Page 296

11—
Screening and Monitoring

There are no generally accepted, validated methods to screen for or monitor exposure to chemicals that could cause adverse hormonal activity—largely because of the complexity of the endocrine system. Assays and end points for testing and monitoring have been proposed (Gray et al. 1997; OECD 1997). but many of them are specific for the effects of hormone mimics, especially estrogenlike compounds (ECETOC 1996). and they do not address the full range of putative hormonally active agents (HAAs). HAAs can act directly, at a specific receptor, or indirectly, and their effects can vary, depending on the target tissue, the timing of exposure, and interactions with other HAAs. Because HAAs can mimic or modulate the activity of endogenous hormones, it can be difficult to distinguish altered responses from the range of basal hormone-regulated responses. For this reason, various U.S. and European groups have urged that currently used tests be assessed for their adequacy to detect the effects of HAAs (Kavlock et al. 1996; Ankley et al. 1997: EPA 1997). EPA has organized the Endocrine Disruptor Screening and Testing Advisory Committee (EDSTAC) to recommend screening and testing guidelines appropriate for chemicals with endocrine-disrupting activity. Congress has mandated a response from EPA by March 1998 (see Addendum).

No single assay can accurately predict all the effects of HAAs. In vivo assays are often unsuitable for large-scale screening because of their relatively high cost, low sensitivity, and labor intensiveness. Moreover, in vivo assays that assess highly complex responses can be modulated through other mechanisms and, therefore, might not be selective for the substances of interest. Screening and monitoring approaches for hazard assessment of HAAs should be designed to assess HAAs alone and in combination with other HAAs that modulate the response of the primary compound of interest (Kavlock et al. 1996; Patlak 1996:continue



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 296
Page 296 11— Screening and Monitoring There are no generally accepted, validated methods to screen for or monitor exposure to chemicals that could cause adverse hormonal activity—largely because of the complexity of the endocrine system. Assays and end points for testing and monitoring have been proposed (Gray et al. 1997; OECD 1997). but many of them are specific for the effects of hormone mimics, especially estrogenlike compounds (ECETOC 1996). and they do not address the full range of putative hormonally active agents (HAAs). HAAs can act directly, at a specific receptor, or indirectly, and their effects can vary, depending on the target tissue, the timing of exposure, and interactions with other HAAs. Because HAAs can mimic or modulate the activity of endogenous hormones, it can be difficult to distinguish altered responses from the range of basal hormone-regulated responses. For this reason, various U.S. and European groups have urged that currently used tests be assessed for their adequacy to detect the effects of HAAs (Kavlock et al. 1996; Ankley et al. 1997: EPA 1997). EPA has organized the Endocrine Disruptor Screening and Testing Advisory Committee (EDSTAC) to recommend screening and testing guidelines appropriate for chemicals with endocrine-disrupting activity. Congress has mandated a response from EPA by March 1998 (see Addendum). No single assay can accurately predict all the effects of HAAs. In vivo assays are often unsuitable for large-scale screening because of their relatively high cost, low sensitivity, and labor intensiveness. Moreover, in vivo assays that assess highly complex responses can be modulated through other mechanisms and, therefore, might not be selective for the substances of interest. Screening and monitoring approaches for hazard assessment of HAAs should be designed to assess HAAs alone and in combination with other HAAs that modulate the response of the primary compound of interest (Kavlock et al. 1996; Patlak 1996:continue

