at. ONO-4007 induces specific anti-tumor immunity mediated by tumor necrosis factor-alpha. Anticancer Drugs 1998; 9(3) :273-282. 105. Matsumoto N. Aze Y, Akimoto A ee at. Restoration of immune responses in tumor-bearing mice by ON0-4007. an antitumor lipid A derivative. Immunopharmacology 1997; 36(1):69-78. 106. Akimoro T. Kumazawa E.]imbo T et al. DT-5461a. an antitumor synthetic lipid a analog. causesselective blood flow reduction in tumor tissue. Anticancer Res 1995; 15(1):105-107. 107. limbo T. Akirnoto T. Tohgo A. Systemic administrat ion of a synthetic lipid A derivative. DT-546la. reduces tumor blood flow through endogenous TNF production in hepatic cancer model of VX2 carcinoma in rabbits. Anticancer Res 1996; 16(1):359-364. 108. Saiki 1. Saro K. Yoo YC et al. Inhibition of tumor-induced angiogenesis by the administration of recombinant interferon-gamma followed by a synthetic lipid-A subunit analogue (GLA-60).Int] Cancer 1992; 51(4):641 -645.

CHAPTERS

Lipid A-Mediated Tolerance and Cancer Therapy Cheryl E. Rockwell, David. C. Morrison and Nilofer Qureshi"

Abstract

T

he term "tolerance", from an immunological perspective, broadly encompasses a number of phenomena, but generally refers to a diminished responsiveness to LPS and/or other microbial products. With the discovery that many of the immunological, physiological and/or pathophysiological effects ofLPS can be attributed to the lipid A moiety ofthe LPS molecule , a number of different lipid A analogs were synthesized with the goal of developing a drug that could be used clinically to treat cancer. In many instances, the development of tolerance to the lipid A congeners confounded the utility of these analogs as cancer therapeutics. In certain circumstances, however, the development oftolerance in patients has been utilized therapeutically to protect immunosuppressed patients from sepsis. Although numerous studies have been designed to investigate the development oftolerance. the underlying molecular mechanism remains unclear. Th is may be due, in part, to differences in the experimental models used, the sources and types of microbes and microbial products studied, kinetics ofresponses , and/or other experimental conditions. Nonetheless, a number of different signaling pathways have been identified as potentially modulating and/or triggering the development of tolerance. Though complex and incompletely understood, the capacity of tolerance to impact lipid A-based therapeutics, either positively or negatively,is inarguable, thus underscoring the necessity for further investigation toward elucidating the mechanisms contributing to the development of tolerance to lipid A and its analogs.

Early, Late and Cross Tolerance

The occurrence of tolerance or host unre sponsiveness in animals and humans administered multiple doses ofmicrobe or microbial products has long been recognized by scientists and physicians with published reports appearing in professional journals dating back to the 19th century.1 Many of the very early obser vations focused largely upon the establishment of pyrogen tolerance in which animals treated with a microbe or a microbial product exhibited a refractory state for the development of fever, or a markedly diminished fever response, upon subsequent treatments with the same or a related microbe or product. Following the identification and purification of lipopolysaccharide (LPS) in th e 20th century, it was determined that the microbial product, LPS itself, can be highly tolerogenic.' In this respect , many ofthe pleiotropic effects ofLPS, including fever, induction of cytokine production and even mortality are absent or markedly diminished upon repeated administration ofLPS. From these observations, it has been postulated that tolerance serves to protect the host from the detrimental consequences ofthe robust and extensive inflammatory responses that follow exposure to LPS. Tolerance cannot, however, be characterized as a 'Corresponding Author: Nilofer Qureshi-Department of Basic Medical Science, School of Medicine, ShocklTrauma Research Center, University of Missouri, 2411 Holmes Street, Kansas City, MO 64108, USA. Email: qureshin @umkc.edu

LipidA in Cancer Therapy, edited by jean-Francois Jeannin. ©2009 Landes Bioscience and Springer Science+Business Media.

82

Lipid A in Cancer Therapy

global downregulation ofresponsiveness as some LPS-responsive characteristics remain unchanged or in some cases can actually be upregulated in experimental models oftolerance. In studies carried out by numerous investigators over the years to clarify the concept of tolerance, a variety ofterms have been used to designate the observed alteration in immune responses that occur upon repeated administration of LPS. Many alternative terminologies in lieu of the word "tolerance" have been suggested, these include reprogramming, deactivation, adaptation, refractoriness, hypo responsiveness and desensitization. All are arguably accurate to some extent, although many ofthese should not be used interchangeably, as some ofthese terms refer to tolerance that can be established only under specific conditions. Numerous studies have been dedicated to the elucidation ofLPS structures from various microbes in order to account at the molecular level for LPS tolerance. From these studies, it has been established that the lipid A moiety ofLPS is responsible for almost all ofthe described activities ofLPS.3.gAccordingly, the induction oftolerance is also routinely observed with repeated administration ofhighly purified lipid A; responses to which are virtually indistinguishable from those observed with repeated LPS adminlstration," In addition to tolerance to repeated lipid A/LPS treatment, cross-tolerance can be obtained in which other microbial products, such as lipoteichoic acid, induce tolerance to the effects oflipid A/LPS and vice versa.'?" The concept oftolerance is also distinguished temporally on the basis ofthe time required to establish this phenomenon. In this regard , both early and late tolerance have been described. Early tolerance is usually observed within hours ofLPS administration and may last up to a week or more . During early tolerance, suppression of cytokine production, fever and endotoxin lethality occur independently of antibody production.P" Cross-tolerance is generally associated with early, but not late tolerance. In contrast to early tolerance, late tolerance is induced many days after LPS treatment when early tolerance has generally ceased or markedly waned. Late tolerance is characterized by the production of antibodies directed against LPS, which confers the specificity and lack of cross-tolerance observed in this phase. The following discussion will be driven primarily toward an examination ofthe consequences ofearly lipid A/LPS tolerance. The consequences ofthe development ofearly lipid A/LPS tolerance on the therapeutic use of lipid A is two-fold. While subtoxic doses oflipid A may be used to either mitigate the side effects of lipid A/LPS treatment and/or protect against endotoxin shock and thus may be considered beneficial, the diminished therapeutic efficacy of lipid A in repeated administration over time may also be considered a deleterious effect. Consequently, a number oflipid A analogs have been developed to both decrease tolerance in order to increase efficacy as well as to increase tolerance to either diminish the associated toxicity oflipid A/LPS treatment or protect against endotoxin shock in patients susceptible to sepsis.9•16-20

The Isolation ofVarious Lipid A Structures and Synthesis ofAnalogs

An extensive array of lipid A molecules and structurally-related analogs have been isolated and/or synthesized in an effort to identify molecules that could act as competitive inhibitors and/or be used to reduce the toxicity of LPS. Indeed, comparisons of the biological potencies of lipid A structures derived from different bacterial strains and methods of preparation have been reported over the past several decadesY-27Monophosphoryllipid A (MPL) was among the first structures to be purified and structurally characterized by our laboratory.24.27The results of these seminal studies led to the initial determination of the complete structure of lipid A. The subsequent development of a number of highly purified MPL analogs (with respect to number of fatty acyl groups) led to the analysis of numerous analogs in regard to toxicity and efficacy in tumor regression. MPL was found to be generally less toxic than diphosphoryl lipid A and intact LPS, which generated interest in its potential therapeutic use.26,27The relative efficacy and toxicity of MPL is, however, somewhat dependent upon the purity of the preparation and the strain of bacteria from which it is derived . In general, hexaacylated lipid A structures were shown to be more toxic than the pentacylated or tetraacylated structures and the chain length of the fatty acyl groups was also considered an important factor in roxlciry," Although generally less toxic

LipidA-Mediated Tolerance and Cancer Therapy

83

than most diphosphoryllipid A structures and LPS, MPL exhibits some agonist activity, albeit at a markedly lower level than that of intact LPS.28 In contrast, the pentaacyl diphosphoryllipid A derived from LPS of Rhodobacter sphaeroides (RsDPLA) was purified and characterized in our laboratory for its ability to act as a nontoxic effective antagonist ofLPS and agonist lipid A moieties in both human and murine cells.29'33 In addition to native lipid A structures, a relatively large number ofsynthetic lipid A analogs were later developed and screened for toxicity as well as LPS-mimetic or antagonistic activity.2o.34.3s

Relevance ofTolerance to the Use ofLPS/Lipid A in Cancer

The phenomenon oftolerance in immunotherapy for the treatment ofcancer has been observed as long as LPS/lipid A has been used as a cancer therapeutic. Indeed, to avoid the induction of tolerance, the late nineteenth century physician , Dr. William Coley, found it necessary to use incremental doses ofa toxin formulation prepared from culture filtrates ofbacteria in treating his sarcoma and carcinoma patients. "Coley's toxin," as the aforementioned preparation came to be called, was found to have mixed success, but was used for many years by physicians for a variety of different malignancies. The tolerance induced by Coley's toxin may have been the result of LPS/lipid A tolerance, cross tolerance, or a combination of both as Coley's toxin was comprised of killed bacteria of both gram-positive and gram-negative strains.' While Coley's toxin was comprised ofa variety ofmicrobial products that included LPS, LPS alone was also found to cause tumor regression.' It was later determined that LPS-induced hemorrhagic necrosis of tumors is primarily due to the induction ofa serum factor, termed tumor necrosis factor (TNF).36 Further investigation revealed that the lipid A portion ofLPS was primarily responsible for the induction ofTNFa.37 In the late twentieth century, the results ofseveral clinical trials using LPS as a therapy were reported in cancer patients (Table 1). While purified LPS was confirmed to have positive antitumor activity in humans, both the toxicity ofLPS as well as the relatively rapid induction of tolerance by LPS detracted from its overall utility as a cancer chemotherapeutic. The decreased antitumor activity ofLPS due to tolerance was similar to the reduced antitumor activity observed with multiple administrations ofTNFa, suggesting that the tolerance observed in vivo may well be due to both reduced TNFa activity as well as the diminished induction ofTNFa by repeated LPS administration."

LPS in ClinicalandAnimalStudies

As discussed previously, one ofthe primary problems with the use of LPS/lipid A as a thera peutic intervention in the treatment ofcancer is the rapid induction oftolerance that diminishes the efficacy of the treatment. To address the problem of decreased antitumor activit y of LPS after multiple administrations, Mackensen et al evaluated the effect of endotoxin tolerance upon cytokine production in cancer patients following repeated daily LPS injections. Patients treated a

Table

1.

The role of tolerance in the use of various lipid A moieties in cancer

Lipid A Moiety LPS(lipid A) Tolerin MPL/Detox

Activity

Roleof Tolerance

Tumor regression/antimetastasis Protection against sepsis Adjuvant for cancer vacc ine Tumor regression

Detrimental: diminishes efficacy Beneficial : prevents endotoxi n shock Unknown Detrimental, however induces less tol erance than LPS

DT-5461

Tumor regression/antimetastasis, protection against sepsis

SDZMRL 953

Protect ion against sepsis

Detrimental for anti-tumor activity, but benefi cial for prevention of LPS shock Beneficial : prevent s endotox in shock

ONO -4007

84

Lipid A in Cancer Therapy

second time with LPS, 24 hours after an initial LPS treatment. had significantly decreased levels ofcirculatingTNFa. interleukin (IL)-6, IL-8. granulocyte colony stimulatingfaetor (G-CSF) and macrophage colony stimulating factor (M -CSF). The levels ofmost ofthese mediators continued to decline to baseline levels upon subsequent daily injections ofLPS for 5 days. Interestingly. IL-6 levels plateaued upon the third day ofLPS injections with no further decline in days 4 and S. The authors also reported that side effects, such as fever and chills. increased on the second day ofLPS treatment, which did not correlate with the reported decreases in cytokine levels. In addition to tolerance induced by daily LPS injections. the authors also noted that tolerance can be induced with repeated LPS injections of 1 week and 2 week intervals as well." In initial clinical trials carried out by the same group to evaluate LPS therapeutic benefit. induction of tolerance was reduced by using escalating doses of LPS.40 In a subsequent Phase I trial, tolerance was further reduced through the use ofboth escalating doses ofLPS as well as an increased treatment interval of two weeks. which results in less tolerance induction than shorter treatment intervals. While moderate or strong antitumor activity with this treatment regimen was observed in a few ofthe reported cases. most cases did not demonstrate significant antitumor effects." The authors hypothesized that the relatively limited antitumor effect ofLPS was. at least in part, due to the induction ofLPS rolerance.v -" The same group also conducted a Phase I clinical study in which interferon y (IFNy) was administered in addition to LPS, in an effort to prevent LPS tolerance and thereby increase the therapeutic efficacy of the treatment. Pretreatment with IFNy not only prevented LPS tolerance induction but. in fact, induced higher levels ofTNFa. IL-6 and G-CSF than in the initial administration with LPS alone. Conversely. the downregulation ofIL-8 upon repeated LPS administration was unaffected by IFNy pretreatment. Whether the diminished LPS tolerance observed with IFNy pretreatment correlated with improved antitumor activity was not reported, however," In addition to clinical trials in humans, the induction ofLPS/lipid A-mediated tolerance has also been widely examined in experimental animal models. Numerous tactics have been found to delay or prevent tolerance to LPS/lipid A in animals, including treatment with recombinant IFN~, nitric oxide synthase inhibitors, p38/stress-activated protein kinase-2 inhibitors, administration of flt3 ligand (a growth factor important for dendritic cell differentiation) and many other methods as well.45-48 Interestingly, one group recently attempted to counteract the effects of LPS/lipid A-mediated tolerance by using LPS in combination with cytotoxic drugs, such as 4'-(9-acridinylamino)-methansulfone-m-aniside, cyclophosphamide. l-octadecyl-z-merhoxy-rac-glycero-3-phosphocholine and hexadecylphosphocholine. The authors reported, however, that they were unable to produce sufficient antic ancer therapy with acceptable toxicity using this

approach."

While the anticancer activity ofLPS in its native form has been extensivelyinvestigated, the potential therapeutic efficacyofirradiated LPS in cancer patients has also been reported. Presumably, the irradiation ofLPS produces a variety ofdifferent lipid A structures. The overall goal ofclinical studies with irradiated LPS. which is currently registered under the market name ofTolerin°, was somewhat different than the aforementioned studies with native LPS. Tolerin was designed to be administered with the purpose ofinducing tolerance to LPS and at the same time, boosting natural immunity with the intent ofprotecting highly susceptible immunosuppressed cancer patients from sepsis and subsequent lethal septic shock. The results from these recently published studies have documented that Tolerin was well-tolerated by the patients and increased natural resistance which, importantly, correlated with a decreased incidence ofinfection in these patienrs." Although the majority ofstudies that have investigated LPS/lipid A in the treatment ofcancer have focused upon their effects on inhibition of tumor growth and/or protection against sepsis, other studies have been directed toward the reverse, namely the effects oftumors upon endotoxin lethality. Interestingly, Berendt et al reported an increased sensitivity to endotoxin in mice bearing LPS-sensitive tumors, which the authors maintain is analogous to increased endotoxin lethality in mice infected with pathogens that cause systemic macrophage activation. Furthermore, the authors report systemic activation of macrophages in mice bearing LPS-sensitive tumors which

Lipid A-MediatedTolerance and Cancer Therapy

85

they correlate to the increased endotoxin lethality observed in these anirnals.t' In contrast, numerous studies have reported that many tumors are capable of inducing a tolerant state in immune cells.52.55While the term "tolerance" is routinely used to describe both the hyporesponsive state of leukocytes induced by tumor cellsas well as the diminished cytokine production, fever and lethality following multiple administrations ofLPS/lipid A, these are two separate phenomena. Although there may be similarities between these two different types oftolerance, the two phenomena can be distinguished from one another by the causative agents, which are presumably one or more factors produced by tumor cells and microbial products, respectively. In addition to these direct effects of the tumors themselves upon the ability of immune cells to respond to LPS, surgical removal oftumors can also impact responsiveness ofimmune cells to LPS. A general state ofirnmunosuppression is routinely observed in humans and animals following surgery.56 Moreover, it has been reported recently that cryosurgery oftumor tissue induces tolerance to endotoxin-mediated lethality, suggesting that this surgical procedure may also be protective against septic shock in cancer patienrs." Collectively, these studies suggest that the effects of tumors upon immune cell activation overall and LPS responsiveness specifically, appear to be complex and seem to differ based upon the nature ofthe tumor. In addition, surgical removal oftumors may well contribute to the induction of tolerance to LPS/lipid A , which may have the dual effect ofdiminishing the efficacy oflipid A treatments as well as protecting the patient from septic shock.

Monophosphoryl LipidA andLipidA Adjuvants

Following the isolation, derivatization and synthesis of various lipid A structures, it was quickly established that some of the beneficial effects of the identified nontoxic lipid A moieties had therapeutic potential. For instance, SDZ MRL 953 and MPL were found to enhance host defenses against subsequent bacterial infection through induction of G -CSF, M-CSF and other cytokines.26•58.60 At the same time, however, these compounds induced refractoriness (tolerance) to LPS toxicity through reduced secretion ofinflammatory cytokines, while host cell phagocytic activity was maintained. Some ofthe nontoxic lipid A moieties, such as MPL, also demonstrated tumor regression activity.61 Similar to LPS, however, MPL has also been shown to induce tolerance in both experimental animals as well as in patients enrolled in clinical trials.26.28.62.64 Due to these confounding factors, MPL is not currently in use clinically for its tumor regression activity per se, but remains important in cancer therapy as an immunoadjuvant in cancer vaccines {see chap. 10).65 Because macrophage activation is critical to the adjuvant activity of lipid A (and in the development oftolerance), lipid A adjuvancy may also be susceptible to tolerance, depending upon experimental conditions and treatment administration.

Synthetic LipidA Analog, ONO-4007

In addition to the various native forms oflipid A derived from different bacterial strains, numerous lipid A analogs have been synthesized and screened for antitumor activity. Several synthetic lipid A analogs have demonstrated potential as cancer chemotherapeutics, but the induction of tolerance upon repeated administration has only been reported for a fraction ofthese. 0 N 0-4007, a synthetic triacylated monosaccharidic lipid A analog, has been the focus of numerous studies in both animals and humans, which have shown that it exhibits lower toxicity than LPS and causes tumor regression. 66' 68 While ONO-4007 has been shown to induce tolerance in animals using several different models, it induces less tolerance than that induced by LPS and synthetic E. coli-type lipid A {LA_15_PP)P.69.7o Interestingly, differential tolerance to ON0-4007 was observed in different tissue types. While significant tolerance to ON0-4007-mediated cytokine induction was observed in serum and liver tissue, no tolerance was observed in tumor tissue extracts as assessed by TNFa production in tumor-bearing mice treated with ON0-4007 at 8, 12 and 15 days following tumor Implantation.'? The results of further investigations have established that the tumor tissue did become hyporesponsive to stimulation with ONO-4007 as measured by TNFa induction, but tumoral responses recovered more quickly than responses in liver and serum and were completely responsive by 72 hours. Similarly, tumor infiltrating macrophages recovered from ONO-4007-rnediared

86

Lipid A in Cancer Therapy

hyporesponsiveness within 72 hours after initial exposure. The authors hypothesized that the selective recovery of tumor tissues may. at least in part, be due to constant recru itment of macrophages to tumor tissue. In addition. the authors demonstrate that repeated injections of LA-15-PP into mice enhanced its clearance from blood circulation, whereas the clearance of ON0-4007 was stable even following multiple administrations, suggesting that pharmacokinetics also playa role in the differences in tolerance between these two lipid A structures. Moreover. TNFa tissue levels peaked 1-2 hours following ONO-4007 treatment and then decreased in the spleen and liver, but remained elevated for at least 6 hours in tumor tissue.7l •n

Synthetic Compound, DT-5461

Like LPS and ON0-4007, tolerance to the synthetic tetraacylated lipid A analog, DT- 5461, is also observed. Tolerance to DT-5451 occurs one day after treatment with a return to responsiveness observed 3-5 days later," In experimental animal models of cancer using prostaglandin ~ (PGE2)-producingtumors. a combined treatment ofindomethacin and DT- 5461 wasshown to have significant antitumor activity and an additive effect upon survival. The authors have hypothesized that the antitumor effect of the combined therapy is due to a combination of TNFa activity, as well as the inhibition ofPGE 2 production." Interestingly, PGE2 has also been hypothesized to be a mediator ofLPS/lipid A-induced tolerance." Furthermore and ofsome interest, cyclooxygenase inhibitors have been reported to prevent the induction oftolerance to LPS,76 While the majority ofstudies ofDT-5461 have focused upon its antitumor activity, DT-5461 has also been shown to be protective against endotoxernia. DT-5461 pretreatment induced significant tolerance to lethal LPS exposure in mice." Similar to studies described earlier with Tolerin, the induction of tolerance by DT-5461 was found to be beneficial therapeutically for the protection of immunosuppressed patients from sepsis, including cancer patients unde rgoing chemotherapy or radiation treatment. In addition to the induction oftolerance by DT-5461 itself, LPS also induces tolerance to DT-5461 as well, strongly suggesting that these two compounds act through similar cellular mechanisms . In mice injected with LPS at daily intervals for a week, the antimetastatic activity ofDT-5461 was significantly reduced, further supporting the conclusion that the mechanisms of tolerance by LPS and DT-5461 are identical or cross-tolerance occurs between the two mediators."

Synthetic Compound, SDZ MRL 953

Similar to Tolerin and DT-5461, the LPS tolerance induced by the synthetic triacylated lipid A analog, SDZ MRL 953, has been evaluated for its protective effects against sepsis in cancer parients. " SDZ MRL 953 pretreatment inhibited LPS-induced TNFa, IL-l~, IL-8, IL-6 and G-CSF serum levels,suggestingthat it induces tolerance to LPS and might be effectiveas a prophylactic treatment for patients who may otherwise be susceptible to sepsis." Further clinical studies will be needed, however, to determine the overall efficacyofSDZ MRL 953 in the prevention of gram-negative infections and septic shock in immunosuppressed cancer patients.

Mechanisms of EarlyLPS/Lipid A-Mediated Tolerance

Because many investigators have demonstrated that lipid A is the active component of LPS and is also capable ofinducing LPS tolerance , presumably most ofthe LPS-mediated activity that occurs during early tolerance can be attributed to the lipid A portion ofthe molecule.28•8o,81 These conclusions are also strongly supported by the results ofstudies using a variety oflipid A analogs as described in previous sections. Despite hundreds ofpapers that have been published concerning LPS/lipid A-mediated early tolerance, the underlying mechanism still remains to be determined. Indeed, numerous hypotheses and supportive studies suggest that multiple mechanisms may play a role in tolerance. If multiple pathways do exist, they may be activated concurrently or at temporally distinct points in the activation/deactivation pathway sequences. Alternatively, different mechanisms may operate in a manner that is distinct from one another. The activation of one particular mechanism rather than another may depend upon the specific cell type involved, the local environment within the host, or the experimental conditions in vitro . While the majority

87

Lipid A-Mediated Tolerance and Cancer Therapy

Cytoldne Receptor (I. •. lFNyl

Figure 1. Schematic diagram of hypothetical mechanisms of tolerance within the TLR4 pathway. While ST2, IRAK-M and the splice variant of MyD88 (MyD88s) are thought to suppress association of IRAK-l with MyD88 through competitive inhibition. TaLLlP has been shown to bind to IRAK-l and inhibits its activity. saCS-l and saCS-3 are inhibitors of the jAK/STAT pathway. Wh ile the mechanism by which SHIP induces tolerance is unknown, it is believed to inhibit the NFKB signaling pathway.

of studies of early tolerance to LPS/lipid A have focused upon macrophages, a number of other cell types have also been shown to be susceptible to tolerance. 82•83 Because detailed reviews ofthe numerous studies focusing upon the mechanisms oftolerance have recently been published. only a broad overview will be included in this chapter. The reader is referred to a number of excellent reviewsfor a more comprehensive discussion concerning the role ofdifferent LPS/lipid A signaling pathways in the induction of tolerance. 13.75.84.85

Toll-Like Receptors, Associated Signaling Molecules andNegative Regulators

The members of the toll-like receptor (TLR) family are unique in their ability to recognize pathogen associated molecular patterns (PAMPs) found in variety ofmicrobial products. including LPS/lipid A . LPS has been shown to activate cells through the TLR4 pathway and to a lesser extent, through TLR2. CD 14 and MD-2 are both proteins that have been shown to associate with and also be required for. TLR4-dependent activity and are therefore important for activation of cells by LPS/lipid A (Fig. 1). Upon activation ofTLR4 by LPS, the adapter protein. myeloid differentiation factor 88 (MyD88) is recruited to the cytoplasmic domain ofTLR4, where it associates with and then activates the IL-1 receptor associated kinase-I (IRAK-1). Activated IRAK-1 then dissociates from the TLR4 complex and subsequently binds and activates TNF receptor associated factor -6 (TRAF-6). which in turn activates TGF~-activated kinase-I (TAKl) and nuclear factor of KB-inducing kinase (NIK). The activation ofTAKl and NIK results in the activation ofthe mitogen activated protein kinase (MAPK) and nuclear factor ofKB (NFKB) pathways.7S.85 The reader is referred to previous chapters of this volume for more information concerning the signaling pathways oflipid A/LPS. Numerous studies have reported the role ofTLRs and associated proteins in the induction of tolerance. Although it would be intuitive that downregulation ofTLR4 or another component of

