ToxSci Advance Access originally published online on August 29, 2006
Toxicological Sciences 2006 94(1):46-56; doi:10.1093/toxsci/kfl092
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Published by Oxford University Press 2006.
Nature of the Binding Interaction for 50 Structurally Diverse Chemicals with Rat Estrogen Receptors
* Endocrinology Branch, Reproductive Toxicology Division, National Health and Environmental Effects Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, Mail-drop 72, Research Triangle Park, North Carolina 27711
Physiology-Pharmacology Department, Wake Forest University School of Medicine, Winston Salem, North Carolina 27157
1To whom correspondence should be addressed at. Fax: (919) 541-5138. E-mail: laws.susan{at}epa.gov.
Received May 30, 2006; accepted August 22, 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUPPLEMENTAL DATAREFERENCES |
|---|
This study was conducted to characterize the estrogen receptor (ER)–binding affinities of 50 chemicals selected from among the high production volume chemicals under the U.S. EPA's (U.S. Environmental Protection Agency's) Toxic Substances Control Act inventory. The chemicals were evaluated using the rat uterine cytosolic (RUC) ER-competitive binding assay, with secondary analysis using Lineweaver-Burk plots and slope replots to confirm true competitive inhibition and to determine an experimental Ki. Data from these ER-competitive binding assays represent the types of competitive binding curves that can be obtained when screening chemicals with broad structural diversity. True competitive inhibition was observed in 17 of 50 chemicals. Binding affinities were much lower than that of estradiol (E2) with Ki concentrations ranging from 0.6 to 373µM as compared with that of E2 (0.77nM). Other chemicals that appeared to displace radiolabeled E2 binding to ER were, in fact, found not to be competitive inhibitors in the secondary Ki experiments. These seven chemicals likely altered the stability of the assay by changing the buffer pH, denaturing ER, or disrupting the ER-binding kinetics. Thus, several conditions that may confound interpretation of RUC ER-binding assay data are illustrated. For another group of eight chemicals, neither an IC50 nor Ki could be determined due to solubility constraints. These chemicals exhibited slight (20–40%) inhibition at concentrations of 10–100µM, suggesting that they could be competitors at very high concentrations, yet Ki experiments were not possible as the limit of chemical solubility in the aqueous assay buffer was well above the IC50. An additional 18 of the 50 chemicals were classified as nonbinders because in concentrations up to 100µM they produced essentially no displacement of radiolabeled E2. These results show that although the ER-competitive binding assay is a valuable tool for screening chemicals, secondary tests such as a double reciprocal Lineweaver-Burk experiment with slope replot should be used to confirm true competitive inhibition. This information will be useful for the ongoing validation of the RUC ER-competitive binding assay under the U.S. EPA's Endocrine Disruptor Screening Program, as well as to support research efforts to develop computational models designed to identify chemicals with the ability to bind to ER.
Key Words: estrogen receptors; rat uterine cytosol; environmental chemicals; Ki.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUPPLEMENTAL DATAREFERENCES |
|---|
In response to emerging concerns that environmental chemicals may have adverse effects on human health by altering the function of the endocrine system, the Food Quality Protection Act mandated that the U.S. Environmental Protection Agency (U.S. EPA) develop and implement an Endocrine Disruptor Screening Program (EDSP; U.S. EPA, 2006
). Toward this end, the U.S. EPA is evaluating a number of in vitro and in vivo assays designed to detect chemicals that can interact with the estrogen, androgen, and thyroid systems in human, fish, and wildlife (U.S. EPA, 2006). Following a validation process, some of these assays will be selected for use in a Tier I screening battery.
The estrogen receptor (ER)–competitive binding assay is one of the methods currently being validated. This in vitro assay was recommended by the Endocrine Disruptor Screening and Testing Advisory Committee, a federal advisory group charged with assisting the U.S. EPA in development of a scientifically defensible EDSP (U.S. EPA, 1998a). The assay can be used to determine the relative binding affinities (RBAs) of environmental chemicals for the ER, by comparison with the endogenous hormone, estradiol (E2). The assay measures the ability of radiolabeled E2 to bind with rat uterine cytosolic (RUC) ER in the presence of increasing concentrations of a test chemical (Clark and Gorski, 1969; ICCVAM, NIEATM, 2003a; Markaverich et al., 1983; U.S. EPA, 1998b). Environmental chemicals that compete with estrogen for binding to ER have the potential to either induce (agonist) or block (antagonist) estrogen-dependent transcriptional activity since ER association is the initial step in a series of events that culminate in multiple estrogen-dependent physiological responses (Tsai and O'Malley, 1994; Brzozowski et al., 1997). Using this assay as a screening tool can provide a fairly rapid and inexpensive method for identifying likely endocrine disrupting substances (ICCVAM, NIEATM, 2003b; Laws et al., 2006).
