Jump to main content.


Note: EPA no longer updates this information, but it may be useful as a reference or resource.


Second Interim Report to the EDSTAC

Submitted by The Screening and Testing Work Group
November 24, 1997

The purpose of the Second Interim Report is to provide a description of the Screening and Testing Work Group (STWG) progress to date and identify areas requiring further discussion and work. While the Report has evolved significantly since the first one was submitted in October, the STWG members are still in the process of discussing many components of both Tier 1 Screening (T1S) and Tier 2 Testing (T2T).

After an introduction to the intended purpose of T1S, including the criteria used in designing a program, the report includes a general overview of the value of including both in vitro and in vivo assays in the battery. Three T1S battery alternatives, which include some of the same assays, are then presented. Overviews of the T1S assays in each option are provided, and their relation to the criteria required of the T1S battery will be discussed. For both mammals and other vertebrates, the STWG formally considered many more assays than those presented here. In addition, a section on implementation of T1S, still in need of expansion, is provided.

After the New York EDSTAC plenary meeting, the STWG expanded upon their discussions related to T2T. Included in the Second Interim Report are a rationale for T2T, guidance for selecting tier 2 tests, a rationale for inclusion of the two-generation reproductive toxicity study, a rationale for inclusion of non-mammalian tests, a discussion of the dose considerations for multiple generation reproduction tests, and an introduction to the group's discussion of standardization and validation.

II. General Introduction to a Tier 1 Screening Battery

A. Introduction to T1S

T1S is intended to act as a "gatekeeper." After the T1S battery is applied, a chemical substance or mixture (CSM) will be designated as having either: 1) potential for estrogen, androgen, or thyroid hormone disruption and therefore will go to T2T; or 2) low or no potential for endocrine disruption for estrogen, androgen, or thyroid and therefore will go to the "hold unless ..." box. Criteria for this designation have not yet been agreed upon by the STWG members, however, the discussion of weight of evidence (see Section IV, a.) begins to give general guidance.

The STWG has interpreted the nature of the decision, based on T1S results, to mean T1S must therefore: 1) maximize sensitivity (if need be, over specificity) to minimize false negatives, while permitting an as-of-yet undetermined, but acceptable, level of false positives; 2) encompass a sufficient range of metabolic activity; 3) capture all known areas of activity for estrogen, androgen, and thyroid; 4) capture a sufficient range of taxonomic groups; and 5) incorporate sufficient diversity among the assays to reach conclusions based on weight of evidence. The T1S battery must therefore be structured such that, at the completion of the assays, the STWG and EDSTAC (during the T1S development) and the EPA and other stakeholders (during its implementation) will be scientifically and ethically comfortable in assigning CSMs as either having 1) low or no potential for estrogen, androgen, or thyroid endocrine disruption, or 2) as having such potential. This comfort level in moving a CSM to the "hold unless ..." box requires that T1S "covers the waterfront," (i.e., that T1S includes a sufficient diversity of screening assays, endpoints, and taxa). At the same time, the T1S must be relatively fast and economical while meeting the criteria described below. T1S is not, and should not be, designed to produce definitive answers; that is the goal of T2T.

T1S results should also inform T2T, in terms of guidance on which tests to perform, which endpoints to include, and which organisms to assess to satisfy the goals of T2T. These goals include identifying: at what doses (dose-response), at which life stage(s) (most sensitive), and in which organism(s) (most appropriate, most sensitive, most at risk, etc.), do the adverse effects occur. Based on these concerns and goals, the STWG is considering a tailored approach to T2T, with a necessarily more global approach to T1S.

The STWG has, therefore, spent a great deal of time evaluating existing and potential assays for inclusion in the T1S Battery. The options included in this paper are being evaluated further for their relative strengths and weaknesses. The goal is to create a T1S battery with the necessary breadth, depth, and rigor to allow assignment of CSMs to the "hold unless ..." box or T2T, to inform subsequent T2T as necessary, and to provide a comprehensive detection of hazard, but be as economical as possible (in terms of time and costs) to maximize throughput.

B. The STWG Working Model of Endocrine Disruption and the Implications for the Design of T1S

Xenobiotics can alter endocrine function by affecting the availability of a hormone to the target tissue, and/or affecting the cellular response to the hormone. Mechanisms regulating hormone availability to a responsive cell are complex, including hormone synthesis, serum binding, metabolism, cellular uptake (thyroid), and neuroendocrine mechanisms that control overall function of the endocrine axis. Mechanisms regulating cellular responses to hormones are likewise complex and are tissue-specific. Because the role of receptors is crucial to cellular responsiveness, specific nuclear receptor binding assays are included. In addition, tissue responses that are particularly sensitive and specific to a hormone are included as end-points for Tier 1 screens.

C. Criteria for Tier 1 Screening

The STWG has identified five criteria for constructing a screening battery and used these criteria in discussions of which specific assays to recommend. These criteria were presented to the EDSTAC during the Chicago meeting and derive from the Conceptual Framework. They are presented here again with brief discussion:

    1. Does the T1S battery minimize false negatives while permitting an as-of-yet undetermined, but acceptable, level of false positives? [this criterion expresses the need to 'cast the screening net widely' and not miss potential endocrine disruptors]
    2. Does the T1S battery capture an adequate range of metabolic activity? [the battery should include assays from representative vertebrate classes to reduce the likelihood that important pathways for metabolic activation of parent CSMs, or other events, are not overlooked]
    3. Does the T1S battery have the potential to capture known endocrine endpoints? [all known chemicals affecting the action of estrogen, androgen, or thyroid hormones should be detected]
    4. Does the T1S battery capture a sufficient range of taxonomic groups? [there are known differences at the organismal level (e.g., in detoxification) among taxa that may affect endocrine activity of CSMs]
    5. Does the T1S battery incorporate sufficient diversity among the assays to reach conclusions based on weight of evidence? [we suspect many decisions on further testing of CSMs will require weighing the evidence of several assays]
D. Introduction to in vitro and in vivo Screening for Estrogen, Androgen, and Thyroid

Three T1S batteries are being considered by the STWG. The first two have been discussed by the entire work group while the third was only discussed by a subset of the group, though the specific assays included have all been discussed by the full group. Each of the assays being considered are listed below, followed by a rationale for each one. The STWG recognize these assays require varying levels of development, standardization, and validation. It is critical to acknowledge that the state-of-the-science in this area is evolving rapidly and assays are currently being developed that may offer distinct advantages over some of those discussed below. As they are developed and validated, the use of new assays for screening is strongly encouraged.

General agreement has been reached on the strengths and limitations of most in vitro, in vivo and ex vivo methods for detection of toxicants that act via estrogen receptor (ER), androgen receptor (AR), steroid hormone synthesis inhibition, and altered hypothalamic-pituitary-gonadal (HPG) mechanisms. With this in mind, several short-term in vitro assays for AR, ER, and SIS action were identified by the STWG as quite useful in screening. Examples of large scale high throughput endocrine screening (HTPS) programs exist in the pharmaceutical industry. Several companies involved in drug design are currently screening chemicals for hormonal activity on a large scale (thousands per month) using standardized in vitro binding and transcriptional activation assays (reporter gene) together with receptor binding assays. Their assays detect chemicals that act via AR, ER, thyroid receptor (TR), retinoic acid receptor (RAR), peroxisome proliferating receptor, and other nuclear hormone receptors.

In vivo tests are often apical and therefore, less specific, but more comprehensive, than in vitro tests. In vivo tests can be made more specific if accompanied by target organ/cell dosimetry of biologically active metabolites. In vitro data are enhanced if the actual concentration of the chemicals in the media is determined, to account for metabolism, stability, and solubility, and to determine whether these concentrations match with those that can be achieved in vivo. Cellular assays should determine viability and the specificity and limitations of each assay should be defined. It is clear a combination of in vivo and in vitro assays can detect EAT alterations that act via AR, ER, inhibition of steroid hormone synthesis, or alterations of the HPG axis. In the STWG more than fifty endpoints were discussed, including in vitro, in vivo and ex vivo (in vivo dosing followed by in vitro assessment of function) techniques. In vivo endpoints considered include, reproductive organ weights and histology, serum hormone levels, in vivo gene activation, protein synthesis, behavior, growth, development, pregnancy maintenance and anatomy/morphology. In vitro methods included the assessment of cell proliferation (E-Screen, MCF-7 cells), AR, and ER binding, steroidogenic enzyme/ hormone synthesis, reporter gene assays using transient or stably transfected cell, biochemical assays and in vitro and ex vivo testis/ovary steroid hormone synthesis.

For each endpoint, the sensitivity (defined here as the response of the assay to low concentrations or dosage levels), specificity (pathognomonic for a mechanism of action, since the lack of specificity leads to false positives), relative simplicity, problems, confounds and limitations and test duration were discussed. In addition, we considered how widely accepted the method was in toxicology, how many chemicals had been screened and whether or not this was a relatively new assay (state-of-the-art) or one that had been used for decades. For in vitro assays, the metabolic capacity of cultured cells and the solubility of chemicals in media were important considerations. The inability of many, but not all, in vitro systems to activate toxicants was considered to be a limitation of these methods. Appropriate control groups and special experimental design considerations were also discussed. The work group has discussed the degree to which each assay is, or is not, "validated" for use in screening (see Section IX), but has not finalized those discussions.