88

Lipid A in Cancer lherapy

the TLR4 receptor complex may be responsible for LPS/lipid A-induced tolerance , this remains a somewhat controversial issue. While some studies have reported a downregulation of TLR4 receptor protein expression following LPS treatment, other studies have demonstrated that TLR4 expression is either unaffected or sometimes even increased upon LPS administration.83.86·93 In contrast to TLR4, studies ofthe effects ofLPS on TLR2 expression have been more conclusive, with no downregulation of TLR2 observed in most models of tolerance.P:" In addition, the expression levels of CD 14 and MD-2 have also been evaluated in numerous studies. The results of most studies examining CD14 expression have shown no decrease in protein or transcript levels following LPS treatment.88.91.94.96 Similar to TLR4, however, the results with MD-2 have been inconsistent. While many studies demonstrate little change in MD-2 expression with LPS/ lipid A treatment, some studie s show decreased MD-2 transcription and surface expression of the TLR4/MD-2 complex.89.91.94.97Interestingly, tolerance has been demonstrated in HEK293T cells in which TLR4, CD14 and MD-2 have all been overexpressed, suggesting that decreased CD14/TLR4/MD-2 expression is not required for the induction of tolerance in this model," Overall, these studies suggest that, while there may be decreased expression ofTLR4/MD-2 upon LPS/lipid A administration under certain experimental conditions, it is by no means a universal requirement for tolerance . The role ofa number ofdifferent signaling molecules downstream ofthe CD 14/TLR4/MD-2 complex has also been investigated. Although expression levels of MyD88 have not been found to be decreased in tolerant monocytes, a marked inhibition of MyD88/TLR4 association has been reported." In addition, a splice variant ofMyD88 that is expressed upon LPS treatment and functions to inhibit LPS signaling, has garnered considerable attention as a potential mechanism oftolerance induceion.99•1OO Similarly, IRAK-l activity and MyD88IIRAK-1 association have been reported to be diminished in LPS-tolerant cells.90·101.102 Moreover, upregulation of the inactive kinase, IRAK-M, may also playa role in the induction of tolerance in monoeyces. IRAK-M is markedly upregulated upon LPS treatment, causing inhibition ofLPS signaling, possibly through competitive inhibition ofIRAK-l binding to MyD88. Ofparticular relevance, the induction of tolerance is significantly diminished in macrophages lacking IRAK-M ascompared to macrophages derived from wild-type mice.103.104 In addition to IRAK-M , other negative regulators ofLPS signaling have also been investigated. Suppressor ofcytokine signaling-l (SOCS-l) and SOCS-3 have been shown to be rapidly induced upon LPS exposure and these proteins can serve as negative feedback regulators ofthe janus activated kinase (JAK)/signal transducers and activators of transcription (STAT) signaling cascade, which is the downstream mechanism of signal transduction of many cytokine receptors .IOS•I06 The results of studies examining the role of SOCS-l in tolerance have on ce again been mixed, with one report demonstrating an absence oftolerance to endotoxin lethality and LPS-mediated TN Fa secretion in SOCS-l -/- mice, while a different group demonstrated no difference in the induction ofLPS tolerance to IL-12 secretion in SOCS-l -/- macrophages. As a consequence, the role of SOCS-l in tolerance remains inconclusive. F'{" ST2, a member of the Toll-Ill superfamily, is another inhibitor of LPS signaling. There is evidence to suggest that ST2 suppresses signaling ofboth the TLR4 and the IL-l receptor through the sequestration ofMyD88 and Mal adaptor proteins. Although the role ofST2 in tolerance has not yet been extensively investigated, recent reports demonstrating a failure ofST2-deficient mice to develop tolerance to endotoxin lethality, as well as LPS-mediated IL-6 and IL-l2 production, suggest that ST2 may also playa role in the induction of tolerance.F"!" These promising preliminary findings are likely to provide the basis for additional study ofmechanisms ofLPS/lipid A-mediated tolerance indu ction. SH 2-conraining inositol phosphatase (SHIP) and toll interacting protein (TOLLIP) are also negative regulators ofLPS signaling. While TOLLIP mRNA and protein levels have been reported to be upregulated in LPS-tolerant cells, the role of this protein in tolerance has not been extensively investigated. 1I 1.112 Interestingly, the results of recent stu dies with TOLLIP null mice indicate that TOLLIP can also act as an activator ofLPS signaling, at least under certain

Lipid A-Mediated Tolerance and Cancer Therapy

89

circumsrances.!" In comparison to TOLLIP, there is stronger evidence for involvement of SHIP in endotoxin tolerance. SHIP protein levels are significantly elevated following LPS treatment, a response which is mediated by LPS-induced TGFp secretion. Accordingly, treatment with antibodies against TGFp blocks LPS-induced tolerance. Notably, LPS-mediated tolerance cannot be induced in SHIP null mice."! The suppressive effects ofNO on cytokine production, as well as on immune cell proliferation and growth, have been extensively described.!" As a result, the effect of NO upon tolerance has also been investigated. Supportive evidence for NO involvement in the induction of tolerance includes increased NO production by tolerant peritoneal macrophages, inhibition of tolerance to endotoxin lethality by NO synthase inhibitors and increased survival to a lethal dose ofLPS in rats treated with an NO donor.116-120While many studies report increased NO production and/or iNOS transcription in tolerant macrophages, other studies have shown decreased NO production and NO synthase activity by macrophages in experimental models of tolerance.!" Similarly and perhaps not totally unexpectedly, contradictory evidence for the role of NO in tolerance is also found in investigations with knockout mice. While Dias et al have recently reported that iNOS null mice do not become tolerant to LPS-mediated pyrogenicity, Zingarelli er al earlier demonstrated tolerance to LPS-mediated lethal ity and TNFa production in iNOS null mice. 122.m The differences between the two studies may be dependent upon the different doses ofLPS used, the endpoints measured, or the treatment regimens which also differed substantially between the two sets ofexperiments.

Corticosteroids, Anti-Inflammatory Cytokines andProstaglandins

In addition to participation ofa number ofcomponents ofthe TLR4 pathway and the negative regulators associated with this pathway in the development of tolerance to LPS/lipid A, other mechanisms have also been proposed to mediate this phenomenon. LPS-induced glucocorticoid secretion has been well-documented and results in the suppression ofa variety ofdifferent immune cell types . As such, the role ofglucocorticoids as potential mediators oftolerance has remained an area ofinterest for decades. Despite numerous studies to address this hypothesis, however, strong evidence for the involvement ofglucocorticoids in the induction of tolerance has been relatively sparce." While currently inconclusive, the role of glucocorticoids in tolerance continues to be investigated. Anti-inflammatory cytokines, such as IL-l a and TGFp, have also been suggested to contribute to the induction oftolerance to LPS/lipid A treatment. Support for this conc ept comes from the findings of studies showing significant downregulation of TNFa and other pro-inflammatory mediators by IL_1O.124.12s Furthermore, treatment with an antibody specific for IL-l 0 results in an inhibition ofthe induction ofLPS/lipid A-mediated tolerance. 126 Conversely, tolerance to LPS/ lipid A treatment can be readily induced in IL-l a knockout mice, suggesting that IL-l a cannot be the sole mediator of tolerance.!" Indeed, induction ofIL-lO by LPS itselfis diminished upon subsequent administration ofLPS under certain experimental conditions.F" Similar to IL-la, pretreatment ofisolated human peripheral blood mononuclear cellswith recombinant TGFp also suppresses LPS-mediated TNFa secretion and a blocking antibody against TGFp also diminishes the development of tolerance to LPS. 126.128 In addition to immunosuppressive cytokines, the synt hesis and release ofimmunosuppressive prostaglandins, such as PGE2, have also been hypothesized to cont ribute to the induction oftolerance in some experimental models . The basis for this hypothesis is that PGE 2levels have been shown to be highly elevated in tolerant cells reexposed to endotoxin in a number ofdifferent animal and human models. Moreover, PGE 2inhibits cytokine production in activated macrophages and Iymphocyres.F r!" Further stu dies are needed to more completely elucidate the role ofPGE2 in the induction oftolerance, however.

TranscriptionalMediators

Peroxisome proliferator activated receptor y (PPARy) activation has been correlated with inhibition of macrophage activation by LPS, as assessed by cytokine and NO production. Fr!" Consequently, the role ofPPARy in LPS-mediated tolerance has been examined. Increased PPARy

90

Lipid A in Cancer Therapy

binding to the PPAR response element (PPRE) and PPARy transcriptional activity have been observed in LPS-tolerant cells,suggesting possible involvement ofPPARy in endotoxin tolerance.P'' Furthermore, the use of blocking oligonucleotides for the PPRE inhibited tolerance induced by LPS. Collectively, these preliminary studies suggest a potential role for PPARy in tolerance, but more extensive investigations are required before definitive conclusions can be reached. In part because so many LPS-responsive genes have been shown to be regulated by NFKB, there has been considerable interest in the numerous reports ofdecreased NFKB activation in the development ofLPS tolerance , which has been widely observed in a variety ofexperimental models oftolerance. The results ofsome ofthese studies suggest that the decrease in NFKB activity may be due to an increase in the formation ofpSO/pSOhomodirners, which are transactivationally inactive and therefore would be expected to antagonize the active pSO/p6S heterodimer by competitively binding to KB sites in the promoters ofLPS-responsive genes," This hypothesis is supported by the reportedly increased levels ofpSO/pSO homodimers in tolerant cells and the failure to induce tolerance in pSO-deficient mice .P? Contradictory evidence has, however, also been presented, demonstrating that LPS tolerance could still be induced in pSO - 1- mice, as assessed by IL-12 and TNFa production by splenocyres.P" Collectively, these data suggest that, while decreased NFKB transcriptional activity is likely to be a causative factor in the induction of tolerance in a number of different experimental models, current data suggest th at pSO/pSO homodimers are not likely to be the only mechanism responsible for the diminished NFKB activity in various models of LPS tolerance.

The Role ofthe Proteasome

Our recent data have provided convincing evidence for a prominent role ofthe proteasome, a cytoplasmic organelle with multiple protease activities, in LPS signaling and subsequent development of inflammatory and immune responses . Structurally, proteasomes exist as multi-subunit complexes, consisting of a number of distinct, well-characterized, proteinsy9The so-called 26S proteasome complex (2.5 MDa) is comprised of a 20S proteasome, which exhibits proteolytic activity and a 19S cap, which provides regulatory functions.!" The 20S proteasome has been defined structurally as a hollow, cylindrical, multi-protein structure composed of28 protein subunits that are derived from 14 distinct gene products.141.142 The protein subunits ofthe proteasome are arranged in four heptameric rings shaped approximately as a barrel. The three proteases of the proteasome are X (LMP7) (chymotrypsin-like protease activity) , Y (LMP2) [posc-gluramase protease activity) and Z (MECL-l) (trypsin-like protease activity) and these have been described in detail. 142.143 The protease activities ofthe proteasome have been shown to be regulated by IFNy. Subunits LMP7, LMP2 and MECL-l of the 20S proteasome are recognized as IFNy -inducible proreasome-associated ~-subunits. There is an overproduction ofthese subunits due to IFNy produced early on during an inflammatory response, resulting in the introduction ofthe subunits into newly assembled proteasomes, which have been termed immunoproteasomes. Inununoproteasomes appear to have enhanced capability for generating class I MHC-binding peptides, as compared with "standard" proteasomes, cleaving more efficiently afier basic or hydrophobic residues and less efficiently mer acidic residues.!" The role of the proteasome in LPS-induced inflammation had not been extensively pursued until our demonstration that LPS binds specifically to Al (C2) and B4 (N3) proteins ofthe 20S proteasome complex.!" Mer demonstrating that LPS binds proteasome subunits, we then assessed the potential physiological relevance ofthese interactions. To this end , we first carried out studies to determine the extent to which LPS modulates the proteasorne's proteolytic activity. We demonstrated that the addition ofLPS to partially purified proteasomes in vitro activated the chymotrypsin-like and posr-gluramase activities ofmacrophage proteasomes.145.146 We next sought to determine the extent to which well-defined proteasome inhibitors might block LPS-induced inflammation. To address this question, we pretreated RAW 264 .7 macrophages with the well-characterized proteasome inhibitor, lactacystin, and observed a dose-dependent inhibition of LPS-induced

91

LipidA-Mediated Tolerance andCancer Therapy

NFKB IRAK-1

IRAK-M

"<,

\

SOCS-1 /

----{~~

SOCS-3

Z~~ ~

/ MyD88

\

TLR4 TRAF-6

Figure 2. Schematicdiagram of tolerance-related mediatorsthat are regulated by the proteasome at either the transcriptional or posttranscriptional levels. cytokine secretion.145.146 Furthermore, we found that pretreatment ofprimary murine macrophages with lactacystin inhibited the expression ofa spectrum of LPS-inducible genes, including IL-l~, IL-6, IL-12 p40 and p3S, COX-2 and iNOS. In addition, lactacystin also blocked the LPS-induced upregulation ofTLR2 mRNA and reduced constitutive levels ofTLR4 mRNA expression.Iv Ihe net effect of proteasome activation would appear to be enhancement ofTLR-mediated inflammatory responses, while proteasome inhibition would be predicted to suppress the inflammatory response. Our data demonstrate that more than 90% of LP S-responsive genes in peritoneal macrophages are regulated by the proteasorne.!" Furthermore, studies from our laboratory and others suggest that the proteasome regulates a number ofproteins involved in tolerance, including SOCS-l , SOCS-3, IRAK-M, IRAK-l, MyD88, TLR4 and others (Fig. 2) .147.148 In addition, the proteasome also regulates NFKB, a critical transcription factor for many LPS-responsive genes that has been shown to be dysregulated in LPS-tolerant cells. The role of the proteasome in tolerance remains largely untested thus far, however.

Mechanisms ofTolerance of Other Lipid A Structures and LPS Antagonists

In addition to lipid A moieties with agonist activity, there also exist a variety oflipid A analogs that that can function as LPS antagonists. The mechanism ofthe LPS antagonists is likely through the competitive inhibition ofLPS binding to either LPS binding molecules, such as LPS binding protein (LBP), or the TLRcomplex. Indeed, evidence for this has been presented for RsDPLA, the biologically inactive lipid A molecule from Rhodobacterspbaeroides, using colloidal gold particles to label both LPS and RsD PLA and electron microscopy to monitor cellular binding and internalization. Our studies conducted thus far suggest that RsDPLA competes with LPS in binding to LBP, CD 14 and TLR4.14~·151 Other lipid A analogs, including ONO-4007 and others described earlier, have been postulated to induce tolerance to LPS and lipid A through upregulation ofendogenous corticosteroids/" In addition, the inhibition of suppressor T-cell activity by MPL and RsDPLA has also been p roposed as a mechanism of tolerance for the se two lipid A molecules.Y'' '"

Conclusion

As summarized in this chapter, evidence has been presented to suggest that multiple LPS /lipid A-induced signal transduction intermediates and other mediators are involved in the induction of tolerance. It is likely that no single mechanism will emerge as playing the dominant role in this process. The continued investigation ofthe numerous factors implicated thus far in the development of tolerance to LPS/lipid A will help to determine to what extent and under which circumstances these various factors playa role in this phenomenon. In addition to the aforementioned factors that are currently being investigated, other factors , such as the proteasome that have not yet been widely

92

Lipid A in Cancer Therapy

Table 2. LPS-modulated, proteasome-dependent genes

SaCS3 SaCSl TLR2 NFKB2 STAT2 STATl TLR3 MAP3K8 IRAKM TAKl TLRl My088 C014 TLR4 TRAF6

LPS

LPS/lact

Lact

Description

256.51 52.35 10.32 10.20 9.17 6.77 5.74 4.36 3.46 3.32 2.83 2.50 2.24 -2 .83 -5.56

-4.02 12.87 6.82 3.60 6.68 4.84 2.77 1.86 1.66 -3 .09 -1.69 1.51 -1.22 -6.47 -2.33

-3.49 2.34 -1.74 1.23 -1.57 -1.48 -6.63 -3 .09 -1.38 2.18 - 1.45 -1.08 - 2.31 -3.89 2.28

Suppressor of cytokine signaling 3 Suppressor of cytokine signaling1 Toll-like receptor 2 Nuclear factor of kappa light geneenhancer in B-cells Signal transducer and activator of transcription 2 Signal transducerand activator of transcription 1 Toll-like receptor 3 Mitogen-activated protein kinase kinase kinase 8 Interleukin-l receptor associated kinase3 Mitogen-activated protein kinase kinase kinase7 Toll-like receptor 1 Myeloid differentiation pathway primary response gene C014 antigen Toll-like receptor 4 TNF-receptorassociated factor 6

Thioglycollate-elicited murine macrophages weretreatedwith the compounds,LPS and/or lactacystin (Lact). The gene expressionvalues are reported as average normalization ratios (modified from ref. 165). A data setcontaining geneidentifiers and their corresponding expressionvalueswas uploaded as an excel spreadsheet using the template provided in the application . Each gene identifier was mapped to its corresponding gene object in the ingenuity pathways knowledge base.

studied in tolerance, may also be involved. Studies from this laboratory strongly suggest a key role for the proteasome in LPS/lipid A signaling. This evidence includes the modulation ofproteasomal protease activity by LPS/lipid A, the degradation OfIKB by the proteasome and the subsequent activation ofNFKB that ultimately upregulates inflammatory mediators. In addition to that, proteasomal proteases have also been shown to degrade various mediators of LPS signaling , including TLR4. lRAK-l, etc . Furthermore. inhibition of the proteasome modulates the gene expression ofmany mediators ofLPS signaling, including many associated with tolerance (Table 2). The proteasome may also playa key role in tolerance such that when LP S-induced inflammatory mediators increase to a certain level, compensatory mechanisms are induced to trigger the development oftolerance. possibly by modulating the activities ofindividual proteasomal proteases. Interestingly. there is a great deal of overlap between many of the proposed mechanisms of tolerance and putative targets ofcancer therapy. The roles ofNFKB, lRAK-I , lRAK-M. SaCS-I, saCS-3 and MyD88 have been evaluated both in tolerance as well as in several models ofcancer. for example. Because proteasome inhibitors modulate gene expression ofthese signal transduction intermediates, they have also been investigated for their potential as cancer therapeutics. Indeed. the proteasome inhibitor, Bortezornib, has recently been approved by the FDA for the treatment of refractory multiple myeloma and has also shown promise in the treatment of lung cancer. as well as various types oflymphoma. As our knowledge oflipid A and its derivatives continues to expand, the therapeutic potential ofthese compounds has become evident. MPL has been used as an effective adjuvant for cancer vaccines and will likely continue to be used in future vaccine formulations. The development of nontoxic lipid A analogs. such as SDZ MRL 953 , that induce tolerance to LPS and thereby protect susceptible patients against endotoxin shock also shows therapeutic promise. Moreover, the development oflipid A analogs that induce less tolerance than LPS and exhibit greater efficacy in tumor regression show potential as cancer therapeutics as well. Because lipid A-mediated tolerance is a particularly complicated phenomenon that plays dual and opposite roles in the efficacy

LipidA-Mediated Tolerance and Cancer Therapy

93

of cancer therapeutics, elucidation of the mechanisms of tolerance is essential for the continued development oflipid A compounds into nontoxic, efficacious treatments for cancer patients.

Acknowledgements

This work was supported by NIH grants GMS0870 (N .Q.), AI54962 (N.Q) and AI44936

(D.C.M.).

References

1. Hoption Cann SA , van Nerren JP, van Nerren e. Dr William Coley and tumour regression: a place in history or in the future. Posrgrad Med J 2003; 79(938):672-680. 2. Shear MJ, Turner Fe. Chemical treatment of tumors : isolation of hemorrhagic-producing fraction from Serratia marcescens (Bacillus prodigiosus) culture filtrate. J Nat! Cancer Inst 1943; 4:81-97. 3. Raetz CR, Garrett TA, Reynolds CM et al. Kdoz-Lip id A of Escherichia coli, a defined endotoxin that activates macrophages via TLR-4. J Lipid Res 2006 ; 47(5) :1097-1111. 4. Galanos e. Physical state and biological activity of lipopolysaccharides. Toxicity and immunogenicity of the lipid A component. Z Immunitatsforsch Exp Klin Immuno11975; 149(2-4) :214-229. 5. Kanegasaki S, Kojima Y, Matsuura M er al. Biological activities of analogues of lipid A based chemically on the revised structural model. Comparison of mediato r-Inducing, immunomodulating and endotoxic activities. Eur J Biochem 1984; 143(2) :237-242. 6. Homma JY, Matsuura M, Kanegasaki S et al. Structural requirements of lipid A responsible for the functions: a study with chemically synthesized lipid A and its analogues. J Biochem (Tokyo) 1985; 98(2 ):395-406. 7. Kot ani S, Takada H, Tsujimoto M et al. Synthetic lipid A with endotoxic and related biological activities comparable to those of a natural lipid A from an Escherichia coli rernutant. Infect Imrnun 1985; 49( I):225-237. 8. Galanos C , Luderirz 0, Rierschel ET et al. Synthetic and natural Escherichia coli free lipid A express identical endotoxic activities. Eur J Blochern 1985; 148(1) :1-5. 9. Galanos C , Lehmann V, Luderitz 0 et al. Endotoxic propertie s of chemically synthesized lipid A part structures . Compatison of synthetic lipid A precursor and synthetic analogues with biosynthetic lipid A precursor and free lipid A. Eur J Biochern 1984; 140(2):221-227. 10. Dobrovolskaia MA, Medvedev AE, Thomas KE et al. Induct ion of in vitro reprogramming by Toll-like receptor (TLR)2 and TLR4 agonists in murine macrophages: effects of TLR "homo tolerance" versus "heterorolerance" on NF-kappa B signaling pathway components . J Immuno12003; 170(1):508·519. II. Lehner MD , Mor ath S, Michelsen KS et al. Induction of cross-tolerance by lipopolysaccharide and highly purified lipoteichoic acid via different Toll-like receptors independ ent of paracrine mediators. J Immunol 2001; 166(8):5161-5167. 12. Saro S. Nomu ra F. Kawai T et al. Synergy and cross-tolerance between toll-like receptor (TLR) 2- and TLR4-mediated signaling pathways. J Immunol 2000 ; 165(12) :7096-7101. 13. Schade FU. Flach R. Flohe S er al. Endotoxin tolerance. In: Brade H . Opal SM, Vogel SN. Morrison D, eds. Endotox in in Health and Disease. New York: Marcel Dekker, Inc., 1999: 751-767. 14. West MA, Heagy W. Endotoxin tolerance: a review. Ctit Care Med 2002; 30(1 Suppl) :S64-S73. 15. Grcisman SE, Hornick RB. The nature of endotoxin tolerance. Trans Am Clin Clirnatol Assoc 1975; 86:43-50:43-50. 16. Matsuura M. Nakano M. Mechanisms to induce tolerance by synthetic monosaccharide lipid A analogues against LPS lethality in mice. Prog Clin Bioi Res 1995; 392:539-548. 17. Satoh M, Tsurumaki K, Kagehara H et al. Induction of intrarurnoral tumor necrosis factor by a synthetic lipid A analog, ON0-4007. with less tolerance in repeated administration and its implication in potent antitumor effects with low toxicity. Cancer Immunol Imrnunorher 2002 ; 50(12) :653-662. 18. Qureshi N, Takayama K. Structure and function of lipid A. In: IglewskiBH, Clark VL. eds. The bacteria. New York: Academic Press. 1990: 319-338. 19. Golenbock DT, Hampton RY. Qureshi N er al. Lipid A-like molecules that antagonize the effects of endotoxins on human monocytes, J Bioi Chern 1991; 266(29 ):19490-19498. 20. Stutz P, Liehl E. Lipid A analogs aimed at preventing the detrimental effects of endotoxin . Infect Dis Clin North Am 1991; 5(4):847-873. 21. Ribi E. Haskins WT. Landy Met al. Symposium on bacterial endotoxins. 1. Relationship of chemical composition to biological activity. Bacteriol Rev 1961; 25:427-436. 22. Ribi E, Haskins WT, Milner KC et al. Physicochemical changes in endotoxin associated with loss of biological potency. J Bacteriol1962; 84:803-814. 23. Takayama K, Qureshi N, Raetz CR et al. Influence of fine structure of lipid A on Limulus amebocyte lysate dotting and toxic activities. Infect Immun 1984; 45(2) :350-355.

94

Lipid A in Cancer 1herapy

24. Qureshi N. Takayama K. Heller D et al. Position of ester groups in the lipid A backbone oflipopolysaccharides obtained from Salmonella typhimurium. J Bioi Chern 1983; 258(21):12947-12951. 25. Qureshi N. Mascagni P. Ribi E et al. Monophosphoryl lipid A obta ined from lipopolysaccharides of Salmonella minnesota R595. Purification of the dimethyl derivative by high performance liquid chromatography and complete structural determination. J Bioi Chern 1985; 260(9) :5271-5278. 26. Madonna GS, Peterson JE. Ribi EE er al. Early-phase endotoxin tolerance: induction by a detoxified lipid A derivative. monophosphoryllipid A. Infect Immun 1986; 52(1) :6-11. 27. Qureshi N, Takayama K. Ribi E. Purification and structural determination of nontoxic lipid A obtained from the lipopolysaccharide of Salmonella typhimurium. J Bioi Chern 1982; 257(19) :11808-11815. 28. Henricson BE. Benjamin WR. VogelSN. Differential cyrokine induction by doses of lipopolysaccharide and mono phosphoryl lipid A that result in equivalent early endotoxin tolerance. Infect Immun 1990; 58(8) :2429-2437. 29. Takayama K. Qureshi N, Beutler B et al. Diphosphoryllipid A from Rhodopseudomonas sphaeroides ATCC 17023 blocks induction of cachectin in macrophages by lipopolysaccharide. Infect Immun 1989; 57(4) :1336-1338 . 30. Kirkland TN. Qureshi N. Takayama K. Diphosphoryllipid A derived from lipopolysaccharide (LPS) of Rhodopseudomonas sphaeroides inhibits activation of 70Z/3 cells by LPS. Infect Immun 1991; 59(1) :131-136. 31. Qureshi N. Takayama K. Kurtz R. Diphosphoryl lipid A obtained from the nontoxic lipopol ysaccharide of Rhodopseudomonas sphaeroides is an endotoxin antagonist in mice. Infect Immun 1991; 59(1):441-444. 32. Henricson BE, Perera PY. Qureshi N et al. Rhodopseudomonas sphaeroides lipid A derivatives block in vitro induction of tumor necrosis factor and endotoxin tolerance by smooth lipopolysaccharide and monophosphoryllipid A. Infect Immun 1992; 60(10) :4285-4290 . 33. Qureshi N. Jarvis BW;Takayama K. Nontoxic RsDPLA as a potent antagonist of toxic lipopolysaccharide. In: Brade H. Opal SM. Vogel SN. Morrison DC. eds. Endotoxin in health and disease. New York: Marcel Dekker, Inc.• 1999: 687-698. 34. Rossignol DP. Hawkins LD, Christ WJ et al. Synthetic endotoxin antagonists. In: Brade H. Opal SM, Vogel SN. Morrison DC. eds. Endotoxin in health and disease. New York: Marcel Dekker. Inc.• 1999: 699-717. 35. Persing DH. Coler RN, Lacy MJ et al. Taking roll: lipid A mimetics as adjuvants and immunomodulators. Trends Microbiol2002; 10(10 Suppl):S32-S37. 36. Carswell EA. Old LJ. Kassel RL ct al. An endotoxin-induced serum factor that causes necrosis of tumors. Proc Natl Acad Sci USA 1975: 72(9) :3666-3670. 37. Haranaka K. Satomi N. Sakurai A er al. Role of lipid A in the production of tumor necrosis factor and differences in ant itumor activity between tumor necrosis factor and lipopolysaccharide. Tohoku J Exp Med 1984; 144(4):385-396. 38. Fraker DL . Sheppard BC. Norton JA. Impact of tolerance on antitumor efficacyof tumor necrosis factor in mice. Cancer Res 1990; 50(8) :2261-2267. 39. Mackensen A. Galanos C. Wehr U et al. Endotoxin tolerance : regulation of cyrokine production and cellular changes in response to endotoxin application in cancer patients . Eur Cytokine Nerw 1992; 3(6) :571-579 . 40. Engelhardt R. Mackensen A, Galanos C et al. Biological response to intravenously administered endotoxin in patients with advanced cancer. J Bioi Response Mod 1990; 9(5) :480-491. 41. Engelhardt R. Mackensen A. Galanos C. Phase I trial of intravenously administered endotoxin (Salmonella abortus equi) in cancer patients. Cancer Res 1991; 51(10) :2524-2530. 42. Otto F, Schmid P. Mackensen A et al. Phase II trial of intravenous endotoxin in patients with colorectal and nonsmall cell lung cancer. Eur J Cancer 1996; 32A(10):1712-1718. 43. Engelhardt R, Otto F, Mackensen A er al. Endotoxin (Salmonella abortus equi) in cancer patients . Clinical and immunological findings. Prog Clin Bioi Res 1995; 392:253-261. 44. Mackensen A. Galanos C, Engelhardt R. Modulating activity of interferon-gamma on endotoxin-induced cytokine production in cancer patients . Blood 1991; 78(12) :3254-3258. 45. Kawasaki H , Moriyama M. Ohtani Y et al. Restoration of normal febrile response to endo toxin in pyrogen-tolerant rabbits by injection with human beta interferon. Infect Immun 1987; 55(11) :2574-2578 . 46. Soszynski D. Inhibition of nitric oxide synthase delays the development of tolerance to LPS in rats. Physiol Behav 2002; 76(1) :159-169. 47. Roperr C. Closel M. Chaves AC et al. Inhibition of a p38 /stress-activated protein kinase-2-dependent phosphatase restores function of IL-l receptor-associate kinase-I and reverses Toll-like receptor 2- and 4-dependent tolerance of macrophages. J Immunol 2003; 171(3):1456-1465 .

