22 - reserch needs for the risk assessment of health and environmental effects of.. (epa)

Research Needs for the Risk Assessmentof Health and Environmental Effectsof Endocrine Disruptors: A Reportof the U.S. EPA-sponsored WorkshopRobert J. Kavlock,1 George R Daston,2 Chris DeRosa,3 PennyFenner-Crisp,4 L. Earl Gray,1 Steve Kaattari,5 George Lucier,6Michael Luster,6 Michael J. Mac,7 Carol Maczka,8 Ron Miller,9Jack Moore,10 Rosalind Rolland,"1 Geoffrey Scott,12 DanielM. Sheehan,13 Thomas Sinks,14 and Hugh A. Tilson1The hypothesis has been put forward that humans and wildlife species have suffered adversehealth effects after exposure to endocrine-disrupting chemicals. Reported adverse effects includedeclines in populations, increases in cancers, and reduced reproductive function. The U.S.Environmental Protection Agency sponsored a workshop in April 1995 to bring together interest-ed parties in an effort to identify research gaps related to this hypothesis and to establish priori-

ties for future research activities. Approximately 90 invited participants were organized into workgroups developed around the principal reported health effects-carcinogenesis, reproductive toxi-city, neurotoxicity, and immunotoxicity-as well as along the risk assessment paradigm-hazardidentification, dose-response assessment, exposure assessment, and risk characterization.Attention focused on both ecological and human health effects. In general, the group felt that thehypothesis warranted a concerted research effort to evaluate its validity and that research shouldfocus primarily on effects on development of reproductive capability, on improved exposure

assessment, and on the effects of mixtures. This report summarizes the discussions of the workgroups and details the recommendations for additional research. Environ Health Perspect1 04(Suppl 4):715-740 (1996)

Key words: endocrine disruptors, hormones, risk assessment, carcinogenesis, reproductivetoxicity, developmental toxicity, immunotoxicity, neurotoxicity, exposure assessment,research needs

IntroductionEvidence has been accumulating whichindicates that humans and domestic andwildlife species have suffered adverse healthconsequences from exposure to environ-mental chemicals that interact with theendocrine system (e.g., 1-3). To date,these health problems have been identifiedprimarily in domestic or wildlife species

with relatively high exposures to organo-chlorine compounds, including 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane(DDT) and its metabolites, polychlorinat-ed biphenyls (PCBs) and dioxins, or tonaturally occurring plant estrogens. It isnot known if similar effects are occurringin the general human population, but

again there is evidence of adverse effects inpopulations with relatively high exposures.Several reports (4) of declines in the quali-ty and decreases in the quantity of spermproduction in humans over the last fourdecades and reported increases in inci-dences of certain cancers (breast, prostate,testicular) that may have an endocrine-related basis have led to speculation aboutenvironmental etiologies. However, con-siderable scientific uncertainty remainsregarding the causes of these reportedeffects. Nevertheless, it is known that thenormal functions of all organ systems areregulated by endocrine factors, and smalldisturbances in endocrine function, especial-ly during certain stages of the life cycle suchas development, pregnancy, and lactation,can lead to profound and lasting effects.The critical issue is whether sufficiently highlevels of endocrine-disrupting chemicals existin the ambient environment to exert adversehealth effects on the general population.

Current methodologies for assessinghuman and wildlife health effects (e.g., thegeneration of data in accordance with test-ing guidelines developed by the U.S.Environmental Protection Agency [U.S.EPA]) are generally targeted at detectingeffects rather than mechanisms, and maynot adequately evaluate effects on theendocrine system. This is particularly truefor exposures that occur during criticaldevelopmental periods when the endocrinesystem plays a key role in regulating essen-tial physiological and morphologicalprocesses. Given the potential scope of theproblem, the possibility of serious adverseeffects on the health of human and wildlifepopulations, and the broad occurrence andpersistence of some endocrine-disruptingagents in the environment, it is importantto focus the available resources for researchon the most critical gaps in our knowledge

Disclaimer: This document has been reviewed in accordance with U.S. Environmental Protection Agency policy and approved for publication. Mention of trade names orcommercial products does not constitute endorsement.

Address correspondence to Dr. Robert J. Kavlock, National Health and Environmental Effects Research Laboratory, U.S. EPA, Research Triangle Park, NC 27711.Telephone: (919) 541-2326. Fax: (919) 541-1499. E-mail: [email protected]

1National Health and Environmental Effects Research Laboratory, U.S. EPA, Research Triangle Park, NC; 2Miami Valley Laboratories, The Procter and Gamble Co.,Cincinnati, OH; 3The Agency for Toxic Substance and Disease Registry, Atlanta GA; 4The Office of Prevention, Pesticides, and Toxic Substances, U.S. EPA, WashingtonDC; 5Virginia Institute of Marine Science, The College of William and Mary, Williamsburg, VA; 6National Institute of Environmental Health Sciences, Research TrianglePark, NC; 7National Biological Service, Washington, DC; 8National Research Council, Washington, DC; 9The Dow Chemical Company, Midland, Ml; 1Olnstitute forEvaluating Health Risks, Washington, DC; "World Wildlife Fund, Washington, DC; 12National Marine Fisheries Service, Charleston, SC; 13National Center for ToxicologicalResearch, U.S. FDA, Jefferson, AR; 14Centers for Disease Control and Prevention, Atlanta, GA.

Abbreviations used: CNS, central nervous system; DDE, 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene; DDT, 1,1,1-trichloro-2,2-,bis(p-chlorophenyl)ethane; DES, diethyl-stilbestrol; DHEA, dehydroepiandrosterone; EDC, endocrine-disrupting chemical; EROD, ethoxyresorufin-0-deethylase; LH, luteinizing hormone; MUNG, N-methyl-N-nitroso-N'-nitroguanidine; NTP, National Toxicology Program; NOAEL, no observed adverse effect level; PAHs, polyaromatic hydrocarbons; PBBs, polybrominatedbiphenyls; PCBs, polychlorinated biphenyls; PCDFs, polychlorinated dibenzofurans; QSAR, quantitative structure-activity relationship; SAR, structure-activity relationship;SEER, surveillance epidemiology and end results; TCDD, 2,3,7,8-tetrachlorodibenzo-p-dioxin; TEF, toxic equivalency factor; TEQ, toxic equivalent approach; TIE, toxicityidentification evaluation.

Environmental Health Perspectives - Vol 104, Supplement 4 * August 1996 715

KAVLOCK ET AL.

base so that more informed regulatory andpublic health decisions can be made in thefuture. The broad nature of the problemnecessitates a coordinated effort on boththe national and the international levels.The National Science and TechnologyCouncil, which advises the president and hisCabinet on directions for federal researchand development efforts, has established amilestone for 1995 to 1998 to produce anational research strategy on endocrine-dis-rupting chemicals. Therefore, in response tothe growing public health concerns relatedto chemicals in the environment that havethe potential to act as endocrine disruptors,the Office of Research and Development ofthe U.S. EPA held a workshop on April10-13, 1995, in Raleigh, North Carolina,to begin developing a national researchstrategy related to endocrine-disruptingchemicals. An organizing committee wasformed that consisted of representativesfrom various organizations, including theU.S. EPA, Department of Health andHuman Services, Department of theInterior, Department of Commerce, andDepartment of Agriculture; industrialgroups such as the Chemical ManufacturersAssociation and the American IndustrialHealth Council; independent organizationssuch as the Institute for Evaluating HealthRisks; and public interest groups suchas the World Wildlife Fund and theEnvironmental Defense Fund.

The premise of the workshop was asfollows: because environmental endocrinedisruptors have caused a variety of adversebiological effects in wildlife species, domes-tic animals, and humans, we need to iden-tify specific research that would assist thefederal government in making informeddecisions. An environmental endocrinedisruptor was broadly defined as "an exoge-nous agent that interferes with the produc-tion, release, transport, metabolism, bind-ing, action or elimination of natural hor-mones in the body responsible for themaintenance of homeostasis and the regu-lation of developmental processes." Thisdefinition reflects a growing awareness thatthe issue of endocrine disruptors in theenvironment extends considerably beyondthat of exogenous estrogens and includesantiandrogens and agents that act on othercomponents of the endocrine system suchas the thyroid and pituitary glands.Approximately 90 invited participants andmembers of the organizing committee(Table 1) were asked to discuss researchneeds related to the principal adversehealth effects reported for endocrine

disruptors (carcinogenesis, reproductive[including developmental] toxicity,immunotoxicity, and neurotoxicity) as wellas research to improve specific componentsof the risk assessment paradigm (hazardidentification, dose-response assessment,exposure assessment, and risk characteriza-tion). Attention focused on both ecologicaland human health effects. A series of ques-tions was posed to each group to helpguide the discussions. More than 200observers from academia, industry, govern-mental organizations, public interestgroups, and the press also attended theworkshop. This report summarizes themajor findings of each discussion group.

Dr. Lynn Goldman, Assistant Adminis-trator, Office of Prevention, Pesticides andToxic Substances (OPPTS), U.S. EPA,opened the workshop. She reviewed theimpact of the National PerformanceReview on environmental protection, theprocess of updating the OPPTS testingguidelines, and some regulatory activitiesrelated to endocrine disruptors (e.g., thespecial review of triazines, an evaluationof endosulfan, and the status of thealkylphenol ethoxylate consent order). Herpresentation emphasized the need toaddress the major scientific and policyquestions as the foundation for mitigatingthe potential impact of endocrine disrup-tors. Dr. Howard Bern (University ofCalifornia at Berkeley) gave the keynoteaddress. This historical perspective wasbased largely on the human diethylstilbes-trol (DES) syndrome and on the neonatalmouse model used in its experimentalanalysis. He emphasized long-term perma-nent effects in the adult as a result of expo-sure to agents during development, whichcan occur without apparent birth defects inthe neonate. Dr. Bern also emphasized theconcept of critical periods for epigeneticeffects on different targets and indicatedthe wide range of organs and physiologicalsystems that may be affected-reproductive,endocrine, immune, neural, behavioral,metabolic, skeletal, etc. The particular sen-sitivity of systems to endocrine-disruptingagents during development implies thatembryonic, fetal, and neonatal tissues may"see" estrogens, estrogen-mimics, and otherendocrine disruptors in a different way(perhaps even by different mechanisms)than adult tissues. As a final prelude to thediscussions, representatives from Germany(Dr. Andreas Gies, Umwelt Bundes Amt),the United Kingdom (Dr. Linda Smith,Department of the Environment), andDenmark (Dr. Jorma Toppari, University

of Turku, Turku, Finland, on behalf of theDanish Ministry of Environment andEnergy and the Danish EnvironmentalProtection Agency) presented summaries ofresearch needs identified by their respectivegovernments in recent workshops (4-6).

General Commentsfrom the Work GroupsEach work group consisted of individualswith various backgrounds including fieldecology, epidemiology, basic sciences, ani-mal and human toxicology, exposureassessment, and risk assessment. This mixof experts was perceived as a great advan-tage in enhancing the groups' ability tolook at the overall problem, stimulatingestablishment of common priorities, andidentifying new solutions to existingmethodologic issues. An interdisciplinaryapproach should be maintained in any fol-low-up action to this workshop. Consistentwith the interdisciplinary approach, it isimportant that methods and results foundin one research arena be applied to otherresearch arenas. Basic research (e.g., mech-anistic discoveries in cancer etiology)should be applied to observational research(e.g., use of DNA polymorphisms in epi-demiological or field ecology studies) andvice versa (e.g., identification of the mecha-nisms by which DES increases clear-cellcarcinoma of the vagina in women). Soundscientific information must be the basis ofgood decision making. We should be care-ful not to overinterpret study results whenthe basic scientific techniques have notbeen standardized or validated. The use ofbiologic measurements requires an under-standing of good laboratory practices(QA/QC), reproducibility, accuracy, statis-tical power, and replicability. In addition,interpretation that some measurements arepredictive of adverse biological effectsrequires validation studies. Finally, theunderlying issue of dose response mustalways be considered when results observedover a limited dose range are evaluated .

Many of the gaps in our understandingof dose-response relationships for endocrinedisruptors are the same as those that haveled to much uncertainty and associated con-troversy arising from using default assump-tions made by regulatory agencies in the riskassessment process. The validity of assump-tions used in risk assessments is frequentlychallenged when high-dose data in experi-mental systems are used to estimate effectsat much lower doses in humans. Our grow-ing analytical ability to detect chemicals atcontinually lower concentrations, coupled

Environmental Health Perspectives * Vol 104, Supplement 4 * August 1996716

ENDOCRINE DISRUPTOR RESEARCH NEEDS

Table 1. Work group participants and assignments.

Participanta Affiliationb Groupc

Mel AndersonGerald AnkleyaChristopher J. BayneDavid BellingerHoward A. BernLinda BirnbaumAaron BlairJoanna BurgerCynthia CareyJanice E. ChambersRobert E. ChapinJames R. ClarkaTheodora ColbornaRory ConollyJon CookRalph CooperJohn CouchDavid CrewsSally DarneyGeorge DastonaWilliam P. DavisChris DeRosaaPenelope Fenner-CrispWarren FosterMichel FournierGlen FoxD. Michael FryMichael A. GalloDavid GaylorEllen GoldeyTom GoldsworthyJay GoochL. Earl Gray JraLouis J. GuilletteMaureen C. HatchJerry D. HendricksAndrew G. HendrickxDiane HenshelDavid E. HintonMike HolsappleClaude L. Hughes JrLyndal JohnsonRod JohnsonSteven L. KaattariWilliam KelceCarole Kimmel

ICF KaiserU.S. EPAOregon State UniversityBoston's Children's HospitalUniversity of California-BerkeleyU.S. EPANCIRutgers UniversityUniversity of ColoradoMississipi State UniversityNIEHSExxon Biomedical SciencesWorld Wildlife FundChemical Industry Institute of ToxicologyDupont-Haskell LaboratoryU.S. EPAU.S. EPAUniversity of Texas at AustinU.S. EPAThe Procter and Gamble Co.U.S. EPAATSDRU.S. EPAEnvironmental Health CanadaUniversite du Quebec a MontrealCWS, Environment CanadaUniversity of California-DavisRobert Wood Johnson Medical SchoolNCTRU.S.EPAChemical Industry Institute of ToxicologyThe Procter and Gamble Co.U.S. EPAUniversity of Florida-GainesvilleMt. Sinai School of MedicineOregon State UniversityUniversity of California-DavisIndiana UniversityUniversity of California-DavisDOW Chemical Co.Wake Forest UniversityNorthwest Fisheries Science CenterU.S. EPAViMS, College of William and MaryU.S. EPAU.S. EPA

C,DRR,HD1,DRN,DRR,HDC,EXC,HDN,HDl,EXN,HDR,RCR,RCl,EXC,DRC,HDN,HDC,HDN,RCR,RCRt,DRR,EXC,EXtC,RCtR,RCO,DRl,EXR,HDRCN,RCN,HDC,DRR,HDR,HDtR,HDR,EXC,HDR,RCN,DRC,EX1,RCR,DRR,RCC,RCIt,RCR,HDR,RC

Participanta

Garet LahivsCoral A. LamartiniereJohn F. LeatherlandJonathan J. LiGeorge LucieraMike LusterMichael J. MacaNeil J. MacLuskyCarol MaczkaPeter MathiessenLynne F. McGrathJohn McLachlanSue McMasteraMark S. MeyersDiane MillerRon R. MilleraJohn A. MooreaLarry L. NeedhamReynaldo PatinoRichard E. PetersonWarren P. PorterChristopher J. PortierWalter RoganRosalind M. RollandaLouise M. RyanGeoffrey I. ScottaR. Woodrow SetzerDaniel M. SheehanaBarbara B SherwinEllen K. SilbergeldaThomas SinksaRalph SmialowiczGeorge StancelJohn J. StegemenPeter ThomasDonald TillittHugh TilsonJorma ToppariKamala TripathiaDaniel A. ValleroFrederick S. Vom SaalChris WallerPatricia WhittenElizabeth WilsonJudith T. Zelikoff

Affiliationb Groupc

University of Maryland-BaltimoreUniversity of Alabama-BirminghamUniversity of GuelphUniversity of Kansas Medical CenterNIEHSNIEHSNational Biological ServiceThe Toronto HospitalNational Research CouncilMAFF, United KingdomHoechst-CelaneseTulane UniversityU.S. EPANorthwest Fisheries Science CenterU.S. EPADOW Chemical Co.IEHRCDCNational Biological ServiceUniversity of Wisonsin-MadisonUniversity of Wisconsin-MadisonNIEHSNIEHSWorld Wildlife FundHarvard UniversityNational Marine Fisheries ServiceU.S. EPANCTR/FDAMcGill UniversityUniversity of Maryland-BaltimoreCDCU.S. EPAUniversity of Texas Medical SchoolWoods Hole Oceanographic InstituteUniversity of Texas at AustinNational Biological ServiceU.S. EPAUniversity of Turku, FinlandUSDAU.S. EPAUniversity of Missouri-ColumbiaU.S. EPAEmory UniversityUniversity of North Carolina-CHNew York University Medical Center

aOrganizing committee members (Dick Hill, U.S. EPA, and Jim Reisa, NRC, were unable to attend). bATSDR, Agency for Toxic Substances and Drug Reserch; CDC, Centers forDisease Control; CWS, Canadian Wildlife Service; FDA, Food and Drug Administration; ICF Kaiser; MAFF, Ministry of Agriculture, Fisheries and Food (United Kingdom); NCI,National Cancer Institute; NCTR, National Center for Toxicological Research; NIEHS, National Institute of Environmental Health Sciences; U.S. EPA, U.S. EnvironmentalProtection Agency. CC, carcinogenic effects, R, reproductive effects; N, neurological effects; immunological effects; H D, hazard detection methods; DR, dose-response meth-ods; EX, exposure methods; RC, risk characterization methods; t, chairperson; t rapporteur.

with the availability of molecular approachesto detect chemical interactions with biologi-cal systems, is creating opportunities andmore tractable approaches to improve low-dose risk estimates. Yet considerable workremains to be done at the laboratory andscience policy levels before there is wide-spread acceptance of mechanistic data inquantitative risk assessments. We need to

improve the use of existing data and to havebetter data and predictive models to

strengthen the scientific foundation for esti-

mating dose-response relationships forendocrine disruptors.

Several considerations are essential to

the examination of the effects of potentialendocrine disruptors. First is the consid-eration of the different sensitivities at

different ages. In general, the developingorganism is especially sensitive, for exam-

ple, DES induction of adenocarcinoma ofthe vagina in females exposed before theend of the first trimester of their mother'spregnancy. The second consideration is a

direct consequence of the first. Because thework group recognized the importance ofthe developmental stage at exposure, expo-

sure assessment as defined by the NationalResearch Council (7) was modified (changenoted in italics) to read: "the process ofmeasuring or estimating the intensity,frequency, duration, and the timing ofexpo-sure, of humans and wildlife to an agentcurrently present in the environment or ofestimating hypothetical exposures thatmight arise from releases of new chemi-


1,DRC,EXN,EXC,EXC,DRtIt,DRR,EXtN,RCR,HDtR,EX1,DRC,RCN,RCC,RCN,RCCt,R1,RCtC,EXR,EXR,RCN,RC1,DRN,RCRf,HDR,DRNt,RR,DRR,DRtN,RCN,EXCt,EXI,HDR,DRC,EXR,DRR,EXNt,HDR,IDl,EXN,EXR,DRC,HDR,DRR,HDl,EX

7 17

KAVLOCK ET AL.

cals." The third consideration is that adverseeffects can arise from either primary or sec-ondary disturbances of endocrine function.Thus, an indirect-acting endocrine disruptoraffects a systemic target organ first; theseeffects in turn may influence the endocrinesystem to cause secondary neurotoxicity,reproductive toxicity, and/or immunotoxi-city. Conversely, a direct-acting endocrinedisruptor affects the endocrine system first,which in turn results in toxicity in otherorgan systems. In some aspects it is verydifficult, if not impossible, to separate theendocrine system from the systemic targetorgans, so this distinction should be viewedin an abstract manner. Obviously, somechemicals can adversely affect the structureand/or function of the systemic organswithout any endocrine system involve-ment. Such chemicals may be considereddirect-acting target organ toxicants. Theserelationships are portrayed in Figure 1.Finally, the development of a researchagenda for endocrine disruptors should bea national or international effort by manyagencies, not just the U.S. EPA, so that thelimited resources can be used in the mostefficient manner.

