Volume 75, Issue 11-12

Nuclear receptors (NRs) comprise a family of 49 members that share a common structural organization and act as ligand-inducible transcription factors with major (patho)physiological impact. For some NRs (“orphan receptors”), cognate ligands have not yet been identified or may not exist. The principles of DNA recognition and ligand binding are well understood from both biochemical and crystal structure analyses. The 3D structures of several DNA-binding domains (DBDs),in complexes with a variety of cognate response elements, and multiple ligand-binding domains (LBDs), in the absence (apoLBD)and presence (holoLBD) of agonist, have been established and reveal canonical structural organization. Agonist binding induces a structural transition in the LBD whose most striking feature is the relocation of helix H12, which is required for establishing a coactivator complex, through interaction with members of the p160 family (SRC1, TIF2, AIB1) and/or the TRAP/DRIP complex. The p160-dependent coactivator complex is a multiprotein complex that comprises histone acetyltransferases (HATs), such as CBP,methyltransferases, such as CARM1, and other enzymes (SUMO ligase,etc.). The agonist-dependent recruitment of the HAT complex results in chromatin modification in the environment of the target gene promoters, which is requisite to, or may in some cases be sufficient for, transcription activation. In the absence of ligands, or in the presence of some antagonists, certain NRs are bound to distinct multiprotein complexes through the interaction with corepressors, such as NCoR and SMRT. Corepressor complexes comprise histone deacetylases (HDACs) that have the capacity to condense chromatin over target gene promoters. Ligands have been designed that selectively modulate the interaction between NRs and their coregulators. Both HATs and HDACs can also modify the acetylation status of nonhistone proteins, but the significance in the context of NR signaling is unclear. NRs communicate with other intracellular signaling pathways on a mutual basis, and their functionality may be altered, positively or negatively, by post-translational modification. The majority of NRs act as retinoid X receptor (RXR) heterodimers in which RXR cannot a priori respond autonomously to its cognate ligand to activate target gene transcription. This RXR subordination allows signaling pathway identity for the RXR partner. The corresponding mechanism is understood and reveals cell and NR selectivity, indicating that RXR can, under certain conditions, act autonomously. NRs are regulators of cell life and death,and NR malfunction can be at the basis of both disease and therapy, as is impressively documented in the case of acute promyelocytic leukemia. Recently, several pathways have been uncovered that link NR action with cell proliferation and apoptosis.

It has been postulated that nuclear receptors (NRs) regulate transcription via interactions with chromatin and the basal transcription machinery at the promoters of genes. Coregulators (coactivators or corepressors) are important in mediating these interactions and thereby modulating positive or negative receptor activity. A large number of putative coactivators have been isolated, several of which will be reviewed with respect to certain "criteria" initially proposed for coactivators. We will discuss, with reference to in vitro and in vivo experiments, the main steps in initiation that are influenced by coactivators: (1) initiation (e.g., SRC-1 family, CBP); (2) repetitive transcription (e.g., TRAPs/DRIPs); (3) RNA processing (PGC-1, etc); and (4) termination/turnover (E6-AP, etc). A variety of enzyme functions have been implicated in the coactivator complex including acetylase, methylase, ubiquitin ligase, kinase, and phosphatase activities. Moreover, coactivators and corepressors appear to exist in the steady-state cell as a series of multiprotein complexes referred to collectively as the "coregulatorsome". Different subcomplexes within the coregulatorsome may have different levels of preference for individual receptors or promoters, likely contributing to context-specific functions of NRs in target tissues.

Estrogens and progestins play important roles in regulating growth and differentiation of a wide range of cell types, in both reproductive and nonreproductive tissues. Not surprisingly, therefore, endocrine active substances that mimic the actions of these steroid hormones have been demonstrated to have untoward effects on the reproductive function of a variety of animals. Although direct links between exposure to endocrine active substances of this class and reproductive abnormalities in humans have been difficult to establish, the potential for harm by this class of chemicals warrants further investigation. This chapter summarizes our current understanding of the molecular pharmacology of the estrogen and progesterone receptors, ER and PR, beginning with a historical perspective of ER pharmacology and ending with a comparison of these receptors to the vitamin D receptor (VDR), a nonsteroid hormone nuclear receptor involved in regulating growth and development in nonreproductive tissues. Major topics discussed include receptor subtypes (or isoforms), receptor functional domains, ligand-binding characteristics, receptor structure, cofactor binding, effects of phosphorylation, and nonclassical modes of action. This discussion will demonstrate the need for developing novel screens for potential endocrine disruptors that incorporate our current understanding of nuclear receptor pharmacology.

Chemical contaminants with antiandrogen or androgen activity have been identified in the environment. The mode of action of these endocrine disruptors derives from their ability to bind to the androgen receptor (AR), a member of the steroid receptor family of nuclear receptors. The AR is a ligand-activated transcription factor with properties unique among members of the steroid receptor family. Binding of endocrine disruptors to the AR impacts its ability to activate or inhibit AR-regulated genes. Most notable of these are the agonist-induced and antagonist-inhibited interdomain interactions that influence AR stability and function. Environmental antiandrogens identified thus far are metabolites of pesticides and herbicides, exerting androgen antagonist effects by blocking AR-induced gene transcription required for male sexual development. Environmental androgens can be precursors of the naturally occurring biologically active androgens testosterone and dihydrotestosterone or result from anabolic steroid use in the livestock industry. They have agonist activity by their ability to bind the AR and mimic the natural hormone, increasing AR-mediated transcription of androgen-responsive genes. The presence of masculinized female fish in polluted rivers indicates the presence of androgen-like pollutants in the environment.

Nuclear hormone receptors function to transduce hormonal signals into transcriptional responses by controlling the activity of specific target genes. These target genes comprise a genetic network whose coordinate activity defines the physiological responses to hormonal signals. Dissecting nuclear hormone receptor functions in vivo by gene inactivation and transgenic strategies represents an invaluable and powerful approach to increase our knowledge of these genetic networks and their physiological functions. Glucocorticoids and mineralocorticoids are involved in numerous physiological processes important to maintain metabolic, cardiovascular, central nervous, and immune system homeostasis. Germline and somatic gene targeting as well as an increased dosage of the glucocorticoid receptor (GR) allows the characterization of the various functions and molecular modes of action of this receptor. Most of the effects of the GR are mediated via activation and repression of gene expression. To separate activating from repressing functions of the GR, a point mutation was introduced which allowed us to characterize and distinguish functions dependent on GR binding to DNA from those mediated by protein/protein interaction. Cell/tissue-specific mutations of the gluco- and mineralocorticoid receptor is the basis for the evaluation of their cell-specific functions, including the characterization of target genes of the receptors in order to describe their specific effects on different targets.

Retinoids, the active metabolites of vitamin A, regulate complex gene networks involved in vertebrate morphogenesis, growth, cellular differentiation, and homeostasis. They are used for the treatment of skin disorders and as chemopreventive agents for certain cancers. Molecular biology and genetic studies performed during the last 15 years in vitro, using either acellular systems or transfected cells, have shown that retinoid actions are mediated through heterodimers between the 8 major RARα, β, and γ; isoforms and the 6 major RXRα, β and γ isoforms that belong to the nuclear receptor (NR) superfamily, and act as ligand-dependent transcriptional regulators. Furthermore, RXRs not only heterodimerize with RARs, but also with numerous other members of the NR superfamily. As in vitro studies are carried out under nonphysiological conditions, they only indicate what is possible, but not necessarily what is actually occurring in vivo. Therefore, mutations have been introduced by homologous recombination (HR) in F9 embryonal carcinoma (EC) cells, a cell-autonomous system that differentiates in the presence of RA, in order to disrupt RAR and RXR genes and establish their cellular and molecular functions in RA-induced differentiation. However, genetic approaches in the animal should be used to determine the function of retinoid receptors under truly physiological conditions. HR in embryonic stem (ES) cells, has therefore been used to generate null mutations of the various RARs and RXRs in the mouse germline. As reviewed here, the generation of such RAR and RXR germline mutations, combined with pharmacological approaches to block the RA signaling pathway, has provided many valuable insights on the developmental functions of RA receptors. However, due to (i) the complexity in "hormonal" signaling through transduction by the multiple RARs and RXRs, (ii) the functional redundancies (possibly artefactually generated by the mutations) within receptor isotypes belonging to a given gene family, and (iii) in utero or postnatal lethality of certain germline null mutations, these genetic studies through germline mutagenesis have failed to reveal many of the physiological functions of RARs and RXRs, notably in adults. We conclude that spatio-temporally controlled somatic mutations generated in animal models in given cell-types/tissues and at chosen times during pre- and postnatal life, are required to reveal the physiological and pathophysiological functions of the receptor genes involved in the retinoid signaling pathway throughout the life of the mouse.

Two related nuclear receptors, the pregnane X receptor (PXR) and the constitutive androstane receptor (CAR), act as xenobiotic sensors that protect the body from a multitude of foreign chemicals (xenobiotics) and play a central role in the metabolism and clearance of steroids and toxic endogenous lipids (endobiotics). A structurally diverse array of chemicals including pharmaceutical drugs, steroids, herbal extracts, and pesticides activate PXR or CAR. This activation results in induction of overlapping, but yet distinct drug clearance pathways consisting of cytochrome P450 enzymes, conjugating enzymes, drug transporters, and other related proteins. Similar pathways are also utilized to protect the body from toxic compounds of endogenous origin. Thus, the xenobiotic regulatory circuit contributes both to drug-drug and food-drug interactions as well as endocrine disruption. Consistent with the notion that xenobiotic receptors regulate drug clearance, single nucleotide polymorphisms (SNPs) in either the receptors themselves or receptor-binding sites in the regulatory region of genes encoding metabolic enzymes appear to contribute to the polymorphic expression of components of drug clearance pathways. Together, the xenobiotic receptors PXR and CAR confer metabolic immunity via the ability to control an integrated array of target genes.

Signaling pathways can be linear, but more complex patterns are common. Growth factors and many other extracellular signals cannot directly enter cells and transduce their information via membrane-bound receptors. In contrast, steroid receptors are members of the nuclear receptor superfamily and await their cognate hormones inside the cells. These two types of signaling pathways are extensively intertwined and crosstalk at many different levels. A wide range of extra- and intracellular signals, including a variety of growth factors, can activate the transcriptional activity of steroid receptors in the absence of their cognate hormones. Conversely, steroid receptors lead a double life. By coupling to signaling molecules that mediate signal transduction of extracellular factors, they can elicit very rapid nongenomic responses. The signaling pathways of steroid-independent activation of steroid receptors, on the one hand, and of nongenomic signaling by steroid receptors, on the other, display a remarkable reciprocal relationship suggesting that these two modes of signaling crosstalk may be two faces of the same coin.

The estrogen receptors alpha and beta (ERα and ERβ) mediate the changes in gene expression from physiological and environmental estrogens. Early studies identified classical estrogen response elements (EREs) in the promoter region of target genes whose expression is regulated by estrogen and to which the ERs bind via their DNA-binding domain (DBD). EREs in the pituitary prolactin promoter, for example, mediate an activation by both ERα and ERβ albeit with different affinities for different ligands. Full activation in most cell types requires the integrity of the activation function 2 (AF-2) in the receptors ligand binding domain (LBD), which is engaged by estrogens and disengaged by tamoxifen, raloxifene, and other antiestrogens. However, in some cells and ERE contexts, the AF-1 in the ERα amino terminal domain (NTD) is sufficient. We now know that ERs also regulate expression of target genes that do not have EREs, but instead have various kinds of alternative response elements that bind heterologous transcription factors whose activity is regulated by interactions with ERs. Thus, ERα activates genes, including collagenase and cyclin D1, an important mediator of cellular proliferation, by AP-1 and CRE sites, which bind Jun/Fos or Jun/ATF-2 transcription factors. ERα also activates gene expression through GC-rich elements that bind the SP1 transcription factor. Finally, we also know that ERs mediate inhibition of the expression of many genes. In one well-studied instance, ERs counterexpression of genes involved in the inflammatory response by inhibiting the action at tumor necrosis factor response elements (TNF-REs) that bind the NFkappaB transcription factor. ERβ is especially efficient at this inhibition. ERα activation of AP-1/CRE target genes is of special interest because of the putative role of these target genes in mediating proliferation. The AF-1 and AF-2 functions of ERα are both needed for this activation in most cell types. However, in uterine cells, the AF-1 function is sufficient. Thus, the antiestrogen tamoxifen, which allows AF-1, mimics estrogen and drives activation of AP-1/CRE target genes and proliferation of uterine cells. This estrogen-like action, which can increase the risk of uterine cancer, complicates the use of tamoxifen to prevent breast cancer. Surprisingly, ERβ inhibits AP-1/CRE target genes in the presence of estrogen. When both receptors are present, ERβ efficiently opposes activation by ERα. Moreover, ERβ activates the AP-1/CRE target genes in the presence of antiestrogens especially so-called "complete" antiestrogens raloxifene, and ICI 182, 780. We here review the evidence for different kinds of promoter elements that mediate ER action, for the differential ligand preferences of ERα and ERβ at these different elements, and the potential mechanisms by which they are mediated. One attractive strategy for the investigation and comparison of potential environmental estrogens is to assay their activity in cell culture systems using reporter genes with simplified promoter elements. Thus, the findings of complexity in ERα and ERβ activation at different types of response elements needs to be taken into account in the development and interpretation of assays using simplified promoter elements systems.

