Response-Surface Modeling of the Effect of 5{alpha}-Dihydrotestosterone and Androgen Receptor Levels on the Response to the Androgen Antagonist Vinclozolin

Susan Y. Euling*,1, Chris Gennings{dagger}, Elizabeth M. Wilson{ddagger}, Jon A. Kemppainen{ddagger}, William R. Kelce§ and Carole A. Kimmel*

* National Center for Environmental Assessment, U.S. Environmental Protection Agency (8623-D), Ariel Rios Building, 1200 Pennsylvania Avenue NW, Washington, DC 20460; {dagger} Department of Biostatistics, Virginia Commonwealth University, Richmond, Virginia 23298; {ddagger} Laboratories for Reproductive Biology, Department of Pediatrics, and Department of Biochemistry and Biophysics, University of North Carolina, Chapel Hill, North Carolina 27599; and § Drug Development Toxicology, Pharmacia Corporation (7226–300–228), 7000 Portage Road, Kalamazoo, Michigan 49001

Received March 25, 2002; accepted June 27, 2002


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Androgens secreted by the testes bind the androgen receptor in developing target tissues to induce the expression of genes required for male sexual differentiation and development. Androgen concentration and androgen receptor levels vary in male reproductive target tissues during development. Exposure to environmental androgen antagonists during critical windows of fetal and postnatal development can inhibit male sexual development by blocking transcription of androgen-dependent genes. As the sensitivity to androgen antagonists under conditions of varying androgen concentrations and varying androgen receptor levels is unknown, we used a luciferase reporter assay to investigate the transcriptional effects of a known androgen antagonist (the vinclozolin metabolite M2) at different androgen concentrations and different androgen receptor levels. The ability of M2 to inhibit transcription was greater at lower concentrations of androgen (5{alpha}-dihydrotestosterone) and androgen receptor. The data were modeled to determine the dose-response surface of M2 and androgen receptor concentrations at different 5{alpha}-dihydrotestosterone levels and the relationship of the 3 components to the response. The model and hypothesis testing results suggest that, at 0.01 and 0.1 nM 5{alpha}-dihydrotestosterone concentrations within the expected in vivo range of free androgen levels during development, the response-surface shapes were similar and the interaction of the androgen receptor and M2 concentrations to the response were similarly antagonistic. Thus, two components of the developmental stage, androgen and androgen receptor concentrations, are critical for sensitivity to the inhibitory effects of an androgen antagonist.

Key Words: endocrine disruptor; androgen antagonist; antiandrogen; vinclozolin; critical windows; male development; response-surface modeling; testosterone; androgen receptor.


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Understanding the effect of developmental stage on the response to an endocrine disrupting chemical (EDC) is of interest, since hormone levels vary with the developmental stage, and EDCs effectively add or subtract from the endogenous hormone activity levels. The developmental stage may impact the response to the antiandrogens because androgens are critical for a number of male developmental events. Androgens secreted from the testes are critical for male urogenital tract differentiation in eutherian mammals. In the presence of testes in the males, the Müllerian ducts regress and the Wolffian ducts differentiate into the epididymis, vas deferens, and seminal vesicles (George and Wilson, 1994Go). In experiments with castrated male or female rabbits at the indifferent gonad stage, female development of the urogenital tract occurs; the Müllerian ducts develop into fallopian tubes, uterus, and part of the vagina, while the Wolffian ducts regress (George and Wilson, 1994Go; Jost, 1953Go, 1972Go). Testosterone (T) is transported in blood to target cells where T and its active metabolite, 5{alpha}-dihydrotestosterone (DHT), bind the androgen receptor (AR) to form the androgen-AR complex (Fig. 1AGo). The complex binds as a dimer to androgen response elements, thereby activating or repressing the expression of androgen-dependent gene products. Transcriptional activation also requires the binding of coactivator proteins, proteins that mediate protein–protein interactions with the AR and RNA polymerase molecules at the promoter region (McKenna et al., 1999Go; Fig. 1AGo).