For purposes of clarity the following definitions are utilized in this report. Estrogenic refers to compounds whose effects are mediated through the ER, initiating a cascade of cell/tissue specific effects similar to those initiated by estradiol. Estrogen-like effects, in contrast, resemble those of estrogen but are not or have not been shown to be mediated through the ER. Similarly, androgenic effects are AR mediated, as opposed to androgen-like effects which are not mediated via AR. In contrast, the terms antiandrogenic and antiestrogenic are not specifically limited to AR and ER mediated interactions. Antihormones can act via 1) the steroid hormone receptor, 2) steroid hormone synthesis inhibition, 3) reduction of bioavailability by reducing the amount of free hormone in the serum, 4) increased hormone metabolism leading to reduced serum hormone levels and 5) other mechanisms.

The endpoints discussed by the STWG were considered for their utility in screening endocrine disrupting compounds (EDCs). It was generally agreed that the sole objective of EDC screening is to detect specific endocrine activities and not to determine dose-response relationships, confirm the mechanism of action or determine the adversity of the chemical's effect on reproduction and/or development. However, screening assays must be sensitive enough to detect all known xenobiotics that act via the mechanism of action the assay is intended to reveal.

In vitro evaluations, it is acknowledged, can provide both false positive and false negative results. False positives arise in vitro (active in vitro but not in vivo), when a chemical is not absorbed or distributed to the target tissue, is rapidly metabolically inactivated and excreted, or when some other form of toxicity predominates in vivo. However, false negatives are considered to be of greater concern if in vitro tests were used to the exclusion of in vivo methods. In vitro evaluations can result in false negatives due to their inability or unknown capacity to metabolically activate toxicants. There was consensus that in vivo methods should be utilized in conjunction with in vitro techniques if metabolic data are unavailable. In cases where it is known that certain classes of chemicals do not require metabolic activation, or the metabolites are known, some in vitro tests offer distinct advantages over in vivo assessment.

Advantages of in vitro methods include:

As stated above, inclusion of in vivo methods can help avoid false negatives in the absence of knowledge of absorption, distribution, metabolism, and excretion (ADME).

Advantages of in vivo assays include:

Objectives of Screening are: Objectives of Testing but not Screening: It is important to be aware that the screening battery is being designed to minimize false negatives, based on an assessment of the ability of the battery to detect known EDCs that act via EAT. In this regard, the value of each individual assay cannot be considered in isolation from the other assays in the battery, as they have been combined in a manner such that the weakness of one assay are complemented by another.

The STWG believes the screening battery chosen should be able to detect all of the desired EDCs including xeno(anti)estrogens (that act via the ER or inhibition of aromatase by oral or parenteral administration), xeno(anti)androgens (via AR or hormone synthesis), altered HPG axis and antithyroid action (via synthesis, metabolism and transport, and the TR).

The participants of the STWG noted, however, that the results of even the most specific in vivo assays, can be affected by endocrine mechanisms other than those directly related to ER or AR action. For example, uterine weight in the ovariectomized female rat is affected in an estrogen-like manner by high doses of aromatizable and nonaromatizable androgens and growth factors like EGF. The age at puberty (vaginal opening in the female or preputial separation in the male rat) can be affected by chemicals that act on the hypothalamus, pituitary, or thyroid or alter growth hormone secretion. If gonadally intact females are used, uterine weight can also be affected by toxicants that stimulate hypothalamic-pituitary or gonadal endocrine secretions. Clearly, castration of the treated male or female markedly affects the specificity of the test. The lack of specificity of in vivo assays is a limitation if the goal is to only identify ER, or AR alterations. In contrast, this lack of specificity could be considered an advantage if a broader, more apical screening strategy is desired.

E. Three Proposed T1S Options:

While three options are still being considered by the STWG, they all include some of the same assays. The work group is NOT proposing all three options be run in totality, but rather that one of the options be chosen as the T1S battery. Subsequent work group discussions will be designed to achieve this goal, however, if one battery cannot be agreed upon, the work group will lay out the relative strengths and weaknesses of each option. A decision on exactly which option is chosen may likely need to be predicated upon the results of the standardization and validation program.

Option 1:

Option 2: Option 3: F. General Information on Receptor Binding and Functional in vitro Assays

The STWG has agreed receptor binding and/or functional assays should be included in T1S. The specific assays chosen, whether done at the "bench" or through the high throughput pre-screening process still in development, should have the following characteristics:

The STWG agreed receptor binding assays should be performed for estrogen and androgen receptors, but not thyroid if done at bench (currently only performed at bench level; characteristics a, b, and d above are relevant); and/or, functional assays should be performed for estrogen, androgen, and thyroid receptors (specifically recommended is stably transfected cell like the MVLN-cell if available to assess transcriptional activation. If stably transfected cell lines are not available then transiently transfected reporter gene assays should be used. Although MCF-7 proliferation assays are acceptable; yeast-based assays are not recommended at this time; can be performed either high throughput or bench level; characteristics a, b, c, and d above are relevant).

Receptor binding assays can use rat, mouse, or human ER or AR. These assay the ability of the xenobiotic CSM to displace the radiolabelled endogenous ligand from the binding site, in a cell-free system. Relative potency can be determined for positive CSMs. These systems currently do not have metabolic capability. The limitations are therefore solubility in the culture medium, inability to distinguish agonists from antagonists, and lack of metabolic capability.

The functional assay, specifically transcriptional activation, requires, for agonist or antagonist activity, that the CSM binds to the receptor and that there is a consequence to the binding, i.e., transcription (synthesis of mRNA) of a reporter gene and translation of the mRNA to an identifiable detectable protein such as firefly luciferase (with substrate and cofactors present in the culture, there is a light flash detected from formation of the product when the enzyme is synthesized in response to transcriptional activation and acts on the provided substrate) or bacterial (-galatosidase (with substrate and cofactors present in culture, the product is detected colorimetrically when the enzyme is synthesized in response to transcriptional activation and acts on the provided substrate). This is an intact cell culture assay, which currently requires different cell lines for assessment of effects on EAT binding domains with transfected (transiently or permanently) receptors and reporter gene constructs. This assay can distinguish between agonists and antagonists. The limitations are solubility, toxicity, permeability of the cell membrane, lack of or limited metabolic capability. If a CSM must be metabolized to an active moiety, it will not be detected unless the limited residual metabolic capacity of the cultured cells is sufficient to transform the chemical to its active form. Metabolic activity can be provided by provided by either preincubating the CSM with an S9 fraction (supernatant from 9000g x centrifugation of homogenized liver from a metabolically induced rat), or incorporating the S9 fraction into the treatment mixture. In addition, cell lines are being genetically engineered to incorporate genes for P450 enzymes as a method for extending their metabolic capacity and, perhaps, obviate the need for use of S9.

The performance of both the receptor binding and functional in vitro assays may not be necessary since performance of the functional assay includes assessment of both receptor binding and transcriptional activation (the preferred functional assay). Therefore, performance of a functional assay may be sufficient.

These assays evaluate just one of the possible mechanisms of endocrine disruption; if a CSM does not act through the receptor, it will not be detected.

III. Tier 1 Screening Battery - Option One

A. Mammalian Subjects/Material - in vitro Assays

Following the EDSTAC Plenary direction given on October 8, 1997, that 1) the hypothalamic-pituitary-gonadal (HPG) axis must be assessed in the T1S battery, and that 2) assessments of the postnatal consequences of in utero exposure belong in the T2T battery, the STWG presents the rationale for the first T1S Battery option as follows:

1. Cell-free Receptor Binding

Receptor binding assays (RBAs) are long-standing and relatively simple in vitro assays that detect specific mechanism of endocrine activity. This is important because several xenobiotics display affinity for the estrogen and/or androgen receptors (thyroid hormone receptors exist are structurally and functionally similar to receptors for estrogens and androgens, but there is no information on CSMs binding to these proteins). Binding assays identify, but do not discriminate between, agonists and antagonists. The apical nature of these assays can be an advantage rather than a limitation because either activity can produce adverse endocrine effects. These assays typically are less sensitive than the functional assays, and they lack metabolic activity, which is an advantage if one wishes to identify the specific compound with endocrine activity. However, the lack of metabolic activation is also a limitation because some xenobiotics require metabolic activation. With this limitation in mind, the STWG along with some members of the PSWG, have suggested that in vitro screening assays should screen a metabolically activated fraction, as well as the parent material. For the whole cell assays, discussed below, one could employ two cell lines, one that can and one that cannot metabolize toxicants. In vitro assays conducted in a high-throughput mode, with and without metabolites, could provide an extremely powerful tool for potential receptor binding activity of large numbers of CSMs. Importantly, because thyroid hormone binding is not a known mode of endocrine disruption, activity of chemicals through mechanisms altering thyroid hormone synthesis or removal must be detected with high confidence.

