Cancer Risk Assessment: Chemical Carcinogenesis, Hazard Evaluation, and Risk Quantification

CANCER RISK ASSESSMENT CANCER RISK ASSESSMENT Chemical Carcinogenesis, Hazard Evaluation, and Risk Quantification Ed...

Author: Ching-Hung Hsu | Todd Stedeford

58 downloads 1306 Views 7MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

CANCER RISK ASSESSMENT

CANCER RISK ASSESSMENT Chemical Carcinogenesis, Hazard Evaluation, and Risk Quantification

Edited by CHING-HUNG HSU TODD STEDEFORD

A JOHN WILEY & SONS, INC., PUBLICATION

About the cover: The cover structures are chemicals classified as known human carcinogens in the U.S. National Toxicology Program’s Annual Report on Carcinogens (http://www.ntp.niehs.nih.gov/). The center structure is cyclosporin A (CASRN 59865-13-3). The outer structures going clockwise are benzidine (92-87-5), vinyl chloride (CASRN 75-01-4), tamoxifen (CASRN 10540-29-1), cyclophosphamide (CASRN 50-18-0), benzene (CASRN 71-43-2), and azathioprine (CASRN 446-86-6). These structures were prepared using ACD/ChemSketch (ACD/Labs Release: 11; Product Version: 11.01; http://www.acdlabs.com). Copyright © 2010 John Wiley & Sons, Inc. All rights reserved. Published by John Wiley & Sons, Inc., Hoboken, New Jersey Published simultaneously in Canada No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data: Cancer risk assessment : chemical carcinogenesis, hazard evaluation, and risk quantification / [edited by] Ching-Hung Hsu, Todd Stedeford. p. ; cm. Includes bibliographical references and index. Summary : “With a weight-of-the-evidence approach, cancer risk assessment indentifies hazards, determines dose-response relationships, and assesses exposure to characterize the true risk. This book focuses on the quantitative methods for conducting chemical cancer risk assessments for solvents, metals, mixtures, and nanoparticles. It links these to the basic toxicology and biology of cancer, along with the impacts on regulatory guidelines and standards. By providing insightful perspective, Cancer Risk Assessment helps researchers develop a discriminate eye when it comes to interpreting data accurately and separating relevant information from erroneous”—Provided by publisher. ISBN 978-0-470-23822-6 (cloth) 1. Carcinogens. 2. Health risk assessment. I. Hsu, Ching-Hung. II. Stedeford, Todd. [DNLM: 1. Neoplasms–chemically induced. 2. Risk Assessment–methods. 3. Carcinogens– toxicity. 4. Environmental Exposure. 5. Mutagenicity Tests. QZ 202 C21556 2010] RC268.6.C357 2010 616.99′4071—dc22 2009049268 Printed in the United States of America 10

9

8

7

6

5

4

3

2

1

CONTENTS

PREFACE

xvii

CONTRIBUTORS

xix

ABBREVIATIONS AND ACRONYMS

xxiii

PART I

CANCER RISK ASSESSMENT, SCIENCE POLICY, AND REGULATORY FRAMEWORKS CHAPTER 1

1

CANCER RISK ASSESSMENT

3

Elizabeth L. Anderson, Kimberly Lowe, and Paul Turnham 1.1.

1.2.

1.3.

1.4.

Cancer Risk Assessment 3 1.1.1. Cancer in the United States 3 1.1.2. Historical Perspectives of Cancer Risk Assessment 4 1.1.3. The Defining Steps in Cancer Risk Assessment 9 1.1.4. The Mode of Action (MOA) 11 1.1.5. Accounting for Scientific Uncertainty 11 The Weight of Evidence (WOE) for Determining Carcinogenicity 12 1.2.1. Epidemiologic Studies 12 1.2.2. Animal Models 14 1.2.3. Weight of the Evidence Descriptors 15 Risk Assessment in the 21st Century 16 1.3.1. Using the Advances in Molecular and Computational Biology 1.3.2. Genetic Susceptibility 17 Applications in Risk Management 17 1.4.1. Translating Risk Assessment into Risk Management in the United States 17 1.4.2. International Risk Management 18 1.4.3. Risk–Benefit Analysis 19 1.4.4. Risk Acceptance and Risk Communication 20 References 21

CHAPTER 2

SCIENCE POLICY AND CANCER RISK ASSESSMENT

16

23

Gary E. Marchant 2.1. 2.2.

Introduction 23 Use of Risk Assessment in Regulatory Decision-Making

24

v

vi 2.3. 2.4. 2.5. 2.6. 2.7.

CONTENTS

Role Of Risk Assessment Guidelines 25 Data Quality Requirements 28 Types of Data Used in Risk Assessment 30 Application of “Conservative” Assumptions and Precaution Conclusion 34 References 34

33

HAZARD AND RISK ASSESSMENT OF CHEMICAL CARCINOGENICITY WITHIN A REGULATORY CONTEXT

CHAPTER 3

37

Henk Tennekes, Virginia A. Gretton, and Todd Stedeford 3.1. 3.2. 3.3.

3.4.

3.5.

Overview 37 Risk Assessment 37 3.2.1. Principles of Risk Assessment and Management 38 Regulatory Schemes for Industrial Chemicals and Biocides 42 3.3.1. The U.S. Toxic Substances Control Act (TSCA) 42 3.3.2. The EU Registration, Evaluation, Authorization, and Restriction of Chemicals (REACH) 44 3.3.3. Voluntary Initiatives for Evaluating Industrial Chemicals 45 3.3.4. The U.S. Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) 48 3.3.5. The EU Biocidal Products Directive (BPD) 49 Scientific Aspects of Carcinogenic Risk Assessment 50 3.4.1. Dose–Response Relationships in Carcinogenesis and Mechanisms of Carcinogenic Action 50 3.4.2. Mathematical Model for Carcinogenic Risk Assessment 60 Conclusions 61 References 62

USE OF CANCER RISK ASSESSMENTS IN DETERMINATION OF REGULATORY STANDARDS

CHAPTER 4

Robert A. Howd and Anna M. Fan 4.1. 4.2.

4.3.

4.4.

4.5.

4.6.

4.7.

Introduction 66 Air Standards 70 4.2.1. Scientific Issues 70 4.2.2. Regulatory Considerations 72 Water Standards 73 4.3.1. Scientific Issues 73 4.3.2. Regulatory Considerations 75 Food Standards, Pesticide Tolerances, Additives, and Impurities 4.4.1. Scientific Issues 76 4.4.2. Regulatory Considerations 77 Soil Standards 81 4.5.1. Scientific Issues 81 4.5.2. Regulatory Considerations 81 Consumer Product Standards 82 4.6.1. Scientific Issues 82 4.6.2. Regulatory Considerations 83 Recent Developments and Future Directions 84 References 87

76

66

CONTENTS

vii

PART II

CANCER BIOLOGY AND TOXICOLOGY CHAPTER 5

97

THE INTERPLAY OF CANCER AND BIOLOGY

99

James W. Holder 5.1.

5.2. 5.3.

5.4.

Historical Account of Some Important Events in Understanding Cancer 99 5.1.1. Early Cancer Biology History 99 5.1.2. Near-Recent Cancer Biology History 101 Recent Foundations of Biological Mechanisms of Cancer 103 Cell Biology of Cancer 105 5.3.1. In Vitro Systems 105 5.3.2. Programmed Cell Removal 107 5.3.3. Facilitation of Supporting Cells and Cell-to-Cell Communication 5.3.4. Clonal Aspects of Carcinogenesis 116 5.3.5. Biology of Inflammation and Cancer 124 5.3.6. Stem Cell Biology and Cancer 130 5.3.7. Specific Biological Growth and Growth Control Gene Sets and Their Pathways 139 5.3.8. Epigenetic Biology and Nuclear Traffic 142 5.3.9. Biological Initiation of Chemical Carcinogenesis 147 Some Final Thoughts on Biology and Cancer 152 References 155

CHEMICAL CARCINOGENESIS: A BRIEF HISTORY OF ITS CONCEPTS WITH A FOCUS ON POLYCYCLIC AROMATIC HYDROCARBONS

113

CHAPTER 6

168

Stephen Nesnow 6.1. 6.2. 6.3.

A Brief History of Chemical Carcinogenesis 168 James A. and Elizabeth C. Miller and Their Theory of Metabolic Activation 169 6.2.1. Metabolic Activation of PAH and Tumorigenesis 173 The Concepts of Initiation, Promotion, and Progression: The Origin of Multistage Carcinogenesis 182 References 185

CHAPTER 7

HORMESIS AND CANCER RISKS: ISSUES AND RESOLUTION

191

Paolo F. Ricci and Edward J. Calabrese 7.1. 7.2. 7.3. 7.4.

Introduction 191 Evidence for Regulatory Cancer Risk Assessment 194 Hormesis and Cancer Risk Assessment: Models 198 7.3.1. Answers to Our Question 201 Conclusions 203 References 204

THRESHOLDS FOR GENOTOXIC CARCINOGENS: EVIDENCE FROM MECHANISM-BASED CARCINOGENICITY STUDIES

CHAPTER 8

Shoji Fukushima, Min Wei, Anna Kakehashi, and Hideki Wanibuchi

207

viii 8.1. 8.2. 8.3. 8.4. 8.5. 8.6. 8.7.

CONTENTS

Overview 207 Introduction 208 Low-Dose Carcinogenicity of 2-Amino-3,8-Dimethylimidazo[4,5-f ]-Quinoxaline (MEIQX) in the Rat Liver 209 Low-Dose Hepatocarcinogenicity of N-Nitroso Compounds 215 Low-Dose Carcinogenicity of 2-Amino-1-methyl-6-phenylimidazo[5,6-b]pyridine (PHIP) in the Rat Colon 215 Low-Dose Carcinogenicity of Potassium Bromate, KBrO3 in the Rat Kidney 216 Conclusion 219 References 220

PART III

GENETIC TOXICOLOGY, TESTING GUIDELINES AND REGULATIONS, AND NOVEL ASSAYS

223

CHAPTER 9 DEVELOPMENT OF GENETIC TOXICOLOGY TESTING AND ITS INCORPORATION INTO REGULATORY HEALTH EFFECTS TEST REQUIREMENTS

225

Errol Zeiger 9.1. 9.2. 9.3. 9.4. 9.5. 9.6. 9.7.

Introduction 225 Definitions and Usage 226 The Historical Development of Genetic Toxicity Testing Types of Available Tests 228 Testing Approaches 229 Where Are We Now? 232 Summary 235 References 235

227

GENETIC TOXICOLOGY TESTING GUIDELINES AND REGULATIONS

CHAPTER 10

238

Lutz Müller and Hans-Jörg Martus 10.1. 10.2. 10.3.

10.4. 10.5. 10.6.

10.7.

Historical Overview of Genotoxicity Testing Guidelines 238 Organization for Economic cooperation and Development (OECD) Guidelines for Genotoxicity 243 International Conference of Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) Guidelines for Pharmaceuticals 243 International Workshop on Genotoxicity Tests (IWGT) 248 The International Program on Chemical Safety (IPCS) Under the Auspices of the World Health Organization (WHO) 249 In Vitro Testing 250 10.6.1. In Vitro Tests 250 10.6.2. Evaluation of In Vitro Testing Results 250 10.6.3. Follow-Up to In Vitro Testing 250 In Vivo Testing 251 10.7.1. Follow-Up to In Vivo Testing 251 10.7.2. Strategy for Germ Cell Testing 251

CONTENTS

ix

10.8.

European Union Guideline for Testing of Chemicals Under the Registration, Evaluation, Authorization and Restriction of Chemical (REACH) 252 10.9. Specialty Guidelines for Genotoxicity: Genotoxic Impurities in Pharmaceuticals 256 10.10. The Quintessence for Regulatory Assessment: In Vivo Testing for Risk Assessment 258 10.10.1. Choice of In Vivo Test 261 10.10.2. Evaluation of In Vivo Results 263 10.11. Summary and Outlook 264 References 265

CHAPTER 11

IN VITRO GENOTOX ASSAYS

272

David Kirkland and David Gatehouse 11.1. 11.2. 11.3. 11.4. 11.5. 11.6. 11.7. 11.8. 11.9.

Introduction 272 In Vitro Metabolic Activation 273 In Vitro Tests for Gene Mutation in Bacteria 273 In Vitro Tests for Gene Mutation in Mammalian Cells 276 In Vitro Tests for Chromosome Damage in Mammalian Cells 279 The In Vitro Micronucleus Test 280 In Vitro Test for Unscheduled DNA Synthesis in Rat Hepatocytes In Vitro Comet Assay 284 Strengths and Limitations 285 References 286

CHAPTER 12

IN VIVO GENOTOXICITY ASSAYS

283

289

Véronique Thybaud 12.1.

12.2. 12.3.

12.4.

12.5.

Introduction 289 12.1.1. Endpoints Used for In Vivo Genetic Toxicology Assays 289 12.1.2. Contribution of In Vivo Genetic Toxicology Assays to Risk Assessment 291 Parameters and Criteria for Valid In Vivo Genotoxicity Assays and Implications for Experimental Design 292 In Vivo Genotoxicity Assays Required in the Standard Battery of Tests 303 12.3.1. Mammalian Erythrocyte Micronucleus Test 304 12.3.2. Bone Marrow Chromosome Aberration Test 308 In Vivo Genotoxicity Assays Used Mainly as Complementary or Follow-Up Tests 310 12.4.1. The Comet Assay 311 12.4.2. DNA Adducts 314 12.4.3. Unscheduled DNA Synthesis Test in Liver Cells 324 12.4.4. Sister-Chromatid Exchange Assay 326 12.4.5. Gene Mutation Assays 328 Conclusion and Perspectives 344 References 345

x

CONTENTS

PART IV

ASSESSING THE HUMAN RELEVANCE OF CHEMICAL-INDUCED TUMORS

361

FRAMEWORK ANALYSIS FOR DETERMINING MODE OF ACTION AND HUMAN RELEVANCE

CHAPTER 13

363

R. Julian Preston 13.1. 13.2.

13.3. 13.4.

Introduction 363 Framework Analysis: Mode of Action and Key Events 364 13.2.1. Definitions 364 13.2.2. An Overview of the Framework for Analyzing Mode of Action 365 13.2.3. Framework for Assessing Human Relevance of Animal MOA 365 13.2.4. Establishing and Applying Key Events in Support of MOA 367 Framework Analysis: Human Relevance 372 Future Directions 375 References 376

CHAPTER 14

EXPERIMENTAL ANIMAL STUDIES AND CARCINOGENICITY

378

Mary Elizabeth (Bette) Meek 14.1. 14.2.

14.3.

14.4. 14.5.

Introduction 378 Current Status of Hazard Testing for Cancer for Regulatory Risk Assessment 14.2.1. The Combined Chronic/Cancer Bioassay in Rats and Mice 379 14.2.2. Perinatal Carcinogenicity Studies 382 14.2.3. Limited In Vivo Studies 382 Application in Risk Assessment 383 14.3.1. Hazard Identification 383 14.3.2. Hazard Characterization 386 14.3.3. Dose–Response Analyses; Selection of Points of Departure 388 Evolution of Testing Strategies 390 Discussion: Closing the GAP Between Hazard Testing and Risk Assessment References 393

CHAPTER 15

CANCER EPIDEMIOLOGY

15.3.

15.4.

15.5. 15.6.

Introduction 397 Considerations for the Epidemiologic Study of Cancer 15.2.1. Demographics 398 15.2.2. Other Variables 400 Epidemiologic Study Methods 403 15.3.1. Types of Epidemiologic Studies 403 15.3.2. Meta-analysis and Case Reports 407 Evaluation of Studies and Their Results 407 15.4.1. Quality of Studies 407 15.4.2. Determining Causal Association 408 Substances Causally Associated with Cancer 411 Future for Cancer Epidemiology 414

391

397

Herman J. Gibb and Jessie P. Buckley 15.1. 15.2.

379

398

CONTENTS

15.6.1. The Effect of Exposure at Different Ages 15.6.2. Molecular Epidemiology 415 15.6.3. Infectious Agents 415 References 416 CHAPTER 16

xi

414

RODENT HEPATOCARCINOGENESIS

419

James E. Klaunig 16.1.

16.2.

16.3. 16.4.

Introduction 419 16.1.1. Initiation 420 16.1.2. Promotion 422 16.1.3. Progression 422 Mechanisms of Action of Hepatic Carcinogens 16.2.1. Genotoxic Agents 424 16.2.2. Nongenotoxic Mechanisms of Action Human Relevance Framework 434 Summary 435 References 435

423 425

MODE OF ACTION ANALYSIS AND HUMAN RELEVANCE OF LIVER TUMORS INDUCED BY PPARα ACTIVATION

CHAPTER 17

439

J. Christopher Corton 17.1. 17.2. 17.3.

17.4.

Overview 439 Introduction 440 Mode of Action Analysis in the EPA Risk Assessment Framework 441 17.3.1. Summary of the Mode of Action and Human Relevance of Liver Tumors Induced by PPARα Activation 441 17.3.2. Detailed Evaluation of the Rodent Mode of Action 443 Relevance of PPARα Activator-Induced Rodent Liver Tumor Response to Humans 467 References 467

ALPHA2U-GLOBULIN NEPHROPATHY AND CHRONIC PROGRESSIVE NEPHROPATHY AS MODES OF ACTION FOR RENAL TUBULE TUMOR INDUCTION IN RATS, AND THEIR POSSIBLE INTERACTION

CHAPTER 18

482

Edward A. Lock and Gordon C. Hard 18.1. 18.2. 18.3. 18.4. 18.5. 18.6.

Introduction 482 Chemicals that Increase the Incidence of Renal Tubule Tumors in Male Rats by an α2u-Globulin Mode of Action 483 Chemicals Increasing the Incidence of Renal Tumors Through Exacerbation of Spontaneous Chronic Progressive Nephropathy (CPN) 489 Chemicals Increasing RTT Incidence Through a Mode of Action Involving Exacerbation of CPN 491 Examples Where the α2u-Globulin and Exacerbated CPN Modes of Action May Be Acting in Concert 493 Relevance of Rat α2u-Globulin Nephropathy and CPN to Humans 495 References 496

xii

CONTENTS

CHAPTER 19

URINARY TRACT CALCULI AND BLADDER TUMORS

501

Samuel M. Cohen, Lora L. Arnold, and Shugo Suzuki 19.1. 19.2. 19.3. 19.4. 19.5. 19.6. 19.7. 19.8.

Introduction 501 Direct and Indirect Formation of Urinary Solids 502 Urinary Factors Influencing the Formation of Urinary Solids Collection of Urine for Detection of Urinary Solids 507 Interspecies Comparison of Urine Composition 508 Urinary Solid Carcinogenesis in Rodents 508 Epidemiology 510 Risk Assessment 511 References 512

505

PART V

METHODS FOR INFORMING CANCER RISK QUANTIFICATION CHAPTER 20 (Q)SAR ANALYSIS OF GENOTOXIC AND NONGENOTOXIC CARCINOGENS: A STATE-OF-THE-ART OVERVIEW

515

517

Yin-tak Woo and David Y. Lai 20.1. 20.2.

20.3.

20.4.

20.5.

Introduction 517 Overview of (Q)SAR Analysis and Modeling 518 20.2.1. Types of (Q)SAR 518 20.2.2. Criteria for Assessing Validity and Scientific Soundness of (Q)SAR 519 20.2.3. Difficulties of (Q)SAR Modeling/Prediction of Chemical Carcinogens 520 20.2.4. Importance of Mechanistic Understanding 520 Mechanism-Based SAR Analysis of Chemical Carcinogens, Fibers, and Particles/Nanoparticles 521 20.3.1. Basic Principles 521 20.3.2. SAR of Genotoxic Carcinogens 522 20.3.3. SAR of Nongenotoxic Carcinogens 528 20.3.4. SAR of Fibers, Particles, and Nanomaterials 534 Uses of (Q)SAR in Cancer Hazard/Risk Assessment and Brief Overview of Predictive Systems/Softwares 544 20.4.1. Evolving Uses of (Q)SAR in Cancer Hazard Identification and Risk Assessment 544 20.4.2. Brief Overview of (Q)SAR Systems/Softwares for Predicting Carcinogenic Potential of Chemicals 546 Future Perspectives 548 References 550

PHYSIOLOGICALLY BASED PHARMACOKINETIC (PBPK) MODELS IN CANCER RISK ASSESSMENT

CHAPTER 21

Mathieu Valcke and Kannan Krishnan 21.1. 21.2. 21.3.

Introduction 557 PBPK Modeling: Characteristics and Approaches 558 PBPK Models in Cancer Risk Assessment 563 21.3.1. High-Dose to Low-Dose and Interspecies Extrapolation

565

557

CONTENTS

21.4.

21.5.

21.3.2. Intraspecies Extrapolation 568 21.3.3. Route-to-Route Extrapolation 571 21.3.4. Extrapolation from Individual Carcinogens to Mixtures PBPK Models in Cancer Risk Assessment: Case Studies 574 21.4.1. Dichloromethane (Methylene Chloride) 574 21.4.2. Vinyl Chloride 575 21.4.3. Chloroform 576 21.4.4. Dioxane 577 21.4.5. Trichloroethylene 578 21.4.6. Volatile Organic Chemical Mixtures 578 Concluding Remarks 579 References 580

CHAPTER 22

xiii

571

GENOMICS AND ITS ROLE IN CANCER RISK ASSESSMENT

586

Banalata Sen, Douglas C. Wolf, and Vicki Dellarco 22.1. 22.2. 22.3. 22.4.

22.5. 22.6.

Introduction 586 “-Omics” Technologies 588 Genomics and the New Risk Assessment Paradigm Case Studies 590 22.4.1. Perfluorooctanoic acid (PFOA) 590 22.4.2. Formaldehyde and Glutaraldehyde 591 22.4.3. Conazoles 592 Use of Genomics in Predictive Toxicology 593 Conclusions 594 References 595

CHAPTER 23

589

COMPUTATIONAL TOXICOLOGY IN CANCER RISK ASSESSMENT

597

Jerry N. Blancato 23.1. 23.2. 23.3.

23.4. 23.5.

Introduction 597 Risk Assessment: Historical Perspective 598 Enhancements in Quantitative Risk Assessment 599 23.3.1. Physiologically Based Pharmacokinetic (PBPK) Modeling 23.3.2. Pharmacokinetic Variability and Uncertainty 601 23.3.3. Pharmacodynamic and Dose–Response Modeling 602 Computational Toxicology and Future Risk Assessments 602 23.4.1. 21st-Century Toxicology 603 Conclusion 609 References 610

599

PART VI

GENERAL APPROACHES FOR QUANTIFYING CANCER RISKS CHAPTER 24

LINEAR LOW-DOSE EXTRAPOLATIONS

Michael Dourson and Lynne Haber 24.1. 24.2.

Introduction 615 Historical 616

613 615

xiv 24.3. 24.4.

CONTENTS

Issues Related to Extrapolation from Experimental Data Conclusion 631 References 633

625

QUANTITATIVE CANCER RISK ASSESSMENT OF NONGENOTOXIC CARCINOGENS

CHAPTER 25

636

Rafael Meza, Jihyoun Jeon, and Suresh H. Moolgavkar 25.1.

25.2.

25.3.

Introduction 636 25.1.1. The Hazard or Incidence Function 637 25.1.2. Two-Stage Clonal Expansion (TSCE) Model 637 25.1.3. Multistage Clonal Expansion (MSCE) Model 640 25.1.4. Modeling Dose–Response in the TSCE and MSCE Models 25.1.5. Analysis of Epidemiological Data 643 25.1.6. Analysis of Premalignant Lesions Using the TSCE Model Some Examples and Applications 649 25.2.1. Smoking, Radon, and Arsenic Exposures and Lung Cancer 25.2.2. Folate and Colorectal Cancer 653 25.2.3. Enzyme-Altered Foci in the Rat Liver 653 Concluding Remarks 655 References 655

CHAPTER 26

642 644 649

NONLINEAR LOW-DOSE EXTRAPOLATIONS

659

Ari S. Lewis and Barbara D. Beck 26.1. 26.2.

26.3. 26.4. 26.5.

26.6.

26.7.

Introduction 659 Mechanistic Aspects of Nonlinear Carcinogenesis 661 26.2.1. Pre-DNA Damage Mechanisms 661 26.2.2. Post-DNA Damage Mechanisms 662 26.2.3. Hormesis 663 DNA-Reactive Carcinogens and Nonlinearity 664 Nonmutagenic Carcinogens and Nonlinearity 666 Cancer Risk Assessment 668 26.5.1. Basis for the Linearity Assumption 668 26.5.2. EPA Cancer Risk Assessment and Low-Dose Extrapolation 669 26.5.3. Low-Dose Extrapolation Outside the United States 670 Nonlinearity Principles into Practice 670 26.6.1. Using an RfD or MOE Approach 671 26.6.2. Other Nonlinear Cancer Evaluations: Captan and Chloroform 673 26.6.3. BBDR Modeling 674 26.6.4. Harmonization of Cancer and Noncancer Risks 675 Summary and Conclusion 676 References 677

CANCER RISK ASSESSMENT: MORE UNCERTAIN THAN WE THOUGHT

CHAPTER 27

Edmund A. C. Crouch 27.1. 27.2. 27.3.

Introduction 681 Summary of Previous Analyses 681 Selection of Carcinogenicity Measure—The CD10

684

681

CONTENTS

27.4. 27.5. 27.6. 27.7. 27.8.

The Variation of CD10 Within a Species 684 Extrapolation of the Median CD10 Between Species Extrapolation of the IntraSpecies Variation in CD10 Conclusions 693 Appendix 695 27.8.1. Obtaining the CD10 Estimates 695 27.8.2. Median and Geometric Standard Deviation 27.8.3. Testing Hypotheses about ln(GSD) 696 References 697

xv

686 692

696

COMBINING NEOPLASMS FOR EVALUATION OF RODENT CARCINOGENESIS STUDIES

CHAPTER 28

699

Amy E. Brix, Jerry F. Hardisty, and Ernest E. McConnell 28.1. 28.2. 28.3. 28.4.

28.5.

Introduction 699 Rationale for Combining Neoplasms 701 Usefulness of Differentiating Benign from Malignant Neoplasms and of Subclassifying Neoplasms 702 Criteria for Combining Neoplasms 704 28.4.1. Combinations According to Organ and Tissue 704 28.4.2. Combinations by Site 704 28.4.3. Combining Neoplasms of a Common Cell Type in Different Tissues Summary 711 References 711

CANCER RISK BASED ON AN INDIVIDUAL TUMOR TYPE OR SUMMING OF TUMORS

710

CHAPTER 29

716

Andrew G. Salmon and Lindsey A. Roth 29.1. 29.2. 29.3.

29.4.

29.5.

Introduction 716 Summing of Tumors of Related Types 717 Summing of Unrelated Tumor Types 718 29.3.1. Affected-Animal Count 718 29.3.2. Addition of Independent Potency Values 29.3.3. Distribution-Based Methods 719 Example: 1,3-Butadiene 721 29.4.1. Source Data 721 29.4.2. Affected-Animal Count 722 29.4.3. Distribution-Based Methods 724 Conclusions 732 References 734

718

EXPOSURE RECONSTRUCTION AND CANCER RISK ESTIMATE DERIVATION

CHAPTER 30

Shannon Gaffney, Jennifer Sahmel, Kathryn D. Devlin, and Dennis J. Paustenbach 30.1. 30.2.

Introduction 736 Exposure Reconstruction Methodology 737 30.2.1. Addressing the Goals of the Exposure Reconstruction 738 30.2.2. Organizing and Ranking Available Exposure Information 738

736

xvi

CONTENTS

30.2.3. 30.2.4.

30.3.

30.4. INDEX

Identifying Key Data Gaps in the Available Exposure Information 740 Selecting the Appropriate Methodology to Reconstruct Exposure Values 740 30.2.5. Conducting an Uncertainty Analysis of the Reconstructed Exposure Values 764 Application of Estimated Historical Exposure Values to Cancer Risk Estimates 766 30.3.1. Estimating Dose 767 30.3.2. Estimating Risk 768 30.3.3. Use of Probabilistic Analysis to Refine Dose Estimates 770 Summary 770 References 772 785

PREFACE

Cancer risk assessment is an ever-changing discipline with standard regulatory practices and defaults giving way to ever-increasing breakthroughs in scientific discovery. The scientific literature is, however, replete with reports of toxicantinduced changes, but discriminating between those reports that are irrelevant or relevant to humans and those that are compensatory versus truly adverse can be an arduous task. This book aims to inform and to provide interpretive guidance on evaluating toxicological data and understanding the relevance of such data to hazard evaluation and cancer risk estimation. The topics presented herein begin with Part I, which provides an overview of cancer risk assessment, followed by a discussion on science policy. The regulatory frameworks for industrial chemicals and biocides are presented along with the general approaches for developing standards for chemicals in air, water, food, soil, and consumer products. In Part II, basic concepts in cancer biology, chemical carcinogenesis, hormesis, and experimental evidence of thresholds for genotoxic carcinogens are provided. Thereafter, Part III describes the testing guidelines and regulations for in vitro and in vivo genotoxicity testing, and Part IV offers interpretive guidance on assessing the human relevance of chemical-induced tumors from rodent studies, along with the necessary criteria for evaluating data from epidemiological studies. Commonly observed modes of action from experimental animal studies, including PPAR-α, α2u-globulin, and so on, are then discussed. In Part V, methods for informing cancer risk quantification, including quantitative structure–activity relationships (QSAR), physiologically based pharmacokinetic (PBPK) modeling, “-omics”, and computational toxicology are discussed. Finally, Part VI addresses general approaches for quantifying cancer risks including linear and nonlinear low-dose extrapolations, summing tumors, and exposure reconstruction for cancer risk estimation. The foregoing topics are critical for keeping abreast of changes that are taking place in cancer risk assessment, as well as in the fields of toxicology and risk assessment in general. For example, with the increased emphasis on describing a chemical’s mode of action for both cancer and noncancer endpoints, an understanding of the human relevance framework is essential, as is the role of rapidly developing technologies (e.g., “-omics”) for informing mode(s) of action. Therefore, readers of this text will take away knowledge that is applicable to cancer risk assessment and more broadly to toxicology and risk assessment. The resources that formed the bases for this text include: peer-reviewed scientific articles, regulatory guidance documents, validated test guidelines, and the many years of experience conveyed throughout by the contributing authors. xvii

xviii

PREFACE

The editors are truly grateful to the contributing authors of this text, who provided their expertise on a gratis basis. If it were not for their dedication and commitment to advancing the knowledge and understanding of cancer risk assessment, the extensive coverage provided herein would not have been possible. Taipei, Taiwan Baton Rouge, Louisiana April 2010

Ching-Hung Hsu Todd Stedeford

CONTRIBUTORS

Elizabeth L. Anderson, Ph.D., FATS Group Vice President and Principal Scientist, Exponent, Inc., Alexandria, Virginia Lora L. Arnold, M.S. Assistant Professor, University of Nebraska Medical Center, Omaha, Nebraska Barbara D. Beck, Ph.D., DABT, FATS Principal, Gradient, Cambridge, Massachusetts Jerry N. Blancato, M.S., Ph.D. Acting Director, Office of Administrative and Research Support, Office of Research and Development (ORD), United States Environmental Protection Agency, Research Triangle Park, North Carolina Amy Brix, D.V.M., Ph.D., DACVP Veterinary Pathologist and Contractor for NTP QA, Experimental Pathology Laboratories, Inc., Research Triangle Park, North Carolina Jessie P. Buckley, M.P.H. Ph.D. Candidate, Gillings School of Global Public Health, University of North Carolina, Chapel Hill, North Carolina Edward J. Calabrese, Ph.D., FATS Professor of Toxicology, Department of Public Health, University of Massachusetts, Amherst, Massachusetts Samuel M. Cohen, M.D., Ph.D. Havlik–Wall Professor of Oncology, Department of Pathology and Microbiology, University of Nebraska Medical Center, Omaha, Nebraska J. Christopher Corton, Ph.D. Senior Research Biologist, Integrated Systems Toxicology Division, National Health and Environmental Effects Research Laboratory (NHEERL), Office of Research and Development (ORD), United States Environmental Protection Agency, Research Triangle Park, North Carolina Edmund A. C. Crouch, Ph.D. Senior Scientist, Cambridge Environmental, Inc., Cambridge, Massachusetts Vicki Dellarco, Ph.D. Science Advisor, Office of Pesticide Programs, United States Environmental Protection Agency, Washington, D.C. Kathryn D. Devlin, M.S. Health Scientist, ChemRisk, LLC, Boulder, Colorado Michael Dourson, Ph.D., DABT, FATS President, Toxicology Excellence for Risk Assessment (TERA), Cincinnati, Ohio xix

xx

CONTRIBUTORS

Anna M. Fan, Ph.D., DABT Chief, Pesticide and Environmental Toxicology Branch, Office of Environmental Health Hazard Assessment (OEHHA), California Environmental Protection Agency, Oakland, California Shoji Fukushima, M.D., Ph.D. Director, Japan Bioassay Research Center, Japan Industrial Safety & Health Association, Hadano, Kanagawa, Japan Shannon H. Gaffney, Ph.D., M.H.S., CIH Managing Health Scientist, ChemRisk, LLC, San Francisco, California David Gatehouse, Ph.D., FRCPath Consultant, Buntingford, Hertfordshire, United Kingdom Herman J. Gibb, Ph.D., M.P.H. President, Tetra Tech Sciences, Arlington, Virginia Virginia A. Gretton Regulatory Advisor, SafePharm Laboratories Ltd., Derbyshire, United Kingdom Lynne Haber, Ph.D., DABT Associate Director, Toxicology Excellence for Risk Assessment (TERA), Cincinnati, Ohio Gordon C. Hard, BVSc, Ph.D., DSc, DACVP, FRCPath, FRCVS, FAToxSci Independent Consultant, Tairua, New Zealand Jerry F. Hardisty, D.V.M., DACVP, IATP President and Veterinary Pathologist, Experimental Pathology Laboratories, Inc., Research Triangle Park, North Carolina James W. Holder, Ph.D. Toxicologist/Cancer, National Center for Environmental Assessment, Office of Research and Development, United States Environmental Protection Agency (Retired), Washington, D.C. Robert A. Howd, Ph.D. Chief, Water Toxicology Section, Office of Environmental Health Hazard Assessment (OEHHA), California Environmental Protection Agency, Oakland, California Jihyoun Jeon, M.S., Ph.D. Staff Scientist, Program in Biostatistics and Biomathematics, Fred Hutchinson Cancer Research Center, Seatte, Washington Anna Kakehashi, Ph.D. Lecturer, Department of Pathology, Osaka City University Medical School, Osaka, Japan David J. Kirkland, Ph.D. Consultant and Professor (University of Wales, Swansea, United Kingdom), Tadcaster, North Yorkshire, United Kingdom James E. Klaunig, Ph.D. Professor and Chair, Environmental Health, Indiana University, Bloomington, Indiana Kannan Krishnan, Ph.D., DABT, FATS Professor, Department of Environmental Health and Health at Work, University of Montréal, Montreal, Quebéc, Canada

CONTRIBUTORS

xxi

David Y. Lai, Ph.D., DABT Senior Toxicologist, Risk Assessment Division, Office of Pollution Prevention and Toxics (OPPTS), United States Environmental Protection Agency, Washington, D.C. Ari S. Lewis, M.S. Senior Scientist, Gradient, Cambridge, Massachusetts Edward A. Lock, MIBiol, Ph.D., FRCPath, FBTS, FATS Professor, School of Pharmacy and Biomolecular Sciences, Liverpool John Moores University, Liverpool, United Kingdom Kimberly Lowe, Ph.D., M.H.S. Senior Scientist, Exponent, Inc., Seattle, Washington Gary E. Marchant, Ph.D., J.D. Lincoln Professor, College of Law, Arizona State University, Tempe, Arizona Hans-Jörg Martus, Ph.D. Head, Genetic Toxicology, Preclinical Safety, Novartis Institutes for BioMedical Research, Basel, Switzerland Ernest E. McConnell, D.V.M., M.S., DACVP, DABT President, Tox Path, Inc., Raleigh, North Carolina Mary Elizabeth (Bette) Meek, M.Sc., Ph.D. Associate Director, Chemical Risk Assessment, McLaughlin Centre for Population Health Risk Assessment, University of Ottawa, Ottawa, Ontario, Canada Rafael Meza, Ph.D. Research Scientist, Division of Mathematical Modeling, University of British Columbia Centre for Disease Control, Vancouver, British Columia, Canada Suresh H. Moolgavkar, M.D., Ph.D. Corporate Vice President and Director, Center for Epidemiology, Biostatistics, and Computational Biology, Exponent, Inc., Bellevue, Washington Lutz Müller, Ph.D. Head Full Development Projects, Non-Clinical Drug Safety, F. Hoffmann-La Roche Ltd., Basel, Switzerland Stephen Nesnow, Ph.D. Senior Research Scientist, Integrated Systems Toxicology Division, National Health and Environmental Effects Research Laboratory (NHEERL), Office of Research and Development (ORD), United States Environmental Protection Agency, Research Triangle Park, North Carolina Dennis J. Paustenbach, Ph.D., CIH, DABT President and Founder, ChemRisk, LLC, San Francisco, California R. Julian Preston, Ph.D. Associate Director for Health, National Health and Environmental Effects Research Laboratory (NHEERL), Office of Research and Development (ORD), United States Environmental Protection Agency, Research Triangle Park, North Carolina Paolo F. Ricci, Ph.D., LL.M., M.P.A. Professor, Holy Names University, Oakland, California, and University of Massachusetts, Amherst, Massachusetts

xxii

CONTRIBUTORS

Lindsey A. Roth, M.A. Research Scientist II, Safer Alternatives Assessment and Biomonitoring Section, Office of Environmental Health Hazard Assessment (OEHHA), California Environmental Protection Agency, Oakland, California Jennifer Sahmel, M.P.H., CIH, CSP Supervising Health Scientist, ChemRisk, LLC, Boulder, Colorado Banalata Sen, Ph.D. Science Education and Outreach Program Manager, Environmental Health Perspectives, DHHS, NIH, NIEHS, Durham, North Carolina Andrew G. Salmon, M.A., D.Phil. Chief, Toxicology and Risk Assessment Section, Office of Environmental Health Hazard Assessment (OEHHA), California Environmental Protection Agency, Oakland, California Todd Stedeford, Ph.D., J.D., DABT Toxicology Advisor & In-House Counsel, Health, Safety & Environment, Albemarle Corporation, Baton Rouge, Louisiana Shugo Suzuki, M.D., Ph.D. Postdoctoral Research Associate, University of Nebraska Medical Center, Omaha, Nebraska Henk Tennekes, M.Sc., Ph.D., RT Consultant in Toxicology, Experimental Toxicology Services, Zutphen, The Netherlands Véronique Thybaud, Ph.D. Scientific Advisor, Disposition-Safety and Animal Research, Preclinical Safety, Sanofi Aventis, Vitry sur Seine, France Paul Turnham, B. Eng., M.S., P.E. Senior Managing Scientist, Exponent, Inc., Alexandria, Virginia Mathieu Valcke, M.Sc. Scientific Advisor, National Institute of Public Health of Québec, Montreal, Quebec, Canada Hideki Wanibuchi, M.D., Ph.D. Professor, Department of Pathology, Osaka City University Medical School, Osaka, Japan Min Wei, M.D., Ph.D. Assistant Professor, Department of Pathology, Osaka City University Medical School, Osaka, Japan Douglas C. Wolf, D.V.M., Ph.D., FIATP, ATS Assistant Laboratory Director, National Health and Environmental Effects Research Laboratory (NHEERL), Office of Research and Development (ORD), United States Environmental Protection Agency, Research Triangle Park, North Carolina Yin-tak Woo, Ph.D., DABT Senior Toxicologist, Risk Assessment Division, Office of Pollution Prevention and Toxics (OPPTS), United States Environmental Protection Agency, Washington, D.C. Errol Zeiger, Ph.D., J.D. Principal, Errol Zeiger Consulting, Chapel Hill, North Carolina


AAF 4-ABP ACF ACO ACToR ADAF Ade ADI ADME AFC AHF AhR AI AMS ANOVA AOM apo ARB ARNT ATSDR AUC B[a]A BBDR BDA BE BEEL BEIs BMD BMDL BMR B[a]P BPD BPDE BrDU

2-Acetylaminofluorene 4-Aminobiphenyl Aberrant crypts foci Acyl-CoA oxidase Aggregated chemical toxicity resource Age-dependent adjustment factor Adenine Allowable daily intake Absorption, distribution, metabolism, and excretion Altered foci cells Altered hepatic foci Aryl hydrocarbon receptor Artificial intelligence Accelerator mass spectrometry Analysis of variance Azoxymethane Apolipoprotein Air Resources Board, California EPA Ah receptor nuclear translocator U.S. Agency for Toxic Substances and Disease Registry Area under the curve Benz[a]anthracene Biologically based dose–response Bayesian data analysis Biomonitoring equivalents Biological environmental exposure limit Biological exposure indices Benchmark dose Benchmark dose lower bound Benchmark response Benzo[a]pyrene Biocidal products directive Benzo[a]pyrene diol epoxides 5-Bromo-2-deoxyuridine xxiii

xxiv


b.w. CAA CAF CAG CAM CAR CCA CCl4 CD10 CDC CDC CDK CEBS CEO CEO CEPA CERCLA cGys ChAMP CHMP CIIT CMRs CNDR CoA COPC CPDB CPN CPSC CPT-I CPUM CSA CSF CSR CTM Cx CYP 2-D 3-D 4-DAB DAG DAPI 4-DAST DB[a,l]P DC DCB

Body weight U.S. Clean Air Act Cancer-associated fibroblast Carcinogens Assessment Group Cellular adhesion molecule Constitutive androstane receptor Chromated copper arsenate Carbon tetrachloride 10% of Cancer dose Center for Disease Control U.S. Centers for Disease Control and Prevention Cyclin-dependent kinase Chemical effects in biological systems Chloroethylene oxide Cyanoethylene oxide Canadian Environmental Protection Act Comprehensive Environmental Response, Compensation and Liability Act Centigrays Chemical Assessment and Management Program Committee of Human Medicinal Products Chemical Industries Institute of Toxicology Carcinogens, mutagens, or reproductive toxicants Canadian National Dose Registry Acyl coenzyme A Contaminants of Potential Concern Carcinogenic potency database Chronic progressive nephropathy Consumer Product Safety Commission Carnitine palmitoyl transferase-I Colorado Plateau Uranium Miners Chemical Safety Assessment Cancer slope factor Chemical Safety Report Chinese tin miners Connexon Cytochrome P450 Two-dimensional Three-dimensional 4-Dimethylaminoazobenzene Directed acyclic graph 4′,6-Diamidino-2-phenylindole 4-Dimethylaminostilbene Dibenzo[a,l]pyrene Dendritic cells 1,4-Dichlorobenzene


DCC DCM 1,3-DCP DDT DEEM DEHA DEHP DEN DEN, DENA DEPM dGua DHEW DINP DINP DMA DMBA DMN DMN DQA DSS DSSTox Dt EAF ECHA ECM ECVAM ED EFSA 2-EH EHEN ELISA EMSA ENNG ENU ENU EPA EPI EPIC ER ERK ESR ESTR EU FDA FDCA

xxv

Deleted in colorectal cancer Dichloromethane or methylene chloride 1,3-Dichloropropene Dichlorodiphenyltrichloroethane Dietary Exposure Evaluation Model Di-(2-ethylhexyl)adipate Di-(2-ethylhexyl)phthalate N-Nitrosodiethylamine N,N-Diethylnitrosamine Dietary Exposure Potential Model Deoxyguanosine U.S. Department of Health Education and Welfare Di-(2-isononyl) phthalate Diisononyl phthalate Dimethylarsenic acid 7,12-Dimethylbenz[a]anthracene or 9,10-Dimethyl-1,2-benz[a] anthracene Dimethylnitrosamine N-Nitrosodimethylamine Data Quality Act Dextran sulfate sodium Distributed structure-searchable toxicity Dose metrics Enzyme-altered foci European Chemicals Agency Extracellular matrix European Centre for the Validation of Alternative Methods Effective dose European Food Safety Authority 2-Ethylhexanol Ethyl hydroxyethylnitrosamine Enzyme-linked immunosorbant assays Electrophoretic mobility shift assay N-Ethyl-N′-nitro-N-nitrosoguanidine Ethylnitrosourea N-Nitroso-N-ethylurea U.S. Environmental Protection Agency Exposure potency index European Prospective Investigation into Cancer and Nutrition Estrogen receptor Extracellular signal-regulated kinases Electron spin resonance Expanded Simple Tandem Repeat European Union U.S. Food and Drug Administration Food, Drug and Cosmetic Act

xxvi


FFDCA FGF FGFR3 FIFRA FISH FPG FQPA GAC γ-GGT GI GJIC GJs GLP G6PD GSSG GSH GST GST-P Gua HaSDR HCA HCA HC HCC HCV HEAA HGP HIV HMG-CoA Hmgcr hPPARα HPLC hprt HPV HPV HPVIS HRF HSC HTLV HTS IAEMS IARC ICEM ICCVAM ICH

Federal Food, Drug and Cosmetic Act Fibroblast growth factor Fibroblast growth factor receptor 3 Federal Insecticide, Fungicide and Rodenticide Act Fluorescent in situ hybridization Formamido pyrimidine glycosylase Food Quality Protection Act Genetic alterations in cancer Gamma-glutamyltransferase Gastrointestinal Gap junction intercellular communication Gap junction connections Good laboratory practice Glucose-6-phosphate dehydrogenase Glutathione disulfide Glutathione Glutathione S-transferases Glutathione S-transferase placental form Guanine Health and Safety Data Reporting Hydrocyanic acid High content analysis Health Canada Hepatocellular carcinoma Hepatitis C virus β-Hydroxyacetic acid Human Genome Project Human immunodeficiency virus 3-Hydroxy-3-methylglutaryl-CoA Hydroxymethylglutaryl-CoA reductase Human PPARα High-performance liquid chromatography Hypoxanthine-guanine phosphoribosyl transferase Human papilloma viruses High production volume High Production Volume Information System Human relevance framework Hemocytoblasts Human T-cell lymphotropic virus High-throughput screening International Association of Environmental Mutagen Societies International Agency for Research on Cancer International Conferences on Environmental Mutagens Interagency Coordinating Committee on the Validation of Alternative Methods International Conference on Harmonisation


IDS IKK IL1α IL1β ILSI ILSI RSI IND IPCS IR IRIS IRIS ITER ITC IUR IUR IWGT IWR JaCVAM JECFA JEM JNK Kdis LBD LED01 LED10 LET LFC LMS LMW ln(GSD) LNT LOAEL LSC LSS LTA MAC MACT MAP MC MCL MCMC MDA MEHP MeIQx MIBK

xxvii

Immunodefense system IκB kinase Interleukin-1alpha Interleukin-1beta International Life Science Institute International Life Sciences Risk Sciences Institute Exploratory investigational new drug applications International Programme on Chemical Safety Ionizing radiation Integrated Risk Information System U.S. EPA Integrated Risk Information System International Toxicity Estimates for Risk TSCA Interagency Testing Committee Inhalation unit risk Inventory update reporting International Workshop(s) on Genotoxicity Tests Interaction weighting ratio Japanese Center for the Validation of Alternative Methods Joint FAO/WHO Expert Committee on Food Additives Job exposure matrix c-Jun N-terminal kinases Dissolution rate constants Ligand binding domains Lower limit on effective dose01 Lower 95% confidence limit for the dose giving the animals an increased tumor incidence of 10% Linear-energy-transfer Lowest feasible concentration Linearized multistage Low-molecular-weight protein Logarithm of the geometric standard deviation Linear no-threshold Lowest observed adverse effect level Lymphoblast Life-stage study Local tissue array Apoptosis-induced channel Maximum achievable control technology Mitogen-activated protein Mast cell Maximum contaminant level Markov chain Monte Carlo Malondialdehyde Mono-2-ethylhexyl phthalate 2-Amino-3,8-Dimethylimidazo[4,5-f] quinoxaline Methyl isobutyl ketone

xxviii


miRNA MLA MLE MMP MMS MN MNU MOA MOE MPV MS MSCE MTBE MTD MUP MVK NAS NAS NBR NCEA NCEs NCEH NCoR NDI NF-kB NHANES NIOSH NIOSH-IREP NNG NNM NOAEL NOEL NPCs NRC NRC NSRLs NTP NTP Cmax OECD OEHHA 8-OH-dG 2-OH-TMP OMB OPP

MicroRNAs Mouse lymphoma tk+/− assay Maximum likelihood estimate Matrix metalloprotease Methyl methanesulfonate Micronuclei Methylnitrosourea Mode of action Margin of exposure Medium-production volume Mass spectrometric Multistage clonal expansion Methyl-tert-butyl ether Maximum tolerable dose Mouse urinary protein Moolgavkar–Venzon–Knudson National Academy of Sciences U.S. National Academies of Science NCI Black–Reiter U.S. EPA National Center for Environmental Assessment Normochromatic erythrocytes National Center for Environmental Health Nuclear receptor corepressor National death index Nuclear factor kappa B National Health and Nutrition Examination Survey U.S. National Institute for Occupational Safety and Health Interactive RadioEpidemiological Program Net nuclear grain N-Nitrosomorpholine No observed adverse effect level No observed effect level Nonparenchymal cells National Research Council U.S. National Research Council No significant risk levels National Toxicology Program U.S. National Toxicology Program Maximum or peak concentration Organisation for Economic Co-operation and Development Office of Environmental Health Hazard Assessment, California EPA 8-Hydroxy-2′-deoxyguanosine 2,2,4-Trimethyl 2-pentanol U.S. Office of Management and Budget U.S. EPA Office of Pesticide Programs


OPPTS ORD OSHA OSH Act OSOR OSTP OSWER PAHs PAIR PAPS PBBs PBPK PBTs PCBs PCDD PCE pCi PCNA PD PDF PDGF PEI PELs PFAA PFOA PFOS PGMBE Pgp PHGs PhIP PIR PMR POD PPAR PPAR-α PPL PPREs pRb PRGs PSP PTEN PTL PXR q1* qPCR (Q)SAR

Office of Prevention, Pesticides and Toxic Substances U.S. EPA Office of Research and Development U.S. Occupational Safety and Health Administration U.S. Occupational Safety and Health Act of 1970 One substance, one registration U.S. Office of Science and Technology Policy U.S. EPA Office of Solid Waste and Emergency Response Polycyclic aromatic hydrocarbons Preliminary assessment and information reporting 3′-Phosphoadenosine 5′-phosphosulfate Polybrominated biphenyls Physiologically based pharmacokinetic Persistent, bioaccumulative, and toxic substances Polychlorinated biphenyls Polychlorinated dibenzo dioxin Polychromatic erythrocyte Picocuries Proliferating cell nuclear antigen Cell population growth over time Probability density function Platelet-derived growth factor Polyethyleneimine Permissible exposure limits Perfluoroalkyl acid Perfluorooctanoic acid Perfluorooctanesulfonic acid Propylene glycol monobutyl ether P-glycoprotein Public health goals 2-Amino-1-methyl- 6-phenylimidazo[4,5-b] pyridine Proportionate incidence ratio Proportionate mortality ratio Point of departure Peroxisome proliferator-activated receptor Peroxisome proliferation activating receptor-alpha 32 P-Postlabeling PPARα responsive elements Inactivated retinoblastoma gene product Preliminary remediation goals Poorly soluble particles Phosphatase and tension Priority testing list Pregnane X receptor Upper 95% confidence limit on the cancer potency slope Quantitative polymerase chain reaction Quantitative structure–activity relationships

xxix

xxx


RAGS RBC RBP RCF R&D REACH REDs RELs RfD RFLP RIVM RMM RNS R.O.C. RoC ROS RSD RTG RT-PCR RTT RXRα SA SAB SAR SARA SCE SDWA S9 fraction SDS SEER SHEDS SIEF SIR SMR SOT SPP SPS SS IIC STN SV SWCNT SWP T-90 t/a TAA

Risk assessment guidance for Superfund Red blood cell Risk-based prioritizations Refractory ceramic fibers Research and development Registration, evaluation, authorization, and restriction of chemicals Reregistration eligibility documents Recommended exposure limits Reference dose Restriction fragment length polymorphism The Netherlands National Institute for Public Health and Risk management measures Reactive nitrogen species Receiver operating characteristic Report on carcinogens Reactive oxygen species Risk–specific dose Relative total growth Reverse transcription polymerase chain reaction Renal tubule tumors Retinoid X receptor-alpha Structural alert U.S. EPA Science Advisory Board Structure–activity relationship Superfund Amendments and Reauthorization Act Sister chromatid exchange U.S. Safe Drinking Water Act 9000 g Supernatant Safety data sheet Surveillance epidemiology and end results Stochastic human exposure and dose simulation Substance information exchange forum Standardized incidence ratio Standardized mortality ratio U.S. Society of Toxicology Security and prosperity partnership Sanitary and phytosanitary Stoddard solvent IIC Stochastic transition network Simian virus Single-walled carbon nanotubes Safety working party 90% Clearance time Tonnes per annum Thioacetamide


TBA TBARS TCA TCDD, dioxin TCE TD TD50 TDI TERA TF TGD TGF TGFβ1 tk TLC TMP TNF TNFα ToxRefDB TPA TRAIL TRI TSCE TTC TWA UCL UDS UF USDA UVR VC VLDL VOC WOE vPvBs VSD WHO Wnt WT1/2 WTO

Tert-butyl alcohol Thiobarbituric reactive substances Trichloroacetate 2,3,7,8-Tetrachlorodibenzo-p-dioxin Trichloroethylene Tolerable dose The dose inducing a tumor incidence of 50% in rodents Tolerable daily intake Toxicology Excellence for Risk Assessment Transcription factor Technical guidance document Transforming growth factor Transforming growth factor beta 1 Thymidine kinase Thin-layer chromatography 2,2,4-Trimethylpentane Tumor necrosis factor Tumor necrosis factor alpha Toxicology reference database Tetradecanoyl phorbol acetate the Environment TNF-related apoptosis-inducing ligand Toxics release inventory Two-stage clonal expansion Threshold of toxicological concern Time-weighted average Upper confidence limit Unscheduled DNA synthesis Uncertainty factor U.S. Department of Agriculture Ultraviolet radiation Vinyl chloride Very low density lipoproteins Volatile organic compound Weight of evidence Very persistent and very bioaccumulative substances Virtually safe dose World Health Organization Wingless type Weighted clearance half-time World Trade Organization

xxxi

PART

I

CANCER RISK ASSESSMENT, SCIENCE POLICY, AND REGULATORY FRAMEWORKS

CH A P TE R

1

CANCER RISK ASSESSMENT Elizabeth L. Anderson Kimberly Lowe Paul Turnham

1.1. 1.1.1.

CANCER RISK ASSESSMENT Cancer in the United States

Cancer is a group of diseases that result from abnormal and prolific cellular division. Based on current U.S. National Cancer Institute’s Surveillance Epidemiology and End Results (SEER) of cancer prevalence, it is estimated that more than 10 million people were living with cancer in the United States in 2005 (NCI 2008). The American Cancer Society predicts that 1 in 2 males and 1 in 3 females will develop some type of cancer in their lifetime, and that 1 in 4 males and 1 in 5 females is at risk of dying from this disease (NCI 2007a,b). Cancer is undoubtedly a substantial threat to public health. Understanding the etiology of cancer, identifying methods of prevention or treatment, and determining the carcinogenicity of the chemicals we use in our everyday lives are the objectives of many of our government divisions, academic institutions, and health-care industries. However, for public health agencies charged with quantifying safe levels of exposure to protect public health, these tasks are not simple matters of using biology to inform the standard-setting process; instead, gaps in science must be filled using a number of assumptions that are based both on scientific inferences and policy judgments. Under Congressional delegation, the broad mission of public health agencies is disease prevention. This includes a wide range of activities from providing education about healthy living to regulating the use and dispersion of agents that are known, or suspected, to cause cancer or other diseases. The basic principle of cancer risk assessment is to characterize both the weight of evidence (WOE) that the agent might be capable of causing cancer and the magnitude of risk, given past, current, or future exposure levels. The fundamental objective is to determine the threshold at which exposure to the agent poses no appreciable risk to humans or, in the absence of mechanistic knowledge, to define an acceptable risk for suspect carcinogens.

Cancer Risk Assessment, edited by Ching-Hung Hsu and Todd Stedeford Copyright © 2010 John Wiley & Sons, Inc.

3

4

CHAPTER 1 CANCER RISK ASSESSMENT

1.1.2.

Historical Perspectives of Cancer Risk Assessment

Imagine a time when there was no exposure assessment, no evaluation of dose– response relationships (potency), and no particular attention paid to mechanisms of action to define the relevance of responses in animals to diseases in humans, as well as a time when the science of risk assessment to address environmental carcinogens was not developed. This time existed when the U.S. Environmental Protection Agency (EPA) was created in 1970, and it existed until the first Federal policy to adopt the use of risk assessment and risk management was announced by the Agency in 1976 (Albert et al. 1977; USEPA 1976). This policy was accompanied by the first guidelines for carcinogen risk assessment (USEPA 1976) and the establishment of an Agency group to carry out these assessments (named the Carcinogens Assessment Group, or CAG). The approach was novel at the time; however, it borrowed from the experience of radiation risk assessment, where a common mechanism of action was known and dose–response relationships in humans had been reasonably well characterized. Of course, large knowledge gaps existed. For most agents suspected of causing cancer, evidence was from high-dose studies in animals that relied on two dose levels to define cancer potential for humans who experienced much lower environmental exposures. Although controversial at the time, the science of risk assessment has developed into the internationally accepted approach to evaluate carcinogen risk associated with of exposure to environmental agents, food contaminants, and occupational contaminants. These approaches also have dictated close scrutiny of the scientific principles that lead to improved methods of addressing potency, mechanisms of action, test methods, exposure, and internal dose relationships. This section describes the landmarks and key events in the evolution of this science. Not long after the EPA was established, it began evaluating carcinogenesis data and translating its findings into public policy. These early decisions spawned the necessity to depart from simple qualitative characterization of tumors in humans or animals to incorporate the reality of exposures at low doses, far below those in the studies, and the potential for harm associated with these low-dose exposures. Because the Agency was newly developed, there was no precedent for regulating carcinogens in the environment. The early years of the EPA were a time of enormous zeal to cleanse the environment, especially of carcinogens that were thought to be the principal cause of a “cancer epidemic.” The Food, Drug, and Cosmetic Act (FDCA) had a provision for regulating intentional food additives to a zero-tolerance level, meaning that evidence of cancer by tumor formation in animals or humans was sufficient cause for banning the agent. The same zero-tolerance policy was attempted for a wide range of environmental agents thought to be potential carcinogens, including three major pesticides: dichlorodiphenyltrichloroethane (DDT), aldrin/dieldrin, and chlordane/ heptachlor, although the cancellation of DDT was probably more compelled by ecologic harm (USEPA 1972, 1975). Between 1970 and 1975, the EPA moved to suspend their use. The cancellation of these three pesticides set the zero-tolerance policy in motion and became what was judged to be the Agency’s cancer policy. However, it quickly became evident that a zero-tolerance policy was impractical.

1.1. CANCER RISK ASSESSMENT

5

For many economically important products, it was impossible to remove all exposure to agents suspected of having the ability to cause cancer (e.g., low-level exposure to benzene, a known human carcinogen, in gasoline). The policy was also highly controversial. Using the qualitative evidence of tumors in animals or humans, attorneys at the EPA had summarized the scientific information needed to characterize an agent as carcinogenic in legal briefs at the conclusions of the hearings to cancel the pesticides listed above. These summary statements were referred to in legal motions as “Cancer Principles.” The intent of these statements was to establish the foundation for the EPA’s authority to protect public health from exposure to environmental carcinogens. This approach received substantial criticism from the scientific community, parts of the private sector, and the Congress (Anonymous 1976). The criticism was largely based on the fact that the complex field of carcinogenesis could not be reduced to simple summary statements (USEPA 1976). In addition, there was concern that the Agency would take a broad approach to cancer regulation by labeling agents as carcinogenic in humans if they were carcinogenic in animals, treating all agents as if they had equal potency, or regulating without information about exposure and the specific threat of a particular agent. Given the large number of chemicals to which people are exposed in their everyday lives, there was a substantial need to establish a basis for setting priorities and balancing the risks associated with their use in terms of social and economic factors, as called for by the specific statutes under which public health agencies operated, including the EPA, which had inherited very broad authorities (Anderson 1983). Ultimately, the failure of the zero-tolerance policy led to the development of the risk assessment framework at the EPA. It was not until 1979 that other federal agencies joined the EPA in an effort to establish interagency guidance for conducting carcinogen risk assessments (Albert et al. 1977; IRLG 1979c; USEPA 1976). This initial risk assessment approach was developed to answer two questions (Anderson 1983): 1. How likely is the agent to be a human carcinogen? This step involves evaluating all of the relevant biomedical data to determine the total weight of evidence (WOE). At that time, the WOE was ranked from strongest to weakest in a scientific context. The strongest evidence was obtained from human data that were supported by animal bioassay results. Substantial evidence of carcinogenicity could be obtained from laboratory animal bioassay results showing replication of effects across species related to dose levels, and suggestive evidence could be obtained from weaker associations in animal studies. Other evidence from in vivo or in vitro studies was also considered. 2. On the assumption that an agent is a human carcinogen, what is the magnitude of its public health impact given current and projected exposures? This step is quantitative in nature and involves establishing a dose–response relationship to extrapolate to low levels of exposure, where environmental exposures generally occur, and evaluating the magnitude of the exposures of interest. Its purpose was to provide regulators a sense of the cancer potency of the agent, and some information about the public health impacts associated with exposures. In this step, risks were bracketed between an upper and lower

6


bound approaching zero. The upper bounds were expressed both in terms of the individual increased cancer risks in the exposed population and the nationwide impact in terms of the annual increase in cases. Of particular note: (1) These first guidelines called for revising each risk assessment as better information became available, a goal that has been rarely realized. (2) Gaps in scientific knowledge were to be filled with public health protective assumptions to err on the side of safety, an early application of the precautionary principle. Over the last several decades, the Agency has sought to extend guidelines for carcinogens to incorporate improvements in our understanding of the cancer process. Because risk assessment necessarily relies on both science and policy judgments, these guidelines are essential to ensure that a consistent approach to risk assessment is taken. The effort to bring consistency to risk assessment is evolving and has produced revisions of guidelines and standard practices (examples of which are shown in Table 1.1). The most fundamental endorsement of the risk assessments

TABLE 1.1.

Historical Perspectives of the Development of the Risk Assessment Process

Year

Document

Details

1975

Quantitative Risk Assessment for Community Exposure to Vinyl Chloride (Kuzmack and McGaughy 1975) Interim Procedures and Guidelines for Health Risks and Economic Impact Assessments of Suspected Carcinogens (USEPA 1976) Hazardous substances summary and full development plan. United States. Interagency Regulatory Liaison Group (IRLG 1979a) Publications on toxic substances. United States. Interagency Regulatory Liaison Group (IRLG 1979b) Integrated Risk Information System (IRIS)

This was the first risk assessment document to be completed by the EPA.

1976

1978

1979

1980

1983

Risk Assessment in the Federal Government: Managing the Process (NRC 1983)

This document communicated the EPA’s intent to include “rigorous assessments of health risk and economic impacts” in the regulatory process. This document describes laws and legislation regarding hazardous substances and chemicals. This document reports basic facts about toxic substances and describes the publications that are available from many federal agencies. This database reports human health effects that may be related to chemicals found in the environment. Commonly referred to as the “Red Book,” this document was published by the National Academy of Sciences and described methods for risk assessment in the federal government. The EPA adopted and implemented the risk assessment methods that were outlined in this book.

(Continued)


TABLE 1.1.

Year

(Continued) Document

1984

Risk Assessment and Management: Framework for Decisionmaking (USEPA 1984)

1985

Chemical Carcinogens: A Review of the Science and Its Associated Principles (OSTP 1985)

1986

The Risk Assessment Guidelines of 1986a (USEPA 1986b)

1986

Guidelines for Carcinogen Risk Assessment (USEPA 1986a)

1989

Risk Assessment Guidance for Superfund, Vol. I: Human Health Evaluation Manual (Part A) (USEPA 1989) Proposed Guidelines for Carcinogen Risk Assessment (USEPA 1996)

1996

1997

7

Exposure Factors Handbook. U.S. EPA (USEPA 1997)

Details Published by the EPA, this document illustrated the strengths and weaknesses of the risk assessment process and emphasized the need to make the process as transparent as possible. Published by the U.S. Office of Science and Technology Policy (OSTP), this document provides a complete review of the application of epidemiology in carcinogen risk assessment. This EPA document provided guidelines for evaluating the human and animal evidence of carcinogenicity, as well as a classification scheme for categorizing the level of risk associated with a particular agent (i.e., limited, inadequate, no data, or no evidence). The purpose of these guidelines was to outline a procedure that EPA scientists could use to assess the cancer risk associated with exposure to chemicals in the environment. This document was also used to inform the public about the process of cancer risk assessment. Published by the EPA Office of Solid Waste and Emergency Response (OSWER), this is the first of a series of guidance documents on risk assessment for the Superfund. Because limitations were identified in the 1986 carcinogen risk assessment guidelines, new cancer risk assessment guidelines were set forth that allowed scientists the flexibility to incorporate relevant biological information into the assessment process. The new guidelines were reviewed by the EPA Science Advisory Board (SAB) in 1997. The guidelines were made available for public comment in 2001 and then were reviewed again by the SAB in 2003. Published by the EPA National Center for Environmental Assessment (NCEA) within the EPA’s Office of Research and Development (ORD), this document provides data on exposure activities and other parameters for assessing exposure to contaminants in the environment. The 1997 handbook updates the 1989 original.

(Continued)

8


TABLE 1.1. (Continued)

Year

Document

2002

OSWER Draft Guidance for Evaluating the Vapor Intrusion to Indoor Air Pathway from Groundwater and Soils (Subsurface Vapor Intrusion Guidance) (USEPA 2002) World Trade Center Indoor Environment Assessment: Selecting Contaminants of Potential Concern and Setting Health-Based Benchmarks (USEPA 2003)

2003

2005

Guidelines for Cancer Risk Assessment (USEPA 2005)

2008

Child-Specific Exposure Factors Handbook (USEPA 2008a)

2009

The U.S. Environmental Protection Agency’s Strategic Plan for Evaluating the Toxicity of Chemicals (USEPA 2009)

Details Published by the EPA Office of Solid Waste and Emergency Response (OSWER), this document provides guidance for the evaluation of the vapor intrusion exposure pathway. This document, published by the Contaminants of Potential Concern (COPC) Committee of the World Trade Center Indoor Air Task Force Working Group, provides guidelines and methodologies for setting health based standards for chemicals in settled indoor dust. The formal guidelines for cancer risk assessment were initially developed in 1986 and were finalized in 2005. After almost two decades of scientific input and progress, the final guidelines were designed to be flexible, with the ability to evolve as scientific advancement occurs. Published by the National Center for Environmental Assessment (NCEA) within the EPA’s Office of Research and Development (ORD), this document supplements the 1997 Exposure Factors Handbook with child-specific data on exposure activities and other parameters for assessing exposure to contaminants in the environment. In response to modern advances in computational and molecular biology, the EPA developed a strategic plan in 2009 to outline an approach for transforming and improving toxicity testing and risk assessment over the next 10 years. The premise of the proposed new plan is that risk assessors should consider how genes, proteins, and small molecules interact in the molecular pathways to maintain cellular function and how exposure to agents in the environment could disrupt these pathways. The strategic plan is built upon three components: (1) toxicity pathway identification and chemical screening prioritization, (2) toxicity pathway-based risk assessment, (3) institutional transition.


9

that had been practiced at EPA since 1976, where approximately 150 carcinogen risk assessments had been completed in the first eight years, occurred in 1983 when the National Research Council (NRC) of the U.S. National Academies of Science (NAS) endorsed risk assessment as a proper process and defined specific steps for hazard identification, dose–response assessment, exposure assessment, and risk characterization as the risk assessment paradigm (NRC 1983). This endorsement created wider applications of risk assessment, which rapidly expanded across all federal regulatory agencies and beyond to state agencies and international communities. The specifics of this process are described in the following section. Present-day risk assessment methodologies have an increasing emphasis on physiologically based pharmacokinetics (PBPK) or toxicokinetic models and mode of action (MOA). Such models have been developed to predict exposure levels in target tissues for a large number of agents. PBPK models are especially useful in the risk assessment context because they allow data to be extrapolated across species, dose levels, and routes of exposure.

1.1.3.

The Defining Steps in Cancer Risk Assessment

The NAS has developed risk assessment strategies and guidelines that are used by many agencies in cancer risk assessment to answer four fundamental questions: (1) Is the agent a carcinogenic hazard? (2) At what dose does the agent become a carcinogenetic hazard? (3) What is the current and expected extent of human exposure to the agent? (4) What is the estimated disease burden expected from exposure to the agent? The strategies used to answer these questions are divided into four actions (NRC 1983): • Hazard Identification. The total weight of the evidence from epidemiologic, animal, and toxicological studies is evaluated to determine the toxicity and carcinogenicity of an agent. In addition, as scientists begin to understand the process by which healthy cells transform into malignant cells, the use of mechanistic information is becoming more common in risk assessment. This may involve identifying the precursor events that may lead to increased cancer risk, as well as the specific genetic or cellular processes that occur during carcinogenesis. • Dose–Response Assessment. The toxic effect of an agent is dependent upon many factors, including the amount of agent that is ingested, the route of exposure, and the specific endpoint under evaluation. Dose–response assessments are primarily focused on determining the safe dose for human exposure for noncarcinogens or acceptable risk levels for carcinogens. Because thresholds for carcinogen activity could not be defined as had traditionally been the case for noncarcinogens, the first risk assessment guidelines at the EPA relied on a linear, nonthreshold, dose extrapolation model for placing plausible upper bounds on risk; the real risks at low doses were thought to be lower, even approaching zero. Dose–response assessments are generally conducted in animals and use empirical, physiologically based toxicokinetic,

10


or mechanism-based dose–response modeling techniques. In contrast, safety assessments for noncarcinogens historically relied on (a) establishing a no observed effect level (NOEL) or a lowest observed adverse effect level (LOAEL) in animals and (b) reducing this level by application of various safety or uncertainty factors to arrive at a safe dose for humans. Today, there is a convergence of methods for carcinogens and noncarcinogens, at least academically, to utilize understandings of toxicokinetics and toxicodynamics to arrive at safe exposure levels. • Exposure Assessment. The fate of an agent in the environment and the extent to which humans will be exposed to the agent is determined through exposure assessment. The primary interests in exposure assessments are to determine the magnitude, frequency, and duration of the exposure. This assessment involves determining the environmental fate and transport of the agent, as well as evaluating the routes of potential exposure (i.e., inhalation in the air, ingestion in food or water, and through dermal contact). The most detailed guidance for exposure assessment is found in the EPA’s Risk Assessment Guidance for Superfund, Volume I (USEPA 1989) and the EPA’s Exposure Factors Handbook (USEPA 1997). • Risk Characterization. Using both (a) the results of the qualitative hazard identification to express the WOE that an agent poses a cancer risk and (b) the quantitative information obtained from the dose–response modeling together with the results of the exposure assessment, the risk characterization step fundamentally describes the risk associated with exposure to an agent at various levels of exposure for the circumstances of concern. The fact that there are scientific uncertainties in these steps has long been recognized. While there are no formal methods to fully characterize the uncertainties in the hazard assessment and dose–response stages (USEPA 2005), methodology and mathematical techniques exist for accounting for uncertainty (and variability) in the exposure assessment stages. Monte Carlo risk analysis modeling, for example, is a mathematical tool that can be used to describe the impact of uncertainty in a specific exposure scenario. It provides a probability distribution for each uncertainty parameter in the model and then can calculate thousands of probability scenarios. This tool allows risk assessors to model the unavoidable uncertainties that are inherent in the risk assessment process, including the occasion when conflicting expert opinions needs to be combined (Vose 1997). The NAS also defined a separate step, Risk Management, where the level of acceptable risk is established. For suspected carcinogens, an acceptable risk range of one in a million to one in ten thousand has been chosen by the EPA and most other public health agencies as the acceptable risk range for regulatory purposes, with risk becoming less acceptable as it rises above the presumptively safe level of one in ten thousand (40_CFR_Part_61 1989). In addition, the results of any necessary risk–benefit analyses and scientific uncertainty analyses, as well as other social and economic issues as defined by the enabling statute, may be considered at this stage in the process.


1.1.4.

11

The Mode of Action (MOA)

As described in the Guidelines for Cancer Risk Assessment (USEPA 2005, pp. 1–10), the MOA is defined as “a sequence of key events and processes, starting with interaction of an agent with a cell, proceeding through operational and anatomical changes, and resulting in cancer formation.” In fact, the severity of effect associated with exposure to an agent largely depends on the interaction between the biology of the organism and the chemical properties of the specific agent (USEPA 2005). In terms of cancer risk assessment, theoretically the potential carcinogen effect of an agent can be identified through modes of action that influence mutagenicity, mitogenesis, inhibition of cell death, cytoxicology, and immune function (USEPA 2005). Conclusions about the MOA for a particular agent are based on the following questions (USEPA 2005): (1) Do animal tests sufficiently support the hypothesized MOA? (2) If the MOA is supported by animal models, is the same action relevant to humans? (3) Are there specific populations or life stages in which humans are more vulnerable to the MOA? This information is included in the final risk assessment narrative that summarizes the total weight of the evidence regarding the potential carcinogenicity of an agent. Because the MOA is based on physical, chemical, and biological processes, it is possible for an agent to have more than one MOA at different sites within the body. This makes it impossible to generalize the results obtained for one endpoint to other sites within the body. Information on the MOA often includes tumor data in humans, tumor data in animals and observations from in vitro test systems, and the structural analogue of the agent (USEPA 2005). As with all components of risk assessment, establishing the MOA of an agent can only be defined with confidence where complete data packages, rather than generic assessments or general knowledge of the agent, provide the foundations. When determining if the MOA observed in animal models is relevant to humans, risk assessors must rely on many sources of information including consideration of the tumor type, the number of studies conducted at each site, and the subgroups evaluated (gender, species, etc.), the metabolic activation and detoxification process observed in the animal model and in humans, the route of exposure, the dose, and the effect of dose and time on the progression of the tumor (see Chapter 13) (USEPA 2005). Only rarely are complete data sets available for defining the MOA. Most often the available information can provide only partial certainty about the MOA and its contribution to the WOE evaluation.

1.1.5.

Accounting for Scientific Uncertainty

One of the greatest challenges of risk assessment is to account for and manage the scientific uncertainty associated with each step in the assessment process. Uncertainty is an unavoidable consequence of evaluating the fate of an agent in our dynamic environment and complex human systems. Sources of uncertainty in assessing the carcinogenicity of an agent include: (1) the parameter values resulting from data that are limited or inadequate, (2) the parameter modeling caused by inherent limitation

12


in the models that are used to evaluate exposures and outcomes, and (3) the completeness of the assessment because of the often infeasible task of exhaustively evaluating all possible components of risk (USEPA 1997). In addition, there is uncertainty associated with applying the results of laboratory animal studies to humans (i.e., interspecies extrapolation), estimating the risk of low-dose ambient exposures from high-dose animal studies (i.e., dose extrapolation), and accounting for the needs of susceptible populations (i.e., intraspecies extrapolation). Given these intrinsic challenges, it may be impossible to guarantee that the best outcome identified in the risk assessment process will actually occur; however, it is imperative that public health decisions are made despite these uncertainties. The consequence of not doing so would be paralysis of the public health and regulatory systems (Bean 1988).

1.2. THE WEIGHT OF EVIDENCE (WOE) FOR DETERMINING CARCINOGENICITY 1.2.1.

Epidemiologic Studies

Results from well-conducted epidemiologic studies provide the strongest weight of evidence (WOE) in cancer risk assessment. Epidemiology is the science of understanding the distribution of disease among humans and the factors that increase or decrease the risk of disease incidence (see Chapter 15). Because epidemiologic studies always measure an exposure (i.e., to a toxic agent) and an outcome (i.e., a specific cancer type), they are of great value to the cancer risk assessment process. Nevertheless, most observations in human populations have occurred when populations have been inadvertently exposed at high levels, above those commonly experienced in the environment. Epidemiologic studies are conducted in humans; therefore there are no issues related to species-to-species variation; however, other factors must be considered when estimating how the carcinogen potential of an agent may change when exposures are far lower or when population circumstances are at issue—for example, when lifestyle factors of the individual or population are concurrently assessed. The best evidence comes from well-conducted epidemiologic studies that are sufficiently powered to test a specific hypothesis and are backed up by confirmatory animal studies. However, well-conducted epidemiology studies are available for only a limited number of substances and often have limited uses because of difficulties involved in interpretation. Unlike animal studies that are conducted in a controlled setting within the laboratory, epidemiologic studies seek to evaluate humans in their natural environments. This is both advantageous and challenging for the risk assessment process. Well-conducted epidemiologic studies will often have many of the following attributes (USEPA 2005): The objectives and the hypothesis are clearly stated, the people included in the study have been properly selected, the exposure has been characterized, the length of the study is long enough to ensure adequate time for the disease to occur, design flaws that may bias the results have been identified and minimized, factors that may confound the relationship between the exposure and the outcome have been properly accounted for, enough people have been enrolled in the study

1.2. THE WEIGHT OF EVIDENCE (WOE) FOR DETERMINING CARCINOGENICITY

13

to detect the desired measure of effect, the data have been collected and analyzed using appropriate methods, and the results have been clearly documented. Because it is possible for one or more of these factors to be inadequate, epidemiologic studies that show no association between exposure to an agent and a cancer outcome do not prove that an agent has no carcinogenic potential. Therefore, the limitations of epidemiologic studies that are used in the risk assessment process must be identified and considered. The types of epidemiologic studies used by risk assessors include case–control studies, cohort studies, descriptive epidemiologic studies, and case reports: • Case–control studies enroll people who have the disease (i.e., cases) and people who do not have the disease (i.e., controls) and then look retrospectively to assess the differences in exposure between the two groups. It is possible to determine causality from a well-conducted case–control study; overall evidence of causality is judged as a WOE that takes account of all qualified epidemiologic studies. • Cohort studies enroll people who have been exposed to the agent of interest and people who have not been exposed to the agent, and then they follow the two groups through time to see which group (if either) has a higher incidence of disease. It is possible to determine causality from a well-conducted cohort study; overall evidence of causality is judged as a WOE that takes account of all qualified epidemiologic studies. • Descriptive epidemiologic studies do not have a temporal component like case–control or cohort studies. Rather, this type of study evaluates factors that may influence the incidence of a disease, such as demographic or socioeconomic characteristics. It is not possible to determine causality from a descriptive epidemiologic study. Rather, this type of study is often used to generate a hypothesis that can be tested in case–control or cohort studies. • Case reports are used to describe specific events or outcomes that occurred in a small number of people. It is not possible to determine causality from case reports, but they are useful for identifying unique events, such as the effects of a unique exposure or the incidence of an unusual tumor and for generating hypotheses that may be tested in follow-up, appropriately designed studies. The premise of epidemiology is to determine if there is an association between an exposure and an outcome. However, the goal of risk assessment is to determine if the WOE from all human studies establishes that the agent is known to cause the outcome. In 1965, Sir Bradford Hill developed a list of criteria that is used to help scientists and epidemiologists assess whether the relationship between an exposure and an outcome in epidemiological studies is causal (Hill 1965). Meeting each criterion does not provide a definitive determination of causation, but it does provide substantial information that can be used when the weight of the evidence is evaluated. In addition, the Hill criteria are intended for use in the evaluation of human data, not the combination of human and animal data. As listed below, the EPA has slightly modified the original list that was developed by Hill so it can be used in modern-day risk assessments (USEPA 2005):

14


1. The association is observed across many different independent studies. 2. The magnitude of the association is large. 3. There is specificity in the observed association such that one exposure leads to one outcome. (Note: This is currently believed to be the weakest of all of Hill’s criteria.) 4. The exposure precedes the outcome, which leads to a temporal relationship between the two factors. 5. There is a biological gradient that is the result of a strong correlation between the exposure and the outcome. 6. The relationship between the exposure and the outcome is biologically plausible. 7. The relationship between the exposure and the outcome is observed in animal studies or other types of studies. 8. There is experimental evidence of causation from human populations. (Note: Given the ethical boundaries associated with using humans in experiments, data from these types of studies are rarely generated.) 9. Information of the structural analogues of an agent can provide information about causality. In addition, given the complexity of the risk assessment process and the growing amount of scientific literature on this topic, the use of meta-analyses is becoming a necessary skill of risk assessors. Meta-analysis is a valuable statistical technique, in which the potential health effects of an exposure are quantitatively evaluated across the entire body of relevant epidemiologic literature. Meta-analysis differs from a qualitative review of the literature because it is data-driven rather than narrative-based. Conducting a meta-analysis can be a very time-consuming and tedious process, especially when there is a large body of literature available on a specific topic. However, there are many benefits to applying this tool to cancer risk assessment. First, because the results of epidemiologic studies are sometimes conflicting, meta-analysis allows the scientific experts to formally identify sources of heterogeneity across studies. Second, meta-analysis provides researchers with an opportunity to examine selected subgroups of studies and to determine how specific studies influence the overall trend observed in the literature at large. This is especially valuable in cancer risk assessment because factors beyond exposure to the agent may be influencing the risk of cancer. Additional uses of epidemiology information in cancer risk assessments are described in the later part of this book (Chapter 15).

1.2.2.

Animal Models

Whole-animal test models are commonly used to determine the potential carcinogenicity of an agent (see Chapter 14). Animal models provide a platform to evaluate cancer outcomes after long-term exposure to the agent at various doses, as well as to identify possible modes of action. Although epidemiologic studies are favored

1.2. THE WEIGHT OF EVIDENCE (WOE) FOR DETERMINING CARCINOGENICITY

15

because they are conducted in human population, data from animal studies are often the primary data available and do provide valuable information to the risk assessment process because they allow the relationship between the agent and the cancer to be evaluated in a highly controlled environment. In addition, because ethical considerations are different for animals from humans, it is possible to learn a great deal about the factors that influence the carcinogenicity of an agent (i.e., detrimental doses and lengths of exposure that increase the risk of tumor initiation and promotion in the chosen laboratory model). If the outcome of an animal study is the presence of an uncommon tumor type, tumors at multiple anatomical locations within the same animal, development of tumors by more than one route of entry, tumors in multiple species, tumors in both genders, progression of a preneoplastic lesion to a malignant tumor, metastatic disease, unusual tumor response, a high proportion of malignant tumors, or clear evidence of dose-related increases in tumor incidence in replicated studies, then substantial credence is given to the carcinogenic potential of an agent (USEPA 2005). On the contrary, an agent is reasonably deemed as having no carcinogenic potential if no malignancies develop from well-conducted, long-term animal studies in more than two species.

1.2.3.

Weight of the Evidence Descriptors

As part of the risk assessment process, the total weight of the evidence from the aforementioned studies is used to determine the agent’s carcinogenic potential. In an effort to maintain consistency in the assessment and reporting process, agents are typically categorized in some way. The EPA has defined categories that are very similar to categorical schemes used by the U.S. National Toxicology Program (NTP), the International Agency for Research on Cancer (IARC), and the European Union (EU) (USEPA 2005). The example from EPA is as follows. It is possible for an agent to be classified into more than one group if its association with cancer varies by dose or route of exposure. • Carcinogenic to Humans. There is strong evidence of human carcinogenicity. To meet this classification, there must be evidence of causality from epidemiologic studies. If there is not, an agent can still meet this classification if all of the following conditions are met: (1) There is strong evidence of an association but not enough evidence to show exposure to the agent causes cancer, (2) there is extensive evidence that the agent is carcinogenic to animals, (3) the MOA and precursor have been identified in animals, and (4) there is strong evidence that the key precursor events that initiate the MOA in animals also occur in humans. • Likely to Be Carcinogenic to Humans. There is strong evidence of human carcinogenicity, but the weight of the evidence is not sufficient to meet the conditions of the “Carcinogenic to Humans” category. For example, there is strong evidence to support an association between exposure to the agent and cancer, but epidemiologic causality cannot be confirmed. In this category, the agent has generally been carcinogenic to more than one species of animal.

16


• Suggestive Evidence of Carcinogenic Potential. There is evidence to suggest that an agent is carcinogenic, but the data cannot support strong conclusions about its effect. In this category, there are weak associations (that may or may not be statistically significant) between the agent and the cancer outcome in animal or human studies. • Inadequate Information to Assess Carcinogenic Potential. Agents are categorized into this group if there are inadequate or conflicting data of cancer outcomes associated with exposure to a particular agent. • Not Likely to Be Carcinogenic to Humans. Agents are categorized into this group if there is evidence to suggest that there is no association between exposure to an agent and cancer. In some cases, an agent may be carcinogenic in animals, but the MOA is not similar in humans.

1.3.

RISK ASSESSMENT IN THE 21ST CENTURY

1.3.1. Using the Advances in Molecular and Computational Biology In 2009, the EPA released a strategic plan to use new molecular and computational biology technologies in toxicity testing and risk assessment (USEPA 2009). The goal of the strategic plan is to use knowledge about the toxicity pathway to improve how risk assessments are conducted over the next 10 years. Although the complexity of the human body is well appreciated, specific information about toxicity pathways has been lacking. As a result of scientific and technological advances, valuable information about how genes, proteins, and small molecules interact to form pathways that maintain cellular function is quickly emerging (see Part IV). Understanding the manner in which exposure to agents in the environment disrupt these pathways is of high value to the sustained public health. The goal of the strategic plan is to replace whole-animal studies with in vitro tests in human cell lines. This approach would allow the rapid evaluation of new chemicals, chemical mixtures, different exposure scenarios, and the influence of chemicals on sensitive populations. If successful, this approach will be ideal for areas where data from animal and epidemiologic studies are nearly impossible to obtain and the existing knowledge base is lacking for many substances, such as in the fields of developmental toxicology, neurotoxicology, immunotoxicity, and reproductive toxicity. In the new plan, animal models will be used for evaluating mechanisms and the MOA. The plan is built upon three components (USEPA 2009): • Chemical Screening and Prioritization. There is urgent need for the rapid and cost-efficient screening of chemicals so they can be prioritized for risk assessment. This includes chemicals that are produced in high volumes, toxicants in the air, the drinking water Contaminant Candidate list, and chemicals found at Superfund sites. • Toxicity Pathway-Based Risk Assessment. Current risk assessment strategies are challenged by issues related to species extrapolation, dose extra-

1.4. APPLICATIONS IN RISK MANAGEMENT

17

polation, and quantifying cancer risk in susceptible populations. In the new plan, disruptions in the baseline biological processes that are likely associated with toxicity pathways will be identified, and their association with adverse health effects will be measured. • Institutional Transition. Adopting a new paradigm for toxicity testing and risk assessment will require changes to the EPA’s operations, organization, and outreach. The EPA is expecting that this transition will likely require more than a decade for full implementation.

1.3.2.

Genetic Susceptibility

Carcinogenesis is a complex and multistep process that often cannot be simplified into the basic exposure–outcome matrix. The effect an agent has on cancer risk is dependent upon several factors, including, but not limited to, the nature of the individual who was exposed, the dose the individual received, and the length of the exposure. During the formal risk assessment process, it is relatively straightforward to quantify or model the dose levels and the length of exposure an individual may experience under circumstances with defined parameters. In fact, the exposure assessment process has been well informed by guidelines as well as the availability of exposure factors to be used in determining the average concentration an individual might experience over the applicable duration and frequency of exposure. However, determining the genetic factors that may influence cancer risk and then accounting for these findings during the regulation process is challenging. Furthermore, the role of background genetic factors in cancer causation may be far more important than the role of the agent in question. With the completion of the Human Genome Project (HGP) in 2003, a substantial amount of evidence came to light that illustrated the importance of genetic factors in cancer susceptibility and risk. In fact, a person’s genetic background is now considered to be a major factor in determining their risk of developing cancer. Genetic variants in key DNA repair genes and carcinogen metabolism genes have been associated with an increase in risk for some types of cancer.

1.4.

APPLICATIONS IN RISK MANAGEMENT

1.4.1. Translating Risk Assessment into Risk Management in the United States Risk management and public policy decisions related to the regulation of carcinogenic agents are largely based on quantitative risk assessments and qualitative assessments of the biomedical evidence (Anderson 1983). Risk assessment is now commonly used to set priorities, determine if there is residual risk present after the best available technologies have been implemented, balance the risks and benefits of using a carcinogenic agent, set standards and target levels of risk to protect public health, and provide information regarding the urgency of situations where populations have been inadvertently exposed to toxic agents (Anderson 1983).

18


The determination that an agent has the potential to be labeled a suspect or known human carcinogen does not alone provide the quantitative basis for determining a safe level of exposure. As noted from the early work of the Carcinogen Assessment Group (CAG) at the EPA, there are hundreds of agents that show some evidence of carcinogenic potential; however, the relative potency of these agents has been found to vary enormously (Anderson 1983). In fact, some of the chemicals that have the strongest qualitative evidence of carcinogenicity have a relatively low potency. Consequently, risk managers must be cautious and must consider relative potency in setting quantitative standards. In the absence of the MOA, the quantification of risk has defaulted to a linear nonthreshold dose–response model to establish a public health protective level of risk. The best-defined approaches for evaluating the risk and setting a level of protective risk have been defined under the EPA programs for cleanup of hazardous waste sites. Given the protective nature of the inference judgments, the outcomes of the risk assessment process are intended to be biased toward public health protection, and consequently they are best used as plausible upper bounds on risk (USEPA 2005). The EPA has commonly used an acceptable range of risk of one in a million to one in ten thousand, becoming presumptively less acceptable as risk rises above this level. However, public health agencies across national and international boundaries may arrive at different levels of acceptable risk as a generic matter or for particular agents, depending upon the application of the precautionary principal. For risk management purposes, low risk defined by the linear nonthreshold model in association with conservatively evaluated exposure can define, with a reasonable degree of confidence, when a risk to public health is acceptable and not of concern as a causal agent of disease. However, because these approaches rely partly on science and partly on inference-based public health protective assumptions, they cannot be used to determine causality. Therefore, it is inappropriate to imply that associated levels of exposure are causally related to disease occurrence when the acceptable risk ranges used by public health agencies to quantify standards for exposure and remediation are marginally exceeded (USEPA 2008b).

1.4.2.

International Risk Management

In the United States, the science of risk assessment has evolved out of the necessity to make public health decisions in the face of scientific uncertainty. Risk assessment methodologies have been established over the past three decades, and their applications have impacted virtually every aspect of public health and environmental protection in many countries. An example of the far-reaching applications of risk assessment can be found in the World Trade Organization (WTO) Agreement on Sanitary and Phytosanitary (SPS) (Anderson and St. Hilaire 2004; Measures 1994). This agreement requires counties to either (1) adopt the harmonized international standards or (2) use standards based on risk assessment, scientific principles, and scientific evidence if they choose to adopt stricter regulations than the international standards (GATT 1947; Howse 2000; Measures 1994). The WTO provides a platform for resolving discrepancies that arise over the appropriateness of national

1.4. APPLICATIONS IN RISK MANAGEMENT

19

standards that are more restrictive than other national or international standards. As of July 2008, the WTO had 153 members (www.wto.org). In 2007 the regulation on Registration, Evaluation, Authorization, and Restriction of Chemicals (REACH) was enacted in an effort to improve the framework in which chemicals are regulated in the EU. REACH requires industry to be responsible for the assessment and management of risks that may be posed by chemicals, as well as to provide the necessary safety information to their users. The overall goal of REACH is to enhance the manner in which public health and the environment are protected from the risks that are associated with the use of synthetic chemicals. It requires that companies work together to complete the registration requirements for all substances that are made in or imported into the EU. REACH requires participation in the Substance Information Exchange Forum (SIEF), which obligates companies to share information from vertebrate studies. In addition, REACH promotes the framework of “One Substance, One Registration” (OSOR), which minimizes the administrative issues that can be associated with this type of regulation. REACH has also established parameters for submitting chemical safety reports that encourage the collection, evaluation, and dissemination of all data based on the elements of risk assessment and public health protection (Environment_ Directorate-General_of_the_European_Commission 2009). Most developed countries have developed their own guidelines and practices for risk assessment. The Society for Risk Analysis and its flagship journal, Risk Analysis: An International Journal, serve as an academic forum to share the rapidly advancing sciences in the field. Also, the importance of these sciences and their applications and development is found in the curricula of most major universities.

1.4.3.

Risk–Benefit Analysis

Determining the level of risk associated with an agent may not be the only factor that is evaluated when determining when, how, and where the agent will be used. Risk–benefit analyses may play various roles in risk management, to determine if the risk of an agent outweighs its benefits. The enabling statutory language and a variety of other social and economic factors play roles in risk–benefit analysis. Generally speaking, the risk associated with an agent will be tolerated at a higher level if the agent poses substantial benefit (and vice versa). The U.S. Food and Drug Administration (FDA), the U.S. Occupational Safety and Health Administration (OSHA), and the EPA use risk–benefit analyses as permitted by the applicable statute to determine the standard of regulation for a given agent. For example, if the contraindication for a specific type of heart medication is liver cancer in 1 per 10,000 individuals, the risk associated with its use will likely be deemed as more acceptable if the drug reduces the mortality associated with heart attack by 80% than if it reduces mortality associated with heart attack by only 10%. The EPA’s regulation of pesticides is governed by the Federal Insecticide, Fungicide, and Rodenticide Act (FIRFA). Because there are public health benefits associated with controlling pests as well as risk associated with the chemicals used for this purpose, FIRFA requires that the EPA balance the risk and benefits of an agent when determining how it will be regulated. Resulting decisions include

20


quantifying the risk of disease in the general population that is associated with exposure to the agent after normal use, the risk of disease experienced by the applicators of the agent, and comparative risk for a substitute agent, if available. The challenge of risk–benefit analyses is to ensure that all costs are accounted for at the social and environmental levels. In addition, one must consider risks and benefits at both the individual and population levels. Certainly, the level of risk that a person is willing to accept is a private and personal decision.

1.4.4.

Risk Acceptance and Risk Communication

Information obtained from risk assessments is used to aid public health officials in developing management decisions. However, the public will often view the risks associated with an agent differently than will the scientific experts, even after costly and time-consuming risk assessment efforts have been implemented. These discrepancies may be attributable to difference in how the public and scientific communities define risk, or they may stem from the fundamental lack of trust the public has toward the risk assessment process (Slovic 1991). Regardless, risk perception is an important topic that invariably must be considered before the implementation of regulations or public health management decisions. The manner in which an individual or different cultures perceive risk is often influenced by demographic, psychological, social, or political factors (Slovic 1991). The perception of risk can vary between and within individuals, such that two people may perceive the risk of the same agent differently, and a single person may view the risk of an agent differently depending on the current events in their life. Research in this area has consistently revealed many issues that are known to affect how risk is perceived, including (Asante-Duah 2002b): Are exposures to the risk factor voluntary or involuntary? Are the potential or known effects of exposure to the risk factor immediate or delayed? Is the risk factor natural or manmade? Can the risk factor be controlled? If it is controllable, how does the individual perceive their control over the risk factor? Is the type of risk factor new to the individual or are they familiar with it? Are there benefits associated with the risk factor? Are the consequences of exposure to the risk factor manageable or catastrophic? Is the individual exposed to similar risk factors? Are the effects of the risk factor reversible? Are there alternatives to the risk factor? Does the individual view the distribution of the risk factor as equitable within the population? Is exposure to the risk factor continuous or intermittent? Are the consequences associated with exposure to the risk factor tangible? Understanding and considering these issues is a challenging but essential component of risk management. However, effective risk communication is central to the successful implementation and acceptance of management actions. Risk communication often takes shape in the form of written communication (i.e., newsletters, public notices, warning labels) or verbal communication (i.e., focus groups, public meetings, workshops) (Asante-Duah 2002a). In terms of cancer risk assessment, effective risk communication strategies include, but are not limited to: involving all stakeholders and the public early in the decision-making process; taking the necessary steps to ensure that there is a two-way dialogue between the scientific experts

REFERENCES

21

and the interested parties; anticipating and preparing for the mitigation of controversy; delivering clear, honest, and factual information about the risk factors; and implementing a system to evaluate how all parties perceived the risk communication efforts (Asante-Duah 2002a). The precautionary nature of risk management decisions made by public health authorities can approach a zero risk tolerance that is not based on the outcome of the risk assessment process or the certainty of the data that underlie the assessment process but rather on social and political influences. The original purpose of risk assessment was to separate important from less important risks and provide a basis for making decisions to protect the public health. With the adoption of risk assessment and risk management as a process for making public health decisions, the concept of achieving zero risk for suspect carcinogens was abandoned as a workable, achievable policy. The important role of risk assessment is to inform the public health decision process so that responsible decisions in the interest of public health can be made. Extreme application of the precautionary principle, whether motivated by public expectations or regulatory desire to achieve ever lower risk, can lead to a virtual zero tolerance policy; it is the role of risk assessment founded on scientific principles to advise the reasonableness of these policy decisions.

REFERENCES Anonymous (1976). Editorial: Seventeen principles about cancer, or something. Lancet 13, 571–573. 40_CFR_Part_61 (1989). National Emissions Standards for Hazardous Air Pollutants; Benzene Emissions from Maleic Anhydride Plants, Ethylbenzene/Styrene Plants, Benzene Storage Vessels, Benzene Equipment Leaks, and Coke By-Product Recovery Plants. Albert, R. E., Train, R. E., and Anderson, E. L. (1977). Rationale developed by the environmental protection agency for the assessment of carcinogenic risk. J Natl Cancer Inst 58, 1537–1541. Anderson, E. L. (1983). Quantitative approaches in use to assess cancer risk. Risk Anal 3, 277–295. Anderson, E. L., and St. Hilaire, C. (2004). The contrast between risk assessment and rules of evidence in the context of international trade disputes: Can the U.S. experience inform the process? Risk Anal 24, 449–459. Asante-Duah, K. (2002a). Design of public health risk managment programs. In Public Health Risk Assessment for Human Exposures to Chemicals, Kluwer Academic Publishers, London, pp. 237–256. Asante-Duah, K. (2002b). Principles and concepts in risk assessment. In Public Health Risk Assessment for Human Exposures to Chemicals, Kluwer Academic Publishers, London, pp. 43–70. Bean, M. (1988). Speaking of risk. ASCE Civil Eng 589, 59–61. Environment_Directorate-General_of_the_European_Commission (2009). What is REACH? Vol. 2009. European Commission. GATT (1947). General Agreement of Tariffs and Trade. Art X Oct 30. Hill, A. (1965). The environment of disease: Association or causation? Proc R Soc Med 58, 295–300. Howse, R. (2000). Democracy, science and free trade risk regulation on trial at the World Trade Organization. Michigan Law Rev 98, 23–29. IRLG (1979a). Hazardous Substances Summary and Full Development Plan. United States. Interagency Regulatory Liaison Group, Washington, D.C. IRLG (1979b). Publications on Toxic Substances: United States. Interagency Regulatory Liaison Group, Washington, D.C. IRLG (1979c). Scientific basis for the identification of potential carcinogens and estimation of risk. J Natl Cancer Inst 63, 243–268.

22


Kuzmack, A. M., and McGaughy, R. E. (1975). Quantitative Risk Assessment for Community Exposure to Vinyl Chloride. U.S. Environmental Protection Agency, Washington, D.C. Measures, S. A. P. (1994). Marrakesh Agreement Establishing the World Trade Organization. Art 7 and Annex B. Reprinted from H.R. Doc. No. 103–316 at 69–81. NCI (2007a). SEER cancer statistics review 1975–2004. In Lifetime Risk (Percent) of Being Diagnosed with Cancer by Site and Race/Ethnicity: Males, 17 SEER Areas, 2002–2004 (Table I-15) and Females, 17 SEER Areas, 2002–2004 (Table I-16), National Cancer Institute, ed. NCI (2007b). SEER cancer statistics review 1975–2004. In Lifetime Risk (Percent) of Dying from Cancer by Site and Race/Ethnicity: Males, Total U.S., 2002–2004 (Table I-18) and Females, Total U.S., 2002–2004 (Table I-19), National Cancer Institute, ed. NCI (2008). SEER Cancer Statistics Review 1975–2005, Ries, L. A. G., Melbert, D., Krapcho, M., Stinchcomb, D. G., Howlader, N., Horner, M. J., Mariotto, A., Miller, B. A., Feuer, E. J., Altekruse, S. F., Lewis, D. R., Clegg, L., Eisner, M. P., Reichman, M., and E, B. K., eds., National Cancer Institute, Bethesda, MD. NRC (1983). Risk Assessment in the Federal Government: Managing the Process, National Academy Press, Washington, DC. OSTP (1985). Chemical carcinogens: A review of the science and its associated principles. Fed Reg 50, 10371–10442. Slovic, P. (1991). Risk perception and trust. In Fundamentals of Risk Analysis and Risk Management, Molak, V., ed., Lewis Publishers New York, pp. 233–245. USEPA (1972). Respondents brief in support of proposed findings, conclusions, and order at 63–64. In Re: Stevens Industries Inc. et al. Consolidated DDT hearings (5 April 1972). USEPA (1975). Respondents motion to determine whether or not the registration of mirex should be cancelled or amended. Attachment A (9 September 1975). USEPA (1976). Interim procedures and guidelines for health risk and economic impact assessments of suspected carcinogens. Fed Reg 41, 21402–21405. USEPA (1984). Risk Assessment and Management: Framework for Decisionmaking, EPA/600/9–85/002, Washington D.C. USEPA (1986a). Guidelines for Carcinogen Risk Assessment, US Environmental Protection Agency, Risk Assessment Forum Washington, D.C. USEPA (1986b). The risk assessment guidelines of 1986a. Fed Reg 51, 33992–34005. USEPA (1989). Risk Assessment Guidance for Superfund, Vol. I: Human Health Evaluation Manual (Part A), US EPA Office of Emergency and Remedial Response. USEPA (1996). Proposed guidelines for carcinogen risk assessment. Fed Reg 61, 17960–18011. USEPA (1997). Exposure Factors Handbook, US EPA Office of Research and Development/National Center for Environmental Assessment. USEPA (2002). OSWER Draft Guidance for Evaluating the Vapor Intrusion to Indoor Air Pathway from Groundwater and Soils (Subsurface Vapor Intrusion Guidance), US EPA Office of Solid Waste and Emergency Response. USEPA (2003). World Trade Center Indoor Environment Assessment: Selecting Contaminants of Potential Concern and Setting Health-Based Benchmarks. Contaminants of Potential Concern (COPC), Committee of the World Trade Center Indoor Air Task Force Working Group. USEPA (2005). Guidelines for Cancer Risk Assessment, US Environmental Protection Agency, Risk Assessment Forum, Washington, D.C. USEPA (2008a). Child-Specific Exposure Factors Handbook, US EPA Office of Research and Development/National Center for Environmental Assessment. USEPA (2008b). IRIS Limitations, US Environmental Protection Agency. USEPA (2009). The U.S. Environmental Protection Agency’s Strategic Plan for Evaluating the Toxicity of Chemicals, EPA100/K-09/001, Washington, D.C. Vose, D. (1997). Monte Carlo risk analysis modeling. In Fundamentals of Risk Analysis and Risk Management, Molak, V., ed., Lewis Publishers, New York.

CH A P TE R

2

SCIENCE POLICY AND CANCER RISK ASSESSMENT Gary E. Marchant

2.1.

INTRODUCTION

Cancer risk assessment is primarily a scientific undertaking, but as recognized in 1983 with the publication of the U.S. National Research Council’s (NRC’s) “Red Book” on risk assessment in the federal government (NRC 1983), policy inputs are necessary to bridge the uncertainties and assumptions that are inherent in risk assessment science. As cancer risk assessments are increasingly used to support regulatory decisions with substantial real-world health and economic consequences, the policy inputs into risk assessment become more critical, scrutinized, and contested. Of course, risk management decisions that often utilize risk assessments, such as determining an acceptable level of risk, are also laden with policy issues, but those are beyond the scope of this chapter, which focuses on policy issues relating to how risk assessments are conducted, not on the related issue of how they are used. Policy inputs into risk assessment generally seek to achieve one or more of the following goals: (i) to ensure that risk assessments are scientifically credible and robust, given the inherent uncertainties in risk assessment; (ii) to support a particular policy goal or outcome, such as ensuring greater protection of human health or avoiding inefficient or unwarranted regulatory burdens or liabilities; or (iii) to make risk assessments more efficient, timely, and legally defensible. Many science policy controversies related to risk assessment focus on specific, narrow questions, such as whether to use animal or human data, the shape of the dose-response curve, how data on mechanism or mode of action should affect the risk assessment, and how to treat susceptible subpopulations. Many of these specific questions are discussed elsewhere in this volume. The analysis here, while touching on many of these specific questions as examples, will instead emphasize the broader structural and institutional aspects of how science policy issues affect risk assessment. These broader issues include: • Use of risk assessment in regulatory decision-making • Role of risk assessment guidelines


23

24

CHAPTER 2 SCIENCE POLICY AND CANCER RISK ASSESSMENT

• Data quality requirements • Types of data that can be used in risk assessment • Application of “conservative” assumptions and precaution In discussing these issues, this chapter will draw primarily on U.S. examples and experience. Although all industrial nations have confronted similar issues in their risk assessment and regulatory programs, the United States has generally had the most advanced and express policies for carcinogen risk assessment, although other nations are now more openly addressing similar issues (Millstone et al. 2008).

2.2. USE OF RISK ASSESSMENT IN REGULATORY DECISION-MAKING Regulatory statutes are generally silent on whether regulatory agencies can, or must, use risk assessment in making regulatory decisions. Rather, the role of risk assessment must be inferred from the statutory regime and interpretation by the courts. Three general approaches are specified by statutes for setting regulatory standards: (i) “acceptable” risk; (ii) cost–benefit analysis; and (iii) feasibility (or best available technology). The first two of these approaches are premised on risk assessment: The first (acceptability) involves identifying risks and then determining what types and levels of risk are acceptable, while the second (cost–benefit analysis) weighs the benefits of reducing risks against the costs of those reductions. Both of these approaches require identification, if not quantification, of risks; thus both approaches presumably permit and arguably require risk assessment. The third approach (feasibility) requires the regulatory agency to reduce risks as low as technologically (or perhaps economically) feasible, and it appears to be oblivious to what the actual risks are. Statutory programs utilizing this approach would therefore presumably not require risk assessment and may even prohibit such consideration. In recent years, there has been a trend in the United States away from riskbased regulatory approaches and toward feasibility or “best available technology” approaches that are not based on risk assessment (Wagner 2000). This trend is largely due to the inherent uncertainties and controversies over risk assessment. The sentiment underlying this trend was expressed by U.S. Senator David Durenberger, who, during the 1990 reauthorization of the U.S. Clean Air Act (CAA) in which Congress abandoned the previous risk-based approach for regulating hazardous air pollutants in favor of a technology-based Maximum Available Control Technology (MACT) requirement, stated (Durenberger 1990): “I’d be glad to declare risk assessment dead.” Courts have enforced the distinction between regulatory programs that permit (or require) risk assessment from those which prohibit reliance on risk assessment. In the seminal 1980 case reviewing the benzene standard promulgated by the U.S. Occupational Safety and Health Administration (OSHA), the U.S. Supreme Court held that the U.S. OSHA must use risk assessment to demonstrate that workers were exposed to a “significant risk” before taking regulatory action (IUD 1980). The U.S. OSHA had proposed to reduce exposures to the lowest levels feasible after

2.3. ROLE OF RISK ASSESSMENT GUIDELINES

25

determining that risk assessment was too uncertain to provide a reliable basis for regulatory action, but the Supreme Court held that the extraordinary power that a regulatory agency wielded could only be exercised after a threshold finding of significant risk. Specifically, the Court held that “the risk from a toxic substance must be quantified sufficiently to enable the Secretary to characterize it as significant in an understandable way” (IUD 1980, p. 646). Conversely, the courts have also overturned agencies for relying on risk assessment in statutory programs that do not authorize such assessments, in particular under statutes mandating a best available technology approach. For example, the D.C. Circuit recently overturned a regulation promulgated by the U.S. Environmental Protection Agency (EPA) that provided a “low-risk” exemption from hazardous air pollutant MACT standards for sources that could demonstrate with a risk assessment that their emissions would impose a maximum individual risk of less than 1 in one million (NRDC 2007). The Court held that such a risk-based approach was impermissible because Congress mandated a technology-based approach that left no room for standards based on risk assessment.

2.3.

ROLE OF RISK ASSESSMENT GUIDELINES

Every risk assessment involves a complex mix of data sets, toxicological methods, models, data gaps, uncertainties, and assumptions. Given this complexity, every risk assessment is, on the whole, unique. At the same time, there are common or at least similar issues that arise over and over again in different risk assessments. Risk assessments conducted on a truly individualized basis, in which the appropriate assumptions and methods to apply are determined de novo based on scientific judgment in light of all available data for that particular risk assessment, would be very difficult, if not impossible, for regulatory agencies. Evaluating all the data and then selecting the appropriate assumptions, methods, and data to apply in a risk assessment is a very resource-intensive undertaking, and it is one that will consume much time and resources and inevitably invite scientific disagreement and controversy (Flamm 1989). To address this tension, some regulatory agencies have developed risk assessment guidelines to provide efficiency, consistency, and predictability in cancer risk assessment. The adoption of these guidelines was heavily influenced by the U.S. NRC’s “Red Book,” which recommended that regulatory agencies adopt risk assessment guidelines containing “inference options” to bridge data or theoretical gaps in the risk assessment process (NRC 1983). The inference options would apply as defaults in the absence of adequate data or theoretical information needed in risk assessment. In response to this report, the U.S. EPA issued its initial carcinogen risk assessment guidelines in 1986 (EPA 1986), which consisted largely of a series of intentionally vague “generalities” about the cancer process and the methods to be used for calculating cancer risk (Albert 1994). These generalities, referred to as “default options,” established a presumptive set of assumptions that were to be applied to address the inherent uncertainties in risk assessment. The default options in the U.S. EPA guidelines were not intended to be inflexible, binding rules, but

26


rather presumptive principles that would apply in the absence of sufficient data to establish an alternative assumption. “Convincing proof” was required to depart from a default option in the guidelines. In contrast, OSHA adopted cancer guidelines as binding rules, and other agencies (including regulatory agencies in most other countries) have not adopted risk assessment guidelines (NRC 2007b). Some of the “most important” default options included in the U.S. EPA’s 1986 carcinogen guidelines (NRC 1994) included the following: • Laboratory animals are assumed to be a valid surrogate for humans in assessing risk; thus, positive cancer results in animal bioassays are taken as evidence of human carcinogenicity. • Humans are assumed to be as sensitive as the most sensitive animal species, and strain or sex is evaluated in an appropriately designed animal bioassay. • Benign tumors are assumed to be as significant as malignant tumors. • The dose–response curve of humans to potential carcinogens is assumed to be linear all the way to the zero exposure levels with no threshold. • A given intake of a substance is assumed to have the same effect regardless of the rate or route of intake. • Individual substances are assumed to exert their effect independently of other substances to which the body is exposed. Most of these default options selected by the U.S. EPA were deliberately chosen to be “conservative,” in that they were intended to estimate the plausible upper bound of actual risk. The adoption of standardized default options in the form of explicit guidelines provided many benefits to the Agency. The highly publicized issuance of the guidelines temporarily quelled much of the brewing controversy about the credibility of the U.S. EPA’s risk assessment practices (Albert 1994). The guidelines helped to sanitize risk assessment by removing the suspicion that the U.S. EPA would manipulate risk assessment principles on a case-by-case basis to support predetermined regulatory outcomes (Goldstein 1989). Risk assessment guidelines also encouraged consistency in risk assessment approach and procedure across the broad array of the U.S. EPA regulatory programs that use risk assessment. Risk assessment guidelines also furthered the objective of efficiency, by sparing the Agency the need to revisit the same controversial issues in each successive rulemaking proceeding. Risk assessment guidelines were also justified on the basis that they would provide regulated businesses greater predictability and certainty about regulatory requirements. The administrative convenience and consistency provided by risk assessment guidelines have, however, come at the expense of flexibility and change in response to emerging science. A 1994 study of the U.S. EPA’s risk assessment practices by the NRC, required by the 1990 CAA Amendments, endorsed the U.S. EPA’s use of default options in its carcinogen risk assessment guidelines, but criticized the agency for applying the guidelines too rigidly (NRC 1994). The U.S. NRC report found that the U.S. EPA rarely, if ever, departed from the defaults in the guidelines, had failed to explain the basis for each default in its guidelines, and had developed no clear

2.3. ROLE OF RISK ASSESSMENT GUIDELINES

27

criteria for departing from default options. The report recommended that the U.S. EPA adopt a more “structured approach” by articulating an explicit set of guidelines or principles for deciding when and how to depart from the default options (NRC 1994, p. 91). Such criteria are needed, according to the U.S. NRC report, “to lessen the possibility of ad hoc, undocumented departures from default options that would undercut the scientific credibility of the agency’s risk assessments” (NRC 1994, p. 105). In 1996, two years after the U.S. NRC report, the U.S. EPA undertook a comprehensive rewrite of its carcinogenic risk assessment guidelines that were finalized almost a decade later in 2005 (EPA 2005a). The revised guidelines incorporated a much more case-by-case approach that considers all relevant evidence. As explained by the EPA, the revised guidelines “are intended to be both explicit and more flexible than in the past concerning the basis for making departures from defaults, recognizing that expert judgment and peer review are essential elements of the process” (EPA 1996). The revised guidelines incorporate “a weight-of-the-evidence approach that considers all relevant data in reaching conclusions about the potential human carcinogenicity of an agent” (Wiltse and Dellarco 1996). Controversial risk assessment policy issues often come to the forefront in the development of risk assessment guidelines. For example, a major controversy in the development of the U.S. EPA’s 2005 revised carcinogen risk assessment guidelines was how to deal with susceptible subpopulations. The U.S. EPA initially took the position that because the guidelines generally apply conservative defaults, they will provide a margin of safety that will protect susceptible subgroups. Environmental organizations and the U.S. EPA’s own Science Advisory Board were critical of this approach, and the controversy was responsible for much of the delay between the 1996 proposal for the revised guidelines and the finalization of those guidelines almost 10 years later in 2005. The U.S. EPA addressed this issue of susceptible subpopulations in part in the final guidelines by publishing an accompanying Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens that proposed to apply an additional age-specific safety factor to account for the increased susceptibility of children, but only for carcinogens that exhibit a mutagenic mode of action (EPA 2005b). The U.S. EPA makes explicitly clear that its risk assessment guidelines are not binding rules, and it is free to depart from the guidelines as necessary (EPA 2005a). However, by adhering to its guidelines, the U.S. EPA not only provides some consistency and predictability to its risk assessments, but also provides some immunity in legal challenges to its risk assessments. For example, adoption of risk assessment guidelines provides a standard by which reviewing courts can review the reasonableness of an agency’s risk assessments. As one federal appeals court noted, “EPA’s specific enunciation of its underlying analytical principles, derived from its experience in the area, yields meaningful notice and dialogue, enhances the administrative process and furthers reasoned agency decision making” (EDF 1976). If the U.S. EPA complies with its own guidelines, the Agency’s decision is likely to be upheld (Ausimont 1988; NRDC 1987). Conversely, if the U.S. EPA violates its own guidelines without a reasoned explanation, its action is susceptible to judicial invalidation (CCC 2000).

28

2.4.


DATA QUALITY REQUIREMENTS

The quality, adequacy, and consistency of risk assessments are promoted by both internal and external oversight mechanisms. The risk assessment guidelines discussed in the previous section are an important internal mechanism. Congress, the Executive branch, and courts all impose some external oversight to ensure that agency risk assessments meet minimum requirements for data quality. Although, as discussed above, Congress has not provided specific requirements relating to risk assessment in most regulatory statutes, section 300g-1(b)(3)(A) of the U.S. Safe Drinking Water Act specifies that the U.S. EPA is to use “the best available, peer-reviewed science and supporting studies conducted in accordance with sound and objective scientific practices” (SDWA 1996). The D.C. Circuit relied on this statutory language to reject a U.S. EPA regulation for chlorinated byproducts in drinking water, which applied a linear dose–response model, even though the “best available” scientific evidence suggested a nonlinear relationship (CCC 2000). In 2000, the U.S. Congress enacted the Data Quality Act (DQA; sometimes also known as the Information Quality Act) requiring the U.S. Office of Management and Budget (OMB) to issue guidance for “ensuring and maximizing the quality, objectivity, utility, and integrity of information … disseminated by Federal agencies” (DQA 2000). The U.S. OMB subsequently issued a directive to federal agencies to “adopt a basic standard of quality (including objectivity, utility, and integrity) as a performance goal,” and to help agencies in this endeavor also provided a model guideline describing substantive standards for information quality (OMB 2001). Each federal agency adopted its own standards for ensuring data quality based on the U.S. OMB guidance. Federal agency risk assessments are therefore subject to these data quality guidelines, and they can be challenged if they fail to meet the applicable standards. The Act requires agencies to provide a mechanism for interested parties to challenge agency actions that purportedly fail to meet the data quality standards, and so each agency provides for members of the public to petition the agency under the Act. Several risk assessment documents prepared by the U.S. EPA and other federal agencies have been challenged under these provisions to date. In some cases the agency has revised the document in response to the DQA petition, whereas in other cases the agency has rejected the petition and upheld the risk assessment document in its original form. A key factor in the application of the DQA is whether agency decisions on petitions are judicially reviewable. Although the statute is silent on judicial reviewability, the initial court cases have held that there is no right of judicial review (SI 2006), which significantly limits the force of the DQA statute. The U.S. OMB has promulgated additional measures that seek to further influence agency risk assessments, ostensibly for the purpose of enhancing the scientific credibility and validity of agency actions. In 2005, the OMB issued a bulletin mandating peer review of scientific information disseminated by the federal government (OMB 2005). This bulletin imposed stringent peer review requirements for “influential scientific information” that included “scientific assessments” such as health risk assessments, which are required to be externally peer reviewed prior to dissemination pursuant to stated criteria for “scientific integrity” and “process

2.4. DATA QUALITY REQUIREMENTS

29

integrity.” Scientific integrity was defined as issues such as “expertise and balance of the panel members; the identification of the scientific issues and clarity of the charge to the panel; the quality, focus and depth of the discussion of the issues by the panel; the rationale and supportability of the panel’s findings; and the accuracy and clarity of the panel report.” “Process integrity” was defined as issues relating to “transparency and openness, avoidance of real or perceived conflicts of interest, a workable process for public comment and involvement, and adherence to defined procedures.” The U.S. OMB also issued a highly controversial Proposed Risk Assessment Bulletin in January 2006 specifying uniform government-wide requirements for agency risk assessments with the stated objective of enhancing “the technical quality and objectivity of risk assessments prepared by federal agencies” (OMB 2006). The U.S. NRC published a highly critical review of the proposed bulletin in 2007 and concluded that it is “fundamentally flawed and should be withdrawn.” (NRC 2007b). One key criticism was that the “one size fits all” approach of the draft guidance fails to accommodate the significant differences between agencies in the types and goals of risk assessments. In response, the U.S. OMB decided to not issue its risk assessment bulletin in final form and instead issued an “Updated Principles for Risk Analysis,” which revised an earlier 1995 document issued to federal agencies on general risk analysis principles (OMB 2007). In addition to the legislative and executive branches, courts also exercise some oversight over the quality of risk assessments. Courts are generally at their most deferential in reviewing risk assessments and other science-based decisions. The U.S. Supreme Court has instructed that “when examining this kind of scientific determination, as opposed to simple findings of fact, a reviewing court must generally be at its most deferential” (BGEC 1983). As another court opinion acknowledged, “substantive review of mathematical and scientific evidence by technically illiterate judges is dangerously unreliable” (Ethyl 1976). Notwithstanding this general deferential approach to scientific risk assessments, reviewing courts will occasionally overturn agency decisions on the ground that they are based on risk assessments that are outdated or otherwise flawed. For example, the U.S. EPA’s proposal to list methylene diphenyl diisocyanate (MDI) as a “high-risk” pollutant under the U.S. CAA was rejected by a reviewing court because the agency applied a generic air dispersion model to calculate human exposure that had “no rational relationship to the known properties of MDI” (CMA 1994). The Agency’s model assumed that MDI will behave as a gas under the relevant conditions, whereas in fact the undisputed evidence before the Agency showed that MDI would be a solid at the relevant temperature. Courts view their role as ensuring that regulatory agency practices decisions “must remain attuned to our rapidly expanding knowledge and technology” (EDF 1978) and must “accurately reflect the latest scientific knowledge useful in indicating the kind and extent of all identifiable effects on public health” (LIA 1980). The controversy over the shape of the dose–response curve in risk assessment is an example that shows the influential, yet somewhat sporadic and unpredictable, role of the courts in risk assessment policy issues. In the regulatory context, agencies such as the U.S. EPA have traditionally applied a linear, no-threshold dose–response

30


model [specifically, the linearized multistage (LMS) model] as a conservative default assumption in carcinogen risk assessment. While the U.S. EPA’s new 2005 carcinogen risk assessment guidelines signal a more flexible, data-based approach to the selection of dose–response model, most of the Agency’s risk assessments reviewed by courts to date have employed the linear, no-threshold model (EPA 2005a). Most judicial decisions reviewing the use of linear dose–response models in the regulatory context have upheld the Agency’s reliance on such models (IFI 1992; PCHRG 1986). A few court decisions, however, have been more skeptical of the linear model. For example, the U.S. EPA’s use of the linear, no-threshold model in its risk assessment for drinking water chlorinated byproducts was rejected by the court because it was contrary to evidence suggesting a nonlinear model that had been accepted by both the U.S. EPA and its Science Advisory Board (CCC 2000). On the other hand, the U.S. OSHA’s departure from the linear, no-threshold model in its formaldehyde risk assessment was likewise rejected by the court (IU 1989). The court held that the U.S. OSHA had improperly used the maximum likelihood estimate (MLE) rather than the upper confidence limit (UCL) to calculate risk, and the UCL but not the MLE model was consistent with a linear dose–response assumption. The court held that the U.S. OSHA had failed to justify its departure from its traditional linear, no-threshold dose–response assumption. Judicial decisions in nonregulatory contexts such as toxic tort and product liability suits are likewise inconsistent in their consideration of the linear, no threshold model. As in the regulatory context, most cases find no problem with an expert’s reliance on a risk assessment using the linear model. In a handful of cases, however, the court rejects reliance on a linear dose–response assumption. For example, one court in addressing the cancer risks from a low concentration of benzene in Perrier® held that “there is no scientific evidence that the linear no-safe threshold analysis is an acceptable scientific technique used by experts in determining causation in an individual instance” (Sutera 1997). Another court decision concluded that “[t]he linear non-threshold model cannot be falsified, nor can it be validated. To the extent that it has been subjected to peer review and publication, it has been rejected by the overwhelming majority of the scientific community. It has no known or potential rate of error. It is merely an hypothesis” (Whiting 1995). The inconsistency and unpredictability of judicial review of risk assessments adds an additional element of uncertainty into the risk assessment process.

2.5.

TYPES OF DATA USED IN RISK ASSESSMENT

Another important policy issue for risk assessment is the type of data that can be used in risk assessment. The context in which the risk assessment is used will often dictate what types of data may be used. In the regulatory context, agencies such as the U.S. EPA tend to make determinations based on the “weight of evidence” that considers all available evidence, including human epidemiological and clinical data (when available), animal studies, and cellular and molecular assays (EPA 2005a). While the U.S. EPA states a preference for human data, it recognizes that

2.5. TYPES OF DATA USED IN RISK ASSESSMENT

31

human data are often not available, and thus the Agency most commonly makes regulatory decisions based on risk assessments using animal data. Courts have generally upheld this reliance on animal data, citing the preventive and prophylactic function of regulatory agencies in preventing toxic exposures and risks (IFI 1992; PCHRG 1986). Cancer risk assessments are also sometimes used in toxic tort and product liability litigation. In this context, courts express a much stronger preference for risk assessments based on human data and are more skeptical of animal studies. For example, the U.S. Supreme Court rejected the reliance of plaintiffs’ experts on animal studies showing that polychlorinated biphenyls (PCBs) can cause cancer, holding that the studies were “so dissimilar” to the human exposure and toxicity at issue in that case as to be without any value (GE 1997). This difference in the evidentiary approach of courts and agencies flows from the different institutional objectives (Allen 1996): Regulatory [agencies] … make prophylactic rules governing human exposure. This methodology results from the preventive perspective that the agencies adopt in order to reduce public exposure to harmful substances. The agencies’ threshold of proof is reasonably lower than that appropriate in tort law, which “traditionally make[s] more particularized inquiries into cause and effect” and requires a plaintiff to prove “that it is more likely than not that another individual has caused him or her harm.”

Another policy issue relating to the type of data used in risk assessment concerns the incorporation of new types of data and methods. Regulatory agencies generally require new test methods and types of data to be validated before they can be used in regulatory risk assessments. While this validation requirement has traditionally tended to be ad hoc and informal, there has been a trend in recent years toward more formal validation requirements (Balls and Fentem 1999). The U.S. Congress enacted legislation in 2000 that required a formalized and harmonized validation system for new toxicological methods relied on by federal regulatory agencies, implemented through the Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM) (Congress 2000). The statute requires that a federal agency “that requires or recommends acute or chronic toxicological testing …. shall ensure that any new or revised acute or chronic toxicity test method, including animal test methods and alternatives, is determined to be valid for its proposed use prior to requiring, recommending, or encouraging the application of such test method” (§4(c)). The European Union has likewise created the European Centre for the Validation of Alternative Methods (ECVAM) to validate new toxicological test methods, and the Organization for Economic Cooperation and Development (OECD) has also adopted formal guidelines for the validation of test methods for use in regulatory decision-making (OECD 2005). The need for formal validation of new test methods can help ensure the validity and consistency of risk assessment methods, but it also carries the risk of further slowing the adoption of new methods. As science has rapidly progressed over the past few decades in its understanding of how toxic agents cause cancer and other adverse effects, risk assessment has struggled to keep up. Despite the rapid development of risk assessment methodologies and their underlying science, regulatory

32


agencies have been slow to incorporate the most recent scientific data and methods. This lag results from a variety of factors, including the inflexibility built into many governmental programs, legal risks created by changes in agency practices, and the tendency for most new developments in risk assessment “science” to be exculpatory by usually downgrading the magnitude or even existence of risk from particular agents. An example of the delay in accepting new data and models was the U.S. EPA’s long and torturous process for accepting data suggesting that male rat kidney cancers caused by a variety of compounds including unleaded gasoline via a mechanism involving the male rat-specific protein α2u-globulin may not be relevant to humans (EPA 1991). The acceptance of this alternative assumption involved a process lasting almost ten years, which included a peer review workshop convened by the U.S. EPA and review by different committees of the Agency’s Science Advisory Board (McClellan 1996). In response to criticisms from the U.S. NRC (NRC 1994) and others that agencies take too long and are too conservative in adopting new data and methods, there is increasingly awareness of the importance of creating incentives for risk assessment scientists to develop and use better toxicological methods. An example of this tension between adhering to established approaches and creating incentives for new types of data can be seen with the U.S. EPA’s response to new toxicogenomic data. The U.S. EPA issued an “Interim Genomics Policy” in 2002 which recognizes “that genomics will have an enormous impact on our ability to assess the risk from exposure to stressors and ultimately to improve our risk assessments” (EPA 2002). The Agency stated it would consider genomic data in risk assessment “on a case-by-case basis,” but “these data alone are insufficient as a basis for decisions” at this time. Moreover, the guidance adds that “[b]efore such information can be accepted and used, agency review will be needed to determine adequacy regarding the quality, representativeness, and reproducibility of the data.” Another recent report issued by the U.S. NRC urged EPA and other regulatory agencies to be more aggressive in supporting and utilizing genomic data, suggesting that toxicogenomic data should not be held to more rigorous standards than other types of toxicological data (NRC 2007a). The report (p. 199) recommended: “Although caution, scrutiny, and validation are required to protect against premature, inappropriate, and unethical use of toxicogenomic data in regulatory and litigation contexts, care should also be taken to ensure that a higher standard of proof is not imposed for toxicogenomic data relative to other types of toxicological data used in regulation and litigation.” In toxic tort and product liability litigation, the admissibility of new risk assessment methods and data must be approved by the trial judge before being presented to the jury. The U.S. Supreme Court announced a new standard for the admission of scientific data in 1993 in its Daubert decision (Daubert 1993). Under this new standard, federal judges must serve as a “gatekeeper” to ensure that scientific evidence is reliable and relevant, which includes an assessment of whether the evidence (i) has been empirically tested, (ii) has a known rate of error, (iii) has been peerreviewed and published, and (iv) is generally accepted within the relevant scientific field. Many state courts have adopted a similar standard, although some still apply the earlier standard on admissibility (Frye 1923), which is whether the evidence is

2.6. APPLICATION OF “CONSERVATIVE” ASSUMPTIONS AND PRECAUTION

33

“generally accepted” in the relevant field of expertise. These admissibility standards for scientific evidence are likely to present a barrier to the introduction of new risk assessment methods or data that have not yet been widely accepted.

2.6. APPLICATION OF “CONSERVATIVE” ASSUMPTIONS AND PRECAUTION Another policy issue in cancer risk assessment is the role of “conservative” (i.e., upper-bound or worst-case) assumptions in risk assessment, a long-standing controversy that has been rekindled by the recent adoption and proliferation of the precautionary principle (Marchant 2003). Risk assessment inevitably involves uncertainties, and agencies such as the U.S. EPA have traditionally sought to bridge such uncertainties with conservative assumptions that represent a plausible upperbound of risk. As the U.S. EPA explains its approach, “[o]ur risk estimates are designed to ensure that risks are not underestimated which means that a risk estimate is the upper bound on the estimated risk” (EPA 2004). The use of conservative assumptions has been criticized by some for inserting risk management policies (i.e., err on the side of safety) into the risk assessment process and also because the compounding of multiple worst-case assumptions may produce risk estimates that are implausibly high (Nichols and Zeckhauser 1988). Other experts have expressed concern that while the use of conservative assumptions may be appropriate initially when uncertainty is large, it is problematic if the initial use of such assumptions prevents revision of risk estimates when new data become available for political reasons: “One implication of the inherent conservatism in risk assessment is that the inevitable consequence of most scientific advances related to the assessment of risk for individual chemicals is to lower the calculated risk …. The worst thing that we can do is to set up a situation so that we cannot use this increased scientific information because of the political aspects of changing numbers” (Goldstein 1989). Still others argue that risk assessments are not conservative enough and that they underestimate risks because, for example, they fail to fully account for susceptible subpopulations and the synergistic effects of some toxic exposures (Finkel 1996; Latin 1988). The courts have generally been sympathetic to the use of conservative assumptions in agency risk assessments, although with some limitations. The U.S. Supreme Court, in its 1980 decision on the U.S. OSHA’s occupational health standard for benzene which ushered in the era of regulatory risk assessment by requiring a threshold showing of “significant risk,” wrote that “so long as they are supported by a body of reputable scientific thought, the Agency is free to use conservative assumptions” in calculating cancer risk (IUD 1980, p. 656). A subsequent court decision interpreted that decision to say that “OSHA may use assumptions, but only to the extent that those assumptions have some basis in reputable scientific evidence” (AFL-CIO 1992). Courts have in some cases rejected the use of conservative assumptions when those assumptions are contradicted by available data. In one case, the court held that the U.S. EPA could not ignore accurate information “at hand” on the relevant risk “in favor of blanket, highly conservative assumptions” (LIA 1994).

34


The court added that “[w]hile the EPA ‘may err on the side of overprotection,’ it ‘may not engage in sheer guesswork.’ ” The debate about the use of conservative assumptions in risk assessment has now melded into the global debate on the precautionary principle. The precautionary principle is an ill-defined concept, adopted into the laws of the European Union and several other countries, which calls for greater precaution in controlling unknown risks, in some versions shifting the burden of proof to the proponent of a product or technology to demonstrate its safety (Marchant 2003). The United States has taken the position that the use of conservative assumptions in risk assessment and other existing regulatory protections provide a sufficient exercise of precaution, and the precautionary principle is unnecessary and ill-advised (Graham 2002). The European Union, the global leader in promoting the precautionary principle, has taken the position that the precautionary principle is a risk management tool that does not even apply to risk assessment; rather, the precautionary principle is considered only after a full scientific risk assessment (CEC 2000). Yet a third position is that the precautionary principle requires that risk assessment be revised to further incorporate precaution, such as by, for example, expanding the scope of potential harms and subjects, giving greater weight to early indications of potential harms that have not yet been demonstrated, and paying more attention to synergistic and cumulative effects of toxic exposures (Goldman 2003; Tickner 2002). Finally, a fourth position supports using the precautionary principle to replace risk assessment altogether (O’Brien 2000). Like many of the issues discussed in this chapter, this controversy is likely to rage for some time given the divergent opinions and interests, important stakes, and strong emotions at issue.

2.7.

CONCLUSION

Science policy issues and controversies underlie almost every aspect of cancer risk assessment. These policy issues are primarily a function of the scientific uncertainties inherent in risk assessment. As new scientific methods and data begin to fill in some of the data gaps and uncertainties in risk assessment, the role of policy will gradually recede, although there is no prospect of policy issues being mooted entirely in the foreseeable future. Moreover, the extent to which we substitute novel scientific data and models for preexisting policy inferences is itself an ongoing policy debate, as is the appropriate role of precaution and conservatism in risk assessment.

REFERENCES AFL-CIO (1992). American Federation of Labor and Congress of Industrial Organizations v. Occupational Safety and Health Administration, U.S. Department of Labor, 965 F.2d 962 (C.A. 11). Albert, R. E. (1994). Carcinogen risk assessment in the U.S. Environmental Protection Agency. Crit Rev Toxicol 24, 75–85. Allen (1996). Allen v. Pennsylvania Engineering Corp., 102 F.3d 194 (C.A.5 (La.)). Ausimont (1988). Ausimont USA Inc. v. E.P.A., 838 F.2d 93 (C.A.3).

REFERENCES

35

Balls, M., and Fentem, J. H. (1999). The validation and acceptance of alternatives to animal testing. Toxicol in Vitro 13, 837–846. BGEC (1983). Baltimore Gas and Elec. Co. v. Natural Resources Defense Council, Inc., 462 U.S. 87. CCC (2000). Chlorine Chemistry Council v. E.P.A., 206 F.3d 1286 (C.A.D.C.). CEC (2000). Commission of the European Communities, Communication from the Commission on the precautionary principle, Brussels, 02.02.2000 COM(2000) 1, pp. 1–29. CMA (1994). Chemical Mfrs. Ass’n v. E.P.A., 28 F.3d 1259 (C.A.D.C.). Congress (2000). ICCVAM Authorization Act (Public Law 106-545), 106th Congress. Daubert (1993). Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579. DQA (2000). Data Quality Act, Pub. L. No. 106–554, § 515(a), 114 Stat. 2763A, pp. 153–154. Durenberger, D. (1990). 1990 Clean Air Act Amendments—Conference Report, Cong. Rec. S16895, S16932 (Oct. 27, 1990). EDF (1976). Environmental Defense Fund, Inc. v. Environmental Protection Agency, 548 F.2d 998. EDF (1978). Environmental Defense Fund, Inc. v. Costle, 578 F.2d 337. EPA (1986). Guidelines for carcinogen risk assessment, EPA/630/R-00/004, 1–38. EPA (1991). Alpha2u-globulin: Association with chemically induced renal toxicity and neoplasia in the male rat, EPA/625/3-91/019F, 1–118. EPA (1996). Proposed guidelines for carcinogen risk assessment, EPA/600/P-92/003C, 1–142. EPA (2002). Science Policy Council—Interim policy on genomics. http://www.epa.gov/osainter/spc/ pdfs/genomics.pdf, 1–4. EPA (2004). An examination of EPA risk assessment principles and practices. EPA/100/B-04/001, 1–182. EPA (2005a). Guidelines for carcinogen risk assessment, EPA/630/P-03/001F, 1–166. EPA (2005b). Supplemental guidance for assessing susceptibility from early-life exposure to carcinogens, EPA/630/R-03/003F, 1–126. Ethyl (1976). Ethyl Corp. v. Environmental Protection Agency, 541 F.2d 1 (C.A.D.C.). Finkel, A. M. (1996). Who’s exaggerating? Discover May 1, 48–54. Flamm, W. G. (1989). Critical assessment of carcinogenic risk policy, Regul Toxicol Pharmacol 9, 216–224. Frye (1923). Frye v. U.S., 293 F. 1013 (C.A.D.C.). GE (1997). General Electric. Co. v. Joiner, 522 U.S. 136. Goldman, L. R. (2003). The red book: A reassessment of risk assessment. Hum Ecolo Risk Assess 9, 1273–1281. Goldstein, B. D. (1989). Risk assessment and the interface between science and law. Columbia J. Environ Law 14, 343–355. Graham, J. D. (2002). The role of precaution in risk assessment and management: An American’s view. Remarks prepared for “The US, Europe, Precaution and Risk Management: A Comparative Case Study Analysis of the Management of Risk in a Complex World” Conference Organizers: The European Commission (Group of Policy Advisers), the U.S. Mission to the EU, the German Marshall Fund with the European Policy Centre and the Center for Environmental Solutions, Duke University (January 11–12, 2002). IFI (1992). International Fabricare Institute v. U.S. E.P.A., 972 F.2d 384. IU (1989). International Union, United Auto, Aerospace and Agricultural. Implement Workers of America, UAW v. Pendergrass, 878 F.2d 389. IUD (1980). Industrial Union Department, AFL-CIO v. American Petroleum Institute. 448 U.S. 607. Latin, H. (1988). Good science, bad regulation, and toxic risk assessment. Yale J. Regul 5, 89–148. LIA (1980). Lead Industries Association, Inc. v. Environmental Protection Agency, 647 F.2d 1130 (C.A.D.C.), certification denied, 449 U.S. 1042. LIA (1994). Leather Industries of America, Inc. v. E.P.A., 40 F.3d 392 (C.A.D.C.). Marchant, G. E. (2003). From general policy to legal rule: Aspirations and limitations of the precautionary principle. Environ Health Perspect 111, 1799–1803. McClellan, R. O. (1996). Reducing uncertainty in risk assessment by using specific knowledge to replace default options. Drug Metab Rev 28, 149–179. Millstone, E., van Zwanenberg, P., Levidow, L., Spok, A., Hirakawa, H., and Matsuo, M. (2008). Riskassessment policies: Differences across jurisdictions. EUR 23259 EN, Joint Research Centre, Institute for Prospective Technological Studies, 1–84.

36


Nichols, A. L., and Zeckhauser, R. J. (1988). The perils of prudence: How conservative risk assessments distort regulation. Regul Toxicol Pharmacol 8, 61–75. NRC (1983). Risk Assessment in the Federal Government: Managing the Process Working Papers, The National Academy Press, http://www.nap.edu/openbook.php?isbn=POD115&page=R1, Washington, D.C. NRC (1994). Science and Judgment in Risk Assessment, The National Academies Press, http://books.nap. edu/openbook.php?record_id=2125&page=R1, Washington, D.C. NRC (2007a). Applications of Toxicogenomic Technologies to Predictive Toxicology and Risk Assessment, The National Academies Press, http://books.nap.edu/openbook.php?record_id=12037&page=R1, Washington, D.C. NRC (2007b). Scientific Review of the Proposed Risk Assessment Bulletin from the Office of Management and Budget, The National Academies Press, http://books.nap.edu/openbook.php?record_id=11811& page=R1, Washington, D.C. NRDC (1987). Natural Resources Defense Council, Inc. v. E.P.A., 824 F.2d 1211 (C.A.D.C.). NRDC (2007). Natural Resources Defense Council v. E.P.A., 489 F.3d 1364 (C.A.D.C.). O’Brien, M. (2000). Making Better Environmental Decisions—An Alternative to Risk Assessment, The MIT Press, Cambridge, MA. OECD (2005). Guidance document on the validation and international acceptance of new or updated test methods for hazard assessment, ENV/JM/MONO(2005)14. OECD Series on Testing and Assessment, No. 34, pp. 1–96. OMB (2001). Guidelines for ensuring and maximizing the quality, objectivity, utility, and integrity of information disseminated by federal agencies (October 1, 2001), http://www.whitehouse.gov/omb/ fedreg/final_information_quality_guidelines.html. OMB (2005). Final information quality bulletin for peer review. Federal Register 70, 2664–2677. OMB (2006). Proposed risk assessment bulletin, http://www.whitehouse.gov/omb/inforeg/proposed_ risk_assessment_bulletin_010906.pdf, pp. 1–26. OMB (2007). Updated principles for risk analysis; M-07-24 Memorandum for the heads of executive departments and agenices (September 19, 2007), http://www.whitehouse.gov/omb/memoranda/fy2007/ m07-24.pdf, pp. 1–13. PCHRG (1986). Public Citizen Health Research Group v. Tyson, 796 F.2d 1479 (C.A.D.C.). SDWA (1996). Safe Drinking Water Act, 42 U.S.C. § 300F to 300J-26. SI (2006). Salt Institute v. Leavitt, 440 F.3d 156 C.A.4 (Va.). Sutera (1997). Sutera v. Perrier Group of America Inc., 986 F.Supp. 655 D.Mass. Tickner, J. A. (2002). Precaution, Environmental Science, and Preventive Public Policy, Island Press, Washington, D.C. Wagner, W. E. (2000). The triumph of technology-based standards, U. Ill. L. Rev., pp. 83–113. Whiting (1995). Whiting v. Boston Edison Co., 891 F.Supp. 12 D.Mass. Wiltse, J., and Dellarco, V. L. (1996). U.S. Environmental Protection Agency guidelines for carcinogen risk assessment: Past and future. Mutat Res 365, 3–15.