Sex determination and differentiation in the male is an orderly sequence of events coordinated by genetic and hormonal factors operating in a time- and concentration-dependent manner. The constitutive sex in mammals is female. Disorders of fetal sex development have provided the means to identify testis-determining genes and the molecular mechanisms of hormone action. Thus, the androgen receptor, a nuclear hormone receptor critical for androgen-induced male sex differentiation, displays unique intra-receptor and protein-protein interactions which, when disturbed, can result in extreme forms of sex reversal. Polymorphic variants are associated with milder disorders of sex development. Against this genetic background, endocrine active substances may further contribute to the underlying causes of an increase in male reproductive tract disorders.

The identification of naturally occurring nuclear receptor mutations highlights the critical role that many of these transcription factors play in human endocrine development and function. Inactivating mutations in the ligand-dependent nuclear receptors (TRβ, VDR, ERα, GR, MR, AR) are well characterized in patients with conditions such as androgen insensitivity syndrome (AIS) and vitamin D resistance. On the other hand, mutations in TRβ act in a dominant negative manner to cause hormone resistance. Inactivating mutations in orphan nuclear receptors have also been identified (PPARγ2, HNF4α, PNR, NURR1, SF1, DAX1, SHP) and reveal important developmental and metabolic functions for this group of receptors with previously elusive physiologic roles. In addition to loss of function mutations, receptor activation can result from mutations that confer constitutive activity or altered ligand responsiveness to the receptor (MR, AR), or from genetic duplication (DAX1) or the expression of fusion proteins (RARA, PPARγ1). Together, these naturally occurring mutations provide fascinating insight into key structural and functional receptor domains to reveal the diverse role nuclear receptors play in human biology.

Nuclear receptors function as ligand-regulated transcription factors and modulate the expression of sets of genes in response to varying concentrations of ligands. The ligand modulators can be endogenous metabolites that function as hormones, or they can be exogenous substances, such as pharmaceutical agents or environmental substances of natural or man-made origin, which in some cases can cause endocrine disruption. Ligands modulate nuclear receptor activity by binding to their ligand-binding domains and stabilizing conformations that lead either to transcriptional activation or repression. The ligand-binding pocket is somewhat flexible, and binding affinities can be measured over a 10-million-fold range (i.e., with equilibrium dissociation constant values ranging from ca. 0.01 nM to 100 μM). Thus, it is not surprising that by binding a large variety of structures, some nuclear receptors can appear to be promiscuous; however, when affinity is considered, the binding patterns are more restricted. The spectrum of ligands that bind to the estrogen receptor has been most thoroughly investigated. Those from natural sources include natural products in food, such as soy isoflavones and whole grain lignans, as well as microbial products and components from wood. Aside from pharmaceuticals, man-made estrogen ligands can be found in industrial products, such as alkyl phenols from nonionic detergents, bisphenols from plastics, indicator dye impurities, polymer chemicals, and chlorinated aromatics and pesticides. Exogenous ligands are also known for the androgen and progesterone receptors. While it is possible that endocrine disruption can result from exogenous chemicals acting directly as ligands for the nuclear receptors, endocrine disruption needs to be considered in the broader context; thus, compounds also need to be assessed for their effects at other levels, such as on endogenous hormone production, transport, metabolism, and clearance, and at points in signal transduction cascades that are beyond the ligand-receptor interaction.

2,3,7,8-Tetrachlorodibenzo- p -dioxin (TCDD), one of most toxic man-made chemicals, binds arylhydrocarbon receptor (AhR or dioxin receptor), whose endogenous ligand remains unknown, with an extremely high affinity and expresses pleiotropic biological effects. From analysis of the primary structures, AhR belongs to a distinct group of a super-gene family from that of nuclear receptors. PCB, 3-methylcholanthrene, and benzo(a)pyrene are also ligands to the AhR, and these polycyclic aromatic chemicals are considered to display pleiotropic biological effects such as induction of a variety of drug-metabolizing enzymes, teratogenesis, tumor promotion, immunodeficiency due to thymic involution, and liver damage. Generation of the AhR-deficient mice by gene knock-out technology revealed that these biological effects described are mediated by AhR, because AhR( - / - ) mice lost susceptibility to these effects by TCDD and benzo(a) pyrene. It has recently been revealed that AhR is also involved in reproduction of female mice. Although the detailed mechanisms of involvement of AhR in exerting these effects are not always clarified, AhR is generally considered to function as a transcription factor, which activates the expression of genes by binding directly the XRE sequence in their promoter in a heterodimer form with Arnt. From analysis of DNA transfection and GST pull-down assays, it is revealed that AhR and Arnt interact with various coactivators such as RIP140, SRC-1/NcoA, and CBP/p300 to transmit their transactivation activity to general transcription factors (GTFs). AhR has also been shown to interact with various regulatory factors including Rb, NF-κB, and SP1, resulting in mutual inhibition or synergistic enhancement of their activities depending on the mode of localization of their cognate binding sequences in the target genes. Agonistic and antagonistic properties of various ligands to AhR also are discussed.

Most research to identify endocrine-disrupting chemicals and their impacts has relied on mammalian models or in vitro systems derived from them. But nuclear receptors (NRs), the proteins that transduce hydrophobic hormonal signals and are major mediators of endocrine disruption, emerged early in animal evolution and now play biologically essential roles throughout the Metazoa. Nonmammalian vertebrates and invertebrates, many of which are of considerable ecological, economic, and cultural importance, are therefore potentially subject to endocrine disruption by synthetic environmental pollutants. Are methods that rely solely on mammalian models adequate to predict or detect all chemicals that may disrupt NR signaling? Regulation of NRs by small hydrophobic molecules is ancient and evolutionarily labile. Within and across genomes, the NR superfamily is very diverse, due to many lineage-specific gene and genome duplications followed by independent divergence. Receptors in nonmammalian species have in many cases evolved unique molecular and organismal functions that cannot be predicted from those of their mammalian orthologs. Endocrine disruption is therefore likely to occur throughout the metazoan kingdom, and a significant number of the thousands of synthetic chemicals now in production may disrupt NR signaling in one or more nonmammalian taxa. Many of these endocrine disruptors will not be detected by current regulatory/scientific protocols, which should be reformulated to take account of the diversity and complexity of the NR gene family.

A critical review is made of techniques for analysis of residues of endocrine active substances (EASs) including sampling, extraction cleanup and determination based on GC/MS, LC/MS, ELISA, and bioassays. The growing importance of receptor-based in vitro bioassays is highlighted for integrated monitoring of environmental levels of certain classes of EASs and for establishing exposures. Some recent advances in methods of analysis for each of the key classes of EASs are summarized including for organochlorines, PCBs, dioxins and dioxin-like substances, polybrominated diphenyl ethers, phenolic xenoestrogens, phthalates, organotin compounds, steroidal hormones, and phytoestrogens. The issues raised in interpreting complementary chemical and bioassay data at an effects level are briefly discussed.

Steroidal hormones produced by humans and animals are constantly excreted into the environment in their active forms. The primary steroid hormones are progesterone, estrone, estradiol, testosterone, and cortisol, all of which are lipophilic and poorly soluble in water. The steroids of major concern are estrone and estradiol-17β, since they exert their physiological effects at a lower concentration than other steroids and can be found in the environment in concentrations above their LOEL for fish and plants (10 ng/l). The steroid hormones can be readily measured in run-off, soil, and groundwater, but each steroid has its distinct pathway of transport. Since the major source of steroids in the environment appears to be cattle and chickens, the hormonal steroid input into the environment could be drastically reduced by well-established techniques such as buffer strips and composting.

Phytohormones, which naturally occur in plants, must be taken into consideration for their contribution to the total human exposure to potential endocrine active substances (EASs). Phytohormones are usually divided into two families: phytoestrogens that are mainly occur from soybeans and soy derivatives and phytosterols that occur from vegetables and vegetable oils. The present paper compiles different sources of information about the concentration of phytohormones in foodstuffs in order to assess the current human exposure to those substances via food. A particular attention is given to most exposed groups of consumers, on the one hand, infants and young children fed with soy-based infant formulas for their exposure to phytoestrogens and on the other hand, consumers of fortified foods for their exposure to phytosterols. Available literature shows that the total dietary intake of isoflavones could reach 20 to 25 mg/day/person for the Japanese adult population. For infants and young children, the quantity of phytoestrogen ingested is likely to be 35-50 mg/day/person corresponding on a body weight basis to an exposure 7 to 11 times higher. Regarding phytosterols, an assessment of the exposure via food was done, considering both their natural occurrence and their potential concentration in fortified foods. Results shows that the "natural" exposure is estimated at 340 ± 440 mg/day/person at the mean and at 1040 mg/day/person at the 95th percentile. Considering the potential exposure via fortified foods, it is estimated at 2700 ± 1200 mg/day/person at the mean and at 4700 mg/day/person at the 95th percentile. After their ingestion, isoflavones are absorbed from the intestinal tract before being excreted in the urine and feces. The increasing use of phytohormones in human foodstuffs could increase locally their release into the environment. Nevertheless, considering the weak estrogenic potential of phytohormones in relation to synthetic or endogenous steroids, any introduction of these substances into aquatic ecosystems would probably have comparatively minor effects on aquatic organisms.

A wide variety of chemicals are used in pharmaceuticals. Most of these are already under thorough control for endocrine activity. The main causal agents recognized for endocrine disruption from sewage are substances used in medicine (sex hormones, glucocorticoids, and others), natural substances (estrone and 17β-estradiol), and synthetic estrogens (e.g., 17α-ethinylestradiol). Similar substances are used in anabolic agents (growth hormones) in livestock production in some countries. Although the estimated use of anabolic agents in livestock production is approximately one order of magnitude below the natural release of estrogens from farm animals, their possible significance remains unanswered. At present, no other medical substances are recognized as endocrine disruptors in the environment. However, candidates may be identified on the basis of simple assumptions regarding their use and activity: (1) Nonestrogenic steroids may react with environmental endocrine receptors or metabolize on their way to the environment and thus form endocrine disruptors. (2) Many high-volume drugs released to the environment have not yet been tested for their endocrine properties, and some of these are known to interact with the human endocrine system. (3) Compared to medicinal substances, personal care products and additives in drugs are used in high amounts; from this group, parabens, siloxanes, and other substances are suspected of causing endocrine disruption in the environment.

Of the xenobiotic endocrine active substances (EASs), tributyltin (TBT) has had the clearest link to an impact on aquatic ecology. Its release from marine antifouling paints had a drastic impact on dogwhelk populations in polluted harbors due to a masculization effect. 4- tert -nonylphenol is seen as the most significant of the industrial xenobiotic estrogen mimics, being implicated as the dominant endocrine disruptor in certain industrialized river reaches. Apart from hot spots associated with particular industries, the estrogenic alkylphenols, phthalates, and bisphenol A are present in effluent and receiving water at concentrations below that which would give cause for concern. Other more bioaccumulative compounds such as polybrominated flame retardants, dioxins, furans may possess some endocrine active properties. The possibility of additivity effects may yet mean that low concentrations of xenobiotic EASs will need careful consideration. It is noted that considerable quantities of many of these compounds are often found in sewage sludge and sediments.