OCR for page 296
Page 297 Ankley et al. 1997). In vitro measurements can be made in screening (a priori) and in monitoring (a posteriori) modes to measure or predict the activity of HAAs, and several methods to do that have been proposed (ECETOC 1996). The discussion presented here will not attempt to recapitulate all proposed methods; they are reviewed elsewhere (ECETOC 1996: Gray et al. 1997; OECD 1997; Zacharewski 1997). This chapter focuses on principles and strategies for selecting appropriate in vitro and in vivo testing systems. The discussion is restricted to general methods of screening and monitoring. Special attention is given to HAAs that mimic steroid sex hormones, generally defined as "estrogens" and "androgens," and the emphasis is on vertebrates, especially mammals. As additional information is obtained on the functions and dynamics of the endocrine system (particularly during development), new assays can be developed to identify biologic markers of exposure. Screening Tools The need to screen a large number of compounds for their potential to cause hormonal activity is mandated by law (Safe Drinking Water Act of 1997; Food Quality Protection Act of 1996). Several task forces have stressed the need for tests that predict in vivo responses in humans (Kavlock et al. 1996) and wildlife (Ankley et al. 1997). The methods available to screen for HAAs include biologic and instrumental chemical analyses. The former category includes assays that can be used to screen for potential hormonal activity in individual compounds, formulations, or environmental extracts. Extensive reviews of standard in vitro and in vivo tests to screen for estrogenic activity have been published (Reel et al. 1996; Zacharewski 1997). This section reviews mammalian and nonmammalian in vitro and in vivo assays and focuses on their advantages and limitations for identifying and assessing estrogenic substances. In Vitro Assays In vitro assays available for measuring receptor-mediated activities of HAAs include simple receptor-binding, cell-proliferation, gene-expression, and gene-product assays. Most have been developed for determining estrogenic activity. and there is a need for development and validation of in vitro assays for other hormone-receptor systems. In vitro systems are attractive as screening tools because they are rapid, inexpensive, and reproducible, and estimates of the relative potency of a large number of samples or compounds can be obtained rapidly. In vitro assays are also excellent models for investigating the mechanism of action of HAAs and the interactions between endocrine-response pathways. In vitro assays are limited in that they cannot represent in vivo conditions. The pharmacokinetics, biotransformation, and binding to carrier proteins of a compound might or might not be represented accurately by in vitro systems.continue

OCR for page 296
Page 298 There are methods to minimize some of the limitations and to increase the accuracy of in vitro assays. In addition, all in vitro assays that are based on the biologic properties of established cell lines require special care to maintain the original cell phenotype (Laursen et al. 1990; Ruedl et al. 1990; Zugmaier et al. 1991; Wiese et al. 1992; Villalobos et al. 1995). Some in vitro systems that are used to screen for HAAs are described below. Table 1-1 summarizes the advantages and disadvantages of those most commonly used. The selection of assays should be based on their responsiveness to the HAA of interest. There will never be a single in vitro assay capable of screening for a wide range of HAAs simultaneously. Receptor Binding Binding of agonists or antagonists to a receptor is required for direct-acting receptor-mediated HAAs to exert an effect (Clark and Mani 1994; Kramer and Giesy 1995). Therefore, the receptor-binding affinity of a compound could predict ligand potency relative to the endogenous ligand for the specific receptor. For this reason, receptor-binding assays have been used to screen for putative HAAs (R. White et al. 1994; Kelce et al. 1995). Relative receptor-binding affinities of various ligands are routinely determined in competitive assays in which receptor preparations are coincubated with a high-affinity radioligand and with different concentrations of a test compound or mixture. Analysis of kinetic-binding data can be done to determine half-inhibitory concentration (IC50) values, which define the concentration of the test compound required to displace 50% of the radioligand. Receptor preparations from diverse species and from estrogen-responsive tissues, as well as recombinant proteins, have been used for these assays. In vitro assays, such as estrogen-receptor (ER) binding, are invaluable as screening methods; however, they have several limitations: Structure-receptor binding might not always parallel structure-activity relationships; receptor binding is determined in the absence of pharmacokinetic or metabolic effects (Bitman et al. 1978; Bulger and Kupfer 1985; Kramer et al. 1997); receptor-binding assays have not been developed to distinguish between receptor agonists and antagonists; and the sensitivity of the assay is low (half-maximal effective concentration [EC50] for estradiol is in the 300-pM to l-nM range) (Wooge et al. 1992; Soto et al. 1995). Recent studies have identified a second form of the ERa ERß, which exhibits a pattern of expression different from the ER pattern (Kuiper et al. 1996; Mosselman et al. 1996). There are significant structural differences between ERa and ERß proteins. However, Kuiper et al. (1998) evaluated the relative binding affinities of 60 estrogenic compounds to the two ER subtypes and found that only a few compounds exhibited major differences in binding affinity to ERa and ERß. The multiple and variant (Castles et al. 1995) forms of the ER, as well as different tissue-specific expressions of these forms, could be important determinants in the tissue-specificcontinue