CH A P TE R

3

HAZARD AND RISK ASSESSMENT OF CHEMICAL CARCINOGENICITY WITHIN A REGULATORY CONTEXT Henk Tennekes Virginia A. Gretton Todd Stedeford

3.1.

OVERVIEW

The first section of this chapter provides a discussion of hazard assessment, classification of potentially dangerous substances, and the process of risk assessment. A summary of the mandatory and voluntary initiatives for regulating chemicals and biocides in the United States and Europe is also included together with information on the regulatory aspects of hazard communication. The second section deals with the scientific aspects of hazard identification and risk assessment of carcinogenic chemicals within the regulatory context.

3.2.

RISK ASSESSMENT

To enable materials to be stored and used safely, the risks to human health and the environment must be assessed. Risk assessment of both new and existing substances comprises the following steps (NRC 1983): 1. Hazard Identification. Identification of intrinsic hazardous properties. 2. Dose–response Assessment. Determination of the dose/concentration–response characteristics. 3. Exposure Assessment. Exposure assessment for humans (i.e., workers, consumers, and those exposed indirectly via the environment) and for the different environmental compartments (air, soil, water) likely to be exposed to the substance. Cancer Risk Assessment, edited by Ching-Hung Hsu and Todd Stedeford Copyright © 2010 John Wiley & Sons, Inc.

37

38

CHAPTER 3 HAZARD AND RISK ASSESSMENT OF CHEMICAL CARCINOGENICITY

4. Risk Characterization. Comparison of information on hazardous properties and effective dose levels/concentrations with exposure levels in order to characterize the degree of risk posed by the substance to human health or the environment.

3.2.1.

Principles of Risk Assessment and Management

A particular substance may have several hazards, which are categorized as physicochemical, toxicological, or environmental. Physicochemical hazards arise from intrinsic physical or chemical properties of the substance. Toxicological hazards result from a chemical causing harmful effects when ingested, inhaled, or absorbed through the skin. Toxic effects may be acute or chronic, local or systemic, reversible or irreversible. Environmental hazards relate to persistence, bioaccumulation, and toxicity to terrestrial and aquatic organisms. Test guidelines are available for conducting studies aimed at evaluating the physical/chemical properties, environmental fate, and potential human health and ecological hazards of chemical substances [reviewed by Knight and Thomas (2003)]. The most frequently used guidelines are the harmonized test guidelines published by the U.S. Environmental Protection Agency’s (EPA’s) Office of Prevention, Pesticides, and Toxic Substances (OPPTS) (), the Organization for Economic Co-operation and Development’s (OECD’s) guidelines for the testing of chemicals ( < http://www.oecd.org/document/40/0 ,3343,en_2649_34377_37051368_ 1_1_1_1,00.html>), and the European Commission’s testing methods in Annex V to Directive 79/831/EEC (). Risk management measures (RMMs) are implemented after a risk–benefit evaluation and are in the form of instructions for safe use, labeling, or occupational exposure limits. Hazards of substances and preparations must be communicated to users, both workers and the general public. This is achieved by standardized classification and labeling (e.g., EC 2008) of potentially dangerous chemicals and by providing a Safety Data Sheet (SDS). Most developed countries also have legal provisions for banning or restricting the use of chemicals to safe conditions. 3.2.1.1. Classification of Carcinogens. Carcinogenic chemicals are classified on the basis of the weight of evidence. The quality and nature of the evidence determines the category of the classification, not potency. Various bodies in Europe and the United States have subtly different definitions for their categorizations (Persad et al. 2007; Stedeford and Persad 2007); however, international efforts are under way to harmonize classification schemes. For example, under the Globally Harmonised System of Classification and Labelling of Chemicals Regulation No. 1272/2008 (EC 2008), substances are categorized as known or presumed human carcinogens (category 1) if there is sufficient epidemiological and/or animal data to establish a causal association between human exposure to a substance and development of cancer. Substances are categorized as suspected human carcinogens (category 2) based on evidence obtained from epidemiological and/or animal data that are not sufficiently convincing to warrant category 1. Numerous factors need to be taken into account when assessing the weight of evidence for classification as

3.2. RISK ASSESSMENT

39

category 1 or 2 or nonclassification. These factors include but are not limited to the following: • Carcinogenic effects were noted only at very high dose levels exceeding the maximal tolerated dose. • There was appearance of tumors, especially at high dose levels, only in particular organs of certain species known to be susceptible to a high spontaneous tumor formation. • There was appearance of tumors only at the site of application in very sensitive test systems (e.g., intraperitoneal or subcutaneous application of certain locally active compounds), and the particular target is not relevant to man. • There was lack of genotoxicity in short-term tests in vivo and in vitro. • There is a secondary mechanism of action with the implication of a practical threshold above a certain dose level (e.g., hormonal effects on target organs or on mechanisms of physiological regulation, chronic stimulation of cell proliferation). • There is a species-specific mechanism of tumor formation (e.g., by specific metabolic pathways) irrelevant for humans. Carcinogenicity evaluations by the International Agency for Research on Cancer (IARC) serve as the international benchmark for classifying chemicals as carcinogens. IARC assesses and classifies chemicals according to the following scheme (Illing 2001; IARC 2006): • Group 1. The agent (mixture) is carcinogenic to humans. The exposure circumstance entails exposures that are carcinogenic to humans. • Group 2A. The agent (mixture) is probably carcinogenic to humans. The exposure circumstance entails exposures that are probably carcinogenic to humans. • Group 2B. The agent (mixture) is possibly carcinogenic to humans. The exposure circumstance entails exposures that are possibly carcinogenic to humans. • Group 3. The agent (mixture or exposure circumstance) is not classifiable as to its carcinogenicity to humans. • Group 4. The agent (mixture) is probably not carcinogenic to humans. A list of chemicals assessed by IARC is available at the following URL: . In the United States, the two most prominent authorities for classifying the carcinogenicity of chemicals are the U.S. EPA and the U.S. National Toxicology Program (NTP). The weight-of-evidence descriptors used in the EPA’s Guidelines for Carcinogen Risk Assessment are based on the following (EPA 2005): • Carcinogenic to Humans. Strong evidence of human carcinogenicity when, for example, there is convincing epidemiologic evidence of a causal association between exposure and cancer.

40


• Likely to Be Carcinogenic to Humans. The weight of evidence is adequate to demonstrate carcinogenic potential to humans but does not reach the weight of the above descriptor. The use of the term “likely” as a weight of evidence descriptor does not correspond to a quantifiable probability. • Suggestive Evidence of Carcinogenic Potential. A concern for potential carcinogenic effects in humans is raised, but the data are judged not sufficient for a stronger conclusion. This descriptor covers a spectrum of evidence associated with varying levels of concern for carcinogenicity, ranging from a positive cancer result in the only study on an agent to a single positive cancer result in an extensive database that includes negative studies in other species. • Inadequate Information to Assess Carcinogenic Potential. This descriptor is appropriate when available data are judged inadequate, lack pertinent information, or provide conflicting evidence, for applying one of the above descriptors. • Not Likely to Be Carcinogenic to Humans. The available data are considered for deciding there is no basis for human hazard concern. In some instances, there can be positive results in experimental animals when there is strong, consistent evidence that each mode of action in experimental animals does not operate in humans. In other cases, there can be convincing evidence in both humans and animals that the agent is not carcinogenic. On a biennial basis, the U.S. NTP issues the congressionally mandated Report on Carcinogens (RoC). This document provides a list of known human carcinogens and substances that are reasonably anticipated to be human carcinogens, along with a brief profile for each substance. A summary of the classification criteria used in the 11th RoC is provided below (NTP 2008): • Known to Be Carcinogenic to Humans. Sufficient evidence of carcinogenicity from studies in humans which indicates a causal relationship between exposure to the agent, substance, or mixture, and human cancer. • Reasonably Anticipated to Be a Human Carcinogen. This designation may be made based on either (1) limited evidence of carcinogenicity from studies in humans, (2) sufficient evidence of carcinogenicity from studies in experimental animals, or (3) less than sufficient evidence of carcinogenicity in humans or laboratory animals. 3.2.1.2. Current Principles of Carcinogenic Risk Assessment. Weightof-evidence-based systems which classify carcinogenic hazards are part of, but do not substitute for, the risk assessment process (Di Marco et al. 1998). Carcinogen risk assessment is based on an evaluation of appropriate toxicological and exposure data sets, which should meet certain criteria for data quality and relevance (ECHA 2008a; EPA 2003; Klimisch et al. 1997). However, national policy frameworks can differ to the extent that risk assessment outcomes may be quite different for the same chemical(s). As discussed in Chapter 2, differences in science policy have been greater for cancer risk assessment compared to other toxic endpoints, with a


41

tendency to differentiate cancer risk assessment on the basis of presumed mechanism (i.e., genotoxic or nongenotoxic) and relevance to humans (some carcinogenic responses in animals may be considered irrelevant to human risk assessment) (Di Marco et al. 1998; EPA 1991, 1998; IARC 1999). Historically, risk assessment for noncancer endpoints has been based on the identification of a “no observed adverse effect level” (NOAEL) from a toxicity study with an animal model. The NOAEL is then divided by appropriate uncertainty factors to take potential inter- and intraspecies differences in response into account. However, this approach does not take into account the size of the toxicity study or the shape of the dose–response curve. The benchmark dose (BMD) approach has been suggested as an alternative to a NOAEL (Crump 1984). A BMD is a dose or concentration that produces a predetermined change (e.g., 10% or 1 standard deviation) in response rate of an adverse effect (called the benchmark response or BMR). A BMDL is the statistical lower confidence limit on the dose or concentration at the BMD. The BMD and BMDL are calculated using mathematical dose–response models, which make appropriate use of sample size and the shape of the dose– response curve (EPA 2009b, 2000a). The BMDL is like a NOAEL (i.e., as a point of departure) and is divided by an appropriate composite uncertainty factor to derive a reference value. The European Union (Commission Directive 93/67/EEC, Article 3, paragraph 1; repealed) and WHO (1994) have used the NOAEL/uncertainty factor approach for nongenotoxic carcinogens that are believed to have an effect threshold (WHO 1994). For genotoxic carcinogens, however, the regulatory default is applied that is based on the assumption that if “one hit” could cause a mutation and eventually result in cancer, then any exposure level could be associated with a finite cancer probability. Under such circumstances, a mathematical model (that quantitatively describes the relation between dose [exposure] and cancer [probability]) would be required to determine a “virtually” safe dose (VSD), a dose associated with an insignificantly small cancer risk. The choice of the model has an impact on risk predictions, because it usually involves extrapolation to low doses for which no data may be available, and has remained controversial. The U.S. EPA applies an alternative dose–response evaluation of carcinogens using a low-dose, linear model (EPA 2005). The linear extrapolation is applied under two circumstances: (1) when there are data to indicate that the dose-response curve has a linear component below the point of departure or (2) as a default for a tumor site where the mode of action is not established. For a linear extrapolation, a straight line is drawn from the point of departure to the origin. The slope of the line, known as the slope factor, is an upper-bound estimate of risk per increment of dose that can be used to estimate risk probabilities for different exposure levels. The slope factor is equal to 0.01/LED01, for example, if the LED01 is used as the point of departure. The lower limit on effective dose01 (LED01) is the 95% lower confidence limit of the dose of a chemical needed to produce an adverse effect in 1% of those exposed to the chemical, relative to control. If, however, there are sufficient data to ascertain that a chemical’s mode of action supports modeling at low doses, a reference dose or concentration may be developed in lieu of a cancer slope factor.

42


3.3. REGULATORY SCHEMES FOR INDUSTRIAL CHEMICALS AND BIOCIDES Many countries have schemes requiring new chemical substances to be notified to the national regulatory authority with a standard set of hazardous properties data and an assessment of the hazardous properties. This enables the risks to humans and the environment to be assessed with a view to deciding which RMMs are necessary. There are national inventories of existing substances that can be supplied without notification, by definition, all others are new. There are nevertheless schemes, both national and international, to evaluate high production volume (HPV) existing substances [reviewed in Knight and Thomas (2003)]. For example, the U.S. EPA and the OECD evaluate HPV chemicals produced in amounts greater than or equal to one million pounds per year or 2.2 million pounds per year, respectively. More recently, voluntary and mandatory chemical initiatives have been proposed or implemented that evaluate lower volume chemicals—that is, those produced in amounts greater than 25,000 pounds per year (EPA 2008a) or ∼2200 pounds per year (EC 2006). The discussion that follows provides an overview of laws in the United States and Europe for regulating industrial chemicals, followed by a discussion of voluntary initiatives for evaluating medium- and high-production-volume industrial chemicals. Thereafter, the laws for regulating biocides in the United States and Europe are discussed.

3.3.1.

The U.S. Toxic Substances Control Act (TSCA)

The U.S. Toxic Substances Control Act (TSCA) of 1976 (15 U.S.C. 2601 et seq.) provides the U.S. EPA with the authority to regulate industrial chemicals and mixtures (TSCA 1976). Section 2(b)(1) of TSCA states that it is the policy of the United States that “adequate data should be developed with respect to the effect of chemical substances and mixtures on health and the environment and that the development of such data should be the responsibility of those who manufacture [which is defined by statute to include import] and those who process such chemical substances and mixtures [.]” The core sections of TSCA that provide authority for implementing the above policy are discussed below. Prior to regulating chemicals under TSCA, it was foreseen that the chemicals in commerce would have to be known. Section 8(b) of TSCA addresses this need and grants the U.S. EPA authority to: “…compile, keep current, and publish a list of each chemical substance which is manufactured or processed in the United States.” TSCA Section 3(2)(A) defines a “chemical substance” as “… any organic or inorganic substance of a particular molecular entity, including (i) any combination of such substances occurring in whole or in part as a result of a chemical reaction or occurring in nature and (ii) any element or uncombined radical.” Foods, drugs, cosmetics, tobacco and tobacco products, radioactive materials, and pesticides are generally excluded, under Section 3(2)(B)(i)–(vi) of TSCA. Chemical substances not included on the TSCA inventory are considered new chemical substances, with some exemptions, and require compliance with premanufacturing notification requirements set forth under Section 5(a)(1)(A) of

3.3. REGULATORY SCHEMES FOR INDUSTRIAL CHEMICALS AND BIOCIDES

43

TSCA. The U.S. EPA may also promulgate rules after determining that a “significant new use” is proposed for a chemical on the TSCA inventory, under Section 5(a)(1) (B) and Section 5(a)(2). The U.S. EPA requires the submission of EPA Form 771025 (Rev. 5-95) for premanufacturing notices, significant new use notices, and any applicable exemptions at least 90 days before manufacturing or processing the new chemical substance. Along with this form, the submitter is “… required to submit all test data in [the submitter ’s] possession or control and to provide a description of all other data known to or reasonably ascertainable by [the submitter], if these data are related to the health and environmental effects on the manufacture, processing, distribution in commerce, use, or disposal of the new chemical substance.” Several possible outcomes may occur during the U.S. EPA’s 90-day review period and include the following: (1) No additional information is requested, (2) additional information may be requested, or (3) an administrative order may be issued that prohibits or limits the manufacture, processing, distribution, or disposal of a substance, pending the development of information. If a notice has been completed and the submitter has commenced commercial manufacture, the submitter is required to submit a Notice of Commencement of Manufacture or Import (EPA Form 7710-56 (8-95)) within 30 calendar days of the date the substance is first manufactured or imported for commercial purposes. Continued reporting requirements are placed on persons that manufacture, process, or distribute in commerce any chemical substance or mixture and include: (1) maintaining records of significant adverse reactions to health or the environment, alleged to have been caused by the substance or mixture (Section 8(c) of TSCA) and (2) immediately informing the U.S. EPA of “… information which reasonably supports the conclusion that such chemical substance or mixture presents a substantial risk of injury to health or the environment,” unless the person has actual knowledge that the U.S. EPA has been adequately informed (Section 8(e) of TSCA). Section 4(a)(1) of TSCA mandates that the U.S. EPA require by rule that manufacturers and/or processors of new or existing chemicals substances and mixtures conduct testing if the Administrator of the U.S. EPA finds that “[t]he manufacture, distribution in commerce, processing, use, or disposal of a chemical substance or mixture, or any combination of such activities, may present an unreasonable risk of injury to health or the environment [emphasis added].” A TSCA Section 4 test rule may require manufacturers and processors to conduct testing on environmental fate, ecotoxicity, acute toxicity, genetic toxicity, repeated dose toxicity, or developmental and reproductive toxicity. When a statutory finding under TSCA Section 4(a)(1) (i.e., “may present an unreasonable risk of injury to health or the environment”) cannot readily be made, the U.S. EPA may request that the TSCA Interagency Testing Committee (ITC), an independent advisory committee to the Administrator of the U.S. EPA, add the chemical to the TSCA Section 4(e) Priority Testing List (PTL). Once a chemical is added to the PTL, the U.S. EPA must promulgate a TSCA Section 8(a) Preliminary Assessment and Information Reporting (PAIR) rule and a TSCA Section 8(d) Health and Safety Data Reporting (HaSDR) rule. Section 8(a) PAIR rules require producers and importers to submit to the U.S. EPA one-time reports on production/importation volumes, end uses, and exposure-related data for the listed chemicals. Section 8(d)

44


HaSDR rules require producers and importers to submit to the U.S. EPA copies and lists of certain types of unpublished health and safety studies for the listed chemicals. Submitters under the HaSDR rule are also requested to provide robust summaries of health and environmental effects studies. In addition, when the ITC designates chemicals for testing, the U.S. EPA is required to initiate proceeding under a TSCA Section 4(a) test rule, if the PAIR and HaSDR data trigger a finding of unreasonable risk of injury to health or the environment.

3.3.2. The EU Registration, Evaluation, Authorisation, and Restriction of Chemicals (REACH) EU chemical control legislation has recently (2007–2008) been revised under the new scheme for Registration, Evaluation, Authorisation, and Restriction of Chemicals (REACH) (EC 2006). REACH places a duty on companies that produce, import, and use chemicals to assess the risks arising from their use (with new studies in justified cases) and take the necessary RMMs. The burden of proof for putting safe chemicals on the market has been transferred from the regulators to industry. Animal testing data must be shared to avoid duplication. Registration of information on the properties, uses, and safe handling of chemical substances will be an integral part of the system. A phase-in system lasting up to 11 years is planned for existing chemicals, known as “phase-in substances.” Higher tonnage substances (≥1000 tonnes per annum [t/a]), as well as lower tonnage substances that are very toxic to the aquatic environment (≥100 t/a) or classified as CMRs (i.e., carcinogens, mutagens, or reproductive toxicants; ≥1 t/a), will require more data and have to be registered by November 30, 2010. All other phase-in substances manufactured or imported in quantities ≥100 t/a or ≥1 t/a must be registered by May 31, 2013 or May 31, 2018, respectively. New chemicals or “nonphase-in substances” will be evaluated on an ongoing basis and must be registered before being manufactured or imported into the EU. Under REACH, all substances manufactured in or imported into the EU at ≥ 1 t/a must be registered with the European Chemicals Agency (ECHA). A Chemical Safety Report (CSR) is required for substances registered at 10 t/a unless the substance is present only in a preparation at below 0.1% or below the concentration limit(s) triggering classification of the preparation as dangerous. A CSR is a risk assessment that must follow (a) the general provisions of Annex I of REACH and (b) the ECHA’s guidance for writing a CSR (ECHA 2008c). Substances of very high concern classified as category 1 or 2 CMRs are amongst the substances subject to tighter controls, including an authorization regime, along with persistent, bioaccumulative, and toxic substances (PBTs) and very persistent and very bioaccumulative substances (vPvBs). PBTs and vPvBs are classified as such based on the criteria set forth in Annex XIII of REACH. When substances are classified as CMR, PBT, or vPvB, they shall appear on the list of substances subject to authorization under Annex XIV of REACH. Other substances of concern, such as endocrine disrupters, will also be added to this list (Annex XIV) on an ad hoc basis. Substances subject to authorization will have to be approved for a specific use, with decisions based on a risk assessment and consideration of socioeconomic factors. For existing substances in Annex XIV of REACH, a “Sunset Date” will be


45

set after which the substance may no longer be used or sold, unless an application is submitted to and subsequently approved by ECHA. The classification of a substance as a CMR, PBT, or vPvB will be a factor in deciding what further higher-tier testing is required (e.g., Sections 8 and 9 of Annexes VIII, IX, and X), as will be the mandatory performance of an exposure assessment and subsequent risk characterization (ECHA 2008b). Most of the benefits expected from REACH are based on the expected significant increase in cancer prevention. Postle et al. (2003) estimated that the economic benefits of preventing between 2167 and 4333 deaths per annum due to cancer over a 30-year period will be between œ18 billion and œ54 billion; by comparison, the benefits of preventing deaths from all other diseases combined were between œ23 million and œ225 million (Postle et al. 2003). Clearly, carcinogenicity is of paramount importance within regulatory frameworks. Due to the importance of the EU as a trade bloc, it is expected that other jurisdictions will adapt their legislation to be compatible.

3.3.3. Voluntary Initiatives for Evaluating Industrial Chemicals 3.3.3.1. The U.S. EPA’s Former Chemical Assessment and Management Program (ChAMP). In August 2007, Canada, Mexico, and the United States committed, under the Security and Prosperity Partnership (SPP), to accelerate and strengthen national and regional risk-based assessment and management of chemicals. ChAMP was the name given by the U.S. EPA to identify its efforts to meet the SPP commitments. By 2012, the U.S. EPA, under ChAMP, planned to assess and prepare screening-level characterizations of hazard, exposure, and risk, and to use this information to develop initial risk-based prioritizations (RBPs). The U.S. EPA planned to evaluate over 4000 organic Medium-Production Volume (MPV) chemicals produced at volumes greater than 25,000 pounds per year, but less than one million pounds per year. Health and environmental hazard and environmental fate characterizations would be informed by existing data, Canadian categorization results, U.S. EPA Structure Activity analysis input, and knowledge gained under the U.S. EPA’s HPV Chemical Challenge. The U.S. EPA planned to also evaluate 2750 organic HPV chemicals produced at or above one million pounds per year. HPV challenge submissions were to provide the base hazard data for evaluations under ChAMP, whereas the 2006 Inventory Update Reporting (IUR), under TSCA, was to provide the use and exposure information. Beyond organic MPVs and select HPVs, the U.S. EPA planned to assess ∼750 inorganic HPV chemicals, which were first reported under the 2006 IUR cycle. The general voluntary approach used in the HPV Chemical Challenge was expected to be used (see Section 3.3.2, e.g., sponsorship commitments, development of test plans, public review step, completion of data package, and submission to EPA). The OECD’s inorganic HPV guidance would serve as a benchmark for preparing submissions. During the preparations for implementing ChAMP, the U.S. EPA was in the process of “resetting” the TSCA Inventory (EPA 2008b). The original inventory was compiled in 1979 and consisted of 62,000 chemicals. Since then, ∼21,000 new chemicals have been added to the TSCA inventory. Though the U.S. EPA plans to

46


reset the TSCA Inventory in order to better understand the universe of chemicals actually in commerce, ChAMP was superseded on September 29, 2009, with a regulatory management approach due to concerns about the sufficiency of data from which the U.S. EPA could evaluate these chemicals. This determination may eventually lead to changes in TSCA similar to the approach for chemical management under REACH (i.e., no data, no market, tiered testing, etc). 3.3.3.2. The U.S. EPA’s High Production Volume (HPV) Chemicals Challenge. In 1998, the U.S. EPA launched the voluntary High Production Volume (HPV) Chemicals Challenge Program. This initiative was created to ensure the public availability of a baseline set of data on over 2800 HPV chemicals, which are freely accessible at the following URL: http://iaspub.epa.gov/oppthpv/ public_search.html_page. The U.S. EPA defined HPV chemicals as those being manufactured or imported in amounts greater than or equal to one million pounds per year, based on volumes reported under the TSCA, 1990 IUR. The initiative was called a “challenge” because the U.S. EPA challenged U.S. manufacturers and importers of HPV chemicals to voluntarily sponsor chemicals under the program. The data sought under this program were based on internationally agreed-upon test data known as the Screening Information Data Set (SIDS), as developed by the OECD. SIDS data sets enable regulators to assess human and environmental hazards and are intended to provide enough information to assign a priority for further work, if necessary. These data include the following: acute toxicity; repeated dose toxicity; developmental and reproductive toxicity; mutagenicity (gene mutation and chromosomal aberration/damage assays); ecotoxicity (studies in fish, invertebrates, and algae); and environmental fate [including physical/chemical properties (melting point, boiling point, vapor pressure, n-octanol/water partition coefficient, and water solubility), photolysis, hydrolysis, transport/distribution, and biodegradation]. The U.S. EPA has issued several guidance documents, which aid sponsors with preparing submissions (EPA 2009a). As part of the commitment under the U.S. EPA’s program, sponsors submit data summaries of existing data, along with a test plan that proposes a testing strategy to fill data gaps. Once submitted to the U.S. EPA, the documents are posted on a public database and a 120-day comment period begins whereby all stakeholders (e.g., the U.S. EPA, industry, environmental protection groups, animal welfare groups, private citizens, etc.) have an opportunity to provide input. Comments are intended to provide feedback, which may be used to revise test plans and data summaries. In the event that an HPV chemical lacks necessary testing data, the U.S. EPA may, at their discretion, issue a test rule to obtain the required data. 3.3.3.3. The OECD’s Work on Investigation of HPV Chemicals. The OECD defines HPV chemicals as those manufactured or imported in quantities greater than or equal to 2.2 million pounds per year in at least one member country or an EU region. In an effort to undertake the investigation of HPV chemicals in a cooperative manner, the OECD modified its work on HPV chemicals through an OECD Council Decision in 1991 (OECD 1991). The OECD’s cooperative approach includes involvement from member countries in four basic components: (1) selection of chemicals to be evaluated, (2) collection of data from governments, public


47

sources, and industry, (3) completion of the SIDS dossier, and (4) assessment of the potential hazards for the chemical investigated. The current OECD HPV chemicals list includes 4853 substances (OECD 2004). The status of each chemical in the process may be monitored via the OECD’s publicly available HPV database . The OECD’s program enables a member country to sponsor the HPV chemicals produced by industries within its borders and, in turn, to benefit through data sharing from the sponsorship of other countries. This process eliminates duplicative testing. The OECD has developed extensive guidance, which aids sponsors by (1) following procedures to improve the efficiency of investigating HPV chemicals, (2) data gathering and testing for the SIDS, (3) evaluating the quality of data in the SIDS dossier, (4) assessing the initial hazards of chemicals, (5) preparing the SIDS initial assessment report and SIDS profile, and (6) overseeing any additional postSIDS activities (OECD 2007). 3.3.3.4. The U.S. EPA’s Voluntary Children’s Chemical Evaluation Program (VCCEP). On December 26, 2000, the U.S. EPA announced the Voluntary Children’s Chemical Evaluation Program (VCCEP) and requested sponsorship commitments from manufacturers or importers for 23 chemicals. The VCCEP was designed to provide data to enable the public to better understand the potential health risks to children associated with certain chemical exposures. In support of this pilot program, the U.S. EPA designed a tiered-testing approach, which consisted of the following: Tier 1: acute toxicity, in vitro gene mutation, combined repeated dose toxicity with reproductive and developmental toxicity screens or repeated dose oral toxicity and reproductive toxicity (one generation); Tier 2: 90-day subchronic toxicity in rodents; reproduction and fertility effects; prenatal developmental toxicity (two species); in vivo mammalian bone marrow chromosomal aberrations or in vivo mammalian erythrocyte micronucleus (triggered off results from in vitro mammalian chromosomal aberration test if conducted in tier 1); immunotoxicity; metabolism and pharmacokinetics; Tier 3: carcinogenicity or chronic toxicity/carcinogenicity; neurotoxicity screening battery; developmental neurotoxicity (EPA 2000b). The U.S. EPA chose these studies based on their appropriateness for evaluating the toxicity of chemicals to which children have significant potential for exposure. The VCCEP process consists of following basic steps, outlined below (EPA 2000b): Step 1: Chemical Selection. After receiving comments from various stakeholders, the U.S. EPA selected chemicals judged by the U.S. EPA to present the relatively greatest potential for exposures that may impact children. Step 2: Tier 1 Commitment. A manufacturer or importer of a VCCEP chemical submits a letter to the U.S. EPA indicating their commitment to sponsoring a chemical. Step 3: Submission of Tier 1 Data. A VCCEP chemical sponsor submits to the U.S. EPA a Tier 1 Hazard Assessment, a Tier 1 Exposure Assessment, a Tier 1 Risk Assessment, and a Data Needs Assessment. Step 4: Peer Consultation Regarding Tier 2 Data Needs. At the U.S. EPA’s request, a third-party contractor convenes a Peer Consultation to evaluate

48


whether Tier 1 data needs were met, whether the Tier 1 submission fully characterized the chemical’s potential risk to children, and whether there are remaining Tier 2 data needs. The results and comments from the Peer Consultation are compiled by the third-party contractor and submitted to the U.S. EPA. A possible conclusion is that no more work is needed. Step 5: U.S. EPA Review of Peer Consultation. The U.S. EPA reviews the sponsor ’s submission and the third-party contract report and announces a the Tier 2 Data Needs Decision. If the U.S. EPA disagrees with the conclusions from the third-party peer consultation report, sponsors and other stakeholders are given 60 days to comment on the U.S. EPA’s Tier 2 Data Needs Decision. Following a review of comments, the U.S. EPA mails its final decision to the sponsor and posts the decision on the VCCEP website. If the U.S. EPA requires further testing under Tier 2 or Tier 3, steps 2 through 5, above, are repeated with consideration of the appropriate tier ’s testing requirements. If a chemical is recommended for Tier 2 or Tier 3 testing, but is not sponsored by a manufacturer or importer of the chemical, the U.S. EPA may require the data by issuing a TSCA Section 4 test rule. A summary of the submissions on VCCEP chemicals is available at the following URL: http://www.epa.gov/oppt/vccep/pubs/ chemmain.html.

3.3.4. The U.S. Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) The U.S. Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) (7 U.S.C. 136 et seq.) grants the U.S. EPA the authority to regulate the registration, distribution, sale, and use of pesticides (FIFRA 1972). Under Section 2(u), FIFRA defines a pesticide as “… any substance or mixture of substances intended for preventing, destroying, repelling, or mitigating any pest ….” Biocides are included under a separate definition for antimicrobial pesticides, which are defined under Section 2(mm) as pesticides intended to “(i) disinfect, sanitize, reduce, or mitigate growth or development of microbiological organisms; or (ii) protect inanimate objects, industrial processes or systems, surfaces, water, or other chemical substances from contamination, fouling, or deterioration caused by bacteria, viruses, fungi, protozoa, algae, or slime ….” Under FIFRA Section 3, every pesticide product must be registered with the U.S. EPA or specifically exempted under FIFRA Section 25(b) before being sold or distributed in the United States. An applicant for a new registration or an existing registrant must demonstrate to the U.S. EPA’s satisfaction that, among other things, the pesticide product, when used in accordance with widespread and commonly recognized practice, will not cause “unreasonable adverse effects” to humans or the environment. This safety determination requires the U.S. EPA to weigh the risks of the use of the pesticide against any benefits. A general overview of the core provisions of FIFRA that aid with the U.S. EPA’s weighing of risks versus benefits is provided below, along with provisions specific to antimicrobial pesticides. Under Section 3(c)(2) of FIFRA, the U.S. EPA is granted broad authority to require scientific testing and submission of the resulting data to the U.S. EPA by


49

applicants for registration of pesticide products. An applicant must furnish the U.S. EPA with substantial amounts of data on the pesticide, its composition, toxicity, potential for human exposure, environmental properties, and ecological effects, as well as information on its product performance in certain cases. Section 3(c)(2)(B) authorizes the U.S. EPA to require a registrant to develop and submit additional data to maintain a registration. However, Section 3(c)(2) does not require the U.S. EPA to develop data requirements for an “antimicrobial pesticide” as defined under Section 2(mm). Though Section 3(h) describes the registration requirements for antimicrobial pesticides, the scope is limited to requirements for process improvements, timeframes for review purposes, and other regulatory matters, but does not contain provisions for data requirements. Under Title 40 of the Code of Federal Regulations (CFR) Part 158 et seq., the U.S. EPA’s final rules on the data requirements for conventional pesticides (EPA 2007b), biochemical pesticides (Subpart U) (EPA 2007a), and microbial pesticides (Subpart V) (EPA 2007a) are listed. As part of those rules, the U.S. EPA preserved the original Data Requirements for Pesticides to provide continued regulatory coverage for antimicrobial pesticides and transferred the original 1984 data requirements of Part 158 into a new Part 161 titled “Data Requirements for Antimicrobial Pesticides” (EPA 2007c). Part 161 contains the current data requirements for antimicrobial pesticide chemicals, although it is intended to be replaced with Subpart W of Part 158 once the U.S. EPA issues a final rule. This specific action for antimicrobial pesticides is in process because the U.S. EPA determined that the original data requirements of 1984 failed to adequately address the unique applications, use patterns, and other factors germane to antimicrobial pesticides. Once a product is approved for registration, the registrant is required, under Section 6(a)(2), to inform the U.S. EPA if the registrant obtains “… additional factual information regarding unreasonable adverse effects on the environment of the pesticide ….” Section 2(bb) of FIFRA defines “unreasonable adverse effects on the environment” to include unreasonable risk to humans. Section 4 of FIFRA requires that the U.S. EPA reregister each pesticide that the U.S. EPA first registered before November 1984. This date was chosen because pesticides registered after 1984 were subject to the Part 158 testing requirements. Section 3(g) of FIFRA also requires the U.S. EPA to periodically review the registrations of all pesticides due to changes in science, public policy, and pesticide use practices, which occur over time. The U.S. EPA promulgated a new registration review program in 2006, as detailed in 40 CFR Part 155, Subpart C, which began to replace the U.S. EPA’s reregistration program as the mechanism for systematic review of existing pesticides.

3.3.5.

The EU Biocidal Products Directive (BPD)

Biocide active substances or products are generally exempt from REACH evaluation and authorization procedures because they are regulated in the EU under the Biocidal Products Directive (BPD) (EC 1998). This Directive has an established process for evaluation of active substances (for listing in Annex I of the BPD), followed by national approvals of the formulated biocidal products containing them. However,

50


REACH substance evaluation can apply to the active substance if it has been prioritized and placed on the Community Rolling Action Plan, and active substances used in biocidal products may be included in Annex XIV of REACH if they are classified as CMR, PBT, vPvB, or endocrine disrupters. The dossier for an active substance includes information on the applicant, the identity of the active substance and biocidal product, their physical and chemical properties, methods of detection and identification, effectiveness against target organisms, intended uses, their toxicological profiles, their ecotoxicological and environmental fate and behavior properties, measures necessary to protect humans, animals and the environment, EU classification and labeling, and an overall summary and evaluation. The study reports are rewritten as Robust Summaries, and any deviations from the standard methods have to be explained and justified; there are data waiver forms to justify omitting the standard studies. The common core data set for active substances is specified in Annex IIA of the BPD. Additional data selected from Annex IIIA of the BPD are needed for particular product types, as specified in the data requirements Technical Guidance Document (TGD) (EC 1993) to reflect particular exposures, in order to conduct an adequate risk assessment. Furthermore, additional studies, not necessarily restricted to those listed in Annex IIIA of the BPD, may be needed to investigate further ambiguous findings from the standard data set or as an outcome of the risk assessment. Similarly, the common core data set for biocidal products is specified in Annex IIB of the BPD with additional data selected from Annex IIIB for particular product types according to the TGD. Risk assessment is a key part of the EU approval process for Annex I of the BPD listing of active substances and national authorization of biocidal products. The EU risk assessment procedures for biocides are the same as those used for chemical substances. Under the BPD criteria for Annex I inclusion, an active substance may not be authorized for use by the general public if it is classified as a category 1 or 2 carcinogen. In addition, professional use may only be authorized if exposure to humans is unlikely or exposure is below the threshold for the effect.

3.4. SCIENTIFIC ASPECTS OF CARCINOGENIC RISK ASSESSMENT 3.4.1. Dose—Response Relationships in Carcinogenesis and Mechanisms of Carcinogenic Action Risk assessment frequently involves estimating safe exposure concentrations for exposure durations that were not tested experimentally. Generally applicable biologically based models have to be applied. Before developing such a model, extensive data are needed to build its form as well as to estimate how well it conforms to the observed data to support confidence in results. The first benchmark study of dose–response relationships in chemical carcinogenesis was reported by Druckrey (1943) with 4-dimethylaminoazobenzene (4-DAB), also known as “butter yellow,” in BDIII rats. Within the range of daily dosages from 3 to 30 mg per rat, the time up to the appearance of liver cancer (t) was found to be inversely proportional to the daily dose (d). The product of the

51

3.4. SCIENTIFIC ASPECTS OF CARCINOGENIC RISK ASSESSMENT

TABLE 3.1.

Daily Dose, d (mg/rat) 30 20 10 5 3

Induction of Liver Cancer in BDIII Rats by 4-DAB

Median Tumor Induction Time, t (days)

Total Dose, D (mg/rat)

34 52 95 190 350

1020 1040 950 950 1050

daily dosage and the median tumor induction time, which corresponds to the sum of all daily doses—that is, to the total dose, D—was found to be practically constant (Table 3.1): dt = D ~ 1000 mg = constant

(3.1)

Assuming a linear relationship between the daily dosage (d) and the 4-DAB concentration (c) at the site of carcinogenic action, Eq. (3.1) would read as ct = constant

(3.2)

Equation (3.2), that the product of exposure concentration and duration produces a constant toxic effect, is known as Haber ’s rule (Haber 1924), named after the German chemist Fritz Haber, who in the early 1900s characterized the acute toxicity of gases used in chemical warfare. The smaller the effects (Haber) product, c × t, the greater the toxicity of the gas. However, not all gases have a constant effects product. Flury (1921) pointed out that the effects product for hydrocyanic acid (HCA) does not remain a constant, but instead increases with decreasing concentrations of the toxicant in the inspired air. Apparently, agents such as HCA, which are toxic only after their resorption, are better tolerated as the concentration at which they are inhaled becomes smaller, suggesting that detoxification processes are more efficient at low concentrations than they are at higher concentrations. Flury introduced a constant detoxification factor (e) in Haber ’s rule, which appeared sufficient to describe the observations made:

(c − e) t = constant

(3.3)

Clark (1937) further expanded Haber ’s rule for the action of a number of drugs:

(c − cm ) (t − tm ) = constant

(3.4)

where cm is a threshold concentration and tm is a minimum time of response. Clark commented at the time (Clark 1937): The formula ct = constant is indeed an impossible one in the case of drugs acting on biological material because it implies that an infinitely small concentration of a drug will produce the selected action in infinite time, and conversely that a sufficiently high concentration will produce an instantaneous effect. In some cases ct = constant gives an approximate fit, but this merely implies that cm and tm are so small as not to produce a measurable error.

52


So, the perfect fit of Haber ’s rule to the carcinogenic action of 4-DAB suggested that threshold concentration (cm) and minimum time of response (tm) were (in Clark’s words) so small as not to produce a measurable error. Druckrey and Küpfmüller (1948) provided a theoretical explanation for ct = constant. By denoting the initial concentration of specific receptors that 4-DAB reacts with as R, the concentration of receptors that 4-DAB has reacted with as CR, and the mean 4-DAB concentration at the site of action as C, the reaction kinetics in the case of a bimolecular reaction are dC R dt = K ( R − C R ) C − C R TR

(3.5)

where K is the reaction constant for association and TR is the time constant for dissociation. Druckrey and Küpfmüller then inferred that their experiment had shown that the carcinogenic action of 4-DAB was irreversible, and that as TR → ∞ we obtain dC R dt = K ( R − C R ) C

(3.6)

Now, assuming that up to the time of action we have CR 108 potential folded-sequence specific 3-D variations. This calculation does not include the many combinations also contributed by DNA 2° structures, such as local hairpin loops, and long-range DNA–DNA associations even across chromosomes (Clark 2007). The enormous number of total chromatin potential structures is of great interest because such immense structural diversity offers significantly more information than the 1° (primary sequence) of DNA, which heretofore has been the sole rationale for the genetic code and hence for gene expression. Many of these combinations may give rise to explaining the “if,” “when,” “why,” “how long,” and the timely termination of gene expression “not” (Biel et al. 2005; Wang et al. 2008). Not all chromatin structural combinations will necessarily be useful chemically, but the functional structures that do possess utility give rise to The Histone Code (Jenuwein and Allis 2001; Wynter 2006). Chromatin functional arrangements, selected from such a large number of combinatorial structures, determine the cell’s biological cybernetic vocabulary, maintains cellular memory, and gives rise to specialized transcription events in a timely manner required in physiology. Functional chromatin loci are most often associated with specific attractors: the situation-specific and often environmentally controlled specific regulatory transcription factor proteins and their co-factors (NAD+, GTP, etc.).

144

CHAPTER 5 THE INTERPLAY OF CANCER AND BIOLOGY

Another often-used epigenetic change is DNA methylation modification of cytosine bases at the 5′ ring position with the methyl group donated by Sadenosylmethionine (SAM). 5′-Me-C sequences in DNA structures attract certain specific protein transcription marker factors and co-factors. Many DNA “marks” are maintained throughout the life of a cell and can be passed on to progeny cells via the ovum, which makes the methylation a fundamental part of inheritance (Jablonka and Lamb 2007; Szyf 2007). Methylated cytosines also occur in imprinted genes which are gender-based heritable genes. The methylated cytosines can occur anywhere in DNA where there is a CpG island, but a particular DNA locus affected is at the 5′ terminus of the transcription promoter region of genes (Holliday 1987). Methylated gene 5′-promoter regions usually restrict the gene activity, but these genes can be reactivated with the nuclear enzymes histone demethyltransferases (HDMT) that remove the methyl groups. These HDMT enzymes are highly conserved across the taxa and are critical in maintaining the degree of CpG island methylation on a long-term basis, and this is one way a cell acquires and maintains its identity. Histone acetyltransferases (HATs) and histone deacetylases (HDACs) are susceptible to toxic agents and are targets for drug therapy in a number of diverse diseases (Szyf 2007). Some genes can be turned “on” by cytosine methylations or they can be attenuated or turned “off” in a dynamic process during the cell cycle. Acetylation binding to histone R and K amino acids usually prompts heterochromatin unfolding that spatially allows DNA transcription. Histone deacetylation by HDACs acts oppositely by inducing heterochromatin formation. The organspecific propitious combinatorial binding of methyl and acetyl groups in histones contributes to activated chromatin for the proper DNA transcription (Wang et al. 2008). By chemical direction of specific 3° and 4° chromatin structures, the nucleus efficiently packs about 2 m of DNA (≈109 bases) into a cell nucleus volume whose diameter is about 10 μM. In condensed chromatin (heterochromatin), the nucleosomes are highly folded regions and the nuclear elements in the heterochromatin are turned off because the inaccessibility of the DNA and RNA transcribing machinery. In folded and ordered chromatin, interrelating 3-D functional sites can be created in an activation process (Schones and Zhao 2008). Sometimes the sites are conjoined by long-range chromatin interactions of intrachromasomal complimentary loci and even between chromosomes, all of which are created by physical and chemical associations (Clark 2007). Another fundamental set of changes is the histone acetylation at lysine (K) loci by HAT, which tends to form euchromatin; and the acetyl group can be removed by HDAC, which tends to form heterochromatin. These dynamic states are in functional equilibrium, depending on need. Only a small proportion, about 5–10% of the whole human genome, is stabily transcribed into RNA. Of this transcribed RNA, only about 1% of the genome codes for proteins whereas the remainder (∼9%) is high-molecular-weight RNA types that likely participate in cellular control functions (Ponting et al. 2009). It’s been experimentally observed that transcription occurs in euchromatin regions of chromatin where the chromatin is open for transcription and not condensed (Ting et al. 2006). Euchromatin fosters open reading frames for the (1) chromosomal active genes

5.3. CELL BIOLOGY OF CANCER

145

(mRNAs) and (2) control RNA regions—that is, micro-RNAs (miRNA), interfering RNA (RNAi), Piwi interacting RNA (piRNA), and long nonprotein coding RNA (lncRNA or ncRNA) (Amaral et al. 2008; Ponting et al. 2009). An operative notion of cell biology is arising: Selective production of RNAs is the cell’s main cybernetic set of tasks. It has been long recognized since the 1960s that nuclear RNA is initially synthesized, from the DNA template, as long RNA pieces of about 6500 nucleotides known as heterogeneous nuclear RNA (hnRNA). About 10% of the hnRNA is processed to about the size needed for an average protein coding ≈ 1500 nucleotides. Because the remaining RNA quickly associates with nuclear proteins (and possibly DNA), it may be assumed that some of the non-mRNA DNA sequences function to produce control RNAs and not just “noise” RNA (Ponting et al. 2009). The reading and nonreading chromatin regions move about during the cell cycle presumably according to need. This is called position effect variegation (PEV) (Biel et al. 2005). Genes that encode products which promote transcription are enhancers of PEV and are called e(var), of which HATs are representatives, a large enzyme family from yeasts to humans. The genes that suppress PEV (and thus transcription) are collectively called su(var), of which the histone linker H1 is one example and the highly conserved HDAC gene is another (Moss and Wallrath 2007). An objective case can be made for the existence of biological nongenetic cancer mechanisms. A list of 54 chemicals based on positive carcinogenesis, but no genetic positive findings, makes this case that 54 bioassayed carcinogens are not knowingly mutagenic in their cancer causality (Tennant 1993). It remains that there could be genetic causes that are not measured by current genotoxicity protocols. Some tests, like DNA recombination, have been recorded for some nongenetic carcinogens (Schiestl 1993). Clearly, better biomarkers for mutagenic and nonmutagenic carcinogens are needed to assess the hazard of potential carcinogens. The known nongenotoxic carcinogen list is large, and it is worth noting that some of the most potent carcinogens do not demonstrate mutagenicity. It can be surmised from this observation that some chemicals cause chromatin disturbances, not by classical genetic means but rather by disruption or disorganization of the epigenetic code (Feinberg et al. 2006; Jones and Baylin 2007). Another objective case for epigenetic mechanisms can be found in a carcinogenesis study, which evaluated Swedish, Danish, and Finnish identical twins (Lichtenstein et al. 2000). This epidemiological study made observations of cancer incidences at 11 major organ sites. It showed that identical twins that are separated at birth have vastly different cancer site incidences at these 11 sites that were consistent with their respective rearing environments during development and adulthood. Control twins who remained together in rearing were similar in site and frequency. Because identical twins have identical DNA sequences, gene expressions and/or utilization of DNA in resisting cancer must be different in the separated twins. That is, their respective gene expressions relate to their particular environmental factors that influence their gene expressions. Moreover, it’s a common observation that siblings of the same set of paternal and maternal genes set can be more different phenotypically than random crossing-over events might predict. Environmental conditions in utero and externally for the mother vary from pregnancy to pregnancy; it is widely suspected that epigenetic control of plasticity and reversibility has a role in this sibling variance. Long-term phenotypic traits

146


inherited in families or within tribes can become altered or even lost upon social migration to a different environment where new traits are acquired that better accommodate the exigencies and vagaries of the new environment (Maresca and Schwartz 2006). It is becoming clear that biologic controls used in the cell’s cybernetics are more than the Neurospora-based hypothesis of “one gene produces one mRNA, which in turn produces one protein” (Beadle and Tatum 1941; Holliday 2006; Jablonka and Lamb 2007; Morange 2002). The central dogma was advanced by Francis Crick in 1958 when he posited that the sequence information to make a cell (or the whole soma) moves in one direction from the nuclear DNA to RNA to protein, irreversibly. The dogma may need amending in light of genetic findings since 1958 and the many new epigenomic and intra- and extranuclear RNA controls of gene expression (Crick 1958; Hertel 2008; Morange 2008; Ting et al. 2005; Wolters and MacKeigan 2008; Wynter 2006). The original model has indeed served well in the past for experimentation but does not incorporate the reversible and specific adaptability seen in the rapid (sometimes within one generation) construction or restructuring of biological tissues and a means to negotiate with or be responsive to the ever-changing environments in which the organism exists. Darwinian adaptation by classical genetic mechanisms is just too slow to explain the rapid fixed changes that happen within one or two generations. DNA sequence random variability with Darwinian selection of the “fittest” simply cannot produce such rapid and sometimes altruistic adaptations. Nor does the dogma of Crick account for the all the nonvertical transmission of dynamic phenotypic trait changes including ontogeny (Jablonka 2004; Jablonka and Lamb 2007; Szyf 2007). The human genome project found about 20,000–25,000 transcribed DNA sequences as potential human genes (Stein 2004). This ab initio indicated too few genes to carry out all the known phenotypic tasks executed in and among human cells according to the one gene → one protein sequence model. Also, Drosophila melanogaster has about one-half of the amount of coding DNA that humans have. This implies that humans employ DNA more efficiently (or the fly is very inefficient) than flies because of additional structural and developmental complexity, as well as the degree of encephalization that humans possess. This disjunctive observation suggests that humans, albeit suppositionally, make greater combinatorial uses of human epigenetic processes within their episome. This ability is likely true for all higher evolved organisms. Higher organisms utilize an additional mechanism that includes post-transcription RNA splicing (cis-splicing to remove introns and trans-splicing to ligate various mRNAs), thus forming multiple recombinations into various mRNAs from the same DNA sequence (1∼10 mRNAs/gene) (Hertel 2008; Holliday and Murray 1994; Nigro et al. 1991; Shepard and Hertel 2008). The episome of higher evolved organisms is the basis for more information of higher utility being stored in well-ordered but versatile chromatin structures. There is no doubt that DNA contains the blueprints for complete corporal construction and is the cynosure of the cell’s information repository and is the genesis of all cellular construction. However, it is the machinations of the episome that executes much of the control of DNA usage and transactions—that is, (1) which specific genes are expressed in euchromatin (gene activation), (2) when genes are individually expressed but coordinately to achieve


147

a unitary purpose (synchronization), (3) how many times are genes are transcribed and protected from RNases (gene dosage), and (4) when and how genes can be attenuated, or shut “off,” or turned “off” permanently. All these actions are affected by the controls of genes. In summary, the DNA sequence proposes, the episome and its interactive alterations disposes. The aspect of how the episome, especially in the progression of cancer, links with the cell’s internal networks (Hahn and Weinberg 2002) and with the environment or external networks is a subject of systems biology currently in rapid development (Wang et al. 2007) and is covered elsewhere in this volume. Already this eclectic field promises to press the axiom that the whole is merely the sum of its parts, and it appears that interacting complex systems may generate new or varied traits.

5.3.9.

Biological Initiation of Chemical Carcinogenesis

Epigenetic changes are known to be involved in normal development, the evolutionary process, as well as with human diseases, and chemical-induced cellular changes (Holliday 1987, 1989, 2002; Feinberg 2007; Jablonka and Lamb 2002; Oliveira et al. 2007). Cancer development not only depends on the heretofore much published genetic alterations but also involves abnormal cellular memory, which maintains, executes, and passes genomic information to successor cells. As seen in the past, aberrant cell memory can be maintained by mutated DNA, but also altered cell memory can influence the highly articulated nucleosomal structures and their functions in the episome. Both altered memories are active in the carcinogenesis process during transmission of gene expressions, cytoskeletal patterns, and cell-tocell interactions in the neoplastic I stage and P stage (reviewed earlier in this chapter), malignant stage, then metastasis, and finally colonization at a distal site (see Figure 5.1; and also Bachman et al. 2006; Curtin et al. 2006; Feinberg 2005; Feinberg et al. 2006; Ohlsson et al. 2003; Vasiliev 2004; Verma et al. 2004). Recent data suggest that cancer is generated (i.e., “genesis of cancer” or carcinogenesis) by polyclonal disruption of stem cells and/or their progenitor cells that are found experimentally to be epigenically disrupted in specific tumor-progenitor genes in the stem cells of the niche interface to differentiating progenitor and mature cells [refer back to Figure 5.5; also see Feinberg (2004, 2005) and Jablonka (2006)]. Altered development leading to tumor cell heterogeneity has much to do with epigenetic variation among clones (cf. Section 5.3.4). Variation occurs in progenitor cells that devolve with time acting upon a select set of flawed I*-stage cells as well as other cells in the LTA. Plasticity is the normal ability of a tissue or organism to change and adapt, but epigenetic plasticity together with genetic lesions can drive tumor progression (Chan et al. 2008). The reversible early role for nongenetic epigenetic alterations in neoplasia is a necessary precancerous state that is a prelude to later epigenetic alterations that can substitute for, or cooperate with, induced crucial genetic variations in tumor progression (Figure 5.4). Therefore, non-neoplastic but epigenetically disrupted stem/progenitor cells might be a crucial target for cancer risk assessment, chemoprevention, and prophaxis (Feinberg et al. 2006; Marks et al. 2007; Tai et al. 2005; Trosko et al. 2004). Meaningful biomarkers for environmental detection could be derived from these disruptions, too.

148


These cell-altering epigenetic mechanisms occur in a number of ways that include the following: (1) In chemical carcinogenesis, causation of inappropriate levels or placements of histone methylations and acetylations allows restricted genes to be uncovered, or requires unrestricted genes to be restricted, which presents or removes in some cases an untimely or inappropriate set of activities; (2) chromatin epigenetic changes are also manifested in global hypomethylation in chemical carcinogenesis; (3) sometimes, however, in the LTA, chromosome sites exhibit hypermethylation usually located in transcription 5′-promoter regions; and (4) abnormal chromatin packaging or folding alterations can participate in the cancer process because structure and function are the yin and yang of cell being (Biel et al. 2005; Holliday 1987). It is increasingly apparent that cancer development depends not only on genetic alterations but also on abnormal cellular activities and abnormal cellular memory that dynamically interacts and fails to normally adapt with the environment (e.g., other like-cells; support cells such as stromal; external matrices; tissue fluids; etc.). These environments can influence horizontal and vertical heritable gene expression patterns critical for neoplastic initiation and progression (Scheel et al. 2007). In various diseases a unifying theme of epigenetics is the occurrence of defects in phenotypic plasticity—that is, an inability of the collective abilities of cells to adequately change their behaviors in response to internal or external environmental cues. A basic question in toxicological cancer prevention can be posed: What process starts this loss of homeostasis and initiates aberrant development so that the LTA, as an affected locus, disengages from its community and corporal controls? Over 80 years ago, Otto Warburg observed that cancer cells metabolized glucose (GLU) more than normal cells (Warburg 1925). He further noted that cancer cells metabolized GLU more by anaerobic glycolysis (his term “fermentation”) than by oxidative respiration. That is, cancer cells tended not to completely oxidize glucose to CO2 and H2O, as is done in the oxidative respiration in the TCA cycle, but rather excreted lactic acid, CH3CH2(OH)COOH (abbreviated LACA), a glycolysis byproduct that is produced from pyruvic acid, CH3CHO–COOH (abbreviated PYRA). This excess LACA is known as the Warburg Glycolytic Effect, shown in Figure 5.8. Normally, blood GLU comes from food sources or from glycogen and is specifically transported into the cell by insulin. GLU is then phosphorylated (GLU-P), which traps GLU in the cell, and then GLU-P is converted (by a series of glycolytic cytoplasmic enzymes) to PYRA, which in turn enters the TCA cycle, where, in the presence of respiratory O2, is absorbed into the mitochondrion, which executes complete PYRA oxidation to CO2 and H2O by the TCA cycle. Complete GLU aerobic metabolism to CO2 and H2O releases a total of 38 ATPs, the normal basic driver for energy in the cell. In the presence of ample O2, anaerobic glycolysis is usually depressed (the Pasteur Effect) in favor of the more efficient respiration metabolism. This is the normal state. The Warburg effect occurs in exhaustive exercise or when a tissue is hypoxic and the toxic PYRA accumulates because there is limited or not enough O2 present to drive TCA cycle. The cell has evolved lactic dehydrogenase A to act in hypoxia to convert PYRA to LACA. LACA is easily excreted from the metabolizing cell into extracellular environment. Warburg


Anaerobic dissimilation

149

Inhibit cell-to-cell Interactions Alter chromatin structures H+ H+ + H+ H

1 Glucose

2 Lactic acid ΔpH 2 NAD

2 NADH2 Net yield: 2ATP

1 Glucose 2 Pyruvic acid

Glycolysis

Lack of oxygen condition increases as tumor grows away from arterial blood supply Fermentation

Insufficient Oxidation to CO2 and H2O Figure 5.8. The Warburg anaerobic glycolytic effect. Because of the early growth effects of foci and small neoplasms, the transformed cells separate from the local blood supply and become more anoxic as they expand. Cells at the lead edge show the most effects of O2 deprivation, and this deprivation switches metabolism control in these distal cells from the 38 ATP-rich TCA cycle to glycolysis (only 2 ATPs) and fermentation to lactic acid. Because of the reduced energy yield, the glyclosis pathway shown here cycles rapidly in neoplasms, thus depleting glycogen. As cells accumulate and excrete lactic acid, local acid effects occur which can dirupt pH control intra- and intercellularly. Some of the proposed effects of this hypothesis are shown. It is posited that mutagenesis is more likely under these altered conditions (Warburg 1956a,b).

hypothesized that these metabolic differences were the initial cause and not the result of carcinogenesis (Warburg 1956a,b). The cause of hypoxia can come early in the chemical carcinogenesis process where field effects in the LTA become perturbed by the entrance of a chaotropic xenochemical that produces reactive hyperplasia. Chaotropic agents are known to cause numerous types of physical and chemical molecular and cellular and intercellular alterations (e.g., disruption of actin and E-cadherin with the ECM), or alterations in Na+, K+, or H+ gradients which can in turn affect osmolality by shifting the cellular H2O balance in favor of uptake that leads to swollen cells (hypertrophy). Reactive hyperplasia can occur because of episomal responses and produces increases the compartment’s number of cells beyond the normal N-cell number. This is especially true if exposure to the chaotropic chemical is frequent and of long duration, thus preventing reparation. These disruptive cellular reactions have been characterized by a number of investigators as being some of the earliest of responses even before transformed foci form (Boutwell 1976; Feinberg and Tycko 2004; Foulds 1957; Rubin 1994). Hypertrophic and reactive hyperplastic responses are the tissue’s strategies of quickly diluting and shielding the xenochemical insult while preparing

150


TABLE 5.2.

Ion Λ0

Specific Electrical Conductances of Physiological Ions (Λ0)a

H+ 350

K+ 73

Na+ 50

Li+ 39

Ca2+ 119

Mg2+ 57

a

Extrapolated to infinite dilution to minimize chemical activity interactions, and the temperature of the aqueous solution is normalized to 25 °C with specific aqueous conduction of ions Λ0 and the unitis: 10−4 m2·A·s/mole, where m = meters, A = amps, and s = seconds.

the tissue for chemokine injury signaling (Coussens and Werb 2002; Federico et al. 2007; Maronpot et al. 1989; Pitot et al. 1985; Porta et al. 2007). Hypoxia takes place when the cells swell and are increasing in number such that the expanding cells become ever more distanced from their local blood supply. More new cell displacement causes less available O2 because these expanded regions exceed the passive diffusion limit of O2, which is about 100 μM from the blood supply (Gatenby and Gillies 2004). The hyperplastic new cells become hypoxic, causing the tissue to favor the hypoxic state, and LACA excretion commences (the Warburg anaerobic glycolysis effect; see Figure 5.6). That the H+ ions from LACA ionization can cause such pleiotropy should come as no surprise because the specific electrical conductances of physiologic ions seen in Table 5.2 clearly show that the H+ is by far the most mobile ion of cellular ions. It is a proton with a very large charge/mass ratio. This means that a modest change in pH will make substantial structural changes by interacting with various anionic groups in enzymes, chromatin, DNA, and mitochondrial and plasma membranes. The acid (from LACA) in the extracellular fluid leads to even more environmental perturbance and more reactive hyperplasia, along with activation of sentinel white cells, and finally the affected cells enter into a condition of regional acidosis. The penumbral region of distal cells exhibits the most hypoxicity and, in time, have been observed to become premalignant foci that are increasingly resistant to apoptosis and express increased membrane transporters in order to maintain intracellular pH (Gatenby and Gillies 2004; Klaassen and Lu 2008; Pikarsky et al. 2004). All is still reversible if the chemical is removed or otherwise metabolized in a timely manner. If not, the region proceeds to distort from homeodynamic equilibrium and develops early competing clones for resources such as glucose and O2. Oxygen metabolism and its balance can have much influence on tumor promotion (Troll and Wiesner 1985). In a thoughtful review by DeBerardinis, the authors provide a discussion on reprogramming of local metabolism in hypoxia and in establishing anaerobic glycolysis which can become sustained by the induction of hypoxia-inducible factor (HIF-1) expression and other factors (DeBerardinis et al. 2008). Many reviewers of today not cited here (due to space) still think that Warburg’s idea might be the capital effect of the I stage (Kondoh 2008). The biological sequela to form competing clones of foci is a prelude to oncogenesis and has been covered earlier in this chapter. Proof that Warburg’s hypothesis of GLU mismanagement is provided by the discovery that part of this hypoxia mechanism causes cachexia. This wasting condition is common in about one-half of all cancer patients and is the reason why


151

one-fourth of cancer patients die. Besides gaining anorexia, the cancer patients acquire both GLU intolerance and insulin insensitivity (Lelbach et al. 2007). Metabolic effects analogous to the Warburg Effect start in the earliest part of the I stage. Warburg claims that it precedes clinical discovery only to worsen in advanced stages of cancer (Warburg 1956b). Patient muscular activity leads to the release of epinephrine (adrenaline), which causes the breakdown of glycogen in the muscles to GLU. That is, the muscle-stored glycogen is cleaved from the nonreducing polysaccharide ends of the carbohydrate chain by the enzyme glycogen phosphorylase to produce glucose-1-phosphate that is converted to glucose 6-phosphate which cannot leak out of the cell. The breakdown of glycogen is for the production of ATP that is consumed during muscular activity. Continued activity creates the demand for more ATP. In the beginning of the cancer disease, glycolysis produces PYRA that is converted to acetyl CoA, which is metabolized in the citric acid cycle to make ATP via aerobic metabolism. Later in the disease, however, O2 becomes scarce (see above and Figure 5.6) and anaerobic glycolysis becomes more dominant over TCA cycle or aerobic GLU oxidation with time. An effect called the Cori cycle sets up between the muscles and the liver (Figure 5.9). First, the LACA that the muscle produces is released into the blood, circulated, and then taken up by the liver; and by gluconeogenesis, which consumes 6 ATPs, it converts LACA → PYRA → GLU, which is then released from the liver into the blood. Subsequently, GLU is transported from the blood into the muscle and by the glycolysis process converts GLU → PYRA → LACA while releasing 2 ATPs. This completes the muscle–liver Cori cycle with a net deficit of −4 ATP/cycle. This

The Cori Cycle Blood Glucose

More Oxygenated

Glucose 2 ATP

6 ATP

2 Pyruvate

2 Pyruvate 2 Lactate H+

Liver

H+

2 Lactate

Blood Less Oxygenated

H+

H+

Muscle

Figure 5.9. The Cori cycle. Note that muscle and other tissues can produce excessive lactic acid in early cancer. Some lactate ionizes to H+ ion and thus lowers local pH. These acid effects are thought to disturb cell-to-cell interactions by cellular water and pH imbalances with chromatin/DNA rearrangements and hence is proposed to initiate neoplastic environments according to the Warburg cancer theory. The Cori cycle repeats many times in advancing cancer; and because of the net loss of energy by 2 ATP − 6 ATP = −4 ATPs per each Cori cycle, there is loss in the ability maintain tissue. Hence, there is a wasting of tissue over time. Cachexia is seen in a high percentage of terminal cancer patients.

152


progressive ATP or energy deficit produces increasing malaise and weakness usually seen in terminal patients, and it establishes a one-way catabolism of glycogen, body fat through usage of acetyl-CoA, and finally protein muscle mass: the cachexia syndrome (Tisdale 1997). It is not proved, but is a consistent observation, that LACA production in the Cori cycle is most likely correlated with adverse effects of essential structures because of excessive acid (see above) and the creation of a persistent ATP deficit. These effects of hypoxia and ATP deficits are essential properties inherent in cancer disease progression, and their origins exist from the beginning of carcinogenesis. This metabolic disturbance is a rational candidate for the incipient fatal events in chemical carcinogenesis.

5.4. SOME FINAL THOUGHTS ON BIOLOGY AND CANCER Cancer is a rare event among all cellular dystrophies; but because the number of events and adverse environments are so numerous over a lifetime, the chances are one in four of experiencing some form of cancer. Adults lose about 50–70 billion cells per day and execute about 2 × 1015 apoptoses/lifetime. Of necessity, this must be close to the cell replacement rate given good health at dynamic equilibrium. Because some cells rarely turn over, the total sum for a septuagenarian must be somewhat greater than the 2 × 1015 apoptoses/lifetime. The total number of mitoses/ person has been estimated to be about 1016 cell cycles/lifetime/adult body (Weinberg 1997). For all of this mitotic activity and gene expression (of DNA and chromatin), cell physiology exhibits extremely high fidelity in maintenance and passage of correct information compared to any other process we humans experience. Most of the time, our cells are corrected by various combinations of independent protection mechanisms: immune surveillance, numerous systems of genetic and chromatin repair, intracellular and epigenetic surveillance, and intercellular surveillance (Klein et al. 2007). It is indeed the reason why we do not succumb sooner to the chaos of thermodynamic entropy which acts on all processes. On average, there occur relatively few age-corrected diseases or cancers until old age. Heritable defects, adverse lifestyles, and adventitious exposures to chemical carcinogens are examples that can accelerate defect rates beyond the cell’s normal potential to correct mistakes, lesions, and injuries (cf. Section 5.3.5). The preemption or ablation of pesticide excessive exposures can avoid I-stage events or interrupt early P-stage events and positively affect cancer prophylaxis. If caught early enough, in fact, some mistakes or lesions can reverse or remodel as was reviewed previously, and be eliminated by apoptosis or by cell senescence (cf. Section 5.3.2). If the intervention is not imposed soon enough, however, the progression of carcinogenesis can proceed far enough in the P stage such that the cancer mass can pass a “point of no return” (Table 5.1). This point is characterized by the acquisition of multiple and irreversible critical stage steps that in time become permanent (Figure 5.4). The tumor becomes cancerous and usually grows at the physiologic expense of the affected site. Only through a new set of metastatic signals can the localized cancer become delocalized and spread in metastasis, often with fatal results (Weinberg 2008b).

5.4. SOME FINAL THOUGHTS ON BIOLOGY AND CANCER

153

Stress can have biologically significant effects on carcinogenesis (Godbout and Glaser 2006). Stress has two connotations: There can be chemically induced stress reactions and/or psychologically induced stresses. When an organ is under chemical stress, there are induced heat shock proteins like HSP27, HSP 70, and HSP 90 (Calderwood et al. 2006; Calderwood and Ciocca 2008). These protein factors are highly conserved in evolution and are part of a highly effective immunological response to counter chemical toxicity events. However, one factor of this family (HSP 1) seems to support carcinogenesis (Dai et al. 2007). There are many reviews on reactions to chemical stressors such as carcinogens (Zhang and Vande Woude 2007). The reader is also referred to Chapter 6 in this volume. Less known, but not necessarily less important, is the induction of the stress reaction by unremitting psychological or perceived adverse pressure from psychosocial factors that produce signaling that links pathologic stress responses with chemical carcinogenesis (Murakami et al. 2007). Aging was once thought to be the stochastic accumulation of errors and the statistical eventuality that the individual had reached the design limits of the body. More recent evidence suggests that there may be a specific aging program that, when activated, starts an ordered species-specifc decline (Holliday 2004). We postulate here that both stochastic and determinative programmatic theories may be true. If an “aging program” for higher organisms is operative, one might see biochemical reactions favoring a programmed decline or gradual shutdown. More age-related event biomarkers are needed to establish whether this is the fact. One of the fundamental properties of cells in culture was discovered by Leonard Hayflick in 1965, when he demonstrated that normal human cells in an in vitro culture divide about 52 times before autonomously entering a senescence phase. One reason for this is that each mitotic cycle allows the enzyme telomerase to shorten the ends of chromosomes, called telomeres (Ben-Porath and Weinberg 2004). When telomeres become too short or are ablated, chromosomes become unstable structurally and this leads to cell death. This mechanism acts as a countdown clock recording the number times the cell transacts mitosis (Finkel et al. 2007; Stewart and Weinberg 2006). The longer a cell exists, the more chances there are to accumulate errors within the cellular DNA and/or the histone code because these repair systems collectively have a finite error rate. Apoptosis and rebuilding a rigorous new cell replacement is believed to have been selected in evolution in order to protect the body from creating old defect-ridden cells that could lead to cancer before reproduction is achieved. After a time, it is thermodynamically more precise and accurate to start over by generating a new cell from a well-protected stem cell as a template. Short telomeres have often been observed in the P stage of carcinogenesis. Hence, a cell is born from asymmetrical stem cell mitosis, matures, replicates ↔ functions, and finally reaches the cell time limit set for that organ. Cell senescence does not happen abruptly, as once was thought, but instead cells undergo senescence during the last cycles of the aging cells (Weinberg 1997). There is evidence that the self-recognizing immunologic reactions, established early by 2–3 years of age, begin to break down in old age, giving rise to more defects in normal cells. With advanced age, individuals become more susceptible to disease, including cancer, while also establishing “autoimmune types” of problems (Anisimov 2007). However, there is

154


evolutionary evidence to suggest that long-life animals may have more repair mechanisms that are more efficient, thus accounting for their extended age (Holliday 2004). One plausible theory of aging is that somatic stem cells begin to show their age by their accumulated defects: As the organ generative cells go, so go the organs. There may be incorrect control of autophagy in old age, thus causing organelle damages that would, over time, produce faltering cells (Finkel et al. 2007). With so many known evolved control mechanisms in embryos, neonates, and puberty, it is reasonable to assume that mechanisms might also exist to end life in order to maintain and replenish a rigorous gene pool. If true, hazard evaluation must realistically take into account this essential aspect of human design. This may explain the control incidences of various organ cancers observed in control mice and rats in the U.S. NTP’s cancer bioassays (Maronpot 2007; Finkel et al. 2007; Holliday 2004). Examination of an aging cohort presents excellent opportunities to study the senescence of mechanisms and pathways and possibly adding to quality-of-life issues that the elderly face. The goal of this chapter was to cover many of the essential biological reactions and pathways that can participate in carcinogenesis. There are many protective evolved mechanisms in humans that ward off cancer and other diseases. Section 5.3.4 showed that it is an aberrant I cell in a field of physicochemically disrupted normal cells that provides for poor physiological communication in the local tissue array (LTA). This sets the stage for premalignancy but is reversible as shown in Figure 5.4. The participation of sentinel cells, cytokine-recruited cells, and some distal cells as in the Warburg Effect and the Cori cycle (Figures 5.8 and 5.9) demonstrate that multiple cell types can co-participate in cancer initiation and progression. The number of qualitatively independent stages from initiation to malignancy to metastasis and then to distal colonization has been estimated by many authors to be at least 6–8 separately acting, principal stages that are obligatory for cancer progression (Emmelot and Scherer 1977; Farber 1987; Gatenby and Gillies 2008; Marks et al. 2007; Rangarajan et al. 2004; Weinberg 1989). The overall process is not stochastic but has deterministic characteristics too. Cancer manifestation depends on the completion of biochemically different steps within each stage and the orderdependent 6–8 stages. Each stage contains a number of separate steps; and at this time, only some of these steps are recognized and understood. Unique steps should provide unique markers for each the stages. Not only must all 6–8 essential stages be satisfied in multistaged carcinogenesis, but each stage must also be driven to completion by a combination of duration and concentration of that step’s causal agent(s), host factors, genetic predisposition, lifestyle, stress, age, and so on. It has also been reviewed here that the I-stage duration may be as long as the last stage of colonization, but the middle P-stage and malignancy acquisition can go fast by comparison under sufficient carcinogenic pressure. The time versus cancer event curve is likely bell-shaped response. Preliminary biological modeling suggests that the accumulative dose–response curve of the 6–8 multistages is likely a sigmoid curve (whatever the probabilities of the individual steps are). That is, the dose– cancer response would start slow at low exposures but would show higher exponential rates at higher accumulated exposures. In the latter case, high exposures will force all stage steps to react faster, which allows less time for clonal competition,

REFERENCES

155

repair, and remodeling. At a low dose of chemical carcinogen, the competition among clones is less and chemical pressure is less. Either the malignant clone arises on a long time course or it does not arise at all because it can be within the range of detection, detoxification, repair, remodeling (e.g., cyst formation), or resorption. With less clonal remediation time, the pressure of high-doses force genetic and epigenetic alterations would favor malignant clones with more heterogeneity. The response would rise fast with time but would slow down in the latter stages of malignancy formation, metastasis, and colonization which are respectively less frequent and less probable events. This would complete the plateau region of the sigmoid curve. It has been proposed that malignant regrowth after cancer surgery may not always mean that some residual malignant cells were not removed, but rather there is a systematic and ongoing organ dysfunction within the episome or “conditioned cell memory” that is persistent in all the cells of the organ such that tumor regrowth is inevitable (Ruggiero and Bustuoabad 2006). That is, unless this episomal lesion in this LTA is fixed, the organ cells can self-initiate to produce the I* state—that is, mutagenize themselves because of improper controls. For the particulars reviewed here, we currently have a paucity of verified specific biomarkers to monitor cancer progress. This is especially relevant in light of the new cell advances in recent years. There are reports that miRNA can act as specific transcription factors and RNAi that specifically turn off steps in cancerous processes (He et al. 2007a,b; Ma and Weinstein 2008; Makunin et al. 2007). Functional RNAs, such as miRNA and RNAi, promise not only to be rich areas for biomarkers in toxicology, but also of immense value to society in our war against cancer (Wynter 2006). Because cancer development is varied in the various organs, it is suggested that organ cancer specifics be studied so as to better understand the interplay of biology and cancer: bladder (Luis et al. 2007); brain (Calabrese et al. 2007); gastric (Humar and Guilford, 2008); myeloma (Caers et al. 2008); thyroid (Reisco-Eizaguirre and Santisteban 2007); pancreas (Li et al. 2007); lymphatics (Allan et al. 2006); colon (Hedrick et al. 1994); prostate (Mimeaut and Batra 2006); Prins et al. 2008); and breast (Polyak 2007).

REFERENCES Adler, A. S., and Chang, H. Y. (2006). From description to causality: mechanisms of gene expression signatures in cancer. Cell Cycle 5(11), 1148–1151. Aguirre-Ghiso, J. A. (2007). Models, mechanisms and clinical evidence for cancer dormancy. Nat Rev Cancer 7(11), 834–846. Albor, A., and Kulesz-Martin, M. (2007). Novel initiation genes in squamous cell carcinomagenesis: A role for substrate-specific ubiquitylation in the control of cell survival. Mol Carcinog 46(8), 585–590. Allan, A., George, R., Vantyghem, S., Lee, M. Hodson, N. Engel, J. Holliday, R., Girvan, D., Scott, L., et al. (2006). Role of integrin binding protein osteopontin in lymphatic metastasis of breast cancer. Am J Pathol 169(1), 233–245. Amaral, P. P., Dinger, M. E., Mercer, T. R., and Mattick, J. S. (2008). The eukaryotic genome as an RNA machine. Science 319(5871), 1787–1789. Americal Cancer Society (2008). Cancer Facts and Figures 2008. Accessed at http://www.cancer.org/ docroot/home/index.asp, viewed on 2-23-2009.

156


Ames, B. N., Gold, L. S., and Willet, W. C. (1995). The causes and prevention of cancer. Proc Natl Acad Sci 92, 5258–5265. Angel, J. M., and DiGiovanni, J. (1999). Genetics of skin tumor promotion. Prog Exp Tumor Res 35, 143–157. Anisimov, V. N. (2003). The relationship between aging and carcinogenesis: A critical appraisal. Crit Rev Oncol Hematol 45(3), 277–304. Anisimov, V. N. (2007). Biology of aging and cancer. Cancer Control 14(1), 23–31. Anisimov, V. N. (2008). Carcinogenesis and aging 20 years after: Escaping horizon. Mech Ageing Dev 130(1–2), 105–121. Aslakson, C. J., and Miller, F. R. (1992). Selective events in the metastatic process defined by analysis of the sequential dissemination of subpopulations of a mouse mammary tumor. Cancer Res 52(6), 1399–1405. Bach, S. P., Renehan, A. G., and Potten, C. S. (2000). Stem cells: The intestinal stem cell as a paradigm. Carcinogenesis 21(3), 469–476. Bachman, A. N., Curtin, G. M., Doolittle, D. J., and Goodman, J. I. (2006). Altered methylation in genespecific and GC-rich regions of DNA is progressive and nonrandom during promotion of skin tumorigenesis. Toxicol Sci 91(2), 406–418. Balkwill, F., and Mantovani, A. (2001). Inflammation and cancer: Back to Virchow? Lancet 357(9255), 539–545. Baltzer, F. (1964). Theodor Boveri. Science 15, 809–815. Barrett, J. C., Kakunaga, T., Kuroki, T., Neubert, D., Trosko, J. E., Vasiliev, J. M., Williams, G. M., and Yamasaki, H. (1986). Short-term assays to predict carcinogenicity. Mammalian cell transformation in culture. IARC Sci Publ (83), 267–286. Barrett, J. C., and Ts’o, P. O. (1978a). Evidence for the progressive nature of neoplastic transformation in vitro. Proc Natl Acad Sci USA 75(8), 3761–3765. Barrett, J. C., and Ts’o, P. O. (1978b). Relationship between somatic mutation and neoplastic transformation. Proc Natl Acad Sci USA 75(7), 3297–3301. Beadle, G. W., and Tatum, E. L. (1941). Genetic control of biochemical Reactions in Neurospora. Proc Natl Acad Sci USA 27(11), 499–506. Ben-Porath, I., Thomson, M. W., Carey, V. J., Ge, R., Bell, G. W., Regev, A., and Weinberg, R. A. (2008). An embryonic stem cell-like gene expression signature in poorly differentiated aggressive human tumors. Nat Genet 40(5), 499–507. Ben-Porath, I., and Weinberg, R. A. (2004). When cells get stressed: an integrative view of cellular senescence. J Clin Invest 113(1), 8–13. Berenblum, I. (1941). The mechanism of carcinogensis. Cancer Res 1, 804–814. Berenblum, I. (1954). Carcinogenesis and tumor pathogenesis. Adv Cancer Res 2, 129–175. Berenblum, I., and Shubik, P. (1949). An experimental study of the initiating state of carcinogenesis, and a re-examination of the somatic cell mutation theory of cancer. Br J Cancer 3(1), 109–118. Bernards, R., and Weinberg, R. A. (2002). A progression puzzle. Nature 418(6900), 823. Berwald, Y., and Sachs, L. (1965). In vitro transformation of normal cells to tumor cells by carcinogenic hydrocarbons. J Natl Cancer Inst 35(4), 641–661. Biel, M., Wascholowski, V., and Giannis, A. (2005). Epigenetics—An epicenter of gene regulation: histones and histone-modifying enzymes. Angew Chem Int Ed Engl 44(21), 3186–3216. Boreiko, C., Mondal, S., Narayan, K. S., and Heidelberger, C. (1980). Effect of 12-O-tetradecanoylphorbol13-acetate on the morphology and growth of C3H/10T1/2 mouse embryo cells. Cancer Res 40(12), 4709–4716. Boutwell, R. K. (1964). Some biological aspects of skin carcinogenesis. Prog Exp Tumor Res 4, 207–250. Boutwell, R. K. (1974). The function and mechanism of promoters of carcinogenesis. CRC Crit Rev Toxicol 2(4), 419–443. Boutwell, R. K. (1976). The biochemistry of preneoplasia in mouse skin. Cancer Res 36(7 PT 2), 2631–2635. Boutwell, R. K. (1985). Tumor promoters in human carcinogenesis. In Important Advances in Oncology. Lippincott Williams & Wilkins, Philadelphia, pp. 16–27. Boveri, T. (1929). The Origin of Malignant Tumors, The William & Wilkin Company, Baltimore, pp. 1–119. Burnet, F. M. (1971). Immunological surveillance in neoplasia. Transplant Rev 7, 3–25.

REFERENCES

157

Burns, F., Albert, R., Altshuler, B., and Morris, E. (1983). Approach to risk assessment for genotoxic carcinogens based on data from the mouse skin initiation–promotion model. Environ Health Perspect 50, 309–320. Caers, J., Valckenborgh, E., Menu, E., Van Camp, B., and Vanderkerken, K. (2008). Unraveling the biology of multiple myeloma disease: Cancer stem cells, acquired intracellular changes and interactions with surrounding micro-environment. Bull Cancer 95(3), 301–313. Cairns, J. (1975). Mutation selection and the natural history of cancer. Nature 255(5505), 197–200. Calabrese, C., Poppleton, H., Kocak, M., Hogg, T., Fuller, C., Hamner, B., Oh, E. Gaber, M. W., Finklestein, D., Allen, M., et al. (2007). A perivascular niche for brain tumor stem sells. Cancer Cell 11, 69–82. Calderwood, S. K., and Ciocca, D. R. (2008). Heat shock proteins: stress proteins with Janus-like properties in cancer. Int J Hyperthermia 24(1), 31–39. Calderwood, S. K., Khaleque, M. A., Sawyer, D. B., and Ciocca, D. R. (2006). Heat shock proteins in cancer: Chaperones of tumorigenesis. Trends Biochem Sci 31(3), 164–172. Campbell, H. A., Xu, Y. D., Hanigan, M. H., and Pitot, H. C. (1986). Application of quantitative stereology to the evaluation of phenotypically heterogeneous enzyme-altered foci in the rat liver. J Natl Cancer Inst 76(4), 751–767. Casto, B. C., Janosko, N., and DiPaolo, J. A. (1977). Development of a focus assay model for transformation of hamster cells in vitro by chemical carcinogens. Cancer Res 37(10), 3508–3515. Chan, T. A., Glockner, S., Yi, J. M., Chen, W., Van, N. L., Cope, L., Herman, J. G., Velculescu, V., Schuebel, K. E., Ahuja, N., and Baylin, S. B. (2008). Convergence of mutation and epigenetic alterations identifies common genes in cancer that predict for poor prognosis. PLoS Med 5(5), e114. Chang, C. C., Sun, W., Cruz, A., Saitoh, M., Tai, M. H., and Trosko, J. E. (2001). A human breast epithelial cell type with stem cell characteristics as target cells for carcinogenesis. Radiat Res 155 (1 Pt 2), 201–207. Cheng, L., and Lai, M. D. (2003). Aberrant crypt foci as microscopic precursors of colorectal cancer. World J Gastroenterol 9(12), 2642–2649. Cifone, M. A., and Fidler, I. J. (1981). Increasing metastatic potential is associated with increasing genetic instability of clones isolated from murine neoplasms. Proc Natl Acad Sci 78(11), 6949–6952. Clark, S. J. (2007). Action at a distance: epigenetic silencing of large chromosomal regions in carcinogenesis. Hum Mol Genet 16(Spec No 1), R88–R95. Cook, J. W., Heiger, I., Kennaway, E. L., and Mayneord, W. V. (1932). The production of cancer by pure hydrocarbons—Part 1. R Soc Proc 111(Part B), 455–484. Coussens, L. M., Raymond, W. W., Bergers, G., Laig-Webster, M., Behrendtsen, O., Werb, Z., Caughey, G. H., and Hanahan, D. (1999). Inflammatory mast cells up-regulate angiogenesis during squamous epithelial carcinogenesis. Genes Dev 13(11), 1382–1397. Coussens, L. M., Tinkle, C. L., Hanahan, D., and Werb, Z. (2000). MMP-9 supplied by bone marrowderived cells contributes to skin carcinogenesis. Cell 103(3), 481–490. Coussens, L. M., and Werb, Z. (2002). Inflammation and cancer. Nature 420(6917), 860–867. Coyle-Rink, J., Del, V. L., Sweet, T., Khalili, K., and Amini, S. (2002). Developmental expression of Wnt signaling factors in mouse brain. Cancer Biol Ther 1(6), 640–645. Crick, F. H. (1958). On protein synthesis. Symp Soc Exp Biol 12, 138–163. Criollo, A., Galluzzi, L., Maiuri, M. C., Tasdemir, E., Lavandero, S., and Kroemer, G. (2007). Mitochondrial control of cell death induced by hyperosmotic stress. Apoptosis 12(1), 3–18. Curtin, G. M., Hanausek, M., Walaszek, Z., Zoltaszek, R., Swauger, J. E., Mosberg, A. T., and Slaga, T. J. (2006). Short-term biomarkers of cigarette smoke condensate tumor promoting potential in mouse skin. Toxicol Sci 89(1), 66–74. D’Alessio, A. C., and Szyf, M. (2006). Epigenetic tete-a-tete: The bilateral relationship between chromatin modifications and DNA methylation. Biochem Cell Biol 84(4), 463–476. D’Alessio, A. C., Weaver, I. C., and Szyf, M. (2007). Acetylation-induced transcription is required for active DNA demethylation in methylation-silenced genes. Mol Cell Biol 27(21), 7462–7474. Dai, C., Whitesell, L., Rogers, A. B., and Lindquist, S. (2007). Heat shock factor 1 is a powerful multifaceted modifier of carcinogenesis. Cell 130(6), 1005–1018. Dai, D. L., Martinka, M., Bush, J. A., and Li, G. (2004). Reduced Apaf-1 expression in human cutaneous melanomas. Br J Cancer 91(6), 1089–1095.

158


de Visser, K. E., Eichten, A., and Coussens, L. M. (2006). Paradoxical roles of the immune system during cancer development. Nat Rev Cancer 6(1), 24–37. DeBerardinis, R. J., Lum, J. J., Hatzivassiliou, G., and Thompson, C. B. (2008). The biology of cancer: Metabolic reprogramming fuels cell growth and proliferation. Cell Metab 7(1), 11–20. Deelman, H. T. (1927). The part played by injury and repair in the development of cancer. Br Med J 1, 872. DeNardo, D. G., Johansson, M., and Coussens, L. M. (2008). Immune cells as mediators of solid tumor metastasis. Cancer Metast Rev 27(1), 11–18. Déry, U., and Masson, J. Y. (2007). Twists and turns in the function of DNA signalling and repair proteins by post-translational modifications. Dolberg, D. S., Hollingsworth, R., Hertle, M., and Bissell, M. J. (1985). Wounding and its role in RSVmediated tumor formation. Science 230(4726), 676–678. Doll, R., and Hill, A. B. (1956). Lung cancer and other causes of death in relation to smoking; a second report on the mortality of British doctors. Br Med J 2(5001), 1071–1081. Doll, R., and Hill, A. B. (2004). The mortality of doctors in relation to their smoking habits: A preliminary report. 1954. Br Med J 328(7455), 1529–1533. Dragan, Y. P., Campbell, H. A., Xu, X. H., and Pitot, H. C. (1997). Quantitative stereological studies of a “selection” protocol of hepatocarcinogenesis following initiation in neonatal male and female rats. Carcinogenesis 18(1), 149–158. Druckrey, H. (1967). Quantitative aspects in chemical carcinogenesis. In Potential Carcinogenic Hazards from Drugs, Springer-Verlag, Berlin, pp. 60–78. Druckrey, H., Schildbach, A., Schmaehl, D., Preussmann, R., and Ivankovic, S. (1963). Quantitative analysis of the carcinogenic effect of diethylnitrosamine. Arzneimittelforschung 13, 841–851. Dvorak, H. F. (1986). Tumors: Wounds that do not heal. Similarities between tumor stroma generation and wound healing. N Engl J Med 315(26), 1650–1659. Elenbaas, B., and Weinberg, R. A. (2001). Heterotypic signaling between epithelial tumor cells and fibroblasts in carcinoma formation. Exp Cell Res 264(1), 169–184. Elsasser, W. M. (1984). Outline of a theory of cellular heterogeneity. Proc Natl Acad Sci USA 81(16), 5126–5129. Eming, S. A., Krieg, T., and Davidson, J. M. (2007). Inflammation in wound repair: Molecular and cellular mechanisms. J Invest Dermatol 127(3), 514–525. Emmelot, P., and Scherer, E. (1977). Multi-hit kinetics of tumor formation, with special reference to experimental liver and human lung carcinogenesis and some gneral conclusions. Cancer Res 37(6), 1702–1708. Enomoto, K., and Farber, E. (1982). Kinetics of phenotypic maturation of remodeling of hyperplastic nodules during liver carcinogenesis. Cancer Res 42(6), 2330–2335. Fadeel, B., Ottosson, A., and Pervaiz, S. (2008). Big wheel keeps on turning: Apoptosome regulation and its role in chemoresistance. Cell Death Differ 15(3), 443–452. Fadl-Elmula, I., Gorunova, L., Mandahl, N., Elfving, P., Lundgren, R., Mitelman, F., and Heim, S. (1999). Cytogenetic monoclonality in multifocal uroepithelial carcinomas: Evidence of intraluminal tumour seeding. Br J Cancer 81(1), 6–12. Farber, E. (1984). The multistep nature of cancer development. Cancer Res 44(10), 4217–4223. Farber, E. (1987). Pathogenesis of experimental liver cancer: Comparison with humans. Arch Toxicol Suppl 10, 281–288. Fearon, E. R., Cho, K. R., Nigro, J. M., Kern, S. E., Simons, J. W., Ruppert, J. M., Hamilton, S. R., Preisinger, A. C., Thomas, G., Kinzler, K. W., et al. (1990). Identification of a chromosome 18q gene that is altered in colorectal cancers. Science 247(4938), 49–56. Fearon, E. R., Hamilton, S. R., and Vogelstein, B. (1987). Clonal analysis of human colorectal tumors. Science 238(4824), 193–197. Federico, A., Morgillo, F., Tuccillo, C., Ciardiello, F., and Loguercio, C. (2007). Chronic inflammation and oxidative stress in human carcinogenesis. Int J Cancer 121(11), 2381–2386. Feinberg, A. P. (2004). The epigenetics of cancer etiology. Semin Cancer Biol 14(6), 427–432. Feinberg, A. P. (2005). A genetic approach to cancer epigenetics. Cold Spring Harb Symp Quant Biol 70, 335–341.

REFERENCES

159

Feinberg, A. P. (2007). Phenotypic plasticity and the epigenetics of human disease. Nature 447(7143), 433–440. Feinberg, A. P. (2008). Epigenetics at the epicenter of modern medicine. JAMA 299(11), 1345–1350. Feinberg, A. P., Ohlsson, R., and Henikoff, S. (2006). The epigenetic progenitor origin of human cancer. Nat Rev Genet 7(1), 21–33. Feinberg, A. P., and Tycko, B. (2004). The history of cancer epigenetics. Nat Rev Cancer 4(2), 143–153. Fialkow, P. J. (1976). Clonal origin of human tumors. Biochim Biophys Acta 458(3), 283–321. Fidler, I. J. (1978). Tumor heterogeneity and the biology of cancer invasion and metastasis. Cancer Res 38(9), 2651–2660. Fidler, I. J. (2002). Critical determinants of metastasis. Semin Cancer Biol 12(2), 89–96. Fidler, I. J., Kim, S. J., and Langley, R. R. (2007). The role of the organ microenvironment in the biology and therapy of cancer metastasis. J Cell Biochem 101(4), 927–936. Fidler, I. J., and Kripke, M. L. (1977). Metastasis results from preexisting variant cells within a malignant tumor. Science 197(4306), 893–895. Finkel, T., Serrano, M., and Blasco, M. A. (2007). The common biology of cancer and ageing. Nature 448(7155), 767–774. Flanagan, J. M. (2007). Host epigenetic modifications by oncogenic viruses. Br J Cancer 96(2), 183–188. Folkman, J. (2002). Role of angiogenesis in tumor growth and metastasis. Semin Oncol 29(6 Suppl 16), 15–18. Foulds, L. (1957). Tumor progression. Cancer Res 17(5), 355–356. Furth, J. (1953). Conditioned and autonomous neoplasms: A review. Cancer Res 13(7:1), 477–492. Galluzzi, L., Vicencio, J. M., Kepp, O., Tasdemir, E., Maiuri, M. C., and Kroemer, G. (2008). To die or not to die: That is the autophagic question. Curr Mol Med 8(2), 78–91. Gatenby, R. A., and Gillies, R. J. (2004). Why do cancers have high aerobic glycolysis? Nat Rev Cancer 4(11), 891–899. Gatenby, R. A., and Gillies, R. J. (2008). A microenvironmental model of carcinogenesis. Nat Rev Cancer 8(1), 56–61. Gillies, R. J., and Gatenby, R. A. (2007). Adaptive landscapes and emergent phenotypes: Why do cancers have high glycolysis? J Bioenerg Biomembr 39(3), 251–257. Godbout, J. P., and Glaser, R. (2006). Stress-induced immune dysregulation: Implications for wound healing, infectious disease and cancer. J Neuroimmune Pharmacol 1(4), 421–427. Gort, E. H., Groot, A. J., van der Wall, E., van Diest, P. J., and Vooijs, M. A. (2008). Hypoxic regulation of metastasis via hypoxia-inducible factors. Curr Mol Med 8(1), 60–67. Greten, F. R., Eckmann, L., Greten, T. F., Park, J. M., Li, Z. W., Egan, L. J., Kagnoff, M. F., and Karin, M. (2004). IKKbeta links inflammation and tumorigenesis in a mouse model of colitis-associated cancer. Cell 118(3), 285–296. Grum-Schwensen, B., Klingelhofer, J., Berg, C. H., El-Naaman, C., Grigorian, M., Lukanidin, E., and Ambartsumian, N. (2005). Suppression of tumor development and metastasis formation in mice lacking the S100A4(mts1) gene. Cancer Res 65(9), 3772–3780. Hahn, W. C., and Weinberg, R. A. (2002). Modelling the molecular circuitry of cancer. Nat Rev Cancer 2(5), 331–341. Hanahan, D., and Weinberg, R. A. (2000). The hallmarks of cancer. Cell 100(1), 57–70. Haseman, J. K., Huff, J. E., Zeiger, E., and McConnell, E. E. (1987). Comparative results of 327 chemical carcinogenicity studies. Environ Health Perspect 74, 229–235. He, L., He, X., Lim, L. P., de, S. E., Xuan, Z., Liang, Y., Xue, W., Zender, L., Magnus, J., Ridzon, D., Jackson, A. L., Linsley, P. S., Chen, C., Lowe, S. W., Cleary, M. A., and Hannon, G. J. (2007a). A microRNA component of the p53 tumour suppressor network. Nature 447(7148), 1130–1134. He, X., He, L., and Hannon, G. J. (2007b). The guardian’s little helper: MicroRNAs in the p53 tumor suppressor network. Cancer Res 67(23), 11099–11101. Hedrick, L., Cho, K. R., Fearon, E. R., Wu, T. C., Kinzler, K. W., and Vogelstein, B. (1994). The DCC gene product in cellular differentiation and colorectal tumorigenesis. Genes Dev 8(10), 1174–1183. Heim, S., Mandahl, N., and Mitelman, F. (1988). Genetic convergence and divergence in tumor progression. Cancer Res 48(21), 5911–5916.

160


Heim, S., Mertens, F., Jin, Y. S., Mandahl, N., Johansson, B., Biorklund, A., Wennerberg, J., Jonsson, N., and Mitelman, F. (1989). Diverse chromosome abnormalities in squamous cell carcinomas of the skin. Cancer Genet Cytogenet 39(1), 69–76. Hertel, K. J. (2008). Combinatorial control of exon recognition. J Biol Chem 283(3), 1211–1215. Hill, R., Song, Y., Cardiff, R. D., and Van Dyke, T. (2005a). Heterogeneous tumor evolution initiated by loss of pRb function in a preclinical prostate cancer model. Cancer Res 65(22), 10243–10254. Hill, R., Song, Y., Cardiff, R. D., and Van, D. T. (2005b). Selective evolution of stromal mesenchyme with p53 loss in response to epithelial tumorigenesis. Cell 123(6), 1001–1011. Hockel, M., and Vaupel, P. (2001). Biological consequences of tumor hypoxia. Semin Oncol 28 (2 Suppl 8), 36–41. Hoglund, M., Frigyesi, A., Sall, T., Gisselsson, D., and Mitelman, F. (2005). Statistical behavior of complex cancer karyotypes. Genes Chromosomes Cancer 42(4), 327–341. Hoglund, M., Sall, T., Heim, S., Mitelman, F., Mandahl, N., and Fadl-Elmula, I. (2001). Identification of cytogenetic subgroups and karyotypic pathways in transitional cell carcinoma. Cancer Res 61(22), 8241–8246. Holder, J. W., Elmore, E., and Barrett, J. C. (1993). Gap junction function and cancer. Cancer Res 53(15), 3475–3485. Holliday, R. (1987). The inheritance of epigenetic defects. Science 238(4824), 163–170. Holliday, R. (1989). DNA methylation and epigenetic mechanisms. Cell Biophys 15(1–2), 15–20. Holliday, R. (2002). Epigenetic comes of age in the twenty first century. Indian Acad Sci 81(1), 1–4. Holliday, R. (2004). The multiple and irreversible causes of aging. J Gerontol A Biol Sci Med Sci 59(6), B568–B572. Holliday, R. (2005). DNA methylation and epigenotypes. Biochemistry (Moscow) 70(5), 500–504. Holliday, R. (2006). Epigenetics: A historical overview. Epigenetics 1(2), 76–80. Holliday, R., and Murray, V. (1994). Specificity in splicing. Bioessays 16(10), 771–774. Hu, M. C., Wang, Y., Qiu, W. R., Mikhail, A., Meyer, C. F., and Tan, T. H. (1999). Hematopoietic progenitor kinase-1 (HPK1) stress response signaling pathway activates IkappaB kinases (IKK-alpha/ beta) and IKK-beta is a developmentally regulated protein kinase. Oncogene 18(40), 5514–5524. Humar, and Guilford (2008). Hereditary diffuse gastric cancer and lost cell polarity: A short path to cancer. Future Oncology 4(2), 229–239. Illmensee, K., and Mintz, B. (1976). Totipotency and normal differentiation of single teratocarcinoma cells cloned by injection into blastocysts. Proc Natl Acad Sci USA 73(2), 549–553. Jablonka, E. (2004). Epigenetic epidemiology. Int J Epidemiol 33(5), 929–935. Jablonka, E. (2006). Commentary: Induction and selection of variations during cancer development. Int J Epidemiol 35(5), 1163–1165. Jablonka, E., and Lamb, M. J. (2002). The changing concept of epigenetics. Ann NY Acad Sci 981, 82–96. Jablonka, E., and Lamb, M. J. (2007). Precis of evolution in four dimensions. Behav Brain Sci 30(4), 353–365. Jenuwein, T., and Allis, C. D. (2001). Translating the histone code. Science 293(5532), 1074–1080. Jones, and Baylin (2006). The epigenomics of cancer. Cell 128(4), 683–692. Kam, E., Melville, L., and Pitts, J. D. (1986). Patterns of junctional communication in skin. J Invest Dermatol 87(6), 748–753. Karnoub, A. E., Dash, A. B., Vo, A. P., Sullivan, A., Brooks, M. W., Bell, G. W., Richardson, A. L., Polyak, K., Tubo, R., and Weinberg, R. A. (2007). Mesenchymal stem cells within tumour stroma promote breast cancer metastasis. Nature 449(7162), 557–563. Katoh, Y., and Katoh, M. (2005). Hedgehog signaling pathway and gastric cancer. Cancer Biol Ther 4(10), 1050–1054. Kaufmann, S. H., and Steensma, D. P. (2005). On the TRAIL of a new therapy for leukemia. Leukemia 19(12), 2195–2202. Kaufmann, S. H., and Vaux, D. L. (2003). Alterations in the apoptotic machinery and their potential role in anticancer drug resistance. Oncogene 22(47), 7414–7430. Kaufmann, W. K., MacKenzie, S. A., and Kaufman, D. G. (1985). Quantitative relationship between hepatocytic neoplasms and islands of cellular alteration during hepatocarcinogenesis in the male F344 rat. Am J Pathol 119(2), 171–174.

REFERENCES

161

Kehler, J., Tolkunova, E., Koschorz, B., Pesce, M., Gentile, L., Boiani, M, et al. (2004). Oct-4 is required for primodial germ cell survival. EMBO Rep 5(11), 1078–1083. Kiba, T., Koyama, K., Ishiki, Y., Kimura, S., and Fukushima, T. (1960). On the mechanism of the development of multiple-drug-resistant clones of Shigella. Jpn J Microbiol 4, 219–227. Kim, H. J., Hawke, N., and Baldwin, A. S. (2006). NF-kappaB and IKK as therapeutic targets in cancer. Cell Death Differ 13(5), 738–747. Kim, Y., Sills, R. C., and Houle, C. D. (2005). Overview of the molecular biology of hepatocellular neoplasms and hepatoblastomas of the mouse liver. Toxicol Pathol 33(1), 175–180. Kinzler, K. W., and Vogelstein, B. (1996). Lessons from hereditary colorectal cancer 1. Cell 87(2), 159–170. Kinzler, K. W., and Vogelstein, B. (1998). Landscaping the cancer terrain. Science 280(5366), 1036–1037. Kitchin, K. T., Brown, J. L., and Setzer, R. W. (1994). Dose–response relationship in multistage carcinogenesis: Promoters. Environ Health Perspect 102(Suppl 1), 255–264. Klaassen, C. D., and Lu, H. (2008). Xenobiotic transporters: Ascribing function from gene knockout and mutation studies. Toxicol Sci 101(2), 186–196. Klaunig, J. E., and Kamendulis, L. M. (2004). The role of oxidative stress in carcinogenesis. Annu Rev Pharmacol Toxicol 44, 239–267. Klein, G., Imreh, S., and Zabarovsky, E. R. (2007). Why do we not all die of cancer at an early age? Adv Cancer Res 98, 1–16. Klein, G., and Klein, E. (1985). Evolution of tumours and the impact of molecular oncology. Nature 315(6016), 190–195. Kondoh, H. (2008). Cellular life span and the Warburg effect. Exp Cell Res 314(9), 1923–1928. Kulesz-Martin, M. F., Koehler, B., Hennings, H., and Yuspa, S. H. (1980). Quantitative assay for carcinogen altered differentiation in mouse epidermal cells. Carcinogenesis 1(12), 995–1006. Kuper, H., Boffetta, P., and Adami, H. O. (2002). Tobacco use and cancer causation: Association by tumour type. J Intern Med 252(3), 206–224. Lahad, J. P., Mills, G. B., and Coombes, K. R. (2005). Stem cell-ness: A “magic marker” for cancer. J Clin Invest 115(6), 1463–1467. Lassus, P., Opitz-Araya, X., and Lazebnik, Y. (2002). Requirement for caspase-2 in stress-induced apoptosis before mitochondrial permeabilization. Science 297(5585), 1352–1354. Le, N. H., Franken, P., and Fodde, R. (2008). Tumour–stroma interactions in colorectal cancer: Converging on beta-catenin activation and cancer stemness. Br J Cancer 98(12), 1886–1893. Lelbach, A., Muzes, G., and Feher, J. (2007). Current perspectives of catabolic mediators of cancer cachexia. Med Sci Monit 13(9), RA168–RA173. Li, C., Heldt, D., Dalerba, P., Burant, C., Zhang, L., Adsay, V., Wicha, M., et al. (2007). Identification of pancreatic cancer stem cells. Cancer Res 67(3), 1030–1037. Lichtenstein, P., Holm, N. V., Verkasalo, P. K., Iliadou, A., Kaprio, J., Koskenvuo, M., Pukkala, E., Skytthe, A., and Hemminki, K. (2000). Environmental and heritable factors in the causation of cancer—analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med 343(2), 78–85. Liu, G., Meng, X., Jin, Y., Bai, J., Zhao, Y., Cui, X., Chen, F., and Fu, S. (2008a). Inhibitory role of focal adhesion kinase on anoikis in the lung cancer cell A549. Cell Biol Int 32(6), 663– 670. Liu, S., Dontu, G., and Wicha, M. S. (2005). Mammary stem cells, self-renewal pathways, and carcinogenesis. Breast Cancer Res 7(3), 86–95. Liu, S., Ginestier, C., Charafe-Jauffret, E., Foco, H., Kleer, C. G., Merajver, S. D., Dontu, G., and Wicha, M. S. (2008b). BRCA1 regulates human mammary stem/progenitor cell fate. Proc Natl Acad Sci USA 105(5), 1680–1685. Lobo, N. A., Shimono, Y., Qian, D., and Clarke, M. F. (2007). The biology of cancer stem cells. Annu Rev Cell Dev Biol 23, 675–699. Loeb, L. A., Bielas, J. H., and Beckman, R. A. (2008). Cancers exhibit a mutator phenotype: Clinical implications. Cancer Res 68(10), 3551–3557. Loeb, L. A., Loeb, K. R., and Anderson, J. P. (2003). Multiple mutations and cancer. Proc Natl Acad Sci USA 100(3), 776–781.

162


Loeffler, M., Birke, A., Winton, D., and Potten, C. (1993). Somatic mutation, monoclonality and stochastic models of stem cell organization in the intestinal crypt. J Theor Biol 160(4), 471–491. Luis, N., Lopez-Knowles, E., Real, F. (2007). Molecular biology of bladder cancer. Clin Transl Oncol 9(1), 5–12. Luzzi, K. J., MacDonald, I. C., Schmidt, E. E., Kerkvliet, N., Morris, V. L., Chambers, A. F., and Groom, A. C. (1998). Multistep nature of metastatic inefficiency: Dormancy of solitary cells after successful extravasation and limited survival of early micrometastases. Am J Pathol 153(3), 865–873. Lyman, G. H. (1992). Risk factors for cancer. Prim Care 19(3), 465–479. Ma, L., and Weinstein, R. A. (2008). MicroRNAs in malignant progression. Cell Cycle 7(5), 570–572. Maeda, S., Kamata, H., Luo, J. L., Leffert, H., and Karin, M. (2005). IKKbeta couples hepatocyte death to cytokine-driven compensatory proliferation that promotes chemical hepatocarcinogenesis. Cell 121(7), 977–990. Magee, P. N., and Barnes, J. M. (1967). Carcinogenic nitroso compounds. Adv Cancer Res 10, 163–246. Makunin, I. V., Pheasant, M., Simons, C., and Mattick, J. S. (2007). Orthologous microRNA genes are located in cancer-associated genomic regions in human and mouse. PLoS ONE 2(11), e1133. Malarkey, D. E., and Maronpot, R. R. (2005). Carcinogenesis. In Encyclopedia of Toxicology, 2nd edition Wexler, P., ed., Elsevier, New York, pp. 445–466. Maresca, B., and Schwartz, J. H. (2006). Sudden origins: A general mechanism of evolution based on stress protein concentration and rapid environmental change. Anat Rec B New Anat 289(1), 38–46. Margulis, A., Zhang, W., lt-Holland, A., Crawford, H. C., Fusenig, N. E., and Garlick, J. A. (2005). E-cadherin suppression accelerates squamous cell carcinoma progression in three-dimensional, human tissue constructs. Cancer Res 65(5), 1783–1791. Marks, F., Furstenberger, G., and Muller-Decker, K. (2007). Tumor promotion as a target of cancer prevention. Recent Results Cancer Res 174, 37–47. Maronpot, R. R. (2007). Cancer bioassays. In Cancer Handbook, 2nd edition Alison, M., ed., John Wiley & Sons., West Sussex, England, pp. 1–25. Maronpot, R. R., Pitot, H. C., and Peraino, C. (1989). Use of rat liver altered focus models for testing chemicals that have completed two-year carcinogenicity studies. Toxicol Pathol 17(4 Pt 1), 651–662. Martin, M., Simon-Assmann, P., Kedinger, M., Martin, M., Mangeat, P., Real, F. X., and Fabre, M. (2006). DCC regulates cell adhesion in human colon cancer derived HT-29 cells and associates with ezrin1. Eur J Cell Biol 85(8), 769–783. Mauri, F., McNamee, L. M., Lunardi, A., Chiacchiera, F., Del Sal, G., Brodsky, M. H., and Collavin, L. (2008). Modification of Drosophila p53 by SUMO modulates its transactivation and pro-apoptotic functions. J Biol Chem 283(30), 20848–20856. Meuth, M. (1990). The structure of mutation in mammalian cells. Biochim Biophys Acta 1032(1), 1–17. Mimeaut, M., and Batra, S. (2006). Recent Advances on multiple tumorigenic cascades involved in prostatic cancer progression and targeting therapies. Carcinogenesis 27(1), 1–22. Mintz, B., and Illmensee, K. (1975). Normal genetically mosaic mice produced from malignant teratocarcinoma cells. Proc Natl Acad Sci USA 72(9), 3585–3589. Mondal, S., and Heidelberger, C. (1980). Inhibition of induced differentiation of C3H/10T 1/2 clone 8 mouse embryo cells by tumor promoters. Cancer Res 40(2), 334–338. Moore, R. J., Owens, D. M., Stamp, G., Arnott, C., Burke, F., East, N., Holdsworth, H., Turner, L., Rollins, B., Pasparakis, M., Kollias, G., and Balkwill, F. (1999). Mice deficient in tumor necrosis factor-alpha are resistant to skin carcinogenesis. Nat Med 5(7), 828–831. Morange, M. (2002). The relations between genetics and epigenetics: A historical point of view. Ann NY Acad Sci 981, 50–60. Morange, M. (2007). The field of cancer research: An indicator of present transformations in biology. Oncogene 26(55), 7607–7610. Morange, M. (2008). What history tells us XIII. Fifty years of the central dogma. J Biosci 33(2), 171–175. Moss, T. J., and Wallrath, L. L. (2007). Connections between epigenetic gene silencing and human disease. Mutat Res 618(1–2), 163–174.

REFERENCES

163

Mueller, M. M., and Fusenig, N. E. (2004). Friends or foes—Bipolar effects of the tumour stroma in cancer. Nat Rev Cancer 4(11), 839–849. Murakami, S., Noguchi, T., Takeda, K., and Ichijo, H. (2007). Stress signaling in cancer. Cancer Sci 98(10), 1521–1527. Murphy, J. B., and Sturm, E. (1925). Primary lung tumors in mice following the cutaneous application of coal tar. J Exp Med 42, 693–700. Musracchi, P., and Shimkin, M. B. (1956). Gendron’s enquiries into the nature, knowledge, and cure of cancers. Cancer 9(4), 645–647. Nakayama, J., Yuspa, S. H., and Poirier, M. C. (1984). Benzo(a)pyrene-DNA adduct formation and removal in mouse epidermis in vivo and in vitro: Relationship of DNA binding to initiation of skin carcinogenesis. Cancer Res 44(9), 4087–4095. Nigro, J. M., Cho, K. R., Fearon, E. R., Kern, S. E., Ruppert, J. M., Oliner, J. D., Kinzler, K. W., and Vogelstein, B. (1991). Scrambled exons. Cell 64(3), 607–613. Nishikawa, M. (2008). Reactive oxygen species in tumor metastasis. Cancer Lett 266(1), 53–59. Nowell, P. C. (1976). The clonal evolution of tumor cell populations. Science 194(4260), 23–28. Nowell, P. C. (1986). Mechanisms of tumor progression. Cancer Res 46(5), 2203–2207. Ohlsson, R., Kanduri, C., Whitehead, J., Pfeifer, S., Lobanenkov, V., and Feinberg, A. P. (2003). Epigenetic variability and the evolution of human cancer. Adv Cancer Res 88, 145–168. Oliveira, P. A., Colaco, A., Chaves, R., Guedes-Pinto, H., De-La-Cruz, P. L., and Lopes, C. (2007). Chemical carcinogenesis. An Acad Bras Cienc 79(4), 593–616. Orimo, A., Gupta, P. B., Sgroi, D. C., Renzana-Seisdedos, F., Delaunay, T., Naeem, R., Carey, V. J., Richardson, A. L., and Weinberg, R. A. (2005a). Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion. Cell 121(3), 335–348. Orimo, A., Gupta, P. B., Sgroi, D. C., Renzana-Seisdedos, F., Delaunay, T., Naeem, R., Carey, V. J., Richardson, A. L., and Weinberg, R. A. (2005b). Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion. Cell 121(3), 335–348. Orimo, A., and Weinberg, R. A. (2006). Stromal fibroblasts in cancer: A novel tumor-promoting cell type. Cell Cycle 5(15), 1597–1601. Orimo, A., and Weinberg, R. A. (2007a). Heterogeneity of stromal fibroblasts in tumors. Cancer Biol Ther 6(4), 618–619. Orimo, A., and Weinberg, R. A. (2007b). Heterogeneity of stromal fibroblasts in tumors. Cancer Biol Ther 6(4), 618–619. Peraino, C., Fry, R. J., Staffeldt, E., and Christopher, J. P. (1975). Comparative enhancing effects of phenobarbital, amobarbital, diphenylhydantoin, and dichlorodiphenyltrichloroethane on 2-acetylaminofluorene-induced hepatic tumorigenesis in the rat. Cancer Res 35(10), 2884–2890. Pereira, M. A., and Stoner, G. D. (1985). Comparison of rat liver foci assay and strain A mouse lung tumor assay to detect carcinogens: A review. Fundam Appl Toxicol 5(4), 688–699. Perwez, H. S., and Harris, C. C. (2007). Inflammation and cancer: An ancient link with novel potentials. Int J Cancer 121(11), 2373–2380. Peterson, J. A. (1983). The widespread nature of phenotypic variability in hepatomas and cell lines, in the form of a geometric series. J Theor Biol 102(1), 41–53. Pikarsky, E., Porat, R. M., Stein, I., Abramovitch, R., Amit, S., Kasem, S., Gutkovich-Pyest, E., UrieliShoval, S., Galun, E., and Ben-Neriah, Y. (2004). NF-kappaB functions as a tumour promoter in inflammation-associated cancer. Nature 431(7007), 461–466. Pitot, H. C., Dragan, Y. P., Teeguarden, J., Hsia, S., and Campbell, H. (1996). Quantitation of multistage carcinogenesis in rat liver. Toxicol Pathol 24(1), 119–128. Pitot, H. C., Goldsworthy, T. L., Moran, S., Kennan, W., Glauert, H. P., Maronpot, R. R., and Campbell, H. A. (1987). A method to quantitate the relative initiating and promoting potencies of hepatocarcinogenic agents in their dose–response relationships to altered hepatic foci. Carcinogenesis 8(10), 1491–1499. Pitot, H. C., Grosso, L. E., and Goldsworthy, T. (1985). Genetics and epigenetics of neoplasia: Facts and theories. Carcinog Compr Surv 10, 65–79. Pitts, J. D., Finbow, M. E., and Kam, E. (1988). Junctional communication and cellular differentiation. Br J Cancer Suppl 9, 52–57.

164


Polyak, K. (2007). Breast cancer: Origins and evolution. J Clin Invest 117(11), 3155–3163. Ponting, C. P., Oliver, P. L., Relk, W. (2009). Evolution and functions of long noncoding RNAs. Cell 136(4), 629–641. Popp, J. A., and Goldsworthy, T. L. (1989). Defining foci of cellular alteration in short-term and mediumterm rat liver tumor models. Toxicol Pathol 17(4 Pt 1), 561–568. Porta, C., Subhra, K. B., Larghi, P., Rubino, L., Mancino, A., and Sica, A. (2007). Tumor promotion by tumor-associated macrophages. Adv Exp Med Biol 604, 67–86. Prins,G., Tung, W., Belmont, J., Ho, S. (2008). Perinatal Exposure to oestradiol and bisphenol A alters the prostate epigenome and increases susceptibility to carcinogenesis. Basic Clin Pharmacol Toxicol 102, 134–138. Punnett, R. C. (1907). Mendelism, Bowes and Bowes, London, pp. 1–84. Qiu, W. H., Zhou, B. S., Chu, P. G., Chen, W. G., Chung, C., Shih, J., Hwu, P., Yeh, C., Lopez, R., and Yen, Y. (2005). Over-expression of fibroblast growth factor receptor 3 in human hepatocellular carcinoma. World J Gastroenterol 11(34), 5266–5272. Rakoff-Nahoum, S. (2006). Why cancer and inflammation? Yale J Biol Med 79(3–4), 123–130. Rakoff-Nahoum, S., and Medzhitov, R. (2007). Prostaglandin-secreting cells: A portable first aid kit for tissue repair. J Clin Invest 117(1), 83–86. Rangarajan, A., Hong, S. J., Gifford, A., and Weinberg, R. A. (2004). Species- and cell type-specific requirements for cellular transformation. Cancer Cell 6(2), 171–183. Renehan, A. G., Booth, C., and Potten, C. S. (2001). What is apoptosis, and why is it important? Br Med 322(7301), 1536–1538. Reisco-Eizaguirre, G., and Santisteban, P. (2007). Molecular biology of thyroid cancer initiation. Clin Transl Oncol, 9(11), 686–693. Rosenkranz, M., Rosenkranz, H. S., and Klopman, G. (1997). Intercellular communication, tumor promotion and non-genotoxic carcinogenesis: Relationships based upon structural considerations. Mutat Res 381(2), 171–188. Rubin, H. (1984). Early origin and pervasiveness of cellular heterogeneity in some malignant transformations. Proc Natl Acad Sci USA 81(16), 5121–5125. Rubin, H. (1992). Cancer development: the rise of epigenetics. Eur J Cancer 28(1), 1–2. Rubin, H. (1994). Incipient and overt stages of neoplastic transformation. Proc Natl Acad Sci USA 91(25), 12076–12080. Rubin, H. (1999). Cell damage, aging and transformation: a multilevel analysis of carcinogenesis. Anticancer Res 19(6A), 4877–4886. Rubin, H. (2001). Multistage carcinogenesis in cell culture. Dev Biol (Basel) 106, 61–66. Rubin, H. (2007). Ordered heterogeneity and its decline in cancer and aging. Adv Cancer Res 98, 117–147. Ruggiero, R., and Bustuoabad, O. (2006). The biological sense of cancer: A hypothesis. Theor Bio Med Modelling 3(43), 1–14. Satzinger, H. (2008). Theodor and Marcella Boveri: Chromosomes and cytoplasm in heredity and development. Nat Rev Genet 9(3), 231–238. Scheel, C., Onder, T., Karnoub, A., and Weinberg, R. A. (2007). Adaptation versus selection: The origins of metastatic behavior. Cancer Res 67(24), 11476–11479. Schiestl, R. H. (1993). Nonmutagenic carcinogens induce intrachromosomal recombination in dividing yeast cells. Environ Health Perspect 101(Suppl 5), 179–184. Schleiden, M., and Schwann, T. (1839). Beitrage zur Phytogenesis. Translated along with Schwann’s Mikroskopische Untersuchungen by H. Smith for the Sydenham Society, London, in 1947. Schmidt-Hansen, B., Klingelhofer, J., Grum-Schwensen, B., Christensen, A., Andresen, S., Kruse, C., Hansen, T., Ambartsumian, N., Lukanidin, E., and Grigorian, M. (2004). Functional significance of metastasis-inducing S100A4(Mts1) in tumor–stroma interplay. J Biol Chem 279(23), 24498–24504. Schones, D. E., and Zhao, K. (2008). Genome-wide approaches to studying chromatin modifications. Nat Rev Genet 9(3), 179–191. Schones, D. E., Cui, K., Cuddapah, S., Roh, T. Y., Barski, A., Wang, Z., Wei, G., and Zhao, K. (2008). Dynamic regulation of nucleosome positioning in the human genome. Cell 132(5): 887–898. Sell, S. (2004). Stem cell origin of cancer and differentiation therapy. Crit Rev Oncol Hematol 51(1), 1–28.

REFERENCES

165

Sharma-Walia, N., Raghu, H., Sadagopan, S., Sivakumar, R., Veettil, M. V., Naranatt, P. P., Smith, M. M., and Chandran, B. (2006). Cyclooxygenase 2 induced by Kaposi’s sarcoma-associated herpesvirus early during in vitro infection of target cells plays a role in the maintenance of latent viral gene expression. J Virol 80(13), 6534–6552. Shepard, P. J., and Hertel, K. J. (2008). Conserved RNA secondary structures promote alternative splicing. RNA 14(8), 1463–1469. Shimkin, M. B. (1980). Some Classics of Experimental Oncology, National Institutes of Health (NIH), pp. 1–739. Shimkin, M. B., and Triolo, V. A. (1969). History of chemical carcinogenesis: Some prospective remarks. Prog Exp Tumor Res 11, 1–20. Shipitsin, M., and Polyak, K. (2008). The cancer stem cell hypothesis: In search of definitions, markers, and relevance. Lab Invest 88(5), 459–463. Sieweke, M. H., Stoker, A. W., and Bissell, M. J. (1989). Evaluation of the cocarcinogenic effect of wounding in Rous sarcoma virus tumorigenesis. Cancer Res 49(22), 6419–6424. Slaga, T. J. (ed.). (1983). Tumor Promotion in Internal Organs, Volume I of Mechanisms of Tumor Promotion, CRC Press, Boca Raton, FL, pp. 1–179. Slaga, T. J. (ed.). (1984a). Tumor Promotion and Skin Carcinogenesis, Volume II of Mechanisms of Tumor Promotion, CRC Press, Boca Raton, FL, pp. 1–205. Slaga, T. J. (ed.). (1984b). Tumor Promotion and Carcinogensis in Vitro, Volume III of Mechanisms of Tumor Promotion, CRC Press, Boca Raton, FL, pp. 1–192. Slaga, T. J. (ed.). (1984c). Tumor Promotion in Internal Organs, Volume IV of Mechanisms of Tumor Promotion, CRC Press, Boca Raton, FL, pp. 1–154. Smela, M. E., Currier, S. S., Bailey, E. A., and Essigmann, J. M. (2001). The chemistry and biology of aflatoxin B(1): From mutational spectrometry to carcinogenesis. Carcinogenesis 22(4), 535–545. Smyth, M. J., Dunn, G. P., and Schreiber, R. D. (2006). Cancer immunosurveillance and immunoediting: The roles of immunity in suppressing tumor development and shaping tumor immunogenicity. Adv Immunol 90, 1–50. Soengas, M. S., Alarcon, R. M., Yoshida, H., Giaccia, A. J., Hakem, R., Mak, T. W., and Lowe, S. W. (1999). Apaf-1 and caspase-9 in p53-dependent apoptosis and tumor inhibition. Science 284(5411), 156–159. Soengas, M. S., Capodieci, P., Polsky, D., Mora, J., Esteller, M., Opitz-Araya, X., McCombie, R., Herman, J. G., Gerald, W. L., Lazebnik, Y. A., Cordon-Cardo, C., and Lowe, S. W. (2001). Inactivation of the apoptosis effector Apaf-1 in malignant melanoma. Nature 409(6817), 207–211. Solt, D. B., Cayama, E., Tsuda, H., Enomoto, K., Lee, G., and Farber, E. (1983). Promotion of liver cancer development by brief exposure to dietary 2-acetylaminofluorene plus partial hepatectomy or carbon tetrachloride. Cancer Res 43(1), 188–191. Stein, L. D. (2004). Human genome: End of the beginning, Nature 431, 915–916. Stewart, S. A., and Weinberg, R. A. (2006). Telomeres: cancer to human aging. Annu Rev Cell Dev Biol 22, 531–557. Sun, W., Kang, K. S., Morita, I., Trosko, J. E., and Chang, C. C. (1999). High susceptibility of a human breast epithelial cell type with stem cell characteristics to telomerase activation and immortalization. Cancer Res 59(24), 6118–6123. Süss, R., Kinzel, V., and Scribner, J. D. (1973). Cancer: Experiment and Concepts, Springer-Verlag, New York, pp. 1–285. Szyf, M. (2007). The dynamic epigenome and its implications in toxicology. Toxicol Sci 100(1), 7–23. Tai, M. H., Chang, C. C., Kiupel, M., Webster, J. D., Olson, L. K., and Trosko, J. E. (2005). Oct4 expression in adult human stem cells: Evidence in support of the stem cell theory of carcinogenesis. Carcinogenesis 26(2), 495–502. Talmadge, J. E. (2007). Clonal selection of metastasis within the life history of a tumor. Cancer Res 67(24), 11471–11475. Talmadge, J. E., Wolman, S. R., and Fidler, I. J. (1982). Evidence for the clonal origin of spontaneous metastases. Science 217(4557), 361–363. Tasdemir, E., Maiuri, M. C., Galluzzi, L., Vitale, I., Djavaheri-Mergny, M., D’Amelio, M., Criollo, A., Morselli, E., Zhu, C., Harper, F., Nannmark, U., Samara, C., Pinton, P., Vicencio, J. M., Carnuccio, R., Moll, U. M., Madeo, F., Paterlini-Brechot, P., Rizzuto, R., Szabadkai, G., Pierron, G., Blomgren,

166


K., Tavernarakis, N., Codogno, P., Cecconi, F., and Kroemer, G. (2008). Regulation of autophagy by cytoplasmic p53. Nat Cell Biol 10(6), 676–687. Tejpar, S., Michils, G., Denys, H., Van, D. K., Nik, S. A., Jadidizadeh, A., and Cassiman, J. J. (2005). Analysis of Wnt/Beta catenin signalling in desmoid tumors. Acta Gastroenterol Belg 68(1), 5–9. Tennant, R. W. (1993). A perspective on nonmutagenic mechanisms in carcinogenesis. Environ Health Perspect 101(Suppl 3), 231–236. Tiano, H. F., Loftin, C. D., Akunda, J., Lee, C. A., Spalding, J., Sessoms, A., Dunson, D. B., Rogan, E. G., Morham, S. G., Smart, R. C., and Langenbach, R. (2002). Deficiency of either cyclooxygenase (COX)-1 or COX-2 alters epidermal differentiation and reduces mouse skin tumorigenesis. Cancer Res 62(12), 3395–3401. Ting, A. H., McGarvey, K. M., and Baylin, S. B. (2006). The cancer epigenome—Components and functional correlates. Genes Dev 20(23), 3215–3231. Ting, A. H., Schuebel, K. E., Herman, J. G., and Baylin, S. B. (2005). Short double-stranded RNA induces transcriptional gene silencing in human cancer cells in the absence of DNA methylation. Nat Genet 37(8), 906–910. Tisdale, M. J. (1997). Biology of cachexia. J Natl Cancer Inst 89(23), 1763–1773. Tononi, G., Sporns, O., and Edelman, G. (1999). Measures of degeneracy and redundancy in biological networks. Proc Natl Acad Sci 96, 3257–3263. Trent, J. M., Wiltshire, R., Su, L. K., Nicolaides, N. C., Vogelstein, B., and Kinzler, K. W. (1995). The gene for the APC-binding protein beta-catenin (CTNNB1) maps to chromosome 3p22, a region frequently altered in human malignancies. Cytogenet Cell Genet 71(4), 343–344. Triolo, V. A. (1965). Nineteenth century foundations of cancer research advances in tumor pathology, nomenclature, and theories of oncogenesis. Cancer Res 25, 75–106. Troll, W., and Wiesner, R. (1985). The role of oxygen radicals as a possible mechanism of tumor promotion. Annu Rev Pharmacol Toxicol 25, 509–528. Trosko, J. E. (1988). A failed paradigm: Carcinogenesis is more than mutagenesis. Mutagenesis 3(4), 363–366. Trosko, J. E. (2005). The role of stem cells and cell–cell communication in radiation carcinogenesis: Ignored concepts. BJR Suppl 27, 132–138. Trosko, J. E. (2007). Gap junctional intercellular communication as a biological “Rosetta stone” in understanding, in a systems biological manner, stem cell behavior, mechanisms of epigenetic toxicology, chemoprevention and chemotherapy. J Membr Biol 218(1–3), 93–100. Trosko, J. E., and Chang, C. C. (1989). Stem cell theory of carcinogenesis. Toxicol Lett 49(2–3), 283–295. Trosko, J. E., Chang, C. C., Upham, B. L., and Tai, M. H. (2004). Ignored hallmarks of carcinogenesis: Stem cells and cell–cell communication. Ann NY Acad Sci 1028, 192–201. Trosko, J. E., Dawson, B., and Chang, C. C. (1981). PBB inhibits metabolic cooperation in Chinese hamster cells in vitro: Its potential as a tumor promoter. Environ Health Perspect 37, 179–182. Trosko, J. E., Jone, C., and Chang, C. C. (1984). The use of in-vitro assays to study and to detect tumour promoters. IARC Sci Publ 56, 239–252. Trosko, J. E., Yotti, L. P., Warren, S. T., Tsushimoto, G., and Chang, C. (1982). Inhibition of cell–cell communication by tumor promoters. Carcinog Compr Surv 7, 565–585. Ulrich, C. M., Bigler, J., and Potter, J. D. (2006). Non-steroidal anti-inflammatory drugs for cancer prevention: promise, perils and pharmacogenetics. Nat Rev Cancer 6(2), 130–140. Upham, B. L., Guzvic, M., Scott, J., Carbone, J. M., Blaha, L., Coe, C., Li, L. L., Rummel, A. M., and Trosko, J. E. (2007). Inhibition of gap junctional intercellular communication and activation of mitogen-activated protein kinase by tumor-promoting organic peroxides and protection by resveratrol. Nutr Cancer 57(1), 38–47. Upham, B. L., Masten, S. J., Lockwood, B. R., and Trosko, J. E. (1994). Nongenotoxic effects of polycyclic aromatic hydrocarbons and their oxygenation by-products on the intercellular communication of rat liver epithelial cells. Fundam Appl Toxicol 23(3), 470–475. Upham, B. L., and Trosko, J. E. (2009). Oxidative-dependent integration of signal transduction with intercellular gap junctional communication in the control of gene expression. Antioxidants and Redox Signalling 11(2), 1–11. Vasiliev, J. M. (2004). Cytoskeletal mechanisms responsible for invasive migration of neoplastic cells. Int J Dev Biol 48(5–6), 425–439.

REFERENCES

167

Vasiliev, J. M., and Guelstein, V. I. (1966). Local cell interactions in neoplasms and in the foci of carcinogenesis. Prog Exp Tumor Res 8, 26–65. Verma, M., Maruvada, P., and Srivastava, S. (2004). Epigenetics and cancer. Crit Rev Clin Lab Sci 41(5–6), 585–607. Vicencio, J. M., Galluzzi, L., Tajeddine, N., Ortiz, C., Criollo, A., Tasdemir, E., Morselli, E., Ben, Y. A., Maiuri, M. C., Lavandero, S., and Kroemer, G. (2008). Senescence, apoptosis or autophagy? When a damaged cell must decide its path—A mini-review. Gerontology 54(2), 92–99. von Fournier, F. D., Weber, E., Hoeffken, W., Bauer, M., Kubli, F., and Barth, V. (1980). Growth rate of 147 mammary carcinomas. Cancer 45(8), 2198–2207. Wallace, K. B. (book editor) (1997). Free Radical Toxicology, Target Organ Toxicology series, Hayes A. W., Thomas, J. A., Garner, D. E., series editors, Taylor & Francis, Washington, D.C., pp. 1–442. Wang, E., Lenferink, A., and O’Connor-McCourt, M. (2007). Cancer systems biology: Exploring cancerassociated genes on cellular networks. Cell Mol Life Sci 64(14), 1752–1762. Wang, Z., Zang, C., Rosenfeld, J. A., Schones, D. E., Barski, A., Cuddapah, S., Cui, K., Roh, T. Y., Peng, W., Zhang, M. Q., and Zhao, K. (2008). Combinatorial patterns of histone acetylations and methylations in the human genome. Nat Genet 40(7), 897–903. Warburg, O. (1925). Über de Stoffwechsel der Carcinomzelle. Klin Wochenschr 4, 534–536. Warburg, O. (1956a). On respiratory impairment in cancer cells. Science 124(3215), 269–270. Warburg, O. (1956b). On the origin of cancer cells. Science 123(3191), 309–314. Warren, S. T., Yotti, L. P., Moskal, J. R., Chang, C. C., and Trosko, J. E. (1981). Metabolic cooperation in CHO and V79 cells following treatment with a tumor promoter. Exp Cell Res 131(2), 427–430. Watabe, T. (1983). Metabolic activation of 7,12-dimethylbenz[alpha]anthracene (DMBA) and 7-methylbenz[alpha]anthracene (7-MBA) by rat liver P-450 and sulfotransferase 1. J Toxicol Sci 8(2), 119–131. Weinberg, R. A. (1989). Oncogenes, antioncogenes, and the molecular bases of multistep carcinogenesis. Cancer Res 49(14), 3713–3721. Weinberg, R. A. (1997). The cat and mouse games that genes, viruses, and cells play. Cell 88(5), 573–575. Weinberg, R. A. (2008a). Coevolution in the tumor microenvironment. Nat Genet 40(5), 494–495. Weinberg, R. A. (2008b). Mechanisms of malignant progression. Carcinogenesis 29(6), 1092–1095. Wicha, M. S., Liu, S., and Dontu, G. (2006). Cancer stem cells: An old idea—A paradigm shift. Cancer Res 66(4), 1883–1890. Williams, C. S., Mann, M., and DuBois, R. N. (1999). The role of cyclooxygenases in inflammation, cancer, and development. Oncogene 18(55), 7908–7916. Wogan, G. N., Hecht, S. S., Felton, J. S., Conney, A. H., and Loeb, L. A. (2004). Environmental and chemical carcinogenesis. Semin Cancer Biol 14(6), 473–486. Wolters, N. M., and MacKeigan, J. P. (2008). From sequence to function: Using RNAi to elucidate mechanisms of human disease. Cell Death Differ 15(5), 809–819. Woods, N. T., Yamaguchi, H., Lee, F. Y., Bhalla, K. N., and Wang, H. G. (2007). Anoikis, initiated by Mcl-1 degradation and Bim induction, is deregulated during oncogenesis. Cancer Res 67(22), 10744–10752. Wynter, C. V. (2006). The dialectics of cancer: A theory of the initiation and development of cancer through errors in RNAi. Med Hypotheses 66(3), 612–635. Yamagiwa, K., and Ichikawa, K. (1918). Experimental study of the pathogeneisis of carcinoma. J Cancer Res 3, 1–29. Yamagiwa, K., and Ichikawa, K. (1977). Experimental study of the pathogenesis of carcinoma. CA Cancer J Clin 27(3), 174–181. Yotti, L. P., Chang, C. C., and Trosko, J. E. (1979). Elimination of metabolic cooperation in Chinese hamster cells by a tumor promoter. Science 206(4422), 1089–1091. Yuspa, S. H., Hawley-Nelson, P., Koehler, B., and Stanley, J. R. (1980). A survey of transformation markers in differentiating epidermal cell lines in culture. Cancer Res 40(12), 4694–4703. Yuspa, S. H., and Poirier, M. C. (1988). Chemical carcinogenesis: From animal models to molecular models in one decade. Adv Cancer Res 50, 25–70. Zhang, Y. W., and Vande Woude, G. F. (2007). Mig-6, signal transduction, stress response and cancer. Cell Cycle 6(5), 507–513.

CH A P TE R

6

CHEMICAL CARCINOGENESIS: A BRIEF HISTORY OF ITS CONCEPTS WITH A FOCUS ON POLYCYCLIC AROMATIC HYDROCARBONS Stephen Nesnow

6.1. A BRIEF HISTORY OF CHEMICAL CARCINOGENESIS As discussed in Chapter 5, the first reported example of a carcinogenic exposure that led to human cancer is ascribed to Sir John Percival Pott.* Pott was born in 1714 and became a respected surgeon who practiced at St. Bartholomew’s Hospital, London, Great Britain. In his practice, Dr. Pott observed “sores” on the scrotums of chimney sweeps in London. While other surgeons presumed that these were the results of venereal disease, Dr. Pott realized that they were some kind of skin cancer. He surmised that the cause of the cancers was “a lodgement of soot in the ruggae of the scrotum.” In 1775 he reported these findings in “Chirugical Observations Relative to the Cataract, the Polypus of the Nose, the Cancer of the Scrotum, the Different Kinds of Ruptures, and the Mortification of Toes and Feet.” This publication was the first in epidemiology that related external exposures of coal tar/soot to a human cancer (Pott 1775). Originally termed Pott’s cancer, it is more commonly referred to as chimney-sweep’s cancer. One hundred and forty years later, the first experimental evidence that coal tar was carcinogenic came from two Japanese pathologists, Katsusaburo Yamagiwa and *This manuscript has been reviewed by the National Health and Environmental Effects Research Laboratory at the U.S. Environmental Protection Agency and approved for publication. The views expressed in this chapter are those of the authors and do not necessarily represent the views or policy of the U.S. Environmental Protection Agency. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.


168

6.2. JAMES A. AND ELIZABETH C. MILLER AND THEIR THEORY OF METABOLIC ACTIVATION

169

Koichi Ichikawa at the University of Tokyo in 1915 (Yamagiwa and Ichikawa 1915). This was also the first example of experimental chemical carcinogenesis. Professors Yamagiwa and Ichikawa were testing the irritation theory of cancer proposed by the Danish pathologist, Johannes Andreas Grib Fibiger, based on the earlier writings of Julius Vogel (1814–1880) and Rudolph Virchow (1821–1902). Fibiger achieved the first controlled induction of cancer in laboratory animals in 1913 by feeding mice cockroaches infected with the worm Gongylonema neoplasticum (Fibiger 1913). The larvae of a worm induced a chronic inflammation of stomach tissue, eventually inducing gastric tumors. Fibiger received the Nobel Prize for this research in 1926. Yamagiwa and Ichikawa repeatedly painted coal tar on the ears of rabbits and succeeded in producing multiple squamous cell carcinomas, a range of benign and malignant hyperplastic lesions, and inflammatory changes in the painted areas. They were also the first to describe the complexity and progressive nature of the carcinogenic process as they identified the conversion of less malignant to more malignant tumor cells as well as the regression of benign tumors (Yamagiwa and Ichikawa 1915). Years later the search for the active carcinogenic components in coal tar began in the laboratory of Ian Heiger at the Institute for Cancer Research in Great Britain, who isolated carcinogenic polycyclic aromatic hydrocarbons (PAHs) from coal tar (Cook et al. 1933). Using two tons of coal tar pitch, they isolated several components, one of which was highly carcinogenic on mouse skin. It was identified as benzo[a]pyrene (B[a]P). B[a]P was then synthesized, and the synthetic material was also found to be highly tumorigenic on mouse skin (Kennaway 1955). Since then, B[a]P has remained the archetypical PAH used for the study of the mechanisms of chemical carcinogenesis and is the most widely and thoroughly studied PAH in this chemical class. Because B[a]P is a product of the incomplete combustion of fossil fuels, it is pervasive in the environment. Structurally, B[a]P is a fused pentacyclic PAH and has been found to be tumorigenic in almost every species tested by many different routes of exposure: mice (dermal, subcutaneous, intraperitoneal, feed), rats (subcutaneous, inhalation, intratracheal), hamsters (intratracheal), rabbit (dermal), fish (water), and dogs (endobronchial). The target organs for B[a]P-induced neoplasia include: skin, forestomach, lung, liver, mammary gland, esophagus, and tongue (see references for details) (Grimmer et al. 1987). Based on the preponderance of mechanistic, experimental, and epidemiological data, the International Agency for Research on Cancer (IARC) has recently classified both B[a]P and the occupation of chimney sweeping as human carcinogens (Straif et al. 2005).

