RESEARCH Review Ecological Risk Assessment of Endocrine Disruptors Thomas H. Hutchinson,1 Rick Brown,2 Kristin E. Brugger,3 Pamela M. Campbell,4 Martin Holt,5 Reinhard Länge,6 Peter McCahon,7 Lisa J. Tattersfield,8 and Roger van Egmond 9 1AstraZeneca, Brixham Environmental Laboratory, Brixham, Devon, United Kingdom; 2Exxon Biomedical Sciences Inc., East Millstone, New Jersey, USA; 3Du Pont, Wilmington, Delaware, USA; 4Procter & Gamble, Toronto, Ontario, Canada; 5ECETOC, Van Nieuwenhuyse, Brussels, Belgium; 6Schering AG, Berlin, Germany; 7Aventis, Research Triangle Park, North Carolina, USA; 8ZENECA Agrochemicals, Jealott’s Hill Research Station, Bracknell, United Kingdom; 9Unilever Research, Port Sunlight Laboratory, Bebington, United Kingdom

characterize the ecological risk of selected The European Centre for Ecotoxicology and Toxicology of Chemicals proposes a tiered approach endocrine-active substances. For example, for the ecological risk assessment of endocrine disruptors, integrating exposure and hazard recent progress toward this goal is illustrated (effects) characterization. Exposure assessment for endocrine disruptors should direct specific tests by tributyltin (9). In general, however, further for wildlife species, placing hazard data into a risk assessment context. Supplementing the suite of work is needed to evaluate potential mammalian screens now under Organization for Economic Cooperation and Development endocrine disruptors within the established (OECD) validation, high priority should be given to developing a fish screening assay for detect- ecological risk assessment concept (2,10,11) ing endocrine activity in oviparous species. Taking into account both exposure characterization (Figure 1). Although much will be learned and alerts from endocrine screening, higher tier tests are also a priority for defining adverse through the on-going application of the eco- effects. We propose that in vivo mammalian and fish assays provide a comprehensive screening logical risk assessment paradigm to both nat- battery for diverse hormonal functions (including androgen, estrogen, and thyroid hormone), ural and synthetic endocrine disruptors, we whereas Amphibia should be considered at higher tiers if there are exposure concerns. Higher tier support the views of Kendall et al. (4) in that endocrine-disruptor testing should include fish development and fish reproduction tests, whereas “There is no need to develop a new frame- a full life-cycle test could be subsequently used to refine aquatic risk assessments when necessary. work for ecological risk assessment of For avian risk assessment, the new OECD Japanese quail reproduction test guideline provides a endocrine disrupters.” What is needed, how- valuable basis for developing a test to detecting endocrine-mediated reproductive effects; this ever, is a scientifically and ethically justifiable species could be used, where necessary, for an avian life-cycle test. For aquatic and terrestrial approach to prioritizing endocrine-disruptor invertebrates, data from existing developmental and reproductive tests remain of high value for screening and testing that effectively protects ecological risk assessment. High priority should be given to research into comparative endocrine the diversity of wildlife. physiology of invertebrates to support data extrapolation to this diverse fauna. Key words: ecologi- The established risk assessment frame- cal risk assessment, endocrine disruptor, environment, hormone mimic, screening, testing. work provides a robust tool with which to Environ Health Perspect 108:1007–1014 (2000). [Online 4 October 2000] evaluate the impacts of natural and synthetic http://ehpnet1.niehs.nih.gov/docs/2000/108p1007-1014hutchinson/abstract.html toxicants, endocrine disruptors, and other stressors on ecosystems. This framework bal- ances exposure characterization versus effects Given many reports of contaminant-associ- disruptors and supports the establishment of characterization, taking into account the ated reproductive and developmental new wildlife screening and testing protocols. need for test validation, data acquisition, and impacts in wildlife, often considered to be Our strategy for ecotoxicity screening and field monitoring. caused by endocrine disruption, there is now testing is discussed versus the proposed EDSP Exposure assessment for potential a major global effort to develop ecotoxicity (2), based on the earlier report from the endocrine disruptors is important in directing test guidelines for the hazard assessment of EDSTAC (1). Specifically, our critical review specific-effects testing, such that risk assess- endocrine disruptors. The Endocrine of the EDSTAC considers both the scientific ment and risk management can proceed. Disruptor Screening and Testing Advisory rationale and ethical use of animals for eco- Exposure assessment may be defined as the Committee (EDSTAC) (1) and the U.S. toxicity hazard assessment (5). Throughout contact between the bioavailable fraction of Environmental Protection Agency’s (U.S. this exercise, we support the internationally the compound of interest and the organisms EPA’s) proposed Endocrine Disruptor agreed definition from the 1996 Weybridge of concern. A tiered approach to exposure Screening Program (EDSP) (2) have made workshop, whereby an endocrine disruptor is assessment whereby conservative assumptions key initiatives for hazard assessment per se. “an exogenous substance which causes adverse Critically, however, the strategy for ecologi- effects in an organism, or its progeny, subse- Address correspondence to T. Hutchinson, cal effects characterization of endocrine dis- quent to changes in the endocrine system.” AstraZeneca, Brixham Environmental Laboratory, ruptors also needs to be integrated into the (6). In vitro test systems are not addressed in Freshwater Quarry, Brixham, Devon TQ5 8BA, UK. exposure characterization component of a our present discussion because it has been Telephone: 44-1803-882882. Fax: 44-1803-882974. risk-based laboratory and field approach widely agreed at several international work- E-mail: [email protected] (3,4). shops that endocrine-disruptor assessments in The members of the European Centre for Ecotoxicology and Toxicology of Chemicals Wildlife The European Centre for Ecotoxicology wildlife should primarily focus on in vivo Working Group are R. Länge, R. Brown, K. Brugger, and Toxicology of Chemicals (ECETOC) studies (7,8). P.M. Campbell, S. Einarson, M. Holt, T.H. established the Environmental Oestrogens Hutchinson, P. McCahon, O. Oppen-Berntsen, L.J. Task Force to identify the best ways to eval- The Ecological Risk Tattersfield, R. van Egmond, and S. Zok. uate potential endocrine-disrupting sub- Assessment Context We thank our colleagues in the Chemical stances for human and ecological risk Although evaluation of endocrine disruptors Manufacturers Association and the European Crop Protection Association for their helpful comments assessment. From these efforts, our ECE- is a relatively new area, both field and labora- on this paper. TOC working group argues for an ecologi- tory studies have been conducted to define Received 13 December 1999; accepted 16 June cal risk assessment framework for endocrine ecological effects, determine sources, and 2000.

