[Arnold schecter,_thomas_a._gasiewicz]_dioxins_an(book_fi.org)

• 1. DIOXINS AND HEALTH

2. DIOXINS AND HEALTHSECOND EDITIONEdited byArnold SchecterProfessor of Environmental SciencesUniversity of Texas School of Public HealthDallas Campus, Dallas, TexasThomas A. GasiewiczProfessor and ChairDepartment of Environmental MedicineUniversity of Rochester School of MedicineRochester, New YorkA JOHN WILEY & SONS, INC., PUBLICATION 3. Copyright 6 2003 by 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 anyform 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, withouteither the prior written permission of the Publisher, or authorization through payment of theappropriate 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 tothe 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, e-mail:[email protected] of Liability/Disclaimer of Warranty: While the publisher and author have used their besteorts in preparing this book, they make no representations or warranties with respect to theaccuracy or completeness of the contents of this book and specically disclaim any impliedwarranties of merchantability or tness for a particular purpose. No warranty may be created orextended by sales representatives or written sales materials. The advice and strategies containedherein may not be suitable for your situation. You should consult with a professional whereappropriate. Neither the publisher nor author shall be liable for any loss of prot or any othercommercial damages, including but not limited to special, incidental, consequential, or otherdamages.For general information on our other products and services please contact our Customer CareDepartment within the U.S. at 877-762-2974, outside the U.S. at 317-572-3993 or fax 317-572-4002.Wiley also publishes its books in a variety of electronic formats. Some content that appears inprint, however, may not be available in electronic format.Library of Congress Cataloging-in-Publication Data:Dioxins and health / [edited by] Arnold Schecter, Thomas A. Gasiewicz 2nd ed.p. ; cm.Includes bibliographical references and index.ISBN 0-471-43355-1 (cloth : alk. paper) 1. DioxinsToxicology.[DNLM: 1. Dioxinstoxicity. WA 240 D5953 2003] I. Schecter,Arnold. II. Gasiewicz, Thomas A.RA1242.D55D58 2003615.9 0 512dc212003003817Printed in the United States of America10 9 8 7 6 5 4 3 2 1 4. To our families 5. CONTENTSContributorsixPrefacexiii 1. Overview: The Dioxin Debate1Thomas F. Webster and Barry Commoner 2. Production, Distribution, and Fate of Polychlorinated Dibenzo-p-Dioxins, Dibenzofurans, and Related Organohalogens in theEnvironment 55Roger K. Gilpin, Daniel J. Wagel, and Joseph G. Solch 3. Dioxins and Dioxinlike PCBs in Food 89James R. Startin and Martin D. Rose 4. Toxicology of Dioxins and Dioxinlike Compounds 137Jeanelle M. Martinez, Michael J. DeVito, Linda S. Birnbaum, andNigel J. Walker 5. Health Risk Characterization of Dioxins and Related Compounds159Linda S. Birnbaum and William H. Farland 6. Pharmacokinetics of Dioxins and Related Chemicals191James R. Olson 7. DoseResponse Modeling for 2,3,7,8-Tetrachlorodibenzo-p-Dioxin 247Michael J. DeVito, Amy Kim, Nigel J. Walker, Fred Parham, andChristopher Portier 8. Immunotoxicology of Dioxins and Related Chemicals299Nancy I. Kerkvliet 9. Developmental and Reproductive Toxicity of Dioxins and RelatedChemicals329H. Michael Theobald, Gary L. Kimmel, and Richard E. Peterson vii 6. viiiCONTENTS10. Eects of Polychlorinated Biphenyls on Neuronal Signaling433 Richard F. Seegal11. Experimental Toxicology: Carcinogenesis457 Justin G. Teeguarden and Nigel J. Walker12. Ah Receptor: Involvement in Toxic Responses491 Thomas A. Gasiewicz and Sang-ki Park13. Biochemical Responses to Dioxins: Which Genes? WhichEndpoints? 533 J. Kevin Kerzee, Ying Xia, and Alvaro Puga14. Evolutionary and Physiological Perspectives on Ah ReceptorFunction and Dioxin Toxicity 559 Mark E. Hahn15. Dioxin Toxicity and Aryl Hydrocarbon Receptor Signaling in Fish603 Robert L. Tanguay, Eric A. Andreasen, Mary K. Walker, and Richard E. Peterson16. Exposure Assessment: Measurement of Dioxins and RelatedChemicals in Human Tissues 629 Arnold Schecter, Olaf Papke, Marian Pavuk, and Rachel E. Tobey17. Human Health Eects of Polychlorinated Biphenyls 679 Matthew P. Longnecker, Susan A. Korrick, and Kirsten B. Moysich18. Epidemiological Studies on Cancer and Exposure to Dioxins andRelated Compounds729 Lennart Hardell, Mikael Eriksson, Olav Axelson, and Dieter Flesch-Janys19. Reproductive and Developmental Epidemiology of Dioxins 765 Sherry G. Selevan, Anne Sweeney, and Marie Haring Sweeney20. Health Consequences of the Seveso, Italy, Accident 827 Pier Alberto Bertazzi and Alessandro di Domenico21. The Yusho Rice Oil Poisoning Incident855 Yoshito Masuda22. The Yucheng Rice Oil Poisoning Incident893 Yueliang Leon Guo, Mei-Lin Yu, and Chen-Chin HsuIndex921 7. CONTRIBUTORSEric A. Andreasen, Department of Environmental and Molecular Toxicology, Oregon State University, Corvallis, OR 97331Olav Axelson, Department of Health and Environment, Linkoping University,581 85 Linkoping, SwedenPier Alberto Bertazzi, Department of Occupational and Environmental` Health, Universita degli Studi, 20122 Milan, ItalyLinda S. Birnbaum, National Health and Evironmental Eects Research Lab- oratory, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711Barry Commoner, Center for the Biology of Natural Systems, Queens College,CUNY, Flushing, NY 11367Michael J. DeVito, National Health and Evironmental Eects Research Lab-oratory, U.S. Environmental Protection Agency, Research Triangle Park,NC 27711Alessandro di Domenico, Laboratory of Comparative Toxicology and Ecotox- `,icology, Istituto Superiore di Sanita 00161 Rome, ItalyMikael Eriksson, Department of Oncology, University Hospital, SE-221 85Lund, SwedenWilliam H. Farland, Oce of Research and Development, U.S. EnvironmentalProtection Agency, Washington, DC 20460Dieter Flesch-Janys, Working Group Epidemiology, Institute of Mathematicsand Computational Sciences in Medicine, Winterhuder Weg 29, 22085Hamburg, GermanyThomas A. Gasiewicz, Department of Environmental Medicine, University ofRochester School of Medicine, Rochester, NY 14642Roger K. Gilpin, Brehm Laboratories, Wright State University, Dayton,OH 45435Yueliang Leon Guo, Institute of Environmental and Occupational Health,National Cheng Kung University Medical College, Tainan 70428, Taiwanix 8. xCONTRIBUTORSMark E. Hahn, Biology Department, Woods Hole Oceanographic Institution, Woods Hole, MA 02543Lennart Hardell, Department of Oncology, University Hospital, SE-701 85 Orebro, and Department of Natural Sciences, Orebro University, SE 701 82O rebro, SwedenChen-Chin Hsu, Department of Psychiatry, En Chu Kong Hospital, Taipei,TaiwanNancy I. Kerkvliet, Department of Environmental and Molecular Toxicology,Oregon State University, Corvallis, OR 97331J. Kevin Kerzee, Department of Environmental Health, University of Cincin- nati Medical Center, Cincinnati, OH 45267-00567Amy Kim, Curriculum in Toxicology, University of North CarolinaChapelHill, Chapel Hill, NC 27599Gary L. Kimmel, U.S. Environmental Protection Agency, Washington, DC20460Susan A. Korrick, Department of Medicine, Brigham and Womens Hospital,Harvard Medical School, Boston, MA 02115Matthew P. Longnecker, Epidemiology Branch, National Institute of Envi- ronmental Health Sciences, Research Triangle Park, NC 27709Jeanelle M. Martinez, Environmental Toxicology Program, National Instituteof Environmental Health Sciences, Research Triangle Park, NC 27709Yoshito Masuda, Daiichi College of Pharmaceutical Sciences, 22-1 Tamagawa-cho, Minami-ku, Fukuoka 815, JapanKirsten B. Moysich, Department of Cancer Prevention, Epidemiology, and Biostatistics, Roswell Park Cancer Institute, Bualo, NY 14263James R. Olson, Department of Pharmacology and Toxicology, University atBualo, SUNY, Bualo, NY 14214-3000Olaf Papke, ERGO, Forschungsgesellschaft mbH, Geierstrasse 1, D 22305Hamburg, GermanyFred Parham, Environmental Toxicology Program, National Institute of Environmental Health Sciences, Research Triangle Park, NC 27709Sang-ki Park, Department of Environmental Medicine, University of Roches-ter School of Medicine, Rochester, NY 14642Marian Pavuk, University of Texas School of Public Health, Dallas Campus, Dallas, TX 75235-9128Richard E. Peterson, School of Pharmacy and Environmental ToxicologyCenter, University of Wisconsin, Madison, WI 53705-2222 9. CONTRIBUTORS xiChristopher Portier, Environmental Toxicology Program, National Instituteof Environmental Health Sciences, Research Triangle Park, NC 27709Alvaro Puga, Department of Environmental Health, University of CincinnatiMedical Center, Cincinnati, OH 45267-00567Martin D. Rose, Central Science Laboratory, Sand Hutton, York YO41 1LZ, UKArnold Schecter, University of Texas School of Public Health, Dallas Campus,Dallas, TX 75390-9128Richard F. Seegal, New York State Department of Health and School ofPublic Health, University at Albany, SUNY, Albany, NY 12201-0509Sherry G. Selevan, U.S. Environmental Protection Agency, Washington, DC20460Joseph G. Solch, Brehm Research Laboratories, Wright State University,Dayton, OH 45435James R. Startin, Central Science Laboratory, Sand Hutton, York YO41 1LZ,UKAnne Sweeney, University of TexasHouston School of Public Health,Houston, TX 77030Marie Haring Sweeney, National Institute for Occupational Safety and Health, Cincinnati, OH 45226Robert L. Tanguay, Department of Environmental and Molecular Toxicology,Oregon State University, Corvallis, OR 97331Justin G. Teeguarden, Environ International, Ruston, LA 71270H. Michael Theobald, School of Pharmacy, University of Wisconsin, Madison,WI 53705Rachel E. Tobey, University of Texas School of Public Health, Dallas Cam-pus, Dallas, TX 75390-9128Daniel J. Wagel, Brehm Research Laboratories, Wright State University,Dayton, OH 45435Mary K. Walker, College of Pharmacy, Health Sciences Center, University of New Mexico, Albuquerque, NM 87131Nigel J. Walker, Environmental Toxicology Program, National Institute ofEnvironmental Health Sciences, Research Triangle Park, NC 27709Thomas F. Webster, Department of Environmental Health, Boston UniversitySchool of Public Health, Boston, MA 02118-2526 10. xii CONTRIBUTORSYing Xia, Department of Environmental Health, University of CincinnatiMedical Center, Cincinnati, OH 45267-00567Mei-Lin Yu, Department of Public Health, National Cheng Kung University Medical College, Tainan 70428, Taiwan 11. PREFACEIt is sometimes said that there are at least two sides to every story, andthat the truth is often somewhere in between. In like manner, the literature onthe dioxins is most often dichotomized by interests either directly relevantto human health risks or very focused on the molecular mechanisms bywhich these chemicals act to aect cellular functions. How we bring all of thisinformation together to actually determine, and not estimate, the true risksto human (and wildlife) populations exposed to these chemicals continues toremain a challenge. There is also a literature on the chemical aspects of dioxinsand related synthetic chemicals, and historical events where dioxin