6.2. JAMES A. AND ELIZABETH C. MILLER AND THEIR THEORY OF METABOLIC ACTIVATION By the mid-20th century it was known that a diverse group of chemicals were either experimental carcinogens or were associated with human neoplasia from epidemiological investigations of workplace exposures. The variety of chemical structures with this group of carcinogens defied a unifying, common mechanism that could

170

CHAPTER 6 CHEMICAL CARCINOGENESIS

explain their tumorigenic activities and a common unifying theory which explained their mechanisms of action. The concept of metabolic activation and electrophilicity—now a well-accepted mechanism of action of many, if not all, genotoxic carcinogens—was first described by the late James A. Miller (1915–2000) and Elizabeth C. Miller (1920–1987) based on a number of studies conducted in the late 1950s at the McArdle Laboratory for Cancer Research in Madison, Wisconsin. It had been reported that many classes of chemical carcinogens (PAHs, nitrosamines, aflatoxins, aromatic amines) covalently bound to cellular macromolecules, DNA, RNA, and protein in target tissues based on radiometric techniques. However, none of these classes of carcinogens covalently bound to these isolated macromolecules in the test tube. The Millers reasoned that in cells and tissues there must be enzymes that metabolize these agents to chemically reactive forms which then bind to the most likely macromolecular target for cancer, namely, DNA. It is instructive to review a brief example of their work on aromatic amines. Aromatic amines were first associated with human bladder cancer in 1895 based on observations of aniline dye workers (Rehn 1895). It was later determined that a series of related chemicals were also associated with this cancer: benzidine and β-naphthylamine (2-naphthylamine) (Case et al. 1954). Critical research on metabolic activation carried out by the Millers used 2-acetylaminofluorene (AAF), a potent carcinogen in the liver, bladder, intestine, and mammary gland. They found that AAF was N-hydroxylated to its proximate carcinogenic form N-hydroxy-AAF, which was converted into it ultimate carcinogenic form, a sulfate ester by hepatic 3′-phosphoadenosine 5′-phosphosulfate (PAPS) sulfotransferases (Figure 6.1). This sulfate ester could form an incipient nitrenium ion that bound covalently to DNA, and this initiated the cancer process. Using the model N-acetoxyAAF, the C8 of deoxyguanosine was the target in DNA yielding the N-(deoxyguanosin-8-yl)-AAF adduct. Other AAF adducts have been identified: the nonacetylated N-(deoxyguanosine-8-yl)-AF adduct as well as the 3-(deoxyguanosine-N2-yl)-AAF adduct arising from the nitrenium ion activation of the C3 carbon of AAF (Beland and Kadlubar 1985). In 1960, the Millers reported that a metabolite (N-hydroxy-AAF) proved to be much more carcinogenic than its parent compound (AAF) and produced tumors in tissues including the site of administration (Miller et al. 1960). This research demonstrated that for many carcinogens, the initiation of cancer depended on metabolic activation of parent molecules to electrophiles, a major unifying concept of their research. The conclusions of their research on metabolic activation were as follows: (1) Chemical carcinogens that are not themselves chemically reactive must be metabolically converted into a chemically reactive form. (2) The activated metabolite must be an electrophilic form in order to bind to DNA. (3) The covalent adducts to DNA that formed can initiate the process of carcinogenesis. Moreover, a common feature of many diverse chemical carcinogens is that their reactive forms were electrophiles (Heidelberger 1975). The Millers also found that many of the enzymes that participated in the metabolic activation process were the Phase I microsomal mixed function oxidases and Phase II enzymes whose functions were to detoxify and to make xenobiotics more water soluble so they could be excreted (Miller and Miller 1979). The Millers laid the foundation for our current understanding of chemical carcinogenesis.

171


Ac NH

P450

Ac N

Proximate carcinogen OH

PAPS sulfotransferase

Ac

Ultimate carcinogen

N

+

O-

SO 3

O

O

Ac N

N

NH

N N

-

N

NH

N

NH2

N

dR

dR

N-(deoxyguanosin-8-yl)-AAF

NH HN

Ac

3-(deoxyguanosin-N2-yl)-AAF

O N

NH

NH N

N

NH2

dR

N-(deoxyguanosin-8-yl)-AF

Figure 6.1. The metabolic activation of AAF as described by the Millers and further refined by Beland and Kadlubar (1985).

The concept of metabolic activation developed with AAF was then applied to PAHs, aflatoxins, nitrosamines, nitrosoureas, hydrazines, urethane, and vinyl chloride. Several metabolic activation schemes are presented in Figure 6.2. In each case a highly reactive electrophilic carbocation is formed. We now know that the concept of metabolic activation applies to many genotoxic carcinogens and helps to explain

172

CHAPTER 6 CHEMICAL CARCINOGENESIS O

O

O O

O

O

O

P450 O

O O

O

CH3

O

N

CH3

Aflatoxin B1-2,3-epoxide

Aflatoxin B1 H3C

O

OH

O

N

P450

NH

O

NH

H3C

H3C

Dimethylnitrosamine O N N

CH3

H2O

OH

N N

H3C H2N

CH3

Methyl carbonium ion

O

Methylnitrosourea P450

H3C

N

CH3

O

P450

NH NH

N

H3C

Dimethylhydrazine

N CH3

N

-

+

H3C

Azomethane

O

OH

Methylazoxymethanol

O

O

P450 H2N

+

O

CH3

H2N

Ethyl carbamate (urethane)

O

O

P450 CH2

H2N

Vinyl carbamate

O

Vinyl carbamate epoxide

O

Cl

P450

Cl

H2C Vinyl chloride

Vinyl chloride epoxide

Figure 6.2. Metabolic activation schemes for a group of genotoxic carcinogens. In each example the parent carcinogen is converted into a more reactive electrophilic form that alkylates DNA through carbocation formation.

their mechanisms of action. There are other classes of chemical carcinogens that do not require metabolic activation, namely, the nongenotoxic or epigenetic carcinogens consisting of cytotoxicants, mitogens, peroxisome proliferators, and endocrine disruptors. These chemicals do not bind directly to DNA but generally induce DNA mutations by indirect methods (Williams 2001).


6.2.1.

173

Metabolic Activation of PAH and Tumorigenesis

Probably the most intensive and detailed efforts at delineating metabolic activation mechanisms and their relationships to tumorigenesis have been centered in the study of PAH carcinogenesis. To date, there are four major and several minor theories of the metabolic activation of PAHs. The four major theories will be discussed further in detail using B[a]P as an example, where appropriate. The major hypotheses are the bay- and fjord-region diol epoxide metabolic activation mechanism (Figure 6.3A,B), the radical cation mechanism (Figure 6.3C), the o-quinone/reactive oxygen species (ROS) mechanism (Figure 6.3D), and the cyclopenta-ring oxidation mechanism (Figure 6.3E). 6.2.1.1. Bay- and Fjord-Region Diol Epoxide Metabolic Activation Mechanism. The generalized diol epoxide mechanism was developed from the bay region theory proposed by Jerina et al. (1976) and was based on the earlier observations of the nature of PAH metabolites identified by Boyland and Sims (1964) and the results from a quantum mechanical model. This theory recognized that angular benzo ring fusions on PAHs created a topological indentation on the polycyclic ring structure, called the bay region. For B[a]P the bay region encompasses four carbons (carbons 10, 10a, 10b, and 11) and three carbon–carbon bonds (Figure 6.3A; see Figure 6.4 for carbon numbering). In the example of B[a]P, metabolism by the cytochrome P450 isozymes at the carbon 7-carbon 8 aromatic double bond disrupts the aromatic nucleus by saturating that carbon–carbon bond and creates an arene oxide, B[a]P-7,8-oxide (Figure 6.4). The stereospecific and regiospecific metabolizing activity of each cytochrome P450, in combination with the capability of carbons to form chiral centers through metabolism, can create multiple forms of many PAH metabolites. Therefore, due to the chirality of carbon, two stereoisomeric (enantiomeric) forms of B[a]P-7,8-oxide are created: (−)-B[a]P-7S,8R-oxide and (+)-B[a]P-7R,8S-oxide (Figure 6.4). The (+) and (−) terminology refers to the ability of stereoisomers to rotate polarized light in a clockwise or counterclockwise direction, and the R and S terminology refers to their absolute stereochemistry. B[a]P-7,8-oxide is hydrated by epoxide hydrolase to a form two trans-dihydrodiols (diol): (+)-B[a]P-7R,8R-diol and (−)-B[a]P-7S,8Sdiol (Figure 6.4). The B[a]P-7,8-diols are further metabolized (epoxidized) by the cytochrome P450 isozymes at the carbon 9–carbon 10 double bond to give four bay region diol epoxides (BPDE): (+)-anti-7R,8S,9S,10R-BPDE and its enantiomer (−)-anti-7S,8R,9R,10S-BPDE, as well as (+)-syn-7R,8S,9R,10S-BPDE and its enantiomer (−)-syn-7S,8R,9S,10R-BPDE (Figure 6.4). The anti and syn terminology refers to the spatial relationship between the hydroxyl group on carbon 7 and the oxide on carbons 9 and 10. These diol epoxides possess an inherent activity to undergo carbon–oxygen bond scission or ring opening to form a reactive carbonium ion on carbon 10 (i.e., a positive-charged carbon atom). Carbonium ions are highly reactive electrophilic species that react with nucleophiles such as DNA and proteins to form covalent adducts. Theoretically, each of the four BPDEs can covalently bind to specific atoms (almost always nitrogen) on the DNA bases to give two adducts based on the mechanism of the epoxide ring opening yielding both cis and trans

174


A Bay region O P450/EH

P450 HO

HO

OH

OH trans-B[a]P-7,8-dihydrodiol

B[a]P

anti-trans-B[a]P-7,8-dihydrodiol -9,10-epoxide

B Fjord region

O P450/EH

P450

HO

HO OH

DB[a,l]P

OH

trans-DB[a,l]P-11,12-dihydrodiol

anti-trans-DB[a,l]P-11,12dihydrodiol-13,14-epoxide

C P450 +

CH B[a]P

B[a]P radical cation

D AKR/O2 HO

O OH

O B[a]P-7,8-quinone

trans-B[a]P-7,8-dihydrodiol

O

E P450

Cyclopenta[cd]pyrene

Cyclopenta[cd]pyrene-3,4-oxide

Figure 6.3. The metabolic activation of PAH through diol epoxide, radical cation, o-quinone, and arene oxide activation mechanisms.

adducts, thus giving a potential total of 16 unique BPDE-DNA stereoisomeric adducts for each site on the nucleic acid base. In practice, far fewer metabolically formed adducts are observed because the anti-7R,8S,9S,10R-BPDE tends to be the major diol epoxide formed. For example, only one BPDE adduct was detected in mouse lungs treated with B[a]P, the (+)-anti-trans-7R,8S,9R,10S-BPDE-

6.2. JAMES A. AND ELIZABETH C. MILLER AND THEIR THEORY OF METABOLIC ACTIVATION 12

175

1

11

2

10 3

1 4

8 7

6

5

B[a]P

O

O

(-)-B[a]P-7S,8R-oxide

(+)-B[a]P-7R,8S-oxide

HO

HO

OH

OH

(-) B[a]P-7S,8S-diol

(+) B[a]P-7R,8R-diol

O

O

HO

HO

OH

OH

(+)-anti-7R,8S,9S,10R-BPDE

(-)-anti-7S,8R,9R,10S-BPDE +

+ O

O

HO

HO OH

(-)-syn-7R,8S,9R,10S-BPDE

OH

(+)-syn-7S,8R,9S,10R-BPDE

Figure 6.4. The metabolic activation of B[a]P to BPDE including a complete description of all of the possible stereoisomers.

deoxyguanosine (dGuo) adduct (Figure 6.5, upper left structure) (Ross et al. 1995). In vitro, human bronchoalveolar adenocarcinoma H358 cells treated with (±)anti-BPDE also formed the (+)-anti-trans-7R,8S,9R,10S-BPDE-dGuo adduct as the major adduct, with minor amounts of three other adducts (Figure 6.5) (Ruan et al. 2006).

176


N

N N dR

O

N dR

O

N

HN

N

HN

NH

NH

HO

HO

HO

HO OH

OH

(+)-anti-trans-7R,8S,9R,10S-BPDE-dGuo

(-)-anti-trans-7S,8R,9S,10R-BPDE-dGuo

N

N N dR

O N

HN

N dR

O N

HN NH

NH HO

HO

HO

HO OH (+)-anti-cis-7R,8S,9R,10R-BPDE-dGuo

OH (-)-anti-cis-7S,8R,9S,10S-BPDE-dGuo

Figure 6.5. Structures of BPDE–deoxyguanosine adducts. The BPDE adduct (upper left) is the major DNA adduct found in most mammalian tissues after exposure to B[a]P.

One of the postulated quantitative measures of the reactivity of diol epoxides is ΔEdeloc/β, which is based on perturbational molecular orbital calculations that predict the ease of carbonium ion formation. The greater the ΔEdeloc/β value, the more reactive the carbonium ion and greater values were associated with the PAHs exhibiting higher tumorigenic activities (Jerina et al. 1976). This theory was expanded to include PAH structures with deeper peripheral indentations in their structure, those containing a fjord region. An example of a fjord region PAH is the extremely potent PAH, dibenzo[a,l]pyrene (DB[a,l]P) (Figure 6.3B). The fjord region encompasses five carbons and four carbon–carbon bonds, and in some cases the steric interactions between hydrogen atoms within the fjord region of the PAH forces the PAH ring system out of planarity (Katz et al. 1998). Some PAH fjord region diol epoxides are nonplanar (Lewis-Bevan et al. 1995), and these nonplanar fjord region PAH diol epoxides possess even higher reactivities and tumorigenic activities (presumably due to their nonplanarity) than that predicted by ΔEdeloc/β alone (Xue and Warshawsky 2005). The enzymes primarily responsible for Phase I metabolism of PAHs are (a) the cytochrome P450s (CYPs) CYP1A1, CYP1A2, and CYP1B1, (b) NADPH


177

cytochrome P450 reductase, which converts PAHs to a series of arene oxides, and (c) epoxide hydrolase, which catalyzes the addition of water to the arene oxides to form trans diols. PAH phenols are also formed either by rearrangement of arene oxides or by direct oxygen insertion into a carbon–hydrogen bond, while quinones are formed by further oxidation of phenols or by the enzymatic action of aldo–keto reductases on PAH diols. The Phase II enzymes, UDP-glucuronyl transferase, PAPS sulfotransferase, and glutathione S-transferases conjugate PAH diols, phenols, and epoxides to glucuronic acid, sulfate, and glutathione, respectively (Osborne and Crosby 1987). One of the basic tenets of the theory of metabolic activation is that the proximate carcinogen should have greater biological activity compared to the parent molecule it was derived from. Similarly, the ultimate metabolite of the carcinogen should have greater biological activity than the proximate metabolite from which it was derived. For the bay region or fjord region diol epoxide mechanism, the PAH is metabolically activated in a sequence through the diol to the diol epoxide. This process creates intermediates that generally possess greater biological activities than their precursors. This effect is amply demonstrated in the case of benz[a]anthracene (B[a]A). One of the anti-diol epoxides of B[a]A (anti-B[a]A-3,4-diol-1,2-oxide) possesses greater activity as a mouse skin tumorigen or mouse lung tumorigen compared to its precursor diol (B[a]A-3,4-diol), which in turn possesses greater activity compared to the parent PAH, B[a]A (Levin et al. 1978; Wislocki et al. 1979). While this effect is observed for many PAHs, it is not universal for all PAHs that are metabolized to diols and diol epoxides due to a number of confounding factors (e.g., reactivity with water and biological constituents, or cytotoxicity). Also, an important observation to note is that the formation of a bay region PAH diol epoxide by itself does not confer a tumorigenic potential to that PAH. This is the case for phenanthrene because both phenanthrene and its bay region diol epoxide (anti-phenanthrene-3,4-diol-1,2 oxide) are inactive as tumorigens in newborn mice (Buening et al. 1979). Bay-region and fjord-region diol epoxides possess many biological activities, one of the most important being their ability to form covalent stable adducts with DNA. The nature and sequence specificity of these DNA adducts are based, in part, on the absolute configuration, molecular conformation, and stereochemistry of the diol epoxide, the specific purine (or pyrimidine) base being adducted, the site of adduction, and the nature and sequence of the DNA being adducted (Jerina et al. 1986). PAH–DNA adducts represent a type of DNA damage that can be converted into heritable mutations by misrepair or faulty DNA syntheses (Rodriguez and Loechler 1995; Watanabe et al. 1985). Bay- or fjord-region diol epoxide–DNA adducts can be repaired by nucleotide excision repair (Geacintov et al. 2002). There are numerous examples of bay- and fjord-region PAH diol epoxides that are mutagenic in bacteria and that can induce mutations, damage DNA, and chromosomal damage in rodent and human cells in culture. Many diol epoxides are tumorigenic in mice, thereby inducing skin, lung, and liver tumors. Concomitant with these findings are complementary observations that the parent PAHs and the diol metabolites of these PAHs also induced gene mutation, DNA damage, or chromosomal damage in these bioassay systems and were tumorigenic in mice. Furthermore, PAHs or their

178


bay- or fjord-region diol epoxides induced mutations in critical genes associated with carcinogenesis such as proto-oncogenes (Chakravarti et al. 1998; Prahalad et al. 1997) and tumor suppressor genes (Ruggeri et al. 1993; Ramet et al. 1995). A strong relationship exists between the nature of the diol epoxide DNA adducts and the type of ras proto-oncogene mutations observed in DNA from tumors induced by the PAHs. In general, PAHs that form DNA adducts at deoxyguanosine primarily induce mutations in the ras gene at codons 12 or 13, while those that form DNA adducts at deoxyadenosine induce mutations in the ras gene at codon 61. Those PAHs that form adducts at both purine bases induced both types of mutations (Ross and Nesnow 1999). In addition to their genotoxic effects, some bay- or fjord-region diol epoxides are reported to induce apoptosis and cell cycle arrest in mammalian cells (Chramostova et al. 2004). PAH diol epoxide–DNA adducts have not only been detected in rodent tissues in experimental systems after PAH exposure, but have also been identified in (a) populations exposed to complex mixtures containing PAHs, (b) foundry workers (Perera et al. 1988; Hemminki et al. 1988), (c) coke oven workers (Rojas et al. 1995; Pavanello et al. 1999), (d) cigarette smokers (Lodovici et al. 1998; Rojas et al. 1995), (e) chimney sweeps (Pavanello et al. 1999), and (f) populations exposed to smoky coal combustion mixtures (Mumford et al. 1993). Some bay- or fjord-region diol epoxides form DNA adducts in the human p53 tumor suppressor gene at sites that are hotspots for lung cancer (Smith et al. 2000). There are several variants on the diol epoxide mechanism. Bis-diol epoxide– DNA adducts were formed from dibenz[a,h]anthracene (Platt and Schollmeier 1994; Nesnow et al. 1994a) and dibenz[a,j]anthracene (Vulimiri et al. 1999). A bis-diol epoxide was proposed as a mechanism for carcinogenesis for dibenz[a,h]anthracene, while for dibenz[a,j]anthracene its biological significance is unknown. A phenolic diol epoxide–DNA adduct was formed from benz[b]fluoranthene and was proposed to contribute to the biological activity of benz[b]fluoranthene (Weyand et al. 1993). Finally, a phenolic oxide–DNA adduct of B[a]P has also been described with unknown biological significance (Fang et al. 2001). 6.2.1.2. Radical Cation Mechanism. A radical cation is formed when a single electron is removed from the π electron system of a PAH by an oxidation process. Some of the processes that have been described that can perform this oxidation of PAHs are iodine, electrochemical, horseradish peroxidase, and cytochrome P450 (Hanson et al. 1998; RamaKrishna et al. 1992; Cavalieri et al. 1988, 1990). Radical cations are electrophiles and bind to DNA bases to form covalent adducts. The structures of these adducts are dependent on the charge localization of the PAH radical cation and the charge density on specific atoms within the nucleic acid bases. Charge localization of PAHs of radical cations favors the meso positions of PAHs, and for B[a]P the charge localization favors the radical cation on carbon 6 (Cremonesi et al. 1992). The maximum charge density of guanine is found on N7, while those of adenine are on N7 and N3. B[a]P radical–cation DNA adducts have been characterized in vitro after metabolic activation with horseradish peroxidase, or with 3-methylcholanthrene-induced microsomes, and in vivo from the skin of mice treated with B[a]P (Chen et al. 1996). The major B[a]P radical cation–DNA adducts


O

179

N N

HN N

H2N

N N

N

H2N

N

BP-6-N7-Gua

BP-6-N7-Ade

O N

HN H2N

N

N H

N

N N

N NH2

BP-6-C8-Gua

Figure 6.6.

BP-6-N3-Ade

Structures of B[a]P radical cation DNA adducts from Chen et al. (1996).

obtained by horseradish peroxidase were 7-(benzo[a]pyrene-6-yl)-guanine (B[a] P-6-N7Gua), B[a]P-6-C8Gua, B[a]P-6-N7Adenine (Ade), and B[a]P-6-N3Ade (Figure 6.6). In contrast, microsomal activation of B[a]P gave B[a]P-6-N7Ade, B[a] P-6-N7Gua, and B[a]P-6-C8Gua; these depurinating adducts were also identified in the mouse skin treated with B[a]P. Each of these adducts was formed from an intermediary charged unstable covalent B[a]P–DNA adduct that has undergone bond scission at the glycosidic bond to give a depurinating adduct. The result of this bond scission is an apurinic site in the DNA molecule and a released B[a]P–DNA adduct. The role of depurinating adducts and apurinic sites in the PAH-induced cancer process is controversial and has yet to be fully elucidated. There are lines of evidence that both support and refute this theory. In support of this theory, the levels of depurinating adducts of B[a]P correlated with mutations in the H-ras oncogene in DNA isolated from mouse skin papillomas initiated by this compound (Chakravarti et al. 1995). It is well known that the initiation of skin tumors in mice is associated with the formation of mutations in the H-ras gene [reviewed by Ross and Nesnow (1999)]. DB[a,l]P treatment of mouse skin forms papillomas which contain the H-ras codon 61 (CAA to CTA) mutation. These same mutations were induced in early preneoplastic skin within one day after DB[a,l]P treatment and appear to be related to DB[a,l]P-Ade-depurinating adducts. Studies have shown that apurinic

180


sites generated from depurination may undergo error-prone excision repair in preS-phase cells to induce these mutations. The initiated cells carrying specific oncogenic mutations, formed presumably by misrepair, undergo rapid clonal expansion and regression (Chakravarti et al. 2000, 2001). Other investigators studying PAHs such as B[a]P, 7,12-dimethylbenz[a]anthracene (DMBA), and DB[a,l]P have found results not in accord with unstable DNA adduct formation and apurinic site formation. Studies were conducted in cytochrome P450-expressing mammary carcinoma MCF-7 cells and in leukemia HL-60 cells, which produce a high peroxidase activity but no cytochrome P450-mediated activity. The results from these studies demonstrated that metabolic activation of B[a]P, DMBA, and DB[a,l]P was primarily mediated by the cytochrome P450 enzymes leading to diol epoxides that form predominantly stable DNA adducts. Because only low levels of AP sites were detected, the radical cation pathway was not a major contributor to the metabolic activation scenario (Melendez-Colon et al. 1997, 1999a,b, 2000). Additional studies need to be conducted to resolve the issues surrounding the role of apurinic sites in the PAH carcinogenesis process. 6.2.1.3. o-Quinone/Reactive Oxygen Species Mechanism. PAHs are metabolized to trans-dihydrodiols by CYP1A1 and epoxide hydrolase. Aldo–keto reductases catalyze the NAD(P)-dependent oxidation of non-K-region transdihydrodiols to o-quinones of many PAHs, including phenanthrene, chrysene, 5-methylchrysene, DB[a,l]P, and B[a]P (see Figure 6.3A,C) (Smithgall et al. 1986). Aldo–keto reductases are present in many mammalian species. In humans, AKR1C1– AKR1C4 and AKR1A1 are capable of activating trans-dihydrodiols by converting them to redox-active o-quinones (Palackal et al. 2001). The conversion of dihydrodiols to quinones requires the formation of catechols that undergo air oxidation via two sequential one electron events to yield the o-quinones. Each of these steps produces ROS (Penning et al. 1996). The o-quinones can be reduced back to the catechols by nonenzymatic means such as by NADPH and begin the oxidative cycling process again yielding further quantities of ROS (Burczynski and Penning 2000) (Figure 6.7). ROS forms have been implicated in cytotoxic, mutagenic, and

O2

ROS

AKR HO

HO OH

trans-B[a]P-7,8-dihydrodiol

O HO

O B[a]P-7,8-quinone

7,8-dihydroxy-B[a]P NAD(P)

+

NAD(P)H

Figure 6.7. The metabolic activation of trans-B[a]P-7,8-diol to B[a]P-7,8-quinone by AKR and the generation of ROS from Burczynski and Penning (2000).


181

tumorigenic processes through DNA, lipid, and protein oxidation. ROS can induce DNA breaks or can oxidize DNA to form DNA adducts, the most common being 8-oxo-dG (Park et al. 2005). Not only can ROS damage DNA, it can also alter important cell signaling pathways that are involved in cell proliferation. For instance, ROS have been reported to alter protein kinase C. This receptor is susceptible to oxidative modification at the N-terminal regulatory domain containing a zincbinding cysteine-rich motif. When oxidized by ROS, protein kinase C activity is stimulated and signals downstream to c-fos and c-jun (Gopalakrishna and Jaken 2000). Other stress MAPK kinases such as the JNK/AP1 pathway have been altered by ROS (Benhar et al. 2002). It should be noted that to date the o-quinone/ROS mechanism has only been described in vitro. There are no reports that that PAH o-quinones are formed in vivo from B[a]P and are stable enough to redox cycle and induce ROS. B[a]P-7,8quinone has been the most intensely studied o-quinone and has been found to adduct to DNA in vitro (Balu et al. 2004, 2006), but not in vivo (Nesnow et al. 2005). B[a] P-7,8-quinone induces DNA breaks (Park et al. 2008) and induces ROS in vitro (Flowers-Geary et al. 1996). Current thought is that B[a]P-7,8-quinone mediates its in vitro biological effects through ROS formation in vitro. 6.2.1.4. Cyclopenta-Ring Oxidation Mechanism. The cyclopenta-ring oxidation mechanism involves arene oxide formation at a highly electron-rich isolated double bond located at a five-membered cyclopenta-ring within a cyclopenta-PAH. The cyclopenta ring is an external five-membered carbocyclic ring situated on a carbocyclic hexameric fused ring system. For example, a cyclopenta-ring derivative of benz[a]anthracene is benz[j]aceanthrylene while that of pyrene is cyclopenta[c,d] pyrene (Figure 6.3E). In general, cyclopenta ring derivatives of PAH are more mutagenic than their unsubstituted counterparts. For example, anthracene is nonmutagenic while its cyclopenta-ring counterpart, aceanthrylene, is highly mutagenic (Kohan et al. 1985). Similarly cyclopenta-ring derivatives of PAH are generally more tumorigenic than their unsubstituted counterparts. For example, pyrene is not tumorigenic while cyclopenta[c,d]pyrene is highly tumorigenic (Nesnow et al. 1998). Since the cyclopenta-ring is usually the region of highest electron density, it is a major site of oxidation by the cytochrome P450 isozymes (Nesnow et al. 1984, 1988). Rat and mouse liver preparations, human and rodent cells in culture, human CYP1A1, CYP1A2, and CYP3A4, human liver microsomes, and rats in vivo metabolize cyclopenta-fused PAHs at the cyclopenta ring double bond to give cyclopenta ring oxides and diols (Gold and Eisenstadt 1980; Mohapatra et al. 1987; Kwon et al. 1992; Nyholm et al. 1996; Johnsen et al. 1998a,b; Hegstad et al. 1999). Cyclopenta-ring oxides are reactive intermediates and bind to DNA to form DNA adducts in vitro and in vivo mainly at deoxyguanosine (Beach and Gupta 1994; Hsu et al. 1999). Cyclopenta[c,d]pyrene, a mouse lung tumorigen, formed cyclopentaring–deoxyguanosine–DNA adducts in lung tissues and induced mutations at the Ki-ras protooncogene in lung tumors of treated mice (Nesnow et al. 1994b). Cyclopenta-ring oxides like their parent cyclopenta-PAHs are mutagenic in bacterial and mammalian cells and can morphologically transform immortalized cells in culture (Bartczak et al. 1987; Nesnow et al. 1991). Cyclopenta-ring oxides are

182


hydrated by epoxide hydrolase to diols. Some cyclopenta-ring diols are conjugated to sulfate esters by PAPS-sulfotransferase, and these reactive intermediates are mutagenic and can bind to DNA to form adducts (Surh et al. 1993). 6.2.1.5. Summary of the Mechanisms of Metabolic Activation of PAH. For B[a]P, based on the wealth of data, the diol epoxide metabolic activation mechanism seems to be the dominant mechanism in the induction of lung carcinogenesis in rodents and humans. This conclusion is based on mechanistic data obtained from experimental animal and human biomarker studies (Straif et al. 2005). Both the radical cation and diol epoxide metabolic activation mechanisms can explain the mouse skin tumorigenic activities of B[a]P, but their relative contributions to this tumorigenesis process needs to be defined. To date, the PAH o-quinone/ROS metabolic mechanism has been described only in vitro, and no in vivo studies using PAH o-quinones are available that describe ROS formation, DNA damage, or tumorigenic effects, prerequisites for validating this mechanism. Further research is needed to define the relevance of this mechanism of PAH metabolic activation. For cyclopentaring PAHs, arene oxide formation at the cyclopenta ring is currently accepted as the major route of metabolic activation for this class of PAHs.

6.3. THE CONCEPTS OF INITIATION, PROMOTION, AND PROGRESSION: THE ORIGIN OF MULTISTAGE CARCINOGENESIS Much currently accepted theory of chemical carcinogenesis has evolved from studies using the mouse skin model of tumorigenesis using PAH as model compounds. The terms initiation and promotion essentially were first described by Rous and Kidd based on the application of coal tar to the ears of rabbits (initiate) following this treatment with physical wounding (promote) giving rise to tumors (Rous and Kidd 1941). Note that the development and growth of tumors in general involves three distinct stages: (1) initiation, (2) promotion, and (3) progression. The ideas that chemicals could alter DNA and induce mutations came from a series of investigators in the 19th century leading to a book authored by Karl Heinrich Bauer in 1924 titled Mutational Theory on the Origin of Cancers [see review by Edler and Kopp-Schneider (2005)]. Many of these ideas were codified by Berenblum and Shubik for dividing chemical carcinogenesis into two discrete stages, initiation and promotion (Berenblum and Shubik 1949). Initiation was defined as resulting from the single administration of an agent such as B[a]P at a dose that would not induce cancer. Promotion was defined as subsequent repeated administration of an agent [e.g., croton oil or its active component, tetradecanoyl phorbol acetate (TPA)] such as an irritant that by itself would not induce significant numbers of tumors. However, the combined administration of an initiator and a promoter was effective at tumor induction. These concepts were experimentally studied in depth by Boutwell and his colleagues at the McArdle Institute for Cancer Research in Madison, Wisconsin, U.S.A. (Boutwell 1974). This led to a series of characteristics that defined the two stages. Initiation was found to be irreversible, additive, and dose

6.3. THE ORIGIN OF MULTISTAGE CARCINOGENESIS

A X

No Tumors

B X

Tumors

C X

Tumors

D

X

183

No Tumors

E No Tumors F X

No Tumors

Time X = Application of Initiator

= Application of Promoter

Figure 6.8. A series of experiments using the mouse skin model that demonstrated the concepts of initiation and promotion adapted from Pitot (1986).

responsive. Mice treated with a single low dose of an initiator, such as a PAH, did not develop tumors after one year (Figure 6.8A). It has been shown that these mice had covalent DNA adducts that could lead to mutagenic events. Initiated cells in mouse skin could be promoted either immediately (generally twice weekly) after initiation (Figure 6.8B) or one year after initiation (Figure 6.8C), both protocols yielding tumors. Mice first treated with repeated doses of a promoter and then a single dose of an initiator did not develop tumors (Figure 6.8D). Promotion was reversible, nonadditive and possessed a threshold and saturation (Pitot 1986). Mice treated only with repeat doses of a promoter exhibited either no tumors or a low frequency of papillomas (Figure 6.8E). Promotion has been attributed to the clonal expansion of single initiated cells through changes in gene expression, altered gap junction intercellular communication, modified key receptor interactions, and altered cell proliferative responses. This requires the continued presence of the promoting agent because it is reversible upon withdrawal of the promoting agent. This was shown by treating mice with an initiator and then a promoter once every 4 weeks. No tumors were found (Figure 6.8F). The promotion response is assumed to fix the mutations induced by the tumor initiators. Tumor promotion in mouse skin was further stratified into two stages by Slaga et al. (1982, 1996). Based on the theory of multistage carcinogenesis, chemicals could be classified into initiators, promoters, and complete carcinogens [reviewed by Nesnow et al. (1983)]. A complete carcinogen is a chemical that is able to induce cancer itself; that is, it possesses properties of (a) initiation and promotion or (b) initiation, promotion, and progression. While the majority of studies on tumor promotion have focused on mouse skin, there are cancers in other organs that are applicable to this model. Bladder

184


cancer (Hicks 1980), colon cancer (Pitot et al. 2000), and liver cancer (Pitot and Sirica 1980) are the major examples. The concepts of multistage carcinogenesis based on initiation, promotion, and progression have led in part to the current view that cancer is a multistage process due to series of mutations in oncogenes, tumor suppressor genes, and genetic instability genes. This was described by Fearon and Vogelstein, who proposed a model of colorectal carcinogenesis that correlated specific genetic events with changes in tissue morphology from normal tissue to evolving polyp formation and finally to an invasive cancer (Fearon and Vogelstein 1990). Tumor progression can be defined as a stage of neoplastic development that is characterized by additional mutations in tumor oncogenes and tumor suppressor genes leading to karyotype alterations, increased malignancy, metastases, and tumor aggressiveness (Pitot 2001). The most succinct summary of mouse skin initiation, promotion, and progression is found in Hennings et al. (1993) (Figure 6.9). Carcinogenesis in mouse skin can be divided into three distinct stages: initiation, promotion, and progression (malignant conversion). Initiation, induced by a single exposure to a genotoxic carcinogen, can result from a mutation in a single critical gene (e.g., rasHa), apparently in only a few epidermal cells. The change is irreversible. Promotion, resulting in the development of numerous benign tumors (papillomas), is accomplished by the repeated application of a nonmutagenic tumor promoter. The effects of single applications of tumor promoters are reversible since papillomas do not develop after insufficient exposure of initiated skin to promoters or when the interval between individual promoter applications is increased sufficiently. The reversibility of promotion suggests an epigenetic mechanism. Promoter treatment provides an environment that allows the selective clonal expansion of foci of initiated cells. The conversion of squamous papillomas to carcinomas (termed progression or malignant conversion) occurs spontaneously at a low frequency. The rate of progression to malignancy can be significantly increased by treatment of papilloma-bearing mice with certain genotoxic agents. These progressor agents or converting agents are likely to act via a second genetic change in papillomas already bearing the initiating mutation. Progression in the skin is characterized by genetic changes that result in several distinct changes in the levels or activity of structural proteins, growth factors, and proteases.” A more complex model of multistage carcinogenesis was proposed by Hanahan and Weinberg in their paper “The Hallmarks of Cancer” (Hanahan and Weinberg 2000). They proposed a series of discrete steps: (1) self-sufficiency in growth signals (one of the key characteristics of the tumor cell is its capacity for proliferation without dependence on external growth factors), (2) insensitivity to antigrowth signals (antigrowth signals must be avoided for cancer cells to survive and replicate), (3) tissue invasion and metastasis (cancer cells colonize distant sites to form metastases and overcome the normal suppressors of invasion), (4) limitless potential for

Initiation

Promotion

Progression

Figure 6.9. Multistage carcinogenesis describing the conversion of an initiated cell into malignant tumors adapted from Pitot (1986).

REFERENCES

185

replication (tumor cells must become immortal), (5) sustained angiogenesis (angiogenic ability must be acquired to increase tumor growth and size), and (6) evading apoptosis (cancer cells acquire resistance to apoptosis is to maintain proliferation). In summary, the development and maturation of the concepts of multistep and multistage chemical carcinogenesis have been intertwined with the study of PAH exposures, the concepts of metabolic activation, PAH tumorigenesis, and mechanisms of multistep mouse skin tumorigenesis. Much has been accomplished from the early beginnings of the chimney sweep epidemiological studies of Percival Pott in 1714, to our current understanding of the carcinogenesis process over the last 294 years. The mechanisms of chemical carcinogenesis and the development of human cancer are still, relatively speaking, “black boxes.” However, in the past 294 years these “black boxes” have been shrunk remarkably as new molecular techniques have been applied to these questions. New techniques such as Q-real time RT-PCR, genomics, proteomics, and metabolomics in combination with results from the human genome project will hopefully, in the next decades, remove these “black boxes” and give us a complete understanding of these important processes.

REFERENCES Balu, N., Padgett, W. T., Lambert, G. R., Swank, A. E., Richard, A. M., and Nesnow, S. (2004). Identification and characterization of novel stable deoxyguanosine and deoxyadenosine adducts of benzo[a]pyrene-7,8-quinone from reactions at physiological pH. Chem Res Toxicol 17, 827–838. Balu, N., Padgett, W. T., Nelson, G. B., Lambert, G. R., Ross, J. A., and Nesnow, S. (2006). Benzo[a] pyrene-7,8-quinone-3′-mononucleotide adduct standards for 32P postlabeling analyses: Detection of benzo[a]pyrene-7,8-quinone-calf thymus DNA adducts. Anal Biochem 355, 213–223. Bartczak, A. W., Sangaiah, R., Ball, L. M., Warren, S. H., and Gold, A. (1987). Synthesis and bacterial mutagenicity of the cyclopenta oxides of the four cyclopenta-fused isomers of benzanthracene. Mutagenesis 2, 101–105. Beach, A. C., and Gupta, R. C. (1994). DNA adducts of the ubiquitous environmental contaminant cyclopenta[cd]pyrene. Carcinogenesis 15, 1065–1072. Beland, F. A., and Kadlubar, F. F. (1985). Formation and persistence of arylamine DNA adducts in vivo. Environ Health Perspect 62, 19–30. Benhar, M., Engelberg, D., and Levitzki, A. (2002). ROS, stress-activated kinases and stress signaling in cancer. EMBO Rep 3, 420–425. Berenblum, I., and Shubik, P. (1949). A new quantitative approach to the study of the stages of chemical carcinogenesis in mouse skin. Br J Cancer 1, 383–390. Boutwell, R. K. (1974). The function and mechanism of promoters of carcinogenesis. CRC Crit Rev Toxicol 2, 419–443. Boyland, E., and Sims, P. (1964). Metabolism of polycyclic compounds. 24. The metabolism of benz[alpha]anthracene. Biochem J 91, 493–506. Buening, M. K., Levin, W., Karle, J. M., Yagi, H., Jerina, D. M., and Conney, A. H. (1979). Tumorigenicity of bay-region epoxides and other derivatives of chrysene and phenanthrene in newborn mice. Cancer Res 39, 5063–5068. Burczynski, M. E., and Penning, T. M. (2000). Genotoxic polycyclic aromatic hydrocarbon orthoquinones generated by aldo–keto reductases induce CYP1A1 via nuclear translocation of the aryl hydrocarbon receptor. Cancer Res 60, 908–915. Case, R. A., Hosker, M. E., Mc, D. D., and Pearson, J. T. (1954). Tumours of the urinary bladder in workmen engaged in the manufacture and use of certain dyestuff intermediates in the British chemical industry. I. The role of aniline, benzidine, alpha-naphthylamine, and beta-naphthylamine. Br J Ind Med 11, 75–104.

186


Cavalieri, E., Rogan, E., and Sinha, D. (1988). Carcinogenicity of aromatic hydrocarbons directly applied to rat mammary gland. J Cancer Res Clin Oncol 114, 3–9. Cavalieri, E. L., Rogan, E. G., Devanesan, P. D., Cremonesi, P., Cerny, R. L., Gross, M. L., and Bodell, W. J. (1990). Binding of benzo[a]pyrene to DNA by cytochrome P-450 catalyzed one-electron oxidation in rat liver microsomes and nuclei. Biochemistry 29, 4820–4827. Chakravarti, D., Mailander, P., Franzen, J., Higginbotham, S., Cavalieri, E. L., and Rogan, E. G. (1998). Detection of dibenzo[a,l]pyrene-induced H-ras codon 61 mutant genes in preneoplastic SENCAR mouse skin using a new PCR-RFLP method. Oncogene 16, 3203–3210. Chakravarti, D., Mailander, P. C., Cavalieri, E. L., and Rogan, E. G. (2000). Evidence that error-prone DNA repair converts dibenzo[a,l]pyrene-induced depurinating lesions into mutations: Formation, clonal proliferation and regression of initiated cells carrying H-ras oncogene mutations in early preneoplasia. Mutat Res 456, 17–32. Chakravarti, D., Mailander, P. C., Li, K. M., Higginbotham, S., Zhang, H. L., Gross, M. L., Meza, J. L., Cavalieri, E. L., and Rogan, E. G. (2001). Evidence that a burst of DNA depurination in SENCAR mouse skin induces error-prone repair and forms mutations in the H-ras gene. Oncogene 20, 7945–7953. Chakravarti, D., Pelling, J. C., Cavalieri, E. L., and Rogan, E. G. (1995). Relating aromatic hydrocarboninduced DNA adducts and c-H-ras mutations in mouse skin papillomas: The role of apurinic sites. Proc Natl Acad Sci USA 92, 10422–10426. Chen, L., Devanesan, P. D., Higginbotham, S., Ariese, F., Jankowiak, R., Small, G. J., Rogan, E. G., and Cavalieri, E. L. (1996). Expanded analysis of benzo[a]pyrene-DNA adducts formed in vitro and in mouse skin: Their significance in tumor initiation. Chem Res Toxicol 9, 897–903. Chramostova, K., Vondracek, J., Sindlerova, L., Vojtesek, B., Kozubik, A., and Machala, M. (2004). Polycyclic aromatic hydrocarbons modulate cell proliferation in rat hepatic epithelial stem-like WBF344 cells. Toxicol Appl Pharmacol 196, 136–148. Cook, J. W., Hewitt, C. L., and Heiger, I. (1933). The isolation of a cancer producing hydrocarbon from coal tar. J Chem Soc, 395–405. Cremonesi, P., Rogan, E., and Cavalieri, E. (1992). Correlation studies of anodic peak potentials and ionization potentials for polycyclic aromatic hydrocarbons. Chem Res Toxicol 5, 346–355. Edler, L., and Kopp-Schneider, A. (2005). Origins of the mutational origin of cancer. Int J Epidemiol 34, 1168–1170. Fang, A. H., Smith, W. A., Vouros, P., and Gupta, R. C. (2001). Identification and characterization of a novel benzo[a]pyrene-derived DNA adduct. Biochem Biophys Res Commun 281, 383–389. Fearon, E. R., and Vogelstein, B. (1990). A genetic model for colorectal tumorigenesis. Cell 61, 759–767. Fibiger, J. A. G. (1913). Untersuchungen über eine Nematode Spiroptera sp. n.) und deren Fähigkeit papillomatöse und carcinomatöse Geschwulstbildungen im Magen der Ratte hervorzurufen. Z Krebsforschung 13, 217–280. Flowers-Geary, L., Bleczinki, W., Harvey, R. G., and Penning, T. M. (1996). Cytotoxicity and mutagenicity of polycyclic aromatic hydrocarbon ortho-quinones produced by dihydrodiol dehydrogenase. Chem Biol Interact 99, 55–72. Geacintov, N. E., Broyde, S., Buterin, T., Naegeli, H., Wu, M., Yan, S., and Patel, D. J. (2002). Thermodynamic and structural factors in the removal of bulky DNA adducts by the nucleotide excision repair machinery. Biopolymers 65, 202–210. Gold, A., and Eisenstadt, E. (1980). Metabolic activation of cyclopenta(cd)pyrene to 3,4-epoxycyclopenta(cd) pyrene by rat liver microsomes. Cancer Res 40, 3940–3944. Gopalakrishna, R., and Jaken, S. (2000). Protein kinase C signaling and oxidative stress. Free Radic Biol Med 28, 1349–1361. Grimmer, G., Brune, H., Deutsch-Wenzel, R., Dettbarn, G., Jacob, J., Naujack, K. W., Mohr, U., and Ernst, H. (1987). Contribution of polycyclic aromatic hydrocarbons and nitro-derivatives to the carcinogenic impact of diesel engine exhaust condensate evaluated by implantation into the lungs of rats. Cancer Lett 37, 173–180. Hanahan, D., and Weinberg, R. A. (2000). The hallmarks of cancer. Cell 100, 57–70. Hanson, A. A., Rogan, E. G., and Cavalieri, E. L. (1998). Synthesis of adducts formed by iodine oxidation of aromatic hydrocarbons in the presence of deoxyribonucleosides and nucleobases. Chem Res Toxicol 11, 1201–1208.

REFERENCES

187

Hegstad, S., Lundanes, E., Holme, J. A., and Alexander, J. (1999). Characterization of metabolites of benz(j)aceanthrylene in faeces, urine and bile from rat. Xenobiotica 29, 1257–1272. Heidelberger, C. (1975). Chemical carcinogenesis. Annu Rev Biochem 44, 79–121. Hemminki, K., Perera, F. P., Phillips, D. H., Randerath, K., Reddy, M. V., and Santella, R. M. (1988). Aromatic deoxyribonucleic acid adducts in white blood cells of foundry and coke oven workers. Scand J Work Environ Health 14(Suppl 1), 55–56. Hennings, H., Glick, A. B., Greenhalgh, D. A., Morgan, D. L., Strickland, J. E., Tennenbaum, T., and Yuspa, S. H. (1993). Critical aspects of initiation, promotion, and progression in multistage epidermal carcinogenesis. Proc Soc Exp Biol Med 202, 1–8. Hicks, R. M. (1980). Multistage carcinogenesis in the urinary bladder. Br Med Bull 36, 39–46. Hsu, C. H., Skipper, P. L., and Tannenbaum, S. R. (1999). DNA adduct formation by secondary metabolites of cyclopenta[cd]pyrene in vitro. Cancer Lett 136, 137–141. Jerina, D. M., Lehr, R., Yagi, H., Hernandez, O., Dansette, P., Wislocki, P. G., Wood, A. W., Chang, R. L., Levin, W., and Conney, A. H. (1976). Mutagenicity of benzo(a)pyrene derivatives from the description of a quantum mechanical model which predicts the ease of carbonium ion formation from diol epoxides. In In Vitro Metabolic Activation and Mutagenesis Testing, (deSerres, F. J., Fouts, J. R., Bend, J. R., and Philpot, R. M., eds., Elsevier/North-Holland Biomedical, Amsterdam, pp. 159–177. Jerina, D. M., Sayer, J. M., Agarwal, S. K., Yagi, H., Levin, W., Wood, A. W., Conney, A. H., PruessSchwartz, D., Baird, W. M., Pigott, M. A., et al. (1986). Reactivity and tumorigenicity of bay-region diol epoxides derived from polycyclic aromatic hydrocarbons. Adv Exp Med Biol 197, 11–30. Johnsen, N. M., Brunborg, G., Haug, K., Scholz, T., and Holme, J. A. (1998a). Metabolism and activation of cyclopenta polycyclic aromatic hydrocarbons in isolated human lymphocytes, HL-60 cells and exposed rats. Chem Biol Interact 114, 77–95. Johnsen, N. M., Nyholm, S. H., Haug, K., Scholz, T., and Holme, J. A. (1998b). Metabolism and activation of cyclopenta polycyclic aromatic hydrocarbons in liver tissue from rats and humans. Chem Biol Interact 113, 217–237. Katz, A. K., Carrell, H. L., and Glusker, J. P. (1998). Dibenzo[a,l]pyrene (dibenzo[def,p]chrysene): fjord-region distortions. Carcinogenesis 19, 1641–1648. Kennaway, E. (1955). Identification of carcinogenic compound in coal tar. Br Med J 2, 749–752. Kohan, M. J., Sangaiah, R., Ball, L. M., and Gold, A. (1985). Bacterial mutagenicity of aceanthrylene: a novel cyclopenta-fused polycyclic aromatic hydrocarbon of low molecular weight. Mutat Res 155, 95–98. Kwon, H., Sahali, Y., Skipper, P. L., and Tannenbaum, S. R. (1992). Oxidation of cyclopenta[cd]pyrene by human and mouse liver microsomes and selected cytochrome P450 enzymes. Chem Res Toxicol 5, 760–764. Levin, W., Thakker, D. R., Wood, A. W., Chang, R. L., Lehr, R. E., Jerina, D. M., and Conney, A. H. (1978). Evidence that benzo(a)anthracene 3,4-diol-1,2-epoxide is an ultimate carcinogen on mouse skin. Cancer Res 38, 1705–1710. Lewis-Bevan, L., Little, S. B., and Rabinowitz, J. R. (1995). Quantum mechanical studies of the structure and reactivities of the diol epoxides of benzo[c]phenanthrene. Chem Res Toxicol 8, 499–505. Lodovici, M., Akpan, V., Giovannini, L., Migliani, F., and Dolara, P. (1998). Benzo[a]pyrene diolepoxide DNA adducts and levels of polycyclic aromatic hydrocarbons in autoptic samples from human lungs. Chem Biol Interact 116, 199–212. Melendez-Colon, V. J., Luch, A., Seidel, A., and Baird, W. M. (1999a). Cancer initiation by polycyclic aromatic hydrocarbons results from formation of stable DNA adducts rather than apurinic sites. Carcinogenesis 20, 1885–1891. Melendez-Colon, V. J., Luch, A., Seidel, A., and Baird, W. M. (1999b). Comparison of cytochrome P450- and peroxidase-dependent metabolic activation of the potent carcinogen dibenzo[a,l]pyrene in human cell lines: formation of stable DNA adducts and absence of a detectable increase in apurinic sites. Cancer Res 59, 1412–1416. Melendez-Colon, V. J., Luch, A., Seidel, A., and Baird, W. M. (2000). Formation of stable DNA adducts and apurinic sites upon metabolic activation of bay and fjord region polycyclic aromatic hydrocarbons in human cell cultures. Chem Res Toxicol 13, 10–17. Melendez-Colon, V. J., Smith, C. A., Seidel, A., Luch, A., Platt, K. L., and Baird, W. M. (1997). Formation of stable adducts and absence of depurinating DNA adducts in cells and DNA treated

188


with the potent carcinogen dibenzo[a,l]pyrene or its diol epoxides. Proc Natl Acad Sci USA 94, 13542–13547. Miller, E. C., and Miller, J. A. (1979). Milestones in chemical carcinogenesis. Semin Oncol 6, 445–460. Miller, J. A., Cramer, J. W., and Miller, E. C. (1960). The N- and ringhydroxylation of 2acetylaminofluorene during carcinogenesis in the rat. Cancer Res 20, 950–962. Mohapatra, N., MacNair, P., Bryant, B. J., Ellis, S., Rudo, K., Sangaiah, R., Gold, A., and Nesnow, S. (1987). Morphological transforming activity and metabolism of cyclopenta-fused isomers of benz[a] anthracene in mammalian cells. Mutat Res 188, 323–334. Mumford, J. L., Lee, X., Lewtas, J., Young, T. L., and Santella, R. M. (1993). DNA adducts as biomarkers for assessing exposure to polycyclic aromatic hydrocarbons in tissues from Xuan Wei women with high exposure to coal combustion emissions and high lung cancer mortality. Environ Health Perspect 99, 83–87. Nesnow, S., Easterling, R. E., Ellis, S., Watts, R., and Ross, J. (1988). Metabolism of benz[j]aceanthrylene (cholanthrylene) and benz[l]aceanthrylene by induced rat liver S9. Cancer Lett 39, 19–27. Nesnow, S., Lasley, J., Curti, S., Ross, J., Nelson, G., Sangaiah, R., and Gold, A. (1991). Morphological transformation and DNA adduct formation by benz[j]aceanthrylene and its metabolites in C3H10T1/2CL8 cells: Evidence for both cyclopenta-ring and bay-region metabolic activation pathways. Cancer Res 51, 6163–6169. Nesnow, S., Leavitt, S., Easterling, R., Watts, R., Toney, S. H., Claxton, L., Sangaiah, R., Toney, G. E., Wiley, J., Fraher, P., et al. (1984). Mutagenicity of cyclopenta-fused isomers of benz(a)anthracene in bacterial and rodent cells and identification of the major rat liver microsomal metabolites. Cancer Res 44, 4993–5003. Nesnow, S., Mass, M. J., Ross, J. A., Galati, A. J., Lambert, G. R., Gennings, C., Carter, W. H., Jr., and Stoner, G. D. (1998). Lung tumorigenic interactions in strain A/J mice of five environmental polycyclic aromatic hydrocarbons. Environ Health Perspect 106(Suppl 6), 1337–1346. Nesnow, S., Padgett, W., Balu, N., Nelson, G., Winnik, W., Lambert, G., George, M. A., and Ross, J. S. (2005). Benzo[a]pyrene-7,8-quinone forms covalent-DNA adducts in vitro but none were detected in the lungs or livers of strain A/J mice in vivo. In ISPAC 16, Trondheim, NO. Nesnow, S., Ross, J., Beck, S., Lasley, J., Nelson, G., Lambert, G., Platt, K. L., and Agarwal, S. C. (1994a). Morphological transformation and DNA adduct formation by dibenz[a,h]anthracene and its metabolites in C3H10T1/2CL8 cells. Carcinogenesis 15, 2225–2231. Nesnow, S., Ross, J. A., Nelson, G., Wilson, K., Roop, B. C., Jeffers, A. J., Galati, A. J., Stoner, G. D., Sangaiah, R., Gold, A., et al. (1994b). Cyclopenta[cd]pyrene-induced tumorigenicity, Ki-ras codon 12 mutations and DNA adducts in strain A/J mouse lung. Carcinogenesis 15, 601–606. Nesnow, S., Triplett, L. L., and Slaga, T. J. (1983). Mouse skin tumor initiation-promotion and complete carcinogenesis bioassays: Mechanisms and biological activities of emission samples. Environ Health Perspect 47, 255–268. Nyholm, S. H., Alexander, J., Lundanes, E., Frandsen, H., Andersson, R., Grivas, S., Nesnow, S., and Holme, J. A. (1996). Biotransformation of the cyclopenta-fused polycyclic aromatic hydrocarbon benz[j]aceanthrylene in isolated rat liver cell: Identification of nine new metabolites. Carcinogenesis 17, 1111–1120. Osborne, M. R., and Crosby, N. T. (1987). Benzopyrenes, Cambridge University Press, Cambridge, England. Palackal, N. T., Burczynski, M. E., Harvey, R. G., and Penning, T. M. (2001). The ubiquitous aldehyde reductase (AKR1A1) oxidizes proximate carcinogen trans-dihydrodiols to o-quinones: Potential role in polycyclic aromatic hydrocarbon activation. Biochemistry 40, 10901–10910. Park, J. H., Gopishetty, S., Szewczuk, L. M., Troxel, A. B., Harvey, R. G., and Penning, T. M. (2005). Formation of 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxo-dGuo) by PAH o-quinones: Involvement of reactive oxygen species and copper(II)/copper(I) redox cycling. Chem Res Toxicol 18, 1026–1037. Park, J. H., Mangal, D., Tacka, K. A., Quinn, A. M., Harvey, R. G., Blair, I. A., and Penning, T. M. (2008). Evidence for the aldo–keto reductase pathway of polycyclic aromatic trans-dihydrodiol activation in human lung A549 cells. Proc Natl Acad Sci USA 105, 6846–6851.

REFERENCES

189

Pavanello, S., Favretto, D., Brugnone, F., Mastrangelo, G., Dal Pra, G., and Clonfero, E. (1999). HPLC/ fluorescence determination of anti-BPDE–DNA adducts in mononuclear white blood cells from PAHexposed humans. Carcinogenesis 20, 431–435. Penning, T. M., Ohnishi, S. T., Ohnishi, T., and Harvey, R. G. (1996). Generation of reactive oxygen species during the enzymatic oxidation of polycyclic aromatic hydrocarbon trans-dihydrodiols catalyzed by dihydrodiol dehydrogenase. Chem Res Toxicol 9, 84–92. Perera, F. P., Hemminki, K., Young, T. L., Brenner, D., Kelly, G., and Santella, R. M. (1988). Detection of polycyclic aromatic hydrocarbon–DNA adducts in white blood cells of foundry workers. Cancer Res 48, 2288–2291. Pitot, H. C. (1986). Fundamentals of Oncology, Marcel Dekker, New York. Pitot, H. C. (2001). Pathways of progression in hepatocarcinogenesis. Lancet 358, 859–860. Pitot, H. C., Hikita, H., Dragan, Y., Sargent, L., and Haas, M. (2000). Review article: The stages of gastrointestinal carcinogenesis—Application of rodent models to human disease. Aliment Pharmacol Ther 14(Suppl 1), 153–160. Pitot, H. C., and Sirica, A. E. (1980). The stages of initiation and promotion in hepatocarcinogenesis. Biochim Biophys Acta 605, 191–215. Platt, K. L., and Schollmeier, M. (1994). Bisdihydrodiols, rather than dihydrodiol oxides, are the principal microsomal metabolites of tumorigenic trans-3,4-dihydroxy-3,4-dihydrodibenz[a,h]anthracene. Chem Res Toxicol 7, 89–97. Pott, P. (1775). Chirugical Observations Relative to the Cataract, the Polypus of the Nose, the Cancer of the Scrotum, the Different Kinds of Ruptures, and the Mortification of Toes and Feet, L. Hawkes, W. Clarke, & R. Collins, London. Prahalad, A. K., Ross, J. A., Nelson, G. B., Roop, B. C., King, L. C., Nesnow, S., and Mass, M. J. (1997). Dibenzo[a,l]pyrene-induced DNA adduction, tumorigenicity, and Ki-ras oncogene mutations in strain A/J mouse lung. Carcinogenesis 18, 1955–1963. RamaKrishna, N. V., Gao, F., Padmavathi, N. S., Cavalieri, E. L., Rogan, E. G., Cerny, R. L., and Gross, M. L. (1992). Model adducts of benzo[a]pyrene and nucleosides formed from its radical cation and diol epoxide. Chem Res Toxicol 5, 293–302. Ramet, M., Castren, K., Jarvinen, K., Pekkala, K., Turpeenniemi-Hujanen, T., Soini, Y., Paakko, P., and Vahakangas, K. (1995). p53 protein expression is correlated with benzo[a]pyrene–DNA adducts in carcinoma cell lines. Carcinogenesis 16, 2117–2124. Rehn, L. (1895). Ueber Blasentrumoren bei Fruchsinalbeitern. Arch Kind Chir 50, 588. Rodriguez, H., and Loechler, E. L. (1995). Are base substitution and frameshift mutagenesis pathways interrelated? An analysis based upon studies of the frequencies and specificities of mutations induced by the (+)-anti diol epoxide of benzo[a]pyrene. Mutat Res 326, 29–37. Rojas, M., Alexandrov, K., Auburtin, G., Wastiaux-Denamur, A., Mayer, L., Mahieu, B., Sebastien, P., and Bartsch, H. (1995). Anti-benzo[a]pyrene diolepoxide–DNA adduct levels in peripheral mononuclear cells from coke oven workers and the enhancing effect of smoking. Carcinogenesis 16, 1373–1376. Ross, J. A., Nelson, G. B., Wilson, K. H., Rabinowitz, J. R., Galati, A., Stoner, G. D., Nesnow, S., and Mass, M. J. (1995). Adenomas induced by polycyclic aromatic hydrocarbons in strain A/J mouse lung correlate with time-integrated DNA adduct levels. Cancer Res 55, 1039–1044. Ross, J. A., and Nesnow, S. (1999). Polycyclic aromatic hydrocarbons: Correlations between DNA adducts and ras oncogene mutations. Mutat Res 424, 155–166. Rous, P., and Kidd, J. G. (1941). Conditional neoplasma and subthreshold neoplastic states: A study of tar tumors in rabbits. J Exp Med 73, 365–389. Ruan, Q., Kim, H. Y., Jiang, H., Penning, T. M., Harvey, R. G., and Blair, I. A. (2006). Quantification of benzo[a]pyrene diol epoxide DNA-adducts by stable isotope dilution liquid chromatography/tandem mass spectrometry. Rapid Commun Mass Spectrom 20, 1369–1380. Ruggeri, B., DiRado, M., Zhang, S. Y., Bauer, B., Goodrow, T., and Klein-Szanto, A. J. (1993). Benzo[a] pyrene-induced murine skin tumors exhibit frequent and characteristic G to T mutations in the p53 gene. Proc Natl Acad Sci USA 90, 1013–1017. Slaga, T. J., Budunova, I. V., Gimenez-Conti, I. B., and Aldaz, C. M. (1996). The mouse skin carcinogenesis model. J Investig Dermatol Symp Proc 1, 151–156.

190


Slaga, T. J., Fischer, S. M., Weeks, C. E., Klein-Szanto, A. J., and Reiners, J. (1982). Studies on the mechanisms involved in multistage carcinogenesis in mouse skin. J Cell Biochem 18, 99–119. Smith, L. E., Denissenko, M. F., Bennett, W. P., Li, H., Amin, S., Tang, M., and Pfeifer, G. P. (2000). Targeting of lung cancer mutational hotspots by polycyclic aromatic hydrocarbons. J Natl Cancer Inst 92, 803–811. Smithgall, T. E., Harvey, R. G., and Penning, T. M. (1986). Regio- and stereospecificity of homogeneous 3-alpha-hydroxysteroid-dihydrodiol dehydrogenase for trans-dihydrodiol metabolites of polycyclic aromatic hydrocarbons. J Biol Chem 261, 6184–6191. Straif, K., Baan, R., Grosse, Y., Secretan, B., El Ghissassi, F., and Cogliano, V. (2005). Carcinogenicity of polycyclic aromatic hydrocarbons. Lancet Oncol 6, 931–932. Surh, Y. J., Kwon, H., and Tannenbaum, S. R. (1993). Sulfotransferase-mediated activation of 4hydroxy- and 3,4-dihydroxy-3,4-dihydrocyclopenta[c,d]pyrene, major metabolites of cyclopenta[c,d] pyrene. Cancer Res 53, 1017–1022. Vulimiri, S. V., Baer-Dubowska, W., Harvey, R. G., Zhang, J. T., and DiGiovanni, J. (1999). Analysis of highly polar DNA adducts formed in SENCAR mouse epidermis following topical application of dibenz[a,j]anthracene. Chem Res Toxicol 12, 60–67. Watanabe, M., Maher, V. M., and McCormick, J. J. (1985). Excision repair of UV- or benzo[a]pyrene diol epoxide-induced lesions in xeroderma pigmentosum variant cells is “error free.” Mutat Res 146, 285–294. Weyand, E. H., Cai, Z. W., Wu, Y., Rice, J. E., He, Z. M., and LaVoie, E. J. (1993). Detection of the major DNA adducts of benzo[b]fluoranthene in mouse skin: Role of phenolic dihydrodiols. Chem Res Toxicol 6, 568–577. Williams, G. M. (2001). Mechanisms of chemical carcinogenesis and application to human cancer risk assessment. Toxicology 166, 3–10. Wislocki, P. G., Buening, M. K., Levin, W., Lehr, R. E., Thakker, D. R., Jerina, D. M., and Conney, A. H. (1979). Tumorigenicity of the diastereomeric benz[a]anthracene 3,4-diol-1,2-epoxides and the (+)- and (−)-enantiomers of benz[a]anthracene 3,4-dihydrodiol in newborn mice. J Natl Cancer Inst 63, 201–204. Xue, W., and Warshawsky, D. (2005). Metabolic activation of polycyclic and heterocyclic aromatic hydrocarbons and DNA damage: A review. Toxicol Appl Pharmacol 206, 73–93. Yamagiwa, K., and Ichikawa, K. (1915). Experimentelle Studie über die Pathogenese der Epithelialgeschwülste. Mitt Med Fac Kaiserl Univ Tokyo 15, 295–344.

CH A P TE R

7

HORMESIS AND CANCER RISKS: ISSUES AND RESOLUTION Paolo F. Ricci Edward J. Calabrese

7.1.

INTRODUCTION

This chapter aims to discuss of cancer risks and hormesis within the U.S. regulatory risk assessment practices, which legally require the characterization of causal relations between exposures or doses and responses to quantify the probability of cancer response. Causal reasoning characterizes environmental decision-making (EPA 2005; statement in square brackets added) because: The extent of health protection provided to the public ultimately depends upon what risk managers decide is the appropriate course of regulatory action. … When there are alternative procedures having significant biological support [one of which being the existence of hormetic mechanisms], the Agency encourages assessments to be performed using these alternative procedures, if feasible, in order to shed light on the uncertainties in the assessment, recognizing that the Agency may decide to give greater weight to one set of procedures than another in a specific assessment or management decision.

Managing risks, and deciding on the appropriate causal model, is an essential component of both public and private risk decision-making. It follows that current cancer risk assessments must reflect the state-of-science and rely as little as possible on conjecture. Specifically, the U.S. Environmental Protection Agency (EPA) (EPA 2005) adds that: Encouraging risk assessors to be receptive to new scientific information, NRC discussed the need for departures from default options when a “sufficient showing” is made. It called on EPA to articulate clearly its criteria for a departure so that decisions to depart from default options would be “scientifically credible and receive public acceptance”. … It was concerned that ad hoc departures would undercut the scientific credibility of a risk assessment. U.S. National Research Council (NRC) envisioned that principles for choosing and departing from default options would balance several objectives, including “protecting the public health, ensuring scientific validity,


191

192

CHAPTER 7 HORMESIS AND CANCER RISKS: ISSUES AND RESOLUTION

minimizing serious errors in estimating risks, maximizing incentives for research, creating an orderly and predictable process, and fostering openness and trustworthiness.” …

Although guidelines are unenforceable, they are mountains very few attempt to climb because of cost and the truism that one cannot fight City Hall—unless one is able and willing to spend years in litigation. A well-supported reason for recalibrating the credibility and soundness of decisions can be provided in the public interest and address the question, What is a sufficient evidentiary showing? A premise of this discussion is that cancer is a multifactorial disease that has several endpoints—not just the tumor growth itself (primary or metastatic). For example, although cancer can be a solid tumorigenic cellular mass (e.g., a carcinoma), its effects on a living organism go well beyond the abnormal cellular mass (e.g., the tumor can cause cachexia). The justification for any default regulatory model must be based on the weight of the evidence, understood as the collection of outcomes related to exposure and that manifest themselves as adverse endpoints associated with a specific cancer, and not merely with the observation of the tumor itself. Sufficiency (in the context of necessary and sufficient logical expressions) is demonstrated by empirical and theoretical arguments. The EPA arguments about default causal models (i.e., the linear, nonthreshold at low doses cancer dose–response model) is a logical fallacy: Scientific conjecture trumps facts. This chapter deals with correcting the use of conjectures as defaults in regulatory policy, in the context of experimental evidence of hormesis and causation and alternative probabilistic cancer models. Specifically, we summarize how the combination of mode-of-action and weight-of-evidence supports both J-shaped and U-shaped, rather than the linear, no-threshold (LNT) models. The EPA uses the terms nonlinear for the threshold model and low-dose-linear for the LNT models (meaning that the slope is greater than zero at zero dose), which is well-approximated by a straight line, at very low doses and beginning from zero dose (EPA 2005). The totality of the scientific evidence for a causal default—a fundamental dose–response model, given the state-of-science—now discounts conjectural arguments (the linear, at low-dose, nonthreshold model) or arbitrary ones, such as those based on extrapolation (the threshold model) because both of them eliminate a very large number of experimentally observed health benefits. According to the EPA, the use of defaults is a subjective choice (EPA 2005). As the EPA states: Generally, if a gap in basic understanding exists or if agent-specific information is missing, a default option may be used. If agent-specific information is present but critical analysis reveals inadequacies, a default option may also be used. If critical analysis of agent-specific information is consistent with one or more biologically based models as well as with the default option, the alternative models and the default option are both carried through the assessment and characterized for the risk manager. In this case, the default model not only fits the data, but also serves as a benchmark for comparison with other analyses.

But subjectivism raises the paradox: Poorly understood causation requires more, not less, knowledge. Thus, when default dose–response models become the “benchmark” for comparing results with other models, how can the decision be

7.1. INTRODUCTION

193

properly calibrated? Default-based reasoning is not justifiable—contrary to assertions about being precautionary—when experimentally valid alternatives such as the J-shaped cancer model is available and is superior to default-based argumentations: It puts the cart before the horse. A conjecture cannot serve as a benchmark, particularly when that conjecture cannot be demonstrated at the low dose levels generally of concern to regulators [where the individual lifetime excess cancer risks (probability of response) are between 10−4 and 10−6]. Cancer risks are regulated though guidelines (EPA 2005, citations omitted) where: [the u]se health protective risk assessment procedures … means that estimates, while uncertain, are more likely to overstate than understate hazard and/or risk. NRC (1994) reaffirmed the use of default options as “a reasonable way to cope with uncertainty about the choice of appropriate models or theory.”

This combination of scientific analyses and choices (appropriate models) with policy judgment (reasonable way) is fraught with danger. First, causal defaults are conjectured models: For risk management, they are based either on a single model of causation, such as the linear no-threshold (LNT) dose–response model, or on other models that are linear at low dose and originate at the (0, 0) point on the dose– response axis. Second, their combination affects the optimality of risk management choices by inducing unknown and serious differences between the actual and conjectured risks. The LNT conjecture has the distinct disadvantage of negating any potential health benefits; and so does the threshold model. Society pays heavy economic direct and indirect costs from actions designed to avoid a nonexistent danger that, paradoxically, is a benefit. A seemingly plausible reason for default options to be a reasonable way to cope with uncertainty about the choice of appropriate models lies within the proper concern with meeting the ethical principle that it is better to be safe rather than sorry. It is preferable to be conservative when uncertainties are large and when the magnitude of the potential harm is great or dreaded. For example, according to the U.S. Office of Management and Budget (OMB) (2003): The United States employs precautionary approaches throughout the process of risk assessment and management so that the overall level of precaution in a given regulatory decision is appropriate … When analysts assess risks, they frequently use “conservative” or “default” assumptions or explicitly add safety margins or uncertainty factors to characterize a “plausible” upper bound.

But this unimpeachable precautionary approach, as applied, is illogical. It relies on two models that deny proven direct benefits and fails to meet the very reason for its formulation. Moreover, while fundamentally agreeing with the ethical basis of managing cancer risks by being precautionary, the OMB begs the issue that lies within its statement: Is the causal default scientifically sound, given the best and current state of science? In other words, what if conservative defaults cause more harm than good, when applied to regulate exposure? To answer these questions, we first develop a causal network for risk analysis and management. Figure 7.1 depicts the relationship between possible risk management acts that can decrease the incidence of cancer, taking into account other risk factors in the population.

194


act → Δ(exposure) → Δ(cancer incidence) ↑ ↑ [behaviors] [susceptibility] Figure 7.1. A causal graph for risk analysis. The model depicted in this figure can be formalized using a Bayesian network (Ricci et al. 2006): A probabilistic framework interprets the model described in this figure as a Bayesian belief network or causal graph model. Each variable with inward-pointing arrows is interpreted as a random variable with a conditional probability distribution that depends only on the values of the variables that point into it. The essence of this approach to modeling and evaluating uncertain risks is to sample successively from the (often conditional) distribution of each variable, given the values of its predecessors. Algorithms exist to identify and validate possible causal structures.

Figure 7.1 also depicts changes via behaviors, such as occupation, ambient exposure, and predisposition, such as genetic. Logically, it is correct regardless of the shape of the dose–response model. At low dose or at environmental (ambient) exposures, cancer risk assessment models used in regulatory law are either linear or linearized; that is, each is a cumulative distribution function of lifetime cancer risk and thus is a monotonic function. Hormetic cancer dose–response models are also probabilistic; however, they are nonmonotonic (they are relations). The EPA summarizes the reasons for using statistical and probabilistic methods in risk assessment as follows (EPA 2005): The main aim of statistical evaluation is to determine whether exposure to the test agent is associated with an increase of tumor development …. A statistically significant response may or may not be biologically significant and vice versa. The selection of a significance level is a policy choice based on a trade-off between the risks of false positives and false negatives. A result with a significance level of greater or less than 5% (the most common significance level) is examined to see if the result confirms other scientific information ….

This argument leads to a central point of this chapter. That is, what is the evidence necessary and sufficient for a finding of regulatory causation, in the context of the LNT and the hormetic cancer dose–response models? Legal answers to this question were given by Ricci and Molton (1981) and then placed in the context of international tort and environmental law by Ricci and Gray (1998). The following sections focus narrowly on scientific evidence in the context of regulatory science.

7.2. EVIDENCE FOR REGULATORY CANCER RISK ASSESSMENT In U.S. regulatory science, the scientific evidence most relevant to assessing risky outcomes consists of results from animal and human studies. Although there are many other tests, such as in vitro tests using cell lines, the results from those studies are not generally used by regulatory agencies for causal arguments leading to either

7.2. EVIDENCE FOR REGULATORY CANCER RISK ASSESSMENT

195

a guideline or a standard. Regarding empirical studies and causation, the EPA states the following (EPA 2005): Determining whether an observed association (risk) is causal rather than spurious involves consideration of a number of factors. Sir Bradford Hill (Hill 1965) developed a set of guidelines for evaluating epidemiologic associations that can be used in conjunction with the discussion of causality ….

Thus, the EPA opts for a causal analysis (that parallels and updates Hill’s initial work consisting of nine criteria) for judging causation (EPA 2005): (a) Consistency of the Observed Association. An inference of causality is strengthened when a pattern of elevated risks is observed across several independent studies. The reproducibility of findings is one of the strongest arguments for causality. If there are discordant results among investigations, possible reasons such as differences in exposure, confounding factors, and the power of the study are considered. (b) Strength of the Observed Association. The finding of large, precise risks increases confidence that the association is not likely due to chance, bias, or other factors. (c) Specificity of the Observed Association. Based on our current understanding that many agents cause cancer at multiple sites, and many cancers have multiple causes, this is now considered one of the weaker guidelines for causality. Thus, although the presence of specificity may support causality, its absence does not exclude it. (d) Temporal Relationship of the Observed Association. A causal interpretation is strengthened when exposure is known to precede development of the disease. This is among the strongest criteria for an inference of causality. (e) Biological Gradient (Exposure–Response Relationship). A clear exposure– response relationship (e.g., increasing effects associated with greater exposure) strongly suggests cause and effect, especially when such relationships are also observed for duration of exposure (e.g., increasing effects observed following longer exposure times). (f) Biological Plausibility. An inference of causality tends to be strengthened by consistency with data from experimental studies or other sources demonstrating plausible biological mechanisms. A lack of mechanistic data, however, is not a reason to reject causality. (g) Coherence. An inference of causality may be strengthened by other lines of evidence that support a cause-and-effect interpretation of the association. Information is considered from animal bioassays, toxicokinetic studies, and short-term studies. The absence of other lines of evidence, however, is not a reason to reject causality. (h) Experimental Evidence (from Human Populations). Strong evidence of causality can be provided when a change in exposure brings about a change in disease frequency—for example, the decrease in the risk of lung cancer that follows cessation of smoking.

196


(i) Analogy. Information on mode of action for a chemical, as one of many structural analogues, can inform decisions regarding likely causality. None of these nine criteria (individually and in their aggregate) is antithetic to either the J- or inverted J-shaped dose–response model that characterizes hormetic cancer models and toxic effects related to cancer. These, as all criteria of causation we are familiar with, fully support the empirical and biological basis of hormetic causation. But, they do not and cannot support regulatory standards or guidelines based on conjectures. This view is beginning to take hold. For example, the EPA states (Draft in review, Arsenic, Advisory on EPA’s Assessments of Carcinogenic Effects of Organic and Inorganic Arsenic: An Advisory Report of the EPA Science Advisory Board, Dec. 27, 2005) the following (citations omitted for brevity): One cannot dismiss the possibilities of hormesis effects in humans exposed to low-dose arsenic or the essentiality of arsenic to humans. Evidence for essentiality of arsenic has been reported for a number of mammalian species as well as for chickens. These may explain some of the apparent low-dose benefits seen in a variety of systems …. Low concentrations fuel angiogenesis, while higher concentrations injure endothelial cells and promote the vessels dysfunction seen in ischemic diseases and peripheral vascular diseases …. However, arsenic at high doses has been used to destroy the tumor vasculature. If arsenic is essential for humans and/or if epidemiological data could be strengthened at the low-dose range to demonstrate either a low-dose benefit or no effect at low dose, then a threshold is certain. However, at this time, the data are lacking or problematic with regard to low-dose effects. This is an extremely important issue and should be investigated.

The concept of hormesis has been substantially documented with many thousands of studies having passed peer review in numerous journals over multiple decades. A concerted effort has been made to subject each of the possible examples of a hormetic response to the same rigorous a priori evaluative criteria. Those articles passing the a priori hormesis review criteria have been entered into an extensive database (Calabrese and Blain 2005), with many becoming integratively synthesized into comprehensive biomedical and toxicological reviews, including mutagens and carcinogens (Calabrese and Baldwin 1999), toxic metals (Calabrese and Baldwin 2003a; Calabrese and Blain 2004), chemotherapeutics (Calabrese and Baldwin 2003b), reproductive toxins (Calabrese and Baldwin 2000), neuroprotective agents (Calabrese 2008a), growing neurons (Calabrese 2008b), pain (Calabrese 2008c), memory-enhancing agents (Calabrese 2008d), stress (Calabrese 2008e), pglycoprotein membrane efflux systems (Calabrese 2008f), stroke medications (Calabrese 2008g), anxiolytic drugs (Calabrese 2008h), anti-seizure drugs (Calabrese 2008i), chemical and radiation immune stimulatory responses (Calabrese 2005), chemo-attractants and the effects of numerous natural synthetic agonists—for example, estrogens (Calabrese 2001a), androgens (Calabrese 2001b), dopamine (Calabrese 2001c), serotonin (Calabrese 2001d), nitric oxide (Calabrese 2001e), opiates (Calabrese 2001f), prostaglandins (Calabrese 2001g), and adrenergic agonists (Calabrese 2001h), and many others. These findings reveal that the hormetic concept is generalizable, being independent of biological model, endpoint measured and chemical class, and/or physical stressor agent.

7.2. EVIDENCE FOR REGULATORY CANCER RISK ASSESSMENT

197

The data supporting the hormetic dose–response therefore not only far exceeds normal “proof of concept” criteria but have been employed in the development of drugs for numerous human conditions, thereby satisfying a stronger “proof” of application requirements. In head-to-head direct comparisons the hormesis model far outcompeted the threshold and linear at low dose models (Calabrese and Baldwin 2001, 2003c; Calabrese et al. 2006). In fact, while the threshold model was shown to poorly predict below threshold responses the linear at low-dose models, the LNT models cannot be practically validated in either moderate or large-scale studies. For example, in the mid-1970s, the U.S. Food and Drug Administration (FDA) conducted a long-term study with over 24,000 mice exposed to the carcinogen 2-acetylaminofluorene (AAF) to determine the nature of the dose–response in the low-dose zone (because of its magnitude, this study became known as the megamouse experiment). Despite the very large (and yet to be matched in size) number of animals, the estimated cancer risk was only sensitive for a risk of 1/100, far less than the 1 in 1,000,000 to 1 in 100,000 range currently used by regulatory agencies to set tolerable doses or exposures in the United States. The failure to validate risk predictions below 1 in 100 is a serious limitation of the linear at low dose risk assessment because it makes predictions of risk at the very low doses used in regulatory law solely model-dependent and unverifiable. Importantly, the U.S. Society of Toxicology (SOT) 14-member expert panel reviewed the results of the mega-mouse study and reported that it supported a hormetic dose–response model, when the analysis included a time component based on interim sacrifices (Bruce et al. 1981). The SOT indicated that the 2-AAF induced a J-shaped dose–response for bladder cancers that was consistent in each of the six separate rooms in which the large number of animals were housed, thereby relying on an internal replication of the hormetic findings. This study points to an obvious problem: It is practically impossible to demonstrate dose–response behaviors below an excess risk of about 1/100. Although it is possible to develop bioassays that involve several additional dose– response groups in the low-dose region of the experiment, it is the overwhelming evidence across multiple species and endpoints that should demonstrate the superiority of the J- and inverted J-shaped models. Similar findings of J-shaped dose responses in predictive cancer bioassays designed to test hormetic hypotheses have been reported for several epigenetic liver carcinogens by Japanese investigators (Kang et al. 2006; Kinoshita et al. 2006; Puatanachokchai et al. 2006; Fukushima et al. 2005a,b; Hoshi et al. 2004; Sukata et al. 2002; Masuda et al. 2001). Of particular interest have been detailed findings with the banned pesticide and carcinogen dichloro-diphenyl-trichloroethane (DDT). At high doses the Japanese investigators reported that DDT causes dose-dependent increases in liver foci of the Fischer 344 rat (Fukushima et al. 2005a,b; Sukata et al. 2002). However, at lower doses, decreased frequencies of liver foci have been reliably observed, supporting a hormetic interpretation. Detailed mechanistic studies have provided evidence concerning underlying factors that explain high-dose tumor enhancement and low-dose tumor protection caused by exposure to the DDT. These findings are consistent with a broad range of cancer bioassays supporting the hormetic biphasic dose–responses in fish (Brown-Peterson et al. 1999) and rodent models (Kopp-Schneider and Lutz 2001; Teeguarden et al. 2000). Of particular

198


further interest is that low doses of a large number of chemical mutagens act via J-shaped dose responses at low dose (Maki-Paakkanen and Hakulinen 2008; Wilms et al. 2008; Demsia et al. 2007; Lacoste et al. 2006; Pu et al. 2006; Hoshi et al. 2004; Jagetia et al. 2003; Knasmuller et al. 2002; Sasaki et al. 2002; Hartmann et al. 2001; Kirkland and Muller 2000) further supporting the observations of hormetic dose– responses within cancer bioassays. Other evidence indicates that low doses of various xenobiotics enhance cell-to-cell communication function reducing possible risks of tumor promotion (Rivedal and Witz 2005; Jeong et al. 2001; Rivedal et al. 2000; Mikalsen and Sanner 1994; Mercier et al. 1993; Mikalsen et al. 1992; LochCaruso et al. 1984; Kurata et al. 1982). Furthermore, an extensive literature is available indicating that low doses of numerous immune system active agents display biphasic dose–responses enhancing parameters related to immune surveillance at low concentrations further supporting an hormetic interpretation (Calabrese 2005). The arguments justifying hormetic models as the regulatory defaults rely on fundamental scientific reasoning and meet Hill’s nine criteria and the more current causal criteria. Rather than relying on a conjecture that is not provable and denies true benefits, and thus avoids costly societal errors, the regulatory default should be based on the overall evidence for hormetic behaviors and the resulting estimates obtained by the J-model for cancer, or by its inverted form for toxicological endpoints.

7.3. HORMESIS AND CANCER RISK ASSESSMENT: MODELS The hormetic dose–response model for cancer is J-shaped. It accounts for and resolves several of the issues that cannot be resolved by the practical use of its regulatory alternatives. A direct way to test the validity of a hormetic statement is to assess if there is scientifically sound evidence that demonstrates adaptive, nonadverse, or beneficial events that (a) meet good scientific practices, causal criteria, peer reviews, and open discussions, (b) avoid secrecy, and (c) are independent of funding sources and other determining factors. This suggests framing policy science by answering the following two questions: 1. Theoretical/Empirical Question: Since the LNTs negate any protective response at low dose rates, what is the appropriate science policy to overcome this limitation? 2. Corollary Question: Since the J-shaped hormetic (or biphasic) cancer dose– response model yields empirically demonstrated protective (stimulatory) effects at low doses in one or more species, is biologically plausible, and describes a damaging relationship at higher dose that is consistent with the LNT, which of the two is the logical and prudential default model? Although the answer to the first question is legal, and thus beyond the scope of this chapter, the answer to the second question falls well within our framework. We can begin to frame the answers by a limited review of current well-known cancer dose–response models. The hormetic J-shaped model is depicted in Figure 7.2.

7.3. HORMESIS AND CANCER RISK ASSESSMENT: MODELS

199

Adverse response (%) 100 Experimental results (dots are enlarged for exposition

LNT model Hormetic model Experimental control response

Experimental region of adverse responses, common to LNT and hormetic models

Dose rate

Protection optimized

Figure 7.2. Biphasic (hormetic) dose–response model for cancer incidence (the percent response in the controls must be nonzero). Protection is considered “optimized” because it represents the greatest degree of protection at a dose range furthest away from an adverse effect.

The LNT-based dose–response model for cancer, being a cumulative distribution function, begins at zero and is proportional to doses (i.e., is linear at low doses, resulting in the LNT hypothesis). The early form of the LNT model is the one-hit model: Pr ( D ) = 1 − exp [ − ( qD )] In this model, Pr(D) is the lifetime probability of cancer death from lifetime exposure to dose D (often expressed in units of mg/kg-day, consistent with animals’ exposures). The multistage model is Pr ( D ) = 1 − exp [ − ( ∑ i qi D i )] where the same notation used for the single-hit model applies. This model can account for a threshold, but cannot account for any beneficial effect of exposure. A more recent model is the Moolgalvkar–Dewanjii–Venzon (MVK) model (Moolgalvkar et al. 1988), which is a two-stage stochastic model that accounts for cell growth, death, and differentiation. Figure 7.3 depicts the two-stage MVK cellular process in which two adverse and irreparable events must occur for normal cells to become malignant. The events may be mutations or other effects inherited by the cells. The cellular process consists of two stages (excluding the stage in which cells are normal) and the following transition rates, [cells/time]−1, as follows:

200


Normal cells: N1

μ1

Preneoplastic cells: N2

β1

α1

β2

Malignant cells: N3

μ2

α2

β3

α3

Dead or differentiated cells

Figure 7.3.

Cellular biology of the two-stage MVK model.

1. Cellular death or differentiation, βi, (a carcinogen decreases it) 2. Cellular division into a normal or premalignant cell, μi, (a carcinogen increases it) 3. Cellular division into two normal cells, αi, the mitotic rate, (a carcinogen can increase it) The parameters of the MVK model can be dose-dependent. As Figure 7.3 depicts, the MVK model can represent cell proliferation due to exposure to a chemical that aids such proliferation and can account for different cell division rates. The assumptions include the following: (1) Cancer is a two-stage process, (2) cellular transformations are independent, and (3) once a cell becomes malignant, potentially cancerous cells proliferate independently of the normal cells resulting in a detectable cancer. Despite their conceptual value and practical successes, these stochastic transition models leave some important phenomena unexplained. These include (Cox and Ricci 2005): (a) Importance of Proliferation of Normal Cells in Increasing Cancer Risks. Many chemical carcinogens were found to increase tumor rates in experimental animals only in situations that also cause cytotoxicity and regenerative hyperplasia or compensating proliferation of apparently normal cell populations in response to the toxic injury. Examples include chloroform, diesel exhaust, formaldehyde, and many others. When such compensating proliferation is a prerequisite for chemically induced carcinogenesis, traditional linearized multistage modeling may overestimate risks at low concentrations or predict significant risks at low concentrations even if none truly exists. Thus, dose– response models that better account for the role of normal stem cell proliferation and kinetics following cytotoxic damage may be needed to obtain more realistic risk estimates for some chemicals. (b) Carcinogenic Thresholds in Dose Rate and/or Duration of Exposure Are Arbitrary. A generalization of these stochastic models allows for more than one possible sequence of events (e.g., somatically heritable transformations,

7.3. HORMESIS AND CANCER RISK ASSESSMENT: MODELS

201

epigenetic events) leading from normalcy to malignancy, while keeping the feature that some events must be completed before others can occur. A general process of this type can be described as an event tree, explicitly enumerating the possible sequences of events that take a normal stem cell to a malignant one. Each node in the tree corresponds to the sequence of events that have occurred so far, and the successors of each node are the events that can occur next, corresponding to branches out of the current node. The branch probabilities at each node complete the specification of the process. A directed acyclic graph (DAG) indicating allowed transitions among events, along with their probabilities, [i.e., a stochastic transition network (STN)], often provides a more concise representation of the same information. However, the tree provides a clear, useful conceptual model of multiple alternative paths and precedence partial ordering constraints among events.

7.3.1.

Answers to Our Question

The answer to the question we posed earlier in this essay is clear: The time is ripe for including hormesis as the principal regulatory model because it is not conjectural and is based on data consistent with all criteria put forth in the EPA’s causal arguments. Specifically, this answer is justified by the following findings: • The hormetic dose response can be tested because its low-dose response starts immediately to the left (in the dose–response space) of any hypothetical threshold. Recollecting that the threshold model is the linearized form of the S-shaped toxicological cumulative distribution of responses, this response is generally not within the observations (it is an extrapolation via a probit transformation from the experimental results to a dose intercept). On the other hand, the hormetic dose–response can be either validated or rejected with normal testing protocols, provided that a sufficient number of experimental results are available (five or more). • The hormetic dose–response can predict harm below or above the toxicological threshold, and thus it is consistent with positive and negative outcomes, unlike the LNT or the S-shaped models. • The hormetic dose–response model can predict the occurrence of beneficial responses below the toxicological threshold. This can be seen with endpoints such as enhanced longevity, decreased disease incidence, and improved cognition, unlike the threshold and linear at low-dose (LNT) models. • Chemical interactions can be accounted for. While threshold dose response model can only deal with chemical interactions for responses that exceed a threshold, the hormetic model also does this. These models differ where the interaction occurs in the hormetic stimulatory zone. In the case of the hormetic chemical interactions, the maximum response is still constrained to 30–60% above the control value a characteristic that the threshold and linear at lowdose models do not have.

202


• A hormetic dose–response model has the same dose–response for all biological models, endpoints, and chemical or physical agents. This means that hormesis can harmonize risk assessment procedures for both carcinogens and noncarcinogens. • The hormetic dose–response can account for different biological mechanisms, in different species, and for different endpoints. This is relevant to the EPA “mode of action,” which is a weak form “mechanism of action” (EPA 2005). Hormetic dose–response models thus present a significant new challenge and opportunity for regulatory agencies because they permit those agencies to address the question of what is an “optimized” societal acceptable exposure or dose. That is, it answers the question, What exposure standard yields the greatest overall (societally optimal) health benefits? For example, a dose that may provide a beneficial effect in the normal population may be a harmful to those in a high subgroup. Conversely, a dose that provides a beneficial effect in a high-risk group may not have biological impact on the normal population. Because the normal population may be 95% of the entire population, with the high-risk segment the remainder, the total number of increased years of life for the members of society may occur if the exposure standard were established to maximally protect normal individuals. With the use of the threshold and linear at low-dose models, this situation cannot be assessed. Federal agencies, such as the EPA, typically state that an environmental standard is set at a dose that will protect all normal- and most high-risk members of the population. The hormetic model actually allows the possibility of avoiding hazardous exposures and increasing health benefits with the challenge of estimating the optimal overall response for society. The hormesis databases developed by Calabrese and his colleagues are supportive of the EPA mode of action and weight of evidence. For example, they are consistent (in fact, essential) to meet the EPA’s requirement of a weight-of-evidence narrative that should describe and be intelligible to risk managers and nonexpert readers (EPA 2005) regarding: • The quality and quantity of the data • All key decisions and the basis for these major decisions • Data, analyses or assumptions that are unusual for or new to EPA Specifically, this narrative should include (EPA 2005) the following: • Conclusions about human carcinogenic potential (choice of descriptor(s), …) • A summary of the key evidence supporting these conclusions (for each descriptor used), including information on the type(s) of data (human and/or animal, in vivo and/or in vitro) used to support the conclusion(s) • Available information on the epidemiologic or experimental conditions that characterize expression of carcinogenicity (e.g., if carcinogenicity is possible only by one exposure route or only above a certain human exposure level) • A summary of potential modes of action and how they reinforce the conclusions

ACKNOWLEDGMENTS

203

• Indications of any susceptible populations or lifestages, when available • A summary of the key default options invoked when the available information is inconclusive The mode of action in the assessment of potential carcinogens is a main focus of these cancer guidelines (EPA 2005). Its aspects involve a general acceptance criterion that establishes a particular sequence of biological key events and processes beginning at the cellular level and ending with the tumor. The important characteristic of the mode of action concept is that those key events are empirically observable precursors and thus are necessary, but perhaps not sufficient, stages in the development of cancer. The EPA also defines a mechanism of action, which has an even higher granularity than the mode of action: Molecular events fall under this rubric. The mode of action is, according to the EPA, a data-rich assessment. We add that it is the full description of the biological process being investigated, regardless of the endpoint under study. If so, the compelling empirical evidence in the hormetic databases fulfills the data-rich aspect of any empirical and theoretical analysis of hormetic behaviors. The same cannot be said for the LNT or the threshold models: The former is conjectural; the latter is an extrapolation that disregards the complete biological process.

7.4.

CONCLUSIONS

The first conclusion is that the factual and theoretical evidence points to replacing the classical causal regulatory defaults used to deal with low dose–response, the linear no-threshold, and the linear at low-dose–response models, or monotonic functions, with the J- and inverse J-shaped models—or relations. These models have been demonstrated to apply to toxicological and cancer outcomes for a very wide range of substances and diseases. The classical defaults may still be applicable on a case-by-case basis. The reasons for changing the defaults include the fact that the J-shaped class of models quantifies a wide set of health benefits that are completely excluded from estimations that use monotonic models. We conclude that replacing both a conjecture and an arbitrary model with two theoretically and empirically sound ones leads to rational decision and does not exclude actually demonstrable benefits. Overall, the sum is positive for society.

ACKNOWLEDGMENTS The effort of Edward J. Calabrese was sponsored by the Air Force Office of Scientific Research, Air Force Materiel Command, USAF, under grant number FA9550-07-10248. The U.S. Government is authorized to reproduce and distribute for governmental purposes notwithstanding any copyright notation thereon. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsement, either expressed or implied, of the Air Force Office of Scientific Research or the U.S. Government.

204


REFERENCES Brown-Peterson, N., Krol, R. M., Zhu, Y., and Hawkins, W. E. (1999). N-Nitrosodiethylamine initiation of carcinogenesis in Japanese Medaka (Oryzias latipes): Hepatocellular proliferation, toxicity, and neoplastic lesions resulting from short term, low level exposure. Toxicol Sci 50, 186–194. Bruce, R. D., Carlton, W. W., Ferber, K. H., Hughes, D. H., Quast, J. F., Salsburg, D. S., Smith, J. M. (members of the Society of Toxicology ED01 Task Force), Brown, W. R., Cranmer, M. F., Sielken, J. R., Van Ryzin, J., and Barnard, R. C. (1981). Re-examination of the ED01 study whey the society of toxicology became involved. Fund Appl Toxicol 1, 26–128. Calabrese, E. J. (2001a). Estrogen and related compounds: Biphasic dose responses. Crit Rev Toxicol 31, 503–515. Calabrese, E. J. (2001b). Androgens: Biphasic dose responses. Crit Rev Toxicol 31, 517–522. Calabrese, E. J. (2001c). Dopamine: Biphasic dose responses. Crit Rev Toxicol 31, 563–583. Calabrese, E. J. (2001d). 5-Hydroxytryptamine (serotonin): Biphasic dose responses. Crit Rev Toxicol 31, 553–561. Calabrese, E. J. (2001e). Nitric oxide: Biphasic dose responses. Crit Rev Toxicol 31, 489–501. Calabrese, E. J. (2001f). Opiates: Biphasic dose responses. Crit Rev Toxicol 31, 585–604. Calabrese, E. J. (2001g). Prostaglandins: Biphasic dose responses. Crit Rev Toxicol 31, 475–487. Calabrese, E. J. (2001h). Adrenergic receptors: Biphasic dose responses. Crit Rev Toxicol 31, 523–538. Calabrese, E. J. (2005). Hormetic dose–response relationships in immunology: Occurrence, quantitative features of the dose response, mechanistic foundations, and clinical implications. Crit Rev Toxicol 35, 89–295. Calabrese, E. J. (2008a). Dose–response features of neuroprotective agents: An integrative summary. Crit Rev Toxicol 38, 253–348. Calabrese, E. J. (2008b). Pharmacological enhancement of neuronal survival. Crit Rev Toxciol 38, 349–389. Calabrese, E. J. (2008c). Pain and U-shaped dose responses: Occurrence, mechanisms and clinical implications. Crit Rev Toxicol 38, 579–590. Calabrese, E. J. (2008d). Alzheimer ’s disease drugs: An application of the hormetic dose–response model. Crit Rev Toxicol 38, 419–451. Calabrese, E. J. (2008e). Stress biology and hormesis: The Yerkes–Dodson law in psychology—A special case of the hormesis dose response. Crit Rev Toxicol 38, 453–462. Calabrese, E. J. (2008f). P-glycoprotein efflux transporter activity often displays biphasic dose–response relationships. Crit Rev Toxicol 38, 473–487. Calabrese, E. J. (2008g). Drug therapies for stroke and traumatic brain injury often displays U-shaped dose responses: Occurrence, mechanisms, and clinical implications. Crit Rev Toxicol 38, 557–577. Calabrese, E. J. (2008h). An assessment of anxiolytic drug screening tests: Hormetic dose responses predominate. Crit Rev Toxicol 38, 489–542. Calabrese, E. J. (2008i). Modulation of the epileptic seizure threshold: Implications of biphasic dose responses. Crit Rev Toxicol 38, 543–556. Calabrese, E. J., and Baldwin, L. A. (1999). Can the concept of hormesis be generalized to carcinogenesis. Regul Toxicol Pharmacol 28, 230–241. Calabrese, E. J., and Baldwin, L. A. (2000). Reproductive toxicity and hormetic responses. In Toxicology in Risk Assessment, Salem, H., ed., Taylor & Francis, Philadelphia, p. 106. Calabrese, E. J., and Baldwin, L. A. (2001). The frequency of U-shaped dose-responses in the toxicological literature. Toxicol Sci 62, 330–338. Calabrese, E. J., and Baldwin, L. A. (2003a). Inorganics and hormesis. Crit Rev Toxicol 33, 215–304. Calabrese, E. J., and Baldwin, L. A. (2003b). Chemotherapeutics and hormesis. Crit Rev Toxicol 33, 305–353. Calabrese, E. J., and Baldwin, L. A. (2003c). The hormetic dose response model is more common than the threshold model in toxicology. Toxicol Sci 71, 246–250. Calabrese, E. J., and Blain, R. (2004). Metals and hormesis. J Environ Monit 6, 14N–19N. Calabrese, E. J., and Blain, R. (2005). The occurrence of hormetic dose responses in the toxicological literature, the hormesis database: An overview. Toxicol Appl Pharmacol 202, 289–301.

REFERENCES

205

Calabrese, E. J., Staudenmayer, J. W., Stanek, E. J., and Hoffmann, G. R. (2006). Hormesis outperforms threshold model in NCI anti-tumor drug screening data. Toxicol Sci 94, 368–378. Cox, L. A., and Ricci, P. F. (2005). Causation in risk assessment and management: Models, inference, biases, and a microbial risk–benefit case study. Environ Int 31, 377–397. Demsia, G., Vlastos, D., Goumenou, M., and Matthopoulos, D. P. (2007). Assessment of the genotoxicity of imidaclorpid and metalaxyl in cultured human lymphocytes and rat bone-marrow. Mutat Res 634, 32–39. EPA (2005). Guidelines for carcinogen risk assessment, EPA/630/P-03/001F. Fukushima, S., Kinoshita, A., Puatanachokchai, R., Kushida, M., Wanibuchi, H., and Morimura, K. (2005a). Hormesis and dose–response-mediated mechanisms in carcinogenesis: Evidence for a threshold in carcinogenicity of non-genotoxic carcinogens. Carcinogenesis 26, 1835–1845. Fukushima, S., Wanibuchi, H., Morimura, K., Nakae, D., Tsuda, H., Imaida, K., Shirai, T., Tatematsu, M., Tsukamoto, T., Hirose, M., and Furukawa, F. (2005b). Lack of potential of low dose Nnitrosodimethylamine to induce preneoplastic lesions, glutathione S-transferase placental form-positive foci, in rat liver. Cancer Lett 222, 11–15. Hartmann, A., Kiskinis, E., Fjallman, A., and Suter, W. (2001). Influence of cytotoxicity and compound precipitation on test results in the alkaline comet assay. Mutat Res 497, 199–212. Hill, A. B. (1965). The environment and disease: Association or causation? Proc R Soc Med 58, 295–300. Hoshi, M., Morimura, K., Wanibuchi, H., Wei, M., Okochi, E., Ushijama, T., Takaoka, K., and Shoji, F. (2004). No-observed effect levels for carcinogenicity and for in vivo mutagenicity of a genotoxic carcinogen. Toxicol Sci 81, 273–279. Jagetia, G. C., Venkatesh, P., and Baliga, M. S. (2003). Evaluation of the radioprotective effect of Aegle marmelos (l.) Correa in cultured human peripheral blood lymphocytes exposed to different doses of gamma-radiation: A micronucleus study. Mutagenesis 18, 387–393. Jeong, S.-H., Cho, M.-H., and Cho, J.-H. (2001). Effects of cadmium on gap junctional intercellular communication in WB-F344 rat liver epithelial cells. Hum Exp Toxicol 20, 577–583. Kang, J. S., Wanibuchi, H., Morimura, K., Totsuka, Y., Yoshimura, I., and Fukushima, S. (2006). Existence of a no effect level for MelQx hepatocarcinogenicity on a background of thioacetamideinduced liver damage in rats. Cancer Sci 97, 453–458. Kinoshita, A., Wanibuchi, H., Wei, M., and Fukushima, S. (2006). Hormesis in carcinogenicity of nongenotoxic carcinogens. J Toxicol Pathol 19, 111–122. Kirkland, D. J., and Muller, L. (2000). Interpretation of the biological relevance of genotoxicity test results: The importance of thresholds. Mutat Res 464, 137–147. Knasmuller, S., Steinkellner, H., Majer, B. J., Nobis, E. C., Scharf, G., and Kassie, F. (2002). Search for dietary antimutagens and anticarcinogens: Methodological aspects and extrapolation problems. Food Chem Toxicol 40, 1051–1062. Kopp-Schneider, A., and Lutz, W. K. (2001). J-shaped dose–response relationship for tumor induction by caffeic acid in the rat forestomach, modeled by non-monotonic dose response for DNA damage and cell proliferation. HERA 7, 921–931. Kurata, M., Hirose, K., and Umeda, M. (1982). Inhibition of metabolic cooperation in Chinese hamster cells by organochlorine pesticides. Gann 73, 217–221. Lacoste, S., Castonguay, A., and Drouin, R. (2006). Formamidopyrimidine adducts are detected using the comet assay in human cells treated with reactive metabolites of 4-(methylnitrosamino)-1-(3pyridyl)-1-butanone (NNK). Mutat Res 600, 138–149. Loch-Caruso, R., Trosko, J. E., and Corcos, I. A. (1984). Interruption of cell-to-cell communication in Chinese hamster V79 cells by various alkyl glycol ethers: Implications for teratogenicity. Environ Health Perspect 57, 119–123. Maki-Paakkanen, J., and Hakulinen, P. (2008). Assesssment of the genotoxicity of the rat carcinogen 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX) in rat liver epithelial cells in vitro. Toxicol in Vitro 22, 535–640. Masuda, C., Wanibuchi, H., Otori, K., Wei, M., Yamamoto, S., Hiroi, T., Imaoka, S., Funae, Y., and Fukushima, S. (2001). Presence of a no-observed effect level for enhancing effects of development of the α-isomer of benzene hexachloride (α-BHC) on diethylnitrosamine-initiatied hepatic foci in rats. Cancer Lett 163, 179–185.

206


Mercier, T., Honikman-Leban, E., Chaumontet, C., Martel, P., and Shahin, M. M. (1993). Studies on the modulating effects of retinoic acid and retinol acetate using dye transfer and metabolic cooperation assays. Fundam Appl Toxicol 21, 270–276. Mikalsen, S.-O., Rivedal, E., and Sanner, T. (1992). Heavy metal ions, cytotoxicity and gap junctional intercellular communication in Syrian hamster embryo cells. ATLA 20, 213–217. Mikalsen, S.-O., and Sanner, T. (1994). Increased gap junctional intercellular communication in Syrian hamster embryo cells treated with oxidative agents. Carcinogenesis 15, 381–387. Moolgalvkar, S. H., Dewanjii, A., and Venzon, D. J. (1988). A stochastic two-stage model for cancer risk assessment, 1: The hazard function and probability of tumor. Risk Anal 8, 383–392. National Research Council (NRC). (1994). Committee on Risk Assessment of Hazardous Air Pollutants, US National Academy of Science, Science and Judgment in Risk Assessment, National Academies Press, Washington, D.C. Office of Management and Budget (OMB). (2003). Proposed Risk Assessment Bulletin. http://www. whitehouse.gov/omb/inforeg/proposed_risk_assessment_bulletin_010906.pdf. Pu, X., Kamendulis, L. M., and Klaunig, J. E. (2006). Acrylonitrile-induced oxidative DNA damage in rat astrocytes. Environ Mol Mutagen 47, 631–638. Puatanachokchai, R., Morimura, K., Wanibuchi, H., Oka, M., Kinoshita, A., Mitsuru, F., Yamaguchi, S., Funae, Y., and Fukushima, S. (2006). Alpha-benzene hexachloride exerts hormesis in preneoplastic lesion formation of rat hepatocarcinogenesis with the possible role of hepatic detoxifying enzymes. Cancer Lett 240, 102–113. Ricci, P. F., Cox, L. A. Jr., and MacDonald, T. (2006). Science-policy in environmental and health risk assessment: If we cannot do without, can we do better? Hum Exp Toxicol 25, 29–43. Ricci, P. F., and Gray, N. (1998). Towards a new way to deal with toxic torts: Risks in toxic tort law. Part I: Probabilistic causation and legal causation. Univ New South Wales Law J 21, 787–806. Ricci, P. F., and Molton, L. (1981). Risk and benefits in environmental law. Science 214, 1096. Rivedal, E., Mikalsen, S.-O., and Sanner, T. (2000). Morphological transformation and effect on gap junction intercellular communication in Syrian hamster embryo cells as screening tests for carcinogens devoid of mutagenic activity. Toxicol in Vitro 14, 185–192. Rivedal, E., and Witz, G. (2005). Metabolites of benzene are potent inhibitors of gap–junction intercellular communication. Arch Toxicol 79, 303–311. Sasaki, Y. F., Kawaguchi, S., Kamaya, A., Ohshita, M., Kabasawa, K., Iwama, K., Taniguchi, K., and Tsuda, S. (2002). The comet assay with 8 mouse organs: results with 39 currently used food additives. Mutat Res 519, 103–119. Sukata, T., Uwagawa, S., Ozaki, K., Ogawa, M., Nishikawa, T., Iwai, S., Kinoshita, A., Wanibuchi, H., Imaoka, S., Punae, Y., Okuno, Y., and Fukushima, S. (2002). Detailed low-dose study of 1,1-B IS (p-chlorophenyl)-2,2,2-trichloroethane carcinogenesis suggests the possibility of a hormetic effect. Int J Cancer 99, 112–118. Teeguarden, J. G., Dragan, Y., and Pitot, H. C. (2000). Hazard assessment of chemical carcinogens: The impact of hormesis. J Appl Toxicol 20, 113–120. Wilms, L. C., Kleinjans, J. C. S., Moonen, E. J. C., and Briede, J. J. (2008). Discriminative protection against hydroxyl and superoxide anion radicals by quercetin in human leucocytes in vitro. Toxicol in Vitro 22, 301–307.

CH A P TE R

8

THRESHOLDS FOR GENOTOXIC CARCINOGENS: EVIDENCE FROM MECHANISM-BASED CARCINOGENICITY STUDIES Shoji Fukushima Min Wei Anna Kakehashi Hideki Wanibuchi

8.1.

OVERVIEW

In this chapter, the results of a medium-term rat carcinogenicity bioassay for rapid in vivo detection of carcinogenic potential are presented to examine the carcinogenicity of low doses of five genotoxic carcinogens: 2-amino-3,8-dimethylimidazo [4,5-f ] quinoxaline (MeIQx), a heptocarcinogen contained in seared fish and meat; N-nitrosodiethylamine (DEN) and N-nitrosodimethylamine (DMN), heptocarcinogens synthesized in the stomach through the reaction of secondary amines and nitrites; 2-amino-1-methyl-6-phenylimidazo[4,5-b] pyridine (PhIP), a colon carcinogen contained in seared meat and fish; and potassium bromate, a kidney carcinogen that is a contaminate of tap water and also used as a food additive in some countries. DNA damage, gene mutation, and surrogate endpoints for carcinogenicity were examined: Carcinogenic endpoints were glutathione S-transferase placental form (GST-P) positive foci in the liver, a well-known preneoplastic lesion marker in rat hepatocarcinogenesis, and altered crypt foci (ACF), a well-known surrogate marker of preneoplastic lesions in the colon. Low doses of MeIQx induced formation of DNA-MeIQx adducts; somewhat higher doses caused elevation of 8-hydroxy-2′deoxyquanosine (8-OHdG) levels; at still higher doses, gene mutations occurred; and the very highest dose of MeIQx induced formation of GST-P positive foci. Similarly, only the highest doses of DEN and DMN caused an increase in the number of GST-P positive foci in the liver; the lower doses had no effect. Similar results were obtained with the colon carcinogen PhIP. PhIP–DNA adduct formation was


207

208

CHAPTER 8 THRESHOLDS FOR GENOTOXIC CARCINOGENS

observed after treatment with low doses, while only high doses were found to induce ACF. Finally, in experiments with potassium bromate, 8-OHdG formation, GC-toTA transversions, and gene mutations in the rat kidney were observed only after administration of the highest doses of KBrO3; histopathological changes related to carcinogenicity were not observed at any dose used. These data support the existence of thresholds for the genotoxic carcinogens examined in this chapter.

8.2.

INTRODUCTION

Compounds known to be carcinogenic to humans have primarily been identified by epidemiological methods—for instance, cancer development in factory workers (occupational cancer) (see Chapter 15). Epidemiological data, however, are usually not suitable to establish risk from exposure to different levels of human carcinogens. Also, epidemiological data are available only after exposed humans develop cancers. Carcinogen risk assessment aspires to identify and assess risk from exposure to carcinogens prior to extensive human exposure. Identification and assessment of most carcinogens is done using two-year carcinogenicity tests performed in rodents, particularly rats and mice (see Chapter 14). To assess risk in humans, carcinogenic response curves obtained from these tests are used. Importantly, the carcinogenicity of low doses of carcinogenic compounds is generally extrapolated from the carcinogenicity data obtained using high doses; to obtain statistically acceptable data, carcinogens are generally used in rodent carcinogenicity tests at high doses, including the maximum tolerated dose. The principal method of assessing risk posed to humans by exposure to genotoxic carcinogens uses nonthreshold approaches to model experimental data: The curves generated by nonthreshold approach modeling are S-shaped or linear low-dose straight lines that reach zero (see Chapter 24). This “nonthreshold concept” of genotoxic carcinogenicity reflects the idea that a single event caused by a genotoxic carcinogen can have a positive influence on cancer development in humans. However, the physiology of living organisms suggests that, in practical terms, thresholds can exist, even for genotoxic carcinogens. Most chemical carcinogens must be metabolized within the cell to their active forms, known as the ultimate carcinogen, before they are able to exert their carcinogenic activity. The ultimate carcinogen formed from most genotoxic compounds binds covalently to DNA, forming an adduct. These adducts can interfere with normal DNA metabolism, leading to DNA mutations and carcinogenicity. However, these DNA adducts are efficiently repaired by the cell. Still, for any particular adduct there is the possibility of misrepair or replication of damaged DNA resulting in fixation of a mutation into the cell’s genome. Therefore, there is a finite risk of mutation arising from a single adduct. Next, at the level of DNA mutation, carcinogen–DNA adduct formation is essentially random in the euchromatic DNA (DNA that is not highly condensed); consequently, only a minute fraction of the mutations arising from these adducts will actually occur in a gene and have an affect on the cell, and only a very small fraction of these will be carcinogenic. In practical terms then, only a minute fraction of adducts actually give rise to DNA mutations and only a minute fraction of these mutations will affect the cell. As the number of adducts increases, however, the possibility of mutations occurring increases and mutated cells eventually begin

8.3. LOW-DOSE CARCINOGENICITY OF MEIQX IN THE RAT LIVER

209

to appear (Kuraoka 2008). This is especially relevant as the number of adducts becomes greater than the capacity of the cell to repair this damage. Most damaged or mutated cells will die due to metabolic dysfunctions or be eliminated by irreversible senescence or apoptosis, but it is possible that some will survive and be viable. In the two-stage chemical carcinogenesis model, this sequence of events is thought to occur during the initiation stage. Cell proliferation enhances the ability of initiated cells to form preneoplastic lesions and to develop into tumors, benign and then malignant. Evidence indicates that, before developing into tumors, most preneoplastic lesions disappear spontaneously, presumably due, at least in part, to elimination by the immune system. The development from initiated cells into tumors is the promotion stage of the two-stage chemical carcinogenesis model. Therefore, in a finite population, if physiological functions such as DNA repair, induction of senescence or apoptosis, and immune surveillance are effective, there will be levels of exposure to genotoxic carcinogens below which induction of carcinogenesis is effectively zero (see Chapter 26). For detection of carcinogenicity, the standard method is long-term carcinogenicity testing in two rodent species, such as mice and rats (≥50 animals/sex/group), with at least three dose levels, and in-life study termination at 18 months for mice and 24 months for rats (OECD 1981). However, such tests are extremely timeconsuming, laborious, and expensive. This is particularly true when examining the effects of low doses of suspected carcinogens since many more animals are required to reliably determine whether the low doses used are in fact able to induce an increase in tumor formation. In practical terms, it is currently impossible to examine the carcinogenicity of all suspect compounds using long-term rodent assays. Therefore, recently, an alternative method to long-term carcinogenicity testing in which preneoplastic lesions are accepted as endpoint markers for the assessment of carcinogenicity has been proposed (Tsuda et al. 2003). Results are obtained from this in vivo medium-term bioassay system of carcinogens in a matter of weeks rather than, as with long-term testing, many months. The presence or absence of a threshold will determine the reliability of carcinogenic risk assessment when extrapolated from high-dose rodent testing. Therefore, it is essential to verify scientifically whether the nonthreshold concept is valid. Herein, we provide data from low-dose carcinogenicity studies for genotoxic carcinogens using a medium-term bioassay for carcinogens. In addition to determining no-effect doses for carcinogenicity, we also examined markers that cells typically acquire as they move through the initiation and promotion stages of carcinogenesis. Analysis of all the data strongly support the existence of thresholds for the carcinogenetic effects of the five genotoxic carcinogens examined.

8.3. LOW-DOSE CARCINOGENICITY OF 2-AMINO3,8-DIMETHYLIMIDAZO[4,5-F]QUINOXALINE (MEIQx) IN THE RAT LIVER MeIQx is a heterocyclic amine contained in fried meat and fish. MeIQx at doses of 100–400 ppm in the diet is carcinogenic in the rat liver (Kato et al. 1988). To investigate the effect of exposure to low doses of MeIQx, 1145 21-day-old male F344

210


rats were divided into seven groups and administered MeIQx in the diet at doses of 0, 0.001, 0.01, 0.1, 1, 10 ppm (low-dose groups) and 100 ppm (high-dose group) for 4–32 weeks (Fukushima et al. 2002). The data on the induction of the GST-P positive foci after treatment with various doses of MeIQx for 16 weeks is presented in Table 8.1 (GST-P positive foci is a preneoplastic lesion in rat hepatocarcinogenesis and the endpoint marker in the rat liver medium-term carcinogenicity bioassay). The numbers of GST-P positive foci were not significantly elevated in the 0.001– 10 ppm MeIQx groups, but a statistically significant increase was detected in the 100 ppm group. The same results were observed when the treatments with MeIQx were continued for 32 weeks (Figure 8.1). MeIQx is metabolized in liver cells to an ultimate carcinogen capable of covalently binding DNA. In contrast to GST-P foci induction, the formation of

TABLE 8.1. Induction of GST-P positive Foci in the Liver of Rats Treated with MeIQx for 16 Weeks

Group 1 2 3 4 5 6 7

MeIQx Dose (ppm)

Number of Rats

0 0.001 0.01 0.1 1 10 100

150 150 150 150 150 50 50

Size Distribution of GST-P Positive Foci (No./cm2) 2–4 Cells

5–10 Cells

≥11 Cells

Total

0.12 ± 0.17 0.12 ± 0.18 0.13 ± 0.21 0.14 ± 0.20 0.16 ± 0.20 0.35 ± 0.33 13.86 ± 5.11a

0.05 ± 0.17 0.02 ± 0.06 0.03 ± 0.07 0.04 ± 0.08 0.04 ± 0.08 0.10 ± 0.12 8.85 ± 3.23a

0.02 ± 0.09 0.01 ± 0.05 0.01 ± 0.05 0.02 ± 0.10 0.02 ± 0.07 0.01 ± 0.05 6.51 ± 4.06a

0.18 ± 0.35 0.15 ± 0.19 0.16 ± 0.24 0.19 ± 0.25 0.21 ± 0.24 0.47 ± 0.35 29.2 ± 10.99a

p > 0.01 (vs. group 1).

a

No. of GST-P positive foci (no./cm2)

100

*

10

1

0.1

0

0.01

0.1

1

10

100

MeIQx (ppm, in diet)

Figure 8.1. GST-P positive foci in the livers of F344 rats treated with MeIQx at various doses for 32 weeks. Asterisk (*) indicates p < 0.01 versus 0 ppm group.

211


(/107nds) 100

MeIQx-DNA adduct

A

(/105dG) 10

8-OHdG

B

#

*

10

#

1 #

1 0.1

0.01

0.1

0.001

0

0.1 1 0.001 0.01 MeIQx (ppm, in diet)

10

100

0

0.001

0.01 0.1 1 MeIQx (ppm, in diet)

10

100

Figure 8.2. MeIQx–DNA adduct formation (A) and 8-OHdG formation levels (B) in the liver of F344 rats treated with MeIQx at various doses for 4 weeks. Asterisk (*) p < 0.01 versus 0.01 ppm group; number symble (#) indicates p < 0.01 versus 0 ppm group.

MeIQx–DNA adducts at week 4 was induced by administration of 0.01 ppm and higher doses of MeIQx, and induction was dose-dependent and statistically significant in the 100 ppm dose group (Figure 8.2); adduct formations in the 0 and 0.001 ppm MeIQx groups were below the limit of detection. Similar results were obtained after 16 weeks of MeIQx administration. DNA is subject to constant oxidative damage from endogenous oxidants. 8-Hydroxy-2′-deoxyguanosine (8-OHdG) is a marker for oxidative DNA damage, and 8-OHdG levels rise as a cell becomes more metabolically active. 8-OHdG levels at week 4 were unaffected by treatment with 0.001, 0.01, or 0.1 ppm MeIQx, but became statistically significantly elevated after treatment with MeIQx at doses of 1, 10, and 100 ppm (Figure 8.2). Similar results were obtained after 16 weeks of MeIQx administration. Finally, mutation of the H-ras gene, whose role in rat hepatocarcinogenesis is still unclear, was statistically significantly increased in the liver of rats treated with MeIQx for 2 weeks at 10 and 100, but did not differ at 0.001, 0.01. 0.1 and 1 ppm compared to the control value (unpublished data). We also examined mutation of the lacI gene and induction of GST-P positive foci in the livers of Big Blue® rats (Hoshi et al. 2004). Forty male Big Blue® rats were divided into 7 groups and administered MeIQx in the diet at doses of 0, 0.001, 0.01, 0.1, 1, 10, and 100 ppm for 16 weeks. A statistically significant elevation of lacI gene mutation level was detected in the 10 and 100 ppm groups (Figure 8.3). On the other hand, formation of GST-P positive foci was statistically significantly induced by administration of 100 ppm but not 10 ppm or less MeIQx (Figure 8.3). The results obtained from the experiments described above demonstrate the existence of a no-effect level (the highest dose of MeIQx at which there is no effect) for MeIQx mutagenicity and carcinogenicity. Since there is a no-effect level of MeIQx for gene mutagenicity, the initiation activity of MeIQx was examined in a two-stage carcinogenesis model using phenobarbital as a promoter of hepatocarcinogenesis (Fukushima et al. 2003). A total of 850 21-day-old male F344 rats were

212


(No./10 6)

A

Mutation frequencies

(No. /cm2)

B

GST-P positive foci

100

*

1000

*

10 100

* 1

10

0.1

1 0

0.001

0.01

0.1

1


10

100

0

0.001

0.01

0.1

1

10

100

MeIQx (ppm, diet)

Figure 8.3. Lac I gene mutation frequencies (A) and GST-P positive foci (B) in the liver of Big Blue® rats treated with MeIQx at various doses for 16 weeks. Asterisk (*) indicates p < 0.001 versus 0 ppm group.

divided into seven groups and administered MeIQx at doses of 0, 0.001, 0.01, 0.1, 1, 10, and 100 ppm for 4 weeks. This was followed by administration of 500 ppm phenobarbital in the diet. The numbers of GST-P positive foci were not elevated in the 0.001–1 ppm MeIQx groups, but statistically significant increases in GST-P positive foci formation were detected in the 10 and 100 ppm MeIQx groups. These results indicate the existence of a no-effect level for MeIQx initiation activity and are consistent with the existence of a no-effect level for MeIQx mutagenicity. Little is known about differences in the low dose–response relationship of genotoxic carcinogens among different strains of rat. Therefore, we examined MeIQx hepatocarcinogenicity using GST-P positive foci in both F344 and BN strains, with a total of 180 in each group. The background level of GST-P positive foci in the nontreated F344 rats was statistically significantly lower than that of BN rats, and the numbers of GST-P positive foci in the livers of MeIQx-treated F344 rats were statistically significantly lower in nearly all treatment groups compared with the corresponding BN strain groups (Wei et al. 2006). However, the results of MeIQx induction of GST-P positive foci in these two strains was the same: Lower doses of MeIQx, 0.1–10 ppm, had no statistically significant effect on the number of GST-P positive foci compared to the corresponding controls, while a statistically significant increase was detected at 100 ppm in both strains compared to the respective control groups (Table 8.2). Finally, we examined the carcinogenicity MeIQx in damaged livers (Kang et al. 2006). A total of 280 male F344 rats were divided into 14 groups. Liver damage was induced in 7 of these groups by administration of 0.03% thioacetamide (TAA), a well-known hepatotoxin, in their drinking water for 12 weeks. After cessation of TAA treatment, the rats received 0, 0.001, 0.01, 0.1, 1, 10, and 100 ppm MeIQx in the diet for 16 weeks. In both TAA-treated and untreated groups, the lower doses of MeIQx had no effect on the number of GST-P positive foci, but a statistically


213

TABLE 8.2. Development of GST-P Positive Foci in the Livers of BN and F344 Rats Treated with Various Doses of MeIQx

Group

MeIQx (ppm)

BN Rat 1 0 2 0.1 3 1 4 5 5 10 6 100 F344 Rat 7 0 8 0.1 9 1 10 5 11 10 12 100

Size Distribution of GST-P Positive Foci (No./cm2)

Number of Rats

2–4 Cells

5–10 Cells

≥11 Cells

Total

30 30 30 30 30 30

0.16 ± 0.21 0.14 ± 0.23 0.12 ± 0.21 0.23 ± 0.33 1.17 ± 0.98 13.26 ± 7.07a

0.06 ± 0.12 0.03 ± 0.08 0.04 ± 0.09 0.11 ± 0.19 0.42 ± 0.57 7.37 ± 4.78a

0.02 ± 0.10 0.03 ± 0.19 0.04 ± 0.14 0.02 ± 0.09 0.06 ± 0.14 4.25 ± 3.88a

0.24 ± 0.29 0.19 ± 0.30 0.20 ± 0.33 0.36 ± 0.49 1.64 ± 1.43 24.88 ± 14.67a

30 30 30 30 30 30

0.01 ± 0.05c 0.03 ± 0.08c 0.07 ± 0.15 0.08 ± 0.16c 0.29 ± 0.49c 3.60 ± 2.22b,c

0c 0 0c 0.01 ± 0.05c 0.04 ± 0.12c 1.83 ± 1.33b,c

0 0.01 ± 0.05 0 0 0c 0.99 ± 1.01b,c

0.01 ± 0.05c 0.04 ± 0.10c 0.07 ± 0.15 0.08 ± 0.18c 0.33 ± 0.59c 6.41 ± 4.04b,c

p < 0.01 (vs. group1).

a

p < 0.01 (vs. group7).

b

p < 0.01 (vs. corresponding BN rat group).

c

significant increase was observed in the 100 ppm MeIQx groups (Figure 8.4). Using the method of maximum likelihood to model this data, the numbers of GST-P positive foci, with and without TAA treatment, fitted the hockey stick regression model; that is, no statistically significant differences in foci number were observed in the 0–10 ppm MeIQx groups, whereas a statistically significant increase in foci number was observed in the 100 ppm MeIQx group. In contrast, a linear dosedependent increase of MeIQx–DNA adduct formation was evident from 0.1 to 100 ppm; adduct formation in the 0.001 and 0.01 ppm MeIQx groups were below the limit of detection (Figure 8.4). The formation of MeIQx-DNA adducts was virtually identical in undamaged and damaged livers. These results are consistent with the previous results and support the existence of a no-effect level for MeIQx hepatocarcinogenicity, even on a background of liver damage. A summary of the results obtained in our experiments is presented in Figure 8.5. The formation of DNA–MeIQx adducts was observed at very low doses of MeIQx. Due to limitations in detection of these adducts, we were unable to determine whether a threshold dose of MeIQx was required for MeIQx–DNA adduct formation. Increasing the dose of MeIQx next resulted in an elevation of 8-OHdG formation, then gene mutation and the appearance of initiation activity, and, finally, at the highest dose used, an increase in the endpoint marker for carcinogenicity (GST-P positive foci). Notably, these data demonstrate that increased doses

214


A

B GST-P positive foci

(No. /cm2) 100

MeIQx-DNA adduct

(x10-7nds) 100

TAA → MeIQx MeIQx

*

# ##

10

**

10

# ##

1 #

1

##

0.1 TAA → MeIQx MeIQx

0.1 0

0.001

0.01

0.1

1

10

100

0.01

0

0.001

0.01


0.1

1

10

100


Figure 8.4. GST-P positive foci (A) and formations of MeIQx-DNA adduct (B) in the liver of F334 rats treated with MeIQx with or without thioacetamide. Asterisk (*) indicates p < 0.01 versus TAA intiation alone group; double asterisk (**) indicates p < 0.01 versus nontreatment group; Number symbol (#) indicates p < 0.01 versus 0.1 ppm MeIQx without TAA intiation; double number symbol (##) indicates p < 0.01 versus 0.1 ppm MeIQx with TAA intiation.

Response Liver cancer 8-OHdG H-ras mutation lacI mutation Initiation activity

GST-P positive foci

MeIQx-DNA adduct

Control level MeIQx doses Figure 8.5. Risk of liver cancer: Reaction curves for carcinogenesis markers are dependent on the dose of MeIQx.

of MeIQx were required as MeIQx-mediated effects moved from simple adduct formation to cellular metabolic changes (possibly due in part to increased DNA repair) to gene mutation and cancer initiation to carcinogenesis. These results argue strongly for the existence of a threshold, at least a practical threshold, for MeIQx hepatocarcinogenicity in the rat. In support of this conclusion, our 2-year carcinogenicity test of MeIQx in rats showed no hepatocarcinogenicity at low doses (Murai et al. 2008).

8.5. LOW-DOSE CARCINOGENICITY OF PHIP IN THE RAT COLON

215

8.4. LOW-DOSE HEPATOCARCINOGENICITY OF N-NITROSO COMPOUNDS N-nitroso compounds such as diethylnitroamine (DEN) and dimethylnitrosamine (DMN) are synthesized in the stomach through the reaction of secondary amines and nitrites in the diet. They are also found as contaminants of a variety of manufactured food products. Peto et al. (1991) investigated the carcinogenicity of DEN using 2040 male and 2040 female Colworth rats. DEN at doses of 0.033–16.896 ppm was administered to the rats in their drinking water, induction of liver tumors was found to be dependent on the applied dose of DEN, and at the lower doses a linear dose–tumor incidence relationship was observed (Peto et al. 1991). Therefore, it was concluded that DEN had no threshold for its carcinogenicity in the rat liver. We have reexamined the carcinogenic influence of low doses of DEN (Fukushima et al. 2002). Approximately 2000 21-day-old male F344 rats were administered DEN at doses ranging from 0.0001 to 10 ppm in their drinking water for 16 weeks. No increase in the number of GST-P positive foci was found at DEN doses of 0.0001–0.01 ppm; however, the number of GST-P positive foci was statistically significantly elevated at 0.1 and 1 ppm DEN. In the 10 ppm group, the numbers of GST-P positive foci were so numerous that quanitation was not possible. Therefore, we conclude that there is a no-effect level for DEN hepatocarcinogenicity in the rat. Low-dose carcinogenicity experiments were also performed with DMN (Fukushima et al. 2005). The carcinogen was applied to 540 21-day-old F344 rats at doses ranging from 0.001 to 10 ppm in their drinking water for 16 weeks. No induction of GST-P positive foci was found at doses of 0.001 to 0.1 ppm; however, statistically significant increases in the number of GST-P positive foci were observed at 1 and 10 ppm. Therefore, similarly to DEN, we concluded that there is a no-effect level for DMN hepatocarcinogenicity in the rat.

8.5. LOW-DOSE CARCINOGENICITY OF 2-AMINO-1METHYL-6-PHENYLIMIDAZO[5,6-B]PYRIDINE (PHIP) IN THE RAT COLON The heterocyclic amine PhIP is a carcinogen contained in seared meat and fish, and it exerts its carcinogenicity in the rat colon. We investigated the carcinogenicity of PhIP in the rat colon when applied at doses of 0.001– 400 ppm (Fukushima et al. 2004). A total of 1759 6-week-old F344 male rats were administered PhIP in their diet for 16 weeks. The development of aberrant cell foci (ACF), the surrogate marker of preneoplastic lesions in the colon, was not altered by PhIP administration at 0.001–10 ppm; however, at doses of 50 – 400 ppm, statistically significant increases in ACF were observed (Figure 8.6). Like MeIQx, DEN, and DMN, PhIP is a genotoxic compound and is metabolized in cells to an ultimate carcinogen capable of covalently binding DNA.

216


A

B

ACF

Total ACF/rat 10

* #

PhIP-DNA adduct

(/107nds) 100

*

* 10

*

1

* 1

0.1

*

0.1

*

*

*

0.01

0.01 0

0.001

0.01

0.1

1

PhIP (ppm, in diet)

10

50 100 400

0

0.001

0.01

0.1

1

10

50 100 400

PhIP (ppm, in diet)

Figure 8.6. Aberrant crypt foci (A) and formations PhIP–DNA adduct (B) in the colons of F344 rats treated with PhIP at various doses for 16 weeks. Number symbol (#) indicates p < 0.05 versus 0 ppm group. Asterisk (*) indicates p < 0.01 versus 0 ppm group. Note that PhIP–DNA adduct levels were also statistically significantly increased in the same manner at week 4.

Statistically significant increases in the formation of PhIP–DNA adduct levels were found in the groups treated with 0.01 ppm and higher doses of PhIP at 16 weeks (Figure 8.6). Thus, similarly to MeIQx, DNA adduct formation is observed after administration of low doses of PhIP while doses required to induce ACF are much higher (approximately 50,000 times higher) than that needed for PhIP–DNA adduct formation. These results argue for a no-effect level and a threshold dose for PhIP colon carcinogenicity in the rat. Finally, we assessed the effect of low doses of PhIP in the progression of colon tumors (Doi et al. 2005). A total of 192 6-week-old male F344 rats were subcutaneously injected twice with the colon carcinogen azoxymethane (AOM) with a 1-week interval, and then the animals were continuously fed PhIP at doses ranging from 0.001 to 200 ppm for 16 weeks. Lower doses (0.001–10 ppm) of PhIP had no significant effect on AOM-initiated colon carcinogenesis; higher doses (50–200 ppm) of PhIP caused a statistically significantly enhancement of AOM-initiated colon carcinogenesis (Table 8.3). Results obtained from this initiation–promotion model show a no-effect level of 10 ppm for PhIP promotion of colon carcinogenesis and again argue for a threshold dose for PhIP colon carcinogenicity in the rat.

8.6. LOW-DOSE CARCINOGENICITY OF POTASSIUM BROMATE, KBRO3 IN THE RAT KIDNEY Potassium bromate is a rodent renal carcinogen which can be found as a contaminant of tap water and which is used as a dough conditioner and food additive in some countries. It is a genotoxic carcinogen that is reduced in renal proximal tubular cells to yield bromine oxides and radicals, which are the ultimate carcinogens that specifically cause guanine oxidation, leading to renal mutagenesis and carcinogenesis.

TABLE 8.3.

Induction of Tumors in the Colon of F344 Rats Treated with Azoxymethane Followed by PhIP

Incidences (%) Histologic Findings

0 ppm (n = 16)

0.001 ppm (n = 16)

0.01 ppm (n = 16)

0.1 ppm (n = 16)

1 ppm (n = 16)

10 ppm (n = 16)

50 ppm (n = 16)

200 ppm (n = 14)

Adenoma Adenocarcinoma Totala

2 (12.5) 8 (50) 9 (56.3)

3 (18.8) 7 (43.8) 10 (62.5)

1 (6.3) 10 (62.5) 10 (62.5)

5 (31.3) 5 (31.3) 8 (50)

3 (18.8) 9 (56.3) 11 (68.8)

2 (12.5) 8 (50) 8 (50)

14 (87.5)b 14 (87.5) 16 (100)c

14 (100)d 14 (100)c 14 (100)c

a

Total of adenoma and adenocarcinoma. p < 0.005 (vs. 0 ppm).

b c

p < 0.05 (vs. 0 ppm). p < 0.0001 (vs. 0 ppm).

d

217

218


Thus, the genotoxic mechanism of potassium bromate is different from that of MeIQx, DEN, DMN, and PhIP. The studies described to this point indicate that the genotoxic compounds MeIQx, DEN, DMN, and PhIP have no-effect levels for induction of various carcinogenesis markers and strongly suggest that a threshold dose exists for induction of carcinogenesis by these compounds. In the following experiments, we investigated the relationship between potassium bromate dose and induction of gene mutation, one of the markers of carcinogenesis. A total of 40 male Big Blue® rats were divided into 8 groups and administered potassium bromate in their drinking water at doses of 0, 0.02, 0.2, 2, 8, 30, 125, and 500 ppm for 16 weeks (Yamaguchi et al. 2008). No significant induction of lacl gene mutation was observed in the 0.02–125 ppm groups, but a statistically significant increase in lacl gene mutation was observed in the 500 ppm group (Figure 8.7). Similarly, statistically significantly elevated 8-OHdG levels and GC to TA transversions, a mutation known to occur as a result of 8-OHdG adduct formation, also occurred only at a potassium bromate dose of 500 ppm (Figure 8.7). No preneoplastic or neoplastic lesions were detected in the kidney in these experiments. Therefore, we concluded that there is a no-effect level for potassium bromate-induced 8-OHdG formation and mutagenicity in the rat kidney. Finally, the renal carcinogenicity of potassium bromate was examined using a two-stage carcinogenesis model. A total of 240 male Wistar rats were treated with N-ethyl-N-hydroxyethylnitrosamine for the initiation of kidney carcinogenesis and were thereafter administered potassium bromate at doses of 0, 0.02, 0.2, 2, 8, 30, 125, and 500 ppm in their drinking water for 16 weeks (Wei et al. 2009): Due to

A

Qxidative DNA damage (8-OHdG)

(/105dG)

*

1.6 1.2 0.8 0.4 0

0

0.02

0.2

2

8

30

125

500

KBrO3 (ppm, in drinking water)

B

C

Total lacI mutation frequency

lacI mutation frequency (GC to TA)

(/106 plaques)

(/106 plaques) 80

*

*

30

60 20 40 10

20 0

0

0

0.02

0.2

2

8

30


125

500

0

0.02

0.2

2

8

30

125

500


Figure 8.7. 8-OHdG formation levels (A) and LacI gene mutation frequencies (B, C) in the kidney of Big Blue rats treated with KBrO3 for 16 weeks. (*) p < 0.05 versus 0 ppm group.

ACKNOWLEDGMENTS

219

toxicity, the highest dose, 500 ppm, was reduced to 250 ppm from week 12. Enhancement of a preneoplastic lesion, an atypical tubular hyperplasia, and enhancement of tumorigenesis in the kidney was observed only in the highest dosed group. The results of these two sets of experiments support the conclusion that there is a no-effect level and threshold dose for potassium bromate renal carcinogenicity in the rat.

8.7.

CONCLUSION

For the genotoxic carcinogens examined, the no-effect doses for initiation markers (i.e., DNA adduct formation, 8-OHdG formation, and gene mutation) were much lower than the no-effect doses for promotion marker (i.e., GST-P positive foci and ACF); and, generally, induction of promotion markers occurred at doses of carcinogen which did not induce carcinogenesis. These results strongly suggest that processes such as DNA repair, irreversible senescence, apoptosis, and immune system function operate to inhibit the effects of genotoxic carcinogens and that the inhibition is significant. Therefore, we conclude that there are thresholds, at least practical thresholds, for the carcinogens examined in this study. The genotoxic carcinogens examined in this study can be classified into two types from the viewpoint of mechanism (Hengstler et al. 2003). In one type, the carcinogen is metabolized by the cell to an ultimate carcinogen, which binds covalently to the DNA to form DNA adducts. In the second type, the compound is metabolized by the cell to an ultimate carcinogen, which causes oxidative damage to the DNA. The first type of genotoxic carcinogen encompasses heterocyclic amines (e.g., MeIQx and PhIP) and N-nitrosocompounds (e.g., DEN and DMN). The second type of genotoxic carcinogen is represented by potassium bromate. Notably, the first type of genotoxic carcinogen induces formation of DNA adducts at low doses but higher doses are required for gene mutation, while the second type of genotoxic carcinogen causes DNA damage and gene mutation at equivalent doses. This undoubtedly reflects the different mechanisms by which these two types of genotoxic compounds cause DNA damage. Nevertheless, both types of genotoxic compounds clearly have no-effect doses for initiation, which are lower than the no-effect doses for promotion and carcinogenicity. It is probable, therefore, that other (perhaps most or even all) genotoxic carcinogens also have this pattern of no-effect dose and, consequently, do have thresholds for carcinogenicity.