Biological Effects IssuesCarcinogenic EffectsWhat do we know about the carcinogeniceffects of endocrine-disrupting agents inhumans and wildIife? What are the majorclas5es ofchemicals thought to be responsi-ble for these effects? What are the uncer-tainties associated uith the reported effents?Numerous field studies of teleost fishes inlocalized highly contaminated areas (i.e.,"hot spots") have shown high prevalencesof liver tumors (8-10). The predominantrisk factor that has been associated withthese liver tumors is exposure to poly-aromatic hydrocarbons (PAHs) and to alesser degree, PCBs and DDT. Certainspecies such as carp and fathead minnowsare more resistant, while trout are moresensitive. There has been no indicationthat the liver tumors in fish involve anendocrine modulation mechanism. Otherthan for localized areas of high contamina-tion, field studies have shown no increasingtrends for tumors of any type in fish. Thegroup noted that two tumor registries forwildlife species exist in the United States(the Smithsonian Registry of Tumors inLower Animals under the direction ofDr. John Harshbarger, and the ArmedForces Institute of Pathology's Registry ofComparative Pathology under the direction

Direct actingendocrine disruptor

Direct acting targetorgan toxicant

_ Indirect actingendocrine disruptor

Figure 1. Abstract representation of the interplaybetween the endocrine system and the reproductive,neurological and immunological systems to illustratethe complexity of determining the mode of action forchemicals that cause adverse effects through involve-ment of the endocrine system.

of Dr. Linda Johnson). A variety of dose-related tumors can be produced in fishgiven carcinogens under experimentallaboratory conditions (11). Again, there isno specific evidence that the developmentof these tumors involves a hormonal dis-ruption mechanism. Estradiol and certainhormone precursors (e.g., dehydroepi-androsterone [DHEA]) act as promotorsafter treatment of fishes with carcinogenicsubstances such as aflatoxin and N-methyl-N-nitroso-N'-nitroguanidine (MNNG).Toxicopathic liver lesions have been associ-ated with contaminant exposure in somemarine fish (10).

There is a paucity of carcinogenicitydata for other forms of wildlife. One studyof beluga whales in the St. Lawrence sea-way found that approximately 50% ofdead whales examined had neoplasms, ofwhich about 25% were malignant (12,13).

The hypothesis that endocrine disrup-tion can cause cancer in humans is basedon the causal association between DESexposure of pregnant women and clear-cell adenocarcinoma of the vagina andcervix in their female offspring, hormone-related risk factors for breast and uterinecancer, and limited evidence of an asso-ciation between body burden levels of1, 1-dichloro-2,2-bis(p-chlorophenyl)ethyl-ene (DDE) or PCBs and breast cancerrisk. Young women who developed cancerof the vagina were more likely to have hadmothers who used DES during pregnancyto avoid miscarriage than mothers who didnot use the drug (14). This finding has ledto a number of important conclusions.First, maternal exposures during gestationcan lead to cancer in offspring, and second,it demonstrates that a synthetic estrogencan cause cancer. Some of the male off-spring of women who took DES displaypseudohermaphroditism (15) and genitalmalformations, including epididymal cysts,testicular abnormalities such as small testesand microphallus, and reduced semen

quality (16-18). Follow-up surveys ofDES-exposed male offspring, however, havenot shown impairment in fertility or sexualfunction (19,20), nor is there evidence ofincreased risk of testicular cancer (19).

The most common cancer amongwomen in the United States is breast can-cer. A number of epidemiological studieshave examined the risk factors for breastcancer. Identified risk factors include sever-al that relate to hormonal activity:decreased parity, age at first delivery, age atmenarche, age, race, and unopposed estro-gen therapy. In addition, breast tumors canbe characterized as to their degree of estro-gen-receptor positivity resulting in relevantprognostic information. The evidence sup-ports a causal relationship between femalebreast cancer and hormonal activity.

A number of organochlorine pesticidesor pesticidal metabolites are found inbreast milk and human adipose tissue(21,22). Several recent cross-sectional stud-ies suggest a possible relationship betweenlevels of some organohalide residues inhuman tissues and breast cancer risk,although the observations are not entirelyconsistent across studies, and no clear rela-tionship has been established (23-30). Ingeneral, these studies suggested that levelsofp,p'-DDE and total PCBs were higher infat or serum of women who had breastcancer than in comparison groups. Themeaning of these findings is unclear, inpart, because p,p'-DDE and the few PCBcongeners that have been tested have littleor no discernible estrogenic activity, whilethe short-lived forms of DDT, o,p'-DDTand o,p'-DDE, have only very weak estro-genic properties. Further, a recentcase-control study with historical datafrom serum DDE and PCBs conflicts withthe earlier findings (29). This studyshowed no overall effect of serum residuelevels on breast cancer risk, although sub-categorical analysis did suggest a possibleincrease in risk among black women withhigher levels of serum p,p'-DDE. Thewomen in these studies, except those inthe study by Henderson et al. (30), werenot exposed to high levels of PCBs orDDE and the actual differences in levelsmeasured between cases and controls werenot large. Studies ofwomen occupational-ly exposed to high levels of PCBs have notdemonstrated an excess risk of breast can-cer mortality (31,32). The results of thesestudies therefore are equivocal and furtherresearch is needed (including examinationof effects in subsequent generations fromparental exposures).



Relatively good information exists oncancer occurrence (incidence) and mortali-ty in the United States over the last severaldecades. The best incidence data comefrom the Surveillance Epidemiology andEnd Results (SEER) cancer registries thatare supported by the National CancerInstitute (33). Cancer trend data from1973 to 1991 show that age, race, and sex-adjusted total cancer incidence increasedby 31% in males and 14% in females. Inci-dence rates between 1973 and 1991 forseveral hormone-sensitive tissues haveincreased (female breast 24%, ovarian 4%,testicular 41%, prostate 126%) as have sev-eral other cancer sites: melanoma (116% inmales, 73% in females); non-Hodgkinslymphoma (84% in males, 57% infemales); and liver (55% in males, 27% infemales). However, increases in femalebreast and male prostate cancers accountfor the majority of total cancer increasesexperienced by women (52%) and men(70%) during this time period. Theincreases in testicular and ovarian cancerrepresent only 1% of the total increase incancer incidence. SEER data indicate thatincidences of uterine cancer and malebreast cancer have remained constant ordeclined slightly.

Cancer screening is available for breastand prostate cancer. Recent advances in thescreening process account for some of thereported increases in cancer incidence.White et al. (34) reported that althoughthe increase in breast cancer incidencefor women 25 to 44 years of age can beexplained by screening, screening does notentirely explain the increase for younger orolder women. Feuer and Wun (35) suggestthat screening may also account for all ofthe observed increase in older women. Ithas also been suggested that increaseddetection of prostatic cancer is due toincreased screening for prostate-specificantigen (PSA) (36).

Examination of the U.S. EPA databaseon pesticide registration for organochlo-rines showed no correlation of the spec-trum of tumor types observed in laboratoryanimals with the assertion that theorganochlorines are related to humanbreast cancer. Organochlorines frequentlyincreased the incidence of liver tumors inrats, but did not increase the incidence ofmammary tumors (P. Fenner-Crisp, person-al communication to work group). Onesubclass of herbicides, the chloro-S-triazines,produces an earlier onset of mammarytumors in Sprague-Dawley rats (37,38),but there are no epidemiological studies

that suggest a relationship between expo-sure to triazine herbicide and human breastcancer. An examination of the NationalToxicology Program (NTP) databaseinvolving approximately 450 animal stud-ies showed increased incidences of mam-mary tumors in approximately 10% of thestudies. However, based on evaluation ofchemical structures and other availableinformation, this subset of test substancesis not likely to be estrogenic (39). Thisanalysis only considered the possibility thatthe chemicals were direct estrogen agonists.It is obvious that other possible mecha-nisms exist for induction of endocrine-mediated tumors.

It was noted that good animal or cellu-lar models do not yet exist for study ofsome endocrine-mediated tumors (e.g., tes-ticular), which are reported to be on theincrease in the human population, and thatsuch models would be useful in testingcause-and-effect relationships.

What are the research needs related to thedetection of carcinogenic effects ofendocrine disruptors?Most ecological field studies of cancerincidence have been devoted to examiningfish in polluted waters. Some efforts havebeen made to establish comparison popu-lations by examining nonpolluted waters.One study collected data that could beused for background information (11).There are no ongoing national surveillanceprograms for tracking cancer incidenceand mortality in wildlife, but theSmithsonian Institution does maintain atumor registry for fish and wildlife, andsome marine mammal populations are beingtracked for tumors (12,13). When popula-tions of wildlife are evaluated, it is impor-tant to establish estimates of the expectedoccurrence of cancer. It is also importantto identify special populations at high riskfor cancer on the basis of high exposure orincreased susceptibility. Background datacould then be used for comparisons. Fishhave proven to be the easiest animals tostudy. The occurrence of cancers in otheranimals should be examined as well.

In environmental settings, certain sen-tinel species may be useful. Sentinel specieswould include wildlife, domestic animals,or laboratory animals. The goal would beto identify species that are susceptible todeveloping cancer, easily monitored forcancer, and likely to reflect exposure in asingle ecosystem.

Last, the evaluation of exposures shouldbe expanded beyond polycyclic aromatic

hydrocarbons and look at additional typesof endocrine disruptors.

The study of the relationship betweenendocrine disruption and female breastcancer in humans is an obvious priority,not only because of the high incidence ofthis disease, but also because no primaryprevention is available. Hopefully, the rela-tionship between organochlorine exposureand breast cancer risk will be resolvedsoon, given the large number of ongoingepidemiological studies. Hence, focusshould be placed on additional issues suchas research on other possible environmentalagents, including persistent and nonpersis-tent (e.g., nonorganochlorine pesticidesand phthalates) environmental chemicalsthat may affect the endocrine system.We should evaluate the relationship

between exposure to endocrine disruptorsand other cancer sites, particularly prostate,testicular, ovarian, endometrial, and thy-roid. It is also important to evaluate can-cers of the liver, brain, and lung, as thesecancers are frequently associated with expo-sure to environmental hazards, althoughthere is no evidence that exposure to EDCsis among the known primary risk factorsfor these tumors.

Epidemiological data on exposedhuman populations have proven useful foridentification of human carcinogens.Occupational cohorts and cohorts of per-sons with exceptionally high environmentalexposures [e.g., 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in Seveso, Italy (40);polychlorinated biphenyls (PBBs) inMichigan (41); and PCBs and polychlori-nated dibenzofurans (PCDFs) in Japan (42)and Taiwan (43)] may be important forestablishing a dose-response relationship.The highest priority is to identify, register,and follow populations with documentedand quantitatively verified exposures.

There was discussion regarding theneed to include in utero exposures in stan-dard lifetime cancer bioassays to ensureprotection of the developing offspring. Theavailable evidence (44-47) does not clearlydemonstrate a significant increase in sensi-tivity or induction of tumors compared topostnatal-only exposures. However, othershave suggested that these retrospectivecomparisons are not adequate to determinethe need for such exposures with EDCs,since that mechanism of action is under-represented in the historical database.

Domestic animals and pets may beuseful to identify carcinogenic hazards.

Environmental Health Perspectives - Vol 104, Supplement 4 - August 1996 7 19

KAVLOCK ET AL.

What are the highestpriority research needsfor carcinogenic effects?The following high priority research needswere identified:

a) A systematic comparison of endoge-nous versus exogenous substances in termsof biologic activity, metabolism, struc-ture-activity relationships, etc., with aninitial focus on estrogenicity and rapidexpansion to other steroid hormones. Theexogenous estrogenic substances shouldinclude comparisons of phytoestrogensas well as other types of xenoestrogens inrelation to endogenous estradiol.

b) Basic research to systematically andthoroughly evaluate species-, cellular-, andage-dependent responses, including consid-eration of mixtures of agonists, partial ago-nists, and antagonists, at environmentallyrelevant ratios and doses.

c) Careful evaluation of toxicity andmechanistic end points across species, includ-ing the determination of dose-responserelationships in relation to human risk.

d) Surveillance data on the occurrenceof tumors in wildlife species.

e) Identification and follow-up healthstudies of heavily exposed wildlife andhuman populations.

f) Validation and application of bio-markers that might be useful in identifyingexposures as well as adverse outcomesbased upon mechanistic considerations.

Other additional research recommen-dations included thorough consideration ofthe importance of critical timing (i.e., win-dows) of exposure before and after birth forcarcinogenicity end points; determinationof the role of metabolism (e.g., estradiol) inrelation to certain toxicities such as breastcancer; identification of susceptibility fac-tors such as polymorphisms that mightresult in a predisposition to certain endpoints, and assessment of structure-activityrelationships for hormonal activity.

Reproductive Effects

What do we know about the reproductiveand developmental effects of endocrine-disrupting agents in humans and wildlife?What are the major classes of chemicalsthought responsible for these effects? Whatare the uncertainties associated with thereported effects?Field and laboratory studies of wildlifepopulations and individuals have revealedeffects in offspring that appear to be theresult of endocrine disruption. Examplesinclude reproductive problems in woodducks from Bayou Meto, Arkansas (48);

wasting and embryonic deformities inGreat Lakes fish-eating birds (49-56);feminization and demasculinization ofgulls (57-60); developmental effects inGreat Lakes snapping turtles (61); embry-onic mortality and developmental dysfunc-tion in lake trout and other salminiods inthe Great Lakes (62-64); abnormalities ofsexual development in Lake Apopka alliga-tors (65,66); reproductive failure in minkfrom the Great Lakes area (67); and repro-ductive impairment in the Florida Panther(68). In each case, detectable concentra-tions of chemicals with known endocrine-disrupting effects have been reported in theanimals or in their environment, but anetiological link has been established foronly a few of these observations. In ecolog-ical studies, these effects were not recog-nized until the populations began todecline. However, the observation that apopulation is stable is not an assurance thatendocrine-disrupting chemicals are notaffecting reproduction, development,and/or growth of individuals.

In humans, there is evidence of adversereproductive outcomes in the DES sonsand daughters (see "Carcinogenic Effects" )and in male offspring from the Yu-Chengpoisoning incident (69). In adults, repro-ductive dysfunction has been observed inmales exposed to kepone (70), and theduration of lactation has been reported todecrease as the concentration of DDE inthe milk increases (71,72). Egeland et al.(73) reported decreased serum testosteroneand increased luteinizing hormone (LH)levels in workers exposed to dioxin. Therealso have been reports of declining spermquality in males over the last several decades(74,75), but the etiology is far from certain.The most convincing evidence for a gener-al decline in male reproductive healthin humans is the increase in testicular can-cers noted over the recent past in severalWestern countries (5). Again, the contri-bution of environmental contaminants tothis increased rate is unknown.

Further information on reproductiveeffects of endocrine-disrupting chemicalscan be found in Colborn and Clement (1),Adams (2), Medical Research Council (4),Danish Environmental Protection Agency(5), Umweltbudesamt (6), McArthur et al.(76), Hoffman (77), Kihlstrom et al. (78),Colborn et al. (79), Jansen et al. (80),Jobling and Sumpter (81), Kelce et al.(82), Patnode and Curtis (83), Bergeron etal. (84), Guillette (85), Baldwin et al.(86), Dodson and Hanazato (87), Rollandet al. (88), and Kelce et al. (89). There are

multiple targets for these chemicals, andeffects are organ- and life stage-specific.Still other reports contain evidence of hor-monal activity due to environmental chem-icals without a direct link to reproductiveeffects (90,91).

Translating subtle functional deficitswithin individuals into population-leveleffects is the real challenge and will requirebetter field observations and laboratorystudies to more precisely simulate fieldexposures. The generalizations that canbe made are grouped into the followingcategories:

a) Sensitive stages: Developmentalstages are often the most sensitive to expo-sure. (Development in this instance isdefined as broadly as possible to includeembryonic, fetal, larval, and juvenilestages.) There are specific critical periods ofsensitivity to endocrine disruption. Thesemay be quite short and there may be morethan one. Critical periods vary for differ-ent organs and species. The uniquechanges in physiology during developmentmay increase sensitivity to endocrine-dis-rupting agents. Also, sensitivity may besexually dimorphic or distinct in expres-sion if not sensitivity.

Effects on development can be revers-ible because of effects on maturation orirreversible because of effects on differentia-tion. Often, irreversible effects on differen-tiation have a long latency for expression.

Adult males and females are also affect-ed by endocrine disruptors, and there maybe physiologic states in the adult (e.g., earlypregnancy) that enhance susceptibility.Much of our knowledge of the action ofendocrine disruptors in adult humans isderived from the use of various steroids aspharmacologic agents.

b) Types of effects: Endocrine disrup-tor effects vary by species and life stage atwhich exposure occurs. Endocrine disrup-tion can result in morphologic abnormali-ties of the gonads, reproductive tract, brain,and other organs; functional and behavioralabnormalities; and certain malignancies(particularly of the reproductive system orrelated structures [e.g., breast]). Functionalabnormalities include decreased semenquality, reduced numbers of sperm, infertil-ity, disrupted estrous or menstrual cycling,and premature menopause; behavioralabnormalities include altered sexual behav-ior and decreased libido. Functional andbehavioral abnormalities may occur afterdevelopmental or adult exposure, althoughthe expression is likely to be different forthe various life stages.

Environmental Health Perspectives a Vol 104, Supplement 4 - August 1996720


Circulating hormone concentrationsmay be altered by effects on hormonemetabolism or steroidogenesis. There mayalso be effects on cellular function in theendocrine system or in systems thatrespond to the endocrine system mediatedby toxicant interactions on differentreceptor classes.

At the more fundamental levels of bio-logical organization, i.e., at the biochemicaland cellular level, there tends to be a greatdeal of similarity among vertebrate classesdue to the extreme phylogenetic conserva-tion of hormones and hormone receptorbinding characteristics. However, at higherlevels of biological organization there maybe divergence, as hormone systems mayhave different developmental functions indifferent phylogenetic groups. As an exam-ple, in the absence of androgens, the repro-ductive tract of mammals develops into afemale phenotype, whereas birds requireestrogens to initiate the female phenotype,and reptiles use both androgens andestrogens for sexual differentiation.

c) Agents that act by endocrine disrup-tion: A number of chemicals have pro-duced abnormal development and/orreproductive function via an endocrinepathway in some species. It is not the pur-pose of this report to construct an exhaus-tive list of these. Instead this list willinclude a few illustrative examples andindicate the classes of chemicals thatappear to be of special concern. However,not all chemicals of a given class havesimilar endocrine-disrupting potential;also, it is likely that chemicals in otherclasses not mentioned may have endocrinetoxicity potential.

Examples of agents that have beenshown to alter reproductive development invarious species via an endocrine mechanisminclude:* Hormones and drugs, including DES,

progestogens, androgens, ecdysteroidsand farnesyl hormones

* Metabolic inhibitors, including5-a-reductase inhibitors

* Pesticides, including DDT and itsmetabolites, chlordecone, and vinclo-zolin

* Phytoestrogens and mycotoxins* Other chemicals, including dioxins and

some PCBsThese agents do not need to be persis-

tent to have an effect, particularly if theexposure occurs during a critical develop-mental period. Importantly, prior exposureto persistent chemicals can result in expo-sure during a critical window of sensitivity

even though external exposure had longbeen terminated.

Several other chemical or chemicalclasses (e.g., alkylphenols, some phthalates,bisphenol A) have been shown to interferewith some endocrine-mediated processesin some systems, but evidence for effectson reproductive development in vivo islacking. Although these examples suggestdirect involvement with steroid receptors,nonreceptor pathways are also potentiallyimportant targets.

Last, a discussion of hazard is not com-plete without mentioning that it takessome level of exposure to have an effect.That is, there is a dose-response relation-ship for all chemicals. It is important tofocus on those chemicals for which endo-crine disruption is the most sensitive effect.Agents that have other effects of concern atlower doses can be studied and regulatedbased on those other effects if they areconsidered adverse.

d) Modes of action: Modes of actioninclude agonistic and antagonistic receptorbinding, and effects on hormone synthesis,storage, release, transport, and clearance.There are many receptor-mediated modesof action, including effects on estrogen,androgen, progesterone, thyroxine, gluco-corticoid, and Ah receptors. Others arelikely to be identified.

In addition to direct effects on receptors,there are instances of metabolic inhibitionand induction that affect steroidogenesis,inhibitors of enzymes that modify hormones(e.g., 5-o-reductase), and effects on plasmatransport proteins and neurotransmitterlevels (e.g., effects on the hypothalamic-pituitary axis). Differences between speciesmay be marked, and this may be oneimportant source of lack of concordanceacross species. Last, it will not be a simplematter to predict the action of mixtures.