Examples of the biological relevance and utility of this assay as a screening tool are evident from numerous publications where the ER-binding affinities of a variety of chemicals have been correlated with estrogenic or antiestrogenic physiological responses after in vivo exposures. Chemicals such as 4-tert-octylphenol, 4-nonylphenol, and bisphenol A, that have been shown to stimulate estrogenic activity in vivo, also inhibit the binding of E2 in the ER-competitive binding assay (Ashby and Tinwell, 1998; Bicknell et al., 1995; Danzo, 1997; Hammond et al., 1979; Kuiper et al., 1998; Laws et al., 2000; Lee and Lee, 1996; Routledge and Sumpter, 1997; Takemura et al., 2005; White et al., 1994). Elsby et al. (2000) compared the results from ER-binding and uterotrophic assays for 17-ß–E2, and steroid analogues and found that the chemicals produced similar results in both assays. Similarly, Yamasaki et al. (2004) tested 14 chemicals in both assays, and their results demonstrated that the affinity of the chemicals in the ER-binding assay correlated well with their potency in the uterotrophic assay, although a few agonists had a very low receptor-binding affinity. On the other hand, while methoxychlor has a very low binding affinity for ER, its hydroxylated metabolite, 2,2,-bis(p-hydroxyphenyl)-1,1,1-trichlorethane, exhibits a moderate binding affinity for the ER and produces estrogenic responses in rats following administration in vivo (Bulger et al., 1978a,b; Gray et al., 1989; Kuiper et al., 1998). Thus, while a direct correlation between the concentration of a chemical that can inhibit 50% of ER binding in vitro and a dose that will produce physiological effects in vivo may be obscured by pharmacokinetic factors (absorption, metabolism, tissue disposition, and elimination), the historical use of the ER-competitive binding assay clearly demonstrates its usefulness for initially identifying chemicals and/or their metabolites that can bind to the ER.
As part of the validation process for the RUC ER-competitive binding assay, the reliability of the assay must be demonstrated by its ability to produce similar results in multiple laboratories (ICCVAM, NICEATM, 2003b). To meet this criterion, the U.S. EPA is currently evaluating assay performance in several contract laboratories by testing a set of chemicals with known ER-binding affinities (http://www.epa.gov/scipoly/oscpendo/assayvalidation/status.htm). Data from these studies will be used to set the initial standards for performance criteria. In addition, guidance for data interpretation is needed to facilitate uniform, maximal use of the assay as a screening tool. Since the assay is intended to evaluate chemicals over a broad range of structural and physical properties as currently existing among chemicals under the purview of the U.S. EPA, it is expected that the data obtained from the assay will reflect a wide range of binding affinities for the ER. For example, some chemicals with high to moderate binding affinity will produce full competitive binding curves, while others with extremely low binding affinity will produce only partial curves. It is also likely that evaluation of some of the test chemicals will be limited by their lack of solubility in the aqueous assay buffer and/or their disruption of the biochemical kinetics of the binding assay itself. Thus, it is imperative that the validated protocol for the RUC ER-binding assay includes guidance for data interpretation to address these issues as well as to minimize the number of false-negative or -positive classifications of the test chemicals.
The study reported here was designed to evaluate the performance of the RUC ER-competitive binding assay when testing 50 chemicals selected from the high production volume inventory covered by the U.S. EPA's Toxic Substances Control Act. Chemicals were selected based upon unpublished data from ER-competitive binding assays previously conducted by Battelle Pacific Northwest Laboratories, Richland, WA (U.S. EPA Contract Number 68-W-99-033, Task 6). In the earlier experiments, some chemicals produced competitive binding curves that strongly indicated that they were true competitive inhibitors of ER binding. However, others produced only partial curves or irregular (nonsigmoid) binding curves. For this group of chemicals, additional secondary Ki experiments were needed to understand fully whether or not these were true competitive inhibitors of ER binding (e.g., both the natural ligand and test chemical compete for the same site, Fig. 1). Thus, retesting this set of chemicals provided a unique opportunity to (1) compare the reproducibility of the RUC ER-competitive binding assay between two laboratories, (2) conduct the secondary Ki experiments to demonstrate which chemicals were true competitive inhibitors of ER binding, (3) provide a data set of chemicals for which the ER-binding ability has been well characterized, and (4) develop guidance for appropriate data interpretation when using the assay as a screening tool. Therefore, for each chemical in this study that produced a curve suggestive of competitive binding to the ER, a secondary analysis was conducted to determine a Ki and to distinguish which chemicals were true competitive inhibitors of ER binding.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUPPLEMENTAL DATAREFERENCES |
|---|
Chemical selection.
The Chemical Abstract Service numbers and structures of the chemicals evaluated are shown in Figure 2, and all are currently under the U.S. EPA's purview. Chemicals were received from Battelle Pacific Northwest Laboratories and were identical to the specific lot used in their previous studies (Contract Number 68-W-99-033). Chemical purity ranged from 97 to 99%, and all chemicals were initially dissolved in 100% absolute ethanol (200 proof).
|
ER-competitive binding assay.