OCR for page 296
Page 299 TABLE 11-1 Selected In Vitro Assays Assay Advantages Disadvantages Reference Receptor binding Large-scale screening; direct receptor Cannot distinguish between agonists and Jensen and Jacobson 1960;   interaction: requires 1 d or less; yeast or antagonists: cannot detect proestrogens; Ireland et al. 1980   produce baculovirus system can be used requires radioligand: serum-binding     to milligram quantities of receptor proteins absent; no in vivo       pharmacokinetics; artifactual competition       by chemicals at high concentrations due       to detergent-denaturation effects.   Cell proliferation Simple to perform; large-scale screening; Only a few cell lines are estrogen Soto et al. 1990, 1991, 1992a,   animals not required; direct cell exposure; responsive: 4-6 d incubation period; 1994, 1995; Villalobos et al.   few false positives or negatives; sensitive cannot detect all proestrogens; no in 1995; Welshons et al. 1990   detection; can distinguish between vivo pharmacokinetics     agonists and antagonists (partial and       full); detects some proestrogens     Receptor-dependent gene Large-scale screening; wide variety of Expression of some proteins and enzyme Jobling and Sumpter 1993; expression cell-line origins can be used activities can be restricted to some cell Heppell et al. 1995; Jordan et     lines; response can be susceptible to al. 1985: Littlefield et al. 1990:     mechanisms that do not involve the Pelissero et al. 1993; Soto et al.     estrogen receptor; compounds might not 1995     have consistent tissue- or species-specific       responses; cannot detect proestrogens   Recombinant receptor- Large-scale screening: direct cell No standardized assay system generally Mayr et al. 1992; Miksicek reporter gene constructs exposure; internal control reporter gene available; cannot detect proestrogens; no 1993: R. White et al. 1994   to monitor toxicity; some require 1 d; in vivo pharmacokinetics; some cannot     animals not required; can use transfected distinguish between agonists and     mammalian or yeast cells; some can antagonists: yeast cell walls can affect     distinguish between agonists and the uptake of xenoestrogens     antagonists     SOURCE: Adapted from Reel et al. 1996.

OCR for page 296
Page 300 action of ER agonists and antagonists. This could also be important for the different isoforms of other receptors of other HAAs. Cell Proliferation The ability of estrogens to induce cellular proliferation in target organs is considered a hallmark of estrogen action (Hertz 1985). Therefore. a reliable bioassay for assessing estrogenicity would measure cell proliferation as an end point. This measurement can be done in vitro by using established cell lines derived from estrogen-responsive target organs, including rat pituitary cells and several human breast cancer cell lines, such as MCF7 and T47-D cells. MCF7 cells have been used extensively to screen for estrogen agonists and antagonists (Welshons et al. 1990; Soto et al. 1991, 1992a, 1994; Sonnenschein et al. 1994, 1995). MCF7 cells express the ER (Brooks et al. 1973), which regulates several critical genes as well as estrogen-dependent cell proliferation (Katzenellenbogen et al. 1984; Soto and Sonnenschein 1984). MCF7 cells become quiescent when placed in an estradiol-deficient culture medium supplemented with dextran-coated charcoal-absorbed serum. Estrogens release these cells from quiescence. The estradiol EC50 is in the 10- to 15-pM range (Soto et al. 1997). Results obtained using these cells were considered to be highly predictive of estrogenicity (Andersen et al. 1999). Receptor-Dependent Gene Expression Assays that test the ability of a compound to stimulate receptor-dependent responses (in genes or proteins) are routinely used to determine the estrogenic or antiestrogenic potency of HAAs in several cell lines. Genes or gene products that have been used include progesterone receptor, pS2, alkaline phosphatase, cathepsin D, prolactin, and vitellogenin (Jordan et al. 1985; Littlefield et al. 1990: Adlercreutz et al. 1992; Jobling and Sumpter 1993: Pelissero et al. 1993; Heppell et al. 1995: Soto et al. 1995). Although these genes and gene products arc induced by estrogenic compounds, the induction responses are often specific to a target organ or cell and can be induced by several HAA classes. Molecular biology techniques have been used extensively to develop an array of in vitro assays that are hormone-responsive promoter-reporter constructs that can be transiently or stably transfected into diverse mammalian cell lines. The cells also can be transiently or stably transfected with wild-type or variant hormone receptors or chimeric receptors (Pons et al. 1990; Gagne et al. 1994; Jausons-Loffreda et al. 1994; Mäkelä et al. 1994; Jobling et al. 1995; Miksicek 1995; Ruh et al. 1995; Zacharewski et al. 1995; Ramamoorthy et al. 1997:. Various promoter sequences and reporter genes have been used for these assays, and some of the most sensitive constructs contain multiple or tandem estrogen-responsive-element (ERE) motifs, which are strong enhancer elements.break