2. Functional Assay (Transcriptional Activation or Cell Proliferation)

In addition to RBAs, the STWG recommends the use of whole-cell transcriptional activation to detect endocrine disruptor activity. The work group will have further discussions to determine which, if any, of these assays should be recommended over others in the future. These assays can also be conducted in a manner that allows them to distinguish between receptor agonists and antagonists.

Although these functional assays often provide information similar to the above binding assays, this is not always the case. There are occasionally well-founded biological reasons for a chemical to display positive results in either the binding or the transcriptional activation assays but not both. Hence, the use of both of these would reduce the incidence of false negatives. At present, the STWG feels both sets of assays should be considered. However, due to a higher degree of difficulty, the STWG is concerned that proper execution of whole-cell assays requires a level of skill and training that may not currently exist in the toxicology community. If so, these assays might be more difficult to validate, standardize, and implement than the binding assays, some of which have been used for decades and are less complex.

The information obtained from RBAs is a subset of that obtained from a transcriptional activation assay; thus, transcriptional activation assays would be an acceptable alternative to cell-free RBAs and may even be preferable if such assays can be standardized and validated, and if patent restriction questions can be resolved. Functional assays are more sensitive than cell-free receptor binding assays; however, there are many versions of functional assays that may not be acceptable; specific assays have been identified and defined and we are discussing their relative merits and ability to replace RBAs.

a. Cellular Assays Detecting Estrogen Receptor Agonists and Antagonists: Assays like the MVLN cell assay for ER activity, an MCF-7 cell line stably transfected with a promoter and luciferase reporter gene, are recommended over transiently transfected cells. If the stably transfected cells are unavailable, the transiently transfected cells can be used to assess estrogen and anti-estrogen action. In addition to the MVLN, other stably transfected cell lines have been or are being developed. The MCF-7 cell proliferation assay is not recommended because the proliferative response is indirect (i.e., the presence of functional estrogen receptor is necessary but not sufficient to evoke estrogen-mediated cell proliferation). Instead, the reporter gene assays are a direct manifestation of receptor-mediated responses on gene expression (i.e., the presence of functional estrogen receptor and of a reporter gene are sufficient to express estrogen-mediated induction). However, if one wishes to use the MCF-7 proliferation assay, the data are acceptable and would obviate the necessity of running one of the other transcriptional activation assays mentioned above. Another possibility is to run each of these assay types through a validation process to determine which type best meets the goals of T1S. In contrast, the Yeast Estrogen Screen (YES) assay is not considered acceptable because of its inability to detect the estrogenic activity of chlorinated chemicals (positive YES data are acceptable and useful, but negative data are not). In addition, since the YES assay, when it works, does not distinguish between agonists and antagonists, the results are more equivalent to a binding assay than other functional assays.

Minority Suggestion:

a. Cellular Assays Detecting Estrogen Receptor Agonists And Antagonists: Both cell proliferation assays and transactivation assays are acceptable in principle. The MCF-7 cell proliferation assay has been used extensively for this purpose. More recently, the MVLN cell assay for ER activity was introduced; this is a MCF-7 cell line stably transfected with a promoter and luciferase reporter gene. Transiently transfected cells can also be used to assess estrogen and antiestrogen action. In order to recommend one of these assays over the others, each of these assay types should be run through a validation process to determine which type best meets the goals of T1S. In contrast, the Yeast Estrogen Screen (YES) assay is not considered acceptable because of its inability to detect the estrogenic activity of chlorinated chemicals (positive YES data are acceptable and useful, but negative data are not). In addition, since the YES assay, when it works, does not distinguish between agonists and antagonists, the results are more equivalent to a binding assay than other functional assays.

This "minority report" is based on the fact that the MCF-7 proliferation assay has been in use for a long time, and many papers have been published attesting to its use in screening for estrogen agonists and antagonists. The assay chosen by the majority of the members, the MVLN, was developed recently, and hence, it use has not been as thoroughly documented. Because of this disparity, the only objective way to recommend one assay over the others is to conduct a validation study.

b. Cellular Assays Detecting Androgen Receptor Agonists and Antagonists: For AR-mediated activity, stably transfected cell lines are under development, but not yet available. Assays like the CV-1 cell line transiently co-transfected with hAR and a promoter construct with a luciferinase reporter are recommended. It is noteworthy, that as compared to MCF-7 cells the CV-1 has considerable metabolic capability. Here again, the Yeast Androgen Screen (YAS) is not acceptable as it is unable to detect the AR-mediated activity of chlorinated pesticides, however, positive YAS data are acceptable.

3. Steroidogenesis Using Minced Testis Assay

Anti-androgens and anti-estrogens act via a number of direct mechanisms in addition to those which directly involve the steroid hormone receptors. One prominent mechanism of anti-hormonal activity is inhibition of hormone synthesis by inhibiting the activity of P450 enzymes in the steroid pathway. Such activity could be detected in a fairly simple in vitro procedure with minced testicular tissue obtained from adult male rats. Although for many of the pesticides known to alter this pathway the parent material is active, testing a "metabolically activated" fraction would enhance the utility of this assay. Further development and validation of this assay is required. Leydig cell cultures could be used in place of the minced testis culture; the results are more precise, but the technique is more difficult.

Additional Information: evaluates steroidogenesis from cholesterol to testosterone (T), and assesses effects on three different P450 isozymes in pathway, ± human chorionic gonadotrophin (hCG) challenge. Limitations are solubility, cell toxicity, permeability through cell membranes and only minimal metabolic capability (just testicular, not systemic). This assay involves 50 mg of minced adult male rat testis under suitable culture conditions ± additions such as hCG or enzyme substrates. The assay does not assess effects on 5(-reductase which catalyzes conversion of T to dihydrotestosterone (DHT) since it is not present (to any appreciable extent) in the testis (it is present in male accessory sex organs and will be assessed in the Hershberger assay + T; in vivo assay no. 1). This assay does not assess effects on aromatase (which catalyzes conversion of T to 17(-estradio [E2] since it also is not present in the testis); it is detected in the 14-day pubertal assay, in vivo assay no. 3, since it is present in ovarian (and placental) tissue.

B. Mammalian Subjects - in vivo Assays

In vivo assays are more apical; that is, they incorporate endocrine-specific endpoints but disruption of a number of hormone regulation/delivery mechanisms can be evaluated at once. In addition, the in vivo assays allow evaluation of mechanisms of hormonal action that cannot be isolated in in vitro assays.

1. Female: Uterotrophic Assay

A 3-day Uterotrophic Assay for estrogenicity or anti-estrogenicity in immature (18-21 day old) or adult ovariectomized female rat. Females are injected sc or ip with the chemical for 3 days and necropsied 6 hours after the last dose. This assay, which is one of several used for over 50 years, has been used for thousands of estrogenic chemicals, including hundreds of xenoestrogens. Some of the xenobiotics of current concern, e.g., bisphenol A, were identified in such assays in 1938. The assay detects estrogens and anti-estrogens and should incorporate weighing and histological examination, including the percent of endometrial luminal epithelial cells in the proliferative cycle, of the uterus.

Additional Information: ovariectomized adult female, three-day dosing sc with and without coadministered estradiaol (for anti-estrogens) (some known EDs not active after po dosing; see in vivo assay 3). Six hours after last sc dose, necropsy female, weigh uterus (± fluid); i.e., wet and "dry" (pressed weight); can add assessment for lactoferrin; trigger histological assessment of uterus for epithelial cell height (if uterine weight changes equivocal), etc.; detects whether CSM interacts with uterine receptors (does not assess HPG axis). Limitations are solubility, toxicity; organism provides metabolism.

2. Male: Hershberger Assay

This assay has been used for over 40 years in the pharmaceutical industry and for several xenoantiandrogens as well. An in vivo assay is required to screen chemicals for androgenic and anti-androgenic activity. In the standard Hershberger, weanling male rats are castrated and for five consecutive days are given either: 1) vehicle, 2) testosterone, 3) the xenobiotic (detects androgens as compared to group 1), or 4) the xenobiotic plus testosterone (detects anti-androgens as compared to group 2). In conducting the assay one should necropsy male and weigh seminal vesicle (SV), ventral prostate (VP), and levator ani plus bulbocavernosus muscles (LA), which are all androgen-dependent (T or DHT) tissues. This assay could be extended beyond 5 days to 14 days (or longer) for assessment of thyroid function. However, in the absence of a group of intact males (with and without treatment), this assay would provide no assessment of the hypothalamic-pituitary-gonadal axis.

Additional Information: castrated adult or young male rat, ± silastic implants of T, five day subcutaneous (sc) or oral (po) dosing. The assay evaluates changes in the weights of accessory sex organs (ASOs), specifically: prostate (ventral prostate), preputial glands, coagulating glands, seminal vesicles, levator ani plus bulbocavernosus muscles; detects A/antiA; (does not assess HPG axis); detects whether CSM interacts with receptors in ASOs; also can detect any effect(s) on 5(-reductase, this enzyme (which catalyzes T to DHT) is present in male ASO (if analyze for DHT in blood of males with T-implants). Limitations are solubility, toxicity; organism provides metabolism.