RUC was used as the source of ER (containing both ER-
and -ß) and was prepared using uteri from ovariectomized, adult Sprague-Dawley rats (Harlan Industries, Madison, WI). Following CO2 euthanasia of the animals, the uteri were removed, trimmed of excess fat, and placed in ice-cold TEDG buffer (Tris 10mM, ethylene diamine tetraacetic acid 1.5mM, dithiothreitol 1mM, 10% glycerol, pH 7.4; ratio of 1 uterus/0.5 ml TEDG buffer). Tissues were homogenized with three to five bursts (approximately 5 s per burst; Polytron PT 35/10). The homogenate was transferred to precooled centrifuge tubes and centrifuged for 10 min at 2500 x g at 4°C. The supernatant was transferred to precooled tubes and recentrifuged at 105,000 x g for 60 min at 4°C. All supernatants were pooled, assayed for protein content (BCA Pierce), and frozen in 200 µl aliquots at – 80°C. Methods used for the RUC ER-competitive binding assay followed the protocol that is currently undergoing a multilaboratory validation under the Agency's EDSP program (http://iccvam.niehs.nih.gov/methods/endodocs/final/erbndbrd/erbdappx/B5.pdf), with minor modification. Briefly, assays were conducted using RUC with a final protein concentration in each assay tube of 1.9–2.4 mg/ml and 0.33nM 3H-17ß–E2 (Perkin Elmer NEN, NET-517, specific activity = 170 Ci/mmol) in TEDG buffer. Each test chemical was evaluated over a concentration range of 0.1–100µM. Following an 18-h incubation at 4°C, bound and free ligands were separated by dextran-charcoal extraction, and bound radioligand was quantified by scintillation counting. Nonspecific binding was assessed by adding 200nM radioinert E2 (Sigma E-9875, St Louis, MI). Specific binding was calculated by subtracting nonspecific binding from total binding. For each chemical, the assay was replicated on a second separate day. A full competitive binding curve using 17-ß–E2 (0.002–100nM) was included as a positive control for each daily assay.
Data were analyzed using a nonlinear regression program for fitting one-site competitive binding curves (GraphPad Prism, GraphPad Software, Inc. 2003, San Diego, CA), and estimates of IC50 concentrations (e.g., the concentration of a test chemical required to inhibit 50% of the tracer maximal binding) were calculated. Competition by each test chemical that demonstrated an ability to reduce ER binding by 40% or greater was further evaluated with a Ki experiment. Chemicals that did not inhibit ER binding by more than 20% at a concentration of 100µM were classified as nonbinders.
Ki experiments.
Ki values were experimentally determined by incubating increasing concentrations of 3H-E2 (0.05, 0.1, 0.2, 0.5, and 1.5nM) with ER in the presence of test chemical concentrations that bracketed the IC50 (approximately 0.5, 1, and 2 x IC50). Using RUC as the source of ER (1.9–2.4 mg protein/ml), assays were allowed to incubate for 18 h at 4°C. Bound and free tracers were separated by dextran-charcoal, and ER-bound 3H-E2 was quantified by liquid scintillation counting. Double reciprocal plots (Lineweaver-Burk) were constructed of the bound versus total. A pattern of lines intersecting at the y-axis was considered characteristic of true competitive inhibition. Slopes obtained from each of the double reciprocal plots were replotted versus concentration of test chemical, and a Ki was calculated from the negative intercept of the slope replot. Ki experiments were replicated for chemicals demonstrating true competitive inhibition, and Ki values are reported as mean ± SE (n = 2).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUPPLEMENTAL DATAREFERENCES |
|---|
ER-Competitive Binding Assays
The competitive binding assay measured the ability of 3H-E2 to bind with ER in the presence of increasing concentrations of a test chemical. Data from these assays provided an array of full and partial competitive binding curves. Results were grouped according to the ability to reduce radioligand binding. Chemicals in group A produced an almost complete displacement curve (e.g., maximal ER binding was inhibited by > 75%), group B displaced 50–74%, group C displaced 21–49%, group D displaced 0–20%, and group E produced irregular, nonsigmoidal curves. Chemicals in groups A, B, and E were tested in the secondary Ki experiments. Further characterization of all but one chemical in group C was not possible due to solubility limitations. Chemicals in group D were designated as nonbinders when incubated at concentrations up to 100µM and were not tested further.
Figure 3 presents examples of the variation in competitive binding curves observed among all the test chemicals. 17-ß–E2 was used as the positive control for all the assays. As shown in Figure 3a, increasing concentrations of radioinert E2 over a range of 0.02–100nM produced a full sigmoid curve indicative of a chemical with a single, high-affinity binding site for the ER in accordance with the law of mass action (Motulsky and Christopoulos, 2003). Specifically, tracer binding was reduced from 90 to 10% over an approximately 2 log unit increase in the nonradioactive E2 concentration. Under these conditions, the IC50 for E2 was 0.52 ± 0.01nM and the Ki was 0.77 ± 0.04nM, using the Cheng-Prusoff correction (Cheng and Prusoff, 1973).
|
Group A consists of six chemicals that produced nearly complete competitive binding curves. A curve for 4-(1-methyl-1-phenylethyl)-phenol is shown as an example of this group of chemicals (Fig. 3a). Twelve chemicals were included in group B and produced partial competitive binding curves at slightly more than 50% displacement (e.g., 4,4',4''-ethane-1,1,1-triyltriphenol and phenyl(2,3,4-trihydrophenol)-methanone Figs. 3b and 3c). Examples of two chemicals that at concentrations up to 100µM produced curves with 21–49% displacement (group C) are shown in Figure 3c (5-chloro-2-(2,4-dichlorophenoxyl) phenol, 2,2-dimethoxy-1,2-diphenylethanone). Solubility limitations for eight of the nine chemicals in this group prevented further testing with the secondary Ki assay. Another 18 chemicals did not inhibit ER binding more than 20% at concentrations up to 100µM (group D) and were classified as nonbinders (Fig. 3d, 4,4'-propane-2,2-diylbis-(2-tert-butylphenol)).