Considerable information exists as to the initial concentrations of pesticide residues to be expected in soils, plants, and water. Empirical or theoretical models have been developed for incorporating this data into exposure assessments for humans as well as terrestrial and aquatic wildlife. In addition, monitoring data exists for many older products, especially with respect to typical concentrations observed in food commodities for human consumption and in surface and ground waters. Estimated and observed concentrations of pesticides in these matrices have been routinely employed for more than 30 years in assessing the potential impacts of pesticides on a variety of biologically relevant endpoints. The same data will also prove useful for exposure assessments of endocrine active substances. There are some additional research needs, however. First, further research and development is needed to ensure that estimation and monitoring methods for pesticide concentrations in soil, water, and food are applicable and utilized for all important and relevant cultural, agronomic, and environmental conditions. This is especially true with respect to developing countries and tropical climates, which are often disproportionately ignored in favor of developing countries and temperate climates. Second, methodologies for collection of monitoring data and generation of modeled estimates for pesticide residues in soil, water, and food need to be carefully designed with the requirements of higher-tier, probabilistic exposure assessments in view. Although worst-case, point estimates or analyses may be useful for screening-level assessments, advanced assessments targeted at addressing the likelihood of biologically relevant exposures are urgently required by scientists and regulatory authorities for reaching sound risk assessment and risk management decisions.

This review describes the role of metabolism with endocrine active substances. Many modern synthetic compounds are readily metabolized to more polar forms that often contain hydroxy groups. This presence of polar groups and aromatic moieties in the parent compound or metabolite can play an important role in the mechanism of endocrine disruption. In addition, phase II metabolism (e.g., glucuronidation) can also lead to deactivation of the endocrine properties. In the case of bisphenol A and alkylphenols, metabolism can be considered as a detoxification mechanism as glucuronides decrease of inhibit binding to the estrogen receptors. In the case of phthalate esters, the primary metabolites, the monoesters, and further degraded metabolites do not interact with the estrogen receptor either. In contrast, the demethylation of methoxychlor in fish and other vertebrate species leads to metabolites with an increased affinity for the estrogen receptor. Certain PCB metabolites with hydroxy groups on the para position without vicinal chlorines have estrogenic activity, but these metabolites are not relevant for the environment. PCB metabolites with methylsulfonyl groups are commonly found in environmental biota and have been associated with several endocrine, developmental, and reproductive effects. Some DDT metabolites bind weakly to the estrogen receptor, but the major biotransformation product p,p -DDE is an androgen receptor (AR) antagonist. Vinclozolin is an anti-androgen and this effect appears to caused by two of its more water-soluble metabolites. The chloro- s -triazines exhibit an in vitro induction of aromatase, but their dealkylated metabolites show a decrease or lack of this effect. It is recognized that common metabolic processes can differ strongly among species that complicates ecotoxicological risk assessment of endocrine active substances. In conclusion, the testing of metabolites for endocrine-disrupting properties should be encouraged in the future to establish a better risk assessment process. An appendix containing levels and half-lives of various endocrine-disrupting chemicals in the environment and in wildlife is included at the end of this article.

Multimedia transport, partitioning, and degradation pathways are key processes in the probability of a substance to interact with target organisms. Biotic factors such as toxicokinetics, biotransformation capacity, and behavioral and life-cycle aspects of the organisms are determinants for final concentrations at target organs. The role of metabolites in endocrine disruption can be quite different from those of the parent compounds, and often this requires separate toxicokinetic evaluation. The exposure assessment of endocrine active substances (EASs) suffers from a huge lack of reliable data, of both values that are used as input parameters in exposure models, and field data that are needed for validation purposes. In general, for the more classic EASs, such as PCBs, p,p '-DDE, chlorinated dioxins, some pesticides, and organotins, reliable data are sufficiently available, but careful evaluation of the quality of databases is necessary. Several data quality evaluation systems have been proposed. For the "newer" compounds, only few data have been gathered so far. The latter compounds include alkylphenols, bisphenol A, brominated flame retardants, phytoestrogens, and in particular natural and synthetic hormones, which in view of their high estrogenic potency could be the most important compounds in terms of risk. The suitability of current exposure assessment models for EASs at this moment seems to be restricted to the persistent compounds such as PCBs, PCDDs, and PCDFs. Especially for the compounds subject to biodegradation and biotransformation, the lack of experimental data to derive model-input parameters and perform validation studies at this moment is one of the main obstacles for the further application of generic exposure models to other EASs. Most of the current models do not allow life stage-specific predictions. Although the mechanisms of endocrine disruption involve different types of action, the principle of additivity, based on the equivalent toxicity approach (using estrogen equivalent potencies relative to 17β-estradiol) seems promising for the design of integrated exposure and effect models for EASs. Research programs aimed at the endocrine disruption issue must focus on promoting experimental studies for generation of reliable, high-quality parameter data on the one hand, and surveys or monitoring campaigns for collection of representative field data on the other. The non-specificity of possible effects caused through endocrine mechanisms implies that in order to reveal dose-response relationships all potentially active agents, or at least as many as feasible, need be included in the risk assessments. Current regulatory monitoring programs should further be evaluated and harmonized with validation requirements of models used in exposure assessment.

This main topic of the project and symposium includes the issue of releases of EASs and their monitoring in the environment, food, and feed in order to provide the full set of criteria, relevant for exposure assessment. Much less research has been devoted to these areas as compared to investigations on the effects. Issues of special importance regarding exposure to EASs, both from a research and risk management point of view, predominantly result from the fact that high-potency natural products are released as well as anthropogenic substances. In order to provide reliable information for risk assessment and risk management, substantial research, methodological improvements, and improvements in data interpretation are needed regarding the following: releases and technologies for their mitigation; monitoring; establishment of background levels; transport, partitioning, persistence, degradation, and metabolism of EASs; dealing with "joint toxicity"; and providing reliable analytical methodology conforming to the principles of quality assurance. Phytohormones play a special role in this whole area since they may be used as food amendments.

Environmental contaminants can potentially disrupt endocrine processes by interfering with the function of enzymes involved in steroid synthesis and metabolism. Such interferences may result in reproductive problems, cancers, and toxicities related to (sexual) differentiation, growth, and development. Various known or suspected endocrine disrupters interfere with steroidogenic enzymes. Particular attention has been given to aromatase, the enzyme responsible for the conversion of androgens to estrogens. Studies of the potential for xenobiotics to interfere with steroidogenic enzymes have often involved microsomal fractions of steroidogenic tissues from animals exposed in vivo, or in vitro exposures of steroidogenic cells in primary culture. Increasingly, immortalized cell lines, such as the H295R human adrenocortical carcinoma cell line are used in the screening of effects of chemicals on steroid synthesis and metabolism. Such bioassay systems are expected to play an increasingly important role in the screening of complex environmental mixtures and individual contaminants for potential interference with steroidogenic enzymes. However, given the complexities in the steroid synthesis pathways and the biological activities of the hormones, together with the unknown biokinetic properties of these complex mixtures, extrapolation of in vitro effects to in vivo toxicities will not be straightforward and will require further, often in vivo, investigations.

A number of epidemiologic studies on organochlorines (OCs) and breast cancer risk have been published. The majority ( n = 18) measured OCs in adipose tissue, primarily from mammary biopsies in cancer cases and controls with benign breast disease, and studied incident disease. Seven of these studies each included fewer than 50 cases and controls and had limited capacity for covariate adjustment. Eleven studies used serum samples collected from 6 months to 25 years prior to diagnosis. An additional 13 studies (2 with some overlap) used serum collected at or after diagnosis. Regardless of the medium used to measure OC levels, studies conducted to date do not provide consistent evidence that any of the OCs examined thus far play a role in the initial breast cancer risk. This paper provides a compound specific review with discussion of how the lack of evidence for adverse effects might be explained by factors related to study design, or by variation in risk across subgroups. The sum of the evidence does not implicate any OC compound as significantly related to risk of occurrence of breast cancer. The evidence base is greatest for dichlorodiphenyltrichloroethane (DDT) and dichlorodiphenyldichloroethylene (DDE). Limited research has been done on individual polychlorinated biphenyls (PCBs) and their isomers. The studies of OC exposure reflect current exposure levels of chemicals banned as long ago as 20 years. Although the information is extremely limited and not without major design flaws, the association between OCs and disease severity and progression is interesting and worthy of further examination. More studies are needed on OCs other than DDTs in developing countries where use is more recent or continuing, especially given that most estrogenic OCs are not persistent. It is possible that other pathways may be involved, including activity related to cytochrome P450 (CYP) and glutathione- S -transferase (GST); there is limited research to date on this hypothesis. In developed countries, a body mass index (BMI)/weight loss model may warrant further analysis, perhaps using existing data. OC exposures cannot reliably be related to trends in breast cancer incidence, as other known risk factors for breast cancer, such as childbearing and lactation, have changed along with OC residues during this timeframe. Changes in screening and treatment over time also complicate making such links. Ecologic data relating high-exposure countries to high breast cancer mortality rates do not suggest a strong link. Breast cancer rates are not generally higher in parts of the world with high DDE levels. Countries with relatively similar levels of PCBs, such as Great Britain and Japan, have very different breast cancer rates. As OCs are present in the environment as mixtures of correlated isomers and metabolites, it may be difficult to distinguish possible causal links from associations in which measured compounds are merely markers of other underlying exposures. For highly correlated compounds, traditional adjustment strategies may not be feasible. More complex analytical strategies may help to isolate potentially relevant isomers.

Congenital adrenal hyperplasia (CAH) refers to a family of monogenic inherited disorders of adrenal steroidogenesis most often caused by a deficiency of the 21-hydroxylase enzyme. In the classic forms of CAH (simple virilizing and salt-wasting), androgen excess causes external genital ambiguity in newborn females and progressive postnatal virilization in males and females. Prenatal treatment of CAH with dexamethasone has been successfully utilized for over a decade. This article reports on 595 pregnancies prenatally diagnosed using amniocentesis or chorionic villus sampling between 1978 and 2002 at the New York Presbyterian Hospital-Weill Medical College of Cornell University. No significant or enduring side effects were noted in the fetuses, indicating that dexamethasone treatment is safe. Prenatally treated newborns did not differ in weight from untreated, unaffected newborns. Based on our experience, prenatal diagnosis and treatment of 21-hydroxylase deficiency is effective in significantly reducing or eliminating virilization in the newborn female. Prevention of genital virilization in female newborns with classic CAH avoids the risk of sex misassignment and diminishes the need for corrective surgery and the resulting psychological impact that may extend into adulthood.

This review addresses whether there is a secular increasing trend in male reproductive developmental disorders (cryptorchidism, hypospadias, testis cancer, low sperm counts), and highlights the limitations of available data and how these issues are being addressed. These disorders are considered to represent a syndrome of disorders [testicular dysgenesis syndrome (TDS)] with a common origin in fetal life, and in which "endocrine disruption" plays a central role. The potential involvement of environmental estrogens in the etiology of these disorders is reviewed in light of new understanding about the pathways and dose-effect relationships of estrogen action on male reproductive development. Several new pathways of estrogen action have been identified, including suppression of the production of testosterone and insulin-like factor-3 by fetal/neonatal Leydig cells and suppression of androgen receptor expression in androgen target tissues. It is tentatively concluded that identified environmental chemicals are unlikely to activate these pathways because of their intrinsically weak estrogenicity. However, chemicals that may alter endogenous estrogen production, bioavailability, or inactivation represent a new focus of concern. Additionally, environmental chemicals that alter endogenous levels of androgens in the rat fetus (certain phthalates) induce a similar collection of disorders to TDS. Whether human exposure to such compounds might contribute to TDS remains to be shown, but studies in animals should help to define susceptible pathways for induction of TDS.