3. 14-day Thyroid/Pubertal Assay

Peripheral effects of thyroid hormones in the adult and the developing fetus are complex and require considerable time to manifest. Therefore, the STWG proposes the measurement of circulating thyroxin (T4), thyrotropin (TSH) and thyroid histopathology in an in vivo assay in one sex. Measurement of T4 is recommended because many xenobiotics affect circulating levels of thyroxin by increasing T4 clearance rate and/or by displacing thyroxin from carrier proteins in the blood. Thyrotropin level is the single endpoint used clinically to evaluate thyroid hormone action. Changes in circulating TSH can compensate for changes in T4; thus, both T4 and TSH must be measured. Additionally, changes in the biopotency of TSH that are not detected by RIA will be evaluated by histological examination of the thyroid gland. Finally, the STWG feels that 14 days of dosing are required for manifestation of these effects.

14-day Thyroid/Pubertal Assay. Using the immature female rat, oral dosing is initiated at 21 days of age and continued for two weeks to day 35. Oral dosing allows for the detection of chemicals not active in the 3-day uterotrophic assay due to the fact that they are not effectively absorbed via sc injection (i.e., methoxychlor). Estrogens accelerate vaginal opening, induce vaginal cornification, and alter estrous cyclicity (triggered measures from accelerated VO). Anti-estrogens delay VO. In addition, as puberty is the landmark of hypothalamic-pituitary-gonadal maturation culminating in vaginal opening and ovulation, alterations of these functions can be detected as well. The VO assay has been used for about 80 years and is a part of current Agency test guidelines. Hence, test labs have standard operating procedures (SOPs) and experience in collecting pubertal endpoints in female rats. From these studies it is clear that VO is more sensitive to estrogens than traditional multi-generational endpoints. Effects in thyroid hormone (as discussed above) can also be detected in this assay.

Additional Information: intact weanling rats dosed by gavage (to detect those EDs not active after sc dosing; see in vivo assay no. 2) for 14 days, postnatal day 21-35. During dosing, assess acquisition of vaginal patency; at necropsy assess T4 and TSH, thyroid is weighed and examined histologically; weigh uterus (detect anti-estrogen and weigh ovary; trigger histology of uterus or ovary if weight change is equivocal). This apical in vivo assay in intact rats does detect effects on HPG axis in female (not anticipate major sex differences in sensitivity and/or specificity for HPG or thyroid hormone responses), will detect effects on aromatase and TPO. Can also measure age at first estrus, onset of cyclicity and if cycles are absent. Limitations are solubility in dosing solutions/suspensions, toxicity, absorption into gastrointestinal tract; organism provides metabolism.

C. Other Vertebrate Subjects/Material - in vivo Assays(1)

As mentioned earlier, there are known differences at the organismal level (e.g., in detoxification) among taxa that may affect the endocrine activity of chemical substances and mixtures (CSMs), but comparative physiology is not extensive enough to predict which taxa are the most sensitive. Hence, we have evaluated in vivo assays using representatives of selected vertebrate classes. All the assays in this category have been used previously to investigate basic mechanisms of development or reproduction and the influence of hormones on these processes. The assays vary in the amount of data extant relevant to the action of CSMs. Each assay needs at least some of the following: more data to demonstrate their efficacy with CSMs; validation and standardization relative to estrogen, androgen, or thyroid hormone; and/or modification to improve its suitability for screening. The STWG believes the taxonomic range represented by these procedures is a vital characteristic of the battery, and that, further, the evaluation, necessary modifications, standardization, and validation for each assay should be accomplished as soon as possible.

Although several short-term in vitro and in vivo non-mammalian assays of EAT were identified, the STWG did not feel it was necessary to have a separate comprehensive battery like that presented for mammals. The group generally felt estrogens in mammalian assays would also be estrogenic in most non-mammalian systems as well. Rather than unnecessarily duplicate the assays in the screening process, it was our objective to include non-mammalian assays which could complement, rather than duplicate, the mammalian assays. For example, it is unlikely that in vitro binding assays of avian ER would yield results different from those obtained with mammalian ER. In contrast, the proposed amphibian metamorphosis assay provides an example where a simple developmental process, one which has no mammalian equivalent, can facilitate identification of chemicals that display (anti)thyroid activity. No simple mammalian assay of (anti)thyroid activity exists that is as specific as the amphibian metamorphosis assay.

A different strategy was adopted for the development of a fish EAT assay. This vertebrate class is farthest removed phylogenetically from mammals than the other vertebrate classes. The committee felt if significant differences existed between mammals and any other vertebrate class with respect to EAT activities then they would be reflected in fish. For example, fish ER differs from mammalian ER more than the ER of other classes. In addition, fish have some unique androgens and possibly different hormone receptors. Not only is there less conservation of structure but the function of the hormones often differs greatly between these two vertebrate classes. Part of the difficulty in designing a screening battery specifically for EAT in fish lies with the fact that this class is very diverse and in many cases, little is known about the role of these hormones. With this in mind, the STWG adopted the screening approach recommended by the EPA/KC Workshop on this issue, which was to develop a simple, yet comprehensive apical test, that would detect EAT in selected species. This assay includes exposure by injection and evaluation of gonadosomatic index (GSI), histology of the reproductive tract, and induction of vitellogenin in males.

The committee is also considering (anti)androgen assays in avian species to complement the mammalian assays. Two short-term assays for androgenicity in birds have been considered, including crop growth in chickens and proctodeal gland size in Japanese Quail. Several additional assays for non-mammalian vertebrates have been discussed by the STWG, but their role in screening versus testing has not been finalized. In avian and reptilian species, oviductal weight has been used to screen chemicals for estrogenicity in the same manner that the uterotrophic assay is used in rodents. This is a short-term assay that can detect estrogenicity and could be modified to detect anti-estrogens as well. Two fairly long-term developmental assays that are responsive to estrogens and aromatase inhibitors are being discussed. One involves in ovo injection in the Japanese quail, while the other examines the effects of xenobiotics on turtle temperature-dependent sex differentiation. In contrast to mammalian sex differentiation where androgens predominate, in these species estrogens and the aromatization of androgens normally plays a key role in the development of the female phenotype. Another consideration is to extend the length of exposure in the amphibian metamorphosis assay which would enable the detection of (anti)androgens and estrogens in a single assay.

1. Amphibians

Frog Metamorphosis: Tail resorption in frogs is under thyroid control with selective cell death specifically controlled by thyroxin. Thus, monitoring patterns of tail resorption, as well as the rate at which tail resorption occurs is a suitable indicator of thyroid hormone function. The South African clawed frog (Xenopus laevis) has been utilized in a short term (14-day) assay to evaluate the rate and behavior of tail resorption. This assay provides a comparatively simple procedure for evaluating thyroid active CSMs. Preliminary trials demonstrate this assay detects thyroid and anti-thyroid active compounds and, as a logical extension of the FETAX assay, is suitable for standardization. Modifications of this assay to include other morphological endpoints such as hind limb growth and eye migration, histopathology, and deiodinase activity are also under consideration. This assay could be extended to detect anti-androgens as well.

Additional Information: intact larval (tadpole) stage of amphibian Xenopus sp. (laevis), CSM in water; assess tail resorption, forelimb emergence, hindlimb development; duration in-life 14 days; assay is specific and sensitive for thyroid hormone, will detect thyroid/anti thyroid CSMs (it will miss certain classes of known thyroid hormone disruptors, e.g., PCBS; those classes are detected in the 14-day mammalian pubertal assay; in vivo no. 3). Limitations are solubility, toxicity, permeability to respiratory (gills, mouth, skin in adult) and/or gastrointestinal tract cells; organism provides amphibian-specific metabolism.

2. Fish

Fish In Vivo Screening Assay: Adult male and female fish are exposed to test chemicals in the water for 3-4 weeks at the beginning of gonadal recrudescence. The following primary measures will be taken 1) gonadosomatic index (GSI); 2) secondary sexual characteristics; 3) final oocyte maturation (FOM)/ovulation/ spermiation; 4) plasma sex steroids; and plasma vitellogenin. These endpoints were selected on the basis of their broad applicability and ease of routine measurement. This assay will respond to a broad range of EDCs including estrogen, anti-estrogen, anti-androgen, and anti-progestin. It is expected that this assay should detect endocrine disruption at any site of chemical interference on the hypothalamus-pituitary-gonadal-liver axis, including neuroendocrine effects.