Environmental Health Perspectives • VOLUME 108 | NUMBER 11 | November 2000 1007 Review • Hutchinson et al. in the estimate are progressively refined is In comparing the degree of toxicity with Key elements of our approach are, first, likewise appropriate for endocrine disruptors the level of exposure, the risk of the com- to focus testing needs through a greater con- as well as compounds that may be active via pound may then be characterized and any sideration of exposure characterization data other mechanisms. necessary risk management can be conducted when deciding which types of chronic tests Potential for exposure should drive the (Figure 1). Ultimately, potential impacts (e.g., avian, fish, or invertebrate) may be selection of appropriate test organisms in from synthetic substances should be consid- required at the higher tiers (15). Second, our hazard assessment. For example, where a ered in the context of natural stressors to scheme offers a pragmatic alternative to the pesticide is sprayed directly onto crops, it is help define what is an acceptable ecological technical and ethical concerns raised by the reasonable to expect potential exposure to impact. The concept of acceptability under- EDSTAC recommendation to move to aquatic organisms and birds via spray drift. pins the regulatory programs for pesticides extensive multispecies testing directly after The expected environmental concentrations in many countries. For example, the U.S. the Tier 1 screening, without first recogniz- should then be compared with the toxic EPA characterizes “unacceptable” as “wide- ing the value of partial life-cycle (PLC) tests, concentration to aquatic organisms and spread and repeated mortality in the face of which use fewer animals. The following con- birds to determine the potential ecological minor economic benefits to society” cepts have proved useful in consideration risk. For natural or synthetic substances dis- (11,13). Finally, it is essential that new test of new wildlife protocols for endocrine-disrup- charged via wastewater into rivers, aquatic methods should provide data that can be tor screening and testing: a) conceptual organisms are expected to be exposed if the related directly to the field monitoring protocol, which refers to an idea for a novel substance is not degraded during wastewater method. The successful field and laboratory protocol that has not yet been actually con- treatment. In addition, the bioaccumulation measurement of vitellogenin (VTG) induc- ducted and reported in the peer-reviewed sci- and biomagnification potential of the sub- tion in fish illustrates the biological linkage entific literature; b) developmental protocol, a stance should be assessed to determine if concept (14). protocol that has been published in the peer- fish-eating birds and mammals might also reviewed scientific literature and forms the be at risk. Overview of Conceptual basis for further research (e.g., using a range of Once potential for exposure is deter- Testing Framework reference endocrine-disruptor substances) mined, suitable effects tests should then be Although we fully support the principle of toward the development and prevalidation of a selected and the hazard assessed. The types of EDSTAC’s tiered screening and testing pro- reliable method; and c) validated protocol, the tests used should address the sensitivity of the gram, we suggest that this can be enhanced, as appropriate term for a protocol that has been faunal populations in the context of diverse shown in Figure 3, to provide more ecologi- demonstrated to be reliable and reproducible as exposure scenarios (Figure 2), adapted from cally relevant and responsible use of limited a result of appropriate intralaboratory and Solomon (12). animals for testing. interlaboratory comparisons. Protocols need to achieve the validated status before they can be reliably employed for regulatory purposes, for Ecological risk assessment example, within the Organization for

Planning Problem formulation Economic Cooperation and Development (Risk assessor/risk ( manager/interested OECD) test guideline program. parties dialogue)

A

A Exposure concentration n Population sensitivity a Characterization Characterization of As necessary l of exposure ecological effects acquire data, y iterate process, s monitor results i s B

Risk characterization C

Communicating results to the risk manager Time Figure 2. Exposure versus population sensitivity scenarios for persistent or nonpersistent endocrine disruptors, adapted from Solomon (12). (A) Exposure and sensitivity periods overlap with high probability of population effect. (B) The expo- Risk management and communicating results to sure period does not overlap the sensitivity period, interested parties so there is a low probability of population effect. (C) The exposure period overlaps the early devel- opmental period of population sensitivity, so there Figure 1. The U.S. EPA framework for ecological risk assessment (10). is an intermediate probability of population effect.

1008 VOLUME 108 | NUMBER 11 | November 2000 • Environmental Health Perspectives Review • Ecological risk assessment

Screening for Interactions amphibian habitats, then a suitably validated activity during amphibian metamorphosis with Endocrine Systems frog larval metamorphosis test could be use- (enhancement of tail resorption) (22–24). In accordance with EDSTAC, we agree on fully deployed at a higher tier (Figure 3). Also, studies in populations of Great Lakes the need for in vivo screening to identify the In many vertebrates, a number of physio- salmon have also demonstrated that fish can ability of a compound to interact with the logic and morphologic processes are influ- be affected by thyroid-disrupting chemicals endocrine system, primarily for androgen, enced by the thyroid (20). Mechanisms of (25,26). These observations in wild fish estrogen, or thyroid hormone activities. thyroid action may vary between species and suggest that there is an important opportuni- Importantly, the purpose of Tier 1 screening among tissues, with alpha- and beta-variant ty to also detect thyroid disruption in fish (Figure 3) should be to provide alerts to nuclear receptors having being identified in partial and full life-cycle tests. In conclusion, endocrine-active substances, whereas higher- mammals (21), whereas frog thyroid hormone potential effects on the thyroid hormone sys- tier tests should provide adverse effects data receptors (THRs) are highly homologous to tem should in principle be detected by for risk assessment application. Previous inter- THRs of other vertebrates. In Amphibia dur- rodent mammalian Tier 1 screening. The national workshops have agreed that screen- ing normal development, thyroid hormone decision to evaluate chemical effects in the ing with laboratory mammals is sufficiently levels rise, resulting in growth and differenti- frog metamorphosis assay would likely be comprehensive to cover wild mammalian ation of the limbs, with tail resorption being best based on a positive result from the mam- species; however, because of their distinct the final gross morphologic change within malian thyroid screen, and only if exposure reproductive physiology, the screening battery metamorphosis. Metamorphosis is initiated characterization predicts significant potential should include oviparous species (7,8). It is by the binding of thyroid hormone to exposure of amphibian habitats. The devel- also essential to undertake an efficient num- α-THR, whereas β-THR is presumed to play opment of a peer-reviewed database on the ber and range of Tier 1 screens in order to a later role in developmental programs such sensitivity of fish and amphibians to thyroid develop a mechanistic rationale that will allow as tail resorption. Notably, steroids such as disruptors is an important research priority. differentiation between substances which corticoids can modulate thyroid hormone directly impact the endocrine system (primary endocrine effects) versus those substances Set priorities for Tier 1 screening which primarily affect other target organs Exposure-related information before causing secondary endocrine effects Effects-related information for mammals and wildlife (16). As argued by Purchase (17), there is a need for specific screening assays that use a minimum number of test animals, taking into account the environment compartment of concern for the risk assessment. The candi- Tier 1 screening: wildlife Tier 1 screening: mammals date screening assays included in the oviparous species - androgen and estrogen activity Androgen, estrogen and EDSTAC proposals are discussed in the sub- in juvenile fish screen (new method) thyroid activity sequent section. Frog metamorphosis screening assay. EDSTAC’s main objective of the inclusion Potential exposure Go to risk assessment of the frog (Xenopus laevis) metamorphosis No assay as a Tier 1 screen is to detect thyroid Yes agonists and antagonists, rather than to rep- resent amphibians as a taxonomic group (1). Tier 2 testing: wildlife EDSTAC outlines a protocol for Amphibia Amphibian metamorphosis test (new method)a in which the larvae (tadpoles) of Xenopus lae- Avian reproduction 9 (based on 6-week Japanese quail study) vis are exposed to the test substance for 14 Fish development test and fish reproduction test (new models) Invertebrate reproduction test(s) (EDIETA)b days, with the tail resorption rate as the main end point (18). Data are urgently required for a range of thyroid active and negative Potential control substances to evaluate the specificity exposure Go to risk assessment No and sensitivity of the frog metamorphosis Yes assay before international validation. Failure to address these needs, especially assay speci- Tier 3 testing: wildlife ficity, will have serious implications for the Amphibian extended development test (new method)c numbers of animals tested to detect thyroid Avian reproduction (consider longer term study with Japanese quail)d disruption (17,19). Fish full life-cycle (based on existing U.S. EPA protocol)e More broadly, thyroid hormones are crit- ical to the normal processes of development and reproduction in many animals. The Ecological risk assessment amphibian assay described by EDSTAC does Consider exposure data from box 1 to refine exposure assessment, not define the necessary experimental design together with effects data from wildlife screening and testing to specifically detect thyroid disruption. We therefore recommend that thyroid hormone Figure 3. ECETOC strategy for the hazard and ecological risk assessment of endocrine disruptors. In all disruption may be detected in Tier 1 mam- cases, protocols used for both screening and higher tier testing should have gone through international malian screens. If (anti-)thyroidal activity was validation before use. aOnly required if mammalian screens show evidence of thyroid disruption. bAddressed at the SETAC-OECD invertebrate detected in mammals and if exposure charac- workshop (66). cOnly required if the substance is active in the larval amphibian metamorphosis test at Tier 2. dOnly terization suggested significant exposure in required if positive in avian reproduction test at Tier 2. eOnly required if the substance is active in fish PLC tests at Tier 2.