LipidA-Mediated Tolerance and Cancer Therapy

95

48. Wysocka M. Montaner L], Karp CL. FIr3 ligand treatment reverses endotoxin tolerance-related irnmunoparalysis.] Immuno12005 : 174(11):7398-7402. 49. Berger MR , Petru E, Schmahl D. Therapeutic ratio of mono or combination bacrerial lipopolysaccharide therapy in merhylnitrosourea-induced rat mammary carcinoma. ] Cancer Res Clin On col 1987; 113(5):437-445. SO. Bertok L. Radio-detoxified endotoxin activates natural immunity : a review. Pathophysiology 2005; 12(2):85-95. 51. Berendt M], Newborg ME North R]. Increased toxicity of endotoxin for rumor-bearing mice and mice responding to bacterial pathogens: macrophage activation as a common denominator. Infect Immun 1980; 28(2) :645-647. 52. Serafini P, Borrello I, Bronte V. Myeloid suppressor cells in cancer: recruitment, phenot ype, properties and mechanisms of immune suppression. Semin Cancer Bioi 2006; 16(l):53-65. 53. Kusmartsev S. Gabrilovich DI. Role of immature myeloid cells in mechanisms of immune evasion in cancer. Cancer Immunol Immunother 2006; 55(3):237-245. 54. Linehan DC , Goedegebuure PS. CD25+ CD4+ regulatory T-cells in cancer. Immunol Res 2005; 32(1-3) :155-168. 55. Frey AB, Monu N. Effector-phase tolerance : another mechanism of how cancer escapes antirumor immune response.] Leukoc BioI 2006; 79(4) :652-662. 56. Page GG . Surgery-induced immunosuppression and postoperative pain management. AACN C1in Issues 2005 ; 16(3):302-309. 57. Joosten ]], van Muijen GN, Wobbes T et al. Cryosurgery of rumor tissue causes endotoxin tolerance through an inRammatory response. Anticancer Res 2003; 23(IA):427-432. 58. Chase j], Kubey W, Dulek MH et al. Effect of monophosphoryllipid A on host resistance to bacterial infection. Infect Immun 1986: 53(3):711-712. 59. Schutze E, Hildebrandt], Liehl E et al. Protection of mice from mortality caused by livingand heat-killed bacteria by SDZ MRL 953. Circ Shock 1994; 42(3) :121-127. 60. Knopf HP, Otto F, Engelhardt R er al. Discordant adaptation of human peritoneal macrophages to stimulation by lipopolysaccharide and the synthetic lipid A analogue SDZ MRL 953. Down-regulation ofTNF-alpha and IL-6 is paralleled by an up-regulation ofIL-l beta and granulocyte colony-stimulating factor expression.] Immunol1994; 153(1):287-299. 61. Takayama K. Ribi E, Cantrell]L. Isolation of a nontoxic lipid A fraction containing rumor regression activity. Cancer Res 1981; 41(7 ):2654-2657. 62. Astiz ME, Rackow EC, Still]G et al. Pretreatment of normal humans with monopho sphoryl lipid A induces tolerance to endotoxin: a prospective. double-blind, randomized, controlled trial. Crit Care Med 1995; 23(1) :9-17. 63. Henricson BE. Manthey CL, Perera PY et al. Dissociation of lipopolysaccharide (LPS)-inducible gene expression in murine macrophages pretreated with smooth LPS versus monophosphoryllipid A. Infect Immun 1993; 61(6) :2325-2333. 64. Astiz ME, Saha DC, Brooks K et al. Comparison of the induction of endotoxin tolerance in endotoxernia and peritonitis by monophosphoryllipid A and lipopolysaccharide. Circ Shock 1993; 39:194-198. 65. Ribi E. Cantrell ]L, Takayama K er al. Lipid A and immunotherapy. Rev Infect Dis 1984 ; 6(4) :567-572. 66. Matsushita K. Kuramitsu Y. Ohiro Y et al. ON0-4007. a synthetic lipid A analog, induces Th l-rype immune response in tumor eradication and restores nitric oxide production by peritoneal macrophages. Int] Oncol 2003; 23(2):489-493 . 67. Matsushita K. Kobayashi M. Torsuka Y ec al. ON0-4007 induces specific anti-rumor immunity mediated by rumor necrosis factor-alpha. Anticancer Drugs 1998; 9(3):273-282. 68. Kurarnitsu Y. Matsushita K, Ohiro Y et al. Therapeutic effects of a new synthetic lipid A analog. ONO-4007, on rat hepatoma KDH-8 depend on rumor necrosis factor-sensitivity of the rumor cells. Anticancer Drugs 1997; 8(9):898-901. 69. Ishida H, Irie K, Suganuma T et al. A lipid A analog ONO-4007 induces tolerance to plasma leakage in mice. InRamm Res 2002; 51(1):38-43. 70. Matsumoto N, Oida H, Aze Y et al. Intrarumoral rumor necrosis factor induction and rumor growth suppression by ONO-4007, a low-toxicity lipid A analog. Anticancer Res 1998; 18(6A):4283-4289. 71. Veda H , Yamazaki M. Induction of rumor necrosis factor in a murine rumor by systemic administration of a novel synthetic lipid A analogue, ONO-4007.] Irnrnunother 1997; 20(1) :65-69. 72. Morita S. Yamamoto M, Kamigaki T et al. Synthetic lipid A produces antitumor effect in a hamster pancreatic carcinoma model through production of rumor necrosis factor from activated macrophages. Kobe] Med Sci 1996; 42(4) :219-231.

96

Lipid A in Cancer Therapy

73. Akimoto T, Kumazawa E. limbo T er al. Antitumor effect of DT-546la, a synthetic low-toxicity lipid A analog, involves endogenous tumor necrosis factor induction subsequent to macrophage activation. Inr ] Immunopharmacol1994: 16(11) :887-893. 74. limbo T, Akimoto T, Tohgo A. Effect of combined administration of a synthetic low-toxicity lipid A derivative, DT-546la and indomethacin in various experimental tumor models of colon 26 carcinoma in mice. Cancer Immunol Immunother 1995: 40(1) :10-16. 75. Fan H . Cook ]A . Molecular mechanisms of endotoxin tolerance.] Endotoxin Res 2004: 10(2):71-84. 76. Chu E, Casey LC, Harris ]E er al. Suppression of the development of tumoricidal function in gamma interferon -treated human peripheral blood monocytes by lipopolysaccharide: the role of cyclooxygenase merabolires.} Clin Immuno11993: 13(1):49-57. 77. Sato K, Yoo YC. Fukushima A er al. A novel synthetic lipid A analog with low endotoxiciry, DT-5461 , prevents lethal endotoxem ia. Infect Immun 1995: 63(8) :2859-2866. 78. Sato K, Yoo YC. Matsuzawa K et al. Tolerance to the anti-metastatic effect of lipopolysaccharide against liver metastasis in mice. Int ] Cancer 1996: 66( I) :98-103. 79. Kiani A, Tschiersch A, Gaboriau E et al. Downregulation of the proin/lammatory cytokine response to endotoxin by pretreatment with the nontoxic lipid A analog SDZ MRL 953 in cancer patients. Blood 1997: 90(4) :1673-1683. 80. Saha DC. Barua RS, Astiz ME er al. Monophosphoryl lipid A stimulated up-regulation of reactive oxygen intermediates in human monocytes in vitro.] Leukcc Bioi 2001: 70(3) :381-385. 81. Kasravi FB, Lee DH, Weisgraber K et al. Lipoprotein-bound endotoxin exerts an immunomodulatory effect on hepatocytes through the lipid A domain of LPS.] Endotoxin Res 2005: 11(1):19-24. 82. Ogawa H . Rafiee P, Heidemann] et al. Mechanisms of endotox in tolerance in human intestinal microvascular endothelial cells. J Immunol 2003: 170(12):5956-5964. 83. Parker LC , Jones EC. Prince LR et al. Endotoxin tolerance induces selective alterations in neutrophil function. J Leukoc BioI 2005: 78(6) :1301-1305 . 84. Fujihara M, Muroi M, Tanamoto K et al. Molecular mechanisms of macrophage activation and deactivation by lipopolysaccharide: roles of the receptor complex. Pharmacol Ther 2003: 100(2):171-194. 85. Medvedev AE, Sabroe I, Hasday JD ct aI. Tolerance to microbial TLR ligands: molecular mechanisms and relevance to disease. J Endotoxin Res 2006: 12(3):133 -150. 86. Nomura F, Akashi S, Sakao Y et al. Cutting edge: endotoxin tolerance in mouse peritoneal macrophages correlates with down-regulation of surface toll-like receptor 4 expression. J Immunol 2000: 164(7) :3476-3479 . 87. Medvedev AE, Kopydlowski KM, VogelSN. Inhibition of lipopolysaccharide-inducedsignal transduction in endotox in-rolcrized mouse macrophages: dysregulation of cytokine, chemokine and toll-like receptor 2 and 4 gene expression. J Immunol 2000: 164(11) :5564-5574. 88. Martin M, Katz J, Vogel SN et al. Differential induction of endotoxin tolerance by lipopolysaccharides derived from Porphyromonas gingivalis and Escherichia coli.] Immunol 2001: 167(9) :5278-5285. 89. Zarember KA. Godowski PJ. Tissue expression of human Toll-like receptors and differential regulation of Toll-like receptor mRNAs in leukocytes in response to microbes, their products and cytokines. J Immunol 2002: 168(2):554-561. 90. Medvedev AE, Lcntschat A, Wahl LM er al. Dysregulation of LPS-induced Toll-like receptor 4-MyD88 complex formation and IL-l receptor-associated kinase 1 activation in endotoxin-tolerant cells.J Immunol 2002: 169(9) :5209-5216 . 91. Oshikawa K. Sugiyama Y. Gene expression of Toll-like receptors and associated molecules induced by in/lammatory stimuli in the primary alveolar macrophage. Biochem Biophys Res Commun 2003: 305(3) :649-655. 92. Moreno C, Merino J, Vazquez B et al. Anti-in/lammatory cytokin es induce lipopolysaccharide tolerance in human monocytes without modifying toll-like receptor 4 membrane expression. Scand J Immunol 2004 : 59(6) :553-558. 93. W ittebole X, Coyle SM, Kumar A ec al. Expression of tumour necrosis factor receptor and Toll-like receptor 2 and 4 on peripheral blood leucocytes of human volunteers after endotoxin challenge: a comparison of /low cytornetric light scatter and immunofluorescence gating. Clin Exp Immunol 2005: 141(1) :99-106. 94. Tarnai R, Sugawara S, Takeuchi 0 et al. Synergistic effects of lipopolysaccharide and interferon-gamma in inducing interleukin-S production in human monocytic THP-l cells is accompanied by up-regulation of CDI4, Toll-like receptor 4, MD-2 and MyD88 expression. J Endotoxin Res 2003: 9(3) :145-153. 95. Ziegler-Heirbrock HW; Wedel A, Schraut W et al. Tolerance to lipopolysaccharide involves mobilization of nuclear factor kappa B with predominance of p50 homodimers. J Bioi Chern 1994: 269(25):17001-17004.

Lipid A-MediatedTolerance and Cancer Therapy

97

96. Mathison J. Wolfson E, Steinemann S et al. Lipopo lysaccharide (LPS) recognition in macrophages. Participation of LPS-binding protein and CD14 in LPS-induced adaptation in rabbit peritoneal exudate macrophages. J Clin Invest 1993; 92(4) :2053-2059 . 97. Akashi S. Shimazu R, Ogata H et al. Curting edge: cell surface expression and lipopolysaccharide signaling via the toll-like receptor 4-MD-2 complex on mouse peritoneal macrophages. J Immunol 2000 ; 164(7) :3471-3475 . 98. Medvedev AE. Vogel SN. Overexpression of CDI4. TLR4 and MD -2 in HEK 293T cells does not prevent induction of in vitro endotoxin tolerance. J Endotoxin Res 2003 ; 9(1) :60-64. 99. Burns K. Janssens S, Brissoni B er al. Inhibition of interleukin 1 receptor /Toll-like receptor signaling through the alternatively spliced. short form of MyD88 is due to its failure to recruit lRAK-4. J Exp Med 2003 ; 197(2) :263-268. 100. Janssens S. Burns K, Vercammen E er al. MyD88S. a splice variant of MyD88. differentially modulates NF-kappa B- and AP-l-dependent gene expression. FEBS Lett 2003; 548(I-3):103-107. 101. Li L, Cousart S. Hu J er al. Characterization of interleukin-I receptor-associated kinase in normal and endotoxin-tolerant cells. J Bioi Chem 2000; 275(30) :23340-23 345. 102. Noubir S. Hmama Z . Reiner NE . Dual receptors and dist inct pathways mediate interleukin-I receptor-associated kinase degradation in response to lipopolysaccharide. Involvement of CD 14/TLR4. CR3 and phosphatidylinositol 3-kinase. J Bioi Chern 2004 ; 279(24) :25189-25195 . 103. Kobayashi K. Hernandez LD. Galan JE er al. IRAK-M is a negative regulator of Toll-like receptor signaling. Cell 2002; 110(2):191-202. 104. Lopez-Collazo E. Fuentes-Prior P. Arnalich F et al. Pathophysiology of interleukin-I receptor-associated kinase-M: implications in refractory state. CUrt Opin Infect Dis 2006 ; 19(3):237-244. 105. Stoiber D. Kovarik P. Cohney S et al. Lipopolysaccharide induces in macrophages the synthesis of the suppressor of cytokine signaling 3 and suppresses signal transduction in response to the activating factor IFN-gamma. J Immuno11999; 163(5):2640-2647. 106. Berlaro C. Cassatella MA. Kinjyo I et al. Involvement of suppressor of cytokine signaling-3 as a mediator of the inh ibitory effects of IL-l 0 on lipopolysaccharide-induced macrophage activation. J Immunol 2002 ; 168(12) :6404-6411. 107. Nakagawa R. Naka T. Tsutsui H er al. SOCS-l part icipates in negative regulation of LPS responses. Immunity 2002 ; 17(5):677-687. 108. Gingras S. Parganas E. de Pauw A et al. Re-examination of the role of suppressor of cytokine signaling 1 (SOCSl) in the regulation of toll-like receptor signaling. J Bioi Chern 2004; 279(52) :54702-54707. 109. Brint EK. Xu D. Liu H er al. ST2 is an inhibitor of interleukin 1 receptor and Toll-like receptor 4 signaling and maintains endotoxin tolerance. Nat Immunol 2004 ; 5(4) :373-379. 110. Liew FY, Liu H. Xu D. A novel negative regulator for IL-l receptor and Toll-like receptor 4. Immunol Lett 2005; 96(1) :27-31. Ill. Zhang G. Ghosh S. Negative regulation of toll-like receptor-mediated signaling by Tollip. J Bioi Chern 2002; 277(9) :7059-7065 . 112. Li T. Hu J. Li 1. Characterization of Tollip protein upon Lipopolysaccharide challenge. Mol Immunol 2004: 41(1) :85-92. 113. Did ierlaurenr A. Brissoni B. Velin D et al. Tollip regulates proinflammatory responses to interleukin-I and lipopolysaccharide. Mol Cell Bioi 2006: 26(3) :735-742. 114. Sly LM, Rauh MJ. Kalesnikoff J er al. LPS-induced upregulation of SHIP is essential for endotoxin tolerance. Immunity 2004 ; 21(2):227 -239. 115. Guzik TJ. Korbut R. Adamek-Guzik T. Nitric oxide and superoxide in inflammation and immune regulation. J Physiol Pharmacol 2003; 54(4) :469-487. 116. Zhang X. Morrison DC. Lipopolysaccharide-induced selective priming effects on tumor necro sis factor alpha and nitric oxide production in mouse peritoneal macrophages. J Exp Med 1993; 177(2):511-516. 117. Tominaga K. Saito S. Matsuura M et al. Role of IFN-gamma on dissociation between nitric oxide and TNF/ IL-6 production by murine peritoneal cells after restirnulation with bacterial lipopolysaccharide. J Leukoc Bioi 1999; 66(6) :974-980. 118. Spitzer JA. Zheng M. Kolls JK er 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. 119. Fahmi H . Ancuta P, Perrier S er al. Preexposure of mouse peritoneal macrophages to lipopolysaccharide and other stimuli enhances the nitric oxide response to secondary stimuli. Inflamm Res 1996; 45(7) :347-353. 120. Zingarelli B. Halushka PY, Caputi AP et al. Increased nitric oxide synthesis during the development of endotoxin tolerance. Shock 1995; 3(2) :102-108.

98

Lipid A in Cancer Therapy

121. Severn A. Xu D. Doyle J et al, Pre-exposure of murine macrophages to lipopolysaccharide inhibits the induction of nitric oxide synthase and reduces lcishmanicidal activity. Eur J Immunol 1993; 23(7) :1711-1714. 122. Dias MB. Almeida MC. Carnio EC et al. Role of nitric oxide in tolerance to lipopolysaccharide in mice. J Appl Physiol2005; 98(4) :1322-1327. 123. Zingarelli B, Hake PW, Cook JA. Inducible nitric oxide synthase is not required in the development of endotoxin tolerance in mice. Shock 2002 ; 17(6):478-484. 124. Bogdan C. Vodovorz Y, Nathan C. Macrophage deactivation by interleukin 10. J Exp Med 1991; 174(6):1549-1555. 125. de Waal MR. Abrams J. Bennett B er aJ. Interleukin 10(IL-I0) inhibits cyrokine synthesis by human monocytes: an autoregulatory role of lL -10 produced by monocytes. J Exp Med 1991; 174(5) :1209-1220 . 126. Randow F, Syrbe U. Meisel C et al. Mechanism of endotoxin desensitization : involvement of inrerleukin 10 and transforming growth factor beta, J Exp Med 1995; 181(5):1887-1892. 127. Berg DJ. Kuhn R, Rajewsky K et al, Inrerleukin-Ifl is a central regulator of the response to LPS in murine models of endotoxic shock and the Shwartzrnan reaction but not endotoxin tolerance. J Clin Invest 1995; 96(5) :2339-2347. 128. Chantry D. Turner M. Abney E et al, Modulation of cyrokine production by transforming growth Factor-bera. J Immunol 1989; 142(12) :4295-4300 . 129. Hafenrichter DG. Roland CR . Mangino MJ et al, The Kupffer cell in endotoxin tolerance : mechanisms of protection against lethal endotoxemia, Shock 1994; 2(4) :251-256. 130. Li MH. Scatter SC. Manthei R er al, Macrophage endotoxin tolerance : effect of TNF or endotoxin pretreatment. J Surg Res 1994; 57( 1):85-92. 131. Scatter SC, Li MH. Bubrick MP et al, Endotoxin pretreatment of human monocytes alters subsequent endotoxin-triggered release of inflammatory mediators . Shock 1995; 3(4) :252-258. 132. Bulger EM. Maier RY. Lipid mediators in the pathophysiology of critical illness. Crit Care Med 2000; 28(4 Suppl) :N27-N36. 133. Jiang C. Ting AT. Seed B. PPAR-gamma agonises inhibit production of monocyte inflammatory cytokines, Nature 1998; 391(6662):82-86. 134. Ricote M. Li AC, Willson TM et al, The peroxisome proliferaror-acrivared receptor-gamma is a negative regulator of macrophage activation. Nature 1998; 391(6662):79-82. 135. Uch imura K. Nakamura M. Enjoj i M er al. Activation of retinoic X receptor and peroxisome proliferaror-activared receptor-gamma inhibits nitric oxide and tumor necrosis factor-alpha production in rat Kupffer cells. Hepatology 2001; 33(1) :91-99. 136. Von Knerhen AA. Brune B. Delayed activation of PPARgamma by LPS and IFN-gamma attenuates the oxidative burst in macrophages. FASEBJ 2001; 15(2):535-544. 137. Bohuslav J. Kravchenko vv. Parry GC ct aJ. Regulation of an essential innate immune response by the p50 subunit ofNF-kappa B. J Clin Invest 1998; 102(9) :1645-1652 . 138. Wysocka M. Robertson S. Riemann H et al, lL-12 suppression during experimental endotoxin tolerance: dendritic cell loss and macrophage hyporesponsiveness. J Immuno12001 ; 166(12):7504-7513 . 139. H irsch C. Ploegh HL. Intracellular targeting of the proteasome. Trends Cell Bioi 2000; 10(7) :268-272. 140. Voges D. Zwick! P, Baumeister W. The 26S proteasome : a molecular machine designed for controlled proteolysis. Annu Rev Biochem 1999; 68:1015-1068. 141. Elenich LA. Nandi D. Kent AE et aJ. The complete primary structure of mouse 20S protcasornes. Immunogenetics 1999; 49(10) :835-842. 142. Dahlman B. Hendil KB. Kristensen P er al, Subunit arrangement in the human proteasome. In: Hilt W; Wolf DH. eds, Proteasomes: The world of regulatory proteolysis. Georgetown : Landes Bioscience. 2000 : 37-47. 143. Groettrup M, Khan S. Schwarz K et aI. Interferon-gamma inducible exchanges of 20S proteasome active site subunits: why? Biochimie 2001 ; 83(3-4) :367-372. 144. Gaczynska M. Goldb erg AL. Tanaka K et aJ. Proreasome subun its X and Y alter peptidase activities in opposite ways to the interferon-gamma-induced subunits LMP2 and LMP7 . J Bioi Chern 1996; 271 (29) :17275-17280 . 145. Qureshi N. Perera PY. Shen J et al, The proteasome as a lipopolysaccharide-binding protein in macrophages: differential effects of proteasome inhibition on lipopolysaccharide-induced signaling events. J Immunol2003; 171(3):1515-1525. 146. Qureshi N. Vogel SN. Van WC. 111 et al, The proteasome : a central regulator of inflammation and macrophage function. lrnmunol Res 2005; 31(3) :243-260. 147. SheriJ. ReisJ. Morrison DC et al, Key inflammatory signaling pathways arc regulated by the proteasome. Shock 2006; 25(5):472-484.

Lipid A-Mediated Tolerance and Cancer Therapy

99

148. Husebye H , Halaas 0 , Stenmark H et al. Endoc yric pathways regulate Toll-like receptor 4 signaling and link innate and adaptive immunity. EMBO J 2006; 25(4):683-692. 149. Kutuzova GD, Albrecht RM, Erickson CM et al. Diphosphoryllipid A from Rhodobacter sphaeroides blocks the bind ing and internalization of lipopol ysaccharide in RAW 264.7 cells. J Immunol 2001 ; 167(1 ):482-489. 150. Lien E. Means TK, Heine H et al. Toll-like receptor 4 imparts ligand-specific recognition of bacterial lipopolysaccharide. J Clin Invest 2000; 105(4):497-504. 151. Jarvis BW; Lichensteln H, Qureshi N. Diphosphoryl lipid A from Rhodobacrer sphaeroides inhib its complexes that form in vitro between lipopolysaccharide (LPS)-binding protein , soluble CD14 and spectrally pure LPS. Infect Immun 1997; 65(8):3011 -3016. 152. Baker PJ, Taylor CE, Srashak PW et al. Inactivation of suppressor Tvcell activity by the nontoxic lipopolysaccharide of Rhodopseudomonas sphaeroides, Infect Immun 1990; 58(9):2862-2868. 153. Baker PJ, H iernauxJR. Fauntleroy MB et al. Ability of monophosphoryllip id A to augment the antibody response of young mice. Infect Immun 1988; 56(J2):3064-3066 . 154. Baker PJ, Hiernaux JR , Fauntleroy MB ct al. Inactivation of suppressor T-cell activity by nontoxic rnonophosphoryl lipid A. Infect Immun 1988; 56(5 ):1076-1083 .

CHAPTER 9

Lipid A in Cancer Therapies Preclinical Results Daniele Reisser and jean-Franeois Jeannin"

Abstract

S

tu dies in an imal models showed th at the antitumoral effect ofLPS and oftheir biologically active moiety, lipid A, is indirect and relies on the induction of an immune response both innate and specific, leading to cytokine production. They also affect tumor development by inhibiting tumor blood flow and induce necrosi s as well as apoptosis of tumor cells. Lipids A have been tested in animals, either alone or as adjuvant in therapeutic vaccines. The efficacy of treatments depends on the type of molecule and on the protocol. In general . increased survival was obtained, accompanied in some cases by tumor regression and cure.

Introduction

The first synth etic lipids A were chemically synth esized by Kusumoto and collaborato rs in 1985.1 Since then, nontoxic derivatives oflipid A were prepared and tested against tumors. Th is chapter will only consider in vivo data obtained with curative treatments beginning after the establishment of tumors. As the antitumoral effect of LPS is due to lipid A, it is difficult to dissociate the mechanisms ofaction oftreatments performed with th ese compounds. Th is chapter will thus mention the antitumoral mechanisms deciphered in LPS treatments, but will insist on treatments using lipid A and particularly DT·5461 , ON0-4007 and OM·174 that have been more widely documented as to th eir mechanisms.