contamina-tion was noteworthy and of concern. On the social side there is extensiveliterature on legal aspects of environmental pollutants and on response of in-dividuals and organizations to incidents posing perceived or actual physical orpsychological risks.During the years since the rst edition of this book in 1994, an extensiveamount of new policy and scientic literature has been published related todioxins and dioxin-like chemicals. For the purposes of this text these chemicalsinclude the halogenated dioxins and dibenzofurans, certain polychlorinatedbiphenyls (PCBs), and other compounds that are structurally and toxicologi-cally similar to 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD, or justTCDD), the most extensively studied and most potent member of this group.Much of this literature has continued to rene our understanding of the eectsof human exposures, possible eects on human and wildlife populations, doseeect relationships, and the mechanisms whereby these chemicals work at themolecular level, in particular as mediated by a transcription factor the arylhydrocarbon receptor (AhR). In several cases, and in part due to new techno-logical advances, substantial new and striking strides have been made in ourability to detect sensitive endpoints of toxicity, to measure, evaluate, and pre-dict with greater accuracy eects of low exposure levels, to examine on a cell-and tissue-specic basis the genes that may be altered following exposure,and to begin to understand the normal function of the AhR. Thus, it appearsthat the scientic community is at the brink of nally making some connectionsbetween the molecular actions of TCDD and those biological and toxic eectsin animals and humans. Real issues such as What concentrations are toxic tohumans?, What eects are more likely to be observed and at what doses?, andWhat subpopulations might be most sensitive?, are beginning to be addressed.This text was written to foster further these connections by oering a perspec- xiii 12. xivPREFACEtive as to how recent scientic data may relate to very relevant issues in humanand environmental health. The chapter authors are scientists with international reputation in their par-ticular area of the dioxin arena. Many of these individuals have been responsi-ble for generating the original scientic data that they discuss. They have beenasked to address their chapters to an audience of well-educated and intelligentlay persons and professionals who may not necessarily be familiar with thedetails of the dierent specialties. As such, the authors have included high-lights of their respective elds with a sampling, without being encyclopedic, ofimportant references. Furthermore, each author has been asked to discuss therelevance of these scientic data to possible human exposures and health risks.Through this approach, the book gives a meaningful and accessible presenta-tion to the broadest range of health professionals and nonhealth professionalsinterested in dioxins, as well as policy makers and the general public. Since the middle of the 20th century, dioxins have been widespread andpersistent synthetic environmental contaminants in the United States and otherindustrial counties. Because of their ubiquity, persistence, and extreme toxicityin laboratory animals, considerable concern arose regarding their presence inthe food chain and in human tissue. The fact that these chemicals have beenshown to cause cancer, immune system disorders, reproductive and devel-opmental abnormalities, neurological and endocrine system alterations in lab-oratory animals at very low doses has further fueled this concern. While thereexists controversy regarding extrapolation of laboratory animal data to humanrisks, more and more data is available that may allow us within the next fewyears to draw better conclusions about the actual risks to human populations. In their introductory overview of the continuing dioxin debate, ThomasWebster and Barry Commoner provide a summary of some of the major cur-rent dioxin controversies. Chapters 2 and 3, by Roger Gilpin, Daniel Wagel,and Joseph Solch, and James Startin and Martin Rose, respectively, provide aperspective on the sources of the dioxins, how they move in the environment,and how and why humans are exposed predominantly by dietary intake. In separate chapters, several scientists present our current understanding ofthe information we have obtained, mainly from studies of experimental ani-mals, for the eects of these chemicals on certain health outcomes. Chapter 4,by Jeanelle Martinez, Michael DeVito, Linda Birnbaum, and Nigel Walker,gives a brief overview of the toxicology of the dioxins and the predominantissues that bench scientists are attempting to address. More detailed considerations of the subdisciplines are presented in subse-quent chapters. Chapter 5, by Linda Birnbaum and William Farland, summa-rizes the approaches used to consider the available animal and human datafor possible human health risks. Here, while the good news is presented thathuman body burdens of these chemicals are, in general, declining in moreindustrialized countries, an analysis of the toxicity data may suggest that thesebody burdens are still at or near concentrations where some eects might beexpected to occur. In Chapter 6, James Olson summarizes what is known about 13. PREFACE xvthe fate of these chemicals in animal and human tissues. These data are partic-ularly important for dening body burdens at dierent life stages and exposureconditions. One of the major issues that remains to be better characterized isthe determination of what body burdens cause what toxic eects. The discus-sion by Michael DeVito, Amy Kim, Nigel Walker, Fred Parham, and ChrisPortier in Chapter 7 indicates that most responses to these chemicals do nothave the same doseresponse relationships and clearly some responses are moresensitive than others. Notably this conclusion is mirrored in a later chapterwhere gene responses are discussed. Chapters 8 through 11, by Nancy Ker-kvliet; Michael Theobald, Gary Kimmel, and Richard Peterson; Richard See-gal; Justin Teeguarden and Nigel Walker, respectively, discuss in greater detailsome of the most consistent and sensitive toxic responses to these chemicals inthe immune system, on developing and reproductive tissues, in the nervoussystem, and for carcinogenesis. Here, much of the focus is on dening the cel-lular and biochemical alterations that lead to these responses. Chapters 12, 13, and 14, by Tom Gasiewicz and Sang-ki Park; KevinKerzee, Ying Xia, and Alvaro Puga; and Mark Hahn, respectively, are newchapters in this edition. These have been written to summarize the most recentdata at the molecular level, indicating that the Ah receptor and its ability tomodulate the expression of genes is ultimately responsible for the toxic eectsobserved. In particular, these studies open avenues to explore the possibility ofdeveloping molecular biomarkers of susceptibility and exposure. These chap-ters also present particularly exciting ndings concerning possible normalfunctions of the receptor, our understanding of which could add much to de-termining how and at what concentrations the dioxins may be acting to causetoxicity, and why these chemicals elicit such tissue- and species-specic eects.As Robert Tanguay, Eric Andreasen, Mary Walker, and Richard Petersonindicate in Chapter 15, sh have been found to be particularly sensitive to theeects of these chemicals. The newest models using Zebra sh may also be ex-tremely useful in dissecting relationships between molecular actions and eectsof the dioxins on several physiological systems. Since one cannot purposefully dose humans with the dioxins, it is more dif-cult to come to conclusions on health consequences from studies on people.Yet, epidemiology, the study of human populations and health outcomes, hasthe advantage of dealing with the human species. A number of chapters discusswhat is known from epidemiology, including cancer epidemiology. In Chapter16 Arnold Schecter, Olaf Papke, Marian Pavuk, and Rachel Tobey review ex- posure assessment of dioxins, with special emphasis on high resolution gaschromatographyhigh resolution mass spectroscopy, the current gold standardof exposure assessment. They note that only in the 1980s did this become pos-sible, and it was only in the 1980s that data showed that all humans carry abody burden of chlorinated dioxins and dibenzofurans. In Chapter 17 Matthew Longnecker, Susan Korrick, and Kirsten Moysichreview the epidemiology of polychlorinated biphenyls or PCBs. In Chapter 18Lennart Hardell, Mikael Eriksson, Olav Axelson, and Dieter Flesch-Janys 14. xviPREFACEreview the epidemiology of dioxins and cancer, including evidence of humancarcinogenicity. Chapter 19, by Sherry Selevan, Anne Sweeney, and Marie H.Sweeney, discuss the reproductive and developmental epidemiology of dioxins.In Chapter 20 a major dioxin incident that took place in Seveso, Italy in 1976 isdescribed by Pier Alberto Bertazzi and Alessandro di Domenico, who alsonoted the health consequences detected to date. Chapter 21 oers a descriptionof the health consequences of Japan rice oil, or Yusho, poisoning with PCBs,dibenzofurans, and small amounts of other chemicals in 1968, by one of thekey scientists who studied the incident, Yoshito Masuda. In the last chapter,Chapter 22, Yueliang Leon Guo, Mei-Lin Yu, and Chen-Chin Hsu describe asimilar rice oil poisoningthe Yucheng incidentthat occurred in Taiwan in1979 and its consequences on exposed persons. This book presents policy and science from the molecule to whole animals,and to human epidemiology in a selective fashion within one volume. Hope-fully, it will continue the eorts of the rst edition in presenting a relativelylarge but not overwhelming amount of material useful to experts, policy mak-ers, and the general public.Arnold Schecter Thomas A. Gasiewicz 15. CHAPTER 1Overview: The Dioxin DebateTHOMAS F. WEBSTERBoston University, Boston, MassachusettsBARRY COMMONERQueens College, CUNY, Flushing, New York1.1 INTRODUCTIONTo the general public, dioxin is the archetype of toxic chemicals, a substancethat in minute amounts causes cancer and birth defects. Raised to a high levelof visibility by the use of Agent Orange in Vietnam, it continues to generateenvironmental issues that capture public attention: Times Beach, Seveso, LoveCanal, herbicide spraying in the United States, waste incineration, and foodcontamination.Public fear engendered counter-reactions. Some claimed that dioxin causesno harm to humans other than chloracne, a disguring skin disease.1,2 Otherscompared the public attitude toward dioxin with witch hunts. Dioxin, theysaid, is a prime example of chemophobia, the irrational fear of chemicals.3,4U.S. Assistant Surgeon General Vernon Houk claimed that the evacuation ofTimes Beach, Missouri had been a mistake.5,6 Administrator William Reilly ofthe U.S. Environmental Protection Agency (USEPA) ordered a reassessmentof the toxicity of dioxin. He stated: I dont want to prejudge the issue, but weare seeing new information on dioxin that suggests a lower risk assessment fordioxin should be applied.6In our opinion, the public fears are largely justied. The current scienticevidence argues not only that dioxin is a potent carcinogen, but also that thenoncancer health and environmental hazards of dioxin may be more seriousthan believed previously. Indeed, dioxin appears to act like an extremely per-sistent synthetic hormone, perturbing important physiological signaling sys-tems. Such toxic mimicry leads to a host of biological changes, especiallyaltered cell development, dierentiation, and regulation. Perhaps the mostDioxins and Health, Second Edition, Edited by Arnold Schecter and Thomas A. GasiewiczISBN 0-471-43355-1 6 2003 John Wiley & Sons, Inc.1 16. 