Environmental Health Perspectives • VOLUME 108 | NUMBER 11 | November 2000 1009 Review • Hutchinson et al.

Fish Screening Assays screening assay has many advantages and been extensive use of avian reproductive tox- Fish gonadal recrudescence assay. EDSTAC should be seriously considered as a practical icity testing in several countries and regions originally recommended the concept of the alternative to the conceptual fathead minnow (37–39). More recently, the OECD is taking fish gonadal recrudescence assay (GRA), GRA. Also, using juvenile fish avoids the a leading role in promoting international whereby breeding fathead minnows potential problem associated with adult fish harmonization of avian testing, including (Pimephales promelas) of both sexes would first reproductive and territorial behavior, for reproductive toxicity test guidelines and the be transferred to “winter phase” (short day example, where the presence of sexually dom- new question of the endocrine-disruption in length and low temperature) conditions to inant males (or females) may lead to behav- birds (40). This is warranted given the dis- inhibit spawning. Once fish were synchro- ioral and pheromone modulation of other tinct reproductive endocrinology of birds, nized in this nonbreeding phase, fish would fish in the laboratory population (30). Once whereby reproductive dysfunction may be be transferred to a “summer phase” (increas- endocrine baselines have been established for expressed in terms of impaired egg laying or ing day length and temperature regime) to a given fish model, it is possible to use sensi- abnormal hormonal imprinting of males and stimulate gonadal maturation and develop- tive and specific end points (e.g., VTG females. Importantly, the “default sex” for ment of secondary sexual characteristics. induction) to screen for endocrine activity. many bird species is the male, whereas in Proposed end points include assessment of Using this approach, a number of mammals the “default sex” is female (14). gamete quality, fecundity, gross pathology of research groups have shown that juvenile fish Avian sex ratios may be more plastic, howev- gonads, measurement of relative gonad are very sensitive to estrogens and antiestro- er, than was traditionally thought, implying weight (gonadal somatic index), and measure- gens (32–34). Current research sponsored by that sex determination is affected by envi- ment of vitellogenin levels (a widely used bio- the European chemical industry is now ronmental factors (41,42). marker of estrogen exposure in fish) (8,27). extending the evaluation of this juvenile fish The EDSTAC proposal appears to sug- The disadvantages of the fish GRA assay for (anti) androgens and aromatase gest that for chemicals which are positive in include the absence of any published data on inhibitors (35). Adapted from the OECD the Tier 1 screens, avian reproduction testing gonadal recrudescence in OECD test species; test guideline 204 (36), the juvenile fish assay is necessary at Tier 2, although the triggers for unlike other approaches that seek to enhance has significant potential for practical use, a requiring such an avian study are not defined. existing regulatory test guidelines for testing potential that is currently being critically Specifically, EDSTAC recommends conduct- endocrine disruptors (28), the GRA has no evaluated as part of the OECD prevalidation ing reproduction tests with two avian species, clear link to an existing regulatory protocol, exercise. northern bobwhite quail (Colinus virginianus) and the practical use of end points such as and mallard duck (Anas platyrhynchos), using gamete quality and fecundity in such a screen- Higher Tier Tests for Hazard existing protocols with possible modifications ing assay incurs a high degree of interindivid- Assessment to enhance the ability to detect endocrine- ual variability because cyprinid fish have Higher tier tests should serve to identify related effects. The existing avian reproduc- widely variable fecundity between breeding effects of concern (e.g., development and tion protocols are well known, and it is pairs, even within control populations. Also, reproduction) that are a consequence of estimated that they have been used approxi- considering extending the international use of endocrine disruption, supporting the effects mately 300–500 times (43,44). These meth- this approach beyond the fathead minnow, characterization within ecological risk assess- ods are standardized and regarded as other OECD fish species (e.g., medaka and ment. Working within the ecological effects adequately reproducible; hence, protocols for zebrafish) have different breeding require- and exposure characterization phases of the both northern bobwhite quail and mallard ments. For example, some zebrafish popula- ecological risk assessment guideline (11), if duck have de facto acceptance by the scientif- tions have been reported to display a normal significant exposure is predicted for a sub- ic community as being valid. The EDSTAC juvenile hermaphrodite stage (29), making stance that is active in the Tier 1 mammalian recommendations for suggested modifications baselines potentially uncertain. Although the or fish screens, then higher tier testing for to the current one-generation avian reproduc- sexes in fathead minnows are usually very dis- potential adverse effects (endocrine disrup- tion protocol include extending the current tinct, in other fish species secondary sexual tion) is warranted. At this stage, the key study design to two generations, measuring characteristics in males develop differently as terms “adverse effects” and “organisms and circulating sex steroid and thyroid hormones, dominant (territorial) males appear during their progeny” become critical. Although this and monitoring a suite of morphologic and group culturing, whereas others show less is not apparent within EDSTAC (1), such functional parameters in offspring. obvious secondary sexual characteristics (30). testing should be targeted at the specific Higher tier testing with Japanese quail. In zebrafish, for example, the sex of the fish is ecosystem and wildlife populations of concern The use of Japanese quail (Coturnix coturnix difficult to determine when the fish is not in a (e.g., aquatic animals or seed-eating birds) japonica) promises to be a cost-effective alter- reproductive stage. On the other hand, using data on exposure characterization, rather native for endocrine-disruptor testing, and zebrafish can produce eggs even at suboptimal than unnecessary higher tier testing on all the draft revised OECD test guideline 206 is conditions; therefore, it is difficult to main- wildlife groups due to endocrine-disruptor currently being adapted for this species (45). tain this species in a nonreproductive stage concerns per se. In contrast, if the mammalian In general, reproductive parameters such as (31). Overall, there is little prospect that a and fish Tier 1 screens are all negative, then no egg production and egg quality, fertility, small fish GRA could be robust or practical, further testing for potential endocrine disrup- embryo survival, hatching success, and or could specifically detect endocrine-active tion is justified and the risk characterization growth and development are part of the cur- substances. phase should proceed, taking into account all rent draft. Gonad histology is at present not Juvenile fish screening assay. An alterna- information (both endocrine and nonen- included as a requirement for the draft- tive approach using the juvenile stage of docrine related). The candidate higher-tier revised OECD test guideline 206, but histol- OECD fish species was originally proposed at testing assays included in the EDSTAC pro- ogy is mentioned in the EDSTAC proposals the second Duke workshop, Screening posals are discussed in the subsequent section. as a possible further end point, subject to Methods for Detecting Potential (Anti-) Avian reproduction testing. Since the research. Also, gonadal gross morphology Estrogenic/Androgenic Chemicals in Wildlife concerns raised over pesticide poisoning of and accessory organ development of offspring (8). The use of juvenile fish as the basis of a birds during the 1950s and 1960s, there has may be a valuable addition to the current