ACKNOWLEDGMENTS The authors would like to acknowledge the help of Masao Hirose (Division of Pathology, National Institute of Health Sciences), Yoichi Konishi (Department of Oncological Pathology, Cancer Center, Nara Medical University), Dai Nakae (Tokyo Metropolitan Institute of Public Health), Shuzo Otani (Department of Biochemistry, Osaka City University Graduate School of Medicine), Tomoyuki Shirai (Department Pathology, Nagoya City University Graduate School of Medicine), Michihito

220


Takahashi (Div. Pathology, National Institute of Health Sciences), Masae Tatematsu (Division of Oncological Pathology, Aichi Cancer Center Research Institute), Hiroyuki Tsuda (Department of Molecular Toxicology, Nagoya City University Graduate School of Med.), and Keiji Wakabayashi (Cancer Prevention Research Division, National Cancer Center Research Institute). The authors would also like to acknowledge the encouragement of Dr. Nobuyuki Ito (Professor Emeritus, Nagoya City University Medical School, Nagoya, Japan) and Dr. Tomoyuki Kitagawa (Institute Director Emeritus, Japanese Foundation for Cancer Research, Tokyo, Japan). These studies were supported by a grant from the Japan Science and Technology Corporation, included in the Project of Core Research for Evolutional Science and Technology (CREST), and by a grant from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

REFERENCES Doi, K., Wanibuchi, H., Salim, E. I., Morimura, K., Kinoshita, A., Kudoh, S., Hirata, K., Yoshikawa, J., and Fukushima, S. (2005). Lack of large intestinal carcinogenicity of 2-amino-1-methyl-6phenylimidazo[4,5-b]pyridine at low doses in rats initiated with azoxymethane. Int J Cancer 115, 870–878. Fukushima, S., Wanibuchi, H., Morimura, K., Wei, M., Nakae, D., Konishi, Y., Tsuda, H., Uehara, N., Imaida, K., Shirai, T., Tatematsu, M., Tsukamoto, T., Hirose, M., Furukawa, F., Wakabayashi, K., and Totsuka, Y. (2002). Lack of a dose–response relationship for carcinogenicity in the rat liver with low doses of 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline or N-nitrosodiethylamine. Jpn J Cancer Res 93, 1076–1082. Fukushima, S., Wanibuchi, H., Morimura, K., Wei, M., Nakae, D., Konishi, Y., Tsuda, H., Takasuka, N., Imaida, K., Shirai, T., Tatematsu, M., Tsukamoto, T., Hirose, M., and Furukawa, F. (2003). Lack of initiation activity in rat liver of low doses of 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline. Cancer Lett 191, 35–40. Fukushima, S., Wanibuchi, H., Morimura, K., Iwai, S., Nakae, D., Kishida, H., Tsuda, H., Uehara, N., Imaida, K., Shirai, T., Tatematsu, M., Tsukamoto, T., Hirose, M., and Furukawa, F. (2004). Existence of a threshold for induction of aberrant crypt foci in the rat colon with low doses of 2-amino-1-methyl6-phenolimidazo[4,5-b]pyridine. Toxicol Sci 80, 109–114. Fukushima, S., Wanibuchi, H., Morimura, K., Nakae, D., Tsuda, H., Imaida, K., Shirai, T., Tatematsu, M., Tsukamoto, T., Hirose, M., and Furukawa, F. (2005). Lack of potential of low dose Nnitrosodimethylamine to induce preneoplastic lesions, glutathione S-transferase placental form-positive foci, in rat liver. Cancer Lett 222, 11–15. Hengstler, J. G., Bogdanffy, M. S., Bolt, H. M., and Oesch, F. (2003). Challenging dogma: thresholds for genotoxic carcinogens? The case of vinyl acetate. Annu Rev Pharmacol Toxicol 43, 485–520. Hoshi, M., Morimura, K., Wanibuchi, H., Wei, M., Okochi, E., Ushijima, T., Takaoka, K., and Fukushima, S. (2004). No-observed effect levels for carcinogenicity and for in vivo mutagenicity of a genotoxic carcinogen. Toxicol Sci 81, 273–279. Kang, J. S., Wanibuchi, H., Morimura, K., Totsuka, Y., Yoshimura, I., and Fukushima, S. (2006). Existence of a no effect level for MeIQx hepatocarcinogenicity on a background of thioacetamideinduced liver damage in rats. Cancer Sci 97, 453–458. Kato, T., Ohgaki, H., Hasegawa, H., Sato, S., Takayama, S., and Sugimura, T. (1988). Carcinogenicity in rats of a mutagenic compound, 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline. Carcinogenesis 9, 71–73. Kuraoka, I. (2008). Effects of DNA lesions on transcription elongation by RNA polymerases. Genes Environ 30, 63–70. Murai, T., Mori, S., Kang, J. S., Morimura, K., Wanibuchi, H., Totsuka, Y., and Fukushima, S. (2008). Evidence of a threshold-effect for 2-amino-3,8-dimethylimidazo-[4,5-f]quinoxaline liver carcinogenicity in F344/DuCrj rats. Toxicol Pathol 36, 472–477.

REFERENCES

221

OECD (1981). Carcinogenicity Studies. OECD Guideline for Testing of Chemicals 451, 1–17. Peto, R., Gray, R., Brantom, P., and Grasso, P. (1991). Effects on 4080 rats of chronic ingestion of Nnitrosodiethylamine or N-nitrosodimethylamine: A detailed dose–response study. Cancer Res 51, 6415–6451. Tsuda, H., Fukushima, S., Wanibuchi, H., Morimura, K., Nakae, D., Imaida, K., Tatematsu, M., Hirose, M., Wakabayashi, K., and Moore, M. A. (2003). Value of GST-P positive preneoplastic hepatic foci in dose–response studies of hepatocarcinogenesis: Evidence for practical thresholds with both genotoxic and nongenotoxic carcinogens. A review of recent work. Toxicol Pathol 31, 80–86. Wei, M., Hori, T. A., Ichihara, T., Wanibuchi, H., Morimura, K., Kang, J. S., Puatanachokchai, R., and Fukushima, S. (2006). Existence of no-observed effect levels for 2-amino-3,8-dimethylimidazo[4,5-f ] quinoxaline on hepatic preneoplastic lesion development in BN rats. Cancer Lett 231, 304–308. Wei, M., Hamoud, A.S., Yamaguchi, T., Kakehashi, A., Morimura, K., Doi, K., Kushida, M., Kitano, M., Wanibuchi, H., and Fukushima, S. (2009). Potassium bromate enhances N-ethyl-Nhydroxyethylnitrosamine-induced kidney carcinogenesis only at high doses in Wistar rats: indication of the existence of an enhancement threshold. Toxicol Pathol 37, 983–991. Yamaguchi, T., Wei, M., Hagihara, N., Omori, M., Wanibuchi, H., and Fukushima, S. (2008). Lack of mutagenic and toxic effects of low dose potassium bromate on kidneys in the Big Blue rat. Mutat Res 652, 1–11.

PART

III

GENETIC TOXICOLOGY, TESTING GUIDELINES AND REGULATIONS, AND NOVEL ASSAYS

CH A P TE R

9

DEVELOPMENT OF GENETIC TOXICOLOGY TESTING AND ITS INCORPORATION INTO REGULATORY HEALTH EFFECTS TEST REQUIREMENTS Errol Zeiger

9.1.

INTRODUCTION

Genetic toxicology testing—the testing for the ability of substances to produce mutations or chromosome aberrations, or otherwise damage DNA—has been central to the safety evaluation of chemicals since the mid- to late 1970s. Concern for induction of genetic damage began with concern for heritable gene and chromosomal germ cell mutations in the offspring of exposed individuals. This concern was reflected in the early guidance documents (see, e.g., Crow 1968; DHEW 1969; Drake et al. 1975; EPA 1975, 1979, 1980; Flamm et al. 1977; NRC 1983) that were produced by various agencies and scientific societies. However, with the accumulating evidence that mutagenesis was an early step in the development of a tumor and that carcinogenic chemicals were mutagenic (Ames 1971; Ames et al. 1973; McCann et al. 1975; Sugimura et al. 1976; Purchase et al. 1978), the genetic toxicology testing emphasis switched from heritable mutations to carcinogenesis. This concern for heritable effects, although valid, but will not be addressed here. The continuing concern for heritable mutations in addition to carcinogenesis is reflected in the US Environmental Protection Agency’s Office of Prevention, Pesticides and Toxic Substances (EPA OPPTS) testing schemes (Auletta et al. 1993) and in the international Organization for Economic Co-operation and Development (OECD) international harmonized test guidelines (OECD 2001), which classifies mutagens based on their potential for causing heritable mutations in humans. One reason behind this shift in concern is because there are a number of demonstrated human carcinogens but no demonstrated human germ cell mutagens. Another reason is that the chemicals


225

226

CHAPTER 9 DEVELOPMENT OF GENETIC TOXICOLOGY TESTING

that are known germ cell mutagens tend to also be in vivo somatic cell mutagens and carcinogens. Therefore, although the public health concerns for germ cell mutagenicity is high and taken into consideration by the regulatory authorities, it is easier and more practical to regulate mutagens based on their potential carcinogenicity. As a consequence of testing compilations first published in the 1970s which showed high cancer predictivities for the in vitro genetic toxicity tests, the United States and other regulatory authorities began requiring premarket genetic toxicity testing for chemicals and drugs. The in vitro genetic toxicity assays used internationally for regulatory approval of chemicals are the bacterial (Salmonella; E. coli), mammalian cell mutagenicity (L5179Y mouse lymphoma cells; CHO cells), and/or mammalian cell chromosome damage (L5178Y, CHO, CHL cells, or human lymphocytes) assays (see Chapter 11). In vivo testing uses primarily the rodent bone marrow cell chromosome aberration or micronucleus assay (Chapter 12). Substances that are positive in the in vitro tests are considered to be of the most concern for inducing cancer or genetic mutations in rodents and, by extension, in humans. As a consequence of these early studies, formal guidelines for conducting the tests were developed and recommended by U.S. and international organizations (EPA 2008; FDA 2000; OECD 2008; ICH 2008). These in vitro positives are then tested in rodents to determine if they have the capability of inducing genetic damage in the animal. In vivo genetic toxicity testing currently is also a prerequisite for identifying germ cell mutagens—that is, those that have the potential to mutate sperm or egg cells resulting in offspring either expressing or only carrying a mutant gene (see Chapter 10). The apical endpoint, cancer or germ cell mutagenicity, is currently demonstrated and quantified by extensive animal experiments because the genetic toxicity assays and the structure–activity relationship models are not sufficiently accurate predictors of these effects, or of their dose–responses, to support human health and safety decisions. In many cases, the high cost and lengthy nature of the follow-up in vivo tests, coupled with the high probability that the chemical would be tumorigenic or produce germ cell mutations, will lead companies to drop the chemical from further consideration without performing the confirmatory in vivo test. This chapter is designed to present a brief overview of the development of genetic toxicity testing for regulatory purposes; specifically the identification and characterization of carcinogens. It also presents supporting rationales for the types of tests mandated and the use of the data, and it identifies scientific and practical issues that will need to be resolved in the near future. More detailed descriptions of the tests, testing strategies, and decision processes are addressed elsewhere in this volume.

9.2.

DEFINITIONS AND USAGE

The terms mutagenic and genotoxic are often used interchangeably, although they are not the same, so that it is important to clarify the distinction between genotoxicity and mutagenicity. Mutagenicity includes gene mutations (either point mutations or deletions), chromosome breaks and rearrangements, and aneuploidies. A mutagenic

9.3. THE HISTORICAL DEVELOPMENT OF GENETIC TOXICITY TESTING

227

event is, by definition, heritable and will be passed to daughter (F1) cells (somatic cell tests) or to the offspring (germ cell tests). This means that genetic damage that is not compatible with cell survival or reproduction will not lead to a mutant organism. In contrast, genotoxicity is a broad term that includes mutagenicity, but also includes interaction with or damage to DNA, adduct formation, interference with the DNA replication or repair processes, and other nonspecific DNA-related effects. Genotoxic events do not always lead to mutagenicity and, if they are not toxic, may have no noticeable or lasting effects on the cell. Point mutations—that is, changes in single DNA bases or intragenic deletions and rearrangements—are considered to be heritable effects because they are typically measured in the post-treatment generation cells. Chromosome breakage (clastogenicity) is typically measured in the treated cells. These effects can be heritable and are the cause of many genetic diseases, although most clastogenicity seen in genetic toxicology tests is not compatible with cell survival and would therefore not result in a heritable effect. However, for testing purposes, the presence of chromosome breaks or rearrangements is evidence that the substance will cause heritable effects even though the test, itself, does not measure whether the effects seen will allow the cell to divide. An exception to this is the measurement of micronuclei (MN), which can be the effect of chromosome breaks or aneuploidy and are measured in the post-treatment-generation cells. Mutagens can also be classified as direct or indirect. This approach to classification has become an area of great interest in the regulatory agencies and industry. Direct mutagens are those that directly interact with and damage DNA, either as the parent compound or as a metabolite. Indirect mutagens act via two different means. They act either through the generation of intermediate molecules, such as active oxygen species, that subsequently react with DNA, or by interfering with the cell’s replicative proteins (either DNA synthesis or repair) or with the mitotic spindle. The determination of whether or not a substance is mutagenic as a result of its direct interaction with DNA, or as a secondary effect of other cellular interactions or reactions, is central to the regulation of chemicals causing mutation and cancer—that is, whether the cancer dose extrapolation should be linear, nonlinear, or threshold—and is addressed further elsewhere in this volume. In practice, in the absence of positive human epidemiological studies, the cancer response in the rat or mouse by a specific chemical is considered to be definitive for the identifying the chemical as a presumptive human carcinogen and for determining the relative carcinogenic potency in the exposed individuals. The exceptions to this practice are situations where it can be shown that the rodent carcinogenicity occurs by a mechanism(s) that is not operative in humans.

9.3. THE HISTORICAL DEVELOPMENT OF GENETIC TOXICITY TESTING A history of the development of genetic toxicity testing can be found in Zeiger (2004). Briefly, in the early 1950s, chemicals were tested for mutagenicity in E. coli using suspension and plate tests (Demerec et al. 1951; Hemmerly and Demerec

228


1955). Subsequently, Szybalski (1958) reported on the screening of more than 400 chemicals for mutagenicity by applying the chemicals on a filter paper disk in petri dishes containing an E. coli strain spread on the agar. The E. coli strain used at that time responded only to a single base change and was relatively insensitive. This was also prior to the use of liver enzyme systems to provide mammalian metabolism. Ames (1971) adopted the spot-test method for screening mutagens using Salmonella typhimurium histidine mutant strains, which was followed by the development of in vitro metabolic activation systems and the currently used plate test. Protocols for performing the test were subsequently published (Ames et al. 1973; Maron and Ames 1983; Mortelmans and Zeiger 2000). The initial publication by McCann et al. (1975) that approximately 90% of all rodent carcinogens and noncarcinogens could be predicted by the Salmonella test was followed by similar studies and values in other laboratories (e.g., Purchase et al. 1978; Sugimura et al. 1976). Some of the differences in responses among the different compilations can be attributed to different classes of chemicals tested. The early (i.e., pre-1985) results led to the incorporation of this, as well as mammalian cell mutagenicity tests, into regulatory and industrial decision making, although similar data were not always available for the mammalian cell tests. Later studies of the effectiveness of the Salmonella and the mammalian cell tests produced lower predictivity values (e.g., Dunkel et al. 1985; Tennant et al. 1987; Zeiger et al. 1990; Zeiger 1998; Kirkland et al. 2005), but not low enough to remove them from their status as cancer-predictive tests. Unfortunately, a number of the compilations used datasets with very high frequencies of carcinogens, so that tests tending to be positive appeared to be highly effective, although high proportions of noncarcinogens were also detected as positive; that is, the specificities were low. The in vitro mammalian cell tests currently in use were also developed and/ or refined in the late 1960s to early 1970s (Zeiger 2004), and they comprised tests for gene mutation, chromosome aberration, and sister chromatid exchanges (SCE). The SCE tests were initially viewed as an alternative to the chromosome aberration assays because they were easier and less expensive to perform. However, they subsequently dropped out of favor based primarily on their performance in National Toxicology Program (NTP) validation studies (Tennant et al. 1987; Zeiger et al. 1990) and because of questions concerning their relevance to heritable genetic effects and cancer initiation. More recently, the in vitro MN test in mammalian cells has been proposed as an alternative measure of chromosome damage to the aberration test because, like the SCE test, it is easier and less expensive to perform than the chromosome aberration test, but it can also be used to distinguish between MN caused by chromosome aberrations and nondisjunction (Kirsch-Volders et al. 2000; Parry et al. 1996).

9.4.

TYPES OF AVAILABLE TESTS

By the late 1970s a large number of diverse tests had been developed or adapted for carcinogen screening in the hope that they would be useful as a replacement or adjunct to the Salmonella test (now called the Ames test) which had become the

9.5. TESTING APPROACHES

229

benchmark. A compilation by Hollstein et al. (1979) identified 119, while a later compilation (IARC 1987) identified 173, test systems or endpoints and included plant, insect, microbial, and mammalian in vitro and in vivo tests. Some of these tests were fairly widely used, with reports of the testing of many chemicals, while most had seen limited use (often only in the laboratory of the test developer) and had a database of relatively few chemicals which tended to be potent mutagens and clastogens. A number of other mammalian or microbial cell lines, and in vivo systems, have been proposed since then, and the endpoints measured have extended to new molecular effects. It has been estimated (Zeiger, unpublished) that there are 200–300 such test systems available at the present time, or reported in the literature. The majority of these test systems have not been systematically examined for their ability to discriminate between carcinogens and noncarcinogens. The currently used test systems for genetic toxicity for health effects testing for regulatory submissions, along with their EPA and OECD Test Guidelines, where they exist, are listed in Table 9.1. The selection of these tests does not necessarily signify that they are the best, or the only ones available for the particular endpoints, but was made based on information available in the 1970s–1980s and familiarity with their use and was encouraged by the reputations or persistence of the individual scientists or agencies advocating the tests. With time, with the exception of the Salmonella and E. coli tests, the plant tests and nonmammalian tests (e.g., yeast, Drosophila) were considered to be less relevant for human health prediction than were the mammalian tests and are no longer performed for regulatory submissions.

9.5.

TESTING APPROACHES

In order to make sense of the large number of available genetic toxicity tests and to simplify their use for supporting regulatory decisions, Bridges (1973) proposed a tier testing scheme for identifying potential carcinogens and germ cell mutagens, which was further elaborated on by Flamm (1974) and Bridges (1976). This tier approach forms the basis for the majority of the current regulatory testing schemes. In its early form, the initial tier would comprise in vitro tests for gene mutation and chromosome damage that are highly sensitive so as not to miss any potential in vivo mutagens. One requirement for the initial testing tier was that it not produce too many false negatives, with the consideration that the higher, in vivo, tiers would be capable of distinguishing between the “true” and the “false” positives. The second tier would consist of in vivo mammalian tests for the same endpoints to confirm the in vitro positive findings and/or to ensure that high exposure substances that were negative in vitro would also be negative in vivo. These first two tiers would be used to provide qualitative data on potential mutagenicity or clastogenicity in somatic and/or germ cells. The third, and final, tier would comprise apical in vivo rodent germ cell tests that could be used for quantitative genetic risk assessment of chemicals that were positive in tier 1 and/or 2 but considered sufficiently valuable for further development or study despite their potential mutagenicity and carcinogenicity. It is worth noting that these testing schemes were proposed at the time that the predictivity of the short term in vitro tests for carcinogenicity was believed to be

230


TABLE 9.1. Tests Currently Used for Genetic Toxicity Screening and for Regulatory Approval of Commercial Chemicalsa

Test Guideline No. Test

Example(s)

Effect Measured

OECD

U.S. EPA

Gene mutations

471

870.5100

DNA damage repair Gene mutations

476

870.5500 870.5300

Chromosome damage; nondisjunction

473 487b 487b

870.5375 — —

DNA damage

482 —c —

870.5550 — —

475 474 — —

870.5385 870.5395 — —

486d

—

Chromosome damage

483

870.5380

Chromosome damage incompatible with embryo survival Heritable (to F1) chromosome rearrangements Gene or chromosome damage in F1

478

870.5450

485

870.5460

In Vitro—Bacterial and Mammalian Cells Bacterial mutagenicity Bacterial DNA damage Mammalian cell mutation Mammalian cell cytogenetics

Mammalian cell DNA damage

Ames (Salmonella) test; E. coli test SOS test Mouse lymphoma test; CHO-hprt test CHO, CHL, or human lymphocyte chromosome aberration or MN test UDS; comet assay; adduct formation

In Vivo—Rodent Somatic Cells Bone marrow cytogenetics Transgenic rodent gene mutation DNA damage

Aberrations; micronuclei; aneuploidy BigBlue mouse; MutaMouse Liver UDS; comet assay; DNA adducts

Chromosome damage; nondisjunction Gene mutations in various tissues DNA damage leading to strand breaks

In Vivo—Rodent Germ Cells Male germ cell cytogenetics Sperm cell chromosome damage Heritable sperm cell chromosome damage Heritable gene mutations

Spermatogonial, spermatocyte cytogenetics Dominant lethal assay

Heritable translocation test Mouse-specific locus test

—

870.5195; 870.5200

a

This listing is not exhaustive, but includes the test systems currently addressed by formal test guidelines, or which may be recommended for health effects screening or subsequent testing. The use of these tests for regulatory submissions is addressed in more detail elsewhere in this volume. b

OECD Guideline (No. 487) for in vitro MN and aneuploidy tests is being developed, but is not approved at the time of this writing. It is not anticipated to be formalized before 2010.

c

—, No Test Guideline available.

d

Test Guideline only for liver UDS test.

9.5. TESTING APPROACHES

231

approximately 90%. Although the performance of germ cell mutagenicity tests was not as well quantified, this endpoint was considered to be of equal importance to carcinogenicity. Such a tier system is based on a number of premises about the relationships among the different tests and endpoints and cancer. The basic premises derived from the above publications, along with their explicit or implicit rationales, can be summarized as follows: Premise #1. The Salmonella mutation test is a necessary component of genetic toxicity testing schemes. Gene mutations are a necessary, if not sufficient, inducer of the tumorigenic process. The test is mechanistically simple and the easiest to perform, and it has been validated more extensively than the other tests. It is also less susceptible than the in vitro mammalian cell tests to artifactual positive results. Despite the fact that the bacterial chromosome is structurally and functionally different from the mammalian chromosome, substances that directly damage or adduct nucleotides in the DNA helix would be expected to act similarly in both chromosome types. E. coli mutation tests are performed in addition to Salmonella for some regulatory needs (Gatehouse et al. 1994). Premise #2. Tests for chromosome aberrations in mammalian cells are needed in addition to gene mutation tests. Chromosome aberrations are the classical genotoxic response, are involved in the tumor initiation and development processes, and are associated with a large proportion of human genetic diseases. Chromosome aberrations can also be used as a biomonitor of exposure; thus, such results in test systems can be correlated with chromosome damage events in humans. Additionally, there are genotoxic and carcinogenic chemicals that produce chromosome aberrations but not gene mutations, which would be not be identified if only gene mutation tests were used. Premise #3. A mammalian cell mutagenicity test is needed to confirm or complement the Salmonella mutation test. Mammalian cell tests are considered to be more relevant for mammalian carcinogenesis than are microbial or other nonmammalian tests because of the similarity of mammalian chromosomes and DNA repair and replication processes across species. Positive results in mammalian cells for a substance that is positive in bacterial cells ensure that the result seen in bacteria was not unique to the bacterial chromosome or bacterial metabolism. However, a negative result in mammalian cell tests would not necessarily negate the implications of a positive response in the bacterial test. Premise #4. An in vivo test is needed to confirm a positive in vitro test. In vivo tests are more relevant than in vitro tests because they integrate the relevant factors of test chemical absorption, distribution, metabolism, and excretion. As a result of these considerations, and because they use the animal’s metabolism rather than a surrogate metabolic system (i.e., a liver homogenate with cofactors),

232


in vivo tests should be less likely than in vitro tests to produce “false” positives. The ability to administer the test substance by the routes and doses relevant to human exposures allows for extrapolation of the results to humans. In addition, in vivo tests provide the potential to identify genetic damage in all tissues of interest as well as in germ cells, which cannot be adequately studied in vitro. Premise #5. Results from test batteries have a higher predictivity for cancer induction than results from the individual component tests. Gene mutations, chromosome aberrations, and nondisjunction are initial steps in the initiation of cancer; therefore chemicals causing any of these events are also capable of initiating cancer. Some chemicals induce only one type of relevant genetic damage—that is, gene mutations, chromosome aberrations, or nondisjunction—and therefore tests are needed for all endpoints. In order to be effective as a test battery, the combination of tests should be more effective than the individual tests for identifying chemicals of concern, and the tests must be complementary. That is, they must measure different effects and not duplicate each other to any large extent. In an ideal test battery, “false” negatives in one test will be correctly identified by one or more of the other tests. The accumulated data from the above testing schemes has shown that the in vitro tests are not complementary. The chemicals that are positive in the Salmonella test tend to also be positive in the mammalian cell tests, regardless of whether they are true positives (i.e., carcinogens) or “false” positives (i.e., noncarcinogens) (Zeiger 1998, 2001). Even more relevant to the interpretation of the test battery results has been the showing that the in vitro mammalian cell tests produce a high rate of false positives (Zeiger et al. 1990; Zeiger 1998, 2001; Kirkland et al. 2005). Every test has its false positives and false negatives; these values are generally quantified during validation tests and are partly a function of the chemicals being tested. When a battery of tests is used, each test that is added to the battery brings along its own, unique, true and false positives. Therefore, the more tests that are added to the battery to fill in the “gaps” left by the other tests, the more true positives that will be detected. This comes with a price: A higher proportion of false positives will also be detected. The additional false positives have the potential to overwhelm the number of true positives in a screening situation where the tests have a high sensitivity and the majority of test chemicals are not anticipated to be positive; that is, there is a low prevalence of noncarcinogens in the tested population.

9.6.

WHERE ARE WE NOW?

The current regulatory testing schemes are based on tier or battery testing. The initial tests used are the in vitro bacterial and mammalian cell gene mutation and mammalian cell cytogenetics tests. Chemicals that are negative at this level are typically not tested further. Chemicals that produce genetic effects in vitro are generally tested in short-term in vivo somatic cell tests to address the simple question of whether

9.6. WHERE ARE WE NOW?

233

or not the in vitro genetic toxicity can be translated to the animal and therefore, presumably, be more of a risk for carcinogenicity or heritable damage than chemicals that are not detected in the in vivo tests. However, the in vivo tests (i.e., bone marrow cytogenetics, in vivo/in vitro unscheduled DNA synthesis, and the transgenic mouse mutation test) generally used tend to be less sensitive than the in vitro tests. As a consequence, carcinogenic, DNA-reactive chemicals that are readily detected in vitro are often not detected in the in vivo assays, which is why a negative in vivo test is not sufficient, by itself, to negate the implications of the positive in vitro test. Similar difficulties arise when attempting to confirm in vitro positives as germ cell mutagens. In general, it is assumed that chemicals that do not produce somatic cell mutations or chromosome aberrations in vivo will not produce germ cell genetic effects. This is why a negative in vivo somatic cell test is sufficient to conclude that the substance will not be positive in germ cell tests. As a result, the potential carcinogenicity of the in vivo chemicals often dominates the implications of heritable mutations. There are a number of potential reasons for this lack of concordance between in vitro and in vivo responses and the relatively high rate of positives in the in vitro tests compared to the in vivo tests, primary of which are: The in vivo doses to the target cells are often lower than are reached in vitro, or the active metabolite may not be sufficiently stable to reach the target cells. Other possibilities include chemicals that may be uniquely positive in bacteria as a result of bacterial metabolic pathways not found in mammalian cells (e.g., sodium azide; Owais et al. 1979), differences in activation (or inactivation) activities, and metabolite profiles between the in vitro S9 preparation and in vivo metabolism (Ku et al. 2007). Additionally, in vitro mammalian cell systems, specifically those measuring chromosome damage, can produce artifactual positives as a secondary effect of high toxicity, high osmolality, or changes in pH (Brusick 1986; Galloway et al. 1987). A number of studies since the early 1990s have shown that the testing approach directed by the above-mentioned premises is not as effective as originally thought for identifying potential carcinogens. The reduced concordance of the genetic toxicity tests with rodent carcinogenesis is not unexpected because these genetic toxicology tests measure gene mutations, chromosome aberrations, and other chromosome damage; they do not measure cancer. Justification for their initial and continued use comes from the mechanistic relationship between mutations, chromosome damage, and cancer, as well as from the empirical correlations developed in the 1970s using model carcinogens. The high proportions of carcinogens that are not mutagenic in vitro (which became an issue in the l980s) led to the category of nongenotoxic carcinogens, which included chemicals that initiated the carcinogenic process by other than direct DNA damage—for example, hormonal, epigenetic cytotoxicity with subsequent cell proliferation. The question always arises as to why the recent test performance values (e.g., Tennant et al. 1987; Zeiger et al. 1990; Zeiger 1998; Kirkland et al. 2005) are poorer than the initial compilations in the 1970s, which showed that the Ames test could correctly identify 90% or more of carcinogens and noncarcinogens (McCann et al. 1975; Sugimura et al. 1976; Purchase et al. 1978), and other in vitro tests were similarly, or slightly less, effective (Preston et al. 1981; Clive et al. 1983).

234


The known carcinogens prior to 1980 comprised primarily alkylating agents and substances that were carcinogenic following short-term administration (i.e., less than 1 year) to the test animal. Simultaneous to these genetic toxicology test performance evaluations, the U.S. National Cancer Institute began a carcinogen testing program designed to identify carcinogenic chemicals by treating mice and rats for up to two years at doses up to what was described as a maximum tolerable dose (MTD; also defined as the minimally toxic dose) and performing an extensive histopathological evaluation of the animals at the end of this time (This testing program was subsequently incorporated into the National Toxicology Program). As opposed to many of the earlier carcinogenicity tests that exposed the animals for up to 1 year, the animals in this program were exposed for up to 2 years, which led to the appearance of tumors that are not normally expressed in less than 2 years of exposure. As a consequence of this expanded testing protocol, a number of chemicals were identified as carcinogens that would not have been identified in the shorter-term studies. Unlike the original listing of carcinogenic chemicals that were primarily DNAreactive chemicals, such as direct alkylating agents, aromatic amines, polycyclic aromatic hydrocarbons, and nitrosamines, many of the chemicals identified in the longer-term studies included such non-DNA reactive chemicals as chlorinated hydrocarbons, phthalates, and hormonally active substances. These latter classes of chemicals do not adduct DNA, tend to be less toxic, and tend not to induce gene mutations. These chemicals have contributed to the shift in the performance of the Ames test from a sensitivity of ≥90% to a sensitivity of 50–60% (i.e., a false-negative rate of 40–50%). The existence of this category of nongenotoxic carcinogens not detected by the Ames or other genetic tests has been well established. These substances will only be addressed by the development and validation of tests for other precancer mechanistic endpoints. In summary, the genetic toxicity tests routinely used for identifying potential carcinogens have not performed as effectively as their original promise, partially because the initial validation studies used potent carcinogens that were known or suspected to be DNA-reactive. Subsequent to those studies, a large number of rodent carcinogens have been identified that are not DNA-reactive and therefore are not detected, or poorly detected, in genetic toxicity tests. Similarly, a relatively high proportion of substances that are positive in mammalian cell systems have been shown to be noncarcinogenic. Despite these apparent deficiencies, the tests are widely used for screening chemicals to presumptively identify carcinogens. Although the sensitivity and specificity of the tests are not as high as originally anticipated, positive results in the genetic toxicity tests are highly predictive for rodent carcinogenicity. As a consequence of these issues, extensive efforts are underway to identify tests that are more predictive than the standard tests, or as predictive but with fewer false positives, which can be used to supplement or replace the tests currently used. Similarly, protocol modifications have been proposed for mammalian cell systems— for example, reducing the toxicity levels or test chemical concentrations that must be achieved for a test to be considered valid, which are designed to reduce the number of artifact-induced positive responses.

REFERENCES

9.7.

235

SUMMARY

The initial hopes and aspirations for the short-term genetic toxicology tests was the accurate prediction of carcinogenicity and the ability to distinguish between carcinogens and noncarcinogens. The in vitro tests remain the basis of carcinogen screening tests despite the knowledge that they are far from accurate for predicting carcinogenicity and are ineffective for identifying potential noncarcinogens (i.e., the specificities of the tests currently used are about 50%). In retrospect, it was naive to expect in vitro tests that measured point mutations and chromosome breakage to accurately reflect the multiple genetic and nongenetic steps between the induction of the initial DNA damage and the development of a tumor. The predictive ability of these short-term tests for mutation and chromosome breakage needs to be placed into context. The tests are designed and used to identify chemicals that cause cancer, with the ideal being the correct identification of carcinogens (by their genetic toxicity) and noncarcinogens (by their lack of genotoxicity). This search for the ideal should be compared with the interspecies predictivity of the in vivo cancer tests. In a compilation of tests performed by the NTP on rats and mice, in parallel, typically in the same laboratory, and using a larger database that went beyond the NTP studies and where the rat and mouse studies were not always performed in the same labs, the correspondence between rat and mouse carcinogenicity was 70–75% (Haseman et al. 1987; Gold et al. 1997). This level of interspecies predictivity of carcinogenicity under highly controlled conditions puts an upper limit on the predictivity of in vitro, single-cell systems and systems that measure clastogenicity or mutagenicity in single, typically nontarget, tissues of a single rodent species.

REFERENCES Ames, B. N. (1971). The detection of chemical mutagens with enteric bacteria. In Chemical Mutagens: Principles and Methods for Their Detection, Vol. 1, Hollaender, A., ed., Plenum Press, New York, pp. 267–282. Ames, B. N., Durston, W. E., Yamasaki, E., and Lee, F. D. (1973). Carcinogens are mutagens: A simple test system combining liver homogenates for activation and bacteria for detection. Proc Natl Acad Sci USA 70, 2281–2285. Auletta, A. E., Dearfield, K. L., and Cimino, M. C. (1993). Mutagenicity test schemes and guidelines: US EPA Office of Pollution Prevention and Toxics and Office of Pesticide Programs. Environ Mol Mutagen 21, 38–45. Bridges, B. A. (1973). Some general principles of mutagenicity screening and a possible framework for testing procedures. Environ Health Perspect 6, 221–227. Bridges, B. A. (1976). Short term screening tests for carcinogens. Nature 261, 195–200. Brusick, D. (1986). Genotoxic effects in cultured mammalian cells produced by low pH treatment conditions and increased ion concentrations. Environ Mutagen 8, 879–886. Clive, D., McCuen, R., Spector, J. F. S., Piper, C., and Mavournin, K. H. (1983). Specific gene mutations in L5178Y cells in culture. A report of the US Environmental Protection Agency Gene-Tox Program. Mutat Res 115, 225–251. Crow, J. F. (1968). Chemical risk to future generations. Scientist and Citizen, June–July, 113–117.

236


Demerec, M., Bertani, G., and Flint, J. (1951). A survey of chemicals for mutagenic action on E. coli. Am Naturalist 85, 119–136. DHEW (US Department of Health, Education, and Welfare). (1969). Report of the Secretary’s Commission on Pesticides and Their Relationship to Environmental Health. Parts I and II. US GPO, December 1969. Drake, J. W., Abrahamson, S., Crow, J. F., Hollaender, A., Lederberg, S., Legator, M. S., Neel, J. V., Shaw, M. W., Sutton, H. E., Von Borstel, R. C., and Zimmering, S. (1975). Environmental mutagenic hazards. Science 187, 503–514. Dunkel, V. C., Zeiger, E., Brusick, D., McCoy, E., McGregor, D., Mortelmans, K., Rosenkranz, H. S., and Simmon, V. F. (1985). Reproducibility of microbial mutagenicity assays: II. Testing of carcinogens and noncarcinogens in Salmonella typhimurium and Escherichia coli. Environ Mutagen 7(Suppl 5), 1–248. EPA (1975). Pesticide Program. Guidelines for Registering Pesticides in United States. Fed Reg (Part II) 40(123), 26802–26928 (Part VII, Subpart A, Methods for Studying Mutagenicity, pp. 26899–26900). Wednesday, June 25, 1975. EPA (1979). Environmental Protection Agency. Proposed Health Effects Test Standards for Toxic Substances Control Act Test Rules and Proposed Good Laboratory Practice Standards for Health Effects. Fed Reg (Part IV) 44(145), 44054–44093. (Subpart E, Mutagenic Effects §772.144, pp. 44080–44087). Thursday, July 26, 1979. EPA (1980). Mutagenicity Risk Assessments; Proposed Guidelines. Fed Reg 45(221), 74984–74988. Thursday, November 13, 1980. EPA (2008). OPPTS Harmonized Test Guidelines. Series 870 Health Effects Test Guidelines— Final Guidelines. http://www.epa.gov/opptsfrs/publications/OPPTS_Harmonized/870_Health_Effects_ Test_Guidelines/Series/ FDA (2000). Toxicological Principles for the Safety Assessment of Food Ingredients Redbook 2000, July 2000. IV.C.1. Short-Term Tests for Genetic Toxicity. http://vm.cfsan.fda.gov/∼redbook/red-ivc1. html. Flamm, W. G. (1974). A tier system approach to mutagen testing. Mutat Res 26, 329–333. Flamm, W. G., Valcovic, L. R., D’Aguanno, W., Fishbein, L., Green, S., Malling, H. V., Mayer, V., Prival, M., Wolff, G., and Zeiger, E. (1977). Approaches to determining the mutagenic properties of chemicals: Risk to future generations. J Environ Pathol Toxicol 1, 301–352. Galloway, S. M., Deasy, D. A., Bean, C. L., Kraynak, A. R., Armstrong, M. J., and Bradley, M. O. (1987). Effects of high osmotic strength on chromosome aberrations, sister-chromatid exchanges and DNA strand breaks, and the relation to toxicity. Mutat Res 189, 15–25. Gatehouse, D., Haworth, S., Cebula, T., Gocke, E., Kier, L., Matsushima, T., Melcion, C., Nohmi, T., Ohta, T., Venitt, S., and Zeiger, E. (1994). Mutat Res 312, 217–233. Gold, L. S., Slone, T. H., and Ames, B. N. (1997). Overview and update of analyses of the carcinogenic potency database. In Handbook of Carcinogenic Potency and Genotoxicity Databases, Gold, L. S., and Zeiger, E., eds., CRC Press, Boca Raton, FL, pp. 661–685. Haseman, J. K., Huff, J. E., Zeiger, E., and McConnell, E. E. (1987). Comparative results of 327 chemical carcinogenicity studies. Environ Health Perspect 74, 229–235. Hemmerly, J., and Demerec, M. (1955). XIII. Tests of chemicals for mutagenicity. Cancer Res Suppl 3, 69–75. Hollstein, M., McCann, J., Angelosanto, F., and Nichols, W. (1979). Short-term tests for carcinogens and mutagens. Mutat Res 65, 133–226. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans (1987). Genetic and Related Effects: An updating of selected IARC Monographs from Volumes 1 to 42, Supplement 6, Lyon, France. ICH (International Conference on Harmonisation) (2008). S2(R1): Guidance on Genotoxicity Testing and Data Interpretation for Pharmaceuticals Intended for Human Use; S2A: Guidance on Specific Aspects of Regulatory Genotoxicity Tests for Pharmaceuticals; S2B: Genotoxicity: A Standard Battery for Genotoxicity Testing for Pharmaceuticals. http://www.ich.org/cache/compo/276-254-1.html. Kirkland, D., Aardema, M., Henderson, L., and Muller, L. (2005). Evaluation of the ability of a battery of three in vitro genotoxicity tests to discriminate rodent carcinogens and noncarcinogens: I. Sensitivity, specificity and relative predictivity. Mutat Res 584, 1–256.

REFERENCES

237

Kirsch-Volders, M., Sofuni, T., Aardema, M., Albertini, S., Eastmond, D., Fenech, M., Ishidate, M., Jr., Lorge, E., Norppa, H., Surralles, J., von der Hude, W., and Wakata, A. (2000). Report from the in vitro micronucleus assay working group. Environ Mol Mutagen 35, 167–172. Ku, W. W., Bigger, A., Brambilla, G., Glatt, H., Gocke, E., Guzzie, P. J., Hakura, A., Honma, M., Martus, H.-J., Obach, R. S., and Roberts, S. (2007). Strategy for genotoxicity testing—Metabolic considerations. Mutat Res 627, 59–77. Maron, D. M., and Ames, B. N. (1983). Revised methods for the Salmonella mutagenicity test. Mutat Res 113, 173–215. McCann, J., Choi, E., Yamasaki, E., and Ames, B. N. (1975). Detection of carcinogens in the Salmonella/ microsome test: Assay of 300 chemicals. Proc Natl Acad Sci USA 72, 5135–5139. Mortelmans, K., and Zeiger, E. (2000). The Ames Salmonella/microsome mutagenicity assay. Mutat Res 455, 29–60. NRC (National Research Council) (1983). Identifying and Estimating the Genetic Impact of Chemical Mutagens. Committee on Chemical Environmental Mutagens, National Academies Press, Washington D.C., 295 pages. OECD (Organization for Economic Co-operation and Development) (2001). OECD Series on Testing and Assessment, No. 33. Harmonised Integrated Hazard Classification System for Chemical Substances and Mixtures. Chapter 2.5. Harmonised System for the Classification of Chemicals which Cause Mutations in Germ Cells. OECD (Organization for Economic Co-operation and Development) (2008). OECD Guidelines for the Testing of Chemicals. Section 4: Health Effects. [specifically, Guideline nos. 471–486] http://www. oecd.org/document/55/0,3343,en_2649_34377_2349687_1_1_1_1,00.html. Owais, W. M., Kleinhofs, A., and Nilan, R. A. (1979). In vivo conversion of sodium azide to a stable mutagenic metabolite in Salmonella typhimurium. Mutat Res 68, 15–22. Parry, J. M., Parry, E. M., Bourner, R., Doherty, A., Ellard, S., O’Donovan, J., Hoebee, B., de Stoppelaar, J. M., Mohn, G. R., Onfelt, A., Renglin, A., Schultz, N., Soderpalm-Berndes, C., Jensen, K. G., KirschVolders, M., Elhajouji, A., Van Hummelen, P., Degrassi, F., Antoccia, A., Cimini, D., Izzo, M., Tanzarella, C., Adler, I.-D., Kliesch, U., Schriever-Schwemmer, G., Gasser, P., Crebelli, R., Carere, A., Andreoli, C., Benigni, R., Leopardi, P., Marcon, F., Zijno, Z., Natarajan, A. T., Boei, J. J. W. A., Kappas, A., Voutsinas, G., Zarani, F. E., Patrinelli, A., Pachierotti, F., Tiveron, C., and Hess, P. (1996). The detection and evaluation of aneugenic chemicals. Mutat Res 353, 11–46. Preston, R. J., Au, W., Bender, M. A., Brewen, J. G., Carrano, A. V., Heddle, J. A., McFee, A. F., Wolff, S., and Wassom, J. S. (1981). Mammalian in vivo and in vitro cytogenetic assays: A report of the US EPA’s Gene-Tox Program. Mutat Res 87, 143–188. Purchase, I. F. H., Longstaff, E,, Ashby, J., Styles, J. A., Anderson, D., Lefevre, P. A., and Westwood, F. R. (1978). An evaluation of 6 short-term tests for detecting organic chemical carcinogens. Br J Cancer 37, 873–959. Sugimura, T., Sato, S., Nagao, M., Yahagi, T., Matsushima, T., Seino, Y., Takeuchi, M., and Kawachi, T. (1976). Overlapping of carcinogens and mutagens. In Fundamentals of Cancer Prevention, Magee, P. N., Takayama, S., Sugimura, T., and Matsushima, T. eds., University Park Press, Baltimore, MD, pp. 191–215. Szybalski, W. (1958). Special microbiological systems. II. Observations on chemical mutagenesis in microorganisms. Ann NY Acad Sci 76, 475–489. Tennant, R. W., Margolin, B. H., Shelby, M. D., Zeiger, E., Haseman, J. K., Spalding, J., Caspary, W., Resnick, M., Stasiewicz, S., Anderson, B., and Minor, R. (1987). Prediction of chemical carcinogenicity in rodents from in vitro genetic toxicity assays. Science 236, 933–941. Zeiger, E. (1998). Identification of rodent carcinogens and noncarcinogens using genetic toxicity tests: premises, promises and performance. Regul Toxicol Pharmacol 28, 85–95. Zeiger, E. (2001). Mutagens that are not carcinogens: Faulty theory or faulty tests? Mutat Res 492, 29–38. Zeiger, E. (2004). The history and rationale of genetic toxicity testing—An impersonal, and sometimes personal, view. Environ Mol Mutagen 44, 363–371. Zeiger, E., Haseman, J. K., Shelby, M. D., Margolin, B. H., and Tennant, R. W. (1990). Evaluation of four in vitro genetic toxicity tests for predicting rodent carcinogenicity: Confirmation of earlier results with 41 additional chemicals. Environ Mol Mutagen 16(Suppl 18), 1–14.

CH A P TE R

10

GENETIC TOXICOLOGY TESTING GUIDELINES AND REGULATIONS Lutz Müller Hans-Jörg Martus

10.1. HISTORICAL OVERVIEW OF GENOTOXICITY TESTING GUIDELINES As early as in the late 1940s, it was Auerbach who demonstrated that chemicals could be powerful mutagens (Auerbach and Robson 1946). Over the years, this created a concern that exposure to environmental chemicals could introduce deleterious alterations in the DNA of human beings in the environment. These concerns included damage to the germ line that could cause heritable disease and genetic alterations to individuals via somatic DNA damage (DHEW 1977; Meselson 1971; Wassom 1989). These concerns led to formation of the Environmental Mutagen Society in 1969 (Wassom 1989) and to the introduction of requirements for testing for mutagenic properties of chemicals in the 1970s. In this context, the U.S. Toxic Substances Control Act of 1976 specifically required the U.S. Environmental Protection Agency (EPA) to establish standards for the assessment of health and environmental effects associated with mutagenesis (TSCA 1976). During this period, the primary focus was on the potential of any chemical to induce germ-line mutations and to the development of appropriate testing methodologies for assessment of heritable mutations (Ehling et al. 1978; Meselson 1971). The thinking of the field at this key stage, when the recognition of the need for genetic toxicology testing had led to the initial formulation of testing requirements for genotoxicity, is illustrated by a key report of the department-wide working group of the U.S. Department of Health Education and Welfare (DHEW) (present name Department of Health and Human Services) issued in 1977 (DHEW 1977). This working group, formed by the DHEW Committee to Coordinate Toxicology and Related Programs, Subcommittee on Environmental Mutagenesis, was established in 1974 to develop a background document on mutagenicity test procedures and approaches to testing chemicals for mutagenic activity. The intent was “… to aid officials of regulatory agencies who have the responsibility for deciding:


238

10.1. HISTORICAL OVERVIEW OF GENOTOXICITY TESTING GUIDELINES

239

(1) advisability of promulgating test requirements for mutagenicity at the present time under any of their legislative authorities; (2) the appropriateness of mutagenicity tests for a wide range of product use and exposure categories; and (3) the reliability and interpretation of data from mutagenicity tests developed on substances of commerce within their regulatory purview in spite of the absence of formal testing requirements.” This report, entitled “Approaches to Determining the Mutagenic Properties of Chemicals: Risk to Future Generations,” emphasized two key points— first, that the primary concern about genotoxic damage was the potential to cause heritable genetic alterations in the human germ-line and, second, that quantitative assessment of the risk of heritable damage was necessary and that mere hazard identification was insufficient. The importance of quantitative risk assessment was emphasized: “It is not sufficient merely to identify substances which may pose a genetic hazard to the human population. Many such compounds will have a significant benefit factor and hence cannot reasonably be eliminated from use. Therefore, it is necessary to obtain quantitative data from relevant animal model systems from which extrapolation to humans can be made to predict virtually safe or tolerable levels of exposure.” Additionally, the association of mutagenesis with other toxicological endpoints such as carcinogenesis, teratogenesis, and aging was also noted. In the mid-1970s, the landmark publication of McCann et al. on the detection of carcinogens as mutagens based on an analysis of 300 chemicals, demonstrated a strong correlation of mutagenic activity in Salmonella with animal carcinogenicity (Ames et al. 1975; Maron and Ames 1983; McCann et al. 1975). This report generated great enthusiasm that inexpensive in vitro mutagenesis screening tests could be used to identify chemical carcinogens and hence control of exposure to such agents could potentially lower the human tumor burden. As regulatory guidelines were implemented during the 1970s and 1980s, there was a shift in focus from concern over germ-line mutagenesis to control of chemical carcinogens (MacGregor 1994). Though these early results in Salmonella were highly promising, it was already recognized at that time that mutations could arise by multiple mechanisms, some of which would not be detected in a nutritional reversion assay such as the Salmonella his reversion test. In particular, chromosomal interchanges, DNA strand breaks, and large chromosomal deletions, all characteristic of damage induced by ionizing radiation, which was one of the environmental mutagens of most concern during this period, are not efficiently detected in the Ames assay. Thus, an in vitro and in vivo test battery was devised that would detect the major classes of damage known to result in heritable mutations (NRC 1983). These concepts underlie the batteries currently in use (Brusick 1987; Hoffmann 1998). The types of lesions expected to be detected by the test systems most commonly used for mutagenesis screening at the present time are in line with our knowledge about the types of lesions involved in modifying the activity of oncogene products and tumor suppressor gene products. In the meantime, changes in such genes are widely accepted to be associated with cancer risk. Guidelines for testing environmental chemicals in the United States were delineated during the 1970s and 1980s (Auletta et al. 1993; Waters and Auletta 1981) and for food additives in 1982 (FDA 1982). Classically, the first batteries included (1) a bacterial test for gene mutation, (2) either an in vitro test for chromosomal

240

CHAPTER 10 GENETIC TOXICOLOGY TESTING GUIDELINES AND REGULATIONS

aberrations (based on the knowledge that ionizing radiation and radiomimetic chemicals produced high rates of chromosomal aberrations even when induced mutation rates were relatively low) or a mammalian cell mutagenesis test, and (3) a general test for DNA damage (FDA 1982). An in vivo test was generally encouraged, with preference for a test for bone marrow chromosomal aberrations or micronucleus induction, based on the knowledge of a few chemicals that were uniquely active in vivo (ICH 1997b; Tweats et al. 2007a). Much research effort was focused on development of appropriate mutagenicity testing methods that would detect a broad array of mutagenic chemicals. The classical series initiated by Hollaender, Chemical Mutagens: Principles and Methods for their Detection, was devoted to summarizing these methodologies (Hollaender 1971). By the time of the 1993 draft revision of the U.S. Food and Drug Administration’s (FDA) guidance on testing requirements for food and color additives, the U.S. FDA-recommended “core” testing battery consisted of the following: (1) a test for gene mutation in bacteria (S. typhimurium), (2) a test for gene mutation in mammalian cells in vitro, with the recommendation that the endpoint be based on an autosomal locus (so that events related to chromosomal interchanges could be detected), and (3) an assay for cytogenetic damage in vivo, with preference for a rodent bone marrow assay (FDA 1993). By the year 2000, these so called “Redbook” guidelines were finalized (FDA 2000). At this same time, the European, Japanese, and Canadian recommendations were similar. However, there were distinct differences in requirements both among regions and within different regulatory agencies within each region (DOH 1991; Purves et al. 1995; Shelby and Sofuni 1991). For example, the European recommendations generally included both an assay for gene mutation and an assay for chromosomal aberrations in mammalian cells (Kirkland 1993), while the Japanese relied on an in vitro mammalian cell chromosomal aberration assay and did not necessarily include the in vitro mammalian cell mutagenesis assay (Shirasu 1988). Test practices regarding potential genotoxicity of pharmaceuticals, including test quality and assessment issues, have been delineated in a series of publications communicated by members of the German regulatory authorities (Madle et al. 1987; Müller and Kasper 2000; Müller et al. 1991). The evaluation spans the period between 1982 and 1997 and addresses nearly 600 new pharmaceutical entities. These publications summarize changes in test selection, improvements in test quality and shifts in the focus of test interpretation and assessment. The initial review (Madle et al. 1987) as well as its update (Müller et al. 1991) focused on deficiencies in test quality which was at that time considered to be a major issue. By the 1990s, this was no longer considered a major issue. In addition, some genotoxicity systems which played a considerable role in the 1970s and 1980s, such as assays using yeast as indicator organisms, host-mediated assays, sister chromatid exchange (SCE) tests in vitro or in vivo, chromosomal aberration analysis in bone marrow or spermatogonia, and dominant lethal assays, were little used by the 1990s. In part this reflects changes in test philosophy including a move away from assays involving cells of the germ line (Müller and Kasper 2000). A similar evaluation addressing the experience in Germany with the review of tests for 776 new chemicals reviewed between 1982 and 1997 has been published

10.1. HISTORICAL OVERVIEW OF GENOTOXICITY TESTING GUIDELINES

241

(Broschinski et al. 1998). This evaluation focuses on the rates of positives in various standard in vitro systems. These data are correlated with chemical structure characteristics and genotoxicity effects versus cytotoxicity. A later review of pharmaceuticals on the U.S. market appeared to indicate a similar tendency including the fact that ∼20–30% of marketed pharmaceuticals seem to possess some kind of genotoxic potential especially in mammalian cells under in vitro conditions (Snyder and Green 2001). Because such data seem to indicate indirect means of genotoxicity in vitro, which may lack relevance in vivo, further evaluations have focused in a broader context on this issue. Kirkland et al. have published an updated comparison of in vitro genotoxicity assay results with The Carcinogenicity Potency Database, the most comprehensive carcinogenicity database available (CPDB 2007; Kirkland et al. 2005a). This evaluation showed in general terms that a battery of in vitro tests for genotoxicity can be pushed to high levels of sensitivity for detection of rodent carcinogens (sensitivity), but this sensitivity came at the price of inappropriately increasing the likelihood of obtaining a positive genotoxicity result for noncarcinogens (specificity) (Figure 10.1). Matthews et al. (2006) obtained confirmation of these results, which included the use of proprietary data. In the views of many scientists, there is a growing lack of confidence in the results that come out of regulatory in vitro genotoxicity tests. On the one hand, some scientists see a benefit to

100

Specificity and Sensitivity (%)

90 80 70

MN

Ames + MLA + MN MLA + MN Ames + MN MLA + Cab Ames + MLA + Cab Ames + MLA Ames + Cab

Ames MLA Cab

60

Sensitivity

Ames

Specificity

50 Cab

40 30

MLA MN

Ames + Cab Ames + MLA MLA + Cab Ames + MLA + Cab

20 10

Ames + MN

MLA + MN Ames + MLA + MN

0

Figure 10.1. Correlation data for the sensitivity and specificity for single in vitro genotoxicity assays and assay combinations. Ames, Salmonella reverse mutation test introduced by B. Ames; MLA, tk assay using the L5178Y mouse lymphoma cell line; Cab, chromosome aberration test with different mammalian cell types; MN, in vitro micronucleus assay with different mammalian cell types. [Data according to Kirkland et al. (2005a).]

242


approach maximum sensitivity for hazard identification followed by a weight-ofevidence approach for risk assessment (Bergman et al. 1996; Cimino 2006; Dearfield and Moore 2005; FDA 2006; Kasper et al. 2007; Müller et al. 2003). Others, however, see a need to refine the conditions for in vitro mammalian cell genotoxicity tests to optimize their predictivity or to introduce new assays (Kasper et al. 2007; Kirkland et al. 2005a,b, 2007a; MacGregor et al. 2000; Müller et al. 2003). In part, this is driven by the economic circumstances of restricted resources in the industry and the need to prioritize expenditures. In Europe, efforts in this context culminated in a publication from a workshop held under the auspices of the European Center for Validation of Alternative Methods (ECVAM) entitled “How to reduce false positive results when undertaking in vitro genotoxicity testing and thus avoid unnecessary follow-up animal tests: Report of an ECVAM Workshop” (Kirkland et al. 2007a). In addition, the International Life Sciences Institute (ILSI) has instituted a working group that tackles “Relevance and follow-up of positive results in in vitro genetic toxicity assays: An ILSI-HESI initiative” (Thybaud et al. 2007a,b). Furthermore, recent changes to regulatory genotoxicity testing have been communicated with the revised ICH Guideline “Guidance on Genotoxicity Testing and Data Interpretation for Pharmaceuticals Intended for Human Use” (ICH 2008). These changes include (a) a reduction in the top concentrations for in vitro mammalian cell tests from 10 mM to 1 mM and (b) more stringent criteria for acceptable ranges of cytotoxicity for evaluation of test compounds. The scientific reasoning for these recommendations can be found later in this chapter. With this historical overview and short elaborations on recent regulatory developments on genotoxicity testing, interpretation, and risk assessment, this chapter will now focus on major sets of internationally relevant guidelines and scientific efforts to support the concepts of regulatory testing. In the views of the authors of this chapter, there are two major sets of internationally applicable regulatory guidelines for genotoxicity testing and two major international scientific processes that dominate the regulatory landscape: 1. The Organisation for Economic Cooperation and Development (OECD) test guidelines 2. The International Conference of Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) guideline(s) for genotoxicity testing of pharmaceuticals 3. The International Workshop(s) on Genotoxicity Tests (IWGT) 4. The International Program on Chemical Safety (IPCS) under auspices of the World Health Organization (WHO) These guidelines and processes are selected because they have been driving the scientific process (IWGT and IPCS) and have set internationally acknowledged standards of testing that go beyond country or regional borders (ICH and OECD). This chapter does not focus on genotoxicity guidelines for compounds for other purposes such as pesticides, new chemicals, food additives, and so on. The reader is referred to review articles that cover these regulations in more detail (e.g., Cimino 2006; Kirkland et al. 2005b) or to the specific guidelines such as those from the

10.3. HUMAN USE ICH GUIDELINES FOR PHARMACEUTICALS

243

United Kingdom’s Committee on Mutagens (UKCOM 2000), the U.S. Environmental Protection Agency’s (EPA’s) Office of Prevention, Pesticides and Toxic Substances’ health effects test guidelines series 870 (EPA 2008), and the European Union’s (EU) guidelines for the testing of chemicals under Annex V, Part B (EC 2008). It might be worthwhile to mention here that regulatory guidance documents on genotoxicity are partly influenced by general considerations prevailing in the society such as the three R’s (replace, refine, reduce) to optimize use of animals in safety testing and to reduce their burden. For example, in the European Union this has led to a ban of animal testing for decorative cosmetics, which require a focus on in vitro genotoxicity testing only (Kirkland et al. 2005b). This approach will require work on more predictive in vitro tests for the purpose of risk assessment (Kirkland et al. 2008).

10.2. ORGANIZATION FOR ECONOMIC COOPERATION AND DEVELOPMENT (OECD) GUIDELINES FOR GENOTOXICITY The OECD has played a major role in developing recommendations for internationally harmonized testing protocols. The protocols developed by OECD are very influential, because parties to the OECD treaty, which include most of the major industrialized nations of the world, have agreed to accept testing protocols developed by the OECD consensus process. These guidelines also served as the bases from which most of the U.S. EPA’s health effects guidelines and the EU’s Annex V testing guidelines were developed. OECD testing guidelines are updated periodically and are available from the OECD in Paris, France and from the following URL: http://www.oecd.org/sourceoecd/. The OECD guidelines embrace a number of test guidelines for genetic toxicity, some of which are little used in practice nowadays (OECD 1997). A very recent addition to the collection of OECD Guidelines is a draft guideline on the in vitro micronucleus test (OECD 2006) on the basis of various validation exercises for this test (Corvi et al. 2008; Kirsch-Volders et al. 1997, 2003). There is a great interest in this guideline since this test approach is simpler in evaluation than the cytogenetic evaluation of chromosome aberrations that are normally required by guidelines.

10.3. INTERNATIONAL CONFERENCE OF HARMONIZATION OF TECHNICAL REQUIREMENTS FOR REGISTRATION OF PHARMACEUTICALS FOR HUMAN USE (ICH) GUIDELINES FOR PHARMACEUTICALS On a worldwide basis, the resources that are spent to identify new chemicals with the potential to cure diseases or help to cope with symptoms of diseases are enormous and far exceed expenses in any other area of chemistry. Enormous is also

244


the economic pressure of the health-care systems to give patients access to pharmaceuticals at a reasonable price while still maintaining scientific progress. In this context, any unreasonable regulatory requirements that result in redundancies in testing of pharmaceutical candidates and their registration are counterproductive. Hence, the ICH was established in the early 1990s to reduce such redundancies and to strive for international harmonization. This unique project brought together the regulatory authorities of Europe, Japan, and the United States and experts from the pharmaceutical industry in these three regions. Recommendations (ICH guidances) for the economical use of human, animal, and material resources in quality, safety, efficacy, and multidisciplinary (“Q”, “S”, “E” and “M” guidances) testing were developed by these parties and are now accepted internationally as the standards for evaluation of pharmaceuticals (http://www.ich.org/). Genotoxicity was established as an ICH guidance topic at the first ICH conference in November 1991 in Brussels. Within the subsequent years of negotiation between the parties involved, two ICH genotoxicity guidances were developed: (1) the guidance on “Specific Aspects of Regulatory Genotoxicity Tests for Pharmaceuticals” (ICH 1996) and (2) a guidance on “A Standard Battery for Genotoxicity Testing of Pharmaceuticals” (ICH 1997b). These two ICH guidances are complementary and are the principal guidances on genotoxicity studies for pharmaceuticals in the three ICH regions. Detailed information on the background of these guidances, along with their text, has been published (Müller et al. 1999). The ICH guidances address test procedures, as well as strategy and test interpretation. The ICH guidance “A Standard Battery for Genotoxicity Testing of Pharmaceuticals” recommends a core battery of tests “core” testing battery for pharmaceutical registration, which consists of the following: (1) a test for gene mutation in bacteria, (2) a test for chromosomal aberrations in mammalian cells in vitro or the L5178Y mouse lymphoma mammalian cell mutagenesis test, and (3) an in vivo test for chromosomal damage in rodent hematopoietic cells. Compounds giving negative results in this battery, performed and evaluated in accordance with current recommendations, will usually be considered to have a sufficient level of assurance of safety to allow product approval. Within 10 years of the ICH guidelines for genotoxicity testing in operation, it was realized that advances in genotoxicity testing and interpretation would require a maintenance process. This process was initiated at the end of 2006 and has resulted in a new single draft ICH “Guidance on Genotoxicity Testing and Data Interpretation for Pharmaceuticals Intended for Human Use” (ICH 2008). In the following, the main principles of this new ICH guideline are laid down and discussed. It is understood that some of these principles will bring about changes to genetic toxicology testing and its regulatory use, and we may see genetic toxicology guidelines for purposes other than testing of pharmaceuticals follow these rationales. I. The Ames test continues to be an elementary and indispensable part of regulatory testing. However, there is no continued need to repeat negative Ames tests in an independent experiment. While it is clear that there are some differences between mammalian cells and bacteria in regard to metabolism and DNA repair processes, it continues to be


245

acknowledged that there is no suitable alternative for the bacterial reverse mutation test (Ames test). The Ames test is the most widely used test for genotoxic activity with unparalleled easiness, cross-laboratory robustness, and specificity for mutagenic carcinogens (Gatehouse et al. 1994; Kirkland et al. 2006). It appears logical that, provided that the appropriate metabolic pathways are incorporated, the ability of a chemical to produce DNA damage that result in mutations, which conversely initiate cells to outgrow to tumors, is most easily measured in bacteria. II. The in vitro micronucleus test is endorsed as an alternative option to the in vitro chromosome aberration test and the mouse lymphoma tk assay. Many years of protocol evaluation and validation exercises imply that the in vitro micronucleus test has reached a status of reliability that is comparable with the mouse lymphoma tk assay in L5179Y cells or the chromosome aberration test with various cell lines of primary human lymphocytes (Corvi et al. 2008; Lorge et al. 2006, 2007; OECD 2006). Hence, it can be used interchangeably with these assays in the regulatory world. Since many industrial laboratories already screen for genotoxic activity in the in vitro micronucleus test in early stages of nonregulatory activity, this should now enable a seamless transition from early non-GLP screening activities into the regulatory GLP testing phase. III. An extensive review of exposure data to pharmaceuticals suggests that testing to a concentration of 1 mM (instead of 10 mM) for nontoxic compounds in mammalian cells in vitro is sufficient. Traditionally, in vitro tests for genotoxicity have been viewed as hazard identification tests to be followed up by in vivo tests for risk identification or risk assessment. Under such a view, a maximal sensitivity approach has often been the goal for in vitro tests, and an upper limit of 10 mM (or 5 mg/ml) for test material in the cell culture has been applied for testing of compounds that were nontoxic. It is understood that this level somehow represented worst-case assumptions that human cells might be at risk to ∼5% of the foreign test material in their in vivo environment. This upper limit was also borne out of early testing experience that some mutagenic carcinogens appeared to require such high concentrations to elicit a chromosome damaging response in mammalian cells in vitro (Scott et al. 1991). Consequently, very often, positive results in vitro were viewed as potentially relevant for a chronic low-level exposure in vivo because of the stochastic element in genetic toxicology and mutation induction where fully linear dose–response characteristics are thought to prevail. In practice, evidence for genotoxic activity in vitro has led in numerous cases to the conduct of extensive in vivo evaluations without ever reaching a conclusion that was acceptable to regulatory review. Another consequence was the discontinuation of development of potentially useful products very early before wasting resources on further activities with uneconomic delays. Over the years, however, there was growing evidence in the applied science for nonlinearity of many aspects of genotoxic activity. Thus, mistrust has been building up in the judgment of in vitro positive findings. An essential element of in vivo testing and risk assessment is the comparison of concentrations that are positive in vitro and the exposure that can be

246


reached under in vivo conditions. In this context, human pharmaceuticals offer the best possible judgment basis because exposure in animals and in humans is measured by default and into high dose ranges usually employed in animals studies. A limit of 1 mM maintains the element of hazard identification, being higher than clinical exposures to known pharmaceuticals, including those that concentrate in tissues (Hardman et al. 2001), and is also higher than the levels generally achievable in preclinical studies in vivo. Certain drugs are known to require quite high clinical exposures—for example, nucleoside analogs and some antibiotics. While comparison of potency with existing drugs may be of interest to sponsors, perhaps even above the 1 mM limit, it is ultimately the in vivo tests that determine relevance for human safety. IV. Concerns over growing numbers of nonrelevant positive findings in mammalian cell tests in vitro will also be counterbalanced by limiting the levels of cytotoxicity to “at most 50%” for in vitro chromosome aberration and micronucleus tests. This proposal is supported by an extensive review of results obtained with in vitro hazard identification testing and in vivo risk assessment testing in the pharmaceutical industry. Though some genotoxic carcinogens are not detectable with in vitro genotoxicity assays unless the concentrations tested induce some degree of cytotoxicity, particularly when measured by colony forming assays, DNA damaging agents are generally detectable with only moderate levels of toxicity (e.g., 30% reduction in growth measured at the time of sampling in the chromosome aberration assay) (Greenwood et al. 2004). As cytotoxicity increases, mechanisms other than direct DNA damage by a compound or its metabolites can lead to ‘positive’ results that are related to cytotoxicity and not genotoxicity. Such indirect induction of DNA damage secondary to damage to non-DNA targets is more likely to occur above a certain concentration threshold. The disruption of cellular processes is not expected to occur at lower, pharmacologically relevant concentrations. In cytogenetic assays, even weak clastogens that are known to be carcinogens are positive without exceeding a 50% reduction in cell counts. On the other hand, compounds that are not DNA damaging, mutagenic, or carcinogenic can induce chromosome breakage at toxic concentrations. For cytogenetic assays in cell lines, measurement of cell population growth over time by measuring the change in cell number during culture relative to control (e.g., by the method referred to as population doubling) has been shown to be a useful measure of cytotoxicity (Greenwood et al. 2004), because it is known that cell numbers can underestimate toxicity (Kirkland et al. 2007a). For lymphocyte cultures, an inhibition of mitotic index (MI) not exceeding about 50% is considered sufficient. For the in vitro micronucleus assay, a limit of about 50% cytotoxicity is also appropriate. Moreover, for the in vitro micronucleus assay, since micronuclei are scored in the interphase subsequent to a mitotic division, it is important to verify that cells have progressed through the cell cycle. This can be done by use of cytochalasin B to allow nuclear division but not cell division, so that micronuclei can be scored in binucleate cells (the preferred method for lymphocytes). Other methods to demonstrate cell proliferation, including cell population growth over time (PD)


247

as described above, may be used for cell lines (Kirsch-Volders et al. 2003; Lorge et al. 2006, 2007). For the mouse lymphoma tk+/− assay (MLA), appropriate sensitivity is achieved by limiting the top concentration to one with close to 20% relative total growth (RTG) both for soft agar and for microwell methods. This is based on reviews of published data using the current criteria described by Moore et al. (2006), which found very few chemicals that were positive in MLA only at concentrations with less than 20% RTG and that were rodent carcinogens, and convincing evidence of genotoxic carcinogenesis for this category is lacking. The consensus is that caution is needed in interpreting results when increases in mutation are seen only below 20% RTG, and a result would not be considered positive if the increase in mutant fraction occurred only at ≤10% RTG (Moore et al. 2006, 2007). Because of the inherent difficulties to obtain an almost exact value of 20% RTG in an MLA, it is acceptable to approach a range of 10–20 RTG for a valid assay with a compound that produces toxicity. In conclusion, caution is appropriate in interpreting positive results obtained as reduction in growth/survival approaches or exceeds 50% for cytogenetic assays or 80% for the MLA. It is acknowledged that the evaluation of cells treated at these levels of cytotoxicity/clonal survival may result in greater sensitivity, but bears an increased risk of nonrelevant positive results (Kirkland et al. 2007a). The battery approach for genotoxicity is designed to ensure appropriate sensitivity without the need to rely on single in vitro mammalian cell tests at high cytotoxicity. To obtain an appropriate toxicity range, a preliminary range-finding assay over a broad range of concentrations is useful, but in the genotoxicity assay it is often critical to use multiple concentrations that are spaced quite closely (less than twofold dilutions). Extra concentrations may be tested, but not all concentrations will need to be evaluated for genotoxicity. It is not intended that multiple experiments be carried out to reach exactly 50% reduction in growth, for example, or exactly 80% reduction in RTG. V.

Because pharmaceuticals are normally tested for toxicity in rodent repeated dose toxicity tests and because there is no longer a requirement for an acute high dose rodent toxicity test, the assessment of genotoxicity (e.g., bone marrow micronucleus test or other tissue/endpoint) should be integrated, if feasible, into the rodent repeated dose toxicity study to optimize animal usage. VI. The options for a standard battery of genotoxicity tests are expanded by the possibility to choose to conduct an in vivo test with investigation of genotoxic damage in two tissues instead of conducting an in vitro test with mammalian cells followed by an in vivo test. In conjunction with the respective ICH carcinogenicity guidances (see under http://www.ich.org), the ICH genotoxicity guidances are setting a new standard for genetic toxicology and carcinogenesis testing, assessment, and interpretation which is applicable in most parts of the world. Different criteria are proposed for so-called exploratory investigational new drug applications (exploratory IND). This is the case for clinical investigations in which volunteers will receive a low number of doses of an investigational drug at

248


relatively low or micro-doses, with the intent of collecting limited human information. In this scenario, it is acceptable to conduct a standard bacterial mutation assay (ICH 2008).

10.4. INTERNATIONAL WORKSHOP ON GENOTOXICITY TESTS (IWGT) Four workshops have been organised previously under the auspices of the IWGT. The International Association of Environmental Mutagen Societies (IAEMS) formalized these workshops in 2002 under the IAEMS umbrella and agreed that they would be held on a continuing basis in conjunction with the International Conferences on Environmental Mutagens (ICEM) that are held every four years (Kirkland et al. 2007b,c). In this way, an ongoing process of international scientific discussion and harmonization of testing methods and testing approaches has been established that can take advantage of the international experts who attend these meetings. These ongoing workshops have proven to be useful to ensure that different recommendations for methodology in these new assays do not arise in different parts of the world (Kirkland et al. 2007b,c) and thus avoid situations that could lead to (i) unnecessary duplication of testing to satisfy local requirements, (ii) variations in the test performance, (iii) potential differences in test outcome, and (iv) unjustified differences in the use of test data for description, assessment, and management of risk. The IWGT process is implemented through working groups of recognized international experts from industry, academia, and the regulatory sectors, with due attention to geographical, disciplinary, and sector balance. For each working group, a chairperson, deputy chair, and rapporteur are appointed. Experts in the science of each topic are invited to bring experimental data to bear on the discussions; the remit of each group is to derive recommendations based on data, and not on unsupported opinion or anecdotal information (Kirkland et al. 2007b). There are several objectives sought in bringing together representatives from around the world to share their experiences in generating and evaluating genotoxicity data from a variety of methodologic and strategic approaches. The IWGT strives to (i) attain a greater understanding of true test performance from a wide database, (ii) provide recommendations that minimise misinterpretation, (iii) recognize that no single assay can detect every genotoxicant, and (iv) achieve compromise for the sake of harmonisation or acceptance that more than one approach is both reasonable and valid. Because of the IWGT approach—in particular, development of data-driven consensus by the key global experts from academia, government, and industry—IWGT recommendations have been seen as state of the art and have high credibility. These recommendations serve as important supplements to established regulatory guidelines and provide a sound basis for updating those guidelines as the state of science advances. With OECD and ICH guidelines constituting the two major sets of internationally harmonised genotoxicity guidelines in regulatory use, the IWGT process and working group recommendations are of particular help in supplementing test design

10.5. THE IPCS UNDER THE AUSPICES OF THE WORLD HEALTH ORGANIZATION

249

and interpretation of genotoxicity test packages that are based on these guidelines. For further information on the IWGT recommendations, the reader is referred to the various special issues of the peer-reviewed journal Mutation Research that have emerged from the IWGT workshops (Kirkland et al. 2007b).

10.5. THE INTERNATIONAL PROGRAM ON CHEMICAL SAFETY (IPCS) UNDER THE AUSPICES OF THE WORLD HEALTH ORGANIZATION (WHO) The first harmonized scheme for mutagenicity testing on behalf of the IPCS was published in 1996 (Ashby et al. 1996). Similar to the maintenance process of ICH guidances (see Section 10.3), the WHO has recognized important developments in the field and has decided to update this IPCS Harmonized Scheme as part of the IPCS Harmonization of Approaches to the Assessment of Risk from Exposure to Chemicals. A draft for public and peer review was prepared by an International Drafting Group Meeting of experts held at the Fraunhofer Institute for Toxicology and Experimental Medicine in Hanover, Germany, on 11–12 April 2007 (IPCS 2007). The approach presented by this IPCS expert group focused on the identification of mutagens and genotoxic carcinogens. The approach is shortly presented below. For further details, the reader is referred to updates in the harmonized scheme emerging from this IPCS group. The term “mutation” as understood in this document comprises gene mutations, as well as structural and numerical chromosome alterations. The group is aware of other mechanisms leading to carcinogenicity and other heritable diseases, but their identification requires additional types of mechanistic studies. The group proposed to use a weight-of-evidence approach at various stages of the outlined testing strategy. However, it is also stated that a clear positive result at a single mutagenicity endpoint, even when multiple negative results in other endpoints have been reported, is generally sufficient for the classification “positive.” Most short-term tests in bacteria and mammalian cell cultures have been designed primarily for hazard identification and, thus, can represent only the starting point in the process of risk assessment. Whether or not the observed effects are relevant for human exposure depends on bioavailability, absorption, metabolism, half-lives, and other factors that require investigation in vivo. Especially when choosing in vivo assays and when proceeding into germ cell mutagenicity studies, expert judgment is required to select the appropriate test system(s) and to avoid uninformative and thus unnecessary animal experiments. Before initiating mutagenicity testing on a particular compound, the following aspects should be considered: (i) chemical structure and class of the agent (possible structure–activity relationships) and physicochemical properties, such as solubility and stability; (ii) expected routes of metabolism, chemical and biological reactivity/ activity, and relationship to known genotoxic chemicals; and (iii) routes of exposure, bioavailability, and target organ(s).

250


10.6.

IN VITRO TESTING

Usually two or three different tests in bacteria and mammalian cells are selected to cover the endpoints of gene mutations, clastogenicity (structural chromosome aberrations), and aneuploidy (numerical chromosome aberrations), taking into account physicochemical properties of substances under consideration (see Chapter 11).

10.6.1.

In Vitro Tests

Screening should be based on a limited number of tests that are well-validated and informative. Genetic toxicity test batteries generally include the following: (1) A test for gene mutation in bacteria (bacterial reverse mutation assay): Organisation for Economic Cooperation and Development (OECD) Guideline 471 recommends the use of at least five strains of bacteria: (i) Salmonella typhimurium TA1535, (ii) S. 12 typhimurium TA1537 or TA97 or TA97a, (iii) S. typhimurium TA98, (iv) S. typhimurium TA100, and (v) Escherichia coli WP2 or E. coli WP2uvrA or S. typhimurium TA102. The choice of additional tests depends on the chemical structure and class of the agent. (2) In vitro mammalian assays: These assays should evaluate the potential of a chemical to produce point mutations, clastogenicity, and/or aneugenicity, by using either mammalian cell lines or primary human cell cultures such as fibroblasts or lymphocytes (e.g., mouse lymphoma TK assay or cytogenetic evaluation of chromosomal damage in mammalian cells via in vitro micronucleus test).

10.6.2.

Evaluation of In Vitro Testing Results

Evaluation of results and classification into: (i) positive results, (ii) negative results, and (iii) inconsistent, conflicting, or equivocal results. Positive: Substance is positive at one or more endpoints of mutagenicity. Negative: Substance is negative in all test systems under appropriate in vitro conditions; the substance is not mutagenic (genotoxic) in vitro and is predicted not to be mutagenic in vivo [for exceptions, see Tweats et al. (2007a,b)]. Inconsistent, conflicting, or equivocal (i.e., borderline biological or statistical significance): All other substances.

10.6.3.

Follow-Up to In Vitro Testing

In case of positive results: Conduct an in vivo test with selection of an appropriate endpoint; if necessary, conduct further in vitro studies to optimize in vivo testing (e.g., kinetochore staining as an addition in the micronucleus assay of in vitro aneugens). In case of negative results: Further in vivo testing is required only in the case of “high” or “moderate and sustained” exposure, or for chemicals of high concern.

10.7. IN VIVO TESTING

251

In case of inconsistent, conflicting, or equivocal results: Further in vitro testing to clarify positive or negative results; depending on whether the situation is resolved by further in vitro testing, proceed according to “Positive” or “Negative.”

10.7.

IN VIVO TESTING

In vivo tests should be chosen carefully to avoid an uninformative outcome. Therefore, toxicokinetics, metabolism, chemical reactivity, and mode of action have to be considered carefully. Typically, a bone marrow micronucleus or clastogenicity test is conducted. However, if there are indications that point to a more appropriate assay, then this assay should be conducted instead (e.g., mutagenicity study with transgenic animals; comet assay in stomach/small intestine/colon, if there is no uptake via gastrointestinal tract; comet assay in the liver if there is metabolism to toxic species) (see Chapter 12).

10.7.1.

Follow-Up to In Vivo Testing

In case of positive results: The compound is an “in vivo somatic cell mutagen” and testing for germ cell mutagenicity may be required. In case of negative results: Further in vivo testing is required only in the case of positive in vitro studies; again, the second in vivo test is chosen on a case-by-case basis as stated above. If the test is negative, it is concluded that there is no evidence for in vivo mutagenicity. In case of equivocal results: Equivocal results may be due to low statistical power, which can be improved by increasing the number of treated animals and/or scored cells. If the situation is unresolved, a second in vivo test is required, chosen on a case-by-case basis (ordinarily on a different endpoint or in a different tissue, depending on toxicokinetics, metabolism, and mode of action); proceed according to “Positive” or “Negative.”

10.7.2.

Strategy for Germ Cell Testing

When information on the risk to the offspring of exposed individuals is important, the following germ cell testing strategy is recommended. For substances that give positive results for mutagenic effects in somatic cells in vivo, their potential to affect germ cells should be considered. If there is toxicokinetic or toxicodynamic evidence that germ cells are actually exposed to the somatic mutagen, it is reasonable to conclude that the substance may also pose a mutagenic hazard to germ cells and thus a risk to future generations. Where germ cell testing is required, judgment should be used to select the most appropriate test strategy. There are a number of tests available, which fall into two classes: (1) tests in germ cells per se (“class 1”) and (2) tests to detect effects in the offspring (or potential offspring) of exposed animals (“class 2”). Three

252


internationally recognized OECD guidelines are available for such studies (OECD 1997): (1) clastogenicity in rodent spermatogonial cells (class 1), (2) the dominant lethal test (class 2), and (3) the mouse heritable translocation assay (class 2). In order to minimize the use of animals in germ cell testing, it is advisable to start with tests that detect effects in germ cells per se (class 1). These methods include (but are not limited to) gene mutation tests in transgenic animals, gene mutations in the Expanded Simple Tandem Repeat (ESTR) assay (Dubrova et al. 1998), chromosomal assays including those using fluorescence in situ hybridization (FISH) (Hill et al. 2003), comet assay (Burlinson et al. 2007; Hartmann et al. 2003; Merk and Speit 1999; Pfuhler and Wolf 1996), and DNA adduct analysis (Phillips et al. 2000). Following the use of such tests, if quantification of heritable effects is required (class 2), an assay for ESTR mutations can be performed with the offspring of a low number of exposed animals. Tests used historically to investigate transmitted effects (i.e., the heritable translocation test and the specific locus test) can also be performed; however, they use large numbers of animals.

10.8. EUROPEAN UNION GUIDELINE FOR TESTING OF CHEMICALS UNDER THE REGISTRATION, EVALUATION, AUTHORIZATION AND RESTRICTION OF CHEMICAL (REACH) REACH is a new EU policy aimed at evaluating the health risks of chemicals marketed (i.e., produced in or imported) in the EU. It was proposed by the European Commission at the end of 2003, and it came into effect in June 2007 (EC 2006; EU 2003). The objective of REACH is to realize a more rapid and less expensive method to identify risks for exposed humans and at the same time minimize the use of laboratory animals (van der Jagt et al. 2004). Also, clearly the improvement of the competitiveness of the European chemicals industry is defined as another goal. In this context, it is important to understand that of the ∼30,000 chemicals which are in regular human use, and of which many are high-tonnage chemicals, only ∼3% are sufficiently tested for toxicological risk (EC 2001). Therefore, one of the aims of REACH is to provide guidance for a retrospect safety evaluation of existing chemicals, as well as for an evaluation of newly developed and marketed chemicals. The REACH regulation requires toxicological information from every chemical with a marketing volume of >1 tonnes per annum (t/a). In the context of testing for a genotoxic potential, REACH considers genotoxicity testing also a surrogate for a mechanistic link to carcinogenicity, although under some circumstances germ cell mutagenicity—that is, the risk of induction of heritable diseases—is evaluated in addition. The amount of information required is dependent on the tonnage in that a higher production volume means that more information is needed. A technical dossier is required for all registered chemicals; however, only chemicals with a production volume greater than 10 t/a require a chemical safety assessment documented in a chemical safety report. No new toxicological information is required for substances between 1 and 10 t/a, unless they are considered a priority substance (i.e., carcinogenic, mutagenic, or toxic to reproduction [CMR]) (EC 2006). Column

10.8. EUROPEAN UNION GUIDELINE FOR TESTING OF CHEMICALS UNDER THE REACH

253

1 “Standard Information Required” of the Annexes of REACH describe the standard information requirements for substances produced or imported in quantities of ≥1 t/a (Annex VII), ≥10 t/a (Annex VIII), ≥100 t/a (Annex IX), and ≥1000 t/a (Annex X); different tonnage classes have defined testing requirements (EC 2006). Also, in the context of REACH, it should be understood that existing chemicals (called “phasein” substances) are considered equivalent to newly introduced chemicals (called “non phase-in” substances) with respect to the toxicological testing data required. This means that a major part of information obtained will be on phase-in substances. In general, like for other areas of genotoxicity testing, information on gene mutations, structural chromosome aberrations (clastogenicity), and numerical chromosome aberrations (aneugenicity) is required (Aardema et al. 1998). Also, the guidance specifically defines the terms mutagenicity and genotoxicity in a way that is also acceptable for other areas of genetic toxicology (ECHA 2008a): Mutagenicity refers to the induction of permanent transmissible changes in the amount or structure of the genetic material of cells or organisms. These changes may involve a single gene or gene segment, a block of genes or chromosomes. The term clastogenicity is used for agents giving rise to structural chromosome aberrations. A clastogen can cause breaks in chromosomes that result in the loss or rearrangements of chromosome segments. Aneugenicity (aneuploidy induction) refers to the effects of agents that give rise to a change (gain or loss) in chromosome number in cells. An aneugen can cause loss or gain of chromosomes resulting in cells that have not an exact multiple of the haploid number. For example, three number 21 chromosomes or trisomy 21 (characteristic of Down syndrome) is a form of aneuploidy. Genotoxicity is a broader term and refers to processes which alter the structure, information content or segregation of DNA and are not necessarily associated with mutagenicity. Thus, tests for genotoxicity include tests which provide an indication of induced damage to DNA (but not direct evidence of mutation) via effects such as unscheduled DNA synthesis (UDS), sister chromatid exchange (SCE), DNA strand breaks, DNA adduct formation or mitotic recombination, as well as tests for mutagenicity.

Testing of chemicals under REACH is done in a tiered fashion, in that the different tonnage classes define the necessary data for which testing may be required. This is one of the fundamental differences between REACH and other regulated substances (e.g., regulation of pharmaceuticals, where the production volume is not considered at all). The second fundamental difference between chemical testing under REACH and testing of chemicals for human use (e.g., pharmaceuticals or cosmetics) lies in the fact that chemical testing is mainly conducted for the purpose of labeling and defining appropriate protection measures and exposure limitations, whereas particularly for drug ingredients, the definition and quantification of a benefit–risk ratio and a risk assessment derived therefrom is the predominant goal of a test strategy. This is mainly justified by the fact that exposure to pharmaceuticals generates a therapeutic benefit and that—in contrast to the chemicals under REACH—consideration, administration, and exposure scenarios for pharmaceuticals are well-defined. Therefore, the tonnage principle may be viewed as a surrogate for the lack of reliable exposure information, so that a higher tonnage is considered to result in the higher likelihood of human exposure in general.

254


For chemicals, the main product of a risk evaluation under REACH is the Chemical Safety Assessment (CSA), which is prepared for all chemicals above 10 t/a to establish the safe conditions of manufacture and use of a substance for all lifecycle stages, submitted to the European Chemicals Agency (ECHA) as part of the registration dossier (ECHA 2008b,c). Although the minimum information requirements under REACH is primarily determined by tonnage triggers, it may be adapted due to hazard, exposure, or risk considerations, as well as technical difficulties in testing the substance. In addition, the level of follow-up testing is defined by the result in one of the preceding genotoxicity assays, an approach that is reminiscent to as the regulation of other substances (e.g., pharmaceuticals). All eligible chemicals under REACH require the conductance of a bacterial reverse mutation (Ames) test (Annex VII, §8.4.1) (EC 2006). For compounds ≥10 t/a, an additional mammalian in vitro cytogenetic test is required (Annex VIII, §8.4.2) (EC 2006). For the latter, both the chromosome aberration test and the micronucleus test in vitro are considered acceptable. In the case of a negative result in the mammalian in vitro cytogenetic test and in the in vitro gene mutation test in bacteria, data from an in vitro gene mutation study in mammalian cells are required (Annex VIII, §8.4.3) (EC 2006). A positive in vitro cytogenetic result should be evaluated for the possibility that the chemical tested acts via an aneugenic mechanism, which is appropriately addressed by an in vitro assay, should the endpoint not be covered already by the cytogenetic test conducted. Appropriate in vivo mutagenicity studies should be considered in case of a positive result in any of those genotoxicity studies (Annex VIII, §8.4) (EC 2006). For compounds marketed at ≥100 t/a, an in vivo somatic cell mutagenicity test is requested (Annex IX, §8.4) (EC 2006). Dependent on the data obtained in the above tests, a second in vivo genotoxicity test may be required. If there is a positive result obtained in any in vivo somatic cell study, the potential for germ cell mutagenicity should be considered on the basis of all available data, including toxicokinetics. If no clear conclusion about germ cell mutagenicity can be made, additional investigations shall be considered (Annex IX, §8.4) (EC 2006). Obviously, the major aim of the in vivo genotoxicity tests is to confirm the positive results of in vitro genotoxicity tests in vivo and to identify those compounds where a positive in vitro result does not translate into a positive in vivo finding—that is, in vitro effects defined as false-positive. REACH suggests that this should be accomplished with only one in vivo test. A second in vivo test is considered necessary when the compound has shown both a clastogenic and a gene mutagenic potential in vitro and the in vivo test is considered inadequate to address both endpoints simultaneously. In addition, evidence for specific genotoxicity to germ cells may trigger the performance of a second in vivo study. Eventually, the assessment of the totality of data obtained after the REACH-compliant test strategy will be the classification (with corresponding labeling) into the following categories*: *In the EU, dangerous substances and preparations must be classified and labeled according to Directives 67/548/EEC and 1999/45/EC, respectively. These Directives will be repealed and replaced with the EU Regulation on classification, labeling, and packing of substances and mixtures, implementing the Globally Harmonised System (GHS) in the EU.

10.8. EUROPEAN UNION GUIDELINE FOR TESTING OF CHEMICALS UNDER THE REACH

255

Category 1 (label T R46): Substances known to be mutagenic to man. In this category, substances would fall with sufficient evidence to establish a causal association between human exposure to a substance and heritable genetic damage. Human mutation epidemiology studies are needed for chemicals to be placed in this category (EC 1967, 1999). However, so far no chemical has received this label (Appendix 3, Point 29—Mutagens: category 1) (EC 2006). Category 2 (label T R46): Substances which should be regarded as if they were mutagenic for man (EC 1967, 1999). For compounds in this category, there is sufficient evidence to provide a strong presumption that human exposure to the substance may result in the development of heritable genetic damage, generally on the basis of appropriate animal studies or other relevant information (Appendix 4, Point 29—Mutagens: category 2) (EC 2006). Category 3 (label Xn R68 or R40): Substances that cause concern to man owing to possible mutagenic effects. For compounds in this category, there is evidence from appropriate mutagenicity studies, but this is insufficient to place the substance in Category 2 (EC 1967, 1999). Obviously, those requirements will necessitate the conductance of additional animal experiments. Therefore, in order to fulfill the intention to reduce the need for animal experiments, REACH requests that all available in vitro data, in vivo human data and animal–human data, data from in silico SAR systems, and data from structurally related substances be evaluated before additional tests be carried out or that arguments be provided to waive additional testing. In addition to pharmaceuticals, for those substances regulated under REACH, an independent determination of genotoxic risk to germ cells is a requirement, whereas for pharmaceuticals, which are regulated mainly according to ICH guidances, germ cell genotoxicity tests are not foreseen. However, also under REACH it is acknowledged that heritable mutation is the consequence of a general mutagenic effect elicited in germ cells. The product of a mutagenic potential and the ability of a chemical to be distributed into the germ line is considered and therefore is not an independent genotoxic endpoint. As a consequence, all available toxicokinetic and toxicodynamic properties of the test substance are taken into consideration to estimate whether there is sufficient information to conclude that the substance poses a mutagenic hazard to germ cells. If this is the case, it can be concluded that the substance may cause heritable genetic damage and no further testing is justified. If no clear conclusion can be drawn, additional toxicokinetic experiments or tests for inheritable mutations may be conducted, although toxicokinetic investigations are preferred. The trigger to consider a compound a potential germ cell mutagen would be a positive in vivo genotoxicity test; and depending on this dataset, the appropriate germ cell genotoxicity test (i.e., germ cell clastogenicity of dominant lethal test for chromosome damage, transgenic animals for gene mutations, DNA binding, or the comet assay for direct DNA damage) would be selected. Also, like for pharmaceuticals, evidence for a relevant genotoxic potential to male or female germ cells will, without otherwise convincing

256


evidence, led to the assumption that a chemical is also leading to reproductive toxicity, given the mechanistic association of both endpoints. In summary, the newly introduced REACH guidance, which is regulating the safety of chemicals marketed or imported into the EU and which has been finalized recently, is, for the purpose of genetic toxicology, based on a stepwise strategy of sequential tests, each based on the results of the respective previous step and leading to a labeling of the chemical which in turn spikes measures to limit the exposure of humans (or the environment) during production, transport, or use of the chemical. The new regulation treats already existing chemicals equally to those brought newly into the market (produced or imported) in the EU. So far, it remains to be seen how the reduction in the number of animal experiments, one of the goals of REACH, will be achieved, while at the same time compiling a comprehensive database on all the existing compounds.

10.9. SPECIALTY GUIDELINES FOR GENOTOXICITY: GENOTOXIC IMPURITIES IN PHARMACEUTICALS In recent years, the testing and control of pharmaceuticals for the presence of genotoxic impurities have been under regulatory scrutiny. Because of their dedicated use—in many cases chronic use—in humans, pharmaceuticals are expected to be of high purity and little batch-to-batch variability is allowed. This focus is justified compared to chemicals of other use areas, for which exposure may be dedicated but limited (e.g., cosmetics), more of an accidental type (e.g., household chemicals), or low and chronic under workplace or use conditions (e.g., pesticides, industrial chemicals). The relevant ICH guidelines concerning the qualification of impurities in commercial manufacture are Q3A(R) and Q3B(R) that focus on impurities in drug substances and drug products, respectively, while Q3C recommends limits for residual solvents in the drug product (ICH 1997a, 2002, 2003). The guidance given in these regulatory documents is considered to be applicable at the time of registration of a new pharmaceutical entity. The first two guidelines describe threshold levels above which impurities are required to be reported, identified, and qualified either in toxicological investigations or in the clinic. The threshold levels vary according to the maximum daily dose of a drug. For drug substances, the identification thresholds are within the range of 500 and 1000 ppm (i.e., 0.05 and 0.1%). While in general very high purity of more than 98% is attained for pharmaceuticals, the presence of impurities even at low levels of 0.1% or lower may cause unwanted effects or may be of concern for chronic intake. This is of particular importance when the drug is taken at a high daily dose. Hence, the ICH Guidelines Q3A(R) and Q3B(R) take precaution for this case and state that although identification of impurities is not generally necessary at levels less than or equal to the identification threshold, “analytical procedures should be developed for those potential impurities that are expected to be unusually potent, producing toxic or pharmacological effects at a level not more than (≤) the identification threshold.” Thus in the case of impurities where a potential safety concern for genotoxicity exists, the guidelines imply that the routine identification threshold is not considered to be applicable.

10.9. SPECIALTY GUIDELINES FOR GENOTOXICITY

257

One may ask the following questions: (I) Why do genotoxic impurities exist at all in pharmaceuticals? (II) Why can’t they be totally eliminated? I: The synthesis of often complex chemical structures that bear in themselves sufficient specificity and selectivity for a pharmacological action is only manageable via reactive steps of chemistry. Such reactions often involve or result in electrophilic intermediates, which may possess genotoxic activity. It is a dream of every chemist to create reaction conditions that fully create a product from its educts. This dream can hardly be put into reality. II: Purification and control are part of the means to manage processes in a way that they can be economically reliable and result in safe products. Purification steps often result in significant loss of the product, and a balance between safety requirements and economic conditions must be found. Since genotoxic compounds are usually considered as potentially carcinogenic with a linear dose–response relationship, genotoxic impurities are considered separately from the existing ICH Guidelines and limits of acceptability have to be set. Because there was no general guidance on how to do this, considerable differences occurred between regulatory authorities, even within the same region. To reduce the differences in judgement on genotoxic impurities between EU Member States, the Committee of Human Medicinal Products (CHMP) decided to ask the Safety Working Party (SWP) to develop a guideline on genotoxic impurities. This guideline was released after much discussion in June 2006 (CHMP 2006). The central idea in this draft guideline is the concept of the Threshold of Toxicological Concern (TTC). A TTC value of 1.5 μg/day intake of a genotoxic impurity is derived from a large database of animal carcinogenicity studies. It is estimated that the intake of any genotoxic impurity, with a few exceptions of highly potent genotoxic carcinogens, below this TTC is associated with an acceptable risk of less than one excess cancer in a population of 100,000 people (i.e., 1–3

>3–6

>6–12

>12

120

40

20

10

1.5

a

For details see Müller et al. (2006).

TABLE 10.2. Generic Staged TTC Values for Genotoxic Impurities During the Clinical Trial Stage for Pharmaceutical Candidate Compounds Proposed by the CHMP in the EU

Duration of Exposurea

Allowable Daily Intake (ADI) in μg/person/day:

Single Dose

≤1

≤3

≤6

≤12

>12

120

60

20

10

5

1.5

a

Note: The U.S. FDA has proposed adoption of a similar set of values for investigational new drug applications, with the exception that the ADI of 120 μg/day be used for exposures of less than 14 days and that 60 μg/day be used for exposures from 14 days to 1 month (FDA 2008).

chronic lifetime exposure cannot be excluded; (ii) in the clinical trial stage, the benefits of a pharmaceutical candidate are not fully established. Hence, the generic TTC value of 1.5 μg/day is proposed for any intake duration of more than 12 months, and all staged TTC values for shorter duration of exposure are calculated using an acceptable risk of less than one excess cancer in a population of 1,000,000 (i.e., 50% for lymphocytes. However, there are growing concerns as to the relevance of genotoxic effects that are found only at highly cytotoxic concentrations (Galloway 2000), and this may be reflected by the poor specificity of mammalian cells tests (Kirkland et al. 2005).

11.6.

THE IN VITRO MICRONUCLEUS TEST

Analysis of structural chromosomal aberrations requires considerable training. Normal chromosomes in metaphase preparations can display a variety of appearances, and a full understanding of these different manifestations of normality is necessary before abnormal chromosomes can be satisfactorily identified. The demands of training and the diligent approach needed for thorough analysis means that only 100 cells per replicate culture (usually 2 replicates per test concentration) are scored for aberrations. A more recent technique that will also detect structural chromosomal damage, but requires much less training, is more rapid and allows analysis of larger numbers of cells is the in vitro micronucleus test. Another reason for the increased interest in this test is its ability to detect aneugens. The need for a specific assay to detect genome mutation (i.e., chromosome loss/gain) has been considered by a number of regulatory authorities. Aneuploidy is considered to be a condition in which the chromosome number of a cell or individual deviates from a multiple of the haploid set. The maintenance of karyotype during cell division depends upon the fidelity of chromosome replication and the accurate segregation of chromosomes to daughter cells. In turn, these events depend upon different cell organelles functioning correctly and a number of metabolic activities related specifically to cell division (e.g., synthesis of nuclear spindle, proteins, etc.). Aneuploidy can occur through errors of many types; hence there are numerous cellular targets that can lead to chromosome gain or loss. Briefly, the mechanisms by which aneuploidy can occur fall into several categories, including damage to the mitotic spindle and associated elements, damage to chromosomal substructures, chromosome rearrangements, alterations to cellular physiology, and mechanical disruption. The importance of aneuploidy to adverse human health is well accepted, and the effects of aneuploidy include birth defects, spontaneous abortions, and infertility. Tumor cells frequently have alterations in chromosome number, and several specific aneuploidies have been associated with tumor development, although whether this is the cause or the effect of tumorigenesis is not clear.

11.6. THE IN VITRO MICRONUCLEUS TEST

281

When chromosomes fail to segregate correctly, this process of nondisjunction can lead to the production of both monosomic and trisomic progeny cells. If chromosomes are lost from the dividing nucleus, they produce monosomic progeny without the reciprocal trisomic cell and the expelled chromosomes become membrane-bound and are detected as micronuclei outside the main progeny nuclei. Consequently, chromosome loss can be measured by monitoring micronucleus formation, and the in vitro micronucleus assay using mammalian cells provides such a technique (Fenech and Morley 1985). As mentioned above, this methodology has also been developed and validated as a simpler method for detecting structural chromosome damage because micronuclei may also arise from acentric chromosome fragments (lacking a centromere), which are unable to migrate with the rest of the chromosomes during the anaphase of cell division. Because micronuclei in interphase cells can be assessed much more objectively than chromosomal aberrations in metaphase cells, there is not such rigorous a requirement for training personnel and slides can be scored more quickly. This makes it practical to score thousands instead of hundreds of cells per treatment and thus imparts greater statistical power to the assay. Recently, methodology has been published that allows micronucleus analysis to be conducted using flow cytometry, thus further enhancing the utility of this assay (Avlasevich et al. 2006). The in vitro micronucleus assay may employ cultures of established cell lines, cell strains, or primary cultures, including human and Chinese hamster fibroblasts, mouse lymphoma cells, and human lymphocytes. Guidelines have been published recommending suitable protocols (Kirsch-Volders et al. 2000, 2003). To analyze the induction of micronuclei, it is essential that nuclear division has occurred in both treated and untreated cultures. It is therefore important to provide evidence that cell proliferation has occurred after test chemical exposure. Analysis of the induction of micronuclei in human lymphocytes has indicated that the most convenient stage to score micronuclei in this cell system is the binucleate interphase stage. Such cells have completed one mitotic division after chemical treatment and are thus capable of expressing micronuclei. Treatment of the cells with the inhibitor of actin polymerization cytochalasin B inhibits microfilament assembly and cytokinesis, thus preventing the separation of daughter cells after mitosis and trapping them at the binucleate stage. A schematic for this method is shown in Figure 11.6, and an example of a binucleate cell with a micronucleus is shown in Figure 11.7. The principle of the method is to expose cell cultures to the test substance in both the presence and absence of an in vitro metabolizing system. After exposure, the cultures are grown for a period sufficient to allow chromosome damage or chromosome loss to lead to the formation of micronuclei in interphase cells (usually 1.5–2 normal cell cycles after the start of treatment). Harvested and stained interphase cells are then analyzed microscopically for the presence of micronuclei. If the cytokinesis-block technique is used, micronucleus analysis is restricted to binucleate cells, and at least 1000 lymphocytes per duplicate culture should additionally be classified as mononucleates, binucleates, or multinucleates to estimate the cytokinesis-block proliferation index, which is a measure of cell cycle delay. Micronuclei formed by aneuploidy induction can be distinguished from those produced by clastogenic activity by the presence of centromeric DNA or kinetochore

282

CHAPTER 11


Figure 11.6. The basis for the in vitro micronucleus assay using cytochalasin B.

Figure 11.7. Example of a binucleate cell with a micronucleus. See insert for color representation of this figure.

proteins in the micronuclei. Fluorescent in situ hybridization (FISH) with pancentromeric DNA probes can be used to detect the former, whereas specific antibodies can be used to detect the presence of kinetochores (Migliore et al. 1996; Schuler et al. 1997). If aneuploidy is the suspected cause of micronucleus induction, further analysis of binucleate cells can be performed with chromosome-specific probes and the mal-segregation of chromosomes between daughter nuclei can be studied. An

11.7. IN VITRO TEST FOR UNSCHEDULED DNA SYNTHESIS IN RAT HEPATOCYTES

283

Figure 11.8. Example of nondisjunction of 2 chromosomes induced by an aneugen in a binucleate cell using fluorescence in situ hybridization (FISH). See insert for color representation of this figure.

example of a cell showing nondisjunction of 2 chromosomes as a result of aneugen treatment is shown in Figure 11.8. Until recently, most regulatory guidelines have focused mainly on tests for gene mutations and structural chromosome damage. However, the validation of the in vitro micronucleus test and the development of an OECD guideline indicates that its use will become much more widespread (OECD 2007). It will continue to be of considerable importance to establish a specific role for chromosome loss in tumor development. The analysis of aneuploidy in interphase cells of solid tumors using FISH will be greatly advantageous in this respect. For cancer risk assessment purposes, results from aneuploidy assays can be considered particularly useful when the mode of action of a chemical is known to result in chromosome loss or nondisjunction.

11.7. IN VITRO TEST FOR UNSCHEDULED DNA SYNTHESIS IN RAT HEPATOCYTES Because the liver is the major organ of xenobiotic metabolism, hepatocytes are an appropriate cell in which to conduct genotoxicity tests. Because they do not divide readily in culture, they are not so useful in tests for chromosomal aberrations or mutations, but they can be used to detect DNA damage or repair where dividing cells are not needed. In studies with fresh hepatocytes, it is not necessary to include S9 mix.

284

CHAPTER 11


UDS assays quantify the resultant excision repair of DNA following a permanent change such as covalent binding of an activated mutagen or a reactive chemical species generated intracellularly. Cells undergoing such repair synthesize DNA at stages of the cell cycle other than S-phase, where normal replicative (scheduled) DNA synthesis takes place, hence the term “unscheduled DNA synthesis.” This technique is potentially highly sensitive because the whole genome is theoretically a target for chemical reaction. A detailed description of the methodology was given by Madle et al. (1994). Briefly, viable hepatocyte populations are prepared by perfusing the livers of rats with collagenase (Madle et al. 1994). The hepatocytes are treated in vitro for 16–20 hours in the presence of the radiolabeled nucleotide [3H]thymidine. Uptake of the tritium into nuclear or cytoplasmic DNA is identified by autoradiography. Cells undergoing repair are identified by increases in the number of silver grains overlying the nuclei compared to those overlying the cytoplasm. S-phase cells exhibit extremely high numbers of nuclear silver grains and are excluded from analysis.

11.8.

IN VITRO COMET ASSAY

A useful way to measure direct damage to DNA is the single-cell gel electrophoresis assay or “comet” assay. This is a rapid and simple system for measuring alkali labile sites and overt strand breaks in the DNA of mammalian cells (Fairbairn et al. 1995). During electrophoresis, damaged (fragmented) DNA penetrates further than undamaged DNA into the agar gel in which the cells are embedded. The basis for this assay is represented in Figure 11.9. The technique can be applied to virtually any cell culture from which a single-cell suspension can be prepared (McKelvey-Martin et al. 1993). After treatment, the cells are suspended in agar and exposed to strong alkali, which denatures the proteins and permits DNA unfolding. Electrophoresis is then performed, during which time the supercoiled DNA relaxes and fragmented DNA is pulled toward the anode. After electrophoresis, the slides are neutralized and stained with a DNA-specific stain such as propidium iodide or ethidium bromide, when the cell ghosts with damaged DNA are visible as comets of various sizes

Figure 11.9. The theoretical basis for the formation of DNA comets.

11.9. STRENGTHS AND LIMITATIONS

Damaged cells

285

Control cells

Figure 11.10. Visualization of comets using ethidium bromide. See insert for color representation of this figure.

(hence the name), whereas those with undamaged DNA are visible as round images (examples are shown in Figure 11.10). DNA can be determined visually by the categorization of comets into different “classes” of migration or by using an eyepiece micrometer to estimate image or tail length. However, image analysis is recommended with the measurement of parameters such as the percentage of DNA in the tail (percent migrated DNA), tail length, and tail moment (fraction of migrated DNA multiplied by some measure of tail length). Of these, tail moment and/or tail length measurements are the most commonly reported, but there is much to recommend the use of percent DNA in tail, because this gives a clear indication of the appearance of the comets and is linearly related to the DNA break frequency over a wide range of levels of damage (Collins 2004). Cell death is associated with increased levels of DNA strand breaks, and in the comet assay the microscopic image resulting from necrotic or apoptotic cells can be comets with small or nonexistent heads and large diffuse tails, commonly called “clouds” or “hedgehogs.” However, such cells can also be seen after treatment with high concentrations of strong mutagens indicating that these images are not uniquely diagnostic for apoptosis/necrosis (Collins 2004). For this reason, it is recommended that relatively high levels of viability are achieved (e.g., no more than 30% cytotoxicity) at the end of treatment. Further validation and development of this methodology is continuing and recommended guidance on the correct conduct of the comet assay have been published (Hartmann et al. 2003; Tice et al. 2000).

11.9.

STRENGTHS AND LIMITATIONS

Analysis of the ability of in vitro genotoxicity tests to predict carcinogenicity in rodents has shown them to have a high sensitivity (i.e., giving positive results with carcinogens), particularly when combined in batteries of complementary tests

286

CHAPTER 11


(Kirkland et al. 2005; Tennant et al. 1987). Predictivity rises above 90% for batteries of 2 or 3 tests, and only carcinogens with a known nongenotoxic mode of cancer induction, or those chemicals that are extremely weak carcinogens, are missed. However, apart from the Ames test, the specificity (i.e., ability to give negative results with noncarcinogens) of the in vitro tests is poor. For the mammalian tests (e.g., mouse lymphoma, chromosomal aberration, micronucleus) used alone, the specificity was less than 50%, and when combined in batteries of two or three tests the specificity fell to 33% or 25%, respectively (i.e., the chances of a wrong prediction of carcinogenicity rose to 2/3 or 3/4). Snyder and Green (2001) also reported that from a total of 467 pharmaceuticals examined in the 1999 Physicians Desk Reference and from the open literature, 75% were positive in at least one in vitro assay (Snyder and Green 2001). Experimental culture conditions such as changes in pH and high osmolality are known to cause false-positive results in in vitro mammalian assays. However, other biochemical and physiological stresses such as inhibition of protein synthesis, inhibition of DNA synthesis or repair, inhibition of topoisomerases, overload of metabolism, and so on, can also lead indirectly to DNA damage and genotoxic responses, particularly in mammalian cells. Consequently a positive result in any one in vitro assay does not necessarily mean that the chemical poses a genotoxic/carcinogenic hazard to humans. It is believed that some of these irrelevant positives result from the lack of functional p53 in many cell lines, from artifacts resulting from high levels of cytotoxicity or high concentrations that overload normal metabolism and defense mechanisms. These have been discussed at length in Kirkland et al. (2007). There have therefore been calls for in vitro tests in mammalian cells in particular to be more robust using p53 and DNA repair-proficient cells, with a stable karyotype, avoiding excessive cytotoxicity and test chemical concentrations. Further investigation in relevant in vivo assays is usually required to put any positive in vitro results into perspective. The in vivo tests have advantages in terms of relevant metabolism, and so on, and also allow the influence of detoxification mechanisms to be assessed. However, some understanding of the mode of action leading to the genotoxic response, and whether this is relevant for humans or may have a threshold, is important. Nonetheless, positive results from in vitro tests that are wrong or irrelevant predictors of in vivo mutagenic or carcinogenic hazard will lead to either abandoning the development of certain products or the unnecessary use of animals in follow-up studies. Thus while in vitro genotoxicity tests are very useful, we need new or modified tests for the future that demonstrate improved specificity without compromising sensitivity.

REFERENCES Ames, B. N. (1971). The detection of chemical mutagens with enteric bacteria. In Chemical Mutagens, Principles and Methods for Their Detection, Vol. 1, ed. H. Hollaender. Plenum Press, New York, pp. 267–282. Ames, B. N., Durston, W. E., Yamasaki, E., and Lee, F. D. (1973). Carcinogens are mutagens: A simple test system combining liver homogenates for activation and bacteria for detection. Proc Natl Acad Sci USA 70, 2281–2285.

REFERENCES

287

Ashby, J., and Tennant, R. W. (1988). Chemical structure, Salmonella mutagenicity and extent of carcinogenicity as indicators of genotoxic carcinogenesis among 222 chemicals tested in rodents by the U.S. NCI/NTP. Mutat Res 204, 17–115. Avlasevich, S. L., Bryce, S. M., Cairns, S. E., and Dertinger, S. D. (2006). In vitro micronucleus scoring by flow cytometry: Differential staining of micronuclei versus apoptotic and necrotic chromatin enhances assay reliability. Environ Mol Mutagen 47, 56–66. Clive, D., Caspary, W., Kirby, P. E., Krehl, R., Moore, M., Mayo, J., and Oberly, T. J. (1987). Guide for performing the mouse lymphoma assay for mammalian cell mutagenicity. Mutat Res 189, 143–156. Cole, J., Fox, M., Garner, R. C., McGregor, D. B., and Thacker, J. (1990). Gene mutation assays in cultured mammalian cells. In Basic Mutagenicity Tests: UKEMS Recommended Procedures Kirkland, D. J., ed., Cambridge University Press, Cambridge, pp. 87–114. Collins, A. R. (2004). The comet assay for DNA damage and repair: Principles, applications, and limitations. Mol Biotechnol 26, 249–261. Elliott, B. M., Combes, R. D., Elcombe, C. R., Gatehouse, D. G., Gibson, G. G., Mackay, J. M., and Wolf, R. C. (1992). Alternatives to Aroclor 1254-induced S9 in in vitro genotoxicity assays. Mutagenesis 7, 175–177. Fairbairn, D. W., Olive, P. L., and O’Neill, K. L. (1995). The comet assay: A comprehensive review. Mutat Res 339, 37–59. Fenech, M., and Morley, A. A. (1985). Measurement of micronuclei in lymphocytes. Mutat Res 147, 29–36. Galloway, S. M. (2000). Cytotoxicity and chromosome aberrations in vitro: Experience in industry and the case for an upper limit on toxicity in the aberration assay. Environ Mol Mutagen 35, 191–201. Galloway, S. M., Aardema, M. J., Ishidate, M., Jr., Ivett, J. L., Kirkland, D. J., Morita, T., Mosesso, P., and Sofuni, T. (1994). Report from working group on in vitro tests for chromosomal aberrations. Mutat Res 312, 241–261. Gatehouse, D., Haworth, S., Cebula, T., Gocke, E., Kier, L., Matsushima, T., Melcion, C., Nohmi, T., Ohta, T., Venitt, S., et al. (1994). Recommendations for the performance of bacterial mutation assays. Mutat Res 312, 217–233. Guengerich, F. P., Dannan, G. A., Wright, S. T., Martin, M. V., and Kaminsky, L. S. (1982). Purification and characterization of liver microsomal cytochromes p-450: Electrophoretic, spectral, catalytic, and immunochemical properties and inducibility of eight isozymes isolated from rats treated with phenobarbital or beta-naphthoflavone. Biochemistry 21, 6019–6030. Hartmann, A., Agurell, E., Beevers, C., Brendler-Schwaab, S., Burlinson, B., Clay, P., Collins, A., Smith, A., Speit, G., Thybaud, V., and Tice, R. R. (2003). Recommendations for conducting the in vivo alkaline comet assay. 4th International Comet Assay Workshop. Mutagenesis 18, 45–51. Kirkland, D., Aardema, M., Henderson, L., and Muller, L. (2005). Evaluation of the ability of a battery of three in vitro genotoxicity tests to discriminate rodent carcinogens and noncarcinogens I. Sensitivity, specificity and relative predictivity. Mutat Res 584, 1–256. Kirkland, D., Pfuhler, S., Tweats, D., Aardema, M., Corvi, R., Darroudi, F., Elhajouji, A., Glatt, H., Hastwell, P., Hayashi, M., Kasper, P., Kirchner, S., Lynch, A., Marzin, D., Maurici, D., Meunier, J. R., Muller, L., Nohynek, G., Parry, J., Parry, E., Thybaud, V., Tice, R., van Benthem, J., Vanparys, P., and White, P. (2007). How to reduce false positive results when undertaking in vitro genotoxicity testing and thus avoid unnecessary follow-up animal tests: Report of an ECVAM Workshop. Mutat Res 628, 31–55. Kirsch-Volders, M., Sofuni, T., Aardema, M., Albertini, S., Eastmond, D., Fenech, M., Ishidate, M., Jr., Kirchner, S., Lorge, E., Morita, T., Norppa, H., Surralles, J., Vanhauwaert, A., and Wakata, A. (2003). Report from the in vitro micronucleus assay working group. Mutat Res 540, 153–163. Kirsch-Volders, M., Sofuni, T., Aardema, M., Albertini, S., Eastmond, D., Fenech, M., Ishidate, M., Jr., Lorge, E., Norppa, H., Surralles, J., von der Hude, W., and Wakata, A. (2000). Report from the In Vitro Micronucleus Assay Working Group. Environ Mol Mutagen 35, 167–172. Madle, S., Dean, S. W., Andrae, U., Brambilla, G., Burlinson, B., Doolittle, D. J., Furihata, C., Hertner, T., McQueen, C. A., and Mori, H. (1994). Recommendations for the performance of UDS tests in vitro and in vivo. Mutat Res 312, 263–285.

288

CHAPTER 11


McKelvey-Martin, V. J., Green, M. H., Schmezer, P., Pool-Zobel, B. L., De Meo, M. P., and Collins, A. (1993). The single cell gel electrophoresis assay (comet assay): A European review. Mutat Res 288, 47–63. Migliore, L., Cocchi, L., and Scarpato, R. (1996). Detection of the centromere in micronuclei by fluorescence in situ hybridization: Its application to the human lymphocyte micronucleus assay after treatment with four suspected aneugens. Mutagenesis 11, 285–290. Moore, M. M., Honma, M., Clements, J., Awogi, T., Bolcsfoldi, G., Cole, J., Gollapudi, B., HarringtonBrock, K., Mitchell, A., Muster, W., Myhr, B., O’Donovan, M., Ouldelhkim, M. C., San, R., Shimada, H., and Stankowski, L. F., Jr. (2000). Mouse lymphoma thymidine kinase locus gene mutation assay: International Workshop on Genotoxicity Test Procedures Workgroup Report. Environ Mol Mutagen 35, 185–190. Moore, M. M., Honma, M., Clements, J., Bolcsfoldi, G., Cifone, M., Delongchamp, R., Fellows, M., Gollapudi, B., Jenkinson, P., Kirby, P., Kirchner, S., Muster, W., Myhr, B., O’Donovan, M., Oliver, J., Omori, T., Ouldelhkim, M. C., Pant, K., Preston, R., Riach, C., San, R., Stankowski, L. F., Jr., Thakur, A., Wakuri, S., and Yoshimura, I. (2003). Mouse lymphoma thymidine kinase gene mutation assay: International Workshop on Genotoxicity Tests Workgroup report—Plymouth, UK 2002. Mutat Res 540, 127–140. Moore, M. M., Honma, M., Clements, J., Harrington-Brock, K., Awogi, T., Bolcsfoldi, G., Cifone, M., Collard, D., Fellows, M., Flanders, K., Gollapudi, B., Jenkinson, P., Kirby, P., Kirchner, S., Kraycer, J., McEnaney, S., Muster, W., Myhr, B., O’Donovan, M., Oliver, J., Ouldelhkim, M. C., Pant, K., Preston, R., Riach, C., San, R., Shimada, H., and Stankowski, L. F., Jr. (2002). Mouse lymphoma thymidine kinase gene mutation assay: Follow-up International Workshop on Genotoxicity Test Procedures, New Orleans, Louisiana, April 2000. Environ Mol Mutagen 40, 292–299. Mortelmans, K. E., and Dousman, L. (1986). Mutagenesis and plasmids. In Chemical Mutagens, Principles and Methods for Their Detection, Vol. 10, de Serres, F. J., ed., Plenum Press, New York, pp. 469–508. OECD (2007). Draft proposal for a new guideline 487: In vitro mammalian cell micronucleus test (MNvit). OECD Guideline for the Testing of Chemicals 487, 1–21. Schuler, M., Rupa, D. S., and Eastmond, D. A. (1997). A critical evaluation of centromeric labeling to distinguish micronuclei induced by chromosomal loss and breakage in vitro. Mutat Res 392, 81–95. Snyder, R. D., and Green, J. W. (2001). A review of the genotoxicity of marketed pharmaceuticals. Mutat Res 488, 151–169. Storer, R. D., Kraynak, A. R., McKelvey, T. W., Elia, M. C., Goodrow, T. L., and DeLuca, J. G. (1997). The mouse lymphoma L5178Y Tk+/− cell line is heterozygous for a codon 170 mutation in the p53 tumor suppressor gene. Mutat Res 373, 157–165. Tennant, R. W., Margolin, B. H., Shelby, M. D., Zeiger, E., Haseman, J. K., Spalding, J., Caspary, W., Resnick, M., Stasiewicz, S., Anderson, B., et al. (1987). Prediction of chemical carcinogenicity in rodents from in vitro genetic toxicity assays. Science 236, 933–941. Tice, R. R., Agurell, E., Anderson, D., Burlinson, B., Hartmann, A., Kobayashi, H., Miyamae, Y., Rojas, E., Ryu, J. C., and Sasaki, Y. F. (2000). Single cell gel/comet assay: Guidelines for in vitro and in vivo genetic toxicology testing. Environ Mol Mutagen 35, 206–221. Tweats, D., and Gatehouse, D. (1999). Mutagenicity. In General and Applied Toxicology, Vol. 2, Ballantyne, B., ed., Macmillan Press, London, pp. 1017–1078.

CH A P TE R

12

IN VIVO GENOTOXICITY ASSAYS Véronique Thybaud

12.1.

INTRODUCTION

12.1.1. Endpoints Used for In Vivo Genetic Toxicology Assays Genetic changes affecting cell-cycle control or genome integrity, such as oncogene activation and inactivation of DNA repair or tumor-suppressor genes, are key events in the multistep process of carcinogenesis and are also associated with other human diseases and with aging (Hanahan and Weinberg 2000). In addition, germ-line mutations can lead to inheritable diseases. A variety of in vivo genotoxicity assays (Figure 12.1) have been developed over the last 35 years to detect genotoxicity, and no single assay is currently able to detect all genotoxic agents (MacGregor et al. 2000; Müller et al. 2003). In vivo genotoxicity assays can use either (a) somatic cells for the prediction of cancer and aging or (b) germ cells for inheritable diseases. When a compound is found to be genotoxic in somatic cells and is able to reach germ cells, it is reasonable to conclude that it may also pose a mutagenic hazard to germ cells and thus a risk to future generations (Waters et al. 1993, 1994; Shelby 1996). Genotoxic evaluation on germ cells (OECD 1984, 1986a, 1997c; Bishop and Kodell 1980; Russell and Matter 1980; Russo 2000) can use either germ cells themselves (e.g., spermatogonial cells) or the offspring (or potential offspring) of exposed animals (dominant lethal test and mouse heritable translocation assay). This chapter will mainly focus on (a) the evaluation of genetic toxicity in somatic cells and (b) its use in cancer risk assessment. The ability of chemicals to damage DNA, either directly or after metabolic activation, can be evaluated by detecting covalent DNA binding (for example, by measuring DNA adduct formation) or by detecting single or double DNA strandbreaks, crosslinking, and apurinic sites with methods such as the comet assay. Primary DNA damage (DNA adducts and strand-breaks) is considered a biomarker of exposure because it provides an integrated measurement of compound absorption, metabolic activation, and delivery to target organ macromolecules (Swenberg et al. 2008). Primary DNA damage can be repaired and is not therefore systematically


289

290

CHAPTER 12


Direct Carcinogen

Indirect Carcinogen Metabolic activation

Electrophilic compound/metabolite interactions with DNA Markers of exposure

DNA PRIMARY DAMAGE

Apoptosis Cell death

No DNA repair, errors in repair or proliferation before DNA repair

DNA repair

DNA primary damage assays: • DNA strand breaks (Comet Assay) • DNA adducts Indicator assays: • Unscheduled DNA synthesis • Sister Chromatid Exchanges

Markers of exposure Gene mutation assays: • Endogenous genes • Transgenes Chromosome damage assays: • Chromosome aberrations • Micronuclei

Figure 12.1.

MUTATIONS

Back to normal

TUMOR

End-points detected with the available in vivo genotoxicity assays.

transmitted to daughter cells (Norbury and Hickson 2001; Baute and Depicker 2008; Fousteri and Mullenders 2008; Hegde et al. 2008). Thus, primary DNA damage can also be assessed in indicator assays, such as those detecting increased DNA excision repair or DNA recombination activity [e.g., unscheduled DNA synthesis (UDS) and sister chromatid exchange (SCE)]. If DNA damage is not repaired and/or the cell bearing DNA damage does not undergo apoptosis or death through other mechanisms, then the damage may lead to replication errors and to irreversible changes in DNA structure. These changes are then transmitted to the progeny of the mutated cell and can potentially lead to inheritable genetic changes. Gene mutations and chromosome damage, because they are stable and transmissible genetic changes, are considered as biomarkers of effect (Swenberg et al. 2008). Gene mutation assays measure mutagenicity—that is, the ability of a product to induce point mutations in single genes or blocks of genes resulting from basepair substitutions, frameshifts, and small deletions or insertions. These methods generally rely on reporter genes, which may be endogeneous (e.g., hprt, tk, aprt, Dlb-1, and Pig-a) or transgenes (e.g., lacZ, lacI, and gpt), based on the assumption that a product able to induce mutations in a reporter gene can also provoke mutations in genes involved in the initiation and progression of cancer, such as oncogenes (e.g., ras) and tumor suppressor genes (e.g., p53) (Nestmann et al. 1996; Hemminki et al. 2000). Clastogenicity—the ability to induce structural chromosome damage— is measured by detecting gross chromosome abnormalities (e.g., chromatid breaks, and chromosome rearrangements) with standard cytogenetic methods. Such changes are rarely compatible with cell viability or with transmission to daughter cells. It is assumed that if such abnormalities are detected, then more discrete stable

12.1. INTRODUCTION

291

rearrangements (translocations or small deletions) are also generated (Obe et al. 2002). Changes in chromosome numbers (aneuploidy) can also be associated with cancer and other human diseases (Chi and Jeang 2007). The capacity of the different assays to detect relevant genetic changes, and especially those associated with a risk of cancer, is controversial. Gene mutations and chromosome damage are considered more relevant than primary DNA damage, and it is generally accepted that only gene mutation and chromosome damage assays can establish a genotoxic mode of action (Kasper et al. 2007).

12.1.2. Contribution of In Vivo Genetic Toxicology Assays to Risk Assessment As initially recommended in the 1970s and 1980s (MacGregor et al. 2000), in vivo genotoxicity assays are generally combined with two in vitro genotoxicity tests (a bacterial gene mutation assay and a chromosome damage assay using mammalian cells)(Cimino 2006)(for discussion, see Chapter 11). This is because some genotoxic carcinogens provoke genetic damage in animals, of a nature that cannot readily be detected in vitro (Tweats et al. 2007). In vivo genetic toxicology assays are part of the standard battery of regulatory tests required for the development and registration of products like pharmaceuticals, food additives, and pesticides [for review see Cimino (2006) and Chapter 10]. For other products, such as cosmetics and chemicals, in vivo genetic toxicology assays do not necessarily belong to the minimal battery of regulatory tests, at least in Europe. For ethical reasons, animal testing of the latter products is mainly recommended (1) if in vitro genetic toxicity tests are positive, or (2) in case of “high” or “moderate and sustained” human exposure, or (3) when large amounts of the product are to be released on the market and, potentially, in the environment. In this regulatory context, in vivo genetic toxicity assays contribute to genotoxic hazard identification and to human risk assessment, by predicting carcinogenic activity and inheritable genetic changes. They are also crucial for mechanistic interpretation of 2-year bioassay findings (Kasper et al. 2007). One or more in vivo genetic toxicology assays may be necessary when positive results are obtained in vitro, in order to increase the weight of evidence and to better evaluate the human risk. A major advantage of in vivo genetic toxicology assays over in vitro tests is that they take into account not only intrinsic genotoxic potential, but also toxicokinetic parameters such as bioavailability, absorption, tissue distribution, metabolism (activation, detoxification, and excretion), and other factors that only exist in vivo. Individually, they are generally considered less sensitive (more false-negatives) but more specific (fewer false-positives) than in vitro assays. They are also thought to be more relevant to human exposure, being less prone to experimental artifacts, confounding factors, and irrelevant results. Consequently, a positive result in an in vivo genetic toxicology assay represents strong evidence for genotoxic carcinogenicity. Moreover, in vivo genetic toxicology assays are helpful for understanding and interpreting the results of carcinogenicity studies. In the case of genotoxic carcinogens, gene mutations and chromosome damage are generally necessary but not sufficient for carcinogenesis, because additional key events such

292

CHAPTER 12


as cell proliferation are required for tumor formation. In other words, most genotoxic compounds are carcinogens, especially those showing genotoxic activity in vivo, but not all carcinogens exhibit genotoxic activity: Some have other mechanisms of action (e.g., hormone imbalance, enzymatic induction, cell proliferation, and inhibition of apoptosis). Importantly, genetic changes can also secondarily occur after exposure to nongenotoxic carcinogens as a result of genetic instability (Moore et al. 2008). It is generally agreed that a no-effect dose level can be determined for nongenotoxic carcinogens that act via well-understood thresholded mechanisms (KirschVolders et al. 2003; Müller and Kasper 2000; Pratt and Barron 2003). This notion is not yet fully accepted for genotoxic carcinogens, despite some evidence of thresholds in metabolic activation, repair, and other key mechanisms involved in the formation of gene mutations and chromosome damage, along with data recently reported for ethyl methanesulfonate (Gocke et al. 2009; Gocke and Muller 2009). Therefore, no dose below which no tumorigenic effect would occur can usually be determined for genotoxic carcinogens, unless mechanistic studies demonstrate that (a) the primary target is not DNA itself but other cell components such as proteins and (b) DNA damage occurs secondarily. For example, effects on components of the mitotic apparatus are responsible for chromosome gain or loss (aneuploidy) as a result of mechanisms such as improper attachment of chomosomes to the mitotic spindle, failed cytokinesis, and an abnormal number of mitotic spindle poles (Chi and Jeang 2007). In this case the primary target is generally not DNA but instead multiprotein complexes involved in chromosome segregation (Bharadwaj and Yu 2004; Chi and Jeang 2007). Thus, a no-effect level can be determined and a safety margin can be estimated compared with actual human exposure. Consequently, understanding the mode of action is essential for risk assessment (Dearfield and Moore 2005; Kasper et al. 2007). Numerous in vivo genotoxicity assays have been developed for research purposes. The present chapter does not aim to provide an exhaustive list of all models described in the literature. Rather, it focuses on the main assays using somatic cells, which may be part of the standard battery of genetic toxicology tests or be requested as part of a weight-of-evidence approach, mechanistic investigations, or more accurate human risk assessment. General considerations applicable to all tests are presented first. Then, for each test system, the principles, the protocol design, the advantages and limitations, data interpretation, and regulatory status are discussed.

12.2. PARAMETERS AND CRITERIA FOR VALID IN VIVO GENOTOXICITY ASSAYS AND IMPLICATIONS FOR EXPERIMENTAL DESIGN The general criteria for the selection of parameters used in in vivo genotoxicity assays are listed below and are indicated for each type of assay in Table 12.1. Selection of the Top Dose. In the absence of toxicity, the highest dose tested is generally 2000 mg/kg/day for treatments up to 2 weeks and 1000 mg/kg/day for

12.2. PARAMETERS AND CRITERIA

293

treatments longer than 2 weeks. In the absence of toxicity and genotoxicity, effective tissue exposure should generally be documented. For the most commonly used tissues (e.g., bone marrow and liver), plasma levels are used. No specific measurements are therefore needed to demonstrate tissue exposure. For other tissues, such as site-ofcontact tissues, specific toxicokinetic data may be required. If toxicity is noted, the highest dose should be the maximal tolerated dose, usually defined as “the dose producing signs of toxicity such as higher dose levels, based on the same dosing regimen, would be expected to produce lethality” (Mackay 1995). Observation of cytotoxicity in a given tissue (e.g., impact on erythopoeisis in bone marrow, hepatotoxicity, etc.) can contribute to limiting the top dose. For example, at least a 50% reduction in the mitotic index in the bone marrow chromosome aberration test is required. When used to interpret results obtained in carcinogenicity studies, the doses selected for in vivo genotoxicity assays can be those used in two-year bioassays. Number of Doses. A maximum of three doses are generally required for cytogenetic assays in bone marrow and peripheral blood. In the case of complex and labor-intensive assays (e.g., gene mutation assays in transgenic animals, liver UDS test), two doses might be considered sufficient. In the absence of toxicity, a “limit test” using a single dose of 1000 or 2000 mg/kg/day is considered acceptable for some assays (see OECD guidelines). Duration of Exposure. In most cases, acute treatment with one or two administrations of high doses is recommended. For chronic administration, there is a balance between the duration of treatment and the dose levels, because the doses evaluated during long-term treatment are frequently lower because of more pronounced toxicity. When DNA lesions, mutations, and damaged cells do not accumulate over time (because of efficient damage repair or elimination of damaged cells by apoptosis or cell turnover), longer treatment might result in lower sensitivity. When DNA lesions accumulate over time and damaged cells are not eliminated (e.g., transgenic gene mutation assays in slowly proliferating tissues), multiple administrations may improve assay sensitivity. In vivo genetic toxicity endpoints can be evaluated after multiple administrations in other toxicology studies—for example, 14- or 28-day general toxicity studies. After more than 28 days (e.g., 3-month studies), confounding effects such as oxidative stress, enzymatic induction, and preneoplasic lesions might be expected, leading to the measurement of secondary rather than primary effects on DNA. Genotoxicity findings obtained after long-term exposure should therefore be interpreted with caution. Sampling Time. The optimal sampling time for a given assay in a given tissue should correspond to the likely maximal effect. The time needed for the compound to reach the tissue of interest, for metabolic activation, for the formation and accumulation of DNA lesions and their fixation into mutations, as well as for the possible elimination of damaged cells and cellular turnover, should be taken into account. Sampling times vary from one assay to another: 3–24 hours after treatment in the UDS and comet assays, 24–48 hours for the detection of chromosome damage, and up to several weeks for the measurement of gene mutations.

294

CHAPTER 12


TABLE 12.1. Summary Table of Principle, Study Design, Parameters and Criteria Used in the In Vivo Genotoxicity Assays

In Vivo Assays

Micronucleus Assay in Bone Marrow and Peripheral Blood

Bone Marrow Chromosome Aberration Assay

Comet Assay

Endpoints

Structural and numerical chromosome damage. Request cell proliferation.

Chromosome damage (structural and some indication of potential numerical damage, e.g., polyploidy). Request cell proliferation.

DNA primary damage: DNA single and double strand-breaks, alkali-labile sites, incomplete excision repair sites, DNA-DNA and DNA-protein crosslinks. Does not request cell proliferation.

Parameters measured

Chromosome damage: Incidence of micronucleated polychromatic erythrocytes (PCE) or normochromatic erythrocytes (NCE).

Chromosome damage: Incidence of cells with structural chromosome aberrations including and excluding gaps (including number and type of aberrations), as well as polyploid cells and cells with endoreduplicated chromosomes.

DNA migration: In case of image analysis: tail length, % tail DNA, tail moment. In case of manual scoring: tail length, or incidence of cells with and without DNA migration, from undamaged to highly damaged (∼4 categories).

Regulatory acceptance

Protocol described in OECD 474 (OECD 1997c).

Protocol described in OECD 475 (OECD 1997b).

No OECD guideline available. International validation ongoing.

Part of standard battery and/or of follow-up testing

Standard battery, or first follow-up when in vivo test are not part of the minimal battery of test.

Standard battery. Equally acceptable alternative to the in vivo bone marrow micronucleus test.

Follow-up testing.


DNA Adducts

Unscheduled DNA Synthesis Assay in Liver Cells

Sister-Chromatid Exchange Assay

Repair of DNA lesions and unscheduled DNA synthesis in response to DNA primary damage. Does not request cell proliferation.

Repair of DNA lesions by homologous recombination (interchange between sister chromatids). Request cell proliferation.

Net nuclear grain count (nuclear grain count minus cytoplasm grain count). Proportion of cells “in repair.”

Incidence of sister chromatid exchange (SCE) per cell.

No OECD guideline available.

Protocol described in OECD 486 (OECD 1997d).

Follow-up testing.

Follow-up testing.

No OECD guideline available for in vivo SCE, OECD guideline only available for in vitro SCE (OECD 1997a). Follow-up testing.

DNA primary damage: DNA adducts (alkyl and bulky adducts) are nucleotide bases covalently modified by reactive electrophilic chemical intermediates or free radicals. Does not request cell proliferation. Number of adducts per normal nucleotides. Chemical structure of adducts, only for some methods. For more details see Table 12.2 for the in vivo measurement of DNA adducts.

295

Gene Mutation Assays Gene mutations: Point mutations such as base pair substitutions, frame shifts, small deletions or insertions. Request DNA replication and cell proliferation for fixation of DNA primary damage into stable gene mutations. Mutant frequency (e.g., number of mutants per million cells). Mutation spectrum for some assays (e.g., transgenic models, hprt). For more details see Table 12.3 for In vivo gene mutation assays in transgenic models, and Table 12.4 for in vivo gene mutation assays in endogenous genes of somatic cells. No OECD guideline available. OECD detailed review paper recently released. Follow-up testing.

(Continued)

296

CHAPTER 12

TABLE 12.1.


(Continued)

In Vivo Assays



Comet Assay

Practicability

Widely used. Easy cell sampling and preparation. Easy to score and possible automation of micronucleated cell scoring (image analysis or flow cytometry).

Less widely used than the bone marrow micronucleus test. Easy cell sampling and preparation. Time-consuming and tedious scoring of chromosome aberrations.

Relatively widely used. Still under validation. Easy cell/nuclei sampling and preparation. Easy to score and possible automation of comet scoring.

Integration in general toxicity

Possible if criteria for top dose selection acceptable.

Possible if criteria for top dose selection acceptable. Administration of mitotic inhibitor shortly before tissue sampling might impact the measurement of other parameters.

Sampling times not compatible with general toxicity studies. Would require extra treatment 2–6 hours before sampling.

Selection of top dose

Maximal tolerated dose, or a dose inducing bone marrow cytotoxicity, or in the absence of toxicity 1000 mg/kg if more than 14 days of treatment; and 2000 mg/kg/day if less than 14 days of treatment.

Maximal tolerated dose, or a dose inducing bone marrow cytotoxicity (at least 50% reduction in mitotic index), or in the absence of toxicity 1000 mg/kg if more than 14 days of treatment; and 2000 mg/kg/day if less than 14 days of treatment.

Maximal tolerated dose, or a dose inducing cytotoxicity in the selected tissue, or in the absence of toxicity 1000 mg/ kg if more than 14 days of treatment; and 2000 mg/kg/ day if less than 14 days of treatment.


DNA Adducts Rarely used. Different methods available, some being more straightforward than the others (for more details see Table 12.2).

Possible for methods that do not request animal treatment with radio-labeled compounds.

Maximal tolerated dose, or maximal dose used in 2-year bioassay, if study conducted as a follow-up of carcinogenicity studies.

Unscheduled DNA Synthesis Assay in Liver Cells Rarely used. Labor-intensive. Treatment and sampling times not compatible with normal working day. Possible automation of grain scoring. Require handling of radio-labeled thymidine, and appropriate authorization. Sampling times not compatible with general toxicity studies. Extra treatment 2–6 hours before sampling, and liver perfusion required. Maximal tolerated dose, or a dose inducing liver cytotoxicity (e.g., pyknotic nuclei) or 2000 mg/kg/day.


297

Gene Mutation Assays

Rarely used. Easy cell sampling and preparation, as for chromosome aberrations. Relatively easy to score.

Rarely used. Generally labour intensive. Different methods available, some being more straightforward than the others (for more details see Table 12.3 and 12.4).

Possible if criteria for top dose selection acceptable. Administration of mitotic inhibitor and thymidine analogue shortly before tissue sampling might impact the measurement of other parameters. Maximal tolerated dose, or a dose inducing cytotoxicity in the selected tissue.

Possible for a few endogenous genes. Most often specific strain and/or study design requested.

Maximal tolerated dose, or maximal dose used in 2-year bioassay, if study conducted as a follow-up of carcinogenicity studies.

(Continued)

298

CHAPTER 12

TABLE 12.1.


(Continued)

In Vivo Assays



Comet Assay

Number of doses

3 (only 1 acceptable, in absence of toxicity, i.e., limit test using 1000 or 2000 mg/kg).


2 (limit test not acceptable).

Duration of exposure

Single or multiple treatments.

Single or multiple treatments.

One or two administration. Multiple administrations under evaluation.

Sampling time

In bone marrow, generally two samplings after a single administration (24 and 48 hours after treatment) or one sampling, 24 hours after the last treatment in case of multiple administrations. Sampling should not be before 24 hours and after 48 hours. In peripheral blood, generally 48-hour sampling time Sampling time should not be before 36 and after 72 hours.

One sampling time, 12–18 hours (i.e., 1.5 cell cycle) after a single or last administration. One additional sampling 24 hours later is optional.

Two sampling times 2–6 and 16–26 hours after treatment. Multiple administrations: 2–6 hours after the last treatment, under evaluation.


DNA Adducts No clear recommendations. As dose–response is generally linear at low doses, an evidence of dose response is an important confirmation of a positive response. Single or multiple administrations.

No clear recommendations. Should take place before repair and adduct removal after single administration and at steady state after repeated administrations (e.g., 10 days to 1–2 months).



299



Generally three doses.

No clear recommendations. Two to three for transgenic models.

Single administration.

Single or multiple administrations.

Two sampling times: 12–16, and 2–4 hours after single administration.

Depends on the cell turn over in the evaluated tissue, generally second cell cycle after the last treatment.

Depends on the model and cell turn over in the evaluated tissue, e.g., 28 days recommended for any tissue in transgenic models. Depends on the model and cell turn over in the evaluated tissue, e.g., for transgenic models, 3 and 28 days after the last administration are recommended.

(Continued)

300

CHAPTER 12

TABLE 12.1.


(Continued)

In Vivo Assays



Comet Assay

Species

Mainly rodents, potentially applicable to nonrodent species, including human.


Applicable to rodent and nonrodent species, including human.

Genders

Males are sufficient if no gender difference anticipated.


One gender is sufficient if no gender difference anticipated.

Tissue(s)

Mainly bone marrow and peripheral blood. Other tissues described in the literature. 5

Mainly bone marrow. Other tissues described in the literature.

Any tissue/organ from which cell/ nuclei can be properly isolated.

5

4–5

2000 for chromosome damage. 200 and 1000 for cytotoxicity, in bone marrow and peripheral blodd, respectively.

100 for chromosome damage, 1000 for cytotoxicity

100–150 cells, preferably 150 (depends on the number of animals per group).

Number of animal per group

Number of cells per animal


DNA Adducts



Applicable to rodent and nonrodent species, including human.

Mainly rodents, potentially applicable to nonrodent species.


No clear recommendations. One gender should be sufficient if no gender difference anticipated. Any tissue/organ from which DNA can be properly isolated.


No clear recommendations. One gender should be sufficient if no gender difference anticipated.

Mainly liver cells, but other tissues described in the literature.

Any dividing tissue/ organ from which cell suspensions can be properly isolated.

No clear recommendations, depends on the statistical power of the method used for the detection.