OCR for page 296
Page 301 Estrogen responsiveness in cells that express wild-type ERs depends on the interaction of activation function 2 (AF-2), AF-1, or both domains of the ER with other nuclear proteins. The complex interactions of hormone receptors and nuclear coactivators, repressors, and accessory proteins are important in the ligand-dependent and cell-specific induction of various promoter-reporter constructs and their related genes (Katzenellenbogen et al. 1996). Ligand-dependent differences in the functions of wild-type and ER variants have been reported for different classes of ER agonists and antagonists (McDonnell et al. 1995), and it is likely that application of these techniques will demonstrate functional differences between estrogenic HAAs (Gould et al. 1998). Chimeric-receptor and reporter-gene assays also have been developed to bypass the requirement for many nuclear factors. For example, Zacharewski et al. (1995) used a Gal4-HEGO chimeric receptor, which contains the ER ligand-binding domain fused to the DNA-binding domain of the Gal4 yeast transcription factor, and a Gal4-regulated luciferase reporter-gene construct, which contains copies of the Gal4 DNA-binding motif. These constructs have been used in transiently and stably transfected cells to detect estrogenic and antiestrogenic HAAs (Zacharewski et al. 1995; Connor et al. 1996; Moore et al. 1997). Recombinant estrogen-based yeast assays also can be used for rapid screening of hormone-receptor agonists (Klein et al. 1994; Routledge and Sumpter 1996; Ramamoorthy et al. 1997). Cotransfection of genes encoding various P450 enzymes involved in metabolic activation and inactivation of HAAs will increase the utility of these receptor assays. Prediction of In Vivo Effects from In Vitro Assays The usefulness of in vitro techniques is hampered by their inherent simplicity. In vivo, there are multiple cell types in tissues that communicate via extracellular signals, such as hormones, growth factors, and cytokines. However, in vitro assays use cloned cell lines; that rules out interactions among cell types (Kao et al. 1996) and thus modifies the potential of cells to respond to hormones. Additionally, the cell types used in in vitro systems might not have or might have lost the ability to metabolize certain compounds. In vivo, some compounds need to be activated to cause hormonal modulation, and others can be deactivated by being metabolized, congregated, and or excreted. Most in vitro tests are based either on a cellular response, such as cell proliferation, or on gene expression. Historically, the understanding of steroid or thyroid hormonal function has been based on the genomic theory that hormones function through regulation of nuclear transcription complexes by intracellular steroid-binding proteins. That results in a steroid-effect cascade. However, recent studies have shown that nongenomic effects can occur (Wehling 1994). For example, o,p'-DDT is a weak ER agonist (i.e., it rapidly dissociates from the receptor) in vitro (Soto et al. 1994), and it has been difficult to reconcile thecontinue

OCR for page 296
Page 302 estrogenlike effects of o,p'-DDT observed in wildlife (Fry et al. 1987; Fry 1995) with the concentrations that have been shown to elicit estrogenlike effects in vitro (Donahoe and Curtis 1996). Ligand-independent activation of ER-mediated responses via tyrosine-kinase pathways have been reported (Kato et al. 1995), and steroid hormonal action might also involve post-transcriptional regulation of mRNA (Hadcock and Malbon 1991; Martin et al. 1994). Another mechanism that can lead to differences between in vitro and in vivo potency of HAAs relative to estradiol is differential binding to plasma proteins. For example, o,p'-DDT shows little binding to estrogen-binding plasma proteins (Skalsky and Guthrie 1978), a factor that will increase the effective free fraction of o,p'-DDT in vivo relative to estradiol (Nagel et al. 1998). Thus, because of the limitations of in vitro assays and because the precise mechanisms of hormonal action have not been determined, it is not always possible to predict in vivo effects from results of individual in vitro assays. Therefore, in vitro bioassays should utilize a battery of simple assays based on different mechanisms of action. In Vivo Assays Despite their limitations for use in large-scale screening of compounds or extracts, estrogen-responsiveness assays in rodents are important test procedures. In vivo assays measure several end points, including organ weight, cell differentiation, protein and gene expression, and enzyme activity (Lan and Katzenellenbogen 1976; Lyttle and DeSombre 1977; Dix and Jordan 1980; Cooper et al. 1992; Branham et al. 1993; Medlock et al. 1994; Sheehan et al. 1994; Heppell et al. 1995; Sumpter and Jobling 1995; Teng 1995). The advantages and disadvantages of various in vivo tests are discussed below. In vivo tests for detecting HAA-induced effects on reproduction and development use a complete biologic system that accounts for pharmacokinetics (i.e., absorption, disposition, metabolism, and excretion) of a test compound. Moreover, in vivo testing is done in the presence of physiologically relevant types and concentrations of endogenous hormones, hormone-binding proteins, and accessory factors. Repair or defense systems that might be absent in an in vitro system will be present in an in vivo system. Thus, results of in vivo assays are relevant for extrapolating to wildlife or humans. Many testing protocols have been standardized, validated, and in some cases, published as official guidelines (EPA 1996; OECD 1997). The regulatory protocols that relate particularly to HAAs are the two-generation reproduction, developmental toxicity, subchronic toxicity, and chronic toxicity studies (Stevens et al. 1997). These tests are designed to detect adverse effects on reproductive function or development, regardless of mechanism. Response end points monitored in multigenerational reproduction studies include fertility; litter size; and weight, survival, and growth of offspring. All ofcontinue