1010 VOLUME 108 | NUMBER 11 | November 2000 • Environmental Health Perspectives Review • Ecological risk assessment

draft in order to evaluate the potential effects (FFLC) test as standardized by the U.S. EPA, development and validation before they can on the endocrine system of birds (7). fathead minnows are exposed to the test be considered useful for regulatory purposes. Two-generation avian testing. There are compound from the egg (F0 generation) to Subsequently, end points may best be target- a number of scientific and practical challenges early development of the F1 generation off- ed at the endocrine mechanisms of interest in moving directly to an avian two-generation spring, with a test duration of approximately on a case-by-case basis, depending on results study following Tier 1 screening with other 9–10 months. At maturation of the F1 gen- from lower tier assays. The ability to relate species. These include the extension of the eration, breeding pairs or groups are formed end points at testing Tiers 2 and 3 to those northern bobwhite quail or mallard duck to promote spawning (47). The end points used at the Tier 1 screens would be benefi- test protocol to a two-generation study, analyzed in the existing FFLC study include cial, offering a mechanistic basis to support makes these tests even longer and more cost- spawning frequency, number of eggs pro- ecological risk assessments (Figure 3) (16). ly, thus increasing international interest in duced, viability of embryos, hatching success, Therefore, the use of a short-term test proce- the Japanese quail (45) and the modification growth, and development. The EDSTAC dure on fish using selected biochemical para- mentioned in the EDSTAC report (1) for recommends the addition of other parame- meters (e.g., VTG) would be desirable to extending the current U.S. EPA protocol to ters such as gonad weight and histopatholo- make such correlations on a clearer mecha- include end points such as plasma steroid gy, sperm motility and egg maturation, and nistic basis. This approach offers a practical levels, organ weights, and histopathology. plasma VTG and steroid hormone levels. means to developing linkage tools for itera- An extensive validation program with Although multigenerational studies are tion between field and laboratory that can be known endocrine disruptors is needed to the ultimate test for evaluating the effects of used as necessary to support the predictive identify the potential relevancy of recom- compounds that may be bioaccumulative or risk assessment process (11,14). mended biochemical and neurobehavioral have specific modes of action posing hazards Currently, all fish early life-stage study parameters for developmental and reproduc- to progeny, there are also significant limita- (ELS), PLC, and FFLC tests include critical tive effects testing in birds. Extensive control tions to the current regulatory FFLC elements of growth and development, and data in the various avian species of potential method. A major disadvantage of FFLC tests thereby are amenable to examination of thy- interest are also essential to allow reliable is their technical complexity. In fact, because roid hormone-related effects (26). In the interpretation of the many parameters in the of these problems and high cost, FFLC tests future, determination of such effects in fish EDSTAC proposals, and the specificity of are not routinely conducted for any class of may be sufficient to detect potential effects the various endocrine, behavioral, and mor- substance. FFLC tests have, for example, of thyroid-active compounds (with reference phologic parameters are unknown among been used to address concerns over persistent to the mammalian Tier 1 data for thyroid bird species or across chemical groups. pesticides (48,49) or endocrine-active phar- activity) and potentially negate the need for Natural variations in blood hormone maceuticals that affect the reproductive duplicative testing with Amphibia. This pos- concentrations in most avian species is health of fish (50). EDSTAC proposes that sibility should be reviewed as data become unknown; thus, comparisons of experimen- the FFLC test should be conducted after a available; comparative research into this tally derived laboratory data to baseline nor- trigger from either the sorting of initial data aspect is highly desirable. Determination of mal values would not be possible without or from the screening tier (1). A direct tran- genetic sex differentiation in fish may also further research. Because biological signifi- sition to such a high level definitive test prove useful and would enable the sex ratio cance of experimentally derived effects in apparently ignores the practical necessity to of phenotypically similar offspring to be birds may be difficult to define, a large conduct a chronic range-finding test with determined. Knowledge of the genotypic sex research effort is needed to provide a scien- the same fish species before conducting an ratio relative to the phenotypic sex ratio tific context in which to interpret the regula- FFLC test. We therefore recommend incor- would be useful, but markers for genetic sex tory meaning of new end points proposed poration of an intermediate testing tier need to be developed for this purpose (53). for the avian testing tier. The key reason for (Figure 3), including a fish development In conclusion, we support the conduct of using the mallard duck is the demonstrated (extended embryo-larval) test (51) and a fish higher tier tests using fathead minnows as the ability to show egg-shell thinning after expo- reproduction test based on pair-breeding preferred test species for chronic studies of sure to contaminants (46); however, this adults (52). Our recommendation reflects compounds that are active in Tier 1 and that effect has only been observed with a very the fact that for most chemicals, fish are predicted to enter aquatic ecosystems. We small percentage of test compounds. Other embryo-larval tests are predictive of effects in suggest, however, that the EDSTAC recom- end points (e.g., behavioral tests or steroid FFLC tests (48,49); a fish development test mendation to default to the conduct of levels) may be added at a later stage, but not will cover the most sensitive stage of sex FFLC tests at Tier 2 is neither scientifically before fundamental research and validation determination, whereas an adult fish repro- optimal nor ethically responsible in terms of efforts. Therefore, we propose that for both duction test will address both specific and the numbers of animals involved. Instead, we scientific and practical purposes, there is a nonspecific impacts on fecundity and related recommend that a functionally equivalent useful role in higher Tier 2 for the Japanese parameters. The number of substances tested approach be considered, incorporating fish quail reproduction test, with the multigener- could be greatly increased, taking into ELS and PLC tests at Tier 2 and the FFLC ational avian study being more appropriate account the duration and technical chal- test at Tier 3 (Figure 3). Additional end for higher Tier 3 (Figure 3). lenges of the studies and the capacities of points may need to be incorporated into the Fish higher-tier testing. The EDSTAC laboratories with experience in FFLC tests. protocols for these various tests, and these proposals include the use of a full life-cycle We estimate that probably 10 times more end points should be validated and their use- test on the freshwater fathead minnow as a compounds can be tested worldwide using fulness assessed before regulatory implemen- definitive Tier 2 test for endocrine disrup- these PLC tests versus the EDSTAC propos- tation. Subsequent measurements may be tion in fish. Additionally, it is suggested that al for FFLC tests (approximately 200 com- targeted at different end points if the specific other species such as the sheepshead minnow pounds per year as opposed to 20 per year, mode of action of a compound is known should be used when estuarine or marine respectively). (e.g., androgens, aromatase inhibitors, thy- environments are expected to be exposed to The addition of new endocrine disrup- roid-active chemicals). To be optimal, this the test compound. In the fish full life-cycle tor-relevant end points requires considerable will require expert judgment and dialogue in