It is noteworthy that some agents inter-act with more than one receptor type (e.g.,estrogen, androgen, and progesteronereceptors) and that there may be multiplemechanisms of action for a single agent,leading to different dose-response curvesfor different outcomes.

e) Species affected: In vertebrates, thereare examples of endocrine disruption fromMammalia (including humans and labo-ratory animals) to Pisces. Examples ininvertebrates include gastropod mollusks(exposed to alkyltin), insects (exposed toinsect growth hormones), and crustaceans.It is likely that other examples have yet tobe identified. Sensitivity varies by speciesand population.

f) Species concordance and divergence:As noted above, the level of concordancebetween species depends on the biologicallevel of organization being examined. Thegreatest homology tends to occur at themost fundamental levels of organizationand less so at the level of the organism.Knowledge of mechanisms of action and ofbasic comparative endocrinology andembryology will greatly enhance our abilityto extrapolate among species.

Although we have considerable infor-mation on perturbations of reproductivedevelopment as well as on direct effects onthe adult, after exposure to chemicals thatdisrupt the normal functioning of theendocrine system, there remain importantuncertainties. In particular the work groupnoted the following:

a) Species-to-species extrapolation:There is a great deal to learn about basicaspects of comparative endocrinology andembryology. Concerning the former, it isknown that birds rely on estrogen for sexualdifferentiation, whereas mammals rely onandrogens, and reptiles utilize both estro-gens and androgens. While mammals relyon androgens, it is important to rememberthat the androgen serves as a precursor forestrogen, which is the effective moleculefor sexual differentiation in brain morphol-ogy, endocrine secretion, and behavior.Differences have also been observed in thespecificity of serum-binding proteins. It islikely that other differences will be foundthat will help explain the diverse responsesto endocrine disruptors.

In comparative toxicology, three factorsmust be considered. First, there may bepharmacokinetic differences that affect theconcentration of the active agent at the tar-get site. Second, there may be pharmaco-dynamic differences in the interactions ofthe agent with the molecular target andsubsequent responses. Third, superimposedon these are differences in the genetic tem-plate across species.

b) Availability and reliability of historicaldata: Historical trend data are the basis foridentifying adverse health effects by inter-preting the extent to which the incidenceof these effects is changing and/or is associ-ated with other time-related trends. Thesedata are also extremely useful in formulat-ing hypotheses on causation. Therefore, arobust historical database is important forfurther field and epidemiological research.However, the historical database on humanand animal reproductive end points is notadequate for our needs. More prospectiveand retrospective studies are needed on a

Environmental Health Perspectives * Vol 104, Supplement 4 * August 1996 721

KAVLOCK ET AL.

variety of reproductive parameters. Compre-hensive evaluation of reproductive parame-ters in highly exposed populations (forexample, sons and daughters of mothersexposed to DES or PCBs) would be usefulto identify the most critical effects forwhich historical trend data should be ana-lyzed. Finally, statistical analysis of histori-cal trend data should be rigorous, as pat-terns of change may suggest potentialunderlying causes (e.g., one might suspectdifferent causes for a single cataclysmicdecrease in sperm count versus a continuousgradual decline).

c) Effects at low levels of exposure:Effects at low levels of exposure may bequalitatively different from effects athigh levels for several reasons, includingmultiple mechanisms of action, each ofwhich takes effect at a different dose level.Furthermore, an agent may have multipleeffects, each having a qualitatively andquantitatively different dose-response curve.

d) Latency: It is known that develop-mental exposure to an endocrine disruptorcan have long latency periods beforeexpression of an adverse effect, the hall-mark being DES-induced vaginal adeno-carcinoma. Efforts should be made toestablish prospective registries of highlyexposed populations, such as sons anddaughters of mothers exposed to DES, forother latent effects, particularly breast,endometrial, and prostate cancers, as thisgroup is reaching the age when problemsbecome more prevalent. Other latenteffects of potential concern include cardio-vascular, premature menopause, andaltered reproductive behavior in birds andother fauna.

e) Relevance of bioaccumulation andbiomagnification: Although it is intuitivethat materials which bioaccumulate andbiomagnify are of special concern to thosespecies that consume them, the relativecontribution of these processes to toxicityis dependent on trophic level in the foodweb, life stage, physiological conditionsfavoring lipid mobilization (e.g., pregnan-cy, lactation, egg laying), and reproductivestrategy. More work needs to be donebefore the contribution of bioaccumulationand biomagnification to the toxicity of aparticular agent in a particular species canbe fully appreciated.

f) Behavior of mixtures of ligands:Additivity, synergism, and antagonism ofcomponents of mixtures acting at thesame receptor are complicated by severalfactors; for example, the interactions maybe qualitatively different depending on

the concentration and ratio of each compo-nent of the mixture and whether they areagonists, antagonists, or combinations. Itmay be possible to construct a set of rules toexplain the behavior of any mixture, givensatisfactory knowledge of the activities andmechanisms of action of each componentand validation of the system. This approachlimits testing to a large but finite number ofchemicals in order to characterize virtuallyinfinite numbers of mixture componentsand concentration. This should be contrast-ed with the impossibility of empirically test-ing every possible mixture.

Risk assessment approaches for mix-tures also need further refinement. In par-ticular, the toxic equivalency factor (TEF)approach needs improvement. A commonmechanism of action must be clearlydemonstrated before initiating the process.When calculating TEFs, the lowest experi-mentally measurable response of the mostappropriate effect in the most appropriatespecies should be used consistently.

g) Basic research: Although much isknown about the mechanisms of develop-mental effects of endocrine disruptors,much more needs to be learned. Two areasthat deserve particular attention areimproved understanding of dose at thetarget site and the investigation of addi-tional potential modes of action involvingparacrine and autocrine signaling pathways.

h) Sensitive populations: Although weare generally aware of the life stages mostsensitive to endocrine disruption, there isstill some uncertainty about the relativesensitivity of populations. Marked differ-ences in the responsiveness of differentinbred mouse strains to the toxicity ofTCDD suggest that there may be substan-tial variability of response in other species,and for other endocrine disruptors. Inwildlife populations, there is uncertaintyas to whether differences in susceptibilityof subspecies has led to a decrease ingenetic variability.

What are the research needs related to thedetection of reproductive and developmen-tal effects ofendocrine disruptors?While numerous wildlife and ecosystempollution studies have assessed reproduc-tive end points and a few well-documentedexamples of the effects of EDCs in wildlifehave been identified, little is known aboutendocrine-disrupting chemical effects inmost wildlife populations. Structuredresearch on the status and trends of popu-lations and communities including naturalvariations is needed to detect population

changes that might not otherwise be appar-ent until significant losses occur. Thegroup constructed the following list ofresearch needs:

a) Establishing cause and effectrelationships: Hypotheses generated fromfield observations must be tested in the lab-oratory and in controlled field studies. Thiswill require collection of more completefield data on hormone levels, tissue burdensof chemicals, and more measures of markerexpression (e.g., vitellogenin) to guide lab-oratory studies. Existing long-term ecologi-cal field studies should be identified thatcan be integrated into the endocrine dis-ruptor research agenda to test hypotheses.This research must be multi-disciplinary innature, and improved mechanisms forinformation exchange between field andlaboratory researchers must be established.Better statistical models to predict riskfrom observations of exposure and effectsare also needed.

b) Biomarkers: Better biomarkers areneeded that reflect both exposure to andeffects of endocrine disruptors. Thesebiomarkers need to correlate with the mostsensitive end points associated withendocrine disruptors. These biomarkersmust address species differences, sexualdimorphism characteristics, and be lifestage specific. They must be applied inlong-term, transgenerational studies toidentify biomarkers in offspring that can bemeasured shortly after exposure and thatare predictive of long-term or latent effects.

c) Data collection: More extensivestudies are needed of both highly exposedwildlife and human populations (forexample, sons and daughters of mothersexposed to DES or PCBs) as well as popu-lations exposed to a contemporary ambi-ent level of endocrine disruptors.Throughout these studies, researchersmust consider that there may be no unex-posed populations. Research hypothesesshould be evaluated in these populations.More information on normal populationvariation, as well as regional and seasonaleffects, should be gathered. Both prospec-tive and retrospective historical trendinformation is required for wildlifeand human populations in order forresearchers to identify adverse reproductivehealth trends (e.g., reduced semenquality and quantity in humans, reducedreproductive success in wildlife) and todevelop and test hypotheses about theircausation when possible.

More data on exposure monitoringcoupled with integrative bioassays of



effects are needed, which would be facili-tated by coordination of different agencyefforts (e.g., National Oceanographic andAtmospheric Administration [NOAA],U.S. Fish and Wildlife Service [USFWS],U.S. Geological Service [USGS], NationalBiological Service [NBS]). Traditionally,exposure data have focused on adults. Thatfocus must shift to collecting data for themost sensitive life stages, recognizing thateffects may have prolonged latency periodsbefore they are manifested.

d) Basic research: Studies on basicdevelopmental biology need to address theontogeny of receptor systems. Classicalapproaches to the study of receptor-basedmechanisms may not apply to early devel-opmental periods. Research should beconducted to identify the end points inmultigenerational studies that are mostsensitive to endocrine disruption. Bothreceptor and nonreceptor-mediated mech-anisms of endocrine disruption should bestudied, and the normal hormonal envi-ronment of the developing organism andthe adult must be better characterized.Better and cheaper analytical tools needto be developed, including molecular probesfor gene expression products, and immuno-assays for biological tissues. Research effortswould be greatly facilitated by establishmentof a repository of radiolabeled compounds,antibodies, and cDNAs available at minimalcost to researchers. Laboratory studies mustfocus on low-dose exposures reflecting real-istic environmental levels, environmentalmixtures, and pharmacokinetic parame-ters, including modeling of nonlineardose-response relationships.

e) Mixtures: Because little is knownabout the hazards of chemicals in environ-mentally relevant mixtures, a scientificallybased risk assessment approach is needed todeal with mixtures. Chemical interactionscan be very complex and need to be charac-terized at a number of environmentally rele-vant dose levels. The model of receptorinteraction used in these studies needs to becarefully examined, as a single receptor canactivate a number of genes, which is a partic-ular concern with environmental mixtures.A toxic equivalents approach (TEQ) is

potentially useful for assessing the risk posedby multiple chemicals with a commonmechanism of action. Validation of thisapproach should consider multiple variablessuch as species differences, sensitivity of theend point(s) measured, mechanism ofaction, nutritional status, co-occurrence ofinfection, and doses. The most usefulapproach would use the most appropriate

end point in the most appropriate life stageof the appropriate sensitive species.

f) Screening methods: Given the needto test many more chemicals for endocrine-disrupting potential, short-term in vivo andin vitro tests that are rapid, reliable, andinexpensive must be developed to screenchemicals for relevant hormonal activity.Screening methods will only be useful ifthey are sensitive to a variety of mecha-nisms of action by endocrine disruption.The most sensitive end points need to beidentified. Behavioral and growth parame-ters are end points to explore further. Atesting approach incorporating both invitro and quantifiable in vivo tests that arevalidated is needed. Where the mode ofaction is known, research on the molecularbasis of the effect of endocrine disruptorson gene expression is needed to identifymolecular probes reflecting the morpholog-ic and/or functional changes in the repro-ductive system. Where the particularmechanism(s) of action of a compound isunknown, more long-term in vivo assaysare needed. Early life-stage testing inrodents, fish, and amphibians seemspromising as an in vivo screening method.Quantitative structure-activity relationship(QSAR) models need to be developed fur-ther to help prioritize compounds for moreextensive testing.

g) Population heterogeneity: Thepossibility needs to be studied thatendocrine-disrupting chemicals may bedecreasing the genetic variability ofpopulations through selection pressure.Comparisons of liver tumor frequencies,PAH burdens, age, and length characteris-tics of brown bullheads collected in theearly 1980s from two tributaries of LakeErie strongly support the hypothesis thatthe bullheads in the Black River weresubjected to an age-selective mortality asso-ciated with a high incidence of PAH-associ-ated liver carcinoma (8). Genetic diversityestimates for the mitochondrial genome ofthis species at these and seven other sites inthe lower Great Lakes in the late 1980swere always much lower in populationsfrom the contaminated sites than in nearbyreference sites (92), apparently due to sto-chastic reductions in population size.

What are the highestpriority research needsfor reproductive anddevelopmental effects?Developmental and reproductive toxicity isa problem with important implications forpublic health and ecosystem health.Research is needed to delineate the contri-bution of endocrine-disrupting chemicals

to the observations of adverse effects onreproduction and development in humansand in wildlife populations. Because of thepotential long-term impacts on both indi-viduals and populations, this area deservesa high research priority.

Among the highest priority needs inthis area are the following:

a) Controlled laboratory tests ofhypotheses generated from field studies.

b) More extensive studies on wildlifeand human populations exposed to highlevels of endocrine-active toxicants (e.g.,the DES and PCB cohorts) to identify theadverse health effects most likely to occurfrom developmental endocrine disruption.

c) Better definition of normal variabili-ty in reproductive parameters and morecomprehensive temporal data (bothprospective and retrospective) so that poten-tial trends can be identified more readilyand reliably, and hypotheses tested regard-ing their causation.

d) Characterization of the interactionof mixtures of endocrine-active toxicants,and the development and validation of riskassessment methods that adequatelyaccount for these interactions.

e) Development of QSAR and short-term screening approaches to identifypotential endocrine-active hazards to repro-duction and development.

Neurological Effects

What do we know about the neurologicaleffects of endocrine-disrupting agents inhumans and wildlife? What are the majorclasses ofchemicals thought to be responsi-ble for these effects? What are the uncer-tainties associated uwith the reported effects?The work group determined that neuroen-docrine disruption can be induced by multi-ple mechanisms. Direct effects on endocrineglands (e.g., the thyroid) may alter the hor-monal milieu, which in turn can affect thenervous system, resulting in neurotoxicity.Conversely, EDCs may initially act on thecentral nervous system (CNS) (e.g., neu-roendocrine disruptors), which in turn caninfluence the endocrine system. It wasnoted that exposure to chemicals canadversely affect the structure and function ofthe nervous system without any endocrinesystem involvement. The group consideredseveral examples of effects produced by dis-ruption of the endocrine system and agreedthat alterations in the following would beindicative of neuroendocrine disruption:reproductive behaviors mediated by alter-ations in the hypothalamic--pituitary axis

Environmental Health Perspectives * Vol 104, Supplement 4 - August 1996 723

KAVLOCK ET AL.

(e.g., courtship and parental behavior inavian species); alterations in metabolic rate,which could indirectly affect behavior;altered sexual differentiation in the brain,which could affect sexually dimorphicreproductive and nonreproductive neuralend points; and some types of neuroterato-genic effects. The group concluded thatthere were clear examples in the humanand animal literature in which exposure toendocrine disruptors had occurred andeffects on behavior, learning and memory,attention, sensory function, and psy-chomotor development were observed(60,76,79,93-110). Some of these effects,however, can also be produced by develop-mental neurotoxicants having little or noknown endocrine-disrupting propertiesand, therefore, cannot be regarded asspecific to the endocrine-disrupting class ofchemicals. It was also pointed out thatexposure to a number of nonchemical fac-tors (e.g., food or oxygen deprivation, infec-tions, and temperature) could also adverselyaffect the nervous system resulting in effectssimilar to those produced by endocrine dis-ruptors. These nonchemical factors mayalso interact in as yet unpredictable wayswith chemical stressors. Therefore, consid-erable care should be taken to eliminatenonchemical causes before concluding thatneurotoxicity is causally related to theeffects of a chemical acting on theendocrine system.

The group concluded that there wereseveral examples of chemicals or classes ofchemicals that produce neurotoxicity by anendocrine mechanism. It was agreed thatenvironmental toxicologists should consid-er the dose at which neuroendocrine dys-functions are produced relative to the con-centrations existing in the environmentand relative to dose levels at which othertoxic effects occur, the relationship betweenexposure and effect, and the role of natu-rally occurring chemicals with endocrine-mimicking properties. With these caveatsin mind, examples of directly or indirectlyacting neuroendocrine disruptors includesome PCBs, dioxins, DDT and relatedchlorinated pesticides and their metabo-lites, some metals (methylmercury, lead,organotins), insect growth regulators,dithiocarbamates, synthetic steroids,tamoxifen, phytoestrogens, and triazineherbicides. Identification of chemicals asneuroendocrine disruptors should be basedon mechanistic information at the cellularor molecular level in the endocrine systemor defined functionally in terms of activityon responses known to be mediated by or

dependent on hormones. All definitions ofneuroendocrine disruptors should be inter-preted specifically with respect to gender,hormonal status, and developmental stage,since the expression of toxicities of chemi-cals may change significantly depending onthese variables.

The group identified a number ofuncertainties critical to understanding thesignificance of the effects of neuroendocrinedisruptors, including:

a) Chemicals occur as mixtures in theenvironment, thereby making it difficult toassign cause and effect for specific agents. Itis possible that the parent chemical maynot affect the endocrine system but ismetabolized to an active form. The toxico-kinetics of and relative tissue distributioninto the nervous system are generallyunknown for most chemicals; little isknown about the metabolic interactionbetween chemicals in mixtures.

b) There are ranges of possible specificand nonspecific effects that could be mea-sured. Research to date has used only asmall number of techniques and methods,and it is likely that many neuroendocrineeffects may be subtle and not easily detectedwith currently available procedures. It isalso a concern that the functions most sen-sitive to chemically induced alterations inneuroendocrine function are the mostdifficult to measure in the field.

c) It is critical to know when exposureoccurred relative to when the effects aremeasured. Observed effects could be depen-dent on a number of extrinsic factors suchas seasonal variability and intrinsic factorssuch as hormonal status. In addition, thenervous system is known to be differentiallysensitive to chemical perturbation at variousstages of development. A chemical mayhave a significant effect on neuroendocrinefunction if exposure occurs at a criticalperiod of development but have little or noeffect at other stages of maturation.

d) Several issues related to extrapola-tion are critical to understanding neuroen-docrine disruptors. For example, it isdifficult to evaluate the significance tohuman health of a chemically inducedchange in a behavior that does not natural-ly occur in humans, i.e., there are concernsabout the appropriateness of some animalmodels for toxicologic studies. In addition,there are uncertainties about extrapolatingfrom species to species and from experi-ments conducted in the laboratory to thoseperformed in the field.

e) There are uncertainties about theshape of the respective dose-response

curves for many neuroendocrine effects. Itis likely that some chemicals may havemultiple effects occurring at differentpoints on the dose-response curve.

f) The group concluded that basicinformation concerning the mechanism ofaction of chemicals on the developing ner-vous system and the neurological role ofhormones during development wouldgreatly reduce uncertainties about risk ofexposure to neuroendocrine disruptors.Furthermore, it is also important to under-stand the consistency of the effects relativeto the hypothesis that chemicals are affect-ing the nervous system.

What are the research needs related to thedetection ofneurological fects ofendocrinedisruptors?The group considered a number of researchneeds:

a) Opportunistic field studies:Coordinated epidemiological research toexploit human and wildlife populations,which have known exposures to neuroen-docrine disruptors, to define the biologicaleffects most likely to occur:

b) Laboratory/field studies: Systematicfield and laboratory studies that focus oncritical experimental uncertainties, e.g.,dose-response determinations, effects ofdifferent duration of exposures, time ofexposure during development, and ageof assessment and integration of variousend points.

c) Mixtures: Systematic research toaddress the principle of additivity in deter-mining the risk associated with exposure tomixtures.

d) Toxicokinetics: Studies to determinethe age group-dependent toxicokineticsand toxicodynamics of environmentallyrelevant chemicals, with an emphasis onproviding better exposure assessments tocorrelate biological effects with target ortissue dose.

e) Mechanisms of action: Research atthe cellular and molecular levels to providea better understanding of the mechanismsof action for known neuroendocrinedisruptors.

f) Basic research: Research to betterunderstand the normal development of thenervous system and the role that endocrinesystems may play in that development.

g) Sentinel species and biomarkers:Identification of sentinel species and devel-opment of biomarkers of exposure andeffects for neuroendocrine disruptors.

h) Identification of sensitive subpopu-lations: Research to determine if there are



populations or individuals that may bedifferentially sensitive to neuroendocrinedisruptors.

i) Multigeneration assay development:Multigenerational research in both inverte-brates and vertebrates to assess the possibletransmission of effects across generations

What are the highestpriority research needsfor neurologaleffcts?The group agreed that research concerningthe effects of chemicals on the neuroen-docrine system should have a relativelyhigh priority. Because the developing ner-vous system is differentially sensitive tochemicals, there is great concern that low-level exposure to environmentally relevantmixtures of chemicals could have subtle,long-lasting effects on nervous system func-tion in a number of animal and humanpopulations. Neurotoxicologic effects havebeen documented in children exposed to anumber of environmentally relevant chemi-cals, including the PCBs, methylmercury,and lead (111), although these do not nec-essarily act primarily via endocrine-mediat-ed mechanisms. Finally, the group notedthat the nervous system interacts with orcontrols other potential targets (immuneand reproductive systems) and serves as aninterface between the environment and theinternal milieu. It seems likely that effectson other target systems and the manner inwhich organisms perceive and respond tothe environment may involve the neuroen-docrine system to some degree. Thus, stud-ies on the mechanisms of neuroendocrinedisruption should provide crucial informa-tion concerning mechanisms of other bio-logical effects.