2 OVERVIEW: THE DIOXIN DEBATEtroubling consequence is the possibility of reproductive, developmental, andimmunological eects at the levels of dioxinlike compounds now present in thebodies of the average person. Observation of such phenomena in wildlife sug-gests that the environment is overburdened with these dangerous compounds. The pendulum of ocial opinion has swung back. Contradicting Houk,U.S. Assistant Surgeon General Barry Johnson testied in June 1992 that theevacuation of Times Beach was not a mistake.7 The USEPAs reassessment,although still incomplete at this writing, indicates that the danger from dioxinmay be broader and more serious than thought previously.8 In this overview,we discuss the basis for this dramatic turnaround and its logical implication: apolicy directed toward exposure reduction and pollution prevention.1.2 DIOXIN AND DIOXINLIKE COMPOUNDSThe polychlorinated dibenzo-p-dioxins are a group of 75 structurally relatedcompounds (congeners), including the well-known 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD or, as we shall refer to it, TCDD). Based on toxicitysimilar to that of TCDD, a wider group of halogenated aromatic compoundshave been recognized as dioxinlike. These include certain polychlorinateddibenzofurans (PCDFs), polychlorinated biphenyls (PCBs), polychlorinateddiphenyl ethers, polychlorinated naphthalenes, and others. Brominated andchlorobromo versions of these compounds may be dioxinlike as well.9 Membership in the class is dened biologically: dioxinlike compounds pro-duce a similar spectrum of toxic eects thought to be caused by a commonmechanism. The key step in the presumed mechanism is binding of the dioxin-like compound to a receptor protein, the Ah (aryl hydrocarbon) receptor(AhR). The molecules planar shape facilitates binding to the receptor, and itsrelative potency depends to a large degree on its persistence and how well it tsthe receptor. TCDD binds the Ah receptor with a very high anity and isextremely potent. Other planar molecules of about the same size and shape,including a number of the polyhalogenated dibenzo-p-dioxins and dibenzo-furans, t almost as well and are also very active. Although certain types ofpolychlorinated biphenyls bind to the receptor only weakly, their relativeabundance in the environment nevertheless makes them biologically important.PCBs with chlorines in positions that prevent the molecule from assuming aplanar position do not bind to the Ah receptor and are not dioxinlike in theirbiological eects. Some of these PCBs can exert toxicity through other mecha-nisms, however.101.3 SOURCESLarge-scale industrial production of the dioxinlike polychlorinated naph-thalenes began during World War I. Production of polychlorinated biphenyls 17. SOURCES 3(PCBs) followed in the late 1920s (Table 1.1). The thermal and chemical sta-bility of PCBs, among other properties, led to their widespread use in trans-formers, capacitors, heat transfer and hydraulic uids, as well as carbonlesscopy paper, plasticizers, and numerous other applications. Health and envi-ronmental problems led to curbs on their industrial production, but not untildecades later. In the meantime, about 650,000 metric tons were produced in theUnited States and about 1.5 million metric tons worldwide.11 It is estimatedthat about 20 to 30% of this amount has entered the environment. Much of theremainder is still in stock or in uses such as capacitors and transformers.12In contrast, the polychlorinated dibenzo-p-dioxins (PCDDs) and poly-chlorinated dibenzofurans (PCDFs) are unwanted by-products. Knowledgeof their origins has increased considerably, beginning with the identica-tion of TCDD as an unwanted by-product of the production of certaintrichlorophenols and herbicides,13 in particular, Agent Orange, a 1 : 1 mixtureof the n-butyl esters of 2,4,5-trichlorophenoxyacetic acid (2,4,5-T) and 2,4-dichlorophenoxyacetic acid (2,4-D). TCDD formation simply requires com-bining two molecules of 2,4,5-trichlorophenol under the right conditions.14More highly chlorinated PCDDs and PCDFs are formed during the productionof pentachlorophenol, a compound still used in the United States and elsewhereas a wood preservative. PCDFs also occur as low-level contaminants of PCBs.Much higher levels are generated by heating PCBs under the right conditions ofheat and oxygen.15 This phenomenon was a major contributor to the Yushoand Yucheng (oil disease) tragedies, where cooking oil was contaminated withPCBs.16,17 It also occurred in numerous incidents involving capacitor andtransformer res18,19 and the contamination of Belgian food in 1999.20It was thought for a while that the dioxin problem was limited to a fewreactions of closely related chemicals. Unfortunately, this is not the case.PCDD and PCDF were discovered in ash from trash-burning incinerators in1977 and later in their air emissions.21 It was not known at rst whether theemissions were due to unburned PCDD and PCDF in the fuel, formation fromchlorinated organic precursors, or de novo synthesis.21,22 Based on the obser-vation of y-ash-catalyzed chlorination of organic residues,23 it was hypothe-sized that PCDD and PCDF were synthesized as exhaust gases cooled down inthe boiler and air pollution control devices.24,25 This was soon conrmed bytests conducted at a Canadian incinerator. Little or no PCDD and PCDF werefound in the gases leaving the furnace, but these compounds were detected inthe cooler stack gases.26 Laboratory studies indicate signicant formation atabout 300 C.27 A number of U.S. incinerators equipped with electrostatic pre-cipitators running at these temperatures were very large dioxin sources: twoemitted at rates of roughly 1 kg per year of TCDD equivalents (TEQs),28 anamount of dioxinlike compounds considered equivalent in toxicity to TCDD.9Various mechanisms have been postulated for the synthesis reactions, in whichmetals play an important role (e.g., Ref. 29). Both organic and inorganicsources of chlorine may contribute to the formation of PCDD and PCDF.30Although there undoubtedly exists a connection between total emissions of 18. 4OVERVIEW: THE DIOXIN DEBATETABLE 1.1 Selected History of Dioxin and Dioxinlike Compounds18971899Chloracne characterized1918 Outbreaks of chloracne following exposure to chlorinated naphthalenes19201940Dramatic increase of PCDD and PCDF levels in North American lake sediments (reported 1984)1929 U.S. commercial production of PCBs begins1947 X disease described in cattle in the United States19491953Chemical accidents at Monsanto, Boehringer, and BASF1957 TCDD identied as unwanted contaminant in the manufacture of trichlorophenol1957 Chick edema disease outbreak in poultry in southeastern United States19621970Agent Orange used in Southeast Asia19651966Holmesburg prison experimentsMid 1960sOutbreaks of reproductive and developmental eects in Great Lakes sh-eating birds1968 Yusho oil disease (Japan)1971 TCDD found to cause birth defects in mice Contamination of Times Beach and other Missouri sites19721976Ah receptor hypothesis developed1973 Polybrominated biphenyls accidentally added to cattle feed in Michigan1974 TCDD detected in human breast milk from South Vietnam1976 Accident in Seveso, Italy1977 U.S. commercial production of PCBs halted TCDD found to cause cancer in rats Discovery of dioxin emissions from trash incinerators1978 Kociba et al. cancer study of rats exposed to TCDD1979 USEPA emergency suspension of some 2,4,5-T uses Yucheng oil disease (Taiwan) Association of soft tissue sarcoma with TCDD and phenoxyacetic acid herbicides1979TCDD found to modulate hormones and their receptors1980 Evacuation of Love Canal1981 Transformer re in Binghamton, New York state oce building19831985General public found to be contaminated with PCDD and PCDF1985 USEPA health assessment of TCDD1986 Production of dioxin by chlorine-bleached paper mills discovered, al- though proposed earlier1988 First USEPA reassessment of TCDD Die-o of Baltic seals1990 Second Banbury conference on dioxins1991 NIOSH cancer mortality study of U.S. chemical workers Second USEPA reassessment begins1992 U.S.Canadian International Joint Commission 6th Biennial Report1993 AhR is a member of bHLH PAS family1997 IARC classies TCDD as a human carcinogen1998 WHO reduces tolerable daily intake1999 Food contamination in Belgium2000 POPs treaty USEPA reassessment completed? 19. SOURCES 5these compounds and levels of chlorine in fuel, the nature of this relationship isstill not understood completely.31,32Such ndings imply that dioxinlike compounds may be formed during vir-tually any combustion process when chlorine is present. This idea was proposedin the Trace chemistry of re hypothesis, which stated that PCDD formationwas a natural consequence of combustion.33 This led, in turn, to the claim thatforest res and other nonindustrial sources are potentially signicant or eventhe dominant source of dioxin.3436 However, this claim is at odds with anumber of observations. Levels of PCDD and PCDF are higher in people fromindustrialized countries than in residents of less industrialized nations.37,* Lakesediments from North America and Europe show that PCDD and PCDF levelswere very low until approximately 19201940.38 Similarly, the levels of thesecompounds in ancient mummies and 100- to 400-year-old frozen bodies are farlower than those currently found in the average resident of an industrializedcountry.3941 Hence, although trace amounts of PCDD and PCDF may havebeen present in preindustrial times, the current levels represent a huge increaseover these low values.In North America, the dramatic increase of PCDD and PCDF in lake sedi-ment matches the beginnings of large-scale industrial chlorine chemistryandcombustion of its productsduring the period 19201940. Concentrations ofPCDD and PCDF in stored vegetation and soil samples from the UnitedKingdom increased at the turn of the century,42 perhaps reecting the advanceddevelopment of industrial processes in that country. Coal burning has beensuggested as a major contributor; however, large-scale coal burning antedatesthe increase of PCDD/PCDF in North America sediments. PCDD and PCDFhave been detected in the emissions from coal combustion, but at fairlylow levels. It is possible that the relatively high levels of sulfur in coal inhibitsformation of PCDD/PCDF.43,44 Low levels of PCDD and PCDF found inBritish soil and vegetation samples from the mid-nineteenth century may par-tially reect even earlier industrial activity. For instance, the Leblanc processfor producing alkali, a forerunner of modern industrial chlorine chemistry, rstfound widespread application in the United Kingdom at this time.45PCDD and PCDF are emitted when other chlorine-containing fuelsare burned, including chemical waste, hospital waste, and sewage sludge.46Dioxin-containing wastes have not proven as easy to destroy in hazardouswaste incinerators as once claimed, perhaps because of resynthesis.47, Theexhaust from automobiles burning leaded gasoline contains both chlorinatedand mixed halogenated dioxins and dibenzofurans, apparently arising fromethylene dichloride and ethylene dibromide used as lead scavengers. Muchlower levels of PCDD and PCDF have been found in exhaust from vehicles* Levels of PCDD and PCDF, especially TCDD, are elevated in human tissues of southern Viet-nam relative to northern Vietnam, reecting both dierences in industrialization and the millions ofgallons of Agent Orange sprayed in the south.37y The USEPAs response was that the regulation requiring an incinerator to achieve 99.9999%destruction or removal of dioxin-containing wastes does not actually apply to the dioxin itself.48 20. 