Environmental Health Perspectives • VOLUME 108 | NUMBER 11 | November 2000 1011 Review • Hutchinson et al.

the problem formulation and risk characteri- Currently, however, there are relatively disruptors have yet to be established. In other zation phases of the ecological risk assessment few data on the influence of endocrine dis- invertebrate taxa, however, there is clear evi- paradigm (11) (Figure 1). ruptors on development and reproduction in dence of hormonal disruption. For example, Higher tier invertebrate testing. EDSTAC Crustacea, and further research is necessary. in gastropod molluscs, the biocide tributyl (1) and the consequent EDSP proposal (2) Indeed, mysids and daphnids, like many tin is suggested to inhibit the aromatization both recommend the use of crustacean life- other arthropods, have a specific hormone of testosterone in female snails, thus increas- cycle assays for higher tier testing, with daph- regulation for growth and reproduction, ing the endogenous androgen level in these nids or mysid shrimps as the suggested test namely the ecdysteroid system (54). This is organisms (9,66). organisms. For daphnids, the EDSP proposes very different from the androgen- and estro- The international Society of Environ- a 3-week chemical exposure period (flow- gen-based regulation of reproduction in mental Toxicology and Chemistry (SETAC)- through conditions), and the end points cur- mammals and oviparous vertebrates. The OECD expert workshop on Endocrine rently evaluated include length of time for the EDSTAC suggestion that estrogen or andro- Disruption in Invertebrates: Endocrinology, appearance of the first brood, body length of gen disruptors are likely to interfere with the Testing and Assessment (EDIETA) held in F0 females, cumulative number of young pro- ecdysteroid activity is speculation and does the Netherlands in December 1998 has duced per female, and effects on the F1 gener- not reflect the other hormonal mechanisms recently published its conclusions (67). Key ation (number of offspring, offspring lengths, known to be important in arthropods. aspects of invertebrate endocrinology and and cumulative mortality). For the mysid Additionally, juvenile hormones should be physiology are discussed, together with labo- shrimp (Americamysis bahia), the test protocol considered in insects and crustaceans (55). ratory test methods and the use of aquatic involves a 6-week chemical exposure period Although changes in estrogen and andro- and terrestrial invertebrates for environmen- (flow-through conditions), and the end points gen metabolism have been measured in tal monitoring and assessment. Many of the currently evaluated include length of time for Daphnia magna exposed to high doses of standard methods for conducting toxicity the appearance of the first brood, sex determi- xenobiotics (56), the putative role of these tests with invertebrates were reviewed at this nation (sex ratio), body length of males and steroids in daphnid physiology remains workshop: it was concluded that although no females, cumulative number of young pro- unclear. Baldwin et al. (56) also found that methods were designed to specifically duced per female, and effects on the F1 gener- high doses of diethylstilbestrol affected the evaluate endocrine disruption, many inverte- ation (number of males and females, body reproduction and growth of D. magna, brate tests include end points that may be length of males and females, and cumulative whereas Kopf (57) found a decrease in repro- endocrine responsive (e.g., development, mortality). duction after exposure of D. magna to 17β- growth, reproduction). It was highlighted, Given the scope of EDSTAC and the estradiol in the low micrograms-per-liter however, that basic research on invertebrate EDSP, the scientific criteria for selecting an range. In contrast, Schweinfurth et al. (58) endocrinology for diverse taxa is urgently arthropod species to test for (anti)androgenic found no effect of 17α-ethynylestradiol in D. needed to remedy our ignorance of mecha- or estrogenic effects is scientifically unclear, magna exposed at a level of several hundred nisms of action, physiologic control, and hor- although it is a pragmatic use of existing regu- micrograms per liter. In conclusion, although mone structure and function in invertebrates. latory test guidelines. Nevertheless, we sup- daphnids may show developmental and repro- Research was called for to evaluate known port the view that where aquatic invertebrates ductive effects associated with toxic effects of endocrine disruptors with a variety of inver- may be exposed to endocrine-active sub- vertebrate endocrine disruptors (56,59), there tebrate bioassays using a suite of designated stances, and because laboratory daphnid pop- is little evidence that such effects are mediated reference compounds (67). ulations consist of parthenogenetic females, via the ecdysteroid or juvenile hormones. In In conclusion, for Crustacea and other an alternative species to daphnids may be nec- addition to daphnids, a number of other arthropods, juvenile and ecdysteroid hor- essary for the assessment of chronic effects of aquatic invertebrate species have been used to mones appear to be more important than endocrine disruptors in sexually reproducing evaluate development and reproductive effects vertebrate-type steroids in influencing sexual Crustacea. For example, the EDSTAC pro- of vertebrate endocrine disruptors. Recent differentiation, growth, and reproduction posal to consider mysid shrimp may be a valid work has shown inhibition of development (55). We therefore suggest that before imple- option for a chronic invertebrate species in and reproduction in marine crustaceans (Tisbe menting the proposal to employ daphnid or endocrine effects testing. The advantages are battagliai) by 20-hydroxyecdysone and mysid life-cycle tests for endocrine disruptors, that mysids have been widely used in regula- diethylstilbestrol (60). In contrast, life-cycle it is essential to address the serious gaps in our tory ecotoxicology, they are a sexually dimor- (21 day) studies with this copepod species did basic knowledge of endocrine function in phic species, and there is an extensive not show significant effects after exposure at these and other invertebrate taxa. In terms of reference database on reproductive toxicity to up to 100 mg/L of either 17β-estradiol, invertebrates often used in ecological effects this species. Unfortunately, the chronic mysid estrone, or 17α-ethynylestradiol (61). characterization, further activity in this area test includes several technical disadvantages, Protocols for developmental and reproductive should be assisted by the recommendations which include major knowledge gaps regard- end points have also been produced for chi- contained in the EDIETA workshop report. ing the basic endocrinology of mysids; identi- ronomids (62) and have been applied to Amphibian extended development and fication of the sexes is technically difficult, phthalates and other important classes of reproduction. Although EDSTAC proposed a and parental cannibalism upon offspring can industrial chemicals known to be endocrine higher-tier amphibian assay incorporating compromise assessment of reproductive out- active (63,64). As part of the ongoing research extended development and reproduction, no put. Given these aspects, reproduction studies program of the European Chemical Industry, specific assay was described (1). Subject to in other invertebrate species may prove to be a review addressing the use of aquatic inverte- adequate research, we suggest that the more suitable for this test. Positive features of brates to detect endocrine activity potential inclusion of such an amphibian chronic test alternative invertebrate test guidelines would has been carried out in collaboration with the might be a technically viable option, but only include much easier culturing and testing pro- U.K. Environment Agency (65). It was con- if triggered by lower tier tests and by exposure cedures than those for the saltwater mysid and cluded that the role of sex steroid hormones in characterization (Figure 3). From an ethical use of freshwater or synthetic sea water for development and reproduction of arthropods perspective, research efforts need to address culturing and chronic toxicity testing. and the potential sites of action of endocrine the comparative effects of a range of