Additional Information: intact male fish (species?) reared under short day conditions (primary and secondary sex organs/characteristics atrophied; "winter" status) exposure to CSM in water, evaluate recrudescence of male sex organs (organ weights), secondary sex characteristics, gonadosomal index (GSI), vitellogenin synthesis in liver. This assay is in an intact fish; therefore it is apical and sensitive to HPG axis effects. It evaluates a member of the most distant class, Osteichthyes, from Mammalia in vertebrates, in the T1S battery, so the degree of homology would be expected to be least with regard to E/A and/or T receptors, endogenous ligands, mechanism of action, etc. This assay will detect A/anti-A, and E. Limitations are solubility, toxicity, permeability to respiratory (i.e., gill and mouth) and/or gastrointestinal tract cells, organism provides fish-specific metabolism. This assay requires selection of appropriate species, and more validation and standardization than the other in vivo assays. The STWG strongly recommends development of this assay for inclusion in T1S battery. P> 3. Non-mammalian in vivo Assays Not Included

    i. The assay involving effects on temperature-dependent sex determination in reptiles (turtles) by "painting" the CSM on the eggs was not selected. This assay is comparably sensitive to E/A - anti E/A CSMs as the in vitro mammalian assays, can only be performed when eggs are laid (four months out of the year) and takes a long time in-life (approximately four months). It is a sensitive, specific assay for E/A - anti E/A in Reptilia, involving exposure during in ovo development, and is being considered for inclusion in the T2T battery.
    ii. Assays in Avian species, e.g., development of primary and secondary sex characteristics, including reproductive structures, after exposure during in ovo development (egg injection) were not selected. These assays are comparably sensitive to E/anti-E, A/anti-A as the mammalian assays and take a long time in-life (one-two months). They are being considered for inclusion in the T2T battery.
D. Invertebrate Assays

No invertebrate assays have been evaluated for use in a screening battery for detecting estrogen, androgen, or thyroid hormone disruption. It is recommended that a workshop of invertebrate endocrinologists and toxicologists be convened to address first, the suitability of invertebrate assays for estrogen and androgen (not thyroid) for use in a screening battery, and second, future improvements to the broader consideration of endocrine disruption in the environment and the utility of invertebrates as surrogate test organisms.

There are two aspects to considering endocrine disruption for invertebrates; one is relevance to the health of invertebrate organisms themselves and the other is relevance of invertebrates as surrogates for investigating vertebrate-related phenomena. Conventional risk assessment of toxic chemicals such as outdoor-use pesticides and high volume industrial chemicals generally include a crustacean reproduction or life cycle test in the data set used in the assessment. Although specific endocrine system endpoints are not considered, the apical nature of these tests may be adequate to detect the adverse consequences of an endocrine disrupting chemical in crustacean arthropods. Additional information is needed to determine what is most useful beyond these conventional tests for the wider invertebrate taxa. As surrogates, more information on the correlation of endocrine phenomena between invertebrates and vertebrates would be helpful. For instance, to what degree does a substance which disrupts ecdysteroid metabolism in crustacea disrupt sex steroid metabolism in vertebrates? Perhaps good correlations may be found, but more comparative information is needed before recommendations of specific invertebrate tests useful for evaluating potential endocrine disrupting activity relevant to vertebrates can be made. Please see Attached A, Endocrine Disruption and Invertebrates, for further discussion of invertebrate issues.

IV. Tier 1 Screening Battery - Option Two

1. 3-day Uterotrophic Assay (in vivo)

Assay for estrogenicity in vivo: uterine weight. Same as in the current version, except the route of administration would be intraperitoneal (ip). Also, uterine epithelial height should be measured in cases of equivocal results.

2. 14-day Intact Male Assay (in vivo)

Assay for thyroid, HPG axis, aromatase, steroid synthesis, 5-alpha-reductase, and anti-androgens. Young adult (70-90 day old) male rats will be used. They will be dosed daily with the test agent for 14 days, and sacrificed on the 15th. Measurements will include the following: 1) for male effects: testis weight; accessory sex organ weights; testis and epididymus histology; epididymal sperm concentration and motility; serum T, DHT, and estradiol; 2) for higher order effects: serum LH; and 3) for thyroid effects: thyroid histology; serum T4; and TSH. This assay is sensitive to steroidogenesis, therefore, there is no need to include an in vitro steroidogenesis assay.

In addition to these two in vivo assays, option three would include two in vitro assays (ER binding or reporter gene assays and AR binding or reporter gene assays); and, two non-mammalian in vivo assays (amphibian metamorphosis and fish in vivo assay).

V. Tier 1 Screening Battery - Option Three

1. Placental Aromatase Assay (in vitro)

One critical enzyme present at very low levels in the testis is aromatase, which converts testosterone to estradiol and is another P450 enzyme. Human placental aromatase is commercially available and could be used in vitro to assess the effects of toxicants on this enzyme fairly easily.

2. 20-day Thyroid/Pubertal Male Assay (in vivo)

In conducting the assay one should: initiate oral dosing of weanling male rats at 30 days of age (10 per group, selected for uniform body weights at 29 days of age to reduce variance); dose daily, seven days a week and examine daily for PPS; continue dosing until 50 days of age and necropsy males; optional endpoints include weighing body, liver (Ah and etc.), heart (thyroid), adrenal, testis (sperm counts), portions of epididymis (do sperm counts), seminal vesicle plus coagulating glands (with fluid), ventral prostate, levator ani plus bulbocavernosus muscles (as a unit); save thyroid for histopathology; take serum for T4, T3, testosterone and dihydrotestosterone analyses; optional biochemical and gene activation measurements could be taken as well; if T4 or T3 is altered, then measure TSH; and could run minced testis culture, ex vivo. One advantage of using the pubertal assay in the male rather than the female is that both anti-androgens and estrogens delay puberty in the male, while the female is unresponsive to anti-androgens. Pubertal alterations also result from chemicals that disrupt HPG function. For example, alterations of prolactin, growth hormone, GNRH, LH, or FSH secretion can alter pubertal maturation in the rat.

In addition to the two assays above, option three would include two other in vitro assays (ER binding or reporter gene assays and AR binding or reporter gene assays); one other mammalian in vivo assay (3-day Uterotrophic Assay); and two non-mammalian in vivo assays (amphibian metamorphosis and fish in vivo assay).

VI. Implementation of the Tier 1 Screening Battery

A. General Principles in the Application of a Weight of Evidence Approach to Evaluating the Results of Tier 1 Assays for Endocrine Disrupters

A weight of evidence (WOE) approach permits one to take into account, multiple features of test data, and results from multiple test methods, in order to reach a judgment on how likely a test substance is to be an endocrine disrupter and/or how likely it is to pose a hazard to exposed populations. WOE can be applied to results from a single test or to results from multiple tests.

Results of a single test may provide information on 1) the qualitative results (positive or negative), and if positive, 2) the doses at which an effect is induced, 3) the shape of the dose response curve, 4) the level of effect induced, and 5) the reproducibility of the effect observed. Positive test results showing reproducible, high levels of effects at relatively low doses (environmental or human exposure levels) are likely to be of greater concern than weak effects that are only observed at very high, perhaps toxic, levels of exposure.

Perhaps more importantly, WOE is useful in reaching a judgment when confronted with results for a substance that has been subjected to testing in multiple assays that involve different biological endpoints, in vitro and in vivo methods, different organisms, and varying proximity to the organism(s) and adverse health effects of concern. In such cases, factors routinely taken into consideration in determining the WOE include:

Regardless of the final recommendations of the STWG, it is clear that substances tested for their potential to be endocrine disruptors will be subjected to multiple tests, both in vitro and in vivo. Apart from substances that yield negative results in all assays, it is likely that each substance tested will produce a unique array of test results with regard to the factors listed above. Rather than attempting to contrive a set of fixed rules for determining a WOE, it is more practical to establish some general principles for reaching a WOE determination, including: ( greater weight = >)

Within the context of the EDSTAC/STWG activities, this WOE approach may be applied in at least two situations. First, based on the application of these principles to T1S objectives, substances for which endocrine disruption data already exist may be evaluated using a WOE approach for their need to proceed to T2T. Second, the WOE approach should be applied routinely to results emanating from the T1S scheme eventually adopted by the EPA. Results of the WOE would determine if the substance in question is not likely to be an endocrine disrupter, or should, as a potential endocrine disrupter, proceed to T2T.

B. Skipping Tier 1 Screening

The endocrine disruptor screening and testing program should be sufficiently flexible to allow a test sponsor to skip screening assays and initiate testing directly. Circumstances in which skipping screening may be advantageous to industry include the following: where the 2-generation reproduction and fertility assay is already required by regulation, e.g., for food pesticide registration; in cases for which screening results are highly likely to be positive; and for chemicals for which marketing considerations dictate the earliest definitive answer to the question regarding their status with respect to endocrine disruption. An advantage to the public and government accrues in these cases as testing would provide the ultimate answer regarding hazard due to endocrine disruption and could accelerate management of potential risks.

There is some concern about skipping screening assays whose function may not be captured in testing or which may be necessary to set dose levels for tests. The concern that the screening assays and tests may not be of equal sensitivity should not be of concern so long as the tests are conducted at the proper dose level (see section VII, D. on dose considerations). In the hierarchy of weight-of-evidence, the results of in vivo assays outweigh those of in vitro assays and tests outweigh screens. The tests account for absorption, distribution, and metabolism of the xenobiotic and are designed to register adverse effects from chemical exposure. More sensitive measures can frequently be added to standard test protocols to obtain information normally obtained through screening.