From an environmental point of view, an increasing important group of organohalogen compounds are the brominated flame retardants (BFRs), which are widely used in polymers and textiles and applied in construction materials, furniture, and electronic equipment. BFRs with the highest production volume are the polybrominated diphenyl ethers (PBDEs), tetrabromobisphenol A (TBBP-A), and hexabromocyclododecane (HBCD). Because of their persistence and low biodegradation profile, several of the PBDE congeners accumulate in biota and are widely found in the aquatic food chain. Their levels in the environment and in humans have increased during the last decades, in contrast to compounds such as polychlorinated biphenyls (PCBs) and dichlorodiphenyltrichloroethane (DDT), for example. Humans may be exposed to PBDEs mainly through consumption of fatty food of animal origin (e.g., fish), but exposure through skin contact with textiles protected with flame retardants or through inhalation of BFRs volatilized from electronic and electric equipment may also occur. The levels of PBDEs in Swedish human milk showed a doubling in concentration every five years over the period 1972 to 1997 (2,2',4,4'-tetraBDE being the predominant congener). The levels of penta- and hexa-BDEs increased at the same rate in ringed seals collected in the Canadian Arctic from 1981 to 2000. PBDEs exhibit a great variety of biological effects, depending on the bromine substitution pattern. PBDEs are potential endocrine disrupters, based on shared toxicity with the structurally related PCBs, polychlorinated dibenzofurans (PCDFs), and polychlorinated dibenzo- p -dioxins (PCDDs) (partial aryl hydrocarbon- [Ah-] receptor agonist and antagonist activity in vitro, thyroid toxicity, and immune effects), including developmental toxicity. The potency of TBBP-A to interact with thyroid hormone homeostasis is indicated from in vitro studies in which the compound competes with thyroxin (T4) for binding to transthyretin (TTR). So far, the toxicological profile of many BFRs is too incomplete and insufficient to perform an adequate risk assessment, and further information is required regarding the potential for endocrine disruption of these compounds that are of increasing environmental concern.

Phytoestrogens are nonsteroidal plant-derived compounds possessing estrogenic activity. These include two major classes: isoflavonoids and lignans. Phytoestrogens have received recently great attention because of their beneficial effects, which include the prevention of cancer, atherosclerosis, and bone density loss. However, they have estrogenic activity and may affect as endocrine disruptors. In this review, we pay attention to both the toxic and beneficial effects of phytoestrogens. Epidemiological data support that isoflavonoids help prevent cancer of breast, prostate, stomach, and lung. However, there have been some reports about a positive association between some phytoestrogens (e.g., campesterol and stigmasterol) and prostate cancer risk. Animal experiments and in vitro experiments have shown that the biological effects of phytoestrogens may be organ-specific, inhibiting cancer development in some sites, yet showing no effect or an enhancing effect on tumorigenesis at other sites. Also, their effects may be dependent on the timing and duration of exposure. For example, several studies in rodents have established that the favorable effect of an isoflavone-rich diet on breast cancer risk may be significant only if consumption occurs before puberty or during adolescence. On the other hand, prenatal exposure of genistein was reported to have a carcinogenic effect on the uterus of rodents. Phytoestrogen, especially soy products, has been used as hormone-replacement therapy, reducing such symptoms as hot flashes, vaginal dryness, and mood changes while protecting women from osteoporosis and heart disease. However, the animal data suggest that the timing of exposure to such compounds is crucial, with neonatal exposure having the most pronounced effects. Given the exposure of neonates to phytoestrogens, this should be a cause for concern.

Although known to regulate growth and development, cellular metabolism, the use of oxygen, and basal metabolic rate, thyroid hormones have been only minimally evaluated in neonatal rodents at critical times of development. Despite some modulation of metabolic rate by other hormones, such as testosterone, growth hormone, and norepinephrine, 3,5,3'-triiodothyronine (T3) and 3,5,3',5'-tetraiodothyronine (T4) are the most important metabolic rate modulators. Endpoints used for thyroid function assessment in neonatal and adult rats include thyroid-stimulating hormone (TSH), T3, and T4 levels and histopathology. In rodents, decreased serum levels of T3 and T4 and increased serum TSH levels, with sustained release of TSH and resultant follicular cell hypertrophy/hyperplasia, are typical hormonal and histopathological findings attributable to compounds altering thyroid function. Hypothyroidism early in the neonatal period can affect reproductive endpoints in both male and female rats, with the critical period of exposure being the first two weeks postnatal. Hypothyroidism has been shown to reduce gonadotrophin levels and delay pubertal spermatogenesis in male rats and to block gonadotropin-induced first ovulation in immature female rats by decreasing FSH and luteinizing hormone (LH) serum concentrations. Inclusion of evaluations of TSH, T3, and T4 assays in multigeneration and developmental neurotoxicity protocols may assist in risk assessments.

Wildlife and humans are exposed to a complex mixture of endocrine active chemicals. The activity of a specific chemical in any mixture can be modified through interactions with other components of the mixture. The toxic equivalency factor (TEF) approach for risk assessment was developed for chemicals such as halogenated aromatics that induce their effects through ligand-activated receptors. For persistent halogenated aromatic AhR agonists, this approach has some utility. However, the use of the TEF approach for endocrine active compounds is confounded by the unique tissue- and response-specific activities of these structurally diverse compounds. The term selective receptor modulator describes the ability of a natural or synthetic receptor ligand to manifest agonist activity in one tissue or for one response and antagonist activity in other tissues or for another response in the same tissue. Thus, it is possible for chemicals in a mixture to behave in an additive manner for one response and an antagonist manner for another response. A mechanisms-based hazard risk assessment of endocrine active chemical mixtures must account for these multiple variables.

Concerns about the effects of endocrine disruptors on humans and wildlife have resulted in revised governmental testing guidelines (e.g., U.S. Environmental Protection Agency, Organization for Economic Cooperation and Development, U.S. Food and Drug Administration), adding endpoints to enhance their capability to detect endocrine active compounds. Based on experience with these testing guidelines, I present my opinions and data on study design, performance, results, endpoints, interpretation, and recommendations for improvement. New regulatory endpoints must be reproducible, robust, sensitive, relevant, and consistent. These new endpoints are appropriate: anogenital distance, examination of culled pups on postnatal day 4, examination for retained nipples/areolae in preweanling males, weanling necropsy, acquisition of puberty in offspring, prebreed estrous cyclicity, reproductive organ weights and histopathology, and andrology. Endpoints considered not appropriate are ovarian primordial follicle counts, stage of estrous for parental females at demise, and single blood sample at necropsy to measure circulating levels of relevant hormones. In conclusion: (1) regulatory guidelines represent only the minimum requirements and should/will change as science improves and scientific and societal concerns arise; (2) scientists must interact to identify and refine new methods, endpoints, and mechanisms; (3) the objective of these studies is to provide good science and useful information for risk assessment. Regulatory agencies should convene workshops with stakeholders to discuss issues such as new fundamental/applied information, mechanisms, and current and new endpoints. Interactions will lead to enhanced science, periodic revisions of testing guidelines, and improved risk assessment.

The last 40 years have seen many reports that man-made chemicals and environmental pollutants cause adverse effects in humans and wildlife; however, actually linking an exposure with a mechanism and an effect has yet to be done for endocrine disruption. Certainly, studies in experimental animals have shown that sufficient doses of select compounds can disrupt the endocrine system and produce the attendant adverse outcomes. The purpose of this contribution is to evaluate some of the recent reports of the adverse effects on reproduction and development, the immune system, and the nervous system that have been observed in experimental animals after treatment with man-made chemicals and environmental pollutants. Space limitations prevent us from presenting a comprehensive review of all reported endocrine active chemicals and their effects. Instead, we have focused on drawing conclusions as to the scope and etiology of the adverse effects in experimental animals using examples from the scientific literature, and on suggesting a path forward for further work.

The U.S. Environmental Protection Agency (USEPA) is developing an endocrine disruptor screening and testing program to detect chemicals that alter hypothalamic-pituitary-gonadal (HPG) function, estrogen, androgen, and thyroid (EAT) hormone synthesis or metabolism and induce androgen (AR) and estrogen (ER) receptor-mediated effects in mammals and other animals. The utility of this approach is based upon the knowledge that mechanisms of endocrine-disrupting chemical (EDC) action are highly conserved at the cellular and molecular levels among vertebrates. Some EDC mechanisms also are shared with invertebrates. High-priority chemicals would be evaluated in a Tier 1 screening (T1S) battery, and chemicals that are positive in T1S would then be tested in Tier 2 (T2). T1S includes in vitro ER and AR receptor binding and/or gene expression, an assessment of steroidogenesis and mammalian (rat) and nonmammalian (fish) in vivo assays. In vivo, the uterotropic assay detects estrogens and antiestrogens, while steroidogenesis, antithyroid activity, antiestrogenicity, and HPG function are assessed in a pubertal female assay. Antiandrogens are detected in the Hershberger assay (weight of androgen-dependent tissues in castrate-immature-male rats). Fish and amphibian assays are also being developed to identify EDCs. Several alternative mammalian in vivo assays have been proposed. Of these, a short-term pubertal male rat assay appears most promising. T1S is designed to be sensitive to EAT activities, but many of the effects detected at the screening level would not be considered adverse, the dosage levels may be high, and the route of administration used may not be the most relevant. However, issues of adversity, dose response, and route(s) of exposure would be resolved in the testing phase. In addition to using an enhanced multigenerational test for Tier 2, an in utero-lactational screening protocol is also being evaluated by USEPA for use in T2 or T1S. For T2, the numbers of endocrine-sensitive end-points and offspring (F1) examined in multigenerational tests need to be expanded for EDCs in a thoughtful manner, based in part upon the results of T1S. In addition, for some chemicals histological examination of 10 adult F1 per sex in only the control and high-dose groups provides inadequate statistical power to detect low-dose lesions induced during development. In these cases, we propose that all the offspring be examined after puberty for gross and histological reproductive abnormalities. Since EDCs, like the phthalates and AR-antagonists, produce characteristic profiles, or syndromes, of adverse effects, data need to be reported in a manner that clearly identifies the proportion of animals displaying one or more of the abnormalities in a syndrome. Consideration should be given to tailoring T2, based on the results of T1S to assure that all of the effects in such chemically induced developmental syndromes are included in the study.

Regulatory authorities and other scientific organizations around the world have developed hazard/risk assessment practices that involve the derivation of numerical values said to represent "acceptable" or "safe" levels of human exposure to individual chemical substances. Human data would be preferable for this purpose since interspecies extrapolation would not be necessary. In most cases, however, these data are not available or are inadequate for this purpose. Therefore, results from studies conducted in non-human mammalian species are used as the principal or sole basis for hazard/risk assessment. Existing risk assessment frameworks developed for the evaluation of endpoints of toxicity for chemical substances are sufficiently flexible to incorporate knowledge concerning the mode(s)/mechanism(s) of action by which these endpoints occur, including those involving disturbance of normal endocrine status ("endocrine disruption"). Background on nomenclature, traditional practices using non-human animal models, and the nature and adequacy of data sets for deriving acceptable human exposures for chemicals, including endocrine active substances, are described, as is how/why the existing frameworks are adequate for the purpose of deriving numerical values for endocrine active substances. Projections on how assessment practices for these substances may evolve further in the future also are offered.

The benchmark dose (BMD) is the dose of a substance that is expected to result in a prespecified level of effect, the benchmark response level or BMR. It is a general approach to characterizing dose response, applicable to any toxicant and endpoint. A BMD is conceptually superior to a "no observed adverse effect level" (NOAEL) for this purpose because of being less determined by experimental design, because it is a precisely defined entity, and because its precision can be estimated. Since a BMD is a single number, just as an NOAEL, it is tempting to use the BMD as a straightforward replacement for the NOAEL in the assessment process for calculating allowable daily intakes. However, the level of toxic response at a NOAEL is unknown, while that at a BMD is well defined. Use of the BMD approach potentially adds consistency and objectivity to the process of deriving reference values (RfDs, RfCs, or ADIs) for setting regulatory levels. To take advantage of this, BMRs need to be selected in a consistent way across studies and endpoints. This paper discusses some issues affecting the selection of BMRs, and presents an example of a BMD calculated for the effects of peripubertal exposure to the fungicide vinclozolin.

Dose-response characteristics for endocrine disruption have been major focuses in efforts to understand potential impacts on human and ecological health. Issues include assumptions of thresholds for developmental effects, effects at low doses with nonmonotonic (e.g., "U-shaped") behaviors, population vs. individual responses, and background exposures (e.g., dietary phytoestrogens). Dose-response analysis presents a challenge because it is multidisciplinary, involving biologists and mathematicians. Statistical analyses can be valuable for evaluating issues such as the reproducibility of data as illustrated for contradictory findings on low-dose effects. Mechanistically based modeling provides insights into how perturbations of biological systems by endocrine active substances can create different dose-response behaviors. These analyses have demonstrated that higher order behaviors resulting from the interaction of component parts may appear highly nonlinear, thresholded, low-dose linear, or nonmonotonic, or exhibit hysteresis. Some effects need to be evaluated as population impacts. For example, alterations in male:female ratio may be important at the population level even though not adverse for the individual. Descriptions of the contributions of background exposures to dose-response behaviors are essential. The challenge for improving dose-response analyses is to better understand how system characteristics create different dose-response behaviors. Such generalizations could then provide useful guidance for developing risk assessment approaches.