Among the highest priority researchneeds in this area are

a) The initial focus on identifying anddocumenting effects of concern to humansand wildlife that are possibly mediated bythe neuroendocrine system

b) Follow-up studies to determine theparameters of exposure to specific chemi-cals and the time of assessment of effectson the neuroendocrine system

c) Demonstration of biological plausi-bility between exposure to chemicals andobserved effects in humans and wildlife

d) Studies on the potential mechanismof action of observed effects

e) Attention to the roles of mixture inthe effects produced by neuroendocrinedisruptors

f) Better understanding of the interac-tion of the nervous system with otherpotential targets of endocrine disruptors.

Immunological Effects

What do we know about the immunologi-cal effects ofendocrine-disrpting agents inhumans and wildlif? What are the majorclaes ofchemicals thought responsibleforthese effeets? What are the uncertaintiesassociaed uith the reported effects?Published studies have demonstratedassociations between autoimmune syn-dromes and DES exposure (112). A rela-tionship is well established between physio-logical estrogen levels and autoimmune dis-eases in women (113-115). The observa-tions that exposure of humans to DES,TCDD, PCBs, carbamates, organo-chlorines, organometals, and certain heavymetals alters immune phenotypes or func-tion are suggestive of immunosuppressionand potential disease susceptibility(116-119). Experimental animal studiessupport these observations (e.g., 120-126),although dose-response information isneeded to clarify whether these are directly-or indirect-acting agents. With respect tofish and wildlife, it was also noted that sev-eral of the agents listed above induceimmune suppression or hyperreactivitysimilar to that reported in experimentalanimals and humans. Embryonic exposureof trout to aflatoxin has led to alterationsin adult immune capacity (127-129).With regard to disease susceptibility andexposure, there have been examples such asthe dolphin epizootic of 1987 to 1988(130). In this case there was an associationwith PCBs and DDT in the blood,decreased immune function, and increasedincidence of infections among affectedindividuals (131). Impairment in immunefunction has been reported in bottlenosedolphins exposed to PCBs and DDT (132)and in harbor seals fed fish from pollutedwaters (133,134). From 1991 to 1993,specific immune functions and generalhematologic parameters were measured inherring gull and Caspian tern chicks froma number of study sites in the Great Lakeschosen across a wide range of organochlo-rine contamination (primarily PCBs). Asthe hepatic activity of ethoxyresorufin-0-deethylase (EROD), an index of exposure,increased, thymus mass decreased. At high-ly contaminated sites both gull and ternchicks showed marked reductions in T-cell-mediated immunity as measured bythe phytohemagglutinin skin test (135).A variety of immunoassays have been

used to demonstrate effects in experimen-tal laboratory animals, humans, fish, andwildlife. These include modulation of

antibody responses (both in vivo and invitro), the phytohemagglutinin skin test,mitogenesis, phagocytosis, levels of com-plement or lack of acute phase reactants,cytotoxic T-lymphocyte reactivity, and nat-ural killer cell activity (136-142).

Evidence of an increased rate ofautoimmunity associated with prenatalDES exposure suggests the possibility thatother endocrine-disrupting chemicals(EDCs) may induce a similar pathologicstate. Studies are needed to determine ifthere has been an increase in cases ofimmune dysregulation in areas or siteswhere EDC exposures have occurred.Evidence indicates that the incidences ofallergy and asthma (which are forms ofhypersensitivity) are increasing in humans(143-145). It is not known whether EDCexposures are responsible for some part ofthis development. Alteration of sex-steroidbalance has been shown to lead to increasedor accelerated onset of autoimmune syn-dromes in mice (114). In rats and mice,heavy metals such as lead, mercury, andgold enhance autoimmune syndromes(146,147). There have also been reports ofexposures of fish to EDCs in the environ-ment that lead to immune enhancement.Although autoantibodies have been reportedin sharks and trout (148-150), no attemptshave been made to correlate exposure toEDCs with incidences of autoantibodies. Introut, embryonic exposure to aflatoxin B1can lead to immune stimulation or sup-pression in the adult, depending upon theimmune parameter analyzed (129). Otherdata suggest that small changes in physio-logic levels of estrogens can affect theimmune system, and studies in gull andtern chicks in the Great Lakes clearly indi-cate that the findings are associated withdevelopmental exposures (135).

Concerning direct-acting EDCs,although it would appear that these agentsdirectly affect the immune system, it isunknown whether there may be disruptiveeffects on the endocrine-immune axis.Since the immune and endocrine systemsare linked via various cytokine signalingprocesses (IL-1, ACTH, catecholamines,prolactin, and endorphins), it is likely thatEDC effects on the immune systemmodulate elements of the endocrine or ner-vous systems or vice versa. Too little isknown about the dose-response curves forimmunotoxicity, neurotoxicity, or endo-crine effects to decipher the independent orinteractive effects on these systems. Becauseof the high degree of intercommunicationbetween these systems, there is a need for


KAVLOCK ET AL.

coordinated and cooperative studies amonglaboratories in all these disciplines.

Although the most forceful argumentsfor the overall consequences of immune dys-function would be increased disease inci-dence, this is difficult to assess in humansor wildlife populations. Furthermore, onlycertain subpopulations (the very young orelderly) may be affected. Disease may onlybe manifested as a population decline.Disease trials can be undertaken, but theyrequire controlled laboratory experimentsemploying populations of wild animals orfish that can easily be maintained in thelaboratory. In humans, the variability with-in a population makes it difficult to deci-pher exogenously triggered effects.

The fact that employment of a varietyof in vitro assays has been successful in cor-relative exposure studies leads to the ques-tion of whether those immune parameterscan be correlated with the increased risk ofdisease. A number of immune parametersoperate independently (i.e., lysozyme lev-els, complement activity, phagocytosis,induction of cytotoxic T lymphocytes,plaque-forming cells, etc.). Which combi-nation of these assessments would make foran optimal predictive suite of assays?

Knowledge of normal baseline valuesfor wildlife species, and in most caseshumans, is lacking. If these populations areto be screened for perturbations in immunefunction, control populations must bedefined and standardized control valuesobtained. Also, the types of exposures mustbe well documented (i.e., dose, length ofexposure, timing).

What are the research needs related to thedetection of immunological effects ofendocrine disruptors?The work group identified the followingresearch needs related to the assessment ofimmunological risk to human and wildlifepopulations:

a) Epidemiological analyses: Humanpopulations with known exposure shouldreceive greater attention, concentrating onthe possible association of exposure withautoimmune symptomology, hypersensi-tivity, and disease incidence. In wildlifepopulations, particularly marine mammals,analysis must be associative and wouldrequire more rapid and inexpensive meth-ods to quantify or identify the presence ofEDCs. This will require additional livecapture research on marine mammals usingnoninvasive sampling techniques to estab-lish appropriate baselines of normality.It is especially important to coordinate

immunological research activities withreproductive, neurological, and carcinogen-esis research. The ability to quickly mobi-lize groups to address environmental prob-lems, with a tested suite of immunoassays,would be advantageous when new environ-mental exposures are detected.

b) Mechanisms: Studies on mechanismsof action must be conducted for agents thatcause endocrine disruption after initial inter-action with the immune system. Studies arealso needed that determine the dose level atwhich immune effects occur secondary toendocrine disruption. Exclusive immunedysfunction at low concentrations wouldindicate that the primary target of thespecific EDC would be the immune system.The endocrine effects caused by many dis-ruptors have not been examined forimmune system effects. Studies are alsoneeded to characterize potential effects onthe endocrine system of direct-actingimmunotoxicants. The general feeling wasthat many of the endocrine disruptors iden-tified thus far act directly upon the immunesystem; however, the group did not exten-sively discuss agents that may affect theimmune system via endocrine disruption(e.g., ammonia).

c) Mixtures: Identification of theimmunotoxic elements within such mix-tures would be a primary aim, althoughmore information would be gained if coor-dinated studies were conducted betweenendocrinology, developmental, and neuro-biology laboratories. In addition, inexpen-sive analytical tools are needed to analyzecontaminant body burdens. Althoughthere was some amount of uncertainty inthe group as to the value of bioassays togrossly quantify classes or groups of EDCs,this avenue might provide an inexpensivemeans of assessing exposure. However,most samples would be comprised by thepresence of mixture and the contributionsof agonists and antagonists to the EDCwould further complicate analysis.

d) Critical periods: Because studies arelacking on developmental exposure, theywill have to be conducted in the laborato-ry, in long-term human epidemiologicalstudies, or with individually markedwildlife populations. To date, work withagents that have an initial impact on theendocrine system has demonstrated thatthe most compelling evidence is related todevelopmentally acquired dysfunction.Furthermore, and perhaps more important,the effects on the developing immune sys-tem may be more persistent or longer last-ing than those that might occur with adult

exposure. The doses needed to elicit thesedevelopmental dysfunctions are not as largeas those for acute exposures of adults; thus,a much greater percentage of populationsmay be affected in this way. For somewildlife species, studies on normal develop-mental immunobiology would have to beconducted prior to any extensive EDCtesting on model species.

e) Sentinel species: The choice of modelor sentinel species for the assessment ofEDC effects on the immune systemrequires a great deal of forethought. Thesechoices should be well suited to answerspecific questions (both of an immunologi-cal and endocrinological nature). It was feltthat a logical first step might be greatlyexpedited by agencies such as the U.S. EPA.Their access to information concerning theimpact of the environment on populationsthroughout the United States could providevaluable information for determiningpotentially good model/sentinel species.These species should also possess the fol-lowing characteristics: ease of maintenancein the laboratory; ubiquitous distribution inthe environment-broad range; accessibili-ty of large numbers; easily bred within alaboratory environment (permits analysis ofgenetic basis of susceptibility required fordevelopmental analysis and eliminatescarry-over environmental effects in dose-response experiments); substantial databaseon physiologic parameters; inexpensive costof procurement and maintenance; ecologi-cal relevance; short generation time andrepresentative of a large number of speciesor, at least, groups of species

Upon selection of model species, thereshould be coordinated optimization andstandardization of the immunological assaysand their analyses. Major gaps in the endo-crine and immune databases of selectedspecies must be addressed immediately.The entire repertoire of sentinel speciesshould account for the variety of environ-mental conditions that may modulateeither the exposure or response to EDCs,including temperature (if an ectotherm),trophic level, and other conditions of theparticular niche (i.e., salinity, exposure tosediment, water column, atmosphere.

f) Ecological monitoring: Studies ofmarked wildlife populations are neededto determine the effects of contaminantson the demographics and to monitor theage and accumulation of EDCs withintissues. Some of the sentinel species usedin these studies need to be relativelyhardy while others may need to be rathersensitive. For example, trout are often the



first to disappear from a contaminated site,thus suggesting the first indication of aproblem. Although this feature in and ofitself may be a good marker attribute for asentinel species, it is also advantageous forthe animal to be hardy enough to remainon the site so sampling can reveal thenature of the disorders that occur.

What are the highestpriority research needsfor immunologic effects?The work group agreed that the highestresearch priorities should go to areas wherethe risk to human and animal life is thegreatest. If the impact is on the immunesystem, the risk to life could obviously betremendous and would necessitate a com-mitment to immunotoxicologic research inthis area. However, before such action istaken, it was felt that the existence of suchproblems should be determined by con-ducting epidemiological analyses ofhuman, wildlife, and sentinel species, asdescribed in the previous section. If riskappears considerable, research should thenbe directed toward developmental effects aswell as toward acute effects, and workshould begin as outlined with the mostappropriate sentinel/model species.

Among the highest priority researchneeds are

a) Epidemiological studies in bothhuman and wildlife populations are neededto establish the incidence of immune-relat-ed diseases, including immunosuppression,hypersensitivity, and autoimmunity associ-ated with exposure to endocrine-disruptingagents and to determine the impact ofthese effects on clinical diseases such asinfections and cancer.

b) Basic research is required to identifythe mechanisms and individual substancesthat alter the immune system, and todetermine whether they act directly orindirectly through alterations of endocrinesystem function. Included in this researchshould be determination of dose-responserelationships.

c) Methods need to be developed,particularly for wildlife species, to identifyand validate sensitive assays that detectimmune effects including the selection ofappropriate sentinel species. Included inthis process should be the developmentof biomarkers.

d) Studies are needed to determinewhether sensitive populations exist. Basedupon existing evidence, the immunologiceffects of these substances on the very young(i.e., during the developmental phase) areof particular concern.

Risk Methodology IssuesHazard Identification

Hazard identification includes the collec-tion and evaluation of toxicity data fromtest systems, epidemiological studies, case

reports, and field observations. The numberof species evaluated, the number of studiesconducted, the quality of the studies, endpoints evaluated, and other factors are

assessed in the context of dosage, route, tim-ing, and duration of exposure. These dataare used to determine whether the agent inquestion poses a hazard, and the context inwhich it poses a hazard (i.e., is it route

specific, species specific, life stage specific,etc.). Discussion focused mainly on prospec-

tive hazard detection, whereas from the eco-

logical perspective, retrospective analysisrelated to post-environmental release of con-taminants is often the more important issue.This dichotomy was explored further by thesubsequent U.S. EPA-sponsored workshopdevoted to ecological issues (151).

What are the existingguidelines/testingpro-tocols that evaluate endocrine-relatedeffects?Ecological test guidelines. The followingecological tests are commonly conductedduring the evaluation of industrial chemi-cals and pesticides. Tests marked with an

asterisk offer the opportunity to detecteffects relevant to endocrine disruption. Itshould be noted that the end points ofgrowth and reproduction measured inthese tests are apical, and hence possiblyreflect effects caused by an underlyingendocrine-linked mechanism. However,because of this integration, an explicitendocrine mechanism would not be impli-cated. The tests include a) short- and long-term algal toxicity; b) acute and chronicreproduction* tests in aquatic and terrestri-al invertebrates; c) acute, chronic*, early(embryo, larval*) and full life-cycle* tests infish; d) acute, 14-day, and longer-termreproduction studies*, and egg-dosingstudies of hatchabilty and teratology* inavian species; and e) acute and chroniceffects in plants. Ecological testing uses a

small number of surrogate species to repre-

sent the environment; therefore, the diver-sity of reproductive and developmentalstrategies contained within aquatic andterrestrial organisms may not be fullyaddressed. It was noted that some of thesetests require development of appropriatepositive and negative controls and standard-ization. Furthermore, they typically do not

address sublethal effects (e.g., hormone

levels, behavior, transgenerational effectson the offspring) often consideredfundamental in mammalian testing.

Human health testing guidelines.Various forms of the a) 2-year cancerbioassay; b) 90-day subchronic toxicitystudy; c) multigenerational reproductionstudy; d) developmental toxicity study;e) developmental neurotoxicity tests; andf) immunotoxicity tests were discussed.Although these tests are generally used toevaluate the impact on human health, theyalso play a role in assessing effects in mam-malian wildlife species.

It was the consensus of the work groupthat these tests are intended to detect effects,not to identify mechanisms. For endocrinedisruption, the current tests in many casesmay fail to determine the appropriate noobserved adverse effect level (NOAEL) andmay fail to detect the reproductive toxicityof estrogenic pesticides [e.g., methoxychlor(152)]. Implementation of the HarmonizedReproductive and Developmental ToxicityTest Guidelines (153) from the U.S. EPAwill improve the ability of these teststo detect the appropriate NOAELs, includ-ing those related to endocrine effects.However, none of the current or proposedtest guidelines require the measurement ofserum hormone levels. In addition, thesetests are poorly designed to evaluate latenteffects that result from exposure early inlife. It was stated by the work group thatthe Immunotoxicity and DevelopmentalNeurotoxicity Test Guidelines were notspecifically designed to detect the effects ofendocrine disruptors on immune functionor CNS development, and, although thebasic two-year cancer bioassay can and doesdetect tumors of endocrine organs, transpla-cental carcinogenesis is not evaluatedbecause the study design does not includeexposure during critical stages of develop-ment (i.e., prenatal, neonatal, infantile, andpubertal stages of life). It was also notedthat current tests generally examine only asingle chemical at a time, which does notreflect the real world in which exposure isto complex mixtures of chemicals.

What areas within the present guidelinesneed refining and improvingfor the ade-quate evaluation ofeffects ofendocrine dis-ruptors?In light of the previously mentioned short-comings of the current testing guidelines,the work group proposed that research beconducted to develop and validate apicalmethods to detect endocrine disruptors.Once validated, these methods could be


KAVLOCK ET AL.

included in the appropriate test guidelines.Examples of areas where improvements arewarranted are as follows:

Ecological tests. Because of the lack of anassessment for transgenerational effects inecological tests, a number of additions wereproposed to the current ecological testguidelines. Research is needed to developand validate new testing procedures.

a) Avian reproduction studies shouldinclude an assessment of growth, viability,fecundity, reproductive morphology, andbehavior in birds exposed in ovo.

b) Some fish species with a short lifecycle may prove to be useful models inwhich to examine EDCs in the laboratory.They can be studied from prefertilization,through fertilization, hatching, recruitmentinto the breeding population, and fecundi-ty during adulthood. Their hormonal sys-tems are well characterized and they aresusceptible to hormonally induced cancers.Assay methods for hormonal systems ofnonmammalian vertebrates need to be vali-dated and made available to field biologistson a much larger scale than currentlyexists. Additional life-cycle studies forinvertebrates as well as fish are needed toprovide a more holistic hazard assessmentfor EDCs.

c) Sexual differentiation studies in inver-tebrates, fish, reptiles, birds and mammals.

Human health and mammalian tests(also for other vertebrates, as appropriate).a) Evaluation of weights, histology, andhormone production (in vivo withendocrine challenge tests, or in vitro after invivo exposure) of endocrine organs (testis,ovary, thyroid, adrenal, pituitary, etc.) inlong-term tests [e.g., (154)].

b) Determination of testicular, epididy-mal, and ejaculated sperm counts.

c) Establishment of landmarks ofpuberty and reproductive senescence inmale and female rats and other species.

d) Assessment of sexually dimorphicbehaviors and other CNS functions.

e) Expansion of developmental toxicitystudies to include perinatal exposure and apostnatal evaluation of reproduction func-tion of the offspring [i.e., the U.S. EPAAlternative Reproductive Test (ART) pro-tocol (155)]. Studies of the effects ofEDCs on sexual differentiation have beenconducted on members of various classes ofvertebrates including, mammals, birds,fish, and reptiles. Some invertebrate specieshave been examined as well. Such proce-dures could be used to screen chemicals forEDC activity. While such tests would notbe short term, they would be shorter and

less expensive than the current long-termrodent studies.

f) Evaluation of the need for transpla-cental carcinogenesis tests for potentendocrine disruptors, such as DES andTCDD, as well as for less potent phytoe-strogens and synthetic environmental estro-gens. Examine the utility of short-term teststo predict developmental carcinogenicity,including the methods for estrogenic chem-icals (156-157).

g) Determination whether mixtures ofendocrine-disrupting toxicants act in anadditive or nonadditive fashion at low,environmentally relevant exposure con-centrations.

Are there, or can there be, developed, short-term in vivo and in vitro techniques toscreen toxicants for endocrine disruptoractivity?While it was acknowledged that the pro-posed U.S. EPA harmonized test guidelineswould detect many or most EDCs, theseare long-term, expensive studies, and forthese reasons, short-term tests are requiredto screen more efficiently the large numberof environmental agents.