Unlike the screens which are primarily mammalian, the testing assays attempt to cover all taxa. Thus, given the current state of the science, the STWG is recommending the 2-generation assay in mammals, a fish life-cycle test (a one generation test spanning young adult to young adult), an avian reproduction test, and an invertebrate life cycle test. Since the T1S assays, in aggregate, provide preliminary information on the presence and nature (mechanism) of the potential endocrine disruptor and the species and sex at risk, in the absence of any such information (i.e., if T1S were not done) the broadest coverage in T2T would be mandated. Performance of these tests would need to be consistent with the principles governing T2T, i.e., all T2T would be performed if criteria were satisfied (e.g., presence in water and/or soil, persistence, in the environment, bioaccumulation, etc.).

VII. General Introduction to a Tier 2 Testing Battery

A. Introduction to Tier 2 Testing

The purpose of T2T is to characterize the nature, likelihood, and dose-response relationship of endocrine disruption in humans and wildlife. T2T is a complement to T1S. As already discussed, T1S is composed of a battery of in vitro and in vivo assays designed to detect the potential of a chemical to disrupt the endocrine system. The in vitro screening assays are highly sensitive and quite selective for a particular mode of action. They are, however, quite far removed from the biological complexity of an intact animal and may have a tendency to give false positive readings because, for instance, not all substances which bind to a receptor will cause an adverse biological effect, and not all endocrine disruptors act via the receptor. In vivo screens encompass the metabolic and response capability of a whole organism but focus on such a short time frame that the full effects of exposure to a chemical substance can not be identified and characterized. Since there is considerable biological conservation in the endocrine system, it is not necessary to screen in every major taxonomic group. Screens based on mammalian cell lines or intact animals will indicate the potential for adverse effects which must be characterized in longer term studies in a variety of species.

Tests are longer term studies designed to encompass critical life stages and processes so that the full biological consequences of chemical exposure can be identified and related to the dose or exposure which caused them. Effects associated with endocrine disruption may be latent and not visible until later in life or may not appear until the reproductive period is reached. Tests for endocrine disruption will encompass at least one generation including effects on fertility and mating, in utero or in ovo development, sensitive neonatal growth and development, and transformation from the juvenile life stage to sexual maturity. Unless there is evidence to limit the mammalian test to one-generation, mammals will be exposed to determine the effects of in utero and juvenile exposure to effects in a second generation.

While the STWG believes the two-generation tests are necessary to fully characterize potential effects of concern, it also acknowledges there may be instances when less comprehensive study designs would be adequate (dependent on available prior information). Some of the considerations for determining whether the comprehensive two-generation tests or alternative tests would be conducted include an understanding of: mechanisms of action; exposure scenarios; use patterns; and populations at risk.

B. Guidance for Selecting Tier 2 Tests

The Conceptual Framework believes existing information on biological effects and exposure and the results of T1S should be used to inform decisions regarding the selection and design of Tier 2 tests. T1S information may be of use in determining whether to add a satellite assay for thyroid effects for example but may be of limited value in determining the selection of assays (e.g. bird reproduction and/or fish life cycle, etc.) if only mammalian in vivo assays and mammalian cell lines are used in T1S. In this case, the choice of which Tier 2 tests to perform will hinge much more upon the physico-chemical and environmental release and exposure characteristics of the substance to be tested, rather than biological data from T1S.

As general guidance we can establish the following principles.

    1) Where use, exposure and release of a substance are well known, it may be possible to tailor T2T for particular exposure scenarios. The converse also holds. If the use, exposure and release information is poorly characterized, all Tier 2 tests would be triggered by positive results in T1S unless biological data indicate that certain taxa will not be affected.
    2) If a chemical is released to or can be predicted to potentially reach streams, rivers or freshwater lakes, a fish life-cycle test with a freshwater species and the invertebrate lifecycle test should be conducted. If the release is to a marine estuary, marine species should be substituted instead. If release is to both, the freshwater species is preferred.
    3) Pesticides with agricultural or other outdoor use, and chemicals that might be expected to bioaccumulate and biomagnify through the food chain or that present a potential risk to birds (e.g., water contamination) should be tested in the bird reproduction study.
VIII. Tier 2 Testing Battery

A. Rationale for Inclusion of the Two-Generation Reproductive Toxicity Study in T2T

The two-generation reproductive toxicity study in rats (TSCA 799.9380 [August 15, 1997]; OPPTS 870.3800 [Public Draft, February 1996]; OECD no. 416 [1983]; FIFRA Subdivision F Guidelines - 83-4 ) is designed to comprehensively evaluate the effects of a chemical on gonadal function, estrous cycles, mating behavior, fertilization, implantation, pregnancy, parturition, lactation, weaning, and the offsprings' ability to achieve adulthood and successfully reproduce. Through two generations, one litter per generation. Administration is usually oral (dosed feed, dosed water, or gavage) but other routes are acceptable with justification (e.g., inhalation). In addition, the study also provides information about neonatal survival, growth, development, and preliminary data on possible teratogenesis. The experimental design for a two-generation reproductive toxicity study is presented in Figure 1.

In the two-generation reproductive toxicity test, a minimum of three treatment levels and a concurrent control group are required. At least 20 males and sufficient females to produce 20 pregnant females must be used in each group as prescribed in this current guideline. The highest dose must induce toxicity but not excessive (i.e., < 10%) mortality. In this study, potential hormonal effects can be detected through behavioral changes, ability to become pregnant, duration of gestation, signs of difficult or prolonged parturition, sex ratio of the offspring, feminization or masculinization of offspring, number of pups, stillbirths, gross pathology and histopathology of the vagina, uterus, ovaries, testis, epididymis, seminal vesicles, prostate, and any other identified target organs. Table 1, attached, provides a summary of the endpoints that are evaluated within the framework of the experimental design of the two generation reproductive toxicity test (and some additions, still under consideration, to cover EAT concerns).

These observations are comprehensive and cover every phase of reproduction and development. Tests that measure only a single dimension or component of hormonal activity, (e.g., in vitro or short term assays) provide supplementary and/or mechanistic information, but cannot provide the breadth of information in Table 1, which is critical for risk assessment.

Additionally, in this study type, hormonally-induced effects such as abortion, resorption, or premature delivery as well as abnormalities and anomalies such as masculinization of the female offspring or feminization of male offspring can be detected. Substances such as the phytoestrogen, coumesterol, and the anti-androgen, cyproterone acetate, which possess the potential to alter normal sexual differentiation, were similarly detected in this study test system (i.e., 1982 Guideline). The initial prebreed exposure period (10 weeks) of the two-generation reproductive toxicity test also provides information on subchronic exposures which can be used for other regulatory purposes.

B. Less Comprehensive Study Design

While the STWG believes the two-generation reproductive toxicity test is necessary to fully characterize potential effects of concern, it also acknowledges there may be instances when a less comprehensive study design would be adequate (dependent on available prior information). Some of the considerations for determining whether the comprehensive two-generation reproductive toxicity test or an alternative test would be conducted include an understanding of: mechanisms of action; exposure scenarios; use patterns, and populations at risk. Some alternative, less comprehensive protocols are currently under consideration by the STWG.

C. Rationale for Inclusion of Non-mammalian Tests

T2T is the definitive phase of the screening and testing program and is intended to provide definitive information regarding endocrine disruption activity of a tested CSM. Primarily, this tier should assess the concentrations which elicit endocrine disruption (sensitivity) and the consequences of such disruption (relevance) to inform risk assessments. Whether the concept of "dose-response" is valid for the test material in question or not, should be addressed. For the non-mammalian tests proposed, the over riding consideration for their selection was tests which require little modification from existing standardized methods. Such studies are found which have been utilized for investigating developmental and reproductive endpoints for conventional toxicants, such as pesticides, for several years. These tests are considered adequate for evaluating the most obvious and relevant consequences of endocrine disruption, but they are limited in addressing the specific mechanisms involved. Developmental and reproductive toxicants operating through non-endocrine mechanisms may not be differentiable from those that do, but this would be immaterial to a risk assessment per se.

The STWG agrees that T2T should address at least four taxonomic groups, including birds, amphibians, fish, and invertebrates. It is recommended that the following standardized tests be used as a basic non-mammalian battery:

    1. Avian reproduction (with bobwhite quail and mallard)
    2. Fish life cycle (fathead minnow)
    3. Mysid life cycle (Americamysis)
    4. Amphibian development and reproduction (Xenopus)
Except for the amphibian study, these tests are routinely performed for chemicals with widespread outdoor exposures and expected to affect reproduction. The amphibian test is considered warranted because of the extensive fundamental knowledge base on amphibian development and reproduction.

1. Avian Reproduction (OECD 206; OPPTS 850.2300)

Two species are routinely used in evaluating avian reproduction due to the demonstrated differential sensitivity across avian species. The bobwhite quail and the mallard are the two species most commonly employed. The test evaluates the effects of a chemical exposure prior to the onset of maturation and egg laying, through the egg laying period, and then through the early development of the offspring. Endpoints include: eggs laid, eggs cracked, eggs set, viable embryos, live 3-week embryos, normal hatchlings, 14-day old survivors, eggshell thickness, body weight and food consumption, clinical signs, and gross pathological changes. The test can be extended with additional observations made for circulating steroid titers, sex ratio of offspring, organ and gland weights, histochemistry/histopathology, and reproductive viability of offspring.