OCR for page 296
Page 303 these responses can be altered by exposure to HAAs. Other end points, such as vaginal cyclicity, reproductive-organ weight, gonadal morphology, accessory-sex-organ weight, sperm count, and anogenital distance could be affected by HAA activity, but the changes are not necessarily specific to any particular hormone or portion of the endocrine system. Hormone-Specific Tests Hormone-specific assays determine whether a chemical acts through a particular hormonal mechanism (such as estrogenicity), regardless of whether the effect is thought to be harmful. Specific assays can be used to test the hormonal-agonist or -antagonist activity of a chemical. Mammalian estrogen biologic assays that measure reproductive-tract responses are summarized in Table 11-2. These assays use rodent vaginal and uterine tissues, and they are sensitive, reproducible, and biologically relevant for comparing the estrogenic potential of various HAAs. Typically, endogenous sources of estrogen are eliminated or reduced by using immature or ovariectomized animals, thus eliminating the problem of mistaking responses caused by endogenous estrogen for those caused by exogenous compounds. Measures of estrogenic activity include vaginal cornification, vaginal epithelial-cell proliferation and tetrazolium reduction, vaginal opening. vaginotrophic response, uterine fluid imbibition, uterotrophic response, uterine glycogen deposition, and uterine estrogen-withdrawal bleeding. In the vaginal-cornification assay (Allen and Doisy 1923), estrogen agonists cause the vaginal epithelium to divide and differentiate from cuboidal cells to pseudo-stratified columnar cells to keratinized stratified squamous epithelial cells within 48-72 hr of exposure. Vaginal keratinization and cornification are among the most specific in vivo end points available for determining the estrogenic character of a compound (Edgren 1994). In addition, the vaginal-cornification assay requires only microscopic readings of vaginal smears. Vaginal opening is an estrogenic response and a developmental end point. At birth, the external rat vagina is still a solid cord of cells (Allen and Doisy 1924). Before the first estrus, and about 5 d before ovulation, the vaginal lumen is formed. This process can be accelerated from the average of 40 d of age to 30 d, or even earlier, by exposure to estradiol or other estrogens (Edgren et al. 1966). It has been proposed that this end point be included routinely in the multigenerational reproduction study (EPA 1996). Uterine fluid imbibition (Astwood 1938) is rapid (6 hr), but it is not a graded response. Therefore, this assay does not distinguish potent from relatively weak estrogens or short-acting from long-acting estrogens. The uterotrophic assay typically uses a regimen in which animals are given daily doses of a compound for 3-4 d to measure true uterine growth (Bulbring and Burn 1935; Dorfman and Dorfman 1954; Edgren et al. 1966). Short-acting estrogens, such as estriol, do not produce the same dose-response curves, as docontinue