Some participants felt that in vitromethods could be used for such screening.However, because large numbers of in vitrotests would be needed due to the plethoraof mechanisms by which EDCs act (i.e.,altering hormone synthesis, transport,receptors, metabolism), there was alsostrong support for in vivo testing. In addi-tion, it is not clear how in vitro data couldor would be used in the risk assessmentprocess. It was noted that when the in vivoconcentration of active moieties of anEDC in maternal serum reached the Kifrom a receptor binding assay, then all ofthe pups would likely be severely mal-formed. Thus, in vitro data could be usedfor risk assessment if additional researchindicates that delivered doses in the rangeof the Ki are associated with adverse devel-opmental effects. It was also noted that invitro potency may not correlate well within vivo toxicity because of mechanistic andpharmacokinetic factors. Many experts inthe work group stated that in vivo screen-ing systems should be implemented, whileothers felt that a mixture of in vivo, invitro, and QSAR techniques could be usedmost efficiently to screen toxicants. It waspointed out that screening strategies wouldvary greatly based on what EDC activitythe regulatory agencies decide to screen for(i.e., all mechanisms of action versus a lim-ited subset, like estrogenicity; or all EDCs

versus those that cause developmental alter-ations). The group recognized that theseissues warrant further discussion within thescientific community to reconcile thedisparity in opinions. The work group dis-cussed selected examples of in vivo, in vitroand QSAR methods that might beemployed to screen for endocrine-disruptingactivity. It was noted that in vivo tests areoften apical while in vitro tests can be morespecific. In vivo tests are more useful ifaccompanied by target organ/cell dosimetryof the biologically active moieties. In vitrotests also need to determine the actual ver-sus administered concentration of thechemical to account for metabolism, stabili-ty, and solubility. Cellular assays must alsodetermine cell viability after toxicantadministration. The specificity and limita-tions of each assay must be clearly defined.Research is needed to develop and validatethese assays and to define a testing strategy.

Examples of in vivo methods applicableto the detection of endocrine disruptioninclude a) rodent models of transplacentalcarcinogenesis; b) sexual differentiationstudies in mammals, reptiles, fish, andbirds (in ovo); c) acute and subacute stud-ies of EDCs on endocrine systems (includ-ing the hypothalamic-pituitary-gonadalaxis, adrenal and other endocrine axes,uterine weight and biochemical responsesto estrogens, epididymal and sex accessorygland function, LH surge and ovulation,and pregnancy maintenance including invitro ovarian and placental steroidogenesis);d) pubertal alterations induced in male andfemale rats by EDCs (including landmarksof puberty, serum reproductive, adrenal andthyroid hormones, reproductive organweights, histology, and in vitro hormoneproduction from ovarian, testicular, thyroidand adrenal tissues); and e) in vivo bioassaysin rodent systems (e.g., uterine weight assay(157). Clearly, examination is needed ofthe rich diversity of species within the ani-mal kingdom for useful biomarkers of expo-sure and effect [i.e., vitellogenesis (158)].Finally, a comprehensive literature searchand further discussion should be conductedto expand the list of existing in vivo screen-ing methods used by organizations such asthe the World Health Organization and thepharmaceutical industry for the detection ofEDCs.

Examples of in vitro methodologiesapplicable to detection of endocrinedisruption include a) the MCF-7 cell pro-liferation assay for estrogen and otherreceptors (159); b) competitive receptorbinding assays for estrogens, androgens,

Environmental Health Perspectives - Vol 104, Supplement 4 * August 1996728


and progestins (89,160,161); c) transfectedcell assays for hormonal and antihormonalactivity (162-164); and d) in vitro alter-ations of whole and minced adrenal, ovari-an, testicular, and placental steroidogenesisor pituitary and hypothalamic hormoneproduction [e.g., (165)]. Where possible,attention should be focused on characteriz-ing the difference between developing andadult tissues in these assays.

In addition to direct biological assays,the work group noted that 3-D QSARmodels are being developed for ligand-receptor interactions based upon the Ki orIC50 data derived from androgen, estrogen,and progesterone competitive binding assays(166-168). These data are being used as atraining set to develop the QSAR model.Such models are then tested with chemicalswith known activity, and when validated,the model can be used to screen libraries ofcompounds with unknown activities.Similar approaches have been successful forAh receptor-ligands. Research is needed tovalidate and expand the training sets ofthese models for the above steroids; similarefforts need to be initiated for other hor-mone-receptor interactions. In theory, thisapproach could also be expanded to includetoxicant-enzyme interactions. It was notedthat although this technique has incrediblepotential utility to screen for EDCs, it canresult in false negatives with chemicalswhose structures lie outside the training set.

Summary of research needs for hazardidentification ofendocrine disruption.The emphasis for most of the followingresearch needs is on their application in thecontext of identifying the effects of EDCsin developing organisms. They are

a) Additional validated end points tosupplement current test guidelines andexpand the availability of hormonal assay,especially for nonmammalian species.

b) Tests and biomarkers to identifyacute and latent effects such as testicular andprostatic cancer, premature death, shortenedreproductive lifespan, etc.

c) Transplacental toxicology studies forcancer and noncancer end points.

d) In vivo and in vitro studies of com-plex mixtures (TEFs for EDCs).

e) Expanded developmental toxicitystudies to include postnatal observations.f ) Measurement of target organ dosime-

try in vivo to determine if adverse effects canbe predicted from in vitro IC50 or Ki values.

g) Expanded training set data for cur-rent QSAR models and development ofnew models.

h) Correlation of aquatic and wildlifemodels with mammalian models.

i) Better coordination of researchamong multidisciplinary labs.

j) Improved links between laboratoryand field studies to test hypotheses.

k) Examination of multiple end pointsas well as performance of multiple tests onEDCs.

I) Critical review of additional short-term in vitro and in vivo tests to screenEDCs in terms of cost, ease of implemen-tation, specificity, and limitations.

m) Continued discussion on the devel-opment of short-term in vitro and in vivotests because a complete assessment ofendocrine disruption likely will requirea battery of in vitro and in vivo tests. Suchinformation will facilitate the use ofmechanistic data in risk assessment.

Dose-Response AssessmentThroughout the discussions on human andecological effects of endocrine disruptors,the workshop participants consistentlyagreed that timing of exposure was criticalto the understanding of dose-responserelationships. This is true for the effects ofcancer as well as for the developmental,reproductive, immunologic, and neurologi-cal effects. Numerous examples were dis-cussed, including hormonal influences onbreast cancer where age at exposure is aknown risk factor. Similarly, endocrine dis-ruption of the developing brain can perma-nently alter behavior, whereas similar expo-sures to a fully differentiated brain could bewithout effect. Ecological and wildlifeeffects are also strongly influenced by thetiming of exposure (e.g., during the breed-ing season). Research is critical on how tim-ing of exposure to endocrine disruptorsinfluences dose-response relationships.

Before addressing the specific questionspresented to the work group, it is worth-while to review some of the general recom-mendations targeted at filling knowledgegaps that create uncertainty in dose-response evaluations. These overarchingissues include:

a) Risk assessment issues should beexplicitly considered when studies aredesigned for health or ecological effects. Ofparticular relevance is the issue of doseselection. Ideally, the doses used shouldspan a wide range to identify both toxic andmechanistic end points, and there shouldbe sufficient numbers of animals and rangeof doses to track the various end points.

b) While there may never be completeknowledge on the mechanism(s) of action

for any chemical, some knowledge onkey events could be sufficient to justify theuse of mechanistic information in dose-response evaluations.

c) Mechanistic information is most use-ful when it is linked to adverse outcomes incases in which several discrete events (mol-ecular and biological) are part of the mech-anism of action. Within technical and eco-nomic limits, dose-response informationshould be obtained on as many relevantevents as possible. Identification of mecha-nistic information on the rate-limitingsteps in the induction of toxicity is a mostuseful outcome of this line of research.

d) Population heterogeneity needs tobe characterized to improve risk assessmentdecisions. For human health, a number offactors contribute to a wide range of risks,including genetic predisposition, age(embryos, fetuses, and children are not justsmall adults), gender, diet, disease condi-tions, and past exposures. The range of riskmodulators may be even greater for com-plex ecosystems but little information isavailable in this area.

e) Effects of endocrine disrupters onecological and human health have bothdistinct and common features. Studies toidentify their common features werestrongly recommended by the dose-response work group.

f) Evaluation of the health and envi-ronmental effects of endocrine disruptorswill be most credible when information isavailable at several levels such as toxicity,studies, mechanistic and epidemiologicalstudies, and field studies. Well-plannedand coordinated multidisciplinary studiesare encouraged.

g) Increased reliance on biology andmechanisms will create an increased needto reevaluate potential risks as has occurredwith dioxin.

h) Information exchange systems forthe scientific, regulatory, medical, publicinterest, and community sectors needimprovement. The Internet affords anexcellent opportunity to disseminate rele-vant information to a broad audience.

What are the existing test guidelines andmethodologies used to evaluate endocrine-related effects? Are there areas within ourpresent guidelines that need to be refinedand improvedfor the adequate evaluationof dose-response effects of endocrinedisruptors?There are no specific guidelines to estimatedose-response relationships for endocrinedisruptors. Endocrine disruption is inferred


KAVLOCK ET AL.

from results from studies such as those oncancer or reproduction. The work group feltthat such inferences of endocrine disruptionshould lead to experimentation (toxicity,mechanistic, epidemiological, field studies)designed to characterize endocrine activity.

General recommendations for experimen-tal studies. The work group recommendedan aggressive program of improving experi-mental design that should emphasizethe following:

a) Doses should span a wide range,including environmentally relevant dosesfor human, wildlife, and ecosystem expo-sures. A sufficient number of doses shouldbe used to characterize dose-response rela-tionships for each relevant end point.

b) Studies should quantify and evaluatemultiple end points. End points should bemechanistic and biological and should bebased on existing knowledge about thestructural properties and effects of thechemical being studied.

c) End points should be quantified andlinked in the same animal or ecosystem, tothe extent possible, as well as across speciesand ecosystems to facilitate the develop-ment of biologically based models forestimating dose-response relationships.

d) The use of expanded protocolsshould be selective and based on resources,mechanistic knowledge, and risk assess-ment needs.

General recommendations for humanstudies. The work group felt that thereare opportunities to use existing humandata and to conduct new research toimprove risk assessments for endocrinedisruptors. Specifically, the work grouprecommends that

a) Existing records on occupationaland medical exposures to putative andknown endocrine disruptors should bescrutinized for associations with alteredincidences of adverse health effects.

b) It would be helpful if consensus couldbe reached on the most useful questions tobe included during medical examinations.

c) Biomarker studies need to addresscomparative responses between experi-mental systems and humans.

d) Heterogeneity in dose response,based on genetics, age, gender, and nutri-tion, needs to be characterized.

Are there unifying dose-response conceptsfor endocrine disruptors (i.e., threshold,linear, sublinear)?The work group unanimously felt that acommon dose response for all effects andfor all endocrine disruption should not be

expected. This conclusion was based on themany different kinds of hormonal actionsof chemicals categorized as endocrine dis-ruptors. These activities include estrogenic,antiestrogenic, antiandrogenic, growth fac-tor modulation, cytokine modulation,modulation of hormone metabolism, andmany others. The conclusion was alsobased on the knowledge that there are sev-eral steps in hormone action and that dif-ferent environmental agents may intervenein different processes.

Moreover, there is considerable cellspecificity in hormone action such that thesame hormone and the same receptor canproduce quantitatively and qualitativelydifferent responses depending on cell type,age, and other factors. The diverse mecha-nisms of hormone action are an active areaof research; it appears that factors such asreceptor number, DNA response elements,signal amplification, desensitization, inter-actions with transcription factors, ligandmetabolism, and the presence of cellularagonists could significantly modify dose-response relationships for any givenendocrine disruptor (169,170).

What are the research needs in theevaluation ofdose-response relationships?The work group felt that research onmechanisms of hormone action could havespin-off benefits for improving evaluationof dose-response relationships. For exam-ple, the sensitivity of cells, tissues, or devel-opmental stages to an environmental hor-mone could be predicted with improvedaccuracy. Also, it would be helpful tomodel common steps in hormone actionsuch as ligand-receptor interactionswith responsive genes. Additionally, chemi-cals that share common pharmacologicproperties may share a common responsemodel. The work group cautions, however,that endocrine disruption may occurthrough mechanisms other than binding tocellular receptors (i.e., inhibition ofenzymes of steroid hormone metabolism).The existence of multiple mechanisms bothcomplicates evaluation of dose-responserelationships and offers opportunities tobetter predict low-dose effects for differentkinds of mechanisms of endocrine disrup-tion. In addition to studies of known recep-tors, there may be unknown receptors,including orphan receptors, on which newresearch is needed.

There was considerable discussion ofknowledge gaps in mechanisms of hor-mone action, as these gaps create uncer-tainty in the evaluation of dose-response

relationships for endocrine disruptors.Although a great deal is known about theinitial steps in hormone action such asbinding to receptors and transcriptionalactivation of responsive genes, very little isknown about how those changes in geneexpression lead to biological effects such ascell proliferation. This lessens the credibilityof biologically based models for predictingdose-response relationships for endocrinedisruptors. However, several work groupmembers felt that such models still repre-sent an improvement over the defaultapproaches currently used in risk assess-ments. For example, it may be possible tocompare potencies of chemicals apparentlyacting through common mechanisms (i.e.,binding to a specific receptor) and to betterpredict targets or potentially sensitive cellsor tissues. It is also important to distin-guish steps in the mechanism that can alterthe slope of the dose-response curve,potency, or target organ specificity.

The role of interactions betweenendogenous and exogenous hormones wasdiscussed. This led to the recommendationthat baseline data on endogenous hor-mones, including cyclicity and differencesin tissue concentrations, are needed notonly to improve the ability to predict theenvironmental or health consequences ofendocrine disruptors but also as input intomathematical models of development andendocrine system function.

Finally, issues related to the dose-response assessment of mixtures wereaddressed. Considerable debate existsregarding about dose-response relation-ships for single compounds even when areasonable amount of toxicologic data isavailable. Yet humans and ecosystems areexposed to a vast array of chemicals thatmay interact by potentiation, synergism,inhibition, and antagonism of toxic effects.Although progress will be slow in riskassessment of mixtures, the followingrecommendations were developed:

a) Include environmentally relevantdoses and ratios of chemicals in ecological,wildlife, and experimental animal studies

b) Use the knowledge of mechanismsand appropriate biomarkers (sensitive andspecific) to dissect major contributors totoxicity

c) Determine the pharmacokinetics oftoxic chemicals, within mixtures, and thedose at target tissue

d) Improve TEF estimates for environ-mental agents that interact with specificendogenous receptors (e.g., Ah and estrogenreceptor ligands) by systematically testing

Environmental Health Perspectives - Vol 104, Supplement 4 - August 1996730


assumptions inherent to the TEQ approach.These assumptions include the similarity ofmechanism and the role of persistence andaccumulation. The work group also felt thatTEF estimates would be improved by care-ful selection of experimental systems used todetermine relative potencies and the use ofappropriate dose ranges, including low dosesto separate direct from indirect effects onthe endocrine system.

Are there particular elements to beconsidered in dose-response evaluationin wildlife species?The work group felt that evaluation ofdose-response relationships to assess ecosys-tem and wildlife effects will be extremelydifficult but would benefit by implementa-tion of the following considerations:

a) Improved understanding of the fac-tors, such as genetic diversity, that arecritical to the maintenance of integratedecosystems. It will be difficult to character-ize dose-response relationships withoutimproved knowledge of normal fluctuationsin ecosystem constituents.

b) Laboratory and field studies need tobe better coordinated. In addition, betterdefinition is needed of what constitutesfield validation of a laboratory finding aswell as laboratory verification of an effectobserved in the field.

c) Guidelines need to be developed forselection of sentinel species and evaluationof dose-response relationships. Guidelinesfor dose-response assessment should begeneral and not overly prescriptive.

Is itfeasible to develop complete biologicallybased dose-response models for endocrinedisruptors?The work group encouraged the develop-ment of biomathematical models forendocrine systems, recognizing that therewill not be a single model applicable to allhormones. However, the shared steps forsome systems (e.g., hormone receptor bind-ing) could be modeled using the same set ofequations. These models should incorpo-rate data from humans, rodents, and theecological species. Development of thesemodels must include time dependence ofhormonal action.

Statistical and mathematical models fortoxic end points such as cancer, reproduc-tion, or development should be helpful inidentifying knowledge gaps that createuncertainty in dose-response estimates and,in this way, can focus available resources inthe most productive way. These efforts willrequire increased resources for fostering the

multidisciplinary research needed for thedevelopment of models for health andenvironmental effects. Establishing cross-disciplinary programs for mathematicaland simulation training for biologists andbiological training for statisticians andmathematicians would enhance the credi-bility of biologically based models forendocrine disruptors.

Exposure AssessmentWhat chemicals in the environment are ofconcernfor endocrine disruption?

The work group agreed that developing acomprehensive list of putative endocrinedisruptors would take longer than the timeavailable at the workshop. Any list so con-structed would not be entirely accurate orcomprehensive at this time. Nonetheless,from the viewpoint of exposure assess-ment, it was agreed that EDCs should becategorized as follows:

Use pattern-for example, herbicides,insecticides, fungicides, hormones, etc.

Chemical structure or class-forexample, dioxins, halogenated biphenyls,alkyloxyphenols, etc.

Biological function and mechanism ofaction-for example, estrogen mimic,androgen inhibitor, etc.

What do we know about the status andtrends ofputativeEDC?The initial discussion focused on what weknow and what we need to know about thestatus and trends of putative EDCs.Environmental and tissue levels of someEDCs, such as DDT and its analogs, andPCBs, have declined in some countries andin most areas of the United States inresponse to regulation (171-178). Uncer-tainty still exists regarding future trends ofthese compounds, however, because of off-shore inputs and releases from stored mate-rials. It also appears that for many EDCs,the environmental concentrations thatdeclined from the mid-1970s through theearly 1980s have now reached a plateau(179,180). This is a cause for concern. Formost other EDCs, particularly new chemi-cals or chemicals that have not been rou-tinely monitored, the trends are unknown.

The information needs regarding statusand trends of EDCs are considerable. Abetter understanding is needed of the fateand transport of new and existing chemi-cals, particularly among the different envi-ronmental compartments (water, sediment,biota). Key concerns have been raised aboutair and water serving as transport media

and exposure routes for EDCs for bothhumans and wildlife. The importance ofwater use practices and how they impacton exposure must be more thoroughlyinvestigated. Water use practices that con-tribute to EDC exposure, such as agricul-ture, sewage discharge, unfinished waters,and unregulated drinking water sources,require review. Any research on water usemust consider factors that affect flow anddilution such as season, diversion, andregulated release of impounded waters.

Exposure assessment, particularly as itinvolves human health, must focus on vul-nerable groups, both in terms of life stageand lifestyle. Exposure assessment for thecritical development stages is a highresearch priority. This includes pregnancy,gestation, lactation, adolescence, and senes-cence. Vulnerability of different groups inthe population will be affected by lifestylefactors such as subsistence hunting andfishing and avid sportsmen who consumefish and wildlife, or host factors such asmetabolic differences among polymorphicgroups, special dietary habits, and age (e.g.,the types and rates of food consumption inchildren). While the work group agreedthat diet would likely be the major exposureroute, an approach based on integratedexposure assessment needs to be taken. Allroutes should be examined (e.g., dermal,inhalation, and ingestion). The work groupstressed the importance of a global perspec-tive on exposure. EDCs that may haverestricted or no use in the United States arestill used in other countries and may becomesources of exposure through either importedfood or atmospheric transport and deposi-tion. Further, the potential for human orwildlife exposure to multiple chemicals thatmay function as EDCs should be includedin any exposure assessment.

The most critical need on status andtrends is for the continuation and improve-ment of monitoring of the environment forthe presence and magnitude of contami-nants. Existing programs that furnishrepeated measures of chemical contamina-tion in the environment or in food provideour only indication of whether exposure isincreasing or decreasing, and to what mag-nitude. Therefore, the work group stronglyrecommended continuation of existingprograms such as the National HumanHealth and Nutrition Examination Survey(NHANES) (181); the National HumanExposure Assessment Survey (NHEXAS)(182); the Market Basket Analysis of theU.S. Food and Drug Administration; thePesticide Data Program of the U.S.


KAVLOCK ET AL.

Department of Agriculture; the NationalContaminant Biomonitoring Program of theNational Bureau of Standards (173,183);and the NOAA Status and Trends Program(184). Existing international monitoringefforts must be continued. For new pro-grams or for improving existing programs,the work group felt analytical expensescould be reduced through specificity ofanalysis. Where markers or bioassays areable to replace chemical analysis, theyshould be used, and analytes should be tar-geted to meet the specific monitoring needof the designed program.

More research needs to be focused ondevelopment and validation of monitoringtools. Current results with caged organismshave been successful because they providebiologic relevance to the estimation of expo-sure. Despite the drawbacks inherent inusing caged organisms, such as stress associ-ated with captivity and the absence of expo-sure during the sensitive reproductive anddevelopmental stages, the work group feltthis tool warranted further development.Research also should continue on othermonitoring tools such as in situ samplersthat mimic biological tissue, or assays usedwith field grab samples. As field tools tomeasure exposure assessment become moreavailable, concurrent research in toxicityidentification evaluation (TIE) procedures(185) must occur to enable verification ofcausative agents.