Examples of the four chemicals (group E) that produced nonsigmoidal binding curves are shown in Figures 3c and 3d. N,N-dimethyl-N-(3-trimethoxysilyl)propyl-octadecan-1-aminium chloride and N,N,N-trimethyl-octadecan-1-aminium chloride (Fig. 3d) were among the three chemicals for which the maximal binding of 3H-E2 dropped dramatically over a single log increase in concentration of either of these chemicals. 4-Dodecylbenzene-sulfonic acid produced a U-shaped curve as the concentrations were increased from 1.0 to 100µM (Fig. 3c). The shapes of all these curves suggested that the biochemical stability of the binding assay may have been compromised, and all these chemicals warranted further testing by secondary Ki experiments.
Ki Experiments
Ki concentrations (means ± SEM, n = 2) for the 17 chemicals that exhibited true competitive inhibition are shown in Table 1. Five of six chemicals in group A and 11 of 12 chemicals in group B were confirmed to be true ER-competitive binders by reanalysis of competitive binding properties using a series of Lineweaver-Burk plots. Data are shown in Figure 4 for (1) 4-(1-methyl-1-phenylethyl)-phenol, (2) 4,4'4''-ethane-1,1,1-triyltriphenol, (3) phenyl(2,3,4-trihydroxyphenyl)-methanone, and (4) 2,2-dimethoxy-1,2-diphenylethanone. None of the chemicals in these groups exhibited an ability to bind to ER at concentrations approximating E2 itself (Ki = 0.77nM). Ki concentrations for this group of chemicals ranged from 0.5 to 373µM or by 1000–850,000 times the Ki for E2.
|
|
All the four chemicals that produced nonsigmoid competitive binding curves also failed to exhibit true competitive inhibition when evaluated in the secondary Ki experiments. For all chemicals where the competitive binding curve dropped dramatically over a single log increase in concentration, the follow-up Ki experiment demonstrated that each was not a true competitive inhibitor (N,N,N-trimethyloctadecan-1-aminium chloride [Fig. 5a], octanal, and N,N-dimethyl-N-(3-trimethoxysilyl)propyl-octadecan-1-aminium chloride [Fig. 5b]). In these cases, the reduction in the maximal binding of tracer was likely due to an interference with the assay conditions by denaturing the ER, altering pH, and disrupting the receptor conformation such that it was no longer capable of binding E2. 4-Dodecylbenzenesulfonic acid, a chemical that produced a U-shaped binding curve, was also confirmed to be a nonbinder (Fig. 5c).
|
Finally, there were other chemicals for which the secondary Ki experiments facilitated the proper classification as ER binders or nonbinders. N,N-dimethyltetradecan-1-amine produced a nearly complete competitive binding curve, and sodium-5-chloro-2-(4-chloro-2-(((3,4-dichlorophenyl)-carbamoyl)amino)phenoxy) benzene sulfonate hydrate and 5-chloro-2-(2,4-dichlorophenoxyl)phenol produced partial competitive binding curves similar to other chemicals that were ER binders. However, the secondary binding assays indicated that these chemicals were nonbinders (Figs. 3c and 5d; 5-chloro-2-(2,4-dichlorophenoxyl)phenol exemplifies data from these chemicals). Of the chemicals in group C, further testing by a Ki experiment for eight of nine was not possible because the limit of solubility of each was well above the IC50. These chemicals are currently being reevaluated using DMSO rather than ethanol as the initial stock solvent.
Comparison of Data Interpretations Using Only ER-Competitive Binding Curves as Opposed to the Ki Experiments
Results from the present competitive binding assays were compared with data obtained from initial tests at Battelle Pacific Northwest Laboratories. In the earlier study, a chemical was categorized as an ER binder if it displaced the maximal binding of 3H-E2 by 50% or greater. The assay was considered inconclusive for chemicals that displaced 20–49%, while chemicals whose inhibition dropped rapidly over a single log concentration unit, or that displaced < 20%, were classified as nonbinders. Comparison of results from the two laboratories indicated that there was 94.4% concordance among the 17 chemicals that were designated in this study as true competitive binders using the Ki experiments. Two chemicals not detected in the earlier study were classified as "inconclusive" in the present assay when based solely on the competitive binding assay.