The meaning of the term "low dose" is discussed in relation to endocrine toxicity data for chemicals. Consideration is also given to experimental conditions likely to impinge on the interpretation and extrapolation of such low-dose effects, and the importance of gathering appropriate control data is emphasized. In the specific case of bisphenol A (BPA), it is concluded that despite the extensive endocrine disruptor (ED) database available for this chemical, it is still not possible to locate a single study that passes the most rudimentary scientific requirements-that the observations are capable of independent confirmation. Two possible explanations for this are considered. First, that BPA possesses subtle low-dose ED toxicities that only become evident under certain undefined experimental conditions. Until these conditions are defined and understood, it will be a matter of chance what individual investigators observe experimentally for BPA or any other chemical. Second, that the general failure of investigators to define and understand natural variability among control parameters monitored in ED studies allows artefactual positive results to be encountered for chemicals, especially in limited and nonreproduced studies. Whichever of these conclusions is correct, the positive low-dose data currently available for BPA cannot be extrapolated to humans with any confidence.

The term "environmental estrogen" refers to chemical substances that exhibit some degree of estrogen-like activity. The primary emphasis for potential adverse effects resulting from exposure to environmental estrogens is on in utero exposure because such exposure can occur during critical periods of organogenesis. Assessment of biological plausibility can be based, in part, on the extensive data on the effects of diethylstilbestrol (DES). The available evidence is too limited to judge with any confidence whether sperm counts have declined during the past 50 years. Based on both animal and human data with DES, it is biologically plausible that in utero exposures to exogenous estrogenic compounds are capable of reducing sperm production in adult men. However, the apparent existence of a maternal dose threshold for DES-induced effects on sperm counts undermines the likelihood that environmental estrogens, which are substantially less potent, are capable of causing similar effects.

Endocrine disruption (ED) as a named field of research has been very active for over 10 years, but effects in wildlife that would now be labeled as ED have been studied since the 1940s. This paper briefly surveys the progress in wildlife studies that has been made to date and draws out the major themes and issues that have been identified. In particular, it discusses information concerning causative substances, modes of action, ubiquity of effects across taxa, individual- and population-level impacts, and the importance of low-dose and mixture effects. The main conclusion is that while most wildlife taxa are showing some ED effects at some locations, good evidence for population-level impacts is still limited to a few groups. In order to improve both the interpretation of field observations and the way in which environmental risk assessments are conducted, we need to develop an enhanced ability to predict effects on populations and communities from a knowledge of effects on individuals.

Recent reports have shown that a number of xenobiotics in the environment are capable of interfering with the normal endocrine function in a variety of animals. The overwhelming majority of the studies on the effects of hormone-mimetic industrial chemicals were focused on findings in vertebrates. More detailed information about the effects on and mechanisms of action in invertebrates has only been obtained from a few cases, although invertebrates represent more than 95 % of the known species in the animal kingdom and are extremely important with regard to ecosystem structure and function. The limited number of examples for endocrine disruption (ED) in invertebrates is partially due to the fact that their hormonal systems are rather poorly understood in comparison with vertebrates. Deleterious endocrine changes following an exposure to certain compounds may easily be missed or simply be unmeasurable at present, even though a number of studies show that endocrine disruption has probably occurred. The well-documented case studies of tributyltin effects in mollusks and of insect growth regulators, the latter as purposely synthesized endocrine disruptors, are explained to support this view. According to our present knowledge, there is no reason to suppose that such far-reaching changes are in any sense unique. The additional existing evidence for ED in invertebrates from laboratory and field studies are summarized as an update and amendment of the EDIETA report from 1998. Finally, conclusions about the scale and implications of the observed effects are drawn and further research needs are defined.

Endocrine disruption has been reported in freshwater fish populations around the world. This phenomenon ranges from subtle changes in the physiology and sexual behavior of fish to permanently altered sexual differentiation and impairment of fertility. Despite widespread reports of endocrine disruption in fish (and this is well characterized at the individual level), few studies have demonstrated population-level consequences as a result of exposure to endocrine-disrupting chemicals (EDCs). An exception to this is in Lake Ontario Lake trout where precipitous declines in the population have been linked with periods of high exposure to organochlorine chemicals (known EDCs). Recently, it has been established that roach ( Rutilus rutilus ) exposed to treated sewage effluent (that contains complex mixtures of EDCs) in UK rivers, have a reduced reproductive capacity. This, in turn, may have population-level consequences. Evidence for a link between exposure to effluents from kraft mill (BKME) and sewage treatment works (STWs) and altered reproductive function in freshwater fish is compelling. In most cases, however, a causal link between a specific chemical and a physiological effect has not been established. Indeed, identifying specific chemical(s) responsible for adverse effects observed in the wild is difficult, given that tens of thousands of man-made chemicals enter the aquatic environment and that mixtures of chemicals can have combination (e.g., additive) effects. Some EDCs are known to act at a number of different body targets to affect a variety of physiological processes, further complicating the identification of the causative agent(s). Endocrine disruption appears to be particularly widespread in freshwater fish populations. There is little evidence, however, to suggest fish are more susceptible to EDCs relative to other wildlife. Notwithstanding this, there are some features of the endocrine physiology of fish that may be particularly susceptible to the effects of EDCs, including the processes of sex-determination and smoltification (in salmonids). Furthermore, their aquatic existence means that fish can be bathed constantly in a solution containing pollutants. In addition, uptake of chemicals readily occurs via the gills and skin, as well as via the diet (the major exposure route for most EDCs in terrestrial animals). The exposure of fish early life stages to the cocktail of EDCs present in some aquatic environments may be of particular concern, given that this is an especially vulnerable period in their development. The challenge, from the point of view of ecological risk assessment, is to determine effects of EDCs on freshwater fish populations and freshwater ecosystems. In order to meet this challenge, high-quality data are required on the population biology of freshwater fish, on the effects of EDCs on their various life history characteristics, and comprehensive and appropriate population models. Basic information on the population biology of most species of wild freshwater fish is, however, extremely limited, and needs significant improvement for use in deriving a sound understanding of how EDCs affect fish population sustainability. Notwithstanding this, we need to start to undertake possible/probable predictions of population level effects of EDCs using data derived from the effects found in individual fish. Furthermore, information on the geographical extent of endocrine disruption in freshwater fish is vital for understanding the impact of EDCs in fish populations. This can be derived using published statistical associations between endocrine disruption in individual fish and pollutant concentration in receiving waters. Simplistic population models, based on the effects of EDCs on the reproductive success of individual fish can also used to model the likely population responses to EDCs. Wherever there is sufficient evidence for endocrine disruption in freshwater fish and the need for remediation has been established, then there is a need to focus on how these problems can be alleviated. Where industrial chemicals are identified as causative agents, a practical program of tighter regulation for their discharge and/or a switch to alternative chemicals (which do not act as EDCs) is needed. There are recent examples where such strategies have been adopted, and these have been successful in reducing the impacts of EDCs from point source discharges on freshwater fish. Where EDCs are of natural origin (e.g., sex steroid hormones from human and animal waste), however, remediation is a more difficult task. Regulation of the release of these chemicals can probably be achieved only by improvements in treatment processes and/or the implementation of systems that specifically remove and degrade them before their discharge into the aquatic environment.

In the last few decades, various studies have shown that aquatic mammals are sensitive to the toxicological effects of certain xenobiotic compounds, including the large class of endocrine-disrupting chemicals (EDCs). Since some EDCs, particularly organochlorines, tend to bioaccumulate and biomagnify in the aquatic food chain, various aquatic mammals, particularly those high in the food chain, such as pinnipeds, odontocete cetaceans, and polar bears, are potentially "at risk". The main aim of this chapter is to define the state of the art on effects of endocrine disruptors in aquatic mammals, both freshwater and marine. Another aim is to formulate recommendations for future research in this field and finally to define what can be done internationally for hazard/risk assessment and communication of the findings.

This topic reviews the whole field of endocrine disruption (ED) in marine fish and compares this with our knowledge of the situation in freshwater species. In broad terms, similar types of ED have been observed in the two groups, although effects in the marine environment tend to be less marked, presumably due to dispersion and dilution. There are, however, some data which suggest that marine fish that are top-predators can experience ED due to biomagnification of organochlorines. Processes such as smoltification, metamorphosis, and hermaphroditism, which are common in some marine species, may be particularly susceptible to ED, but have as yet been scarcely studied. As with freshwater fish, firm links to population-level effects have not yet been demonstrated, although it is not unreasonable to suppose that they are occurring in some locations. The topic concludes with some recommendations for future research.

Since the early 1990s, a substantial number of deformed frogs have been observed in North America, particularly in the upper Midwest and Canada. Attempts to understand the etiology of the deformed frog problem have met with limited success to date with nearly as many proposed explanations as research groups working on the problem. Models for the mechanism underlying the development of deformed frogs include parasite/predation, ultraviolet radiation, and chemical exposure. Each model has its strengths and weaknesses. Despite contentious debate among researchers, there is an overall consensus that the increasing prevalence of deformed frogs is the result of a water-borne contaminant that has recently appeared, or reached a critical concentration. Our detailed analysis of malformed frogs collected in Minnesota ponds and lakes suggested that limb patterning was being modified by the disruption of a retinoid-sensitive developmental signaling pathway. Accordingly, we focused in the identification and characterization of bioactive retinoids from lake water and showed that retinoid treatment of frog embryos at sensitive times of development could recapitulate the full spectrum of limb abnormalities observed in field specimens in the laboratory. These data have led to the conclusion that inappropriate modulation of retinoid signaling by environmental contaminants is the mechanism underlying the increased incidence of frog malformations.

Many chemicals introduced into the environment by humans adversely affect embryonic development and the functioning of the vertebrate reproductive system. It has been hypothesized that many developmental alterations are due to the endocrine-disruptive effects of various environmental contaminants. The endocrine system exhibits an organizational effect on the developing embryo, altering gene expression and dosing. Thus, a disruption of the normal hormonal signals can permanently modify the organization and future functioning of the reproductive and endocrine system. We have worked extensively with contaminant-exposed and reference populations of the American alligator ( Alligator mississippiensis ) as well as performed a number of experimental studies exposing developing embryos to various persistent and nonpersistent pesticides. Using this species, we have described altered steroidogenesis, circulating hormone levels, and hepatic transformation of androgen and endocrine organ (gonad, thyroid) morphology in juvenile alligators living in polluted environments. Given the adverse observations reported to date, we recommend several important future needs: Further development of "receptor zoos" and other molecular tools that include key reptiles from various major ecosystems, in addition to freshwater ecosystems. Global studies extending the current knowledge base on crocodilians and freshwater turtles to comparable ecosystems on other continents, such as linked studies examining and extending current molecular to population level studies in Florida (USA) to tropical and temporate regions of Africa, Australia, and South America. Further studies of actual exposure, assimilation and excretion of contaminants by ectothermic vertebrates, especially reptiles that occupy high levels of the food chain.

There have been several case studies of the impact of chemical contaminants on birds at the level of individuals or populations. While many of the chemicals involved in these incidents have been classified as endocrine-disrupting chemicals or endocrine active substances (EASs) the mechanisms by which these chemicals affect birds are not clearly or fully understood.

This paper will review briefly the use of wildlife as models in the study of how mixtures, low doses, and the embryonic environment modulate the action of endocrine active substances (EASs). In so doing it will show how the issue of low dosages must be considered within the context of mixtures present in the environment and the endocrine background of the exposed individual. That is, in nature, EASs usually are found in mixtures in which the constituent parts are in concentrations well below their no observed adverse effect level (NOAEL) as determined in single compound studies in the laboratory. In addition, exposure always occurs on organisms in various endocrine states. Thus, the issue of mixtures and dosages must always be considered within the context of the endocrine background. Finally, the effects of exposure are passed down through the generations. The question of exposure then at the level of the individual becomes very complicated, as it must take into account that at every life stage, the naturally occurring endocrine milieu of the organism (or tissue), any EAS burden inherited from the mother or built up over the individual's life, and the social environment in which the individual develops and interacts as an adult, will influence the response to acute exposure.