View larger version (27K):
[in this window]
[in a new window]
 
FIG. 1. (A) Androgen action in an in vitro reporter gene assay. The AR expression vector, pCMVhAR and the mouse mammary tumor virus promoter-luciferase reporter vector were cotransfected into CV-1 cells (see Materials and Methods). The AR expression vector produces AR protein. The AR-DHT or AR-T complexes bind as dimers to the androgen response element DNA regulatory regions to activate androgen-dependent transcription. As a result, the cells express the luciferase protein, whose activity is measured by the emission of light after the cells are extracted and substrate is added. (B) The effect of endogenous AR ligand, an AR agonist, and an AR antagonist on luciferase expression in an in vitro assay. The endogenous AR ligands, DHT and testosterone, bind to the AR and the complex activates luciferase transcription via the steps outlined in Figure 1AGo. An AR agonist binds to the AR, and the AR agonist-AR complex binds to DNA, leading to transcriptional activation of the luciferase reporter vector. An AR antagonist binds to the AR, decreasing formation of DHT-AR complexes, which leads to reduced luciferase reporter gene transcription. AR, androgen receptor; DHT, dihydrotestosterone; ARE, androgen response element; T, testosterone; MMTV-Luc, mouse mammary tumor virus promoter-luciferase reporter vector.

 
Androgen action is required for a number of male development events. In humans, urogenital tract sexual differentiation occurs between 8 and 18 weeks of gestation, and external genitalia differentiation occurs between 9 and 12 weeks of gestation. Male pubertal development is also androgen-dependent, beginning at approximately 12 years of age. The critical window for exposure to chemicals that affect a developmental event typically precedes and overlaps the time of visible structural development. Exposure to chemicals that inhibit androgen action during these critical windows may interfere with these male sexual developmental events (Euling and Kimmel, 2001Go). One class of androgen inhibitors is the androgen antagonists, which block androgen action by binding to the AR ligand binding domain, thereby excluding endogenous androgens from regulating androgen-dependent transcription (Fig. 1BGo). The androgen antagonist vinclozolin is a fungicide that prevents spore germination (Tomlin, 1997Go). The two vinclozolin metabolites M1 (2-[[3,5-dichlorophenyl)-carbamoyl]oxy]-2-methyl-3-butenoic acid) and M2 (3', 5'-dichloro-2-hydroxy-2-methylbut-3-enanilide) have been shown to bind to the AR in vitro and to alter androgen-dependent transcription in vivo (Kelce et al., 1994Go, 1997Go; Wong et al., 1995Go). While a number of uses are being phased out, vinclozolin is currently used on a few U.S. and imported fruit and vegetable crops as well as on golf course turf grass, suggesting that human exposure to vinclozolin may occur by ingestion via pesticide residues on foods, inhalation via proximity to pesticide application, or dermal exposure via pesticide application and contact with turf and plants (U.S. EPA, 2001Go).

In vivo dose-response data for vinclozolin male rat developmental effects have revealed at least two critical windows of exposure during prenatal and prepubertal development. Vinclozolin exposure to male rats via the pregnant dam during the time of sexual differentiation, gestation day 14 to postnatal day 3, at 100 mg/kg/day resulted in a female-like anogenital distance, nipple retention, cleft phallus with hypospadias, suprainguinal ectopic scrota/testes, vaginal pouch, epididymal granulomas, and small to absent sex accessory glands (Gray et al., 1994Go). Statistically significant effects on anogenital distance were observed after treatment with vinclozolin doses as low as 12.5 mg/kg/day using the same treatment regimen (Gray et al., 1999Go). Prenatal vinclozolin treatment studies suggest that the critical window for exposure is GD 14–19 (Wolf et al., 2000Go). Additional antiandrogenic effects have been observed following prepubertal exposure: vinclozolin (100 mg/kg/day) treatment beginning on postnatal day (PND) 22 (day of weaning) led to delayed pubertal onset (as measured by preputial separation), decreased weight of the epididymis, ventral prostate, and seminal vesicles as well as increased serum testosterone, 5 {alpha}-androstanediol, and luteinizing hormone in the male rat (Monosson et al., 1999Go). In vitro and in vivo results predict that the observed developmental effects are a consequence of inhibited or blocked expression of androgen-dependent genes in androgen-responsive tissues (Kelce et al., 1995).