What are the assumptions in estimatingexposure to endocrine disruptors?Valid exposure assessment requires that cer-tain assumptions be either accepted orproven. The work group assumed that envi-ronmental concentrations did not equate toexposure because of several factors. Activitypatterns such as migratory behavior, bioavail-ability of compounds, and lifecycle stage areexamples of factors that limit the use of envi-ronmental concentrations. Body burdensmay be much better indicators of exposure,especially for persistent chemicals, but theyalso may vary seasonally and with age (186).However, research is needed to develop tis-sue-specific body burdens and determinehow they translate to specific dose. This willrequire full understanding of the physiologyand toxicokinetics of the compound and itshost. Assessing bioavailability is complicatedby transplacental boundaries, interactionwith transport molecules, and matrix binding(e.g., sediment particles, dissolved organiccarbon) and should be a focus of research.

Another critical assumption can bemade based on knowledge of the actions of

EDCs; that is, a population of breedingadults with apparently healthy offspring, oryoung in the population does not necessar-ily indicate health. The ability of someEDCs to evoke transgenerational effectsrequires a more in-depth look at the healthof a population. Reproductive capacity andnormal reproductive functioning of theoffspring may be the most sensitive test ofpopulation health. Research is needed todevelop preclinical indicators of transgen-erational impacts so that they can be linkedto strategies that limit or eliminate exposure.

To what degree can body burdens in thegeneralpopulations ofhumans and wildlifebe identified by exposure or hostfactors?One of the research needs that wasidentified and given the highest priority wasthe further development and validation ofindicators of exposure to EDCs. Whetherthese indicators are biomarkers or bioassays,they must be sensitive and specific to bothpersistent and nonpersistent compounds.These indicators must be validated on aspecies-by-species basis. Examples of somemeasures that have been applied includevitellogenin production in male fish(158,187) and induction of cytochromeP450 (188), although in both instances,lack of response does not necessarily equateto lack of exposure.

In addition to bioindicators, otherindices that may be termed population mea-sures need to be applied to EDC exposure.Sperm counts, sex ratios, and incidence of aspecific tumor or abnormality are examples(74,189-191). If these prove to be indicesof effect, changes in these population mea-sures could be used to identify populationspotentially exposed to EDCs. Exposureindicators then could be used to quantifythe dose-response relationship. The workgroup proposed a scenario in which indica-tors could be used in a tiered hierarchy thatwould demonstrate evidence of exposure toEDCs; clarification of the mechanism ofeffect (e.g., estrogen mimic, androgeninhibitor); and TIE methods to identifyresponsible compounds. Given these needs,indicators should be sensitive at ambientlevels of contamination and specific tochemicals and end points that would enableidentification of mechanism of action.

Are there specific confounding factors ortemporalpatterns ofeposure to be consid-ered in the assessment ofpublk and wildlfehealth implications ofendocrine disruptors?Estimating the exposure of humans andwildlife to potential EDCs creates a unique

set of confounding factors. While this list iscertainly not comprehensive, the work groupidentified the following significant factors:

a) Time lags between exposure andeffect: The transgenerational nature ofsome EDC effects may be the single mostcomplicating factor. All of the potentiallatent effects that may occur from short-term exposures during critical developmen-tal windows have not yet been identified.

b) Seasonality: Because of the sensitivi-ty of reproductive stages to EDCs, season-ality will be extremely important to wildlife.In addition, the association of EDCs withthe aquatic environment is complicated byseasonal rainfall, storm runoff, and waterreleases.

c) Species variability: Perhaps this is nomore important for EDC exposure thanwith any other toxicant effect. However,the work group particularly stressed thatmore basic research is needed on generalendocrine physiology of target species.

d) Multiple chemical exposures: This,too, is a confounding factor for any toxi-cant. It is especially identified here becauseof the potential for joint toxic action andthe presence of naturally occurringphytoestrogens.A comprehensive list would identify a

number of general confounding factors suchas past exposure history, occupational expo-sures, nutritional status, trophic structureand other relevant factors.

Summary of research needs for exposureassessment.A number of research needs were identifiedin the previous section. The following listwas deemed to be of the highest priority:

a) Monitoring: Maintain and increasemonitoring efforts that help identify statusand trends of EDCs, including chemicalmonitoring programs and population mea-sures such as wildlife reproductive success,sperm count studies, etc. These programsare critical to the identification of popula-tions at risk.

b) Biomarkers: Develop and validatebiological indices for use as screening tools.More biomarkers, bioassays, and popula-tion measures are needed that are sensitiveat relevant environmental concentrationsand are mechanism specific.

c) Exposure hierarchies: Develop anapplied, integrative iterative hierarchy ofexposure indicators. Components shouldrange from screening level tests to muchmore predictive indicators. This approachshould be designed to avoid false negativesin initial tiers and progress in final tiers to



more predictive indicators that are linkedto effects.

Risk CharacterizatonThe National Academy of Sciences' RiskAssessment paradigm was used as a tool toidentify research needs. In this case, theapproach had limited success since themajority of the group members had limitedexperience with the specific composition ofa risk characterization. The work group,however, did identify several major pointsfor which there was broad agreement.

Normal functioning of the endocrinesystem encompasses a wide fluctuation ofhormone and other biological indices,which reflect among other things, circadi-an rhythm, season (temperature, light),age, and gender. For example, concentra-tion extremes of sex hormones occur atspecified times for normal physiologic func-tions; such periods include sexual differ-entiation, puberty, reproductive cycles, par-turition, lactation, and menopause. Thepresence and magnitude of certain hor-mones are critical to normal ontogenicdevelopment, including some neurobehav-ioral programming, in a broad range ofbiota that include insects and other inver-tebrates, amphibia, reptiles, fish, birds,and mammals. Studies in wildlife and inthe laboratory have demonstrated thatcertain chemicals may perturb specificendocrine functions at specified periodsduring the life span. The consequences ofexposure during these windows of sensi-tivity may manifest themselves at laterperiods of life. To better understand andpredict specific circumstances underwhich adverse effects may occur, it isessential that there be better identificationand characterization of these critical expo-sure windows. The homology of these sen-sitive periods across species is also impor-tant for accurate risk prediction.

To understand the nature and degree towhich an adverse effect reflects response toa nonendogenous chemical with endocrineproperties or alters the type and magnitudeof endogenous hormones often requiresconsideration of a broad range of biologicinteractions. For example, the role of agentsfrom environmental sources such as phy-toestrogens or estrogenic products fromfungi should be considered. It was recom-mended that the current literature on phy-toestrogens be reviewed as a prerequisitefor determining additional data needs inthis area. A comprehensive bibliographyfor phytoestrogens is available on theInternet (unpublished data).

The traditional risk assessment para-digm can be used to adequately identify andcharacterize chemicals that cause adverseeffects by altering endocrine processes.However, the assessment paradigm shouldbe tailored to permit consideration of suchfactors as normal fluctuations of endoge-nous levels of hormones, impact of otheragents (phytoestrogens), the existence ofcritical windows of sensitivity, and the needto understand the significance of subtleeffects at low doses. In other words, doseeffect may need to be expressed in severalcontexts such as adult effects, adult doseeffect in a particular physiologic state (suchas immediately postpartum or during lacta-tion), or at a particular window of sensitivityfor causing developmental toxicity.

Endocrine effects often occur subse-quent to receptor-ligand binding. A varietyof agents may interact with a binding siteacting as agonists, partial agonists, or antag-onists. Since exposures may entail simulta-neous interaction with various endogenousand exogenous substances, it is critical thatreceptor theory be used to develop or refinequantitative models for estimating theeffects of such exposures.

The limited utility, to date, of SAR topredict biological effect for estrogenicagents was noted. Perhaps a useful criterionfor screening EDCs for more robust studyis whether they can elicit (or inhibit in thecase of antagonists) a transcriptional event.Such a criterion would only serve to screenout agents; agents that met this sort of func-tional definition would still need to be char-acterized before determining if they haveactual toxicological potential. However,such tests would still be limited in theirabilities to detect the action of metabolites.

The presumed receptor-based mecha-nisms, responsible for at least some adverseendocrine-modulated effects, present aunique opportunity to establish a commonbiologically relevant risk assessment processfor all effects, i.e., developmental, immuno-logic, neurological, and carcinogeniceffects. The group is not aware of a biolog-ical basis for selecting different models toquantitatively estimate cancer or non-cancer effects for chemicals that act byendocrine-mediated mechanisms.

Summary RecommendationsThe majority of the participants at theworkshop agreed that the endocrine dis-ruptor hypothesis was of sufficient concernto warrant a concerted research effort. Inparticular, the study of potential effects onthe development of reproductive capability

at multiple phylogenetic levels was deemedthe most important area in need of atten-tion. It was repeatedly emphasized that thedeveloping embryo, fetus, and neonateshould not be viewed as small adults andthat the processes of development areespecially vulnerable to brief periods ofendocrine disruption. However, for manyof the effects reported in both wildlife andhumans that have been attributed to, orassociated with, endocrine disruption, expo-sure assessment has generally been inade-quate for quantitative risk assessment.Because of this, some participants felt it wasdifficult to critically evaluate and establishthe level of priority relative to other researchtopics. Still other participants reminded thework group not to lose sight of the presenceof naturally occurring endocrine disruptors(e.g., phytoestrogens) as the effects of man-made chemicals are studied.

Several general comments emanatedfrom the discussions. These include therecognition that there was a great advan-tage in bringing together a multidiscipli-nary group of scientists representing boththe human health and ecological healthviewpoints to help identify common issuesand that this interaction must be nurturedas the research agenda unfolds. The workgroup noted that some key similarities anddifferences exist between endocrine disrup-tors and other chemicals that can causeadverse biological effects. Two of the keydifferences are the presence of natural lig-ands within the body that must interact atsome level with the exogenous chemical;and that the concentrations of the naturalligands within the body fluctuate duringthe life cycle and must be maintained with-in narrow limits at key times during devel-opment. This latter point indicates thattiming of exposure is a very significant fac-tor in any assessment. Last, the mechanisticbasis of the interaction with biological sys-tems presages the induction of subtleeffects at low doses that must be interpretedas to whether or not the effects are adverse.As the level of organization at which bio-logic responses to endocrine disruptors areobserved decreases (e.g., from physiologicto cellular to molecular), the challenge todescribe the effects as adverse at the level ofthe individual and the population increases.In this regard, endocrine disruptors are notunlike other types of chemicals for whichtoxicologic information is amassed.

In general, it was felt that linkingspecific exposures to specific effects in thegeneral environment would often be diffi-cult because of the complexities of exposure,

Environmental Health Perspectives - Vol 104, Supplement 4 - August 1996 733

KAVLOCK ET AL.

the latency of the effects, and, at times, thesubtle nature of the outcomes. Therefore,confirmation of the validity of the hypoth-esis will rely heavily on application of theHill criteria (192,193) for causality(strength of the association, presence of adose-response relationship, specificity ofthe association, consistency across studies,biological plausibility, and coherence of theevidence). Such considerations will havesignificant impact on the types of researchactivities necessary to adequately confirm

or refute the central hypothesis. A compos-ite representation of the identified needs isprovided in Table 2. Ten broad categoriesof research needs were identified: basicresearch, biomarkers, database develop-ment, exposure determination, exposurefollow-up, hazard identification, mixtures,multidisciplinary studies, risk assessmentmodels, and sentinel species. Withineach of these categories, the work groupsthat identified each need are indicated.The neurological, immunological and

carcinogenic work groups all noted thecomplexity of identifying whether effects ofxenobiotics on those systems were theresult of primary or secondary aspects ofendocrine disruption. This concept isdeveloped in Figure 1, which portrays notonly the interaction between the endocrinesystem and the target organs but also theinteractions among the target organs them-selves. For these reasons, identifying agentsas direct or indirect endocrine disruptors isproblematic, and necessitates research to

Table 2. Composite research needs identified by the eight work groups.

Research area Work group origina Research need

Basic research R, N, 1, DR

Biomarkers

Database development

Exposure determination

Exposure follow-up

Hazard identification

Mixtures

Multidisciplinary studies

Risk assessment models

Sentinel species

RC

R, N, DRDR, RC

RCDR

C, R, N, /, DREXHDRHDRRCC

EX

C, R, N, IHD

R, HDRC

R, HDN, 1, HD

NR, HD, DR

R, I,DRCN

R, HD, DRHD

N, I,DRHDR

DREX

C, DRN

DR, RCDR

C, N, /, DRC

Understanding the cellular and molecular mechanisms, including nonreceptor mechanisms, for EDCsSensitive, inexpensive, and widely available analytical toolsAnimal and cellular models of endocrine-mediated tumorsOntogeny of receptor-based systems and role in regulating developmentUnderstanding of the mechanisms and biological significance of subtle low-dose effectsIdentify and characterize critical windows of susceptibility across speciesCharacterize source of population heterogeneity in responsiveness (age, gender, nutrition, etc.)Development of biomarkers of exposure and effects of EDCsDevelop and validate biological indices as screening tools for exposure assessmentDevelopment of biomarkers for latent effectsInformation on normal population variation, regionnal, and seasonal effectsCritical review of short-term tests for EDCsProspective and retrospective reproductive health trendsField data on hormone levels, body burdens, and gene expression markersSystematic potency comparison of endogenous versus exogenous substancesSurveillance systems for cancer incidence and mortality in wildlifeRapid and inexpensive exposure monitoring methods for use in wildlife populationsIncreased monitoring efforts to identify status and trends of EDCsMultidisciplinary teams to study exposed populationsAutoimmune symptomology, hypersensitivity, and disease in EDC-exposed humansCoordinated research on exposed humans, wildlife, and sentinel speciesTarget organ dosimetry for comparison with ligad-binding affinitiesExpanded development of QSAR models for hazard detection and rankingIdentification of transcriptional events after ligand binding as QSAR inputDevelopment and validation of apical methods to detect EDCsPerinatal and multigenerational exposure toxicity studies for cancer and noncancer effectsResearch to address the additivity principle for mixturesIn vitro and in vivo studies of complex mixtures to evaluate validity of TEFsIdentification and testing of environmentally relevant mixturesSystematic evaluation of species-, cellular-, and age-dependent response to mixtures of EDCsSystematic field and laboratory studies focused on critical uncertaintiesLaboratory-field hypothesis-based studies and improved information exchangeExamination of correlation of effects between wildlife and mammalian modelsMultidisciplinary studies on effects of endocrine disruptionExamination of multiple end points and multiple tests of ED actionStatistical models to predict risk from exposure and effectsImprovements in study design (dose selection, end points, end point linkages)Development of applied integrative iterative hierarchy of exposure indicatorsEvaluation of toxicity and mechanistic end points across species common steps and chemical classes)Toxicokinetics and toxicodynamic studies of environmentally relevant chemicalsQuantitative dose-response models based upon receptor theory and biochemical interactionsEstablishment of training programs in biomathematics for BBDR model constructionIdentification and monitoring of differentially susceptible sentinel speciesCancer studies in domestic animals and pets


"BBDR, biologically based dose-response model; C, carcinogenesis; R, reproductive toxicity; N, neurotoxicity; 1, immunotoxicity; HD, hazard detection; DR, dose response; EX,exposure assessment; RC, risk characterization. Designations in bold and italics were among the highest priority needs identified by the respective work groups. In general,carcinogenic and reproductive studies were considered to be higher priorities for biological effects research, while exposure assessment was a recognized deficiency in mostpopulation studies.

734


carefully define dose-response relationshipsacross multiple end points, to delineate theproximate mechanism of action, and toascertain the complete organismal responseto an environmental exposure. However,in terms of protection of human or wildlifepopulations, it is less important to deter-mine whether effects are primary orsecondary once they are observed in thegeneral environment.

Workshop participants were in generalagreement that the highest priorities forbiologic effect research lie in the areas ofdevelopment of reproductive capabilityand carcinogenesis, as these end pointshave had greater documentation of beingadversely affected by alterations inendocrine function than have either thenervous or immune system. However, theincidence of effects on the nervous andimmune systems may be underestimatedat this point because of incompletecharacterization of the biologic effects ofendocrine disruptors. In addition, itshould be noted that many reproductiveeffects, especially those involving a behav-ioral component, are mediated by effectson neuroendocrine function. In particular,identification and characterization ofeffects on the developing reproductive sys-

tem were considered to be high priorityfor additional research because of the highsensitivity and frequent irreversibility ofeffects after even brief exposures. Morerefined exposure assessments and researchon the toxicology of mixtures were alsoconsidered to be of great importance.Special emphasis was placed on the uniquechallenges endocrine disruptors mightpose to the risk assessment paradigm.Interestingly, understanding the basicmechanisms of endocrine disruptioninduced by various chemicals was seen asan advantage in that this knowledge mayresult in a common, biologically based,risk assessment process for all effects (i.e.,both cancer and noncancer).

At least three outcomes are expectedfrom the workshop. The first of these isthe publication of this report in thescientific literature so that it is readilyavailable to both researchers and the pub-lic. Second, within the U.S. EPA, theOffice of Research and Development inconjunction with input from the ProgramOffices will be developing an augmentedresearch initiative beginning in Fiscal Year1996 to implement some of the recom-mendations of the workshop. To assist infurther focusing the research needs for

endocrine disruptors for research support-ed by the U.S. EPA, a workshop on ecolog-ical research needs was held in Duluth,Minnesota, in June 1995 (151). Personsseeking information from that effortshould contact Dr. Gary Ankley of theMED/NHEERL/USEPA in Duluth.Finally, an Endocrine Disruptor ResearchCoordination Workshop has been formedunder the auspices of the National Scienceand Technology Council that will a) devel-op a federal research strategy that addressesthe key scientific uncertainties forendocrine disruptors, b) inventory relatedongoing federal research efforts, c) identifyresearch gaps between ongoing researchprograms and needs identified in the strate-gic plan, and facilitate cooordination andcooperation across the federal governmentto address them, d) initiate outreachefforts to engage public interest, privatesector, and international groups with inter-est in this issue, and e) promote education-al activities such as symposia and work-shops to disseminate endocrine disruptorinformation across the scientific communi-ty. Persons wishing to find out more aboutthis effort should contact the senior authorof this report.

REFERENCES

1. Colborn T, Clement C, eds. Chemically induced alterations insexual and functional development: the wildlife/human connec-tion. Advances in Modern Environmental Toxicology, Vol XXI.Princeton, NJ:Princeton Scientific Publishing, 1992;401 pp.

2. Adams NR. Organizational and activational effects of phyto-estrogens on the reproductive tract of the ewe. Proc Soc ExpBiol Med 208:87-91 (1995).

3. Stob M. Naturally occurring food toxicants: estrogens. In:Handbook of Naturally Occurring Chemicals (Recheigl M,ed). Boca Raton, FL:CRC Press, 1983;81-102.

4. Medical Research Council. IEH Assessment on EnvironmentalOestrogens: Consequences to Human Health and Wildlife.Leicester, England:University of Leicester, 1995; 105 pp.

5. Danish Environmental Protection Agency. Male ReproductiveHealth and Environmental Chemicals with Estrogenic Effects.Miljoprojekt 20. Copenhagen:Ministry of the Environmentand Energy, 1995.

6. Umweltbundesamt. Endocrinically Active Chemicals in theEnvironment. Texte Series 3, Vol 36. In: Expert Round, 9-10March 1995, Berlin, Germany. Berlin:Umweltbundesamt,1995;151 pp.

7. Natural Research Council. Risk Assessment in the FederalGovernment: Managing the Process. Washington:NationalAcademy Press, 1983; 191 pp.

8. Baumann PC, Harshbarger JC, Hartman KJ. Relationshipbetween liver tumors and age in brown bullhead populationsfrom two Lake Erie tributaries. Sci Total Environ 94:71-87(1990).

9. Harshbarger, JC, Clark, JB. Epizootiology of neoplasms in bonyfish ofNorth America. Sci Total Environ 94:1-32 (1990).

10. Meyers MS, Stehr CM, Olson AP, Johnson LL, McCain BB,

Chan S-L, Varanasi U. Relationships betweeen toxicopathichepatic lesions and exposure to chemical contaminants inEnglish sole (Pleuoronectes vetulus), starry flounder (PlatichtysStelatus), and white croaker (Genyonemus lineatus) from selectedmarine sites on the Pacific Coast, USA. Environ HealthPerspect 102:200-215 (1994).

11. Couch JA, Harshbarger JC. Effects of carcinogenic agents onaquatic animals: an environmental and experimental overview.Environ Carcinog Rev 3:63-105 (1985).

12. Beland P, DeGuise S, Girard C, Lagace A, Martineau D,Michaud R, Muir ECG, Norstrom RJ, Pelletier E, Ray S, ShugartLR. Toxic compounds and health and reproductive effects in St.Lawrence beluga whales. J Great Lakes Res 19: 766-775 (1993).