The extent to which nonreproductive aspects of the endocrine system are affected by environmental contaminants is to a large extent unknown. However, an emerging body of data demonstrates that the neuroendocrine stress response is a sensitive target for disruption by a range of environmental contaminants, at a number of discrete loci. Several mechanisms are responsible for generating and sustaining the corticosteroid response to a stressor, including synthesis of the steroid, negative feedback at the pituitary and hypothalamus, and clearance via metabolism and conjugation in peripheral tissues and the liver. Laboratory and field studies provide evidence that these elements of the stress response are susceptible to interference by EAS. The functional significance to the individual of interference with this important adaptive mechanism remains to be established.

Responses to endocrine active substances (EASs) in animals are various, and differences between the responses among individuals, populations and species are well known. These differences are observed not only in EASs but in most environmental chemicals including synthetic and naturally occurring ones. The basic differences in sensitivity to EASs are attributed to that of affinity or specificity of the receptors to EASs at the cellular level. Although the nucleotide sequences encoding for estrogen receptor proteins have been documented in several species and the functions of the receptors are the same, the ability to bind the natural hormones and the estrogenic xenobiotics is not necessarily identical. The reproductive endocrine system is basically common among vertebrates, but chemical types of hormones, physiological roles of hormones and the basal blood levels of hormones differ among each species, especially in sex steroids. These differences cause various types of responses and sensitivity to EASs among animal species. Xenobiotic metabolism is important for the genetical, biochemical and physiological factors concerning the influence of EASs. Some EASs directly inhibit cytochrome P450 (CYP) activity as was reported in tributyltin that inhibits CYP19 (aromatase) activity causing imposex in neogastropods. Some organochlorines including dioxins stimulate aryl hydrocarbon (Ah) receptor-mediated xenobiotic metabolism, and result in the metabolic disruption of steroid hormones such as estrogen as were reported in eggshell thinning in birds of prey and uterus occlusion in seals. CYP activity greatly differs among wildlife species in both terrestrial and aquatic organisms, and these differences are significantly responsible for the multiple effects or toxicity of EASs. Sex and age differences also cause different responses to EASs and are largely due to the differences in xenobiotic metabolizing activities.

International concern over endocrine active substances (EASs) has led to intensive research programmes to establish fish reproductive and developmental toxicity tests for use in environmental (ecological) risk assessment. This chapter gives an overview of key themes of in vivo ecotoxicology research, including fish screening assays, partial life-cycle tests (the draft OECD fish reproduction test and the new fish development test) and fish full life-cycle tests. In the context of the OECD test guidelines program, fish species of primary interest include fathead minnow, medaka and zebrafish, while guppy, rainbow trout, sheepshead minnow, and three-spined stickleback are also of scientific importance. Critical factors for evaluation include baseline reproductive biology and definition of EAS sensitive life-stages. For regulatory applications, a critical review of existing fish EAS data suggests that apical adverse effect endpoints, namely development, growth and reproduction (e.g., fecundity, fertilization rates, and hatching success) should be used to derive predicted no-effect concentrations (PNECs) for the environmental risk assessment of EASs. In support of these apical adverse effect endpoints, biomarker responses (e.g., vitellogenin, gonadal-somatic index, and gonad histopathology) should be used to provide mechanistic data, compare species (e.g., cyprinids vs. salmonids), and allow extrapolation between laboratory and field studies.

Probably the only thing that can be said with certainty about the future of this field of ecotoxicology is that predicting it is foolish; the chances of being right are very slim. Instead, it seems to me likely that unexpected discoveries will probably have more influence on the field of endocrine disruption than the outcomes of all the planned experiments. It is certainly true that chance discoveries, such as masculinized fish in rivers receiving paper-mill effluent, imposex in molluscs due to exposure to tributyltin and feminized fish in rivers receiving effluent from sewage-treatment works, have been pivotal in the development of the field of endocrine disruption in wildlife. I consider that further such discoveries are likely, but I do not know which species will be affected, what effects will be found, what chemical(s) will be the cause, or what endocrine mechanism(s) will underlie the effects. The recent realization that many pharmaceuticals are present in the aquatic environment only underscores the range of effects that could, in theory at least, occur in exposed wildlife. What is somewhat easier to predict is the research that will be conducted in the immediate future, which will build upon what is known already. For example, it is clear that wildlife is rarely, if ever, exposed to single chemicals, but instead is exposed to highly complex, ill-defined mixtures of chemicals, including many that are endocrine active in various ways. We need to understand much better how chemicals interact, and what overall effects will occur upon exposure to such mixtures. We also need to move from assessing effects at the individual organism level, to understanding the consequences of these effects at the population level. Then, we need to determine the significance of any population-level effects due to endocrine disruption in comparison with the impact of many other significant stressors (e.g., over-exploitation, habitat loss, climate change) that also negatively impact wildlife. Such research will be difficult, and time-consuming, and will probably produce many surprises. All I can be fairly certain about is that the next few years are likely to be as interesting and exciting as the last few have been.

Mechanism-based QSAR models for interactions between various ligands and the estrogenic receptor are developed by using well-developed physicochemical parameters. Common features of these QSARs are identified, and deficiencies in some datasets are highlighted. The relative binding affinities of various substituted hexestrols to estrogen receptors are examined in terms of their steric, electronic, and hydrophobic attributes. Different QSARs for hexestrols and tamoxifens reveal that steric effects are of overriding importance in variations of binding affinity. In the few cases where a large number of diverse substituents are located on aromatic rings, electronic effects emerge and suggest that electron-donating groups enhance binding to the receptor while hydrophobicity plays a marginal role in decreasing binding affinity. With substituted phenols bearing alkyl-type substituents and substituted hydroxy-biphenyls, the binding is strictly dependent on hydrophobicity and size. These QSAR models are described in detail and examined together to illustrate the utility of lateral validation in mechanistic interpretation.

Some seven years have passed since the U.S. legislature mandated the EPA to develop and implement a screening and testing program for chemicals that may disrupt the delicate endocrine system. The envisioned EPA program has evolved to incorporate a tiered scheme of in vitro and in vivo assays, and considered QSAR as a viable method to set testing priorities. At the U.S. FDA's National Center for Toxicological Research (NCTR), the Endocrine Disruptor Knowledge Base Project has developed models to predict estrogen and androgen receptor binding. Our approach rationally integrates various QSAR models into a sequential "Four-Phase" scheme according to the strength of each type of model. In four hierarchical phases, models predict the inactive chemicals that are then eliminated from the pool of chemicals to which increasingly precise but more time-consuming models are subsequently applied. Each phase employs different models selected to work complementarily in representing key activity-determining structure features in order to absolutely minimize the rate of false negatives, an outcome we view as paramount for regulatory use. In this paper, the QSAR models developed at NCTR, and particularly how we integrated these models into the "Four-Phase" system will be discussed for a number of datasets, including 58 000 chemicals identified by the U.S. EPA.

Binding affinity between chemicals and the estrogen receptor (ER) serves as an indicator of the potential to cause endocrine disruption through this receptor-mediated endocrine pathway. Estimating ER-binding affinity is, therefore, one strategic approach to reducing the costs of screening chemicals for potential risks of endocrine disruption. While measuring ER binding with in vitro assays may be the first choice in prioritizing chemicals for additional in vitro or in vivo estrogenicity testing, the time and costs associated with screening thousands of chemicals is prohibitive. Recent advances in 3D modeling of the reactivity of flexible structures make in silico methods for estimating ER binding possible. One technique, the common reactivity pattern (COREPA) approach, was applied to development of reactivity patterns for ER relative binding affinity based on global nucleophilicity, interatomic distances between nucleophilic sites, and local electron donor capability of the nucleophilic sites. The reactivity patterns provided descriptor profiles for order-of-magnitude RBA ranges of training set chemicals. An exploratory expert system was subsequently developed to predict RBA and rank chemicals with respect to potential estrogenicity. A strategy is presented for extending initial exploratory 3D QSAR models beyond current training sets to increase applicability to more diverse structures in large chemical inventories.

The ligand-binding pockets of estrogen receptor alpha and beta (ERα and ERβ) appear to have subpockets of different size and flexibility. To find ligands that will discriminate between the two ER subtypes on the basis of affinity or efficacy, we have prepared compounds of varying size, shape and structure. We have evaluated the binding affinity of these compounds and their potency and efficacy as transcriptional activators through ERα and ERβ. In this manner, we have identified a number of ligands that show pronounced ER subtype selectivity. These studies also highlight the eclectic structure–activity relationships of estrogens and the challenges inherent in developing computational methods for the prediction of estrogenic activity.

We examined the published data for the binding affinity of typical ligands to the α-subtype of the human estrogen receptor with use of an approximate molecular orbital method applicable to interacting molecular clusters. An ab initio procedure for "molecular fragments" proposed recently to deal with such macromolecules as proteins was applied to the molecular orbital calculations. The receptor protein was primarily modeled using 50 amino acid residues surrounding the ligand. For a few ligand-receptor complexes, the binding energy was also calculated with use of 241 amino acid residues contained in the entire binding domain. No significant difference was found in the calculated binding energy between the complex modeled with ligand-surrounding 50 amino acids and that with residues of the entire domain. The calculated binding energy was correlated very well with the published relative binding affinity for typical ligands.

Toxicology is the science of adverse effects of chemicals, drugs, environmental agents, and stressors. Genomics defines the structure, sequence (code), and function of the entire DNA complement of organisms. The interface of these diverse disciplines is called toxicogenomics and is based upon the application of genomic technologies to define globally the changes in gene expression (both mRNA and proteins) as a consequence of exposures. DNA microarray technology enables the simultaneous measurement of transcription of thousands of genes using microchips containing thousands of probes of complementary DNA (cDNA) immobilized in a predetermined array. The ultimate application of this technology to toxicology holds great promise but faces several formidable problems. With the solution to these problems, it will be possible to develop a substantial database of Chemical Effects in Biological Systems (CEBS). Such a database will provide the capacity to relate specific changes in gene expression to specific adverse effects and to look for similar pathways in different organisms. Such data will provide an objective way of assessing surrogate systems for reporting or predicting potential adverse effects in humans. While the potential for toxicogenomics is thus very high for studying active substances, the task must be approached in a deliberate, incremental manner to insure that only high-quality data are compiled and analyzed. This workshop will present the latest results of research conducted in leading international laboratories studying endocrine-mediated toxicity.

The estrogen receptor acts primarily as an estrogen-inducible transcription factor in target organs, however, until recently the identification of individual responsive transcripts has been cumbersome. In the present study, oligonucleotide GeneChip microarrays were used to describe and analyze the levels of approximately 6500 transcripts in the uteri of immature, ovariectomized mice at various times following oral exposure to 100 μg/kg of ethynyl estradiol. The most reproducible responses were identified and subjected to K-means clustering, and functional annotations were obtained for the transcripts within each cluster. Connections were made between the observed transcriptional responses and the known physiological responses to estrogen exposure, focusing here on effects on uterine cell cycle progression leading to estrogen-stimulated uterine growth.

Toxicogenomics can be expected to be a useful method for detecting the carcinogenic potential of endocrine active substances (EASs) in the short term with the generation of understanding of mode-of-action and mechanisms when a reliable database with information about proteomics and informatics is established. At present, there are no concrete epidemiological data supporting any exogenous EAS contribution to hormone-related organ carcinogenesis in humans. However, with the establishment of appropriate animal models and analysis of genomic-scale gene expression, risk identification and evaluation should be facilitated within a relatively short period, and this approach eventually promises to contribute a great deal of risk management regarding EASs.

Genomics can be applied in toxicology as a means of identifying modes of action, for generating hypotheses on the relationships of gene activity and toxicity, and for better characterizing the nature of dose-response relationships at low doses. This paper provides examples from our research on endocrine active compounds that illustrate each of these applications. We have determined that agents that bind estrogen receptors produce a characteristic transcript profile in estrogen-responsive tissues of the fetal and juvenile rat. The transcript profile is diagnostic of the mechanism of action. The transcripts that are up- or down-regulated by estrogens belong to a number of functional groups, such as growth factors, pro-apoptotic factors, transcription factors, and steroid metabolizing enzymes. There are a number of testable hypotheses that can be generated from these data regarding the relationships between changes in these genes and developmental or physiological responses to estrogens. We have determined that the sensitivity of gene expression changes is high, making it possible to define the shape of the dose-response curve at dose levels of estrogen several orders of magnitude below those that cause a physiological response (in this case, a uterotrophic response). The shape of the dose response is monotonic: both the number of genes changed and the intensity of the up- or down-regulation decreases with decreasing dose.