3

No clear recommendations, depends on the statistical power of the method used for the detection. Generally at least three.

Not appropriate.

100 liver cells.

Generally 25 to 50.

301

Gene Mutation Assays Mainly rodents, especially for transgenic models. Endogenous genes potentially applicable to nonrodent species, including human. No clear recommendations. One gender should be sufficient if no gender difference anticipated. Single or limited number of tissues (most of endogenous genes). Any tissue (transgenic models). No clear recommendations, depends on the statistical power of the method used for the detection. From about 5 for most models to ∼50 dams and 100’s offsprings for spot tests. Depends on the models. Not appropriate when DNA is extracted.

302

CHAPTER 12


Species. In vivo genotoxicity assays mainly use rodents (rats and mice), because they are small (standard animal facilities and low amount of test article needed) and because abundant toxicologic and metabolic data are available on these species. Most assays are also theoretically applicable to nonrodent animals, but limited data have been reported in the literature. Most of the assays are applicable to wild-type and readily available animal strains. It should nevertheless be noted that a few models require specific strains (e.g., some gene mutation assays). For follow-up testing, the choice of species should take previous data into consideration (e.g., organ-specific generation of DNA-reactive metabolites, tumor findings, etc.). Selection of Gender. Unless clear sex differences (in e.g., metabolism, toxicity, pharmacological activity, two-year bioassay findings, gender specificity) are anticipated, only one gender needs to be tested. Males are generally considered most sensitive for genotoxicity studies. Selection of Tissues. The in vivo genotoxicity assay required for the standard battery is generally a chromosome damage test performed with erythrocytes from bone marrow or peripheral blood. Other in vivo models are mostly used as complementary or follow-up tests. They can also be used as first-line tests if considered appropriate for the compound, in terms of its known properties, metabolism, exposure, target organ, and endpoints. For example, when the compound is known to be chemically unstable, contact tissues might be preferred (e.g., skin for dermal application, lung for inhalation, and gastrointestinal tract for ingested compounds). Similarly, if the compound is metabolized into toxic species, the liver might be preferred. Depending on the assay limitations, genotoxic endpoints can be evaluated either in a restricted number of tissues (e.g., micronucleus and UDS assays) or in almost any tissue (e.g., comet assay, transgenic gene mutation models). Moreover, some assays are dependent on cell proliferation status: The UDS and comet assays can be conducted with nonproliferating cells such as liver cells soon after exposure, while cell proliferation is needed for chromosome damage and gene mutation tests. Depending on the tissues of interest, in vivo assays can be used to detect primary DNA damage, gene mutations, and/or chromosome damage. When in vivo assays are conducted as a follow-up to positive 2-year bioassays, the selected tissues are preferably those in which the tumors arose. In other cases, genotoxicity is evaluated in surrogate tissues. Number of Animals. For statistical reasons, about five to ten animals per dose group are generally used, except in the case of labor-intensive tests where a minimum of three (e.g., for the UDS test in liver cells) or four animals is recommended. A much larger number of animals must be used in some assays (e.g., up to 50 dams treated and several hundred F1 animals examined in mouse spot tests). For ethical reasons, such assays are seldom used. Animals should be evaluated individually,

12.3. IN VIVO GENOTOXICITY ASSAYS REQUIRED IN THE STANDARD BATTERY OF TESTS

303

and pooled samples should be avoided, if possible. Both individual and mean group values are usually reported. Negative and Positive Control Animals. Negative control animals should only receive the vehicle. If an unknown “exotic” vehicle is used, an absolute negative control group of animals receiving no treatment might be necessary. A positive control group of animals treated with a well-known genotoxic carcinogen, preferably requiring metabolic activation to express its genotoxic activity, is generally used. While positive control groups are mandatory for new assays and for laboratories with limited experience, it is increasingly recommended, for ethical reasons, to generate positive control samples every 6–12 months (for example) and to use them for different studies (e.g., DNA samples for DNA adducts, slides for chromosome damage, comet, and UDS assays). A study is considered valid if the results obtained with positive and negative controls are consistent with the laboratory’s historical data and with the literature. Statistical analysis is usually applied to compare treated and negative control groups. Both pairwise and linear trend tests can be used. Because of the low background and Poisson distribution, data transformation (e.g., log transformation) is sometimes needed before using tests applicable to normally distributed data. Otherwise, nonparametric analyses should be preferred. The 3Rs. The three Rs—Reduce, Refine, and possibly Replace the use of animals—are increasingly being taken into consideration before initiating in vivo studies, including genotoxicity assays. It is recommended (1) to evaluate only one gender unless gender differences are anticipated, (2) not to include positive control groups in all studies (see above), (3) to avoid assays requiring large numbers of animals when alternatives exist, (4) to evaluate multiple genotoxicity endpoints in a single animal whenever possible, and (5) to integrate genotoxicity studies with other toxicology studies (organ toxicity, etc.). In specific cases, it may also be advisable to only rely on data obtained from in vitro assays.

12.3. IN VIVO GENOTOXICITY ASSAYS REQUIRED IN THE STANDARD BATTERY OF TESTS Given that the genotoxicity of some compounds can only be detected in vivo, in vivo genotoxicity assays—generally those able to detect chromosome damage in bone marrow or peripheral blood—are often recommended in the standard battery as a complement to in vitro genotoxicity tests (Brambilla and Martelli 2004; Cimino 2006). They either directly quantify and analyze different types of chromosome aberrations in metaphase cells (chromatid and chromosome deletions and exchanges) or indirectly measure the induction of chromosome damage by scoring micronuclei resulting from chromosome breaks, chromosome rearragements, and chromosome lagging (Mateuca et al. 2006). The two methods are considered equally acceptable and interchangeable (Shelby and Witt 1995).

304

CHAPTER 12

12.3.1.


Mammalian Erythrocyte Micronucleus Test

Purpose. The purpose of the in vivo micronucleus test is to evaluate the potential of the test substance to cause chromosomal damage (clastogenicity) or damage to the mitotic apparatus (aneugenicity) by the analysis of micronuclei in erythrocytes sampled in bone marrow and/or in peripheral blood of experimental animals (usually rodents) (Schmid 1975; Heddle 1973; Heddle and Salamone 1981). Regulatory Acceptance. The rodent erythrocyte micronucleus assay is relatively easy to conduct and is considered able to detect the vast majority of clastogens and rodent carcinogens (when combined with the Ames test) and all human genotoxic carcinogens (Shelby and Zeiger 1990; Rosenkranz and Cunningham 2000). This test is well-validated and widely accepted by regulatory agencies as part of the standard battery of genetic toxicity assays, in addition to the two in vitro assays (Brambilla and Martelli 2004; Cimino 2006). Because of the mechanism of micronuclei formation, the micronucleus test does not in principle detect gene mutations. It should therefore be considered complementary to in vitro gene mutation assays, and not as a follow-up assay to confirm in vitro gene mutations. The experimental conditions and data interpretation are described in OECD guideline 474. It is the most commonly used in vivo assay. Many laboratories are highly experienced, and a large database has been generated for comparison with carcinogenicity and other effects. Principle. During the anaphase of mitosis, acentric chromosome fragments and/ or unseparated chromosomes lag and fail to become incorporated into daughter cell nuclei. After telophase, most of these fragments and/or lagging chromosomes are not included in the nuclei of the daughter cells, but condense to form one or several micronuclei (smooth-boundaried bodies that stain strongly and specifically for chromatin, one-fifth to one-twentieth the size of the main nucleus, also named Howell– Jolly bodies in hematology). During erythropoiesis, when the erythroblast develops into an immature or polychromatic erythrocyte (PCE), the main nucleus is expelled while micronuclei are retained in the cells, facilitating their detection (Heddle et al. 1991). Typically, micronucleus induction is measured in bone marrow PCEs after acute treatment. After repeated administration, it is also advisable to evaluate the induction of micronuclei in mature or normochromatic erythrocytes (NCEs). As erythrocytes persist for about one month in peripheral blood (named reticulocytes), the measurement of micronucleated reticulocytes in peripheral blood is considered equally acceptable in any species, provided that the spleen does not remove the micronucleated erythrocytes from blood and that both aneugens and clastogens are efficiently detected (see more in “assays limitations and confounding factors” section). Micronuclei are generally analyzed in the youngest (i.e., immature) reticulocytes in peripheral blood. Because bone marrow is a relatively well-perfused tissue, its exposure to systemically distributed compounds is generally adequate and can be extrapolated from the plasma concentration. Moreover, the high rate of cell proliferation during erythropoeisis in bone marrow facilitates the formation and detection of micronuclei.


305

Study Design. The sampling time should take into account the interval between the last mitosis in erythroblasts and the formation of polychromatic erythrocytes (i.e., about 6–8 hours) and the lifespan of polychromatic erythrocytes (i.e., 18–24 hours) (Mavournin et al. 1990). Following a single administration, the incidence of micronucleated PCEs is generally scored at 24 and 48 hours after the administration (not earlier than 24 hours or later than 48 hours) in bone marrow and at 36 and 72 hours in peripheral blood (not earlier than 36 hours or later than 72 hours). After multiple administrations (2 or more), the bone marrow should be sampled at least once between 18 and 24 hours, and peripheral blood should be sampled between 36 and 48 hours after the last administration. In the acute version of the bone marrow micronucleus assay, the bone marrow is generally sampled 24 hours after two administrations 24 hours apart in order to save animals (only one group of animals per dose) without impacting the sensitivity (Salamone et al. 1980; Ashby et al. 1985; CSGMT 1990). The incidence of micronucleated PCEs in bone marrow and in peripheral blood (also termed reticulocytes) should be evaluated among at least 2000 erythrocytes per animal, in order to take into account the low spontaneous rate of micronucleated erythrocytes (0–3/4 micronucleated PCEs per thousand PCEs) and to ensure adequate statistical power (Adler 1984; Hayashi et al. 1994). After continuous treatment for 4 weeks or more, micronuclei should also be analyzed in bone marrow NCEs among at least 2000 erythrocytes per animal. Data suggest that the in vivo micronucleus assay can be integrated into a 28-day toxicological study. For the detection of a majority genotoxic compounds, the incidence of micronucleated erythrocytes should be determined both 4 and 28 days after the beginning of treatment in the peripheral blood and at 28-day necropsy time for bone marrow (Hamada et al. 2001). In addition, a reduction in the ratio of PCEs to NCEs is usually considered to indicate inhibition of erythroblast proliferation or maturation, or destruction of nucleated cells. The cytotoxic effect on bone marrow is therefore measured as PCE/ (PCE + NCE) by counting a total of at least 200 erythrocytes in bone marrow and 1000 erythrocytes in peripheral blood per animal. The distinction between polychromatic and normochromatic erythrocytes is based on the presence of RNA in PCEs, and it is visualized under the microscope after differential labeling with a nonfluorescent stain such as Giemsa (blue for PCEs and pink for NCEs), or a fluorescent stain such as acridine orange (e.g., orange fluorescence in PCEs and no staining in NCEs) (Krishna and Hayashi 2000). For reticulocytes, acridine orange supravital staining is used to distinguish the youngest reticulocytes (i.e., type I and II reticulocytes) based on RNA content (RNA disappears from reticulocytes older than 3 days) (MacGregor et al. 1980, 1987; Hayashi et al. 1990). Immediately after sacrifice, bone marrow cells are collected from the femurs or tibias, spread on slides and fixed. Similarly, slides can be prepared from blood samples. Slides are then stained (as briefly described above) and visually scored under the microscope. Because manual scoring is time-consuming, the scoring can also be done automatically with validated methods (Hayashi et al. 2007). Bone marrow micronucleated cells can be analyzed by image analysis following cellulose column separation of rodent bone marrow samples to remove nucleated cells (Romagna and Staniforth 1989; Frieauff and Romagna 1994) and to make the

306

CHAPTER 12


automatic scoring more efficient. Automatic scoring of micronuclei in bone marrow erythrocytes and peripheral blood reticulocytes can also be done by flow cytometry using fluorescein-conjugated monoclonal antibodies against the transferrin receptor (anti-CD71-FITC) to stain polychromatic erythocytes and young (immature) reticulocytes, and propidium iodide with RNase treatment is used to identify micronuclei (Dertinger et al. 1997; Hayashi et al. 2000; MacGregor et al. 2006; Weaver and Torous 2000). A three-color labeling method, using antiplatelet-PE antibody in addition to anti-CD71-FITC and propodium iodide, was recently developed to improve the flow cytometry method, especially for human blood samples (Dertinger et al. 2004). Both automatic methods result in more objective measurements, allow more cells to be analyzed than with microscopic scoring (up to 10-fold more in the case of flow cytometry), and improve the statistical power (Torous et al. 2000, 2003, 2005; MacGregor et al. 2006; Kissling et al. 2007). The flow cytometric procedure was first developed for the analysis of micronuclei in mouse peripheral blood reticulocytes (Dertinger et al. 1996, 1997) before being applied to rat, dog, monkey, and human reticulocytes, and rodent bone marrow (Torous et al. 2000; Dertinger et al. 2002; MacGregor et al. 2006; Harper et al. 2007; Hotchkiss et al. 2008). The main advantage of the micronucleus test in peripheral blood is the possibility of using small volumes of blood (a few microliters) obtained during other studies such as general toxicity studies, in any species. Interpretation. A significant increase in the number of micronucleated polychromatic erythrocytes or young reticulocytes is usually considered indicative of structural and/or numerical chromosome damage caused by exposure to a clastogenic and/or aneugenic substance. In order to distinguish between clastogenic and aneugenic effects and to improve the risk assessment, it can be useful to conduct additional mechanistic investigations. Because the formation of micronuclei containing whole chromosomes results from an impact on the mitotic apparatus and not from a direct effect on DNA, this effect is considered to exhibit a threshold dose–response (Aardema et al. 1998), while this is generally not considered to be the case of micronuclei containing acentric fragments resulting from direct interaction with DNA. Whole chromosomes can be detected in micronuclei by using specific kinetochore antibodies and immunofluorescent CREST staining, or pancentromeric DNA probes and fluorescence in situ hybridization (FISH) (Iarmarcovai et al. 2006). A few authors have recommended the use of flow cytometry because, as noted previsouly, it increases the statistical power and allows better assessment of low doses (Grawé et al. 1998; Asano et al. 2006). Assay Limitations and Confounding Factors. In mice, the spleen does not destroy circulating micronucleated erythocytes, and the results obtained in bone marrow and peripheral blood are similar. In other species, and in particularly rats, the spleen efficiently removes circulating micronucleated reticulocytes, especially those induced by aneugens, owing to the large size of the micronuclei (Cammerer et al. 2007a,b). The sensitivity of the micronucleus test on rat peripheral blood has been reported to be lower than on rat bone marrow, from both a qualitative and


307

quantitative point of view (Wakata et al. 1998). Moreover, published data on the effect of aneugens on rat reticulocytes are conflicting. Recent data show that aneugens such as colchicine, vincristine, and vinblastin are barely detected after a single administration (Cammerer et al. 2007a), while colchicine was properly detected after five consecutive administrations (Cammerer et al. 2007b; MacGregor et al. 2006). Numerous studies are being conducted to further evaluate this problematic and controversial issue (Hayashi et al. 2007). Recently, the scoring of large numbers of reticulocytes with flow cytometry was described as being able to compensate for the low rate of micronucleated cells in peripheral blood resulting from spleen removal of micronucleated erythrocytes in some species, including rats (Torous et al. 2000; Witt et al. 2008). Provided that the ability of the flow cytometric approach to solve the sensitivity issue is confirmed, the micronucleus test on peripheral blood reticulocytes represents a promising noninvasive in vivo assay for the detection of chromosome damage in any species, including rat (MacGregor et al. 2006), dog (Harper et al. 2007), monkey (Hotchkiss et al. 2008), and human (Dertinger et al. 2002). Some toxic compounds (e.g., mitomycin C and dimethylhydrazine-2HCl), while they clearly increase the incidence of micronucleated erythrocytes after acute treatment, were not readily detected after multiple administrations, when integrated in general toxicity studies because the doses reached are much lower doses (Hamada et al. 2001). Even if the bone marrow is a well-perfused tissue, chemically unstable compounds and/or metabolites may not reach it in sufficient quantities to induce detectable effects (Brambilla and Martelli 2004; Morita et al. 1997). In a large collaborative study of IARC carcinogens (groups 1, 2A, and 2B), the in vivo bone marrow micronucleus test easily detected compounds able to induce tumors in hematopoietic tissues and lung. In contrast, it predicted only 40% of liver carcinogens (Morita et al. 1997). To solve this issue, chromosome damage, especially micronuclei, can be measured in tissues other than bone marrow and peripheral blood erythrocytes (Hayashi et al. 2007). The measurement of micronuclei on lymphocytes (from spleen and peripheral blood) consists of animal exposure, cell isolation, in vitro lymphocyte stimulation and micronucleus evaluation (Ren et al. 1991; Benning et al. 1992, 1994). The in vivo micronucleus test has also been developed in skin (Nishikawa et al. 1999, 2002) and gastrointestinal tract (e.g., colon) (Vanhauwaert et al. 2001; Ohyama et al. 2002). When evaluated in liver, cell proliferation is provoked by partial hepatectomy (Tates et al. 1980; Cliet et al. 1989; Igarashi and Shimada 1997) or by treatment with products able to induce cell proliferation such as 4-acetlyaminofluorene (Braithwaite and Ashby 1988) and carbon tetrachloride. Recently, a hepatocyte micronucleus assay in young rats was reported. Because hepatocytes are still able to proliferate in young animals, no mitogen stimulus is required. Moreover, young animals have a metabolic capacity similar to that of adults (Suzuki et al. 2004; Hayashi et al. 2007). Despite promising results, these models present technical difficulties (e.g., the need to induce liver cell proliferation) and are only used in specific cases (Hayashi et al. 2007). Confounding factors can lead to irrelevant findings in the in vivo micronucleus test on bone marrow [for review, see Tweats et al. (2007)]. They include changes

308

CHAPTER 12


in core body temperature, because hypothermia and hyperthermia can disrupt chromosome binding to mitotic spindles and cause chromosome loss. Similarly, an increase in erythropoeisis, as a result of toxicity for erythroblasts or through direct stimulation of cell division (e.g., after bleeding, hemolysis, or erythropoietin production), can enhance the incidence of micronucleated erythrocytes. It is suggested that acceleration of erythroblast maturation and proliferation can lead to errors in erythrocyte enucleation (expulsion of the main nucleus from erythroblasts) or differentiation, as well as errors in genetic repair processes, resulting in higher rates of micronucleated cells (Tweats et al. 2007). Induction of apoptosis can also be a confounding factor, but it is generally easy to recognize because the micronuclei are much more numerous or pyknotic as compared to those induced by clastogens and aneugens.

12.3.2.

Bone Marrow Chromosome Aberration Test

Purpose. The purpose of the in vivo chromosome aberration test is to evaluate the potential of the test substance to cause chromosomal aberrations in bone marrow cells of experimental animals (usually rodents) (Tice et al. 1994; Preston et al. 1987). Regulatory Acceptance. The rodent erythrocyte chromosome aberration assay is well-validated and widely accepted by regulatory agencies as part of the standard battery of genetic toxicity assays, in addition to the two in vitro assays. It is considered as an equally acceptable alternative to the in vivo micronucleus test on rodent erythrocytes (Shelby and Witt 1995). The experimental conditions and data interpretation have been published in OECD guideline 475. Nevertheless, it is less widely and commonly used than the in vivo micronucleus test, because it is less simple. Chromosome aberration scoring is more tedious and time-consuming than micronucleus scoring, and requires skilled and experienced personnel. Principle. Chromosome aberrations are scored in bone marrow cells in first metaphase after compound administration—that is, 1.5 normal cell cycles. In order to take into account possible delays in absorption, metabolism, and cell cycling, bone marrow cells can also be collected 24 hours after the first sample. The structural aberrations are classified into two types (i.e., chromatid and chromosome aberrations) and three different categories (i.e., gaps, deletions, and rearrangements or exchanges) (Preston et al. 1987; Tice et al. 1994). Even if this test is not specifically designed to detect aneuploidy, an increased incidence of polyploid cells and of cells with endoreduplicated chromosomes suggests a potential to induce numerical chromosome aberrations. As in the in vivo micronucleus test, bone marrow is preferred because it is a well-perfused and rapidly dividing tissue. Study Design. Bone marrow cells are sampled 1.5 normal cell cycles (12–18 hours) after compound administration, and they are sampled optionally 24 hours later in additional groups of animals. In case of multiple administrations, cells are sampled 1.5 normal cell cycles (12–18 hours) after the last treatment. In order to accumulate cells in metaphase and make the scoring easier, the animals receive a


309

metaphase arresting agent (e.g., colchicine) by the intraperitoneal route 3–5 hours before sampling, depending on the species (OECD 1997b). Immediately after sacrifice, bone marrow cells are collected from femurs or tibias, exposed to hypotonic treatment, fixed, spread on slides, and stained. The slides are then scored visually under the microscope. Structural chromosome aberrations should be analyzed in at least 100 metaphase cells per animal. To this end, for each cell, the number and type of aberrations (chromatid or chromosome breaks and gaps, as well as the different types of chromatid and chromosome rearrangements) should be recorded. The incidence of polyploid cells and of cells with endoreduplicated chromosomes should also be reported because they potentially reflect the induction of numerical chromosome aberrations and/or inhibition of cell cycle progression. In addition, the mitotic index, used as a cytotoxicity parameter, is calculated for 1000 cells per animal. With cytotoxic compounds, the highest dose level should induce at least a 50% reduction in the mitotic index (OECD 1997b; Richold et al. 1990; Tice et al. 1994). Interpretation. A significant increase in the number of cells with structural chromosome aberrations, excluding gaps (gaps are reported separately), is usually considered as indicative of structural chromosome damage caused by exposure to a clastogenic agent. Moreover, an increased incidence of polyploid cells or cells with endoreduplicated chromosomes suggests that the compound is potentially an aneugen (Kirsch-Volders et al. 2002). The biological significance of polyploidy and endoreduplication (i.e., DNA replication without cell division) is controversial (Storchova and Pellman 2004). Formation of polyploid cells in normal tissues is far from negligible in nature. Moreover, polyploid cells that possess more than two sets of homologous chromosomes are generated by various mechanisms (endoreduplication, cell fusion, abortive cell cycle, mitotic slippage, and cytokinesis failure) in case of cellular stress, aging, and diseases, because they are thought to confer a metabolic advantage. While polyploid cells are normally blocked in G1 cell cycle arrest or eliminated by apoptosis in the case of cells bearing functional ploidy-sensing (e.g., p53, and Rb proteins) checkpoints, aneuploidy can arise from genetically instable tetraploid intermediates (Ganem et al. 2007). Finally, polyploidy cells are formed during stress conditions and therefore do not always reflect genotoxicity (Storchova and Kuffer 2008). In order to confirm the compound ability to provoke aneuploidy, as well as to improve the risk assessment, it can be useful to conduct an additional mechanistic evaluation and/or an in vivo micronucleus test including the presence of kinetochores in micronuclei (Aardema et al. 1998; Kirsch-Volders et al. 2002). Assay Limitations and Confounding Factors. Owing to the relatively small number of cells evaluated (100 cells per animal) and the low background (0–5% of cells with chromosome aberrations), the statistical power is lower than with the in vivo micronucleus test (Adler 1984). Moreover, chromosomes can be lost during metaphase spread. Therefore, aneuploidy cannot be directly assessed by counting the number of chromosomes per cell, but only by looking at polyploid cells and cells with endoreduplicated chromosomes (Aardema et al. 1998; Kirsch-Volders et al. 2002). As described above, not all compounds with such effects are aneugens. Some

310

CHAPTER 12


might only impact cell division, without inducing chromosome loss or nondisjunction. Therefore, while the in vivo chromosome aberration test is considered equally acceptable for the detection of structural aberrations, it is less sensitive than the in vivo micronucleus test for the detection of numerical changes. This issue is nevertheless offset by the ability of this test to reveal the type of structural aberrations and to provide mechanistic information. As in the bone marrow micronucleus test, chemically unstable compounds and/or metabolites that cannot reach the bone marrow in sufficient quantities to induce detectable effects cannot induce chromosome aberrations in the bone marrow and are not appropriately detected in this test. Chromosome aberration can also be measured in peripheral blood or spleen lymphocytes after acute or chronic administrations (e.g., as part of general toxicity studies). Because lymphocytes remain quiescent in G0 cell-cycle phase for about one month in rodent blood and spleen, they have the opportunity to accumulate DNA lesions. Chromosome aberrations are analyzed after an in vitro 48-hour stimulation of lymphocyte proliferation by mitogenic compounds [e.g., phytohemagglutinin or concanavalin A (Kligerman et al. 1984)] and addition of a metaphase arresting agent (e.g., colcemid) for the last 2–3 hours. S-dependent genotoxins would not be detected with this method, as the compound is not in contact with the cells during the S-phase.

12.4. IN VIVO GENOTOXICITY ASSAYS USED MAINLY AS COMPLEMENTARY OR FOLLOW-UP TESTS More than one in vivo genotoxicity assay is generally required when positive results are obtained in in vitro assays and 2-year bioassays (Cimino 2006; Kasper et al. 2007). In case of positive results in in vitro assays, the second assay is generally conducted in a different tissue (e.g., the most exposed tissue, a target organ for toxicity, or a tissue with a high capacity for metabolic activation). The endpoint can be primary DNA damage (in order to further evaluate DNA reactivity) and/or the endpoint found to be impacted in in vitro assays and not yet evaluated in vivo (i.e., gene mutations), in order to better evaluate the relevance of the in vitro findings. The most promising approach to determining whether tumors found in specific rodent tissues are attributable to genotoxic events is the assessment of genotoxicity in cancer target tissues (Kasper et al. 2007; Kirkland and Speit 2008; Lambert et al. 2005). Genotoxicity assays applicable to any tissue comprise rodent transgenic mutation assays, the in vivo comet assay, and determination of DNA adducts; other models such as the liver UDS and micronucleus assay in tissues other than bone marrow and peripheral blood are also valuable, but they are restricted to one or a limited number of tissues. The following section presents (1) primary DNA damage assays that detect the ability of compounds to interact with DNA and to cause primary DNA damage such as DNA strand-breaks (comet assay) and DNA adducts, (2) indicator assays (UDS and SCE assays) that reveal DNA repair or recombination in response to DNA lesions, and (3) gene mutation assays.

12.4. IN VIVO GENOTOXICITY ASSAYS USED MAINLY AS COMPLEMENTARY OR FOLLOW-UP TESTS

12.4.1.

311

The Comet Assay

Purpose. The purpose of the comet assay is to evaluate the potential of the test substance to induce primary DNA damage—that is, DNA strand-breaks in treated animals, usually rodents (Ostling and Johanson 1984; Singh et al. 1988; Olive et al. 1990a,b; 1991; Olive 2002; Collins 2002). It is applicable to any tissue from which a sufficient amount of cells or nuclei can be isolated without damaging DNA or triggering DNA repair processes. Regulatory Acceptance. The comet assay is widely accepted by regulatory agencies, even though it was only relatively recently developed and is still being validated. No OECD guidelines are yet available, but several publications provide internationally agreed protocol recommendations (Tice et al. 2000; Hartmann et al. 2003; Burlinson et al. 2007). The standard protocol for regulatory purposes is currently being refined. These validation exercises and collaborative efforts are currently coordinated by the Japanese Center for the Validation of Alternative Methods or JaCVAM. The comet assay is mentioned or clearly recommended as a complementary assay in the vast majority of regulatory documents: (1) in case of positive in vitro results not confirmed in the in vivo bone marrow micronucleus test, in order to improve the weight of evidence and risk assessment, (2) in case of negative results in the standard battery of genotoxicity tests and tumor findings in 2-year bioassays, in order to better understand the mechanisms responsible for carcinogenesis, to confirm the absence of a genotoxic impact in a target organ of carcinogenicity, and to allow the compound to be classified as a nongenotoxic carcinogen provided that the nongenotoxic mechanism is elucidated (Kasper et al. 2007), and (3) to evaluate the genotoxic impact in the first contact tissues in the case of (a) poorly absorbed compounds giving little or no systemic exposure, and (b) chemically unstable short-lived compounds and metabolites. Examples of application are given in Brendler-Schwaab et al. (2005) and Hartmann et al. (2004). Principle. The comet assay (or single-cell gel electrophoresis assay) is a rapid and simple method for the detection of DNA breakage in mammalian cells. The recommended and most commonly used method, the alkaline version (pH > 13) of the comet assay (Singh et al. 1988, 1994; Burlinson et al. 2007), detects in individual cells double strand-breaks, single strand-breaks (including those resulting from alkali-labile sites and incomplete excision repair), and both DNA–DNA and DNA– protein crosslinks. Thus, the comet assay detects most types of primary DNA damage that could later become fixed as either gene mutations (e.g., resulting from alkali-labile or abasic sites and damage not eliminated by excision repair), or structural chromosome damage (e.g., owing to DNA strand-breaks). Study Design. The compound is administered by the most appropriate route to at least four to five animals per group and at a minimum of two doses (MTD and 25–50% MTD). After a single administration, the tissues are sampled respectively 2–6 hours (preferably 3) and 16–26 hours (preferably 21) after treatment, in order

312

CHAPTER 12


to detect the effects of (a) rapidly absorbed, unstable and directly acting compounds and (b) compounds requiring more time for absorption, distribution, and metabolic activation. After multiple treatments (2 or more) at 24-hour intervals, the tissues/ organs should be collected once, 2–6 hours (preferably 3) after the last administration (Hartmann et al. 2003). The cells are isolated from solid tissues by using digestive enzymes (trypsin or collagenase), by brief mincing with scissors, or by pushing the tissue through a mesh (Brendler-Schwaab et al. 1994). The nuclei are obtained by tissue mincing and homogenization (Miyamae et al. 1998; Sasaki et al. 1997a–d). During cell or nuclei isolation, EDTA and radical scavengers (e.g., dimethylsulfoxide) can be added to prevent degradation by endonucleases and oxidative DNA damage. The isolated cells or nuclei are embedded in agarose gel and layered on microscopic slides. The slides are incubated first in lysis buffer containing detergents and high salt concentrations in order to release the DNA. The next steps are DNA unwinding in the presence of alkaline buffer in order to produce single-stranded DNA and to express the DNA alkali-labile sites as single strand-breaks, followed by electrophoresis in alkaline conditions. At the end of the electrophoresis step, the slides are neutralized, dried, and stained with fluorescent dyes (propidium iodide, ethidium bromide, SYGR green, or yoyo-1) in order to visualize the DNA. When examined under the microscope, the cells look like comets, with (a) a head corresponding to undamaged DNA in the nuclear region and (b) a tail containing DNA strands and fragments or loops (Shaposhnikov et al. 2008) resulting from DNA breakages, which have migrated in the direction of the anode. DNA migration should be measured with 100–150 cells per animal (Lovell et al. 1999; Wiklund and Agurell 2003) on two to three different slides by manual scoring or by using an interactive or fully automated image analysis systems (Böcker et al. 1999; Frieauff et al. 2001; Schunck et al. 2004; Dehon et al. 2008). The experimental conditions should allow some DNA migration in vehicle controls in order to detect a delay in DNA migration after treatment with crosslinking agents (Hartmann et al. 2003). Because cell death, necrosis, and apoptosis can lead to DNA fragmentation and to possibly irrelevant findings, cytotoxicity should be evaluated, especially in case of a positive result. First, comets with small or nonexistent heads and large diffuse tails (Fairbairn et al. 1996; Olive et al. 1993), named “hedgehogs,” “ghost cells,” “clouds,” or “nondetectable nuclei cells,” are either excluded from the analysis or scored separately as potentially apoptotic/necrotic cells. The presence of low-molecular-weight DNA fragments in apoptotic/necrotic cells can be evaluated by omitting the electrophoresis step after alkaline unwinding, in a neutral diffusion assay (Tice et al. 2000). Finally, it is strongly recommended to collect samples of tissues/organs for histopathological examination (Burlinson et al. 2007). Interpretation. Manual scoring generally quantifies the distance of DNA migration, the percentage of cells with and without DNA migration, and/or the proportion of cells in four to five different categories, from undamaged to heavily damaged cells (Miyamae et al. 1998). When the scoring is done by image analysis, in addition to the tail length, the percentage of DNA in the tail (% tail DNA) is also measured. The product of the tail length and the % tail DNA, named the tail moment, is also calculated. The values of the three parameters (tail length, % tail DNA, and tail


313

moment) should be reported for each individual cell, along with the mean and/or median values for each animal and treatment group. The % tail DNA is the preferred parameter, because it is less dependent than tail length on the technology used, is linear with respect to the dose, and is more easily reproduced between laboratories (Dehon et al. 2008). Moreover, the distribution of migration among cells from each animal and group are useful for data interpretation (Olive and Durand 2005; Burlinson et al. 2007). An increase in DNA migration parameters indicates that the test substance has induced DNA strand-breaks, while a decrease suggests DNA–DNA and DNA– protein crosslinks. Advantages. The main advantage of the comet assay is that it allows simple, rapid (hours to days after sampling), and cost-effective DNA damage evaluation. It has the added advantage of detecting low levels of DNA damage in single cells of any organ (Burlinson et al. 2007; Anderson et al. 1998). Its ability to visualize and quantify DNA strand-breaks in individual cells is clearly seen as an advantage as compared to alkaline elution method measuring breaks in DNA from a cell pool (Kohn and Grimek-Ewig 1973; Kohn et al. 1976). It does not require cell division, contrary to the in vivo chromosome aberration and micronucleus tests, or manipulation of radiolabeled compounds as required by the UDS test and some DNA adduct assays. Moreover, in contrast to the micronucleus, chromosome aberration, and UDS tests, mainly conducted with bone marrow and liver, the comet assay is applicable to any tissue and can be performed on a limited number of cells or nuclei. The ability to measure DNA strand-breaks in site-of-contact tissues is particularly important in case of low systemic exposure and chemically unstable compounds. The tissue selection should take into consideration all available information on structural analogues, absorption, distribution, metabolism, excretion, and/or toxicology. When no information is available, one or preferably two tissues should be examined: the liver for orally absorbed compounds and a site of first contact tissue—that is, the gastrointestinal tract, respiratory tract, and skin for the oral route, inhalation, and dermal application, respectively (Hartmann et al. 2003). Tremendous amounts of data have already been obtained with the comet assay in vivo (McKelvey-Martin et al. 1993; Fairbairn et al. 1995; Burlinson et al. 2007; Rojas et al. 1999; Kirkland and Speit 2008). Some publications focus on its ability to detect organ specificity (Burlinson et al. 2007). In general, these publications confirm the ability of the comet assay to detect DNA damage in carcinogenicity target tissues, when the appropriate route of administration is used—for example, when first site-of-contact tissues (stomach, skin, and basal mucosa) are examined. Other authors assessed a large number of compounds on a selected set of tissues (stomach, colon, liver, kidney, urinary bladder, lung, brain, and bone marrow, using intraperitoneal injection or oral gavage) in order to assess its sensitivity and specificity for the detection of carcinogens. These data, summarized by Sasaki et al. (2000), show that among 208 compounds selected from IARC monographs and the US NTP Carcinogenicity Database, 94% of rodent genotoxic carcinogens were positive in the mouse comet assay and 80% of rodent noncarcinogens were negative (Sasaki et al. 2000). Moreover, 91% of the carcinogens that did not induce

314

CHAPTER 12


micronuclei in hematopoietic cells in bone marrow and/or peripheral blood were positive in the comet assay. The authors also pointed out that DNA damage was detected not only in almost all carcinogenicity target organs, but also in nontarget organs. Therefore, the comet assay can be used to confirm the induction of DNA damage in target organs, but not to predict where tumors will occur. When DNA migration is evaluated at the two recommended sampling times, 62% of genotoxic carcinogens are detected in liver while 69% are detected in colon and stomach. The detection rate is vastly improved when liver and stomach (86%) or liver and colon are combined (87%). Limitations and Confounding Factors. The standard alkaline method does not differentiate between the different types of damage (single and double strand-breaks, alkali-labile and incomplete excision repair sites). Additional evaluations and/or modifications are needed to obtain mechanistic information. These include (1) comparing strand-breaks obtained with the alkaline and neutral versions of the protocol to potentially differentiate double from single strand-break induction, (2) using DNA repair enzymes such as UV-specific endonucleases, endonuclease III for oxidized pyrimidines, or formamido pyrimidine glycosylase (FPG) for 8-hydroxyguanines (Collins et al. 1993, 2008) to provoke specific breaks, and (3) combining the comet assay with DNA-damage-specific antibodies revealed by immunofluorescence or with chromosome painting probes measured by fluorescence in situ hybridization (COMET-FISH) to visualize DNA damage and specific genomic regions in the comet (Santos et al. 1997; Sauvaigo et al. 1998; Rapp et al. 2005). It should also be emphasized that rapidly repaired damage can be missed unless early sampling times and high doses, at which the repair capacity is generally overwhelmed, are appropriately selected. One of the main potential confounding factors is the formation of DNA strandbreaks through cell death, necrosis, or apoptosis. It is therefore important to confirm the absence of toxicity in the case of positive results (Tice et al. 2000; Burlinson et al. 2007), using the neutral diffusion assay and histological tissue evaluation. Another confounding effect is the induction of DNA strand-breaks through indirect mechanisms, such as production of free radicals in the case of enzymatic induction, inflammation, and preneoplastic lesions. Therefore, positive results obtained after long-term administration, during which tissue remodeling may have taken place, should be considered with caution.

12.4.2.

DNA Adducts

Purpose. DNA adducts are nucleotide bases (i.e., purines and pyrimidines) that have been covalently modified by reactive electrophilic chemical intermediates or free radicals. The chemical structures of DNA adducts are diverse and vary from simple alkyl adducts induced by alkylating agents to complex bulky adducts such as those resulting from metabolic activation of polycyclic aromatic hydrocarbons, aromatic amines, and aflatoxins (Dipple 1995; Chiarelli and Jackson 1992; Rundle 2006; Xue and Warshawsky 2005). The purpose of measuring DNA adducts is to determine whether a DNA-reactive compound or a metabolically activated


315

intermediate can covalently bind to nucleophilic sites of DNA nucleotide bases. If replication takes place before DNA adducts are removed—for example, through base and nucleotide excision repair or dealkylation mechanisms—or if DNA adducts are misrepaired, they can be fixed into point mutations (Hemminki et al. 2000) or lead to chromosome damage. For many years, DNA adducts have been used as biomarkers of exposure: If measured shortly after treatment, they provide an integrated marker of compound intake (absorption and distribution), metabolic activation, delivery to target macromolecules in target tissues, and interaction with DNA (Farmer 2004a; Koc and Swenberg 2002; Swenberg et al. 2008; Farmer and Singh 2008). The level of adducts measured at a given time point also depends on the formation and stability of electrophilic entities, adduct stability and DNA repair, and tissue turnover (i.e., cell proliferation and cell death, including apoptosis). DNA damage such as DNA adducts is considered to be necessary but not sufficient for tumorigenesis (Poirier et al. 2000). Mutagenesis and cell proliferation must also take place. However, DNA adducts are considered to be early key events in carcinogenesis induced by genotoxic carcinogens. Their measurement contributes to understanding the metabolism and action of carcinogens through molecular dosimetry across species, including humans [e.g., aflatoxins, tamoxifen (Gamboa da Costa et al. 2003), benzo[a]pyrene (Beland et al. 2005), and 2-amino-3,8-dimethylimidazo [4,5-f]quinoxaline (MeIQx) (Mauthe et al. 1999)], and provides useful information for cancer risk assessment (Poirier and Beland 1992; Weston 1993). Regulatory Acceptance. The detection of DNA adducts is not part of the standard battery of tests, but can be recommended as a follow-up investigation, in the case of positive results in in vitro genetic toxicity assays and negative results in the bone marrow chromosome damage test, as well as in the case of negative results in the standard battery of genotoxicity tests and tumor findings in 2-year bioassay (Reddy 2000; Phillips et al. 2000). In the latter case, DNA adducts are sought in the target organ of carcinogenicity. No guidelines are available, but several publications provide protocol recommendations [for general review, see Farmer (2004b), Farmer and Singh (2008), Hemminki et al. (2000), Phillips et al. (2000), Poirier et al. (2000), Reddy (2000), Singh and Farmer (2006), Garner (1998)]. Principle. After animal treatment with single or multiple administrations, depending on the method used for DNA adduct detection and the purpose of the study, DNA is isolated from the tissue(s) of interest. DNA is preferably analyzed in tissue from individual animals and is pooled only if necessary to make up the required quantity (Phillips et al. 2000). The measurement of DNA adducts generally consists of four main steps: (1) DNA isolation from the tissue or organ of interest, (2) DNA hydrolysis or digestion, (3) DNA adduct enrichment and isolation, and (4) DNA adduct analysis with or without the addition of a standard. Qualitative and quantitative analyses of DNA adducts have gradually improved over the last 40 years. In the past two decades, significant efforts have been made to elucidate the chemical structures of DNA adducts by using chemically specific techniques such as mass, fluorescence, and nuclear magnetic resonance spectrometry (Poirier 2004). The main methods for the detection and analysis of DNA adducts are, from the oldest to the

316

CHAPTER 12


most recent: (1) administration of radiolabeled compounds and measurement of radioactive decay by scintillation counting, (2) immunoassays using antibodies to carcinogen-modified DNA, also used to locate DNA adducts in tissues by immunohistochemistry, (3) 32P-postlabeling (PPL), (4) high-performance liquid chromatography (HPLC) combined with physicochemical detection methods such as fluorescence or electrochemical detection, (5) various chromatographic techniques with mass spectrometric (MS) detection, and (6) administration of radiolabeled compounds and measurement of isotope ratios with accelerator mass spectrometry (AMS). A brief description of the principle, study design, sensitivity, specificity, limitations, and strengths of the different techniques is given in Table 12.2. Interpretation. The biological significance of DNA adducts is still controversial (Nestmann et al. 1996; Phillips et al. 2000). The formation of DNA adducts is considered to be a key event in human cancer formation. However, it has been clearly demonstrated only for a limited number of compounds—for example, aflatoxin and aromatic hydrocarbons (see examples in Poirier et al., 2000). Comparison of DNA adduct formation and tumorigenesis after chronic administration of genotoxic carcinogens to animals showed that steady-state DNA adduct levels are generally reached after 1–2 months of chronic administration (Poirier et al. 2000). No tumors were observed in the absence of DNA adducts, but, on the other hand, the presence of DNA adducts was not always synonymous with tumor formation, suggesting either that a threshold level of DNA adducts is needed or that other key events such as cell proliferation are necessary for tumour formation. Ottender and Lutz (1999)

TABLE 12.2.

In Vivo Measurement of DNA Adducts

Methods Used for the Measurement of DNA Adducts References

Radiolabeling Method Coupled with Liquid Scintillation Counting Baird (1979), Buss et al. (1990), Lutz (1979, 1986), Martin et al. (1993), Phillips et al. (2000), Reddy (2000), Swenberg et al. (2008).

Immunoassays Müller et al. (1982), Müller and Rajewsky (1980, 1981), Den Engelse et al. (1990), Farmer (2004a,b), Hsu et al. (1981), Kriek et al. (1984), Phillips etal. (2000), Poirier et al. (2000), Poirier and Beland (1992), Poirier (1981, 1993, 2004), Reddy (2000), Santella (1999), Strickland and Boyle (1984), Wild (1990).

32

P-Postlabeling Assay

Farmer (2004a,b), Gupta et al. (1982), Phillips (1997), Phillips and Arlt (2007), Phillips and Castegnaro (1999), Phillips et al. (2000, 2005), Poirier et al. (2000), Randerath et al. (1981), Randerath and Randerath (1994), Reddy (2000), Reddy and Randerath (1986), Reddy et al. (1984), Shibutani et al. (2006), Terashima et al. (2002), Whong et al. (1992).


317

concluded from data obtained with 27 carcinogens that a 50% tumor incidence rate was associated with 53–5543 adducts per 108 nucleotides, depending on the compound. Evaluation of the risk associated with DNA adducts depends on the chemical nature, quantity, and stability of the adducts, the rate of cell proliferation and adduct fixation into stable mutations, adduct mutagenic and repair efficiencies, and the extent of changes to critical genes (Nestmann et al. 1996). Some DNA adducts, including minor adducts, are associated with mutagenesis and tumorigenesis, while others are not (Yuspa and Poirier 1988; Hemminki et al. 2000). Current data, obtained with highly sensitive methods able to detect one adduct among 1010–1012 nucleotides, suggest that DNA adduct formation is linear at low doses. There may thus be adduct levels (e.g., 1 adduct among 1010 normal nucleotides) at which the risk of mutations and tumors is indistinguishable from the background risk (Nestmann et al. 1996). Therefore, adducts are mainly used as a marker of exposure, given that the DNA binding of many compounds is linear over the dose range. Because DNA binding efficiency does not strictly correlate with the incidence of mutations and tumors, the detection of DNA adducts does not necessarily predict (a) tumorigenicity for a given tissue or (b) the human cancer risk (Hemminki et al. 2000). Thus, DNA adducts should be interpreted in view of other in vivo endpoints—that is, stable mutations and carcinogenicity (Nestmann et al. 1996). Assay Limitations and Advantages. Each method mentioned above has its strengths and weaknesses (see Table 12.2) that impact its specificity, sensitivity, cost, and practicality. The radiolabeling method coupled with liquid scintillation

High-Performance Liquid Chromatography with UV, Fluorescence, or Electrochemical Detection

Mass Spectrometry Coupled with Liquid Chromatography, Gas Chromatography or Capillary Electrophoresis

Kriek et al. (1984), Farmer (2004a,b), Poirier et al. (2000), Poirier (2004), Reddy (2000), Weston et al. (1989), Weston (1993).

Beland et al. (2005), Chiarelli and Jackson (1992), Farmer (2004a,b), Farmer et al. (2005), Gamboa da Costa et al. (2003), Koc and Swenberg (2002), Phillips et al. (2000), Poirier (2004), Poirier et al. (2000), Reddy (2000), Singh and Farmer (2006).

Radiolabeling Method Associated with Accelerator Mass Spectrometry Dingley et al. (1998, 2005), Farmer (2004a,b), Farmer et al. (2005), Garner (1998), Goldman et al. (2000), Mauthe et al. (1999), Phillips et al. (2000), Poirier (2004), Poirier et al. (2000), Reddy (2000), Tompkins et al. (2006), Turteltaub and Dingley (1998).

(Continued)

318 TABLE 12.2.

CHAPTER 12

(Continued)

Methods Used for the Measurement of DNA Adducts Principle


Radiolabeling Method Coupled with Liquid Scintillation Counting Measurement of radioactive decay after single administration of [14C]labeled or [3H]-labeled compounds. A few hours to days after exposure, DNA is extracted from tissue of interest and purified. Increase in radioactivity, measured by liquid scintillation counting, as compared to control DNA is considered to reflect DNA adduct formation. Further characterization can be done on hydrolyzed DNA after isolation of modified nucleotides by highperformance liquid chromatography (HPLC).

Immunoassays DNA extraction from the tissue of interest after compound administration. Competitive or direct immunoassays are conducted and inhibition of antibody binding is measured using antibodies against DNA adducts or modified DNA obtained from immunized rabbit. Changes in antibody binding are indicative of the presence of DNA adducts.

32


DNA extraction from the tissue of interest after compound administration followed by (1) enzymatic digestion of DNA in 3′-monophosphates of normal and adducted nucleotides, (2) optional enrichment step to select or isolate adducted nucleotides (typically butanol extraction or nuclease P1 treatment), (3) radio-labeling of the adducts by incorporation of 32P-orthophosphate at nucleotides 5′-end using a polynucleotide kinase and [γ32P]-ATP, (4) separation of DNA adducts using multidirectional thin-layer chromatography (TLC) on polyethyleneimine (PEI) cellulose, HPLC or electrophoretic separation, (5) radioactivity quantification by measurement of radioactive decay using autoradiography or electronic imaging. Intensity changes in background spots and observation of additional spots are indicative of a positive response.


319



Radiolabeling Method Associated with Accelerator Mass Spectrometry

DNA extraction from the tissue of interest after compound administration. High-performance liquid chromatography (HPLC) separation of DNA adducts, followed by fluorescence or electrochemical detection and quantification. Observation of additional peaks/signals is indicative of the presence of adducts.

DNA isolation from any tissue of interest after compound administration followed by (1) addition of stable isotopelabeled internal standard for quantification, (2) hydrolysis/ digestion of DNA, (3) enrichment of DNA adducts of interest (e.g., solid-phase extraction, immunoaffinity chromatography, or DNA repair enzymes), (4) quantification by mass spectrometry (MS) using gas chromatography–MS (after derivization) with electron impact ionisation, or liquid chromatography–MS using electrospray or other ionspray capillary electrophoresis interfaces. Tandem MS–MS is often used to increase the specificity of the assay and is the favored technique. A signal at the correct chromatographic retention time with intensity greater than a specified signal-to-noise ratio is indicative of the presence of DNA adducts.

Measurement of radioactive decay (accelerated mass spectrometry) after single administration of [14C]-labeled or [3H]-labeled compounds (other isotopes also applicable). After exposure (hours to days), DNA is extracted from tissue of interest and purified. Incorporated radioactivity, is measured by separation of isotope based on mass followed by quantification in gas ionization detector to measure isotope ratio. Another alternative is the use of postlabeling, i.e., digestion of DNA to nucleotides, depurination, HPLC separation to isolate adducts, [14C]-acetylation reaction, HPLC to collect [14C]-labeled adducts, and AMS analysis. Increase in radioactivity, measured by AMS, as compared to control DNA is considered to reflect the presence of adduct.

(Continued)

320 TABLE 12.2.

CHAPTER 12


(Continued)

Methods Used for the Measurement of DNA Adducts

Radiolabeling Method Coupled with Liquid Scintillation Counting

Immunoassays

Sensitivity

Typically 1 adduct per 106 nucleotides to as low as 1 adduct per 109 nucleotides.

Typically 1 adduct out 106 nucleotides. Depends on the method used for the detection of antiserum bound in microtiter plates, i.e., colorimetry, fluorescence, or chemiluminescence. Could be as low as 1 adduct per 108–109 nucleotides with chemiluminescence.

Specificity

Limited specificity as no information on chemical structure of the adducts.

Limitations

Difficult and expensive synthesis of quite large amount of radio-labeled compounds with high and stable specific activity, i.e., millicurie amounts per animal needed. Multidose treatment is difficult, if not impossible, because of the cost and difficulty to use large volume of radiolabeled compounds. Need to ensure that the location of labeling is resistant to loss during metabolism and adduct formation. Need to verify that radioactivity is not due to contamination or metabolism, i.e., to avoid possible artifacts, HPLC of DNA hydrolysates and digestion with proteases and ribonucleases can be used.

Potentially highly specific. Depends on antibody preparation, and specificity. Preparation of large quantities of specific antibodies is needed. Relatively large amounts of DNA (up to 100 μg) required. Information on DNA adduct structure is needed for the preparation of specific antibodies, when highly specific antibodies are prepared. Therefore not applicable for unknown compounds, unless whole damaged DNA is used for antiserum preparation.

32


Highly sensitive: Generally as low as 1 adduct in 1010 nucleotides, when butanol extraction or nuclease P1 treatment are applied. Very efficient for bulky adducts, and N7alkylguanines and other positively charged adducts, less suitable for other nonaromatic, small or depurinating adducts (1 adduct out of 105–106 nucleotides). Sensitivity can be improved by about one order of magnitude using HPLC combined with radioisotope detector after TLC or PAGE. Moderate specificity, as it does not provide information on chemical structure of the adducts. Use of high quantities of 32 P radioactivity (25–50 μCi per sample) with very high specific activity. Relatively labor intensive and low throughput (completed in 3 days), except with PAGE method that allows the concomitant analysis of a large number of samples in parallel within a few hours. The efficiency of the enrichment and labelling steps varies according to the adduct structures and can result in adduct loss or low sensitivity for some types of adducts.




321

Radiolabeling Method Associated with Accelerator Mass Spectrometry

As low as 1 adduct per 107–108 nucleotides.

As low as 1 adduct per 108–109 nucleotides. Sensitivity is increasing with new equipments.

Extremely sensitive: As low as 1 adduct per 1011–1012 nucleotides following administration of [14C]-labeled compounds, i.e., less than one adduct per cell. Lower sensitivity generally obtained when [3H]-labeled compounds are used.

Highly specific.

Very high specificity and accurate information on chemical structure of adducts.

Limited as it provides no information on the chemical structure of the adducts.

Only applicable to adducts chemically characterized possessing fluorophore (e.g., polycyclic aromatic hydrocarbons) or electrochemically active groups (e.g., some oxidative DNA lesions). Relatively large amounts of DNA (up to 100 μg) needed.

Inability to screen unknown mixtures. High purity of DNA is required to avoid artefacts due to protein and RNA. Relatively large amounts of DNA (up to 100 μg) needed.

Very expensive instrumentation and limited number of laboratories equipped. Requests administration of radio-labeled compounds, unless post-labeling is applied. Requires synthesis of radio-labeled compounds ([14C]-labeled or [3H]-labeled compounds). Artifacts can result from metabolism and DNA contamination by proteins and RNA, and crosscontamination between samples. Further work is needed to verify that isotopes are integrated in DNA adducts.

(Continued)

322

CHAPTER 12


TABLE 12.2. (Continued) Methods Used for the Measurement of DNA Adducts Strengths

Radiolabeling Method Coupled with Liquid Scintillation Counting Simple and straightforward method as compared to the other techniques.

Immunoassays No need for radiolabeled compounds. Useful for highthroughput analysis (e.g., ELISA) of specific adducts, because cost effective and relatively easy to conduct. Antibodies also used for immunohistochemistry in order to get information on morphological localization of adducts, and for immunoaffinity chromatography used for preliminary steps of DNA adduct enrichment before any other techniques.

32


Radio-labeled compounds or specific antibodies are not requested. Limited amount of DNA needed (less than 10 μg DNA). Ability to detect adducts from chemicals with various structures, e.g., polycyclic aromatic hydrocarbons, aromatic amines, heterocyclic amines, small aromatic compounds, alkylating agents, unsaturated aldehydes from lipid peroxidation, reactive oxygen species and UV radiation. Attempts to define a standardized protocol. Applicable after multiple administrations and for complex mixtures.

counting and AMS has the drawback of involving radioactive manipulations. AMS is the only method able to detect DNA adducts after exposure to very low doses (Dingley et al. 2005; Brown et al. 2005). It also requires much lower isotope doses than liquid scintillation counting. This is a clear advantage for risk assessment, because it eliminates the need to extrapolate from high dose levels (Turteltaub and Dingley 1998). Moreover, the very low isotope dose levels required for AMS make the method applicable in humans (Turteltaub and Dingley 1998; Mauthe et al. 1999; Farmer et al. 2005). Immunoassays (in the case of specific antibodies) and methods associated with physicochemical detection methods such as fluorescence or electrochemical detection can only be used if the structure of the DNA adducts is known. Similarly, depending on the chemical nature of the adducts, different enrichment methods are used in PPL such as further enzymatic digestion with nuclease P1 for aromatic hydrocarbon-like bulky adducts, immunoaffinity chromatography with specific antibodies, and anion-exchange column chromatography for methyl- and ethylsubstituted compounds bearing a positive charge. There is no PPL method available for other simple alkylated adducts. Moreover, false-negative results or underestimation of adduct levels can result from the use of an inappropriate method. Indeed, the efficiency of the enrichment and labeling steps varies according to the adduct




Radio-labeled compounds or specific antibodies are not requested.

The most promising method for the detection of DNA adducts, especially as the new equipments are less expensive and more easy to use, and more the method readily automated. No need for radio-labeled compounds or specific antibodies. Use of chemical-specific stable isotope internal standards ensures very accurate quantification, and confirmation with certainty of the chemical nature of adducts. Possible association with 32 P-postlabeling and immunochemical methodologies. Detection of very different chemicals, from bulky adducts to modified DNA bases.

323

Radiolabeling Method Associated with Accelerator Mass Spectrometry Quantification of DNA adducts in samples at very low exposure levels, and after administration of very low quantities of isotopes/ radiolabeled compounds (as low as 1 μCi/40 kBq). Measurement independent of radioactive decay. Being improved by postlabeling techniques, i.e., incorporating 14 C into specific DNA adducts after formation.

structure, and this can result in adduct loss or in poor sensitivity for some types of adduct (e.g., adducts formed at N7 positions of purines). Because PPL protocols may underestimate adduct levels, several methods (i.e., two enrichment methods) should be used in the case of unknown adducts, and preliminary studies with standards are therefore very important (Whong et al. 1992). AMS, immunoassays, HPLC combined with fluorescence, and mass spectrometry using stable isotope-labeled internal standards can accurately quantify DNA adducts (Koc and Swenberg 2002). PPL and AMS are extremely sensitive (detecting 1 adduct per 1012 nucleotides) but lack specificity, and they provide no information on the chemical nature of the adducts. MS is less sensitive than PPL or AMS but is highly specific and can provide structural information (Farmer and Singh 2008). In addition, mass spectrometers are becoming cheaper and simpler to use. MS methods are also improving with the development of high-resolution mass spectrometry and tandem MS, for example, and are expected to become the most specific and sensitive approaches in the near future. All this means that MS is playing an increasingly important role in DNA adduct analysis (Farmer and Singh 2008; Farmer et al. 2005; Koc and Swenberg 2002; Singh and Farmer 2006). The available methods have rarely been compared but appear to give similar qualitative results (Weston et al. 1989; Eide et al. 1999).

324

CHAPTER 12


The main artifacts and possible confounding effects result from DNA contamination, lack of purity, and a lack of specificity of certain steps of DNA adduct preparation or analysis (Phillips et al. 2000). Recommendations for DNA isolation and storage have been made (Phillips and Castegnaro 1999). Endogenous DNA adducts and adducts formed by nongenotoxic carcinogens (e.g., peroxisome proliferators and estrogens) can sometimes interfere with the detection of DNA adducts, especially when highly sensitive methods (e.g., PPL) are used.

12.4.3.

Unscheduled DNA Synthesis Test in Liver Cells

Purpose. The purpose of the ex vivo unscheduled DNA synthesis (UDS) test on mammalian liver cells is to evaluate the potential of the test substance to induce DNA excision repair in liver cells of treated animals (usually rodents and preferably rats). An increase in UDS activity is indicative of primary DNA damage and subsequent excision repair (Butterworth et al. 1987; Mirsalis and Butterworth 1980; Mirsalis et al. 1982). Regulatory Acceptance. The UDS test on mammalian liver cells is wellvalidated and widely accepted by regulatory agencies. The optimal experimental conditions and rules for data interpretation have been published in OECD guideline 486. The UDS test has never been widely used because it is relatively timeconsuming (in vivo and in vitro experimental steps, and treatment and sampling times that do not fit with usual working day). In addition, it requires the use of radiolabeled compounds with relatively high specific activity and, therefore, special authorization. For many years, this assay has been recommended as an in vivo follow-up assay, in the case of positive results in the in vitro genetic toxicity assays and negative results in the bone marrow chromosome damage test, or findings in 2-year bioassays, especially in liver. Its potential lack of sensitivity has been highlighted, and other, more recent in vivo assays (e.g., the comet assay) are now preferred (Kirkland and Speit 2008). Principle. During nucleotide excision repair (NER), a stretch of 20–30 nucleotides (up to 100 nucleotides) containing the DNA damage (e.g., bulky adducts) is removed and replaced; but during base excision repair (BER), only 1 to about 10 nucleotides, including the modified nucleotide, are excised (Shuck et al. 2008; Fousteri and Mullenders 2008; Baute and Depicker 2008; Hegde et al. 2008). The UDS test detects the incorporation of radiolabeled thymidine during DNA resynthesis in the excised region (Butterworth et al. 1987; Mirsalis and Butterworth 1980; Mirsalis et al. 1982). Liver is the preferred tissue for UDS measurement because it is well-perfused and the main metabolic site for absorbed compounds (during first-pass of compounds administered by the oral and intraperitoneal routes). Moreover, it is a slow-dividing tissue, and only a small proportion of cells undergo replicative DNA synthesis. Therefore DNA synthesis in most liver cells is limited to DNA repair (Butterworth et al. 1987; Mirsalis and Butterworth 1980).


325

Study Design. This is an ex vivo assay. The compound is administered once, to at least three animals per group, then the liver cells are isolated and UDS activity is measured in primary culture. The amount of tritiated thymidine incorporated into liver cell DNA is measured in individual cells by autoradiography; liquid scintillation is not recommended. Respectively 2–4 and 12–16 hours after treatment, the liver cells are isolated by liver perfusion with collagenase, in order to detect the effects of (a) rapidly absorbed and/or metabolized compounds (e.g., methyl methanesulfonate, and dimethyl nitrosamine) and (b) compounds requiring more complex metabolic activation (e.g., 2-acetylamidofluorene). Amphlett et al. (1996) showed that in order to save animals and simplify the assay the animals can be treated twice, 2–4 and 12–16 hours, before sampling. After liver cell attachment to culture dishes, the cells are incubated for 3–8 hours with tritiated thymidine. At the end of the incubation period, which can be followed by a cold chase (incubation with unlabeled thymidine), the cells are fixed, dipped into radiographic emulsion, and kept in the dark for 7–14 days. Grains are counted over the nuclei (nuclear grain count) and cytoplasm (cytoplasm grain count), in 100 cells per animal. The average net nuclear grain (NNG) count (cytoplasmic grain count subtracted from nuclear grain count) and the percentage of cells undergoing repair (increase in the NNG value over spontaneous background) are then calculated. The grains are counted with an interactive or fully automated image analysis system. Cells undergoing DNA replication are easily identified (completely covered by grains) and are excluded from the analysis, because they contain extremely large numbers of grains over the nucleus. Interpretation. Statistical analysis is not generally used, and each laboratory should consider the distribution of its negative control values to define an NNG cutoff value for positive results. In general, it is considered that negative controls always have NNG values lower than zero. An increase in the mean NNG value over the threshold value (usually zero) obtained for a given animal and for a given dose group is indicative of enhanced DNA repair activity (Hamilton and Mirsalis 1987). The proportion of cells undergoing repair (i.e., with NNG values over the threshold) is also taken into consideration. Therefore, a compound that increases the average NNG value above zero, as well as the number of cells undergoing repair, would be considered positive in this test. Assay Limitations. The UDS test is only an indirect measurement of primary DNA damage. It does not provide information on the nature of the primary damage or on the fidelity of repair. Moreover, its sensitivity depends on the number of nucleotides removed and replaced and, thus, the amount of tritiated thymidine incorporated into DNA. Compounds inducing bulky adducts and long-patch repair via NER are generally more efficiently detected by the UDS test than those removed by short-patch repair or BER. Moreover, the UDS assay can only detect DNA damage repaired via the excision repair process: Single strand-breaks and oxidative base damage, for example, are not detected. Like tests done with bone marrow, the liver UDS test has limited value for the detection of labile direct-acting compounds, which cannot readily reach the liver.

326

CHAPTER 12


Site-of-contact tissues might be preferable for such compounds (Burlinson 1989; Furihata et al. 1984; Furihata and Matsushima 1987; Mori et al. 1999; Sawyer et al. 1988), if sufficiently validated. The main technical limitation of this assay with respect to other tissues is the need to isolate the cells after in vivo treatment and to get them to incorporate tritiated thymidine in vitro. Because UDS measurement does not require cell division, it can potentially be applied to many different tissues, provided that the cells can be isolated and maintained in primary culture for the few hours required for tritiated thymidine incorporation. The literature contains reports of UDS-based studies of stomach, colon, kidney, pancreas, tracheal epithelium, nasal epithelium, epidermis, keratinocytes, and spermatocytes (Burlinson 1989; Furihata et al. 1984; Furihata and Matsushima 1987; Sawyer et al. 1988; Loury et al. 1987; Mori et al. 1999; Latt et al. 1981; Helleday 2003).

12.4.4.


Purpose. The purpose of the sister chromatid exchange (SCE) assay is to evaluate the potential of the test substance to induce repair of DNA lesions by homologous recombination in cells of treated animals (potentially all species, usually rodents) (Latt et al. 1981; Helleday 2003). It can easily be applied to any dividing tissue, such as bone marrow and peripheral blood, from which cell suspensions can be isolated and analyzed. Regulatory Acceptance. A regulatory OECD guideline (476) describes the protocol of the in vitro SCE assay conducted with mammalian cells, but no OECD guideline exists for the in vivo assay. Though popular (especially for biomonitoring) and widely used to detect exposure to mutagens and carcinogens (Perry and Evans 1975; Latt 1981; Tucker et al. 1993), the use of SCE data for risk assessment is more controversial because the mechanism of SCE formation and the biological significance of the increased incidence of SCE are not fully elucidated (Latt 1981; Tucker et al. 1993; Wilson and Thompson 2007). Principle. SCE consist of an interchange of DNA replication products and parental DNA strands between two sister chromatids at homologous sites. They require DNA breakage and rejoining steps (Wilson and Thompson 2007; Latt 1981; Latt and Schreck 1980). They generally do not alter chromosome morphology. SCE are conservative and error-free end-products of homologous recombination associated with the repair of persisting single strand breaks. They do not occur during double strand-break repair by homologous recombination, which mostly results in gene conversion, deletions, and tandem duplications (Helleday 2003; Wilson and Thompson 2007). SCE take place during S-phase of the cell cycle and can therefore only be evaluated in actively dividing cells. Any agent or mechanism that stalls the replication progression fork is able to induce SCE. In contrast, inhibition of replication initiation (by X rays and bleomycin for example) rarely and only mildly affect the incidence of SCE (Latt 1981; Painter 1980). Study Design. Chromatids can be visualized in the late prophase and early metaphase of mitosis, before chromatid segregation in daughter cells. In order to increase


327

the number of cells in metaphase and thus facilitate SCE scoring, a mitotic inhibitor such as colchicine is administered a few hours before tissue sampling (e.g., about 3 hours for bone marrow), and cells are then prepared in the same way as for chromosome aberration analysis. Techniques for SCE visualization take advantage of the semiconservative nature of DNA replication (Wilson and Thompson 2007). In the 1950s, the first method developed for the detection of SCE was based on the incorporation of tritiated thymidine in newly synthesized DNA. Subsequently, the most commonly used techniques consist of injecting a thymidine base analogue (e.g., 5-bromo-2-deoxyuridine or BrdU) to animals, shortly after test compound administration (e.g., 2–3 hours after administration in the case of bone marrow). Alternatively, small tablets of BrdU can be implanted subcutaneously (e.g., 8 hours before treatment) for sustained and continuous base analogue release (Allen et al. 1977; Latt 1981; King et al. 1982; Madrigal-Bujaidar and Sanchez-Sanchez 1991). Tissues— for example, rodent bone marrow—are generally sampled 21 hours after BrdU infusion, or 24 hours after tablet implantation. BrdU is only incorporated into the nascent daughter strand of each DNA duplex, and after the second division the two sister chromatids bear different amounts of BrdU. One of the two sister chromatids has the original template strand that contains no BrdU and the second strand that has incorporated BrdU. The other sister chromatid has BrdU incorporated on both strands. SCE are visualized as asymmetrically stained chromosomes, or “harlequin” chromosomes, after differential chromatid staining with nonfluorescent dyes (e.g., Giemsa), or fluorescent dyes (e.g., Hoechst dyes), the staining being generally less pronounced in the chromatid or fragments of chromatid containing BrdU on both strands (Korenberg and Freedlender 1974; Latt 1974; Latt and Schreck 1980). One of the standard techniques, named the fluorescent plus Giemsa method, combines fluorescent (e.g., Hoescht) and nonfluorescent (e.g., Giemsa) dyes (Perry and Wolff 1974; Speit 1984; Speit and Haupter 1985; Spencer and Butler 1987). The sister chromatid with BrdU incorporation shows a less pronounced staining with Giemsa. SCE can also be visualized with fluorescent BudR antibodies that specifically label BrdU-substituted DNA (Natarajan et al. 1986). DNA is generally counterstained with 4′,6-diamidino-2-phenylindole (DAPI) or propidium iodide (Pinkel et al. 1985). The parameter measured in this assay is the incidence of SCE per cell. Assay Limitations and Interpretation. Incorporation of BrdU itself can contribute to SCE, because it results in single strand-breaks and alkali-labile sites. Immunofluorescence methods that only necessitate small doses of BrdU for SCE visualization yield a lower SCE background than other techniques (Wilson and Thompson 2007). A method using Biotin-dUTP instead of BrdU was recently reported to overcome this technical issue (Wojcik et al. 2004). The SCE indicator assay measures error-free homologous recombination occurring in case of fork collapse and replication-blocking lesions. An increased incidence of SCE is considered as a potential marker of exposure (Wilson and Thompson 2007). It has also been linked to the induction of single strand breaks. Numerous studies have described the induction of SCE after exposure to DNAdamaging compounds such as alkylating agents, crossslinking agents, heavy metals, agents that form bulky adducts, and UV (Helleday 2003).

328

CHAPTER 12


Most clastogenic and mutagenic compounds are SCE inducers, although some qualitative and quantitative differences have been reported. However, it has been suggested that SCE can also be provoked by nongenotoxic mechanisms able to disturb DNA synthesis. Moreover, recent epidemiological studies suggest that studies of SCEs cannot replace the analysis of structural chromosome damage (Gebhart 1981) and that SCEs might not be indicative of cancer risk (Liou et al. 1999; Hagmar et al. 2001; Norppa et al. 2006). Consequently, even if a large number of genotoxic products can enhance the incidence of SCE, an increased incidence of SCE does not always reflect a genotoxic mechanism (Bradley et al. 1979) and is therefore difficult to interpret.

12.4.5.


Purpose. The purpose of in vivo gene mutation assays is to detect point mutations, such as base-pair substitutions, frameshifts, small deletions, and insertions. As mentioned in Tables 12.3 and 12.4, a few assays are also able to detect large deletions and reciprocal recombination. TABLE 12.3. In Vivo Gene Mutation Assays

In Vivo Gene Mutation Assays in Endogenous Genes of Somatic Cells

Mouse (Coat) Spot Assay

Mouse Retinoblast (Eye Spot) Assay

References

OECD Test Guideline 484 (OECD 1986b), Fahrig (1975, 1995), Fahrig and Neuhauser-Klaus (1985), Russell (1977), Russell and Major (1957).

Bishop et al. (2000), Gondo et al. (1993), Searle (1977), Schiestl et al. (1997).

Casciano et al. (1999), Deubel et al. (1996), Dobrovolsky et al. (1999b, 2005), Jones et al. (1985), Aidoo et al. (1991), Tates et al. (1994), van Dam et al. (1992), Walker et al. (1999).

Endpoints

Gene mutations (point mutations, small deletions/insertions), reciprocal recombination and chromosome aberrations.

Gene deletions resulting from intrachromosomal recombination (intrachromosomal crossing-over, single-strand annealing, unequal sister-chromatid exchange and sister-chromatid exchange).

Gene mutations: point mutations, small deletions/ insertions.

Hprt Assay


329

Regulatory Acceptance. Detection of gene mutations is not generally part of the standard battery of genotoxicity tests. The mouse spot test (Fahrig 1975, 1995; Fahrig and Neuhauser-Klaus 1985; Russell 1977) is the only gene mutation test on somatic cells described in an OECD guideline (OECD 1986b), but it is rarely used because it is restricted to a single tissue and requires a very large number of animals. No official guidelines are yet available for assays applicable to any tissue, such as transgenic gene mutation assays, but several publications provide protocol recommendations (Heddle et al. 2000; Nohmi et al. 2000; Thybaud et al. 2003). Use of transgenic gene mutation assays is frequently recommended by international guidelines as follow-up tests, in the case of positive results in in vitro genetic toxicity assays and negative results in the bone marrow chromosome damage test, as well as in the case of negative results in the standard battery of genotoxicity tests and tumor findings in 2-year bioassays. A general review has been published by Lambert and colleagues (Lambert et al. 2005; OECD 2009). Principle, Interpretation, Limitations, and Strengths. Tables 12.3 and 12.4 summarize the main characteristics of the different gene mutation assays that are further compared below.

Aprt Assay Gupta et al. (1994), Van Sloun et al. (1998).

Gene mutations (point mutations, small deletions/insertions) and events conducting the loss of heterozygosity (large deletions, mitotic nondisjunctions, mitotic recombination and gene conversions).

Dbl-1 Assay Cosentino and Heddle (1995), Tao et al. (1993a,b), Tao and Heddle (1994), Uiterdijk et al. (1986), Winton et al. (1990), Winton and Ponder (1990). Gene mutations (point mutations, small deletions/ insertion) and events conducting the loss of heterozygosity (large deletions, mitotic nondisjunctions, mitotic recombination and gene conversions).

Tk Assay

Pig-a Assay

Dobrovolsky et al. (1999a,b, 2005), Tischfield et al. (1994).

Bryce et al. (2008).

Gene mutations (point mutations, small deletions/ insertion) and events conducting the loss of heterozygosity (large deletions, mitotic nondisjunctions, mitotic recombination and gene conversions).

Gene mutations: point mutations, small deletions/insertions.

(Continued)

330

CHAPTER 12

TABLE 12.3. In Vivo Gene Mutation Assays Principle

Reporter gene

Species, strain


(Continued) Mouse (Coat) Spot Assay


Detection in F1 animals of spots on fur/coat resulting from point mutations at c locus or recombination between in c and p genes, induced in melanoblasts in utero. Dam mice (∼50 per group) are treated on day 8, 9, and 10 of gestation. 3–4 weeks after birth, ∼300 offsprings are examined for coat spots, resulting from somatic mutations considered genetically relevant (i.e., gene mutations at c locus in melanocytes). c locus

Detection in F1 animals of spots on eyes resulting from deletion mutations at p locus, induced in utero in precursor cells of retinal pigment epithelial cells. C57Bl/6J pun/pun dam mice are treated at about day 10 of gestation. F1 animals are euthanized 20 days after birth. A large number of dams per group needs to be exposed, and numerous offsprings should be examined for eye spots at single-cell level.

Detection of hprt mutant in peripheral blood T lymphocytes or splenocytes, using 6-thioguanine (6TG)resistant phenotype.

pun locus (pink-eye unstable mutation) in the tandem duplication. Mouse C57Bl/6J pun/pun carrying an autosomal recessive mutation that produces pink eyes.

hprt locus. Located on X chromosome. Only one active copy per cell.

Mouse F1 animals that are heterozygous for different recessive coat color genes, as a result of mating of T-strain mice with HT or C57/Bl mice.

Hprt Assay

Wild-type animals, mainly rodents. Also applicable in other species, including human.


331

Aprt Assay

Dbl-1 Assay

Tk Assay

Detection of aprt mutant in T lymphocytes from peripheral blood or splenocytes, and fibroblasts, using 8-azaadenine (AZA) or 2,6-diaminopurine (DAP) resistant phenotype.

Detection of Dlb-1b mutant cells in vertical stripes of small intestine and colon villi in heterozygous mice (Dlb-1a/ Dlb-1b) at Dlb-1 locus, using differential staining.

Detection of tk mutant in peripheral blood T lymphocytes or splenocytes, using 5-bromo-2′deoxyuridine (BrdU)-resistant phenotype.

Detection of Pig-a mutant cells in red blood cells, showing CD59− phenotype.

aprt locus. Located in chromosome 8 in mouse.

Dlb-1 locus. Located in chromosome 11 in mouse. Mouse heterozygous at Dlb-1 locus (Dlb-1a/Dlb-1b). Dlb-1b is an autosomal dominant gene that determines the expression of the binding site for the lectin Dolichos biflorus agglutinin in small intestine and colon epithelium. (Dlb-1a leads to its expression in vascular epithelium.)

Tk locus. Located in chromosome 11 in mouse.

Pig-a locus. Located on X chromosome. Only one active copy per cell. Wild-type animals, mainly rodents. Also applicable in other species, including human.

Aprt+/− C57BL/6 heterozygous mouse.

Tk+/− C57BL/6 heterozygous mouse.

Pig-a Assay

(Continued)

332

CHAPTER 12

TABLE 12.3. In Vivo Gene Mutation Assays


(Continued) Mouse (Coat) Spot Assay


Tissues

Target cells are melanoblasts in utero, and scoring is done in melanocytes in F1 animals.

Method for mutant selection

Changes in fur color: brown, gray, black, and nearly white spots randomly distributed over the whole coat. Different types of mutations can be identified. For example, gene mutations in c (albino) and c (chinchilla) alleles lead to brown spots, while reciprocal recombination by crossing over involving p (pink-eyed dilution) and c (albino) loci result in black and near white black spots.

Target cells are precursor cells of retinal pigment epithelium cells in utero, and scoring is done in retinal pigment epithelium cells in F1 animals. Changes in pigmentation of retinal pigment epithelium, appearance of pigmented mutant cells. Retinas are examined under microscope and spots of pigmented cells are counted. Deletion of one of two pun copies in the tandem duplication causes reversion from pun to wild-type p. The wild-type cells are easily identified as black pigmented cells/spots.

Hprt Assay Any tissue that can be subcultured in vitro. Mainly T-lymphocytes from spleen or peripheral blood.

Isolation of T-lymphocytes, followed by an in vitro mitogen stimulation and selection of mutant cells in the presence of a selective agent, such as 6TG, cytotoxic for nonmutant cells. The wild-type hprt gene encodes for hypoxanthineguanine phosphoribosyl transferase that phosphorylates 6TG into a cytotoxic compound. In the presence of 6TG only cells bearing a mutated hprt gene survive. The mutant frequency is the number of cell clones in the presence of 6TG versus the number cell clones in the absence of 6TG.


Aprt Assay

333

Dbl-1 Assay

Tk Assay

Pig-a Assay

Any tissue that can be subcultured in vitro. Mainly T-lymphocytes from spleen, or skin fibroblasts.

Small intestine and colon epithelial cells.

Any tissue that can be subcultured in vitro. Mainly splenic lymphocytes.

Peripheral blood erythrocytes.

Isolation of T-lymphocytes or fibroblasts, followed by in vitro mitogen stimulation (for lymphocytes) and selection of mutant cells in the presence of a selective agent, such as AZA or DAP, cytotoxic for nonmutant cells. The wild-type aprt gene encodes for adenine phosphoribosyltransferase that phosphorylates the selective agent into a cytotoxic compound. In the presence of the selective agent only cells bearing a mutated aprt gene survive. The mutant frequency is the number of cell clones in the presence of selective agent versus the number cell clones in the absence of selective agent.

Differential staining of Dlb-1b mutant cells. Dlb-1b mutant cells do not express the lectin binding site, and can easily be identified as they remain unstained in the presence of peroxidase conjugate of Dolichos biflorus agglutinin, while nonmutant cells are dark-brown stained.

Isolation of splenocytes, followed by an in vitro mitogen stimulation and selection of mutant cells in the presence of a selective agent, such as BrdU, cytotoxic for nonmutant cells. The wild type tk gene encodes for thymidine kinase that phosphorylates the selective agent into a cytotoxic compound. In the presence of the selective agent only cells bearing a mutated tk gene survive. The mutant frequency is the number of cell clones in the presence of selective agent versus the number cell clones in the absence of selective agent.

Blood samples are analyzed by flow cytometry to numerate the CD59-negative erythrocytes. The Pig-A gene product is involved in the first step in glycosylphosphatidylinositol (GPI) anchor biosynthesis. Mutations in Pig-a gene impact the cell surface expression of all GPI-anchored proteins, including CD59. The incidence of GPI-anchor deficient cells, i.e., Pig-a mutant cells, is quantified by flow cytometry, using anti-CD59-Pe and other fluorescent reagents to differrentiate the blood cell populations.

TABLE 12.4.

In Vivo Gene Mutation Assays in Transgenic Models

Transgenic Gene Mutation Assays

Muta™Mouse

Big Blue® Models

Plasmid lacZ

References

Gossen and Vijg (1993), Gossen et al. (1989, 1991, 1992), Douglas et al. (1994), Dean and Myhr (1994), Blakey et al. (1995), Heddle et al. (2000), Lambert et al. (2005), Mientjes et al. (1994), Nohmi et al. (2000), Piegorsch et al. (1997), Vijg and Douglas (1996), Thybaud et al. (2003).

Dolle et al. (1996), Gossen et al. (1995), Gossen and Vijg (1993), Heddle et al. (2000), Lambert et al. (2005), Nohmi et al. (2000), Thybaud et al. (2003), Vijg and Douglas (1996).

Endpoints


Principle

Detection of gene mutations in the lacZ bacterial reporter gene. Main steps are: (1) administration of the compound to mouse, (2) isolation of highmolecular-weight genomic DNA from the tissues of interest, (3) rescue of bacteriophage DNA bearing the bacterial reporter gene using packaging kits, (4) infection of host bacteria and (5) quantification of lacZ gene mutations in host bacteria in medium containing either X-Gal and P-Gal.

Dycaico et al. (1994), Heddle et al. (2000), Kohler et al. (1990, 1991a,b), Lambert et al. (2005), Nohmi et al. (2000), Piegorsch et al. (1995), Stiegler and Stillwell (1993), Thybaud et al. (2003), Wyborski et al. (1995). Gene mutations: point mutations, small deletions/ insertions. Detection of gene mutations in the lacI bacterial reporter gene. Main steps are: (1) administration of the compound to mouse or rat, (2) isolation of high molecular weight genomic DNA from the tissues of interest, (3) rescue of bacteriophage DNA bearing the bacterial reporter gene using packaging kits, (4) infection of host bacteria and (5) quantification of lacI gene mutations in host bacteria in medium containing X-Gal.

Gene mutations (point mutations, small deletions/insertions) and large deletions. Detection of gene mutations in the lacZ bacterial reporter gene. Main steps are: (1) administration of the compound to mouse, (2) isolation of high-molecularweight genomic DNA from the tissues of interest, (3) rescue of plasmid DNA bearing the bacterial reporter gene using HindIII digestion, absorption on lac represssor coated magnetic beads and plasmid recircularization, (4) electroporation of plasmids into the host bacteria and (5) quantification of lacZ gene mutations in host bacteria in medium containing P-Gal.

cII lambda phage Gene

Gpt delta (6-thioguanine)

Gpt Delta (spi)

Heddle et al. (2000), Jakubczak et al. (1996), Lambert et al. (2005), Nohmi et al. (2000), Swiger et al. (1999, 2001), Thybaud et al. (2003).

Hayashi et al. (2003), Heddle et al. (2000), Lambert et al. (2005), Masumura et al. (1999), Nohmi et al. (1996, 2000), Swiger et al. (2001), Thybaud et al. (2003).

Gunther et al. (1993), Hayashi et al. (2003), Heddle et al. (2000), Lambert et al. (2005), Nohmi and Masumura (2004), Nohmi et al. (1996, 1999, 2000), Thybaud et al. (2003).


Gene mutations: point mutations, small deletions/insertions.

Small and large deletions.

Detection of gene mutations in the cII lambda phage gene. Main steps are: (1) administration of the compound to mouse or rat, (2) isolation of high-molecular-weight genomic DNA from the tissues of interest, (3) rescue of bacteriophage DNA bearing the bacterial reporter gene lacZ or lacI using packaging kits, (4) infection of host bacteria and (5) quantification of cII gene mutations in host bacteria maintained at 25 °C.

Detection of gene mutations in the gpt bacterial reporter gene. Main steps are: (1) administration of the compound to mouse or rat, (2) isolation of high-molecular-weight genomic DNA from the tissues of interest, (3) rescue of bacteriophage DNA bearing the bacterial reporter gene using packaging kits, (4) infection of host bacteria and (5) quantification of gpt point mutations in host bacteria in medium containing 6-thioguanine.

Detection of deletion leading to inactivation of both redBA and gam genes. Main steps are: (1) administration of the compound to mouse or rat, (2) isolation of highmolecular-weight genomic DNA from the tissues of interest, (3) rescue of bacteriophage DNA bearing the bacterial reporter gene using packaging kits, (4) infection of host bacteria and (5) quantification of deletions with spi selection (spi stands for sensitive to P2 interference).

(Continued)

336

CHAPTER 12



Transgenic Gene Mutation Assays

Muta™Mouse

Big Blue® Models

Plasmid lacZ

Reporter gene

lacZ bacterial gene in λgtı0 bacteriophage vector. 40 copies in head-to-tail manner in chromosome 3 of all mouse cells.

LacZ bacterial gene in pUR288 plasmid. ∼20 copies per genome in multiple chromosomes, e.g., chromosomes 3 and 4 in mouse line 60.

Species, strains

Muta™Mouse: CD2F1 (BALB/CxDBA2) mouse strain 40.6

lacI bacterial gene in λLIZα bacteriophage vector, also containing α-lacZ gene. The commercially available Big Blue® mouse (mouse lineage A1) and rat contains 40 and 15–20 copies in a head-to-tail manner in chromosome 4, i.e., 80 and 30–40 copies per genome in homozygous strain, respectively. Big Blue® mouse and rat: B6C3F1 and C57BL/6 background for mice, and Fischer 344 background for rat.

Tissues

Any tissue from which DNA can be properly extracted.

Any tissues from which DNA can be properly extracted.


LacZ plasmid mouse: C57BL/6 background mouse strain 60.


337



Gpt Delta (spi)

λ phage cII gene. CII encodes for a protein that regulates the λ phage lysogenic/lytic cycle.

gpt E. coli bacterial gene in λEG10 bacteriophage vector. 80 copies in head-to-tail manner in mouse chromosome 17, i.e., 160 copies per genome, the mouse strain being maintained as homozygous. About 10 copies in rat chromosome 4q24-q31, the strain being only available as hemizygote.

Red and gam genes in λEG10 bacteriophage vector. 80 copies in head-to-tail manner in mouse chromosome 17, i.e., 160 copies per genome, the mouse strain being maintained as homozygous. About 10 copies in rat chromosome 4q24-q31, the strain being only available as hemizygote.

Applicable to all assays that use λ phage as shuttle vector (e.g., Muta™Mouse, and Big Blue® mouse and rat), except gpt delta models. The λEG10 phage used as vector in the later model bears a mutation in the chiC gene, involved in positive mutant selection. Any tissues from which DNA can be properly extracted.

gpt delta mouse (C57BL/6J background), and gpt delta rat (Sprague– Dawley background).

gpt delta mouse (C57BL/6J background), and gpt delta rat (Sprague–Dawley background).



(Continued)

338

CHAPTER 12



Transgenic Gene Mutation Assays Method for mutant selection

Muta™Mouse

Big Blue® Models

Plasmid lacZ

Two methods: (1) Colorimetric method: Infection of E. coli C (ΔlacZ) lacZ deficient host bacteria with the λgt10 bacteriophage vectors containing mutated or nonmutated lacZ genes rescued from the mouse genomic DNA. Mutant bacteria are detected on medium supplemented with X-Gal (a substrate of β-galactosidase that produces a blue product). Blue and white plaques contain wild-type lacZ, and mutant lacZ− genes, respectively. The mutant frequency is the frequency of white plaques out of the total number of plaques. (2) Positive selection method: Infection of E. coli C (ΔgalE ΔlacZ) host bacteria, i.e., deficient in galE and lacZ, with the λgt10 bacteriophage vectors containing mutated or nonmutated lacZ genes rescued from the mouse genomic DNA. Mutant bacteria are detected on medium containing P-Gal toxic for lacZ+ galE− bacteria. Only lacZ− phages lead to plaque formation after bacteria infection. The mutant frequency is expressed as the number of plaques observed in the presence of P-Gal out of the total number of plaques formed in the absence of P-Gal.

Colorimetric method: Infection of E. coli SCS-8 (lacZΔM15) host bacteria, i.e., lacI-deficient, with the λLIZα bacteriophage vectors containing mutated or nonmutated lacI genes rescued from the mouse genomic DNA. Mutant bacteria are detected on medium supplemented with X-Gal (a substrate of β-galactosidase that produces a blue product). No β-galactosidase is synthetized when lacZ gene is repressed by lacI product. White and blue plaques contain wild-type lacI, and mutant lacI− genes, respectively. The mutant frequency corresponds to frequency of blue plaques out of the total number of plaques.

Positive selection method: Electroporation of plasmids containing mutated or nonmutated lacZ genes in E. coli C (ΔgalE ΔlacZ) host bacteria, i.e., deficient in galE and lacZ. Mutant bacteria are detected on medium containing P-Gal toxic for lacZ+ galE− bacteria. Only bacteria that integrate lacZ− plasmid are able to survive. The mutant frequency is expressed as the number of colonies observed in the presence of P-Gal out of the total number of colonies formed in the absence of P-Gal.


339



Gpt Delta (spi)

Positive selection method: Infection of E. coli G1225 (hfl−) host bacteria with λgt10 bacteriophage vectors containing mutated or nonmutated cII genes rescued from the mouse genomic DNA. When E. coli hfl− host bacteria are infected by phages bearing nonmutated CII genes, the CII protein is not degraded by Hfl protease and induces cI and int gene expression that leads to phage lysogeny: No plaques are formed. In contrast, phages bearing cII mutated gene do not produce CII protein and enter in lytic instead of lysogenic cycle: plaques are formed. The mutant frequency is expressed as the number of plaques observed in hfl− host bacteria versus the number of plaques obtained in hfl+ host bacteria after incubation at 25 °C.

6-thioguanine (6TG) selection: Infection of E. coli YG6020 (cre+) host bacteria with λEG10 bacteriophage vectors containing mutated or nonmutated gpt genes rescued from the mouse genomic DNA. In the bacteria expressing Cre recombinase λEG10 DNA is converted in multi-copy-number of circularized plasmids carrying gpt and gene for chloramphenical acetyltransferase (CAT). The wild-type gpt gene encodes for guanine phosphoribosyltransferase that phosphorrylates 6TG into a cytotoxic compound for the bacteria. In the presence of 6TG and chloramphenicol, only bacteria bearing mutated gpt gene survive. Total number of infected bacteria is evaluated in medium containing chloramphenicol only. The mutant frequency is the number of colonies in medium with 6TG and chloramphenicol versus the number colonies observed in medium containing chloramphenicol only.

Spi selection: Infection of E. coli XL1 Blue host bacteria, carrying P2 phage DNA (i.e., P2 lysogen) with λEG10 bacteriophage vectors containing or not deletions in red/gam region. Only mutant phages deficient in both red and gam genes, as a result of a deletion in this region, are able to grow in the P2 lysogen host bacteria and display spi− phenotype (as long as they carry chi site and the host bacteria is recA+). Mutant frequency is the number of spi− plaques out of the total number of rescued plaques measured in E. coli XL1 Blue host bacteria not carrying the P2 phage DNA.

340

CHAPTER 12


Some publications report the measurement of mutations in cancer-related genes such as p53, but these models are currently only used for research purposes [e.g., McKinzie et al. (2001), McKinzie and Parsons (2002), and Parsons and Heflich (1998)]. The most frequently used in vivo gene mutation assays evaluate mutations not in genes specifically involved in carcinogenesis but in surrogate reporter genes. Only a few assays detect mutations in endogenous genes (see Table 12.3), and most are restricted to a limited number of tissues, species, and developmental stages, such as melanocytes in the mouse (coat) spot test, mouse retinal epithelial cells for the retinoblast (eye spot) assay, mouse small intestine and colon for mutations in Dlb-1 gene, mostly splenocytes for mutations in Hprt gene, and lymphocytes for mutations in Tk gene. During the past 20 years, transgenic animal models have been developed to detect gene mutations in any organ or tissue (see Table 12.4), provided that high-molecular-weight genomic DNA can be properly extracted (e.g., lacZ in Muta™mouse, lacI in Big Blue® mouse and rat, and gpt delta mouse and rat). These models can be used to measure gene mutations in the most appropriate tissues, according to the mode of administration, absorption, distribution, and tissue-specific metabolism. They facilitate the evaluation of gene mutations at site-of-contact tissues after inhalation, topical application, or oral administration. They are also claimed to be able to specifically detect site-of-contact mutagens (Dean et al. 1999). Most endogenous genes are transcriptionally active. DNA lesions are therefore actively removed by transcription-coupled repair and cannot accumulate over time. By contrast, the bacterial reporter genes in transgenic animals, as well as the endogenous gene Pig-A, are neutral and nontranscribed. In neutral genes, DNA lesions accumulate over time in reporter genes (Tao and Heddle 1994), and especially in nondividing or slow-dividing tissues. The study design (i.e., treatment and sampling times) should take this information into account in order to optimize the detection of gene mutations (Thybaud et al. 2003). First, the duration and number of treatments, (i.e., the administration time) should be sufficient to permit the accumulation of primary DNA damage. Then, a second period after compound administration, generally called the fixation time, is needed to allow the distribution and metabolization of the compound, the formation of primary DNA damage, and its fixation into stable mutations during DNA replication or repair. A short fixation time is sufficient for highly proliferative tissues (e.g., 3–10 days for bone marrow and the gastrointestinal tract), while a longer period is required for slowly proliferating tissues (e.g., 28 days or more for liver or mammary gland). Finally, in the case of actively transcribed endogenous genes, the expression time is the time required for the turnover and disappearance of preexisting nonmutated protein in the tissues. This is especially important when selective agents are used to visualize the mutants (e.g., 6–8 days of in vitro culture before adding the selective agent in the case of Aprt assay in splenocytes). These three periods—the administration, fixation and expression times—determine the optimal sampling time for each organ in a given model, also named the manifestation time (Heddle 1999). For endogenous genes (e.g., hprt, aprt and tk), up to 5 weeks may be required between animal exposure and analysis of mutant cells in order to ensure optimal evaluation of the mutant frequency. Preliminary data suggest that Pig-a mutations could be detected more rapidly after


341

the last administration, a plateau in mutant frequency being reached after 1–2 weeks (Bryce et al. 2008). In transgenic models, and most probably in all models using neutral genes, multiple administrations increase mutant frequency in an approximately additive manner, and the time needed to reach a mutant frequency plateau varies from one organ to another—from a few days with highly proliferative tissues to several weeks with slowly proliferative tissues (Tao and Heddle 1994; Thybaud et al. 2003). An increased mutant frequency obtained only after chronic administration (3 months or more) should be interpreted with caution, because it can result from secondary nongenotoxic effects, such as clonal expansion, genomic instability inherent to preneoplastic foci and tumors, and oxidative damage resulting from chronic induction of cytochrome P-450 monooxygenases (Heddle et al. 2000). A 28-day administration period with sampling 3 and/or 28 days after the last administration is considered appropriate for most tissues (except maybe for germ cells, for which the timing of cell development should also be taken into consideration; more than 50 days may be necessary to detect gene mutations in sperm) and is, by default, considered the optimal design for transgenic models (Thybaud et al. 2003). Another issue with gene mutation assays is how they allow mutant cells to be identified, visualized, and counted, in order to calculate the mutant frequency. Available methods for endogenous genes consist of using (1) a selective agent to select mutant cells in vitro after cell isolation and transfer to culture medium (e.g., 6-thiogunaine for hprt, 2,6-diaminopurine for aprt and bromodeoxyuridine for tk), (2) histopathological colorimetric methods (Dlb-1), and (3) pigmentation differences to visualize mutant cells in the whole animal (e.g., spot tests). Transgenic models use transgenes bearing a bacterial reporter gene (e.g., lacZ, lacI, or gpt) integrated in a shuttle vector (e.g., a phage or plasmid) in order to allow DNA exposure and the formation of gene mutations in animals, the rescue of reporter gene vectors from animal genomic DNA by in vitro packaging of shuttle vectors or excision/relegation of integrated plasmids, and the measurement of mutations in reporter genes (lacI, lacZ, or gpt) in host bacteria after phage infection or plasmid electroporation (see Table 12.4 for more details). The selection of mutants in the presence of X-Gal, the first method developed for the Muta™Mouse and Big Blue® models (Dycaico et al. 1994), is quite labor-intensive and time-consuming, because at least 100,000 plaques per animal have to be scored. A more straightforward positive selection system using P-Gal and E. coli galE- is now available for the Muta™Mouse model (Gossen et al. 1992; Gossen and Vijg 1993; Mientjes et al. 1994; Dean and Myhr 1994; Vijg and Douglas 1996), but is not applicable to Big Blue®. Alternatively, mutations can easily be detected in the cII gene of the lambda phage used as shuttle vector in the Muta™ Mouse and Big Blue® models (Jakubczak et al. 1996; Swiger et al. 2001): results obtained with cII are similar to those obtained with lacZ and lacI (Zimmer et al. 1999). Gene mutation assays measure mutant frequency, which is generally expressed as the incidence of mutant cells per million cells. A higher mutant frequency in treated animals than in controls indicates that the compound has the potential to induce gene mutations. The background mutant frequency obtained in transgenic models (∼5 × 10−5 range) is generally 5- to 10-fold higher than that of endogenous genes (as low as 10−6) (Cosentino and Heddle 2000; Lambert et al. 2005).

342

CHAPTER 12


This relatively high mutation background might have an impact on the sensitivity of these assays, as more induced mutations would be needed to detect an effect. This disadvantage is thought to be at least partially overcome by lesion accumulation after multiple administrations, as reporter genes are neutral, not transcribed, and therefore inefficiently repaired (Swiger et al. 2001). With some models, in addition to measuring the incidence of mutant cells, DNA sequencing is used to determine the mutation frequency. Mutations in endogeneous genes can only be measured if the mutants can be isolated and if the gene sequence is known (e.g., hprt and tk). The ability to isolate the mutants and to sequence them relatively easily is one advantage of transgenic models. This is particularly true for lacI (1080 bp) (Stiegler and Stillwell 1993), gpt (456 bp) (Masumura et al. 1999) and cII (294 bp) (Kohler et al. 1991a,b). Because of its length (3021 bp), molecular analysis of the lacZ gene is more complex and is generally done after genetic complementation analysis to determine in which of the three complementation regions (α, β, Ω) the mutation occurred (Douglas et al. 1994). Molecular analysis of mutations is not considered essential, but is useful for understanding the mechanism of mutation formation and for further evaluation of interindividual differences owing to potential jackspot mutations or clonal expansion (Heddle et al. 2000; Lambert et al. 2005). In this case, it can be necessary to sequence 10–25 mutants. Moreover, whenever possible, the mutational spectra can be analyzed in different genes in the same tissues (e.g., both lacZ and cII in the Muta™Mouse, and lacI and cII Big Blue® models). Assays of autosomal endogenous genes generally reveal both gene mutations and chromosome damage (e.g., aprt, tk, Dlb-1), while those using nonautosomal genes (e.g., hprt on the X chromosome) are limited to the detection of point mutations and small deletions (Tao et al. 1993b). In somatic cells, only one copy of the X chromosome gene is active. Males have only one copy of the X chromosome, and in females the second copy is inactivated. As a result, the loss of essential genes adjacent to the reporter gene cannot be compensated for by the homologous region present on the second copy of the gene. Thus, large deletions and chromosome rearrangements that impact adjacent essential genes are generally lethal for cells. The mouse (coat) spot test uses F1 animals to detect point mutations at the c locus or recombination between the c and p genes induced in melanoblasts in utero. The mouse (eye) spot test detects eye spots resulting from induction of deletions at the p locus in utero, in precursor cells of retinal pigmented epithelial cells. This assay is considered useful for specifically detecting deletions in vivo, as well as for mechanistic studies and research purposes. In transgenic models the large deletions that inactivate essential phage sequences (e.g., cos sites at both ends of the bacteriophage lambda required for phage DNA packaging) and the large deletions or insertions that strongly affect phage size prevent phage packaging and reporter gene recovery. Thus, transgenic models using a lambda phage as shuttle vector (e.g., Muta™Mouse and Big Blue®) are only able to reveal point mutations and small deletions (Heddle et al. 2000). The lacZ plasmid and gpt delta (spi) models do not have this disadvantage, and were designed to detect large deletions (Hayashi et al. 2003). The mutants in the lacZ plasmid mouse model result from both point mutations and large deletions (Gossen


343

and Vijg 1993; Gossen et al. 1995; Dolle et al. 1996, 1999). In the gpt delta mouse and rat models (Hayashi et al. 2003; Nohmi et al. 2000; Nohmi and Masumura 2004), two different selection methods are used in parallel to detect (a) point mutations by means of 6-thioguanine selection and (b) large deletions by means of spi selection (Gunther et al. 1993). Another important aspect of in vivo gene mutation assays is the number of animals used per group. Spot tests are seldom used, partly because they require (a) large numbers of treated animals in each group (up to 50 dams) and (b) the observation of up to 300 offspring. The use of such a large number of animals, when other alternatives exist, raises clear ethical issues. The other gene mutation assays, including transgenic models, generally require no more than 5–10 animals per group (Heddle et al. 2000; Lambert et al. 2005). A collaborative study has shown the reproducibility of data obtained with the Muta™Mouse and BigBlue® models across laboratories. The same study showed that, despite the multiple technical steps (DNA isolation, phage packaging, bacterial infection and mutant selection), the principal source of variability in these assays is inter-animal variability, but that 5–10 animals per group are sufficient (Piegorsch et al. 1995, 1997). It is nevertheless strongly recommended to use a block design protocol and to make sure that all samples are handled in parallel at the different steps of the analysis (Piegorsch et al. 1995, 1997; Heddle et al. 2000). Only a few assays (e.g., hprt and pigA) can be done with readily available wild-type animals. Recently published data identify the endogenous gene pigA as a potentially useful model gene for the detection of mutations in peripheral blood erythrocytes of any wild-type species, the gene being conserved across species (Bryce et al. 2008). Moreover, the analysis requires only a small volume of blood and does not require in vitro mitogen stimulation for mutant selection. Therefore, once fully validated, this assay could easily be integrated in all toxicology studies. The other models (e.g., both spot tests, aprt, tk, and Dlb-1 assays, and transgenic models) can only be performed with a specific strain of mouse or rat. Some animal strains are commercially available but quite expensive (Muta™Mouse and BigBlue®), whereas others are more difficult to purchase and are mainly used for research purpose. Other transgenic models have been developed for research purposes and are not described in this chapter. The majority of in vivo gene mutation data have been obtained with transgenic models, and especially with Muta™Mouse and Big Blue® (about 80%). In a review issued in 2005, Lambert and colleagues indicated that 165 agents have been evaluated in transgenic models (Lambert et al. 2005). In a recent update (OECD 2009), the same authors analyzed data for 228 different types of exposure (chemical, radiation, diet, infectious agents and mixtures). Among the 165 agents described in the first paper, 105 have already been evaluated in carcinogenicity studies (92 carcinogens and 13 noncarcinogens). In the updated report, 141 agents have been evaluated in 2-year bioassays (118 carcinogens and 23 noncarcinogens). It should be noted that the vast majority of the gene mutation results were obtained after short-term treatment and that the recommended optimal protocol was not applied (i.e., 28 days of treatment and sampling 3 and/or 28 days later). Nevertheless, for the 105 compounds analyzed in 2005, the concordance with carcinogenicity data is

344

CHAPTER 12


77% (same value in the recent update), with 78% sensitivity and 69% specificity. Transgenic in vivo gene mutation assays show a slightly stronger correlation with the in vivo UDS and comet assays (about 85%), than with in vivo chromosome damage tests (about 70–80%). They are therefore a useful complement to in vivo chromosome damage tests (i.e., bone marrow micronucleus and chromosome aberration tests), in case of conflicting results. At least 70% of carcinogen compounds considered to induce tumors through a genotoxic mechanism, and previously found to be positive in in vitro genotoxicity assays and negative in in vivo chromosome damage tests conducted with bone marrow, are properly detected in gene mutation assays (Lambert et al. 2005; Kirkland and Speit 2008). Moreover, these models are particularly appropriate for site-of-contact effects (Dean et al. 1999). It was recently recommended to use gene mutation assays to evaluate the effect of low doses (Moore et al. 2008) and to better assess the shape of the dose–response curve (i.e., the existence of a threshold).

12.5.

CONCLUSION AND PERSPECTIVES

In vivo genotoxicity assays provide an integrated and pertinent approach for evaluating genetic changes. Numerous in vivo assays have been developed and continuously improved during the past 30–40 years in order to detect different genotoxicity endpoints in different tissues. However, no currently available in vivo assays are able to detect all genotoxic carcinogens. Some assays are used as markers of exposure— to detect the ability of the compound to interact with DNA in tissues—while other assays represent markers of effect, able to reveal stable genetic changes. The latter are considered more relevant to risk assessment. In the weight-of-evidence approach, more weight is generally given to positive results obtained in in vivo assays than in in vitro assays. In this context it is important to select in vivo assays that accurately identify genotoxic carcinogens, with a low rate of false-negative results. More than one in vivo endpoint and/or tissue might be necessary to reach an appropriate level of sensitivity (e.g., combination of the in vivo bone marrow micronucleus test and the liver comet assay). In order to optimize the use of animals and to respect the 3R’s engagement, the choice of assays and the data interpretation should take into account all available information (e.g., physicochemical properties, metabolism, potential pharmacological and toxicological activities, in vitro data and specificities related to the tissue, gender and/or species). They can also be integrated in studies conducted for other purposes, such as organ toxicity. Moreover, models that can be applied both in vitro and in vivo (e.g., micronucleus test and comet assay, and more recently gene mutations in the lacZ reporter gene), as well as endpoints/tissues that can be evaluated in both animals and humans (e.g., comet assay, DNA adducts, micronucleus in peripheral blood reticulocytes and chromosome aberrations in blood lymphocytes, and, maybe soon, Pig-a gene mutations), contribute to a better understanding of (a) conflicting results and (b) extrapolation to humans. A mode-of-action approach is being developed that uses all available information to understand carcinogenicity mechanisms and to identify for a given compound the key events involved in this multistage process, in order to better evaluate the risk for humans. Even if

REFERENCES

345

genetic changes are necessary for tumor induction by genotoxic carcinogens, they are often not sufficient; other mechanisms such as cell proliferation could be the key events and limiting factors for tumor development. In this context, the development of “omic” techniques (e.g., proteomics, toxicogenomics, etc.) might contribute to a more global analysis of mechanisms involved in early and late steps of carcinogenicity, as well as to the identification of new markers. In vivo genotoxicity studies are also being conducted to better define the impact of low doses and to determine if a threshold dose may also exist for genotoxic carcinogens. In vivo genotoxicity assays, despite their limitations and the need to consider the results together with all other available information, remain crucial for genotoxicity and carcinogenicity risk assessment.

REFERENCES Aardema, M. J., Albertini, S., Arni, P., Henderson, L. M., Kirsch-Volders, M., Mackay, J. M., Sarrif, A. M., Stringer, D. A., and Taalman, R. D. (1998). Aneuploidy: A report of an ECETOC task force. Mutat Res 410, 3–79. Adler, I. D. (1984). Cytogenetic tests in mammals. In Mutagenicity Testing: A Practical Approach, Venitt, S., and Parry, J. M., eds., IRL Press, Oxford, pp. 275–306. Aidoo, A., Lyn-Cook, L. E., Mittelstaedt, R. A., Heflich, R. H., and Casciano, D. A. (1991). Induction of 6-thioguanine-resistant lymphocytes in Fischer 344 rats following in vivo exposure to N-ethyl-Nnitrosourea and cyclophosphamide. Environ Mol Mutagen 17, 141–151. Allen, J. W., Shuler, C. F., Mendes, R. W., and Latt, S. A. (1977). A simplified technique for in vivo analysis of sister-chromatid exchanges using 5-bromodeoxyuridine tablets. Cytogenet Cell Genet 18, 231–237. Amphlett, N. W., Mitchell, I. D., Rees, R. W., and Haynes, G. A. (1996). A proposal for a two-dose/ single-sample in vivo/in vitro rat hepatocyte unscheduled DNA synthesis assay. Mutagenesis 11, 19–26. Anderson, D., Yu, T. W., and McGregor, D. B. (1998). Comet assay responses as indicators of carcinogen exposure. Mutagenesis 13, 539–555. Asano, N., Torous, D. K., Tometsko, C. R., Dertinger, S. D., Morita, T., and Hayashi, M. (2006). Practical threshold for micronucleated reticulocyte induction observed for low doses of mitomycin C, Ara-C and colchicine. Mutagenesis 21, 15–20. Ashby, J., Lefevre, P. A., Burlinson, B., and Penman, M. G. (1985). An assessment of the in vivo rat hepatocyte DNA-repair assay. Mutat Res 156, 1–18. Baird, W. M. (1979). The use of radioactive carcinogens to detect DNA modification. In Chemical Carcinogens and DNA, Grover, P. L., ed., CRC Press, Boca Raton, FL, pp. 59–83. Baute, J., and Depicker, A. (2008). Base excision repair and its role in maintaining genome stability. Crit Rev Biochem Mol Biol 43, 239–276. Beland, F. A., Churchwell, M. I., Von Tungeln, L. S., Chen, S., Fu, P. P., Culp, S. J., Schoket, B., Gyorffy, E., Minarovits, J., Poirier, M. C., Bowman, E. D., Weston, A., and Doerge, D. R. (2005). Highperformance liquid chromatography electrospray ionization tandem mass spectrometry for the detection and quantitation of benzo[a]pyrene–DNA adducts. Chem Res Toxicol 18, 1306–1315. Benning, V., Brault, D., Duvinage, C., Thybaud, V., and Melcion, C. (1994). Validation of the in vivo CD1 mouse splenocyte micronucleus test. Mutagenesis 9, 199–204. Benning, V., Depasse, F., Melcion, C., and Cordier, A. (1992). Detection of micronuclei after exposure to mitomycin C, cyclophosphamide and diethylnitrosamine by the in vivo micronucleus test in mouse splenocytes. Mutat Res 280, 137–142. Bharadwaj, R., and Yu, H. (2004). The spindle checkpoint, aneuploidy, and cancer. Oncogene 23, 2016–2027.

346

CHAPTER 12


Bishop, A. J., Kosaras, B., Sidman, R. L., and Schiestl, R. H. (2000). Benzo(a)pyrene and X-rays induce reversions of the pink-eyed unstable mutation in the retinal pigment epithelium of mice. Mutat Res 457, 31–40. Bishop, J. B., and Kodell, R. L. (1980). The heritable translocation assay: Its relationship to assessment of genetic risk for future generations. Teratog Carcinog Mutagen 1, 305–332. Blakey, D. H., Douglas, G. R., Huang, K. C., and Winter, H. J. (1995). Cytogenetic mapping of lambda gt10 lacZ sequences in the transgenic mouse strain 40.6 (Muta Mouse). Mutagenesis 10, 145–148. Böcker, W., Rolf, W., Bauch, T., Müller, W.-U., and Streffer, C. (1999). Automated comet assay analysis. Cytometry 35, 134–144. Bradley, M. O., Hsu, I. C., and Harris, C. C. (1979). Relationship between sister chromatid exchange and mutagenicity, toxicity and DNA damage. Nature 282, 318–320. Braithwaite, I., and Ashby, J. (1988). A non-invasive micronucleus assay in the rat liver. Mutat Res 203, 23–32. Brambilla, G., and Martelli, A. (2004). Failure of the standard battery of short-term tests in detecting some rodent and human genotoxic carcinogens. Toxicology 196, 1–19. Brendler-Schwaab, S., Hartmann, A., Pfuhler, S., and Speit, G. (2005). The in vivo comet assay: use and status in genotoxicity testing. Mutagenesis 20, 245–254. Brendler-Schwaab, S. Y., Schmezer, P., Liegibel, U., Weber, S., Michalek, K., Tompa, A., and PoolZobel, B. L. (1994). Cells of different tissues for in vitro and in vivo studies in toxicology: Compilation of isolation methods. Toxicology in Vitro 8, 1285–1302. Brown, K., Dingley, K. H., and Turteltaub, K. W. (2005). Accelerator mass spectrometry for biomedical research. Methods Enzymol 402, 423–443. Bryce, S. M., Bemis, J. C., and Dertinger, S. D. (2008). In vivo mutation assay based on the endogenous Pig-a locus. Environ Mol Mutagen 49, 256–264. Burlinson, B. (1989). An in vivo unscheduled DNA synthesis (UDS) assay in the rat gastric mucosa: Preliminary development. Carcinogenesis 10, 1425–1428. Burlinson, B., Tice, R. R., Speit, G., Agurell, E., Brendler-Schwaab, S. Y., Collins, A. R., Escobar, P., Honma, M., Kumaravel, T. S., Nakajima, M., Sasaki, Y. F., Thybaud, V., Uno, Y., Vasquez, M., and Hartmann, A. (2007). Fourth International Workgroup on Genotoxicity testing: Results of the in vivo Comet assay workgroup. Mutat Res 627, 31–35. Buss, P., Caviezel, M., and Lutz, W. K. (1990). Linear dose–response relationship for DNA adducts in rat liver from chronic exposure to aflatoxin B1. Carcinogenesis 11, 2133–2135. Butterworth, B. E., Ashby, J., Bermudez, E., Casciano, D., Mirsalis, J., Probst, G., and Williams, G. (1987). A protocol and guide for the in vivo rat hepatocyte DNA-repair assay. Mutat Res 189, 123–133. Cammerer, Z., Elhajouji, A., Kirsch-Volders, M., and Suter, W. (2007a). Comparison of the peripheral blood micronucleus test using flow cytometry in rat and mouse exposed to aneugens after single-dose applications. Mutagenesis 22, 129–134. Cammerer, Z., Elhajouji, A., and Suter, W. (2007b). In vivo micronucleus test with flow cytometry after acute and chronic exposures of rats to chemicals. Mutat Res 626, 26–33. Casciano, D. A., Aidoo, A., Chen, T., Mittelstaedt, R. A., Manjanatha, M. G., and Heflich, R. H. (1999). Hprt mutant frequency and molecular analysis of Hprt mutations in rats treated with mutagenic carcinogens. Mutat Res 431, 389–395. Chi, Y. H., and Jeang, K. T. (2007). Aneuploidy and cancer. J Cell Biochem 102, 531–538. Chiarelli, M. P., and Jackson, O. L., Jr. (1992). Mass spectrometry for the analysis of carcinogen–DNA adducts. Mass Spectrom Rev 11, 447–493. Cimino, M. C. (2006). Comparative overview of current international strategies and guidelines for genetic toxicology testing for regulatory purposes. Environ Mol Mutagen 47, 362–390. Cliet, I., Fournier, E., Melcion, C., and Cordier, A. (1989). In vivo micronucleus test using mouse hepatocytes. Mutat Res 216, 321–326. Collins, A. R. (2002). The comet assay. Principles, applications, and limitations. Methods Mol Biol 203, 163–177. Collins, A. R., Duthie, S. J., and Dobson, V. L. (1993). Direct enzymic detection of endogenous oxidative base damage in human lymphocyte DNA. Carcinogenesis 14, 1733–1735.

REFERENCES

347

Collins, A. R., Oscoz, A. A., Brunborg, G., Gaivao, I., Giovannelli, L., Kruszewski, M., Smith, C. C., and Stetina, R. (2008). The comet assay: Topical issues. Mutagenesis 23, 143–151. Cosentino, L., and Heddle, J. A. (1995). The induction of dominant somatic mutations at the Dlb-1 locus. Mutat Res 346, 115–119. Cosentino, L., and Heddle, J. A. (2000). Differential mutation of transgenic and endogenous loci in vivo. Mutat Res 454, 1–10. CSGMT (1990). Single versus multiple dosing in the micronucleus test: the summary of the fourth collaborative study by CSGMT/JEMS.MMS. Collaborative Study Group for the Micronucleus Test, the Mammalian Mutagenesis Study Group of the Environmental Mutagen Society, Japan (CSGMT/JEMS. MMS). Mutat Res 234, 205–222. Dean, S. W., Brooks, T. M., Burlinson, B., Mirsalis, J., Myhr, B., Recio, L., and Thybaud, V. (1999). Transgenic mouse mutation assay systems can play an important role in regulatory mutagenicity testing in vivo for the detection of site-of-contact mutagens. Mutagenesis 14, 141–151. Dean, S. W., and Myhr, B. (1994). Measurement of gene mutation in vivo using Muta Mouse and positive selection for lacZ- phage. Mutagenesis 9, 183–185. Dearfield, K. L., and Moore, M. M. (2005). Use of genetic toxicology information for risk assessment. Environ Mol Mutagen 46, 236–445. Dehon, G., Catoire, L., Duez, P., Bogaerts, P., and Dubois, J. (2008). Validation of an automatic comet assay analysis system integrating the curve fitting of combined comet intensity profiles. Mutat Res 650, 87–95. Den Engelse, L., van Benthem, J., and Scherer, E. (1990). Immunocytochemical analysis of in vivo DNA modification. Mutat Res 233, 265–287. Dertinger, S. D., Camphausen, K., Macgregor, J. T., Bishop, M. E., Torous, D. K., Avlasevich, S., Cairns, S., Tometsko, C. R., Menard, C., Muanza, T., Chen, Y., Miller, R. K., Cederbrant, K., Sandelin, K., Ponten, I., and Bolcsfoldi, G. (2004). Three-color labeling method for flow cytometric measurement of cytogenetic damage in rodent and human blood. Environ Mol Mutagen 44, 427–435. Dertinger, S. D., Torous, D. K., Hall, N. E., Murante, F. G., Gleason, S. E., Miller, R. K., and Tometsko, C. R. (2002). Enumeration of micronucleated CD71-positive human reticulocytes with a single-laser flow cytometer. Mutat Res 515, 3–14. Dertinger, S. D., Torous, D. K., and Tometsko, K. R. (1996). Simple and reliable enumeration of micronucleated reticulocytes with a single-laser flow cytometer. Mutat Res 371, 283–292. Dertinger, S. D., Torous, D. K., and Tometsko, K. R. (1997). Flow cytometric analysis of micronucleated reticulocytes in mouse bone marrow. Mutat Res 390, 257–262. Deubel, W., Bassukas, I. D., Schlereth, W., Lorenz, R., and Hempel, K. (1996). Age dependent selection against HPRT deficient T lymphocytes in the HPRT+/− heterozygous mouse. Mutat Res 351, 67–77. Dingley, K. H., Roberts, M. L., Velsko, C. A., and Turteltaub, K. W. (1998). Attomole detection of 3H in biological samples using accelerator mass spectrometry: Application in low-dose, dual-isotope tracer studies in conjunction with 14C accelerator mass spectrometry. Chem Res Toxicol 11, 1217–1222. Dingley, K. H., Ubick, E. A., Vogel, J. S., and Haack, K. W. (2005). DNA isolation and sample preparation for quantification of adduct levels by accelerator mass spectrometry. Methods Mol Biol 291, 21–27. Dipple, A. (1995). DNA adducts of chemical carcinogens. Carcinogenesis 16, 437–441. Dobrovolsky, V. N., Casciano, D. A., and Heflich, R. H. (1999a). Tk+/− mouse model for detecting in vivo mutation in an endogenous, autosomal gene. Mutat Res 423, 125–136. Dobrovolsky, V. N., Chen, T., and Heflich, R. H. (1999b). Molecular analysis of in vivo mutations induced by N-ethyl-N-nitrosourea in the autosomal Tk and the X-linked Hprt genes of mouse lymphocytes. Environ Mol Mutagen 34, 30–38. Dobrovolsky, V. N., Shaddock, J. G., and Heflich, R. H. (2005). Analysis of in vivo mutation in the hprt and tk genes of mouse lymphocytes. Methods Mol Biol 291, 133–144. Dolle, M. E., Martus, H. J., Gossen, J. A., Boerrigter, M. E., and Vijg, J. (1996). Evaluation of a plasmidbased transgenic mouse model for detecting in vivo mutations. Mutagenesis 11, 111–118. Dolle, M. E., Snyder, W. K., van Orsouw, N. J., and Vijg, J. (1999). Background mutations and polymorphisms in lacZ-plasmid transgenic mice. Environ Mol Mutagen 34, 112–120.

348

CHAPTER 12


Douglas, G. R., Gingerich, J. D., Gossen, J. A., and Bartlett, S. A. (1994). Sequence spectra of spontaneous lacZ gene mutations in transgenic mouse somatic and germline tissues. Mutagenesis 9, 451–458. Dycaico, M. J., Provost, G. S., Kretz, P. L., Ransom, S. L., Moores, J. C., and Short, J. M. (1994). The use of shuttle vectors for mutation analysis in transgenic mice and rats. Mutat Res 307, 461–478. Eide, I., Zhao, C., Kumar, R., Hemminki, K., Wu, K., and Swenberg, J. A. (1999). Comparison of (32) P-postlabeling and high-resolution GC/MS in quantifying N7-(2-hydroxyethyl)guanine adducts. Chem Res Toxicol 12, 979–984. Fahrig, R. (1975). A mammalian spot test: induction of genetic alterations in pigment cells of mouse embryos with x-rays and chemical mutagens. Mol Gen Genet 138, 309–314. Fahrig, R. (1995). The mouse spot test. In Environmental Mutagenesis, Phillips, D. H., and Venitt, S., eds., Bios Scientific Publishers, Oxford, pp. 180–199. Fahrig, R., and Neuhauser-Klaus, A. (1985). Similar pigmentation characteristics in the specific-locus and the mammalian spot test. A way to distinguish between induced mutation and recombination. J Hered 76, 421–426. Fairbairn, D. W., Olive, P. L., and O’Neill, K. L. (1995). The comet assay: A comprehensive review. Mutat Res 339, 37–59. Fairbairn, D. W., Walburger, D. K., Fairbairn, J. J., and O’Neill, K. L. (1996). Key morphologic changes and DNA strand breaks in human lymphoid cells: Discriminating apoptosis from necrosis. Scanning 18, 407–416. Farmer, P. B. (2004a). DNA and protein adducts as markers of genotoxicity. Toxicol Lett 149, 3–9. Farmer, P. B. (2004b). Exposure biomarkers for the study of toxicological impact on carcinogenic processes. IARC Sci Publ, 71–90. Farmer, P. B., Brown, K., Tompkins, E., Emms, V. L., Jones, D. J., Singh, R., and Phillips, D. H. (2005). DNA adducts: mass spectrometry methods and future prospects. Toxicol Appl Pharmacol 207, 293–301. Farmer, P. B., and Singh, R. (2008). Use of DNA adducts to identify human health risk from exposure to hazardous environmental pollutants: The increasing role of mass spectrometry in assessing biologically effective doses of genotoxic carcinogens. Mutat Res 659, 68–76. Fousteri, M., and Mullenders, L. H. (2008). Transcription-coupled nucleotide excision repair in mammalian cells: Molecular mechanisms and biological effects. Cell Res 18, 73–84. Frieauff, W., Hartmann, A., and Suter, W. (2001). Automatic analysis of slides processed in the Comet assay. Mutagenesis 16, 133–137. Frieauff, W., and Romagna, F. (1994). Technical aspects of automatic micronucleus analysis in rodent bone marrow assays. Cell Biol Toxicol 10, 283–289. Furihata, C., and Matsushima, T. (1987). Use of in vivo/in vitro unscheduled DNA synthesis for identification of organ-specific carcinogens. Crit Rev Toxicol 17, 245–277. Furihata, C., Yamawaki, Y., Jin, S. S., Moriya, H., Kodama, K., Matsushima, T., Ishikawa, T., Takayama, S., and Nakadate, M. (1984). Induction of unscheduled DNA synthesis in rat stomach mucosa by glandular stomach carcinogens. J Natl Cancer Inst 72, 1327–1334. Gamboa da Costa, G., Marques, M. M., Beland, F. A., Freeman, J. P., Churchwell, M. I., and Doerge, D. R. (2003). Quantification of tamoxifen DNA adducts using on-line sample preparation and HPLCelectrospray ionization tandem mass spectrometry. Chem Res Toxicol 16, 357–366. Ganem, N. J., Storchova, Z., and Pellman, D. (2007). Tetraploidy, aneuploidy and cancer. Curr Opin Genet Dev 17, 157–162. Garner, R. C. (1998). The role of DNA adducts in chemical carcinogenesis. Mutat Res 402, 67–75. Gebhart, E. (1981). Sister chromatid exchange (SCE) and structural chromosome aberration in mutagenicity testing. Hum Genet 58, 235–254. Gocke, E., Ballantyne, M., Whitwell, J., and Müller, L. (2009). MNT and MutaMouse studies to define the in vivo dose response relations of the genotoxicity of EMS and ENU. Toxicol Lett 190, 286–297. Gocke, E., and Müller, L. (2009). In vivo studies in the mouse to define a threshold for the genotoxicity of EMS and ENU. Mutat Res 678, 101–107. Goldman, R., Day, B. W., Carver, T. A., Mauthe, R. J., Turteltaub, K. W., and Shields, P. G. (2000). Quantitation of benzo[a]pyrene–DNA adducts by postlabeling with 14C-acetic anhydride and accelerator mass spectrometry. Chem Biol Interact 126, 171–183.

REFERENCES

349

Gondo, Y., Gardner, J. M., Nakatsu, Y., Durham-Pierre, D., Deveau, S. A., Kuper, C., and Brilliant, M. H. (1993). High-frequency genetic reversion mediated by a DNA duplication: The mouse pink-eyed unstable mutation. Proc Natl Acad Sci USA 90, 297–301. Gossen, J. A., de Leeuw, W. J., Tan, C. H., Zwarthoff, E. C., Berends, F., Lohman, P. H., Knook, D. L., and Vijg, J. (1989). Efficient rescue of integrated shuttle vectors from transgenic mice: A model for studying mutations in vivo. Proc Natl Acad Sci USA 86, 7971–7975. Gossen, J. A., de Leeuw, W. J., Verwest, A., Lohman, P. H., and Vijg, J. (1991). High somatic mutation frequencies in a LacZ transgene integrated on the mouse X-chromosome. Mutat Res 250, 423–429. Gossen, J. A., Martus, H. J., Wei, J. Y., and Vijg, J. (1995). Spontaneous and X-ray-induced deletion mutations in a LacZ plasmid-based transgenic mouse model. Mutat Res 331, 89–97. Gossen, J. A., Molijn, A. C., Douglas, G. R., and Vijg, J. (1992). Application of galactose-sensitive E. coli strains as selective hosts for LacZ-plasmids. Nucleic Acids Res 20, 3254. Gossen, J. A., and Vijg, J. (1993). A selective system for lacZ-phage using a galactose-sensitive E. coli host. Biotechniques 14, 326, 330. Grawé, J., Abramsson-Zetterberg, L., and Zetterberg, G. (1998). Low dose effects of chemicals as assessed by the flow cytometric in vivo micronucleus assay. Mutat Res 405, 199–208. Gunther, E. J., Murray, N. E., and Glazer, P. M. (1993). High efficiency, restriction-deficient in vitro packaging extracts for bacteriophage lambda DNA using a new E. coli lysogen. Nucleic Acids Res 21, 3903–3904. Gupta, P. K., Sahota, A., Boyadjiev, S. A., Bye, S., O’Neill, J. P., Hunter, T. C., Albertini, R. J., and Tischfield, J. A. (1994). Analysis of in vivo somatic mutations at the APRT locus. Adv Exp Med Biol 370, 653–656. Gupta, R. C., Reddy, M. V., and Randerath, K. (1982). 32P-postlabeling analysis of nonradioactive aromatic carcinogen–DNA adducts. Carcinogenesis 3, 1081–1092. Hagmar, L., Stromberg, U., Tinnerberg, H., and Mikoczy, Z. (2001). The usefulness of cytogenetic biomarkers as intermediate endpoints in carcinogenesis. Int J Hyg Environ Health 204, 43–47. Hamada, S., Sutou, S., Morita, T., Wakata, A., Asanami, S., Hosoya, S., Ozawa, S., Kondo, K., Nakajima, M., Shimada, H., Osawa, K., Kondo, Y., Asano, N., Sato, S., Tamura, H., Yajima, N., Marshall, R., Moore, C., Blakey, D. H., Schechtman, L. M., Weaver, J. L., Torous, D. K., Proudlock, R., Ito, S., Namiki, C., and Hayashi, M. (2001). Evaluation of the rodent micronucleus assay by a 28-day treatment protocol: Summary of the 13th Collaborative Study by the Collaborative Study Group for the Micronucleus Test (CSGMT)/Environmental Mutagen Society of Japan (JEMS)-Mammalian Mutagenicity Study Group (MMS). Environ Mol Mutagen 37, 93–110. Hamilton, C. M., and Mirsalis, J. C. (1987). Factors that affect the sensitivity of the in vivo–in vitro hepatocyte DNA repair assay in the male rat. Mutat Res 189, 341–347. Hanahan, D., and Weinberg, R. A. (2000). The hallmarks of cancer. Cell 100, 57–70. Harper, S. B., Dertinger, S. D., Bishop, M. E., Lynch, A. M., Lorenzo, M., Saylor, M., and MacGregor, J. T. (2007). Flow cytometric analysis of micronuclei in peripheral blood reticulocytes III. An efficient method of monitoring chromosomal damage in the beagle dog. Toxicol Sci 100, 406–414. Hartmann, A., Agurell, E., Beevers, C., Brendler-Schwaab, S., Burlinson, B., Clay, P., Collins, A., Smith, A., Speit, G., Thybaud, V., and Tice, R. R. (2003). Recommendations for conducting the in vivo alkaline Comet assay. 4th International Comet Assay Workshop. Mutagenesis 18, 45–51. Hartmann, A., Schumacher, M., Plappert-Helbig, U., Lowe, P., Suter, W., and Mueller, L. (2004). Use of the alkaline in vivo comet assay for mechanistic genotoxicity investigations. Mutagenesis 19, 51–59. Hayashi, H., Kondo, H., Masumura, K., Shindo, Y., and Nohmi, T. (2003). Novel transgenic rat for in vivo genotoxicity assays using 6-thioguanine and Spi− selection. Environ Mol Mutagen 41, 253–259. Hayashi, M., MacGregor, J. T., Gatehouse, D. G., Adler, I. D., Blakey, D. H., Dertinger, S. D., Krishna, G., Morita, T., Russo, A., and Sutou, S. (2000). In vivo rodent erythrocyte micronucleus assay. II. Some aspects of protocol design including repeated treatments, integration with toxicity testing, and automated scoring. Environ Mol Mutagen 35, 234–252. Hayashi, M., MacGregor, J. T., Gatehouse, D. G., Blakey, D. H., Dertinger, S. D., Abramsson-Zetterberg, L., Krishna, G., Morita, T., Russo, A., Asano, N., Suzuki, H., Ohyama, W., and Gibson, D. (2007). In vivo erythrocyte micronucleus assay III. Validation and regulatory acceptance of automated scoring

350

CHAPTER 12


and the use of rat peripheral blood reticulocytes, with discussion of non-hematopoietic target cells and a single dose-level limit test. Mutat Res 627, 10–30. Hayashi, M., Morita, T., Kodama, Y., Sofuni, T., and Ishidate, M., Jr. (1990). The micronucleus assay with mouse peripheral blood reticulocytes using acridine orange-coated slides. Mutat Res 245, 245–249. Hayashi, M., Tice, R. R., MacGregor, J. T., Anderson, D., Blakey, D. H., Kirsh-Volders, M., Oleson, F. B., Jr., Pacchierotti, F., Romagna, F., Shimada, H., et al. (1994). In vivo rodent erythrocyte micronucleus assay. Mutat Res 312, 293–304. Heddle, J. A. (1973). A rapid in vivo test for chromosomal damage. Mutat Res 18, 187–190. Heddle, J. A. (1999). Mutant manifestation: The time factor in somatic mutagenesis. Mutagenesis 14, 1–3. Heddle, J. A., Cimino, M. C., Hayashi, M., Romagna, F., Shelby, M. D., Tucker, J. D., Vanparys, P., and MacGregor, J. T. (1991). Micronuclei as an index of cytogenetic damage: Past, present, and future. Environ Mol Mutagen 18, 277–291. Heddle, J. A., Dean, S., Nohmi, T., Boerrigter, M., Casciano, D., Douglas, G. R., Glickman, B. W., Gorelick, N. J., Mirsalis, J. C., Martus, H. J., Skopek, T. R., Thybaud, V., Tindall, K. R., and Yajima, N. (2000). In vivo transgenic mutation assays. Environ Mol Mutagen 35, 253–259. Heddle, J. A., and Salamone, M. F. (1981). The micronucleus assay: In vivo. In Proceedings of the International Workshop on short term tests for chemical carcinogens, Stich, H., and San, R. H. C., eds., Springer-Verlag, New York, pp. 243–249. Hegde, M. L., Hazra, T. K., and Mitra, S. (2008). Early steps in the DNA base excision/single-strand interruption repair pathway in mammalian cells. Cell Res 18, 27–47. Helleday, T. (2003). Pathways for mitotic homologous recombination in mammalian cells. Mutat Res 532, 103–115. Hemminki, K., Koskinen, M., Rajaniemi, H., and Zhao, C. (2000). Dna adducts, mutations, and cancer 2000. Regul Toxicol Pharmacol 32, 264–275. Hotchkiss, C. E., Bishop, M. E., Dertinger, S. D., Slikker, W., Jr., Moore, M. M., and Macgregor, J. T. (2008). Flow cytometric analysis of micronuclei in peripheral blood reticulocytes IV: An index of chromosomal damage in the rhesus monkey (Macaca mulatta). Toxicol Sci 102, 352–358. Hsu, I. C., Poirier, M. C., Yuspa, S. H., Grunberger, D., Weinstein, I. B., Yolken, R. H., and Harris, C. C. (1981). Measurement of benzo(a)pyrene–DNA adducts by enzyme immunoassays and radioimmunoassay. Cancer Res 41, 1091–1095. Iarmarcovai, G., Botta, A., and Orsiere, T. (2006). Number of centromeric signals in micronuclei and mechanisms of aneuploidy. Toxicol Lett 166, 1–10. Igarashi, M., and Shimada, H. (1997). An improved method for the mouse liver micronucleus test. Mutat Res 391, 49–55. Jakubczak, J. L., Merlino, G., French, J. E., Müller, W. J., Paul, B., Adhya, S., and Garges, S. (1996). Analysis of genetic instability during mammary tumor progression using a novel selection-based assay for in vivo mutations in a bacteriophage lambda transgene target. Proc Natl Acad Sci USA 93, 9073–9078. Jones, I. M., Burkhart-Schultz, K., and Carrano, A. V. (1985). A method to quantify spontaneous and in vivo induced thioguanine-resistant mouse lymphocytes. Mutat Res 147, 97–105. Kasper, P., Uno, Y., Mauthe, R., Asano, N., Douglas, G., Matthews, E., Moore, M., Mueller, L., Nakajima, M., Singer, T., and Speit, G. (2007). Follow-up testing of rodent carcinogens not positive in the standard genotoxicity testing battery: IWGT workgroup report. Mutat Res 627, 106–116. King, M. T., Wild, D., Gocke, E., and Eckhardt, K. (1982). 5-Bromodeoxyuridine tablets with improved depot effect for analysis in vivo of sister-chromatid exchanges in bone-marrow and spermatogonial cells. Mutat Res 97, 117–129. Kirkland, D., and Speit, G. (2008). Evaluation of the ability of a battery of three in vitro genotoxicity tests to discriminate rodent carcinogens and non-carcinogens III. Appropriate follow-up testing in vivo. Mutat Res 654, 114–132. Kirsch-Volders, M., Vanhauwaert, A., De Boeck, M., and Decordier, I. (2002). Importance of detecting numerical versus structural chromosome aberrations. Mutat Res 504, 137–148. Kirsch-Volders, M., Vanhauwaert, A., Eichenlaub-Ritter, U., and Decordier, I. (2003). Indirect mechanisms of genotoxicity. Toxicol Lett 140–141, 63–74.

REFERENCES

351

Kissling, G. E., Dertinger, S. D., Hayashi, M., and MacGregor, J. T. (2007). Sensitivity of the erythrocyte micronucleus assay: Dependence on number of cells scored and inter-animal variability. Mutat Res 634, 235–240. Kligerman, A. D., Erexson, G. L., and Wilmer, J. L. (1984). Development of rodent peripheral blood lymphocyte culture systems to detect cytogenetic damage in vivo. Basic Life Sci 29(Pt B), 569–584. Koc, H., and Swenberg, J. A. (2002). Applications of mass spectrometry for quantitation of DNA adducts. J Chromatogr B Analyt Technol Biomed Life Sci 778, 323–343. Kohler, S. W., Provost, G. S., Fieck, A., Kretz, P. L., Bullock, W. O., Putman, D. L., Sorge, J. A., and Short, J. M. (1991a). Analysis of spontaneous and induced mutations in transgenic mice using a lambda ZAP/lacI shuttle vector. Environ Mol Mutagen 18, 316–321. Kohler, S. W., Provost, G. S., Fieck, A., Kretz, P. L., Bullock, W. O., Sorge, J. A., Putman, D. L., and Short, J. M. (1991b). Spectra of spontaneous and mutagen-induced mutations in the lacI gene in transgenic mice. Proc Natl Acad Sci USA 88, 7958–7962. Kohler, S. W., Provost, G. S., Kretz, P. L., Fieck, A., Sorge, J. A., and Short, J. M. (1990). The use of transgenic mice for short-term, in vivo mutagenicity testing. Genet Anal Tech Appl 7, 212–218. Kohn, K. W., Erickson, L. C., Ewig, R. A., and Friedman, C. A. (1976). Fractionation of DNA from mammalian cells by alkaline elution. Biochemistry 15, 4629–4637. Kohn, K. W., and Grimek-Ewig, R. A. (1973). Alkaline elution analysis, a new approach to the study of DNA single-strand interruptions in cells. Cancer Res 33, 1849–1853. Korenberg, J. R., and Freedlender, E. F. (1974). Giemsa technique for the detection of sister chromatid exchanges. Chromosoma 48, 355–360. Kriek, E., Den Engelse, L., Scherer, E., and Westra, J. G. (1984). Formation of DNA modifications by chemical carcinogens. Identification, localization and quantification. Biochim Biophys Acta 738, 181–201. Krishna, G., and Hayashi, M. (2000). In vivo rodent micronucleus assay: Protocol, conduct and data interpretation. Mutat Res 455, 155–166. Lambert, I. B., Singer, T. M., Boucher, S. E., and Douglas, G. R. (2005). Detailed review of transgenic rodent mutation assays. Mutat Res 590, 1–280. Latt, S. A. (1974). Sister chromatid exchanges, indices of human chromosome damage and repair: Detection by fluorescence and induction by mitomycin C. Proc Natl Acad Sci USA 71, 3162–3166. Latt, S. A. (1981). Sister chromatid exchange formation. Annu Rev Genet 15, 11–55. Latt, S. A., Allen, J., Bloom, S. E., Carrano, A., Falke, E., Kram, D., Schneider, E., Schreck, R., Tice, R., Whitfield, B., and Wolff, S. (1981). Sister-chromatid exchanges: A report of the GENE-TOX program. Mutat Res 87, 17–62. Latt, S. A., and Schreck, R. R. (1980). Sister chromatid exchange analysis. Am J Hum Genet 32, 297–313. Liou, S. H., Lung, J. C., Chen, Y. H., Yang, T., Hsieh, L. L., Chen, C. J., and Wu, T. N. (1999). Increased chromosome-type chromosome aberration frequencies as biomarkers of cancer risk in a blackfoot endemic area. Cancer Res 59, 1481–1484. Loury, D. J., Smith-Oliver, T., and Butterworth, B. E. (1987). Assessment of unscheduled and replicative DNA synthesis in rat kidney cells exposed in vitro or in vivo to unleaded gasoline. Toxicol Appl Pharmacol 87, 127–140. Lovell, D. P., Thomas, G., and Dubow, R. (1999). Issues related to the experimental design and subsequent statistical analysis of in vivo and in vitro comet studies. Teratog Carcinog Mutagen 19, 109–119. Lutz, W. K. (1979). In vivo covalent binding of organic chemicals to DNA as a quantitative indicator in the process of chemical carcinogenesis. Mutat Res 65, 289–356. Lutz, W. K. (1986). Quantitative evaluation of DNA binding data for risk estimation and for classification of direct and indirect carcinogens. J Cancer Res Clin Oncol 112, 85–91. MacGregor, J. T., Bishop, M. E., McNamee, J. P., Hayashi, M., Asano, N., Wakata, A., Nakajima, M., Saito, J., Aidoo, A., Moore, M. M., and Dertinger, S. D. (2006). Flow cytometric analysis of micronuclei in peripheral blood reticulocytes: II. An efficient method of monitoring chromosomal damage in the rat. Toxicol Sci 94, 92–107. MacGregor, J. T., Casciano, D., and Müller, L. (2000). Strategies and testing methods for identifying mutagenic risks. Mutat Res 455, 3–20.