OCR for page 296
Page 304 TABLE 11-2 Mammalian Estrogen Biologic Assays Assay Advantages Disadvantages Reference Vaginal cornification Specificity; can detect proestrogens; Not for large-scale screening; requires 3 Allen and Doisy 1923   multiple routes of exposure; in vivo d or more; typically uses immature     pharmacokinetics, including animals or ovariectomized adult animals     bioaccumulation and repair; animals can       be reused     Vaginal epithelial cell Specificity: can detect proestrogens; Not for large-scale screening; labor Martin and Claringbold 1958; proliferation and multiple routes of exposure: in vivo intensive; typically uses ovariectomized Martin 1960 tetrazolium reduction pharmacokinetics, including adult animals: requires excision of the     bioaccumulation and repair vagina for histologic examination   Vaginal opening Can detect proestrogens: multiple routes Lacks specificity; not for large-scale Allen and Doisy 1924: Edgren   of exposure: in vivo pharmacokinetics, screening; labor and time intensive; uses et al. 1966   including bioaccumulation and repair immature animals   Vaginotrophic response Can detect proestrogens: multiple routes Lacks specificity: not for large-scale Folman and Pope 1966   of exposure: in vivo pharmacokinetics. screening; requires 3 d; typically uses     including bioaccumulation and repair immature animals or ovariectomized adult       animals; requires excision of the vagina       for weighing   Uterine fluid imbibition Simple. rapid response; can detect Not for large-scale screening; uses Astwood 1938   proestrogens: multiple routes of exposure; immature animals; requires excision of     in vivo pharmacokinetics, including the uterus for weighing and histologic     bioaccumulation and repair examination; no graded response   Uterotrophic response Sensitive; can detect proestrogens; Lacks specificity; not for large-scale Bulbring and Burn 1935;   multiple routes of exposure; in vivo screening: requires 3 d; typically uses Dorfman et al. 1936; Lauson et   pharmacokinetics. including immature animals or ovariectomized adult al. 1939; Evans et al. 1941;   bioaccumulation and repair animals; requires excision of the uterus Rubin et al. 1951     for weighing   (table continued on next page)break

OCR for page 296
Page 305 (table continued from previous page)break Assay Advantages Disadvantages Reference Uterine glycogen Simple, rapid response: can detect Not for large-scale screening; uses Bitman and Cecil 1970 deposition proestrogens; multiple routes of exposure; immature animals; requires excision of     in vivo pharmacokinetics, including the uterus for biochemical analysis:     bioaccumulation and repair specificity unknown   Uterine estrogen- Can detect proestrogens; multiple routes Lacks specificity; not for large-scale Eckstein et al. 1952; Schane et withdrawal bleeding of exposure; in vivo pharmacokinetics, screening: costly; typically uses al. 1972; Kassis et al. 1984;   including bioaccumulation and repair: ovariectomized rhesus monkeys; possible Tiemann and Tuchscherer 1995   and animals can be reused confounding by compounds with       extremely long half-lives   SOURCE: Adapted from Reel et al. 1996.

OCR for page 296
Page 306 such long-acting estrogens as diethylstilbestrol (DES), estradiol. and estrone. Weight increases are measured, and cell proliferation and hypertrophy are confirmed by histopathology. Although organ-weight responses are comparable for the vagina and uterus in these tests, vaginotrophic activity has not been used often as a quantitative end point (Folman and Pope 1966), because it is more difficult to excise and weigh vaginal tissue than uterine tissue. Estrogens also stimulate biochemical responses in the reproductive tract. For example, estrogens stimulate uterine glycogen concentrations within 6 hr of administration, and maximal concentrations are achieved within 24-48 hr of a single treatment (Galand et al. 1987). Biologic Markers of Exposure and Effect Several functional biologic markers have been used or could be used to monitor exposures to HAAs in wildlife and human populations. Functional biologic markers include population-level responses, changes in secondary-sex characteristics, changes in concentrations of plasma steroid hormones, ex vivo responses, changes in enzyme activity, histologic changes in endocrine-responsive tissues, vitellogenin response, and zona-radiata-protein response. Population-Level Responses and Secondary Sex Characteristics There is no specific end point or group of end points that can be used exclusively to monitor exposure to HAAs and the effects of HAAs. Organisms are simultaneously exposed to many chemical stressors, and for many conditions, it is difficult to determine a single causative agent or group of agents (Ludwig et al. 1993). Biochemical markers of exposure are useful as an early-warning system, because they change before histologic, organism, or population-level effects are observed. Unfortunately, many of these biochemical markers are fairly specific to individual compounds or at least to classes of compounds. In wildlife monitoring, a tiered approach has been used. In general, population-level responses are monitored, and if the populations are reproducing successfully it is inferred that they are healthy. If adverse effects are observed, then a combination of biologic and instrumental chemical techniques is often applied to determine the cause (Giesy et al. 1994a). Because monitoring nonmammalian vertebrates for potential effects of HAAs is complicated by the limited information about what constitutes normal endocrine function for many species, surrogate species are used for laboratory testing and for field monitoring exposure to HAAs. Secondary sexual characteristics have been suggested as possible functional biologic markers of exposure to compounds that affect sex steroid hormones. These markers are best observed in cold-blooded vertebrates, such as fish and reptiles, and they include the size andcontinue