1012 VOLUME 108 | NUMBER 11 | November 2000 • Environmental Health Perspectives Review • Ecological risk assessment endocrine disruptors on fish versus amphibian and their active metabolites before blindly addressing the risk characterization and risk development and reproduction in order to testing xenobiotics as putative endocrine dis- management of natural and synthetic establish potential redundancy between the ruptors. This approach is being evaluated at endocrine disruptors. Within these guidelines, higher-tier fish and amphibian test methods. present using 20-hydroxyecdysone in a num- we propose that although the exposure char- The merits and objectives of a specific test ber of crustacean systems, including daphnids acterization phase remains in principle the with amphibian species on endocrine-dis- (68) and harpacticoid copepods (60). same for endocrine disruptors as for other rupting effects was discussed in the “Frog Selection of reference chemicals. A num- substances, there are convincing scientific metamorphosis screening assay” section. ber of strategies for endocrine-disruptor arguments for seeking to strengthen the A validated frog chronic test for develop- screening and testing include proposals for effects characterization (hazard assessment) mental and reproductive effects may be justi- protocols that are currently in the conceptual component to efficiently deal with potential fied at Tier 3 (Figure 3) if exposure is likely phase of development. To ensure scientific endocrine disruptors and their wildlife to be significant and if the substance is active dialogue in protocol development from the impacts. In response to recommendations in the Tier 2 frog metamorphosis test. For start and to avoid any party having to go from the EDSTAC (1) and proposed EDSP relevance to the field situation on such a back in the future to repeat key studies, it is (2), we suggest a refined set of screening and case-by-case basis, this type of amphibian essential that the international scientific testing tiers to optimize scientific iteration chronic test should consider an ecologically community agree on a list of reference between characterization of both exposure representative species (for the protection of chemicals for use in research and develop- and ecological effects, to minimize the num- wild amphibian populations) together with ment as part of an integrated scientific effort ber of animals required for testing, and to exposure characterization. (69). The chemical selection process should make optimal use of public and commercial consider reference substances for each hor- testing resources. With a scientific flexibility Protocol Development and mone end point (namely, androgens, estro- to allow incorporation of ideas from the Validation gens, and thyroid hormones; both negative OECD and other scientific groups active in Validation can be defined as the input control and positive controls), natural and man- the endocrine-disruptors area, the ECETOC technique used to detect any data that are made substances, composition and definition approach (Figure 3) is considered to offer an inaccurate, incomplete, or scientifically unrea- of reference substance purity, and verification optimal way forward for the development sonable. The basic challenge is to demonstrate of chemical stability in the test systems. In and validation of new endocrine-disruptor that a given assay is both biologically mean- the current European Chemical Industry screening and testing protocols using select- ingful (ecologically relevant) and is repro- aquatic research program, the endocrine-dis- ed reference chemicals. Once validated tests ducible in the hands of the international ruptor compounds selected include diethyl- are available, the emphasis should be on the scientific community. Screening assays should stilbestrol, 17α-ethynylestradiol, fadrozole, use of higher-tier data for the ecological risk also be able to resolve whether endocrine flutamide, genistein, methoxychlor, methyl- assessment of both natural and synthetic interactions are due to the (anti)estrogenic, testosterone, 4-tert-pentylphenol, and the endocrine-disrupting substances. androgenic, and thyroid activity of a given pure antiestrogen ZM189,154 (33,70). substance. The critical objective of the Tier 1 These compounds were principally chosen REFERENCES AND NOTES screening assays is to detect chemicals that in view of their published database from 1. U.S. EPA. Endocrine Disruptor Screening and Testing may be active in vivo, thereby taking into mammalian systems, and the intention is to Advisory Committee (EDSTAC): Final Report. Washington, account bioavailability, comparative metabo- generate comparative data for a range of ani- DC:U.S. Environmental Protection Agency, 1998. 2. U.S. EPA. Endocrine Disruptor Screening Program: pro- lism, and excretion of a substance in animals. mal species looking at a range of diverse posed statement of policy. Fed Reg 63:71542–71568 (1998). Such in vivo screens should ideally avoid false endocrine-disruptor end points. For inverte- 3. Kendall RJ, Dickerson RL. Principles and processes for positives, but it is essential that the potential brates, the recent EDIETA workshop recom- evaluating endocrine disruption in wildlife. Environ Toxicol Chem 15:1253–1254 (1996). for false negatives is minimized. mended comparative research using a variety 4. Kendall RJ, Brouwer A, Giesy JP. A risk-based field and lab- Interspecies comparisons. The validation of reference compounds including metho- oratory approach to assess endocrine disruption in wildlife. of in vivo assays should provide definition of prene, precocene, 20-hydroxyecdysone, lute- In: Principles and Processes for Evaluating Endocrine the similarities and differences between the olin, fadrozole, methyltestosterone, flutamide, Disruption in Wildlife (Kendall RJ, Dickerson RL, Giesy JP, Suk WP, eds). Pensacola, FL:SETAC Press, 1998;1–16. different animal species involved and be rele- 4-tert-pentylphenol, ethynylestradiol, and the 5. ECETOC. Screening and Testing Methods for vant to wildlife populations of concern for a anti-estrogen ZM189,154. In summary, the Ecotoxicological Effects of Potential Endocrine valid ecological risk assessment process (11). method development and validation process Disrupters: Response to the EDSTAC Recommendations and a Proposed Alternative Approach. ECETOC Wherever possible, compounds that have should be made a joint activity between gov- Document No. 39. Brussels:European Centre for the been selected for the validation of mammalian ernmental regulators, international harmo- Ecotoxicology and Toxicology of Chemicals, 1999. protocols should also be considered for the nization organizations such as the OECD, 6. European Commission. European Workshop on the Impact of Endocrine Disrupters on Human Health and validation of protocols with egg-laying and industry research groups. Such positive Wildlife. Proceedings from a Workshop, 2–4 December species. The development of endocrine-dis- collaboration will help ensure that we make 1996, Weybridge, UK. Report reference EUR 17549. ruption tests with novel ecotoxicity test best use of existing data, initiated research Brussels:European Commission, 1996. 7. Tattersfield L, Matthiessen P, Campbell PM, Grandy N, species, for example Amphibia, is more prob- programs, and collaborative activities in Länge R, eds. SETAC Europe/OECD/EC Expert Workshop lematic because the baseline data for their order to avoid duplication and conflicting on Endocrine Modulators and Wildlife: Assessment and reproductive biology and toxicology is far less approaches (71). Science is developing rapid- Testing (EMWAT), Veldhoven, Netherlands, 10–13 April 1997. Brussels:SETAC Europe, 1997. developed than is the case for existing OECD ly in the endocrine-disruptor arena; we all 8. Ankley G, Mihaich E, Stahl R, Tillitt D, Colborn T, test species. This basic problem should be rec- need to take such developments into account McMaster S, Miller R, Bantle J, Campbell P, Denslow N, ognized in drawing together a timetable of during the development of prospective haz- et al. Overview of a workshop on screening methods for validation for the wildlife protocols and par- ard and risk assessment programs (72). detecting potential (anti-) estrogenic/androgenic chemi- cals in wildlife. Environ Toxicol Chem 17:68–87 (1998). ticularly for the protocols on thyroid-active 9. Matthiessen P, Gibbs PE. Critical appraisal of the evi- developments. A similar challenge exists for Conclusions dence for tributyltin-mediated endocrine disruption in invertebrates; arthropod tests should first We advocate the adoption of the existing eco- mollusks. Environ Toxicol Chem 17:37–43 (1998). 10. U.S. EPA. Proposed guidelines for ecological risk establish the effects of endogenous hormones logical risk assessment guidelines (11) for assessment. Fed Reg 61:47551–47631 (1996).