Use of the Ki experiments was most helpful in differentiating between the chemicals with partial binding curves of up to 50% displacement as well as for those that presented nonsigmoidal curves. For example, both 5-chloro-2-(2,4-dichlorophenoxyl)phenol and 2,2-dimethoxyl-1,2-diphenylethanone (Fig. 3c) produced similar competitive binding curves with slightly less than 50% displacement of maximal binding. However, only the latter was actually a true competitive binder. All Ki experiments for the chemicals with the rapid drop in binding over a single log unit were demonstrated to be nonbinders using the Ki experiments. Finally, there were two chemicals in this study that would have been incorrectly designated as an ER binder if evaluated solely with competitive binding curves (N,N-dimethyltetradecan-1-amine and sodium-5-chloro-2-(4-chloro-(((3,4-dichlorophenl)-carbamoyl)amino)phenoxy) benzene sulfonate hydrate). Interestingly, both of these chemicals produced binding curves with sharp drops in binding when tested in the Battelle Pacific Northwest Laboratories.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUPPLEMENTAL DATAREFERENCES |
|---|
The purpose of the RUC ER-competitive binding assay is to identify chemicals that have the potential to compete with endogenous estrogens for binding to the ligand-binding domain of ER. The assay is conducted by measuring the binding of a single concentration of 3H-17ß–E2 to RUC ER in the presence of increasing concentrations of radioinert E2 or a test chemical. It is based upon the assumptions that (1) the radioligand and the test chemical compete for the same receptor site so that binding of one or the other is mutually exclusive (Fig. 1), (2) the reaction is reversible and reaches equilibrium such that the rate of new ER ligand association equals the rate of dissociation, (3) neither the ligand nor ER is altered by binding. If these assumptions are met, the data can be analyzed using simple models based upon the law of mass action (Motulsky and Christopoulos, 2003
) which predicts the fraction of receptors that will be occupied at equilibrium as a function of ligand concentration. Plotting the data as the percent 3H-E2 bound versus the concentration of radioinert E2 or a test chemical produces a sigmoid curve with the top plateau equal to the maximal binding capacity in the absence of a competing chemical and the bottom of the curve equal to the nonspecific binding. Theoretically, the maximal ER binding should drop from 100 to 0% over a 2-log increase in chemical concentration if the chemical is competing with 3H-E2 for a single receptor site. In addition, the slope factor (e.g., steepness of the binding curve, Hill slope) for a one-site competitive binding curve should be –1.0. Chemicals that can inhibit binding at nanomolar concentrations have a higher affinity for ER compared with those that require micromolar or millimolar concentrations. Since the IC50 is dependent upon the concentration of 3H-E2 used in the assay, the RBA (percent of test chemical IC50, compared with that of E2) is commonly reported. Alternatively, another method for reporting binding affinity is to calculate a Ki value (e.g., the concentration of the test chemical that will bind to one half of the ER in the absence of radioligand or other competitors) either by using the correction of Cheng and Prusoff (1973) to convert an IC50 to Ki or by determining a Ki experimentally. Here we report the Ki values for 17 of 50 test chemicals that demonstrate true competitive binding for the ER. All produced full or partial competitive binding curves, and 16 of 17 chemicals inhibited at least 50% of the radioligand binding at concentrations up to 100µM. Further characterization of ER binding using Lineweaver-Burk double reciprocal plots and slope replots confirmed that these chemicals were true competitive inhibitors at the ER ligand-binding domain. Ki concentrations for this group of chemicals ranged from 0.5 to 373µM. While all the competitors exhibited far less affinity for the ER than E2, the data support a conclusion that high concentrations could potentially act as ER agonists or antagonists. Since the RUC ER-competitive binding assays are limited to detecting the ability to interact with ER, additional tests, such as ER reporter gene assays, are needed to confirm agonist or antagonist activity, as well as the potential for in vivo activity.
The occurrence of nonsigmoid curves for some of the chemicals was not unexpected given the diversity in chemical structures and physical properties. Those curves that dropped dramatically over a single log unit change in the concentration clearly demonstrated the biochemical limitations of the RUC ER-competitive binding assay. Changes in the assay pH in the presence of high concentrations of test chemical can result in subtle to dramatic alterations in ER confirmation that would alter receptor-binding characteristics. What may initially appear as inhibition of binding can actually be an inability of ER to bind any ligand. In some cases, the chemical may totally denature the receptor. This was the case for ethanol when used in excess as a primary solvent for the test chemicals. In our experiments, we observed reduced binding capacity beginning at ethanol concentrations of 2%, while concentrations above 20% essentially destroyed all ER binding (data not shown).
Any chemical in excessive concentrations is likely to disrupt the biochemical stability of the assay and diminish the maximal binding of radiolabeled ligand. This is not a fundamental failure of in vitro binding assays but rather a requirement for optimal biochemical conditions to maintain receptor integrity at the ligand-binding domain. In addition, some chemicals may be capable of forming submicrometer aggregates that can then inhibit enzymatic and receptor proteins via a nonspecific mechanism that would not occur if individual molecules were present (McGovern et al., 2003; Ryan et al., 2003; Seidler et al., 2003). Unfortunately, the limiting concentration will vary among the vast number of structurally diverse chemicals that may be screened in the RUC ER-competitive binding assay. Therefore, it is imperative that data which do not produce the expected sigmoid curve be viewed with circumspect until further evaluation. As expected, the Ki experiments using the three chemicals that produced sudden drops in ER binding demonstrated that these chemicals were not true competitive inhibitors.