OCR for page 296
Page 307 shape of sex organs, as exemplified in studies with male fish (Howell et al. 1980) and alligators (Guillette et al. 1995b). Secondary sexual characteristics are greatly pronounced in some species. such as the fathead minnow (Pimephales promelas). Males have dark coloration and a body shape that is very different from that of females. In addition, the male has a fat pad on the top of its head and breeding tubercles on the snout. These characteristics can be altered by exposure to estrogen (estradiol). When males were exposed to 2 nM of estradiol in the water for 5 wk, the size of the breeding tubercles and fat pads was decreased (Miles-Richardson et al. in press). Changes in secondary sexual characteristics occurred at exposures that also caused significant histologic effects but at concentrations that were less than those required to cause significant decreases in fecundity. Thus, it is thought that such secondary characteristics could be used as indicators of exposure to some HAAs. Other adverse effects on secondary sexual characteristics attributed to exposure to HAAs have been seen in mammals (Facemire et al. 1995). fish (Bortone and Davis 1994; Purdom et al. 1994; Harries et al. 1996), birds (Fry and Toone 1981; Fry et al. 1987), amphibians (Hayes 1997), and reptiles (Guillette et al. 1994). They include masculinization of females, feminization of males, deformities, and altered behavior (Colborn and Clement 1992; Giesy et al. 1994b). See Chapter 5 for more detail on these studies. Compromised gonadal development in fish has been attributed to exposure to HAAs, primarily environmental estrogens (see Chapter 5). In the United Kingdom, a relatively high incidence of intersexing has been observed in fish living in the vicinity of wastewater treatment plants (Purdom et al. 1994). The gonad histologic effects were thought to be estrogenic in nature because vitellogenin was induced in males of the roach (Rutilis rutilis). Subsequent studies with caged rainbow trout (Oncorhynchus mykiss) indicated that indeed the compounds causing the effects were environmental estrogens (Jobling et al. 1996). It is believed that the most likely cause of the observed effects in these fish were steroidal estrogens from domestic sewage and alkylphenol ethoxylates from industrial effluent (R. White et al. 1994; Jobling et al. 1995; Nimrod and Benson 1996). Histologic and Biochemical Responses Exposure of animals to HAAs can result in pathologic tissue changes that can be observed histologically. For example, exposure of fathead minnows (Pimephales promelas) to 2 nM of estradiol causes histologic and ultrastructural changes that are related to impaired reproductive performance as seen in seminiferous tubules, hypertrophy and hyperplasia of Sertoli cells, and degeneration of spermatozoa (Miles-Richardson et al. in press). In females in the same study, the relative proportions of oocytes were altered to more primary cells and fewer graffian cells. These histologic effects were observed at exposures that did notcontinue

OCR for page 296
Page 308 cause any induction of vitellogenin or decrease in the number of viable eggs produced. Steroid hormonal concentrations in blood also could signal exposure to HAAs. Several pathways are possible, including inhibition or induction of key enzymes in steroidogenesis and in ovo exposure to HAAs during critical periods that results in abnormal steroidogenesis in offspring (Guillette et al. 1994). Vitellogenin, a glycolipophosphoprotein synthesized by the liver in response to estrogen stimulation, is the precursor of egg yolk in oviparous animals, such as fish, birds, and crocodilians (Korsgaard and Petersen 1979). The term vitellogenin is applied to all such proteins, even though the exact structure can vary among species. Vitellogenin is a good biologic marker of exposure to estrogenic substances because it is regulated by the ER, it is a plasma protein, and it can be readily detected and quantitated in plasma samples. Although exogenous estrogen agonists have minimal effects on plasma vitellogenin in female fish, male fish express low to nondetectable concentrations of the protein, which is readily induced by estrogenic compounds and secreted into plasma. Exposure to environmental estrogens has been shown to induce production of vitellogenin in the blood plasma of male fish (Jobling et al. 1995). Although it is not known whether such production has any adverse effect, measuring plasma vitellogenin does provide a useful biologic marker of exposure. At least one study posited a connection between reproductive outcome and serum vitellogenin concentration (Kramer et al. 1997). The results suggest that induction in male fish is inversely correlated with the number of viable eggs produced by pairs of fish. Based on tests with nonylphenol and effluent from an oil refinery treatment plant, there is some evidence that the zona-radiata-protein response could be as sensitive as the vitellogenin response in detecting exposure to estrogenic HAAs (Arukwe et al. 1997) and that it could be a more useful biologic marker for hormonal activity because critical population parameters, such as offspring survival and recruitment, are more directly affected. However, the advantages of the zonaradiata-protein as a biologic marker need to be further assessed through long-term exposure studies of fish exposed to low concentrations of xenoestrogens. Instrumental Chemical Techniques Hormonally active agents can occur everywhere in the environment, but the two primary vectors of exposure for wildlife and humans are water and food. The presence of known HAAs can be determined either by directly measuring concentrations of identifiable compounds or by screening for a range of possible HAAs. For compounds that have an established reference dose (RfD), a hazard quotient (HQ) can be calculated to determine whether adverse effects would be expected. The HQ is the ratio of the RfD to the measured concentration. Thecontinue