13. DeGuise S, Martineau D, Beland P, Fournier M. Possiblemechanisms of action of environmental contaminants on St.Lawrence beluga whales (Delphinapterus leucas) Environ HealthPerspect 103:73-77 (1995).

14. Herbst AL, Ulfelder H, Poskanzer DC. Adenocarcinoma of thevagina. Association of maternal diethylstilbestrol therapy withtumor appearance in young women. N Engl J Med284:878-881 (1971).

15. Kaplan NM. Male pseuodhemaphrodism: report of a case withobservations on pathogenesis. N Engl J Med 261:641-644(1959).

16. Gill, WH, Schumacher, FB, Bibbo, M., Straus, FH,Schoenberg, HW. Association of diethylstilbestrol exposure inutero with cryptochidism, testicular hypoplasia and semenabnormalities. J Urol 122:36-39 (1979).

17. Driscoll SG, Taylor SH. Effects of prenatal estrogen on themale urogenital system. Obstet Gynecol 56:537-542 (1980).

18. Penny R. The effcts of DES on male offspring. West J Med


KAVLOCK ET AL.

136:329-330 (1982).19. Leary FJ, Resseguie LJ, Kurland LT, O'Brien PC, Emslander

RF, Noller KL. Males exposed in utero to diethylstilbestrol.JAMA 252:2984-2989 (1984).

20. Wilcox AJ, Baird DD, Weinberg CR, Hornsby PP, Herbst AL.Fertility in men exposed prenatally to diethylstilbestrol. N EnglJ Med 332:1411-1415 (1995).

21. Jensen AA, Slorach SA. Chemical contaminants in humanmilk. Boston:CRC Press, 1991.

22. Sonawane BR. Chemical contaminants in human milk: anoverview. Environ Health Perspect 103(Suppl 6):197-205(1995).

23. Unger M, Olsen J, Clausen J. Organochlorine compounds inthe adipose tissue of deceased persons with and without cancer:a statistical survey of some potential confounders. Environ Res29:371-376 (1984).

24. Mussalo-Rauhamaa H, Hasanen E, Pyysalo H, Antervo K,Kauppila R, Pantzar P. Occurrence of 3-hexachlorocyclohexanein breast cancer patients. Cancer 66:2124-2128 (1990).

25. Austin H, Keil JE, Cole P. A prospective follow-up of cancermortality in relation to serum DDT. Am J Public Health79:43-46 (1989).

26. Falck F, Ricci A, Wolff MS, Godbold J, Deckers P. Pesticidesand polychlorinated biphenyl residues in human breast lipidsand their relation to breast cancer. Arch Environ Health47:143-146 (1992).

27. Wolff MS, Toniolo PG, Lee EW, Rivera M, Dubin N. Bloodlevels of organochlorine residues and risk of breast cancer.J Natl Cancer Inst 85:648-652 (1993).

28. DeWaily E, Dodin S, Verreault R, Ayotte P, Sauve L, Morin J,Brisson J. High organochlorine body burden in women withestrogen receptor-positive breast cancer. J Natl Cancer Inst86:232-234 (1994).

29. Krieger N, Wolff MS, Hiatt RA, Rivera M, VogelmanJ, Orentreich N. Breast cancer and serum organochlorines: aprospective study among white, black, and Asian women. JNatl Cancer Inst 86(9):589-599 (1994).

30. Henderson AK, Rosen D, Miller GL, Figgs LW, Zahm SH,Sieber SM, Humphrey HB, Sinks T. Breast cancer amongwomen exposed to polybrominated biphenyls. Epidemiology6(5):544-546 (1995).

31. Brown DP. Mortality of workers exposed to polychlorinatedbiphenyl-an update. Arch Environ Health 42:333-339 (1987).

32. Sinks T, Steele G, Smith AB, Rinsky R, Watkins K. Risk fac-tors associated with excess mortality among polychlorinatedbiphenyl exposed workers. Am J EpidemioI 136:389-398(1992).

33. Reis LAG, Miller GA, Hankey BF, Kosary CL, Harras A,Edwards BK, eds. SEER Cancer Statistics Review, 1973-1991:Tables and Graphs. NIH Publ No 94-2789. Bethesda,MD:National Cancer Institute, 1994.

34. White E, Lee CY, Kristal AR. Evaluation of the increase inbreast cancer incidence in relation to mammography use. J NatlCancer Inst 82:1546-1552 (1990).

35. Feuer EJ, Wun LM. How much of the recent rise in breast can-cer incidence can be explained by increases in mammographyutilization? A dynamic population model approach. Am JEpidemiol 136:1423-1436 (1992).

36. Potosky AL, Miller BA, Albertson PC, Kramer BS. The role ofincreasing detection in the rising incidence of prostate cancer.JAMA 273(7):548-552 (1995).

37. Wetzel LT, Leumpert LG III, Breckenridge CB, Tisdel MO,Stevens JT, Thakur AK, Extrom PJ, Eldridge JC. Chroniceffects of atrazine on estrus and mammary tumor formation infemale Sprague-Dawley and Fischer 344 rats. J Toxicol EnvironHealth 43:169-182 (1994).

38. Stevens JT, Breckenridge CB, Wetzel LT, Gillis JH, LuempertLG III, Eldridge JC. Hypothesis for mammary tumorigenesisin Sprague-Dawley rats exposed to certain trizaine herbicides. JToxicol Environ Health 43:139-153 (1994).

39. Dunnick JK, Elwell MR, Huff J, Barrett JC. Chemically-induced mammary gland cancer in the National Toxicology

Program's Carcinogenesis Bioassay. Carcinogenesis 16:173-179(1995).

40. Bertazzi PA, Pesatori AC, Consonni D, Tironi A, Landi MT,Zoccheti C. Cancer incidence in a population accidentallyexposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Epidemiology4:398-406 (1993).

41. Landrigan PL, Wilcox KR, Silva J, Humphrey HEB, KauffmanC, Heath CW. Cohort study of Michigan residents exposed topolybrominated biphenyls: epidemiologic and immunologicfindings. Ann NYAcad Sci 320:284-294 (1979).

42. Kuratsuna M, Nakamura Y, Ikeda M, Hirohita T. Analysis ofdeaths seen among patients with Yusho. A preliminary report.Chemosphere 16:2085-2088 (1987).

43. Hsu CC, Yu MM, Chen YJ, Guo YL, Rogan WJ. The Yu-Cheng rice-oil incident. In: Dioxin and Health (Schecter A,ed). New York:Plenum Press;1994, 661-684.

44. McConnell EE. Comparative response in carcinogenesis bioas-says as function of age at first exposure. In: Similarities andDifferences between Children and Adults: Implications forRisk Assessment (Guzelian PS, Henry CJ, Olin SS,eds).Washington:ILSI Press, 1992;66-78.

45. Chhabra RS, Eustis JK, Haseman JK, Kurtz PJ, Carlton BD.Comparative carcinogenicity of ethylenethiourea with andwithout prenatal exposures in rats and mice. Fundam ApplToxicol 18:405-417 (1991)

46. Chhabra RS, Bucher JR, Haseman JK, Elwell MR, Kurtz PJ,Carlton BD. Comparative carcinogenicity of 5,5-diphenylhy-dantoin with and without prenatal exposures in rats and mice.Fundam Appl Toxicol 21:174-186 (1993).

47. Chhabra RS, Bucher JR, Haseman JK, Elwell MR, Kurtz PJ,Carlton BD. Comparative carcinogenicity of polybrominatedbiphenyl with and without prenatal exposures in rats and mice.Fundam Appl Toxicol 21:451-460 (1993).

48. White DH, Hoffman DJ. Effects of polychlorinated dibenzo-p-dioxins and dibenzofurans on nesting wood ducks (Aix sponsa)at Bayou Meto, Arkansas. Environ Health Perspect 103:37-39(1995).

49. Peakall DB, Fox GA. Toxicological investigations of pollutant-related effects in Great Lakes gulls. Environ Health Perspect77:187-193 (1987).

50. Hoffman DJ, Rattner BA, Sileo L, Docherty D, Kubiak TJ.Embryotoxicity, teratogenicity and arylhydrocarbon hydroxy-lase activity in Foster's terns on Green Bay, Lake Michigan.Environ Res 42:176-184 (1987).

51. Colborn T. Epidemiology of Great Lakes bald eagles. J ToxicolEnviron Health 33:395-453 (1991).

52. Gilbertson M, Kubiak T, Ludwig J, Fox GA. Great Lakesembryo mortality, edema, and deformities syndrome (Glemeds)in colonial fish-eating birds: similarity to chick-edema disease.J Toxicol Environ Health 33:455-520 (1991).

53 Fox GA, Collins B, Hayakawa E, Weseloh DV, Ludwig JP,Kubiak TJ, Erdman TC. Reproductive outcomes in colonialfish-eating birds: a biomarker for developmental toxicants inGreat Lakes food chains. J Great Lakes Res 17:158-167 (1991).

54. Fox GA, Gilbertson M, Gilman AP, Kubiak TJ. A rationale forthe use of colonial fish-eating birds to monitor the presence ofdevelopmental toxicants in Great Lakes fish. J Great Lakes Res17:151-152 (1991).

55. Giesy JP, Ludwig JP, Tillitt DE. Deformities in birds of theGreat Lakes region. Environ Sci Technol 28:128-135 (1994).

56. Bowerman WW, Giesy JP, Best DA, Kramer VJ. A review offactors affecting productivity of bald eagles in the Great Lakesregion: implications for recovery. Environ Health Perspect 103:51-59 (1995).

57. Boss WR, Witschi E. The permanent effects of early stilbestrolinjections on the sex organs of the herring gull (Larusargentatus). J Exp Zool 94:181-209 (1943).

58. Fry DM, Toone CK. DDT-induced feminization of gull emby-ros. Science 213: 922-924 (1981).

59. Fry DM, Toone CK, Speich SM, Peard RJ. Sex ratio skew andbreeding patterns of gulls: demographic and toxicological con-siderations. Stud Avian Biol 10:26-43 (1987).

736 Environmental Health Perspectives - Vol 104, Supplement 4 * August 1996


60. Fox GA, Gilman AP, Peakall DB, Anderka FW. Aberrantbehavior of nesting gulls. J Wildl Manage 42:477-483 (1978).

61. Bishop CA, Brooks RJ, Carey JH, Ng P, Norstrom RJ, LeanDRS. The case for a cause-effect linkage between environmen-tal contamination and development in eggs of the commonsnapping turtle (Chelydra S. serpentina) from Ontario, Canada.J Toxicol Environ Health 33:521-547 (1991).

62. Mac MJ, Edsall CC. Environmental contaminants and thereproductive success of lake trout in the Great Lakes: an epi-demiological approach. J Toxicol Environ Health 33:375-394(1991).

63. Mac MJ, Schwartz TR, Edsall CC, Frank AM. Polychlorinatedbiphenyls in Great Lakes lake trout and their eggs: relations tosurvival and congener composition 1979-1988. J Great LakesRes 19:752-765 (1993).

64. Leatherland JF. Field observations on reproductive and devel-opmental dysfunction in introduced and native salmonids fromthe Great Lakes. J Great Lakes Res 19:737-751 (1993).

65. Guillette LJ Jr, Gross TS, Masson GR, Matter JM, PercivalHF, Woodward AR. Developmental abnormalities of thegonad and abnormal sex hormone concentrations in juvenilealligators from contaminated and control lakes in Florida.Environ Health Perspect 102: 680-688 (1994).

66. Guillette LJ Jr, Gross TS, Gross DA, Rooney AA, Percival HF.Gonadal steroidogenesis in vitro from juvenile alligatorsobtained from contaminated or control lakes. Environ HealthPerspect 103:31-36 (1995).

67. Wren CD. Cause-effect linkages between chemicals and popu-lations of mink (Mustela vison) and otter (Utra canadensis) inthe Great Lakes Basin. J Toxicol Environ Health 33:549-585(1991).

68. Facemire CF, Gross TS, Guillette LJ Jr. Reproductive impair-ment in the Florida panther: nature or nurture? Environ HealthPerspect 103:79-86 (1995).

69. Guo YL, Lai TJ, Ju SH, Chen YC, Hsu CC. Sexual develop-ments and biological findings in Yucheng children. DIOXIN'93: 13th International Symposium on Chlorinated Dioxinsand Related Compounds, 20-24 September, Vienna, Austria.Vienna:Austrian Federal Environmental Agency 14:235-238(1993).

70. Guzelian PS. Comparative toxicology of chlordecone (Kepone)in humans and experimental animals. Ann Rev PharmacolToxicol 22:89-113 (1982).

71. Rogan WJ, Gladen BC, McKinney JD. Polychlorinatedbiphenyls (PCBs) and dichlorodiphenyl dichloroethane (DDE)in human milk: effects on growth, morbidity and duration oflactation. AmJ Public Health 77:1294-1297 (1987).

72. Gladen BC, Rogan WJ. DDE and shortened duration of lacta-tion in a northern Mexican town. AmJ Public Health85:504-508 (1995).

73. Egeland GM, Sweeney MH, Fingerhut MA, Wille KK,Schnorr TM, Halperin WE. Total serum testosterone andgonadotropins in workers exposed to dioxin. Am J Epidemiol139(3):272-281 (1994).

74. Carlsen E, Giwereman A, Keiding N, Skakkebaek NE.Evidence for decreasing sperm quality of semen during the past50 years. Br Med J 304:609-613 (1992).

75. Auger J, Kunstman JM, Czyglik F, Jouannet P. Decline insemen quality among fertile men in Paris during the past 20years. N Engl J Med 332:281-285 (1995).

76. McArthur MLB, Fox GA, Peakall DB, Philogene BJR.Ecological significance of behavioral and hormonal abnormali-ties in breeding ring doves fed an organochlorine chemical mix-ture. Arch Environ Contam Toxicol 12:343-353 (1983).

77. Hoffman DJ. Embryotoxicity and teratogenicity of environ-mental contaminants to bird eggs. Rev Environ ContamToxicol 115:40-82 (1990).

78. Kihlstrom JE, Olsson M, Jensen S, Johansson A, Ahibom J,Bergman A. Effects of PCB and different fractions of PCB onthe reproduction of the mink (Mustela vison). Ambio 2:663(1992).

79. Colborn T, vom Saal FS, Soto AM. Developmental effects of

endocrine-disrupting chemicals in wildlife and humans.Environ Health Perspect 101:378-384 (1993).

80. Jansen HT, Cooke PS, Porcelli J, Liu T-C, Hansen LG.Estrogenic and antiestrogenic actions of PCBs in the female rat:in vitro and in vivo studies. Reprod Toxicol 7:237-248 (1993).

81. Jobling S, Sumpter JP. Detergent components in sewageeffluent are weakly oestrogenic to fish: an in vitro study usingrainbow trout (Oncorhynchus mykiss) hepatocytes. AquatToxicol 27:361-372 (1993).

82. Kelce WR, Monosson E, Gamcsik MP, Laws SC, Gray, LE.Environmental hormone disruptors: evidence that vinclozolindevelopmental toxicity is mediated by antiandrogenic metabo-lites. Toxicol Appl Pharmacol 126:276-285 (1994).

83. Patnode KA, Curtis LR. 2,2',4,4',5,5'- and 3,3',4,4',5,5'-Hexachlorobiphenyl alteration of uterine progesterone andestrogen receptors coincides with embryotoxicity in mink(Mustela vison). Toxicol Appl Pharmacol 127:9-18 (1994).

84. Bergeron JM, Crews D, McLachlan JA. PCBs as environmentalestrogens: turtle sex determination as a biomarker of environ-mental contamination. Environ Health Perspect 102:780-781(1994).

85. Guillette LJ Jr. Endocrine-disrupting environmental contami-nants and reproduction: lessons from the study of wildlife. In:Women's Health Today: Perspectives on Current Research andClinical Practice (Popkin DR, Peddle LJ, eds). NewYork:Parthenon Publ Group, 1994;201-207.

86. Baldwin WS, Milam DL, LeBlanc GA. Physiological and bio-chemical perturbations in Daphnia azna following exposure tothe mode environmental estrogen diethylstilbestrol. EnvironToxicol Chem 14:945-952 (1995).

87. Dodson SI, Hanazato T. Commentary on effects of anthro-pogenic and natural organic chemicals on development, swim-ming behavior, and reproduction of Daphnia, a key member ofaquatic ecosystems. Environ Health Perspect 103:7-11(1995).

88. Rolland R, Gilbertson M, Colborn T, eds. EnvironmentallyInduced Alterations in Development: A Focus on Wildlife.Environ Health Perspect 103(Suppl 4):3-100 (1995).

89. Kelce WR, Stone CR, Laws SL, Gray LE, Kemppaainen JA,Wilson EM. Persistent DDT metabolite p,p'-DDE is a potentandrogen receptor antagonist. Nature 375:581-585 (1995).

90. Krishnan AV, Stathis P, Permuth SF, Tokes L, Feldman D.Bisphenol-A: an estrogenic substance is released from polycar-bonate flasks during autoclaving. Endocrinology 132:2279-2286 (1993).

91. White R, Jobling S, Hoare SA, Sumpter JP, Parker MG.Environmentally persistent alkylphenolic compounds are estro-genic. Endrocrinology 135:175-182 (1994).

92. Murdoch MH, Hebert PDN. Mitochondrial DNA diversity ofbrown bullhead from contaminated and relatively pristine sites inthe Great Lakes. Environ Toxicol Chem 13:1281-1289 (1994).

93. Masuda YU, Kagawa R, Kuroki H, Taki I, Kusada M,Yamashita F, Hayashi M, Kuratsune M, Yoshimura T. Transferof polychlorinated biphenyls from mothers to fetuses andinfants. Fd Cosmet Toxicol 16:543-546 (1978).

94. Meyeroff WL. Hyperthyroidism and the ear: electro-physiological, morphological and chemical considerations.Laryngoscope 89:1-25 (1979).

95. Collins WT, Capen CC. Fine structural lesions and alterationsin thryroid glands of perinatal rats exposed in utero and by themilk to PCBs. Am J Pathol 99:125-142 (1980).

96. Uziel A, Gabrion J, Ohresser M, Legrand C. The effect ofhypothyroidism on the onset of cochlear potentials in develop-ing rats. Brain Resl82:172-175 (1980).

97. White DH, Mitchell CA, Hill EF. Parathion alters incubationbehavior of laughing gulls. Bull Environ Contam Toxicol31:93-97 (1983).

98. Overman SR, Kostas J, Wilson LR, Shain W, Bruysh B.Neurobehavioral amd somatic effects of perinatal PCB expo-sure in rats. Environ Res 44:56-70 (1987).

99. Tilson HA, Jacobsen JL, Rogan WJ. Polychlorinated biphenylsand the developing nervous system: cross species comparison.Neurotoxicol Teratol 12:239-248 (1991).


KAVLOCK ET AL.

100. Meserve LA, Murry EA, Landis JA. Influence of maternalingestion of Arochlor 1254 (PCB) or Firemaster BP-6 (PBB)on unstimulated and stimulated corticosterone levels in youngrats. Bull Environ Contam Toxicol 48:715-720 (1992).

101. Pluim HJ, Vijilder A, Olie K, Kopk JH, Vulsma T, van TijnDA, van der Slikke JW, Koppe JG. Effects of pre- and postna-tal exposure to chlorinated dioxins and furans on humanneonatal hormone concentrations. Environ Health Perspect102(6):504-508 (1993).

102. Porterfield SP, Hendrich CE. The role of thyroid hormones inprenatal and neonatalal neurological development-currentperspectives. Endocr Rev 14(1): 94-106 (1993).

103. Daly HB. Laboratory rat experiments show consumption ofLake Ontario salmon causes behavioral changes: support forwildlife and human research results. J Great Lakes Res19:784-788 (1993).

104. Jacobson JL, Jacobson SW. A 4-year followup study of childrenborn to consumers of Lake Michigan fish. J Great Lakes Res19:776-783 (1993).

105. Porterfield SP, Stein SA. Thyroid hormones and neurologicaldevelopment: update 1994. Endocr Rev 3(1):357-363 (1994).

106. Schantz SL, Moshtaghian J, Ness DK. Spatial learning deficitsin adult rats exposed to ortho-substituted PCB congeners dur-ing gestation and lactation. Fundam Appl Toxicol 26:117-126(1995).

107. Henshel DS, Martin JW, Norstrom R, Whitehead P, SteevesJD, Cheng KM. Morphometric abnormalities in brains of greatblue heron hatchlings exposed in the wild to PCDDs. EnvironHealth Perspect 103:61-66 (1995).