Environmental Health Perspectives • VOLUME 108 | NUMBER 11 | November 2000 1013 Review • Hutchinson et al.

11. U.S. EPA. Final Guidelines for Ecological Risk Assessment. Toxicology and Chemistry (SETAC), 15–19 November 53. Nakamura M, Kobayashi T, Chang XT, Nagahama Y. Risk Assessment Forum. EPA/630/R-95/002F. Washington 1998, Charlotte, NC. Pensacola, FL:SETAC, 1998;149. Gonadal sex differentiation in teleost fish. J Exp Zool DC:U.S. Environmental Protection Agency, 1998. 34. Thorpe KL, Hetheridge MJ, Hutchinson TH, Tyler CR. 281:362–372 (1998). 12. Solomon KR. Endocrine-modulating substances in the Development of an in vivo screen for oestrogenic chemi- 54. Koolman J. Ecdysone metabolism. Insect Biochem environment: the wildlife connection. Int J Toxicol cals using juvenile rainbow trout (Oncorhynchus mykiss). 12:225–250 (1982). 17:159–172 (1998). Environ Toxicol Chem (in press). 55. Chang ES. Comparative endocrinology of molting and 13. GCPF. Framework for the Ecological Risk Assessment of 35. Länge R, Campbell P, Einarson S, Holt M, Hutchinson T, reproduction: insects and crustaceans. Annu Rev Plant Protection Products. Tech Monogr No. 21. McCahon P, Scholz N, Panter G, Tattersfield L, van Entomol 38:161–180 (1993). Brussels:Global Crop Protection Federation, 1999. Egmond R. Endocrine disrupters in wildlife organisms: an 56. Baldwin WS, Milam DL, LeBlanc GA. Physiological and 14. Campbell P, Hutchinson TH. Wildlife and endocrine dis- approach of the European chemical industry to hazard biochemical perturbation in Daphnia magna following rupters: requirements for hazard identification. Environ identification and testing methods. In: Proceedings of exposure to the model environmental oestrogen diethyl- Toxicol Chem 17:127–135 (1998). the 3rd IBC Conference on Endocrine Disrupters in the stilbestrol. Environ Toxicol Chem 14:945–952 (1995). 15. Tillitt DE, Solomon KR, Mihaich EM, Cobb G, Touart L, Environment, London, 14–15 May 1998. London:IBC UK 57. Kopf W. Wirkung endocriner Stoffe in Biotests mit Kubiak TJ. Role of exposure assessment in characteriz- Conferences Limited, 1998;52–57. Wasserorganismen. In: Stoffe mit endokriner Wirkung im ing risks of endocrine-disrupting substances to wildlife. 36. OECD. OECD Guidelines for the Testing of Chemicals. Wasser. Münchener Beiträge zur Abwasser-, Fischerei- u. In: Principles and Processes for Evaluating Endocrine Guideline 204: Fish, Prolonged Toxicity Test: 14-day Flußbiologie. Bd. 50 0368-8275. München, Germany:R. Disruption in Wildlife (Kendall R, Dickerson R, Giesy J, Study. Paris:Organization for Economic Cooperation and Oldenbourg, 1997;39–54. Suk W, eds). Pensacola, FL:SETAC Press, 1998;39–68. Development, 1992. 57. Schweinfurth H, Länge R, Miklautz H, Schauer G. 16. Solomon KR. Research needs. In: Principles and 37. Fite E. The Environmental Protection Agency’s avian Umweltverhalten und aquatische Toxizität von Processes for Evaluating Endocrine Disruption in Wildlife pesticide assessment model. In: Wildlife Toxicology Ethynylestradiol. In: Stoffe mit endokriner Wirkung im (Kendall RJ, Dickerson RL, Giesy JP, Suk WP, eds), and Population Modelling: Integrated Studies of Wasser. Münchener Beiträge zur Abwasser-, Fischerei- u. Pensacola, FL:SETAC Press, 1998;449–455. Agroecosystems (Kendall RJ, Thomas E, Lacher Jr, Flu§biologie. Bd. 50 0368-8275. München, Germany:R. 17. Purchase I. Ethical review of regulatory toxicology guide- eds). Boca Raton, FL:Lewis Publishers, 1993;519–530. Oldenbourg, 1997; 39–54. lines involving experiments on animals: the example of 38. Fry DM. Reproductive effects in birds exposed to pesti- 59. LeBlanc GA, McLachlan JB. Molt-independent growth endocrine disrupters. Toxicol Sci 52:141–147 (1999). cides and industrial chemicals. Environ Health Perspect inhibition of Daphnia magna by a vertebrate anti-andro- 18. Fort DJ, Stover EI. Development of short-term, whole 103(suppl 7)165–171 (1995). gen. Environ Toxicol Chem 18:1450–1455 (1999). embryo assays to evaluate detrimental effects on amphib- 39. European Commission. Ecotoxicology Data Requirements - 60. Hutchinson TH, Pounds NA, Hampel M, Williams TD. ian limb development and metamorphosis. In: Annexes II and III 91/414/EEC. Commission Directive 96/12/EC; Life-cycle effects of 20-hydroxyecdysone and diethyl- Environmental Toxicology and Risk Assessment: Modeling 8th March 1996. Brussels:European Commission, 1996. stilbestrol on the marine copepod Tisbe battagliai. and Risk Assessment (Dwyer FJ, Doane TR, Hinman MI, 40 OECD. Report of the SETAC/OECD Workshop on Avian Environ Toxicol Chem 18:2914–2920 (1999). eds). 6th ed. ASTM STP 1317. Philadelphia, PA:American Toxicity Testing. OECD/GD(96)166. Inter-Organisation 61. Hutchinson TH, Pounds NA, Hampel M, Williams TD. Society for the Testing of Materials, 1997;376–390. Programme for the Sound Management of Chemicals. Impact of ecdysteroids and oestrogens on developmen- 19. DeVito M, Biegel L, Brouwer A, Brown S, Brucker-Davis Paris:Organization for Economic Cooperation and tal and reproductive parameters in the marine copepod F, Cheek AO, Christensen R, Colborn T, Cooke P, Crissman Development Environment Directorate, 1996. Tisbe battagliai. Sci Total Environ 233:167–179 (1999). J, et al. Screening methods for thyroid hormone disrup- 41. Komdeur J, Daan S, Tinbergen J, Mateman C. Extreme 62. ASTM. Standard Guide for Conducting Sediment Toxicity tors. Environ Health Perspect 107:407–415 (1999). adaptive modification in sex ratio of the Seychelles war- Tests with Freshwater Invertebrates. E1383-94. 20. Norris DO. Vertebrate Endocrinology. 3rd ed. San Diego, bler’s eggs. Nature 385:522–525 (1997). Philadelphia, PA:American Society for Testing and CA:Academic Press, 1997. 42. Smallwood PD, Smallwood JA. Seasonal shifts in sex Materials, 1994. 21. McNabb FMA. Thyroid Hormones. Englewood Cliffs, ratios of fledgling American kestrels (Falco sparverius 63. Brown D, Thompson RS, Stewart KM, Croudace CP, NJ:Prentice Hall, 1992. paulus): the early bird hypothesis. Evol Ecol 12:839–853 Gillings E. The effect of phthalate ester plasticisers on 22. Kanamori A, Brown DD. The analysis of complex devel- (1998). the emergence of the midge Chironomus riparius from opmental programmes: amphibian metamorphosis. 