OCR for page 296
Page 309 RfD can be tissue-specific or dietary, and it can be related to concentrations in abiotic matrices by using appropriate transfer coefficients (Starodub et al. 1996). In other cases, a combination of functional screening assays and instrumental analyses is applied in what has been defined as bioassay-directed fractionation and identification or toxicant identification and evaluation (TIE). In general, the process is an iterative process that uses fractionation methods and functional assays. Typically, the abiotic or biotic matrix of interest is separated into fractions based on polarity and molecular size. Functional assays are used to test the fractions for hormonal activity. Positive samples are further fractionated with instrumental methods, such as high-pressure liquid chromatography with both fluorescent and ultraviolet-visible detection, gas chromatography with flame-ionization detection, electron-capture detection, or mass-selective detection. Those methods are used repeatedly until the structure of the estrogenic compounds in the active fractions has been determined. Reference standards are used to quantify the mass of material in the sample. Relative potency factors (RPFs) are derived, and total equivalents are determined by multiplying the RPFs by the molar concentrations of the compound in the sample. Then, the predicted and measured activity in the assay are compared in a mass balance of potency. If the values are unequal, selective isomolar additions and selective removal of compounds can be used to confirm the presence or absence of interactions or unidentified compounds. These methods are repeated until all the active compounds are identified. Normally, the identity of the compound is confirmed by several instrumental techniques. A similar protocol has been proposed for use to separate the relative contributions of endogenous and exogenous hormones in plasma (Sonnenschein et al. 1995; Soto et al. 1997). Summary And Conclusions There are no generally accepted, validated methods to screen for or monitor exposure to HAAs, largely because the endocrine system is so complex. Relying on any single screening or monitoring method could result in uncertainty in the risk-assessment process and encourage the use of inappropriate or arbitrary safety factors. The use of simple models or screening methods can result in false-positive and false-negative results (Patlak 1996). Thus, efforts have been made to design a battery of simple tests that are based on different mechanisms of action or that are designed to assess the effects of activation or deactivation or the effects of accessory factors (Shelby et al. 1996). The design of such a battery should account for which mechanisms of action should be included, and it should consider the sensitivity and accuracy of the assays. In any case, it should always be recalled that a host of compounds with no hormonal activity can disrupt reproductive and other organismic functions.break

OCR for page 296
Page 310 Recommendations On the basis of its evaluation of the available data on screening for and monitoring HAAs, the committee recommends the following: —A battery of short-term assays should be developed for rapid and inexpensive screening for putative HAAs. The assays should detect diverse responses that depend on hormone receptors, should detect other indirect responses, and should be readily adapted for use in multiple laboratories. —Short-term assays should be validated, replicated, and deployed in a rational fashion. Investigations should also be conducted to determine whether short-term assays predict in vivo toxicity. —Some potential biomarkers of exposure to HAAs in wildlife and humans are available and should be applied. Additional biomarkers should be developed and validated before application. In particular, assays should be developed that screen for embryonic and fetal events (markers) that predict long-term, delayed effects. —Species- and tissue-specific effects resulting from exposure to environmental HAAs need to be investigated further. —Differences in response to HAAs between adults and developing fetuses need to be investigated with regard to the possibility of unique effects due to exposure during critical periods in development when genetic imprinting is occurring. —Wildlife can serve as environmental sentinels. Populations known to be exposed to HAAs should be monitored for both obvious and subtle responses to exposures. Studies of caged, pinioned, or telemetered animals could provide information about the location, duration, and magnitude of exposure, which could be used to interpret the results of field studies. —Dose-response characteristics of recognized actions of various HAAs should be further investigated in in vitro and in vivo studies using concentrations found in the environment.break