6 OVERVIEW: THE DIOXIN DEBATEburning unleaded gasoline, presumably reecting the low level of chlorine inthis fuel.46 The identication of additional PCDD and PCDF sources strengthens theconnection between these compounds and industrial chlorine chemistry. PCDDand PCDF are formed in the bleaching of pulp and paper with chlorine, notsurprising given the rich aromatic content of the lignin found in wood.49Hypothesized as early as 1974,50 this phenomenon was not conrmed until themid-1980s, when high concentrations of TCDD were found in sh downstreamof bleached pulp mills51,52 and then in the mills themselves.53,54 Dioxin mayalso be formed during chlorine regeneration of metal catalysts used in petro-leum rening.55 Large amounts of PCDD and PCDF are also produced by certain typesof metal processing, perhaps reecting the catalytic properties of a number ofmetals. PCDD and PCDF are emitted by the burning of scrap metal, such ascopper cable coated with PVC plastic insulation.56 Other sources include alu-minum smelting, magnesium and nickel production, scrap metal melting, andiron and steel production.5759 In these manufacturing processes chlorine iseither used or is contained in cutting oils, plastic, and other contaminants. Dioxinlike compounds are formed at the heart of the chlorine industry aswell. Large amounts of PCDD and PCDF have been found in the sludge fromchloralkali plants that used graphite electrodes, once widely employed.* Mostmodern facilities now use other kinds of electrodes.60 PCDD/PCDF have beendetected in some common chlorinated hydrocarbons.61 They are formed duringthe production of ethylene dichloride (EDC).6264 EDC is used primarily toproduce vinyl chloride, the precursor to polyvinyl chloride (PVC) plastic.About 4.2 million metric tons of PVC were produced in the United States in1991,65 making it the single largest use of chlorine in the country. Octachlorinated dioxin may be formed from pentachlorophenol at near-ambient temperatures in sewage sludge66 and in the atmosphere.67 Dioxinshave recently been found in ancient clay deposits.68 Their origin is still myste-rious, but they suggest an unknown natural source. In sum, the range of sources has expanded to the point that virtually allindustrial chlorine chemistry can be suspected of generating dioxinlike com-pounds at some point during production, use, or disposal. The unwantedproduction of PCDD and PCDF may reect the relative stability of thesecompounds. They can be thought of as thermodynamic sinks that are likely toaccumulate in reactions involving chlorine and organic materials and maytherefore be expected to occur in a very wide range of reactions. What are the largest sources? National air emissions inventories havebeen performed in the United States, Canada, Japan, Australia, and a numberof European countries.46,69 Waste incinerationmunicipal, hospital, hazard-* The chloralkali process manufactures chlorine and sodium hydroxide (an alkali) from sodiumchloride brine via electrolysis. The graphite may provide a source of carbon for the generation ofPCDD and PCDF. 21. ENVIRONMENTAL FATE AND EXPOSURE 7ous, and industrialand metal processing are the largest current estimatedsources of emissions to the atmosphere. More dicult to quantify, but poten-tially important, are releases from pentachlorophenol and 2,4-D as well asbackyard waste burning.46 Cases of massive dioxin contamination have beenreported in Russia.70,71 Several groups have attempted to balance emissionsagainst levels found in the environment (e.g., Ref. 72). This is made dicult bythe limitations of current inventories and representativeness of environmentalsamples.28,46 Nevertheless, environmental levels appear to exceed the levelimplied by known sources. It is possible that important sources have not yetbeen accounted for or identied properly.671.4 ENVIRONMENTAL FATE AND EXPOSUREDioxin is primarily a modern problem, a by-product of industrial chlorinechemistry and the combustion of chlorine-containing fuels. The growth of theseprocesses during the twentieth century dramatically increased the levels ofthese compounds in the environment and in biota. Accumulation also dependson the environmental behavior of dioxinlike compounds, a consequence oftheir chemical and physical properties: low vapor pressure and water solubility,high lipophilicity, and relative chemical stability.73 When the metabolic inert-ness of many congeners is added to the list, the prole is complete: Dioxinlikecompounds tend to persist and bioaccumulate. Combustion sources emit large quantities of dioxinlike compounds into theatmosphere, where they are both dispersed and subjected to selective degrada-tion. The more highly chlorinated compounds tend to adsorb onto airborneparticulate at ambient temperature, greatly reducing their rate of degrada-tion.74 Larger fractions of the lower-chlorinated congeners are found as vapor,making them more susceptible to photolysis and attack by hydroxyl radicals.These phenomena may partly explain the shift toward a preponderance of themore highly chlorinated compounds seen in many abiotic environmental sam-ples relative to the typical pattern found in combustion emissions.75 Atmo-spheric residence times of particulate-bound congeners are determined by dryand wet deposition of the particulate.76 Thus, many dioxinlike compounds are suciently stable to travel long dis-tances in the atmosphere. The ubiquitous presence of dioxins in the environ-ment, even in remote locations such as arctic Canada, is probably due to thecumulative impact of many sources.77,78 Deposition from the air contaminatessoil, water, and vegetation. Deposition of both particulate and vapor ontoplants provides a signicant entry into the terrestrial food chain.7982 Humanexposure via milk and beef may be orders of magnitude higher than via inha-lation, making it a major issue in the permitting of air emission sources.79 Because of their low water solubilities and vapor pressures, PCDD andPCDF tend to partition into soil and sediment. The half-life of TCDD may beon the order of a decade or more in soil and probably longer in sediment.15 As 22. 8 OVERVIEW: THE DIOXIN DEBATEa result, these two media can act as reservoirs, leading to recontamination ofother media.In aquatic systems, the highly lipophilic and hydrophobic dioxinlike com-pounds tend to bioconcentrate from water to aquatic animals and then bio-magnify up the multistep food chain.8385 Levels of PCBs found in sh-eatingbirds, animals near the top of the aquatic food chain, can reach concentrationstens of millions of times higher than those of PCBs dissolved in water.83 Thecombined eects of bioaccumulation and the action of sediment as a reservoirmake direct discharge of these compounds into aquatic systems particularlyproblematic.Humans are also high on the food chain, eating the meat and milk of her-bivores as well as sh and plants. The average person in an industrial country isthought to be exposed to PCDD and PCDF primarily via these animal prod-ucts. The average daily dose is about 1 pg/kg per day of TCDD equivalents(TEQs).8 General contamination of the environment and food sources mayexplain the relatively similar levels of PCDD and PCDF found in the averageresidents of industrialized countries.37Dioxinlike compounds accumulate primarily in peoples body lipid, espe-cially adipose tissue. Their elimination depends on metabolic degradationslight or nil for many congenersand on the rate of excretion that is almostcompletely via the feces. As a result, the half-life of TCDD in humans is verylong, on the order of a decade; OCDD may have a half-life as long as 50years.86,87 This biological persistence leads to another route of exposure. Afteraccumulating in the mother over decades, dioxinlike compounds can be passedto the developing fetus in uteroa particularly vulnerable periodor to new-borns via lactation.37 Similarly, birds and sh accumulate these compoundsand pass them to the egg.83,84,88Like the increase of atmospheric chlorine caused by chlorouorocarbonsand certain chlorinated solvents, the dioxinlike compounds represent a pertur-bation of the planets chemistry. This might have been only a curious and littlenoticed sidelight of the industrial age if not for an additional factor: theextraordinarily powerful biological eects of the dioxinlike compounds.1.5 BIOCHEMISTRY AND TOXICITY1.5.1 Biological PersistenceAs noted earlier, the central event that instigates the biological eects ofTCDD and dioxinlike substances is thought to be their binding to a recep-tor protein. This aryl hydrocarbon (Ah) receptor (AhR; see Chapter 12) wasrst postulated during studies of PCDDs and polycyclic aromatic hydrocarbons(PAHs) such as 3-methylcholanthrene.8991 Sequencing of the gene for theAhR revealed a surprise. Rather than being a member of the steroid hormone 23. BIOCHEMISTRY AND TOXICITY 9receptor family as supposed originally, it belongs to a family of basic helix-loop-helix proteins containing a sequence known as the PAS domain.92Once a compound is bound to the AhR, a series of intracellular processesmay ensue, including shedding and binding of other factors [including the arylhydrocarbon nuclear translocator (ARNT)], migration into the nucleus, andbinding of the complex to specic sequences of DNA called AhR-responsiveelements. By inuencing the rate of transcription of specic messenger RNAs,the rate of synthesis of the related proteins is altered. Thus, the binding of anappropriate compound (ligand) to the receptor can change the cellular concen-tration of certain proteins by regulating the expression of genes governing theirsynthesis.93 The presence or absence of cofactors may account for dierencesin gene expressions by tissue. Other molecular mechanisms are of increasinginterest.92,94Among the proteins induced by the Ah receptor are three cytochromeP450 enzymes: CYPIA1, CYPIA2, and CYPIB1. These phase I enzymes oxi-dize