Chemical solubility was another issue reflected in the study results. Because the ER is physiologically active in aqueous intracellular fluid, the binding assays must also be conducted in aqueous buffers. Endogenous estrogens are highly nonpolar but can be suspended in aqueous buffers at biologically active concentrations. However, since the test chemicals typically have a much lower affinity for ER, there is a need to use concentrations of 1000–100,000 times that of E2 in the assays. The overnight assay incubation at 4°C to achieve equilibrium of the reaction also exacerbates the solubility issue. In a review of in vitro ER-competitive binding assays, an expert panel recommended that chemicals be tested over the range of 1.0nM–1mM (ICCVAM, NICEATM, 2003b). However, as demonstrated by the U-shaped and rapidly dropping curves observed in this study, a maximal concentration of 1mM will not be universally suitable for all chemicals. However, the use of a secondary Ki experiment can facilitate the characterization of ER-binding activity for those chemicals that are soluble at concentrations at the IC50 and for which a partial binding curve can be observed.
Chemical solubility issues rendered the competitive binding results inconclusive for 8 of 50 chemicals tested. These chemicals exhibited slight inhibition (21–49%) at concentrations up to 100µM, suggesting that they could be competitors at very high concentrations. This group poses a particularly vexing problem because, although a potential for ER binding was established, none of the chemicals were sufficiently soluble in the assay buffer system to permit a secondary analysis in a Ki experiment. This points to a need to evaluate alternative solvents and approaches in order to properly evaluate these chemicals. Solubility issues will also likely be problematic for other in vitro assays, such as reporter gene assays, that might be employed to evaluate these chemicals.
The physiological relevance for testing chemicals up to 1.0mM concentrations is a question that still remains to be addressed before adopting the RUC ER-competitive binding assay as a routine screening tool. The present study demonstrates that the highest possible concentration for testing will vary among the chemicals. Guidelines for maximal concentrations will need to consider the biochemical limitations of the RUC ER-competitive binding assay regarding the limit of solubility for each chemical, as well as any adverse effects a chemical may exert on the stability of the assay environment. In addition, testing a chemical at concentrations of 1,000,000-fold or greater than the Kd for the ER enhances the risk of reaction with lower affinity binding sites of little or no physiological significance. Thus, safeguards to ensure that the assay is being conducted under proper biochemical conditions that will support the ligand binding to ER, as well as proper data interpretation, need to be employed in order to assess the binding characteristics of a given test chemical with accuracy. Setting an upper concentration limit for the assay will require a consensus on whether chemicals with Ki values of 100µM or greater will pose a significant physiological risk to living organisms. Relative concentrations of xenoestrogens may vary substantially in vivo due to differences in bioaccumulation and/or the ability to bind to serum transport proteins (Bigsby et al., 1997; Nagel et al., 1997, 1998). In addition, the dose that may be effective as an ER agonist or antagonist in vivo may or may not be directly correlated with the Ki concentration (Zacharewski, 1997).
In summary, true competitive inhibition of ER binding was observed for 17 of the 50 chemicals tested in this study. Albeit these chemicals reflected a very weak affinity for the ER, true competitive inhibition was observed even when the Ki was as much as five orders of magnitude higher than that of E2. Importantly, 6% of the chemicals that appeared to be inhibitors in the competitive binding assay were in fact not true competitive inhibitors when further evaluated in the Ki experiments. In addition, irregular, nonsigmoidal curves were observed for 8% of the chemicals. These chemicals may have altered the stability of the assay in some fashion, for example, by changing pH, denaturing ER, precipitating out of solution, or forming aggregates of individual molecules. This study demonstrates the importance of understanding the limitations of the RUC ER-competitive binding assay, as well as the proper interpretation of the data when testing chemicals over a broad range of structural and physiochemical diversity. While the competitive binding assay is itself a valuable tool for screening chemicals for potential estrogenic or antiestrogenic activity, the assay works best when supported by secondary analyses. Here, we used Ki experiments to identify chemicals that were capable of true competitive binding to ER. Alternatively, other in vitro assays, such as an ER reporter gene or vitellogenin induction, might be employed to confirm estrogenic or antiestrogenic capability. Finally, these data show the range of results that can be obtained for a competitive binding assay when screening chemicals with diverse structural and physical properties. Although this study used RUC as the source for the ER, the issues discussed here are directly applicable to other assays utilizing alternative receptor sources, such as human recombinant ER or ER from other species, that might be more amenable to high throughput applications. By thoroughly evaluating the ER-binding characteristics using a double reciprocal Lineweaver-Burk model with a slope replot, our results demonstrated some potential errors that can occur during data interpretation. Such data are useful for the ongoing validation of the RUC ER-competitive binding assay as part of the EPA's EDSP, as well as to support research efforts to develop computational models to identify chemicals with the ability to bind to ER.
|
SUPPLEMENTAL DATA |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUPPLEMENTAL DATAREFERENCES |
|---|
Supplementary data are available online at http://toxsci.oxfordjournals.org/. Chemical structures and Ki data will also be available for public access through the National Center for Computational Toxicology's Distributed Structure-Searchable Toxicity Database Network at http://www.epa.gov/ncct/dsstox/.
|
NOTES |
|---|
Disclaimer: The information in this document has been funded by the U.S. EPA. It has been subjected to review by the National Health and Environmental Effects Research Laboratory, Office of Research and Development and approved for publication. Neither does approval signify that the contents necessarily reflect the views of the Agency nor does mention of trade names or commercial products constitute endorsement or recommendation for use.