The two critical windows for vinclozolin susceptibility likely correspond to the time of androgen-dependent male sexual development and differentiation events. Androgen action is predicted to be affected by the concentrations of androgens, AR, sex hormone binding globulin and accessory proteins required for transcriptional activity. T and AR levels vary throughout development in humans and rodents (reviewed by Euling and Kimmel, 2001Go). The free T concentration ([free T]) ranges during the critical windows for external genitalia differentiation, urogenital tract differentiation, and puberty are shown in Table 1Go. Compared to 0.17–0.73 nM free T in adult males (0.05–0.21 ng/ml; Fisher, 1998Go), [free T] during these three critical windows is lower (Table 1Go) on average possibly leading to an increased susceptibility of the fetus to effects of androgen antagonists compared to the adult. While less is known about human AR concentrations and tissue expression levels during critical windows of male development, rodent data indicate that AR is detected in urogenital tissues coincident with urogenital tract differentiation and the presence of relatively high fetal androgen levels (Bentvelsen et al., 1995Go; Cooke et al., 1991Go).


View this table:
[in this window]
[in a new window]
 
TABLE 1 Free Androgen Concentrations in Serum during 3 Critical Windows of Human Male Development
 
To understand the relationship between AR and androgen concentrations on the sensitivity of AR-mediated transcription to androgen antagonists, we asked whether the potency of the pesticide chemical vinclozolin was dependent on the three-way dose combinations. The data were modeled to allow predictions at untested dose combinations. To this end, an in vitro luciferase reporter gene assay was used to investigate:

This assay allowed for the manipulation of the concentrations of two developmental stage components, AR and DHT, and for direct measurement of gene expression changes resulting from effects on androgen-AR complex transcriptional activation (Fig. 1AGo). This transient cotransfection model system performed in cultured cells has been used previously to characterize the antagonistic effects of vinclozolin (Wong et al., 1995Go) and other AR antagonists (Kemppainen et al., 1999Go) and reflects the in vivo AR activity in terms of high affinity induction of gene transcription by testosterone and DHT and its modulation by coactivator proteins (He et al., 1999Go, 2001Go). Dose-response data were generated at different concentration combinations of M2, AR, and DHT and statistical modeling of the dose-response data was performed to address the four questions. Response-surface modeling of the dose-response data was used to determine whether each component had an additive, synergistic, or antagonistic effect on inhibition of transcription by M2 and to compare the shapes of the response surfaces at each DHT concentration. Statistical analysis (parameter estimation and associated hypothesis testing) was used to determine whether the relationship of AR concentration to the M2 response changed at different DHT concentrations.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Transient transfection assay.
AR transcriptional activity was determined in monkey kidney CV1 cells maintained in Dulbecco’s modified Eagle medium (GIBCOBRL) containing 20 mM Hepes, pH 7.2 (DMEM-H) and antibiotics. Cells were plated at 4.25 x 105 cells/6 cm dish and 24 h later transfected with the human AR expression vector pCMVhAR (10–200 ng DNA/dish as indicated) and the mouse mammary tumor virus promoter luciferase reporter vector (5 µg/dish) using the calcium phosphate DNA precipitation method (Fig. 1AGo; Kemppainen et al., 1999Go). After transfection, cells were placed in DMEM-H phenol-red free media containing 0.2% bovine calf serum and incubated for 24 h at 37°C. Cells were placed in phenol-red free and serum-free media in the absence or presence of the indicated concentrations of DHT and M2, a metabolite of the pesticide vinclozolin (3-(3,5-dichlorophenyl)-5-methyl-5-vinyloxazolidine-2, 4-dione), and incubated for 24 h (Gray et al., 1994Go; Kelce et al., 1994Go, 1998Go; Wong et al., 1995Go). Cells were harvested in 0.5 ml lysis buffer (Ligand Pharmaceuticals) and luciferase activity was determined using a Monolight 2010 luminometer (Analytical Luminescence Laboratory, San Diego). Luciferin, the substrate for the light reaction, was added automatically by the luminometer. The response was measured by the amount of light emitted in optical units and was proportional to the amount of luciferase gene product produced. This response indicates that the DHT-AR complex had bound to the androgen response element DNA and activated luciferase gene transcription (Fig. 1AGo). The mean luciferase light activity was recorded for two replicates at each concentration combination.