LPS Treatments

In animals, th e first in vivo experiments were performed with filtrares of bacteria cultures in a guinea pig sarcoma model. ' Hemorrhage and necrosis were induced within the tumors. Then . purified LPS were used in mouse primary subcut aneous (s.c.) tumors.' The effects ofLPS on the growth of s.c., intramuscular (i.m .) or intraperitoneal (i.p.) tumors have been extensively investigated since the studies of Strausser and Bober.t The efficacies ofLPS and lipid A treatments according to the nature oftumors, their sites ofirnplantation and to treatment schedules were compared by Parr et al.1 In a model of peritoneal carcinomatosis induced by PROb colon cancer cells in syngeneic BDIX rats, we showed that i.p. injections ofLPS from Escherichia coli cured 20% of the rats. whatever the injections were performed daily, or three or five days apart. " Several mechanisms. by no way exclusive, were advanced to explain the antitumoral effect of LPS. *Corresponding Author: lean-Franco is Jeannin - EPHE, D ijon, F-21000, France; Inserm U866, D ijon , F-21000, France. Email: jean-francoi s.jeannin @u-bourgogne.fr

LipidA in Cancer Therapy, edited byjean-Francoisjeannin. ©2009 Landes Bioscience and Springer Science+Business Media.

102

Lipid A in Cancer Therapy

The role of antibodies in leukemia regression in LPS-treated mice was evidenced by Sinkovic et aFParr ec als showed that coagulation was involved in, but was not sufficient for tumor regression in mice. Lymphocytes were necessary since the regression ofsarcoma or fibrosarcoma did not occur in Tscell-deficlent mice.! Since the 1970s, the antitumoral effect ofLPS was mainly attributed to tumor necrosis factor (TNF). Carswell et al9 showed that in BALB/c mice bearing Meth A sarcoma , the antitumoral activity ofLPS was due to a circulating factor, further known as TNF-a. In consequence, several studies targeted cells that can be induced to secrete this cytokine. Historically, macrophages were the first candidates. Mannel et al l O considered them as the effector cells in the regression offibrosarcoma in C3H mice. Moreover, the regression ofMeth A sarcoma in BALB/c mice, MH-l34 hepatoma in C3H/He mice, Lewis Lung carcinomas in C57BL/6 mice and pulmonary Meth A metastases, involves the secretion ofTNF-a by polymorphonuclear neutrophils after intradermal (l.d.) injections ofLPS.11The antitumoral effect ofLPS against Meth A sarcoma was inhibited by injections ofan antibody directed against TNF-a. 12 However, tumor necrosis and regression are to be dissociated, since in SCID (severe combined immuno-deficiency) mice treated with LPS, necrosis did not lead to the rejection of tumors. Furthermore, anti-interferon-y (IFN-y) antibodies inhibited the LPS-induced regression ofMeth A sarcoma in mice, but not the necrotic hemorrhage attributed to TNF-a. 13Thus. both TNF-a and IFN-y are involved in the antitumoral effect ofLPS in the Meth A sarcoma model. In addition to causing vascular obstruction in the tumor, LPS injections decrease tumor irrigation in a systemic way.14 On the other hand, the TNF-a secreted by intratumoral macrophages in response to LPS stimulates angiogenesis in primary prostatic cancer in the rat.'S In the rat model of peritoneal carcinomatosis. our team showed that whatever the origin of the LPS, in vivo efficiency correlated with the in vitro secretion by macrophages ofIL-1 ~ but not that ofNO, TNF-a and IL-6 .!6 The effect of LPS on metastasis is controversial. They may be antirnetastatic by promoting the detachment of tumor cells from lung endothelium.'?On the contrary, in a murine model of experimental pulmonary metastasis, LPS enhanced tumor metastasis by increasing angiogenesis in tumoral nodules through the production ofvascular endothelial growth factor (VEGF) by tumor cells. They facilitated vascular permeability. migration and invasion of tumor cells by increasing nitric oxide synthase (NOS) II and matrix metalloproteinase (MMP2) 2 production by infiltrating host cells.'! However, in the rat model ofperitoneal carcinomatosis, LPS treatments decreased the vascular permeability since they inhibit the ascitis development."The schedules ofthe treatments were important, since injections performed every four days decreased the metastasis induced by lymphomas in mice, while daily injections were without any effect."

Lipids A Treatments

Since 1973, Parr et als showed that a synthetic lipid A given i.p. slowed down the growth ofi.d. L5178Y lymphoma and even inhibited it in three out of nine DBA/2 mice. The antitumoral activity ofa lipid A from Salmonella typhimurium injected into hepatocarcinoma 10 tumors in guinea pigs" was due to a monophosphoryl diglucosamine." Synthetic lipids A were also shown to be active in this model. 22 In BALBIc mice bearing i.d, Meth A fibrosarcoma Shimizu et al23-2S compared treatments with LPS and synthetic lipid A analogs. Intravenous (Lv.) injections oflipids A slowed down tumor growth and the authors linked this antitumoral effect with the TNF-a production by peritoneal cells, which they induced in vitro. In our rat model of peritoneal cardnomatosis," we compared the efficacy of different lipids A.26 We showed that five i.p, injections twice a week cured some rats bearing macroscopic nodules at the beginning ofthe treatment and that three or more hydroxymyristic acids were necessary to obtain the best efficiency, which was correlated with the in vitro secretion ofTNF-a and interleukin-I ~ (IL-1~) by peritoneal macrophages but not with their cytolytic activity. Two intravenous injections of the monoglucosamine GLA-27 as well as of a natural lipid A slowed down the growth ofi.d. RL-d' 1 lymphoma and induced the regression ofMeth A sarcoma

Lipid A in Cancer Therapies PreclinicalResults

103

in BALBIc mice, in correlation with tumor necrosis," A synthetic lipid A which induced TNF-a secretion by mice peritoneal cells slowed down the growth ofhuman pancreatic tumors in nude mice when injected Lp,twice a week to a total ofeight injections and enhanced survival.28 The effect was attributed to TNF-a. Better respon seswere obtained and tumor necrosiswas evidenced earlier in mice receiving lipid A at the dose of20 mg/kg than at the dose of 10 mg/kg. On the contrary, in the rat model ofperitoneal carcinomatosis, our team showed that whatever the lipid A used, in vivo efficiency correlated with the in vitro secretion by macrophages ofIL-l~ and TNF-a but not the capacity to induce macrophage-mediated cytolysls," Moreover, to achieve tumor regression, several injections of the lipids A: SDZ MRL 953, DT-5461 or OM-163 were needed, inducing the tolerizarion of rat macrophages , which no longer produced TNF-a (Onier N, unpublished results). In this model, lymphocytes were involved as treatment was not efficient in nude rats." Spleen cells were protective in a Winn-type assay, e.g., the protection of naive animals against a tumor by the injection ofcells from an animal cured of the same tumor. The inhibition ofpulmonary meta stases ofmelanoma B16 in BD F1 mice by the lipid A analog GL-60 was highly dependent on the treatment schedule ." A single i.v, injection at the dose of 10 f.lg/mouse was not efficient, three injections on days 3, 7 and 10 inhibited metastasis while three injections on days 1,4 and 7 or 5, 9 and 12 did not. Efficiency was not better with GL-60 at the dose of 100 f.lg/ mouse. In a model of i.d. B16 tumors, while i.v,injections of GL-60 at the dose of50 f.lg had no effect on angiogenesis, angiogenesis was decreased when this compound was injected at the same time as 104 U IFN-yand tumor growth was reduced, " Spontaneous metastases caused by intra-footpad inoculation of B16 cells were also inhibited by the combined treatment, by inducing the production ofTNF-a. Ther apeutic cancer vaccinations e.g., performed in a curative and not a preventive purpose, using lipids A as adjuvant with tumor extracts or tumor antigens were seldom documented in preclinical studies. In BALB/c mice bearing i.p TA3-Ha mammary carcinoma cells, three s.c. injections ofRibi adjuvant [a commercial preparation containing monophosphoryllipid A (MPLA)] mixed with Thomsen-Friedenreich or epiglycanin antigens were performed on days 1 or 4 after tumor cell injection. No effect was obtained. However, pretreating mice i.v,with cyclophosphamide (CY) in order to inhibit the suppressive response enhanced survival. Both anti body as well as delayed-type hypersensivity responses were obtained.P Moreover, lymph node cells were protective in a Winn-type assay. In C3H/HeN mice bearing intracutaneous MH134 hepatoma, MPLA from Porphyromonas gingivalis or S. minnesota Re 595 increased the survival ofmice when admini stered intracutaneously in combination with tumor celllysates on day 7 after the injection oftumor cells. When a second vaccine injection was performed on day 21 the rate oftumor regression was 90% on day 30.33 The efficiency ofMUC 1 pep tides encapsulated in liposomes containing MPLA was tested in CB6F1 mice bearing the GZHI A5.3H4 tumor resulting from the transfection of mouse mammary adenocarcinoma with MUC] gene. Subcutaneous injections of the vaccine on days 3 and 17 after the i.v, injection of tumor cells resulted in fewer lung foci." In a postoperative schedule, BALBIc mice received an s.c, injection of C26 colon carcinoma cells, laparatomy was performed on day 5. MPLA mixed with C26 irradiated cells, either in a liquid form or encapsulated in alginate was injected s.c, on day 8. Encapsulated vaccine limited the growth ofestablished tumors more efficiently than liquid one."

DT-5461

The synthetic diglucosamine DT-5461 (Daiichi Pharmaceutical Co, Tokyo, Japan) was tested first on various murine tumors e.g., Meth A fibrosarcoma and colon 26 adenocarcinoma in BALBIc mice, MHI34 hepatoma and MM46 mammary carcinoma in C3H/HeN mice, 3LL Lewis Lung carcinoma and C38 colon carcinoma in C5 7BL/6 mice." Mice bearing solid i.d, or s.c, tumors (Meth A, MH 134, MM46 , 3LL , colon 26 and colon 38) received three i.v,injections of200 f.lg/ mouse of the lipid A on days 7, 12 and 17. The weight of the Meth A, MH 134, MM46, 3LL

104

LipidA in Cancer Therapy

tumors was reduced and hemorrhagic necrosis was evidenced. However colon tumors were not sensitive to the treatment. Two i.v, injections on days 3 and 6 increased survival of mice bearing Meth A malignant ascitis, l.p, injections or i.v, injections on days 1 and 5 were without any effect. Whatever the schedule no treatment had any effect on MH 134 ascitis. Four Lv, injections reduced the number ofmetastatic foci in mice, after i.v,injections ofL5178Y lymphoma. Both i.v, or intranasal treatment reduced lung metastasis after injection ofB16 melanoma cells or resection ofthe primary tumor. No treatment was efficient on ascitic tumors." To inhibit the resistance of C26 tumors, Jimbo et al used a cotreatment with indomethacin to inhibited prostaglandin (PG) E2 production by these cells." When DT-5461 was injected alone, four injections at the dose of200 ug/rnouse on days 1, 5, 9 an 13 or continuous injections till the animal death increased the life span ofmice bearing hepatic metastases, continuous treatment was efficient in mice bearing lung metastases and no treatment had any effect on mice bearing peritoneal metastases. In any cases,a combined treatment with indomethacin given orally allowed an increase of survival. This effect was attributed to TNF-a that peaked in the tumor following lipid A injection and was further increased by indomethacin administration. Though three i.v, injections ofDT-5461 on days I, 5 and 9 after tumor inoculation significantly inhibited liver metastasis ofL5178Y-ML25 lymphoma, a pretreatment by LPS inhibited this effect. 39 This finding was explained by the fact DT-5461 shares receptor site with LPS.40 Thirteen days after Meth A s.c. inoculation in BALB/c mice an injection ofDT-5461 at the dose of200 !!g/ml decreased the blood flow in tumor nodules. This phenomenon was reversed by pretreating the animals with antisera directed against TNF-a and IFN-a/~, or TNF-a and IFN_y.41 When treatment schedules of Meth A fibrosarcoma were extensively studied, administration on days 7, 12 and 17 proved to be of the best efficiency by i.v, or intratumoral route. Intraperitoneal route gave better results than s.c. or i.rn. ones and oral treatment, even at a dose of 2,400 !!g/mouse, was ofno efficacy.42 A significant antitumoral effect was obtained with lipid A at doses ~50 ug/mouse. By i.v, route , DT-5461 induced the production ofTNF-a in s.c, tumors where TNF-a activity lasted at least during 16 hours while it decreased in serum. Two injections ofmurine anti-TNF-a serum, one ofthem by i.p. route 3 hours before and the second one by i.v, route just before lipid A treatment partially reduced the antitumoral effect ofDT-5461. Though IFN-y was not induced in the tumor nodules, the ant itumoral activity ofDT-5461 was decreased by two injections ofanti-IFN -a/~ or ofanti-IFN-y, one just before and the second one 10 hours after each injection ofDT-5461. By i.v, route, DT-5461 injected 4 days after the s.cinoculation ofBI6-BL6 cells in BDFI mice decreased the neo -vascularization in tumor nodules. The inhibition of tumor angiogenesis was likely mediated by the TNF-a production induced in tumors. Intravenous injections of lipid A on days I , 5 and 9 after tumor cell inoculation which caused a high TNF-a production at the primary tumor site, reduced the number ofspontaneous metastases. However, though anti-TNF-a antiserum abrogated the anti-angiogenic effect ofDT-5461 , it did not prevent inhibition oftumor

growrh."

The efficiency ofD5461 was not limited to mouse models . In New Zealand white rabbits, four i.v,injections ofDT-5461 at the dose of 10 mg/kg on days 10, 14, 17 and 21 afterthe implantation in the liver ofhepatic VX2 carcinoma, inhibited tumor growth." This inhibition was associated with TNF-a production in tumor tissue and with blood flow reduction in the tumor area that could be reproduced by i.v, injections ofTNF-a. In conclusion, in several models, DT-5461 proved to display antitumoral effect linked with TNF-a and IFN induction and the reduction of tumor blood flow and angiogenesis in tumors. Several i.v, injections slowed down tumor growth, inhibited metastasis, increased life span but no cure was reported in any of the models.

ONO-4007

Studies using the monoglucosamine ONO-4007 (Ono pharmaceutical, Osaka, Japan) were published slightly later than those reporting on DT-5461. Interestingly the first one documented

Lipid A in Cancer Therapies PreclinicalResults

105

another type oftumors: leukemia." WKA rats bearingc-WRT-7 myelomonocytic leukemia were treated from day 3 after tumor cell injection with daily i.v, injections of ON0-4007 at the doses of2.5 mg/kg or 5 mg/kg, One injection was without any effect, but seven injections increased life span. The prolongation ofsurvivalwas explained by a differentiating effect that was not reproduced in vitro by a treatment with cytokines (IL-1 a , IL-6 or TNF-a). The antitumoral effect ofON0-4007 was then studied on solid tumors. A week after the implantation ofMM46 mammary carcinoma cells s.c, or i.d., C3H/He mice were administered five weekly i.v,injections ofONO-4007 at the dose of 10 mg/kg.The tumor growth was suppressed and respectively 2 out of 6 mice bearing s.c. tumors and 4 out of 6 mice bearing i.d, tumors were cured. Though MM46 are TNF-a-resistant, the inhibition ofthe growth ofs.c. tumors correlated with TNF-a secretion by intratumoral macrophages. 46 In the same model, one ON0-4007 injection induced TNF-a to peak earlier in serum than in tumor.The antitumoral effect of three i.v,injections of 10 mglkg ON0-4007 twice a week correlated with intraturnoral TNF-a activiry," Though a systemic tolerance was achieved, the TNF-a level remained high in tumors, due to the rapid retraction of the tolerance oftumor-infiltrating macrophage . In BALB/c and C3H/He mice bearing respectively i.d, Meth A fibrosarcoma and MH 134 hepatoma, ON0-4007 was used in combination with CY. On day 14 after tumor cell injection, mice received an i.p, injection of CY at the dose of 100 mg/kg and then three i.d, injections of ON0-4007 at the dose of30 mglkg on days 21, 24 and 27. While the drugs administered alone had no effect on tumor growth, the combined treatment led to the regression of both tumors in the five mice tested per group ." Transient TNF-a activity vas induced in serum after ONO-4007 administration, while it lasted 24 hours at least in tumors. Combined treatment with CY led to a further increase in TNF-a level in tumor. After an injection of ONO-4007 , the levels of IL-1 ~, TNF-a, IL-12 and Il-6 increased in Meth-A tumors, while the level ofTGF-~ decreased. The combined treatment led to increased IL-10 and IL-6Ievels. The level ofFas ligand decreased, thus lowering the ability for the tumor to escape immune surveillance. In another study in the same animal model, combined treatment was administered according to a different schedule. Seven days after Meth A inoculation, CY was injected i.p, at the dose of 100 mg/kg and ONO-4007 was injected i.v, at the dose of30 mg/kg." As in the previous study, CY that shifted the cytokine profile in tumor towards a Th1 type, particularly by increasing IL-12 production, enhanced the antitumoral effect of ONO-4007. In hamsters bearing pancreatic carcinoma too, six i.p, injections of ON0-4007 at the dose of 50 mg/kg, inhibited tumor growth. The TNF-a level increased more in tumor than in serum and other organs. l? The antitumoral effect of0 N 0 -4007 was tested on 13762NF carcinoma implanted in different sites in F-344 rats." Four weekly i.v, injection ofONO-4007 at the dose of2.5 mg/kg cured 50% ofthe rats bearing s.c. tumors when beginning on day 8 and 33% ofthe rats when beginning on day 14. When tumor cells were injected by Lv, route, four ONO-4007 i.v,injections beginning on day 2 were without any effect. So combined treatments with carboplatin injected i.v,on day 10 at the dose of35 mglkgor CY injected i.v, on day 10 at the dose of30 mg/kgwere used, with no more efficiency whatever the schedule tested . In rats bearing intraperitoneal tumors, three i.p. injections cured 2 rats out of 10, while two i.p. injections were inefficient. Two or three i.v, injections were also inefficient. Carboplatin did not improve the efficacy. Treatments with ONO-4007 were also tested in rats bearing different s.c. tumors: KDH-8 hepatocarcinoma and KMT-17 fibrosarcoma in WKAH rats, SST-2 mammary adenocarcinoma in SHR rats. Intravenous injections of ONO-4007 at the doses of 1,2.5, or 5 mg/kg performed on days 7,14 and 21 after s.c, inoculation ofKDH-8 cells dose-dependently increased the survival of rats." No therapeutic effect was evidenced in rats bearing KMT-17 or SST-2 tumors. In KDH-8 tumors, the treatment induced the expression ofIL-1a, IL-6 and TNF-a, but had no effect on IL-2, IL-4, IL-lO and IFN-y levels. When tested in vitro, ON0-4007 was cytotoxic on KDH-8 cells, but neither on KMT-17 nor on SST-2 cells, what could explain in vivo results. Further studies dealt with the KDH-8 cells/WKAH rat model. Kuramitsu et al 53 compared

106

Lipid A in Cancer Therapy

ON0-4007 treatment efficacy on two KDH-8 clones differing by their in vitro sensitivity to TNF-a. Three intravenous injections of ON0-4007 at the dose of2.5 mg/kg led to the regression ofthe s.c, tumors induced by the injection ofthe TNF-a-sensitive KDH-81YK line in three out of five rat, but had no effect on the tumors induced by injection of the TNF-a-resistant line KDH-8/11. In another study using parental KDH-8 cells, the number of injections of ON0-4007 wasincreased to four, on days7, 14,21 and 28 afier tumor cellsinjection. ON0-4007 at the doses of 3 or 10 mg/kg decreased tumor growth , but this effect was no more observed for the dose 000 mg/kg. 54 Total cures were obtained in 2 out of 10 rats treated with ON0-4007 at the dose of 3 mg/kg, Diclofenac, a cyclooxygenase inhibitor, given per os at the dose of 3 mg/kg did not improve the treatment efficiency. An injection of ONO-4007 induced TNFactivity in liver, spleen, serum and tumor. While TNF-a production decreased in liver and serum following repeated injections, it was not reduced in spleen and tumor. An intratumoral injection ofanti-rat-TNF-a-antibody suppressed the antitumoral effect ofON0-4007. However, though ON0-4007 proved to be cytotoxic against KDH-8 cellsin vitro, TNF-a was not. Simultaneous injections of anti-rat-TNF-a-antibody during the treatment of rats abrogated the antitumoral effect of ONO-4007 on KDH-8 tumors .v Cured rats rejected a challenge with the TNF-resistant KDH-8/11 cells and specific immunity was also evidenced in a Winn-type assay. In rats bearing KD H -8 tumor, the treatment by ON0-4007 increased the expressionofIFN-y, 11-1~, IL-6, IL-lO and IL-12. The level ofIL-12 protein was increased and the production ofNO by peritoneal macrophages was resrored" In conclusion , the efficiency of ON0-4007 treatments seems to follow a bell-shaped curve according to doses and i.v, route gave the best results. The efficacyof ON0-4007 may be limited to TNF-a-sensitive tumors in rats.

OM-174

We investigated the antitumoral activity of0 M-174 (0M PHARMA, Meyrin, Switzerland) , a triacylared diglucosamine, reproducing a partial structure of E. coli lipid A ,57 in our rat model of peritoneal carcinomatosis. Without treatment, all rats die of their tumors. Treatment always started 14 days afier the i.p. injection oftumor cells.At this date numerous macroscopic nodules reach 3 mm in diameter, the cumulative volume ofwhich corresponds to a large tumor. Several treatment schedules were rested." Whatever the conditions tested, OM-174 slowed down tumor growth and inhibited ascitis formation. The best efficacywas monitored for the rate oftotal cures. Five i.p. injections twice a week cured 21% of the animals. However, i.v, treatment proved to be ofbetter efficiency. The best dose was 1 rug/kg: the optimal number of injections was 15, with a frequency of2-4 day intervals . Fifteen i.v, injections every two days ofOM-174 at the dose of 1 rng/kg, cured 90-100% ofthe rats. The establishment of a specific immune response was proved in a Winn-type assay. T-Iymphocytes playa role in the tumor regression as OM-174 was less efficient in nude rats (Onier N , personal communication). Treatment efficacydepended on the number and frequency ofinjections that induced the tolerance oftumor-infiltrating macrophages to OM-174, evidenced by the decrease ofTNF-a production in the tumor which peaked following the two first injections, then returned to basal level. TNF-a is thus probably not involved in tumor regression more especiallyas PROb cellsare resistant to TNF-a in vitro. Tumors disappear via apoptotic pathway without inflammatory reaction. Since 0 M-174 is not toxic to tumor cellsin vitro, it does not provoke tumor cell apoptosis directly. During lipid A-induced tumor regression, NOS II mRNA and protein are expressed in tumor cells with concomitant production ofNO. 59 Neither OM-174 nor TNF-a induces NO production by PROb tumor cells in vitro, whereas NOS II expression is induced by IFN-y and IL-1~ in these cells/" Therefore, the in vivo NOS II induction by OM-174 may be indirectly due to the accumulation ofIFN-y and IL-1 ~ evidenced in the tumors, at the mRNA and protein levels.v In tumo rs the apoptosis oftumor cellsand NOS II expression were evidenced in the same areas." The NO thus produced is auto toxic for tumor cells provoking their apoptosis . Besides,the treatment ofrats with 0 M-174 inhibited the synthesis of TGF-~1 by PROb tumor cells,62therefore abrogating its immunosuppressive role.63 Moreover since

Lipid A in Cancer Therapies Preclinical Results

107

the inhibition ofTGF-~l enhances the synthesis ofNOS II62 the autotoxic effect ofNO on tumor cellsthus increases.Meanwhile, treatment with lipid A induces the tolerance ofmacrophages, which produce no further NO.59Without treatment, the NO produced by macrophages upon activation by tumor cells inhibits T-cell proliferation in the spleen, thereby promoting tumor growth,60 on the contrary to the NO produced by tumor cells. A histochemical study showed that chronologically, a high proportion ofapoptotic tumor cells were detected in tumor nodules as soon as on day 15, one day after the first injection ofOM-174. Dendritic cells invaded tumoral nodules first on day 15 and became activated before macrophages could be detected. Then lymphocytes were attracted at the periphery ofthe tumor nodules but were not detected inside and were always separated from tumor cells by fibroblasrs /" Innate immunity may thus precede the specific immunity evidenced in this model." After apoptosis oftumor cells, dendritic cells may phagocytose apoptotic bodies or debris and present the immunogenic peptides to T-lymphocytes thus inducing a specific immune response. The antitumoral effect of OM-174 , alone or in combination was evidenced in mice too . In a model ofexperimental lung metastasis ofEMT6 mammary carcinoma in BALB/c mice, five i.v, injections ofOM-174 at the dose of8 mg/kg on days 5, 10, 15,20 and 25 after Lv, tumor cell injection, cured 67% of mice, which were immunized against another challenge of EMT6 cells. This antitumoral effect involved the production of NO by tumor cells (Nolwenn Gauthier, Thesis, Universite de Bourgogne, Ecole Pratique des Hautes Etudes , Dijon, 2003) . In C57BL/6 mice bearing s.c. B16 melanoma, five i.p, injections ofOM-174 at the dose of 1 mg/kg on days 8, 13, 18,23 and 28 after tumor cell inoculation reduced tumor growth and increased survival. When CYat the dose of200 mg/kg was injected one day before OM-174 treatment, both parameters were Improved.f No total cure was reported. OM-174 treatment enhanced the specific cytolytic activit y ofT-lymphocytes in tumor-bearing mice, which was further increased by CY pretreatment. In conclusion, treatments with OM-174 are efficient both on mouse and rat tumors. Intravenous route is the best way ofadministration, repeated injections are required. Total cures were obtained which were likely not TNF-a-dependent but involved the production of NO by tumor cells. Though specific immunity was induced, the apoptosis oftumor cells began before recruitment of lymphocytes to tumor area which have not direct contact with tumor cells.

Conclusion

Various lipids A have been used to treat tumor-bearing animals. The most effective structure so far consists of diglucosamines acylated by long chain fatty acids and the substitu tion of the diglucosamine backbone is now under investigation . Intratumoral, i.p., Lv. or i.d, routes have been used and most ofthe studies were performed using several i.v, injections. Frequency is an important parameter and the optimal dose is not necessarily the maximal one . Optimum doses range from 1 to 10 mg/kgfor the lipids A DT 5461, ONO 4007 and OM-174 in rats and mice. Lipids A are generally considered to act through TNF-a secretion. For instance the efficacyof both DT-5461 and ONO-4007 in mice and rats depend on TNF-a. Therefore only well-vascularized tumors might be affected. However, in our model, TNF-a is probably not involved since this cytokine peaks after the first two injections and then returns to basal levels before tumor regression. The efficacy of 0 M -174 relies on the indirect induction of an autotoxic production of NO by the tumor cells. To achieve the best efficiency,treatments require several injections oflipids A, but whether tolerance induction is involved is not unequivocal. In rats bearing S.c. KD H -8 hepatocarcinorna, though treatments with ONO-40071ed to systemic desensitization evidenced by the decrease in serum TNF-a level, no tolerance occurred in tumor-infiltrating macrophages and the NO production of peritoneal macrophages was restored. On the contrary, in rats bearing PROb peritoneal carcino matosis , the efficiency of OM-174 treatment was linked with tolerance of the tumor-infiltrating macrophages, evidenced by the collapse oftheir NO and TNF-a secretions.