43. Mineau P, Boersma DC, Collins B. An analysis of avian treated sediments. Chemosphere 32:2177–2187 (1996). Genes Cells 1(5):429–435 (1996). reproduction studies submitted for pesticide registration. 64. ECETOC. Environmental Oestrogens - A Compendium of 23. Brown DD, Wang Z, Kanamori A, Eliceiri B, Furlow JD, Ecotoxicol Environ Saf 29:1–26 (1994). Test Methods. ECETOC Document No. 33. Brussels: Schwartzman R. Amphibian metamorphosis–a complex 44. Montague B. “Brian.xls”, the U.S. EPA Ecotox 1-liner European Centre for the Ecotoxicology and Toxicology of program of gene-expression changes controlled by the database. Washington DC:U.S. Environmental Protection Chemicals, 1996. thyroid-hormone. Recent Prog Horm Res 50:309–315 (1995). Agency, Office of Pesticide Programs, 1999. 65. Pinder LCV, Pottinger TG, Billingshurst Z, Depledge MH. 24. Hayes TB, Wu TH. Role of corticosterone in anuran 45. OECD. Proposed draft OECD Guideline for testing of Endocrine Function in Aquatic Invertebrates and metamorphosis and potential role in stress-induced chemicals. Avian reproduction toxicity test in the Evidence for Disruption by Environmental Pollutants. metamorphosis. Neth J Zool 45(1–2):107–109 (1995). Japanese quail (Coturnix coturnix japonica) and Northern Wallingford, UK:UK Environment Agency, 1999. 25. Leatherland JF. Endocrine and reproductive function in bobwhite (Colinus virginianus). Paris:Organization for 66. Bettin C, Öhlmann J, Stroben E. TBT-induced imposex in Great Lakes salmon. In: Chemically-Induced Alterations in Economic Cooperation and Development Environmental marine gastropods is mediated by an increasing andro- Sexual and Functional Development: The Human/Wildlife Health and Safety Division, 1997. gen level. Helgol Meeresunters 50:299–317 (1996). Connection (Colborn T, Clement C, eds). Princeton, 46. Lundholm CE. DDE-induced eggshell thinning in birds: 67. DeFur P, Crane M, Ingersoll C, Tattersfield L, eds. NJ:Princeton Scientific Publishing Co., 1991;129–145. effects of p,p´-DDE on the calcium and prostaglandin Endocrine Disruption in Invertebrates: Endocrinology, 26. Kime DE. Endocrine Disruption in Fish. London:Kluwer metabolism of the eggshell gland. Comp Biol Physiol Testing and Assessment. Pensacola, FL:SETAC Press, Academic Publishers, 1998. 118C:113–128 (1997). 1999. 27. Panter GH, Thompson RS, Sumpter JP. Adverse repro- 47. U.S. EPA. Fish Life-cycle Toxicity Tests. EPA 540/9–86–137. 68. Baldwin WS, LeBlanc GA. In vivo biotransformation of ductive effects in male fathead minnows (Pimephales Washington DC:U.S. Environmental Protection Agency, testosterone by phase I and II detoxification enzymes promelas) exposed to environmentally relevant concen- Hazard Evaluation Division, Office of Pesticide Programs, and their modulation by 20-hydroxyecdysone in Daphnia trations of the natural oestrogens, oestradiol and 1986. magna. Aquat Toxicol 29:103–117 (1994). oestrone. Aquat Toxicol 42:243–253 (1998). 48. McKim JM. Evaluation of tests with early life-stages of 69. Ashby J, Houtoff E, Kennedy SJ, Stevens J, Bars R, Jekat 28. Tyler CR, van Aerle R, Hutchinson TH, Maddix S, Tripp H. fish for predicting long term toxicity. J Fish Res Board FW, Campbell PM, van Miller J, Carpanini FM, Randall An in vivo testing system for endocrine disrupters in fish Can 34:1134–1154 (1977). GLP. The challenge posed by endocrine-disrupting early life-stages using induction of vitellogenin. Environ 49. McKim JM. Early life-stage toxicity tests. In: Fundamentals chemicals. Environ Health Perspect 105:164–169 (1997). Toxicol Chem 18:337–347 (1999). of Aquatic Toxicology (Rand GM, Petrocelli SR, eds). 1st 70. Dukes M, Chester R, Yarwood L, Wakeling AE. Effects of 29. Takahashi H. Juvenile hermaphroditism in the zebrafish, ed. New York:Hemisphere Publishers, 1985;58–95. a non-steroidal pure antioestrogen, ZM189,154, on Brachydanio rerio. Bull Fac Fish Hokkaido Univ 28:57–65 50. Länge R, Schweinfurth H, Panter G, Croudace C. Growth oestrogen target organs of the rat including bones. J (1977). and reproduction of fathead minnow (Pimephales prome- Endocrinol 141:335–341 (1994). 30. Pyron M, Beitinger LT. Behaviour of male fathead min- las) exposed to the synthetic steroid hormone 71. Stahl RG, Clark JR. Uncertainties in the risk assessment nows in the presence of an additional male or female ethynylestradiol in a life-cycle test. In: Proceedings of the of endocrine-modulating substances in wildlife. In: fathead minnow. Tex J Sci 41:151–154 (1989). 7th Annual Meeting of SETAC-Europe, 6–10 April 1997, Principles and Processes for Evaluating Endocrine 31. Laale HW. The biology and use of zebrafish Brachydanio Amsterdam. Pensacola, FL:SETAC, 1997;32. Disruption in Wildlife (Kendall RJ, Dickerson RL, Giesy JP, rerio in fisheries research. J Fish Biol 10:121–173 (1977). 51. Gray MA, Niimi AJ, Metcalf CD. Factors affecting the Suk WP, eds). Pensacola, FL:SETAC Press, 1998;431–448. 32. Purdom CE, Hardiman PA, Bye VF, Eno NC, Tyler CR, development of testis-ova in medaka Oryzias latipes, 72. U.S. EPA Science Advisory Board. Review of the EPA’s Sumpter JP. Estrogenic effects of effluents from sewage exposed to octylphenol. Environ Toxicol Chem Proposed Environmental Endocrine Disrupter Screening treatment works. Chem Ecol 8:275–285 (1994). 18:1835–1842 (1999). Program by a Joint Subcommittee of the Science 33. Panter GH, Campbell PM, Hutchinson TH, Kime DE, Länge 52. Harries JE, Runnals T, Hill E, Harris CA, Maddix S, Advisory Board and Scientific Advisory Panel. EPA-SAB- R, McCahon P, Sumpter JP, Tattersfield LJ, Tyler CR. Sumpter JP, Tyler CR. Development of a reproductive EC-99-013. Washington, DC:U.S. Environmental Protection Effects of endocrine disrupters in the fathead minnow performance test for endocrine disrupting chemicals Agency, 1999. (Pimephales promelas) [Abstract]. In: Proceedings of the using pair-breeding fathead minnows (Pimephales 18th Annual Meeting of the Society of Environmental promelas). Environ Sci Technol 34:3003–3011 (2000).

1014 VOLUME 108 | NUMBER 11 | November 2000 • Environmental Health Perspectives