Twelve separate experiments were conducted using various concentration combination designs for AR vector, M2, and DHT concentrations. The M2 concentrations ([M2]) that were tested ranged from 50 nM to 1 µM. The DHT concentrations ([DHT]) ranged from 0.001 to 1 nM; the range included the maximal dose response range of luciferase expression in the in vitro assay with the maximum response occurring at approximately 0.1 nM DHT, the equilibrium binding constant (Kd) for the androgen receptor. The response at 0.001 nM DHT was at the limit of detection and at 1 nM DHT sometimes showed a decrease in luciferase response, likely reflecting a limiting concentration of transcription cofactors. The AR expression vector concentration ([AR vector]) ranged from 10 to 200 ng/dish. For each experiment, duplicate data points were performed for each three-way dose combination. For the data analysis and modeling, dose-response data from all 12 experiments was included, with the exception that response data for AR vector at 200 ng/dish were excluded. Of the 12 experiments, four included 25 ng/dish AR vector; 0, 50, 100, 200, 400, or 600 nM M2; and 0.01, 0.1, or 1 nM of DHT. Three experiments used either 10 or 50 ng/dish of AR vector; 0, 50, 100, 250, or 500 nM M2; and 0.01, 0.1, or 1 nM DHT. The remaining 5 experiments used designs with concentrations that were different but overlapped with each other and those in the 7 experiments (above); as a group, these included 0.001, 0.01, 0.1, and 1 nM DHT; 0, 10, 50, 100, 200, 500, and 1000 nM M2; and 10, 20, 25, 50,100, and 200 ng/dish AR vector. The amount of AR expression vector DNA of 1 µg or less was shown to be directly proportional to the level of expressed AR protein, using the method of calcium phosphate precipitation for DNA transfection and by densitometric analysis of protein on immunoblots (Choong et al., 1996Go). AR expression vector concentration was therefore referred to as AR concentration.

The luciferase activity response data were modeled as the fraction of mean control to normalize the response to the mean response of the positive controls (i.e., controls with DHT and without M2). The fraction of mean control was calculated as (mean luciferase response for replicates #1 and 2)/(DHT positive control luciferase response). Data are presented in this manner to remove interexperiment variability in the luciferase response, allowing for data from different experiments to be analyzed together. The sample means and variances for each concentration group were calculated.

Statistical modeling and hypothesis testing.
To determine the relationships among the three components of the system, [DHT], [AR], and [M2], on the luciferase response and to account for the intra- and interexperiment variability, a multivariate nonlinear model with associated hypothesis testing was performed. A modeling scheme was based on fitting the observed sample means across replicate values (Seber and Wild, 1989Go). Thus, the sample mean for each concentration group was the average of two values. The assumption was made that the interexperiment variability (variances across experimental groups) was constant. The model accounted for the inter- and intra-experiment variability. However, it allowed for the assumption that intra-experiment observations were correlated. The correlation structure assumed that the intra-experiment variability was compound symmetric, which assumes a nonzero common correlation for within-experiment observations. Hypothesis testing used a general linear hypothesis with an appropriate F test.