The use of gene expression data in predictive and mechanistic toxicology is hindered by a lack of information on the relationships between transcriptional events and physiological and pathological changes. We discuss the analysis of these relationships using the rodent uterotrophic response as a model experimental system for estrogen-induced uterine growth.

Environmental monitoring programs on endocrine active compounds (EACs) have been used to document the level of exposure and to assess the possible association to the occurrence of developmental and reproductive disorders in wildlife. The establishment of causal links between exposure and effect data, however, was found to be difficult due to, for example, the presence of confounding factors or limited understanding of EAC mechanisms and interactions, but also because of conceptual and methodological limitations of current monitoring strategies. In order to provide plausibility of an EAC etiology for a developmental or reproductive alteration in a wildlife population, integrated monitoring programs are needed that will use a combination of complementary approaches: methods for a targeted search for suspected EACs in an environmental mixture, analysis of internal EAC doses instead of external EAC concentrations, utilization of mechanism-based end-points in bioanalytical and effect monitoring, investigation of the basic biology and physiology of wildlife sentinel species, laboratory replication of field effects, as well as consideration of epidemiological and weight-of-evidence criteria in the design and data evaluation of monitoring programs.

Between January 1999 and December 2001, the European Community project COMPREHEND was performed. The overall aim of COMPREHEND was to assess endocrine disruption in the aquatic environment in Europe, consequent to effluent discharge, with emphasis on estrogenic activity. COMPREHEND demonstrated the widespread occurrence of estrogenic effluents across Europe and presented evidence of impacts on a range of wild fish species. Using a variety of bioassays in combination with chemical analytical methods, estrogenic steroids of human origin from domestic wastewater effluents were identified as the most pervasive problem, although alkylphenols may be important estrogenic components of some industrial effluents. New tools have been developed for the identification of estrogenic effluents, and recommendations are made for the improvement of existing techniques. We have shown that individual fish within natural populations may be feminized to varying degrees, but it has not been possible to show, using traditional fish population parameters, that the survival of fish populations is threatened. However, laboratory-based fish life-cycle studies demonstrate the sensitivity of fish to estrogen (and androgen) exposure and how this might lead to complex (and potentially damaging) genetic changes at the population level. New approaches to this problem, utilizing recent advances made in the field of molecular and population genetics, are recommended. Finally, a study of estrogenic and androgenic activity of waste waters during the treatment process has shown that some of the existing wastewater treatment technologies have the potential to eliminate or minimize the hormonal activity of the final effluent.

Biomonitoring is currently performed at two levels, assessing exposure to pollutants and effects monitoring by bioassays. As an example for the first approach, vitellogenin (VTG) in male fish of Abramis brama as an endpoint for estrogen exposure is discussed. However, similar changes of VTG or VTG-like proteins in the hemolymph of mussels could not be detected. Enzyme-linked receptor assays for monitoring estrogenic effects at the molecular level serve as an example for the second category. Applications of the enzyme-linked receptor assay (ELRA) developed in our laboratory are presented. Detection limits of 0.02 μg/l 17β-estradiol were recently achieved with the chemiluminescent format. Although effect monitoring provides information in terms of toxicity equivalents, it is not possible to relate the signals to specific pollutants and their concentrations. For this purpose, chemical analysis is required. New approaches are reported for the direct coupling of bioassays and chemical analysis. This concept is defined as bioresponse-linked instrumental analysis. It combines biomolecular recognition, initiating a biological effect, and chemical analysis. In addition to the classical bioanalytical approaches, new strategies in genomics and proteomics have been developed. This may lead to multimarker approaches opening this area to environmental analytics.

The toxicity identification and evaluation (TIE) approach that identified unconjugated natural and synthetic steroid estrogens to be the major estrogenic component in domestic sewage treatment work effluent is described. Given the increasing popularity of TIE-based approaches in ecotoxicology research, a number of important considerations, limitations, and possible future directions of TIE studies are also discussed.

A major challenge to contaminant monitoring programs is the selection of an appropriate suite of measurements for assessing exposure and effects. Early monitoring programs relied solely on residue analysis to detect the organochlorine compounds that were in use at that time. A shift to the use of more transient, less persistent chemicals required that a new set of tools be developed to determine if an organism had been exposed. This led to the development of cellular and biochemical assays that could indicate the presence of these types of chemicals in biota and the environment. However, it was recognized that measures of contaminant presence alone were insufficient to assess the health of biota. As a result, considerable research began to be directed toward development of diagnostic tools for measuring chemical effects in fish and wildlife. Today, contaminant monitoring programs follow a paradigm for study design that emphasizes not only the use of measures of exposure, but also measures of effect. Using data from our monitoring and research studies for hormonally active substances, we discuss a variety of metrics of exposure and effects and their application to specific chemicals, and the current information gaps. We conclude that although several bioindicators of exposure and effect have been promoted and used, to date there continues to be a poor association between cause and effect for endocrine active substances. In part, this is due to the limited number of diagnostic tools that are available and to a lack of basic toxicological information concerning toxicokinetics and mechanisms of action of hormonally active chemicals in fish and wildlife species. In the foreseeable future, both tissue and environmental residue data, despite the many limitations, will continue to be an important component of monitoring programs for hormonally active chemicals as we continue to develop and validate more specific bioindicators of exposure and effects.

Study protocols for the characterization of endocrine active compounds presented in Workshop 4 included the enhanced Organization for Economic Cooperation and Development (OECD) test guideline (TG) 407, the medium-term rat liver and rat multi-organ carcinogenicity assays, and an enhanced one-generation reproduction study. The outcome of rat studies on flutamide and ethinylestradiol indicated that these strongly active compounds can readily be detected even with a low animal number using the enhanced OECD TG 407. Both newly added (such as male accessory sex organ weights, histology of pituitary, vagina and male mammary gland) and already included parameters contributed to the detection of endocrine effects. Thorough evaluation of the results of 20 studies conducted with 10 compounds thought to interfere with the endocrine system by different mechanisms will identify the most appropriate enhancements to the current OECD TG 407. Medium-term rat liver and rat multi-organ carcinogenicity assays are well recognized in the International Conferences on Harmonization for Pharmaceutical Chemicals. They have been successfully used to detect carcinogenic and modifying potentials of new chemicals within a relatively short time and can be applied to endocrine active compounds. Dose-response studies on nonylphenol, bisphenol A, and styrene using the rat liver carcinogenicity assay did not reveal effects of any of these compounds on the development of preneoplastic lesions in rat liver. The enhanced one-generation reproduction study protocol included treatment of pregnant female rats from gestation day 0 through to lactation day 21, and examination of all offspring. Half of the animals were necropsied at weaning, the remaining animals were examined for vaginal opening, preputial separation, estrous cyclicity, and sperm characteristics and were necropsied at adulthood. In a pilot study ethinylestradiol inhibited maternal fertility at dose levels similar to those effective in the uterotrophic assay. It is recommended to rapidly evaluate the conducted enhanced OECD TG 407 studies and to enhance the current OECD TG 407 appropriately. Further compounds with different mechanisms of action should be studied in the one-generation reproduction study to further investigate the usefulness of this protocol. The established medium-term carcinogenicity assays can be used to study carcinogenic potential rapidly. Use of female animals and inclusion of carcinogens targeting at breast and uterus should be considered in order to explore further the predictibility of this model.

Groups of five male and five female Wistar rats were treated by gavage with 0, 1, 10, and 100 mg/kg body weight (b.w.) flutamide (FLU) or 0.01, 0.05, and 0.2 mg/kg b.w. of ethinylestradiol (EE2) for at least 28 days according to an enhanced Organization for Economic Cooperation and Development (OECD) test guideline (TG) 407 to investigate which of the current and/or additional parameters would detect effects on the endocrine system and to provide data on intralaboratory variability. Two identical studies were performed in parallel on each compound. Common enhancements were determination of thyroid hormones (T3, T4) and thyroid stimulating hormone (TSH), of the stage of the estric cycle to ensure necropsy of females in diestrus, of the number and morphology of epididymal sperm, and of additional organ weights (e.g., male accessory sex organs, MASO) and histopathology of additional organs (e.g., pituitary, vagina). Endocrine-mediated findings consistently observed in these studies were decreased relative weights of MASO at 100 mg/kg FLU and at 0.2 mg/kg EE2, histological changes in pituitary and testes at > or = 10 mg/kg and in MASO, epididymis and adrenals at 100 mg/kg in FLU-treated males, histological changes in the mammary gland at > or = 0.05 mg/kg and in testes, MASO and adrenals at 0.2 mg/kg in EE2-treated males, estrogenization of uterus and vagina (despite necropsy in diestrus) at > or = 0.01 mg/kg EE2, and changes in the ovary at 0.2 mg/kg EE2. Spermatology was insensitive (EE2) or revealed changes only at the maximum tolerated dose (MTD). Determination of T3, T4, and TSH did not contribute to the detection of the endocrine effects (FLU) or provided equivocal results. Doubling the group size to 10 animals by combining the studies run in parallel did not increase the sensitivity of detection of endocrine-mediated effects above the level obtained by histopathological examination of groups of five animals. Only some of the proposed enhancements evaluated were helpful in detecting the endocrine-mediated effects of FLU and EE2. Evaluation of studies according to an enhanced TG 407 on 10 compounds with different endocrine activities will identify the most appropriate enhancements.

Two in vivo bioassay methods, a rat medium-term liver bioassay and a rat multiorgan bioassay, can be used for detecting carcinogenic or modifying potentials of endocrine active substances (EASs) on endocrine disruption (ED). The first bioassay, the rat mediumterm liver bioassay, is fundamentally based on the two-step hypothesis of liver carcinogenesis; initiation with diethylnitrosamine (DEN, 200 mg/kg b.w., ip) is followed by test chemical administrations during the second stage, in combination with 2/3 partial hepatectomy. It requires only eight weeks for animal experimental treatment and a further few weeks for quantitative analysis of immunohistochemically demonstrated gluthathione- S -transeferase placental form positive hepatic foci. A total of 313 chemicals/substances have already been analyzed, and the efficacy of the system for hepatocarcinogenesis has thereby been well established. This bioassay also provides information concerning dose responses and inhibitory potentials of test chemicals. Several possible EASs, most of them categorized as pesticides, have already been examined in this bioassay, and dose-response studies of nonylphenol, bisphenol A, and styrene have also been tested. Another bioassay, a medium-term, multiorgan bioassay system, using five different chemical carcinogens -DEN, MNU, BBN, DMH, and DHPN- has also been established for rapid detection of not only hepatocarcinogens, but also other organ-targeted carcinogens. These medium-term bioassays are particularly useful and reliable methods for detecting carcinogenic or modifying potentials of low doses of test chemicals, such as EASs, and these methods can be used for the effects of chemical mixtures of EASs.

An enhanced one-generation reproductive toxicity study in rats without adjusting a litter size during the lactation period is proposed as a rapid and reliable bioassay for providing the data concerning adverse and/or low-dose effects of suspected endocrine disruptors. In this study, pregnant females are treated with the test substance from gestation day 0 through lactation day 21, in principle. F1 offspring from one-half of the litters in each dose group are killed and necropsied at weaning, while those from the remaining litters are examined for sexual maturation, estrous cyclicity, and/or sperm production. A series of pilot studies with ethynylestradiol as a reference chemical have suggested that the exposure of estrogenic chemicals during the early gestation period is critical for detecting effects on fertilization and/or implantation of eggs and survival of implants, and that expression of some genes including AR in the prostate and IGF-1 in the uterus of F1 offspring may be sensitive markers for monitoring potential estrogenic effects of the test compound.

Traditionally, science has progressed by slow steps involving the accumulation of studies showing particular effects, leading eventually to a general consensus. However, with increasing development and industrialization, environmental problems have escalated faster than the ability to collect sufficient data to form clear consensus among scientists. Since managers require scientific information to make decisions about management, regulation, and public policy, the gap has been partially filled by two approaches: weight of evidence and the precautionary principle. I suggest that both are useful for making decisions about endocrine active substances, although few papers in the refereed literature link the precautionary principle with endocrine active substances. As with most public policy decisions, these involve an iterative process whereby scientific inquiry must continue to fill data gaps, and to determine if the decisions made by these processes are still appropriate and protective of human and ecological health. The precautionary principle is most useful when it continues to inform and help direct research to fill data gaps in our understanding of environmental problems, such as the effect of endocrine active substances on endocrine disruption.