108

Lipid A in Cancer Therapy

In most studies treatments with lipid A increase the survival of tumor-bearers and slow down the growth ofestablished tumors in mice, rats and rabbits . ON0-4007 cures s.c. tumors both in mice and rats, but not l.p, or lung ones in rats. OM-174 cures peritoneal carcinomatosis in rats and lung metastases in mice. With a same lipid A, the best treatment schedule changes from one animal model to the other. Comparative studies have not been performed to elucidate ifspecies and tumor origin and/or the immunogenicity of tumor cells and their immunosuppressive effect are important parameters at the origin ofthese differences. References

1. Imoro M, Yoshimura H , Sakaguchi N et al. Total synthesis of Escherichia coli lipid A. Tetrahedron Lett 1985; 26:1545-1548. 2. Gratia A, Linz R. Le phenomene de Shwartzm an dans le sarcome du cobaye. C R Soc Bioi 1931; 108:427-428 . 3. Shear M], Turner Fe. Chemical treatment of tumo rs: isolation of hemorrhagic-producing fraction from Serratia marcescens (Bacillus prod igiosus) culture filrrate.} Nat! Cancer Inst 1943; 4:81-97. 4. Strausser HR, Bober LA. Inhibition of tumor growth and survival of aged mice inoculated with Moloney tumor transplants and treated with endotoxin. Cancer Res 1972; 32:2156-2159. 5. Parr I, Wheeler E, Alexander P. Similarities of the anti-tumour actions of endotoxin, lipid A and double-stranded RNA. Br] Cancer 1973; 27:370-389. 6. Lagadec P, ]eannin ]F, Reisser D er al. Treatment with endotoxins of peritoneal carcinomatosis induced by colon tumor cells in the rat. Invasion Metastasis 1987; 7:83-95. 7. Sinkovics]G, Ahearn M], Shirato E et al. Viral leukemogenesis in immunologically and hematologically altered mice. ] Rericuloendothel Soc 1970; 8:474-492. 8. Berendt M], North R], Kirstein DP. The immunological basis of endotoxin-induced tumor regression. Requirement for Tvccll-mcdiared immunity. ] Exp Med 1978; 148:1550-1559. 9. Carswell EA, Old L], Kassel RL et al. An endotoxin -induced serum factor that causes necrosis of tumors. Proc Nat! Acad Sci USA 1975; 72 :3666 -3670. 10. Mannel DN, Rosensrreich DL, Mergenhagen SE. Mechanism of lipopolysaccharide-induced tumor necrosis: requirement for lipopolysaccharide-sensitive lymphoreticular cells. Infect Immun 1979; 24:573-576. 11. Inagawa H , Nish izawa T, Takagi K et al. Ant itumor mechanism of int radermal administration of lipopolysaccharide. Anticancer Res 1997; 17:1961-1964. 12. Inagawa H , Nishizawa T, Noguchi K ct al. Anti-tumor effect of lipopolysaccharide by intradermal administration as a novel drug delivery system. Anticancer Res 1997; 17:2153-2158. 13. Dighe AS, Richards E, Old L] et al, Enhanced in vivo growth and resistance to rejection of tumor cells expressing dominant negative IFN gamma receptors . Immun ity 1994; 1:447-456. 14. MacPherson GG, North R]. Endotoxin-mediated necrosis and regression of established tumours in the mouse. A correlative study of quant itative changes in blood Rowand ultrastructu ral morphology. Cancer Immunol Imrnunother 1986; 21:209-216 . 15. Vukanovic], Isaacs ]T. Linomide inhibits angiogenesis, growth , metastasis and macrophage infiltration within rat prostatic cancers. Cancer Res 1995; 55:1499 -1504. 16. Blondiau C, Lagadec P, Lejeune P er al. Correlation between the capacity to activate macrophages in vitro and the antitumor activity in vivo of lipopol ysaccharides from different bacterial species. Immunobiol 1994; 190:243-254. 17. ]ibu T, Koike Sando K et al. Antirnetasratic activity of lipopolysaccharide against a NK-resistant murine fibrosarcoma. Clin Exp Metastasis 1993; 11:306-312. 18. Harmey]H, Bucana CD, Lu W et al. Lipopolysaccharide-induced metastatic growth is associated with increased angiogenesis, vascular permeability and tumo r cell invasion. Int ] Cancer 2002; 101:4 15-422. 19. Sato K, Yoo YC, Matsuzawa K et al. Tolerance to the anti-metastatic effect of lipopolysaccharide against liver metastasis in mice. Int ] Cancer 1996; 66:98-103. 20. Ribi E, Amano K, Cantrell] er al. Preparation and antitumor activity of nontoxic lipid A. Cancer Immunol Immunother 1982; 12:91-96. 21. ~eshi N, Takayama K, Ribi E. Purification and structural determ ination of nontoxic lipid A obtained from the lipopolysaccharide of Salmonella typhimurium.] Bioi Chem 1982; 257:11808-11815. 22. Ukei S, Iida ], Shiba T er al. Adjuvant and antitumour activities of synthetic lipid A analogues. Vaccine 1986; 4:21-24. 23. Shimizu T, Sugiyama K, Iwamoto Y et al. Biological activities of chemically synthesized N accrylared serine-linked lipid A analog in mice. Int ] Immunopharmacol 1994; 16:659-665.

Lipid A in Cancer Therapies Preclinical Results

109

24. Shimizu T, Iida K, Iwamoto Y et al. Biological activities and antirumor effects of synthetic lipid A analog linked N acerylated serine. Inr ] Immunopharmacol 1995; 17:425-431. 25. Shimizu T, Iwamoto Yo YanagiharaY et al. Comparison of the biological activity of synthetic Nvacylared asparagine or serine linked monosaccharide lipid A analogs. Immunobiology 1996; 196:321-331. 26. jeannin ]-P, Onier N, Lagadec P ct aJ. Antitumor effect of synthetic derivatives of lipid A in an experimental model of colon cancer in the rat. Gastroenterology 1991; 101:726-733. 27. Nakatsuka M, Kumazawa Yo Ikeda S er al. Antitumor and antimicrobia l activities of lipid A-subunit analogue GLA-27.] Clin Lab Immunol1988; 26:43-47. 28. Sarake K, YokomatsuH, Hiura A. Effects of a new synthetic lipid A on endogenous tumor necrosis factor production and antitumor activity against human pancreatic cancer cells. Pancreas 1996; 12:260-266. 29. Onie r N, Lejeune P, Martin M er al, Involvement of T-Iymphocytes in curative effect of a new irnmunornodulator OM-163 on rat colon cancer metastases. Eur] Cancer 1993; 29A:2003-2009. 30. Nakatsuka M, Kumazawa Yo Homma]Y et al. Inhibition in mice of experimental metastasis of B16 melanoma by the synthetic lipid A-subunit analogue GL-60. Inr ] Immunopharmacol1991; 13:11-19. 31. Saiki 1,Sato K, Yoo yc. Inhibition of rumor-induced angiogenesis by the administration of recombinant interferon-gamma followed by a synthetic lipid-A subunit analogue (GLA-60 ). Int ] Cancer 1992; 51:641-645. 32. Fung PY, Madej M, Koganty RR et al. Active specific immunotherapy of a murine mammary adenocarcinoma using a synthetic tumor-associated glycoconjugate. Cancer Res 1990; 50:4308-4314. 33. Ogawa T, Nakazawa M, Masui K. Immunopharmacological activities of the nontoxic monophosphoryl lipid A of Porphyromonas gingivalis. Vaccine 1996; 14:70-76. 34. Samuel ]. Budzynski WA, Reddish MA et al. Immunogenicity and antitumor activity of a liposomal MUC1 peptide-based vaccine. Ine ] Cancer 1998; 75:295-302. 35. Kirman I, Asi Z. Carter] et al. Combined whole tumor cell and monophosphoryl lipid A vaccine improved by encapsulation in murine colorectal cancer. Surg Endosc 2002; 16:654-658. 36. Kumazawa E, Tohgo A, Soga T et al. Significant antitumor effect of a synthetic lipid A analogue, DT-5461, on murine syngeneic rumor models. Cancer Immunol Immunother 1992; 35:307-314. 37. Saro K, Saiki I, Yoo YC et al. DT-5461 , a new synthetic lipid A analogue, inh ibits lung and liver metastasis of rumor in mice. ]pn] Cancer Res 1992; 83:1081-1087. 38. limbo T, Akimoto T, Tohgo A. Effect of combined admin istration of a synthetic low-toxicity lipid A derivative. DT-546la and indomethacin in various experimental rumor models of colon 26 carcinoma in mice. Cancer Immunol Immunother 1995; 40:10-16. 39. Sato K, Yoo YC, Matsuzawa K et al, Tolerance to the anti-metastatic effect of lipopolysaccharide against liver metastasis in mice. Int ] cancer 1996; 66:98-103. 40. Akahane K, Sorneya K, Tsutomi Y er al. Antitumor synthetic lipid A analog DT-5461 upregulares cytokine expression in a murine macrophage cell line through LPS pathway. Cell Immunol 1994;155: 42-55 . 41. Akirnoto T, Kumazawa E, limbo T er al, DT-546la, an ant itumo r synthet ic lipid a analog. causes selective blood /low reduct ion in rumor tissue. Anticancer Res 1995; 15:105-107. 42. Kumazawa E. Akimoro T, Kita Y et al. Intr atumoral production of tumor necrosis factor augmented by endogenous interferons results in potent anritumor effects of DT-5461. a synthetic lipid A analog. J Irnrnunother Emphasis Tumor Immunol1995; 17:141-150. 43. Sato K, Yoo YC, Moch izuki M et al. Inhibition of rumor-induced angiogenesis by a synthetic lipid A analogue with low endotoxicity, DT-5461. ]pn] Cancer Res 1995; 86:374-382. 44. l imbo T, Akimoto T. Tohgo A. Systemic administration of a synthetic lipid A derivative, DT-546la, reduces rumor blood /low through endogenous TNF production in hepatic cancer model of VX2 carcinoma in rabbits. Anticancer Res 1996; 16:359-364. 45. Kobayashi M, NagayasuH , Hamada] ct al. ONO-4007, a new synthetic lipid A derivative, induces different iation of rat myelomonocytic leukemia cells in vitro and in vivo. Exp Hemato11994; 22:454-459. 46. Yang D, Sacoh M, Ueda H et al. Activation of rumor -infiltrating macrophages by a synthetic lipid A analog (ON0-4007) and its implication in antitumor effects. Cancer Immunol Immunother 1994; 38:287-293. 47. Sacoh M, Tsurumaki K, Kagehara H et al. Induction of intrarumoral rumor necrosis factor by a synthet ic lipid A analog (ONO-4007) with less tolerance in repeated administration and its implication in potent antirumor effects with low toxicity. Cancer Immunol Immunorher 2002; 50:653-662. 48. Inagawa H. Nishizawa T. Honda T et al, Mechan isms by which chemotherapeutic agents augment the ant itumor effects of tumor necrosis factor : involvement of the pattern shili of cytokines from Th2 to Th1 in tumor lesions. Anticancer Res 1998; 18:3957-3964. 49. Takiguchi K, Nakamoto T, Inagawa H et al. Profile of cytokines produced in rumor tissue alier administration of cyclophosphamide in a combination therapy with rumor necrosis factor. Anticancer Res 2004 ; 24:1823-1828.

110

Lipid A in Cancer Therapy

50. Morita S, Yamamoto M, Kamigaki T er al. Synthetic lipid A produces antitumor effect in a hamster pancreatic carcinoma model through production of tumor necrosis factor from activated macrophages. Kobe J Med Sci 1996; 42:219-231. 51. Mizushima Y, SassaK, FujishitaT er a1. Therapeutic effect of a new synthetic lipid A analog (ONO-4007) on a tumor implanted at diffetent sites in rats. J Imrnunother 1999; 22:401-406. 52. Kurarnitsu Y, Nishibe M, Ohiro Y et al. A new synthetic lipid A analog, ON0-4007, stimulates the production of tumor necrosis factor-alpha in tumor tissues, resulting in the rejection of transplanted rat hepatoma cells. Anticancer Drugs 1997; 8:500-508. 53. Kurarnirsu Y, Matsushita K, Ohiro Y er al. Therapeutic effect of a new synthetic lipid A analog, ON0-4007, on rat hepatoma KDH-8 depend on tumor necrosis facror-senslnvlry of the tumor cells. Anti-cancer Drugs 1997; 8:898-901. 54. Matsumoto N, Oida H, Aze Y er al. Intrarumoral tumor necrosis factor induction and tumor growth suppression by ON0-4007, a low-toxicity lipid A analog. Anticancer Res 1998; 18:4283-4289. 55. Matsushita K, Kobayashi M, Totsuka Y et al. ONO-4007 induces specific anti-tumor immunity mediated by tumor necrosis factor-alpha. Ant icancer Drugs 1998; 9:273-282. 56. Matsushita K, Kuramitsu Y, Ohiro Y er al. ON0-4007, a synthetic lipid A analog, induces Thl -rype immune response in tumor eradication and restores nitric oxide production by peritoneal macrophages. Int J Oncol 2003; 32:489-493. 57. Brandenburg K, Lindner B, Schromm A et al. Physicochemical characteristics of rriacyl lipid A partial structure OM-174 in relation to biological activity. Eur J Biochem 2000; 267:3370-3377. 58. Onier N, Hilpert S, Arnould L er al. Cure of colon cancer metastasis in rats with the new lipid A OM 174. Apoptos is of tumor cells and immunization of rats. Clin Exp Merastas 1999; 17:299-306. 59. Onier N, Hilpert S, Reveneau S et al. Expression ofinducible nitric oxide synthase in tumors in relation with their regression induced by lipid A in rats. Inr J Cancer 1999; 81:755-760. 60. Lejeune P, Lagadec P, On ier N er al. Nitric oxide involvement in tumor- induced immunosuppression. J Immunol1994; 152:5077-5083. 61. Reisser D, Pance A, Jeannin JE Mechanisms of the antitumoral effect of lipid A. Bioessays 2002; 24:284-289. 62. Lagadec P, Raynal S, Lieubeau B et al. Evidence for control of nitric oxide synthesis by intracellular transforming growth factor-beta1 in rumor cells. Implications for tumor development. Am J Pathol 1999; 154:1867-1876. 63. Lagadec P, Reveneau S, Lejeune P er al. Immunomodulator OM 163-induced reversal of tumor-mediated immunosuppression and downregulation of TGF -beta 1 in vivo. J Pharmacol Exp Ther 1996; 278:926-933. 64. Larmonier CB, Arnould L, Larmonier N et al. Kinetics of tumor cell apoptosis and immune cell activation during the regression of tumors induced by lipid A in a rat model of colon cancer. Inr J Mol Med 2004 ; 13:355-61. 65. D'Agostini C, Pica F, Febbraro G er al. Antitumour effect of OM -174 and cyclophosphamide on murine B16 melanoma in different experimental conditions. Inr Immununopharmacol 2005; 5:1505-1212.

CHAPTER 10

Monophosphoryl Lipid A (MPL) as an Adjuvant for Anti..Cancer Vaccines: Clinical Results

Christopher W. Cluff·

Abstract

A

s technological advances allow for the identification oftumor-associated antigens (TAAs) again st which adaptive immune responses can be raised , efforts to develop vaccines for the treatment ofcancer cont inue to gain momentum. Some ofthese vaccines target differentiation antigens that are expressed by tumors derived from one particular tissue (e.g., Melan -A/ MART-I , tyrosinase , gp 100). Some target antigens are specifically expressed in tumors ofdifferent types but not in normal tissues (e.g., MAGE-3), while other possible targets are antigens that are expressed at low level in normal tissues and are over-expressed in tumors of different types (e.g., HER2, Muc 1). Oncogenes (H ER2 / neu, Ras, E7 HPV 16), tumor supp ressor genes (pS3 ) or tumor-specific post-translational modified proteins (underglycosylated Muc 1) can also be used as cancer vaccine candidates. In either case, th ese antigens tend to be poorly inmmunogenic by themselve s and vaccines containing them generally require the inclu sion ofpotent imm un ological adjuvants in order to generate robust anti-tumor immune responses in humans. Man y adjuvants currently under evaluation for use in cancer vaccines activate relevant antigen pre senting cells,such as dendritic cells and macrophages, via toll-like receptors (TLRs) and promote effective uptake, processing and presentation ofantigen to T-cells in draining lymph nodes. Lipid A, the biologically active portion of the gram-negative bacterial cell wall consti tu ent lipopolysaccharide (LPS) , is known to possess strong immunostimulatory properti es and has been evaluated for more than two decades as an adjuvant for promoting immune respon ses to minimally immunogenic antigens, including TAAs. The relatively recent d iscovery ofTLRs and the identification ofTLR4 as the signaling receptor for lipid A have allowed for a better understanding ofhow this immunostimulant functions with regard to induction ofinnate and adaptive immune response s. Although several lipid A species, including LPS and synth etic analogs, have been developed and tested as monotherapeutics for the treatment ofcancer,}·8 only 3-0-desacyl-4'-monophosphoryllipid A (MPL) has been evaluated as a cancer vaccine adjuvant in published human clinical trials. MPL comprises the lipid A portion of Salmonella minnesota LPS from which the (R)-3-hydroxytetrade canoyl group and the l -phosphate have been removed by successive acid and base hydrolysis," LPS and MPL induce similar cytokine profiles, but MPL is at least 1OO-fold less toxiC. 9,IO MPL has been administered to more than 300 ,000 human subject s in studies of next-gen eration vaccines.I I In this chapter, published clinical trials conducted to evaluate the safety and/or efficacy of various cancer vaccines containing MPL, eith er alone or combined with other immunostirnulants, *ChristopherW. C1uff-GlaxoSmithKline Biologicals, 553 Old Corvallis Road, Hamilton, MT 59840 , USA. Email: christopher.w.c1 [email protected]

LipidA in Cancer Therapy, edited by jean-Francoisjeannin, ©2009 Landes Bioscience and Springer Science+Business Media.

112

Lipid-A in Cancer Therapies

such as cell wall skeleton (CWS) of Mycobacterium pble; (in the adjuvant Detox"; Biomira, Inc.), the saponin QS-21 (in the adjuvants ASOIB and AS02B; GSK Biologicals) or with QS -21 and CpG oligonucleotides (in the adjuvant AS 15; GSK Biologicals) will be summarized. Combining MPL with other immunostimulants has been demonstrated to be advantageous in many cases and may be required to elicit the full complement of activities necessary to achieve an effective immune response and overcome the ability of tumors to evade attack by the immune system . In this chapter, information relating to vaccines targeting specific cancers will be presented in the first section, while information relating to vaccines targeting multiple tumor types by the induction of immune responses to shared TAAs is presented in the second section.

Section 1: Vaccines Targeting Specific Cancer Types Vaccines Targeting Colorectal Carcinoma

Colorectal carcinomas have been demonstrated to express TAAs that are shared among tumor-bearing individuals. In an effort to induce immune responses to common TAAs, a vaccine consisting offour irradiated allogeneic colon carcinoma cell lines combined with Detox" adjuvant (25 flg MPL, 250 flg CWS, squalane oil and Tween-80 as emulsifier) was administered three times at three week intervals by intradermal injection to eight patients with low-volume metastatic colorectal carcinomaY The hypothetical basis for this strategy derived from the fact that common TAAs are present in a vast majority of colorectal carcinomas, suggesting that immunization with allogeneic carcinoma cell lines would immunize against autologous tumors. Seventeen more patients were treated similarly, but also received subcutaneous interleukin-Ia (IL-Ia), 0.3-0.5 flg/ m 2 per day for 8 days after each vaccination. In the patients receiving vaccine only, toxicity was limited to transient erythema, induration and pruritis at the injection site. Addition ofIL-Ia caused fevers, chills and rigors that began within 4 hours ofadministration and lasted 1-2 hours. IL -I a treatment also caused fatigue by day 8 of the treatment regimen. Six patients were DTH-posirive to one or more of the colon cancer cell lines before vaccination. Nineteen oftwenty-two assessed patients had positive skin reactions to one of the vaccine cell lines (HCTII6) at the completion of treatment, indicating that immune responses were generated to the vaccine antigens. KSA (also known as Ep-CAM, GA733, COI7-IA, EGP and KSI-4) is a 40 kDa protein expressed on the cell surface of multiple human cancers, including colorectal carcinoma. KSA is considered an excellent target for cancer vaccines designed to treat colorectal cancer, since it is expressed by more than 85% ofmetastatic lesions and more than 80% ofcells within this type of rumor.'? A Phase I safety and immunogenicity trial of a liposomal vaccine containing recombinant baculovirus-derived KSA and MPL (OncoVax-CLb;Jenner Biotherapies, Inc.) formulated in an oil-in-water emulsion was conducted in patients with histologically documented Stage IV KSA-positive colorectal cancer," The vaccine was administered via the subcutaneous route every four weeks for a total of four immunizations. In addition, some of the vaccinated patients were randomly assigned to receive recombinant granulocyte-macrophage colony stimulating factor (rGM-CSF; Sargramostim, Imrnunex Corp.). Eleven patients were treated and vaccinations with or without addition of rGM-CSF treatment were tolerated equally well, with no injection site reactions or regional lymphadenopathy noted. Seven of eleven patients developed significant KSA-specific cellular immune responses (as measured by IFNy ELISPOT and lymphoproliferation assays), nine of nine patients tested developed DTH responses to KSA and eight of eleven patients developed KSA-specific antibody responses. There were no significant difference in immune responses between vaccine only patients and those receiving vaccine plus rGM-CSF. No significant clinical responses to the vaccination regimen were observed.

Vaccines Targeting Prostate Cancer

Prostate-specific antigen (PSA) is normally secreted into the seminal fluid by prostate cells and is present at low levels in the serum of disease-free individuals. I I Because PSA is expressed exclusively on prostate tissue and is over-expressed in nearly all advanced prostate cancers.P:" it is a logical choice as a vaccine target. A liposomal vaccine consisting ofrecombinant PSA and MPL

Monophosphoryl Lipid A (MPL) asan Adjuvantfir Anti-CancerVaccines: ClinicalResults

113

(JBT 1001 or OncoVax-P ;Jenner Biotherapies, San Ramon, CA) was tested in pilot studies of prostate cancer patients. 17 Patients with histologically confirmed prostate cancer received subcutaneous injections of the vaccine three times (days 0,30 and 60) , combined with subcutaneous administration ofrGM-CSF (Sargmostim, Irnmunex Corp., Seattle , WA) on days 0-4 , 30-34 and 60-64. A second set ofpatients received intramuscular injections ofthe vaccine in an oil-in-water emulsion according to the same schedule. The vaccine induced high titer anti-PSA IgM and IgG in a majority of patients evaluated, while there was considerable heterogeneity in terms of the magnitude and timing of cellular responses (measured by ELISPOT assay).18The small number ofpatients in the trial precluded evaluation ofthe vaccine for clinical efficacy. Prostein is a prostate tissue-specific protein that is uniquely and abundantly expressed in normal and cancerous prostate tissues. Due to this expression profile , prostein is considered a vaccine candidate for prostate cancer. GlaxoSmithKline (GSK) Biologicals is currently testing a vaccine consisting ofa recombinant prostein protein and a GSK proprietary adjuvant AS 15 in a European Phase 1111 safety and efficacystudy in prostate cancer patients. The Zl-patient study is fully enrolled, but no results are available at this time. The AS 15 adjuvant is a liposome formulation containing QS-21 and ligands for both TLR4 (i,e., MPL) and TLR9 (i.e., CpG oligonucleotide).