The definition of additivity used is given by Berenbaum (1985) and is based on the classical isobolograms for the combination of two chemicals (Loewe, 1953Go; Loewe and Muischnek, 1926Go). For a combination of c chemicals, let Xi represent the concentration of the ith component alone that yields a fixed response, y0, and let xi represent the concentration of the ith component in combination with the c agents that yields the same response. According to this definition of additivity, if the substances combine with zero interaction, then

(1)

The left side of Equation 1Go is equal to 1 for an additive relationship (either a linear addition or subtraction of response) between the chemical components and the response. The left side of Equation 1Go is less than 1 for a synergistic relationship between the components measured and the response. The left side of Equation 1Go is greater than 1 for an antagonistic relationship between the components and the response.

Analysis using a Gompertz Nonlinear Model.
A nonlinear model was chosen for these data as a sigmoid-shaped relationship was expected for dose-response data. The form of the model was based on a Gompertz model parameterized as follows:

(2)
where yij is the average log (fraction of mean control +1) from the jth experimental group and the ith experiment, j = 1, ... , Ji; and i = 1, ... , 12; /{gamma} is a user-specified parameter associated with the maximum plateau effect in the experiment, the p-dimensional vector ß consists of unknown parameters associated with the intercept, slopes, and interaction terms on the complementary log-log scale, xj is a p-dimensional vector denoting the [M2], [DHT] and [AR] in the jth group, and {varepsilon}i is the unobserved random error vector from the ith experiment, assumed to be independent and normally distributed with mean vector zero and variance {Sigma}.

For these data, /{gamma} was assumed known and was specified at two values that were above the observed responses (2 and 5) for comparison. Here, the parameter /{gamma} was fixed as the response-surface did not plateau within the experimental region, and the maximum response was not observed. To account for the intra-experimental variability of the luciferase response in this assay, within-experiment observations could be correlated. Observations across experiments were assumed to be independent. The response means from duplicates of each dose combination and variances for each group were calculated. The model given in Equation 2Go was fit to the sample means. The variance of each group was assumed to be constant across concentrations. In addition, the correlation of observations within an experiment was assumed to be constant. Thus, a compound symmetric correlation structure was assumed for (yij, yij', where j != j'). The method of maximum likelihood was used to estimate unknown model parameters using the MIXNLIN macro in SAS version 6.12 (SAS, 1999).

The DHT concentrations tested were 0.001, 0.01, 0.1, and 1 nM. As the model given in Equation 2Go is based on the arithmetic concentration scale and not the log scale, the concentrations of DHT do not uniformly span the experimental region. One concern was that the responses at 1 nM DHT may overly influence the model fit and associated inference. Therefore, to address the effect of the fit of DHT concentrations, we considered a model that allowed for a response surface associated with AR and M2 at each fixed DHT level. More specifically, x'ß in Equation 2Go was parameterized as follows:

(3)
where k denotes the [DHT] (0.001, 0.01, 0.1, or 1) (nM)
x1 is the [M2] (nM)
x2 is the [AR] (ng/dish)
ß0_k is an unknown parameter associated with the intercept on the complemen tary log-log scale at the kth fixed [DHT]
ß1_k is an unknown parameter associated with the slope of M2 at the kth fixed [DHT]
ß2_k is an unknown parameter associated with the slope of AR at the kth fixed [DHT]
ß12_k is an unknown parameter associated with the interaction between M2 and AR (M2*AR) at the kth fixed [DHT].
Carter et al. (1988) and Gennings (2000) demonstrated that when the interaction or cross-product term (ß12_k) is zero, Berenbaum’s (1985) interaction index is equal to 1 and the isobologram is linear and coincident with the "line of additivity," and indicates an additive or "no interaction" case. Otherwise, the algebraic sign of the interaction term can be used to characterize the interaction as synergistic or antagonistic.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Dose-response data.
To investigate whether the concentrations of DHT and AR alter the inhibitory effect of M2 on DHT-dependent luciferase transcription, the luciferase response was assessed in various three-way dose combinations of [M2], [DHT], and [AR] using the reporter gene transfection assay. In the presence of 50 or 100 ng AR DNA and in the absence of DHT, a small background response level was observed (Fig. 2Go). In the presence of DHT and AR, a large positive luciferase response was observed. At a fixed level of AR (50 ng pCMVhAR expression vector/dish), the inhibitory effect of M2 on DHT-dependent luciferase expression was observed (Figs. 2B–2DGo). Inhibition by M2 was difficult to quantitate at 0.001 nM DHT and 50 ng AR DNA/dish sinceexpression levels were low in the positive controls. At 0.001 nM DHT and 100 ng AR DNA/dish, a greater transcriptional response was observed, which was inhibited by 200 nM M2 (Fig. 2AGo). At 0.01 nM DHT and 50 ng/dish of AR, 50 nM M2 led to ~50% reduction in luciferase response (Fig. 2BGo). At 0.1 and 1 nM DHT at the same AR level, 50 nM M2 had no significant effect on luciferase activity relative to the positive control, and a higher M2 concentration was required to significantly reduce the luciferase response (Figs. 2C and 2DGo). The lowest tested M2 concentration that led to a significant reduction in luciferase response was 50 nM M2 at 0.01 nM DHT, 100 nM M2 at 0.1 nM DHT, and 250 nM M2 at 1 nM DHT. Figure 3Go (A, D, G, and J) presents the two-way plots of average fraction of mean luciferase response as a function of concentrations of M2 and AR, at the four tested concentrations of DHT. As [AR] was increased, the inhibitory effect of [M2] on luciferase expression decreased (Figs. 3A, 3D, 3G, and 3JGo). The effect of DHT concentration can also be visualized by noting the differences of response between the four plots; as [DHT] increases, the luciferase response is greater and inhibition by the same [M2] is less.



View larger version (24K):
[in this window]
[in a new window]
 
FIG. 2. Mean luciferase expression response, in optical units, after addition of different concentrations of DHT and M2 in the reporter gene assay. Data shown are from representative experiments. The mean luciferase values and standard error were calculated from 2 duplicates for each three-way dose combination. The error bars indicate the standard deviation in both directions. In all experiments, 5 µg/dish of the MMTV luciferase reporter vector was transfected (see Materials and Methods). The DHT and transfected AR expression vector (pCMVhAR) concentrations are given below each graph.

 


View larger version (49K):
[in this window]
[in a new window]
 
FIG. 3. Experimental response data and predicted response plots at different M2 and AR concentrations shown for different DHT concentrations. A, B, and C are at 0.001 nM DHT; G, H, and I are at 0.01 nM DHT; J, K, and L are at 1 nM DHT. A, D, G, and J are 3-D graphs of the average luciferase activity (% control) as a function of [AR] and [M2] at fixed [DHT]. The 3-D plot illustrates the coverage of the data points for different M2 and AR concentrations (excluding 200 ng AR vector/dish) at different DHT concentrations tested in the in vitro assays. Each circle represents the average response from 1–5 duplicate assays tested at the particular AR and M2 dose combination. The total number of data points (each dish counted) at each DHT concentration is 30 at 0.001 nM, 122 at 0.01 nM, 152 at 0.1 nM, and 124 at 1 nM. B, E, H, and K are three-dimensional predicted response surfaces and C, F, I, and L are contour plots for different predicted response levels based upon the model. The 2-D contour plots, which allow for comparisons of dose combinations by response level, and 3-D response surfaces are different graphical representations of the model predictions at the same DHT concentration. The model was based on Equation 2Go and the parameters given in Equation 3Go (see Materials and Methods) and was tested using the data from AR concentrations <200 ng/dish (231 data points).