352

CHAPTER 12


MacGregor, J. T., Heddle, J. A., Hite, M., Margolin, B. H., Ramel, C., Salamone, M. F., Tice, R. R., and Wild, D. (1987). Guidelines for the conduct of micronucleus assays in mammalian bone marrow erythrocytes. Mutat Res 189, 103–112. MacGregor, J. T., Wehr, C. M., and Gould, D. H. (1980). Clastogen-induced micronuclei in peripheral blood erythrocytes: The basis of an improved micronucleus test. Environ Mutagen 2, 509–514. Mackay, J. M. (1995). Dose selection in in vivo genetic toxicology assays. Environ Mol Mutagen 25, 323–327. Madrigal-Bujaidar, E., and Sanchez-Sanchez, M. A. (1991). Sister-chromatid exchange analysis in vivo using different 5-bromo-2’-deoxyuridine-labeling systems. Mutat Res 262, 15–19. Martin, C. N., Lawley, P. D., Phillips, D. H., Venitt, S., and Waters, R. (1993). Measurement for covalent binding to DNA in vivo. In Supplementary Mutagenicity Tests: UKEMS Recommended Procedures, Kirkland, D. J., and Fox, M., eds., Cambridge University Press, Cambridge, pp. 78–104. Masumura, K., Matsui, M., Katoh, M., Horiya, N., Ueda, O., Tanabe, H., Yamada, M., Suzuki, H., Sofuni, T., and Nohmi, T. (1999). Spectra of gpt mutations in ethylnitrosourea-treated and untreated transgenic mice. Environ Mol Mutagen 34, 1–8. Mateuca, R., Lombaert, N., Aka, P. V., Decordier, I., and Kirsch-Volders, M. (2006). Chromosomal changes: Induction, detection methods and applicability in human biomonitoring. Biochimie 88, 1515–1531. Mauthe, R. J., Dingley, K. H., Leveson, S. H., Freeman, S. P., Turesky, R. J., Garner, R. C., and Turteltaub, K. W. (1999). Comparison of DNA–adduct and tissue-available dose levels of MeIQx in human and rodent colon following administration of a very low dose. Int J Cancer 80, 539–545. Mavournin, K. H., Blakey, D. H., Cimino, M. C., Salamone, M. F., and Heddle, J. A. (1990). The in vivo micronucleus assay in mammalian bone marrow and peripheral blood. A report of the U.S. Environmental Protection Agency Gene-Tox Program. Mutat Res 239, 29–80. McKelvey-Martin, V. J., Green, M. H., Schmezer, P., Pool-Zobel, B. L., De Meo, M. P., and Collins, A. (1993). The single cell gel electrophoresis assay (comet assay): A European review. Mutat Res 288, 47–63. McKinzie, P. B., Delongchamp, R. R., Heflich, R. H., and Parsons, B. L. (2001). Prospects for applying genotypic selection of somatic oncomutation to chemical risk assessment. Mutat Res 489, 47–78. McKinzie, P. B., and Parsons, B. L. (2002). Detection of rare K-ras codon 12 mutations using allelespecific competitive blocker PCR. Mutat Res 517, 209–220. Mientjes, E. J., van Delft, J. H. M., Op’t Hof, B. M., Gossen, J. A., Vijg, J., Lohman, P. H., and Baan, R. A. (1994). An improved selection method for lambda lacZ-phages based on galactose sensitivity. Transgenic Res 3, 67–69. Mirsalis, J. C., and Butterworth, B. E. (1980). Detection of unscheduled DNA synthesis in hepatocytes isolated from rats treated with genotoxic agents: An in vivo–in vitro assay for potential carcinogens and mutagens. Carcinogenesis 1, 621–625. Mirsalis, J. C., Tyson, C. K., and Butterworth, B. E. (1982). Detection of genotoxic carcinogens in the in vivo–in vitro hepatocyte DNA repair assay. Environ Mutagen 4, 553–562. Miyamae, Y., Yamamoto, M., Sasaki, Y. F., Kobayashi, H., Igarashi-Soga, M., Shimoi, K., and Hayashi, M. (1998). Evaluation of a tissue homogenization technique that isolates nuclei for the in vivo single cell gel electrophoresis (comet) assay: A collaborative study by five laboratories. Mutat Res 418, 131–140. Moore, M. M., Heflich, R. H., Haber, L. T., Allen, B. C., Shipp, A. M., and Kodell, R. L. (2008). Analysis of in vivo mutation data can inform cancer risk assessment. Regul Toxicol Pharmacol 51, 151–161. Mori, M., Kobayashi, H., Sugiyama, C., Katsumura, Y., and Furihata, C. (1999). Induction of unscheduled DNA synthesis in hairless mouse epidermis by skin carcinogens. J Toxicol Sci 24, 217–226. Morita, T., Asano, N., Awogi, T., Sasaki, Y. F., Sato, S., Shimada, H., Sutou, S., Suzuki, T., Wakata, A., Sofuni, T., and Hayashi, M. (1997). Evaluation of the rodent micronucleus assay in the screening of IARC carcinogens (groups 1, 2A and 2B) the summary report of the 6th collaborative study by CSGMT/JEMS MMS. Collaborative Study of the Micronucleus Group Test. Mammalian Mutagenicity Study Group. Mutat Res 389, 3–122. Müller, L., Blakey, D., Dearfield, K. L., Galloway, S., Guzzie, P., Hayashi, M., Kasper, P., Kirkland, D., MacGregor, J. T., Parry, J. M., Schechtman, L., Smith, A., Tanaka, N., Tweats, D., and Yamasaki, H.

REFERENCES

353

(2003). Strategy for genotoxicity testing and stratification of genotoxicity test results—Report on initial activities of the IWGT Expert Group. Mutat Res 540, 177–181. Müller, L., and Kasper, P. (2000). Human biological relevance and the use of threshold-arguments in regulatory genotoxicity assessment: Experience with pharmaceuticals. Mutat Res 464, 19–34. Müller, R., Adamkiewicz, J., and Rajewsky, M. F. (1982). Immunological detection and quantification of carcinogen-modified DNA components. IARC Sci Publ, 463–479. Müller, R., and Rajewsky, M. F. (1980). Immunological quantification by high-affinity antibodies of O6-ethyldeoxyguanosine in DNA exposed to N-ethyl-N-nitrosourea. Cancer Res 40, 887–896. Müller, R., and Rajewsky, M. F. (1981). Antibodies specific for DNA components structurally modified by chemical carcinogens. J Cancer Res Clin Oncol 102, 99–113. Natarajan, A. T., Rotteveel, A. H., van Pieterson, J., and Schliermann, M. G. (1986). Influence of incorporated 5-bromodeoxyuridine on the frequencies of spontaneous and induced sister-chromatid exchanges, detected by immunological methods. Mutat Res 163, 51–55. Nestmann, E. R., Bryant, D. W., and Carr, C. J. (1996). Toxicological significance of DNA adducts: Summary of discussions with an expert panel. Regul Toxicol Pharmacol 24, 9–18. Nishikawa, T., Haresaku, M., Adachi, K., Masuda, M., and Hayashi, M. (1999). Study of a rat skin in vivo micronucleus test: Data generated by mitomycin C and methyl methanesulfonate. Mutat Res 444, 159–166. Nishikawa, T., Haresaku, M., Fukushima, A., Nakamura, T., Adachi, K., Masuda, M., and Hayashi, M. (2002). Further evaluation of an in vivo micronucleus test on rat and mouse skin: Results with five skin carcinogens. Mutat Res 513, 93–102. Nohmi, T., Katoh, M., Suzuki, H., Matsui, M., Yamada, M., Watanabe, M., Suzuki, M., Horiya, N., Ueda, O., Shibuya, T., Ikeda, H., and Sofuni, T. (1996). A new transgenic mouse mutagenesis test system using Spi− and 6-thioguanine selections. Environ Mol Mutagen 28, 465–470. Nohmi, T., and Masumura, K. (2004). Gpt delta transgenic mouse: A novel approach for molecular dissection of deletion mutations in vivo. Adv Biophys 38, 97–121. Nohmi, T., Suzuki, M., Masumura, K., Yamada, M., Matsui, K., Ueda, O., Suzuki, H., Katoh, M., Ikeda, H., and Sofuni, T. (1999). Spi(−) selection: An efficient method to detect gamma-ray-induced deletions in transgenic mice. Environ Mol Mutagen 34, 9–15. Nohmi, T., Suzuki, T., and Masumura, K. (2000). Recent advances in the protocols of transgenic mouse mutation assays. Mutat Res 455, 191–215. Norbury, C. J., and Hickson, I. D. (2001). Cellular responses to DNA damage. Annu Rev Pharmacol Toxicol 41, 367–401. Norppa, H., Bonassi, S., Hansteen, I. L., Hagmar, L., Stromberg, U., Rossner, P., Boffetta, P., Lindholm, C., Gundy, S., Lazutka, J., Cebulska-Wasilewska, A., Fabianova, E., Sram, R. J., Knudsen, L. E., Barale, R., and Fucic, A. (2006). Chromosomal aberrations and SCEs as biomarkers of cancer risk. Mutat Res 600, 37–45. Obe, G., Pfeiffer, P., Savage, J. R., Johannes, C., Goedecke, W., Jeppesen, P., Natarajan, A. T., MartinezLopez, W., Folle, G. A., and Drets, M. E. (2002). Chromosomal aberrations: Formation, identification and distribution. Mutat Res 504, 17–36. OECD (1984). Genetic toxicology: Rodent dominant lethal test. OECD Guideline for the Testing of Chemicals 478, 1–6. OECD (1986a). Genetic toxicology: Mouse heritable translocation assay. OECD Guideline for Testing of Chemicals 485, 1–6. OECD (1986b). Genetic toxicology: Mouse spot test. OECD Guideline for Testing of Chemicals 484, 1–4. OECD (1997a). In vitro mammalian cell gene mutation test. OECD Guideline for Testing of Chemicals 476, 1–10. OECD (1997b). Mammalian bone marrow chromosome aberration test. OECD Guideline for Testing of Chemicals 475, 1–8. OECD (1997c). Mammalian erythrocyte micronucleus test. OECD Guideline for Testing of Chemicals 474, 1–10. OECD (1997d). Unscheduled DNA synthesis (UDS) test with mammalian liver cells in vivo. OECD Guideline for Testing of Chemicals 486, 1–8.

354

CHAPTER 12


OECD (2009). Detailed review paper on transgenic rodent mutation assays. ENV/JM/MONO(2009)7 103, 1–281, http://www.olis.oecd.org/olis/2009doc.nsf/linkto/ENV-JM-MONO(2009)7. Ohyama, W., Gonda, M., Miyajima, H., Kondo, K., Noguchi, T., Yoshida, J., Hatakeyama, S., Watabe, E., Ueno, Y., Hayashi, M., and Tokumitsu, T. (2002). Collaborative validation study of the in vivo micronucleus test using mouse colonic epithelial cells. Mutat Res 518, 39–45. Olive, P. L. (2002). The comet assay. An overview of techniques. Methods Mol Biol 203, 179–194. Olive, P. L., Banath, J. P., and Durand, R. E. (1990a). Detection of etoposide resistance by measuring DNA damage in individual Chinese hamster cells. J Natl Cancer Inst 82, 779–783. Olive, P. L., Banath, J. P., and Durand, R. E. (1990b). Heterogeneity in radiation-induced DNA damage and repair in tumor and normal cells measured using the “comet” assay. Radiat Res 122, 86–94. Olive, P. L., and Durand, R. E. (2005). Heterogeneity in DNA damage using the comet assay. Cytometry A 66, 1–8. Olive, P. L., Frazer, G., and Banath, J. P. (1993). Radiation-induced apoptosis measured in TK6 human B lymphoblast cells using the comet assay. Radiat Res 136, 130–136. Olive, P. L., Wlodek, D., and Banath, J. P. (1991). DNA double-strand breaks measured in individual cells subjected to gel electrophoresis. Cancer Res 51, 4671–4676. Ostling, O., and Johanson, K. J. (1984). Microelectrophoretic study of radiation-induced DNA damages in individual mammalian cells. Biochem Biophys Res Commun 123, 291–298. Ottender, M., and Lutz, W. K. (1999). Correlation of DNA adduct levels with tumor incidence: carcinogenic potency of DNA adducts. Mutat Res 424, 237–247. Painter, R. B. (1980). A replication model for sister-chromatid exchange. Mutat Res 70, 337–341. Parsons, B. L., and Heflich, R. H. (1998). Detection of a mouse H-ras codon 61 mutation using a modified allele-specific competitive blocker PCR genotypic selection method. Mutagenesis 13, 581–588. Perry, P., and Evans, H. J. (1975). Cytological detection of mutagen–carcinogen exposure by sister chromatid exchange. Nature 258, 121–125. Perry, P., and Wolff, S. (1974). New Giemsa method for the differential staining of sister chromatids. Nature 251, 156–158. Phillips, D. H. (1997). Detection of DNA modifications by the 32P-postlabelling assay. Mutat Res 378, 1–12. Phillips, D. H., and Arlt, V. M. (2007). The 32P-postlabeling assay for DNA adducts. Nat Protoc 2, 2772–2781. Phillips, D. H., and Castegnaro, M. (1999). Standardization and validation of DNA adduct postlabelling methods: Report of interlaboratory trials and production of recommended protocols. Mutagenesis 14, 301–315. Phillips, D. H., Farmer, P. B., Beland, F. A., Nath, R. G., Poirier, M. C., Reddy, M. V., and Turteltaub, K. W. (2000). Methods of DNA adduct determination and their application to testing compounds for genotoxicity. Environ Mol Mutagen 35, 222–233. Phillips, D. H., Hewer, A., and Arlt, V. M. (2005). 32P-postlabeling analysis of DNA adducts. Methods Mol Biol 291, 3–12. Piegorsch, W. W., Lockhart, A. C., Carr, G. J., Margolin, B. H., Brooks, T., Douglas, G. R., Liegibel, U. M., Suzuki, T., Thybaud, V., van Delft, J. H., and Gorelick, N. J. (1997). Sources of variability in data from a positive selection lacZ transgenic mouse mutation assay: An interlaboratory study. Mutat Res 388, 249–289. Piegorsch, W. W., Margolin, B. H., Shelby, M. D., Johnson, A., French, J. E., Tennant, R. W., and Tindall, K. R. (1995). Study design and sample sizes for a lacI transgenic mouse mutation assay. Environ Mol Mutagen 25, 231–245. Pinkel, D., Thompson, L. H., Gray, J. W., and Vanderlaan, M. (1985). Measurement of sister chromatid exchanges at very low bromodeoxyuridine substitution levels using a monoclonal antibody in Chinese hamster ovary cells. Cancer Res 45, 5795–5798. Poirier, M. C. (1981). Antibodies to carcinogen–DNA adducts. J Natl Cancer Inst 67, 515–519. Poirier, M. C. (1993). Antisera specific for carcinogen–DNA adducts and carcinogen-modified DNA: Applications for detection of xenobiotics in biological samples. Mutat Res 288, 31–38. Poirier, M. C. (2004). Chemical-induced DNA damage and human cancer risk. Nat Rev Cancer 4, 630–637.

REFERENCES

355

Poirier, M. C., and Beland, F. A. (1992). DNA adduct measurements and tumor incidence during chronic carcinogen exposure in animal models: Implications for DNA adduct-based human cancer risk assessment. Chem Res Toxicol 5, 749–755. Poirier, M. C., Santella, R. M., and Weston, A. (2000). Carcinogen macromolecular adducts and their measurement. Carcinogenesis 21, 353–359. Pratt, I. S., and Barron, T. (2003). Regulatory recognition of indirect genotoxicity mechanisms in the European Union. Toxicol Lett 140–141, 53–62. Preston, R. J., Dean, B. J., Galloway, S., Holden, H., McFee, A. F., and Shelby, M. (1987). Mammalian in vivo cytogenetic assays. Analysis of chromosome aberrations in bone marrow cells. Mutat Res 189, 157–165. Randerath, K., and Randerath, E. (1994). 32P-postlabeling methods for DNA adduct detection: Overview and critical evaluation. Drug Metab Rev 26, 67–85. Randerath, K., Reddy, M. V., and Gupta, R. C. (1981). 32P-labeling test for DNA damage. Proc Natl Acad Sci USA 78, 6126–6129. Rapp, A., Hausmann, M., and Greulich, K. O. (2005). The Comet-FISH technique: A tool for detection of specific DNA damage and repair. Methods Mol Biol 291, 107–119. Reddy, M. V. (2000). Methods for testing compounds for DNA adduct formation. Regul Toxicol Pharmacol 32, 256–263. Reddy, M. V., Gupta, R. C., Randerath, E., and Randerath, K. (1984). 32P-postlabeling test for covalent DNA binding of chemicals in vivo: Application to a variety of aromatic carcinogens and methylating agents. Carcinogenesis 5, 231–243. Reddy, M. V., and Randerath, K. (1986). Nuclease P1-mediated enhancement of sensitivity of 32P-postlabeling test for structurally diverse DNA adducts. Carcinogenesis 7, 1543–1551. Ren, L., Yang, J. P., and Zhang, H. X. (1991). Use of the cytokinesis-block micronucleus method in mouse splenocytes. Mutat Res 262, 119–124. Richold, M., Chandley, A., Ashby, J., Gatehouse, D. G., Bootman, J., and Henderson, L. (1990). In vivo cytogenetic assays. In Basic Mutagenicity Tests, UKEMS Recommended Procedures, UKEMS Subcommittee on Guidelines for Mutagenicity Testing, Report part I revised, Kirkland, D. J., ed., Cambridge University Press, Cambridge, pp. 115–141. Rojas, E., Lopez, M. C., and Valverde, M. (1999). Single cell gel electrophoresis assay: Methodology and applications. J Chromatogr B Biomed Sci Appl 722, 225–254. Romagna, F., and Staniforth, C. D. (1989). The automated bone marrow micronucleus test. Mutat Res 213, 91–104. Rosenkranz, H. S., and Cunningham, A. R. (2000). The high production volume chemical challenge program: The relevance of the in vivo micronucleus assay. Regul Toxicol Pharmacol 31, 182–189. Rundle, A. (2006). Carcinogen–DNA adducts as a biomarker for cancer risk. Mutat Res 600, 23–36. Russell, L. B. (1977). Validation of the in vivo somatic mutation method in the mouse as a prescreen for germinal point mutations. Arch Toxicol 38, 75–85. Russell, L. B., and Major, M. H. (1957). Radiation-induced presumed somatic mutations in the house mouse. Genetics 42, 161–175. Russell, L. B., and Matter, B. E. (1980). Whole-mammal mutagenicity tests: Evaluation of five methods. Mutat Res 75, 279–302. Russo, A. (2000). In vivo cytogenetics: Mammalian germ cells. Mutat Res 455, 167–189. Salamone, M., Heddle, J., Stuart, E., and Katz, M. (1980). Towards an improved micronucleus test: studies on 3 model agents, mitomycin C, cyclophosphamide and dimethylbenzanthracene. Mutat Res 74, 347–356. Santella, R. M. (1999). Immunological methods for detection of carcinogen–DNA damage in humans. Cancer Epidemiol Biomarkers Prev 8, 733–739. Santos, S. J., Singh, N. P., and Natarajan, A. T. (1997). Fluorescence in situ hybridization with comets. Exp Cell Res 232, 407–411. Sasaki, Y. F., Izumiyama, F., Nishidate, E., Matsusaka, N., and Tsuda, S. (1997a). Detection of rodent liver carcinogen genotoxicity by the alkaline single-cell gel electrophoresis (comet) assay in multiple mouse organs (liver, lung, spleen, kidney, and bone marrow). Mutat Res 391, 201–214.

356

CHAPTER 12


Sasaki, Y. F., Izumiyama, F., Nishidate, E., Ohta, T., Ono, T., Matsusaka, N., and Tsuda, S. (1997b). Simple detection of in vivo genotoxicity of pyrimethamine in rodents by the modified alkaline singlecell gel electrophoresis assay. Mutat Res 392, 251–259. Sasaki, Y. F., Nishidate, E., Izumiyama, F., Matsusaka, N., and Tsuda, S. (1997c). Simple detection of chemical mutagens by the alkaline single-cell gel electrophoresis (Comet) assay in multiple mouse organs (liver, lung, spleen, kidney, and bone marrow). Mutat Res 391, 215–231. Sasaki, Y. F., Sekihashi, K., Izumiyama, F., Nishidate, E., Saga, A., Ishida, K., and Tsuda, S. (2000). The comet assay with multiple mouse organs: Comparison of comet assay results and carcinogenicity with 208 chemicals selected from the IARC monographs and U.S. NTP Carcinogenicity Database. Crit Rev Toxicol 30, 629–799. Sasaki, Y. F., Tsuda, S., Izumiyama, F., and Nishidate, E. (1997d). Detection of chemically induced DNA lesions in multiple mouse organs (liver, lung, spleen, kidney, and bone marrow) using the alkaline single cell gel electrophoresis (comet) assay. Mutat Res 388, 33–44. Sauvaigo, S., Serres, C., Signorini, N., Emonet, N., Richard, M. J., and Cadet, J. (1998). Use of the singlecell gel electrophoresis assay for the immunofluorescent detection of specific DNA damage. Anal Biochem 259, 1–7. Sawyer, T. W., Gill, R. D., Smith-Oliver, T., Butterworth, B. E., and DiGiovanni, J. (1988). Measurement of unscheduled DNA synthesis in primary cultures of adult mouse epidermal keratinocytes. Carcinogenesis 9, 1197–1202. Schiestl, R. H., Aubrecht, J., Khogali, F., and Carls, N. (1997). Carcinogens induce reversion of the mouse pink-eyed unstable mutation. Proc Natl Acad Sci USA 94, 4576–4581. Schmid, W. (1975). The micronucleus test. Mutat Res 31, 9–15. Schunck, C., Johannes, T., Varga, D., Lorch, T., and Plesch, A. (2004). New developments in automated cytogenetic imaging: Unattended scoring of dicentric chromosomes, micronuclei, single cell gel electrophoresis, and fluorescence signals. Cytogenet Genome Res 104, 383–389. Searle, A. G. (1977). The use of pigment loci for detecting reverse mutations in somatic cells of mice. Arch Toxicol 38, 105–108. Shaposhnikov, S. A., Salenko, V. B., Brunborg, G., Nygren, J., and Collins, A. R. (2008). Single-cell gel electrophoresis (the comet assay): Loops or fragments? Electrophoresis 29, 3005–3012. Shelby, M. D. (1996). Selecting chemicals and assays for assessing mammalian germ cell mutagenicity. Mutat Res 352, 159–167. Shelby, M. D., and Witt, K. L. (1995). Comparison of results from mouse bone marrow chromosome aberration and micronucleus tests. Environ Mol Mutagen 25, 302–313. Shelby, M. D., and Zeiger, E. (1990). Activity of human carcinogens in the Salmonella and rodent bonemarrow cytogenetics tests. Mutat Res 234, 257–261. Shibutani, S., Kim, S. Y., and Suzuki, N. (2006). 32P-postlabeling DNA damage assays: PAGE, TLC, and HPLC. Methods Mol Biol 314, 307–321. Shuck, S. C., Short, E. A., and Turchi, J. J. (2008). Eukaryotic nucleotide excision repair: From understanding mechanisms to influencing biology. Cell Res 18, 64–72. Singh, N. P., McCoy, M. T., Tice, R. R., and Schneider, E. L. (1988). A simple technique for quantitation of low levels of DNA damage in individual cells. Exp Cell Res 175, 184–191. Singh, N. P., Stephens, R. E., and Schneider, E. L. (1994). Modifications of alkaline microgel electrophoresis for sensitive detection of DNA damage. Int J Radiat Biol 66, 23–28. Singh, R., and Farmer, P. B. (2006). Liquid chromatography–electrospray ionization–mass spectrometry: The future of DNA adduct detection. Carcinogenesis 27, 178–196. Speit, G. (1984). Considerations on the mechanism of differential Giemsa staining of BrdU-substituted chromosomes. Hum Genet 67, 264–269. Speit, G., and Haupter, S. (1985). On the mechanism of differential Giemsa staining of bromodeoxyuridine-substituted chromosomes. II. Differences between the demonstration of sister chromatid differentiation and replication patterns. Hum Genet 70, 126–129. Spencer, T., and Butler, L. J. (1987). A rapid modified fluorescence-plus-Giemsa technique for sister chromatid exchanges and DNA replication patterns. Cytobios 50, 145–150. Stiegler, G. L., and Stillwell, L. C. (1993). Big Blue transgenic mouse lacI mutation analysis. Environ Mol Mutagen 22, 127–129. Storchova, Z., and Kuffer, C. (2008). The consequences of tetraploidy and aneuploidy. J Cell Sci 121, 3859–3866.

REFERENCES

357

Storchova, Z., and Pellman, D. (2004). From polyploidy to aneuploidy, genome instability and cancer. Nat Rev Mol Cell Biol 5, 45–54. Strickland, P. T., and Boyle, J. M. (1984). Immunoassay of carcinogen-modified DNA. Prog Nucleic Acid Res Mol Biol 31, 1–58. Suzuki, H., Shirotori, T., and Hayashi, M. (2004). A liver micronucleus assay using young rats exposed to diethylnitrosamine: Methodological establishment and evaluation. Cytogenet Genome Res 104, 299–303. Swenberg, J. A., Fryar-Tita, E., Jeong, Y. C., Boysen, G., Starr, T., Walker, V. E., and Albertini, R. J. (2008). Biomarkers in toxicology and risk assessment: informing critical dose–response relationships. Chem Res Toxicol 21, 253–265. Swiger, R. R., Cosentino, L., Masumura, K. I., Nohmi, T., and Heddle, J. A. (2001). Further characterization and validation of gpt delta transgenic mice for quantifying somatic mutations in vivo. Environ Mol Mutagen 37, 297–303. Swiger, R. R., Cosentino, L., Shima, N., Bielas, J. H., Cruz-Munoz, W., and Heddle, J. A. (1999). The cII locus in the MutaMouse system. Environ Mol Mutagen 34, 201–207. Tao, K. S., and Heddle, J. A. (1994). The accumulation and persistence of somatic mutations in vivo. Mutagenesis 9, 187–191. Tao, K. S., Urlando, C., and Heddle, J. A. (1993a). Comparison of somatic mutation in a transgenic versus host locus. Proc Natl Acad Sci USA 90, 10681–10685. Tao, K. S., Urlando, C., and Heddle, J. A. (1993b). Mutagenicity of methyl methanesulfonate (MMS) in vivo at the Dlb-1 native locus and a lacI transgene. Environ Mol Mutagen 22, 293–296. Tates, A. D., Neuteboom, I., and Hofker, M. (1980). A micronucleus technique for detecting clastogenic effects of mutagens/carcinogens (DEN, DMN) in hepatocytes of rat liver in vivo. Mutat Res 74, 11–20. Tates, A. D., van Dam, F. J., de Zwart, F. A., van Teylingen, C. M., and Natarajan, A. T. (1994). Development of a cloning assay with high cloning efficiency to detect induction of 6-thioguanineresistant lymphocytes in spleen of adult mice following in vivo inhalation exposure to 1,3-butadiene. Mutat Res 309, 299–306. Terashima, I., Suzuki, N., and Shibutani, S. (2002). 32P-Postlabeling/polyacrylamide gel electrophoresis analysis: Application to the detection of DNA adducts. Chem Res Toxicol 15, 305–311. Thybaud, V., Dean, S., Nohmi, T., de Boer, J., Douglas, G. R., Glickman, B. W., Gorelick, N. J., Heddle, J. A., Heflich, R. H., Lambert, I., Martus, H. J., Mirsalis, J. C., Suzuki, T., and Yajima, N. (2003). In vivo transgenic mutation assays. Mutat Res 540, 141–151. Tice, R. R., Agurell, E., Anderson, D., Burlinson, B., Hartmann, A., Kobayashi, H., Miyamae, Y., Rojas, E., Ryu, J. C., and Sasaki, Y. F. (2000). Single cell gel/comet assay: Guidelines for in vitro and in vivo genetic toxicology testing. Environ Mol Mutagen 35, 206–221. Tice, R. R., Hayashi, M., MacGregor, J. T., Anderson, D., Blakey, D. H., Holden, H. E., Kirsch-Volders, M., Oleson, F. B., Jr., Pacchierotti, F., Preston, R. J., et al. (1994). Report from the working group on the in vivo mammalian bone marrow chromosomal aberration test. Mutat Res 312, 305–312. Tischfield, J. A., Engle, S. J., Gupta, P. K., Bye, S., Boyadjiev, S., Shao, C., O’Neill, P., Albertini, R. J., Stambrook, P. J., and Sahota, A. S. (1994). Germline and somatic mutation at the APRT locus of mice and man. Adv Exp Med Biol 370, 661–664. Tompkins, E. M., Farmer, P. B., Lamb, J. H., Jukes, R., Dingley, K., Ubick, E., Turteltaub, K. W., Martin, E. A., and Brown, K. (2006). A novel 14C-postlabeling assay using accelerator mass spectrometry for the detection of O6-methyldeoxy-guanosine adducts. Rapid Commun Mass Spectrom 20, 883–891. Torous, D. K., Dertinger, S. D., Hall, N. E., and Tometsko, C. R. (2000). Enumeration of micronucleated reticulocytes in rat peripheral blood: A flow cytometric study. Mutat Res 465, 91–99. Torous, D. K., Hall, N. E., Illi-Love, A. H., Diehl, M. S., Cederbrant, K., Sandelin, K., Ponten, I., Bolcsfoldi, G., Ferguson, L. R., Pearson, A., Majeska, J. B., Tarca, J. P., Hynes, G. M., Lynch, A. M., McNamee, J. P., Bellier, P. V., Parenteau, M., Blakey, D., Bayley, J., van der Leede, B. J., Vanparys, P., Harbach, P. R., Zhao, S., Filipunas, A. L., Johnson, C. W., Tometsko, C. R., and Dertinger, S. D. (2005). Interlaboratory validation of a CD71-based flow cytometric method (Microflow) for the scoring of micronucleated reticulocytes in mouse peripheral blood. Environ Mol Mutagen 45, 44–55. Torous, D. K., Hall, N. E., Murante, F. G., Gleason, S. E., Tometsko, C. R., and Dertinger, S. D. (2003). Comparative scoring of micronucleated reticulocytes in rat peripheral blood by flow cytometry and microscopy. Toxicol Sci 74, 309–314.

358

CHAPTER 12


Tucker, J. D., Auletta, A., Cimino, M. C., Dearfield, K. L., Jacobson-Kram, D., Tice, R. R., and Carrano, A. V. (1993). Sister-chromatid exchange: Second report of the Gene-Tox Program. Mutat Res 297, 101–180. Turteltaub, K. W., and Dingley, K. H. (1998). Application of accelerated mass spectrometry (AMS) in DNA adduct quantification and identification. Toxicol Lett 102–103, 435–439. Tweats, D. J., Blakey, D., Heflich, R. H., Jacobs, A., Jacobsen, S. D., Morita, T., Nohmi, T., O’Donovan, M. R., Sasaki, Y. F., Sofuni, T., and Tice, R. (2007). Report of the IWGT working group on strategies and interpretation of regulatory in vivo tests I. Increases in micronucleated bone marrow cells in rodents that do not indicate genotoxic hazards. Mutat Res 627, 78–91. Uiterdijk, H. G., Ponder, B. A., Festing, M. F., Hilgers, J., Skow, L., and Van Nie, R. (1986). The gene controlling the binding sites of Dolichos biflorus agglutinin, Dlb-1, is on chromosome 11 of the mouse. Genet Res 47, 125–129. van Dam, F. J., Natarajan, A. T., and Tates, A. D. (1992). Use of a T-lymphocyte clonal assay for determining HPRT mutant frequencies in individual rats. Mutat Res 271, 231–242. Van Sloun, P. P., Wijnhoven, S. W., Kool, H. J., Slater, R., Weeda, G., van Zeeland, A. A., Lohman, P. H., and Vrieling, H. (1998). Determination of spontaneous loss of heterozygosity mutations in Aprt heterozygous mice. Nucleic Acids Res 26, 4888–4894. Vanhauwaert, A., Vanparys, P., and Kirsch-Volders, M. (2001). The in vivo gut micronucleus test detects clastogens and aneugens given by gavage. Mutagenesis 16, 39–50. Vijg, J., and Douglas, G. R. (1996). Bacteriophage lambda and plasmid lacZ transgenic mice for studying mutations in vivo. In Technologies for Detection of DNA Damage and Mutations, Pfeifer, G. P., ed., Plenum Press, New York, pp. 391–410. Wakata, A., Miyamae, Y., Sato, S., Suzuki, T., Morita, T., Asano, N., Awogi, T., Kondo, K., and Hayashi, M. (1998). Evaluation of the rat micronucleus test with bone marrow and peripheral blood: Summary of the 9th collaborative study by CSGMT/JEMS. MMS. Collaborative Study Group for the Micronucleus Test. Environmental Mutagen Society of Japan. Mammalian Mutagenicity Study Group. Environ Mol Mutagen 32, 84–100. Walker, V. E., Jones, I. M., Crippen, T. L., Meng, Q., Walker, D. M., Bauer, M. J., Reilly, A. A., Tates, A. D., Nakamura, J., Upton, P. B., and Skopek, T. R. (1999). Relationships between exposure, cell loss and proliferation, and manifestation of Hprt mutant T cells following treatment of preweanling, weanling, and adult male mice with N-ethyl-N-nitrosourea. Mutat Res 431, 371–388. Waters, M. D., Stack, H. F., Jackson, M. A., and Bridges, B. A. (1993). Hazard identification: Efficiency of short-term tests in identifying germ cell mutagens and putative nongenotoxic carcinogens. Environ Health Perspect 101(Suppl 3), 61–72. Waters, M. D., Stack, H. F., Jackson, M. A., Bridges, B. A., and Adler, I. D. (1994). The performance of short-term tests in identifying potential germ cell mutagens: A qualitative and quantitative analysis. Mutat Res 341, 109–131. Weaver, J. L., and Torous, D. (2000). Flow cytometry assay for counting micronucleated erythrocytes: development process. Methods 21, 281–287. Weston, A. (1993). Physical methods for the detection of carcinogen-DNA adducts in humans. Mutat Res 288, 19–29. Weston, A., Rowe, M. L., Manchester, D. K., Farmer, P. B., Mann, D. L., and Harris, C. C. (1989). Fluorescence and mass spectral evidence for the formation of benzo[a]pyrene anti-diol-epoxide–DNA and –hemoglobin adducts in humans. Carcinogenesis 10, 251–257. Whong, W. Z., Stewart, J. D., and Ong, T. (1992). Comparison of DNA adduct detection between two enhancement methods of the 32P-postlabelling assay in rat lung cells. Mutat Res 283, 1–6. Wiklund, S. J., and Agurell, E. (2003). Aspects of design and statistical analysis in the Comet assay. Mutagenesis 18, 167–175. Wild, C. P. (1990). Antibodies to DNA alkylation adducts as analytical tools in chemical carcinogenesis. Mutat Res 233, 219–233. Wilson, D. M., 3rd, and Thompson, L. H. (2007). Molecular mechanisms of sister-chromatid exchange. Mutat Res 616, 11–23. Winton, D. J., Gooderham, N. J., Boobis, A. R., Davies, D. S., and Ponder, B. A. (1990). Mutagenesis of mouse intestine in vivo using the Dlb-1 specific locus test: Studies with 1,2-dimethylhydrazine,

REFERENCES

359

dimethylnitrosamine, and the dietary mutagen 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline. Cancer Res 50, 7992–7996. Winton, D. J., and Ponder, B. A. (1990). Stem-cell organization in mouse small intestine. Proc Biol Sci 241, 13–18. Witt, K. L., Livanos, E., Kissling, G. E., Torous, D. K., Caspary, W., Tice, R. R., and Recio, L. (2008). Comparison of flow cytometry- and microscopy-based methods for measuring micronucleated reticulocyte frequencies in rodents treated with nongenotoxic and genotoxic chemicals. Mutat Res 649, 101–113. Wojcik, A., Bruckmann, E., and Obe, G. (2004). Insights into the mechanisms of sister chromatid exchange formation. Cytogenet Genome Res 104, 304–309. Wyborski, D. L., Malkhosyan, S., Moores, J., Perucho, M., and Short, J. M. (1995). Development of a rat cell line containing stably integrated copies of a lambda/lacI shuttle vector. Mutat Res 334, 161–165. Xue, W., and Warshawsky, D. (2005). Metabolic activation of polycyclic and heterocyclic aromatic hydrocarbons and DNA damage: A review. Toxicol Appl Pharmacol 206, 73–93. Yuspa, S. H., and Poirier, M. C. (1988). Chemical carcinogenesis: from animal models to molecular models in one decade. Adv Cancer Res 50, 25–70. Zimmer, D. M., Harbach, P. R., Mattes, W. B., and Aaron, C. S. (1999). Comparison of mutant frequencies at the transgenic lambda LacI and cII/cI loci in control and ENU-treated Big Blue mice. Environ Mol Mutagen 33, 249–256.

PART

IV

ASSESSING THE HUMAN RELEVANCE OF CHEMICALINDUCED TUMORS

CH A P TE R

13

FRAMEWORK ANALYSIS FOR DETERMINING MODE OF ACTION AND HUMAN RELEVANCE R. Julian Preston

13.1.

INTRODUCTION

The overall aim of a cancer risk assessment is to characterize the risk to humans from environmental exposures.* This risk characterization includes both qualitative and quantitative components and relies on the development of separate hazard, dose–response and exposure assessments. The specific approach currently used by the U.S. Environmental Protection Agency (EPA) can be found in its Guidelines for Carcinogen Risk Assessment (EPA 2005). A similar approach is applied by other national and international organizations. In general terms, the risk characterization summarizes, in a narrative form, the analyses of hazard, dose–response, and exposure assessment. These three assessments are summarized in light of “the extent and weight of evidence, major points of interpretation and rationale for their selection, strengths and weaknesses of the evidence and the analysis, and [a discussion] of alterative conclusions and uncertainties that deserve serious consideration” (EPA 2000). This summary serves as the starting materials for the overall risk characterization process that completes the risk assessment. This chapter will concentrate on a specific feature of this risk characterization process, namely the importance of developing approaches for incorporating mechanistic data into the hazard, dose– response, and exposure assessments to reduce uncertainties in the process and thereby reduce the reliance on default factors that are used in the absence of reliable data. Given that the risk characterization is for the estimation of risks to humans from low, environmental exposures, then the issues that cover the necessary defaults are as follows (see Part I, this volume): *This chapter has been reviewed in accordance with the policy of the U.S. Environmental Protection Agency and approved for publication, although it does not necessarily reflect Agency policy.


363

364

CHAPTER 13 FRAMEWORK ANALYSIS FOR DETERMINING MODE OF ACTION

• Is the presence or absence of effects observed in a human population predictive of effects in another exposed human population? • Is the presence or absence of effects observed in an animal population predictive of effects in exposed humans? • How do metabolic pathways relate across species and among different age groups and between sexes in humans? • How do toxicokinetic processes relate across species and among different age groups and between sexes in humans? • What is the relationship between the observed dose–response relationship to the relationship at lower doses? These issues are ones of extrapolation and as noted by Preston (2005) such extrapolations are “the Achilles heel of risk assessment” (Preston 2005). The U.S. EPA, The International Program on Chemical Safety (IPCS), and The International Life Sciences Institute (ILSI), for example, have proposed a framework based on the mode of action of a chemical, the key events that define a particular mode of action, and a human relevance framework for assessing the plausibility of an animal mode of action to humans. It is this approach that will be described and discussed in this chapter.

13.2. FRAMEWORK ANALYSIS: MODE OF ACTION AND KEY EVENTS 13.2.1.

Definitions

The following definitions are taken from the U.S. EPA Guidelines for Carcinogen Risk Assessment (EPA 2005). Mode of action (MOA) is defined as “a sequence of key events and processes, starting with interaction of an agent with a cell, proceeding through operational and anatomical changes, and resulting in cancer formation.” MOA is contrasted with Mechanism of Action, which implies a more detailed understanding and description of key events, often at the molecular level, than for MOA. Examples of MOA are DNA-reactivity, mitogenicity, inhibition of cell death, cytotoxicity with regenerative cell proliferation, immune suppression, and epigenetic effects, such as changes in gene expression and DNA methylation patterns. A key event is an “empirically observable precursor step that is itself a necessary element of the mode of action or is a biologically based marker for such an element.” In this regard and for this chapter, a biomarker is considered to be a surrogate marker of exposure or an early biological marker of effect (e.g., mutations in reporter genes, total chromosome alterations). In contrast, a biological marker of effect that is itself a key event along the pathway from a normal cell to a transformed one is described as a bioindicator (e.g., mutation in a critical gene for cancer, cancerspecific chromosome translocation). This distinction is useful for considering those cellular events that can be used only in a qualitative way for predicting tumor responses (biomarkers) and those that can be both qualitative and quantitative endpoints in a tumor dose–response assessment (bioindicators).

13.2. FRAMEWORK ANALYSIS: MODE OF ACTION AND KEY EVENTS

365

A point of departure (POD) is a point on a dose–response curve at which the range of data is extended from the observable range to lower dose ranges by extrapolation. Such extrapolation can be by default linear, predicted linear, or predicted nonlinear (to also include a threshold).

13.2.2. An Overview of the Framework for Analyzing Mode of Action The EPA Cancer Guidelines (EPA 2005) provide an analytical framework for assessing if available evidence supports a predicted MOA for a particular agent. As noted by Wiltse and Dellarco (2000), this “framework is based on considerations for causality in epidemiological investigations originally articulated by Hill (1965) but later modified by others and extended to experimental studies (DHHS, 1982; Faustman et al., 1997)” (Wiltse and Dellarco 2000). In outline, this framework first requires a description of a postulated MOA (or MOAs). This postulated MOA is then queried by addressing the pertinent available empirical data and experimental observations. These specific queries are: • • • • • •

What are the key events that lead to the postulated MOA? What is the strength, consistency, and specificity of association? What are the dose–response relationships? What are the temporal relationships? What is the biological plausibility and coherence? Are there other MOAs that are supportable by the available data?

Each of these queries will be discussed in more detail in Section 13.2.4.1.

13.2.3. Framework for Assessing Human Relevance of Animal MOA The majority of the tumor data available for conducting cancer risk assessments for exposure to environmental chemicals come from 2-year cancer bioassays using rats and mice. Thus, a MOA based on key events is inevitably developed for laboratory animals and not humans. There are, of course, a few exceptions for which human tumor data are available (NTP 2005). These human data are generally used together with rodent tumor data as part of dose–response characterization. Thus, the need in all cases is to demonstrate that the animal MOA is plausible in humans. This can be accomplished by use of a human relevance framework (described below in this section and in Table 13.1 and in Figure 13.1). It is of note that this considerable reliance on laboratory animal data for risk assessment purposes for environmental chemicals is in sharp contrast to the situation with ionizing radiation. The cancer risk estimates for ionizing radiation (X rays and γ rays) are based to a very great extent on human tumor data obtained from the Life Stage Study (LSS) of the atomic bomb survivors in Hiroshima and Nagasaki, Japan

366


TABLE 13.1.

Framework for Evaluation of an Animal MOA

1. Postulated MOA. Brief description of the sequence of measured effects, starting with chemical administration, to cancer formation at a given site. 2. Key events. Clear description of each of the key events (measurable parameters) that are thought to underlie the MOA. 3. Dose–response relationships. Dose-response relationships identified for each key event, and comparisons presented of dose–response relationships among key events and with cancer. 4. Temporal association. Sequence of key events over time that lead to tumor formation. 5. Strength, consistency, and specificity of association of key events and tumor response. Complete assessment and presentation of the relationships among the key events, precursor lesions, and tumors. Portrayal of the consistency of observations across studies of different designs. 6. Biological plausibility and coherence. Determination of whether key events and the sequence of events are consistent with current biological thinking, both regarding carcinogenesis in general and for the specific chemical under review. 7. Other MOAs. Alternative MOAs that may be applicable for the chemical under review. Comparison of their likelihood vis-à-vis the proposed MOA. 8. Conclusion about the MOA. Overall indication of the level of confidence in the postulated MOA. 9. Uncertainties, inconsistencies, and data gaps. Identification of information deficiencies in the case; description of inconsistent findings in the data at large; evaluation of uncertainties; proposal of pointed research that could significantly inform the case. Source: Meek et al. (2003); adapted from EPA (1999) and Sonich-Mullin et al. (2001).

Animal MOA (and related endpoints) not relevant to humans

Animal MOA relevant or potentially relevant to humans

Is the weight of evidence sufficient to establish the MOA in animals? No Yes

•MOA: Data insufficient to characterize animal MOA

Are key events in the animal MOA plausible in humans? •MOA: Species-specific protein •MOA: Species-specific hormone suppression

No

Yes

•MOA: Species-specific enhanced hormone clearance rate

No need to continue risk assessment for this endpoint

No

Taking into account kinetic and dynamic factors, are key events in the animal MOA plausible in humans?

Yes

•MOA: Comparable cytotoxicity and cell proliferation response •MOA: Comparable tissue response (different animalhuman exposure potential) Continue risk assessment, including dose-response human exposure analysis, and risk characterization

Figure 13.1. General schematic illustrating how the Human Relevance Framework can be used to assess whether or not an animal MOA has a human counterpart, thereby indicating if a quantitative risk assessment is required. [Adapted from Meek et al., (2003).]


367

(Preston et al. 2007, 2008). This study has been ongoing for over 50 years and utilizes a large study cohort. An extremely thorough dose reconstruction program was used to estimate individual doses (reviewed in Cullings et al. (2006)). Data on cancer mortality and cancer incidence have been obtained for a very wide range of age groups at the time of the bombing, including in utero. Additional epidemiological data for medically and occupationally exposed groups are used for enhancing the lower dose tumor database. Animal tumor data and mechanistic studies are not used directly in the risk assessment process, but really only to provide general support to the conclusions from the human tumor data (ICRP 2008; NRC 2006). Risk estimates at low dose and low dose rates are currently obtained by extrapolation of the dose–response curve for observed tumors over the medium to high dose ranges. The extrapolation used is a linear no threshold (LNT) one. Thus, for radiation an MOA and key events could be developed for human tumors without the need for a human relevance framework. The Risk Science Institute (RSI) of the International Life Sciences Institute (ILSI) developed a logical framework (Human Relevance Framework, HRF) for deciding if a mode of action by which a particular chemical induces tumors in an animal model could plausibly be acting in humans based upon the available human data and considerations of kinetic and dynamic factors (Meek et al. 2003). The HRF analysis has an initial focus on key events in the animal MOA (Table 13.1) and then considers the weight of evidence for the relevance to humans of the animal tumors being studied (Figure 13.1). This approach relies upon addressing three overriding questions: (1) Is the weight of evidence sufficient to establish the MOA in animals? (2) Are key events in the animal MOA plausible in humans? (3) Taking into account kinetic and dynamic factors, is the animal MOA plausible in humans? The outcome is presented in a Yes/No format. If the answer to the third question is No, the decision is “No need to continue risk assessment for this endpoint”; if the answer is Yes, the decision is “Continue risk assessment including dose–response, human exposure analysis, and risk characterization.” Further discussion of the components is presented in the Section 13.3.

13.2.4. Establishing and Applying Key Events in Support of MOA This section will build upon the overview information presented in the previous sections to lay out for known rodent carcinogens, what specific information is required, and how it is used to establish an MOA in animals and its relevance to humans. Specific examples are provided also for illustrating the application of the framework approach leading to risk characterization. 13.2.4.1. Is the Weight of Evidence Sufficient to Establish the MOA in Animals? The following information is used as a guide for evaluating each hypothesized carcinogenic MOA (EPA 2005; Meek et al. 2003) and is, in a sense, a corollary to the queries that were presented in the framework overview above.

368


There are a number of different MOAs (or combinations of MOAs) whereby a chemical exerts its carcinogenicity. In general, these can be characterized as DNAreactivity and nonDNA-reactivity [to include oxidative stress, cytotoxicity and associated regenerative proliferation, mitogenicity, receptor mediation, and epigenetic effects (e.g., changes in gene expression, DNA methylation, chromatin organization)]. This broad classification is pertinent to the subsequent discussion because a chemical with a so-called mutagenic MOA (used synonymously with DNA-reactive) is considered to follow a default linear dose–response extrapolation for tumors whereas chemicals with non-DNA-reactive MOAs are not considered to be a default linear in this regard (EPA 2005). 13.2.4.2. Key Events for Characterizing MOA. A critical component of MOA characterization is the establishment of a set of key events that define how a normal cell can be converted into a malignant one and ultimately to a metastatic tumor as a consequence of exposure to a DNA-reactive or non-DNA-reactive chemical carcinogen. The enormous enhancement of our knowledge of the cancer process over the past decade or so has greatly enhanced the process of identifying key events (Weinberg 2006). In general terms, carcinogenesis is a multistep process that requires an integrated set of genetic and epigenetic alterations to produce the cancer phenotype. This allows for key events to be described in terms of these steps, with any particular event being a specific genetic or other cellular change that characterizes the step. An example has been developed and applied by Preston and Williams (2005). The particular approach presented by Preston and Williams (2005) was for DNA-reactive carcinogens, although as shown here it is readily adaptable to nonDNA-reactive chemicals. The framework is the description of key events for tumor development and is shown in Table 13.2. There is an essential temporal sequence

TABLE 13.2.

Key Events for Tumor Development: DNA-Reactive MOAs

1. Exposure of target cells (e.g., stem cells) to ultimate DNA-reactive and mutagenic species; in some cases this requires metabolism. 2. Reaction with DNA in target cells to produce DNA damage. 3. Misreplication on damaged DNA template or misrepair of DNA damage. 4. Mutations in critical genes in replicating target cell. 5. These mutations result in initiation of new DNA/cell replication. 6. New cell replication leads to clonal expansion of mutant cells. 7. DNA replication can lead to further mutations in critical genes. 8. Imbalanced and uncontrolled clonal growth of mutant cells may lead to preneoplastic lesions. 9. Progression of preneoplastic cells results in emergence of overt neoplasms, solid tumors (which require neoangiogenesis), or leukemia. 10. Additional mutations in critical genes as a result of uncontrolled cell division results in malignant behavior. Note: Key events along the pathway to tumor development for DNA-reactive carcinogens can be assessed both qualitatively and quantitatively by experimental and human studies. For each of the chemicals selected for the case studies, the available data were matched to one or more of these key events to help establish a MOA for human cancer.


369

to these key events since each one (except the first) is dependent on the occurrence of the previous ones. This becomes of particular importance when considering quantitative risk assessments. For non-DNA-reactive chemicals, the only differences would be that there would be no need for binding of the parent chemical or a metabolite to DNA—all other steps could be the same. There might be additional differences in cases where a chemical does not interact with DNA directly or indirectly (for example, via reactive oxygen species). The DNA replication errors as a key event would still hold for non-DNA-reactive chemicals; it is just that the chemically produced mutations arise from an enhanced probability of replication errors as a result of increased cell proliferation in response to cytotoxicity as opposed to being from a more damaged DNA template, which is the case for DNA-reactive chemicals. The initial step in using key events in an MOA framework is a qualitative one, as described in Table 13.2, namely to match the key events for a particular chemical with those for a particular MOA (DNA-reactivity in Table 13.2). However, it is also possible to utilize key events to develop informative biomarkers of exposure and effect as well as bioindicators of disease outcome that can be utilized in a quantitative risk assessment process. A distinction is made between biomarkers and bioindicators because they can be used in quite different ways and for different purposes in a risk assessment context. As mentioned above, a biomarker is considered to be a surrogate marker of exposure or an early biological marker of effect (e.g., mutations in reporter genes, total chromosome alterations). In contrast, a biological marker of effect that is itself a key event along the pathway from a normal cell to a transformed one is described as a bioindicator (e.g., mutation in critical gene for cancer, cancer-specific chromosome translocation). Biomarkers can be used to inform the dose–response for tumors in a qualitative manner. Bioindicators can be used in a qualitative and quantitative way to inform tumor dose–response curves. Use of these biomarkers and bioindicators can make it feasible to characterize a dose–response curve at exposure levels below those at which increases in tumor frequency can be assessed. Recent advances in knowledge of the underlying mechanisms of carcinogenesis and the ever-increasing portfolio of whole genome molecular assay techniques have made it much more feasible to identify and select informative bioindicators of disease processes, especially cancer (Block et al. 2008; Conrad et al. 2008; Preston 2005). The emphasis for the key events in tumor development is on the essential ingredients for driving the process, namely alterations in critical genes (e.g., oncogenes and tumor suppressor genes) and enhanced cell proliferation. Carcinogenesis can be viewed as an evolutionary process that requires an accumulation of genomic alterations that then provide a substrate for selection of an advantageous phenotype that is usually related to a growth advantage (Cahill et al. 1999; Gatenby and Vincent 2003; Maley et al. 2004; Vincent and Gatenby 2008). However, it is not necessary to characterize the precise mechanism of tumor formation in response to an exposure to an environmental agent but rather to establish that certain key genetic and phenotypic changes take place. This difference between precise mechanism and key changes can be exemplified by considering two major cancer models. Fearon and

370


Vogelstein (1990) described a model for colorectal tumorigenesis in which each step along the multistep process from a normal cell to a malignant tumor was characterized at the phenotypic level as specific preneoplastic lesions and at the genome level by specific genetic alterations (gene mutations and chromosomal changes). In contrast, Hanahan and Weinberg in their “Hallmarks of Cancer” (Hanahan and Weinberg 2000) described tumors, irrespective of site or species, as acquiring six characteristics (self-sufficiency in growth signals, insensitivity to anti-growth signals, evading apoptosis, limitless replicative potential, sustained angiogenesis, and tissue invasion/metastasis) without the requirement to know precisely which genetic or epigenetic alteration resulted in the development of the characteristic. It is the latter approach which best exemplifies the application of the framework for MOA and key events. 13.2.4.3. Dose–Response Characteristics. This is the area of the risk assessment process that engenders the most debate. This is to a large extent because the need is to estimate the human risks at environmental exposure levels and yet there are few data at these exposure levels that can be used directly in a qualitative, let alone quantitative, way to support a particular form of dose–response curve (see Part VI, this volume). The approach has to rely upon extrapolation from tumor data (usually for rodents) to predict responses outside the range where tumor data themselves are available. This has in the past been done in a relatively pragmatic fashion, relying to a great extent on default positions for dose–response curve shape. With the advent of the incorporation of mechanistic data into the process, it is possible to utilize data on key events to enhance the process for interspecies, high to low dose, acute to chronic exposure extrapolations. This enhancement can be at a qualitative level (shape of dose–response curve) or a quantitative one, depending on the predictive value of the particular biological marker of the key event for tumor outcomes. The onus is on research investigators to develop linkages between the types of data they collect and their characteristics for use in a risk assessment framework. Such an approach will clearly help to reduce uncertainty in the risk assessment process and will lead to a reduction in the use of default approaches that have to be used in the absence of appropriate datasets. These considerations will be enhanced by the use of systems biology approaches, namely to treat normal tissues and altered ones as systems for comparison and selection of key event bioindicators (Edwards and Preston 2008; NRC 2007). Since the key events that define an animal MOA are in a temporal sequence, the dose–response curve for each successive key event can be viewed as the probability of converting one key event into the next one—for example, for converting a DNA adduct into a mutation by an error of replication. In the absence of any rate-limiting step, the overall probability of inducing a tumor from a chemical exposure is the integration of the probabilities for all key events. A ratelimiting step such as absence of exposure to the target cells (Table 13.2) would lead to a threshold dose–response curve because no other key events could be manifest. This is a simple example, but others are clearly feasible based upon, for example, fidelity of DNA repair processes or interacting (offsetting) gene expression changes.


371

13.2.4.4. Temporal Sequence of Key Events. As noted above, the key events represent a progression from a normal cell to a metastatic tumor, with each key event being dependent on the previous ones in order. Thus, there is a requirement for a temporal sequence to the induction of key events. Such a temporal sequence is most likely to be met when the exposures are chronic, given the propensity for continued induction of key events. 13.2.4.5. Additional MOAs. Even though a specific MOA has been established, based on the use of key events, it is possible that a particular chemical can operate through additional MOAs. This feature could impact the shape of the dose– response curve and the quantitative assessment of risk, depending on the nature of the multiple MOAs. There is, in fact, an expectation that chemicals will function through more than one MOA; for example, a mutagenic chemical is very likely also to be cytotoxic as a consequence of the induction of DNA damage—both features can enhance mutation rate above background and thus act additively or synergistically. 13.2.4.6. Examples of the Use of Key Events in an MOA Framework. There are a number of available examples for which the MOA framework has been used; published examples co-authored by the author of this chapter have been selected. (a) Dichloromethane (from Preston and Williams, 2005). The details for the conduct of the MOA analysis can be found in the publication. For this example, and indeed for many cases, it is possible to identify data for key events 1–3 (Table 13.2) and for key event 9 (carcinogenicity) and this is sufficient for describing a MOA as DNA-reactivity. In summary, Table 13.3 identifies the key events in the animal MOA for dichloromethane. This analysis leads to the conclusion that dichloromethane acts in rodents by a DNA-reactive MOA. The human relevance of this conclusion is considered in Section 13.3, paragraph (a).

TABLE 13.3.

Key Events in the Animal MOA for Dichloromethane

Key Events Metabolism by GSTTI-I (Key Event I) DNA damage induced (Key Event 2) Genotoxicity and mutagenicity (Key Event 3) Carcinogenicity (Key Event 9)

Evidence in Animals In mice metabolism by GSTTI-I in target tissues for tumor formation. In rats metabolism by GSTTI-I much lower in target tissues than for mice DNA–protein crosslink and single strand-breaks induced in mouse cells in vivo and in vitro but either not in rat cells or a lower frequency Reduced DNA damage in presence of GST-depleting agent Mutagenic in bacteria Genotoxic/mutagenic in mouse cells in vitro and in vivo Less genotoxic or nongenotoxic in rat cells in vitro and in vivo Carcinogenic in mice by inhalation (lung and liver); carcinogenic, at a reduced level, in rats by inhalation (mammary gland)

372


TABLE 13.4.

Key Events in the Carcinogenicity of 4-Aminobiphenyl in Animals

Key Event 1:

Key Event 2: Key Event 3: Key Event 9:

TABLE 13.5.

Key Event 1: Key Event 2: Key Event 3: Key Event 4: Key Event 9:

Metabolic activation (a) N-Hydroxylation (b) N-Esterification (gluconoride, acetyl, sulfate) (c) Hydrolysis to nitroniun ion DNA adduct formation (dG–C8, dA–C8, dG–N2) in pluripotent cells of target organs DNA mutation in critical gene(s) leading to cancer Cancer

Key Events in the Animal MOA for 1,3-Butadiene

Target cells (including bone marrow) exposed to ultimate DNA-reactive species. DNA adducts induced in tumor target tissues (liver, lung and testis), with dGNT and dAN6 being the most frequent. Mutations induced in target cells. Mutations induced in critical genes for cancer. Cancer.

(b) 4-Aminobiphenyl. A thorough case study has been developed by Cohen et al. (2006) for assessing the cancer MOA for 4-aminobiphenyl using the framework described above (Cohen et al. 2006). The key events identified in support of the MOA are presented in Table 13.4. Again, the human relevance is considered using the IPCS Human Relevance framework and is discussed in the Section 13.3, paragraph (b). (c) 1,3-Butadiene (Preston, 2007). The details for this example can be found in Preston (2007). An assessment was made of the key events defining a DNA-reactive MOA in rodents for 1,3-butadiene (with an emphasis on contrasting the much greater effectiveness in mice than in rats). These key events are summarized in Table 13.5. The production of these key events is sufficient to conclude that 1,3-butadiene acts via a DNA-reactive MOA. The human relevance is discussed in the original manuscript (Preston 2007) and in summary in the Section 13.3, paragraph(c).

13.3.

FRAMEWORK ANALYSIS: HUMAN RELEVANCE

Once a MOA for an animal model (most frequently rodent) has been developed, an assessment needs to be made to establish if this same MOA cannot be reasonably excluded or is plausible in humans. This has to be done for most chemical carcinogens because, as mentioned above, only a small minority of chemicals have been shown to be carcinogenic in humans. This plausibility was, until recently, conducted

13.3. FRAMEWORK ANALYSIS: HUMAN RELEVANCE

373

through a somewhat arbitrary approach largely involving a search for relevant data. More recently the process has been formalized through the development of a Human Relevance Framework (Meek et al. 2003) that builds upon the US EPA Cancer Guidelines (EPA 2005) and the IPCS MOA framework (Sonich-Mullin et al. 2001). The general approach is shown in Figure 13.1. The initial step is to use a weight of evidence approach that is based on key events, as discussed above, to establish the MOA in animals. On the assumption that there is sufficient evidence to establish the MOA, the key event approach is used to determine if the same MOA cannot reasonably be excluded is plausible in humans. The measure is “cannot reasonably be excluded” (Boobis et al. 2006) or “plausible” because there is generally much less data on key events for humans than rodents, and so it has been agreed that plausibility is sufficient. If the key events are reasonably excludable or are not plausible in humans, then the conclusion is that the chemical under review is rodentor species-specific. For example, it can be metabolized in a rodent kidney, but not in humans, to the ultimate carcinogenic species. On the assumption that the animal MOA is plausible in humans, it is necessary to take into account key toxicokinetic and toxicodynamic factors to further establish plausibility in humans. For example, is it possible in humans to achieve target tissue exposures to the parent compound or an active metabolite that are as high as those required in the animal model to produce the adverse outcome? If the conclusion is that there are no prohibitive kinetic or dynamic issues for plausibility of the animal MOA in humans, then the need is to proceed to the next steps of the risk assessment process: dose–response, human exposure analysis, and risk characterization. If there is clear evidence that there are kinetic or dynamic differences between the pertinent animal model and humans for the particular chemical being considered that would result in no response in humans at exposure levels much lower than those required in animals to produce the adverse outcome, then there is no need to continue the risk assessment for the outcome being assessed (i.e., cancer in the context of this review). This HRF has been applied to DNA-reactive and nonDNA-reactive carcinogens (Meek et al. 2003; Preston and Williams 2005) and to noncancer effects (Seed et al. 2005). It has also been applied to the three examples in Section 13.2.4.6 that were used to exemplify the use of the MOA framework and key events. (a) Dichloromethane (Preston and Williams 2005). A DNA-reactive MOA was demonstrated for dichloromethane carcinogenicity using key events 1–3 and 9. There were reported differences in metabolism, DNA damage, and genotoxicity between mice and rats, with rats being much less sensitive. When the same four key events were considered for human plausibility, it was found that humans were much more similar to rats as regards metabolism, DNA damage, and mutagenicity (Table 13.6). This led to a prediction that a DNA-reactive MOA was operational in mice for inducing tumors but that lower GSST1 levels and its distribution in humans (and rats), together with the weak genotoxicity and mutagenicity in vitro for human cells, would result in a lower carcinogenic potential in humans (and in rats) compared to mice. This conclusion was borne out by the available epidemiological data (and rat tumor studies).

374


TABLE 13.6.

Concordance Analysis of Key Events in Humans for Dichloromethane

Key Events Metabolism by GSTTI-I (Key Event I)

DNA damage induced (Key Event 2)

Genotoxicity and mutagenicity (Key Event 3)

Carcinogenicity (Key Event 9)

Evidence in Animals

Evidence in Humans

In mice metabolism by GSTTI-I in target tissues for tumor formation In rats metabolism by GSTTI-I much lower in target tissues than for mice DNA–protein crosslink and single strand-breaks induced in mouse cells in vivo and in vitro but either not in rat cells or at a lower frequency Reduced DNA damage in presence of GST-depleting agent Mutagenic in bacteria Genotoxic/mutagenic in mouse cells in vitro and in vivo Less genotoxic or nongenotoxic in rat cells in vitro and in vivo Carcinogenic in mice by inhalation—lung and liver Carcinogenic, at a reduced level, in rats by inhalation—mammary gland

Enzyme is present, at lower levels than in mouse target tissues, metabolism by GSTT1-1 predicted by PBPK models DNA–protein crosslinks not observed in human hepatocytes in vitro Low levels of GSTT1-1 suggest that DNA damage induction is plausible but at low levels Inconsistent results for mutagenicity in humans cells in vitro

Human epidemiological data suggest that risks associated with dichloromethane exposure, if any, are small and limited to rare cancers

TABLE 13.7. Concordance Evaluation of Key Events of 4-Aminobiphenyl-Induced Urinary Bladder Carcinogenesis between Species

Key Event 1. Metabolic activation to reactive electrophile 2. DNA adduct formation 3. Mutagenesis 4. Carcinoma

Mouse

Dog

Human

+ + + +

+ + + +

+ + + +

(b) 4-Aminobiphenyl. A DNA-reactive MOA was demonstrated for mice using key events 1–3 and 9. A concordance evaluation of the key events for 4aminobiphenyl-induced bladder carcinogenesis between species (mice, dogs, humans) demonstrated that all these key events were either observed and/or plausible in humans (Table 13.7). Quantitative differences among species do exist, but they do not exclude the DNA-reactive MOA in mice and dogs from being operational in humans.

13.4. FUTURE DIRECTIONS

375

(c) 1,3-Butadiene. For 1,3-butadiene a DNA-reactive MOA was established based upon key events 1–4 and 9. Using the human relevance framework, it can be established that these key events are plausible in humans, even though they have not necessarily been confirmed for target cells or for specific endpoints (Preston 2007). This is based, in part, on (a) the mutagenic effectiveness of butadiene and its metabolites in many in vivo and in vitro assay systems and (b) the link between butadieneinduced DNA adducts (that have been observed in butadiene exposed humans) and mutation spectra. It is reasonable to conclude, based on kinetic considerations, that the levels of active metabolites that would reach the target tissues in a majority of humans are probably unlikely to be sufficiently high to induce mutations. In summary, having established an animal MOA and human relevance for this MOA, it is appropriate to address dose–response assessment, human exposure analysis, and risk characterization. Thus, the purpose of the human relevance framework is to establish which chemicals (or chemical mixtures) should be considered for a quantitative risk assessment and which do not require further consideration because they present a minimal risk or no risk to humans. Several thoroughly worked examples are presented in Meek et al. (2003).

13.4.

FUTURE DIRECTIONS

The framework analysis for determining MOA and human relevance appears to be well-established and can continue to provide sound risk assessment guidance. What will change over a relatively short timeframe will be the ability to collect information that can be used to develop key events for describing the MOA in both laboratory animals and humans. A major reason for this is that there is available the capability of using genome-wide approaches for assessing both responses to exposures to environmental chemicals and for describing diseases at a molecular level (i.e., DNA, RNA, and protein changes) (see Chapters 22 and 23) (Chen et al. 2008; Edwards and Preston 2008). This whole genome assessment capability has allowed for responses of cells or organisms to environmental chemical exposures and disease processes to be characterized in terms of key events and toxicity pathways. This in turn has allowed for the development of much more informative biological indicators of response that can be used as surrogates for adverse outcomes, cancer in the present context. It is important to emphasize that it is most effective to use several bioindicators of disease outcome for helping define the shape of the dose–response curve for cancer or for defining human relevance, because it is unlikely, based on underlying mechanism of tumor formation, that any single indicator can define the doseresponse characteristics. Despite this caution, it remains frequently the case in molecular epidemiology studies for single biomarkers or bioindicators to be assessed. Another level of enhanced effort will be in the area of epidemiology, both traditional and molecular, and human in vitro research in support of the identification of key events in humans and of the human relevance of an animal MoA. The recently published National Research Council (NRC) Report, Toxicity Testing in the 21st Century: A Vision and a Strategy (NRC 2007), provides a set of research options for enhancing the use of in vitro test systems with an emphasis on human cells. This

376


appears to be a viable approach, with the proviso that maintaining some emphasis on in vivo laboratory animal studies remains essential. In this regard, there is an increasing emphasis on the use of informative animal models of human disease for establishing quantitative data for dose–response for a number of diseases, especially cancer (Jacks 2005). However, it needs to be emphasized that reduction in a reliance on extrapolation approaches will greatly reduce the uncertainty in a quantitative risk assessment. This can be achieved through the use of the informative and sensitive technologies that are available or are being developed for the detection and characterization of reliable disease markers (Conrad et al. 2008; Costa et al. 2008; Shim et al. 2008). It will be most interesting to follow the use of the Risk Assessment Framework described in this chapter for a broad range of chemical carcinogens and to see how new methods and data can enhance this use.

REFERENCES Block, T. M., Marrero, J., Gish, R. G., Sherman, M., London, W. T., Srivastava, S., and Wagner, P. D. (2008). The degree of readiness of selected biomarkers for the early detection of hepatocellular carcinoma: Notes from a recent workshop. Cancer Biomark 4, 19–33. Boobis, A. R., Cohen, S. M., Dellarco, V., McGregor, D., Meek, M. E., Vickers, C., Willcocks, D., and Farland, W. (2006). IPCS framework for analyzing the relevance of a cancer mode of action for humans. Crit Rev Toxicol 36, 781–792. Cahill, D. P., Kinzler, K. W., Vogelstein, B., and Lengauer, C. (1999). Genetic instability and darwinian selection in tumours. Trends Cell Biol 9, M57–M60. Chen, Y., Zhu, J., Lum, P. Y., Yang, X., Pinto, S., MacNeil, D. J., Zhang, C., Lamb, J., Edwards, S., Sieberts, S. K., Leonardson, A., Castellini, L. W., Wang, S., Champy, M. F., Zhang, B., Emilsson, V., Doss, S., Ghazalpour, A., Horvath, S., Drake, T. A., Lusis, A. J., and Schadt, E. E. (2008). Variations in DNA elucidate molecular networks that cause disease. Nature 452, 429–435. Cohen, S. M., Boobis, A. R., Meek, M. E., Preston, R. J., and McGregor, D. B. (2006). 4-Aminobiphenyl and DNA reactivity: Case study within the context of the 2006 IPCS Human Relevance Framework for Analysis of a cancer mode of action for humans. Crit Rev Toxicol 36, 803–819. Conrad, D. H., Goyette, J., and Thomas, P. S. (2008). Proteomics as a method for early detection of cancer: A review of proteomics, exhaled breath condensate, and lung cancer screening. J Gen Intern Med 23 (Suppl 1), 78–84. Costa, J. L., Meijer, G., Ylstra, B., and Caldas, C. (2008). Array comparative genomic hybridization copy number profiling: A new tool for translational research in solid malignancies. Semin Radiat Oncol 18, 98–104. Cullings, H. M., Fujita, S., Funamoto, S., Grant, E. J., Kerr, G. D., and Preston, D. L. (2006). Dose estimation for atomic bomb survivor studies: Its evolution and present status. Radiat Res 166, 219–254. Department of Health and Human Services (DHHS). (1982). The Health Consequences of Smoking: Cancer. A Report of the Surgeon General. Washington, DC, pp. 17–20 et seq. Edwards, S. W., and Preston, R. J. (2008). Systems biology and mode of action based risk assessment. Toxicol Sci. 106, 312–318. EPA (U.S. Environmental Protection Agency) (1999). Guidelines for Carcinogen Risk Assessment. Risk Assessment Forum. SAB review draft. U.S. Environmental Protection Agency, Washington, DC. http:// www.epa.gov/ncea/raf/crasab.htm. EPA (2000). Science Policy Council Handbook: Risk Characterization, Office of Science Policy, Office of Research and Development, US Environmental Protection Agency, Washington, D.C., EPA 100B-00-002, http://www.epa.gov/osa/spc/pdfs/rchandbk.pdf. EPA (2005). Guidelines for Carcinogen Risk Assessment, Risk Assessment Forum, US Environmental Protection Agency, Washington, D.C., EPA/630/P-03/001F, http://oaspub.epa.gov/eims/eimscomm. getfile?p_download_id=439797.

REFERENCES

377

Faustman, E. M., Ponce, R. A., Seeley, M. R., and Whittaker, S. G. (1997). Experimental approaches to evaluate mechanisms of developmental toxicity. In Handbook of Developmental Toxicology, Hood, R., ed. CRC Press, Boca Raton, FL, pp. 13–41. Fearon, E. R., and Vogelstein, B. (1990). A genetic model for colorectal tumorigenesis. Cell 61, 759–767. Gatenby, R. A., and Vincent, T. L. (2003). An evolutionary model of carcinogenesis. Cancer Res 63, 6212–6220. Hanahan, D., and Weinberg, R. A. (2000). The hallmarks of cancer. Cell 100, 57–70. Hill, A. B. (1965). The environment and disease: Association or causation? Proc R Soc Med 58, 295–300. ICRP (2008). ICRP Publication 103: Recommendations of the ICRP. Ann ICRP 37, 2–4. Jacks, T. (2005). Modeling cancer in the mouse. Harvey Lect 101, 1–19. Maley, C. C., Galipeau, P. C., Li, X., Sanchez, C. A., Paulson, T. G., and Reid, B. J. (2004). Selectively advantageous mutations and hitchhikers in neoplasms: p16 lesions are selected in Barrett’s esophagus. Cancer Res 64, 3414–3427. Meek, M. E., Bucher, J. R., Cohen, S. M., Dellarco, V., Hill, R. N., Lehman-McKeeman, L. D., Longfellow, D. G., Pastoor, T., Seed, J., and Patton, D. E. (2003). A framework for human relevance analysis of information on carcinogenic modes of action. Crit Rev Toxicol 33, 591–653. NRC (2006). Health Risks from Exposure to Low Levels of Ionizing Radiation: BEIR VII Phase 2, Board on Radiation Effects Research (BRER), National Research Council of the National Academies, Washington, D.C., http://books.nap.edu/openbook.php?record_id=11340&page=R1. NRC (2007). Toxicity Testing in the 21st Century: A Vision and a Strategy, Board on Environmental Studies and Toxicology (BEST), Institute for Laboratory Animal Research (ILAR), National Research Council of the National Academies, Washington, D.C., http://books.nap.edu/openbook.php? record_id=11970&page=R1. NTP (2005). Report on Carcinogens, 11 edition, National Toxicology Program, Public Health Service, U.S. Department of Health and Human Services, Research Triangle Park, NC, http://ntp.niehs.nih.gov/ index.cfm?objectid=32BA9724-F1F6-975E-7FCE50709CB4C932. Preston, D. L., Cullings, H., Suyama, A., Funamoto, S., Nishi, N., Soda, M., Mabuchi, K., Kodama, K., Kasagi, F., and Shore, R. E. (2008). Solid cancer incidence in atomic bomb survivors exposed in utero or as young children. J Natl Cancer Inst 100, 428–436. Preston, D. L., Ron, E., Tokuoka, S., Funamoto, S., Nishi, N., Soda, M., Mabuchi, K., and Kodama, K. (2007). Solid cancer incidence in atomic bomb survivors: 1958–1998. Radiat Res 168, 1–64. Preston, R. J. (2005). Extrapolations are the Achilles heel of risk assessment. Mutat Res 589, 153–157. Preston, R. J. (2007). Cancer risk assessment for 1,3-butadiene: Data integration opportunities. Chem Biol Interact 166, 150–155. Preston, R. J., and Williams, G. M. (2005). DNA-reactive carcinogens: Mode of action and human cancer hazard. Crit Rev Toxicol 35, 673–683. Seed, J., Carney, E. W., Corley, R. A., Crofton, K. M., DeSesso, J. M., Foster, P. M., Kavlock, R., Kimmel, G., Klaunig, J., Meek, M. E., Preston, R. J., Slikker, W., Jr., Tabacova, S., Williams, G. M., Wiltse, J., Zoeller, R. T., Fenner-Crisp, P., and Patton, D. E. (2005). Overview: Using mode of action and life stage information to evaluate the human relevance of animal toxicity data. Crit Rev Toxicol 35, 664–672. Shim, S. Y., Lim, D. K., and Nam, J. M. (2008). Ultrasensitive optical biodiagnostic methods using metallic nanoparticles. Nanomed 3, 215–232. Sonich-Mullin, C., Fielder, R., Wiltse, J., Baetcke, K., Dempsey, J., Fenner-Crisp, P., Grant, D., Hartley, M., Knaap, A., Kroese, D., Mangelsdorf, I., Meek, E., Rice, J. M., and Younes, M. (2001). IPCS conceptual framework for evaluating a mode of action for chemical carcinogenesis. Regul Toxicol Pharmacol 34, 146–152. Vincent, T. L., and Gatenby, R. A. (2008). An evolutionary model for initiation, promotion, and progression in carcinogenesis. Int J Oncol 32, 729–737. Weinberg, R. A. (2006). The Biology of Cancer, Garland Science, New York, pp. 1–850. Wiltse, J. A., and Dellarco, V. L. (2000). U.S. Environmental Protection Agency’s revised guidelines for carcinogen risk assessment: Evaluating a postulated mode of carcinogenic action in guiding dose– response extrapolation. Mutat Res 464, 105–115.

CH A P TE R

14

EXPERIMENTAL ANIMAL STUDIES AND CARCINOGENICITY Mary Elizabeth (Bette) Meek

14.1.

INTRODUCTION

The mainstay of experimental studies on carcinogenicity in animals has been the long-term combined chronic/cancer bioassay in rats and mice, which has been designed principally as a basis to identify hazard (i.e., what is the intrinsic potential of the substance to induce cancer?). Results of such assays are generally combined in a weight of evidence approach with those of short-term principally in vitro assays, which identify potential for interaction with DNA, including the propensity to cause mutation (see Chapters 10 and 11). Cancer bioassays in animals, screening assays for genotoxicity, or the two of them together are conducted at relatively high doses as a basis to identify hazard. These studies fail, however, to provide robust dose– response information or even a minimum amount of the kinetic and dynamic information in a mode of action context, which would most meaningfully contribute to estimation of human risk. The requirement of mandates worldwide to systematically assess much larger numbers of chemicals necessitates more efficient and effective toxicity testing. This includes intelligent testing strategies to focus early on endpoints of interest, to consider “chemical space” in targeted investigation, to prioritize shorter-term in vivo assays of a range of intermediate endpoints based on consideration of mode of action of the chemical(s), and to better tailor mutagenicity testing as a basis to consider mode of action for cancer. Acquisition in traditional cancer bioassays or shorter-term investigations of mechanistic data on key events to better inform modes of induction of tumors as a basis for more accurate extrapolation of risks between doses, species, and subgroups of the population is also a priority. These intermediate approaches to better tailor and target testing as a basis for more informative characterization of risk in humans are essential prerequisites to meeting a broader, longer-term strategy for greater reliance on computational modeling and in vitro data in humans


378

14.2. CURRENT STATUS OF HAZARD TESTING FOR CANCER FOR REGULATORY RISK ASSESSMENT

379

envisaged, for example, in the US National Research Council’s (NRC)’s report entitled Toxicity Testing in the 21st Century—A Vision and a Strategy (NRC 2007b). In this chapter, the nature of experimental animal testing for cancer will be reviewed, trends examined, and recommendations for more progressive testing strategies included, based on consideration of evolving regulatory pressures and scientific advances in the context specifically of their application in risk assessment. Interim and pragmatic strategies to advance common understanding in both the research and risk assessment communities in potential appropriate application of evolving data, as well as the implications for testing strategies, are also considered.

14.2. CURRENT STATUS OF HAZARD TESTING FOR CANCER FOR REGULATORY RISK ASSESSMENT 14.2.1. The Combined Chronic/Cancer Bioassay in Rats and Mice While a range of methods has been explored as a basis to identify substances with potential carcinogenic potential, none has been sufficiently developed and/or validated to be able to supplant the combined chronic/cancer bioassay in rats (24 months) and mice (18 months), which continues to be the mainstay of chemical carcinogenicity testing. It involves combined evaluation of potential carcinogenicity and noncancer chronic toxicity, by a highly standardized method, which has been widely adopted throughout the world. These studies have been conducted since the 1960s with only very limited development of their protocols. Reliance on the combined chronic/cancer bioassay has been predicated predominantly on the basis of positive results for compounds that are known human carcinogens. Indeed, positive results in one or more adequately investigated animal species have been observed for all known human carcinogens (Vanio and Wilbourn 1994). Given their predominance in hazard identification for cancer, this section focuses principally on the objective and design of the combined chronic/cancer bioassay in rats and mice. Specifics of the design of these bioassays are presented in Table 14.1. The combined chronic/cancer bioassay is carried out almost TABLE 14.1.

Conditions: Route: Experimental animals: Number of animals: Dose levels:

Duration of exposure:

Design of Carcinogenicity Studiesa

• Chemical identification of substance, its purity and chemical characteristics • Oral (gavage, diet, drinking water or capsules), inhalation, or dermal • Rat, mouse, hamster, dog, or monkey • 50 rodents per sex per group; for nonrodents usually not more than 7–20 animals per dose group • Control and at least three dose levels, more dose levels for quantitative risk assessment • Satellite groups may be added • Majority of expected lifespan • Inhalation: intermittent (e.g., 6 hr/day, 5 days/week) or continuous

(Continued)

380

CHAPTER 14 EXPERIMENTAL ANIMAL STUDIES AND CARCINOGENICITY

TABLE 14.1.

Examinations:

Tissues normally collectedb:

Results:

(Continued) • Physical measurements: Temperature, humidity, homogeneity and stability of test substance, food and water consumption, and, for inhalation studies, air flow, concentrations, particle size • Clinical observations: Body weight; changes in skin, fur, eyes, mucous membranes, occurrence of secretions and excretions, behavior, respiratory, circulatory, autonomic and central nervous systems, somatomotor activity, sensory reactivity to stimuli; assessment of grip strength and motor activity; ophthalmologic examinations (90 d/chronic) • Hematology: Hematocrit, hemoglobin concentration, erythrocyte count, total and differential leukocyte count, platelet count, measure of blood clotting time • Clinical biochemistry: Investigation of organ function, carbohydrate metabolism, electrolyte balance, serum salts (Ca, P, Na, K, Cl), serum enzymes (such as alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, gamma glutamyl transpeptidase, sorbitol dehydrogenase, ornithine decarboxylase), cholesterol, glucose, urea, creatinine, total protein, albumin, total bilirubin (may be extended to lipids, hormones, acid/base balance, methemoglobin, cholinesterase activity) • Urinalysis (not routinely in 28-day tests): Appearance, volume, osmolality or specific gravity, pH, protein, glucose, blood cells • Pathology: Gross necropsy including external surfaces, orifices, cranial, thoracic and abdominal cavities and contents, organ weights • Recommended for microscopic examinations: 1. All grossly visible tumors or lesions suspected of being tumors in all groups; 2. (a) All preserved organs and tissues of all animals that die or are killed during the study. (b) All preserved organs and tissues of animals of the highest dose group and controls. (c) If a significant difference is observed in hyperplastic, preneoplastic, or neoplastic lesions between the highest dose and control groups, microscopic examination of that particular organ or tissue of all animals in the study. (d) In case the results of the experiment indicate substantial alteration of the animals’ normal longevity or the induction of effects that might affect a neoplastic response, the next lower dose level should be examined as described above. (e) The incidence of tumors and other suspect lesions normally occurring in the strain of animals used (under the same laboratory conditions—i.e., historical control) is desirable for assessing the significance of changes observed in exposed animals. • Adrenal glands, aorta, bone (femur, sternum), bone marrow, brain, carcass, cecum, colon, cervix, duodenum, ear canal, epdidymis, esophagus, eyes and optic nerves, Harderian gland, heart, ileum, jejunum, kidney, larynx, liver, lymph nodes (mandibular and mesenteric), lungs, mammary glands, nose/turbinates, oviducts, ovaries, pancreas, parathyroid, pituitary gland, prostate, salivary gland, sciatic nerve, seminal vescile, skin, skeletal muscle, spinal chord, spleen, stomach, testes, thymus, thyroid glands, tongue, trachea, urinary bladder, uterus, vagina, and Zymbal gland • Information on carcinogenic properties, tumor incidences in relation to dose, latency period, tumor multiplicity, potential for metastasis

Source: Modified from Vermeire et al. (2007)a or Hamm (1994).b

14.2. CURRENT STATUS OF HAZARD TESTING FOR CANCER FOR REGULATORY RISK ASSESSMENT

381

exclusively in rats and mice (currently most often F344 rats and B6CF1 mice), where both sexes are exposed. In addition to tumors, preneoplastic lesions and other indications of chronic toxicity providing evidence of treatment-related effects are also investigated (EPA 2005a; Vermeire et al. 2007). The bioassay involves exposure of weanling or post-weanling animals (beginning generally at 6–8 weeks of age) for the majority of their lifespan (24 months in rats or 18 months in mice). Fifty animals per sex per group are exposed to at least three dose levels and controls. The dose and route of administration for the chronic/ cancer bioassay is based largely on data obtained from prechronic toxicity studies with the same compound, normally encompassing both 14-day and 90-day exposures at a wide range of doses with evaluation of histopathological endpoints similar to those in the chronic study. Traditionally, since small numbers of animals per group in cancer bioassays are used as surrogates for a much larger human population, doses have exceeded considerably those associated with most human exposures. The lowest dose is normally selected so as not to interfere with growth and development nor to cause effects, whereas the highest dose is selected to result in signs of toxicity. With the exception of macronutrients, the highest dose should not exceed a concentration of 5% of the diet or 1 g/kg body weight for oral gavage studies (OECD 1981). The study is designed to include one dose in addition to the control(s) that is not expected to elicit adverse effects. The intermediate dose is normally within the mid range between the high and the low doses. The middle and lowest doses are selected (adequately spaced) to characterize the shape of the dose–response curve as much as possible. It is not uncommon to add a satellite high-dose group (20 animals per sex) to induce frank toxicity and a satellite control group (10 animals per sex per group) to evaluate effects other than neoplasia (usually after 12 months experimentation). Caging, care, feed, and water supply (diet) must be optimum and wellcontrolled. The rate of exposure to the substance is normally comparable to the anticipated human exposure, with frequency dependent on the route. In oral studies, the substance is administered daily unless by gavage, in which event exposure is usually restricted to 5 times a week, as is also characteristic of inhalation studies, where exposure is generally limited to 6 hours per day. Careful daily clinical examination is required and appropriate action is taken to minimize loss of animals during the study due to autolysis or cannibalization. Body weights are measured daily during the first 13 weeks and once every 4 weeks thereafter. Food and water intake are determined weekly during the first 13 weeks and then quarterly for the remainder of the study. Blood tests are performed after 3, 6, 18, and 24 months on 20 animals per sex per group, and a differential blood count is performed on samples of animals from the highest dose group and the controls and is performed at lower dose levels when indicated. Urine analysis of 10 animals per sex per group is conducted at the same intervals. Every 6 months, clinical chemical analysis is conducted (see Table 14.1). At the end of the experiment, a 50% survival rate is expected for rats at 24 months and mice at 18 months. Complete gross examination is performed, and histopathological examination is carried out on all tissues and organs from the

382


highest dose group and the control group (see Table 14.1). Where indicated, the tissues and organs of lower dose groups are examined and all tumors or tumor-like lesions are examined histopathologically. Statistical analysis is performed for each tumor type separately. The incidence of benign and malignant lesions of the same cell type, usually within a single tissue or organ, are considered separately but may be combined when scientifically defensible (McConnell et al. 1986). Trend tests such as the Cochran–Armitage test are those recommended for determining whether chance, rather than a treatment-related effect, is a plausible explanation for an apparent increase in tumor incidence (Cochran and Snedecor 1972). By convention, a statistically significant comparison is generally one for which p is less than 0.05 that the increased incidence is due to chance. Significance in either kind of test is sufficient to reject the hypothesis that chance accounts for the result.

14.2.2.

Perinatal Carcinogenicity Studies

Current standardized long-term carcinogenesis bioassays involve initiation of dosing of animals at 6–8 weeks of age and throughout the lifespan of the animal (18–24 months). This protocol has been modified in some cases to investigate the potential of the test agent to induce transplacental carcinogenesis or potential differences following perinatal and adult exposures. However, standard protocols to investigate these aspects have not been developed, a function (in part) of experience that exposure during the perinatal period rarely identifies carcinogens that are not detected in standard animal bioassays, although it may increase the incidence of a given type of tumor or reduce the latency period for tumor development.

14.2.3.

Limited In Vivo Studies

The cost and duration of the combined chronic/cancer bioassay has limited its conduct to small numbers of selected chemicals. As a result, several short-term methods aimed at increasing predictive accuracy to enable testing of larger numbers of chemicals have been developed in attempts to successfully correlate their results with evidence of carcinogenicity (or lack of carcinogenicity). This includes investigation of potential to promote tumor development, several model systems in transgenic and knockout models, and consideration of the predictive potential of traditional toxicity endpoints in shorter-term studies. Limited, medium-term in vivo studies have been developed to investigate the tumor enhancing properties of chemicals. These involve administration of a known initiator or a genotoxic carcinogen in a subcarcinogenic dose, followed by exposure to the substance being examined. Several organ systems have been investigated (Kroes 1987) in these assays such as skin, lung, stomach, mammary gland, kidney, thyroid, pancreas, intestines, and urinary bladder (Feron et al. 1999). Short- and medium-term assays in transgenic models as a basis to provide essential information about the predisposing factors to specific genetic alterations in carcinogenesis include the rat liver foci model, the XPA−/− and the p53+/− knockout mouse models, the Tg.AC and Tg.rasH2 transgenic mouse models, and the neonatal

14.3. APPLICATION IN RISK ASSESSMENT

383

mouse model (Vermeire et al. 2007). These models have a number of potential advantages in carcinogen identification, including reduction of both the necessary periods of exposure and numbers of animals. Assay length is generally in the range of 24–26 weeks, significantly shorter than the standard chronic/cancer rodent bioassay. Furthermore, with appropriate model selection, relevant to mode of action of the substance, it is possible to more accurately predict the human response, contributing directly to their relevance to risk assessment and regulatory decision making (Gulezian et al. 2000). The capacity of shorter-term in vivo assays to predict carcinogenicity by investigating traditional toxicity endpoints has also been investigated (Ashby and Tennant 1994). Allen et al. (2004) evaluated the correlation of prechronic liver lesions and liver tumor formation from studies performed by the U.S. National Toxicology Program (NTP) in mice (83 compounds) and rats (87 compounds). Lesions considered included hepatocellular necrosis, hepatocellular hypertrophy, hepatocellular cytomegaly, bile duct hyperplasia, and hepatocellular degeneration, along with increased liver weight. Results indicated that pooling of the prechronic data on hepatocellular necrosis, hepatocellular hypertrophy, and hepatocellular cytomegaly was predictive of carcinogenicity in the 2-year study (p < 0.05) (Allen et al. 2004). To study tumor enhancing properties, various in vitro tests have also been proposed (Yamasaki 1990). They are based on the determination of clinical properties common to a group of promoting agents, such as loss of cell-to-cell communication and outgrowth of partially transformed cells.

14.3.

APPLICATION IN RISK ASSESSMENT

Risk assessment (i.e., the characterization of the potential adverse effects of human exposures) is the requisite basis for the development and implementation of control measures that are protective of public health (i.e., risk management). Traditionally, risk assessment has been considered to be composed of four elements, namely hazard identification, dose–response assessment, exposure estimation, and risk characterization (NRC 1983), with the latter being a synthesis of relevant data from all of the component steps with a clear delineation of uncertainties and their implications for risk management (see Chapter 1). While traditionally chronic/cancer bioassays have been designed to address hazard identification (i.e., the intrinsic capacity of a substance to cause harm), there is an increasing need to revise testing guidelines to integrate more hierarchical and predictive mode of action based approaches that will make them much more relevant to hazard characterization and subsequently, risk characterization.

14.3.1.

Hazard Identification

Hazard identification as determined from an adequate assessment of data from chronic/cancer bioassays in animals has been reviewed in Maronpot (1994), Health

384


Canada (HC) (HC 1994), Meek et al. (1994) and the U.S. Environmental Protection Agency (EPA) (EPA 2005a). In considering the adequacy of specific investigations as a basis to identify hazard in risk assessment, several features of study design are considered including the purity of the compound administered, the size of the study (i.e., numbers of exposed and control animals), whether the study was performed under Good Laboratory Practice standards, the relevance of the route of exposure to that of humans, duration of exposure, the number and suitability of the dose levels administered, the extent of examination of various toxicological endpoints, and the statistical analysis of the data (HC 1994; Meek et al. 1994). Criteria for the technical adequacy of animal carcinogenicity studies have been published (e.g., Chhabra et al. 1990; NTP 1984; OSTP 1986). For specific chemicals, all available studies of carcinogenicity in whole animals are considered, at least preliminarily, with those being judged to be wholly inadequate in protocol, conduct, or results being discarded. Current standards of adequacy as well as those that were contemporaneous with the study are consulted. Care is taken to include studies that provide some evidence bearing on carcinogenicity or that are relevant to interpretation of effects noted in other bioassays, even if these investigations have some limitations of protocol or conduct. The findings of long-term rodent bioassays are interpreted in conjunction with results of prechronic studies along with toxicokinetic studies and other pertinent information, if available. Evaluation of tumor effects takes into consideration both biological and statistical significance of the findings (EPA 2005a). Among the many criteria for consideration of technical adequacy of animal carcinogenicity studies is the appropriateness of dose selection. This has been particularly important where results are negative, since traditionally it has been considered that lack of a sufficiently high dose reduces the sensitivity of the studies. A scientific rationale for dose selection is normally articulated based on relevant toxicologic information from prechronic, mechanistic, and toxicokinetic studies. It has generally been considered that an adequate high dose would be one that produces some toxic effects without unduly affecting mortality from effects other than cancer or producing significant adverse effects on the nutrition and health of the test animals (NRC 1993; OECD 1981). If the test agent does not appear to cause any specific target organ toxicity or perturbation of physiological function, an adequate high dose can be specified in terms of a percentage reduction of body weight gain over the lifespan of the animals. The high dose would generally be considered inadequate if neither toxicity nor changes in weight gain is observed. On the other hand, significant increases in mortality from effects other than cancer generally have been considered to indicate that an adequate high dose has been exceeded. Other signs of treatment-related toxicity associated with an excessive high dose may include: (a) significant reduction of body weight gain (e.g., greater than 10%), (b) significant increases in abnormal behavioral and clinical signs, (c) significant changes in hematology or clinical chemistry, (d) saturation of absorption and detoxification mechanisms, or (e) marked changes in organ weight, morphology, and histopathology. Overt toxicity or qualitatively altered toxicokinetics due to excessively high doses may result in tumors secondary to toxicity. Moreover, a lack of


385

tumorigenic response at exposure levels that cause significant impairment of animal survival may not be acceptable as a basis to consider results as negative, in view of the reduced sensitivity of the bioassay. There has been continuing controversy about the use of the “maximum tolerated dose” as the highest dose in chronic/cancer bioassays. Essentially, many consider that observation of cancer under conditions that are often unique to the experimental conditions provides little relevant information even in the context of hazard identification for humans, which are exposed to much lower doses (often six orders of magnitude). For dietary studies, weight gain reductions are considered in the context of whether there may be an issue of palatability. In the case of inhalation studies with respirable particles, evidence of impairment of normal clearance of particles from the lung should be considered along with other signs of toxicity to the respiratory airways to determine whether the high exposure concentration has been appropriately selected (EPA 2001). For dermal studies, evidence of skin irritation may indicate that an adequate high dose has been reached (EPA 1989). Statistical versus biological significance is also necessarily taken into account in interpreting the results of cancer bioassays. A statistically significant response may or may not be biologically significant and vice versa. The selection of significance levels to distinguish positive results is a matter of policy based on a trade-off between the risks of false positives and false negatives; the value most commonly adopted is 5%. A result with a significance level of greater or less than 5% (or other selected value), then, is examined to see if it is consistent with other scientific information. A two-tailed test or a one-tailed test can be used. In either case, a rationale is provided. Statistical power affects the likelihood that a statistically significant result could reasonably be expected. This is especially important in studies or dose groups with small sample sizes or low dose rates. Consideration of the statistical power is often essential for reconciling positive and negative results from different studies. The impact of multiple comparisons should also be taken into account. Based on analysis of typical bioassays in which both sexes of two species were included, studies in which there is only one significant result that falls short of the 1% level for a common tumor should be treated with caution (EPA 2005a; Haseman 1983). While the statistical significance of tumor incidence is judged based principally on comparison in dosed versus concurrent control animals, consideration of historical control data provides additional insight concerning both statistical and biological significance of uncommon tumors types or those with high spontaneous incidence in particular strains (Haseman 1995; Tarone 1982). It can be particularly helpful for cases where there are small increases (not reaching statistical significance) in uncommon tumors in treated groups compared with concurrent controls. However, since they do not take into account differences in survival of animals among studies, caution must be exercised in the interpretation of ranges of historical responses, the most relevant of which are derived from studies in the same laboratory with the same supplier conducted within 2 or 3 years of the study under review. Moreover, the degree of confidence in historical control data is necessarily related to the number of studies in the database. Aspects that need to be addressed

386


in comparisons with historical control data include genetic drift in the laboratory strains, differences in pathology examination at different times and in different laboratories (e.g., in criteria for evaluating lesions; variations in the techniques for the preparation or reading of tissue samples among laboratories), and comparability of animals from different suppliers. The description and nature of peer engagement in review of the combined chronic/cancer bioassay is also critical in evaluation of the outcome. It is necessitated because of the judgmental processes that accrue in cascading fashion during the evaluation of the study in relation to the evaluation of the pathology but also in relation to the relative strength of the experimental evidence provided by the study (Maronpot 1994). The consistency of the results of the principal studies are also considered in the assessment of the weight of evidence for an effect (e.g., have similar effects been observed in studies in other species or would such effects have been expected based on the structure or properties of the chemical?), taking into account traditional criteria for weight of evidence including consistency, specificity, and biological plausibility. The types, site, incidence, and severity of effects and the nature of the exposure– or dose–response relationship are also taken into account. In assessing potential to induce tumors in humans, aspects that add to the weight of evidence include (a) observation of uncommon tumor types, (b) occurrence of tumors at multiple sites by more than one route of administration in multiple strains, sexes, and species, and (c) progression of lesions from preneoplastic to benign to malignant, including metastases and comparatively short latency periods. Traditionally, weight of evidence descriptors such as “carcinogenic to humans,” “probably carcinogenic to humans,” and so on, for cancer hazard have been developed by a number of agencies, as a basis for distinguishing approaches to dose–response analysis in subsequent risk characterization and also as a basis to communicate hazard (see Chapters 1 and 3). Increasingly, however, there is trend to providing more narrative and accurate descriptors, which include reference to the conditions under which cancer is observed, as a basis to avoid misinterpretation.

14.3.2.

Hazard Characterization

Hazard characterization necessarily takes into account not only results of guidelines studies designed to identify hazard but additionally, mechanistic data. Characterization of hazard involves a weight of evidence determination (i.e., a comprehensive, integrated judgment of all relevant information supporting conclusions regarding a toxicological effect, including human relevance, which takes into account traditional criteria for weight of evidence). While the standard combined chronic/cancer bioassay is helpful in hazard identification, it contributes in a more limited extent to hazard characterization (i.e., the likelihood of causing adverse effects in humans). However, with some modification in the context of evolving integrated and hierarchical test strategies for groups of chemicals or individual substances, carcinogenicity bioassays have potential to contribute considerably additionally in this context. For example, as discussed


387

above, traditionally in combined chronic/cancer bioassays, exposure to the test material has been maximized. This is helpful in contributing to confidence in the context of hazard identification that the study is negative for carcinogenicity if there is no increase in tumor incidence at a toxic high dose nor toxicity or tumors at appropriately spaced lower doses. However, this necessitates appropriate caution in interpreting results associated with excessive dosage levels that confound the interpretation of study results. Studies that result in tumors only at excessive doses may be compromised and may or may not carry weight, depending on the interpretation in the context of other study results and lines of evidence, including those informing mode(s) of action. Mechanistic data for cancer are considered in hazard characterization in the context of “mode” of induction of toxic effects. A postulated mode of action is a biologically plausible sequence of key events leading to an observed effect supported by robust experimental observations and mechanistic data. It describes key cytological, genetic, and biochemical events—that is, those that are both measurable and necessary to the observed effect. Mode of action is contrasted with mechanism of action, which generally involves a much greater understanding of the molecular basis for an effect. In 2001, as part of its efforts to harmonize risk assessment practices, the International Programme on Chemical Safety (IPCS) (WHO/ILO/UNEP) published a framework for assessment of mode of action for carcinogenesis in laboratory animals (animal mode of action). This was based on consideration of specific aspects of data analysis developed much earlier by Sir Austin Bradford Hill as a basis for considering causality of observed associations in epidemiological studies (Hill 1965). Relevant factors include dose–response and temporal concordance between key and end events, consistency, biological plausibility, and coherence (SonichMullin et al. 2001). More recently, the IPCS framework has been expanded to address human relevance (Boobis et al. 2006, 2008), based on previous work of the International Life Sciences Institute (ILSI) (Meek et al. 2003; Seed et al. 2005). The human relevance framework (HRF), which was developed and refined originally through its application in case studies for principally nonDNA reactive carcinogens, has been extended more recently to DNA-reactive carcinogens, noncancer endpoints, different life stages, and combined exposures to multiple chemicals. Development of the HRF for mode of action has involved engagement of more than 150 scientists internationally. It has also been widely incorporated into international and supra-national guidance and is being applied in this context as a basis to increase transparency concerning uncertainty, promote consistency in decision-making, facilitate peer engagement, and identify critical research needs (EC 2003; EFSA 2006; IPCS 2006; JMPR 2006; OECD 2002; UNECE 2007). It has also been extensively, even routinely, adopted in risk assessments by the U.S. EPA (Dellarco and Baetcke 2005; EPA 2000a,b, 2005b, 2007), the United Kingdom (COC, 2004), HC [see, for example, Liteplo and Meek (2003)], and other governmental organizations. The Society of Toxicology’s 2006 awards for Best Paper in Fundamental and Applied Toxicology and Toxicological Sciences, provides evidence of peer recognition of the contribution of the framework (Green et al. 2005; Pastoor et al. 2005).

388


In this framework for mode of action analysis, the weight of evidence for a hypothesized mode of action for a particular response observed in experimental animals is considered in the context of key events along the causal pathway. Once established in experimental animals, the HRF provides an analytical tool to enable the systematic evaluation of the data in order to consider its human relevance based often on consideration of more generic information, such as anatomical, physiological, and biochemical variations among species. In this manner, the framework encourages maximum use of both chemical-specific and more generic information. The International Agency for Research on Cancer (IARC) has identified key biochemical and histopathological events as a basis to characterize specific modes of action (IARC 1999a,b). Established mechanistic/key events are additionally being identified in recent IARC monographs for individual chemicals and groups (Straif et al. 2009). Information on mode of action is relevant not only to determine whether tumors observed in animals are relevant to humans but also to consider dose transitions and potentially susceptible subgroups. It is also critical as a basis to address whether or not there is likely to be site concordance of tumors between animals and humans. While there is evidence that growth control mechanisms at the level of the cell are homologous among mammals, there is no evidence supporting nor reason to believe that these mechanisms are site-concordant. Instead, information on likely variations between animals and humans in kinetic and dynamics, based on some understanding of mode of induction of tumors, will inform in the context of potential sites of cancer induction in humans. This information is essential to interpret (particularly) the significance of negative epidemiological data, taking into account the sensitivity of the study to detect cancers at most likely sites. In the development of relevant biomarkers in epidemiological studies, it is also critical to increase the utility of the latter as a basis for consideration of the risks to exposure to chemical in both the occupational and general environments. While their use in hazard identification is necessarily limited owing to limitations such as relevance of the mutation in one pathway to the specific tumor, careful selection and interpretation of data from transgenic models in a mode of action context has considerable potential to contribute to hazard characterization.

14.3.3. Dose–Response Analyses; Selection of Points of Departure While the dose–response relationship observed in cancer bioassays is commonly used as the basis for risk characterization for substances that are considered as carcinogens, the extent to which it meaningfully informs risk is limited by the small number of dose groups and the magnitude of the variation between exposure of humans and administered doses. The limited numbers of doses examined is necessarily a function of the costs associated with close-to-lifetime observation of groups of (commonly) 50 animals each. Normally, characterization of dose–response analyses as a basis for comparison with exposure estimation in risk characterization is based on only those tumors where available data indicate that the mode of action is relevant to humans. Tumors


389

from similar tissues of origin may be combined as a basis of analysis. Where data indicate that there are significant differences in absorption, distribution, metabolism, and elimination of the compound in different animal species, wherever possible, studies in which the species and strain of animal are most similar to humans in this regard are used and differences quantitatively taken into account to the extent possible, through physiologically based pharmacokinetic or biologically based modeling. Generally, mathematical models are used to extrapolate the data on the exposure– or dose–response relationship derived from carcinogenicity bioassays to estimate the risk at concentrations to which the general population is exposed in the absence of more biologically based kinetic or dynamic models. There are numerous uncertainties in such approaches, which often involve linear extrapolation of results over several orders of magnitude, commonly in the absence of relevant data on mode of action for tumor induction or differences in toxico-kinetics and -dynamics between the relevant experimental animal species and humans. Concentrations or doses associated with a negligible or de minimis level of risk (such as a lifetime cancer risk of 1 in 1,000,000) by low-dose extrapolation procedures are often compared with exposure to determine whether risks are acceptable. Selection of appropriate de minimis levels constitutes science policy (i.e., making a societal judgment about what level constitutes de minimis risk). There is no single “correct” value that adequately characterizes de minimis risk associated with a concentration or dose below which risks are acceptable and above which they are not; rather, the risk at low doses or concentrations is assumed to be a continuum, with reduction of exposure leading to an incremental reduction of risk and increases in exposure leading to incremental increases in risk (see Chapter 26). In addition, in view of the considerable uncertainties of current low-dose extrapolation procedures, specification of risks in terms of predicted incidence or numbers of excess deaths per unit of the population is highly inaccurate and open to misinterpretation, particularly without specification of the bounds of uncertainty (HC 1994; Meek et al. 1994). Indeed, low-dose risk estimates based on empirical modeling and extrapolation of the dose–response curve over ranges of as much as six orders of magnitude have meaning in a relative (i.e., one to another) rather than absolute sense. For assessment of substances under the Canadian Environmental Protection Act (CEPA)—for example, for compounds that are carcinogenic involving direct interaction with DNA, where data are judged sufficient—quantitative estimates of the carcinogenic potency are compared to (a) the estimated daily intake of the priority substance by the general population (or certain high-exposure subgroups) in Canada or (b) concentrations in specific relevant environmental media [referred to as the exposure/potency index (EPI)]. Potency is expressed as the concentration or dose, which induces a 5% increase in the incidence of, or deaths due to, tumors or heritable mutations considered being associated with exposure. The tolerable dose (TD)0.05 is not based on the confidence limit but, rather, is computed directly from the curve. This was considered to be appropriate in view of the stability of the data in the experimental range and to avoid unnecessarily conservative assumptions. Also, use of a point estimate or confidence limit does not affect the relative magnitude of the potency estimates for different compounds. The estimates of potency are

390


generally restricted to effects for which there has been a statistically significant increase in incidence and a dose-response relationship, characterized using appropriate mathematical models (e.g., multistage).

14.4.

EVOLUTION OF TESTING STRATEGIES

Conduct of the chronic bioassay by a defined protocol in similar rodent strains permits direct comparisons among chemicals with diverse structural and/or biological properties. However, there are considerable limitations, which restrict the number of substances evaluated with such studies. These include the high cost of a full 2-year bioassay in both sexes of rats and mice and the length of time required for testing and interpretation. These limitations necessarily reduce the throughput of chemicals. The significant numbers of animals sacrificed in 2-year bioassays to consider limited numbers of chemicals are also inconsistent with increasing pressure worldwide to reduce, refine, and replace animal testing. As a result (in part) of these considerations and increasing regulatory pressures to prioritize and consider potential risks associated with much larger numbers of substances more efficiently, toxicity testing continues to evolve from the use of prescribed protocols of whole-animal bioassays to greater emphasis on understanding the underlying pathways that lead to carcinogenesis, or other endpoints. Notable in this context is the content of the U.S. NRC report on toxicity testing (NRC 2007b), which advocates the identification and use of toxicity pathways for both testing and in dose response modeling. Much of the testing envisioned in this report entails in vitro studies (particularly, using tests based on high-throughput assays). These assays aim to characterize cellular processes and toxicity pathways more accurately by testing different levels of cellular function, including (a) genomics, the study of genes and their function as a whole; (b) proteomics, the large-scale study of proteins and their function; and (c) metabolomics, the study of all metabolites in a biological system that are being used to describe toxicant responses (see Part V, this volume). Computational biology techniques can be applied to these “-omics” data to link toxicity pathways and to identify patterns characteristic of specific toxicants. A key challenge to the use of findings from such tests will be the extrapolation of findings from in vitro studies to better understand and estimate human risks. One of the most important contributions of this new strategy is that it attempts to integrate exciting developments in toxicogenomics to increase efficiency and relevance of toxicity testing to risk assessment (NRC 2007a). The advocated use of human cells or tissues has potential to eliminate the need for interspecies extrapolation, to increase efficiencies in testing, and to reduce the use of animals. However, a pragmatic and seemingly essential first step in addressing this reevaluation of adversity would be a recommendation to relate (a) early perturbations to apical endpoints in frameworks designed to systematically address consideration of key events in modes of action and (b) their subsequent implications for dose– response in risk assessment [see, for example, Meek (2008)]. This would be instrumental in advancing common understanding in both the research and risk assessment

14.5. DISCUSSION: CLOSING THE GAP BETWEEN HAZARD TESTING AND RISK ASSESSMENT

391

communities in potential appropriate application of data on early events in a toxicity pathway. Increasing experience in this context could provide the necessary basis for revisiting guidelines for toxicity testing. Moving forward, it will also be critical to link and integrate this vision with other ongoing activities in regulatory risk assessment, including pragmatic developments in several jurisdictions (in particular in Canada and Europe) to (a) address progressive regulatory requirements to efficiently consider much larger numbers of chemical substances and (b) address critical challenges in moving the regulatory community toward the use of this approach and better balance of the focus on hazard with that on exposure. This includes (a) tools developed to consider priorities from amongst the 23,000 compounds included on the Domestic Substances list under the CEPA (Meek and Armstrong 2007) and (b) intelligent or integrated hierarchical testing strategies being developed in Europe for implementation of the legislation for Registration, Evaluation, and Authorization and restriction of Chemical substances (REACH) (see Chapter 3) (Van Leeuwen et al. 2007). Objectives of initiatives under these programs include maximally drawing upon existing data on toxicity, as a basis to increase efficiency. The former also considered prioritization on the basis of much simpler and more discerning data and tools for the significantly potentially more influential component of risk assessment, namely exposure estimation. While the predictive capacity of current computational technologies such as (quantitative) structure–activity relationship analysis (including the threshold of toxicological concern) (see Chapter 4) (Renwick et al. 2003) is necessarily limited currently owing principally to the nature of available toxicological data, their meaningful consideration has important implications for the design of future toxicity testing strategies including focus on coverage of “chemical space” versus individual substances as a critical criterion to increase efficiency and focus on in vitro testing strategies for particular modes of action for specific endpoints. These approaches also require limited new resources and promote more effective and efficient use of existing data as a basis to meaningfully contribute to early risk management.

14.5. DISCUSSION: CLOSING THE GAP BETWEEN HAZARD TESTING AND RISK ASSESSMENT Ultimately, toxicological testing for cancer aims to predict possible adverse effects in humans when exposed to chemicals. Currently, it is designed principally to identify hazards at relatively high doses. As a result, data derived from animal studies are limited with regard to informing the potential risks to humans. Given the limited relevance of output for considerable investment of resources, the development of more predictive testing strategies is inevitable and essential. Fundamentally, the assumptions inherent in the use of the results of long-term chronic/cancer bioassays in rodents as a basis for assessment of risk in human populations consist of the following: (1) If the agent causes cancer in rodents, it can cause cancer in humans (interspecies extrapolation), and (2) if the agent significantly increases cancer incidence when administered at high dose, it will also cause cancer, albeit likely at lower incidence, at low doses (interdose extrapolation). Unfortunately, the considerable

392


amount of data generated in regulatory testing is of less than optimum relevance for informing these default assumptions. In the U.S. NRC report on toxicity testing, four major objectives were identified for future testing strategies (NRC 2007b): “depth, providing the most accurate, relevant information possible for hazard identification and dose–response assessment; breadth, providing data on the broadest possible universe of chemicals, end points, and life stages; animal welfare, causing the least animal suffering possible and using the fewest animals possible; and conservation, minimizing the expenditure of money and time on testing and regulatory review.” In the context of risk assessment, there is a need to additionally evolve this thinking to focus on hazard characterization rather than hazard identification. Obstacles to meeting the above-mentioned objectives with the current toxicity-testing paradigm identified by the U.S. NRC include issues surrounding interspecies extrapolation along with the animal welfare objective noted above. In addition, the breadth and conservation objectives are inconsistent with the time-consuming and expensive process of toxicity testing in animals. Early evolution of animal testing for cancer is also essential to meet immediate regulatory pressures. This necessarily requires increased understanding of the objectives and application of testing data from relevant bioassays in risk assessment/ regulatory programs. In particular, much more iterative and integrated testing strategies which include early consideration of mode of action more relevant to risk characterization and assessment than to hazard identification are required. In fact, there is a need for a paradigm shift to move in a scientifically credible and transparent manner from that which requires extensive hazard (animal) testing to one in which a hypothesis- and risk-driven approach can be used to identify the most relevant in vivo information (Van Leeuwen et al. 2007). Thus, it is critically important to efficiently and credibly predict toxicity drawing upon available information as a basis to facilitate reasonable decisions as to whether experimental studies are required to refine risk assessment further. The underlying rationale is to (1) minimize animal testing through introduction of alternative methods, (2) apply shorter-term and less expensive methods before labor-intensive ones, (3) design studies to address hazard characterization relevant to risk assessment, (4) enable early consideration of potential for exposure as a key determinant of testing strategies and risk assessment, (5) maximize the use of up-to-date information from different sources in an integrated manner, (6) allow greater flexibility in introducing new tools and scientific knowledge, and (7) allow more robust and focused regulatory decisions using testing and nontesting approaches. For cancer, it will require, additionally, much greater emphasis on testing in a mode of action context for tumors rather than reliance on chronic/cancer bioassays and principally screening assays for genotoxicity. These screening methodologies typically include as a minimum a battery of three assays: (a) a test for gene mutation in bacteria, (b) an in vitro test for mutation and/or chromosomal damage in mammalian cells, and (c) an in vivo test for chromosomal damage using rodent hematopoietic cells. Performance of these studies satisfies the aim for which they were first developed (i.e., to identify genotoxic agents that might pose cancer risk in humans as a basis to determine whether or not there should be additional testing or development); however, these studies are less informative in the context of

REFERENCES

393

consideration of mode of action for tumor induction at much lower doses. Experimental studies on mutagenesis and carcinogenesis are usually confined to exposures covering one to two orders of magnitude, and these doses are frequently high to establish hazard identification. In contrast, risk assessment extrapolations frequently cover up to six orders of magnitude. Better data on the dose–response for mutations and the subsequent utilization of these data in quantitative risk assessment will provide critical scientific information on dose–response relationships that are highly relevant for carcinogenesis (Swenberg et al. 2008). As a basis to better integrate regulatory objectives in toxicity testing, weightof-evidence frameworks that provide for transparent integration of hazard and modeof-action information such as those developed by the IPCS/ILSI [see, for example, Meek (2008)] are critically important. In the context of testing strategies for carcinogenicity, they illustrate the types of information that would inform to the greatest extent in the context of risk assessment. As we move forward to develop more integrative test strategies, early assimilation of the information in a mode-of-action context as emphasized by these frameworks will be essential. It is expected to encourage collection of information on toxicokinetics and early toxicodynamic key events at interim periods in animals exposed in similar fashion (e.g., satellite groups in which biochemical and histopathological evaluations are performed). Encouragement of much more intelligent testing of this nature is not new [see, for example, Hamm (1994) and Hill (1994)]. This is consistent with testing strategies to more meaningfully inform hazard characterization based on existing information and integrated testing strategies, which take into account results of computational predictive approaches. Application of the HRF increases the transparency of delineation of the relative degrees of uncertainty associated with various options for consideration in assessment of risk for impacted populations. HRFs are also instrumental in acquiring transparency on critical data gaps that will further reduce uncertainty. They force distinction of choices made on the basis of science policy versus those that are science judgment related, including reliance on default, based on the erroneous premise that it is always health-protective (Meek and Doull 2009). They focus on early events in a toxicity pathway through relation of early perturbations to apical endpoints in frameworks designed to systematically address (a) consideration of key events in modes of action and (b) their subsequent implications for dose-response in risk assessment [see, for example, Meek (2008)].

REFERENCES Allen, D. G., Pearse, G., Haseman, J. K., and Maronpot, R. R. (2004). Prediction of rodent carcinogenesis: An evaluation of prechronic liver lesions as forecasters of liver tumors in NTP carcinogenicity studies. Toxicol Pathol 32, 393–401. Ashby, J., and Tennant, R. W. (1994). Prediction of rodent carcinogenicity for 44 chemicals: Results. Mutagenesis 9, 7–15. Boobis, A. R., Cohen, S. M., Dellarco, V., McGregor, D., Meek, M. E., Vickers, C., Willcocks, D., and Farland, W. (2006). IPCS framework for analyzing the relevance of a cancer mode of action for humans. Crit Rev Toxicol 36, 781–792.

394


Boobis, A. R., Doe, J. E., Heinrich-Hirsch, B., Meek, M. E., Munn, S., Ruchirawat, M., Schlatter, J., Seed, J., and Vickers, C. (2008). IPCS framework for analyzing the relevance of a noncancer mode of action for humans. Crit Rev Toxicol 38, 87–96. Chhabra, R. S., Huff, J. E., Schwetz, B. S., and Selkirk, J. (1990). An overview of prechronic and chronic toxicity/carcinogenicity experimental study designs and criteria used by the National Toxicology Program. Environ Health Perspect 86, 313–321. COC (Committee on Carcinogenicity) (2004). Guidance on a Strategy for the Risk Assessment of Chemical Carcinogens. London, UK, Department of Health. Cochran, W. G., and Snedecor, G. W. (1972). Statistical Methods, 6th edition, Iowa State University Press, Ames, IA. Dellarco, V. L., and Baetcke, K. (2005). A risk assessment perspective: application of mode of action and human relevance frameworks to the analysis of rodent tumor data. Toxicol Sci 86, 1–3. EC (2003). Technical Guidance Document on Risk Assessment, European Commission Joint Research Centre, Italy, EUR 20418 EN/1. EFSA (2006). Opinion of the Scientific Panel on Plant Health, Plant Protection Products and Their Residues on the scientific principles in the assessment and guidance provided in the field of human toxicology between 2003 and 2006. EFSA J 346, 1–13, http://www.efsa.europa.eu/EFSA/Scientific_ Opinion/ppr_op_ej346_summary-tox_summary_en1.pdf?ssbinary=true. EPA (1989). Summary of the second workshop carcinogenesis bioassay with the dermal route, May 18–19, 1988, Research Triangle Park, NC, EPA/560/6-89/003. EPA (2000a). Atrazine: Hazard and Dose–Response Assessment and Characterization, FIFRA Scientific Advisory Panel Meeting June 27–29, 2000, Held at the Sheraton Crystal City Hotel, Arlington, Virginia, SAP Report No. 2000-05, 1–44, http://www.epa.gov/oscpmont/sap/meetings/2000/june27/finalatrazine.pdf. EPA (2000b). Review of the draft chloroform risk assessment, EPA-SAB-EC-00-009, 1–33, http:// yosemite.epa.gov/sab/sabproduct.nsf/D0E41CF58569B1618525719B0064BC3A/$File/ec0009.pdf. EPA (2001). OPPTS 870.8355 Combined chronic toxicity/carcinogenicity testing of respirable fibrous particles, EPA712-C-01-352, 1–15, http://www.epa.gov/opptsfrs/publications/OPPTS_Harmonized/ 870_Health_Effects_Test_Guidelines/Series/870-8355.pdf. EPA (2005a). Guidelines for carcinogen risk assessment, EPA/630/P-03/001F, 1–166. EPA (2005b). Science Issue Paper: Mode of carcinogenic action for cacodylic acid (Dimethylarsinic acid, DMA[v]) and Recommendations for dose response extrapolation, Prepared by: Health Effects Division, Office of Pesticides Programs, US Environmental Protection Agency. 1–201, http://www.epa.gov/ oppsrrd1/reregistration/cacodylic_acid/dma_moa.pdf. EPA (2007). Advisory on EPA’s assessments of carcinogenic effects of organic and inorganic arsenic: A report of the US EPA Science Advisory Board, EPA-SAB-07-008, 1–88. Feron, V. J., Schwartz, M., Krewski, D., and Hemminki, K. (1999). Long- and medium-term carcinogenicity studies in animals and short-term genotoxicity tests. In Quantitative Estimation and Prediction of Human Cancer Risks, Vol. 131, Moolgavkar, S., Krewski, D., Zeise, L., Cardis, E., and Moller, H. eds., International Agency for Research on Cancer, World Health Organization, Lyon, pp. 103–112. Green, T., Toghill, A., Lee, R., Waechter, F., Weber, E., Peffer, R., Noakes, J., and Robinson, M. (2005). Thiamethoxam induced mouse liver tumors and their relevance to humans. Part 2: Species differences in response. Toxicol Sci 86, 48–55. Gulezian, D., Jacobson-Kram, D., McCullough, C. B., Olson, H., Recio, L., Robinson, D., Storer, R., Tennant, R., Ward, J. M., and Neumann, D. A. (2000). Use of transgenic animals for carcinogenicity testing: Considerations and implications for risk assessment. Toxicol Pathol 28, 482–499. Hamm, T. E. (1994). Design of a long-term animal bioassay for carcinogenicity. In Handbook of Carcinogen Testing, 2nd edition, Milman, H. A., and Weisburger, E. K., eds., William Andrew Publishing/Noyes, Norwich, NY, pp. 1–893. Haseman, J. K. (1983). A reexamination of false-positive rates for carcinogenesis studies. Fundam Appl Toxicol 3, 334–339. Haseman, J. K. (1995). Data analysis: Statistical analysis and use of historical control data. Regul Toxicol Pharmacol 21, 52–59; discussion 81–86. Health Canada HC (1994). Human health risk assessment for priority substances, En40-215/41E, 1–41, http://www.hc-sc.gc.ca/ewh-semt/alt_formats/hecs-sesc/pdf/pubs/contaminants/approach/approacheng.pdf.

REFERENCES

395

Hill, A. B. (1965). The environment and disease: association or causation? Proc R Soc Med 58, 295–300. Hill, R. N. (1994). Regulatory implications: Perspective of the U.S. Environmental Protection Agency. In Handbook of Carcinogen Testing, 2nd edition, Milman, H. A., and Weisburger, E. K., eds., William Andrew Publishing/Noyes, Norwich, NY, pp. 1–893. IARC, ed. (1999a). Species Differences in Thyroid, Kidney and Urinary Bladder Carcinogenesis, International Agency for Research on Cancer, World Health Organization, Lyon. IARC, ed. (1999b). The Use of Short- and Medium-Term Tests for Carcinogens and Data on Genetic Effects in Carcinogenic Hazard Evaluation, International Agency for Research on Cancer, World Health Organization, Lyon. IPCS (2006). Tetrachloroethene. Concise International Chemical Assessment Document 68, 1–123, http:// www.who.int/ipcs/publications/cicad/cicad68.pdf. JMPR (2006). Report of the Joint Meeting of the FAO Panel of Experts on Pesticide Residues in Food and the Environment and WHO the Core Assessment Group, FAO Plant Production and Protection Paper, 187 (Thiacloprid), ftp://ftp.fao.org/docrep/fao/010/a0888e/a0888e00.pdf. Kroes, R. (1987). Contribution of toxicology towards risk assessment of carcinogens. Arch Toxicol 60, 224–228. Liteplo, R. G., and Meek, M. E. (2003). Inhaled formaldehyde: Exposure estimation, hazard characterization, and exposure-response analysis. J Toxicol Environ Health B Crit Rev 6, 85–114. Maronpot, R. R., ed. (1994). Considerations in the Evaluation and Interpretation of Long-Term Animal Bioassays for Carcinogenicity, William Andrew Publishing/Noyes, Norwich, NY. McConnell, E. E., Solleveld, H. A., Swenberg, J. A., and Boorman, G. A. (1986). Guidelines for combining neoplasms for evaluation of rodent carcinogenesis studies. J Natl Cancer Inst 76, 283–289. Meek, B., and Doull, J. (2009). Pragmatic challenges for the vision of toxicity testing in the 21st century in a regulatory context: Another Ames test? … or a new edition of “the Red Book”? Toxicol Sci 108, 19–21. Meek, M. E. (2008). Recent developments in frameworks to consider human relevance of hypothesized modes of action for tumours in animals. Environ Mol Mutagen 49, 110–116. Meek, M. E., and Armstrong, V. C. (2007). The assessment and management of industrial chemicals in Canada. In Risk Assessment of Chemicals: An Introduction, Van Leeuwen, K., and Vermeire, T., eds., Kluwer Academic Publishers, Dordrecht, the Netherlands, pp. 591–621. Meek, M. E., Bucher, J. R., Cohen, S. M., Dellarco, V., Hill, R. N., Lehman-McKeeman, L. D., Longfellow, D. G., Pastoor, T., Seed, J., and Patton, D. E. (2003). A framework for human relevance analysis of information on carcinogenic modes of action. Crit Rev Toxicol 33, 591–653. Meek, M. E., Newhook, R., Liteplo, R. G., and Armstrong, V. C. (1994). Approach to assessment of risk to human health for priority substances under the Canadian environmental protection act. J Environ Sci Health, Part C: Environ Carcinog Ecotoxicol Rev 12, 105–134. NRC (1983). Risk Assessment in the Federal Government: Managing the Process Working Papers, National Academies Press, http://www.nap.edu/openbook.php?isbn=POD115&page=R1, Washington, D.C. NRC (1993). Issues in Risk Assessment, National Academies Press, http://books.nap.edu/openbook. php?record_id=2078&page=R1, Washington, D.C. NRC (2007a). Applications of Toxicogenomic Technologies to Predictive Toxicology and Risk Assessment, National Academies Press, http://books.nap.edu/openbook.php?record_id=12037&page=R1, Washington, D.C. NRC (2007b). Toxicity Testing in the 21st Century: A Vision and a Stategy, National Academies Press, Washington, D.C., pp. 1–196. NTP (1984). Report of the NTP Ad Hoc Panel on chemical carcinogenisis testing and evaluation. Board of Scientific Counselors, National Toxicology Program, pp. 1–280. OECD (1981). Carcinogenicity studies. OECD Guideline for Testing of Chemicals 451, 1–17. OECD (2002). Guidance Notes for Analysis and Evaluation of Chronic Toxicity and Carcinogenicity Studies. OECD Series on Testing and Assessment No. 35 and Series on Pesticides No. 14, ENV/JM/ MONO(2002)19. http://www.olis.oecd.org/olis/2002doc.nsf/LinkTo/env-jm-mono(2002)19. OSTP (1986). Chemical carcinogens: a review of the science and its associated principles. U.S. Interagency Staff Group on Carcinogens. Environ Health Perspect 67, 201–282.

396


Pastoor, T., Rose, P., Lloyd, S., Peffer, R., and Green, T. (2005). Case study: Weight of evidence evaluation of the human health relevance of thiamethoxam-related mouse liver tumors. Toxicol Sci 86, 56–60. Renwick, A. G., Barlow, S. M., Hertz-Picciotto, I., Boobis, A. R., Dybing, E., Edler, L., Eisenbrand, G., Greig, J. B., Kleiner, J., Lambe, J., Muller, D. J., Smith, M. R., Tritscher, A., Tuijtelaars, S., van den Brandt, P. A., Walker, R., and Kroes, R. (2003). Risk characterisation of chemicals in food and diet. Food Chem Toxicol 41, 1211–1271. Seed, J., Carney, E. W., Corley, R. A., Crofton, K. M., DeSesso, J. M., Foster, P. M., Kavlock, R., Kimmel, G., Klaunig, J., Meek, M. E., Preston, R. J., Slikker, W., Jr., Tabacova, S., Williams, G. M., Wiltse, J., Zoeller, R. T., Fenner-Crisp, P., and Patton, D. E. (2005). Overview: Using mode of action and life stage information to evaluate the human relevance of animal toxicity data. Crit Rev Toxicol 35, 664–672. Sonich-Mullin, C., Fielder, R., Wiltse, J., Baetcke, K., Dempsey, J., Fenner-Crisp, P., Grant, D., Hartley, M., Knaap, A., Kroese, D., Mangelsdorf, I., Meek, E., Rice, J. M., Younes, M., and International Programme on Chemical, S. (2001). IPCS conceptual framework for evaluating a mode of action for chemical carcinogenesis. Regul Toxicol Pharmacol 34, 146–152. Straif, K., Benbrahim-Tallaa, L., Baan, R., Grosse, Y., Secretan, B., El Ghissassi, F., Bouvard, V., Guha, N., Freeman, C., Galichet, L., and Cogliano, V., on behalf of the WHO International Agency for Research on Cancer Monograph Working Group (2009). A review of human carcinogens—part C: metals, arsenic, dusts, and fibres. Lancet Oncol 10, 453–454. Swenberg, J. A., Fryar-Tita, E., Jeong, Y. C., Boysen, G., Starr, T., Walker, V. E., and Albertini, R. J. (2008). Biomarkers in toxicology and risk assessment: Informing critical dose-response relationships. Chem Res Toxicol 21, 253–265. Tarone, R. E. (1982). The use of historical control information in testing for a trend in proportions. Biometrics 38, 215–220. UNECE (2007). Amendments to the Globally Harmonized System of classification and labelling of chemicals (GHS). United Nations, Geneva. Document ST/SG/AC.10/34/Add.3, http://www.unece.org/ trans/danger/publi/ghs/ghs_rev01/01amend_e.html. Van Leeuwen, C. J., Patlewicz, G. Y., and Worth, A. P. (2007). Intelligent testing strategies. In Risk Assessment of Chemicals: An Introduction, van Leeuwen, K., and Vermeire, T., eds., Kluwer Academic Publishers, Dordrecht, the Netherlands, pp. 467–504. Vanio, H., and Wilbourn, J., eds. (1994). International Perspectives on Carcinogenicity Testing—A Brief Overview. William Andrew Publishing/Noyes, Norwich, NY. Vermeire, V. G., Baars, B. J., Bessems, J. G. M., Blaauboer, B. J., Slob, W., and Muller, J. J. A. (2007). Toxicity testing for human health risk assessment. In Risk Assessment of Chemicals: An Introduction, Van Leeuwen, K., and Vermeire, T., eds., Kluwer Academic Publishers, Dordrecht, the Netherlands, pp. 467–504. Yamasaki, H. (1990). Gap junctional intercellular communication and carcinogenesis. Carcinogenesis 11, 1051–1058.

CH A P TE R

15

CANCER EPIDEMIOLOGY Herman J. Gibb Jessie P. Buckley

15.1.

INTRODUCTION

Cancer is the leading cause of death worldwide (WHO 2008) and the second leading cause of death in the United States (Kung et al. 2008). There were 7.9 million deaths and 11.3 million new cases of cancer worldwide in 2007 with the number of cases expected to increase over the next 20 years (WHO 2008). Lung, liver, stomach, colon, and breast cancer are the five leading global causes of cancer mortality. The World Health Organization (WHO) has estimated that 40% of cancer deaths worldwide are preventable (WHO 2008). Although overall cancer incidence and mortality in the United States is declining, it has been estimated that there were 1.4 million new cases of cancer and 565,650 cancer deaths in 2008 (NCI 2007a, 2008a). In addition, the incidence of several types of cancer is increasing: non-Hodgkin’s lymphoma, leukemia, multiple myeloma, liver cancer, pancreatic cancer, kidney cancer, thyroid cancer, esophageal cancer, testicular cancer in men, melanoma and cancers of the brain and bladder in women, and childhood cancer (NCI 2007a). Cancer epidemiology is designed to identify cancer risks in human populations and determine causal links between cancers and specific exposures. A vast number of etiological agents are of interest in cancer epidemiology, including environmental and occupational factors, infectious agents, lifestyle factors such as nutrition, smoking, or exercise, and genetic factors. Epidemiology studies investigate what makes one group of people at higher risk than another group, determine whether observed relationships are causal, and measure the strength of association between exposure and disease. In 1775, in what is probably the earliest reported evidence of occupationally associated cancer, Percival Pott reported that cancer of the scrotum was particularly prevalent and occurred at an unusually early age among chimney sweeps (Lilienfeld et al. 1967). Pott also postulated that some characteristic of chimney sweeps was relevant to the production of the disease. Following Pott’s observations, other studies reported increased risks of cancer among certain occupationally and environmentally


397

398

CHAPTER 15 CANCER EPIDEMIOLOGY

exposed populations, but it was not until the middle of the 20th century that a substantial body of knowledge on cancer risk factors began to accumulate. Also, about that time, the increase in mortality from lung cancer had become such an issue that the disease took on the characteristics of an epidemic, and it was apparent that something more than industrial causes was to blame. By 1950, reports were published on the relationship between lung cancer and cigarette smoking which led to the intensive study of the carcinogenic effects of tobacco and eventually to the active discouragement of cigarette smoking as an integral part of preventive medicine. Today virtually every type of cancer has received some attention from an epidemiologic perspective. Because of the variation in the incidence of different forms of cancer, it became recognized that environmental, occupational, and lifestyle factors play a major role in cancer risk. As early as 1964, the WHO declared that 75% of all human cancer was affected by extrinsic factors. In 1965, the International Agency for Research on Cancer (IARC) was created to focus on human cancer and the relationship of humans to their environment. Although in vitro and animal studies may indicate that a chemical is carcinogenic, human data are the highest standard of evidence for determining the association between exposure to a hazard and development of disease. Regulatory agencies such as the U.S. Environmental Protection Agency (EPA), IARC, and the National Research Council (NRC) have all stated that epidemiologic studies are the most convincing evidence that a hazard exists (EPA 2005; IARC 1999; NRC 1983). This chapter discusses (1) important issues relating to the study of cancer, (2) types of epidemiology studies, (3) the determination of causal association from epidemiologic evidence, and (4) the future for cancer epidemiology.

15.2. CONSIDERATIONS FOR THE EPIDEMIOLOGIC STUDY OF CANCER Cancer is not a single disease, but rather a process common to a very heterogeneous group of diseases, differing widely in etiology, in frequency, in pattern of occurrence, and in clinical manifestations, as well as in the diagnostic and therapeutic problems that they present (Lilienfeld et al. 1967). These considerations, while presenting considerable challenges to the epidemiologist, make the study of the disease fascinating in its intricacy.

15.2.1.

Demographics

The most commonly used demographic variables in epidemiology are age, race, and gender. Because these variables are relatively easy to study, a considerable amount of data has been amassed on these factors as they relate to cancer risk. 15.2.1.1. Age. Cancer has often been called a disease of old age. For most cancers, the incidence does increase with age, but that is not the case for all cancers. For example, the median age at diagnosis of testicular cancer is 34, and incidence

15.2. CONSIDERATIONS FOR THE EPIDEMIOLOGIC STUDY OF CANCER

399

TABLE 15.1. Incidence Rate per 100,000 of Selected Cancers by Age, Race/Ethnicity, and Sex—United States, 2001–2005 (SEER, 2008)

Age 3 weeks. Noted in the references below: Studies were carried out with mice (M) or rats (R) or both species (M,R). The endpoint is indicated for studies that measured oxidative stress. If there are inconsistent effects, the possible origin of the inconsistency is indicated. In vitro studies are also noted. References: 1Calfee-Mason et al. (2004) (R); 2Fischer et al. (2002) (increase in TBARS but not conjugated dienes) (R); 3Rusyn et al. (2000b,c) (M,R); 4Nilakantan et al. (1998) (M); 5Rusyn et al. (1998) (R); 6Wada et al. (1992) (lipofuscin) (R); 7Marsman et al. (1992) (lipofuscin; trend for increase in cell proliferation by clofibrate) (R); 8Conway et al. (1989) (lipofuscin—positive for both WY-14,643, and DEHP but only WY-14,643 positive for conjugated dienes) (R); 9Reddy et al. (1982) (lipofuscin) (R); 10Cattley et al. (1987) (lipofuscin) (R); 11Tharappel et al. (2003) (M); 12Tharappel et al. (2001) (consistent changes with WY-14,643, but only one condition resulted in increases in NF-kB activation after gemfibrozil treatment) (R); 13Menegazzi et al. (1997) (R); 14Rao et al. (1987) (lipofuscin) (R); 15 Lake et al. (1987) (lipofuscin) (R); 16Rao et al. (1982) (lipofuscin) (R); 17Rao et al. (1991) (lipofuscin) (R); 18Goel et al. (1986) (lipid peroxidation and hydrogen peroxide) (R); 19Li et al. (1996) (R); 20 Hinton et al. (1986) (lipofuscin) (R); 21Stanko et al. (1995) (lipofuscin) (R); 22Lake et al. (1989a) (increases in oxidized glutathione and decreases in vitamin E) (R); 23Tomaszewski et al. (1990) (in vitro cultures; oxidized dienes) (R); 24Cai et al. (1995) (lipid peroxidation—trend increases for PFOA, nafenopin and clofibrate) (M); 26Bility et al. (2004) (in vitro trans-activation assays) (M); 27Corton and Lapinskas (2005) (review of in vitro trans-activation data) (M,R); 28Isseman and Green (1990) (in vitro trans-activation assays) (M); 29Gottlicher et al. (1992) (in vitro trans-activation assays) (R); 30Corton et al. (2000) (review) (M,R); 31Marsman et al. (1988) (R); 32Smith-Oliver and Butterworth (1987) (R); 33Tanaka et al. (1992) (R); 34Yeldandi et al. (1989) (chronic increases in cell proliferation) (R); 35 Lake et al. (1993) (R); 36Schulte-Hermann et al. (1981) (R); 37Chen et al. (1994) (R); 38Dwivedi et al. (1989) (M); 39Barrass et al. (1993) (R); 40Seo et al. (2004) (malondialdehyde) (R); 41Isenberg et al. (2001) (M,R); 42Isenberg et al. (2000) (M,R); 43Hasmall et al. (2000) (R, in vivo (DEHP) and in vitro (MEHP)); 44Soames et al. (1999) (R); 45Busser and Lutz (1987) (R); 47Reddy and Qureshi (1979) (R); 48 Svoboda and Azarnoff (1979) (R); 49Reddy and Rao (1977) (R); 50Rao et al. (1986) (R); 51Elliott and Elcombe (1987) (malondialdehyde—significant change for DEHP and clofibrate but trend increase for methyl clofenapate) (R); 54James and Roberts (1996) (in vitro) (M,R); 55Youssef et al. (2003) (R); 59Thottassery et al. (1992) (R); 62Abdellatif et al. (1990) (initiated with DEN, 2-acetylaminofluorene and carbon tetrachloride) (R); 66Nicholls-Grzemski et al. (2000) (TBARS) (M); 71Soliman et al. (1997) (F2-isoprostanes) (R); 72Fischer et al. (2002) (TBARS increase with treatment but conjugated dienes do not) (R); 73Marsman and Popp (1994) (R); 74Rose et al. (1999a) (R); 75O’Brien et al. (2001b) (decreases in vitamin E) (R); 76Qu et al. (2000); 80Price et al. (1992) (R); 83Bursch et al. (1984) (R, in vivo); 84 Dostalek et al. (2008) (M, increases in hydrogen peroxide, malondialdehyde, and urine F2-isoprostanes but not liver F2-isoprostanes); 87Ohmura et al. (1996) (R).