Vaccines Targeting Melanoma In 1985, Dr. Malcolm Mitchell began clinical testing of a melanoma vaccine containing

homogenates oftwo melanoma cell lines combined with Detox"adjuvant . 19.20 Aswith the colorectal cancer vaccines made from allogeneic carcinoma cell lines, the melanoma vaccine strategy arose from the hope that common TAAs present on allogeneic melanomas would immunize against autologous tumors. In the initial Phase I and II trials, the vaccine was found to be tolerable and approximately 20% oftreated patients had partial (15%) or complete (5%) remission. The strongest correlate of clinical response in each of the trials was an increase in cytolytic Tdymphocyre precursors (pCTL), which were measured by a mixed lymphocyte-tumor cell assay adapted from Vose." Retrospective statistical analysis of patients treated with the vaccines revealed an association between the expression of three HLA Class I alleles (HLA-A2, HLA-BI2 and HLA-C3) and clinical response." Positive results from seven open-label Phase II trials of Melacine" (a total of 139 Stage III/VI melanoma patients) confirmed Mitchell's initial findings" and prompted Ribi ImmunoChem to initiate a multicenter Phase III clinical trial of Melacine~versus the "Dartmouth" combination chemotherapy regimen (dacarbazine, cisplatin, carmustine (BNCU) and tamoxifen) in 1991,24 Although objective responses to the vaccine were not significantly different from chemotherapy and survival times were similar for the two treatments, quality-of-life during treatment significantly favored Melacine-' The results led to submission and approval ofMelacine~in Canada for treatment ofmetastatic melanoma. At this time, Melacine"remains the only cancer vaccine to pass regulatory review (Canada) and be marketed commercially (note: the vaccine was recently removed from the market for business reasons) . In 1992, the Southwest Oncology Group (SWOG) initiated a Phase III clinical trial to evaluate Meladne't u: patients with intermediate thickness melanoma and tumor-free regional lymph nodes (i.e., Stage II disease) in a postsurgical resection setting.25•27 The rationale for targeting Stage II melanoma derived from the theory that cancer vaccines often fail because higher grade, disseminated tumors (Stage III and IV) actively evade immune surveillance and/or suppress immune responses in general (hence, the high rate of infection in advanced cancer patients). If the vaccine was administered early in the disease proces s when the tumor had not metastasized and the immune system was still functional, induction ofadaptive immune responses that can destroy residual tumor should have a better chance of occurring. In add ition, since surgical resection alone is the standard of care for Stage II melanoma, inclusion of relatively nontoxic vaccination with Melacine~waseasily justified, whereas treatment with chemotherapy or high-dose IFNa -2b treatment, both of which are usually associated with significant side-effects, was not. Treatment withMelacine~appearedto provide little benefit versus surger y alone (p = 0.34 corrected for all

114

Lipid-A in CancerTherapies

prognostic factors) , although retrospective analysis of the data revealed a statistically significant benefit for patients expressing HLA-A2 and/or HLA-C3. There have been no further trials to prospectively test Melacine"in patients expressing these MHC haplotypes. At approximately the same time as the Ribi ImmunoChem-sponsored clinical trials were taking place, Mitchell tested the efficacy of Melacine~combinedwith an approved immunotherapy for Stage III/IV melanoma, namely interferon-alpha-2b (Intron A"', Schering Plough Corp.).28.29The large proportion of patients whose disease stabilized upon treatment in these trials led Mitchell to initiate a Phase III trial ofMelacine~plusIntron A~versus Intron A~ alone in 253 patients with Stage IV melanoma." While overall survival and objective response rates were not different for the two arms, the duration of response favored the combination therapy (p = 0.038) . A second Phase III trial (604 patients) was subsequently initiated to compare Meladne" plus Intron A~ to high-dose Intron A~alone . Accrual onto the trial was completed in 2003 and preliminary results published in abstract form showed no significant difference in relapse-free or overall survival between the two arms." In 1995, Schultz et al reported the results ofa Phase 1 clinical trial conducted to evaluate the immunogenicity and efficacy of a polyvalent melanoma antigen vaccine admixed with Detox" adjuvant at two doses (20 I-tg MPL/200 I-tg CWS or 10 I-tg MPLIlOO I-tg CWS).32The antigens were prepared from material shed into serum-free culture medium by four melanoma cell lines . Nineteen patients with resected Stage III melanoma were immunized with the mixtures (7 with low-dose and 12 with high-dose Detox~) every 3 weeks for a total of4 vaccinations . A nonrandomized control group (35 patients) was administered the vaccine with alum as the adjuvant. While Detox" markedly potentiated antibody responses to the antigens, DTH responses were minimal and the vaccine did not appear to improve disease-free survival when compared to the preparation containing alum . The clinical activity, toxicity and immunological effects of immunotherapy with ultraviolet B-irradiated autologous tumor cells plus Detox" adjuvant was tested in melanoma patients in the mid-1990s.J3 Non-anergic patients (i.e., positive for DTH reaction to foreign antigens) who had undergone surgical resection ofmetastatic melanoma were administered 107 irradiated tumor cells plus Detox" by intradermal injection every 2 weeks for six weeks, then monthly until progression ofdisease or exhaustion ofvaccine supply. Toxicity was limited to irritation at the vaccine injection site and, in 10% ofthe patients, malaise lasting 24-48 hours. Small granulomas (<2 ern diameter) occurred at some injections sites and became lessapparent over weeks to months. The induction of cell-mediated cytotoxicity against autologous tumor in 10 of35 assessablepatients indicated that the intradermal treatments resulted in tumor-specific systemic immune responses and detection of this immune reactivity correlated significantly with survival (p =0.009). Among 24 patients with indicator lesions, the vaccination regimen resulted in two major responses in soft-tissue sites and in a bone lesion. The results indicated that there was a small, but clinically significant, therapeutic effect with this vaccination strategy (<20% response, with 95% confidence). Beginning in 2001, the immunogenicity of a vaccine consisting of aT-helper epitope of MART-l and AS02B adjuvant (100 I-tg MPL and 100 I-tg QS21 in an oil/water emulsion, GSK Biologicals, Rixensarr, Belgium) was tested in three Stage III /IV metastatic melanoma panenrs." Patients were immunized once per month for 6 months, then once every 3 months, for a total of8 immunizations in one year. Treatment-associated toxicity was minimal and included injection site discomfort, as well as transient flu-like symptoms and fatigue. Immune reactivity to the peptide was demonstrated using D R4-peptide tetramer staining and enzyme-linked immunospot assayoffresh and restimulated CD4+ T-cells obtained from the patients throughout the course of treatment. MART-I-specific CD4+T-cells obtained postvaccination generally exhibited a mixed Th I/Th 2 phenotype. The T-cells proliferated when cultured with peptide bound to syngeneic antigen-presenting cells. CD4+ T -cells from one patient secreted granzyme B and exhibited MHC-restricted cytolysis when incubated with HLA-matched antigen-expressing tumor cells. At the time ofthe report (August, 2004), two of the three patients were alive and disease-free at 25 and 34 months after entering the trial. The third patient relapsed at 17 months and underwent additional therapy.

Monopbosphoryl Lipid A (MPL) asan Adjuvantfor Anti-Cancer Vaccines: ClinicalResults

115

The results demonstrated that a class II -restricred epitope ofa melanoma-associated antigen is immunogenic in humans and suggests that inclusion of such epitopes in future melanoma vaccines may be worthwhile. A study published in mid-2004 was conducted to evaluate the safety, immunogenicity and efficacyofMelan-A-peptide-based vaccines containingeither AS02B adjuvant or incomplete Freund's adjuvant (IFA) in HLA-A2-positive patients with Melan-A-positive Stage III/IV melanoma.v Ihe vaccines were well tolerated and in vivo expansion of Melan-A-specific CD8+ Tcells, measured using HLA-A2IMelan-A peptide tetrarners and IFNy ELISPOT assays,was observed in 13 ofthe 49 patients (1112 in the AS02B group and 12117 in the IFA group). Four patients had stable disease for a period oftime after treatment and the two patients with the strongest Melan-A-specific T-cell responses (1 from the AS02B group and 1 from the IFA group) experienced regression ofmetastases. Both patients, however, relapsed and died at 14 and 21 months after entering the study.

Vaccines Targeting HER2-Positive Breast Carcinoma

The HER2 protein is a cell-surface glycoprotein receptor that is expressed in many types of epithelial cells and is over-expressed in many types ofcancer, including approximately 25% ofbreast cancers. The high level ofexpression on such cancers makes HER2 a good target for antibody and vaccine therapies. GSK Biologicals conducted a Phase I trial with a recombinant HER2 protein -containingvaccine (plus AS 15 adjuvant) in breast cancer patients. Multiple doses (20,100,500 !!g) ofantigen and two immunization schedules (6 doses versus 3 doses +/- boost) were tested in Stage II patients with >4 positive regional lymph nodes or in patients with locally advanced Stage III disease. The vaccine was well tolerated and seroconversion was demonstrated in a majority ofthe patients, the analysis of the T-cell response is ongoing. A Phase II trial ofthe recombinant Her2 protein (500 f-tg) with ASI 5 adjuvant is currently underway, with transient clinical response observed in 2/5 patients treated thus far.

Section 2: Vaccines Targeting Specific TAAs Expressed on Multiple Tumor Types Vaccines Targeting MAGE-3-Expressing Cancers

MAGE-3 is a cancer testis antigen that is normally expressed only on testicular germ cells. MAGE-3 is aberrantly expressed on a high percentage of a broad variety of tumors, making it a potential target for vaccine immun otherapy. The safety and efficacyofa cancer vaccine containing MAGE-3 protein fused to the N terminal portion ofa protein derived from H influenzae (protein D) and combined with AS02B adjuvant was recently tested in 57 patients with MAGE-3-positive tumors (51 with Stage III /IV melanoma, 3 with transitional carcinoma of the urinary bladder, 2 with nonsmall cell lung carcinoma, 2 with esophageal cancer and 1 with head and neck carcinoma, all Stage IV) .36.37As inclusion criteria. all patients were required to express at least 1 ofthe 3 HLA class I haplorypes known to present MAGE-3 peptides (l.e., HLA-Al, HLA-A2 and/or HLA-B44).The vaccination schedule comprised 4 injections at 3 week intervals, with 2 additional vaccinations administered at 6-week intervals to patients whose tumors stabilized or regressed. Escalating doses ofthe antigen combined with a fixed dose ofthe adjuvant were tested in the trial. The vaccine was well tolerated. A significant MAGE-3 specific IgG response was elicited in most vaccinated patients (96%) and all patients generated IgG antibodies specific for protein D. Using IFNy and IL-5 production as the readout, T-cell responses were detected in approximately 30% of evaluable patients. Five of thirty-three evaluable melanoma patients had partial responses or disease stabilization lasting 4 to 29 months, these five responder patients were all part of the 12 patients with lessadvanced melanoma (5/12 Stage III with non visceral disease). A partial response lasting 10 months was observed in one metastatic bladder cancer patients. A vaccine containing recombinant MAGE-3 protein (+/- AS02B adjuvant) was tested recently in the US by the Ludwig Institute for Cancer Research (LICR) in 17 patients with

Lipid-A in Cancer Therapies

116

MAGE-3-expressing Stage I or II nonsmall cell lung carcinoma who had undergone surgical resection ofthe primary tumor." The patients had no evidence ofdisease at the onset ofthe trial. Three ofnine patients treated with the vaccine in the absence ofadjuvant developed modest, but significant, anti-MAGE-3 antibody responses and one patient in that cohort had a CD8+ T-cell response to the MAGE-3 peptide 243-258. In contrast, seven ofeight patients in the group receiving adjuvanted vaccine developed a marked increase in anti-MAGE-3 antibodies. The investigators used IFNy elispot and tetramer staining to show that inclusion of adjuvant in the vaccine was necessary to generate significant CD4+ and CD8+ T-cell responses to MAGE-3. No assessment of objective clinical response was conducted since the patients had no evidence ofdisease at the onset of the trial . Various tumor types have been targeted by MAGE-3-containing cancer vaccines produced by GSK Biologicals, with emphasis on non-small cell lung carcinoma (NSCLC) and cutaneous metastatic melanoma. 39,40 A Phase lIb, double-blind, placebo-controlled clinical trial was initiated in 2002 to test a MAGE-3 vaccine containing AS02B adjuvant in patients with MAGE-3-positive Stage IB and Stage II NSCLC (post surgical resection). Patients are receiving five intramuscular injections ofplacebo (n 60) or vaccine (300 !J.gMAGE-3; n 122) at three week intervals, followed by eight maintenance immunizations spaced three months apart. The primary endpoints are time to recurrence, disease-free and overall survival, recurrence rate at different timepoints, toxicity and tolerability. Humoral and cellular imm une responses are secondary endpoints. Although the safety aspect ofthe study is still blinded, the vaccine appears to be well tolerated, with mild grade 1 or 2 local or systemic reactions lastingless than 24 hours post-injection being the most common sequale. There have been only three grade 3 adverse events potentially related to treatment. Interim efficacy data was presented at the 2006 American Society of Clinical Oncology (ASCO) annual meeting showing a clear signal (although not significant) with a benefit of33% in recurrence rate. Core analysis ofthe results are expected by October 2006.

=

Vaccines Targeting MUCI-Expressing Cancers

=

Mucins (e.g., MUC-1) are high molecular weight (>500 kDa) glycoproteins expressed on the surface of normal and malignant cells.41•42 The core protein has an extracellular N-terminal domain, a transmembrane region and a C -terminal cytoplasmic domain. The oligosaccharides are attached to serines or threonines of the core peptide by O-glycosidic bonds. In normal cells, mucins are only expressed on the luminal surface and are therefore not exposed to the immune system. Mucins from cancer cells, on the other hand, are exposed uniformly on the cell surface, exposing them to immune surveillance." More importantly, mucins expressed by cancer cells are often underglycosylated, with shorter and simpler carbohydrate chains. Three cancer-specific carbohydrate epitopes have been identified: Thomsen-Friedenreich epitope, a precursor known as Tn and sialyl-Tn (STn). All three epitopes are commonly found on the cell surface ofepithelial cancers or as part ofthe mucins secreted by them, but not on normal epithelial cells, thereby maleing them potential TAAs.44 Most adenocarcinomas express MUC-1 on the cell surface and secrete underglycosylated MUC-1 mucin (expressing repeating STn epitopes) into the serum, resulting in exposure ofthe immune system to multiple tandem repeat core peptide epitopes. A vaccine consisting of a 100-amino acid peptide corresponding to five 20-amino acid long repeats ofMUC-1 and AS02B adjuvant was tested in a Phase I safety and immunogenicity study in 15 patients with resected and 1 patient with locally advanced pancreatic cancer." Patients (4 per dose) were vaccinated by intramuscular injection with one offour doses ofpeptide (100 !J.g, 300 /!g, 1000 !J.gor 3000 !J.g) admixed with the AS02B adjuvant every 3 weeks for a total ofthree doses. The vaccine was well tolerated, with toxicities limited to transient flu-like symptoms and tenderness/ erythema at the injections sites. Vaccination led to an increase in the percentage ofCD8+ cells in the peripheral blood and an increase in MUC-1-specific antibody was seen in some patients. Two ofthe 15 patients with resected tumors were alive and disease-free at 32 and 61 months. BLP25 liposome vaccine iStimuoax", Biomira, Inc.) is a cancer vaccine designed to elicit a cellular immune response to the exposed core peptide of MUCI. The vaccine is a lyophilized

Monophosphoryl Lipid A (MPL) asan Adjuvantfor Anti-CancerVaccines: ClinicalResults

117

liposomal preparation containing BLP25 lipopeptide (25 amino acid peptide), MPL and the lipids cholesterol, dimyristoyl phosphatidylglycerol and dipalmitoyl phosphatidylcholine. MPL and BLP25 are present in the lipid bilayer of the liposome once the dry powder is rehydrated. Following a Phase I safety study ofBLP25 liposomal vaccine in Stage IIIB/IV nonsmall-cell lung cancer patients.t" clinical trials were conducted to test the safety and efficacy ofBLP25 vaccine in prostate" and nonsmall cell lung carcinoma patients (NSCLC).48 A single intravenous dose of cyclophosphamide (300 rng/rn' , maximum of 600 mg) was administered 3 days prior to vaccination. Such treatment has been demonstrated to augment delayed-type hypersensitivity responses, increase antibody responses, abrogate tolerance and potentiate anti-tumor immun ity in preclinical modeIs.4~.so For primary treatment, the vaccine dose (1000 IJ.g ofBLP25 and 25 IJ.g MPL) was divided and administered subcutaneously at four sites weekly for eight weeks. In the prostate study, patients were reassessed following the primary treatment and those who did not require a change in therapy were eligible to continue with BLP25 vaccination every 6 weeks for up to 1 year. At the investigator's discretion, patients in the NSCLC trial continued to receive maintenance vaccinations every 6 weeks starting at week 13. In the study testing BLP25 vaccine in prostate cancer patients with biochemical failure (increasing prostate specific antigen; PSA) after radical prostatectomy, a total of sixteen individuals with a median age of60 were enrolled.? Fifteen ofsixteen completed the primary treatment and ten completed the maintenance period. After primary treatment, eight of sixteen patients had stable or decre ased PSA. Although only one patient maintained stable PSA by the last on-study measurement, six of sixteen patients had greater than 50% prolongation of PSA doubling time compared to prestudy measurements. To evaluate the effect of BLP25 vaccine on survival and toxicity in individuals with Stage IIIB or IV NSCLC, patients were randomly assigned to the BLP25 arm or to the best supportive care (BSC) arm. Overall, the 88 patients assigned to the BLP25 arm had a median survival time (MST) that was 4.4 months longer than the 83 patients in the BSC arm, though the difference was not statistically significant (p = 0.112). The greatest vaccine effect was observed in Stage IIIB locoregional disease, where the MST for 30 BSC patients was 13.3 months while the MST for 35 vaccinated patients was more than 30 months (p = 0.069, though MST had not been reached by end of study for vaccinated patients. A subsequent press release confirmed an MST of 30.6 months for the vaccine group). No significant toxicit y was associated with BLP25 treatment and quality oflife was maintained longer in vaccinated patients.

Vaccines Targeting STn-Expressing Carcinomas

Aberrant carbohydrate molecules, such as sialyl-Tn (STn) , are often expressed as part ofglycoproteins present on the surface ofcarcinomas.As discussed in the previous section, these molecules are not found on normal tissues and are considered TAAs. Antibodies specific for these molecules can be induced that mediate tumor cell lysis by complement or antibody-dependent cellular cytotoxiciry." When expressed on tumors, the mucin-associated STn epitope is a predictor of poor prognosis and is associated with increased metastatic potential.52·54STn is expressed on a significant proportion ofbreast cancer s (16-80%, depending on detection method and laboratory),55'58 with a tendency for higher expression in metastatic disease compared to primary tumors.l" For these reasons , STn is considered a promising target for a cancer vaccine. In the late 1980s, a vaccine consisting ofpartially desialylared ovine submaxillary gland mucin (modified OSM), which contained both Tn and STn epitopes, was tested in patients with metastatic colorectal cancer.S~ Six patients were administered the vaccine alone, eight received the vaccine combined with Detox" adjuvant and six were treated with the vaccine plus BCG as adjuvant. Anti-STn antibody titers increased in 4 of8 patients in the Detox" group, 5 of6 in the BCG group and 0 of6 in the vaccine onl y group. Toxicity was limited to inflammatory skin reactions at the injection sites in patients receiving vaccine plus adjuvant. The result s demonstrated that STn-containingvaccines can induce specific humoral immune responses in cancer patients and that vaccines containing these molecules can be administered safely with immunological adjuvants.

118

Lipid-Ain Cancer 'Iberapies

Tberatope" (Biomira, Inc.) is an investigational cancer vaccine consisting of synthetic STn conjugated to keyhole limpet hemocyanin (KLH) combined with Detox" adjuvant (early trials) or Enhanzyn adjuvant (later trials). Like Detox", Enhanzyn contains MPL and CWS, but is formulated as a stable emulsion with squalene oil instead ofsqualane oil. KLH is a carrier protein that promoted enhanced S'In -specific antibody responses in preclinical models. Because of the promiscuous expression ofSTn on adenocarcinomas derived from numerous tissues, Tbemtope" has been tested clinically as a treatment for a variery of tumor types (breast. colon, ovarian or pancreatic). The vast majority of patients (>1100 of 1500) have been treated for breast cancer. Preclinical testing in mice demonstrated that treatment with low-dose cyclophosphamide prior to immunization leads to enhanced antigen-specific antibody responses." presumably via the inhibition of suppressor cells. Some of the early clinical trials of the vaccine. therefore, also evaluated whether cyclophosphamide pretreatment would have the same effect in humans. In an initial dose-escalation study. 12 metastatic breast cancer patients were administered low-dose (300 mg/m 2) IV cy1cophosphamide on day -3 and four doses ofSTn-KLH (25 /-lg. 100 /-lg or 500 ug) combined with Detox" adjuvant at 2 week intervals, with four monthly injections thereafier." The 500 flg dose caused excessiveDTH reactions and was therefore eliminated from the protocol. At the lower doses. vaccine-associated toxicity was minimal. with side-effects limited to granulomas/ulcerations at the injections sites in 5 patients. All patients developed anti-STn IgG and IgM responses. Two patients had partial remissions lasting 6 months, while several others demonstrated disease stability for 3 -10 months. In a second trial. 100 /-lg ofthe STn-KLH vaccine combined with Detox"was administered to 23 metastatic breast cancer patients (with or without cyclophosphamide pretreatment) at weeks 0.2,5 and 9, with four additional monthly injections administered to patients with responding or stable disease after the first round of vaccinations.S Toxicity was again limited to granuloma formation at the injection sites. All patients developed anti-STn IgG and IgM responses and IgM responses were significantly higher in patients pretreated with cyclophophamide. Because 5 patients developed progressive disease during the initial 12 week period. only 18 patients received all four injections. ofwhich rwo had minor responses. Subsequent prospective trials evaluated the effect of different low-dose cyclophosphamide pretreatments on the response to STn-KLH combined with Detox" .63 Individuals receiving intravenous cyclophosphamide generated stronger anti-STn antibody responses and lived significantly longer than those administered the vaccin e with oral or no cyclophosphamide. An inverse correlation was observed between anti-STn antibody titers and tumor growth. and a lower percentage ofthe patients receiving intravenous cyclophosphamide pretreatment had progressive disease at 9 weeks. As a result of these trials, addition of cyclophosphamide pretreatment to the Theratope~vaccination protocol has become standard practice. In a randomized. double-blind Phase III trial, metastatic breast cancer patients whose tumors did not progress after front-line chemotherapy were treated with intravenous cyclophosphamide followed by Theratope~( STn-KLH and Enhanzyn adjuvant) or intravenous cyclophosphamide followed by KLH alone combined with the Enhanzyn adjuvant (negative control).64.66An improved formulation of 1heratope~thatgenerated higher anti-STn antibody titers was used in this study, such that the STn/KLH ratio was increased compared to the preparation used in earlier studies. Concomitant endocrine hormone therapy was permitted and patients were stratified with regard to hormone therapy and response to initial therapy. The study was large (1028 patients) and subset analysis was conducted on patients that received hormone treatment. With respect to time to progression. there was no statistical difference berween 1heratope~-treated and control arms in the study. regardless ofwhether selective estrogen -receptor modulators or arornatase inhibitors were used as adjunct hormonal therapy. Those patients who developed high titer antibody responses. however. had significantly longer survival times (41.1 months versus 25.4 months. p = 0.01). The failure of the Phase III trial may be related to the observations from early trials that induction of immune response to Theratope ~ takes, on average, 17 weeks, while the time to disease progression in untreated patients is appoximately 12 weeks. Thus, the increased tumor burden after

Monopbospboryl Lipid A (MPL) asan Adjuvantfor Anti-CancerVaccines: ClinicalResults

119

disease progression may suppress S'In -specific immune responses (via production ofimmunosuppressive molecules such as mucin) or may simply overwhelm the nascent immune response that is being generated to the vaccine. To test the vaccine in a situation where tumor burden was low during the vaccination regimen, Theratope-was evaluated in breast and ovarian cancer patients in conjunction with chemotherapy and autologous stem cell rescue. Trial designers reasoned that immune responses to the vaccine would have a chance to develop prior to disease progression in these patients. Since low tumor burden correlated with stronger immune responses and stronger immune responses correlated with longer survival times in initial trials, the strategy appeared sound. Beginning in 1995, 53 breast and 17 ovarian cancer patients were treated with Tberatope" vaccine beginning 30 to 151 days post-stem cell infusionY ·70Toxicity was limited to induration and erythema at injection sites, with some patients experiencing transient flu-like symptoms. Most patients developed elevated anti-STn IgG titers , usually after the third vaccination and approximately halfofthe patients developed S'In-specific T-cell proliferative and cytolytic responses. Induction of strong (versus weak) cytolytic T-cell responses by the vaccine was associated with longer remission times (p =0.047), suggesting that development ofanti-STn cytotoxic cells may also correlate with a clinical effect. Phase III trials to evaluate Theratope "in transplant patients have not been initiated at this time , though pilot studies are planned to determine whether addition ofIL-2 or GM-CSF to Theratope-treatment is beneficial in tran splant patients.

Vaccine Targeting Ras-Expressing Tumors The mutant oncogene ras is a well characterized TAA and is present in more than 20% of all

solid tumors, making it a prospective target for cancer immunotherapy. Taking advantage of the fact that a single point mutation (codon 12) in the rasgene accounts for 90% ofall ras mutations, investigators from the National Cancer Institute developed and tested vaccines containing one of the three B -mer mutant ras peptides expressed by tumor cells and Detox" adjuvant." Fifteen patients with a variety of cancers (colon, pancreas, nonsmall cell lung , duodenal, rectal and appendix) that carried defined ras mutations received three monthly subcutaneous vaccinations with the appropriate peptide at one of five doses (100 , 500 , 1000, 1500 or 5000 !!g) of peptide with a constant dose of Deiox: (25 !!g MPL and 250 !!g CWS). The vaccines were well tolerated, regardless ofdose, with local reaction (sterile granuloma) at the injection site in one patient after the third vaccination. T-cell proliferative responses directed against the vaccine peptide, but not against wild-type or other mutant ras peptides, developed in 3 of 10 evaluable patients as a result ofvaccination. HLA-A2/ras peptide-specific CD8+ cytotoxic T-cell responses also developed in two of three patients evaluated. These cellular responses were not dependent on dose. No major therapeutic responses were observed in any ofthe elevenpatients who received allthree vaccinations. One patient who showed stable disease after the three-dose regimen was administered three additional vaccinations and continued to show no evidence oftumor progression 10 months later.

Conclusion

A variety ofvaccines designed for cancer immunotherapy have been tested in clinical trials for more than two decades. Investigators realized early on that add ition ofadjuvants to cancer vaccines would be required to overcome the poor immune responses that are generally elicited to antigens contained within these vaccines. Although the effectiveness ofLPS as an immunomodulator has long been known, the pharmacologic use ofpurified LPS (or lipid A) asan adjuvant is precluded by its toxicity. In this regard , LPS is highly pyrogenic and promotes systemic inflammatory response syndrome." In an effort to uncouple the immunomodulatory effects oflipid A from its toxicity, Ribi er al developed 3-0-desacyl-4'-monophosphoryl lipid A (MPL), which comprises the lipid A portion ofLPS from which the (R)- 3-hydroxytetradecanoyl group and the l-phosphate have been removed? by successive acid and base hydrolysis. LPS and MPL induce similar cytokine profiles, but MPL is at least 100-fold less toxiC.9•1O MPL, as the active ingredient in MPI:' adjuvant or as one of the active ingredients in Detox"" adjuvant (with CWS and oil), AS02B adjuvant (with QS21 in an oil in water emulsion) or AS15 (a liposomal formulation with QS21 and CpG), has

Lipid-Ain Cancer Therapies

120

been administered to more than 300,000 hwnan subjects in studies ofnext-generation vaccines," including the patients enrolled in the clinical trials discussed in this chapter. At this time , MPL is the only lipid A derivative that has been clinically tested as an adjuvant for cancer vaccines. The reasons for the limited clinical success ofcancer vaccines in general are not clear, but may relate to the fact that most trials have been conducted in patients with late-stage disease. where bulky metastatic tumors may evade or suppress the immune system and prevent induction ofefficacious anti-tumor cellular and/or hwnoral adaptive immune responses. The very encouraging data obtained in an adjuvant setting in early stages NSCLC with the MAGE3 protein formulated in AS02B support that hypothesis. Another constraint is the weakly immunogenic nature ofmany TAAs, which are often "self" in nature and. therefore, not good targets for the induction ofeffector T-cell responses . Efforts to move beyond these limitations include adding potent adjuvants, such as MPL, to the vaccines, as well as combining immunotherapy with surgery. chemotherapy. radiation therapy and/or autologous stern-cell/dendritic cell therapy. The hope is that chemotherapyI radiation therapy will reduce tumor burden and deplete suppressive Tvcells, while optimized vaccination protocols will allow for enhanced induction, proliferation and activity oftumor-specific effector T-cells that can eliminate residual tumor, Such combination therapies are currently in the early stages of clinical testing and may lead to better options for the treatment ofcancer.

References I.