 
Tests of additivity using statistical modeling.
To determine the relationships among the three components of the system, [DHT], [AR], and [M2], to the luciferase response, accounting for the intra- and interexperiment variability, a statistical model with associated hypothesis testing was performed. In order to determine the statistical significance and direction of the effect of [M2], [AR], and the interaction between M2 and AR (M2*AR) on the luciferase response at different levels of DHT, the model given in Equations 2 and 3GoGo was fitted to the data and the resulting parameter estimates were generated (Table 2Go). The parameter /{gamma} is associated with the maximum level of the response surface. Since the response data continues to rise at the upper boundaries of the experimental data region (Figs. 3A, 3D, 3G, and 3JGo), the maximum response is an extrapolation of the data; /{gamma} was fixed at 2 and 5 because these values were above the observed response level. The p values were similar when /{gamma} = 2 or 5 (data not shown). Conditional estimates of the other parameters were calculated and these estimates are presented for /{gamma} = 2 (Table 2Go). The intercept parameter was added to the model to allow for flexibility; i.e., the intercept was not required to go through zero. The intercept is associated with the background response at a given DHT concentration, i.e., in the absence of M2.


View this table:
[in this window]
[in a new window]
 
TABLE 2 Estimated Model Parameters for the Response Surface of the Full Model Given in Equations 2 and 3GoGo Where /{gamma} is Fixed at 2
 
An overall test of additivity evaluates whether the interaction terms (i.e., the ß12_k parameters) are simultaneously equal to zero, indicating that the response relationship between the components [M2], [AR], and [DHT] is the same (additive) at all four DHT levels tested. Four response surfaces associated with the fixed levels of DHT were estimated simultaneously (Table 3Go). The null hypothesis of overall additivity among the three components was rejected since the associated p value was less than 5% (p < 0.001), indicating that additivity at all doses tested was unlikely. Therefore, either a synergistic or antagonistic relationship among the three components on the response is likely.


View this table:
[in this window]
[in a new window]
 
TABLE 3 Hypothesis Testing Results for Overall Additivity and Comparison of Response-Surface Shapes
 
As the overall test of additivity was rejected, the response surfaces at the different DHT concentrations were compared to assess the relationship of M2 and AR at each [DHT]. The response surface at 0.1 nM DHT was selected as the reference surface since it is the tested concentration closest to the Kd for DHT (Wilson and French, 1976Go). The response surfaces at the other DHT concentrations were compared on the basis of slope and interaction parameters to the reference surface. The null hypothesis that the slope and interaction terms are the same in the comparison and reference surfaces was tested. The null hypothesis was rejected for the comparison between the 0.001 nM DHT response surface and the reference surface and between the 1 nM DHT response surface and the reference surface. However, the null hypothesis was not rejected when comparing the 0.01 nM DHT response surface to the reference surface (p = 0.073, Table 3Go). Thus, the response surface shapes at 0.001 and 1 nM DHT were significantly different from that at 0.1 nM DHT, whereas the response surface shapes at 0.01 nM, and 0.1 nM DHT were not significantly different.

Effect of [M2], [AR] and their interaction at fixed levels of DHT.
To determine whether [M2] and [AR] had an additive effect on the response, the fitted four-dimensional relationships are described as three-dimensional surfaces at fixed DHT concentrations (Figs. 3B, 3E, 3H, and 3KGo). Following the work of Carter et al. (1988) and Gennings (2000), the detection of an M2*AR interaction is equivalent to finding an isobologram with a curvilinear relationship different from the line of additivity. If an M2*AR interaction was observed, then the null hypothesis of additivity was rejected.