The precautionary principle has been central to many of the debates concerning the appropriate approach to the threat posed by endocrine active substances (EASs). This newly emerging principle has been applied to issues as diverse as persistent organic pollutants and the European trade barrier on beef from hormone-treated cattle.

The precautionary principle is controversial, and critics invoke the need to wait for "sound science" before taking "costly" regulatory action. For human health effects, epidemiologic results are often considered more valuable than toxicologic studies in animals. Direct human evidence on the endocrine effects of environmental chemicals has been slow to accumulate because of inherent sample size limitations of exposed populations and over-conservative hypothesis testing approaches. Moreover, human health outcomes may take decades to emerge. Indeed, even huge population-based studies of hormone replacement therapy have been inconclusive regarding both benefits and risks. This paper argues that certain intrinsic standard epidemiologic methods are stacked to avoid making a type I error. Moreover, these combine with extrinsic limitations (long latency, high cost), leading me to conclude that reliance solely on epidemiology to provide definitive answers, will almost inevitably delay the discovery of meaningful associations warranting timely action for protection of public health. There are several ways in which the inherent conservatism of epidemiology is arrayed against preventative regulatory steps, hence a precautionary approach is warranted while awaiting the results of additional studies which for various reasons may be very long delayed or even impossible.

The U.S. Environmental Protection Agency's Office of Pesticide Programs evaluates human health risk associated with exposure to pesticide chemicals. Chemical hazard and exposure assessment are components of the risk assessment process. For the risk assessment of single chemical conventional-type pesticides, there may be multiple exposure scenarios depending on the use pattern. Examples include acute and chronic dietary, and short-, intermediate-, and long-term occupational/residential exposures. For hazard assessment, available toxicity data and a weight-of the-evidence approach are used in the process of selecting appropriate toxicity endpoints for relevant exposure scenarios. The pesticide registration process requires that certain types of supporting toxicity data be submitted by the registrant depending in part on the chemical use pattern (e.g., food use). Types of toxicity data that might be submitted and used in hazard assessment include acute, subchronic, chronic, carcinogenicity, mutagenicity, metabolism, reproduction, developmental, neurotoxicity, and mechanistic studies. There may be data from multiple exposure routes (e.g., oral, dermal, inhalation) and from the scientific literature to consider. Dose-response information is also taken into account. In endpoint selection for a chemical, endocrine system-related effect(s) and dose-response relationship(s) are assessed in context of other types of effects, toxicities, and dose-response relationships noted. Endocrine system-related endpoints may include frank effects (e.g., endocrine organ hyperplasia or cancer) or precursor events (blood hormone level elevations). Endocrine system-related endpoints are generally treated like other cancer or non-cancer toxicity endpoints (e.g., hepatic cancer, neurotoxicity) in the risk assessment process. For chemicals with evidence of endocrine system interaction(s), an endocrine system-related effect may or may not be the most sensitive or relevant endpoint for a particular risk assessment exposure scenario. Some chemical examples will be presented. In the final risk assessment, hazard assessment information is integrated with exposure information. The assessment may be adjusted, at some point, for uncertainties in hazard or exposure data. An aggregate risk assessment, in which multiple sources or routes of exposure are considered, is typically performed for occupational and residential exposure scenarios. A cumulative risk assessment may be considered for groups of chemicals with a common mechanism of toxicity.

This paper discusses the practical implementation of the precautionary principle in the area of management of industrial chemicals in the European Union. An analysis of a number of recent cases where the precautionary principle was invoked shows that the main reason for doing so were the uncertainties in the risk assessment (or the underlying effects or exposure data), which were, according to the scientific experts, so high that the "normal" level of certainty could not be obtained. The challenge for the future is to try to develop general guidance or rules that will support the policymakers in their decision as to whether this uncertainty is so large that action is warranted or whether it is acceptable to wait until further information has become available.

Different governments and agencies are approaching the use of scientific evidence, weight of evidence, and the precautionary principle in different ways. The European community has used the precautionary principle in situations where the consequences are great, data are unavailable or will be costly (in terms of money and time) to obtain, or data are difficult or impossible to obtain. Other countries, such as the United States, have a risk assessment process that has built-in safety or uncertainty factors which are themselves precautionary. Risk management decisions can be made on the basis of adequate studies, risk assessment, weight-of-evidence approaches, and the application of the precautionary principle. While weight of evidence has been used in the United States for increased research funding and regulator action with respect to some chemicals that are hormonally active, the European community has applied the precautionary principle.

This workshop was convened to address common issues and concerns associated with risk management of endocrine-disrupting chemicals (EDCs). The talks described the tools and policies for key Japanese, Australian, German, and U.S. regulatory agencies. The agencies participating in the workshop were responsible for the regulation of various substances including: chemicals, pesticides, environmental contamination, pharmaceuticals, and food additives. The panel also described the role of the Organization for Economic Cooperation and Development (OECD) in standardizing the tools and validation of testing and screening methods. The panel also included nongovernmental organizations presenting the views of the World Wildlife Fund, and the chemical industry from industrialized nations; each organization described its concerns and proposed approaches to risk management of EDCs. This summary highlights the most important areas of common points of view of government, industry, and environmentalists. We also try to identify issues upon which viewpoints diverge.

A hormonally active compound was first identified in the book Silent Spring by Rachel Carson in 1962, implicating the effect of pesticides such as DDT and the derivatives. Nearly four decades later, the book Our Stolen Future by Theo Colborn et al., and other pertinent publications have revisited and broadened the issue regarding a variety of possible chemicals and the area exposed. Translation and publication became available in Japan within the last four years. Since then, Japan joined the member countries involved in the global issue of endocrine disruptors, the "environmental hormone". Although a significant number of chemicals possessing a hormone-like action have been recognized for many years, and the action of their biological plausibility related to the receptor-mediated effects strongly suggests possible human effects comparable to hormonal changes in wildlife, little is known about evidences or adversities in experimental animals and humans. The most essential key to resolving these dilemmas may be to understand the mechanism of actions (i.e., a possible low-dose issue). In other words, the mechanism at the low-dose effect may be resolved simultaneously by the mechanism of three major questions linked to the low-dose issue; namely, threshold, possible oscillation, and additive and/or synergistic action.

Like hardly any other issue in ecotoxicology, endocrine disruption has given rise to public concern. Reproductive, behavioral, and immunological effects in wildlife were publicly not only understood as possible threats to wildlife populations, but also as early warning signals that human health could be at risk. Above all, the public has been concerned about negative outcomes in reproductive health, and effects like feminization in fish were regarded as evidence for the biological plausibility of the hypothesis that environmental levels of hormonally active chemicals are high enough to affect human reproductive health. Public concern has been mirrored by several parliamentary and governmental decisions emphasizing the need for extensive research and rapid measures to reduce the risk associated with endocrine-disrupting substances. Endocrine disruption in wildlife is clearly a priority issue. At least in densely populated areas like Europe, symptoms of endocrine disruption in wildlife cannot only be detected in areas with abnormally high levels of pollution, but have also occurred in main river systems, estuaries, and even in the open sea. Imposex in mollusks and feminization in fish that were clearly related to disturbances in the hormonal system of these organisms by exogenous substances have been used as markers in monitoring programs. Though symptoms of endocrine disruption can be clearly identified, it is much more difficult to link these outcomes to causative chemicals or mixtures of substances. Natural and pharmaceutical hormones, phytoestrogens, pesticides, and industrial chemicals may all play a role to a different degree depending on the site under study. This means that several different risk-reduction strategies have to be applied, including bans of substances, use restrictions, and installation and optimization of sewage treatment works embedded in a strategy for the overall reduction of chemical input into the environment. It should be noted that, in addition to national and international regulatory actions taken by state authorities, a considerable reduction of the environmental input could be achieved in several countries by voluntary actions taken by industry. Regulatory bodies are still facing major problems in the field of risk assessment and risk reduction. Association between effects and causative agents or mixtures are in many cases weak. Important tools for risk assessment such as dose-response relationships or the existence of thresholds are not yet agreed on. These uncertainties are the reason that many national governments and the European Commission have identified precaution as the main element in chemicals policy for the management of endocrine disruptors. This paper is based on documents of the German Federal Environmental Agency, but solely represents the view of the author from a regulatory perspective and emphasizes the wildlife aspects of endocrine disruption.

_ Manufactured chemicals are essential to the vast array of goods and services that contribute to modern life. Their benefits are innumerable, and society is entirely dependent upon them. At the same time, there is an increasing awareness of the concept of environmental impacts. The challenge is to achieve the appropriate balance between the benefits and risks from chemicals, so that we all may enjoy the benefits of chemicals without significant detriment to current and future human and wildlife health. Ecological risk assessment is the mechanism that allows potential environmental chemical exposure to be benchmarked against hazardous properties so that risk is acceptable and environmental health is not impaired. Chemical management decisions based on such assessments are said to be risk-based. Within the context of environmental risk assessment practice for endocrine disruption, industry would support a position that: endorsed the risk assessment process; recognized that endocrine disruption is not an adverse effect per se, but rather a potential mechanism of action; gave precedent to population-level effects instead of individual-level effects; employed a tiered approach to hazard assessment; emphasized standardized and validated effects testing methodologies; recognized that exposure per se does not necessarily constitute a risk; considered relative potency (i.e., evaluation of the dose levels and mechanisms producing toxic adverse effects and determining whether the critical effect arises via an endocrine mechanism or another mechanism); benchmarked risk against loss of benefits; and evaluated risk within the context of overall risk from both natural and anthropogenic substances with common modes of action. To help address uncertainty surrounding the risk from Endocrine Active Substances (EAS) to wildlife, the chemical industry -- via the Long-Range Research Initiative (LRI) -- has implemented a research program aimed at identifying and addressing knowledge gaps and establishing internationally harmonized testing methodologies in cooperation with other stakeholders. Details of individual projects within the current LRI research program are presented.

This paper reviews the existing concerns and presents conclusions and recommendations for action. The first section outlines an environmentalist's perspective and predicts that endocrine active substances (EASs) pose a high risk for wildlife populations. The second section explains that risk assessment, as currently practiced for individual chemicals, does not adequately protect the environment, and suggests several improvements that should be made. However, it is argued that the properties of endocrine disruption lead to increased uncertainty in the risk assessment, such that countries wishing to achieve a high level of protection would be justified in implementing precautionary controls. The third section outlines some conclusions and additionally provides an appraisal of the responses of governments and industry. Furthermore, several recommendations are made, particularly the imperative for research funding to be substantially increased and guaranteed for many years to come. Also highlighted is the need to hasten the development of screens and tests for ecotoxicity, and for governments to commit to a comprehensive sorting, screening, and toxicity testing program for all chemicals to which significant exposure occurs.

Drugs may have intended or unintended endocrine effects. Drug evaluation may include both in vitro and in vivo evaluations of toxicity and developmental/reproductive effects. After a signal is identified, human relevance is of utmost concern. An integration "tool" that formalizes a weight-of-evidence approach has been developed to assess concern about reproductive/ developmental toxicity to humans. This approach can be used to assess concern about an endocrine disruption signal. A signal alone does not mean a concern for humans. An effect needs to have biologic relevance, and exposure thresholds for effects may exist. Risk/benefit for a particular drug is a clinical decision and may vary by the drug indication. Risk management for an identified concern could include wording in patient communications, tracking distribution or limited distribution, and patient or pregnancy registries.

Similar to other countries, issues concerning endocrine disruptors have been attracting people's attention, as well as concerns, in Japan. Consequently, relevant ministries have been conducting several activities and projects. A summary of the current activities related to endocrine disruptors being conducted by Japanese government agencies is provided in this paper.

This paper describes the establishment of a number of expert groups in the Organization for Economic Cooperation and Development (OECD) to manage the work related to the development and validation of test methods to identify and characterize the human health and aquatic environmental hazards of endocrine-disrupting chemicals. In addition, the concept of testing and assessment, using the most adequate testing tools without being prescriptive, is explained, and a coordinated approach for international work sharing under the OECD umbrella is described.