|
ACKNOWLEDGMENTS |
|---|
This project was funded by the National Center for Computational Toxicology, U.S. EPA, Research Triangle Park, NC. This work was conducted under U.S. EPA Contract 3D-5967-NTEX by J.C.E. and S.Y., the Xenoendocrine Laboratory, Department of Physiology and Pharmacology, Wake Forest University School of Medicine, Winston Salem, NC. We gratefully acknowledge Dr Ann Richard (National Center for Computational Toxicology, U.S. EPA) and Marti Wolf (Lockheed Martin contractor to the U.S. EPA) for providing the chemical structures; Battelle Pacific Northwest Laboratories for their preliminary data from ER-binding assays (U.S. EPA Contract Number 68-W-99-033, Task 6) that led to the selection of chemicals for this study; Jim Kariya and Gary Timm, EDSP, Office of Science, Coordination and Policy, OPPTS, U.S. EPA, for their assistance with obtaining the test chemicals and assistance with the study design; Janet Ferrell, U.S.EPA, for assistance with data analysis; Ellen Lorang, U.S. EPA, for assistance with the graphics; and Drs Robert Kavlock, John Rogers, Shirlee Tan, and Jim Stevens for reviewing the article.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUPPLEMENTAL DATAREFERENCES |
|---|
Ashby J and Tinwell H. (1998) Uterotrophic activity of bisphenol A in the immature rat. Environ. Health Perspect. 106:11719–720.[ISI][Medline]
Bicknell RJ, Herbison AE, Sumpter JP. (1995) Oestrogenic activity of an environmentally persistent alkyphenol in the reproductive tract but not the brain of rodents. J. Steroid Biochem. Mol. Biol. 54:1–27–9.[CrossRef][ISI][Medline]
Bigsby RM, Caperell-Grant A, Madhukar BV. (1997) Xenobiotics released from fat during fasting produce estrogenic effects in ovariectomized mice. Cancer Res. 57:865–869.
Brzozowski AM, Pike ACW, Dauter X, Hubbard RE, Bonn T, Engstrom O, Ohman L, Greeen GL, Gustafsson J-A, Carlquist M. (1997) Molecular basis of agonism and antagonism in the oestrogen receptor. Nature 389:753–758.[CrossRef][Medline]
Bulger WH, Muccitelli RM, Kupfer D. (1978a) Interactions of methoxychlor, methoxychlor-base soluble contaminant, and 2,3,-bis(p-hydroxypnenyl)-1,1,1-trichlorethane with rat uterine estrogen receptor. Toxicol. Environ. Health 4:881–893.[ISI][Medline]
Bulger WH, Muccitelli RM, Kupfer D. (1978b) Studies on the in vivo and in vitro estrogenic activities of methoxychlor and its metabolites. Role of hepatic mono-oxygenase in methoxychlor activiation. Biochem. Pharmacol. 27:2417–2423.[CrossRef][ISI][Medline]
Cheng Y and Prusoff WH. (1973) Relationship between the inhibition constant (Ki) and the concentration of inhibitor which causes 50 percent inhibition (IC50) of an enzymatic reaction. Biochem. Pharmacol. 22:3099–3108.[CrossRef][ISI][Medline]
Clark JH and Gorski J. (1969) Estrogen receptors: An evaluation of cytoplasmic-nuclear interactions in a cell-free system and a method for assay. Biochem. Biophys. Acta 192:508–515.[Medline]
Danzo BJ. (1997) Environmental xenobiotics may disrupt normal endocrine function by interfering with the binding of physiological ligand to steroid receptors and binding proteins. Environ. Health Perspect. 105:3294–301.[ISI][Medline]
Elsby R, Ashby J, Sumpter A, Brooks N, Pennie WD, Maggs JL, Lefevre PA, Odum J, Beresford N, Paton D, et al. (2000) Obstacles to the prediction of estrogenicity from chemical structure: Assay-mediated metabolic transformation and the apparent promiscuous nature of the estrogen receptor. Biochem. Pharmacol. 60:1519–1530.[CrossRef][ISI][Medline]
Gray LE, Ostby J Jr,, Ferrell J, Rehnberg G, Linder R, Cooper R, Goldman J, Slott V, Laskey J. (1989) A dose-response analysis of methoxychlor-induced alterations of reproductive development and function in the rat. Fundam. Appl. Toxicol. 12:92–108.[CrossRef][ISI][Medline]
GraphPad Software Inc. Introduction to Radioligand Binding Available at: http//www.graphpad.com/curvefit/introduction9e.htm. (accessed September 2006).
Hammond B, Katzenellenbogen BS, Krauthammer H, McConnell J. (1979) Estrogenic activity of the insecticide chlordecone (Kepone) and interaction with the uterine estrogen receptors. Proc. Natl. Acad. Sci. U.S.A. 76:6641–6645.
ICCVAM, NIEATM. (2003a) Background Review Document. Current Status of Test Methods for Detecting Endocrine Disruptors: In Vitro Estrogen Receptor Binding Assays NIH Publication No. 03-4504. Available at: http://iccvam.niehs.nih.gov/methods/endodocs/backgrnd.htm (accessed September 2006).