F, Schmid P, Mackensen A et al. Phase II trial of intr avenous endotoxin in patients with colorecral and nonsmall cell lung cancer. Eur] Cancer 1996; 32:1712-1718. 2. Engelhardt R. Mackensen A. Galanos C. Phase I trial of intravenously administered endotoxin (Salmonella abortus equi) in cancer patients. Cancer Res 1991; 51:2524-2530. 3. Mackensen A, Galanos C. Engelhardt R. Modulating activity of interferon-gamma on endotoxin-induced cyrokine production in cancer patients. Blood 1991; 78:3254-3258 . 4. Mackensen A. Galanos C. Wehr U et al. Endotoxin tolerance : regulation of cytokine production and cellular changes in response to endotox in application in cancer patients. Eur Cytokine Nerw 1992; 3:571-579. 5. Goto S. Sakai S. Kera] et al. Intradermal administration of lipopolysaccharide in treatment of human cancer. Cancer Immunol Imrnunother 1996; 42(4) :255-261. 6. Vosika G]. Barr C. Gilbertson D. Phase-I study of intravenous modified lipid A. Cancer Immunol Immunother 1984: 18(2):107-112. 7. Kiani A, Tschiersch A. Gaboriau E et al. Downregulation of the pro inflammatory cyrokine response to endotox in by pretreatment with the nontoxic lipid A analog SDZ MRL 953 in cancer patients . Blood 1997; 90(4) :1673-1683 . 8. de Bono ]S. Dalgleish AG, Carmichael] et al. Phase 1 study of ONO-4007. a synthetic analogue of the lipid A moiety of bacterial lipopolysaccharide. Clin Cancer Res 2000; 6(2) :397-405. 9. Myers KR. Truchot AT, Ward] et al. A critical determinant of lipid A endoroxic activity. In: Nowotny A. Spitzer ]]. Ziegler E], editors. Cellular and Molecular Aspects of Endotoxin Reactions. Amsterdam : Elsevier Sciences Publishers B V, 1990:145-156. 10. Ulrich ]T, Masihi KN . Lange W. Mechanisms of nonspecific resistance to microbial infections induced by trehalose dimycolate (TDM) and monophosphoryllipid A (MPL) . In: Masihi KN. Lange W; editors. Advances in the Bioscicnces, Great Britain : Pergamon Journals Ltd.• 1988:167-178. II. Evans]T. Cluff CW; Johnson DA et al, Enhancement of antigen-specificimmunity via the TLR4ligands MPL adjuvant and Ribi 529. Expert Review Vaccines 2003 ; 2(2) :219-229. 12. Woodlock T]. Sahasrabudhe DM . Marqu is DM er al. Active specific immunotherapy for metastatic colorectal carcinoma: Phase I study of an allogeneic cell vaccine plus low-dose interleukin-la.] Irnmunother 1999; 22(3) :251-259. 13. Sherye SF, Frodin ]-E. Christens son B. Immunohistochemical mon itoring of metastatic colorectal carcinoma in patients treated with monoclonal antibodies (MAb 17-01A). Cancer Immunol Imrnunorher 1988; 27:154 -162. 14. Neidhart]. Allen KO. Barlow DL er al. Immun ization of colorecral cancer patients with recombinant baculovirus-derived KSA (Ep-CAM) formulated with monophosphoryllipid A in liposomal emulsion, with and without granulocyte-macrophage colony-stimulating factor. Vaccine 2004; 22(5-6):773-780. 15. Papsidero LD. Kuriyama M. Wang MC er al. Prostate antigen : a marker for human prostate epithelial cells.] Nat! Cancer Insr 1981; 66(1) :37-42. 16. Oesterling ]E . Prostate specific antigen : a critical assessment of the most useful tumor marker for adenocarcinoma of the prostate.] Uro11991; 145(5):907-923. 0[[0

Monophosphoryl Lipid A (MPL) asan Adjuvantfor Anti-CancerVaccines: ClinicalResults

121

17. Harris DT, Matyas GR, Mastrangelo M] et al. Inducing immunity to prostate specific antigen (PSA) in prostate cancer patients. Proc ASCO 1999; 18:1693. 18. Meidenbauer N, Harris DT, Spitler LE et al. Generation of PSA-reactive effector cells after vaccination with a PSA-based vaccine in patients with prostate cancer. Prostate 2000; 43(2):88-100. 19. Mitchell MS, Kan-Mitchell ], Kempf RA et al. Active specific immunotherapy for melanoma: Phase I trial of allogeneic lysates and a novel adjuvant. Cancer Res 1988; 48(20):5883-5893. 20. Mitchell MS, Harel W; Kempf RA et al. Active-specific immunotherapy for melanoma. ] Clin Oncol 1990; 8(5):856-869. 21. Vose BM. Quantitarion of proliferative and cytotoxic precursor cells directed against human tumours: limiting dilution analysis in peripheral blood and at the tumour site. Inr ] Cancer 1982 ; 30(2):135-142. 22. Mitchell MS, Harel W; Groshen S. Association of HLA phenotype with response to active specific immunotherapy of melanoma.] Clin OncoI1992; 10(7):1158-1164. Sosman ]A, Sondak VK. Melacine: an allogeneic melanoma tumor cell lysate vaccine. Expert Rev Vaccines 2003 ; 2(3) :353-368. 23 . Elliott GT, Mcl.eod RA, Perez] er al. Interim results of a phase II multicenter clinical trial evaluating the activity of a therapeutic allogeneic melanoma vaccine (rheraccine) in the treatment of disseminated malignant melanoma. Semin Surg Onco11993; 9(3) :264-272. 24. Mitchell MS, Von Eschen KB. Phase III trial of Melacine melanoma theraccine versus combination chemotherapy in the treatment of stage IV melanoma. Proceedings of the American Society of Clinical Oncology 1997; 16,494a. 25 . Sosman ]A, Unger ]M, Liu PY et al. Adjuvant immunotherapy of resected, intermediate-thickness, node-negative melanoma with an allogeneic tumor vaccine : impact of HLA class I antigen expression on outcome, ] Clin Oncol 2002; 20(8):2067-2075. 26. Sondak VK, Sosman ]A . Results of clinical trials with an allogeneic melanoma tumor cell lysate vaccine: Melacine', Semin Cancer BioI 2003; 13:409-415. 27. Sosman ]A, Sondak VK. Melacine : an allogeneic melanoma tumor cell lysate vaccine. Expert Rev Vaccines 2003 ; 2(3) :353-368. 28. Mitchell MS, ]akowatz ], Harel W et al. Increased effectiveness of interferon alfa-2b following active specific immunotherapy for melanoma. ] Clin Onco11994; 12(2):402-411. 29. Vaishampayan U, Abrams ], Darrah D er al. Active immunotherapy of metastatic melanoma with allogeneic melanoma lysates and interferon alpha . Clin Cancer Res 2002; 8(12):3696-3701. 30. Mitchell MS. Immunotherapy as part of combinations for the treatment of cancer. Inc Immunopharmacol 2003; 3(8):1051-1059. 31. Mitchell MA, Abrams ], Kashani-Saber M et al. Interim analysis of a phase III stratified randomized trial of Melacine + low-dose Intron-A versus high-dose Intron-A for resected stage III melanoma. Am Soc Clin Oncol 2003; 22:709a . 32. Schultz N, Oratz R, Chen D et al. Effect of DETOX as an adjuvant for melanoma vaccine. Vaccine 1995 ; 13(5):503-508. 33. Eron 0 , Kharkevirch DD, Gianan MA er al. Active immunotherapy with ultraviolet B-irradiated autologous whole melanoma cells plus DETOX in patients with metastatic melanoma. Clin Cancer Res 1998; 4(3):619-627. 34. Wong R, Lau R, Chang] et al. Immune responses to a class II helper peptide epitope in patients with stage III /IV resected melanoma. Clin Cancer Res 2004; 10(15):5004-5013. 35. Lienard D, Rimoldi D, Marchand M er al. Ex vivo detectable activation of Melan-Avspecific T'cells correlating with inflammatory skin reactions in melanoma patients vaccinated with pepcldes in IFA. Cancer Immun 2004; 4:4 . 36. Marchand M, Punt C], Aamdal S et al. Immunisation of metastatic cancer patients with MAGE-3 protein combined with adjuvant SBAS-2: a clinical report. Eur] Cancer 2003; 39(1):70-77 . 37. Vantomrne V, Dantinne C, Amrani N. Immunologic analysis of a phase lIII study of vaccination with MAGE-3 protein combined with the AS02B adjuvant in patients with MAGE-3 -positive tumors. ] Immunorher 2004; 27:124-135. 38. Atanackovic D, Altorki NK , Stockert E et al. Vaccine-induced CD4+ T-cell responses to MAGE-3 protein in lung cancer patients.] Immunol 2004 ; 172(5) :3289-3296. 39. Brichard V. Development of cancer vaccines with the MAGE-3 protein. Cancer Immunity 2005; 5(1) :16. 40 . Brichard V. CVADD 2005; Portugal. 41. Zotter S, Hageman PC, Lossnitzer A et al. Tissue and tumor distribution of human polymorphic epithelial mucin . Cancer Rev 1988; 11-12:55-101. 42. Ho SB, Niehans GA, Lyfiogr C et al. Heterogeneity of mucin gene expression in normal and neoplastic tissues. Cancer Res 1993 ; 53(3) :641-651.

122

Lipid-Ain Cancer Therapies

43 Burchell J, Gendler S, Taylor-Papadimitriou J er al. Development and characterization of breast cancer reactive monoclonal antibodies directed to the core protein of the human milk mucin. Cancer Res 1987; 47(20) :5476-5482 . 44. Kjeldsen T, Clausen H, Hirohashi S et al. Preparation and characterization of monoclonal antibodies directed to the tumor-associated O-linked sialosyl-2-6alpha-N-acetylgalactosaminyl(sialosyl-Tn) epitope. Cancer Res 1988; 48(8) :2214-2220. 45. Ramanathan RK, Lee KM, McKolanis Jet al. Phase I study of a MUCI vaccine composed of different doses of MUCI peptide with SB-AS2 adjuvant in resected and locally advanced pancreatic cancer. Cancer Immunol Immunother 2005; 54(3) :254-264. 46. Palmer M, Parker J, Modi S er al. Phase I study of the BLP25 (MUCl peptide) liposomal vaccine for active specific immunotherapy in stage IIIB/IV nonsmall-cell lung cancer. Clin Lung Cancer 2001; 3(1):49-57. 47. North SA, Graham K, Bodnar D er al. A pilot study of the liposomal MUC 1 vaccine BLP25 in prostate specific antigen failures atter radical prostatectomy. J Urol 2006; 176(1) :91-95. 48. Butts C, Murray N, Maksymiuk A er al. Randomized phase lIB trial of BLP25 liposome vaccine in stage IIIB and IV nonsmall-celliung cancer. J Clin Onco12005; 23(27):6674-6681. 49. Machiels JP, Reilly RT, Emens LA et al. Cyclophosphamide, doxorubicin and paclitaxel enhance the antitumor immune response of granulocyte/macrophage-colony stimulating factor-secreting whole-cell vaccines in HER-2/neu tolerized mice. Cancer Res 2001; 61(9) :3689-3697. SO. Bass KK, Mastrangelo M]. Immunoprotentiation with low-dose cyclophosphamide in the active specific immunotherapy of cancer. Cancer Immunollmmunother 1998; 47:1-12. 51. Longenecker BM, MacLean G. Prospects for mucin epitopes in cancer vaccines. Immunologist 1993; 1:89-93. 52. Itzkowitz SH, Bloom EJ, Kokal WA et al. Sialosyl-Tn. A novel mucin antigen associated with prognosis in colorecral cancer patients . Cancer 1990; 66(9) :1960-1966. 53. Springer GF. T and Tn, general carcinoma autoantigens. Science 1984; 224(4654):1198-1206. 54. Kobayashi H , Terao T, Kawashima Y. Serum sialyl Tn as an independent predictor of poor prognos is in patients with epithelial ovarian cancer. J Clin Oncol 1992; 10(1):95-101. 55. Thor A, Ohuchi N, Szpak CA er al. Distribution of oncofetal antigen tumor-associated glycoprotein-72 defined by monoclonal antibody B72.3. Cancer Res 1986; 46(6) :3118-3124. 56. Nuti M, Teramoto YA, Mariani-Costantini Ret al. A monoclonal antibody (B72.3) defines patterns of distribution of a novel tumor-associated antigen in human mammary carcinoma cell populations. Inr J Cancer 1982; 29(5) :539-545. 57. Yoneawa S, Tachidawa T, Shin S. Sialosyl-Tn antigen : its distribution in normal human tissues and expression in adenocarcinomas. Am] Clin Parhol 1992; 98:167-174. 58. Longenecker BM, Reddish M, Miles D et al. Synthetic Tumor-Associated Sialyl-Tn Antigen as an Immunotherapeutic Cancer Vaccine. Vaccine Res 1993; 2(3) :151-162. 59. O'Boyle K, Zamore R, Adluri S et al. Immunization of colorectal cancer patients with modified ovine submaxillary gland mucin and adjuvants induces IgM and IgG antibodies to sialylated Tn. Cancer Res 1992; 52(20) :5663 -5667. 60. Berendt MJ, North R]. Tvcell-mediared suppression of anti-tumor immunity. An explanation for progressive growth of an immunogenic tumor. J Exp Med 1980; 151(1) :69-80. 61. MacLean GD, Reddish M, Koganty RR er al. Immunization of breast cancer patients using a synthetic sialyl-Tn glycoconjugate plus Derox adjuvant. Cancer Immunol Imrnunother 1993; 36(4) :215-222. 62. Miles DW; Towlson KE, Graham R et al. A randomised phase II study of sialyl-Tn and DETOX-B adjuvant with or without cyclophosphamide pretreatment for the active specific immunotherapy of breast cancer. Br J Cancer 1996; 74(8):1292-1296. 63. MacLean GD, Miles DW; Rubens RD er al. Enhancing the effect of THERATOPE Stn-KLN cancer vaccine in patients with metastatic breast cancer by pretreatment with low dose intravenous cyclophosphamide. J Immunorher 1996; 19(4):309-316. 64. Miles D, Ibrahim N, Roche H . An international randomized phase III clinical trial of STn-KLH (Therarope) therapeutic cancer vaccine in metastic breast cancer patients . Proceedings 27th Annual San Antonio Breast Cancer Symposium 2003;(36). 65. Ibrahim NK, Murray J, Parker]. Humoral immune-response to naturally occurring STn in metastatic breast cancer (MBC prs) treated with STn-KLH vaccine. Am Soc Clin Oncol 2004; 22:S174. 66. Majordomo J, Tres A, Miles D. Long-term follow-up of patients concomitantly treated with hormone therapy in a prospective controlled randomized multicenter clinical study comparing STn-KLH vaccine with KLH control in Stave IV breast cancer following front-line chemotherapy. Proc Am Soc Clin Oncol 2004 ; 22:1882S.

Monophosphoryl Lipid A (MPL) asan Adjuvantfor Anti-Cancer Vaccines:Clinical Results

123

67. Holmberg LA. Oparin DV. Gooley T er al. Clinical outcome of breast and ovarian cancer patients treated with high-dose chemotherapy. autologous stem cell rescue and THERATOPE STn-KLH cancer vaccine. Bone Marrow Transplant 2000: 25(12) :1233-1241. 68. Holmberg LA. Sandmaier BM. Theratope(R) vaccine (STn-KLH). Expert Opin BioI Ther 2001 : 1(5):881-891. 69. Holmberg LA. Oparin DV. Gooley T er al. The role of cancer vaccines following autologous stern cell rescue in breast and ovarian cancer patients: experience wirh the STn-KLH vaccine (Therarope). Clin Breast Cancer 2003: 3 SuppI4:S144-S151. 70. Holmberg LA. Sandmaier BM. Vaccination wirh Theratope (STn-KLH) as treatment for breast cancer. Expert Rev Vaccines 2004: 3(6) :655-663. 71. Khleif SN. Abrams SI. Hamilton JM et al. A phase 1 vaccine trial with pepridcs reflecting ras oncogene mutations of solid rumors. J Imrnunother 1999: 22(2):155-165. 72. Johnson AG. Molecular adjuvants and immunomodularors: new approaches to immunization. Clin Microbiol Rev 1994: 7(3):277-289.

CHAPTER

11

Antitumoral Effects ofLipids A, Clinical Studies Marc Bardon" and Daniele Reisser

Abstract

C

ancer remains the second leading causeofdeath. after cardiovasculardiseases.in industrialized countries. The first goal to achieve is to prevent cancer occurrence or to diagnose it at an early and curable stage. Some screening strategies have been developed, with controversies across countries. for several cancer type; colorectal, breasts or prostate cancer for example. Treatment ofcancer is generally based on surgery and radiotherapy for localized and attainable tumors. associated. in some cases, with adjuvant chemotherapy. Chemotherapy can also be used as first line treatment for disseminated diseases. The formulation of therapeutic strategies to enhance immune-mediated tumor destruction is a central goal ofcancer immunology. Substantive progress toward delineating the mechanisms involved in innate and adaptive tumor immunity has improved the prospects for crafting efficacious treatments LPS and their active component lipid A, have been used in tumor therapy since the 19th century. Studies in animal models have shown promising results on different models ofcancer but data from human trial are scarce. The published Phase-I cancer studies have shown that lipid A analogues are usually well tolerated. most ofthe side effects being likely related to immune response . i.e.•fever. chills and rigor. The administration ofseveral lipids A analogues was shown to result in a significant increase in circulating levels ofseveral cytokines but no objective antitumor responses were observed. Therefore clinical activity of such molecules deserves further experiments, likely in conjunction with chemotherapy.

Introduction

Even ifrecent data suggest a downward trend ofmortality rates for all cancers combined based on declining rates for many individual sites. including colon.' with only few exceptions affecting mainly females (e.g.• lung cancer)2.3 or specific sites such as liver," cancer remains the second cause of death. after cardiovascular diseases. in industrialized countries. together accounting for over half of all deaths.' Most cancer patients are treated by a combination of surgery. radiation and/or chemotherapy. Whereas the primary tumor can, in most cases, be efficiently treated by a combination ofthese standard therapies. preventing the metastatic spread ofthe disease through disseminated tumor cells is ofien not effective. The eradication ofdisseminated tumor cells present in the blood circulation and micro-metastases in distant organs therefore represents another promising approach in cancer immunotherapy. When the immune system is not destroyed by *Corresponding Author: Marc Bardou-Clinical Pharmacology Unit & Laboratory of Cardiovascular Experimental Physiology and Pharmacology, Faculty of Medicine, 7 Bd Jeanne d'Arc, BP87900, 21079 Dijon, France. Email: [email protected]

LipidA in Cancer Therapy, edited by jean-Francois jeannin. ©2009 Landes Bioscience and Springer Science+Business Media.

126

Lipid A in Cancer Therapy

chemotherapy, it is able to recognize tumor-specific antigens and eventually eliminate cancer cells. Furthermore a recent review highlights the fact that some cancers, such as colorectal cancer, cause direct inhibition ofthe host's immune response with a detrimental effect upon prognosis, suggesting that immunotherapy offers a therapeutic strategy to counteract these effects.6 Comprehensive analysis oftumor immunology and new immunization protocols suggest that immunotherapy can become an efficacious treatment in the near future. Combination with radiotherapy or chemotherapy should be investigated? One ofthe means to reverse this cancer-induced immune tolerance and to stimulate immune system might be the use oflipopolysaccharides (LPS). These are components ofthe outer membrane of Gram-negative bacteria, composed of a polysaccharide, an oligosaccharide core and a lipid A. These compounds have the property ofinducing the secretion ofvarious cytokines such as tumor necrosis factor (TNF), interferon (IFN)y, interleukin (IL)1~ and IL6,8,9, as well as activation of immune cells including neutrophils," rnacrophages" and both CD4 and CD8 T-Iymphocytes, which infiltrate the tumors. I 2.13 Furthermore, these compounds have been shown to decrease suppressive cytokines like tumor growth factor (TGF)~.IO·l4 The first assaysusing bacterial extracts containingLP S as a treatment for cancer were performed in 1898 as clinical trials." Half a century later, the antitumoral effect was attributed to LPS in mouse subcutaneous tumors" and finally it was demonstrated that this effect is due to the lipid A component ofLPS.1 7Since lipids A and their derivatives are less toxic than LPS, most anti-cancer treatments aimed at activating an immune response against tumors have been developed using natural lipids A or synthetic analogs, either alone or as adjuvant to enhance the efficacy of therapeutic anticancer vaccines. Animal models permit the investigation ofthe mechanisms ofthe antitumoral effect oflipids A (for review, see Reisser et al l8and the corresponding chapter ofthis book); therefore this chapter will summarize the results obtained in clinical trials.

Immunological Background Underlying the Clinical Potential Interest

Recent advances in molecular biology and immunology have contributed to a better understanding oftumor growth and tumor-host interactions and have shown that tumor progression is favored by the host tolerance to his tumor. In order to overcome the immunological unresponsiveness of patients to their growing tumors, endotoxins and their active component, lipid A , or synthetic analogs, may favor the settlement of, or increase the antitumoral immune response . Emphasis has been put on the role of dendritic cells on the onset of an efficient immune response.'? In the presence of a danger signal frequently delivered by bacterial produces," they release costimulatory factors that allow specific lymphocytes to differentiate and kill tumor cells they would have otherwise ignored. In vitro, LPS stimulate the production of cytokines by hu man dendritic cells" and monophosphoryllipid A (MPL) induces their maturation and their ability to activate T -cells, as does a lipid A analogue." However, no data document this question in cancer patients. Few studies dealt with the ex vivo activity ofimmune cells from patients treated with LPS or lipids A. Vosika et al23did not detected any clear effect ofa treatment with MPL on immune cell activity. On the contrary, monocytes from patients treated with LPS from Salmonella abortus equi(S.abortus equi) displayed an activated phenotype. 23 ln vitro , the lipid A OM-174 increases natural killer (NK) cytotoxicity ofperipheral blood cells from cancer patients," The number ofwhite blood cells (WBC) generally decreased, albeit transiently, after LPS or lipid A Injection," in particular, monocytes'? and lymphocytes.P'" The CD4:CD8 ratio was enhanced." Meanwhile the number ofgranulocytes always increased after the injections.23.26-28 .3o In spite ofthe decrease in the number ofWBC, in most ofthe studies published so far, treatments ofpatients with LPS or lipid A enhanced the levelofblood cytokines.Indeed the production ofcytokines is a biological marker ofthe activation ofimmune cells and LPS and lipid A have the property ofinducing the secretion ofvarious cytokines in humans."

AntitumoralEffictsofLipidsA, Clinical Studies

127

LPS from S. abortus equi increased the concentration ofTNF and the activity ofIL6, IL8, granulocyte colony-stimulating factor (G-CSF) and macrophage (M)_CSP6.27 while there was no change in ILI~ and IL2 and IL7 was not detected. LPS from Pantoeaagglomerans increased the serum concentrations ofTNF, ILIa, IL6 and G-CSY SDZ MRL 953 increased, though not significantly, the serum concentrations ofTNF, ILI~, IL8, G-CSF and IL6. 30 TNF and IL6 production was induced after a treatment by ON0-4007,28 while this lipid A analog had no effect on GM-CSF, IFN-y and neopterin. Mononuclear cells from patients treated with LPS were primed for cytokine production ex vivo." Lipids A induce the in vitro production ofcytokines by human peripheral blood cells and monocytes v'" and MPL that ofreactive oxygen species. 36 At the difference with LPS, ON0-4007 induced few, if at all, in vitro production of cytokines by monocytes or total peripheral blood cells unless they had been primed with GM-CSF.37The real impact of these cytokines on tumor evolution was not deciphered and one can regret no study was interested in the production of IFNy which is a marker oflymphocyte acrivation." In humans, LPS can induce coagulation via TNF involvemenr.'? but its importance for tumor irrigation and therefore its effect on tumor growth is difficult to evaluate in human. Treatment with S. abortus equP4 and ON0-4007 27led to no changes in coagulation parameters, disseminated intravascular coagulation and clotting parameters The hallmark ofa bacterial infection is fever due to LPS pyrogenicity. Ie may have some relevance in cancer as hyperthermia is tested as a cancer treatment on its own, as reviewed by Christophi et al.40 This effect is mediated by cyrokines (mainly ILI~b),41 which production was inconsistently found after LPS or lipid A treatment. Indeed, treatments with LPS or lipid A induced fever in at least half the patients.22·24.26.29 Fever is generally opposed with anti-inflammatory drugs and Engelhardt? verified ibuprofen did not affect cytokine production. However the role of fever in LPS treatment efficacy was recently discussed by Hoption Cann et al43 who wondered whether fever inhibition might impede treatment efficiency.

Clinical Studies

Few Phase -I trials were performed with intravenous (i.v.) injections oflipids A, from S. typhimurium or S.minnesota,22 or the synthetic lipids A SDZ MRL 95Y9and ONO-4007. 27 Vosika et al22 have conducted a Phase -I trial with MPL prepared from S. typhimurium and S minnesota.