foreign (xenobiotic) substances, including PAHs, plant constituents suchas avones, aromatic amines, and some pharmaceutical drugs. One conse-quence of this metabolic conversion is that the xenobiotic substance, typicallylipid- rather than water-soluble, is then subject to further enzymatic conver-sion. Phase II enzymes add hydrophilic groups, enhancing water solubility andexcretion from the body. In this manner, AhR-induced enzymes can reduce thebiological eect of some environmental agents by facilitating their metabolicdegradation.95,96On the other hand, the oxidative transformation of a xenobiotic com-pound by the AhR-induced cytochrome P450 enzymes may greatly enhanceits biological activity.95 A classical example is 2-acetylaminoourene (AAF),a potent liver carcinogen in rats. A number of studies have shown that theproximal carcinogen is not AAF but an oxidized metabolite produced viaCYPIA2.95,97 Preparations containing these enzymes are used in microbialmutagenesis tests to activate otherwise inert genotoxins.When the oxidative degradation of the xenobiotic compound facilitates itsexcretion so that the intracellular concentration is reduced, a negative feedbackis established: Binding of the compound to the receptor is also reduced, tran-scription decreases, the level of the cytochrome P450 enzymes diminishes, andthe system returns to its initial condition. Other sources of feedback may alsobe present.92TCDD and related halogenated compounds strongly induce CYPIA1 andCYPIA2 but are not readily oxidized by these enzymes. They are apparentlyprotected from attack by the presence of halogen atoms in certain positions ofthe molecule. Hence, they are excreted very slowly, resulting in a prolongedand amplied response. In eect, a feedback system that governs behavior ofother AhR-binding substances (such as the PAHs) is inoperative in the case ofTCDD. Thus, the extraordinary biological potency of dioxinlike substancesmay be due to the consequence of their unique combination of two properties: 24. 10OVERVIEW: THE DIOXIN DEBATEa high anity for the Ah receptor and biological persistence. The relevance ofpersistence is evident from a comparison of the behavior of dioxinlike com-pounds and PAHs.* Although some PAHs bind to the Ah receptor with ananity almost equal to that of TCDD, their in vivo potency (as measured byenzyme induction) is many orders of magnitude less.981.5.2 Perturbation of Hormones and Growth FactorsAlthough the induction of CYPIA1 is the best characterized of the biochemicaleects of dioxinlike compounds, it is by no means the only one. The expressionof a growing number of genes are thought to be regulated by the Ah recep-tor.99,100 This may provide one mechanism whereby dioxinlike compoundsperturb the regulation of hormones, growth factors, and other molecular mes-sengers that control growth and dierentiation with diverse and potentiallydevastating impact. Some examples follow. TCDD may alter the levels of certain hormones through its inuence on theenzymes that metabolize primarily xenobiotic compounds. For instance,TCDD induces one form of UDP-glucuronosyltransferase (UDPGT), a phaseII enzyme that increases a chemicals solubility by adding glucuronic acid.In addition to xenobiotics, UDPGT also conjugates and enhances the excretionof thyroxine (T4), causing reduced serum levels of this thyroid hormone inrats.101 Among the resultant complications is a perturbation of an importantbiological feedback system: The pituitary responds to low T4 with increasedsecretion of thyroid-stimulating hormone (TSH). When prolonged, this maylead to thyroid tumors,102 a sensitive endpoint in TCDD-exposed rats.103 TCDD does not bind to steroid hormone receptors, and steroid hormonesdo not bind to the Ah receptor.104 Nevertheless, TCDD aects steroid hor-mone regulation in more subtle ways. Thus, TCDD decreases (downregulates)the number of estrogen receptors in certain organs of the female rodent, mak-ing tissues less responsive to this hormone.105 This may decrease both fertilityand incidence of tumors of these organs, as has been suggested in rats exposedto TCDD postnatally.106 TCDD reduces testosterone levels in adult male rats by decreasing the pro-duction of testosterone from cholesterol in the testes at a critical rate-limitingstep. The pituitary (and/or hypothalamus) normally responds to low testos-terone concentration by increasing secretion of luteinizing hormone, causingincreased production of testosterone. TCDD interferes with this feedback sys-tem, preventing the compensatory increase of luteinizing hormone.107110 TCDD also aects growth factors, a class of extracellular signaling mole-cules. In the female rat liver, TCDD may increase migration of epidermal* Recent work shows that changing a single nucleotide anking the AhR-responsive element candetermine whether a gene is transcribed by AhR bound to PAH or TCDD. This nding may haveimportant implications for both understanding molecular mechanisms and extension of TEFs toother classes of compounds [see T. Matikainen et al., Nature Genetics 28: 355360 (2001)]. 25. BIOCHEMISTRY AND TOXICITY11growth factor receptors (EGFRs) internally from the cell membrane, providinga stimulus for mitosis.103,104,111 This eect appears to depend on ovarian hor-mones; interactions between EGFRs and estrogen receptors have been noticedelsewhere.112 TCDD may aect EGFRs by increasing the levels of trans-forming growth factor a (TGFa), a ligand for EGFRs.103 In mice, TCDDalters the dierentiation of certain tissues in the developing palate. This may becaused by perturbation of growth factors and their receptors, includingEGFRs. The palatal shelves come into contact but fail to fuse, resulting in cleftpalate.113 TCDD can lead to increased phosphorylation of amino acids. Protein kin-ases often play important roles in transducing signals across cell membranes,regulation of growth factor receptors, and cell dierentiation.114 TCDD mayalter regulation of the cell cycle.94,115 TCDD also inuences a number of otherchemical messengers, including the glucocorticoid hormone receptor, plasmi-nogen activator inhibitor, protein kinase C, interleukin-1b, and other cyto-kines.99 TCDD has been called a persistent environmental hormone.116 One of itsmolecular mechanismsbinding to a receptor that regulates gene expressionhas certain similarities to the action of steroid hormones117 as well as impor-tant dierences.92 It alters cell growth and dierentiation. It aects otherhormones and growth factors, including altering the levels of their receptors.Finally, like hormones, TCDD causes signicant eects at very low doses. Thisknowledge of dioxins biochemistry increases our concern over its widespreadoccurrence in the environment. Does the body possess some unidentied hormone that binds to the Ahreceptor, serving an important but unknown function? Such a situation is notunprecedented. A number of such orphan receptors (i.e., receptors withoutknown ligands) have been found.118 Dioxin may be a case of toxic mimicry,possessing a molecular shape similar to that of its putative natural counterpart.The long residence time of TCDD in the body may alter expression of AhR-regulated genes for an inappropriately long period of time. It is also possiblethat the supposed natural ligand of the Ah receptor might normally functionduring a specic period of development; TCDD may activate the system at thewrong time.119 Knockout miceanimals without a functional AhRare viablebut show defects in the development of the liver, immune system, and repro-duction.120122 The question of the normal function of the AhR is a major goalof molecular research in the eld.1.5.3 ToxicityFrom the foregoing account it is apparent that TCDD is capable of disruptinga wide variety of biochemical processes which are likely to lead to an equallybroad spectrum of macroscopic toxic eects in animals. The latter includeacute toxicity, wasting and death, atrophy of the thymus, liver damage, epi-dermal changes, immunotoxocity, birth defects, reduced fertility, endome- 26. 12OVERVIEW: THE DIOXIN DEBATEtriosis, and cancer.8,98 It is generally thought that the Ah receptor mediatesthese eects,8,9 although there may be exceptions.123 Hence, it is assumed thatother dioxinlike compounds will also cause these eects. The relative sensitivity of various toxic endpoints appears to vary with tissueand species, implying that humans may be less sensitive than laboratory ani-mals for some eects and more sensitive for others.8,119 In this review we focuson cancer, reproductive/developmental eects, and immunotoxicity. They haveformed the primary basis for regulatory eorts by a number of agencies and arecentral to the current USEPA reassessment of TCDD toxicity.1.6 CANCER1.6.1MechanismsThere is no doubt that TCDD causes cancer in animals. This has been shownin both sexes of several species103,124 (see Chapter 18). The contentiouspoints surround its carcinogenic mechanism(s) and their implications forhuman exposure at low doses. The development of cancer is generally thoughtto proceed in several steps: (1) an initial permanent alteration of a cell, typicallysome kind of genetic damage; (2) clonal proliferation of the altered cell; and (3)another permanent alteration in at least one of these cells, followed by morecell replication.104 This last step may repeat several times, a process calledtumor progression. Two-stage cancer experiments attempt to partially dissect these steps:An animal is given a dose of a DNA-damaging (initiating) agent followed bychronic exposure to a promoting agent. In such an experiment, a classic pro-moter greatly enhances the number of tumors and precancerous lesions butcauses little or no cancer by itself. Its action is considered reversible (i.e.,removal of the promoter causes the tumor to regress). Initiation and promotionare operationally dened by this experimental protocol and are not necessarilysynonymous with DNA damage and cell proliferation.103,125 In two-stage experiments involving rat liver and mouse skin, TCDDis an extremely potent promoter and displays little or no initiating activ-ity.103,126128 The latter nding is puzzling given TCDDs ability to generatesubstantial numbers of tumors in the rat liver (and other organs) in long-termanimal experiments (bioassays) when given alone (i.e., without a known initia-tor).106 These divergent results explain why some researchers consider TCDDa promoter, whereas others consider it a complete carcinogen, able to inducetumors by itself.124,129 The discrepancy might be related to dierences in thefood or environment of the test animals or the shorter length of exposure in thetwo-stage experiments. Although it is possible that TCDD promotes background initiated cellsdamaged by some independent process, other evidence suggests that TCDDmay act throughout the carcinogenic process (Figure 1.1). TCDD does notdirectly cause mutations in several common assays and therefore appears to 27. CANCER 13DNA damage andFigure 1.1 Some possible roles of TCDD in carcinogenesis of the female rat liver.