Birds are not included in the screening battery as proposed and any CSMs with endocrine relevant activity detected, need to be evaluated in the testing battery with avian species to determine their activity and consequences of such activity in these homeothermic oviparous animals.

The reproduction test provides the most comprehensive evaluation of endpoints considered relevant to the estrogen, androgen, and thyroid hormone activities screened for in tier one. Other avian assays including the Japanese quail androgenic (foam, proctodeal, cloacal) gland, egg injection, draft OECD Japanese quail reproduction, and 2 generation avian reproduction tests were considered but not selected because of limited endpoints addressed or lack of accepted and standardized methods.

2. Fish Life Cycle Test (OPPTS 850.1500)

The fathead minnow is the recommended species to be used and is exposed from fertilization through development, maturation and reproduction, and early development of offspring with a test duration of approximately 300 days. The fathead minnow is recommended for use in the screening battery in the fish gonadal recrudescence assay, and as such, the significance of any activity detected in the screening assay should be evaluated in the definitive testing tier.

Fish are the most diverse and least homologous to the mammals of vertebrate animals. Reproductive strategies extend from oviparity, to ovoviviparity, to true viviparity. The consequences of an endocrine disruptor may be quite different across the many families of fishes. As a first step though, only a fathead minnow life cycle test is suggested to confirm and quantify any effects detected by the tier one battery. Subsequent tests with other species will then be a function of the risk assessment and nature of the hormones involved and effects expected.

3. Mysid Life Cycle Test (OPPTS 850.1350)

Invertebrates constitute the majority of the fauna, but the relevancy of EAT actions to these organisms and the availability of tests to evaluate such is limited. However, it is plausible that a CSM which interferes with estrogen or androgen actions could interfere with ecdysteroid activity which is an important steroid in arthropods. The mysid life cycle test would allow a determination of the relevancy of an EAT active material to the development, growth, and reproduction in this important group of invertebrates. The saltwater mysid is preferred to the freshwater daphnid because this species undergoes a full sexual reproductive cycle where the daphnid is parthenogenic in the standardized assay. Other invertebrates, such as the molluscs and echinoderms do have EA systems, but again relevant standardized tests for evaluating the consequences of interfering with these systems are not available.

4. Amphibian Development And Reproduction

For those CSMs which are detected in the tier one screen, a test for amphibians which basically begins with the standardized FETAX assay and extends through metamorphosis and reproduction would be necessary to evaluate the consequences of endocrine disruption in a poikilothermic oviparous vertebrate distinct from fishes.

D. Dose Considerations for Multiple Generation Reproduction Studies

The STWG, by general agreement, highly recommends additional new research to resolve existing controversy about the nature of the dose response curve for endocrine active substances, particularly with regard to the low dose region. A collaborative program involving government, industry and appropriate individuals in academia is recommended. Several endocrine active substances should be selected to systematically evaluate:

    1. Selected potential effects on males (e.g., prostate weight, sperm counts, etc.) and females (e.g., ovarian follicular development) exposed in utero over a wide dosage range;
    2. Whether or not effects (if any) persist throughout the entire lifetime of the species tested and the long-term significance of any effects observed (i.e., are they adverse to the long-term health of the animals);
    3. Nature of the dose-response curve for any effects observed, with a particular focus on the low dose region
    4. Species and sex comparisons between rats and mice
The study would require a multi-stakeholder group involving representatives from the various stakeholder organizations to design the protocol, be kept abreast of the conduct of the study, evaluate results, and develop overall conclusions and recommendations.

In addition, during the STWG discussions, the group discussed a number of additional proposals that might help to address the controversy about dose response curves on an interim basis prior to completion of the above research. The proposals were discussed in relation to mammalian tests, however, the same proposals might need to be given some consideration when discussing non-mammalian tests. Numerous implementation questions exist for each of these options and would need to be identified in detail, and subsequently addressed, before a recommendation could be forwarded to the EDSTAC. All of the following recommendations would include formation of the multi-stakeholder process to evaluate the low dose issue (discussed above) until sufficient data exist to determine future recommendations:

IX. Standardization, Validation, Methods Development, and Research

A. Categorization of Assays and Tests

As mentioned throughout the Report, the STWG believes each assay still under consideration for inclusion in T1S or Testing T2T needs some level of standardization, validation, methods development, or further research before the assay can be recommended as a regulatory toxicity screen or test. The level of standardization and validation varies according to a variety of criteria applied to each of the assays including: period of time in use; existing level of general acceptance in the endocrine toxicology field; and, existing understanding of relevancy and reliability. Though the STWG has not developed final recommendations on a standardization and validation program, STWG has reached agreement that the assays proposed for T1S and T2T will fall into four general categories with regard to the level of standardization, validation, or methods development required.

Category I: is comprised of those assays which have been in use for a sufficient period of time and which have gained sufficient general acceptance within the field of endocrine toxicology to be considered de facto validated (reliable AND relevant). These assays measure relevant endpoints, are responsive to endocrine active compounds with a high degree of specificity, are sufficiently sensitive to identify all known active agents, and can reasonably be expected to give reproducible results from laboratory to laboratory, assuming a general level of competence and expertise. Nonetheless, nuances of protocol may exist within the field for these assays. Therefore, standardization of the protocol to be recommended for these assays should be accomplished by EPA before these assays are implemented as screening requirements for endocrine activity or disruption.

Category II: is comprised of those assays which have sufficiently broad use to have generally considered relevant OR reliable to either screening for endocrine activity (tier 1) or to testing for adverse endocrine mediated effects (tier 2). These assays cannot, however, be generally considered to be both relevant AND reliable. The level of performance that can be expected of these assays with respect to identifying endocrine active agents or endocrine disruptive effects of chemicals must be clarified. Therefore, these assays should undergo further but focused validation and standardization to define their relevance and reliability for the task of endocrine disruptor screening or testing. The validation required may be focused to answer specific questions about relevance and to provide information regarding specificity and sensitivity.

Category III: is comprised of those assays which may have relevance to the task of either screening for endocrine activity or testing for endocrine disruptive effects, but whose performance in identifying endocrine active agents or endocrine disruptive effects of chemicals has not been tested. Questions as to whether these assays measure endpoints that are relevant to endocrine activity or endocrine disruptive effects, whether these assays respond with specificity and sensitivity to known endocrine active agents or to identify endocrine disruptive effects cannot be addressed with information currently available. In addition, questions regarding the specific protocols and conditions under which the assays should be conducted must be answered before relevance and reliability can be assessed. Nonetheless, the work group feels that these assays would have sufficient utility, if further developed and validated, to enhance or augment the screening and testing program. Therefore, the work group recommends that resources be made available to pursue methods development and validation and standardization of these assays.

Category IV: is comprised of those assays which if available, could have an important utility in the screening and testing program. However, such assays have not actually been conducted. Therefore, the work group recommends that research be conducted to determine whether such assays can be developed, and if so, what purpose the assays could fulfill within the endocrine disruptor screening and testing program. The STWG has identified a number of assays and tests that they would recommend be placed in this category.

Category IV Assays: Further Research, Development, Validation & Standardization

B. Instituting a Validation Program

The STWG believes one key step in instituting a validation program for T1S, is the identification of a set of chemicals to be used as the Òtest materialsÓ in each assay. The STWG is currently considering whether they should develop such a list; however, an integral step to accomplish prior to releasing such a list is identification of criteria by which the chemicals would be chosen. Some work group members have suggested the following criteria might be helpful in developing a list:

In addition, the STWG recognizes the need to carefully define the expected use of the list in order to avoid inappropriate use. STWG members believe such a list, developed with the already mentioned criteria in mind, would be used in the validation program to assist in defining their relevance and reliability for the task of endocrine disruptor screening, i.e., to identify whether a specific CSM is a potential endocrine disruptor or can be placed in the Òhold unless ...Ó box.

In addition, as was also stated earlier, it is critical to acknowledge the state-of-the-science in this area is evolving rapidly and assays currently being developed, or may be developed in the future, may offer distinct advantages over some included in the current options. As they are developed, validated, and standardized, the use of these new assays for screening is strongly encouraged.

Other pieces to be added in this section of the Report include an overview of suggested steps required for standardization, and components of the validation process based upon the report of The Ad Hoc Interagency Coordinating Committee on the Validation of Alternative Methods (NIEHS, 1997).

Figure 1
Two - Generation Reproduction Toxicity Design
..... Direct exposure via diet, drinking water, inhalation, etc.
..... Possible indirect exposure from transplacental and/or translactational exposure
..... Both direct and possible indirect exposure if in feed or water (nursing pups also self-feeding and drinking)

Table 1
Two-Generation Endpoints

Below are lists of endpoints measured in current EPA guideline studies, plus additions proposed by the STWG, which will detect estrogen, androgen, and thyroid hormone perturbations.