ICCVAM, NICEATM. (2003b) ICCVAM Evaluation of In Vitro Test Methods for Detecting Potential Endocrine Disruptors: Estrogen Receptor and Androgen Receptor Binding and Transcriptional Activation Assays NIH Publication No. 03-4503. Available at: http://iccvam.niehs.nih.gov/methods/endocrine.htm (accessed September 2006).
Kuiper GGJM, Lemmen JG, Carlsson B, Corton JC, Safe SH, Van Der Sagg PT, Van Der Burg B, Gustafsson J. (1998) Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor ß. Endocrinology 139:104252–4263.
Laws SC, Carey SA, Ferrell JM, Bodman GJ, Cooper RL. (2000) Estrogenic activity of octylphenol, nonylphenol, bisphenol A and methoxychlor in rats. Toxicol. Sci. 54:154–167.
Laws SC, Stoker TE, Goldman JM, Wilson V, Gray LE Jr, Cooper RL. (2006) The U.S. EPA Endocrine Disruptor Screening Program: In vitro and in vivo mammalian Tier 1 screening assays. In Hood RD (Ed.). Developmental and Reproductive Toxicology: A Practical Approach(CRC Taylor & Francis, New York) pp. 489–524.
Lee PC and Lee W. (1996) In vivo estrogenic action of nonylphenol in immature female rats. Bull. Environ. Contam. Toxicol. 57:341–348.[CrossRef][ISI][Medline]
Markaverich BM, Roberts RR, Finney RW, Clark JL. (1983) Preliminary characterization of an endogenous inhibitor of 3H-estradiol binding in rat uterine nuclei. J. Biol. Chem. 258:11663–11671.
McGovern SL, Helfand BT, Feng B, Schoichet BK. (2003) A specific mechanism of nonspecific inhibition. J. Med. Chem. 46:4265–4272.[CrossRef][ISI][Medline]
Motulsky H and Christopoulos A. (2003) The law of mass action. Fitting Models to Biological Data Using Linear and Nonlinear Regression(GraphPad Software Inc., San Diego, CA) pp. 187–191 www.graphpad.com/manuals/Prism4/RegressionBook.pdf (accessed September 2006).
Nagel SC, con Saal FS, Thayer KA, Dhar MG, Boechler M, Welshons WV. (1997) Relative binding affinity-serum modified access (RBA-SMA) assay predicts the relative in vivo bioactivity of the xenoestrogens bisphenol A and octylphenol. Environ. Health Perspect. 105:70–76.[ISI][Medline]
Nagel SC, vom Saal FS, Welshons WV. (1998) The effective free fraction of estradiol and xenoestrogens in human serum measured by whole cell uptake assays: Physiology of delivery modifies estrogenic activity. Proc. Soc. Exp. Biol. Med. 217:300–309.[Abstract]
Ryan AJ, Gray NM, Lowe PN, Chung C. (2003) Effect of detergent on "promiscuous" inhibitors. J. Med. Chem. 46:3448–3451.[CrossRef][ISI][Medline]
Routledge EJ and Sumpter JP. (1997) Structural features of alkylphenolic chemicals associated with estrogenic activity. J. Biol. Chem. 272:63280–3288.
Seidler J, McGovern SL, Thompson HD, Schoichet BK. (2003) Identification and prediction of promiscuous aggregating inhibitors among known drugs. J. Med. Chem. 46:4477–4486.[CrossRef][ISI][Medline]
Takemura H, Ma J, Sayama K, Terao Y, Zhu BT, Shimoi K. (2005) In vitro and in vivo estrogenic activity of chlorinated derivatives of bisphenol A. Toxicology 207:2215–221.[CrossRef][ISI][Medline]
Tsai M and O'Malley BW. (1994) Molecular mechanisms of action of steroid/thyroid receptor superfamily members. Annu. Rev. Biochem. 63:451.[CrossRef][ISI][Medline]
U.S. Environmental Protection Agency (U.S. EPA). (1998a) EDSTAC Final Report. Chapter 5. Screening and Testing. Available at: http://www.epa.gov/scipoly/oscpendo/edspoverview/finalrpt.htm (accessed September 2006).
U.S. Environmental Protection Agency (U.S. EPA). (1998b) EDSTAC Final Report. Appendix L. Protocols for Tier I Screening Assays. Available at: http://www.epa.gov/scipoly/oscpendo/edspoverview/finalrpt.htm (accessed September 2006).
U.S. Environmental Protection Agency (U.S. EPA). (2006) Endocrine Disruptor Screening Program. EDSP Background. Available at: http://www.epa.gov/scipoly/oscpendo/edspoverview/background.htm (accessed September 2006).
White R, Jobling S, Hoare SA, Sumpter JP, Parker MG. (1994) Environmentally persistent alkylphenolic compounds are estrogenic. Endocrinology 135:1175–182.[Abstract]
Yamasaki K, Noda S, Imatanaka N, Yakabe Y. (2004) Comparative study of the uterotrophic potency of 14 chemicals in a uterotrophic assay and their receptor-binding affinity. Toxicol. Lett. 146:2111–120.[CrossRef][ISI][Medline]
Zacharewski T. (1997) In vitro bioassays for assessing estrogenic substances. Environmental Science and Technology 31:613–623.[CrossRef]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||