(Adapted from Ref. 104.)lack the direct genetic eect characteristic of an initiator. However, it is pos-sible that TCDD contributes indirectly to DNA damage.103 For example, byinducing CYPIA1 and CYPIA2, dioxinlike compounds may in some casesincrease the conversion of other compounds into mutagens.130 Lymphocytesfrom people exposed previously to dioxinlike compounds show increased fre-quency of sister chromatid exchangesuggesting genetic damagewhen cul-tured with a-naphthoavone (ANF); ANF is metabolized to an active form byCYPIA1.131 On the other hand, some experiments show reduced DNA dam-age and cancer in animals dosed with certain polycyclic aromatic hydrocarbons(PAHs) after exposure to TCDD compared with animals exposed to PAHalone.132 The balance between metabolic activation and deactivation maydepend on the compound and the dosing regime.103 Two-stage studies of the female rat liver support the hypothesis that TCDD-induced cell proliferation is involved in promotion, although this may not bethe only mechanistic step. Preneoplastic lesions were greatly increased in ratsexposed to diethylnitrosamine (DEN) followed by TCDD, but not in animalsexposed to TCDD alone. Cell proliferation was greatly increased in animalsexposed to TCDD. These elegant studies also demonstrate the involvementof estrogenic hormones. Cell proliferation and preneoplastic lesions were sig-nicantly higher in intact rats than in ovariectomized animals.* TCDD may* In these experiments, lung cancer was seen in ovariectomized females, but not in the intact ani-mals. In long-term bioassays, TCDD signicantly increases liver tumors in female but not malerats.106,133 However, TCDD produces liver tumors in both sexes of mice.133 The reason for thisdierence is unknown. 28. 14OVERVIEW: THE DIOXIN DEBATEenhance cell proliferation via the epidermal growth factor pathway. Internal-ization of epidermal growth factor receptor (EGFR), a mitogenic stimulus, isenhanced by TCDD in intact but not ovariectomized animals.111,128 The ini-tiator DEN is metabolically activated by P450 enzymes other than CYPIA1and CYPIA2.134,135 Induction of these enzymes was unaected by removal ofthe ovaries. Evidence from rat liver studies suggests that TCDD may also play a rolein tumor progression. Some precancerous lesions regress when administrationof TCDD ends, but others continued to grow in size.136 One possible explana-tion is that additional permanent alterations occurred in these lesions makingthem promoter-independent. TCDD may contribute indirectly to these eventsby another estrogen-dependent mechanism. Metabolism of the ovarian hor-mone 17b-estradiol by CYPIA2 can lead to the production of DNA-reactivespecies.103,104,111 Although most of the experimental work has been done on the liver,TCDD alters tumor incidence at numerous sites in long-term bioassays. Thisnding implies that a number of mechanisms may be involved. As discussedearlier, prolonged secretion of thyroid-stimulating hormone in response toTCDD-induced degradation of T4 may increase thyroid tumors in rats.TCDD decreases the number of estrogen receptors in the uterus and breast; thismay be connected to the apparent reduction of tumors in these organs in therat exposed postnatally.* Interestingly, prenatal exposure of rats to TCDDretarded the development of mammary tissue, making it more susceptible tocarcinogen exposure later in life.137 TCDD may contribute to cancer in other ways, including interactionwith viruses,138 increased expression of protooncogenes,139 decreased expres-sion of tumor suppressor genes, oxidative stress,94 altered regulation of the cellcycle,115 and suppression of cell-mediated immunity.140 EpsteinBarr virus,which is widespread in the human population, may cause B-cell proliferationand immortalization. Impairment of cell-mediated immunity by TCDD andother chemicals may allow continued proliferation and development into non-Hodgkins lymphoma.141 In one epidemiologic study, the combination ofPCBs and EpsteinBarr virus had a strong synergistic eect on non-Hodgkinslymphoma.142 In sum, TCDD may act at a number of steps in the carcinogenic processin conjunction with endogenous hormones, exogenous compounds, and virusesin an organ-specic fashion. Many aspects of its carcinogenic mechanismremain unknown.1.6.2Cancer Risk Assessment and ReassessmentDebate over the mechanism of TCDD carcinogenicity has played a central rolein regulation of exposure and in the USEPA reassessments of its potency. The* Reduction in body weight gain is another suggested explanation.103 29. CANCER15TABLE 1.2 Some Acceptable or Tolerable Daily Doses of TCDDLevel (pg/kgOrganizationper day)MethodologyBasis a Ref. bUSEPA (1985)0.006 LMS model Kociba 143ATSDR (1998) c1 Safety factor Neurotoxicity144in rhesusWHO (1998) 14Safety factor and Various145body burdensCanada/Ontario (1985)10 Safety factor Kociba, Murray 146Washington 2080Safety factor Receptor occu- 147Department of pancyHealth (1991) daOlder values relied on the rat cancer study of Kociba et al.106 and/or the rat reproduction study ofMurray et al.148bFor an upper-bound lifetime cancer risk of 106 . Increase of cancer potency was proposed in draftreassessment.8cATSDR is the U.S. Agency for Toxic Substances and Disease Registry.dWithdrawn after discounting of receptor threshold hypothesis.general goal of these eorts has been the identication of a safe or accept-able daily dose of dioxin. As Table 1.2 shows, the values used by a number ofcountries and government agencies have ranged from 0.006 to over 20 pg/kgbody weight per day, a factor of several thousand. Even the high end reectsthe extreme toxicity of TCDD: A picogram is a mere trillionth of a gram.These risk assessments have generated intense controversy. The average res-ident of the industrialized countries is exposed to about 1 pg/kg per day ofTCDD equivalents (TEQs).8 If values from the upper end of Table 1.2 areused, the average dose is considered acceptable. On the other hand, averageexposure greatly exceeds the lower estimates of acceptable dose, a situationsome interpret as requiring remedial action.149 In practice, the USEPA regu-lates incremental exposure only from a single source or medium. Nevertheless,some sources fail to meet even these standards.One can immediately see one reason for the political pressure placed on theUSEPA: Its estimate of the acceptable dose is one of the lowest in Table 1.2. Anumber of dioxin-generating industries and owners of dioxin-contaminatedsites have, perhaps not surprisingly, maintained that higher values are moreappropriate.Another fundamental reason for disagreement lies in scientic dierencesabout how to construct an acceptable dose. Practical considerations restrict thenumber of animals that can be used in a cancer bioassay. This imposes a limiton the ability to detect an increased number of tumors. To avoid false neg-atives, the doses employed are typically much larger than those commonlyexperienced by people. Judging the safety of the latter requires extrapolationfrom high to low dose and from animal to human. Depending on how these 30. 16OVERVIEW: THE DIOXIN DEBATEFigure 1.2 Two possible doseresponse curves.extrapolations are carried out, the same data can readily give rise to very dif-ferent conclusions. The primary theoretical dierence between the high and low estimates ofacceptable dose is the shape of the doseresponse curve at low doses, in par-ticular the presence or absence of a threshold (Figure 1.2). The high values arederived from the view that there is a dose of TCDD below which there is noeect. They are typically estimated by applying a safety factor to an experi-mentally dened no observed adverse eect level (NOAEL) or lowest observedadverse eect level (LOAEL). Such levels depend on both the biology of thephenomenon and the methodology and statistics of the experiment. Thus, aneect might occur at some low dose, but the particular experimental designmay not be powerful enough to distinguish it from the control value. Ontariostolerable average daily intake, 10 pg/kg per day, is based on a presumed no-eect level in animals of 1000 pg/kg per day for cancer and reproductive eectswith a safety factor of 100.146 In contrast, the low acceptable dose is based on the theory that there isno threshold for cancer. The probability of cancer is assumed to be directlyproportional to dose at low dose. Since there is no completely safe dose, theslope factor or potency of the chemical is used to calculate the dose resulting ina certain lifetime risk of cancer that is regarded as acceptable. The USEPAfollowed this procedure in its landmark 1985 assessment of TCDD carcinoge-nicity.143 The slope factor was estimated from rat tumor data from Kocibaet al.106 using the linearized multistage (LMS) model of cancer. A small addi-tional factor adjusted for possible interspecies dierences. According to thisestimate, an average daily dose of 0.006 pg/kg per day corresponds to anupper-bound excess lifetime cancer risk of 1 in a million (106 ). Hence, if 1million people received this level of exposure over their lifetime, less than oneadditional case of cancer would be expected. The USEPA has considered thislevel of risk acceptable (de minimus) as a matter of policy, although the agencyoften uses or approves higher values. Assumptions about the shape of the doseresponse curve for cancer dependcritically on the mechanism of carcinogenicity. The linear multistage model 31. CANCER17(LMS) used by the USEPA in 1985 was derived from the theory that cancerinvolves a sequence of irreversible stages.150 There have been numerous criti-cisms of this approach as it applies to TCDD. One group has argued thatTCDD is a promoter, not an initiator, and is therefore subject to a threshold.They contend that the NOAEL/safety factor approach is more appropriate.Others argue that the LMS model is not appropriate because it does not takecell proliferation into account. However, other work indicates that even a purepromoter can act linearly at low dose if its eect is additive to backgroundprocesses.151A third source of controversy concerns the role of scientic uncertainty inregulatory policy. Whereas some argue that pollution should be allowed untilproof of harm is certain, others advocate more precautionary approaches. TheUSEPA has in the past relied on linear models, assuming that this approachwill be more protective of public health.152Some of these arguments were raised during the USEPAs 1988 reassess-ment of TCDDs slope factor. USEPA concluded that TCDD may causecancer through a variety of mechanisms and that the LMS model would beretained, in part because there was no adequate alternative model. Neverthe-less, the agencys Dioxin Workgroup argued that the 1985 slope factor estimateis likely to have led to an overestimate of risk.153 Although the degree ofoverestimation was unspecied, they proposed raising the acceptable dose to0.1 pg/kg per day, simply as a matter of policy. This proposal was rejected bythe agencys Science Advisory Board because no new scientic evidence hadbeen presented to justify the change in the risk estimate.154 However, the boardexpressed concern about the applicability of the LMS model to TCDD andencouraged development of new risk models that would incorporate additionalmechanistic research into risk assessment, in particular, receptor mediation oftoxicity.1.6.3 Reassessment IIBy the late 1980s a new political factor entered the dioxin arena: the interest ofthe paper and allied industries. Dioxin had been discovered in euent andproducts from pulp mills using chlorine as a bleaching agent. The industry wasconcerned about possible legal action and impending surface water qualitystandards. The paper industry began a campaign to get EPA to rethinkdioxin risk assessment.155,156 The chlorine-producing industry became anally, presumably because the paper industry consumed a signicant fraction oftheir output.A new challenge to the USEPAs cancer slope factor came in 1989 duringconsideration of a water quality standard for the state of Maine. Female livertissue samples from the 1978 Kociba experiment were reviewed and reclassiedbased on new criteria for the presence of tumors.157 Tumors from other sitesand bioassays were not examined, although somemale rat thyroid and malemouse liveralso produce high slope factor estimates.8,129 The revised liver 32. 