In this study patients received i.v.MPLA, twice weekly for a total of4 weeks, at the following dose levels of 10,25, 50, 100 and 250 micrograms/rn! body surface area. One additional patient was treated at the dose level of500 rnicrograrns/m', In this study, lipid A was generally well tolerated, the major clinical coxicity being fever, chills and rigor, which occurred in over 50% of the treatments at doses of 250 micrograms/m'. Two instances of bronchospasm occurred in one patient who received 250 rnicrograrns/rrr'. The patient who received 500 micrograrns/rn" became hypotensive. Sequential clinical data showed no evidence of renal or hepatic toxicity, Interestingly a transient decrease in the WBC and platelets was observed during the first 24 hours after therapy. Immune function testing suggested a shift in monocyte populations with activated cells moving into the tissue . Unfortunately no direct objective antitumor activity or necrosis was observed in this group ofpatients. From this trial, Vosika et al 22 recommended a dose of up co 100 micrograms/rn' co be used with acceptable toxiciry for further evaluation of its clinical activity as a single agent in combination with other immunomodulacors. More than ten years later Kiani et al 29 conducted another phase one trial on cancer patients using the synthetic lipid A analog SDZ MRL 953 which has been shown co be protective against endotoxic shock and bacterial infection in preclinical in vivo models. This authors, as part ofa trial ofunspecific imrnunostimulation in cancer patients, conducted a double-blind, randomized, vehicle-controlled Phase-I trial ofSDZ MRL 953 to investigate its

128

Lipid A in Cancer Therapy

biologic effects and safety ofadministration in humans and its influence on reactions to a subsequent challenge ofendotoxin (S. abortus equi). In this Phase-I trial, twenty patients were treated i.v, with escalating doses ofSDZ MRL 953 or vehicle control, followed by an i.v, application ofendotoxin (2 ng/kg ofbody weight [BW]). The first result of this study is that administration of the lipid A analogue, SDZ MRL 953, was safe and well-tolerated. On the contrary to the transient decrease in WBC count that was observed in the stu dy by Vosika et a122SDZ MRL 953 itselfincreased granulocyte counts. This increase was associated with and likely related to , an increase in serum levels of G-CSF and IL6. Surprisingly, the proinflarnmatory eytokines TNF, III ~ and IL8 were not significantlyincreased by SDZ MRL 953 administration. In the second part ofthe trial , pretreatment with SDZ MRL 953 markedly reduced, compared with vehicle control, the release ofTNF, ILl~, IL8, IL6 and G-CSF, but augmented the increase in granulocyte counts to endotoxin. Once again this study suggested an overall safety oflipid A analogue in cancer patients. More recently de Bono et aJ27 published the results of a Phase -I trial using the lipid A ON0-4007, that is a synthetic analogue ofthe lipid A moiety ofbacterial LPS, which, in animal models , exhibits antitumor activity by the induction of intratumoral TNF, the potentiation of tumor-infiltrating macrophages and the inhibition ofangiogenesis. ILla, IL6 and ILl2 induction by ONO-4007 activates NK cells to up-regulate IFNy and nitric oxide synthase activity. In this trial ON0-4007 was given to 24 patients (13 males and 11 females; median age, 53 years) as a 30-min i.v, infusion on day 1, followed on day 15 by a first treatment cycle consisting of three weekly infusions at the same dose, followed by a rest period of 1 week. Cohorts of six pat ients received up to a maximum offour treatment cycles at increasing dose levels (75, 100 and 125 mg). De Bono et a1 27 found a maximum tolerated dose of 125 mg, with grade 3 National Cancer Institute Common Toxicity Criteria toxicity (NCICTC; rigors with cyanosis) occurring in two ofsix patients at thi s dose level. An additional sixpatients were treated at 100 mg, the dose below the maximum tolerated dose. Other toxicities included grade 2 NCICTC myalgia, nausea and hypotension. In this study the pharmacokinetics of ON0-4007 appeared to be independent of dose and showed linearity with respect to time. ON0-4007 was described to have a low systemic clearance (approximately 1.3 ml/min) and a small volume ofdistribution (5-8 liters) with a long half-life of74-95 hours and the administration of this drug was shown to result in a significant increase in circulating levels ofTNF and IL6. As it was the case with the previously reported two phase one trials,22,29no objective antitumor responses were observed. Seven patients maintained stable disease for at least two cycles,whereas five patients maintained stable disease for the full four -cycle duration ofthe study. Limited toxicity was observed with Salmonella lipids A or ONO-4007 and the maximal tolerated dose (MTD) was not reached for SDZ MRL 953. One Phase-I study on cancer patients was conducted with 0 M-174 another analogue oflipid A, by Viens P and aI (unpublished results up to date) . This study was aimed to assessthe tolerance of and the biological response to, incr easing doses of OM-174 administered as single i.v, infusion. The first dose that was administered was half ofthe no adverse event level obtained in a phase one study conducted in healthy volunteers, i.e., 1.25 ug/kg or 50 !J.g/m2. Only one serious adverse event (SAE) was reported at the dose of800 ug/rn' with rigor hypotension , cyanosis and hypothermia. After i.v, acetaminophen this patient experienced full recovery within 24 hours and was discharged from hospital as initially scheduled. Three additional patients were created at the 800 ug/rn! and the study was continued up to the 1300 !J.g/m2dose without reaching the maximal tolerated dose. A biological response was assessed by cyrokine measurement. To summarize the findings that will be published as a full paper a TNF increase was noted, with very high levels (up to 4000 pg/

Antitumoral Effects ofLipidsA, ClinicalStudies

129

ml) for some patients without clear correlation with the dose ofOM-174 that was administered. Furthermore even ifthe patient who experienced the SAE had a high levelofplasmatic TNF, some other patients had even higher levels ofTNF without experiencing SAE. Once again no antitumor activity was reported in this clinical trial . There is another ongoing (as of early 2007) Phase-I study with OM-174 in cancer patients aimed to assess safety, efficacy and biological response of OM-174 with an incremented doses (600,800 and 1000 flg/m 2) and injections (5,10 or 15) protocol. To summarize the results ofthese four clinical trials,with only three published as full paper.lipid A analogues administration in cancer patients was associated with a biological response, evidenced most consistently by a TNF and an IL6 production. But no clear dose-response was observed for cytokine production and no objective ant itumor activity was observed in any of thi s four trials.

Conclusion

From a general point ofview, treatments with lipids A alone have the advantage of bypassing the investigation of specific tumor antigens and are therefore easier to establish on a broader range ofcancers. However, despite the encouraging results obtained, most human trials use lipids A as adjuvant . This is mainly due to the need of minimal quantities oflipids A in vaccines, thus minimizing their eventual toxicity and the goal of generating a more specific immune response against particular tumors. Only hints were afforded concerning the mechanisms of action of treatments with LPS and lipids A in cancer therapy and further stu dies are needed to decipher how to take advantage of the different ways lipid A could oppose tumor growth. Treatments generally lead to endotoxin tolerance ; the beneficial or detrimental role is discussed in another chapter in the present book. None of the published studies with lipids A used as anticancer therapy has been able to show any beneficial effect in terms of tumor control. However, in a field which often lacks efficient therapy, the encouraging results obtained so far indicate that treatments with lipids A could represent a hope. In the situations where no effective therapy is available, it could be ofinterest to test immunotherapy and particularly lipid A, on small tumors. Immunotherapy with lipid A has already been used to activate or reactivate memory cells after chemotherapy. It may also be used after tumor burden reduction by surgery or radiotherapy as a complementary therap y to stimulate or recover the immune system. It is therefore worthwhile to expand further the stu dies and tr ials with these compounds to determine the best circumstances and conditions for the application of these treatments in view ofclarifying whether lipids A will become a clinical approach for the treatment of cancer.

References

1. Cress RD, Morris C, Ellison GL et al. Secular changes in colorectal cancer incidence by subsite, stage at diagnosis and racc/ethniciry, 1992-2001. Cancer 2006 ; 107(5 Suppl):1142-1152. 2. Becker N, Altenburg HP, Stegmaier C et al. Report on trends of incidence (1970-2002 ) of and mortalit y (1952-2002) from cancer in Germany. J Cancer Res Clin Oncol 2006. 3. Levi F, Lucchini F, La Vecchia C. Trends in cancer mortality in Switzerland, 1980-2001. Eur J Cancer Prev 2006; 15(1):1-9. 4. West J, Wood H. Logan RF er al. Trends in the incidence of primary liver and biliary tract cancers in England and Wales 1971-2001. Br J Cancer 2006 ; 94(11 ):1751-1758. 5. Ho yerr DL , Heron MP, Murph y SL er al. Deaths: final data for 2003 . Nat! Vital Stat Rep 2006 ; 54(13):1-120. 6. Evans C, Dalgleish AG, Kumar D. Review article: immune suppression and colorectal cancer. Aliment Pharmacol Ther 2006; 24(8):1163-1177. 7. Carpenti er AF, Meng Y. Recent advances in immunotherapy for human glioma. Curt Op in Onc ol 2006; 18(6):631-636. 8. van de Wiel PA, van der Pijl A. Bloksma N. Role of tumour necrosis factor in the tum our-necrot izing activity of agents with diverging toxicity. Cancer Immunol Irnrnunorher 1991; 33(2):115-120 . 9. Blondiau C, Lagadec P, Lejeune P er al. Correlation between the capacity to activate macrophages in vitro and the antitumor activity in vivo of lipopolysaccharides from different bacterial species. Immuno biology 1994; 190(3):243-254.

130

LipidA in Cancer Therapy

10. Inagawa H . Nishizawa T. Takagi K et al. Antitumor mechanism of intradermal administration of lipopolysaccharide. Anticancer Res 1997; 17(3C):1961-1964. 11. Shimizu T, !ida K, Iwamoto Y et al. Biologicalactivities and antitumor effects of synthetic lipid A analog linked N-acylated serine. Int J Immunopharmacol1995; 17(5):425-431. 12. Onier N, Lejeune P, Martin M et al. Involvement of T-Iymphocytes in curative effect of a new imrnunornodulator OM 163 on rat colon cancer metastases. Eur J Cancer 1993; 29A(14) :2003-2009 . 13. On ier N, Hilpert S, Arnould L et al. Cure of colon cancer metastasis in rats with the new lipid A OM 174. Apoptosis of tumor cells and immunization of rats. Clin Exp Metastasis 1999; 17(4):299-306. 14. Lagadec P, Raynal S. Lieubeau B et al. Evidence for control of nitric oxide synthesis by intracellular transforming growth factor-beta1 in tumor cells. Implications for tumor development. Am J Pathol 1999; 154(6):1867-1876. 15. Coley WB. The treatment of inoperable sarcoma with the mixed toxins of Erysipelas and bacillus prodigiosus. j Am Med Assoc 1898; 31:589-595. 16. Shear MJ, Turner Fe. Chemical treatment of tumors. V. Isolation of the hemorrhage-producing fraction from Serratia rnarcescens (Bacillus prodigiosus) culture filtrate. J Nat! Cancer Inst 1943; 4(81-97). 17. Parr I, Wheeler E, Alexander P. Similarities of the anti-tumour actions of endotoxin , lipid A and double -stranded RNA. Br J Cancer 1973; 27(5) :370-389. 18. Reisser D, Pance A, Jeannin JF. Mechanisms of the antitumoral effect of lipid A. Bioessays 2002; 24:284-289. 19. Matzinger P. Tolerance, danger and the extended family. Annu Rev Immunol1994; 12:991-1045. 20. Verhasselc V. Buelens C, Willems F ec al. Bacterial lipopolysaccharide stimulates the production of cytokines and the expression of costimulatory molecules by human peripheral blood dendritic cells: evidence for a soluble CDI4-dependent pathway.J Immuno11997; 158(6):2919-2925. 21. Ismaili J, Rennesson J, Aksoy E et al. Monophosphoryllipid A activates both human dendritic cells and T-cells. J Immunol 2002; 168(2):926-932. 22. Vosika GJ, Barr C, Gilbertson D. Phase-I study of intravenous modified lipid A. Cancer Immuno l Imrnunother 1984; 18(2):107-112. 23. Rcisser D, Arnould L, Maynadie M et al. Lipid A OM-174 increases the natural killer activity of peripheral blood cells from breast cancer patients. J Endotoxin Res 1999; 5(4) :189-195. 24. Engelhardt R. Mackensen A, Galanos C. Phase 1 trial of int ravenously administered endotox in (Salmonella abortus equi) in cancer patients. Cancer Res 1991; 51(10) :2524-2530 . 25. Mackensen A, Galanos C, Wehr U ec al. Endotoxin tolerance : regulation of cytokine production and cellular changes in response to endotoxin application in cancer patients . Eur Cytokine Netw 1992; 3(6) :571-579 . 26. Mackensen A, Galanos C, Engelhardt R. Modulating activity of interferon-gamma on endotoxin-induced cytokine production in cancer patients. Blood 1991; 78(12):3254-3258. 27. de Bono JS, Dalgleish AG, Carmichael Jet al. Phase I study of ON0-4007. a synthetic analogue of the lipid A moiety of bacterial lipopolysaccharide. Clin Cancer Res 2000; 6(2) :397-405. 28. Otto F, Schmid P, Mackensen A et al. Phase II trial of intravenous endotoxin in patients with colorectal and nonsmall cell lung cancer. Eur J Cancer 1996; 32A(1O):1712-1718 . 29. Kian i A, Tschiersch A, Gaboriau E er al. Downregulation of the proin£lammatory cytokine response to endotoxin by pretreatment with the nontoxic lipid A analog SDZ MRL 953 in cancer patients. Blood 1997; 90(4) :1673-1683. 30. van Devenrer SJ, Buller HR, ten Care JW et al. Experimental endotoxemia in humans: analysis of cyto kine release and coagulation, fibrinolytic and complement pathways. Blood 1990; 76(12) :2520-2526. 31. Goto S, Sakai S, Kera J er al. Intradermal administration of lipopolysaccharide in treatment of human cancer. Cancer Immunol Immunother 1996; 42(4) :255-261. 32. Engelhardt R, Otto F, Mackensen A ct al. Endotoxin (Salmonella abortus equi) in cancer patients . Clinical and immunological findings. Prog Clin Bioi Res 1995; 392:253-261. 33. Matsuura M, Kiso M, HasegawaA. Activity of monosaccharide lipid A analogues in human monocytic cells as agonises or antagonists of bacterial lipopolysaccharide. Infect Immun 1999; 67(12 ):6286-6292. 34. Tarnai R, Asai Y, Hashimoto M er al. Cell activation by monosaccharide lipid A analogues utilizing Toll-like receptor 4. Immunology 2003; 110(1):66 -72. 35. Martin M. Michalek SM, Katz J. Role of innate immune factors in the adjuvant activity of monophosphoryllipid A. Infect Immun 2003; 71(5) :2498-2507. 36. Saha DC, Barna RS. Astiz ME et al. Monophosphoryllipid A stimulated up-regulation of reactiveoxygen intermediates in human monocytes in vitro. J Leukoc Bioi 2001; 70(3):381-385 . 37. Matsumoto N, Aze Y, Akimoto A et al. ON0-4007. an antitumor lipid A analog, induces tumor necrosis factor-alpha production by human monocyres only under primed state : different effects of ON0-4007 and lipopolysaccharide on cytokine production. J Pharmacol Exp Ther 1998; 284(1) :189-195.

AntitumoralEffects ofLipidsA, ClinicalStudies

131

38. Walzer T. Dalod M. Robbins SH et al. Natural-killer cells and dendritic cells: "l'union fait la force" Blood 2005 ; 106(7) :2252-2258. 39. Magder S. Neculcea J. Neculcea V et aL Lipopolysaccharide and TNF-alpha produce very similar changes in gene expression in human endothelial cells. J Vase Res 2006 ; 43(5):447-461. 40. Christophi C. Winkworth A. Muralihdaran V er al. The treatment of malignancy by hyperthermia . Surg Onco11998; 7(1-2):83-90. 4 I. Dinarello CA. Infection. fever and exogenous and endogenous pyrogens: some concepts have changed. J Endotoxin Res 2004; 10(4):201-222. 42. Engelhardt R. Mackensen A, Galanos C et al. Biological response to intravenously administered endotoxin in patients with advanced cancer. J Bioi Response Mod 1990; 9(5) :480-491. 43. Hoption Cann SA, van Nerren JP, van Nerren C. Dr Will iam Coley and tumour regression: a place in history or in the future . Postgrad Med J 2003; 79(938) :672-680.

CHAPTER 12

Conclusion ]ean-Franftois]eannin

O

ne of the major advances in the treatment of cancer with lipid A is the possibility of chemically synthesizing, lipid A analogs that are both biologically active in cancer patients and very well tolerated. Biotechnological production of different lipid A forms is also possible but the chemical route is preferable as it avoids LPS contamination. This significant step was somewhat unexpected as in the 1980s the debate had centered on the question "Is it possible to separate the toxic activity oflipid A from its anti-tumor activity?" and I will come back to this question here. We know very little about the anti -tumor mechanisms oflipid A analogs. Their effects are indirect, in the inflammatory pathway and probably mediated by cells ofthe innate immune system and the TLR4 receptor; however they induce a specific immune response and immunization without an auto-immune reaction. The action oflipid A analogs fits into the Danger model (see Introduction). Antigen-presenting cells (APC, mainly dendritic cells) have to receive (i) activating signals from injured tumor cells, but not from normal cells and (ii) activating signals to induce co-stimulation of helper T cells. It has been shown in this book that lipid A analogs activate APC, especially dendritic cells, to produce co-stimulatory molecules either directly via TLR4 or via cytokine produced by stroma cells. To get a specific immune response without an auto-immune response, only tumor cells should be killed while normal cells should be preserved, thus lipid A analogs must kill tumor cells specifically. Lipid A analogs are not cytotoxic by themselves, the inflammatory response they generate is not specific, so they cannot be involved in the cell injury. On the other hand, lipid A analogs are able to activate innate immune cells to become selectively toxic to tumor cells. NK cells and macrophages are able to discriminate between tumor and normal cells, killing tumor cells and leaving normal cells unharmed. Therefore the most likely hypothesis is that lipid A analogs activate innate immune cells (macrophages or NK cells) to kill tumor cells, then fragments oftumor cellswith antigens are phagocytosed. processed and presented to helper T cells by APC (dendritic cells) which at the same time are stimulated (directly or not) by lipid A analogs to co-stimulate helper T cells. Lipid A analogs induce active mechanisms and also inhibit suppression by inducing the switch from Tumor associated macrophages-2 (TAM-2, suppressor macrophages) to TAM-1 (anti-tumor macrophages) by inhibiting regulatory T cells. So coming back to the question "Is it possible to separate toxic activity from antitumor activiryr" we only have a partial answer. It is possible to induce anti -tumor activity without toxicity. Toxicity is due to huge inflammatory reactions of the innate immune system in which pro-inflammatory cascades are activated and the various pro-inflammatory mediators appear in an amplification loop. Anti-tumor activity comes with an inflammatory response that is local and moderate and probably supportS the anti-tumor response by creating a favorable micro-environment. Although toxicity and anti-tumor activity start by following the same inflammatory response mechanism, they differ in their intensity then diverge, toxicity going on to produce a huge and systemic inflammatory reaction , and the anti-tumor response to tumor cell-mediated toxicity, antigen presentation and a specific 'Corresponding Author: lean-Francois Jeannin-EPHE, Dijon, F-21 000, France; Inserm U866 , Dijon, F-21000, France. Email: [email protected]

LipidA in Cancer Therapy, edited by jean-Francoisjeannin. ©2009 Landes Bioscience and Springer Science+Business Media.

134

Lipid A in Cancer Therapy

immune response. Structure-function analysis of lipid A signalingindicates that the length and number ofacylsidechainsarecriticalforsignalingviaTLR4. Decreasingthe numberor lengthofthe attached fatty acidsor alteringthe overallchargeof lipidA can reducethe magnitude of the signal,' Preclinicalstudiesshowthat the most effective lipidA analogsin termsof antitumor activityin vivo do not havethe optimum structure in terms ofthe length and number of acylsidechainsto induce signaling. Is the signalingthe reasonfor differences between toxicityand the anti-tumor response? Do the lipid analogsbind TLR4 in a differentwaythan natural lipid A forms?Or is the difference upstream of the signaling, in the transport or in the transfer from LBP/CD14 to TLR4? The danger signal model is a good framework in which to assess non specific activeimmunotherapywith compounds isolatedfrom bacteriaor viruses. Until recentlythe mechanismsofaction ofthis class ofsubstanceremainedunknown.IncreasedunderstandingofTLR signalinghasallowed the molecularbasisof suchcompoundsto beelucidated. Themost effective treatment for superficial bladder cancer, and the only reallyeffective immunotherapyto date,ismediated byboth TLR2 and TLR4. 2The biologicalresponse modifierOK-342 isdependent on TLR-4 for itsanti-tumor effect.3 Double-strandedRNAI activates TLR3 on dendritic cells to release Type 1IFN that inducestumor cellapoptosisand NK-mediatedtumor cyroroxiciry," Unmethylateddeoxycytidyl-deoxyguanosine (CpG) dinucleotides are recognizedby the Toll-like receptor 9 (TLR9) and are effective in treating leukemia.' Bytargeting TLR these unrelated compounds may conform to the Danger model, inducing the innate immune systemthen activatingthe tumor adaptiveimmune response. Based on preclinical results, it is not possibleto define the type of cancer to treat. The lipid A analogs tested were effective in treating severaltypes of cancer (carcinoma, sarcoma, erc.) in different animal species (mice,rats, guinea pigs, rabbits, etc.). However there are very few results ofpreclinicalstudieswith hematological cancers. Does this type of cancernot respond to lipid A analogsor doesthe lackofresultssimplyreflectthe lackofexperiments? However, wecan conclude from preclinical results that it is necessary to repeat the injections of lipid A analogsand that the rhythm ofinjections is as important as the route, with the intra venous route being more effective than the intra peritoneal, intra muscullaror sub cutaneous routes. An important issueis addressedin this book-the tolerance induced bylipid A analogs. Is this tolerance necessary or not in the anti-tumor effectiveness of lipid A analogs? Toleranceis induced and maintained in vivowhen lipid A is administered with a specific rhythm ofinjecrions. In rats this rhythm is the best wayofproducing an anti-tumor effect,but it does not followthat tolerance is necessaryfor the anti-tumor effect.The question remainsopen. The rationale for usinglipid A analogsin clinicaltrials is to activateTLR4 on innate immune cells to induce adangersignal(tumor celldeath,antigenpresentationand APC activationto express co-stimulatorymolecules). Somecancers stillhaveno effective treatment,othersarecuredonlywhen they are treated earlyand in the absence of metastases, so lipid A analogs represent a significant sourceofhope for cancerpatients.

References

1. Miller SI, Ernst RK, and Bader MW. LPS, TLR4 and infectious disease diversity. Nat Rev Microbiol 2005; 3:36-46. 2. Tsuji S, Matsumoto M, Takeuchi 0 et al. Maturation of human dendritic cells by cell wall skeleton of Mycobacterium bovis bacillus Calmene-Guerin: involvement of toll-like receptors. Infect Immun 2000; 68:6883-6890. 3. Okamoto M, Oshikawa T, Tano T et al. Involvement of Toll-like receptor 4 signaling in interferon -gamma production and antitumor effect by streptococcal agent OK-432. j Narl Cancer Insr 2003; 95:316-326. 4. Whitmore MM, DeVeerM], Edling A et al. Synergistic activation of innate immunity by double-stranded RNA and CpG DNA promotes enhanced antitumor activity. Cancer Res 2004; 64: 5850-5860. 5. jahrsdorfer B. Muhlenhoff L, BlackwellSE et al. B-celilymphomas differ in their responsivenessto CpG oligodeoxynucleotides. Clin Cancer Res 2005 ; 11: 1490-1499 .

INDEX A Activeprincipleof endotoxin 17 Acylchain fluidity 33, 42 Adjuvant 2,3,71,83,85,92, 101, 103, 111-120,125,126,129 Aggregate structure 25,27-30,32,33, 42-45 Albumin 41,42,69,71 Animal model 3, 71,72, 76,84,86, 101, 105, 108, 125, 126, 128 Antagonist 12, 18,31,32,41,43,69,76, 83,91 Antitumoral activity 102, 104, 106

B Bactericidal/Permeability-Increasing Protein 40,42,45,70 B cell 2, 53, 56 Biosynthetic precursor to lipid a 6, 7 Blood cytokine 126

C C3H/He] 54, 59 Cancer 83-86,92,93,101-103,111,112, 113,115-120,125-129,133,134 Cancer immunotherapy 119, 125 Cancer vaccine 83,85,92,111-113, 115-120 Cathelicidin 40, 45, 46 CD14 2,34,39-41,43,45,46,53,55,59, 70,87,88,91,92, 134 Chemicalsynthesis 1,6,17,18,69 Clinical toxicity 127 Colorectal cancer 112, 113, 117, 126 Critical micellarconcentration 25, 26 Cytokine 28, 30, 32, 34,42, 43, 45, 46, 56, 59,60,62,63,70,71,73-76,81-85, 88-90,92, 101, 102, 105, 107, 111, 119, 125-129, 133

D

D-Glucosamine2-amino-2-deoxy-d-glucose (GleN) 5-11, 14-16,31 Dendritic cell 53,56,62,63, 71,84, 107, 111,120, 126, 13~ 134

E Endotoxicprinciple 5, 25, 33, 39 Endotoxin 1, 5, 9, 17,25,27,30,32-34, 36,39,41,42,46, 53, 56, 59, 82-85, 88-90,92, 126, 128, 129

F Freelipid A 8,9,27,28,30,41

G Gel to liquid crystallinephasetransition 44 GleN3N (3-dideoxy-d-glucose) 8, 11, 14-16 Glycolipid 5, 8, 17

H Heat shock protein 33, 54, 75 Hydrophilic backbone 5, 6, 8, 15, 16

K Kdo (3-deoxy-d-manno-octulosonicacid) 9, 14,18,27,41,44

L Lactoferrin 42, 44, 70 Leptospira interrogan 15, 54 LipidA 9,12-17,25-27,32,33,39,41,42, 44,46,53,54,59,65,69,70,72,73, 75,81-83,85,86,91,101,111,133 Lipid bilayer 5, 117 LipidX 7

Lipid A in Cancer Therapy

136

Lipopolysaccharide (LPS) 1,2, 5-9, 12, 14, 18, 19,25,27-34,39-46,53-56,59,60, 62-64,69,70,72-75,81 -92, 101, 102, 104, Ill, 119, 125-129, 133 Lipopolysaccharide-bindingprotein (LBP) 2,31,33,34,36,39-41,43-46,53,59, 70,91, 134 Lipoprotein 40, 42-44, 69

SDZ MRL 953 72, 83, 85, 86, 92, 103, 127, 128 Secondaryacylgroup 6, 8, 9, 11, 12, 16, 17 Sepsis 39,41,69,81-84,86 Supramolecular structure 3, 25, 28, 32, 43, 45

M

T

S

Pharmacokinetic 86, 128 Polar substituent 13-15, 19 Prirnaryacylgroup 6,8,9,11,17 Proteasome 63, 90-92

Taxol 54 Tcell 2,133 TlRAP 61-63 TLR2 54,62,87, 88,91,92, 134 TLR4 2,3,33,34,40,41,53,-56, 59-65, 69,70,72,87-89,91 ,92,111,113, 133,134 Tolerance 3,62,63,69,72-74,81-93, 105-107, 117, 126, 128, 129, 134 Toll-like receptor (TLR) 2, 33, 40, 53, 59-65,87,91, 134 Toll receptor 2, 54 TRAM 61-63 TRIF 60-64 Tumor 1,2,25,39,54,69-76,82-86,92, 101-108,111-120,125-129,133,134 Tumor associated antigens (TAA) 111-113, 115-117, 119, 120 Tumor growth 71,73-75,84, 102-107, 118, 126, 127, 129 Tumor inununity 117, 125

R

u

RP105 53,56

Unsaturatedacylgroup 12, 18

Massspectrometry(MS) 6, 16, 19 Molecularconformation 12,25,27,29, 31,32 Molecularmodelling 25, 31 Monophosphoryllipid A (MPL) 28,29, 32,69-72,82,83,85,91,92, 103, 111-114,117-120,126,127 MyD88 59-65,87,88,91,92

N NFKB 72,87 Nuclearmagneticresonance spectroscopy (NMR) 6, 16, 17, 19

p