17.3. MODE OF ACTION ANALYSIS IN THE U.S. EPA RISK ASSESSMENT FRAMEWORK

445

in (Klaunig et al. 2003)); and (3) importantly, the majority of studies using PPARαnull mice do not show hepatocyte-specific changes associated with hepatocarcinogenesis (discussed below). There are examples where PPARα activation does not consistently lead to liver cancer, and these have been summarized in Klaunig et al. (2003). Weak PPARα activators (i.e., compounds that minimally induce markers of PPARα activation) would not necessarily increase liver tumor incidence, because a sufficient level of receptor activation is needed for induction of key events (Klaunig et al. 2003). Pharmacokinetic differences between susceptible and nonsusceptible rodents that lead to differences in tissue chemical concentration could also contribute to discrepancies between the ability of chemicals to activate PPARα in trans-activation assays and tumor induction. For example, trichloroacetate (TCA) exposure in mice leads to increases in PCO activity at doses similar to or below those that induce liver tumors whereas in rats TCA, even at high doses, only marginally increases PCO in the absence of increases in liver tumors (Corton 2008). PPARα regulates lipid homeostasis and peroxisome proliferation through the modulation of genes involved in fatty acid uptake, activation, and oxidation as well as peroxisome assembly (the Pex genes) (Desvergne et al. 1998; Desvergne and Wahli 1999; Schoonjans et al. 1996; Wahli et al. 1995). Collectively, these changes result in increased ability to metabolize fatty acids leading to the therapeutic lowering of lipid levels in mice, rats, Syrian hamsters, guinea pigs, monkeys, and humans. These changes have been shown to be PPARα-dependent in mice (summarized in (Peters et al. 2005)). Alterations in lipid metabolism and peroxisome proliferation genes are not thought to be involved in the hepatocarcinogenic effects of PPARα activators (Klaunig et al. 2003). 17.3.2.2. Role of Oxidative Stress in PPARα Activator-Induced Hepatocarcinogenesis. Linkages exist between increases in ROS and increased incidence of liver cancer by PPARα activators. Overproduction of oxidants might cause DNA damage leading to mutations and cancer (Reddy and Rao 1989; Yeldandi et al. 2000). In whole liver of both rats and mice, markers of oxidative stress were increased by PPARα activators (Table 17.1), as determined by measuring indices of lipid peroxidation (e.g., conjugated dienes, lipofuscin, malondialdehyde, F2isoprostanes, etc.), oxidized glutathione, or hydrogen peroxide. A few studies failed to detect increases in markers of oxidative stress, but these are difficult to interpret because other key events were not simultaneously analyzed (e.g., Huber et al. 1991, 1997). There were other studies in which one assay for oxidative stress was positive but another negative (e.g., Conway et al. 1989; Fischer et al. 2002). In spite of these minor discrepancies, the WOE demonstrates that PPARα activators increase oxidative stress. Possible sources of ROS in the livers of rodents exposed to PPARα activators include enzymes that generate and degrade hydrogen peroxide and other ROS. Hydrogen peroxide can oxidize DNA, lipids, and other molecules, and PPARα activators regulate the expression of many enzymes that produce hydrogen peroxide as a byproduct of metabolism including the peroxisomal, mitochondrial, and microsomal oxidases in hepatocytes such as fatty acyl-CoA oxidase (ACO) (Becuwe and

446

CHAPTER 17 HUMAN RELEVANCE OF LIVER TUMORS

Dauca 2005). Administration of PPARα activators can also lead to decreased levels of some enzymes, which degrade ROS that may contribute to the increases in oxidative stress upon exposure (Glauert et al. 1990; O’Brien et al. 2001a,b). The individual contributions of these enzymes to increases in oxidative stress and downstream key events leading to liver tumor induction has not been comprehensively addressed but will likely be complex. In one example, Reddy and co-workers originally proposed that peroxisomal ACO (Acox1) is the enzyme responsible for oxidative stress by PPARα activators (Nemali et al. 1988). However, ACO was later found to be dispensable for increases in oxidative stress. Control ACO-null mice exhibited the phenotype of wild-type mice exposed to PPARα activators including increases in oxidative stress and induction of liver tumors that are dependent on PPARα (Fan et al. 1998; Hashimoto et al. 1999). The role of other ACO family members (Acox2, Acox3) has not been determined in this Acox1-independent induction of oxidative stress and liver tumors. Extensive testing of PPARα activators has shown that these compounds do not consistently induce direct DNA damage. However, indirect DNA damage from oxidative stress has been hypothesized to be a common pathway for many nongenotoxic chemical carcinogens including PPARα activators (Klaunig et al. 1998). Relationships exist between chemical exposure, DNA damage, and cancer based on measurement of 8-hydroxy-2′-deoxyguanosine (8-OH-dG), a highly mutagenic lesion, in DNA isolated from livers of animals treated with PPARα activators (Kasai 1997; Qu et al. 2001; Takagi et al. 1990). However, subsequent studies showed that the increases in oxidative DNA damage may have originated in the way in which the genomic DNA was prepared (Cattley and Glover 1993; Sausen et al. 1995). Experiments measuring other indicators of DNA damage—that is, abasic sites or single strand-breaks in genomic DNA from rats and mice treated with WY-14,643 for one month—failed to show increases over controls (Rusyn et al. 2004). Only in the livers of wild-type but not PPARα-null mice treated with WY-14,643 for five months were there increases in abasic sites in genomic DNA (Woods et al. 2007a), indicating that exposure times longer than one month were necessary to observe increases in DNA damage. The relationship between the increases in abasic sites and subsequent tumor yield has not been determined. DNA repair mechanisms might compensate for increases in DNA damage and may explain the lack of consistent evidence for DNA damage from PPARα activator-induced oxidative stress. PPARα activators increased the expression of liver genes involved in the long-patch base excision DNA repair pathway in a timedependent manner; the degree of induction roughly correlated with the dose and carcinogenic potency of the PPARα activators tested (Rusyn et al. 2000a). Additionally, expression of enzymes that do not repair oxidative DNA damage was not changed. This induction of DNA base excision repair genes may be an indicator that DNA damage is occurring. Evidence that DNA damage caused by PPARα activator-induced oxidative stress is not involved in hepatocarcinogenesis comes from recent work with Ogg1null mice. Ogg1 encodes an 8-oxoguanine DNA glycosylase that repairs one of the major DNA lesions generated by ROS. Control Ogg1-null mice show elevated levels of oxidative DNA damage and exhibit increased spontaneous mutation rates in the


447

absence of chemical exposure (Klungland et al. 1999). Ogg1-null mice, when exposed to WY-14,643 in the diet, did not show additional oxidative DNA damage but exhibited increased numbers and total volumes of preneoplastic lesions in the liver compared to similarly treated wild-type mice (Trapp et al. 2007). The authors concluded that the increase in preneoplastic lesions associated with WY-14,643 exposure did not arise from induced oxidative damage, but instead arose from the promotion of spontaneous mutations generated by endogenous oxidative DNA damage. Overall, PPARα activators increase the level of oxidative stress through multiple mechanisms. There is little direct evidence that increases in oxidative stress generated after PPARα activator exposure leads to direct or indirect DNA damage. The Ogg1-null mouse studies indicate that PPARα activators promote hepatocytes that have been spontaneously initiated. The WOE suggests that direct or oxidatively induced DNA damage is not part of the MOA. 17.3.2.3. Role of NF-kB in the PPARα Activator MOA. Central to the PPARα activator MOA is NF-kB activation. NF-kB transcription factors play critical roles in cancer development and progression (Arsura and Cavin 2005; Karin 2006). A wealth of data demonstrates that NF-kB is activated under conditions of inflammation and oxidative stress (Czaja 2007; Gloire et al. 2006). Consistent with this, studies with PPARα activators demonstrate linkages between oxidative stress and NF-kB activation. Activation is usually assessed by the ability of nuclear NF-kB (usually a heterodimer composed of p50 and p65 subunits) to bind to a NF-kB response element in an electrophoretic mobility shift assay (EMSA). In whole liver of both rats and mice, activity of NF-kB was increased by PPARα activators including WY-14,643, ciprofibrate, and gemfibrozil but not nafenopin (Table 17.1). The fact that nafenopin did not induce NF-kB may be due to differences in the EMSA procedures carried out by that lab (Menegazzi et al. 1997; Ohmura et al. 1996). NF-kB is activated in Kupffer cells and in hepatocytes at different times after exposure. After a single in vivo dose of WY-14,643, NF-kB activity was increased first in Kupffer cells (at 2 hours), and only ∼6 hours later was NF-kB activity increased in hepatocytes. Activation in hepatocytes never achieved the level observed in Kupffer cells (Rusyn et al. 1998). The increase in NF-kB activation in hepatocytes could be due to increases in mitogenic cytokines produced by Kupffer cells that activate signal transduction pathways ultimately impinging on NF-kB. Alternatively, NF-kB can be activated directly by a PPARα activator in the H4IIEC3 rat hepatoma cell line, responsive to the proliferative effects of PPARα activators (Li et al. 2000a). Increased NF-kB activity may be secondary to the action of hydrogen peroxidegenerating enzymes, such as ACO, since overexpression of ACO in COS-1 cells, in the presence of a hydrogen peroxide-generating substrate, can activate a NF-kBregulated reporter gene (Li et al. 2000b). 17.3.2.4. Alteration of Cell Proliferation/Apoptosis Balance by PPARα Activators. PPARα activators produce multiple tumor precursor effects including liver hyperplasia, and altered growth in preneoplastic foci. Increased cell replication induced by PPARα activators may increase the frequency of spontaneous mutations

448


by increasing the frequency of errors in DNA repair or replication and can lead to silencing of tumor suppressor genes or increased expression of oncogenes (Cattley et al. 1998; Huber et al. 1991). Alternatively, PPARα activators can promote the growth of spontaneously initiated hepatocytes. All PPARα activators at a sufficient dose produce a strong, albeit transient, increase in replicative DNA synthesis during the first few days of exposure (Table 17.1). After this initial burst in replication, baseline levels of hepatocyte replication are approached while the liver remains enlarged. Many PPARα activators exhibit measurable sustained or chronic increases in cell proliferation, although the levels are much lower than that observed after acute exposures (Table 17.1). There are some PPARα activators that do not induce chronic cell proliferation; this may be due to the dose used in the experiment and because weak increases above variable background levels of cell proliferation are difficult to detect. PPARα activators promote the growth of chemically and spontaneously induced lesions through enhanced cell replication (Cattley et al. 1987; Isenberg et al. 1997; Marsman et al. 1988). Once early lesions are formed, continued exposure to PPARα activators causes a selective increase in DNA replication of up to ∼40% in these liver foci, while replication of hepatocytes in the normal surrounding liver is increased only slightly (Grasl-Kraupp et al. 1993). Furthermore, preneoplastic foci respond to the cell replicative effects, rather than the peroxisome proliferation effects of PPARα activators, suggesting that the growth stimulus but not the peroxisome proliferation effect is of particular significance for the carcinogenic action of this class of compounds (Grasl-Kraupp et al. 1993). Increases in cell proliferation alone are not sufficient to increase liver tumors. The response of mice transgenic for hepatocyte-specific expression of a constitutively activated form of PPARα (VP16PPARα) was compared to wild-type mice treated with WY-14,643 (Yang et al. 2007). Expression of VP16PPARα led to increases in hepatocyte proliferation in the absence of nonparenchymal cell proliferation, in contrast to WY-14,643 treatment in wild-type livers in which both hepatocytes and nonparenchymal cells exhibited increased replication. Importantly, chronic activation of VP16PPARα did not increase liver tumors (Yang et al. 2007). These results indicate that nonparenchymal cell activation is important for hepatocarcinogenesis and that PPARα-mediated hepatocyte proliferation by itself is not sufficient to induce liver cancer. Taken together, the results indicate that it is the combination of events in hepatocytes and NPC that are important for induction of liver tumors by PPARα activators. Nongenotoxic carcinogens, in general, and PPARα activators in particular suppress hepatocyte apoptosis. Suppression of apoptosis could inhibit the ability of the liver to remove DNA-damaged, pre-neoplastic hepatocytes (Bayly et al. 1994; James and Roberts 1996; Oberhammer and Qin 1995; Schulte-Hermann et al. 1981). Most of the evidence for apoptosis suppression comes from in vitro studies because of the difficulty in measuring the suppression of already low levels of apoptosis in vivo. Studies conducted in vitro show that the PPARα activators nafenopin, methylclofenapate, and WY-14,643 suppress spontaneous hepatocyte apoptosis as well as that induced by a negative regulator of liver growth, transforming growth factor beta 1 (TGFβ1) (Bayly et al. 1994; Oberhammer and Qin 1995) (Table 17.1).


449

In addition, PPARα activators can suppress apoptosis in vitro induced by diverse stimuli such as DNA damage or ligation of Fas, a receptor related to the tumor necrosis factor alpha (TNFα) family of cell surface receptors (Gill et al. 1998). A limited number of in vivo studies also showed suppression of apoptosis after acute dosing with nafenopin, DEHP or WY-14,643 (Bursch et al. 1984; James et al. 1998b; Youssef et al. 2003). Suppression of apoptosis by PPARα activators occurs under acute exposure conditions when the liver is increasing in size. However, once a steady state of liver enlargement is reached, levels of apoptosis are likely to return to background levels or to levels which balance the low level of cell proliferation that occurs for some PPARα activators. Consistent with this, two reports suggest that chronic exposure of rats and mice to the PPARα activator WY-14,643 results in an increase in apoptosis (Burkhardt et al. 2001; Marsman et al. 1992). Furthermore, PPARα activators alter the ability of the liver to respond to apoptosis inducers. Sensitivity to two apoptosis inducers (Jo2 antibody and conconavalin A) was dramatically increased in wild-type but not PPARα-null mice exposed for 1 week to WY-14,643 (Xiao et al. 2006). Lastly, both cell proliferation and apoptosis increase in parallel in PPARα activator-induced tumors in the rat compared with normal surrounding tissue, suggesting that cell turnover is increased in tumorigenic lesions (GraslKraupp et al. 1997). To summarize, alterations in the balance between hepatocyte proliferation and apoptosis have been observed after exposure to multiple PPARα activators at different stages of carcinogenesis including under acute and chronic exposure conditions and in the preneoplastic and tumorigenic lesions. 17.3.2.5. Mechanisms of Cell Growth Alterations. Extensive work has been carried out to identify the mechanistic events that lead to alterations in cell growth by PPARα activators. There are a number of excellent reviews on the subject of signal transduction and downstream events that lead to alterations in cell growth (Burns and Vanden Heuvel 2007; Gonzalez and Shah 2008; Rusyn et al. 2006). Early studies focused on the regulation of individual growth genes that respond to growth promoting stimuli. More recent studies capitalized on technological advancements in assessing global changes in gene expression or assessing the role of individual genes/pathways in the intact animal using transgenic technologies. Many studies focused on growth factors derived from the Kupffer cell. Activated NPCs, particularly Kupffer cells, produce cytokines such as TNFα, interleukin-1α, and interleukin-1β (IL1α, IL1β). These cytokines affect the fate of neighboring hepatocytes. TNFα is able to increase hepatocyte proliferation and suppress apoptosis in cultured rodent hepatocytes (Holden et al. 2000; Rolfe et al. 1997). In intact animals hepatocyte growth can be prevented by injection of antibodies to either TNFα (Bojes et al. 1997; Rolfe et al. 1997) or TNFα receptor 1 (West et al. 1999). PPARα activators increased TNFα mRNA more than twofold (Bojes et al. 1997; Rolfe et al. 1997). Because increases in TNFα expression have not been consistently observed by others (Anderson et al. 2001; Holden et al. 2000), treatment with PPARα activators may not result in de novo TNFα expression, but

450


rather bioactivation or release of preexisting TNFα protein from Kupffer cells (Holden et al. 2000). Other studies suggest that the cell proliferation response to PPARα activators is TNFα pathway-independent. Cell proliferation remained intact in TNFα-null and in TNFα receptor-null mice given a PPARα activator (Anderson et al. 2001; Lawrence et al. 2001). In addition, IL-1 receptor-null mice retained the ability to respond to the induction of hepatocyte proliferation to WY-14,643 (Corton et al., unpublished observations). There remains the possibility that loss of TNFα or IL-1 signaling results in compensation by other genes/pathways, including other cytokine-mediated pathways, because multiple growth modulators secreted by Kupffer cells have been suggested to play a role in hepatocyte proliferation after diethylnitrosamine (DEN) exposure (Maeda et al. 2005). Thus, studies with various nullizygous mice do not necessarily refute the role TNFα or IL-1 may play in PPARα activator-induced cell proliferation. MicroRNAs (miRNA) play important roles in complex processes such as development through the regulation of gene expression. Recent global analysis of the miRNA expression pattern after WY-14,643 exposure has uncovered a signaling pathway which culminates in increased expression of the c-Myc growth regulatory gene (Shah et al. 2007). Expression of let-7C, an miRNA important in cell growth, was down-regulated following acute or chronic treatment with WY-14,643 in wild-type mice. Because let-7C down-regulates the expression of c-Myc, the down-regulation of let-7C by WY-14,643 resulted in increased expression of cMyc. These molecular events did not occur in PPARα-null mice. These studies reveal a let-7C signaling cascade critical for PPARα activator-induced hepatocyte proliferation. Other growth signaling pathways may be involved in PPARα activator growth responses, but overall the data supporting their role is usually confined to gene expression data. Due to the lack of useful genetic models, there is little mechanistic data, which shows causal links between specific pathways and modulation of cell fate except for the role of PPARα and NF-kB activation (discussed below). 17.3.2.6. Genetic and Biochemical Inhibition Studies Support the MOA. Genetic and biochemical inhibition studies have highlighted the relationships between the key events of the PPARα activator MOA (Table 17.2). These studies showed that when a key event is inhibited genetically or biochemically, the downstream but not upstream event(s) are inhibited as well. Genetically modified mice have been useful to show the relationships between the key events in the PPARα MOA. PPARα-null mice provided critical evidence establishing the rodent MOA for PPARα activator-induced hepatocarcinogenesis. Evidence that a particular compound induces key events in wild-type mice but not in mice lacking PPARα would be considered strong support for a PPARα MOA for that particular compound. To date, three chronic bioassays have been conducted in these mice (Hays et al. 2005; Ito et al. 2007; Peters et al. 1997). A greater body of data exists in which precursor events for cancer have been assessed in wild-type and PPARα-null mice after acute or subacute exposures.

TABLE 17.2.

Effects of Inhibition of Key Events in the PPARα Activator MOA

Key Event Mechanism of Inhibition Genetic Inhibition PPARα-null Catalase transgenic P47Phox-null P50-null Biochemical Inhibition Antioxidants in diet Dexamethasone Glycine Methylpalmitate Diphenyleneiodonium

PPARα Activation

Oxidative Stress

NF-kB Activation

Alteration in Hepatocyte Growth

Clonal Expansion

↓ (by definition) NC11 NC1,27,29 NC8,9

↓1,27↑32

↓1 ↓11 ↓29, NC27 ↓8

↓2,3,4 ↓11 ↓29, NC27 ↓8,9

↓2,4

↓2,4↑32

↓9

↓9

NC6,10 NC14,15 ↓23 NC19,21 NC30 NC29

↓7,16↑10

↑10

↓5,12,13↑10

↓29, NC1

↓6,7,26 ↓17,18,24,25

↓20,22 ↓29

↓29

↓ ↓19, NC21 ↓30 ↓29

Liver Tumors

14,15,23

↓21

↓, inhibited; NC, no change; ↑, increases in the parameters measured. For studies in which antioxidants were co-treated with PPARα activators, the antioxidant is indicated in parentheses. References: 1Woods et al. (2007a); 2Peters et al. (1997); 3Peters et al. (1998); 4Hays et al. (2005); 5Rao et al. (1984) [ethoxyquin, 2(3)-tertbutyl-14-hydroxyanisole]; 6Calfee-Mason et al. (2004) (vitamin E); 7Li et al. (2000a) (in vitro studies with vitamin E-treated H4IIE3C cells); 8Tharappel et al. (2003); 9Glauert et al. (2006); 10Glauert et al. (1990) (vitamin E increases the number of tumors while depleting glutathione reserves); 11Nilakantan et al. (1998); 12Rao and Subbarao (1999) (dimethylthiourea); 13Rao and Subbarao (1997a) (deferoxamine—iron chelator); 14 Lawrence et al. (2001c); 15Rao and Subbarao (1997b) (dexamethasone); 16Stanko et al. (1995) (vitamin E); 17Ray and Prefontaine (1994); 18Widen et al. (2003); 19Rose et al. (1997a,b); 20Rose et al. (1999a) (superoxide production in Kupffer cells); 21Rose et al. (1999b); 22Rusyn et al. (2001) (free radicals in bile); 23Ohmura et al. (1996) (measured peroxisomal bifunctional enzyme as PPARα marker); 24Chang et al. (1997); 25De Bosscher et al. (2006) (review); 26Rusyn et al. (1998) (allopurinol); 27Woods et al. (2007b); 29Rusyn et al. (2000b,c); 30Rose et al. (1997b); 32Ito et al. (2007).

451

452


Two studies assessed markers of oxidative stress in wild-type and PPARαnull mice. In the first study, abasic sites (i.e., sites that lack either a purine or a pyrimidine) in genomic DNA were used as a measure of oxidative stress. These sites were increased in wild-type but not PPARα-null mice after exposure to WY-14,643 for 5 months (Woods et al. 2007a). In the second study, electron spin resonance (ESR) identified increases in free radicals in the bile of wild-type but not PPARαnull mice after up to 3-week exposures to WY-14,643 or DEHP. NF-kB activation was observed in the livers of wild-type but not PPARα-null mice after exposure to WY-14,643 (Woods et al. 2007a,b). Using global gene expression profiling, alteration of gene expression by WY-14,643, PFOA, or ciprofibrate was almost completely abolished in PPARα-null mice at multiple time points (Anderson et al. 2004a,b; Corton et al. 2004; Rosen et al. 2008a,b; Sanderson et al. 2008; Woods et al. 2007c; Corton et al., unpublished). The up-regulation of the cell cycle components cyclin-dependent kinase (CDK)-1, CDK-2, CDK-4 and proliferating cell nuclear antigen (PCNA) proteins and CDK-1, CDK-4 and cyclin D1 mRNA was observed in wild-type but not PPARα-null mice fed WY-14,643 (Peters et al. 1998). Wild-type mice exhibited increased hepatocyte proliferation compared to untreated controls while no increases in hepatocyte proliferation were observed in PPARα-null mice after exposure to WY-14,643, DINP, PFOA, or trichloroethylene (Laughter et al. 2004; Peters et al. 1997, 1998; Valles et al. 2003; Wolf et al. 2008; Corton et al., unpublished). The ability of PPARα activators to suppress apoptosis was lost in hepatocytes isolated from PPARα-null mouse livers (Hasmall et al. 2000a). Importantly, chronic treatment with WY-14,643 or bezafibrate resulted in 100% incidence of hepatocellular neoplasia in wild-type mice while the PPARα-null mice were unaffected (Hays et al. 2005; Peters et al. 1997). An additional bioassay in which DEHP induced liver tumors in PPARα-null but not wild-type mice (Ito et al. 2007) is discussed below. Although the WY-14,643 and bezafibrate chronic exposure studies were carried out for relatively short exposure periods (up to a year), the PPARα-null mice did not exhibit any of the precursor events associated with carcinogenesis (Hays et al. 2005; Peters et al. 1997, 1998), making it unlikely that longer-term exposure would result in liver tumors. These studies demonstrate that all of the key events in the MOA are dependent on PPARα. Two transgenic mouse models have been used to determine the relationships between different sources of oxidative stress and downstream events. Catalase converts hydrogen peroxide to water and oxygen. In catalase-transgenic mice that exhibit increased liver expression and activity of catalase, there were decreased levels of NF-kB activation and decreased hepatocyte proliferation upon exposure to ciprofibrate (Nilakantan et al. 1998). NADPH oxidase in Kupffer cells plays an important role in generating superoxide radicals in response to Kupffer cell activators (De Minicis et al. 2006). NADPH oxidase is activated by PPARα activators and is important in cell proliferation after short-term PPARα activator exposure. Mice that lack one of the subunits of NADPH oxidase (the p47Phox-null mice) did not exhibit increases in oxidative stress, NF-kB activation, and hepatocyte proliferation after short-term PPARα activator exposure (Rusyn et al. 2000b,c). However, after exposure of mice to WY-14,643 for three weeks, there were increases in indicators of oxidative stress (including PCO activity), NF-kB activation and cell proliferation,


453

independent of the status of the p47Phox gene; these key events were dependent on PPARα (Woods et al. 2007a,b). Longer-term exposure may allow bypass of p47Phox dependence including increases in oxidative stress through activation of enzymes that produce hydrogen peroxide. NF-kB activation is involved in modulation of hepatocyte fate in response to inducers of oxidative stress [e.g., Maeda et al. (2005)] including PPARα activators. Wild-type mice and mice deficient in the p50 subunit of NF-kB (p50-null mice) were fed a diet with or without 0.01% ciprofibrate for 10 days. NF-kB DNA binding activity was increased after ciprofibrate treatment in wild-type mice but not p50-null mice. The apoptotic index was low in wild-type mice in the presence or absence of ciprofibrate. Consistent with NF-kB acting as a negative regulator of apoptosis (Arsura and Cavin 2005; Karin 2006), apoptosis was higher in untreated p50-null mice compared to wild-type mice (Tharappel et al. 2003). Apoptosis was reduced in p50-null mice after ciprofibrate feeding but was still higher than wild-type levels. The untreated p50-null mice had a higher level of hepatic cell proliferation, as measured by bromodeoxyuridine (BrdU) labeling, than did untreated wild-type mice possibly as a mechanism to compensate for the higher levels of apoptosis. However, ciprofibrate-fed p50-null mice had lower levels of cell proliferation than comparatively treated wild-type mice (Tharappel et al. 2003). A chronic (38-week) exposure study provides direct evidence that NF-kB activation is necessary for hepatocarcinogenesis induced by a PPARα activator (Glauert et al. 2006). Wild-type mice receiving only DEN developed a low incidence of tumors (25%). The majority of wild-type mice receiving both DEN + WY-14,643 developed tumors (63%). However, no tumors were seen in the DEN or DEN + WY14,643-treated p50-null mice, demonstrating that the p50 subunit of NF-kB was required for the promotion of hepatic tumors by WY-14,643. Treatment with DEN + WY-14,643 increased both cell proliferation and apoptosis in wild-type and p50-null mice. Consistent with the tumor levels, cell proliferation and apoptosis were lower in the p50-null mice than in wild-type mice (Glauert et al. 2006). This study shows direct dependence on the p50 subunit of NF-kB for liver tumor induction by a PPARα activator. Biochemical inhibition studies using compounds that inhibit oxidative stress or inflammation also highlight linkages of the key events in the PPARα MOA. In these studies, animals were pretreated with the inhibitor before PPARα activator exposure or co-treated with a PPARα activator and the inhibitor. The free radical scavenger and xanthine oxidase inhibitor allopurinol inhibited the activation of NF-kB in the livers of WY-14,643-treated rats (Rusyn et al. 1998). In in vitro studies, the anti-oxidants vitamin E or N-acetylcysteine blocked the ability of NF-kB to activate a reporter gene in ciprofibrate-treated HIIE3C cells (Li et al. 2000b). Co-treatment with ciprofibrate and one of two anti-oxidants, 2(3)-tert-butyl-14hydroxyanisole or ethoxyquin, decreased the incidence and size of liver tumors compared to ciprofibrate treatment alone (Rao et al. 1984). Studies using either dimethylthiourea or deferoxamine as antioxidants decreased the incidence of liver tumors in rats fed the PPARα activator ciprofibrate (Rao and Subbarao 1997a, 1999). When co-treating rats with the PPARα activator ciprofibrate and the antioxidant vitamin E, the levels of the antioxidant glutathione were paradoxically

454


depleted, and the animals exhibited increased tumor numbers (Glauert et al. 1990). In other studies vitamin E inhibited clofibrate-induced increases in lipofuscinlike products and ciprofibrate-induced increases in NF-kB activation in the absence of effects on markers of PPARα activation (Calfee-Mason et al. 2004; Stanko et al. 1995). Inhibition of key events by compounds that alter inflammatory states including Kupffer cell activation has been observed in multiple studies. The glucocorticoid receptor agonist dexamethasone is an anti-inflammatory agent that decreases the ability of NF-kB to be activated under a variety of inflammatory conditions (Chang et al. 1997; De Bosscher et al. 2006; Ray and Prefontaine 1994; Widen et al. 2003). Dexamethasone decreased PPARα activator-induced hepatocyte proliferation after acute exposures (Lawrence et al. 2001; Ohmura et al. 1996; Rao and Subbarao 1997b) while having either no effect (Lawrence et al. 2001; Rao and Subbarao 1997b) or modest decreases (Ohmura et al. 1996) on markers of PPARα activation. Compounds that inhibit Kupffer cell activation (e.g., glycine, methylpalmitate) or inhibit NADPH oxidase (e.g., diphenyleneiodonium) inhibited increases in oxidative stress and NF-kB activation after exposure to PPARα activators but had no effects on markers of PPARα activation (Rose et al. 1997a,b, 1999a,b; Rusyn et al. 2000b,c, 2001). While pretreatment with diphenyleneiodonium, glycine or methylpalmitate decreased acute cell proliferation (Rose et al. 1997a,b, 1999b; Rusyn et al. 2000b,c), glycine had no effect on chronic cell proliferation but did decrease the size and number of tumors (Rose et al. 1999a). Taken together, these biochemical and genetic inhibition studies demonstrate the linkages of the key events in the PPARα activator MOA. 17.3.2.7. Some PPARα Activators Exhibit Complex MOAs. Before a PPARα activator MOA can be defined as the primary MOA, alternative MOA(s) must be considered. Comparison of wild-type and PPARα-null mice have provided opportunities to determine if additional key events are necessary in addition to PPARα activation. In one example, PFOA was analyzed for liver effects in wild-type and PPARα-null mice. At two doses tested (1 and 3 mg/kg/day), PPARα-null mice lacked increases in cell proliferation but retained increases in liver to body weights. At the highest dose tested (10 mg/kg/day), PPARα-null mice exhibited increases in cell proliferation (Wolf et al. 2008). Microarray analysis using full-genome gene chips showed that PFOA altered ∼85% of the total number of genes in a PPARαdependent manner at 3 mg/kg/day. The PPARα-independent genes exhibited signatures of activation of other nuclear receptors. In particular, the PPARα-independent genes significantly overlapped with those regulated by the constitutive activated receptor (CAR), which regulates cell growth and xenobiotic metabolism genes including Cyp2b family members (Rosen et al. 2008a,b). These CAR signature genes were more robustly regulated in PFOA-treated PPARα-null mice compared to wildtype mice. These findings indicate that CAR activation may be a key event in the transcriptional and cell proliferation effects in PPARα-null mice. In wild-type mice, there were relatively minor alterations of CAR signature genes compared to the strong changes in PPARα-dependent genes indicating that CAR plays a minor role in mediating PFOA effects in wild-type mice (Rosen et al. 2008b).


455

The carcinogenic effects of DEHP were examined in wild-type and PPARαnull mice for 22 months (Ito et al. 2007). A low level of liver tumors was observed in PPARα-null mice but not in wild-type mice. These data suggest that an additional biological event may be operating in DEHP-induced rodent liver tumors.* The tumors in PPARα-null mice most likely arose through a mechanism that is not dominant in wild-type mice. Wild-type and PPARα-null mice did not exhibit equivalent levels of tumor induction. There were no statistically significant increases in liver tumors in the wild-type mice under these exposure conditions, indicating that the biological effects of exposure were not equivalent in these two strains. Expression of growth control genes showed responses in PPARα-null mice but not in wild-type mice at equivalent doses. In follow-up work from the same lab (Takashima et al. 2008), transcript profiling and reverse transcription polymerase chain reaction (RTPCR) showed highly dissimilar changes in gene expression in the liver tumors from the two strains. These data indicate that although DEHP can induce marginal increases in liver tumors in PPARα-null mice, the MOA is different from that in wild-type mice. DEHP is a inducer of Cyp2b family members in wild-type mice (Currie et al. 2005; Ren et al., 2010) suggesting that in the absence of PPARα, DEHP activates CAR, as the rate-limiting key event resulting in increases in liver tumors by a CAR-dependent pathway. In summary, chemicals may produce similar PPARα-independent effects defined in part as effects observed in PPARα-null mice. These effects may suggest additional key events that become the main control points in the absence of PPARα. A determination of the relative contribution of each proposed key event would require comparison of signature genes and biomarkers representing each key event in the two strains. 17.3.2.8. The PPARα Activator MOA Is Chemical-Independent. Mode of action is a series of key events that together result in an adverse health effect such as a liver tumor and as such is chemical-independent (Boobis et al. 2008; Holsapple et al. 2006; Meek 2008). Consistent with this the MOA for PPARα activators is an endogenous series of events that can occur independent of chemical exposure. Livers from ACO-null mice exhibit severe steatosis, increases in markers of PPARα activation (i.e., genes involved in β- and ω-fatty acid oxidation), increases in hydrogen peroxide levels, increases in cell proliferation and liver tumors (Fan et al. 1998). The increases in the markers of PPARα were shown to be PPARα-dependent as the changes were abolished in a double ACO-/PPARα-null mouse (Hashimoto et al. 1999). Microarray analysis of the tumors spontaneously induced in ACO-null mice showed extensive similarity with the liver tumors induced by the PPARα activator ciprofibrate, indicating the mechanism leading to the induction of the tumors was similar (Meyer et al. 2003). Additional mouse models nullizygous for other genes involved in fatty acid oxidation exhibit phenotypes indicative of constitutive PPARα activation (Jia et al. 2003). A mouse model of hepatitis C virus (HCV)-induced *It should be noted that the authors combined different types of liver tumors in their analysis, a nonstandard method of analyzing tumor data leaving open the possibility that the increase in tumor response is actually not statistically significant.

456


hepatocellular carcinoma (HCC) which overexpresses the HCV core protein was used to show that induction of oxidative stress, increases in cell proliferation, and liver tumor induction were PPARα-dependent (Tanaka et al. 2008a,b). The authors conclude that there “is the absolute requirement of persistent PPARα activation for the development of HCV core protein-induced steatosis and HCC” (Tanaka et al. 2008b). All of these mouse models exhibit disruption of fatty acid transport and metabolism resulting in increases in endogenous activators of PPARα including fatty acids (Fan et al. 1998; Tanaka et al. 2008a,b). Taken together, the PPARα MOA is operational in the absence of chemical exposure. Chemical PPARα activators will persistently activate this MOA resulting in liver tumors. 17.3.2.9. Species Differences in Responsiveness of Key Events in the PPARα MOA. Studies conducted in numerous test species indicate that while some rodents (mice and rats) are highly responsive to PPARα activatorinduced hepatocarcinogenicity and associated responses, other species (e.g., Syrian hamsters, dogs, guinea pigs, New and Old World primates, and humans) are less sensitive (Ashby et al. 1994; Bentley et al. 1993; Cattley et al. 1998; Doull et al. 1999). This difference is likely based in large part on differing levels of PPARα expression among species. In a side-by-side comparison, mice had ∼10-fold more PPARα expression than guinea pigs and ∼3-fold more than Syrian hamsters (Choudhury et al. 2004). Humans exhibited ≥10-fold lower expression than mice and rats (described in greater detail below). Thus, guinea pigs may be the more relevant model for PPARα activator effects in the human liver based solely on expression levels of the full-length active PPARα. Table 17.3 summarizes PPARα MOA key events in responsive species (rats and mice summarized from Table 17.1) compared to Syrian hamsters, guinea pigs, Cynomolgus monkeys, and humans. Due to the relative paucity of data for key events, other endpoints associated with exposure to PPARα activators are included (i.e., liver weight to body weight, hypolipidemic effects). Syrian hamsters and guinea pigs exhibit a partial PPARα activator response even though they are considered “nonresponsive species” compared to rats and mice. Fatty acid metabolism genes/proteins are only weakly activated after PPARα activator exposure in the livers of these species. Diminished responsiveness in guinea pigs is not due to a defective PPARα because when overexpressed in cell lines, PPARα from guinea pigs activates reporter genes to levels comparable to rats and mice (Bell et al. 1998; Macdonald et al. 1999; Tugwood et al. 1998). PPARα activators WY14,643 or methylclofenapate decrease triglycerides and very low density lipoproteins (VLDL) in Syrian hamsters and guinea pigs. Five out of the six PPARα activators examined increased liver to body weights in Syrian hamsters, but only one chemical (i.e., perfluorodecanoic acid) out of seven examined increased liver to body weights in guinea pigs; but for perfluorodecanoic acid, there was conflicting evidence of increases. WY-14,643 does not activate NF-kB in hamsters, indicating that this response is species-specific. Differences were also seen between species in relationship to cell proliferation. Studies measuring changes in cell proliferation in Syrian hamsters showed no response, a weak response, or inconsistent responses. Multiple studies showed guinea pigs did not exhibit increases in cell proliferation to four

TABLE 17.3.

Species Differences in Responses to PPARα Activators

Response

PPARα Activation Species

Rats

Relative PPARα expression Likely similar to mice

Mice

10

Syrian hamster

3

Guinea pig

1

Hypolipidemic Effect (Decreases in Triglycerides or VLDL-Triglycerides)

Increases in Liver Weight

Oxidative Stress

NF-kB Activation

Increases in Acute Cell Proliferation

Decreases in Apoptosis

Liver Tumors

Chemical

See table 1 for chemical and reference See table 1 for chemical and reference Nafenopin

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

−1,2,17

+17

−1

WY-14,643 DEHP Methyl clofenapate Ciprofibrate Bezafibrate Methylclofenapate Ciprofibrate WY-14,643 Nafenopin

+1,7,8 (+)4,27 +8 +8,22 +24 −8,21 +8,18 −15,22 +9,11,16 −8 +16,23 −10,12

+1,2,23

+1,2,23 +7,9 +7,9

+9,11 +9,11

+1,7,8,9 +4,27 +7,8,9 +8,22 −24 −8 −8,22 −8 −12,23

−5

−1,8 (+)4 (+)28 +25 −8 −8 −8,25 −8 −8 −12,14,17

−1 −46

−13

+32

457

(Continued)

TABLE 17.3.

Species Differences in Responses to PPARα Activators (Continued)

458

Response

PPARα Activation

Cynomolgus monkey

Humans30

?

≤1

Fenofibrate Perfluorodecanoic acid Bezafibrate DEHP DINP Clofibrate Fenofibrate Ciprofibrate See references for compound used

Hypolipidemic Effect (Decreases in Triglycerides or VLDL-Triglycerides)

Increases in Liver Weight

Oxidative Stress

NF-kB Activation

Increases in Acute Cell Proliferation

−19 −20,26

+26 −20

−24 −3

−24 −3

−3

−3 −3 −6 +6 −31

−3 −3 −6 −6 −41–44,45

− +6 +6 (+)33 +34 −35–39,45 3

−6 +40

−6 −6,33

Decreases in Apoptosis

Liver Tumors

−12,42,43,45

Comments: PPARα activation is a summary of trans-activation data as well as response of markers such as ACO and CYP4A gene, along with protein and enzymatic activity, which are indicators of PPARα activation and are dependent on level of PPARα expression. The endpoint examined in these studies is indicated below. + indicates a strong response, (+) indicates a weak response, and − indicates no response. Spaces left blank indicate no data available. It should be noted that the table does not include PCO data from monkey species other than Cynomolgus monkeys; other monkey data (which is almost universally negative) are summarized in Klaunig et al. (2003). PCO, palmitoyl-CoA oxidase. References: 1Lake et al. (1993) (ACO); 2Price et al. (1992) (ACO); 3Pugh et al. (2000) (peroxisomal fatty acid beta-oxidation); 4Isenberg et al. (2000); 5Tharappel et al. (2001); 6Hoivik et al. (2004) (lipofuscin, peroxisome number, PCO); 7Choudhury et al. (2004) (CYP4A increases); 8Lake et al. (2000) (peroxisome proliferation, CYP4A and carnitine acetyl transferase); 9Choudhury et al. (2000) (trans-activation assay); 10Macdonald et al. (1999) (trans-activation assay); 11Bell et al. (1998) (trans-activation assay); 12Hasmall et al. (1998) (nafenopin); 13Plant et al. (1998) (in vitro apoptosis assay); 14 Elcock et al. (1998) (in vitro); 15Caira et al. (1998) (multifunctional protein, ACO, thiolase); 16Tugwood et al. (1998) (trans-activation assay); 17James and Roberts (1996); 18Pacot et al. (1996) (ACO increases only 1.6-fold); 19Cornu-Chagnon et al. (1995) (ACO in vitro); 20Chinje et al. (1994) (CYP4A); 21Bell et al. (1993) (CYP4A13); 22Makowska et al. (1992) (ACO, CYP4A); 23Lake et al. (1989b) (ACO, CYP4A); 24Watanabe et al. (1989) (slight increases in ACO); 25Styles et al. (1988); 26Van Rafelghem et al. (1987) (peroxisome proliferation); 27Lake et al. (1987) (ACO in vivo and in vitro); 28 Styles et al. (1990); 30Compounds used to treat humans or human primary hepatocytes are indicated in the references; 31Gariot et al. (1987) (fenofibrate); 32James and Roberts (1996); 33Cariello et al. (2005) (fatty acid β-oxidation genes); 34Hanefeld et al. (1983) (clofibrate); 35Hanefeld et al. (1980) (clofibrate); 36De La Iglesia et al. (1982) (gemfibrozil); 37Blumcke et al. (1983) (fenofibrate); 38Gariot et al. (1983) (fenofibrate); 39Bentley et al. (1993) (review); 40Klaunig et al. (2003) (review); 41Perrone et al. (1998) (clofibric acid, diprofibrate); 42Goll et al. (1999) (clofibrate, ciprofibrate, bezafibrate, nafenopin, DEHP); 43Hasmall et al. (1999) (MEHP, MINP, primary metabolite of DINP); 44Hasmall et al. (2000) (MEHP); 45Shaw et al. (2002) (MINP); 46Schmezer et al. (1988).


459

chemicals. Syrian hamsters exhibited suppression of apoptosis after exposure to nafenopin, and guinea pigs exhibited suppression of apoptosis with nafenopin but no change with methylclofenapate. Cancer bioassays performed in Syrian hamsters with nafenopin, WY-14,643 and DEHP were negative (Lake et al. 1993; Schmezer et al. 1988). In summary, although Syrian hamsters and to a lesser extent guinea pigs exhibit changes in endpoints associated with PPARα activation (hypolipidemic effects and changes in fatty acid metabolizing enzymes), they do not exhibit consistent changes in the key events associated with the PPARα activator MOA for liver cancer in rats and mice. In vitro and in vivo data on Cynomolgus monkeys (Table 17.3) and other species of monkeys (i.e., marmoset and Rhesus) indicate that the key events in the PPARα activator MOA are relatively nonresponsive in monkeys. Palmitoyl-CoA oxidase activity was evaluated in monkeys after in vivo exposure to a variety of PPARα activators [e.g., bezafibrate, clofibrate, DEHP, mono-2-ethylhexyl phthalate (MEHP), fenofibrate, nafenopin, and LY171883], and changes were minimal or nonexistent relative to controls (Klaunig et al. 2003). Moreover, Cynomolgus monkeys exposed to DEHP, DINP, or clofibrate failed to exhibit an increase in cell proliferation (Doull et al. 1999; Pugh et al. 2000). Cynomolgus monkeys treated for two weeks with clinically relevant doses of the PPARα activators fenofibrate or ciprofibrate exhibited increases in the number of hepatic peroxisomes (Hoivik et al. 2004). In this study ciprofibrate but not fenofibrate increased liver to body weights in the absence of hepatocyte proliferation. In a follow-up to this study, transcript profiling was used to characterize the genes altered by ciprofibrate exposure (Cariello et al. 2005). Many genes involved in fatty acid metabolism and mitochondrial oxidative phosphorylation were up-regulated, reflecting the known hypolipidemic effects of exposure. However, the magnitude of induction in the β-oxidation pathway was substantially less in monkeys compared to mice and rats. Consistent with the lack of hepatocyte proliferation, there were a number of key regulatory genes that were down-regulated, including members of the JUN, MYC, and NF-kB families. In contrast, JUN and MYC gene expression were up-regulated after PPARα activator treatment in rats (Hsieh et al. 1991). No transcriptional signal for DNA damage or oxidative stress was observed. Lastly, marmosets exposed for 6.5 years to clofibrate at relatively high doses (94 mg/kg/day or higher) did not develop liver tumors over the duration of this study (Tugwood et al. 1996).* Taken together, the key events after PPARα activation in the rodent MOA for liver tumors were not observed in primates treated with PPARα activators. Humans are generally nonresponsive to the effects of PPARα activators. Liver weights were not increased in patients treated with fenofibrate (Gariot et al. 1987). Liver biopsies from humans treated with hypolipidemic drugs or primary human hepatocytes treated with PPARα activators were almost uniformly negative for peroxisome proliferation [reviewed in Bentley et al. (1993)]. In only one out of five studies was there a statistically significant increase in peroxisome number (∼50%), but there was no corresponding increase in volume of peroxisomes (Blumcke et al. 1983; De La Iglesia et al. 1982; Gariot et al. 1983; Hanefeld et al. 1980, 1983). *It should be noted that the duration of this study did not represent a lifetime exposure.

460


Exposure to PPARα activators alters different PPARα gene targets in rodents and humans, including the ACO gene. Unlike the large increases in the expression of marker mRNAs and proteins that are found in rodent primary hepatocytes treated with PPARα activators in vitro, very minor increases, if any, are observed in human primary hepatocytes (Bichet et al. 1990; Cornu-Chagnon et al. 1995; Duclos et al. 1997; Elcombe 1985; Elcombe et al. 1996; Goll et al. 1999; Hasmall et al. 2000; Hasmall et al. 1999; Perrone et al. 1998). ACO mRNA in liver samples from 48 patients treated with one of several fibrates (bezafibrate, fenofibrate or gemfibrozil) was not induced despite significant induction of hepatic apolipoprotein A-I mRNA and lowering of serum lipids following treatment (Roglans et al. 2002). The relatively weak increases in ACO observed in human primary hepatocytes are in stark contrast to the robust inductions observed in the livers of mice and rats exposed to PPARα activators [summarized in Klaunig et al. (2003)]. In summary, there is no evidence that the ACO gene exhibits more than minor inductions in humans. Species differences in sensitivity to PPARα activators may be explained in part by differences in the structure of the promoter regions that regulate the expression of target genes. The lack of ACO induction in human livers and primary human hepatocytes may be attributable to an inactive PPRE. Evidence that a functional PPRE exists in the human ACO gene promoter (Varanasi et al. 1996), was challenged by subsequent studies which showed that the PPRE is inactive in in vitro transactivation assays and that the sequence differs from that originally reported at three positions (Woodyatt et al. 1999). Little heterogeneity exists within the human ACO PPRE because the same altered PPRE sequence was found in all 22 unrelated humans that were investigated as well as in the human hepatocellular carcinoma cell line HepG2 (Woodyatt et al. 1999). A nonfunctional PPRE in the ACO promoter would be consistent with studies showing little, if any induction of the ACO gene/protein expression upon exposure to PPARα activators in human primary hepatocytes. PPARα ligands do not induce cell proliferation or suppress apoptosis in human hepatocytes in vitro (Goll et al. 1999; Hasmall et al. 1999; Perrone et al. 1998; Williams and Perrone 1996). Many of these studies included a positive control to ensure that human hepatocytes were of sufficient quality to mount a positive growth response. In comparison, rat or mouse primary hepatocytes exposed to PPARα activators exhibit up to 8-fold induction in cell proliferation (summarized in (Klaunig et al. 2003)). There are no data on human hepatocyte proliferation in vivo, although in vivo and in vitro data from nonhuman primates show cell proliferation is not induced by PPARα activators [Table 17.3 and reviewed in Doull et al. (1999)]. In summary, available data suggest that PPARα activators are unlikely to alter apoptosis and proliferation in human hepatocytes. 17.3.2.10. Molecular Basis of Species Differences. In the following section, the properties of PPARα and associated responses in the livers of rodents and primates are compared with an emphasis on human data, if available. The weight of evidence demonstrates that humans respond to PPARα activators differently than rodents in that many of the typical markers of PPARα activator exposure associated with hepatocarcinogenesis in rodents are absent in humans. Differences in the


TABLE 17.4.

461

Properties of Rodent (Rat and Mouse) PPARα Versus Human PPARα in Liver

Property Allelic variants

Rodent (Rat/Mouse) None identified

Human L162V

V227A “6/29”

Truncated PPARα (deleted exon 6) Inducibility by environmentally relevant ligands

Below 10% of total PPARα Chemical-specific range of responsiveness

Basal expression of PPARα Regulation of hypolipidemic response

High in liver Intact

10–50% of total PPARα Some differences with rodent activation noted leading to decreased activation 20-fold less than mouse liver). A 3-fold variation in the expression of the full-length PPARα mRNA between human samples was noted. The data indicate that hPPARα in liver is expressed at levels far below that expressed in rodent liver. Additional studies evaluating expression and function of PPARα in human liver are needed to more definitively determine the relative expression of PPARα in rodents and humans. Such studies would benefit from better assessment of the degree of protein and mRNA degradation in the samples. Truncated PPARα. A truncated PPARα variant has been identified in a number of labs and is called hPPARα-8/14 (Tugwood et al. 1996), hPPARSV (Palmer et al. 1998), PPARαtr (Gervois et al. 1999), and PPARα2 (Hanselman et al. 2001). This truncated form lacks exon 6 due to alternative splicing, resulting in a hPPARα lacking the hinge region and ligand binding domain. This form acts as a dominant negative, inhibiting the ability of the wild-type receptor to activate transcription, possibly by titrating out limiting amounts of co-activators (Gervois et al. 1999). The level of the mRNA of this form ranges from 10–50% of full-length hPPARα mRNA (Gervois et al. 1999; Hanselman et al. 2001; Palmer et al. 1998; Roberts et al. 2000) similar to Cynomolgus monkeys (Hanselman et al. 2001). In comparison, this level is below 10% in mice and rats (Hanselman et al. 2001). A more definitive role for this truncated form awaits studies in which the levels of full-length and truncated hPPARα forms are simultaneously measured with well-characterized hPPARα target genes in primary human hepatocytes exposed to PPARα activators. Differences in Transcriptional Networks Controlled by Human and Rodent PPARα. There is overwhelming evidence that the transcriptional networks controlled by PPARα are different between humans and rodents and underlie species-specific differences in key events in the PPARα MOA. Humans and rodents do share hypolipidemic effects of PPARα activators but may achieve this beneficial effect through regulation of different gene sets. A number of genes are likely responsible for the therapeutic hypolipidemic effects of PPARα activators in humans. Many of these genes have functional PPREs that are transcriptionally regulated by human PPARα, including apolipoprotein (apo) C-III (Hertz et al. 1995), lipoprotein lipase (Schoonjans et al. 1996), apo A-I (Vu-Dac et al. 1994), apo A-II (Vu-Dac et al. 1995), and carnitine palmitoyl transferase-I (Mascaro et al. 1998). Human PPARα activation of apolipoprotein A-II and lipoprotein lipase transcription and suppression of apolipoprotein C-III expression are key to lowering serum triglycerides (Auwerx et al. 1996; Staels et al. 1997; Vu-Dac et al. 1995). Human apolipoprotein C-III can be down-regulated by fibrates in cultured human hepatocytes in the absence of changes in PPARα target genes encoding peroxisomal enzymes

464


including ACO, bifunctional enzyme, and thiolase (Lawrence et al. 2001b). Furthermore, stably transfected HepG2 cells expressing either human or murine PPARα at levels similar to rodent liver respond to fibrates by increased expression of 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) synthase and carnitine palmitoyl transferase-I (CPT-I) but lack the typical robust induction of typical PPARα targets—that is, ACO, bifunctional enzyme, or thiolase (Hsu et al. 2001; Lawrence et al. 2001a; Tachibana et al. 2005). In a global analysis of gene expression, genes of the cytosolic, microsomal, and mitochondrial pathways involved in fatty acid transport and metabolism were up-regulated by clofibrate in both rodent and human hepatocyte cultures, whereas genes of the peroxisomal pathway of lipid metabolism were up-regulated only in rodents (Richert et al. 2003). Thus, PPARα activation may lower lipid levels in humans and rodents through regulation of different sets of genes. The human PPARα does not possess all of the functions of the rodent PPARα including the ability to regulate cell proliferation. Two mouse strains have been created which express the hPPARα in the absence of mPPARα (“humanized” hPPARα mice). In the TRE-hPPARα mouse, PPARα is under the control of a liverspecific promoter and is preferentially expressed in hepatocytes (Cheung et al. 2004); the cellular location of hPPARα expression in the humanized PPARα mouse corresponds to the location of mPPARα expression in wild-type mice—that is, in hepatocytes but not Kupffer cells (Peters et al. 2000). The hPPARαPAC mouse contains a 211-kilobase region encoding the regulatory and structural regions of the human PPARα gene. The hPPARα is expressed in the same tissues as those of the mouse PPARα (Yang et al. 2008). The humanized PPARα mouse strains do not respond to a PPARα activator (WY-14,643) in the same manner as wild-type mice even though both strains express hPPARα to levels comparable to mPPARα in wild-type mice. The humanized mice exhibit increases in peroxisome proliferation, decreases in serum total triglycerides and normal activation of lipid metabolism genes including those involved in peroxisome proliferation. However, these mice do not exhibit increased expression of cell cycle genes or increased hepatocyte proliferation in response to a PPARα activator as do wild-type mice (Cheung et al. 2004; Morimura et al. 2006; Yang et al. 2008). In a 38- to 44-week exposure study with the PPARα activator WY-14,643, the TRE-hPPARα mice were also resistant to PPARα activator-induced liver cancer. Wild-type mice but not humanized mice exhibited a significant increase in liver tumors despite the fact that the humanized mice were exposed 6 weeks longer than the wild-type mice to the compound (Morimura et al. 2006). These studies show that hPPARα is pharmacologically active but does not regulate the full spectrum of responses necessary for hepatocarcinogenesis in rodents. The molecular basis for differences between mouse and human PPARα may be differences in the ability of the receptors to interact with transcriptional coactivators or to regulate miRNA cascades. Co-activators convey the transcriptional activation of the ligand-induced nuclear receptor to the transcriptional machinery. Elegant biochemical and crystallographic analyses have shown key interactions between co-activators and the ligand binding domains of nuclear receptors including PPAR family members (Li et al. 2008; Xu and Li 2008). The mouse and rat PPARα


465

ligand binding domains (LBD) do possess amino acid differences with human PPARα LBD (Mukherjee et al. 1994; Sher et al. 1993; Tugwood et al. 1996). Amino acid differences in the LBD between mice and humans may uncouple receptor coactivator interactions in humans required for cell proliferation and gene regulation, while retaining those important in lipid metabolism gene regulation. Alternatively, differences in miRNA regulation may contribute to species differences, as the ability to regulate the let-7c cascade is lost in humanized mice in response to a PPARα activator (Yang et al. 2008). Further studies are needed to define the specific mechanistic basis for species differences. 17.3.2.11. Summary of Key Data that Support the MOA. The PPARα MOA describes the sequence of events beginning with PPARα activation and leading to an increased incidence of liver tumors in rats and mice. This MOA exists independent of exposure to any particular chemical but has been shown to be triggered by chemicals collectively referred to as PPARα activators. The overall WOE supports a MOA that involves five key events. First, PPARα activators activate PPARα. Second, PPARα activation leads to alterations in the expression of genes that regulate oxidative stress and increases in oxidative stress. Third, oxidative stress activates the transcription factor NF-kB. Fourth, NF-kB activation leads to increased cell proliferation and decreased apoptosis in the liver. Fifth, sustained growth signaling upon chronic exposure causes clonal expansion of initiated cells leading to preneoplastic foci and tumors—that is, hepatocellular adenomas and carcinomas. Table 17.5 summarizes the specificity and WOE of the PPARα activator MOA. The WOE strongly supports the MOA due to the large number of studies that have been carried out since the discovery of peroxisome proliferation by these chemicals in 1965 (Hess et al. 1965). PPARα activation is by definition specific, because this key event is distinct from other initiating events such as CAR activation or increases in cytotoxicity. The other key events by themselves are considered to

TABLE 17.5.

PPARα Activators: Mode of Action (MOA) Key Events

Causal Key Eventa 1. 2. 3. 4. 5.

Activation of PPARα Increases in oxidative stress NF-kB activation Perturbation of cell growth and survival Clonal expansion of preneoplastic foci

Specificityb

Evidencec

High Low Low Low Low

Strong Strong Strong Strong Strong

Causal key event is a required step for PPARα MOA, based on empirical evidence.

a

Specificity of each key event to PPARα-induced rodent hepatic tumors is considered high if it is unique to this MOA and low if not. The key events other than PPARα activation by themselves are considered to have low specificity, because these events are observed with other carcinogens. However, the key events when linked are considered to have high specificity because they are dependent on PPARα. b

c

Evidence was determined to be strong if several studies support that key event as part of the MOA, preferably with multiple PPARα activators from multiple laboratories, with limited evidence of contradiction. Evidence is considered weak if only a single study with a single PPARα activator from a single laboratory supports that key event or if a significant amount of contradiction appears in the literature.

466


TABLE 17.6.

Comparative Analysis of Rodent and Human Data—Liver Tumors

Causal Key Events

Plausible in Humans?

Taking into Account Kinetic and Dynamic Factors, Is the Key Event Plausible in Humans?

1. Activation of PPARα 2. Increases in oxidative stress

Yes

Yes

Yes

Unknown

3. NF-kB activation

Yes

Unknown

4. Perturbation of cell growth and survival

Yes

Not likely

5. Selective clonal expansion of preneoplastic foci 6. Liver tumors

Yes

Not likely

Yes

Not likely

Comments PPARα is a target of human hypolipidemic drugs. Gene products that produce oxidative stress in rodents exist in humans but are not induced to the same extent in humans or monkeys. More traditional methods of measuring oxidative stress have not been used. NF-kB exists in humans but has not been measured in human liver or primary hepatocytes after exposure to PPARα activators. Not seen in independent studies of human hepatocytes in vitro; not measured in vivo; not seen in nonhuman primates in vivo or in vitro; not seen in hamsters or guinea pigs. No response in nonhuman primates. Not measured in livers of humans exposed to PPARα activators; no tumors in hamsters with expression of PPARα intermediate between mice/rats and humans.

have low specificity, because these events are observed with other carcinogens. However, the key events when linked are considered to have high specificity because they are dependent on PPARα. Table 17.1 provides examples of chemical-specific data evaluating whether the key events occur after exposure to five different PPARα activators.

REFERENCES

467

Evidence showing the mechanistic linkages between the key events of the MOA is summarized in Table 17.2. Studies that inhibit key events by genetic or biochemical means reveal such relationships because inhibition of one event blocks downstream events. Additional support for the PPARα MOA comes from a comparison of responses in rats and mice to “nonresponsive” species such as Syrian hamsters, guinea pigs, and monkeys. These data are summarized in Table 17.3. Overall, these data show that while all species exhibit a hypolipidemic response and alterations in lipid metabolism and transport genes, Syrian hamsters, guinea pigs and monkeys exhibit little, if any, changes in oxidative stress markers, NF-kB activation, and alterations of hepatocyte growth or tumor response (Klaunig et al. 2003).

17.4. RELEVANCE OF PPARα ACTIVATOR-INDUCED RODENT LIVER TUMOR RESPONSE TO HUMANS Although humans have been regularly exposed to PPARα activators through administration of hypolipidemic pharmaceuticals, epidemiological studies have not provided evidence of increased incidence of liver neoplasms in humans exposed to PPARα activators for up to 13 years [summarized in Klaunig et al. (2003)]. Species comparisons of key events and other endpoints relevant to the PPARα MOA show that mice and rats are much more responsive than humans (Table 17.6) and other species (e.g., hamsters, guinea pigs, and primates) (Table 17.3). Experimental evidence suggests that the differences in responsiveness among species may be due to differences in promoter structure and/or function of PPARα target genes, sensitivity of PPARα to activation, the expression level of full-length and dominant negative forms of PPARα, and species differences in the ability of PPARα to alter expression of genes involved in cell fate (Table 17.4). Overall, the weight of evidence suggests that although the rodent MOA is plausible in humans, humans would not be expected to respond with a hepatocarcinogenic effect from chronic exposure consistent with the original conclusion by an ILSI workgroup (Klaunig et al. 2003).

REFERENCES Abdellatif, A. G., Preat, V., Vamecq, J., Nilsson, R., and Roberfroid, M. (1990). Peroxisome proliferation and modulation of rat liver carcinogenesis by 2,4-dichlorophenoxyacetic acid, 2,4,5-trichlorophenoxyacetic acid, perfluorooctanoic acid and nafenopin. Carcinogenesis 11, 1899–1902. Anderson, S. P., Dunn, C., Laughter, A., Yoon, L., Swanson, C., Stulnig, T. M., Steffensen, K. R., Chandraratna, R. A., Gustafsson, J. A., and Corton, J. C. (2004a). Overlapping transcriptional programs regulated by the nuclear receptors peroxisome proliferator-activated receptor alpha, retinoid X receptor, and liver X receptor in mouse liver. Mol Pharmacol 66, 1440–1452. Anderson, S. P., Dunn, C. S., Cattley, R. C., and Corton, J. C. (2001). Hepatocellular proliferation in response to a peroxisome proliferator does not require TNFalpha signaling. Carcinogenesis 22, 1843–1851.

468

CHAPTER 17 MODE OF ACTION ANALYSIS AND HUMAN RELEVANCE

Anderson, S. P., Howroyd, P., Liu, J., Qian, X., Bahnemann, R., Swanson, C., Kwak, M. K., Kensler, T. W., and Corton, J. C. (2004b). The transcriptional response to a peroxisome proliferator-activated receptor alpha agonist includes increased expression of proteome maintenance genes. J Biol Chem 279, 52390–52398. Arsura, M., and Cavin, L. G. (2005). Nuclear factor-kappaB and liver carcinogenesis. Cancer Lett 229, 157–169. Ashby, J., Brady, A., Elcombe, C. R., Elliott, B. M., Ishmael, J., Odum, J., Tugwood, J. D., Kettle, S., and Purchase, I. F. (1994). Mechanistically-based human hazard assessment of peroxisome proliferator-induced hepatocarcinogenesis. Hum Exp Toxicol 13 (Suppl 2), S1–S117. Auwerx, J., Schoonjans, K., Fruchart, J. C., and Staels, B. (1996). Transcriptional control of triglyceride metabolism: fibrates and fatty acids change the expression of the LPL and apo C-III genes by activating the nuclear receptor PPAR. Atherosclerosis 124 (Suppl), S29–S37. Barrass, N. C., Price, R. J., Lake, B. G., and Orton, T. C. (1993). Comparison of the acute and chronic mitogenic effects of the peroxisome proliferators methylclofenapate and clofibric acid in rat liver. Carcinogenesis 14, 1451–1456. Bayly, A. C., Roberts, R. A., and Dive, C. (1994). Suppression of liver cell apoptosis in vitro by the non-genotoxic hepatocarcinogen and peroxisome proliferator nafenopin. J Cell Biol 125, 197–203. Becuwe, P., and Dauca, M. (2005). Comparison of cytotoxicity induced by hypolipidemic drugs via reactive oxygen species in human and rodent liver cells. Int J Mol Med 16, 483–492. Bell, A. R., Savory, R., Horley, N. J., Choudhury, A. I., Dickins, M., Gray, T. J., Salter, A. M., and Bell, D. R. (1998). Molecular basis of non-responsiveness to peroxisome proliferators: the guinea-pig PPARalpha is functional and mediates peroxisome proliferator-induced hypolipidaemia. Biochem J 332 (Pt 3), 689–693. Bell, D. R., Plant, N. J., Rider, C. G., Na, L., Brown, S., Ateitalla, I., Acharya, S. K., Davies, M. H., Elias, E., Jenkins, N. A., and et al. (1993). Species-specific induction of cytochrome P-450 4A RNAs: PCR cloning of partial guinea-pig, human and mouse CYP4A cDNAs. Biochem J 294 (Pt 1), 173–180. Bentley, P., Calder, I., Elcombe, C., Grasso, P., Stringer, D., and Wiegand, H. J. (1993). Hepatic peroxisome proliferation in rodents and its significance for humans. Food Chem Toxicol 31, 857–907. Bichet, N., Cahard, D., Fabre, G., Remandet, B., Gouy, D., and Cano, J. P. (1990). Toxicological studies on a benzofuran derivative. III. Comparison of peroxisome proliferation in rat and human hepatocytes in primary culture. Toxicol Appl Pharmacol 106, 509–517. Bility, M. T., Thompson, J. T., McKee, R. H., David, R. M., Butala, J. H., Vanden Heuvel, J. P., and Peters, J. M. (2004). Activation of mouse and human peroxisome proliferator-activated receptors (PPARs) by phthalate monoesters. Toxicol Sci 82, 170–182. Blumcke, S., Schwartzkopff, W., Lobeck, H., Edmondson, N. A., Prentice, D. E., and Blane, G. F. (1983). Influence of fenofibrate on cellular and subcellular liver structure in hyperlipidemic patients. Atherosclerosis 46, 105–116. Bojes, H. K., Germolec, D. R., Simeonova, P., Bruccoleri, A., Schoonhoven, R., Luster, M. I., and Thurman, R. G. (1997). Antibodies to tumor necrosis factor alpha prevent increases in cell replication in liver due to the potent peroxisome proliferator, WY-14,643. Carcinogenesis 18, 669–674. Boobis, A. R., Cohen, S. M., Dellarco, V., McGregor, D., Meek, M. E., Vickers, C., Willcocks, D., and Farland, W. (2006). IPCS framework for analyzing the relevance of a cancer mode of action for humans. Crit Rev Toxicol 36, 781–792. Boobis, A. R., Doe, J. E., Heinrich-Hirsch, B., Meek, M. E., Munn, S., Ruchirawat, M., Schlatter, J., Seed, J., and Vickers, C. (2008). IPCS framework for analyzing the relevance of a noncancer mode of action for humans. Crit Rev Toxicol 38, 87–96. Burkhardt, S., Mellert, W., Reinachet, M., and Bahnemann, R. (2001). Zonal evaluation of proliferation and apoptosis in the liver of mice reveals new mechanistic data for phenobarbital (PB), Wyeth 14,643 (WY), and chloroform (CH). The Toxicologist 60, 286 (Abstract #1362). Burns, K. A., and Vanden Heuvel, J. P. (2007). Modulation of PPAR activity via phosphorylation. Biochim Biophys Acta 1771, 952–960. Bursch, W., Lauer, B., Timmermann-Trosiener, I., Barthel, G., Schuppler, J., and Schulte-Hermann, R. (1984). Controlled death (apoptosis) of normal and putative preneoplastic cells in rat liver following withdrawal of tumor promoters. Carcinogenesis 5, 453–458.

REFERENCES

469

Busser, M. T., and Lutz, W. K. (1987). Stimulation of DNA synthesis in rat and mouse liver by various tumor promoters. Carcinogenesis 8, 1433–1437. Cai, Y., Appelkvist, E. L., and DePierre, J. W. (1995). Hepatic oxidative stress and related defenses during treatment of mice with acetylsalicylic acid and other peroxisome proliferators. J Biochem Toxicol 10, 87–94. Caira, F., Clemencet, M. C., Cherkaoui-Malki, M., Dieuaide-Noubhani, M., Pacot, C., Van Veldhoven, P. P., and Latruffe, N. (1998). Differential regulation by a peroxisome proliferator of the different multifunctional proteins in guinea pig: cDNA cloning of the guinea pig D-specific multifunctional protein 2. Biochem J 330 (Pt 3), 1361–1368. Calfee-Mason, K. G., Spear, B. T., and Glauert, H. P. (2004). Effects of vitamin E on the NF-kappaB pathway in rats treated with the peroxisome proliferator, ciprofibrate. Toxicol Appl Pharmacol 199, 1–9. Cariello, N. F., Romach, E. H., Colton, H. M., Ni, H., Yoon, L., Falls, J. G., Casey, W., Creech, D., Anderson, S. P., Benavides, G. R., Hoivik, D. J., Brown, R., and Miller, R. T. (2005). Gene expression profiling of the PPAR-alpha agonist ciprofibrate in the cynomolgus monkey liver. Toxicol Sci 88, 250–264. Cattley, R. C., Conway, J. G., and Popp, J. A. (1987). Association of persistent peroxisome proliferation and oxidative injury with hepatocarcinogenicity in female F-344 rats fed di(2-ethylhexyl)phthalate for 2 years. Cancer Lett 38, 15–22. Cattley, R. C., DeLuca, J., Elcombe, C., Fenner-Crisp, P., Lake, B. G., Marsman, D. S., Pastoor, T. A., Popp, J. A., Robinson, D. E., Schwetz, B., Tugwood, J., and Wahli, W. (1998). Do peroxisome proliferating compounds pose a hepatocarcinogenic hazard to humans? Regul Toxicol Pharmacol 27, 47–60. Cattley, R. C., and Glover, S. E. (1993). Elevated 8-hydroxydeoxyguanosine in hepatic DNA of rats following exposure to peroxisome proliferators: relationship to carcinogenesis and nuclear localization. Carcinogenesis 14, 2495–2499. Chan, E., Tan, C. S., Deurenberg-Yap, M., Chia, K. S., Chew, S. K., and Tai, E. S. (2006). The V227A polymorphism at the PPARA locus is associated with serum lipid concentrations and modulates the association between dietary polyunsaturated fatty acid intake and serum high density lipoprotein concentrations in Chinese women. Atherosclerosis 187, 309–315. Chang, C. K., Llanes, S., and Schumer, W. (1997). Effect of dexamethasone on NF-kB activation, tumor necrosis factor formation, and glucose dyshomeostasis in septic rats. J Surg Res 72, 141–145. Chen, H., Huang, C. Y., Wilson, M. W., Lay, L. T., Robertson, L. W., Chow, C. K., and Glauert, H. P. (1994). Effect of the peroxisome proliferators ciprofibrate and perfluorodecanoic acid on hepatic cell proliferation and toxicity in Sprague-Dawley rats. Carcinogenesis 15, 2847–2850. Cheung, C., Akiyama, T. E., Ward, J. M., Nicol, C. J., Feigenbaum, L., Vinson, C., and Gonzalez, F. J. (2004). Diminished hepatocellular proliferation in mice humanized for the nuclear receptor peroxisome proliferator-activated receptor alpha. Cancer Res 64, 3849–3854. Chinje, E., Kentish, P., Jarnot, B., George, M., and Gibson, G. (1994). Induction of the CYP4A subfamily by perfluorodecanoic acid: the rat and the guinea pig as susceptible and non-susceptible species. Toxicol Lett 71, 69–75. Choudhury, A. I., Chahal, S., Bell, A. R., Tomlinson, S. R., Roberts, R. A., Salter, A. M., and Bell, D. R. (2000). Species differences in peroxisome proliferation; mechanisms and relevance. Mutat Res 448, 201–212. Choudhury, A. I., Sims, H. M., Horley, N. J., Roberts, R. A., Tomlinson, S. R., Salter, A. M., Bruce, M., Shaw, P. N., Kendall, D., Barrett, D. A., and Bell, D. R. (2004). Molecular analysis of peroxisome proliferation in the hamster. Toxicol Appl Pharmacol 197, 9–18. Conway, J. G., Tomaszewski, K. E., Olson, M. J., Cattley, R. C., Marsman, D. S., and Popp, J. A. (1989). Relationship of oxidative damage to the hepatocarcinogenicity of the peroxisome proliferators di(2-ethylhexyl)phthalate and Wy-14,643. Carcinogenesis 10, 513–519. Cornu-Chagnon, M. C., Dupont, H., and Edgar, A. (1995). Fenofibrate: metabolism and species differences for peroxisome proliferation in cultured hepatocytes. Fundam Appl Toxicol 26, 63–74. Corton, J. C. (2008). Evaluation of the role of peroxisome proliferator-activated receptor alpha (PPARalpha) in mouse liver tumor induction by trichloroethylene and metabolites. Crit Rev Toxicol 38, 857–875.

470


Corton, J. C., Anderson, S. P., and Stauber, A. (2000). Central role of peroxisome proliferator-activated receptors in the actions of peroxisome proliferators. Annu Rev Pharmacol Toxicol 40, 491–518. Corton, J. C., Apte, U., Anderson, S. P., Limaye, P., Yoon, L., Latendresse, J., Dunn, C., Everitt, J. I., Voss, K. A., Swanson, C., Kimbrough, C., Wong, J. S., Gill, S. S., Chandraratna, R. A., Kwak, M. K., Kensler, T. W., Stulnig, T. M., Steffensen, K. R., Gustafsson, J. A., and Mehendale, H. M. (2004). Mimetics of caloric restriction include agonists of lipid-activated nuclear receptors. J Biol Chem 279, 46204–46212. Corton, J. C., and Lapinskas, P. J. (2005). Peroxisome proliferator-activated receptors: mediators of phthalate ester-induced effects in the male reproductive tract? Toxicol Sci 83, 4–17. Currie, R. A., Bombail, V., Oliver, J. D., Moore, D. J., Lim, F. L., Gwilliam, V., Kimber, I., Chipman, K., Moggs, J. G., and Orphanides, G. (2005). Gene ontology mapping as an unbiased method for identifying molecular pathways and processes affected by toxicant exposure: Application to acute effects caused by the rodent non-genotoxic carcinogen diethylhexylphthalate. Toxicol Sci 86, 453–469. Czaja, M. J. (2007). Cell signaling in oxidative stress-induced liver injury. Semin Liver Dis 27, 378–389. De Bosscher, K., Vanden Berghe, W., and Haegeman, G. (2006). Cross-talk between nuclear receptors and nuclear factor kappaB. Oncogene 25, 6868–6886. de Duve, C. (1996). The peroxisome in retrospect. Ann N Y Acad Sci 804, 1–10. De La Iglesia, F. A., Lewis, J. E., Buchanan, R. A., Marcus, E. L., and McMahon, G. (1982). Light and electron microscopy of liver in hyperlipoproteinemic patients under long-term gemfibrozil treatment. Atherosclerosis 43, 19–37. De Minicis, S., Bataller, R., and Brenner, D. A. (2006). NADPH oxidase in the liver: defensive, offensive, or fibrogenic? Gastroenterology 131, 272–275. Desvergne, B., A, I. J., Devchand, P. R., and Wahli, W. (1998). The peroxisome proliferatoractivated receptors at the cross-road of diet and hormonal signalling. J Steroid Biochem Mol Biol 65, 65–74. Desvergne, B., and Wahli, W. (1999). Peroxisome proliferator-activated receptors: nuclear control of metabolism. Endocr Rev 20, 649–688. Dostalek, M., Hardy, K. D., Milne, G. L., Morrow, J. D., Chen, C., Gonzalez, F. J., Gu, J., Ding, X., Johnson, D. A., Johnson, J. A., Martin, M. V., and Guengerich, F. P. (2008). Development of oxidative stress by cytochrome P450 induction in rodents is selective for barbiturates and related to loss of pyridine nucleotide-dependent protective systems. J Biol Chem 283, 17147–17157. Doull, J., Cattley, R., Elcombe, C., Lake, B. G., Swenberg, J., Wilkinson, C., Williams, G., and van Gemert, M. (1999). A cancer risk assessment of di(2-ethylhexyl)phthalate: Application of the new U.S. EPA Risk Assessment Guidelines. Regul Toxicol Pharmacol 29, 327–357. Dreyer, C., Krey, G., Keller, H., Givel, F., Helftenbein, G., and Wahli, W. (1992). Control of the peroxisomal beta-oxidation pathway by a novel family of nuclear hormone receptors. Cell 68, 879–887. Duclos, S., Bride, J., Ramirez, L. C., and Bournot, P. (1997). Peroxisome proliferation and beta-oxidation in Fao and MH1C1 rat hepatoma cells, HepG2 human hepatoblastoma cells and cultured human hepatocytes: effect of ciprofibrate. Eur J Cell Biol 72, 314–323. Dwivedi, R. S., Alvares, K., Nemali, M. R., Subbarao, V., Reddy, M. K., Usman, M. I., Rademaker, A. W., Reddy, J. K., and Rao, M. S. (1989). Comparison of the peroxisome proliferator-induced pleiotropic response in the liver of nine strains of mice. Toxicol Pathol 17, 16–26. Elcock, F. J., Chipman, J. K., and Roberts, R. A. (1998). The rodent nongenotoxic hepatocarcinogen and peroxisome proliferator nafenopin inhibits intercellular communication in rat but not guinea-pig hepatocytes, perturbing S-phase but not apoptosis. Arch Toxicol 72, 439–444. Elcombe, C. R. (1985). Species differences in carcinogenicity and peroxisome proliferation due to trichloroethylene: a biochemical human hazard assessment. Arch Toxicol Suppl 8, 6–17. Elcombe, C. R., Bell, D. R., Elias, E., Hasmall, S. C., and Plant, N. J. (1996). Peroxisome proliferators: species differences in response of primary hepatocyte cultures. Ann N Y Acad Sci 804, 628–635. Elliott, B. M., and Elcombe, C. R. (1987). Lack of DNA damage or lipid peroxidation measured in vivo in the rat liver following treatment with peroxisomal proliferators. Carcinogenesis 8, 1213–1218.

REFERENCES

471

EPA (2005a). Guidelines for carcinogen risk assessment. EPA/630/P-03/001F, 1–166, http://oaspub.epa. gov/eims/eimscomm.getfile?p_download_id=439797. EPA (2005b). Supplemental guidance for assessing susceptibility from early-life exposure to carcinogens. EPA/630/R-03/003F, 1–126, http://oaspub.epa.gov/eims/eimscomm.getfile?p_download_id=459042. Fan, C. Y., Pan, J., Usuda, N., Yeldandi, A. V., Rao, M. S., and Reddy, J. K. (1998). Steatohepatitis, spontaneous peroxisome proliferation and liver tumors in mice lacking peroxisomal fatty acyl-CoA oxidase. Implications for peroxisome proliferator-activated receptor alpha natural ligand metabolism. J Biol Chem 273, 15639–15645. Fischer, J. G., Glauert, H. P., Yin, T., Sweeney-Reeves, M. L., Larmonier, N., and Black, M. C. (2002). Moderate iron overload enhances lipid peroxidation in livers of rats, but does not affect NF-kappaB activation induced by the peroxisome proliferator, Wy-14,643. J Nutr 132, 2525–2531. Flavell, D. M., Pineda Torra, I., Jamshidi, Y., Evans, D., Diamond, J. R., Elkeles, R. S., Bujac, S. R., Miller, G., Talmud, P. J., Staels, B., and Humphries, S. E. (2000). Variation in the PPARalpha gene is associated with altered function in vitro and plasma lipid concentrations in Type II diabetic subjects. Diabetologia 43, 673–680. Gariot, P., Barrat, E., Drouin, P., Genton, P., Pointel, J. P., Foliguet, B., Kolopp, M., and Debry, G. (1987). Morphometric study of human hepatic cell modifications induced by fenofibrate. Metabolism 36, 203–210. Gariot, P., Barrat, E., Mejean, L., Pointel, J. P., Drouin, P., and Debry, G. (1983). Fenofibrate and human liver. Lack of proliferation of peroxisomes. Arch Toxicol 53, 151–163. Gervois, P., Torra, I. P., Chinetti, G., Grotzinger, T., Dubois, G., Fruchart, J. C., Fruchart-Najib, J., Leitersdorf, E., and Staels, B. (1999). A truncated human peroxisome proliferator-activated receptor alpha splice variant with dominant negative activity. Mol Endocrinol 13, 1535–1549. Gill, J. H., James, N. H., Roberts, R. A., and Dive, C. (1998). The non-genotoxic hepatocarcinogen nafenopin suppresses rodent hepatocyte apoptosis induced by TGFbeta1, DNA damage and Fas. Carcinogenesis 19, 299–304. Glauert, H. P., Beaty, M. M., Clark, T. D., Greenwell, W. S., Tatum, V., Chen, L. C., Borges, T., Clark, T. L., Srinivasan, S. R., and Chow, C. K. (1990). Effect of dietary vitamin E on the development of altered hepatic foci and hepatic tumors induced by the peroxisome proliferator ciprofibrate. J Cancer Res Clin Oncol 116, 351–356. Glauert, H. P., Eyigor, A., Tharappel, J. C., Cooper, S., Lee, E. Y., and Spear, B. T. (2006). Inhibition of hepatocarcinogenesis by the deletion of the p50 subunit of NF-kappaB in mice administered the peroxisome proliferator Wy-14,643. Toxicol Sci 90, 331–336. Gloire, G., Legrand-Poels, S., and Piette, J. (2006). NF-kappaB activation by reactive oxygen species: fifteen years later. Biochem Pharmacol 72, 1493–1505. Goel, S. K., Lalwani, N. D., and Reddy, J. K. (1986). Peroxisome proliferation and lipid peroxidation in rat liver. Cancer Res 46, 1324–1330. Goll, V., Alexandre, E., Viollon-Abadie, C., Nicod, L., Jaeck, D., and Richert, L. (1999). Comparison of the effects of various peroxisome proliferators on peroxisomal enzyme activities, DNA synthesis, and apoptosis in rat and human hepatocyte cultures. Toxicol Appl Pharmacol 160, 21–32. Gonzalez, F. J., and Shah, Y. M. (2008). PPARalpha: mechanism of species differences and hepatocarcinogenesis of peroxisome proliferators. Toxicology 246, 2–8. Gottlicher, M., Widmark, E., Li, Q., and Gustafsson, J. A. (1992). Fatty acids activate a chimera of the clofibric acid-activated receptor and the glucocorticoid receptor. Proc Natl Acad Sci USA 89, 4653–4657. Grasl-Kraupp, B., Huber, W., Timmermann-Trosiener, I., and Schulte-Hermann, R. (1993). Peroxisomal enzyme induction uncoupled from enhanced DNA synthesis in putative preneoplastic liver foci of rats treated with a single dose of the peroxisome proliferator nafenopin. Carcinogenesis 14, 2435–2437. Grasl-Kraupp, B., Ruttkay-Nedecky, B., Mullauer, L., Taper, H., Huber, W., Bursch, W., and SchulteHermann, R. (1997). Inherent increase of apoptosis in liver tumors: implications for carcinogenesis and tumor regression. Hepatology 25, 906–912. Hanefeld, M., Kemmer, C., and Kadner, E. (1983). Relationship between morphological changes and lipid-lowering action of p-chlorphenoxyisobutyric acid (CPIB) on hepatic mitochondria and peroxisomes in man. Atherosclerosis 46, 239–246.

472


Hanefeld, M., Kemmer, C., Leonhardt, W., Kunze, K. D., Jaross, W., and Haller, H. (1980). Effects of p-chlorophenoxyisobutyric acid (CPIB) on the human liver. Atherosclerosis 36, 159–172. Hanselman, J. C., Vartanian, M. A., Koester, B. P., Gray, S. A., Essenburg, A. D., Rea, T. J., Bisgaier, C. L., and Pape, M. E. (2001). Expression of the mRNA encoding truncated PPAR alpha does not correlate with hepatic insensitivity to peroxisome proliferators. Mol Cell Biochem 217, 91–97. Hashimoto, T. (1996). Peroxisomal beta-oxidation: enzymology and molecular biology. Ann N Y Acad Sci 804, 86–98. Hashimoto, T., Fujita, T., Usuda, N., Cook, W., Qi, C., Peters, J. M., Gonzalez, F. J., Yeldandi, A. V., Rao, M. S., and Reddy, J. K. (1999). Peroxisomal and mitochondrial fatty acid beta-oxidation in mice nullizygous for both peroxisome proliferator-activated receptor alpha and peroxisomal fatty acyl-CoA oxidase. Genotype correlation with fatty liver phenotype. J Biol Chem 274, 19228–19236. Hasmall, S. C., James, N. H., Macdonald, N., Gonzalez, F. J., Peters, J. M., and Roberts, R. A. (2000a). Suppression of mouse hepatocyte apoptosis by peroxisome proliferators: role of PPARalpha and TNFalpha. Mutat Res 448, 193–200. Hasmall, S. C., James, N. H., Macdonald, N., Soames, A. R., and Roberts, R. A. (2000b). Species differences in response to diethylhexylphthalate: suppression of apoptosis, induction of DNA synthesis and peroxisome proliferator activated receptor alpha-mediated gene expression. Arch Toxicol 74, 85–91. Hasmall, S. C., James, N. H., Macdonald, N., West, D., Chevalier, S., Cosulich, S. C., and Roberts, R. A. (1999). Suppression of apoptosis and induction of DNA synthesis in vitro by the phthalate plasticizers monoethylhexylphthalate (MEHP) and diisononylphthalate (DINP): a comparison of rat and human hepatocytes in vitro. Arch Toxicol 73, 451–456. Hasmall, S. C., James, N. H., Soames, A. R., and Roberts, R. A. (1998). The peroxisome proliferator nafenopin does not suppress hepatocyte apoptosis in guinea-pig liver in vivo nor in human hepatocytes in vitro. Arch Toxicol 72, 777–783. Hays, T., Rusyn, I., Burns, A. M., Kennett, M. J., Ward, J. M., Gonzalez, F. J., and Peters, J. M. (2005). Role of peroxisome proliferator-activated receptor-alpha (PPARalpha) in bezafibrate-induced hepatocarcinogenesis and cholestasis. Carcinogenesis 26, 219–227. Hertz, R., Bishara-Shieban, J., and Bar-Tana, J. (1995). Mode of action of peroxisome proliferators as hypolipidemic drugs. Suppression of apolipoprotein C-III. J Biol Chem 270, 13470–13475. Hess, R., Staubli, W., and Riess, W. (1965). Nature of the hepatomegalic effect produced by ethylchlorophenoxy-isobutyrate in the rat. Nature 208, 856–858. Hinton, R. H., Mitchell, F. E., Mann, A., Chescoe, D., Price, S. C., Nunn, A., Grasso, P., and Bridges, J. W. (1986). Effects of phthalic acid esters on the liver and thyroid. Environ Health Perspect 70, 195–210. Hoivik, D. J., Qualls, C. W., Jr., Mirabile, R. C., Cariello, N. F., Kimbrough, C. L., Colton, H. M., Anderson, S. P., Santostefano, M. J., Morgan, R. J., Dahl, R. R., Brown, A. R., Zhao, Z., Mudd, P. N., Jr., Oliver, W. B., Jr., Brown, H. R., and Miller, R. T. (2004). Fibrates induce hepatic peroxisome and mitochondrial proliferation without overt evidence of cellular proliferation and oxidative stress in cynomolgus monkeys. Carcinogenesis 25, 1757–1769. Holden, P. R., Hasmall, S. C., James, N. H., West, D. R., Brindle, R. D., Gonzalez, F. J., Peters, J. M., and Roberts, R. A. (2000). Tumour necrosis factor alpha (TNFalpha): role in suppression of apoptosis by the peroxisome proliferator nafenopin. Cell Mol Biol (Noisy-le-grand) 46, 29–39. Holsapple, M. P., Pitot, H. C., Cohen, S. M., Boobis, A. R., Klaunig, J. E., Pastoor, T., Dellarco, V. L., and Dragan, Y. P. (2006). Mode of action in relevance of rodent liver tumors to human cancer risk. Toxicol Sci 89, 51–56. Hsieh, L. L., Shinozuka, H., and Weinstein, I. B. (1991). Changes in expression of cellular oncogenes and endogenous retrovirus-like sequences during hepatocarcinogenesis induced by a peroxisome proliferator. Br J Cancer 64, 815–820. Hsu, M. H., Savas, U., Griffin, K. J., and Johnson, E. F. (2001). Identification of peroxisome proliferatorresponsive human genes by elevated expression of the peroxisome proliferator-activated receptor alpha in HepG2 cells. J Biol Chem 276, 27950–27958. Huber, W., Kraupp-Grasl, B., Esterbauer, H., and Schulte-Hermann, R. (1991). Role of oxidative stress in age dependent hepatocarcinogenesis by the peroxisome proliferator nafenopin in the rat. Cancer Res 51, 1789–1792.

REFERENCES

473

Huber, W. W., Grasl-Kraupp, B., Stekel, H., Gschwentner, C., Lang, H., and Schulte-Hermann, R. (1997). Inhibition instead of enhancement of lipid peroxidation by pretreatment with the carcinogenic peroxisome proliferator nafenopin in rat liver exposed to a high single dose of corn oil. Arch Toxicol 71, 575–581. Isenberg, J. S., Kamendulis, L. M., Ackley, D. C., Smith, J. H., Pugh, G., Jr., Lington, A. W., McKee, R. H., and Klaunig, J. E. (2001). Reversibility and persistence of di-2-ethylhexyl phthalate (DEHP)- and phenobarbital-induced hepatocellular changes in rodents. Toxicol Sci 64, 192–199. Isenberg, J. S., Kamendulis, L. M., Smith, J. H., Ackley, D. C., Pugh, G., Jr., Lington, A. W., and Klaunig, J. E. (2000). Effects of Di-2-ethylhexyl phthalate (DEHP) on gap-junctional intercellular communication (GJIC), DNA synthesis, and peroxisomal beta oxidation (PBOX) in rat, mouse, and hamster liver. Toxicol Sci 56, 73–85. Isenberg, J. S., Kolaja, K. L., Ayoubi, S. A., Watkins, J. B., 3rd, and Klaunig, J. E. (1997). Inhibition of WY-14,643 induced hepatic lesion growth in mice by rotenone. Carcinogenesis 18, 1511–1519. Issemann, I., and Green, S. (1990). Activation of a member of the steroid hormone receptor superfamily by peroxisome proliferators. Nature 347, 645–650. Ito, Y., Yamanoshita, O., Asaeda, N., Tagawa, Y., Lee, C. H., Aoyama, T., Ichihara, G., Furuhashi, K., Kamijima, M., Gonzalez, F. J., and Nakajima, T. (2007). Di(2-ethylhexyl)phthalate induces hepatic tumorigenesis through a peroxisome proliferator-activated receptor alpha-independent pathway. J Occup Health 49, 172–182. James, N. H., Gill, J. H., Brindle, R., Woodyatt, N. J., Macdonald, N., Rolfe, M., Hasmall, S. C., Tugwood, J. D., Holden, P. R., and Roberts, R. A. (1998a). Peroxisome proliferator-activated receptor (PPAR) alpha-regulated growth responses and their importance to hepatocarcinogenesis. Toxicol Lett 102–103, 91–96. James, N. H., and Roberts, R. A. (1996). Species differences in response to peroxisome proliferators correlate in vitro with induction of DNA synthesis rather than suppression of apoptosis. Carcinogenesis 17, 1623–1632. James, N. H., Soames, A. R., and Roberts, R. A. (1998b). Suppression of hepatocyte apoptosis and induction of DNA synthesis by the rat and mouse hepatocarcinogen diethylhexylphlathate (DEHP) and the mouse hepatocarcinogen 1,4-dichlorobenzene (DCB). Arch Toxicol 72, 784–790. Jia, Y., Qi, C., Zhang, Z., Hashimoto, T., Rao, M. S., Huyghe, S., Suzuki, Y., Van Veldhoven, P. P., Baes, M., and Reddy, J. K. (2003). Overexpression of peroxisome proliferator-activated receptor-alpha (PPARalpha)-regulated genes in liver in the absence of peroxisome proliferation in mice deficient in both L- and D-forms of enoyl-CoA hydratase/dehydrogenase enzymes of peroxisomal beta-oxidation system. J Biol Chem 278, 47232–47239. Karin, M. (2006). Nuclear factor-kappaB in cancer development and progression. Nature 441, 431–436. Kasai, H. (1997). Analysis of a form of oxidative DNA damage, 8-hydroxy-2′-deoxyguanosine, as a marker of cellular oxidative stress during carcinogenesis. Mutat Res 387, 147–163. Keller, H., Devchand, P. R., Perroud, M., and Wahli, W. (1997). PPAR alpha structure-function relationships derived from species-specific differences in responsiveness to hypolipidemic agents. Biol Chem 378, 651–655. Keller, H., Dreyer, C., Medin, J., Mahfoudi, A., Ozato, K., and Wahli, W. (1993). Fatty acids and retinoids control lipid metabolism through activation of peroxisome proliferator-activated receptor-retinoid X receptor heterodimers. Proc Natl Acad Sci USA 90, 2160–2164. Klaunig, J. E., Babich, M. A., Baetcke, K. P., Cook, J. C., Corton, J. C., David, R. M., DeLuca, J. G., Lai, D. Y., McKee, R. H., Peters, J. M., Roberts, R. A., and Fenner-Crisp, P. A. (2003). PPARalpha agonist-induced rodent tumors: modes of action and human relevance. Crit Rev Toxicol 33, 655–780. Klaunig, J. E., Xu, Y., Isenberg, J. S., Bachowski, S., Kolaja, K. L., Jiang, J., Stevenson, D. E., and Walborg, E. F., Jr. (1998). The role of oxidative stress in chemical carcinogenesis. Environ Health Perspect 106 (Suppl 1), 289–295. Klungland, A., Rosewell, I., Hollenbach, S., Larsen, E., Daly, G., Epe, B., Seeberg, E., Lindahl, T., and Barnes, D. E. (1999). Accumulation of premutagenic DNA lesions in mice defective in removal of oxidative base damage. Proc Natl Acad Sci USA 96, 13300–13305.

474


Lacquemant, C., Lepretre, F., Pineda Torra, I., Manraj, M., Charpentier, G., Ruiz, J., Staels, B., and Froguel, P. (2000). Mutation screening of the PPARalpha gene in type 2 diabetes associated with coronary heart disease. Diabetes Metab 26, 393–401. Lake, B. G., Evans, J. G., Cunninghame, M. E., and Price, R. J. (1993). Comparison of the hepatic effects of nafenopin and WY-14,643 on peroxisome proliferation and cell replication in the rat and Syrian hamster. Environ Health Perspect 101 (Suppl 5), 241–247. Lake, B. G., Evans, J. G., Gray, T. J., Korosi, S. A., and North, C. J. (1989a). Comparative studies on nafenopin-induced hepatic peroxisome proliferation in the rat, Syrian hamster, guinea pig, and marmoset. Toxicol Appl Pharmacol 99, 148–160. Lake, B. G., Gray, T. J., Korosi, S. A., and Walters, D. G. (1989b). Nafenopin, a peroxisome proliferator, depletes hepatic vitamin E content and elevates plasma oxidised glutathione levels in rats. Toxicol Lett 45, 221–229. Lake, B. G., Kozlen, S. L., Evans, J. G., Gray, T. J., Young, P. J., and Gangolli, S. D. (1987). Effect of prolonged administration of clofibric acid and di-(2-ethylhexyl)phthalate on hepatic enzyme activities and lipid peroxidation in the rat. Toxicology 44, 213–228. Lake, B. G., Rumsby, P. C., Price, R. J., and Cunninghame, M. E. (2000). Species differences in hepatic peroxisome proliferation, cell replication and transforming growth factor-beta1 gene expression in the rat, Syrian hamster and guinea pig. Mutat Res 448, 213–225. Laughter, A. R., Dunn, C. S., Swanson, C. L., Howroyd, P., Cattley, R. C., and Corton, J. C. (2004). Role of the peroxisome proliferator-activated receptor alpha (PPARalpha) in responses to trichloroethylene and metabolites, trichloroacetate and dichloroacetate in mouse liver. Toxicology 203, 83–98. Lawrence, J. W., Li, Y., Chen, S., DeLuca, J. G., Berger, J. P., Umbenhauer, D. R., Moller, D. E., and Zhou, G. (2001a). Differential gene regulation in human versus rodent hepatocytes by peroxisome proliferator-activated receptor (PPAR) alpha. PPAR alpha fails to induce peroxisome proliferationassociated genes in human cells independently of the level of receptor expresson. J Biol Chem 276, 31521–31527. Lawrence, J. W., Wollenberg, G. K., and DeLuca, J. G. (2001b). Tumor necrosis factor alpha is not required for WY14,643-induced cell proliferation. Carcinogenesis 22, 381–386. Lawrence, J. W., Wollenberg, G. K., Frank, J. D., and DeLuca, J. G. (2001c). Dexamethasone selectively inhibits WY14,643-induced cell proliferation and not peroxisome proliferation in mice. Toxicol Appl Pharmacol 170, 113–123. Li, Y., Glauert, H. P., and Spear, B. T. (2000a). Activation of nuclear factor-kappaB by the peroxisome proliferator ciprofibrate in H4IIEC3 rat hepatoma cells and its inhibition by the antioxidants N-acetylcysteine and vitamin E. Biochem Pharmacol 59, 427–434. Li, Y., Kovach, A., Suino-Powell, K., Martynowski, D., and Xu, H. E. (2008). Structural and biochemical basis for the binding selectivity of peroxisome proliferator-activated receptor gamma to PGC-1alpha. J Biol Chem 283, 19132–19139. Li, Y., Leung, L. K., Glauert, H. P., and Spear, B. T. (1996). Treatment of rats with the peroxisome proliferator ciprofibrate results in increased liver NF-kappaB activity. Carcinogenesis 17, 2305– 2309. Li, Y., Tharappel, J. C., Cooper, S., Glenn, M., Glauert, H. P., and Spear, B. T. (2000b). Expression of the hydrogen peroxide-generating enzyme fatty acyl CoA oxidase activates NF-kappaB. DNA Cell Biol 19, 113–120. Liu, M. H., Li, J., Shen, P., Husna, B., Tai, E. S., and Yong, E. L. (2008). A natural polymorphism in peroxisome proliferator-activated receptor-alpha hinge region attenuates transcription due to defective release of nuclear receptor corepressor from chromatin. Mol Endocrinol 22, 1078–1092. Luebker, D. J., Hansen, K. J., Bass, N. M., Butenhoff, J. L., and Seacat, A. M. (2002). Interactions of fluorochemicals with rat liver fatty acid-binding protein. Toxicology 176, 175–185. MacDonald, N., Holden, P. R., and Roberts, R. A. (1999). Addition of peroxisome proliferator-activated receptor alpha to guinea pig hepatocytes confers increased responsiveness to peroxisome proliferators. Cancer Res 59, 4776–4780. Maeda, S., Kamata, H., Luo, J. L., Leffert, H., and Karin, M. (2005). IKKbeta couples hepatocyte death to cytokine-driven compensatory proliferation that promotes chemical hepatocarcinogenesis. Cell 121, 977–990.

REFERENCES

475

Makowska, J. M., Gibson, G. G., and Bonner, F. W. (1992). Species differences in ciprofibrate induction of hepatic cytochrome P450 4A1 and peroxisome proliferation. J Biochem Toxicol 7, 183–191. Maloney, E. K., and Waxman, D. J. (1999). trans-Activation of PPARalpha and PPARgamma by structurally diverse environmental chemicals. Toxicol Appl Pharmacol 161, 209–218. Mannaerts, G. P., and van Veldhoven, P. P. (1996). Functions and organization of peroxisomal betaoxidation. Ann NY Acad Sci 804, 99–115. Marsman, D. S., Cattley, R. C., Conway, J. G., and Popp, J. A. (1988). Relationship of hepatic peroxisome proliferation and replicative DNA synthesis to the hepatocarcinogenicity of the peroxisome proliferators di(2-ethylhexyl)phthalate and [4-chloro-6-(2,3-xylidino)-2-pyrimidinylthio]acetic acid (Wy-14,643) in rats. Cancer Res 48, 6739–6744. Marsman, D. S., Goldsworthy, T. L., and Popp, J. A. (1992). Contrasting hepatocytic peroxisome proliferation, lipofuscin accumulation and cell turnover for the hepatocarcinogens Wy-14,643 and clofibric acid. Carcinogenesis 13, 1011–1017. Marsman, D. S., and Popp, J. A. (1994). Biological potential of basophilic hepatocellular foci and hepatic adenoma induced by the peroxisome proliferator, Wy-14,643. Carcinogenesis 15, 111–117. Mascaro, C., Acosta, E., Ortiz, J. A., Marrero, P. F., Hegardt, F. G., and Haro, D. (1998). Control of human muscle-type carnitine palmitoyltransferase I gene transcription by peroxisome proliferatoractivated receptor. J Biol Chem 273, 8560–8563. Meek, M. E. (2008). Recent developments in frameworks to consider human relevance of hypothesized modes of action for tumours in animals. Environ Mol Mutagen 49, 110–116. Menegazzi, M., Carcereri-De Prati, A., Suzuki, H., Shinozuka, H., Pibiri, M., Piga, R., Columbano, A., and Ledda-Columbano, G. M. (1997). Liver cell proliferation induced by nafenopin and cyproterone acetate is not associated with increases in activation of transcription factors NF-kappaB and AP-1 or with expression of tumor necrosis factor alpha. Hepatology 25, 585–592. Meyer, K., Lee, J. S., Dyck, P. A., Cao, W. Q., Rao, M. S., Thorgeirsson, S. S., and Reddy, J. K. (2003). Molecular profiling of hepatocellular carcinomas developing spontaneously in acyl-CoA oxidase deficient mice: comparison with liver tumors induced in wild-type mice by a peroxisome proliferator and a genotoxic carcinogen. Carcinogenesis 24, 975–984. Morimura, K., Cheung, C., Ward, J. M., Reddy, J. K., and Gonzalez, F. J. (2006). Differential susceptibility of mice humanized for peroxisome proliferator-activated receptor alpha to Wy-14,643-induced liver tumorigenesis. Carcinogenesis 27, 1074–1080. Mukherjee, R., Jow, L., Noonan, D., and McDonnell, D. P. (1994). Human and rat peroxisome proliferator activated receptors (PPARs) demonstrate similar tissue distribution but different responsiveness to PPAR activators. J Steroid Biochem Mol Biol 51, 157–166. Nemali, M. R., Usuda, N., Reddy, M. K., Oyasu, K., Hashimoto, T., Osumi, T., Rao, M. S., and Reddy, J. K. (1988). Comparison of constitutive and inducible levels of expression of peroxisomal betaoxidation and catalase genes in liver and extrahepatic tissues of rat. Cancer Res 48, 5316–5324. Nicholls-Grzemski, F. A., Belling, G. B., Priestly, B. G., Calder, I. C., and Burcham, P. C. (2000). Clofibrate pretreatment in mice confers resistance against hepatic lipid peroxidation. J Biochem Mol Toxicol 14, 335–345. Nilakantan, V., Spear, B. T., and Glauert, H. P. (1998). Liver-specific catalase expression in transgenic mice inhibits NF-kappaB activation and DNA synthesis induced by the peroxisome proliferator ciprofibrate. Carcinogenesis 19, 631–637. O’Brien, M. L., Cunningham, M. L., Spear, B. T., and Glauert, H. P. (2001a). Effects of peroxisome proliferators on glutathione and glutathione-related enzymes in rats and hamsters. Toxicol Appl Pharmacol 171, 27–37. O’Brien, M. L., Twaroski, T. P., Cunningham, M. L., Glauert, H. P., and Spear, B. T. (2001b). Effects of peroxisome proliferators on antioxidant enzymes and antioxidant vitamins in rats and hamsters. Toxicol Sci 60, 271–278. Oberhammer, F. A., and Qin, H. M. (1995). Effect of three tumour promoters on the stability of hepatocyte cultures and apoptosis after transforming growth factor-beta 1. Carcinogenesis 16, 1363–1371. Ohmura, T., Ledda-Columbano, G. M., Piga, R., Columbano, A., Glemba, J., Katyal, S. L., Locker, J., and Shinozuka, H. (1996). Hepatocyte proliferation induced by a single dose of a peroxisome proliferator. Am J Pathol 148, 815–824.

476


Pacot, C., Petit, M., Rollin, M., Behechti, N., Moisant, M., Deslex, P., Althoff, J., Lhuguenot, J. C., and Latruffe, N. (1996). Difference between guinea pig and rat in the liver peroxisomal response to equivalent plasmatic level of ciprofibrate. Arch Biochem Biophys 327, 181–188. Palmer, C. N., Hsu, M. H., Griffin, K. J., Raucy, J. L., and Johnson, E. F. (1998). Peroxisome proliferator activated receptor-alpha expression in human liver. Mol Pharmacol 53, 14–22. Perrone, C. E., Shao, L., and Williams, G. M. (1998). Effect of rodent hepatocarcinogenic peroxisome proliferators on fatty acyl-CoA oxidase, DNA synthesis, and apoptosis in cultured human and rat hepatocytes. Toxicol Appl Pharmacol 150, 277–286. Peters, J. M., Aoyama, T., Cattley, R. C., Nobumitsu, U., Hashimoto, T., and Gonzalez, F. J. (1998). Role of peroxisome proliferator-activated receptor alpha in altered cell cycle regulation in mouse liver. Carcinogenesis 19, 1989–1994. Peters, J. M., Cattley, R. C., and Gonzalez, F. J. (1997). Role of PPAR alpha in the mechanism of action of the nongenotoxic carcinogen and peroxisome proliferator Wy-14,643. Carcinogenesis 18, 2029–2033. Peters, J. M., Cheung, C., and Gonzalez, F. J. (2005). Peroxisome proliferator-activated receptor-alpha and liver cancer: where do we stand? J Mol Med 83, 774–785. Peters, J. M., Rusyn, I., Rose, M. L., Gonzalez, F. J., and Thurman, R. G. (2000). Peroxisome proliferatoractivated receptor alpha is restricted to hepatic parenchymal cells, not Kupffer cells: implications for the mechanism of action of peroxisome proliferators in hepatocarcinogenesis. Carcinogenesis 21, 823–826. Plant, N. J., Horley, N. J., Dickins, M., Hasmall, S., Elcombe, C. R., and Bell, D. R. (1998). The coordinate regulation of DNA synthesis and suppression of apoptosis is differentially regulated by the liver growth agents, phenobarbital and methylclofenapate. Carcinogenesis 19, 1521–1527. Price, R. J., Evans, J. G., and Lake, B. G. (1992). Comparison of the effects of nafenopin on hepatic peroxisome proliferation and replicative DNA synthesis in the rat and Syrian hamster. Food Chem Toxicol 30, 937–944. Pugh, G., Jr., Isenberg, J. S., Kamendulis, L. M., Ackley, D. C., Clare, L. J., Brown, R., Lington, A. W., Smith, J. H., and Klaunig, J. E. (2000). Effects of di-isononyl phthalate, di-2-ethylhexyl phthalate, and clofibrate in cynomolgus monkeys. Toxicol Sci 56, 181–188. Qu, B., Halliwell, B., Ong, C. N., Lee, B. L., and Li, Q. T. (2000). Caloric restriction prevents oxidative damage induced by the carcinogen clofibrate in mouse liver. FEBS Lett 473, 85–88. Qu, B., Li, Q. T., Wong, K. P., Tan, T. M., and Halliwell, B. (2001). Mechanism of clofibrate hepatotoxicity: mitochondrial damage and oxidative stress in hepatocytes. Free Radic Biol Med 31, 659–669. Rao, M. S., Lalwani, N. D., Scarpelli, D. G., and Reddy, J. K. (1982). The absence of gamma-glutamyl transpeptidase activity in putative preneoplastic lesions and in hepatocellular carcinomas induced in rats by the hypolipidemic peroxisome proliferator Wy-14,643. Carcinogenesis 3, 1231–1233. Rao, M. S., Lalwani, N. D., Watanabe, T. K., and Reddy, J. K. (1984). Inhibitory effect of antioxidants ethoxyquin and 2(3)-tert-butyl-4-hydroxyanisole on hepatic tumorigenesis in rats fed ciprofibrate, a peroxisome proliferator. Cancer Res 44, 1072–1076. Rao, M. S., and Subbarao, V. (1997a). The effect of deferoxamine on ciprofibrate-induced hepatocarcinogenesis in the rat. In Vivo 11, 495–498. Rao, M. S., and Subbarao, V. (1997b). Effect of dexamethasone on ciprofibrate-induced cell proliferation and peroxisome proliferation. Fundam Appl Toxicol 35, 78–83. Rao, M. S., and Subbarao, V. (1999). Inhibition of ciprofibrate-induced hepatocarcinogenesis in the rat by dimethylthiourea, a scavenger of hydroxyl radical. Oncol Rep 6, 1285–1288. Rao, M. S., Tatematsu, M., Subbarao, V., Ito, N., and Reddy, J. K. (1986). Analysis of peroxisome proliferator-induced preneoplastic and neoplastic lesions of rat liver for placental form of glutathione S-transferase and gamma-glutamyltranspeptidase. Cancer Res 46, 5287–5290. Rao, M. S., Thangada, S., and Subbarao, V. (1991). Peroxisome proliferation in neoplastic nodules and hepatocellular carcinomas induced by ciprofibrate in the rat. Exp Pathol 41, 44–49. Rao, M. S., Usuda, N., Subbarao, V., and Reddy, J. K. (1987). Absence of gamma-glutamyl transpeptidase activity in neoplastic lesions induced in the liver of male F-344 rats by di-(2-ethylhexyl)phthalate, a peroxisome proliferator. Carcinogenesis 8, 1347–1350.

REFERENCES

477

Ray, A., and Prefontaine, K. E. (1994). Physical association and functional antagonism between the p65 subunit of transcription factor NF-kappa B and the glucocorticoid receptor. Proc Natl Acad Sci USA 91, 752–756. Reddy, J. K., and Chu, R. (1996). Peroxisome proliferator-induced pleiotropic responses: pursuit of a phenomenon. Ann NY Acad Sci 804, 176–201. Reddy, J. K., Lalwani, N. D., Reddy, M. K., and Qureshi, S. A. (1982). Excessive accumulation of autofluorescent lipofuscin in the liver during hepatocarcinogenesis by methyl clofenapate and other hypolipidemic peroxisome proliferators. Cancer Res 42, 259–266. Reddy, J. K., and Qureshi, S. A. (1979). Tumorigenicity of the hypolipidaemic peroxisome proliferator ethyl-alpha-p-chlorophenoxyisobutyrate (clofibrate) in rats. Br J Cancer 40, 476–482. Reddy, J. K., and Rao, M. S. (1977). Malignant tumors in rats fed nafenopin, a hepatic peroxisome proliferator. J Natl Cancer Inst 59, 1645–1650. Reddy, J. K., and Rao, M. S. (1989). Oxidative DNA damage caused by persistent peroxisome proliferation: its role in hepatocarcinogenesis. Mutat Res 214, 63–68. Ren, H., Aleksunes, L. M., Wood, C., Vallanat, B., George, M. H., Klaassen, C. D., and Corton, J. C. (2010). Characterization of peroxisome proliferator-activated receptor alpha–independent effects of PPARalpha activators in the rodent liver: di-(2-ethylhexyl) phthalate also activates the constitutiveactivated receptor. Toxicol Sci 113(1), 45–59. Richert, L., Lamboley, C., Viollon-Abadie, C., Grass, P., Hartmann, N., Laurent, S., Heyd, B., Mantion, G., Chibout, S. D., and Staedtler, F. (2003). Effects of clofibric acid on mRNA expression profiles in primary cultures of rat, mouse and human hepatocytes. Toxicol Appl Pharmacol 191, 130–146. Roberts, R. A. (1999). Peroxisome proliferators: mechanisms of adverse effects in rodents and molecular basis for species differences. Arch Toxicol 73, 413–418. Roberts, R. A., Ganey, P. E., Ju, C., Kamendulis, L. M., Rusyn, I., and Klaunig, J. E. (2007). Role of the Kupffer cell in mediating hepatic toxicity and carcinogenesis. Toxicol Sci 96, 2–15. Roberts, R. A., James, N. H., Hasmall, S. C., Holden, P. R., Lambe, K., Macdonald, N., West, D., Woodyatt, N. J., and Whitcome, D. (2000). Apoptosis and proliferation in nongenotoxic carcinogenesis: species differences and role of PPARalpha. Toxicol Lett 112–113, 49–57. Roglans, N., Bellido, A., Rodriguez, C., Cabrero, A., Novell, F., Ros, E., Zambon, D., and Laguna, J. C. (2002). Fibrate treatment does not modify the expression of acyl coenzyme A oxidase in human liver. Clin Pharmacol Ther 72, 692–701. Rolfe, M., James, N. H., and Roberts, R. A. (1997). Tumour necrosis factor alpha (TNF alpha) suppresses apoptosis and induces DNA synthesis in rodent hepatocytes: a mediator of the hepatocarcinogenicity of peroxisome proliferators? Carcinogenesis 18, 2277–2280. Rose, M. L., Cattley, R. C., Dunn, C., Wong, V., Li, X., and Thurman, R. G. (1999a). Dietary glycine prevents the development of liver tumors caused by the peroxisome proliferator WY-14,643. Carcinogenesis 20, 2075–2081. Rose, M. L., Germolec, D., Arteel, G. E., Schoonhoven, R., and Thurman, R. G. (1997a). Dietary glycine prevents increases in hepatocyte proliferation caused by the peroxisome proliferator WY-14,643. Chem Res Toxicol 10, 1198–1204. Rose, M. L., Germolec, D. R., Schoonhoven, R., and Thurman, R. G. (1997b). Kupffer cells are causally responsible for the mitogenic effect of peroxisome proliferators. Carcinogenesis 18, 1453–1456. Rose, M. L., Rivera, C. A., Bradford, B. U., Graves, L. M., Cattley, R. C., Schoonhoven, R., Swenberg, J. A., and Thurman, R. G. (1999b). Kupffer cell oxidant production is central to the mechanism of peroxisome proliferators. Carcinogenesis 20, 27–33. Rosen, M. B., Abbott, B. D., Wolf, D. C., Corton, J. C., Wood, C. R., Schmid, J. E., Das, K. P., Zehr, R. D., Blair, E. T., and Lau, C. (2008a). Gene profiling in the livers of wild-type and PPARalpha-null mice exposed to perfluorooctanoic acid. Toxicol Pathol 36, 592–607. Rosen, M. B., Lee, J. S., Ren, H., Vallanat, B., Liu, J., Waalkes, M. P., Abbott, B. D., Lau, C., and Corton, J. C. (2008b). Toxicogenomic dissection of the perfluorooctanoic acid transcript profile in mouse liver: evidence for the involvement of nuclear receptors PPAR alpha and CAR. Toxicol Sci 103, 46–56. Rusyn, I., Asakura, S., Pachkowski, B., Bradford, B. U., Denissenko, M. F., Peters, J. M., Holland, S. M., Reddy, J. K., Cunningham, M. L., and Swenberg, J. A. (2004). Expression of base excision DNA

478


repair genes is a sensitive biomarker for in vivo detection of chemical-induced chronic oxidative stress: identification of the molecular source of radicals responsible for DNA damage by peroxisome proliferators. Cancer Res 64, 1050–1057. Rusyn, I., Denissenko, M. F., Wong, V. A., Butterworth, B. E., Cunningham, M. L., Upton, P. B., Thurman, R. G., and Swenberg, J. A. (2000a). Expression of base excision repair enzymes in rat and mouse liver is induced by peroxisome proliferators and is dependent upon carcinogenic potency. Carcinogenesis 21, 2141–2145. Rusyn, I., Kadiiska, M. B., Dikalova, A., Kono, H., Yin, M., Tsuchiya, K., Mason, R. P., Peters, J. M., Gonzalez, F. J., Segal, B. H., Holland, S. M., and Thurman, R. G. (2001). Phthalates rapidly increase production of reactive oxygen species in vivo: role of Kupffer cells. Mol Pharmacol 59, 744–750. Rusyn, I., Peters, J. M., and Cunningham, M. L. (2006). Modes of action and species-specific effects of di-(2-ethylhexyl)phthalate in the liver. Crit Rev Toxicol 36, 459–479. Rusyn, I., Rose, M. L., Bojes, H. K., and Thurman, R. G. (2000b). Novel role of oxidants in the molecular mechanism of action of peroxisome proliferators. Antioxid Redox Signal 2, 607–621. Rusyn, I., Tsukamoto, H., and Thurman, R. G. (1998). WY-14 643 rapidly activates nuclear factor kappaB in Kupffer cells before hepatocytes. Carcinogenesis 19, 1217–1222. Rusyn, I., Yamashina, S., Segal, B. H., Schoonhoven, R., Holland, S. M., Cattley, R. C., Swenberg, J. A., and Thurman, R. G. (2000c). Oxidants from nicotinamide adenine dinucleotide phosphate oxidase are involved in triggering cell proliferation in the liver due to peroxisome proliferators. Cancer Res 60, 4798–4803. Sanderson, L. M., de Groot, P. J., Hooiveld, G. J., Koppen, A., Kalkhoven, E., Muller, M., and Kersten, S. (2008). Effect of synthetic dietary triglycerides: a novel research paradigm for nutrigenomics. PLoS ONE 3, e1681. Sapone, A., Peters, J. M., Sakai, S., Tomita, S., Papiha, S. S., Dai, R., Friedman, F. K., and Gonzalez, F. J. (2000). The human peroxisome proliferator-activated receptor alpha gene: identification and functional characterization of two natural allelic variants. Pharmacogenetics 10, 321–333. Sausen, P. J., Lee, D. C., Rose, M. L., and Cattley, R. C. (1995). Elevated 8-hydroxydeoxyguanosine in hepatic DNA of rats following exposure to peroxisome proliferators: relationship to mitochondrial alterations. Carcinogenesis 16, 1795–1801. Schmezer, P., Pool, B. L., Klein, R. G., Komitowski, D., and Schmahl, D. (1988). Various short-term assays and two long-term studies with the plasticizer di(2-ethylhexyl)phthalate in the Syrian golden hamster. Carcinogenesis 9, 37–43. Schoonjans, K., Peinado-Onsurbe, J., Lefebvre, A. M., Heyman, R. A., Briggs, M., Deeb, S., Staels, B., and Auwerx, J. (1996). PPARalpha and PPARgamma activators direct a distinct tissue-specific transcriptional response via a PPRE in the lipoprotein lipase gene. EMBO J 15, 5336–5348. Schulte-Hermann, R., Ohde, G., Schuppler, J., and Timmermann-Trosiener, I. (1981). Enhanced proliferation of putative preneoplastic cells in rat liver following treatment with the tumor promoters phenobarbital, hexachlorocyclohexane, steroid compounds, and nafenopin. Cancer Res 41, 2556– 2562. Seo, K. W., Kim, K. B., Kim, Y. J., Choi, J. Y., Lee, K. T., and Choi, K. S. (2004). Comparison of oxidative stress and changes of xenobiotic metabolizing enzymes induced by phthalates in rats. Food Chem Toxicol 42, 107–114. Shah, Y. M., Morimura, K., Yang, Q., Tanabe, T., Takagi, M., and Gonzalez, F. J. (2007). Peroxisome proliferator-activated receptor alpha regulates a microRNA-mediated signaling cascade responsible for hepatocellular proliferation. Mol Cell Biol 27, 4238–4247. Shaw, D., Lee, R., and Roberts, R. A. (2002). Species differences in response to the phthalate plasticizer monoisononylphthalate (MINP) in vitro: a comparison of rat and human hepatocytes. Arch Toxicol 76, 344–350. Sher, T., Yi, H. F., McBride, O. W., and Gonzalez, F. J. (1993). cDNA cloning, chromosomal mapping, and functional characterization of the human peroxisome proliferator activated receptor. Biochemistry 32, 5598–5604. Shipley, J. M., Hurst, C. H., Tanaka, S. S., DeRoos, F. L., Butenhoff, J. L., Seacat, A. M., and Waxman, D. J. (2004). trans-activation of PPARalpha and induction of PPARalpha target genes by perfluorooctane-based chemicals. Toxicol Sci 80, 151–160.

REFERENCES

479

Smith-Oliver, T., and Butterworth, B. E. (1987). Correlation of the carcinogenic potential of di(2-ethylhexyl)phthalate (DEHP) with induced hyperplasia rather than with genotoxic activity. Mutat Res 188, 21–28. Soames, A. R., Cliffe, S., Pate, I., and Foster, J. R. (1999). Quantitative analysis of the lobular distribution of S-phase in rat liver following dietary administration of di(2-ethylhexyl)phthalate. Toxicol Pathol 27, 436–440. Soliman, M. S., Cunningham, M. L., Morrow, J. D., Roberts, L. J., 2nd, and Badr, M. Z. (1997). Evidence against peroxisome proliferation-induced hepatic oxidative damage. Biochem Pharmacol 53, 1369– 1374. Staels, B., Schoonjans, K., Fruchart, J. C., and Auwerx, J. (1997). The effects of fibrates and thiazolidinediones on plasma triglyceride metabolism are mediated by distinct peroxisome proliferator activated receptors (PPARs). Biochimie 79, 95–99. Stanko, R. T., Sekas, G., Isaacson, I. A., Clarke, M. R., Billiar, T. R., and Paul, H. S. (1995). Pyruvate inhibits clofibrate-induced hepatic peroxisomal proliferation and free radical production in rats. Metabolism 44, 166–171. Styles, J. A., Kelly, M., Pritchard, N. R., and Elcombe, C. R. (1988). A species comparison of acute hyperplasia induced by the peroxisome proliferator methylclofenapate: involvement of the binucleated hepatocyte. Carcinogenesis 9, 1647–1655. Styles, J. A., Kelly, M. D., Pritchard, N. R., and Elcombe, C. R. (1990). Acute hyperplasia and peroxisome proliferation induced by methylclofenapate: a species comparison and implications for liver carcinogenesis. Prog Clin Biol Res 331, 385–393. Svoboda, D. J., and Azarnoff, D. L. (1979). Tumors in male rats fed ethyl chlorophenoxyisobutyrate, a hypolipidemic drug. Cancer Res 39, 3419–3428. Tachibana, K., Kobayashi, Y., Tanaka, T., Tagami, M., Sugiyama, A., Katayama, T., Ueda, C., Yamasaki, D., Ishimoto, K., Sumitomo, M., Uchiyama, Y., Kohro, T., Sakai, J., Hamakubo, T., Kodama, T., and Doi, T. (2005). Gene expression profiling of potential peroxisome proliferator-activated receptor (PPAR) target genes in human hepatoblastoma cell lines inducibly expressing different PPAR isoforms. Nucl Recept 3, 3. Tai, E. S., Demissie, S., Cupples, L. A., Corella, D., Wilson, P. W., Schaefer, E. J., and Ordovas, J. M. (2002). Association between the PPARA L162V polymorphism and plasma lipid levels: the Framingham Offspring Study. Arterioscler Thromb Vasc Biol 22, 805–810. Takacs, M. L., and Abbott, B. D. (2007). Activation of mouse and human peroxisome proliferatoractivated receptors (alpha, beta/delta, gamma) by perfluorooctanoic acid and perfluorooctane sulfonate. Toxicol Sci 95, 108–117. Takagi, A., Sai, K., Umemura, T., Hasegawa, R., and Kurokawa, Y. (1990). Significant increase of 8-hydroxydeoxyguanosine in liver DNA of rats following short-term exposure to the peroxisome proliferators di(2-ethylhexyl)phthalate and di(2-ethylhexyl)adipate. Jpn J Cancer Res 81, 213–215. Takashima, K., Ito, Y., Gonzalez, F. J., and Nakajima, T. (2008). Different mechanisms of DEHP-induced hepatocellular adenoma tumorigenesis in wild-type and Ppar alpha-null mice. J Occup Health 50, 169–180. Tanaka, K., Smith, P. F., Stromberg, P. C., Eydelloth, R. S., Herold, E. G., Grossman, S. J., Frank, J. D., Hertzog, P. R., Soper, K. A., and Keenan, K. P. (1992). Studies of early hepatocellular proliferation and peroxisomal proliferation in Sprague-Dawley rats treated with tumorigenic doses of clofibrate. Toxicol Appl Pharmacol 116, 71–77. Tanaka, N., Moriya, K., Kiyosawa, K., Koike, K., and Aoyama, T. (2008a). Hepatitis C virus core protein induces spontaneous and persistent activation of peroxisome proliferator-activated receptor alpha in transgenic mice: implications for HCV-associated hepatocarcinogenesis. Int J Cancer 122, 124–131. Tanaka, N., Moriya, K., Kiyosawa, K., Koike, K., Gonzalez, F. J., and Aoyama, T. (2008b). PPARalpha activation is essential for HCV core protein-induced hepatic steatosis and hepatocellular carcinoma in mice. J Clin Invest 118, 683–694. Tharappel, J. C., Cunningham, M. L., Spear, B. T., and Glauert, H. P. (2001). Differential activation of hepatic NF-kappaB in rats and hamsters by the peroxisome proliferators Wy-14,643, gemfibrozil, and dibutyl phthalate. Toxicol Sci 62, 20–27.

480


Tharappel, J. C., Nalca, A., Owens, A. B., Ghabrial, L., Konz, E. C., Glauert, H. P., and Spear, B. T. (2003). Cell proliferation and apoptosis are altered in mice deficient in the NF-kappaB p50 subunit after treatment with the peroxisome proliferator ciprofibrate. Toxicol Sci 75, 300–308. Thottassery, J., Winberg, L., Youssef, J., Cunningham, M., and Badr, M. (1992). Regulation of perfluorooctanoic acid–induced peroxisomal enzyme activities and hepatocellular growth by adrenal hormones. Hepatology 15, 316–322. Tomaszewski, K. E., Heindel, S. W., Jenkins, W. L., and Melnick, R. L. (1990). Induction of peroxisomal acyl CoA oxidase activity and lipid peroxidation in primary rat hepatocyte cultures. Toxicology 65, 49–60. Trapp, C., Schwarz, M., and Epe, B. (2007). The peroxisome proliferator WY-14,643 promotes hepatocarcinogenesis caused by endogenously generated oxidative DNA base modifications in repairdeficient Csbm/m/Ogg1-/- mice. Cancer Res 67, 5156–5161. Tugwood, J. D., Aldridge, T. C., Lambe, K. G., Macdonald, N., and Woodyatt, N. J. (1996). Peroxisome proliferator-activated receptors: structures and function. Ann NY Acad Sci 804, 252–265. Tugwood, J. D., Holden, P. R., James, N. H., Prince, R. A., and Roberts, R. A. (1998). A peroxisome proliferator-activated receptor-alpha (PPARalpha) cDNA cloned from guinea-pig liver encodes a protein with similar properties to the mouse PPARalpha: implications for species differences in responses to peroxisome proliferators. Arch Toxicol 72, 169–177. Valles, E. G., Laughter, A. R., Dunn, C. S., Cannelle, S., Swanson, C. L., Cattley, R. C., and Corton, J. C. (2003). Role of the peroxisome proliferator-activated receptor alpha in responses to diisononyl phthalate. Toxicology 191, 211–225. Van Rafelghem, M. J., Mattie, D. R., Bruner, R. H., and Andersen, M. E. (1987). Pathological and hepatic ultrastructural effects of a single dose of perfluoro-n-decanoic acid in the rat, hamster, mouse, and guinea pig. Fundam Appl Toxicol 9, 522–540. Vanden Heuvel, J. P., Thompson, J. T., Frame, S. R., and Gillies, P. J. (2006). Differential activation of nuclear receptors by perfluorinated fatty acid analogs and natural fatty acids: a comparison of human, mouse, and rat peroxisome proliferator-activated receptor-alpha, -beta, and -gamma, liver X receptorbeta, and retinoid X receptor-alpha. Toxicol Sci 92, 476–489. Varanasi, U., Chu, R., Huang, Q., Castellon, R., Yeldandi, A. V., and Reddy, J. K. (1996). Identification of a peroxisome proliferator-responsive element upstream of the human peroxisomal fatty acyl coenzyme A oxidase gene. J Biol Chem 271, 2147–2155. Vu-Dac, N., Schoonjans, K., Kosykh, V., Dallongeville, J., Fruchart, J. C., Staels, B., and Auwerx, J. (1995). Fibrates increase human apolipoprotein A-II expression through activation of the peroxisome proliferator-activated receptor. J Clin Invest 96, 741–750. Vu-Dac, N., Schoonjans, K., Laine, B., Fruchart, J. C., Auwerx, J., and Staels, B. (1994). Negative regulation of the human apolipoprotein A-I promoter by fibrates can be attenuated by the interaction of the peroxisome proliferator-activated receptor with its response element. J Biol Chem 269, 31012–31018. Wada, N., Marsman, D. S., and Popp, J. A. (1992). Dose-related effects of the hepatocarcinogen, Wy-14,643, on peroxisomes and cell replication. Fundam Appl Toxicol 18, 149–154. Wahli, W., Braissant, O., and Desvergne, B. (1995). Peroxisome proliferator activated receptors: transcriptional regulators of adipogenesis, lipid metabolism and more. Chem Biol 2, 261–266. Watanabe, T., Horie, S., Yamada, J., Isaji, M., Nishigaki, T., Naito, J., and Suga, T. (1989). Species differences in the effects of bezafibrate, a hypolipidemic agent, on hepatic peroxisome-associated enzymes. Biochem Pharmacol 38, 367–371. West, D. A., James, N. H., Cosulich, S. C., Holden, P. R., Brindle, R., Rolfe, M., and Roberts, R. A. (1999). Role for tumor necrosis factor alpha receptor 1 and interleukin-1 receptor in the suppression of mouse hepatocyte apoptosis by the peroxisome proliferator nafenopin. Hepatology 30, 1417–1424. Widen, C., Gustafsson, J. A., and Wikstrom, A. C. (2003). Cytosolic glucocorticoid receptor interaction with nuclear factor-kappa B proteins in rat liver cells. Biochem J 373, 211–220. Williams, G. M., and Perrone, C. (1996). Mechanism-based risk assessment of peroxisome proliferating rodent hepatocarcinogens. Ann NY Acad Sci 804, 554–572. Wolf, D. C., Moore, T., Abbott, B. D., Rosen, M. B., Das, K. P., Zehr, R. D., Lindstrom, A. B., Strynar, M. J., and Lau, C. (2008). Comparative hepatic effects of perfluorooctanoic acid and WY 14,643 in PPAR-alpha knockout and wild-type mice. Toxicol Pathol 36, 632–639.

REFERENCES

481

Woods, C. G., Burns, A. M., Bradford, B. U., Ross, P. K., Kosyk, O., Swenberg, J. A., Cunningham, M. L., and Rusyn, I. (2007a). WY-14,643 induced cell proliferation and oxidative stress in mouse liver are independent of NADPH oxidase. Toxicol Sci 98, 366–374. Woods, C. G., Burns, A. M., Maki, A., Bradford, B. U., Cunningham, M. L., Connor, H. D., Kadiiska, M. B., Mason, R. P., Peters, J. M., and Rusyn, I. (2007b). Sustained formation of alpha-(4-pyridyl-1oxide)-N-tert-butylnitrone radical adducts in mouse liver by peroxisome proliferators is dependent upon peroxisome proliferator-activated receptor-alpha, but not NADPH oxidase. Free Radic Biol Med 42, 335–342. Woods, C. G., Kosyk, O., Bradford, B. U., Ross, P. K., Burns, A. M., Cunningham, M. L., Qu, P., Ibrahim, J. G., and Rusyn, I. (2007c). Time course investigation of PPARalpha- and Kupffer cell-dependent effects of WY-14,643 in mouse liver using microarray gene expression. Toxicol Appl Pharmacol 225, 267–277. Woodyatt, N. J., Lambe, K. G., Myers, K. A., Tugwood, J. D., and Roberts, R. A. (1999). The peroxisome proliferator (PP) response element upstream of the human acyl CoA oxidase gene is inactive among a sample human population: significance for species differences in response to PPs. Carcinogenesis 20, 369–372. Xiao, S., Anderson, S. P., Swanson, C., Bahnemann, R., Voss, K. A., Stauber, A. J., and Corton, J. C. (2006). Activation of peroxisome proliferator-activated receptor alpha enhances apoptosis in the mouse liver. Toxicol Sci 92, 368–377. Xu, H. E., and Li, Y. (2008). Ligand-dependent and -independent regulation of PPAR gamma and orphan nuclear receptors. Sci Signal 1, pe52. Yamakawa-Kobayashi, K., Ishiguro, H., Arinami, T., Miyazaki, R., and Hamaguchi, H. (2002). A Val227Ala polymorphism in the peroxisome proliferator activated receptor alpha (PPARalpha) gene is associated with variations in serum lipid levels. J Med Genet 39, 189–191. Yang, Q., Ito, S., and Gonzalez, F. J. (2007). Hepatocyte-restricted constitutive activation of PPAR alpha induces hepatoproliferation but not hepatocarcinogenesis. Carcinogenesis 28, 1171–1177. Yang, Q., Nagano, T., Shah, Y., Cheung, C., Ito, S., and Gonzalez, F. J. (2008). The PPAR alphahumanized mouse: a model to investigate species differences in liver toxicity mediated by PPAR alpha. Toxicol Sci 101, 132–139. Yeldandi, A. V., Milano, M., Subbarao, V., Reddy, J. K., and Rao, M. S. (1989). Evaluation of liver cell proliferation during ciprofibrate-induced hepatocarcinogenesis. Cancer Lett 47, 21–27. Yeldandi, A. V., Rao, M. S., and Reddy, J. K. (2000). Hydrogen peroxide generation in peroxisome proliferator-induced oncogenesis. Mutat Res 448, 159–177. Youssef, J. A., Bouziane, M., and Badr, M. Z. (2003). Age-dependent effects of nongenotoxic hepatocarcinogens on liver apoptosis in vivo. Mech Ageing Dev 124, 333–340.

CH A P TE R

18

ALPHA2U-GLOBULIN NEPHROPATHY AND CHRONIC PROGRESSIVE NEPHROPATHY AS MODES OF ACTION FOR RENAL TUBULE TUMOR INDUCTION IN RATS, AND THEIR POSSIBLE INTERACTION Edward A. Lock Gordon C. Hard

18.1.

INTRODUCTION

It is now well established that a low incidence of renal tubule tumors (RTT) can be produced by certain chemicals in male rats through a mechanism involving proximal tubule accumulation of a rat-specific protein with subsequent sustained compensatory cell regeneration. The protein, α2u-globulin (α2u-g), occurs at much lower concentrations in female rats, and not at all in mice. Hence, chemically induced RTT arising in female rats or male or female mice cannot be explained by this mechanism (Swenberg et al. 1989; Hard et al. 1993; Hard 1998; Lehman-McKeeman et al. 1998; Meek et al. 2003; Lock and Hard 2004). Male rats—and, to a lesser extent, female rats—are also predisposed to developing chronic progressive nephropathy (CPN), and this age-related, spontaneous disease entity appears to convey a slightly increased risk for development of atypical tubule hyperplasia, a preneoplastic lesion, and RTT later in life (Hard 1998, 2002; Seely et al. 2002; Lock and Hard 2004; Hard and Khan 2004). This chapter will do the following: (1) briefly discuss the mechanisms whereby these responses are observed in rat kidney, (2) provide examples of chemicals falling into these two classes of activity, (3) discuss some of the areas of potential conflict


482

18.2. CHEMICALS THAT INCREASE THE INCIDENCE OF RENAL TUBULE TUMORS

483

when endeavoring to classify chemicals into one or the other of these classes, (4) discuss the potential for interaction between these two pathways in leading to tumor development, and (5) discuss the relevance of these findings to humans.

18.2. CHEMICALS THAT INCREASE THE INCIDENCE OF RENAL TUBULE TUMORS IN MALE RATS BY AN α2U-GLOBULIN MODE OF ACTION Conventional male rats, but not female rats, are physiologically proteinuric because of the high urinary excretion of a low-molecular-weight protein (LMW), α2u-g. This protein (molecular weight 18–20 kilodaltons) is synthesized mainly in the liver of the male rat, where hepatic mRNA for α2u-g represents about 1% of total hepatic mRNA. Female and male rats synthesize this protein in much smaller amounts in other locations, such as secondary sex glands, salivary glands, and lachrymal glands (Lock et al. 1987; MacInnes et al. 1986; Mancini et al. 1989). In male rats, α2u-g is freely filtered at the glomerulus into the tubular lumen, with about 40% being excreted in the urine and the remainder endocytosed by cells in the P2 segment of the proximal tubule, where the protein undergoes catabolism within cellular phagolysosomes (Neuhaus et al. 1981; Lehman-McKeeman et al. 1998). Female rats excrete several hundred times less α2u-g in their urine than do males (Vandoren et al. 1983). The function of the urinary protein in male rats appears to be for territorial scent marking, with the protein acting as a scent carrier. A structurally similar protein is found in mouse urine, which has been studied extensively by scientists interested in olfaction and animal ecology (Novotny 2003; Brennan and Kendrick 2006). These rodent proteins have the ability to bind pheromones. A number of strongly odoriferous compounds have been identified bound to mouse urinary protein (MUP), such as dehydro-exo-brevicomin and 2-sec-butyl-4,5-dihydrothiazole, which potentiate aggression in male mice, and the sesquiterpenes α-farnesene and β-farnesene, which signal dominance in males. Other ligands include 2-heptanone, a fairly common metabolic product, and 6-hydroxy-6-methyl-3-heptanone. Though female rats possess the entire complement of hepatic α2u-g genes, estrogen is a very effective repressor of the expression of these genes in the liver (Roy et al. 1975). Masculinisation of female rats will increase the expression of α2u-g, but not to the same levels as in males (Roy and Neuhaus 1967). In male rats, α2u-g expression is regulated by a complex interaction of testosterone, glucocorticoids, insulin, thyroid hormone, and growth hormone, with gene expression being maximal in hormonally intact, sexually mature male rats. Hence, in immature male rats the protein is either absent or present at a low concentration, while in older rats it tends to decline as testosterone levels wane, being absent or at a very low level in 12- to 18-month-old male rats (Roy et al. 1983). A number of chemicals of diverse structure have been shown to produce a specific form of nephropathy in male rats but not in female rats or mice of either sex. The histological features of this syndrome, which has been called “α2u-globulin nephropathy” or “hyaline droplet nephropathy,” are the excessive accumulation of eosinophilic, hyaline droplets in epithelial cells of the P2 segment, an increase in

484

CHAPTER 18 ALPHA2U-GLOBULIN NEPHROPATHY AND CPN

granular casts at the junction of the outer and inner stripes of the outer medulla, and evidence of early exacerbation of CPN. After many months to 2 years of treatment, linear mineralization caused by the accumulation of calcium hydroxyapatite in the thin limbs of Henle is noted in the papilla, and there may be a low incidence of renal tubule hyperplasia and RTT, but in male rats only. By this stage, exacerbation of CPN is more pronounced (Swenberg et al. 1989; Hard et al. 1993). Among the chemicals in this class tested in 2-year carcinogenicity bioassays by the U.S. National Toxicology Program (NTP), the solvent, decalin (NTP 2005), has produced the highest incidence of RTT, at 30%. The food constituent, d-limonene (NTP 1990) also produced a relatively high incidence, at 22%. Other examples of these male rat-specific renal carcinogens include unleaded gasoline and certain jet fuels, the dry-cleaning agent tetrachloroethylene, and the insect repellant 1,4-dichlorobenzene. The initiating step in the mechanism is the noncovalent binding of the xenobiotic chemical or its metabolite to α2u-g (Lock et al. 1987; Lehman-McKeeman et al. 1989). This reversible binding interferes with the intra-renal lysosomal degradation of the protein, by prolonging the naturally very long half-life of 5–8 hr by about 30% (Lehman-McKeeman et al. 1990). The latter estimate of 30% is derived from in vitro experimentation, but serves to pinpoint what is undoubtedly a more severe problem in vivo. In the milieu of the functioning nephron where many LMW proteins are competing for lysosomal catabolism in the P2 tubules at the same time, the prolongation of α2u-g half-life is likely to be very much longer, causing lysosomal congestion. This results in the accumulation of the protein chemical complex in the P2 segment, which is visible microscopically as hyaline droplets. Under fluorescence microscopy, or following Mallory–Heidenhain staining, many of the droplets can be seen to contain large, polyangular crystalline forms (Hard 2008). Lysosomal overload of tubule cells results in single cell detachment into the lumen, and these exfoliated cells pass down the tubule in the filtrate. They become lodged at the junction where the wider lumen of the P3 tubule narrows into the thin descending limb of Henle to form granular casts, which markedly dilate the affected portion of tubule. The cells probably accumulate at this juncture because they are engorged with poorly digestible protein of crystalline nature. In turn, there is a compensatory cell proliferation in the cortex where the cell loss occurred, which persists as long as exposure to the chemical continues (Short et al. 1989), but presumably not beyond the age when liver synthesis of α2u-g has ceased. Thus, the renal tubule injury is a consequence of the perturbation of a physiological process and not due to a direct action of the chemical or a metabolite. A number of key studies have substantiated the tight correlation between the accumulation of α2u-g and the increase in hyaline droplets. The most compelling evidence for the role of this protein in the nephropathy comes from studies with genetically defective rats and transgenic mice. The male NCI Black–Reiter (NBR) rat lacks mRNA for α2u-g in the liver (Chatterjee et al. 1989) and consequently does not develop the nephropathy when challenged with chemicals such a d-limonene and lindane (Dietrich and Swenberg 1990, 1991a), propylene glycol mono-t-butyl ether (Doi et al. 2004), and decalin (NTP 2005). As mentioned earlier, wild-type mice excrete large amounts of a urinary protein (MUP), which is structurally related


485

to α2u-g; however, they do not develop a nephropathy upon exposure to these chemicals. This appears to be due to two main factors: (1) MUP does not bind ligands such as d-limonene-1,2-epoxide, and (2) MUP is not reabsorbed by the proximal tubule (Lehman-McKeeman and Caudill 1992b). Interestingly, MUP binds 2-secbutyl-4,5-dihydrothiazole, which gives the protein and urine its odoriferous smell. This ligand can also bind to α2u-g with a Ki = 2.3 μM and is able to displace dlimonene-1,2-epoxide from its binding site (Lehman-McKeeman et al. 1998). Administration of 2-sec-butyl-4,5-dihydrothiazole to male rats, as anticipated, produces an increase in hyaline droplet formation in the P2 segment of the proximal tubules and increases the concentration of α2u-g in the kidney (Lehman-McKeeman et al. 1998). The crystal structure of α2u-g and its complex with d-limonene-1,2epoxide, at 2.9-Å resolution, has been published and the binding site for these ligands compared with the corresponding mouse protein (Chaudhuri et al. 1999). Transgenic mice modified to express α2u-g exhibit hyaline droplet formation on challenge with d-limonene, whereas wild-type mice do not (Lehman-McKeeman and Caudill 1994). Of importance to human risk assessment are studies showing that other members of the lipocalin superfamily of proteins, including human-derived α1-acid glycoprotein, rat-derived retinol-binding protein, human protein-1, and bovine β-lactoglobulin, do not bind either d-limonene-1,2-epoxide or 2,4,4trimethyl-2-pentanol, both high-affinity ligands for α2u-g. These proteins were, however, able to bind their own ligands such as (a) retinol by retinol binding protein and (b) progesterone by α1-acid glycoprotein. It therefore appears that under conditions where members of the α2u-g superfamily of proteins are known to bind to established, physiological ligands, those proteins do not interact with hyaline droplet inducing agents (Lehman-McKeeman and Caudill 1992a). The link between hyaline droplet nephropathy, renal tubule cell proliferation, and renal tumors comes from studies by Swenberg and co-workers. They demonstrated that unleaded gasoline produced a sustained increase in cell proliferation in renal cortical tubule cells throughout and beyond the period of chemical exposure (Short et al. 1989). In an initiation/promotion model using ethyl hydroxyethylnitrosamine (EHEN) as the initiating agent, they also showed that in contrast to male Fischer 344 rats, male NBR rats did not respond to d-limonene with an increase in renal hyperplasia and did not develop renal tumors beyond that of the background with EHEN alone (Dietrich and Swenberg 1991b). This series of studies has provided a mechanistic basis for the production of male-rat specific renal tumors, by a nongenotoxic mechanism that has no relevance to humans (Hard et al. 1993; Dietrich 1995; Lehman-McKeeman et al. 1998; IARC 1999). Criteria have been defined by regulatory and authoritative bodies, such as the U.S. Environmental Protection Agency (EPA) (EPA 1991) and the International Agency for Research on Cancer (IARC) (IARC 1999), to enable chemicals to be placed in this class. The essential evidence required for establishing a role for α2u-g nephropathy in renal carcinogenesis is as follows: 1. The renal tumors occur only in male rats. 2. Acute exposure to the chemical causes hyaline droplet accumulation in proximal convoluted tubules.