18OVERVIEW: THE DIOXIN DEBATEtumor data might reduce the TCDD cancer slope factor estimate by a factor ofabout 2 to 3, an insignicant amount in view of the underlying uncertainties.158Although some proposed that the liver tumors were a secondary response tohepatotoxicity,159 a review by the USEPA and Food and Drug Administrationdisagreed.103,160 In October 1990, the Chlorine Institute and the USEPA cosponsored a sci-entic meeting entitled the Biological Basis for Risk Assessment of Dioxinsand Related Compounds at the Banbury Center. There was general agree-ment among the toxicologists and biochemists present at the meeting that thecurrently known toxic eects of dioxin are mediated by the Ah receptor.Although this was not particularly radical thinking, some participants drew acontroversial conclusion: Receptor mediation implied a threshold for the bio-logical eects of dioxin. They stated that a certain number of receptors must beoccupied for any biological eect to occur. Rather than being linear at lowdoses, the doseresponse curve was shaped like a hockey stick: at or increas-ing very slightly at rst and becoming linear only at higher doses. Furthermore,a practical threshold for toxic eects could be determined from the dose ofTCDD necessary to induce CYPIA1, the presumed most sensitive endpoint.The resulting rough and rapidly calculated estimate was about 1 to 3 pg/kg perday, much greater than the USEPAs value of 0.006 pg/kg per day.2,161 Apublic relations rm hired by the Chlorine Institute went further, claiming in apress packet that the attendees had formally reached consensus on the thresh-old concept. This was not in fact correct, prompting vociferous protests andcreating a minor scandal in the scientic press.162 This meeting set the stage for another dioxin reassessment by the USEPA,which was announced in April 1991. The primary focus would be the develop-ment of a new biologically based model for dioxin toxicity, developing theideas from the Banbury conference and the earlier comments of the ScienceAdvisory Board.163165 Incorporating the latest scientic ndings, the newmodel would provide an alternative to the safety factor and LMS approachesto risk assessment. News of the reassessment was reported widely, including thenotion that TCDD was much less toxic than previously thought. Meanwhile, the paper industry used the supposed outcome of the BanburyConference to argue for relaxed TCDD water quality standards in a numberof states (e.g., Ref. 166). The Washington State Department of Health issuedrevised guidelines for sh consumption based on a tolerable daily intake in therange 20 to 80 pg/kg per day. This value was calculated by applying a safetyfactor to the dose estimated to give 5% occupation of Ah receptors in the ratliver, a level assumed necessary for any biological response. No references tothe scientic literature were given for this crucial assumption.147 The tolerabledaily intake was later withdrawn. There are good reasons to be skeptical of the claim that involvement ofa receptor requires a threshold. The simple classic model for receptors predictsa linear relationship between low concentrations of TCDD and the amount of 33. CANCER19receptor-bound dioxin.* Biological responses would not have a threshold ifthey are proportional to the amount of receptor-bound TCDD. Of course, thedoseresponse curves of more complicated biological responses might deviatefrom linearity, but this is only a possibility and not a requirement of receptortheory.207 The simple threshold model was seriously weakened at the Eleventh Inter-national Symposium on Chlorinated Dioxins and Related Compounds, heldin North Carolina in September 1991. Studies yielded data on induction ofCYPIA1 and CYPIA2 in the rat liver that were consistent with both thresholdand nonthreshold (low-dose linear) models. The no-threshold models providedthe best mathematical t. Similar results were found for dioxin-induced lossof EGFR from plasma membrane.167, These results relied on extrapolationfrom experimental doses, so it is possible that a deviation from linearity mayexist at lower doses. However, increased messenger RNA for CYPIA1 hasbeen detected in rats at doses corresponding to background tissue levels inhumans.168,169 Hence, if there is a threshold for this eect, human tissues mayalready be above it. If the doseresponse behaviors of CYPIA1, CYPIA2, andEGFR are used as surrogates of toxicity, the cancer risks posed by TCDD maybe as high or higher than previously estimated by USEPA.167,170 In sum, themodel proposed by some at the Banbury meeting is incorrect: Action of TCDDthrough the Ah receptor does not necessitate a threshold. The next question, of course, is whether these biochemical markers are rea-sonable surrogates of cancer or other toxic eects. As noted earlier, CYPIA1,CYPIA2, and internalization of EGFR may be related to cancer in the ratliver. On the other hand, liver cell proliferation and preneoplastic foci show nodetectable increase at low doses of dioxin. The doseresponse curve for thesehigher-level biological responses may be nonlinear, but substantial variabilityamong experimental animals clouds the issue.104,128 The current dioxin reassessment is producing biologically more realisticmodels, but these still contain substantial uncertainty, especially with respectto the gap between biochemical markers and more complicated biologicalresponses. Although much of this work has gone into modeling cancer of therat liver, TCDD causes cancer in other organs as well. Since their mechanisticdetails appear to dier, they will require additional modeling eorts. Bio-logically realistic models will also have to address other complications, such asinteractions with other compounds and viruses.* Some people may have been confused by dierent ways of plotting the fraction of occupiedreceptors as a function of ligand concentration. Plotted on a linear scale, the curve is linear at lowdose; plotted on a log scale the curve looks nonlinear at low doses. The latter version was printed inthe Science story covering the Banbury conference.161y The dierence depends on whether the action of TCDD is additive or independent ofexisting processes.167 Heterogeneity of liver cell response suggests a further complication. Lowdoses of TCDD appear to induce CYP1A1 and CYP1A2 maximally in some cells and little or nonein others; increasing the dose turns on more cells.168 34. 20OVERVIEW: THE DIOXIN DEBATE1.6.4 Human EffectsIs dioxin merely a powerful rodenticide? Are humans somehow exempt or lesssusceptible to the biological eects of dioxin? Is it true that chloracne is [the]only adverse eect associated with human exposure?171 Such arguments aresometimes oered for downgrading the toxicity of dioxin. If correct, the allow-able doses of dioxin overstate the real hazard since they are based on animalresearch.The reasons for the primary reliance on animal research are well known.First, the underlying biology of animals and humans is generally similar. Sec-ond, experimentation on humans is usually considered unethical. Only at Hol-mesburg prison, Pennsylvania, were people (other than the scientists involved)deliberately exposed to TCDD to test toxicity, in this case the dose necessaryto cause chloracne. The fate of most of these unfortunate people is notknown.14,172The inadvertent exposure of people to large amounts of dioxinlike com-pounds began at the turn of the century (see Table 1.1). Chloracnea severe,persistent acnelike diseasewas rst described in 18971899 in workers han-dling tarry wastes from the production of chlorine using graphite electrodes.173Chloracne cases were observed during World War I following exposure topolychlorinated naphthalenes; these halowaxes were used in the production ofgas masks.174 Several of the best known occupational exposures to dioxinoccurred around midcentury in facilities manufacturing TCDD-contaminatedherbicides. In the Yusho and Yucheng incidents (see Chapters 21 and 22,respectively), people consumed rice oil contaminated with PCBs, PCDFs, andrelated compounds. Populations were exposed by the use of the herbicideAgent Orange in Southeast Asia, the spreading of dioxin-contaminated wastesin Missouri and the chemical accident at Seveso, Italy (see Chapter 20). Indeed,it was learned in the 1980s that we are all exposed: The general population ofthe industrialized world carries some quantity of PCDD and PCDF in theirbodies.37,175180Information on the eects of dioxin on humans has been obtained fromsome of these experiences by comparing the rates of disease in exposed andreference populations. Given the uncontrolled nature of human exposure, it isimportant to take note of the strengths and weaknesses of these epidemiologicstudies. Certain eects, such as cancer, may not occur until many years, evendecades, after exposure. Other eects (e.g., subtle neurobehavioral abnormal-ities in children) may be missed unless specically looked for. It can be dicultto exclude confounding factors that might contribute to changes in diseaseincidence. Eects may not be detected unless the exposed population is su-ciently numerous, and the dierence in exposure from control groups is rela-tively large. Estimating who was exposed and at what levels is often quite dif-cult. As a result, many supposedly negative studies are in reality merelyinconclusive. When biologically plausible eects are seen in a number of care-fully performed studies, the implications need to be taken very seriously. 