108. Holene E, Nafstad I, Skaare JU, Bernhoft A, Engen P,Sagvolden T. Behavioral effects of pre- and postnatal exposureto individual polychlorinated biphenyl congeners in rats.Environ Toxicol Chem 14:967-976 (1995).

109. Hontela A, Dumont P, Duclos D, Fortin R. Endocrine andmetabolic dysfunction in yellow perch, Percaflavescens, exposedto organic contaminants and heavy metals in the St. LawrenceRiver. Environ Toxicol Chem 14:725-731 (1995).

110. Goldey ED, Kehn LS, Lau, C, Rehnberg, GL, Crofton, KM.Developmental exposure to polychlorinated biphenyls(Arochlor 1254) causes hypothyroidism and hearing deficits inrats. Toxciol Appl Pharmacol 135:77-88 (1995).

111. Office of Technology Assessment. Neurotoxicity: Identifyingand Controlling Poisons of the Nervous System. OTA-BA-436. Washington:U.S. Government Printing Office, 1990

112. Noller KL, Blair PB, O'Brien PC, Melton L III, Offord JR,Kaufman RH, Colton T. Increased occurrence of autoimmunedisease among women exposed in utero to diethylstilbestrol.Fertil Steril 49(6):1080-1082 (1988).

113. Schuurs AHWM, Verheul HAM. Effects of gender and Sexsteroids on the immune response. J Steroid Biochem35:157-172 (1990).

114. Homo-Delarche F, Fitzpatrick F, Christeff N, Nunez EA, BachJF, Dardenne M. Sex steroids, gluocorticoids, stress andautoimmunity. J Steroid Biochem 40: 619-637 (1991).

115. Grossman CJ. Interactions between the gonadal steroids andthe immune system. Science 227(4684):257-261 (1985).

116. Davis D, Safe S. Immunosuppressive activities of polychlorinat-ed biphenyls in C57BL/6N mice: structure-activity relation-ships as Ah receptor agonists and partial antagonists.Toxicology 63:97-111 (1990).

117. McKinney JD, Chae K, Gupia BN, Moore JA, Goldstein JA.Toxicological assessment of hexachlorobiphenyl isomers and2,3,7,8-tetrachlorodibenzofuran in chicks. Toxicol ApplPharmacol 36:65-80 (1976).

118. Luster MI, Germolec DR, Rosenthal GJ. Immunotoxicology:review of current status. Ann Allergy 64:427-432 (1990).

119. Dean JH, Luster MI, Munson A, Kimoor I, eds.Immunotoxicology and Immunopharmacology. Second ed.Target Organ Toxicology Series. New York:Raven Press, 1994.

120. Loose LD, Pittman KA, Benitz KF, Silkworth JB.Polychorinated biphenyl and hexachlorobenzene inducedhumoral immunosuppression. J Reticuloendothel Soc 22:

253-271 (1977).121. Clark DA, Gauldie J, Szewczuk MR, Sweeney G. Enhanced

suppressor cell activity as a mechanism of immunosuppressionby 2,3,7,8-tetrachlorodibenzo-p-dioxin. Proc Soc Exp BiolMed 168:290-299 (1981).

122. Lundberg K, Gronvik K-O, Goldschmidt TJ, Klareskog L,Dencker L. 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD)alters intrathymic T-cell development in mice. Chem BiolInteract 74(1-2):179-193 (1990).

123. Kerkvliet NI, Baecher-Steppan L, Smith BB, Youngberg JA,Henderson MC, Guhler DR. Role of the Ah locus in suppres-sion of cytotoxic T lymphocyte activity by halogenated aromat-ic hydrocarbons (PCBs and TCDDs): structure-activity rela-tionships and effects in C57B/1/6 mice congenic at the Ahlocus. Fundam Appl Toxicol 14:532-541 (1990).

124. Kerkvliet NI, Steppan LB, Brauner JA, Deyo JA, HendersonMC, Tomar RS, Buhler DR. Influence of the Ah locus on thehumoral immunotoxicity of 2,3,7,8-tetrachlorodibenzo-p-diox-in: evidence for Ah-receptor-dependent and Ah-receptor-inde-pendent mechanisms of immunosuppression. Toxicol ApplPharmacol 110:26-36 (1990).

125. Anderson L, Nikolaids E, Brunstrom B, Bergman A, DenckerL. Effects of polychlorinated biphenyls with Ah receptoraffinity on lymphoid development in the thymus and bursa ofFabricius of chick embryos in ovo and in mouse thymus anla-gen in vitro. Toxicol Appl Pharmacol 107:183-188 (1991).

126. Morris DL, Snyder NK Gokani, V Blair, RE Holsapple MP.Enhanced suppression of humoral immunity in DBA/2 micefollowing subcronic exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Toxicol Appl Pharmacol 112:128-132(1992).

127. Arkoosh MR, Kaattari S. The effect of early aflatoxin Bi expo-sure on in vivo and in vitro antibody response in rainbow trout,Salmo gairdneri. J Fish Biol 31 (Suppl A): 19-22 (1987)

128. Arkoosh MR. Development of immunological memory in rain-bow trout (Oncorhynchus mykiss) and aflatoxin modulation ofthe response [PhD dissertation]. Oregon State University,Corvallis, OR, 1989.

129. Kaattari SL, Adkison M, Shapiro D, Arkoosh AR. Mechanismsof immunosuppression by aflatoxin B 1. In: Modulators of FishImmune Responses, Vol 1 (Stolen J, ed). Fair Haven, NJ:SOSPubl, 1994;151-156.

130. Aguilar, A, Raga, JA . The striped dolphin epizootic in theMediterranean Sea. Ambio 22: 524-528 (1983).

131. Svensson BG, Hallberg T, Nilsson A, Schutz A, Hagmar L.Parameters of immunological competence in subjects with highconsumption of fish contaminated with persistent organochlo-rine compounds. Ing Arch Occup Environ Health 65:351-358(1994).

132. Lahvis GP, Wells RS, Kuehl DW, Stewart JL, Rhinehart HL,Via CS. Decreased lymphocyte responses in free-ranging bot-tlenose dolphins (Tursiops truncatus) are associated withincreased concentrations of PCBs and DDT in peripheralblood. Environ Health Perspect 103: 67-72 (1995).

133 deSwart RL, Ross PS, Vedder LJ, Timmerman HH,Heisterkamp S, Van Loveren H, Vos JG, Reijnders PJH,Osterhaus ADME. Impairment of immune function in harborseals (Phoca vitulina) feeding on fish from polluted waters.Ambio 23: 155-159 (1994).

134. Ross PS, DeSwart RL, Reijnders PJH, Van Loveren H, Vos JG,Osterhaus ADME. Contaminant-related suppression ofdelayed-type hypersensitivity and antibody responses in harborseals fed herring from the Baltic Sea. Environ Health Perspect103:162-167 (1995).

135. Grasman KA. Immunological and hematological biomarkersfrom contaminants in fish-eating birds of the Great Lakes [PhDdissertation]. Virginia Polytechnic University, Blacksburg, VA,1995.

136. Luster MI, Munson AE, Thomas PT, Holsapple, MP, Fenters,JD, White, KL, Lauer, LD, Germolec, DR, Rosenthal, GJ,Dean, JH. Development of a testing battery to assess chemical-induced immunotoxicity: National Toxicology Program's

738 Environmental Health Perspectives * Vol 104, Supplement 4 - August 1996


guidelines for immunotoxicity evaluation in mice. FundamApplied Toxicol 10:2-19 (1988).

137. Schrank CS, Cook ME, Hansen WR. Immune responses ofmallard ducks treated with immunosuppressive agents: anti-body response to erythrocytes and in vivo response to phyto-hemagglutinin. J Wildl Dis 26:307-315 (1990).

138. Zelikoff JT, Enane NA, Bowser D, Squibb K, Frenkel K.Development of fish peritoneal macrophages as a model forhigher vertebrates in immunotoxicological studies. FundamAppl Toxicol 16:576-589 (1991).

139. Bowser D, Frenkel K, Zelikoff JT. Effects of in vitro nickelexposure on macrophage-mediated immunity in rainbow trout.Bull Environ Contam Toxicol 52:367-373 (1994).

140. Dean JH, Cornacoff GF Rosenthal GJ, Luster MI. Immunesystem: evaluation of injury. In: Principles and Methhods ofToxicology (Hayes AW, ed). New York:Raven Press,1994; 1065-1090.

141. Grasman KA, Scanlon PF. Effects of acute lead ingestion anddiet on antibody and T-cell-mediated immunity in Japanesequail. Arch Environ Contam Toxicol 28: 161-167 (1995).

142. Zelikoff JT, Bowser, D, Squibb, KS, Frenkel, K.Immunotoxicity of low cadmium exposure in fish: alternativeanimal models for immunotoxicological studies. J ToxicolEnviron Health 45:235-248 (1995).

143. Buist AS, Vollmer WM. Reflections of the rise in asthma mor-bidity and mortality. JAMA 264:1719-1720 (1990).

144. Weiss KB, Wagener DK. Changing patterns of asthma mortali-ty. Identifying target populations at high risk. JAMA264:1688-1692 (1990).

145. Gergen PJ, Weiss KB. Changing pattern of asthma hospitaliza-tion among children: 1979-1987. JAMA 264:1688-1692(1990).

146. ZelikoffJT, Smialowicz R, Bigazzim PE, Goyer RA, LawrenceDA, Maibach Gardner D. Immunomodulation by metals.Fundam Appl Toxicol 22:1-8 (1994).

147. Pelletier L, Castedo M, Bellon B, Druet P. Mercury andAutoimmunity. In: Immunotoxicology and Immuno-pharmacology (Dean, J, Luster M, Munsen AE Kimber I, eds).New York:Raven Press, 1994;539.

148. Gonzalez R, Matsiola P, Torchy C, de Kinkelin P, Avrameas S.Natural anti-TNP antibodies from rainbow trout interfer withviral infection in vitro. Res Immumol 140:675-684 (1989).

149. Michel C, Gonzalez R, Bonjour E, Avrameas S. A concurrentincreasing of natural antibodies and enhancement to furunculo-sis in rainbow trout. Ann Rech Vet 21:211-218 (1990).

150. Marcholonis JJ, Hohman VS, Thomas C, Schluter SF.Antibody production in sharks and humans: a role for naturalantibodies. Dev Comp Immunol 17:41-53 (1993)

151. Ankley GT, Johnson RD, Toth G, Folmar LC, Defenbeck NE,Bradbury SP. Development of a research strategy for assessingthe ecological risk of endocrine disruptors. Rev Toxicol,B-Environ Toxicol (in press).

152. Gray LE, Ostby J, Ferrell J, Rehnberg G, Linder R, Cooper R,Goldman J, Slott V, Laskey J. A Jose-response analysis ofmethoxychlor-induced alterations of reproductive developmentand function in the rat. Fundam Appl Toxicol 12:92-108(1989).

153. U. S. Environmental Protection Agency. Testing Guidelines forDevelopmental and Reproductive Toxicity. Docket ControlNo OPP-00383. Fed Reg 59:42272 (1994).

154. Laskey JW, Berman E. Steroidogenic assessment using ovaryculture in cycling rats: effects of bis(2-diethylhexyl) phthlate onovarian steroid production. Reprod Toxicol 7:25-33 (1993).

155. Zenick H, Clegg ED, Perreault SD, Klinefelter GR, Gray LE.Assessment of male reproductive toxicity: a risk assessmentapproach. In: Principles and Methods in Toxicology, 3rd ed.(Hayes AW, ed). New York:Raven Press, 1994;937-988.

156. Bern HA. The fragile fetus. In: Chemically-induced Alterationsin Sexual and Functional Development: The Human-WildlifeConnection (Colborn T, Clement C, eds). Princeton,NJ:Princeton Scientific Publishing, 1992.

157. Sheehan DM, Branham WS, Medlock KL. An estrogen devel-

opmental potency and toxicity assay for environmental estro-gens. Teratology 51:196-197 (1995).

158. Sumpter JP, Jobling S. Vitellogenesis as a biomarker for oestro-genic contamination of the aquatic environment. EnvironHealth Perspect 103(Suppl 7):173-178 (1995).

159. Soto AM, Chung KL, Sonneschein C. The pesticides endosul-fan, toxaphene, and dieldrin have estrogenic effects on humanestrogen-sensitive cells. Environ Health Perspect 102:380-383(1994).

160. Kemppainen JA, Lane MV, Sar M, Wilson EM. Androgenreceptor phosphorylation, turnover, nuclear transport, andtranscriptional activation: specificity for steroids and antihor-mones. J Biol Chem 267:969-974 (1992).

161. Yarbrough WG, Quarmby VE, Simental JA, Joseph DR, SarM, Lubahn DB, Olsen KL, French FS, Wilson EM. A singlebase mutation in the androgen receptor gene causes androgeninsensitivity in the testicular feminized rat. J Biol Chem265:8893-8900 (1990).

162. Klein KO, Baron J, Colli MJ, McDonnell DP, Cutler GB.Estrogen levels in childhood determined by an ultrasensitiverecombinant cell bioassay. J Clin Invest 94:2475-2480 (1994).

163. Zhou ZX, Sar M, Simental JA, Lane MV, Wilson EM. A lig-and-dependent bipartite nuclear targeting signal in the humanandrogen receptor: requirement for the DNA-binding domainand modulation by NH2-terminal and carboxyl-terminalsequences. J Biol Chem 269:13115-13123 (1994).

164. Laidlaw IJ Clarke RB, Howel A, Owen AWMC, Potten CS,Anderson E. The proliferation of normal breast tissue implant-ed into athymic nude mice is stimulated by estrogen but notprogesterone. Endocrinology 136:164-171 (1995).

165. Laskey JW, Berman E, Ferrell JM. The use of cultured ovarianfragments to assess toxicant alterations in steroidogenesis in theSprague-Dawley rat. Reprod Toxicol 9:131-141 (1995).

166. McKinney JD, Waller CL. Polychlorinated biphenyls as hor-monally active structural analogs. Environ Health Perspect102:290-297 (1994).

167. Waller Cl, McKinney JD. Comparative molecular field analysisof polyhalogenated dibenzo-p-dioxins, dibenzofurans, andbiphenyls. J Med Chem 36:3660-3666 (1992).

168. Waller CL, McKinney JD. Examination of the estrogen recep-tor binding affinities of polychlorinated hydroxybiphenylsusing three-dimensional quantitative structure-activity relation-sips. Environ Health Perspect 103:702-707 (1994).

169. Means GD, Kilgore MW, Mahendroo MS, Mendelson CR,Simpson ER. Tissue-specific promoters regulate aromatasecytochrome P450 gene expression in human ovary and fetal tis-sues. Mol Endocrinol 5(12):2005-2013 (1991).

170. Webb P, Lopez GN, Uht RM, Kushner PJ. Tamoxifen activa-tion of the estrogen receptor/AP- 1 pathway: potential origin forthe cell-specific estrogen-like effects of antiestrogens. MolEndocrinol 9:443-456 (1995).

171. Hesselberg RJ, Hickey JP, Nortrup DA, Willford WA.Contaminant residues in the bloater (Coregonus hoyt) of LakeMichigan, 1969-1986. J Great Lakes Res 16:121-129 (1990).

172. Hesselberg RJ, Gannon JE. Comtaminant trends in the GreatLakes. In: Our Living Resources: A Report to the Nation on theDistribution, Abundance, and Health of US Plants, Animals andEcosystems (LaRoe ET, Farris GS, Puckett CE, Doran PD, MacMJ, eds). Washington:U.S. Dept Interior, 1995;242-244.

173. Schmitt CJ, Zajicek JL, Peterman PL. National ContaminantBiomonitoring Program residues of organochlorine chemicalsin freshwater fishes of the United States, 1976-1984. ArchEnviron Contam Toxicol 19:748-782 (1990).

174. Schmitt CJ, Bunck CM. Persistent environmental comtami-nants in fish and wildlife. In: Our Living Resouces: A Report tothe Nation on the Distribution, Abundance, and Health of USPlants, Animals, and Ecosystems (LaRoe ET, Faris GS, PuckettCE, Doran PD, Mac MJ, eds). Washington:NationalBiological Service, 1995;413-416.

175. Stow CA, Carpenter SR, Eby LA, Amrhein JF, Hesselberg RJ.Evidence that PCBs are approaching stable concentrations inLake Michigan fishes. Ecol Appl 5:248-260 (1995).

Environmental Health Perspectives - Vol 104, Supplement 4 * August 1996 739

KAVLOCK ET AL.

176. Hebert CE, Norstrom RJ, Simon M, Braune BM, WeselohDV, Macdonald CR. Temporal trends and sources of PCDDsand PCDFs in the Great Lakes: Herring gull egg monitoring,1981-1991. Environ Sci Technol 281:1268 (1994).

177. Borgmann U, Whittle DM. Contaminant concentration trendsin Lake Ontario lake trout (Salvelinus namaycush). J GreatLakes Res 17:368-381 (1991).

178. Blomkvist G, Roos A, Jensen S, Bignert A, Olsson M.Concentrations of DDT and PCB in seals from Swedish andScottish waters. Ambio 2 1:539-. (1992)

179. Council on Environmental Quality. Environmental Quality,23rd Annual Report of the Council on Environmental Qualtiy.Washington:U.S .Government Printing Office, 1992;45 1.

180. Whittle DM, Baumann PC. The status of selected organics inthe Laurentian Great Lakes: an overview of DDT, PCBs, diox-ins, furans, and aromatic hydrocarbons. Aquat Toxicol11:241-257 (1988).

181. Ezatti TM, Massey JT, Waksburg J, Chu A, Maurer KR.Sample Designn: Third National Health and Nutrition Survery.Vital Health Statistics. Publ No PHS 92-1387.Washington:U.S. Dept. Health and Human Services, 1992.

182. Sexton K, Callahan MS, Bryan EF, Saint CG, Wood WP.Informed decisions about protecting and promoting publichealth. Rationale for a National Human Exposure AssessmentSurvey. J Expos Anal Environ Epidemiol 5(3):233-236 (1996).

183. Schmitt CJ, Brumbaugh WG. National ContaminantBiomonitoring Program: concentrations of arsenic, cadmium,copper, lead, mercury, selenium, and zinc in freshwater fishesof the United States 1976-1984. Arch Environ ContamToxicol 19:731-747 (1990).

184. Turgeon DD, Robertson A. Contaminants in coastal fish andmollusks. In: Our Living Resources: a Report to the Nation onthe Distribution, Abundance, and Health of US Plants,

Animals and Ecosystems (LaRoe ET, Farris GS, Puckett CE,Doran PD, Mac MJ, eds). Washington:U.S. Dept. Interior,1995;408-412.

185. Burkhard LP, Ankley GT. Identifying toxicants: NETAC's tox-icity-based approach. Environ Sci Technol 23:1438-1443(1989).

186. Scott GS et al. Correlating field and laboratory toxicity testingwith ecotoxicological biomonitoring in assessing the impact ofagriculutural NPS runoff. Final Report to USEPA/GBERL,Sabine Island, Gulf Breeze, FL 1992.

187. Purdom CE, Hardiman PA, Bye VJ, Eno NC, Tyler CR,Sumpter JP. Estrogenic effects of effluents from sewage treat-ment works. Chem Ecol 8:275-285 (1994).

188. Melancon MJ. Bioindicators used in aquatic and terrestrialmonitoring. In: Handbook of Ecotoxicology ( Hoffman DJ,Rattner BA, Burton GA Jr, Cairns J, eds). Boca Raton,FL:Lewis Publishers 1995;220-239.

189. Sharpe RM. Declining sperm counts in men: is there anendocrine cause? J Endocrinol 136:357-360 (1993).

190. Sharpe RM, Skakkebaek NE. Are oestrogens involved in fallingsperm counts and disorders of the male reproductive tract?Lancet 341:1392-1395 (1993).

191. Davis EP, Bortone SA. Effects of kraft pulpmill effluent on thesexuality of fishes: an environmental early warning? In:Chemically-induced Alterations in Sexual and FunctionalDevelopment: The Wildlife-Human Connection, Vol 21(Colborn T, Clement C, eds). Princeton, NJ:PrincetonScientific Publishing, 1992;1 13-127.

192. Hill AB. The environment and disease: association or caus-ation. Proc Royal Soc Med 58:295 (1965).

193. Fox GA. Practical causal inference for ecoepidemiologists.J Toxicol Environ Health 33:359-373 (1991).

740 Environmental Health Perspectives - Vol 104, Supplement 4 * August 1996

22 - reserch needs for the risk assessment of health and environmental effects of.. (epa)

Documents

toxic equivalency factor

great lakes region

research triangle park

lawrence beluga whales

evaluating health risks

guillette lj jr

historical trend data

phytohemagglutinin skin test