486


3. The protein accumulating in hyaline droplets should be identified as α2u-g. 4. Hallmark histopathologic lesions, including granular casts at subchronic timepoints and linear papillary mineralization at chronic stages, should be observed. 5. There should be an absence of hyaline droplets and other typical histopathological changes in female rats and mice of both sexes. 6. The chemical should be negative in short-term tests for genotoxicity. Additional supporting evidence includes demonstration of (a) reversible binding of the chemical or a metabolite to α2u-g, (b) a sustained increase in cell proliferation in the proximal tubules (P2 segment), and (c) a dose–response relationship between hyaline droplet severity and renal tumor incidence (IARC 1999). The α2u-g hypothesis has stood the test of time, although not without challenge and debate [see Melnick (1993), Borghoff et al. (1993), Ashby (1996), Huff (1996), Melnick et al. (1997), and Dietrich (1997)]. This is due, in part, to the fact that only a small number of chemicals have been shown to fulfill all of the necessary criteria, leading some workers to seek alternative mechanisms of action (Melnick 1992; Kohn and Melnick 1999; Doi et al. 2007). Amongst the few chemicals that have been shown to sufficiently meet the required criteria are: d-limonene (Hard and Whysner 1994), 1,4-dichlorobenzene (Barter and Sherman 1999), and decalin (Dill et al. 2003). Concerns have also been raised because chemicals—for example, 1-(aminomethyl)cyclohexaneacetic acid (GABA-pentane)(Dominick et al. 1991); Stoddard solvent IIC (Doi et al. 2007); p-nitrobenzoic acid (Williams et al. 2001), and lindane (Dietrich and Swenberg 1990)—produce hyaline droplet formation representing accumulation of α2u-g in male rats without an increase in RTT. In some of these cases, the answer may lie in the fact that the severity of the hyaline droplet accumulation has not been sufficient to produce extensive cell degeneration/regeneration. Additionally, if the ligand has a rather low affinity for α2u-g, the protein–chemical complex may not be stable enough to slow lysosomal degradation and lead to cell loss. In other words, chemicals with the potential to bind to α2u-g and cause some degree of hyaline droplet accumulation will not be of equal potency, but will instead show a range of activity from weak to strong. Perhaps only the chemicals with strong activity produce sufficiently sustained regenerative conditions to lead to tumour development. Kohn and Melnick (1999) have attempted to construct a physiologically based pharmacokinetic model using datasets (sometimes incomplete) produced by scientists at the Chemical Industries Institute of Toxicology (CIIT), who worked with the α2u-g ligand 2,4,4-trimethyl-2-pentanol in male rats. The CIIT scientists had shown that male but not female rats administered trimethylpentane developed hyaline droplets and increased α2u-g in the renal P2 segment and that a metabolite of trimethylpentane (i.e., 2,4,4-trimethyl-2-pentanol) was reversibly bound to α2u-g (Charbonneau et al. 1987a; Lock et al. 1987). Subsequent studies showed that male rats administered unleaded gasoline also had 2,4,4-trimethyl-2-pentanol bound to renal α2u-g, suggesting that this metabolite of aliphatic hydrocarbons present in unleaded gasoline was the likely source leading to the renal injury (Charbonneau et al. 1987b). Kohn and Melnick (1999) found that they were unable to model the


487

experimental findings by just reducing the rate of lysosomal proteolysis of α2u-g. However, if they built in a transient increase in the hepatic synthesis of the protein and a consequent increase in its secretion from the liver, they were able to reproduce the time-course data for blood and renal 2,4,4-trimethyl-2-pentanol concentrations and for α2u-g (Kohn and Melnick 1999). They concluded that in addition to a decreased proteolysis of the protein, some increased hepatic synthesis of the protein was required to explain the findings. This seems plausible because one might expect some feedback mechanisms to exist to switch on hepatic synthesis should plasma levels transiently drop due to removal of the chemically bound form from the circulation. These authors also suggested that increased lysosomal activity and the generation of toxic metabolites of trimethylpentane within the tubule cells may have contributed to the nephrotoxicity observed. However, there are a number of key findings that suggest that any toxic metabolite accumulation is not sufficient to cause nephrotoxicity, as illustrated by the lack of renal tubule necrosis in female rats exposed to these chemicals where the metabolism is broadly similar to that of male rats (Charbonneau et al. 1987a). Furthermore, male rats that do not express the protein, such as the NBR strain, do not show any evidence of renal injury when exposed to these chemicals. Moreover, an in vitro study with isolated proximal tubule cells exposed to high concentrations of one model compound showed no cytotoxicity (Wilke et al. 1993). A recent comparative study (Doi et al. 2007) attempted to clarify the relationship between α2u-g nephropathy and RTT development in male rats by reevaluating the data from four chemicals tested by the U.S. NTP, namely, d-limonene (NTP 1990), decalin (NTP 2005), propylene glycol monobutyl ether (PGMBE)(NTP 2004c), and Stoddard solvent IIC (SS IIC)(NTP 2004b). Doi et al. (2007) examined the reported hyaline droplet formation, α2u-g concentration in the kidney, presence of granular casts in the outer medulla, and the extent of renal tubule cell proliferation following 3 months of exposure at five dose levels. They also reexamined the kidneys of 30 animals following 2 years exposure for severity of linear mineralization of the renal papilla, CPN, renal tubule hyperplasia, and incidence of neoplastic lesions. This was done for two dose levels for d-limonene, four for decalin, and three for PGMBE and SS IIC. Because the Doi et al. (2007) results represented a sampling of the total group numbers from the original NTP studies, their incidences of hyperplastic foci and RTT do not reflect numerically the NTP 2-year bioassay data for these four chemicals. All compounds produced an increase in the renal content of α2u-g; for dlimonene, the protein was increased 2- to 2.75-fold over the dose range following 14 doses over 21 days. With decalin, the maximum increase was 4-fold; however, the control value was very low and was not in line with the previous data or with that expected in mature adult male rat kidneys. For PGMBE, again the control value was low with the maximum increase being about 2.25-fold. For SS IIC, the increase was 1.5- to 2-fold over the dose range after 3 months of exposure. The quantitation of hyaline droplet accumulation showed a dose-related increase with these chemicals with a severity band of 2 for the controls and 3–4 for the treated animals. For dlimonene, the severity band was 2–3, the lower response presumably reflecting the shorter duration of exposure. Thus d-limonene and decalin showed a more marked

488


increase in α2u-g, while the response was somewhat lower with PGMBE and SS IIC. Tubule regeneration (defined as clusters of basophilic tubules in the cortex with increased nuclear density and occasional mitotic figures) was observed in the kidneys of all dosed rats showing a 6-fold increase with d-limonene, an 8-fold increase with decalin, an 11-fold increase with SS IIC, while with PGMBE the response was very flat, increasing 2- to 3-fold across the dose range. Reexamination of the data on renal cell turnover (which had been conducted by proliferating cell nuclear antigen [PCNA] immunostaining) following PGMBE and decalin exposure showed a 1.5- to 2-fold increase over control values at 13 weeks. Renal cell turnover had not been determined with d-limonene, while studies with SS IIC had used bromodeoxyuridine (BrDU) infusion with mini-pumps, but the results were not presented in the NTP report. Though there was evidence of renal tubule regeneration with all compounds, the presence of granular casts in the outer medulla was more marked in the dlimonene and decalin treated male rats than in SS IIC-treated rats, while in the PGMBE-treated rats there was only minimal evidence of casts across the dose range. Thus, all four compounds showed many of the hallmarks of α2u-g nephropathy, but the size and severity of the responses measured appeared to be most marked with d-limonene and decalin (Doi et al. 2007). After 2 years of exposure, Doi et al. (2007) noted an increase in linear mineralization of the papilla associated with all compounds in all dose groups. Again, the response was more severe with d-limonene and decalin. Similarly, the severity of CPN exacerbation was more marked with d-limonene, decalin, and PGMBE, while there was no increase in severity with SS IIC. The occurrence of renal tubule hyperplasia, lesions considered to be preneoplastic, was low with all four chemicals. Finally, the incidence of renal tubule adenoma or carcinoma was statistically significantly increased with d-limonene and decalin, with a small but not significant increase in adenomas for PGMBE. Only one adenoma was seen with SS IIC compared to no preneoplastic/neoplastic lesions in any of the four control groups. The Doi et al. (2007) reevaluation is in agreement with other studies showing that it is possible to have α2u-g nephropathy without an increased incidence of RTT (e.g., SS IIC). These authors pointed out that α2u-g accumulation at 3 months and linear mineralization of the papilla at the end of the 2-year study correlated somewhat to the tumor response, while the severity of CPN was, in general, in better agreement with the tumor response. In summary, the position still stands with regard to a link between (a) renal α2u-g accumulation coupled with tubule cell regeneration and (b) the later development of RTT only in male rats. What subsequent work has shown is that the severity of the protein accumulation and the extent of cellular repair can determine the final outcome. In particular, the severity of granular cast formation is a reflection of the degree of tubule cell loss in the cortex coupled with necessity for tubule regeneration. Likewise, the later development of linear papillary mineralisation appears to be an indicator of the amount of preceding tubule cell injury and granular cast formation. It should be noted that granular casts seem to be more easily visualized (or more numerous) in sagittal kidney sections than in transverse sections. If the carcinogenicity study provides only transverse sections, the severity of granular cast formation may be underestimated. Again, sagittal kidney sections often do not

18.3. CHEMICALS INCREASING THE INCIDENCE OF RENAL TUMORS

489

transect the papilla, in which case an absence or low incidence of linear papillary mineralisation may be misleading. Nevertheless, if these histopathological hallmarks of α2u-g nephropathy (granular casts and papillary mineralisation) are severe, they appear to predict development of RTTs. However, it is becoming increasingly clear that the severity of exacerbated CPN may also be a contributory factor.

18.3. CHEMICALS INCREASING THE INCIDENCE OF RENAL TUMORS THROUGH EXACERBATION OF SPONTANEOUS CHRONIC PROGRESSIVE NEPHROPATHY (CPN) CPN is a very common, age-related, spontaneous renal disease affecting conventional strains of rat used in safety evaluation studies—and in particular, the most commonly used strains Fischer 344 and Sprague–Dawley (Gray 1977; Peter et al. 1986; Goldstein et al. 1988; Montgomery and Seely 1990; Hard and Khan 2004). The incidence and severity of CPN can represent a confounding factor in subchronic toxicity and chronic carcinogenicity bioassays, especially if the kidney is the target organ for toxicity (Wolf and Mann 2005). CPN occurs in both sexes of rat, but, because of hormonal factors, it occurs at a higher incidence and with progressively greater severity in males than in females. It is a generally accepted dogma that the disease is due to increased glomerular permeability resulting from protein hyperfiltration. Notwithstanding, the precise basis for the disease is still poorly understood and controversial. It is known that a number of factors, primarily diet-related, can influence the incidence and severity of CPN. Reducing the protein content of the diet is protective, while increasing the protein content exacerbates the disease (Rao et al. 1993). Modification of other dietary components also ameliorates the disease, although restriction of caloric intake is more powerful than any other dietary manipulation (Bertani et al. 1989; Masoro and Yu 1989; Keenan et al. 2000). Long-term administration of androgen can make females more sensitive (Tanaka et al. 1995), indicating that it is the presence of male sex steroids that is associated with the risk of developing CPN, rather than the absence of estrogens (Baylis 1994). It has been recommended that preclinical studies using the common strains of laboratory rat should be conducted under conditions of dietary restriction (Keenan et al. 2000). The reluctance to adopt this approach has primarily been due to the lack of historical background data on lesion incidences with this regimen, compared to the extensive information on conventional studies. It is encouraging to note that the U.S. NTP, which is responsible for the testing of chemicals in the United States, converted some 15 years ago to a rodent diet that was formulated with lower protein content (14%) and higher fiber and fat, namely NTP-2000. Subsequent studies with this diet have reported a reduction in the severity of CPN, renal cortico-medullary tubule mineralization, and cardiomyopathy, without having any major effects on growth or bodyweight (Rao et al. 2001; Rao 2002). It is not the intention in this chapter to discuss the pathology of CPN in rats in detail. This has been done in several comprehensive reviews to which the reader

490


is referred (Gray 1977; Barthold 1979; Hirokawa 1975; Hard and Khan 2004; Peter et al. 1986). Suffice it to say that CPN is first seen at the light microscope level as occasional, small, and discrete foci of basophilic tubules with thickened basement membranes, located in the cortex. This change is associated with eosinophilic hyaline casts in the same tubule downstream in the medulla. With progression, more tubules become affected and foci merge into areas of tubule alteration, accompanied by frank glomerulosclerosis and minor interstitial infiltration of mononuclear inflammatory cells. Thus, the histological hallmarks of the disease are basophilic tubules, thickened basement membranes, hyaline cast formation, and glomerulosclerosis (Peter et al. 1986; Hard and Khan 2004; Hard and Seely 2005). Studies examining cell proliferation in the kidney have shown that CPN is both a degenerative and regenerative disease. The regenerative aspect is supported by studies with 3 H-thymidine, BrDU or PCNA, showing that certain tubules within CPN have a high rate of cell proliferative activity. Because of the increased proliferative activity, it appears that CPN can be a weak risk factor for the spontaneous development of RTT in rats, particularly in males (Hard and Khan 2004). For 90-day toxicity and 2-year carcinogenicity studies performed in accordance with Good Laboratory Practice standards, it is necessary to grade the severity of CPN, which should treat the component lesions as one disease entity. Traditionally, this has been done on a scale of 0–4 by estimating the percentage of parenchyma affected by CPN, recognizing minimal, mild, moderate, and marked stages of CPN. One of the present authors (G. C. Hard) has used a much broader scale (Table 18.1) to (a) enable discrimination of differences between control and treatment groups at an early stage of CPN development in 90-day studies and (b) increase the statistical power of associating CPN grade with tumor incidence at later stages. This is aided by having separate grades for advanced CPN, including end-stage kidney, which signals imminent renal failure and death (Hard and Khan 2004). The schema is described in detail here for the purpose of supporting statements in the succeeding sections. It is not our intention to recommend its application, because the 0–4 grading system is quite adequate for general use.

TABLE 18.1.

The 8-Grade Scale for Semiquantitating CPN

Grade of Lesion Progression

Stage of CPN

0 1 2 3 4 5 6 7 8

Nil Minimal Mild Low-moderate Mid-moderate High-moderate Low-severe High-severe End-stage

Description No CPN lesions Lesions are focal Progressive increase in number of foci from minimal to moderate Foci too numerous to count Foci coalesce into areas Majority of outer parenchyma involved No, or almost no normal parenchyma remains

18.4. CHEMICALS INCREASING RTT INCIDENCE

491

18.4. CHEMICALS INCREASING RTT INCIDENCE THROUGH A MODE OF ACTION INVOLVING EXACERBATION OF CPN Evidence is emerging that certain chemicals can interact with CPN to increase the incidence of CPN-related proliferative lesions (Hard et al. 1997; Hard 2002). Seely et al. (2002) investigated the relationship between CPN severity and the occurrence of renal tumors in male Fischer rats from the NTP database. They found a slight but statistically significant increase in CPN severity in rats with RTT compared to agematched control males without tumors, suggestive of a positive correlation between these two states in untreated animals. This study also revealed that there had been a decrease in the mean incidence of RTT in control male rats from the U.S. NTP’s carcinogenicity bioassays, since the NTP-2000 diet was introduced. The postulate that chemically exacerbated CPN could increase the risk of RTT development was initially investigated by histopathological reevaluations of the U.S. NTP’s carcinogenicity bioassays on two chemicals, hydroquinone (Hard et al. 1997) and ethyl benzene (Hard 2002). With hydroquinone, reevaluation demonstrated that the compound caused exacerbation of CPN such that almost 40% of the high-dose males had end-stage renal disease compared to only 7% in the control males. All of the foci of atypical tubule hyperplasia and adenomas were observed to be arising within areas of CPN and all occurred in rats with either end-stage renal disease or the next highest grade of CPN (high severe). In addition, all of the tumors were adenomas with nearly half being incipient or marginal lesions (Hard et al. 1997). Ethyl benzene showed even more persuasive evidence of renal tumor association with advanced CPN. Exacerbation of end-stage kidney disease involved 68% of the high-dose males versus 12% of control males, while the high-dose females also showed a modest 8% increase compared to none in the controls. The tumors occurred in areas of the parenchyma involved in the CPN process similar to the situation seen with hydroquinone. Though three of the tumors were graded as carcinomas, a high proportion were small or marginal lesions, borderline between atypical hyperplasia and adenoma. In control rats with end-stage CPN, there was an equivalent incidence of renal proliferative lesions (atypical tubule hyperplasia and RTT) as in treated rats with end-stage kidney. Statistical analysis confirmed a highly significant correlation between (a) atypical tubule hyperplasia combined with RTT and (b) severity grade of CPN. Furthermore, when the tumor incidence data were adjusted for end-stage CPN, there was no statistically significant difference between control and treated groups of rats (Hard 2002). Because hormonal influences predispose male rats to CPN more than females, RTT increases linked to chemical exacerbation of CPN are more frequently observed in males than in females, although females can occasionally be prone to this association. It has been proposed that very specific criteria for renal tumor induction need to be met in order to conclude that an increase in RTT incidence has occurred solely through chemical exacerbation of CPN. First and foremost, the chemical must have been shown to exacerbate CPN to a very advanced grade of severity, involving high severe (grade 7) or end-stage kidney (grade 8), or grade 4 in the conventional 0–4

492


grading system, in comparison to control groups in a 2-year carcinogenicity study. The tumors should occur at a very low incidence representing a marginal increase only. For the most part, the tumors should be minimal grade lesions conforming to small adenomas, or lesions borderline between atypical tubule hyperplasia and adenoma. The tumors and any precursor foci of atypical hyperplasia must be located within CPN-affected parenchyma, and they should usually be observed only toward the end of two-year studies. Very importantly, careful microscopic examination of the renal parenchyma not involved in the CPN process should reveal an absence of cytotoxicity that would suggest alternative modes of action. For example, based on many in vitro studies, it has been speculated that hydroquinone might produce RTT via a free radical mechanism (Ramachandiran et al. 2002). However, in the U.S. NTP’s 2-year bioassay of this chemical, the associated renal tumors occurred only in tissue affected by CPN, with no histopathologic evidence of cytotoxicity in renal tubules unaffected by CPN (Hard et al. 1997). Recently, all chemicals or mixtures tested in the U.S. NTP’s carcinogenicity bioassay program, which had some association with an increased incidence of RTT, were placed into categories based on available mechanistic information (Lock and Hard 2004). Ethyl benzene and hydroquinone, discussed above, were the only two chemicals placed in a category of tumor induction consequent upon exacerbated CPN. However, another category included 16 chemicals that induced RTT through an unknown mechanism, but all of these were associated with CPN exacerbation. Some of the 16 chemicals are probably candidates for inclusion in the above CPNrelated category, but the studies would require careful histological reevaluation to determine presence of the necessary criteria. Subsequently, the 2-year study of quercetin (one of the 16 chemicals mentioned above) has been reevaluated, and detailed histological examination showed that the pathology met the criteria proscribed for a CPN pathway (Hard et al. 2007). Quercetin produced a modest increase in renal tubule tumors in male rats, which correlated with CPN exacerbated to endstage (grade 8) in 20% of the high-dose male rats, versus only 2% seen in the control group at 2 years. No renal tumors were present in female rats, which correlated with a lack of exacerbation of CPN. The tumors were mainly adenomas, either borderline lesions with atypical tubule hyperplasia or of a small size. The occurrence of tumors and foci of atypical tubule hyperplasia was predominantly in rats with advanced CPN and located in tissue affected by CPN. Again, there was an absence of any cellular alterations indicative of chemical toxicity in parenchyma that was not involved in the CPN process (Hard et al. 2007). Thus, it is becoming increasingly suggestive that chemical exacerbation of CPN to severe stages (grades 7 and 8) may account for the marginally increased incidence of RTT (small adenomas or lesions borderline between atypical tubule hyperplasia and adenoma) seen in male rats in some carcinogenicity bioassays. In the U.S. NTP’s database of 2-year carcinogenicity studies, the kidney is the second most frequent site for chemically associated tumor induction in male rats, involving almost exclusively, tumors of renal tubule origin (NTP 2004a). This is partly due to the number of chemicals acting through the α2u-g mode of action, but there are an equal number of marginal renal carcinogens that potentially involve only an exacerbation of CPN (Hard et al. 2007; Lock and Hard 2004).

18.5. EXAMPLES WHERE α2U-G AND CPN MODES OF ACTION MAY BE ACTING IN CONCERT

493

18.5. EXAMPLES WHERE THE α2U-G AND EXACERBATED CPN MODES OF ACTION MAY BE ACTING IN CONCERT As discussed in preceding sections, it appears that there are two processes whereby chemicals can produce a small increased incidence of RTT in male rats. These processes are not mutually exclusive, and it is likely that a combination of indirect cytotoxicity via an α2u-g mechanism and exacerbation of CPN could both contribute to a small increase in renal tumor incidence. In fact, α2u-g nephropathy and CPN are linked from an early time-point in their development. Very recently, in a histopathological survey for renal changes in 43 of the U.S. NTP’s 90-day studies that had been conducted over a 10-year period from 1991 to 2001, it was observed that all cases of hyaline droplet nephropathy likely to be due to α2u-g-binding were associated with early exacerbation of CPN (G. S. Travlos and G. C. Hard, unpublished observations). Where the 90-day study was complemented by a 2-year study of the same chemical, CPN had become further exacerbated to advanced stages. In a 1993 survey of chemicals considered to be acting through an α2u-g mode of action at that time, Hard et al. (1993) noted that linear mineralisation of the papilla had been recorded in the male rats from the 2-year studies in all cases of chemicals suspected of acting through this mechanism. Furthermore, attention was drawn to the increase in severity of CPN by 2-years with each one of these chemicals (Hard et al. 1993). The list of chemicals included d-limonene, α-methylbenzyl alcohol, dimethyl methyl phosphonate, 1,4-dichlorobenzene, isophorone, and hexachloroethane (Hard et al. 1993). Thus, there is an intimate association of α2u-g nephropathy with exacerbating spontaneous nephropathy throughout the course of α2u-g nephropathy disease progression. When comparing the U.S. NTP’s 90-day and 2-year results for d-limonene, decalin, SS IIC, and PGMBE, Doi et al. (2007) also demonstrated that the severity of CPN appeared to be somewhat more predictive of renal tumor outcome than hallmark histological markers of a α2u-g response, such as granular cast formation and linear papillary mineralization. In fact, they suggested that α2u-g nephropathy may simply contribute to a weak background tumorigenic stimulus provided by age-related CPN. Though this may have been the case with weak responders like SS IIC and PGMBE, it is most likely that a strong α2u-g response overrides a CPN mode of action, which is a weak promoter of RTT. The renal tumor response by d-limonene and decalin must certainly represent an α2u-g mode of action because these two chemicals are the most potent of the α2u-g-binding class, inducing frequent granular casts at 3 months, severe linear papillary mineralization at 2-years, and RTT induction ranging from a 22% to 30% incidence—that is, much higher tumor incidences than seen where exacerbated CPN is the sole mode of action. One example of a chemical that induces the pathological hallmarks of α2u-g nephropathy in subchronic studies, but shows little in the way of chronic markers at 2-years, is methyl-tert-butyl ether (MTBE). Studies in rats administering MTBE (usually by vapor inhalation) at various time-points of 10 days (Prescott-Mathews et al. 1997), 14 days (Robinson et al. 1990), 28 days (Bird et al. 1997), and 90 days (Lington et al. 1997; Robinson et al. 1990) all recorded (in male rats only) an

494


increase in hyaline droplet nephropathy typical of α2u-g nephropathy. One study at 10 days identified additionally a concomitant increase in proximal tubule cell proliferation, along with a significant increase in α2u-g concentration as measured by enzyme-linked immunoabsorbent assay (ELISA). There was a strong positive correlation between cell proliferation and α2u-g concentration, with MTBE exposure levels (Prescott-Mathews et al. 1997). In these various subchronic studies the α2u-g response was regarded as positive but weak. MTBE was tested for carcinogenicity in a 2-year study by vapor inhalation in Fischer 344 rats (Chun et al. 1992), although the results of this bioassay were not officially published until 1997 (Bird et al. 1997). Modest increases in RTT at the two highest doses of exposure were reported (Bird et al. 1997). Subsequently, at the request of a sponsor, one of the authors of the present chapter has reevaluated the renal histopathology from this 2-year study (Hard 2006). Though mineralization had been reported in many mid- and high-dose male rats by Chun et al (1992), on reevaluation this was found to be mostly basement membrane mineralization of proximal tubules that was a consequence of terminal disease progression and not a direct result of compound exposure. In fact, only one male rat demonstrated linear papillary mineralization typical of the long-term effects of α2u-g nephropathy. In contrast, 86% of the high-dose males and 55% of mid-dose males had end-stage CPN, compared to only 8% for the control males. Furthermore, 85% of the treatment-related adenomas occurred in rats with end-stage CPN (severity grade 8), and the remainder occurred in male rats with grade 7 CPN (i.e., very advanced CPN). So in the case of MTBE, although α2u-g nephropathy is an important early force, it appears that CPN exacerbation takes over as the main mode of action underlying later RTT development. However, a contribution from the α2u-g mechanism in the process cannot be excluded. Tert butyl alcohol (TBA), the primary metabolite of MTBE (McGregor 2006), also cannot be excluded from the short list of chemicals in which both modes of action may be operative. In the 13-week study conducted by the U.S. NTP, hyaline droplet accumulation, associated with angular crystalline structures in some droplets, was observed in male rats, but not in female rats or mice of either sex (NTP 1995). Borghoff et al. (2001) showed that the accumulating protein was immunoreactive for α2u-g (Borghoff et al. 2001). Williams and Borghoff (2001) demonstrated that TBA was capable of binding reversibly to α2u-g, thus fulfilling an important biochemical criterion for an α2u-g mode of action (Williams and Borghoff 2001). In the 2-year study, demonstration of a statistically significant increase in RTT required step-sectioning of the remaining wet tissue to produce an additional 7–8 kidney sections for examination. A tumor increase was observed in the mid-dose males, but a lower incidence in the high-dose males was probably influenced by the high early death rate inflicted mainly by end-stage CPN. Linear mineralization of the papilla was present in the majority of high-dose males and was also present in many of the mid-dose males, but to a lesser degree of severity. TBA therefore represents a case where both the α2u-g and exacerbated CPN modes of action seem likely to play a role in tumor development, with no clear distinction between the two. Methyl isobutyl ketone (MIBK), an industrial solvent, undergoes metabolism to form 4-methyl-2-pentanol, which is the sort of chemical structure that may bind to α2u-g. This suggestion is consistent with short-term inhalation studies with MIBK

18.6. RELEVANCE OF RAT A2U-GLOBULIN NEPHROPATHY AND CPN TO HUMANS

495

where hyaline droplet formation was observed in the kidneys of male rats exposed to 500 and 2000 ppm and epithelial regeneration of proximal convoluted tubules at 2000 ppm (Phillips et al. 1987). In the 2-year inhalation study, the incidence of CPN was increased at the top dose of 1800 ppm and the severity was increased at all dose levels. Demonstration of a statistically significant increase in RTT required stepsectioning, a combination of single and step sections giving a 26% increase at the top dose of 1800 ppm in male rats only (Stout et al. 2008). Linear mineralization of the papilla was present in 58% of the top dose and 44% of the mid-dose males. Thus, MIBK may also be a case where both α2u-g and exacerbated CPN modes of action play a role in tumor development. A recent short-term study with MIBK has confirmed the identity of the accumulating protein to be α2u-g (Borghoff et al. 2009).

18.6. RELEVANCE OF RAT A2U-GLOBULIN NEPHROPATHY AND CPN TO HUMANS From the discussion above, it should be clear that the induction of tumors in α2u-g nephropathy is a male-rat-specific phenomenon, which does not occur in female rats or male or female mice, or in rats where the gene for hepatic synthesis of α2u-g is absent. In addition, the chemicals that bind to α2u-g have been shown not to bind to human members of the lipocalin superfamily of proteins. Thus, it is now recognized that provided the chemical of interest meets the criteria set by the various regulatory or authoritative bodies, such as the U.S. EPA and IARC, then chemicals producing a low incidence of RTT in male rats by this mode of action should be judged as having no relevance for hazard assessment in humans. With respect to a mode of action involving exacerbation of CPN, Hard et al. (2009) have made a detailed comparison of this spontaneous rat disease with the various types of nephropathy that afflict humans. Humans are affected by several different nephropathies of known etiology, but there is no entity in humans that shows the combination or pattern of histological features that characterize CPN. In particular, CPN is not an inflammatory or vascular disease, nor does it have an immunological or autoimmune basis, and hematuria and glucosuria are not clinical findings. Relative to the various human causes of end-stage renal disease, the pattern of histological features confers uniqueness on CPN such that it can be concluded as having no strict counterpart in humans. Furthermore, no chemical that exacerbates CPN in rats is known to cause an increase in severity of any human renal disease (Hard et al. 2009). As a consequence of this reasoning, chemicals that exacerbate rat CPN in carcinogenicity bioassays, linked to a marginal but sometimes statistically significant increase in RTT incidence in treated rats, can be regarded as having no relevance for extrapolation to humans (Hard et al. 2009). Furthermore, in the view of Hard et al. (2009), because so many physiological factors influence the severity of CPN, chemically induced exacerbation of this spontaneous disease process might be regarded as an adverse event and not necessarily as an expression of chemical toxicity. In the few cases where a chemical may be judged as acting through modes of action involving both α2u-g nephropathy and exacerbation of CPN, the judgment

496


should be that any RTT increase has no relevance to humans, as would be the case for each mode of action when considered separately.

REFERENCES Ashby, J. (1996). Alpha 2 mu-globulin nephropathy in white ravens. Environ Health Perspect 104, 1264–1267. Barter, J. A., and Sherman, J. H. (1999). An evaluation of the carcinogenic hazard of 1,4-dichlorobenzene based on internationally recognized criteria. Regul Toxicol Pharmacol 29, 64–79. Barthold, S. W. (1979). Chronic progressive nephropathy in aging rats. Toxicol Pathol 7, 1–6. Baylis, C. (1994). Age-dependent glomerular damage in the rat. Dissociation between glomerular injury and both glomerular hypertension and hypertrophy. Male gender as a primary risk factor. J Clin Invest 94, 1823–1829. Bertani, T., Zoja, C., Abbate, M., Rossini, M., and Remuzzi, G. (1989). Age-related nephropathy and proteinuria in rats with intact kidneys exposed to diets with different protein content. Lab Invest 60, 196–204. Bird, M. G., Burleigh-Flayer, H. D., Chun, J. S., Douglas, J. F., Kneiss, J. J., and Andrews, L. S. (1997). Oncogenicity studies of inhaled methyl tertiary-butyl ether (MTBE) in CD-1 mice and F-344 rats. J Appl Toxicol 17 (Suppl 1), S45–S55. Borghoff, S. J., Hard, G. C., Berdasco, N. M., Gingell, R., Green, S. M., and Gulledge, W. (2009). Methyl isobutyl ketone (MIBK) induction of alpha2u-globulin nephropathy in male, but not female rats. Toxicology 258, 131–138. Borghoff, S. J., Lehman-McKeeman, L. D., Short, B. G., Hard, G. C., and Swenberg, J. A. (1993). Critique of R. Melnick’s “An alternative hypothesis on the role of chemically induced protein droplet (alpha 2u-globulin) nephropathy in renal carcinogenesis”. Regul Toxicol Pharmacol 18, 357–364. Borghoff, S. J., Prescott, J. S., Janszen, D. B., Wong, B. A., and Everitt, J. I. (2001). Alpha 2u-globulin nephropathy, renal cell proliferation, and dosimetry of inhaled tert-butyl alcohol in male and female F-344 rats. Toxicol Sci 61, 176–186. Brennan, P. A., and Kendrick, K. M. (2006). Mammalian social odours: Attraction and individual recognition. Philos Trans R Soc Lond B Biol Sci 361, 2061–2078. Charbonneau, M., Lock, E. A., Strasser, J., Cox, M. G., Turner, M. J., and Bus, J. S. (1987a). 2,2,4-Trimethylpentane-induced nephrotoxicity. I. Metabolic disposition of TMP in male and female Fischer 344 rats. Toxicol Appl Pharmacol 91, 171–181. Charbonneau, M., Short, B. G., Lock, E. A., and Swenberg, J. A. (1987b). Mechanism of petroleum induced sex-specific protein droplet nephropathy and renal cell proliferation in Fischer-344 rats: Relevance to humans. Trace Subst Environ Health 21, 263–273. Chatterjee, B., Demyan, W. F., Song, C. S., Garg, B. D., and Roy, A. K. (1989). Loss of androgenic induction of alpha 2u-globulin gene family in the liver of NIH black rats. Endocrinology 125, 1385–1388. Chaudhuri, B. N., Kleywegt, G. J., Bjorkman, J., Lehman-McKeeman, L. D., Oliver, J. D., and Jones, T. A. (1999). The structures of alpha 2u-globulin and its complex with a hyaline droplet inducer. Acta Crystallogr D Biol Crystallogr 55, 753–762. Chun, J. S., Burleigh-Flayer, H. D., and Kintigh, W. J. (1992). Methyl tertiary butyl ether: Vapor inhalation oncogenicity study in Fischer 344 rats. Bushy Run Research Center, Export, Pennsylvania. Dietrich, D. R. (1995). Alpha 2u-globulin: Species- and sex-specific protein synthesis and excretion, association with chemically induced renal toxicity and neoplasia in the male rat and relevance in human cancer risk assessment. Rev Biochem Toxicol 11, 115–180. Dietrich, D. R. (1997). Doubting nongenotoxic mechanisms of renal cancer: comparing apples and oranges in the alpha2u-globulin hypothesis. Environ Health Perspect 105, 898–902. Dietrich, D. R., and Swenberg, J. A. (1990). Lindane induces nephropathy and renal accumulation of alpha 2u-globulin in male but not in female Fischer 344 rats or male NBR rats. Toxicol Lett 53, 179–181.

REFERENCES

497

Dietrich, D. R., and Swenberg, J. A. (1991a). NCI-Black-Reiter (NBR) male rats fail to develop renal disease following exposure to agents that induce alpha-2u-globulin (alpha 2u) nephropathy. Fundam Appl Toxicol 16, 749–762. Dietrich, D. R., and Swenberg, J. A. (1991b). The presence of alpha 2u-globulin is necessary for dlimonene promotion of male rat kidney tumors. Cancer Res 51, 3512–3521. Dill, J. A., Lee, K. M., Renne, R. A., Miller, R. A., Fuciarelli, A. F., Gideon, K. M., Chan, P. C., Burka, L. T., and Roycroft, J. H. (2003). Alpha 2u-globulin nephropathy and carcinogenicity following exposure to decalin (decahydronaphthalene) in F344/N rats. Toxicol Sci 72, 223–234. Doi, A. M., Hill, G., Seely, J., Hailey, J. R., Kissling, G., and Bucher, J. R. (2007). Alpha 2u-globulin nephropathy and renal tumors in national toxicology program studies. Toxicol Pathol 35, 533–540. Doi, A. M., Roycroft, J. H., Herbert, R. A., Haseman, J. K., Hailey, J. R., Chou, B. J., Dill, J. A., Grumbein, S. L., Miller, R. A., Renne, R. A., and Bucher, J. R. (2004). Inhalation toxicology and carcinogenesis studies of propylene glycol mono-t-butyl ether in rats and mice. Toxicology 199, 1–22. Dominick, M. A., Robertson, D. G., Bleavins, M. R., Sigler, R. E., Bobrowski, W. F., and Gough, A. W. (1991). Alpha 2u-globulin nephropathy without nephrocarcinogenesis in male Wistar rats administered 1-(aminomethyl)cyclohexaneacetic acid. Toxicol Appl Pharmacol 111, 375–387. EPA (1991). Alpha2u-globulin: Association with chemically induced renal toxicity and neoplasia in the male rat. EPA/625/3-91/091F, 1–118. Goldstein, R. S., Tarloff, J. B., and Hook, J. B. (1988). Age-related nephropathy in laboratory rats. FASEB J 2, 2241–2251. Gray, J. E. (1977). Chronic progressive nephrosis in the albino rat. CRC Crit Rev Toxicol 5, 115–144. Hard, G. C. (1998). Mechanisms of chemically induced renal carcinogenesis in the laboratory rodent. Toxicol Pathol 26, 104–112. Hard, G. C. (2002). Significance of the renal effects of ethyl benzene in rodents for assessing human carcinogenic risk. Toxicol Sci 69, 30–41. Hard, G. C. (2006). Expert review of kidney histopathology in the 2-year carcinogenicity study of methyl tert butyl ether (MTBE) administered to Fischer 344 rats by vapor inhalation. Report prepared for Lyondell Chemical Company, October 10th, Houston, Texas. Hard, G. C. (2008). Some aids to histological recognition of hyaline droplet nephropathy in ninety-day toxicity studies. Toxicol Pathol 36, 1014–1017. Hard, G. C., Johnson, K. J., and Cohen, S. M. (2009). A comparison of rat chronic progressive nephropathy with human renal disease—Implications for human risk assessment. Crit Rev Toxicol 39, 332–346. Hard, G. C., and Khan, K. N. (2004). A contemporary overview of chronic progressive nephropathy in the laboratory rat, and its significance for human risk assessment. Toxicol Pathol 32, 171–180. Hard, G. C., Rodgers, I. S., Baetcke, K. P., Richards, W. L., McGaughy, R. E., and Valcovic, L. R. (1993). Hazard evaluation of chemicals that cause accumulation of alpha 2u-globulin, hyaline droplet nephropathy, and tubule neoplasia in the kidneys of male rats. Environ Health Perspect 99, 313–349. Hard, G. C., and Seely, J. C. (2005). Recommendations for the interpretation of renal tubule proliferative lesions occurring in rat kidneys with advanced chronic progressive nephropathy (CPN). Toxicol Pathol 33, 641–649. Hard, G. C., Seely, J. C., Betz, L. J., and Hayashi, S. M. (2007). Re-evaluation of the kidney tumors and renal histopathology occurring in a 2-year rat carcinogenicity bioassay of quercetin. Food Chem Toxicol 45, 600–608. Hard, G. C., and Whysner, J. (1994). Risk assessment of d-limonene: An example of male rat-specific renal tumorigens. Crit Rev Toxicol 24, 231–254. Hard, G. C., Whysner, J., English, J. C., Zang, E., and Williams, G. M. (1997). Relationship of hydroquinone-associated rat renal tumors with spontaneous chronic progressive nephropathy. Toxicol Pathol 25, 132–143. Hirokawa, K. (1975). Characterization of age-associated kidney disease in Wistar rats. Mech Ageing Dev 4, 301–316. Huff, J. (1996). Response: Alpha-2-mu-globulin nephropathy, posed mechanisms, and white ravens. Environ Health Perspect 104, 1264–1267.

498


IARC (1999). Species differences in thyroid, kidney and urinary bladder carcinogenesis. IARC Sci Publ 147, 1–225. Keenan, K. P., Coleman, J. B., McCoy, C. L., Hoe, C. M., Soper, K. A., and Laroque, P. (2000). Chronic nephropathy in ad libitum overfed Sprague–Dawley rats and its early attenuation by increasing degrees of dietary (caloric) restriction to control growth. Toxicol Pathol 28, 788–798. Kohn, M. C., and Melnick, R. L. (1999). A physiological model for ligand-induced accumulation of alpha 2u globulin in male rat kidney: Roles of protein synthesis and lysosomal degradation in the renal dosimetry of 2,4,4-trimethyl-2-pentanol. Toxicology 136, 89–105. Lehman-McKeeman, L. D., and Caudill, D. (1992a). Alpha 2u-globulin is the only member of the lipocalin protein superfamily that binds to hyaline droplet inducing agents. Toxicol Appl Pharmacol 116, 170–176. Lehman-McKeeman, L. D., and Caudill, D. (1992b). Biochemical basis for mouse resistance to hyaline droplet nephropathy: Lack of relevance of the alpha 2u-globulin protein superfamily in this male ratspecific syndrome. Toxicol Appl Pharmacol 112, 214–221. Lehman-McKeeman, L. D., and Caudill, D. (1994). d-Limonene induced hyaline droplet nephropathy in alpha 2u-globulin transgenic mice. Fundam Appl Toxicol 23, 562–568. Lehman-McKeeman, L. D., Caudill, D., Rodriguez, P. A., and Eddy, C. (1998). 2-sec-Butyl-4,5dihydrothiazole is a ligand for mouse urinary protein and rat alpha 2u-globulin: Physiological and toxicological relevance. Toxicol Appl Pharmacol 149, 32–40. Lehman-McKeeman, L. D., Rivera-Torres, M. I., and Caudill, D. (1990). Lysosomal degradation of alpha 2u-globulin and alpha 2u-globulin-xenobiotic conjugates. Toxicol Appl Pharmacol 103, 539–548. Lehman-McKeeman, L. D., Rodriguez, P. A., Takigiku, R., Caudill, D., and Fey, M. L. (1989). dLimonene-induced male rat-specific nephrotoxicity: evaluation of the association between d-limonene and alpha 2u-globulin. Toxicol Appl Pharmacol 99, 250–259. Lington, A. W., Dodd, D. E., Ridlon, S. A., Douglas, J. F., Kneiss, J. J., and Andrews, L. S. (1997). Evaluation of 13-week inhalation toxicity study on methyl t-butyl ether (MTBE) in Fischer 344 rats. J Appl Toxicol 17 (Suppl 1), S37–S44. Lock, E. A., Charbonneau, M., Strasser, J., Swenberg, J. A., and Bus, J. S. (1987). 2,2,4-Trimethylpentaneinduced nephrotoxicity. II. The reversible binding of a TMP metabolite to a renal protein fraction containing alpha 2u-globulin. Toxicol Appl Pharmacol 91, 182–192. Lock, E. A., and Hard, G. C. (2004). Chemically induced renal tubule tumors in the laboratory rat and mouse: Review of the NCI/NTP database and categorization of renal carcinogens based on mechanistic information. Crit Rev Toxicol 34, 211–299. MacInnes, J. I., Nozik, E. S., and Kurtz, D. T. (1986). Tissue-specific expression of the rat alpha 2u globulin gene family. Mol Cell Biol 6, 3563–3567. Mancini, M. A., Majumdar, D., Chatterjee, B., and Roy, A. K. (1989). Alpha 2u-globulin in modified sebaceous glands with pheromonal functions: Localization of the protein and its mRNA in preputial, meibomian, and perianal glands. J Histochem Cytochem 37, 149–157. Masoro, E. J., and Yu, B. P. (1989). Diet and nephropathy. Lab Invest 60, 165–167. McGregor, D. (2006). Methyl tertiary-butyl ether: Studies for potential human health hazards. Crit Rev Toxicol 36, 319–358. Meek, M. E., Bucher, J. R., Cohen, S. M., Dellarco, V., Hill, R. N., Lehman-McKeeman, L. D., Longfellow, D. G., Pastoor, T., Seed, J., and Patton, D. E. (2003). A framework for human relevance analysis of information on carcinogenic modes of action. Crit Rev Toxicol 33, 591–653. Melnick, R. L. (1992). An alternative hypothesis on the role of chemically induced protein droplet (alpha 2u-globulin) nephropathy in renal carcinogenesis. Regul Toxicol Pharmacol 16, 111–125. Melnick, R. L. (1993). Critique does not validate assumptions in the model on alpha 2u-globulin and renal carcinogenesis. Regul Toxicol Pharmacol 18, 365–368. Melnick, R. L., Kohn, M. C., and Huff, J. (1997). Weight of evidence versus weight of speculation to evaluate the alpha2u-globulin hypothesis. Environ Health Perspect 105, 904–906. Montgomery, C. A., and Seely, J. C. (1990). Kidney. In Pathology of the Fischer Rat: Reference and Atlas, Boorman, G. A., Eustis, S. L., Elwell, M. R., and Montogomery, C., eds., Academic Press, San Diego, pp. 127–152. Neuhaus, O. W., Flory, W., Biswas, N., and Hollerman, C. E. (1981). Urinary excretion of alpha 2 muglobulin and albumin by adult male rats following treatment with nephrotoxic agents. Nephron 28, 133–140.

REFERENCES

499

Novotny, M. V. (2003). Pheromones, binding proteins and receptor responses in rodents. Biochem Soc Trans 31, 117–122. NTP (1990). Toxicology and carcinogenesis studies of d-limonene (CAS No. 5989-27-5) in F344/N rats and B6C3F1 mice (Gavage Studies). NTP TR 347; NIH Publication No. 90-2802, 1–165. NTP (1995). NTP technical report on the toxicology and carcinogenesis studies of t-butyl alcohol (CAS No. 75-65-0) in F344/N rats and B6C3F1 mice (Drinking Water Studies). NTP TR 436; NIH Publication No. 95-3167, 1–313. NTP (2004a). Chemicals associated with site-specific tumor induction in kidney tubular cell. Accessed from http://ntp-server.niehs.nih.gov/htdocs/sites/psite_cnt.html on 18 May 2004. NTP (2004b). NTP technical report on the toxicology and carcinogenesis of Stoddard Solvent IIC (CAS No. 64742-88-7) in F344/N rats and B6C3F1 mice (Inhalation Studies). NTP TR 519; NIH Publication No. 04-4453, 1–274. NTP (2004c). NTP technical report on the toxicology and carcinogenesis studies of propylene glycol mono-t-butyl ether (CAS No. 57018-52-7) in F344/N rats and B6C3F1 mice and a toxicology study of propylene glycol mono-t-butyl ether in male NBR rats (Inhalation Studies). NTP TR 515; NIH Publication No. 04-4449, 1–306. NTP (2005). NTP technical report on the toxicology and carcinogenesis studies of decalin (CAS No. 91-17-8) in F344/N rats and B6C3F1 mice and a toxicology study of decalin in male NBR rats (Inhalation Studies). NTP TR 513; NIH Publication No. 05-4447, 1–316. Peter, C. P., Burek, J. D., and van Zwieten, M. J. (1986). Spontaneous nephropathies in rats. Toxicol Pathol 14, 91–100. Phillips, R. D., Moran, E. J., Dodd, D. E., Fowler, E. H., Kary, C. D., and O’Donoghue, J. (1987). A 14week vapor inhalation toxicity study of methyl isobutyl ketone. Fundam Appl Toxicol 9, 380–388. Prescott-Mathews, J. S., Wolf, D. C., Wong, B. A., and Borghoff, S. J. (1997). Methyl tert-butyl ether causes alpha2u-globulin nephropathy and enhanced renal cell proliferation in male Fischer-344 rats. Toxicol Appl Pharmacol 143, 301–314. Ramachandiran, S., Huang, Q., Dong, J., Lau, S. S., and Monks, T. J. (2002). Mitogen-activated protein kinases contribute to reactive oxygen species-induced cell death in renal proximal tubule epithelial cells. Chem Res Toxicol 15, 1635–1642. Rao, G. N. (2002). Diet and kidney diseases in rats. Toxicol Pathol 30, 651–656. Rao, G. N., Edmondson, J., and Elwell, M. R. (1993). Influence of dietary protein concentration on severity of nephropathy in Fischer-344 (F-344/N) rats. Toxicol Pathol 21, 353–361. Rao, G. N., Morris, R. W., and Seely, J. C. (2001). Beneficial effects of NTP-2000 diet on growth, survival, and kidney and heart diseases of Fischer 344 rats in chronic studies. Toxicol Sci 63, 245–255. Robinson, M., Bruner, R. H., and Olson, G. R. (1990). Fourteen- and ninety-day oral toxicity studies of methyl tertiary-butyl ether in Sprague–Dawley rats. J Am Coll Toxicol 9, 525–540. Roy, A. K., McMinn, D. M., and Biswas, N. M. (1975). Estrogenic inhibition of the hepatic synthesis of alpha2u globulin in the rat. Endocrinology 97, 1501–1508. Roy, A. K., Nath, T. S., Motwani, N. M., and Chatterjee, B. (1983). Age-dependent regulation of the polymorphic forms of alpha 2u-globulin. J Biol Chem 258, 10123–10127. Roy, A. K., and Neuhaus, O. W. (1967). Androgenic control of a sex-dependent protein in the rat. Nature 214, 618–620. Seely, J. C., Haseman, J. K., Nyska, A., Wolf, D. C., Everitt, J. I., and Hailey, J. R. (2002). The effect of chronic progressive nephropathy on the incidence of renal tubule cell neoplasms in control male F344 rats. Toxicol Pathol 30, 681–686. Short, B. G., Burnett, V. L., and Swenberg, J. A. (1989). Elevated proliferation of proximal tubule cells and localization of accumulated alpha 2u-globulin in F344 rats during chronic exposure to unleaded gasoline or 2,2,4-trimethylpentane. Toxicol Appl Pharmacol 101, 414–431. Stout, M. D., Herbert, R. A., Kissling, G. E., Suarez, F., Roycroft, J. H., Chhabra, R. S., and Bucher, J. R. (2008). Toxicity and carcinogenicity of methyl isobutyl ketone in F344N rats and B6C3F1 mice following 2-year inhalation exposure. Toxicology 244, 209–219. Swenberg, J. A., Short, B., Borghoff, S., Strasser, J., and Charbonneau, M. (1989). The comparative pathobiology of alpha 2u-globulin nephropathy. Toxicol Appl Pharmacol 97, 35–46. Tanaka, A., Kyokuwa, M., Mori, T., and Kawashima, S. (1995). Acceleration of renal dysfunction with ageing by the use of androgen in Wistar/Tw rats. In Vivo 9, 495–502.

500


Vandoren, G., Mertens, B., Heyns, W., Van Baelen, H., Rombauts, W., and Verhoeven, G. (1983). Different forms of alpha 2u-globulin in male and female rat urine. Eur J Biochem 134, 175–181. Wilke, A. V., Dorman, D. C., and Borghoff, S. J. (1993). Alpha2u-globulin (alpha2u) and 2,4,4-trimethyl2-pentanol (TMP-2-OH) toxicity in proximal tubule cells. The Toxicologist 13, 439. Williams, K. D., Dunnick, J., Horton, J., Greenwell, A., Eldridge, S. R., Elwell, M., and Sills, R. C. (2001). P-Nitrobenzoic acid alpha2u nephropathy in 13-week studies is not associated with renal carcinogenesis in 2-year feed studies. Toxicol Pathol 29, 507–513. Williams, T. M., and Borghoff, S. J. (2001). Characterization of tert-butyl alcohol binding to alpha2uglobulin in F-344 rats. Toxicol Sci 62, 228–235. Wolf, D. C., and Mann, P. C. (2005). Confounders in interpreting pathology for safety and risk assessment. Toxicol Appl Pharmacol 202, 302–308.

CH A P TE R

19

URINARY TRACT CALCULI AND BLADDER TUMORS Samuel M. Cohen Lora L. Arnold Shugo Suzuki

19.1.

INTRODUCTION

Urinary bladder cancer in humans has been associated with exposure to chemicals since the observation by Rehn in 1895 of an increased incidence of bladder cancer in aniline dye workers in Germany (Cohen 1998; Cohen et al. 2000; Johansson and Cohen 1997). As a result of extensive research during the past century, numerous chemicals and mixtures have been identified as causative factors for human bladder cancer, most notably cigarette smoking, which accounts for approximately one-half of the bladder cancer cases in the United States (Cohen et al. 2000; Johansson and Cohen 1997). Most of the chemicals that have been identified as human bladder carcinogens documented by the International Agency for Research on Cancer (IARC) are DNA reactive chemicals, many of which are aromatic amines, such as 2-naphthylamine, 4-aminobiphenyl, and benzidine (IARC 1987, 2008). Aromatic amines are also believed to be the major component of cigarette smoke associated with the high incidences of bladder cancer (Cohen et al. 2000). Most of the chemicals that are known to be human bladder carcinogens also induce increased incidences of bladder cancer in rodent models, although many aromatic amines increase tumors of other organs in addition to the bladder—and occasionally organs other than the bladder, such as the liver and mammary gland (IARC 1987, 2008). Rodents, most commonly rats and mice, have been extensively utilized during the past 50 years as model systems to screen for chemical carcinogens in general, including those with potential effects on the urinary bladder. As a consequence of this testing, numerous chemicals have been identified as carcinogens toward the urinary bladder in rats and/or mice, including several DNA reactive chemicals, as well as several chemicals that are not metabolically activated to reactive electrophiles and thus do not form DNA adducts (Cohen 1998; Gold et al. 2001; NTP 2008). These chemicals are classified as non-DNA-reactive (Cohen 1998).


501

502

CHAPTER 19 URINARY TRACT CALCULI AND BLADDER TUMORS

Chemicals induce bladder cancer in animal models and in humans by a limited number of modes of action (Cohen 1998). For human carcinogens, the most common mode of action is DNA reactivity leading to adduct formation, mutations, and ultimately cancer. However, in animal models, numerous bladder carcinogens are nonDNA-reactive. These chemicals are able to increase the incidence of bladder cancer in these animal models by increasing cell proliferation. Increased cell proliferation of the urothelium can be induced either by direct stimulation of cell proliferation (mitogenesis) or by inducing cytotoxicity with consequent regenerative proliferation and hyperplasia. Cytotoxicity can be induced either by the formation of a reactive, cytotoxic metabolite of the administered chemical (occasionally the parent chemical itself), which is excreted in the urine at sufficiently high concentrations to produce cytotoxicity, or by the formation of urinary solids which act as corrosive, cytotoxic agents to the urothelium, again with consequent regenerative proliferation. There has also been a suggestion that extreme modifications of urinary composition, such as urinary pH, volume, or osmolality, can also produce a cytotoxic effect. However, these extremes in urinary composition frequently are associated with the formation of urinary solids, so it remains unclear whether the urinary compositional changes themselves are actually the cytotoxic stimulus. For all of these non-DNA-reactive chemicals, the evidence strongly supports a dose–response relationship that is nonlinear and likely to involve a threshold (Cohen 1998). A nonlinear dose response and the presence of a true threshold, based on biologic and physical chemical mechanistic considerations, are best demonstrated using examples of cytotoxicity involving the formation of urinary solids (Cohen et al. 2002; IARC 1999; RBCWG 1995). Urinary solids only form when there is sufficient material in the urine to exceed the level of solubility for that agent based on its chemical and physical properties, including its interaction with other components in the urine. This is the clearest example of a true threshold, based on mechanistic considerations. This is of considerable importance since many of the substances that form urinary solids in the urine, whether in rodents or in humans, involve substances that are essential for life, such as calcium, phosphate, and cysteine, in addition to other natural and synthetic chemicals (Table 19.1). As long as the exposure does not lead to a urinary concentration that exceeds the solubility of the chemical, the substance remains in solution and does not pose a toxicological risk. Numerous chemicals have been identified in rodent models which lead to the production of urinary solids, especially in rats. Urinary solids also occur in humans (McPherson et al. 2006). In rodents, urinary solids are frequently associated not only with cytotoxicity and regenerative proliferation, but also with an increased incidence of bladder tumors when the substance is administered in a standard two-year bioassay or longer (Cohen et al. 2002; IARC 1999; RBCWG 1995). The possible risk to humans of such substances is the subject of this chapter.

19.2. DIRECT AND INDIRECT FORMATION OF URINARY SOLIDS There are essentially three types of urinary solids: precipitate (amorphous material), crystals, and calculi. The distinction between crystals and calculi, especially in

19.2. DIRECT AND INDIRECT FORMATION OF URINARY SOLIDS

TABLE 19.1.

503

Substances Leading to the Formation of Urinary Solids

Endogenous Substances (Normal Urinary Constituents) Calcium carbonate Calcium oxalate Calcium phosphate Magnesium ammonium phosphate Urates and uric acid Xanthine Cystine Hippuric acid Tyrosine Uracil Bilirubin Cholesterol Hematin Hemosiderin Vitamin C

Synthetic Chemicals and Pharmaceuticals Sulfonamides Carbonic anhydrase inhibitors HIV protease inhibitors Ampicillin and amoxicillin Radiographic media (meglumine diatrizoate) β3-Adrenoceptor agonists Glafenic acid Terephthalic acid Dimethyl terephthalate Biphenyl Melamine Fosetyl-A1 Sulfosulfuron

rodent urine, is somewhat artificial, based on size (Dominick et al. 2006). Amorphous precipitate and crystals of various types are present normally in the urine of most mammalian species, including humans (Cohen 1998; Cohen et al. 2002; Dominick et al. 2006; IARC 1999; McPherson et al. 2006; RBCWG 1995). In rodents, the most common type of crystalline material is magnesium ammonium phosphate (struvite) crystals (Cohen 1995, 1998). These are also the most common crystals in human urine, but other types of crystals commonly occur in human urine, such as calcium oxalate and calcium phosphate, in addition to several other crystals formed from substances that are normally present in the urine (McPherson et al. 2006). Formation of urinary solids, either (a) qualitatively different from those normally seen in the urine or (b) with an increase in the number or size of solids normally detected in the urine, can be produced either directly by the administered substance or indirectly (Cohen 1998) (Figure 19.1). Direct formation is defined as formation of the urinary solid by the administered substance or by one of its metabolites. An example is the formation of either (a) melamine-containing crystals and calculi in rats administered melamine (Meek et al. 2003) or (b) uracil-containing crystals and calculi in rats (Shirai et al. 1989) or mice (Sakata et al. 1988) administered uracil at high levels of the diet. Indirect formation of urinary solids occurs when administration of a substance, whether natural or synthetic, to the animal causes an increased concentration of substances that are normally present in the urine, resulting in concentrations in the urine in excess of the solubility of the naturally occurring substance and also leading to formation of urinary solids (Cohen 1998). Examples include (a) the administration of extremely high doses of sodium salts (such as saccharin, ascorbate, chloride, or bicarbonate) leading to formation of calcium phosphate-containing amorphous precipitate in the urine (Cohen 1998; IARC 1999) or (b) the administration of PPARγ or dual α/γ agonists leading to formation of calcium phosphate- and calcium

504

A


Chemical

High urinary concentration(s)

Urinary solids

Metabolite

B

Chemical

Altered urinary composition

High concentration of normal urinary constituent(s)

Alteration in endogenous intermediary metabolism

Inherited disorder or surgical procedure

Figure 19.1. Alternative processes for formation of urinary tract solids. (A) Direct formation of solids composed of chronically administered parent chemical or metabolite(s). (B) Indirect formation of urinary tract solids composed of chemicals normally present in the urine. Formation occurs because of significant alterations in urine composition secondary to altered urinary physiology, alteration of normal intermediary metabolism, or secondary to an inherited metabolic disorder (e.g., gout, oxalosis) or surgical procedure (e.g., porta caval shunt).

oxalate-containing crystals and calculi (Dominick et al. 2006). Formation of urinary solids from substances normally present in the urine can be produced not only by administering substances exogenously, but also by changes in normal intermediary metabolism. For example, hyperparathyroidism leads to increased urinary calcium levels and the formation of calcium-containing crystals and calculi (McPherson et al. 2006). Similarly, in patients with gout, associated with hyperuricemia, urate crystals increasingly form in the urine; and if present at sufficiently high levels, urate calculi can form (McPherson et al. 2006). A model of this in rodents involves the surgical formation of a portacaval shunt that produces marked alterations in uric acid metabolism, eventually leading to the formation of urate-containing crystals and calculi (Clayson et al. 1995). Numerous other examples of endogenous alterations in metabolism leading to formation of urinary tract calculi have been identified. In many of the rodent models involving formation of urinary solids, there is associated cytotoxicity, regenerative proliferation, and ultimately the induction of tumors of the urothelium, usually of the urinary bladder but occasionally of the kidney pelvis or ureters (Shirai et al. 1989). Cytotoxicity and regenerative

19.3. URINARY FACTORS INFLUENCING THE FORMATION OF URINARY SOLIDS

505

proliferation is dependent on several factors, including the amount and size of urinary solid and its surface features (rough-surfaced calculi are more abrasive than smooth calculi) (Clayson 1974; Clayson and Cooper 1970; Clayson et al. 1995). Examples have been identified where the presence of a crystal formed from an administered substance, such as sulfosulfuron, is not cytotoxic, but the calculi composed of the same substance are cytotoxic (Arnold et al. 2001). The relationship of urinary solids to bladder carcinogenesis is not associated with the actual chemical composition of the solid, but is reliant entirely on the physical properties of the solid (Clayson 1974; Clayson et al. 1995; DeSesso 1989; IARC 1999; RBCWG 1995). This has been demonstrated by implanting pellets of various substances, including paraffin wax, cholesterol, glass, stainless steel, and wood, directly into the lumen of the mouse or rat bladder (Bryan 1969; Clayson and Cooper 1970; DeSesso 1989). In a classic experiment reported by Jull (1979) involving implantation of paraffin wax pellets into the mouse bladder, he was able to demonstrate that the incidence of bladder tumors after one year was approximately 10% whereas by two years it was approximately 50% (Jull 1979). This was a critical study, because the method involving pellet implantation had been used by numerous investigators until that time as a way of directly exposing the urothelium to known substances by incorporating the substance into the pellet and then implanting the pellet into the bladder to determine a possible direct effect of the substance on the urothelium without involving metabolism (Bryan 1969; Clayson 1974; Clayson and Cooper 1970; DeSesso 1989). The experiment by Jull (1979), as well as research by others, proved that the effects on the rodent urothelium were due to the pellet itself rather than the chemical, although the speed with which the chemical could be leached from the pellet greatly affected the surface characteristics of the pellet and thereby affected the incidence of tumors being induced by the pellet (DeSesso 1989).

19.3. URINARY FACTORS INFLUENCING THE FORMATION OF URINARY SOLIDS The critical factor leading to the formation of urinary solids is the solubility of the substance in the urine. However, this can be influenced by several factors normally present in the urine which can vary considerably, not only between species but also within a given animal or human based on variations in food and water consumption, type of diet, hydration, and alterations in metabolism (Cohen 1995, 1998; McPherson et al. 2006; Pearle and Lotan 2007). The composition of urine varies considerably with diet and drinking, which results in a marked diurnal variation in the composition of the urine (Fisher et al. 1989). This has been demonstrated in rats for several parameters, but is essentially true for all and reflects the role of the urine as an excretory pathway. The relationship of this diurnal variation to food and water consumption can be demonstrated by reversing the light–dark hours in an animal room. Rodents are nocturnal in their eating pattern, so the urine composition varies based on the light and dark variations of their environment. Variations in the urine of humans also are dependent on food

506


and water consumption (Pearle and Lotan 2007). This diurnal variation in the composition of urine also results in significant biological variation of the urothelium itself on a diurnal basis, including the mitotic rate (Tiltman and Friedell 1972). Several major factors can affect the solubility of various solutes, depending on whether they are ionic or not (Cohen 1995, 1998; Pearle and Lotan 2007). A major determinant is urinary pH. For example, in rodents it appears that a urinary pH of 6.5 or greater is necessary for precipitation of calcium salts to occur (Cohen 1998; IARC 1999). In contrast, urate-containing crystals form at low, acidic pH (McPherson et al. 2006; Pearle and Lotan 2007). The formation of these crystals can be prevented by treatments that alter urinary pH. For example, in rats, urinary pH can be increased by administration of a variety of carbonates or bicarbonates, or it can be lowered by administration of ammonium chloride (Cohen 1995, 1998; Cohen et al. 2002; Dominick et al. 2006; IARC 1999; McPherson et al. 2006; RBCWG 1995). Various rodent diets also affect the urinary pH. For example, Purina 5002 diet results in a nearly neutral urinary pH whereas Altromin 1531 produces markedly alkaline urine (Cohen et al. 1994). In contrast, AIN-76A a semi-synthetic diet, which is dependent on the presence of casein as the protein source, produces markedly acidic urine (Okamura et al. 1991). By converting the protein source to albumin, the urinary pH can be increased. These factors also are pertinent to the human situation, where individuals with calcium-containing calculi are given various treatments, such as the consumption of cranberry juice, to acidify the urine (Cohen 1998; Pearle and Lotan 2007), whereas the treatment of gout, prior to the availability of allopurinol, usually was associated with administration of substances that alkalinized the urine (McPherson et al. 2006; Pearle and Lotan 2007). Obviously, a major determinant of the solubility of the substances is their concentration in the urine. If the substances contain calcium, magnesium, phosphate, or other ions, the concentration of these ions in the urine obviously will contribute most significantly to the potential for crystallization and precipitation. However, in addition to pH and concentration, several other factors can contribute to the potential of the substances to actually precipitate in the urine (Cohen 1995; Cohen et al. 2002; Dominick et al. 2006; IARC 1999; McPherson et al. 2006; Pearle and Lotan 2007; RBCWG 1995). Many of these salts are present in the urine as supersaturated solutions, predominantly because of the presence of substances in the urine which act to keep several of these ions in solution rather than precipitate. For example, citrate is a major chelating substance in the urine for calcium and magnesium salts, particularly calcium (Dominick et al. 2006; Pearle and Lotan 2007). Lowering urinary citrate levels can contribute to the potential for precipitation of calciumcontaining salts, without altering the actual levels of the calcium ion. Also, there are several proteins in the urine which bind to calcium and keep it in solution. This includes albumin, a protein referred to as Tamm–Horsfall protein (Marengo et al. 2002; Pearle and Lotan 2007), and others. Thus, merely measuring the concentration of the various substances in the urine that constitute the urinary solid does not represent the entire picture of the potential for crystallization of any given substance. Urine is a complex mixture, and the potential for solubilization versus crystallization is extremely complex. All of these factors need to be taken into account when assessing the potential for the

19.4. COLLECTION OF URINE FOR DETECTION OF URINARY SOLIDS

507

formation of urinary solids as a potential mode of action for the relationship of bladder urothelial proliferation and tumorgenicity.

19.4. COLLECTION OF URINE FOR DETECTION OF URINARY SOLIDS The method used for the collection of urine to detect urinary solids is particularly sensitive to a variety of artifacts and variations in treatment (Cohen et al. 2007). Most of all, it is essential that the animals not be fasted or go without water during the period of collection of urine. Since the excretion of the substances that are included in formation of the urinary solids is dependent on their consumption, fasting the animals changes the urine composition considerably and can lead to a condition in which the solids are no longer formed. Furthermore, urinary solids can be rapidly excreted in the urine and are not retained; so if they are not being constantly formed anew, they will not be detected. This includes urinary tract calculi. Some calculi will be small enough that they will be excreted in the urine, or dissolve with the lowering of the concentration of the solute itself. Furthermore, many of these calculi are actually quite soluble in urine, such as uracil, and rapidly solubilize in the urine. It is essential that urine collection also be performed taking into account the propensity for artifactual changes, as well as, the potential for solubilization or further crystallization of the urine while standing (Cohen et al. 2007). Thus, collecting urine overnight on ice greatly increases the propensity of solids to form in the urine with the lowering of temperature. Also, if the urine is collected and stored in the refrigerator before examination, urinary solids can be artifactually formed. Allowing time for urinary solids to sit in the urine before being examined can lead to their solubilization. The same caution pertains to urine or tissues placed in aqueous fixatives. Artifactual crystallization or solubilization can occur depending on the specific circumstances of the process. Because of the potential for these artifacts, we strongly recommend the collection of fresh void specimens from animals with immediate examination of the urinary sediment for possible presence of urinary solids. Examination of the urinary sediment by light microscopy can detect many of the usual types of crystals and calculi. However, passing urine through a Millipore filter and then examining it by scanning electron microscopy is a much more sensitive method (Cohen et al. 2007). With attached energy dispersive X-ray spectroscopy, the elemental composition of the urinary solid can also be determined. Likewise, collection of fresh void urine specimens for examination of the urine for chemical composition is also strongly recommended (Cohen et al. 2007). It is the actual concentration of the substances, not the overall amount being excreted or its ratio to some normalizing substances such as creatinine that is the critical variable when evaluating urine for the potential formation of solids. The procedures just described for urine collection for evaluation of the presence of urinary solids is in marked contrast to the way that urine is typically collected for assessment of renal function. Furthermore, we have found that examination of rodent urine with dipsticks can also lead to misleading results, particularly with respect to measurement

508


of urinary pH, protein, and the potential for hematuria (Cohen 1995; Cohen et al. 2007). We strongly recommend examination of urine immediately after collection using a microelectrode to determine pH, rather than using a dipstick. Also, for rodent urine, the Bradford method for assaying protein is preferable to biuret-based methodology (Cohen 1995).

19.5. INTERSPECIES COMPARISON OF URINE COMPOSITION Though the substances present in the urine are generally the same between species of mammals, there are marked quantitative differences as well as some critical qualitative differences. A major difference between rodent and human urine is the osmolality (Cohen 1995). Because of extremely high concentrations of ions and urea, rodent urine has a high osmolality, generally in the range of 1500–2500 mosmol. In contrast, human urine generally has an osmolality in the range of 200–300 mosmol, similar to blood. It can become considerably more dilute, but even under conditions of dehydration it rarely attains an osmolality above 600 or 700. It has theoretically been calculated that human urine, based on renal physiology, cannot attain a concentration greater than approximately 1000–1200 mosmol. Also, there are significant differences in protein concentrations between rodents and humans, particularly for male rats and male mice. Male rats and mice excrete unique proteins, α2u-globulin and mouse urinary protein (MUP), respectively, for which there are no human homologues (Hard 1995; Meek et al. 2003; Olson et al. 1990). This is particularly noteworthy in rats, since various substances can bind to the α2u-globulin and greatly affect renal and urinary tract function. These proteins lead to extremely high protein concentrations in the urine, generally in the range of several milligrams per milliliter, in contrast to human urine in which there is usually micrograms of protein per milliliter. Furthermore, both rats and mice gradually develop an aging nephropathy, which leads to increased excretion of albumin with age. This is particularly noteworthy in rats, both males and females.

19.6. URINARY SOLID CARCINOGENESIS IN RODENTS Numerous substances administered to rats and/or mice lead to formation of urinary solids with consequent cytotoxicity, regenerative proliferation, and ultimately the formation of tumors. A variety of specific mechanisms have been demonstrated for the formation of these solids (Clayson et al. 1995; Cohen 1998; IARC 1999; RBCWG 1995). Most readily understood is the direct formation of crystals and calculi composed of the administered substance or metabolite(s). Thus, dietary administration of melamine (Meek et al. 2003), uracil (Shirai et al. 1989), or sulfosulfuron (Arnold et al. 2001) at extremely high levels leads to the formation of crystals composed predominantly of these substances. If the exposure levels do not produce urinary

19.6. URINARY SOLID CARCINOGENESIS IN RODENTS

509

concentrations sufficient to lead to the formation of crystals and/or calculi composed of these substances, then urinary solids are not formed, there is no evidence of cytotoxicity, and no consequent regenerative proliferation or tumorigenicity occurs. Tumors are only produced when the administered dose is sufficient to produce the formation of solids in the urinary tract. This is based on physical chemical properties and is the mechanistic basis for a clearly defined threshold for the carcinogenicity for these substances. Additionally, it appears that these substances have to be administered for an adequate period of time to lead to a detectable incidence of tumors. Thus, short-term administration, even if there is formation of calculi and extensive proliferation, is inadequate for generating a statistically significant incidence of tumors in a standard 2-year bioassay (Cohen 1998; Shirai et al. 1989). Upon ceasing administration of the test substance, any crystals or calculi that have formed gradually are dissolved and/or excreted, and the proliferative response ends. The removal of the hyperplastic, rapidly proliferating urothelial cells appears to be primarily by apoptosis. The urothelium returns to a normal morphologic and cell kinetic state within a few weeks after the disappearance of the urinary solid (Shirai et al. 1989). A recently described example of indirect crystalluria and calculus formation has been demonstrated for the PPAR α/γ agonist, muraglitazar (Dominick et al. 2006). Administration of muraglitazar leads to the formation of urinary calcium phosphate-containing precipitate, as well as calcium and magnesium-containing crystals and calcium-containing calculi. These solids appear to be due to the inhibition by muraglitazar of citrate synthesis leading to hypocitratemia and consequent hypocitraturia. As indicated above, citrate is the major chelating substance for calcium in the urine; and with the decrease of urinary citrate, calcium- and (to some extent) magnesium-containing crystals are able to form in the urine. For reasons that are not entirely clear, precipitation of calcium-containing crystals in the urine requires a pH greater than or equal to 6.5. Co-administration of ammonium chloride in the diet with muraglitazar treatment leads to significant acidification of the urine, generally at pH less than 6.0; this nearly completely inhibits the formation of urinary solids, with complete inhibition of urinary tract cytotoxicity, regenerative proliferation, and tumorigenicity. Numerous substances such as calcium, magnesium, and phosphate, administered to rats and/or mice, have been demonstrated to lead to formation of urinary solids and are listed in Table 19.1. This table includes not only a large number of natural, essential ingredients in our diet, but also a number of substances that are formed from normal intermediary metabolism, such as carbonate, oxalate, cystine, urate, and uracil, which are present in normal urine. Numerous synthetic chemicals also produce urinary solids when administered at very high doses, including agrichemicals (such as sulfosulfuron and Fosetyl-A1), industrial chemicals (such as melamine), and pharmaceuticals (such as sulfonamides, carbonic anhydrase inhibitors, and HIV protease inhibitors). Sodium saccharin and numerous other sodium salts of moderately strong acids, such as ascorbate, glutamate, bicarbonate, aspartate, citrate, and others, administered at very high levels in the diet (≥25,000 ppm) to rats (males > females), result in the production of large amounts of an amorphous calcium phosphate-containing

510


precipitate in the urine (Cohen 1998; IARC 1999; RBCWG 1995). Calcium phosphate is an essential component for cell survival as long as it is present at concentrations that remain soluble. However, when the concentration is sufficiently high for precipitation to occur, the precipitate is cytotoxic to epithelial cells, including urothelium (IARC 1999). The conditions are appropriate in rat urine for the excretion of sufficient concentrations of calcium phosphate for precipitation to occur. The precipitate also contains protein, silicates, and mucopolysaccharides. Precipitation does not occur in mice because they have urinary concentrations of calcium and phosphate approximately 10 times less than that in rats. The precipitate also does not form in primates, including humans. Conditions leading to acidification of the urine, such as co-administration with ammonium chloride, administration in AIN76A semisynthetic diet, or administration of the acid form of the corresponding sodium salt, prevents formation of the calcium phosphate precipitate and completely inhibits the urothelial cytotoxicity, proliferation, and tumorigenicity (IARC 1999). Thus, for these sodium salts, the mechanism leading to the induction of urothelial tumors in rats not only occurs only at high doses, above a threshold, but is speciesspecific. Extensive epidemiological investigations have not shown an increased risk of bladder cancer in individuals consuming saccharin (Elcock and Morgan 1993). Consumption of ascorbate (vitamin C) is associated with a decreased risk of bladder cancer in humans (Cohen 1998; Cohen et al. 2000; Johansson and Cohen 1997). Based on these mechanistic considerations, including species-specificity and a threshold dose–response, the IARC down-classified saccharin from possibly carcinogenic to humans (Group 2B) to not classifiable as to their carcinogenicity to humans (Group 3) (IARC 1999), and the United States National Toxicology Program removed saccharin from its list of carcinogens (NTP 2000).

19.7.

EPIDEMIOLOGY

Numerous studies have examined the relationship of urinary tract solids to toxicity and to bladder cancer in humans (Burin et al. 1995; Cohen et al. 2000; La Vecchia et al. 1991; RBCWG 1995). The evidence suggests that urinary amorphous precipitate and urinary crystals of any kind are not associated with cytotoxicity or deleterious effects in humans. Crystalluria in humans is not associated with any toxicological response (McPherson et al. 2006; Pearle and Lotan 2007). In some instances it can be an indication of the propensity of the individual to form calculi from these substances, such as calcium oxalate, or occasionally it can be an indication of systemic metabolic disturbances, such as gout, oxalosis, or hypercalcemia. The relationship of calculi to human bladder cancer remains unclear (Burin et al. 1995; Cohen et al. 2000; La Vecchia et al. 1991; RBCWG 1995). Several epidemiologic studies have not found any evidence for a relationship of urinary tract calculi to human bladder cancer. However, occasional studies have found a small but statistically significant increased risk of bladder cancer in association with exposure to calculi (Burin et al. 1995). In humans, urinary tract calculi are generally not present for long periods of time, in contrast to rodents (DeSesso 1995). This is because of the normal anatomy


511

of the human lower urinary tract. Formation of calculi in humans generally leads to obstruction of the lower urinary tract, either at the kidney pelvic–ureteral junction, the site at which the ureter crosses the pelvic brim, or where the ureter enters the bladder, as well as at the urethral outlet of the bladder. When calculi are large enough to lead to obstruction, they cannot be readily excreted, leading to excruciating pain. An individual promptly consults a physician on an emergency basis, resulting in either (a) spontaneous evacuation of the calculus by hydration of the patient or (b) treatment by ultrasound or surgery. There are a few situations in which lower urinary tract calculi can be retained for substantial periods of time in humans (DeSesso 1995; Pearle and Lotan 2007). One is the presence of calculi in the kidney pelvis, frequently leading to formation of what are referred to as staghorn calculi. Other situations that can result in the prolonged presence of calculi include bladder diverticuli and the neurogenic bladder associated with paraplegia. However, in circumstances with long-standing urinary tract calculi, patients also have bacterial cystitis (Schaeffer and Schaeffer 2007). Bacterial cystitis is a known risk factor for a slightly increased risk for development of bladder cancer. Thus, it is difficult to ascertain whether the cases of bladder cancer developing in patients with long-standing urinary tract calculi are related to the calculus or whether they are related to the bacterial cystitis, which is associated with the calculi. Moreover, the types of tumors associated with bacterial cystitis and calculi, as well as with other infectious inflammatory processes in the bladder, such as schistosomiasis, frequently are squamous cell carcinomas, in contrast to the usual transitional (urothelial) cell carcinomas that occur in the bladder (Oyasu 1995). In rodents, the tumors associated with urinary tract solids are for the most part transitional (urothelial) cell tumors rather than squamous cell proliferations.

19.8.

RISK ASSESSMENT

From the above review, the overall risk assessment for humans of any chemical producing bladder cancer in rodents based on a mode of action involving the formation of urinary tract solids needs to take into account qualitative and quantitative differences between rodent species and humans and also consideration of a threshold dose response (the necessity for the presence of sufficiently high concentrations of solute to form a precipitate in the urine). Without question, whether involving substances directly or indirectly leading to the formation of calculi or other urinary tract solids, there is a nonlinear, threshold dose response that is based on the physical chemical properties of the solutes and the composition of the urine. Since the composition of human and rodent urine is strikingly different, this must be taken into account in any quantitative comparison between the species. Furthermore, the formation of urinary precipitate or crystals appears to be insufficient for the production of toxicity to the human urothelium in contrast to many instances of cytotoxicity being produced in rodents by such solids, particularly in rats. In addition, the anatomical differences between rodents and humans need to be taken into account in any risk assessment. Humans generally do not retain calculi for long periods of time

512


whereas the horizontal, quadruped rodent can retain calculi in the dome of the bladder essentially for its lifetime since the calculi do not completely obstruct the urethral outlet. Thus, in assessing potential bladder cancer risk for humans based on studies in rodents, consideration of a threshold dose response is the foremost consideration. The differences in composition of the urine, anatomic differences, and especially exposure differences between rodents and humans must be taken into account. Furthermore, the evidence for a relationship for urinary tract calculi to an increased risk of bladder cancer is relatively weak and is complicated by the usual association of bacterial cystitis with the presence of long-standing calculi. Urinary precipitate and crystals are not relevant to human carcinogenesis, in contrast to rodents. The potential for formation of urinary tract crystalluria and calculi in humans can be readily assessed in the clinical setting (McPherson et al. 2006; Pearle and Lotan 2007). This can be accomplished by routine collection of urine for urinalysis, including sediment analysis for crystals, cells, and casts. As indicated above, the presence of abnormal crystals or an increase in crystals is not sufficient to indicate urothelial toxicity in humans. It is only the presence of calculi that poses any potential for urinary toxicity. Since calculi will frequently be associated with obstruction and consequent pain, this will be a clinical observation that is readily made by the patient with corroboration by the clinician. Many substances that can lead to the production of calculi in rodents do not appear to do so in humans. Thus, sulfonamides can produce crystalluria and calculi frequently in rodents, whereas sulfonamide crystalluria is common in humans but calculi are rare (McPherson et al. 2006; Pearle and Lotan 2007). Another example is muraglitazar (Dominick et al. 2006), which leads to formation of a variety of urinary solids in rats, including calculi, but it is not associated with the formation of urinary crystalluria or calculi in humans. Overall, the presence of urinary tract solids can readily be assessed with appropriate studies of urine and with proper collection methodology, so that the mode of action can be established in the rodent model. It clearly represents a threshold phenomenon, and estimation of exposure levels for humans can be made. It appears doubtful, however, whether urinary tract calculi actually pose a cancer hazard to humans, and therefore they are not a cancer risk for humans.

REFERENCES Arnold, L. L., Cano, M., St John, M. K., Healy, C. E., and Cohen, S. M. (2001). Effect of sulfosulfuron on the urine and urothelium of male rats. Toxicol Pathol 29, 344–352. Bryan, G. T. (1969). Pellet implantation studies of carcinogenic compounds. J Natl Cancer Inst 43, 255–261. Burin, G. J., Gibb, H. J., and Hill, R. N. (1995). Human bladder cancer: Evidence for a potential irritationinduced mechanism. Food Chem Toxicol 33, 785–795. Clayson, D. B. (1974). Editorial: Bladder carcinogenesis in rats and mice: possibility of artifacts. J Natl Cancer Inst 52, 1685–1689. Clayson, D. B., and Cooper, E. H. (1970). Cancer of the urinary tract. In Advances in Cancer Research, Vol. 13, Klein, G., and Weinhouse, S., eds., Academic Press, New York, pp. 271–381.

REFERENCES

513

Clayson, D. B., Fishbein, L., and Cohen, S. M. (1995). Effects of stones and other physical factors on the induction of rodent bladder cancer. Food Chem Toxicol 33, 771–784. Cohen, S. M. (1995). Role of urinary physiology and chemistry in bladder carcinogenesis. Food Chem Toxicol 33, 715–730. Cohen, S. M. (1998). Urinary bladder carcinogenesis. Toxicol Pathol 26, 121–127. Cohen, S. M., Cano, M., Johnson, L. S., StJohn, M. K., Asamoto, M., Garland, E. M., Thyssen, J. H., Sangha, G. K., and van Goethem, D. L. (1994). Mitogenic effects of propoxur on male rat bladder urothelium. Carcinogenesis 15, 2593–2597. Cohen, S. M., Johansson, S. L., Arnold, L. L., and Lawson, T. A. (2002). Urinary tract calculi and thresholds in carcinogenesis. Food Chem Toxicol 40, 793–799. Cohen, S. M., Ohnishi, T., Clark, N. M., He, J., and Arnold, L. L. (2007). Investigations of rodent urinary bladder carcinogens: Collection, processing, and evaluation of urine and bladders. Toxicol Pathol 35, 337–347. Cohen, S. M., Shirai, T., and Steineck, G. (2000). Epidemiology and etiology of premalignant and malignant urothelial changes. Scand J Urol Nephrol Suppl, 105–115. DeSesso, J. M. (1989). Confounding factors in direct bladder exposure studies. Comments in Toxicol 3, 317–334. DeSesso, J. M. (1995). Anatomical relationships of urinary bladders compared: Their potential role in the development of bladder tumours in humans and rats. Food Chem Toxicol 33, 705–714. Dominick, M. A., White, M. R., Sanderson, T. P., Van Vleet, T., Cohen, S. M., Arnold, L. E., Cano, M., Tannehill-Gregg, S., Moehlenkamp, J. D., Waites, C. R., and Schilling, B. E. (2006). Urothelial carcinogenesis in the urinary bladder of male rats treated with muraglitazar, a PPAR alpha/gamma agonist: Evidence for urolithiasis as the inciting event in the mode of action. Toxicol Pathol 34, 903–920. Elcock, M., and Morgan, R. W. (1993). Update on artificial sweeteners and bladder cancer. Regul Toxicol Pharmacol 17, 35–43. Fisher, M. J., Sakata, T., Tibbels, T. S., Smith, R. A., Patil, K., Khachab, M., Johansson, S. L., and Cohen, S. M. (1989). Effect of sodium saccharin and calcium saccharin on urinary parameters in rats fed Prolab 3200 or AIN-76 diet. Food Chem Toxicol 27, 1–9. Gold, L. S., Manley, N. B., Slone, T. H., and Ward, J. M. (2001). Compendium of chemical carcinogens by target organ: Results of chronic bioassays in rats, mice, hamsters, dogs, and monkeys. Toxicol Pathol 29, 639–652. Hard, G. C. (1995). Species comparison of the content and composition of urinary proteins. Food Chem Toxicol 33, 731–746. IARC (1987). Overall Evaluations of Carcinogenicity: An Updating of IARC Monographs Volumes 1 to 42. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans Supplement 7, http:// monographs.iarc.fr/ENG/Monographs/suppl7/Suppl7.pdf. IARC (1999). Consensus report. In Species Differences in Thyroid, Kidney and Urinary Bladder Carcinogenesis, Capen, C. C., Dybing, E., Rice, J. M., and Wilbourn, J. D., eds., Vol. 147, International Agency for Research on Cancer, Lyon, France, pp. 5–9. IARC (2008). Some industrial and cosmetic dyes, and related exposures. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans 99, in press. Johansson, S. L., and Cohen, S. M. (1997). Epidemiology and etiology of bladder cancer. Semin Surg Oncol 13, 291–298. Jull, J. W. (1979). The effect of time on the incidence of carcinomas obtained by the implantation of paraffin wax pellets into mouse bladder. Cancer Lett 6, 21–25. La Vecchia, C., Negri, E., D’Avanzo, B., Savoldelli, R., and Franceschi, S. (1991). Genital and urinary tract diseases and bladder cancer. Cancer Res 51, 629–631. Marengo, S. R., Chen, D. H., Kaung, H. L., Resnick, M. I., and Yang, L. (2002). Decreased renal expression of the putative calcium oxalate inhibitor Tamm–Horsfall protein in the ethylene glycol rat model of calcium oxalate urolithiasis. J Urol 167, 2192–2197. McPherson, R. A., Ben-Ezra, J., and Zhao, S. (2006). Basic examination of urine. In Henry’s Clinical Diagnosis and Management by Laboratory Methods, 21st edition, McPherson, R. A., and Pincus, M. R., eds., Saunders, New York, pp. 393–425.

514


Meek, M. E., Bucher, J. R., Cohen, S. M., Dellarco, V., Hill, R. N., Lehman-McKeeman, L. D., Longfellow, D. G., Pastoor, T., Seed, J., and Patton, D. E. (2003). A framework for human relevance analysis of information on carcinogenic modes of action. Crit Rev Toxicol 33, 591–653. NTP (2000). Appendix B. Summary—Actions on the nomination of saccharin for delisting from the Report on Carcinogens. In Report on Carcinogens, 9th edition, National Toxicology Program, Washington, D.C., pp. B3–B4. NTP (2008). Long-term study reports and abstracts. National Toxicology Program, http://ntp.niehs.nih. gov/index.cfm?objectid=D16D6C59-F1F6-975E-7D23D1519B8CD7A5. Okamura, T., Garland, E. M., Masui, T., Sakata, T., St John, M., and Cohen, S. M. (1991). Lack of bladder tumor promoting activity in rats fed sodium saccharin in AIN-76A diet. Cancer Res 51, 1778–1782. Olson, M. J., Johnson, J. T., and Reidy, C. A. (1990). A comparison of male rat and human urinary proteins: Implications for human resistance to hyaline droplet nephropathy. Toxicol Appl Pharmacol 102, 524–536. Oyasu, R. (1995). Epithelial tumours of the lower urinary tract in humans and rodents. Food Chem Toxicol 33, 747–755. Pearle, M. S., and Lotan, Y. (2007). Urinary lithiasis: Etiology, epidemiology, and pathogenesis. In Campbell–Walsh Urology, Vol. 2, 9th edition, Wein, A. J., Kavoussi, L. R., Novick, A. C., Partin, A. W., and Peters, C. A., eds., Saunders, Philadelphia, pp. 1363–1392. RBCWG (1995). Urinary bladder carcinogenesis: Implications for risk assessment. Rodent Bladder Carcinogenesis Working Group. Food Chem Toxicol 33, 797–802. Sakata, T., Masui, T., St John, M., and Cohen, S. M. (1988). Uracil-induced calculi and proliferative lesions of the mouse urinary bladder. Carcinogenesis 9, 1271–1276. Schaeffer, A. J., and Schaeffer, E. M. (2007). Infections of the urinary tract. In Campbell–Walsh Urology, Vol. 1, 9th edition, Wein, A. J., Kavoussi, L. R., Novick, A. C., Partin, A. W., and Peters, C. A., eds., Saunders, Philadelphia, pp. 223–303. Shirai, T., Fukushima, S., Tagawa, Y., Okumura, M., and Ito, N. (1989). Cell proliferation induced by uracil-calculi and subsequent development of reversible papillomatosis in the rat urinary bladder. Cancer Res 49, 378–383. Tiltman, A. J., and Friedell, G. H. (1972). Effect of feeding N-(4-(5-nitro-2-furyl)-2-thiazolyl)formamide on mitotic activity of rat urinary-bladder epithelium. J Natl Cancer Inst 48, 125–129.

PART

V

METHODS FOR INFORMING CANCER RISK QUANTIFICATION

CH A P TE R

20

(Q)SAR ANALYSIS OF GENOTOXIC AND NONGENOTOXIC CARCINOGENS: A STATE-OFTHE-ART OVERVIEW Yin-tak Woo David Y. Lai

20.1.

INTRODUCTION

During the past decade, there has been an explosive growth in the interest of using qualitative as well as quantitative structure–activity relationships analysis or modeling—collectively known as (Q)SAR—in predicting the carcinogenic potential of chemicals. Despite tremendous advancement in predictive technology, carcinogenicity remains to be one of the most difficult toxicological endpoints to predict because of the complexity of its mechanisms of action and the difficulty of obtaining robust, well-balanced databases needed for effective (Q)SAR studies. Meanwhile, the scientific, industrial, and regulatory communities are under increasingly intense pressure to expand the use of (Q)SAR from the traditional research and development (R&D) tool to health and environmental protection and regulatory purposes. The user base has also substantially expanded from experienced research scientists to a broader, heterogeneous base that may include nonscientists with limited knowledge of chemical carcinogenesis. It is important to point out that proper use of (Q)SAR predictions requires basic understanding of (a) the complexity of the toxic endpoint of interest, (b) the validity and applicability of the specific (Q)SAR method for the query chemical, and (c) the limitations of the method, the uncertainty or degree of confidence of the prediction, and the need for supportive evidence. The purposes of this chapter are to provide a background document to address some of these issues along with descriptions of some practical guiding principles and structural alerts or factors for (Q)SAR users. The chapter aims to provide (a) an overview of (Q)SAR analysis or modeling, (b) basic knowledge of the essence of mechanism-based SAR exemplified by genotoxic carcinogens, nongenotoxic carcinogens, and fibers,


517

518

CHAPTER 20

(Q)SAR ANALYSIS OF GENOTOXIC AND NONGENOTOXIC CARCINOGENS

particles, and nanomaterials, (c) descriptions of possible uses of (Q)SAR and some widely used (Q)SAR methods/models, and (d) some future perspectives of (Q) SAR field.

20.2. OVERVIEW OF (Q)SAR ANALYSIS AND MODELING 20.2.1.

Types of (Q)SAR

The carcinogenic potential of chemicals may be assessed by different types of (Q) SAR analysis which include (a) qualitative or semiquantitative SAR, (b) traditional QSAR analysis, (c) formalized computerized models or softwares, and (d) other methods such as biologically based or molecular modelling methods [e.g., Rabinowitz et al. (2008)] or integrative methods. Qualitative or semiquantitative SAR may involve (a) human expert judgment or computer-assisted identification of structural features (e.g., “structural alert”) that may contribute to carcinogenicity, (b) assessment of factors that affect absorption/ distribution/metabolism/excretion (ADME), and (c) consideration of other supportive information as a basis for prediction. The predicted results may be expressed qualitatively as positive/negative or semiquantitatively in a relative scale such as low/moderate/high. For chemicals with abundant data on closely related homologues or analogues, “read across” or “trend analysis” may be conducted to project the carcinogenic potential of the query chemical by comparison to homologues or analogues. In traditional QSAR methods, the carcinogenic potential of the query chemical is predicted quantitatively (usually in TD50, the dose inducing a tumor incidence of 50% in rodents) by using mathematical equations/models that relate the carcinogenic activity of a training set of structurally related chemicals to a combination of their physicochemical properties and other molecular descriptors (e.g., topological, quantum mechanical) using various statistical methods, such as regression analysis, principal component and factor analysis, discriminant and pattern recognition analysis, and similarity analysis. A number of computerized predictive softwares or models have been developed to predict carcinogenic potential of chemicals. These include (a) knowledge rule-based expert systems that capture human expertise, (b) programs that combine human expert decisions with statistical and correlative approaches, and (c) machine learning, neuronal networking, artificial intelligence (AI), or data mining systems to identify molecular fragments of interest, discover SAR features, induce knowledge rules, and/or develop decision logic. Computational methods or approaches are also being used to develop (a) biologically based models such as 2-D or 3-D receptor modeling, docking, and ligand SAR and (b) integrative models that incorporate or combine both chemical and biological information. (Q)SAR may also be classified as (a) statistically based (relying on statistical, deterministic, or probabilistic association) or (b) mechanistically based (e.g., receptor modeling, electrophilicity-based), or they may be a combination of both. Ideally,

20.2. OVERVIEW OF (Q)SAR ANALYSIS AND MODELING

519

(Q)SAR studies should strive to achieve statistical association with mechanistic foundation. The importance of mechanistic understanding in (Q)SAR will be discussed further in the following section. For model development, (Q)SAR studies may be undertaken using structurally closely related sets of chemicals (termed congeneric or homogeneous) or structurally diverse sets of chemicals (termed global, noncongeneric, or heterogeneous). In general, predictive models using congeneric data require fewer chemicals for model development and tend to perform better (presumably because they are more likely to act by a particular mechanism) but are more limited in scope whereas global, noncongeneric models tend to be more versatile but more likely to yield false negative results if the predicted chemical is not well-represented in the training knowledge base or database. Beyond structural homogeneity, (Q)SAR studies may also be conducted on classes of chemicals with similar biological activity/function (e.g., peroxisome proliferators) to identify common factors beyond chemical structure.

20.2.2. Criteria for Assessing Validity and Scientific Soundness of (Q)SAR Most validation studies [e.g., Benigni and Bossa (2008), Benigni and Zito (2004), and Mayer et al. (2008)] tend to focus on the predictive accuracy/concordance of (Q)SAR methods. The predictive accuracy may be separated into two components: (a) sensitivity, which measures the ability to correctly detect positive chemicals (i.e., avoid false negatives), and (b) specificity, which measures the ability to correctly predict negative chemicals (i.e., avoid false positives). The receiver operating characteristic (R.O.C.) curve can graphically express both components and has often been used to compare the relative performance of different methods [e.g., Benigni and Bossa (2008)]. It should be emphasized that the predictive accuracy of any (Q) SAR method is often chemical batch-specific; that is, the accuracy shown for one batch of chemicals is not necessarily applicable to another batch of chemicals. Beyond predictive accuracy, other factors must be considered to assess the validity and scientific soundness of the (Q)SAR methods. For example, for (Q)SAR methods involving expert judgment, the factors that should be considered in assessing the scientific validity and soundness include: (a) knowledge, expertise, and predictive track record of the experts involved, (b) the scope and purpose of the (Q)SAR analysis, (c) the extent of consideration of relevant literature and supportive evidence, (d) the extent of consideration of relevant structural and mechanistic information, and (e) the articulation of the scientific basis for prediction, reasoning rationale, confidence and uncertainty, and knowledge gaps, if any. The critical requirements and pitfalls for conducting (Q)SAR studies have been the subject of many recent publications [e.g., Benigni et al. (2007), Cronin and Schultz (2003), Doull et al. (2007), Helma (2004), and Woo and Lai (2003)]. The Organization for Economic Cooperation and Development (OECD) recently conducted an international workshop at Setubal, Portugal to define the principles (now often referred to as the Setubal principles) for considering a (Q)SAR model for regulatory purposes. Essentially, the workshop panel concluded that, for regulatory

520

CHAPTER 20


uses, (Q)SARs should (a) be associated with a defined endpoint of regulatory importance, (b) take the form of an unambiguous algorithm, (c) have a defined domain of applicability, (d) be associated with appropriate measure of goodness of fit, robustness, and predictivity, and (e) preferably have a mechanistic basis (OECD 2004). It should be emphasized that, for mechanistically complex endpoints such as carcinogenicity, a sound mechanistic basis should be strongly preferred if high confidence is needed. An in-depth discussion of the criteria needed to assess the scientific soundness of (Q)SAR was the focus of an International Life Science Institute (ILSI) and U.S. Environmental Protection Agency (EPA) workgroup that dealt with issues related to (a) selection of proper endpoint or characteristics for (Q)SAR analysis, (b) knowledge/data coverage for knowledge-based/statistically based (Q)SAR models, (c) methodology, selection, and handling of molecular descriptors, (d) validation considerations, (e) transparency and rationale, (f) confidence and uncertainty, and (g) strengths, weaknesses, and limitations. The readers are referred to the workgroup report (Doull et al. 2007) for details.

20.2.3. Difficulties of (Q)SAR Modeling/Prediction of Chemical Carcinogens (Q)SAR studies, no matter how well conducted, are always subject to the limitation of a variety of inherent endpoint-specific difficulties. This is particularly true for the carcinogenicity endpoint. Some of the major difficulties include: (a) mechanistic complexity that hampers (Q)SAR analysis (see discussion below); (b) lack of wellbalanced training database required for good (Q)SAR studies because researchers tend to favor conducting studies that lead to positive findings over those that are likely to be negative; (c) high variability of long-term studies leading to inconsistent or equivocal findings that may require further studies; (d) species, strain, and gender differences complicating interpretation of data and human relevancy; (e) the commonly used quantitative parameter, TD50, does not fully take into account other important consideration such as tumor multiplicity, malignancy, and latency period; (f) complications from the use of maximum tolerated doses often used in many cancer bioassays; and (g) high cost of prospective, external validation [e.g., Benigni and Zito (2004)]. To fully interpret the outcome of (Q)SAR predictions and evaluate reliability, it would be advisable to “test drive” the (Q)SAR models with related chemicals of known activity to ensure that the model is properly trained, to examine the rationale and analogues that led to the predictions, and to peruse the original study data of the key analogs.

20.2.4.

Importance of Mechanistic Understanding

Mechanistic understanding is especially crucial for (Q)SAR analysis/modeling of chemical carcinogens because of the complexity and multistage, multifactorial process of carcinogenesis (see discussion below). The importance of mechanistic considerations in (Q)SAR analysis of chemical carcinogens has been discussed by Woo and Lai (2003). Mechanistic considerations can improve (Q)SAR study by

20.3. MECHANISM-BASED SAR ANALYSIS

521

helping to (a) select the most appropriate molecular descriptors (e.g., electrophilicity versus receptor-based), (b) serve as a criterion to assess whether the training database is suitable for making predictions on the chemicals of interest, (c) stratify the training database into smaller but mechanistically more homogenous subsets to improve predictive capability, (d) interpret outliers, (e) guide hypothesis testing to fill data gaps, and (f) assess the human significance of predictions based on animal data.

20.3. MECHANISM-BASED SAR ANALYSIS OF CHEMICAL CARCINOGENS, FIBERS, AND PARTICLES/NANOPARTICLES 20.3.1.

Basic Principles

Chemical carcinogenesis is a multistage, multifactorial process, which conceptually consists of three operational stages: initiation, promotion, and progression. Initiation involves a mutational event that may include gene mutation, chromosome aberration, translocation, and instability. Promotion involves clonal expansion of initiated cells to reach a critical mass by a variety of means such as cell proliferation, inhibition of programmed cell death, persistent chronic inflammation, inhibition of terminal differentiation, and loss of growth control. Progression may involve a second mutational event, the loss of tumor suppressor gene, impairment of immune surveillance, and acquisition of ability to metastasize. The underlying mechanisms of these three stages differ significantly; therefore, the key molecular descriptors for (Q)SAR analysis differ accordingly. To be a complete carcinogen, a chemical must be able to trigger, either directly or indirectly, activity in all three stages of the process. The relative importance of the chemical’s contribution to each of these three stages differs from chemical to chemical. Based on the predominant mechanism of action, carcinogens may be classified as genotoxic and epigenetic/nongenotoxic. Genotoxic carcinogens, also known as DNA-reactive carcinogens, generally are chemicals that directly interact with DNA either as parent chemicals or as reactive metabolites to form DNA adducts or lesions which, if unrepaired, may initiate carcinogenesis. Epigenetic carcinogens are agents that act through secondary mechanisms that do not involve direct DNA damage. Mechanism-based SAR analysis basically involves comparison of an untested chemical with structurally related compounds for which carcinogenic activity is known. Considering the most probable mechanism(s) of action, the structural features and functional properties of the untested chemical are evaluated and compared with those of the reference compounds with focus on how the differences between the untested chemical and the reference compounds may affect the potential mechanism of action. These include consideration of (a) SAR knowledge, (b) toxicokinetic and toxicodynamic parameters that may affect the delivery of biologically active intermediate to target tissue(s) for interaction with key macromolecules that may contribute to carcinogenesis, and (c) available supportive evidence. For chemicals with limited knowledge base and information, human expert judgment with delineation of rationale and possible knowledge gaps is often needed. For chemicals with

522

CHAPTER 20


abundant SAR information, structural alerts, knowledge rules, decision trees, expert systems, and predictive software may be identified or developed to facilitate the process. Depending on the mechanism of action, the approaches to (Q)SAR analysis may be totally different. In this section, the basic approaches to mechanism-based SAR analysis of (a) genotoxic carcinogens, (b) nongenotoxic carcinogens, and (c) fibers and particles are presented. Some of the approaches and knowledge have been captured in the U.S. EPA’s OncoLogic Cancer Expert System for predicting carcinogenic potential of chemicals (Woo and Lai 2005; Woo et al. 1995).

20.3.2.

SAR of Genotoxic Carcinogens

There are numerous examples of chemical carcinogens that act predominantly by genotoxic mechanisms [see Arcos et al. (1982), Woo and Lai (2003), and Woo et al. (1985b, 1988)]. Virtually all these carcinogens are either electrophiles per se (also called direct-acting genotoxic carcinogens) and can be metabolically activated to electrophilic intermediates. The following are a variety of functional groups that can directly bind covalently to DNA without requirement metabolic activation. The reasons for their direct-acting activity are also discussed along with some of the factors that may affect their ability to exert carcinogenic activity. 20.3.2.1.

Direct-Acting Electrophilic Functional Groups

a. Strained Ring Systems: Strained rings such as (a) epoxide, (b) aziridine (also known as ethyleneimine), (c) lactone, and (d) sultone (see structures below) can readily generate reactive electrophilic intermediates due to their propensity to open up the ring. Upon acidification, epoxides and aziridines can generate carbonium ions. Lactones and sultones can generate acylating intermediates and carbonium ions. The ability of the lactone and sultone rings to open up decreases with the increase in ring size due to reduction in ring strain. The introduction of a double bond to the ring may restore some of the activity especially if adjacent to the carbonyl group.

O

N

(a)

(b)

O

S O (c)

O

O

O (d)

b. Alkyl esters of Moderate and Strong Acids: Alkyl esters of moderate and strong acids such as (a) sulfate, (b) phosphate, (c) tosylate, and (d) methanesulfonate (see below) can serve as alkylating agents. The alkylating activity is dependent on the size of the alkyl group with the relative activity following the order methyl >> ethyl > propyl > butyl; beyond butyl, there is hardly any activity. For alkyl esters of dibasic (e.g., sulfate) and tribasic (e.g., phosphate) acids, the alkylating activity is completely eliminated if any one of the alkyl group is hydrolyzed (e.g., monoalkyl sulfate or dialkyl phosphate).


523

[ RO]2 SO2 [ RO]3 PO ROSO2 C6 H 4 CH3 ROSO2 CH 2 − R = alkyl (c) (a ) (d) (b) c. Haloalkanes and Substituted Haloalkanes: Haloalkanes with one halogen (other than fluorine) atom at the terminal end(s) of the alkyl chain are potential alkylating agents because the halogen is a good leaving group. In general, the alkylating activity of haloalkanes decreases with (a) the decrease in the leaving tendency of the halogen in the order, I > Br >> Cl, and (b) the increase in the size of the alkyl chain. The introduction of additional halogen(s) to the terminal carbon can also decrease the alkylating activity because the electron withdrawing activity of additional halogen(s) may hinder the departure of the first halogen. In contrast, the introduction of either (a) a heteroatom (e.g., N, S, or O), (b) double bond, or (c) aryl group at the carbon bearing the halogen can significantly increase the alkylating activity by facilitating the departure of halogen. XCH 2 − XCH 2 O− XCH 2 S− XCH 2 N− XCH 2 CH=CH− XCH 2 ArX = Cl, Br, or I d. N-Mustards and S-Mustards: N-Mustards (a) and S-mustards (b) are potent alkylating agents. The nitrogen and sulfur atom may facilitate the departure of chlorine (or bromine or iodine) by providing a resonance stabilizing mechanism through cyclization of the carbonium ion to form aziridiuim or episulfonium ion.

[ XCH 2 CH 2 ]2 N− [ XCH 2 CH 2 ]2 S X = Cl, Br, or I (b)

(a )

e. N-Nitrosamides: N-Nitrosamides, which include (a) N-nitrosoureas, (b) Nnitrososguanidines, and (c) N-nitrosourethanes (see below), can generate alkylating intermediates without metabolic activation. Thiols and alkalis can catalyze the process. The alkylating activity is dependent on the size and nature of the R group attached to nitrogen bearing the nitroso group. R′

R N CO

R″

N NO

(a)

H

NH

R

R

R=Alkyl

R'-O-CO N

N

N

O2N

NO

NO (b)

(c)

f. Aldehydes and Substituted Aldehydes: Aldehydes are highly reactive electrophiles capable of crosslinking DNA as well as reacting with protein. Owing to rapid oxidation, either chemically or metabolically, to unreactive carboxylic acids, most of the target organs of aldehydes tend to be at or close to portal of entry. However, individuals deficient with aldehyde dehydrogenase may be more susceptible to aldehydes. The reactivity of aldehydes decreases with the increase in the size of the alkyl chain. The introduction of α,β-double bond may increase the reactivity of aldehydes. O H C

O H C CH CH

524

CHAPTER 20


g. Quinones and Quinoid Compounds: Quinones and quinoid compounds (e.g., 1,4-quinone, 1,2-quinone, 1,4-quinonediminine, 1,4-quinonemethide; see structures below) are reactive electrophiles capable of reacting with –SH compounds. They may be generated during oxidation of ortho/para (but not meta) hydroquinones or aromatic diamines or hydroxytoluene compounds. The oxidation process may involve one electron oxidation that could also generate free radicals. O

O

NH

O

NH

CH2

O

O

h. Michael Addition Acceptors: The presence of a carbonyl, sulfonyl, or phosphoryl group at the α-carbon of a terminal vinyl group (see structures below) can polarize the double bond and impart partial positive charge on the terminal carbon make it electrophilic. The ability to serve as Michael Addition acceptor is highly susceptible to substitution on the vinyl group. For example, methacrylates are substantially less active than acrylates. O

O

O

CH2 CH C

CH2 CH S

CH2 CH P

O i. Arylating Agents: Aryldiazonium compounds can generate arylating agents after departure of nitrogen. Pyridine-type heteroaromatic compounds with halogen at the ortho position, as well as nitroaromatic compounds with halogen ortho/para to the nitro group(s), are also arylating agents (see below). Although the fluoro group is not a good leaving group in haloalkanes, it can be activated when ortho to ring nitrogen and aromatic nitro group(s) and can be even a better leaving group than the other halogens. + Ar N N

X N

X O2N

NO2

j. Acylating Agents and Isocyanates: Benzoyl, acyl, or carbamoyl halides (including fluoride), dihalocarbonyl compounds (e.g., phosgene), and anhydrides are potent acylating agents. Isocyanate group can react with a hydroxyl functional group to form an urethane linkage. They can all be readily hydrolyzed, are very short-lived, and are therefore mainly of portal of entry (e.g., inhalation) concern.


Ar or R

O

O

O X

N

525

O X

X

X

O

N C O

O X = F, Cl, Br, or I; R = alkyl; Ar = aryl

All the above substructures/functional groups can be considered as structural alerts (SA) of genotoxicity. The presence/attachment of one or more of these groups in a molecule is suggestive of carcinogenic potential. Whether their presence may actually impart carcinogenic activity is dependent on a variety of factors such as (a) the nature of the SA (e.g., reactivity versus stability, hard versus soft electrophile, the size of the alkyl group for alkylating agent), (b) the physicochemical properties of the molecule to which the SA is attached (e.g., impeding versus facilitating the ability of SA to reach target tissue), (c) the microenvironment surrounding the SA (e.g., steric hindrance versus resonance stabilization), (d) the exposure scenario (particularly for highly reactive SA that can be readily detoxified), and (e) dosage and frequency of the exposure. Judicious use of SA is needed to avoid oversensitivity and ensure reasonable specificity in predicting carcinogenic potential of chemicals. Depending on the specific conditions of the SA-bearing chemical, a different concern level for carcinogenic potential may be predicted. The following are some of the “rules” that may be used to modify the prediction. Direct-Acting SA Concern-Enhancing or “Boosting” Rules a. Presence of two or more functional groups with high molecular flexibility at terminal positions 2–6 atoms apart (e.g., linear alkyl chain but not cycloalkyl ring) b. Low-molecular-weight volatile compounds c. Attachment to intercalating moiety (e.g., linear 3-ring planar molecule) d. Attachment to molecules that are or resemble normal body constituents (e.g., nucleosides, amino acids) e. Attachment to molecules that contain chemical structure capable of exercising resonance stabilization (e.g., α,β-double bond) f. Attachment of additional genotoxic functional groups g. Anticipated exposure that may lead to direct access to potential targets (e.g., inhalation, parenteral injection) Direct-Acting SA Concern-Mitigating or “Busting” Rules a. Physiochemical properties indicative of negligible bioavailability at the route of exposure of concern/evaluation b. For alkyl esters of dibasic (e.g., sulfuric) and tribasic (e.g., phosphoric) acids, absence of one ester group (e.g., monoallkyl sulfate or dialkyl phosphate)

526

CHAPTER 20


c. Alkylating agents with alkyl group higher than butyl group d. Presence of substituent adjacent to the electrophilic group or bulky substituent(s) on the molecule of attachment e. Positioning of all electrophilic functional groups in the middle of the molecule with none at any terminal position f. Easily hydrolysable reactive compounds (e.g., α-haloethers, acylating agents, isocyanates) via oral route g. Easily oxidizable reactive compounds (e.g., aldehydes) via oral (dietary) route 20.3.2.2. SAR of Genotoxic Carcinogens that Require Metabolic Activation. The classical major structural classes of genotoxic carcinogens that require metabolic activation include homocyclic and heterocyclic polycyclic aromatic hydrocarbons, aromatic amines, N-nitrosamines, aflatoxin type furocoumarins, carbamates, benzene and alkenylbenzenes, and compounds with terminal double bonds. It is obvious that not all chemicals in these classes are carcinogenic. The key common features for most of the potent carcinogens in these classes are: (a) propensity to generate electrophilic intermediates, especially at or near their target organ; (b) availability of a stabilizing mechanism to allow transport of reactive intermediates from the site of activation to the site of interaction for DNA covalent binding; (c) favorable molecular size, shape, and planarity; (d) characteristics of persistent DNA adducts or chromosomal lesions; and (e) ability to act on various stages of carcinogenesis. Knowledge of the key metabolic activation pathway(s) of chemicals in a specific structural class can provide important clues and approaches to effective (Q)SAR analysis and identification of structural features that may contribute to or reduce carcinogenic potential. Some of these principles may be illustrated by the following examples. N-Nitrosamines. The vast majority of the more than 400 N-nitroso compounds that have been tested for carcinogenic activity are carcinogenic [see Arcos et al. (1982) and Lijinsky (1992)]. As a result, N-nitrosamines are often presumed or predicted to be carcinogenic without careful examination of structural features. However, it is well known that, with few exceptions, the predominant initial metabolic activation pathway for N-nitrosamines is α-hydroxylation. Since the presence of an α-hydrogen is needed for α-hydroxylation, it can be mechanistically predicted that substitution(s) that replace α-hydrogen in dialkylnitrosamines can lead to reduction or elimination of carcinogenic potential. This is consistent with the experimental findings that the relative carcinogenic potency of diethylnitrosamine is much greater than di-sec-propylnitrosamine (with fewer α-hydrogen), which, in turn, is greater than the inactive di-tert-butylnitrosamine (with no α-hydrogen). Some of the other substituents that are known or can be expected to reduce/eliminate carcinogenic potential of N-nitrosamines include: (a) acidic group, fluoro group, or any bulky/ unmetabolizable groups at the α-carbon, (b) branching of alkyl groups or bulky substituents in the vicinity of the α-carbon, and (c) large alkyl groups with total exceeding 12 carbons.

527


Polycyclic Aromatic Hydrocarbons (PAHs). Although PAHs are a wellknown structural class of carcinogens, relatively few are really potent [see Richard and Woo (1990)]. Virtually all the potent PAH carcinogens contain 4–6 rings with a bay/fjord region because metabolic activation to bay/fjord ring diol epoxide has been shown to be the key pathway. Opening of the epoxide ring generates an electrophilic carbonium ion that can be stabilized by the ring system to give it time to reach and bind to DNA. bay region 10 9 8

distorted bay region 2 3 CH3 1

fjord region 14 13

4 5 L region K region Benzo[a]pyrene

4

12

7

7

8

11

CH3

Dibenzo[a,l]pyrene

7,12-Dimethylbenz[a]anthracene

O

O

HO

OH OH

OH Reactive dihydrodiol epoxides

Other structural features that may enhance carcinogenic potential include: (a) blocking of the L-region by ring fusion to prevent detoxification and (b) a methyl group at the immediate vicinity of bay region (e.g., 12-methyl of benz[a]anthracene or 5-methyl of chrysene) to slightly distort the bay region. On the other hand, the structural features that may reduce or eliminate carcinogenic potential include (a) ring substitution at each and every bay/fjord region benzo ring(s), (b) bulky substituent(s) at virtually any ring, (c) acidic group at any ring, (d) more than four linear rings, (e) PAHs with high degree of symmetry, and (f) PAHs with less than four rings and no methyl at the L-region. Aromatic Amines. The SAR of aromatic amines has been extensively studied. The predominant activation pathway is oxidation of the amino group to generate electrophilic nitrenium ion which can be resonance-stabilized by the aryl moiety to make it stable enough to travel from site of activation to reach and bind to DNA. Molecular planarity is also favorable for carcinogenicity due to ease of DNA intercalation and binding and more accessible to metabolic activation. The critical structural features can best be illustrated by the following molecule along with reasoning:

5′

6′

6

R 4

A

X

B

4′

3

2

2′

3′

5

N R′

528

CHAPTER 20


Critical Position/Factor

Effect on Carcinogenic Potential

Brief Reasoning

a. Amino group

↑ if R/R′ = H, CH3 ↓↓ if R/R′ = t-alkyl Relative activity: 4- >> 2- ≥ 3↑ if one CH3 ↓↓ if bulky group ↑ if –, –O–, –S– ↓↓ if –CH2– CH2– ↓↓ if bulky group ↑ if –NH2 ↑ if –F ↓↓ if –COOH, –SO3H

Metabolic activation

b. Position of amino group c. 3- and 5d. Intercyclic group (X) e. 2-, 2′-, 6-, 6′f. 4′-

Resonance stabilization Flanking effect Allows conjugation Electron insulating Distorts planarity Extended conjugation Blocks detoxification ↓ absorption, ↑ excretion

The details of these SAR approaches and how these factors can be manipulated to design safer chemicals have been described (Lai et al. 1996).

20.3.3.

SAR of Nongenotoxic Carcinogens

The scientific literature on epigenetic mechanisms of chemical carcinogens has been growing at an accelerated pace in the past several years because of the increasing importance of mechanistic understanding in elucidating the molecular basis of carcinogenesis, considering human relevance of animal data, and modeling quantitative risk assessment. Epigenetic carcinogens are chemicals, which induce cancer without covalently binding to DNA or directly causing DNA damage. They may act via a variety of mechanisms such as (a) receptor-mediated cell proliferation, (b) generation of reactive oxygen species and free radicals to cause oxidative stress and secondary DNA damage, (c) perturbation of DNA methylation leading to aberrant gene expression, (d) hormonal imbalance or disturbance of homeostatic status of cells, (e) cytotoxicity with subsequent compensatory regenerative hyperplasia, (f) persistent chronic inflammation, (g) inhibition of gap junctional intercellular communication, (h) disturbance of signal transduction, (i) reduction of programmed cell death (apoptosis), (j) mitogenesis, (k) tissue/cell overload with foreign body or certain metals, (l) inhibition of microtubulin polymerization, and (m) impairment of immune surveillance (Woo and Lai 2003). As may be expected from the multiple mechanisms, (Q)SAR analysis of epigenetic carcinogens is very difficult and is dependent on the specific mechanism involved. Unlike genotoxic carcinogens that center on DNA as the common initial target molecule, the initial targets of epigenetic carcinogens may be distributed throughout the cell (e.g., nuclear receptor, cell membrane, cytoplasmic protein, organelles, etc). Some mechanisms (e.g., signal transduction) may involve complex molecular and cell biology and, therefore, may necessitate high-throughput assays or toxicogenomic studies and computational biology [e.g., Dix et al. (2007), Kavlock et al. (2008), Vinken et al. (2008)]. By far the most extensively studied mechanism of epigenetic carcinogens is receptor-mediated cell proliferation. Xenobiotic ligand-induced activation of several


529

nuclear receptors (Pascussi et al. 2008; Safe 2001) for enzyme induction (arylhydrocarbon receptor or AhR, pregnane X receptor or PXR, constitutive androstane receptor or CAR) have been shown to contribute to increased cell proliferation as the mode of carcinogenic action of a variety of hepatocarcinogens. Peroxisome proliferation activating receptor α (or PPARα)-mediated cell proliferation and oxidative stress are believed to be the mode of carcinogenic action of peroxisome proliferators. Mechanistically, since receptor binding is noncovalent and therefore reversible, (Q)SAR analysis of receptor-mediated epigenetic carcinogens essentially involves two components: (a) ability of the chemical to fit the receptor and serve as agonist/antagonist and (b) long biological half-life to allow continuous or sustained binding/activation of the receptor. Receptor fitness can be studied by either (a) analyzing the structural features of active compounds to find the favorable molecular size, shape, and thickness/planarity or (b) using a computer-assisted 2D or 3D receptor modeling or docking study. Biological half-life can be experimentally measured by looking for structural features suggestive of resistance to metabolism—for example, fluorination (stable C–F bond), ω-1 branching of fatty acids (inhibition of β-oxidation), presence/absence of two adjacent ring positions in aromatic compounds (which allows ring oxidation). These principles can be illustrated by following examples with emphasis of identification of structural features that are predictive of epigenetic carcinogenic potential. SAR Prediction of Carcinogenic Potential of TCDD-Related Compounds. It is now well-documented that the broad spectrum of toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) is mediated by the AhR (Poland et al. 1976; Safe 2001). Earlier studies showed that TCDD binds to the AhR with extremely high affinity in the picomolar range. The ligand-binding site is hydrophobic and preferentially accommodates planar nonpolar ligands with molecular dimensions approximating a 3 × 10 Å rectangle. Subsequent studies showed a wider range of ligand types including dibenzofurans, biphenyls, naphthalenes, PAHs, and indole carbazoles. The AhR is now predicted to have either a single ligand-binding pocket of 14 × 12 × 5 Å or two ligand-binding sites with one for TCDD-like compounds and the other for the larger PAH-like compounds. A 3D QSAR study provided a detailed characterization of the molecular binding domain of the AhR (Waller and McKinney 1995). From the SAR point of view, the dibenzo-p-dioxin, dibenzofuran, and planar biphenyl rings are approximately isosteric. A comparison of the available carcinogenicity data of unsubstituted dibenzo-p-dioxin and its 2,7-dichloro-, 2,3,7,8tetrachloro-, and 1,2,3,6,7,8/1,2,3,7,8,9-hexachloro- and octachloro- congeners [see Woo et al. (1985b)] suggested that the presence of chlorine substitution at all the four lateral 2,3,7,8-positions can provide TCDD with the most optimal molecular size/shape/planarity and metabolic stability for optimal carcinogenic activity. Lower chlorinated compounds are more likely to be metabolized due to the presence of more adjacent unsubstituted ring positions that favors metabolism, whereas more highly chlorinated compounds may increase the size beyond the optimal range. For the polychlorinated biphenyls (PCBs), the most carcinogenic congener was the one with all six lateral 3,3′,4,4′,5,5′-positions chlorinated. The requirement for planar

530

CHAPTER 20


structure can be observed by the loss of dioxin-like activities of PCBs with ring substitution(s) at one or more of the ortho (2,2′6,6′-) position(s) which can lead to distortion from planar structure. Subsequent AhR binding data indicated that polychlorinated dibenzofuran congeners have binding activity similar to that of their dibenzo-p-dioxin counterparts (Woo and Lai 2003). These SAR considerations have been captured in EPA’s OncoLogic Cancer Expert System for predicting carcinogenic potential of chemicals completed in 1999–2000 (Woo and Lai 2005; Woo et al. 1995). In 2006, the NTP completed the cancer bioassay of a number of new congeners of polychlorinated dibenzofurans and PCBs. The results of these studies are summarized in the table below and compared to the OncoLogic predictions. As can be seen from the table, the mechanism-based SAR approach of the OncoLogic system allowed the system to accurately predict the carcinogenic activities. Cl

O

Cl

Cl

O

Cl

2,3,7,8-TCDD

Cl

Cl

Cl

O

Cl Cl

Cl

Cl

# 521 # 520 # 525 # 529

Cl Cl

2,3,4,7,8-PCDB

2,2′,4,4′,5,5′-HCB

3,3′,4,4′,5′-PCB

NTP TR

Cl

Cl

Cl

Cl

Chemical Agent 2378-TCDD 33′44′5′-PCB 23478-PCDB 22′44′55′-HCB

Cl

Cl

Cl

NTP Bioassay Result in Female Rats Evidencea

Target Organ (s)

Incidenceb

OncoLogic Predictionc

CE CE SE EE

Liver/lung/oral Liver/lung/oral Liver/oral Liver

25/53 13/53 4/53 2/53

H HM M LM

a

CE, clear evidence; SE, some evidence; EE, equivocal evidence.

b

Highest incidence seen in any one specific target organ.

c

H, high; HM, high/medium; M, medium; LM, low/medium.

SAR of Hepatocarcinogenic Peroxisome Proliferators. Another welldocumented example of receptor involvement in epigenetic chemical carcinogenesis is PPARα-mediated cell proliferation and oxidative stress as the mode of hepatocarcinogenic action of peroxisome proliferators (Cattley et al. 1998; Gonzalez and Shah 2008; Klaunig et al. 2003; Yu et al. 2003). PPARα belongs to the steroid receptor superfamily that is involved in physiological functions such as lipid


531

metabolism and energy transfer. The physiological ligands of the receptor are quite diverse and include fatty acids. The SAR of peroxisome proliferators has been reviewed ((Bentley et al. 1993; Lai 2004; Lake and Lewis 1993; Woo and Lai 2003)). Figure 20.1 shows the chemical structures for a variety of hepatocarcinogenic peroxisome proliferatiors. However, at first glance the chemical structures appear to be diverse. From the mechanistic point of view, there are two common features: (a) Virtually all the

(1) Substituted Phenoxyacid Pharmaceuticals and Pesticides Cl

O

CH3

CH3

CH3 COOC2H5

O

CH3

Cl

Clofibrate

COOH CH3

Cl

CH3 COOH CH3

O (CH2)3 H3C

Ciprofibrate

Gemfibrozil Cl

CH3

Cl

O

CH3

COOCH3

O

COOH

CH3

H3C

CH2

COOH

H COO C

O

CF3

2,4-D

SCH2COOH

WY-14.643

Cl O

CH3

Nafenopino

Cl

N N

CH3

Methylclofenapate

Cl

H N

NO

O CH3

O C2H5

Lactofen Cl O

Cl

O

CH2

COOH

Cl

CH3

NH

CH2

CH2

O

COOH CH3

Cl

Bezafibrate

2,4,5-T

(2) Alkyl Carboxylic Acids and Precursors COOH

C2H5

O CH2 CH (CH2)3 CH3 ( CH2 )4

C2H5

Cl Cl

O

CF3 (CF2)6

CH3CH2CH2CH2CHCH2OH

COOH O

Cl

O CH2 CH (CH2)3 CH 3 C2H5

TCA

2-EH

Perfluorooctanoic Acid

DEHA

(3) Phthalate Esters C2H5

O O

CH2 CH (CH2)3 CH3

O

CH2 CH (CH2)3 CH3

O

C2H5

DEHP

Figure 20.1.

CH3

O O

O O

O

CH2

CH

(CH2)5 CH3

CH2

CH

(CH2)5

CH3

DINP

CH3

O

CH2 CH2 CH2 CH 3

O

CH2

O

BBP

Chemical structures of some major classes of peroxisome proiferators.

532

CHAPTER 20


compounds contain (either per se or after oxidation of a hydrolysis product) a hydrophilic group (in most cases, a carboxylic acid) at one end and a nonpolar moiety at the other end, and (b) most of the compounds contain structural features that are suggestive of resistance to metabolic detoxification. For example, for substituted phenoxy acid pharmaceuticals and pesticides, one of the key structural elements for peroxisome proliferative activity is chlorine substitution at the 4position of the phenyl ring. When the chlorine atom is at the 2- or 3-position of the phenyl ring, peroxisome proliferation is totally abolished. The 4-position is usually the most important position for detoxification. Chlorine substitution at the 4-position may discourage metabolic detoxification as well as exert some electronic effect. Most of the more potent compounds also have substitution at the ω-1 carbon (the carbon next to the terminal carboxylic acid). In the substituted n-alkylcarboxylic acids, most of the active compounds are also substituted at the ω-1 carbon, which renders the compound resistant to metabolism; the most effective substituent is an ethyl group. Several 2-ethylhexyl-containing esters such as di-(2-ethylhexyl) adipate (DEHA), di-(2-ethylhexyl)phthalate (DEHP), and di-(2-isononyl) phthalate (DINP) induce peroxisome proliferation because they are readily hydrolyzed to yield 2-ethylhexanol, which is further oxidized to 2-ethylhexanoic acid, an active peroxisome proliferator (Reddy and Lalwai 1983). As compared to 2-ethylhexanoic acid, the 3- and 4-isomers are virtually inactive in peroxisomal responses [cited in Lake and Lewis (1993)]. Perfluorooctanoic acid (PFOA), a rodent hepatocarcinogen with long biological half-life, is also believed to owe its activity to metabolic resistance because of the strength of the C–F bonds (Olsen et al. 2007). The human relevance of PPARα-mediated rodent hapatocarcinogenesis has been a subject of intensive debate for more than a decade. Whereas the relevancy cannot be totally disregarded, considering the totality of the evidence, most regulatory agencies are supportive of the conclusion that this mechanism is not relevant to humans (IARC 1994; Klaunig et al. 2003; Lai 2004). Nevertheless, the SAR features mentioned above can be effectively used to interpret rodent cancer data and assess the significance of human cancer risk. SAR of Rodent Thyroid Carcinogenesis. Hormonal imbalance leading to overcompensation of trophic hormone production and subsequent cell proliferation is an important mechanism of epigenetic carcinogenesis. The rodent thyroid gland is particularly susceptible to chemical carcinogens via disruption of the pituitary– thyroid axis feedback mechanism. A number of possible mechanisms (e.g., inhibition of iodide transport, perturbation of biosynthesis or catabolism or secretion of thyroid hormones) have been identified (Hill et al. 1998; Woo and Lai 2003). Of these, the most well established SAR feature is the association of thiourea/ thionamide substructure with rodent thyroid carcinogens (see structures below). In addition to these compounds, the thyroid carcinogenic activity of a few ethylenbisdithiocarbamate (e.g., mancozeb) in rats has also been attributed to the presence of ethylenethiourea as metabolite and/or contaminant. The molecular mechanism is not clearly understood but is believed to involve inhibition of thyroid hormone biosynthetic enzyme by thiourea moiety via formation of disulfide bridge. It should


533

be noted that the presence of thiourea/thioamide moiety alone may not be sufficient evidence of an epigenetic mechanism of thyroid carcinogenesis. Negative genotoxicity data or lack of genotoxic structural alerts may be needed to ascertain epigenetic mechanism. On the other hand, finding of goitrogenic/antithyroid activity may provide further support. The criteria and science policy for determining whether a thyroid carcinogen should be considered genotoxic or epigenetic and the subsequent use of appropriate quantitative risk assessment methodology have been discussed by the U.S. EPA (EPA 1998).

R

R″ N

R N

R H

O

HN

NH

S

S

N,N′-Dicyclohexylthiourea (R=H; R′=R″=C6H11) N,N'-Diethylthiourea (R=H; R′=R″=C2H5) Trimethylthiourea (R=R′=R″=CH3)

2-Thiouracil (R=H) 6-Methylthiouracil (R=CH3) 6-n-Propylthiouracil (R=C3H7)

HN

NH S

Ethylenethiourea

NH-C(S)-SH NH-C(S)-SH Ethylenebisdithiocarbamate

SAR of Alpha-2u-Nephropathy and Male Rat Kidney Tumor. A growing list of nonmutagenic chemicals and chemical mixtures of diverse structures has been found to induce kidney tubule cell tumors in male rats but not in female rats or other animal species. Mechanistic studies showed that these chemicals or their metabolites/components bind reversibly to a male rat-specific α2u-globlin (an 18,700-dalton protein synthesized under androgenic control in hepatic parenchymal cells of mature male rats of various strains [except inbred NBR strain], secreted into the blood, and excreted in urine) to form a protein complex that is resistant to hydrolytic degradation. An excessive accumulation of α2u-globulin-containing hyaline droplets in renal proximal tubules ensues. The protein overload causes cytotoxicity leading to necrosis of the tubule epithelial cells, followed by sustained regenerative cell proliferation (in the P2 segment of proximal tubules), hyperplasia, and eventual induction of renal tubule tumors. Considering the fact that α2u-nephropathy is mainly a male rat-specific mechanism, both the U.S. EPA (1991) and the International Agency for Research on Cancer (IARC) concluded that chemicals that induce renal tumors solely as a result of α2u-nephropathy should not be of human significance in risk assessment (EPA 1991; IARC 1999a). A possible useful structural alert of α2u-nephropathy-inducing carcinogen may be a small to medium-sized t-alkyl alcohol group. A variety of chemicals (isooctane in unleaded gasoline, methyl isobutyl ketone, methyl t-butyl ether, t-butyl alcohol, propylene glycol t-butyl ether) that have been shown to be α2u-nephropathy-inducing carcinogens by the U.S. NTP either are or can be metabolized to t-alkyl alcohols (see below).

534

CHAPTER 20

CH3


CH3

CH3

O

CH3

HOH3C

CH3

CH3 CH CH2 C CH3 CH3 CH CH2 C CH3 CH3 C CH3 CH3 CH O C CH3 CH3

O CH3

MIBK

Isooctane (TMP)

CH3

CH3

CH3 C CH2 C CH3 OH

CH3

2-OH-TMP

CH3

O

CH3 C CH2 C CH3

CH3 CH3 C CH3

OH Major metabolite of MIBK

CH3 PG t-Butyl ether

MTBE

OH t-Butyl alcohol

Further studies have shown that 2,2,4-trimethyl 2-pentanol (2-OH-TMP), a metabolite of 2,2,4-trimethylpentane (TMP), is capable of binding to α2u-globulin (Lock et al. 1987). The ability of 2-OH-TMP and other compounds to bind α2uglobulin has been investigated using molecular modeling. The results showed that 2-OH-TMP has the highest binding affinity to α2u globulin and that a number of structurally related compounds also have comparable binding affinity to α2u-globulin (Borghoff et al. 1991).

20.3.4.

SAR of Fibers, Particles, and Nanomaterials

Fibers, particles, and nanoparticles are physical agents that can cause so-called “foreign-body” or “solid-state carcinogenesis” (IARC 1999b; Woo et al. 1988)). A number of mineral fibers and particles including asbestos and crystalline silica have been shown to be carcinogenic or possibly carcinogenic to humans (Table 20.1). Although the mechanisms of fiber/particle carcinogenesis are still not clearly understood, studies of asbestos and other fibrous and nonfibrous particles for the last few decades have revealed a number of physical and chemical characteristics that are related to their biological activities. The toxic and carcinogenic potential of other untested fibers, particles, and nanoparticles may also be predicted using mechanismbased SAR analysis with the understanding of their mechanisms of carcinogenic/ toxic action. 20.3.4.1.

Fibers and Particles

20.3.4.1.1. Mechanisms of Fiber and Particle Carcinogenesis/Toxicity. The mechanisms leading to the development of lung cancers and/or malignant mesothelioma by exposure to fibers and particles are not clearly understood. Asbestos fibers may act as direct or indirect carcinogens. Current hypotheses of fiber-induced tumorigenesis based on cellular and molecular mechanisms include (i) the generation of free radicals that damage DNA, (ii) the physical interference of fibers with cell division, (iii) the fiber-stimulated cellular proliferative response of target cells, (iv) the fiber-induced chronic inflammatory response with release of reactive oxygen


TABLE 20.1.

535

Carcinogenicity of Fibers and Particlesa

Fibers Group 1. Carcinogenic to Humans Asbestos Erionite

Group 2B. Possibly Carcinogenic to Humans Refractory ceramic fibers Special purpose glass fibers (e.g., E-glass microfibers) Palygorskite (attapulgite), long fibers (length >5 μm)

Group 3. Not Classifiable as to their Carcinogenicity Glass wool Continuous glass filament Rock (stone) wool Slag wool Para-Aramid Wollanstonite Sepiolite Palygorskite (attapulgite), short fibers (length 20 μm expressed as weighted clearance half-time (WT1/2) and 90% clearance time (T-90) was measured in the rat short-term inhalation studies. Fiber types with a WT1/2 >50 days and a T-90 >200 days induced some degree of pulmonary pathology, while fiber types with biopersistence times below these levels induced no, or only transient, inflammation (Hesterberg et al. 1998). However, for fibers with the same dimension, those with a higher surface charge density have been shown to be more carcinogenic, suggesting that the carcinogenicity of fibers may also be related to surface properties such as surface chemistry, reactivity, and surface area of the fiber (Bonneau et al. 1986a,1986b). It has been suggested that the carcinogenicity of some fibers may be a function of the


539

aspect ratio, the dose, and other chemical properties (e.g., surface charge density) of the fiber/particle; a sufficient quantity of short, thin fibers/particles (e.g., crystalline silica) may also be carcinogenic. For particles, surface activity is the fundamental aspect of their toxicity. Surface reactivity has been shown to be most important for the toxicity/carcinogenicity of some particles such as crystalline silica and TiO2 (Clouter et al. 2001; Fubini 1997; Warheit et al. 2007; Warheit et al. 2006). Particle surface characteristics have been considered to be key factors in free radicals and reactive oxygen species (ROS) formation and in the development of fibrosis and cancer by quartz (crystallized silica). It has been shown that surface modification of silica affects its cytotoxicity, inflammation responses, and fibrogenicity (Duffin et al. 2002; Fubini 1997). Other investigators (Duffin et al. 2002; Oberdörster et al. 1994) have noted that biological effects of particles correlate better with their specific surface area than with their mass. In addition, it has been shown that lung overload of lowsurface-reactivity poorly soluble particles (PSP)(e.g., carbon black, some TiO2) at high lung burden dose expressed as surface area can also elicit inflammatory responses (Duffin et al. 2002; Oberdörster et al. 1994). The relevance of the rat lung response to particle overload for human assessment has been the subject of an international workshop. Because in some human cohorts (e.g., coal workers), very high lung burdens of poorly soluble particles have been observed, the consensus view of the workshop was that there are insufficient data to conclude that the PSP-induced tumor response in the rat model is not relevant for human hazard identification (ILSI 2000).

20.3.4.1.3. Mechanism-Based SAR Analysis of Fibers and Particles. Studies of asbestos and other fibers have shown that the dimension, durability, and dose (the three D’s) of fibrous particles are the key parameters with respect to achieving a sufficient lung burden/biopersistence and their pathogenicity. In general, fibers with a smaller diameter will penetrate deeper in the lungs while long fibers (longer than the diameter of alveolar macrophages) will only be cleared slowly. In addition to fiber length, chemical factors play an important role in fiber durability and biopersistence; fibers with high alkali or alkali earth oxide contents and low contents of Al2O3, Fe2O3, and TiO2 tend to have low durability and hence low biopersistence (Searl 1994). Based on what we know about the mechanisms and key characteristics related to fiber carcinogenicity/toxicity, it appears that those fibers that are durable (e.g., with high contents of Al2O3, Fe2O3, TiO2, or a dissolution rate constant, Kdis

Cancer Risk Assessment: Chemical Carcinogenesis, Hazard Evaluation, and Risk Quantification

Cancer Risk Assessment

Landslide Hazard and Risk

Risk Assessment and Risk Management

Handbook of Chemical Risk Assessment

Cancer Risk Evaluation: Methods and Trends

Risk Assessment

Genetic Toxicology and Cancer Risk Assessment

Risk Assessment

Risk Assessment

Risk Assessment

Risk Assessment

Risk Assessment

Risk Assessment

Risk Assessment

Risk Assessment

Risk Assessment

Risk Assessment

Immunotoxicology and Risk Assessment

Reliabilty and Risk Assessment

Toxicokinetics and Risk Assessment

Coastal and Estuarine Risk Assessment (Environmental and Ecological Risk Assessment)

Landslide risk assessment

Country risk assessment

Torchwood Risk Assessment

Forensic Assessment of Violence Risk: A Guide for Risk Assessment and Risk Management

Applications of Toxicogenomics in Safety Evaluation and Risk Assessment

Chemical Carcinogenesis

Quantification of Operational Risk under Basel II

Cancer Risk Assessment: Chemical Carcinogenesis, Hazard Evaluation, and Risk Quantification