Endpoints Sensitive to Estrogens/Antiestrogens

Proposed Additions Under Consideration: Endpoints Sensitive to Androgens/Antiandrogens

Endpoints Sensitive to Thyroid Hormone Agonists/Antagonists (general)

Proposed Additions Under Consideration:

Attachment A
Endocrine Disruption and Invertebrates

Considerations by the EDSTAC Screening and Testing Work Group (STWG) have predominantly dealt with vertebrate animals for several reasons. The first, and perhaps overriding one, is that the charge given to the work group of focusing on estrogen, androgen, and thyroid hormone actions is not especially relevant to important and well-studied hormones of invertebrates. The purported endocrine disruption effects of public concern are almost exclusively human health or vertebrate wildlife related. The expertise in the work group is, also, predominantly with the vertebrate classes. However, invertebrates represent over 95% of all animals, are ubiquitous, and are tremendously important ecologically and economically. Commercial fisheries of shrimp, crab, and oyster and agriculturally important insect pollination are but a few key examples. Because invertebrates are ubiquitous and are easily adapted for laboratory testing, they can serve as sentinels and surrogates for investigating environmental stress. For these reasons, invertebrates should not be ignored from consideration.

Endocrine disruption has been well studied and well exploited for certain invertebrates, especially the insects. The endocrine systems of insects have been intentionally targeted for insecticidal activity and several insecticides have been developed and used to suppress insect populations by disrupting their normal endocrine functions. Juvenile hormone mimics (e.g., methoprene), antijuvenile hormone analogs (e.g., precocene), chitin synthesis inhibitors (e.g., diflubenzuron), ecdysone analogs (e.g., tebufeno-zide), and molting disruptants (e.g., fenoxycarb) are some examples. These insect growth regulating compounds have also been observed to have adverse effects in related arthropods such as crustaceans, including disrupting normal molting processes, limb regeneration, and reproduction (Christiansen et al. 1977a,b, 1979; Cunningham 1976; Forward and Costlow 1978; Landau and Rao 1980; Nimmo et al. 1980; Touart and Rao 1987). Other substances like the organotin TBT have caused imposex and intersex conditions in gastropods (Gibbs and Bryan 1986; Reijnders and Brasseur 1992) and sewage outfalls have caused intersex conditions in harpacticoid copepods (Moore and Stevenson 1994), conditions indicative of endocrine disruption.

Although the relevance of estrogen and androgen hormones to invertebrates is unclear, invertebrates may be useful as surrogates for investigating phenomena relevant to these hormones in vertebrates. Estrogens have been reported to play a meaningful role in development and reproduction in echinoderms and molluscs (Takeda 1979; Brueggemeier et al. 1988; Shirai and Walker 1988). Daphnids have been used to investigate the effects of xenoestrogens on steroid metabolism (Baldwin et al. 1995; Baldwin et al. 1997) and sex reversal (Shurin and Dodson 1997). Because of their generally shorter life cycles and relative ease of handling many species in the laboratory, invertebrates could be useful for evaluating endocrine disrupting phenomena. However, additional research is needed before this promise is realized.

There are, therefore, two aspects to considering endocrine disruption for invertebrates, one is relevance to the health of invertebrate organisms themselves and the other is relevance of invertebrates as surrogates for investigating vertebrate-related phenomena. Conventional risk assessment of toxic chemicals such as outdoor use pesticides and high volume industrial chemicals generally include a crustacean reproduction or life cycle test in the data set used in the assessment. Although specific endocrine system endpoints are not considered, the apical nature of these tests may be adequate to detect the adverse consequences of an endocrine disrupting chemical in crustacean arthropods. Additional information is needed to determine what is most useful beyond these conventional tests for the wider invertebrate taxa. As surrogates, more information on the correlation of endocrine phenomena between invertebrates and vertebrates would be helpful. For instance, to what degree does a substance which disrupts ecdysteroid metabolism in crustacea disrupt sex steroid metabolism in vertebrates? Perhaps good correlations may be found, but more comparative information is needed before recommendations of specific invertebrate tests useful for evaluating potential endocrine disrupting activity relevant to vertebrates can be made.

No invertebrate assays, therefore, have been evaluated for use in a screening battery for detecting estrogen, androgen, or thyroid hormone disruption. It is recommended that a workshop of invertebrate endocrinologists and toxicologists be convened to address first, the suitability of invertebrate assays for estrogen and androgen (not thyroid) for use in a screening battery, and second, future improvements to the broader consideration of endocrine disruption in the environment and the utility of invertebrates as surrogate test organisms.

Literature Cited

Baldwin, W. S., D. L. Milam, and G. A. LeBlanc. 1995. Physiological and biochemical perturbations in Daphnia magna following exposure to the model environmental estrogen diethylstilbestrol. Environ. Toxicol. Chem. 14:945-952.

Baldwin, W. S., S. E. Graham, D. Shea, and G. A. LeBlanc. 1997. Metabolic androgenization of female Daphnia magna by the xenoestrogen 4-nonylphenol. Environ. Toxicol. Chem. 16:1905-1911.

Brueggemeier, R. W., G. D. Yocum, and D. L. Denlinger. 1988. Estranes, androstanes, and pregnanes in insects and other invertebrates. In: Physiological Insect Ecology, F. Sehnal, A. Zabza, and D. L. Denlinger (eds.) Wroclaw Technical University Press, Wroclaw, Poland pp.885-898.;

Christiansen, M.E., J. D. Costlow, Jr., and R. J. Monroe. 1977a. Effects of the juvenile hormone mimic ZR-515 (Altosid) on larval development of the mud-crab Rhithropanopeus harrisii in various salinities and cyclic temperatures. Marine Biol 39:269-279.

Christiansen, M.E., J. D. Costlow, Jr., and R. J. Monroe. 1977b. Effects of the juvenile hormone mimic ZR-512 (Altozar) on larval development of the mud-crab Rhithropanopeus harrisii at various cyclic temperatures. Marine Biol 39:281-288.

Christiansen, M.E., J. D. Costlow, Jr., and R. J. Monroe. 1979. Effects of the insect growth regulator Dimilin (TH-6040) on the larval development of two estuarine crabs. Marine Biol. 50:29-36.

Cunningham, P.A. 1976. Effects of Dimilin (TH-6040) on reproduction in the brine shrimp Artemia salina. Environm. Entomol. 5:701-706.

Gibbs, P. E., and G. W. Bryan. 1986. Reproductive failure in populations of the dog-whelk, Nucella lapillus, caused by imposex induced by tributyltin from antifouling paints. J. Mar. Biol. Assoc. UK 66:767-777.

Forward, R.B., Jr. and J. D. Costlow, Jr. 1978. Sublethal effects of insect growth regulators upon crab larval behavior. Water, Air, Soil Pollution 9:227-238.

Landau, M. and K. R. Rao. 1980. Toxic and sublethal effects of precocene II on the early developmental stages of the brine shrimp Artemia salina (L.) and the barnacle Balanus eburneus Gould. Crustaceana 39:218-221.

Moore, C. G. and J. M. Stevenson. 1994. Intersexuality in benthic harpacticoid copepods in the Firth of Forth, Scotland. J. Nat. History 28:1213-1230.

Nimmo, D.R., T. L. Hamaker, J. C. Moore, and R. A. Wood. 1980. Acute and chronic effects of Dimilin on survival and reproduction of Mysidopsis bahia. In: Aquatic Toxicology, ASTM 707, J. G. Eaton, P. R. Parrish, and A. C. Hendricks (eds.) American Society for Testing and Materials, Philadelphia, PA. pp. 366-376.

Reijnders, P. J. H. and S. M. J. M. Brasseur. 1992. Xenobiotic induced hormonal and associated disorders in marine organisms and related effects in humans; an overview. In: Chemically Induced Alterations in Sexual and Functional Development: The Wildlife/Human Connection, T. Colburn and T. Clement (eds.) Princeton Scientific Publishing Co., Inc., Princeton, NJ pp. 159-174.

Shurin, J. B. and S. I. Dodson. 1997. Sublethal toxic effects of cyanobacteria on nonylphenol on environmental sex determination and development in Daphnia. Environ. Toxicol. Chem. 16:1269-1276.

Shirai, H. and C. W. Walker. 1988. Chemical control of asexual and sexual reproduction in echinoderms. In: Endocrinology of Selected Invertebrate Types, H. Laufer and G. H. Downer (eds.) Alan R. Liss, Inc., New York pp. 453-476.

Touart, L. W. and K. R. Rao. 1987. The influence of diflubenzuron on survival, molting and limb regeneration in the grass shrimp, Palaemonetes pugio. In: Pollution Physiology of Estuarine Organisms, W. Vernberg, W. Calabrase, F. Thurberg, and J. Vernberg (eds.) University of South Carolina Press, Columbia, SC. pp. 333-349.

1 The relative sensitivities of various hormone responsive endpoints, within each species, needs to be established, and the relative sensitivities of each species needs to be assessed to determine which, if any, are required for T1S assays.


Local Navigation


Jump to main content.