35. CANCER 21Qualitative Evidence for Cancer in Humans Much of the debate abouthuman eects has centered on the question of whether TCDD causes cancer inpeople. That it causes cancer in animals should not be in doubt. Given similarbiology, this strongly suggests that it will cause cancer in humans as well. Asof the late 1980s, the human epidemiological evidence was mixed, includingboth positive and negative studies. Uncertainty was increased by the dicultyof establishing exposure and of separating the possible eects of TCDD fromother chemicals to which people were often coexposed (e.g., phenoxyacetic acidherbicides). Some denied any connection between increased human cancer andexposure to phenoxyacetic herbicides and/or their dioxin contaminants.181183Others, such as the Agent Orange Scientic Taskforcea group of indepen-dent scientists on which one of us (B.C.) servedconcluded that there wassucient evidence to legally qualify Vietnam veterans for compensation forseveral types of cancer and disease,184,185 a position that the U.S. governmentlater adopted.186 At that time, the USEPA and the International Agency forResearch on Cancer (IARC) rated TCDD as a probable human carcinogen,based on what they considered sucient evidence in animals and inadequateevidence in humans.143,187The position at that time can be illustrated with two sets of studies.Beginning in the late 1970s, Hardell and others found increased numbers ofa very rare cancer, soft tissue sarcoma (STS), in Swedish forestry and agricul-tural workers exposed to phenoxyacetic acid herbicides and/or chlorophenolswhich are frequently, but not always, contaminated with dioxins.188192 Onthe other hand, several early studies of chemical workers thought to havebeen exposed to high levels of TCDD during the manufacture of herbicidesand chlorophenols were considered negative. In particular, cancer mortalitywas not increased in workers exposed following an accident in 1949 at a Mon-santo facility in Nitro, West Virginia.193195 According to a 1989 report bythe World Health Organization, this was one of only two such incidents thathave been adequately followed up epidemiologically with matched controlgroups.196 Others have called it a major source of information about theeects of high-level dioxin exposure.181 In retrospect, the pioneering work ofHardell et al. on STS has survived criticism better than the Monsanto studies.The latter are at the very least awed by exposure misclassication.197,198Similar controversy surrounds a number of human health studies.199Establishment of exposure played a crucial role in the debates. A technicalbreakthrough came with the ability to measure PCDD and PCDF in humantissues, rst in breast milk,200 and later in adipose tissue and blood.37 SincePCDDs and PCDFs are persistent in the body, this provides a useful measureof past exposure (although the absence of elevated levels does not necessarilypreclude exposure184,201). The U.S. National Institute of Occupational Safetyand Health (NIOSH) followed this approach in its landmark retrospectivecohort study of male American chemical workers thought to have been exposedto TCDD.202 Over 5000 workers were included from 12 plants (includingMonsantos Nitro, West Virginia plant). Blood serum from a sample of 36. 22 OVERVIEW: THE DIOXIN DEBATEworkers was used to validate estimates of exposure made on the basis of workhistory.NIOSH found a 15% increase [relative risk 1:15, 95% condencelevel (CI) 1:02 to 1.30] in total cancer mortality in the entire cohort. In asubcohort of 1520 men with more than one year of exposure and over 20 yearsof latencya group most likely to show eectsthere were increases inoverall cancer (RR 1:46, CI 1:21 to 1.76) as well as mortality from softtissue sarcoma (RR 9:22, CI 1:90 to 26.95) and respiratory tract cancer(RR 1:42, CI 1:03 to 1.92). NIOSH conservatively concluded that theseresults were consistent with the status of TCDD as a carcinogen.202 Afollow-up to the NIOSH study has been published.203 Similar results on lungcancer and all cancers combined have been observed in other occupationalstudies.204,205 Certain types of cancer were increased in Seveso as well.206The increase in mortality from all cancers combined noted in several occu-pational studies is somewhat unusual, as chemical carcinogens are typicallyassociated with a particular organ. In this respect, the human results appearconsistent with the animal experiments, in which TCDD causes cancer atmultiple sites. There are also some parallels between rodents and humans withrespect to certain sites. However, while liver tumors are observed in TCDD-exposed rodents, they are not generally seen in TCDD-exposed people. IfTCDD-related liver cancer in humans is dependent on ovarian hormones as itis in rats (but not mice), this may be related to the fact that most studies haveexamined men.* Primary human liver cancer is very rare outside sub-SaharanAfrica and Asia.208The newer studies shifted the weight of evidence toward considering TCDDas a human carcinogen. The agreement of several studies of occupationallyexposed men with reasonable checks on exposure provided the strongest evi-dence thus far. On this basis, IARC revised its qualitative ranking of TCDDupward from inadequate to limited human evidence. In conjunction with su-cient animal evidence and mechanistic considerations (newly added to its clas-sication system), IARC declared TCDD a human carcinogen in 1997.205 TheUSEPA proposed a similar reclassication in its draft reassessment.8Quantitative Evidence for Cancer in Humans The publication of thelong-awaited NIOSH ndings in January 1991 provided part of the scienticrationale for USEPAs reassessment.165 The study also played a curious rolein the public discussion of dioxin. Although it strengthened the qualitative evi-dence for cancer in humans, it was often portrayed as showing reduced dangerfrom TCDD.209211 Some had the story completely wrong, reporting (with-out qualication) that cancer mortality was not elevated signicantly in thecohort.212,213 Others reported the NIOSH results as indicating that TCDDcauses cancer in humans only at very high doses. NIOSH found statistically* Ovariectomy increases lung cancer in two-stage experiments with TCDD in the female rat, sug-gesting a partial hormonal dependence for this tumor.111 37. SENSITIVE NONCANCER EFFECTS23signicant increased cancer mortality in the entire cohort as well as in a highlyexposed subcohort. They had also analyzed a less exposed subcohort (less than1 year of exposure and over 20 years of latency). Although cancers of somesites were elevated in this group, none were statistically signicant at a 95%level. These results may have been interpreted by some as showing a threshold,in apparent agreement with the alleged consensus of the Banbury conferenceheld only a few months earlier.214 However, the epidemiological evidence isambiguous on this point; such results could arise merely from lack of statisticalpower (i.e., an insucient number of subjects).215Another argument was that much less cancer was observed in the exposedchemical workers than was expected based on rat experiments. The question ofwhether TCDD is less potent in humans than in animals was also discussedduring USEPAs 1988 reassessment.216 The NIOSH data presented a betteropportunity to test this idea. Several groups estimated the carcinogenic slopefactor of TCDD from the increased cancer mortality and the dose these menreceived, projected from the current serum levels. The results were approxi-mately the same as or higher than those derived by the USEPA in 1985 basedon rat data.217,218 Similar results have emerged from newer studies.8,219 Indeed,a recent draft of the USEPA reassessment proposed an increase in the cancerpotency for TCDD.8Although the epidemiological studies of the last decade have yielded impor-tant new evidence for TCDDs carcinogenicity in humans, arguments continueto rage. In the meantime, other signs of danger arose from a completely dier-ent direction.1.7 SENSITIVE NONCANCER EFFECTS1.7.1Livestock and WildlifeTwentieth-century chlorine chemistry exposed livestock and wildlife as well ashumans to dioxinlike compounds (Table 1.1). A mysterious cattle maladyknown as X disease, marked by thickened skin (hyperkeratosis), was describedin 1947220; its cause was later determined to be chlorinated naphthalenes.221Polybrominated biphenylssome with dioxinlike activitywere inadvertentlyadded to cattle feed in Michigan in 19731974.222 This episode has been calledthe most costly and disastrous accidental contamination ever to occur inUnited States agriculture.223 Sickness and death of horses and other animalswas one of the rst signs of trouble at Times Beach, Missouri, where TCDDwas the toxic agent. Chick edema disease was rst described in 1957 in thesoutheastern United States.224,225 Millions of chickens have since died or hadto be killed as a result of consuming feed contaminated with dioxinlike com-pounds. Several outbreaks were caused by feed containing toxic fat derivedfrom animal hides treated with chlorophenols.221,226 Although these episodes with livestock were due to specic contamina-tions, a much more ominous phenomenon appeared in wildlife during the 38. 24OVERVIEW: THE DIOXIN DEBATE1960s.227,228 Epidemics of reproductive and developmental problems havesince been observed in sh-eating birds from the Great Lakes and elsewhere.There is good evidence that some of these episodes were caused by dioxinlikecompounds. The Great Lakes embryo mortality, edema and deformities syn-drome (GLEMEDS) has very close parallels to chick edema disease. The eectsare consistent with the results of laboratory studies with dioxinlike com-pounds.227,228 A strong correlation was found between the egg mortality ofdouble-crested cormorants and the levels of dioxinlike compounds found in theeggs, as measured by enzyme induction.229 Forsters terns from Green Bayhatched fewer eggs in 1983 than those from a less contaminated area. Eggsfrom the contaminated area had signicantly higher levels of dioxinlike com-pounds. A cross-fostering experiment in which eggs were switched between thetwo areas showed that parental behavior played a role as well. Adults from thecontaminated area took less care of their eggs.230 The neurotoxic eects ofcertain nondioxinlike PCBs may have contributed to the latter problem.10Dioxinlike compounds may cause bill deformities and embryonic abnormalitiesin double-crested cormorants and Caspian terns.231233 Although certain locations, such as Green Bay, had a particularly high inci-dence of GLEMEDS, the problem appeared to be relatively widespread in theGreat Lakes. Egg mortality and bill defects of double-crested cormorants weregenerally more prevalent in Great Lakes birds than in reference areas.229,231Persistent dioxinlike PCBs were probably the major problem,229,230 althoughoutbreaks among Lake Ontario herring gulls in the 1970s were probably causedmainly by TCDD discharged from chemical manufacturing and waste dumpson the Niagara River.227 Reproductive and developmental problems have also been observed in laketrout, which are particularly sensitive to TCDD during early development.Failure of restocked sh to reproduce in Lake Ontario during the 1970s wasprobably related to TCDD and dioxinlike compounds.84,234 The decreasednumber of mink and otter found around the Great Lakes may have been con-nected to PCBs and dioxinlike compounds.235 Great blue heron chicks livingnear a pulp mill in British Columbia suered depressed growth and greateredema than did birds from a less contaminated area.236 Crossed bills hav