At 0.001 nM DHT, the parameter associated with the effect of [M2] was positive but not significant (p = 0.081, Table 2Go) and the parameter associated with [AR] was negative but not significant (p = 0.459, Table 2Go). These results indicate that the effect of [AR] and [M2] did not significantly alter the response at 0.001 nM DHT. M2*AR was negative and borderline significant at 0.001 nM DHT (p = 0.057, Table 2Go), indicating that increasing concentration combinations of AR and M2 decrease the luciferase response to DHT. All contour lines bow below the line of additivity (Fig. 3CGo) indicating a consistently synergistic relationship between M2 and AR at 0.001 nM DHT.

At 0.01 nM DHT, the parameters associated with the effect of [M2] and [AR] on the response were negative and significant (M2, p < 0.001; AR, p = 0.025, Table 2Go), suggesting that [M2] and [AR] exhibited an inhibitory effect on the response. The M2*AR interaction was positive and significant (p = 0.033, Table 2Go) suggesting that M2 and AR have an antagonistic interaction. This result indicates that the response to M2 inhibition of luciferase activity decreases as [AR] increases. While some contour lines bow up and others bow down (Fig. 3FGo), all contours are above their corresponding line of additivity and thus are consistent with an antagonistic relationship between M2 and AR at 0.01 nM DHT.

At 0.1 nM DHT, the parameter associated with the effect of [M2] was negative and significant (p < 0.001, Table 2Go), indicating that [M2] exhibited an inhibitory effect on the response. The parameter associated with the effect of [AR] was negative and not significant at 0.1 nM DHT (p = 0.965, Table 2Go), indicating that the effect of [AR] on the response to DHT did not significantly alter the response to M2. The M2*AR was positive and significant (p < 0.001, Table 2Go), indicating that the inhibiting effect of [M2] on luciferase response diminished with an increase in [AR] (Figs. 3H and 3IGo). This result suggests an antagonistic relationship between [AR] and [M2]. The negative slope of M2 can be visualized in Figure 3HGo at the edge of the surface corresponding to 10 ng/dish AR. The edge at 100 ng/dish AR demonstrates a less steep, negative slope and this change in slope is due to the M2*AR. The corresponding contour plot (Fig. 3IGo) shows all response contours bow above the line of additivity, indicating a consistently antagonistic M2*AR relationship at 0.1 nM DHT.

At 1 nM DHT, the parameters associated with the effect of [M2] and [AR] were negative but not significant (M2, p = 0.131; AR, p = 0.749, Table 2Go) indicating that neither the effect of [AR] nor [M2] significantly altered the response to DHT. The parameter associated with M2*AR did not show a significant effect on the fitted response surface (p = 0.674, Table 2Go; Figs. 3K and 3LGo), indicating that their interaction cannot be determined with the modeled data at 1 nM DHT.


    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
We were interested in determining the effect of developmental stage on the response to an EDC. This study tested the effect of two components of the developmental stage, androgen and androgen-receptor concentrations, on the response to the antiandrogen vinclozolin in an in vitro AR-dependent transgene assay. The transgene expression response was measured after treatment with various three-way dose combinations of the antiandrogen, androgen, and androgen receptor concentrations. The results indicate that as [DHT] increases, the inhibitory effect of M2 decreases, which was expected since the AR has a high binding affinity for DHT compared to M2 (AR, Kd = 0.2–0.5 nM, Wilson and French, 1976Go; M2, Ki = 9.7 µM, Kelce et al., 1994Go). The data were modeled to predict responses at untested dose combinations. The model found the relationship between the antiandrogen concentration and the AR concentration to be similarly antagonistic at 0.01 and 0.1 nM DHT, whereas the relationships and response surfaces behaved differently at 0.001 and 1 nM DHT. The model suggests that androgen and androgen receptor concentrations can affect the response to an antiandrogen.

The relevance of the DHT concentrations tested in vitro to physiological androgen levels was explored by identifying androgen concentrations during critical windows for human male urogenital tract, external genitalia, and pubertal development in the literature. Free androgen concentrations in vivo were considered comparable to the DHT concentrations used in this study because serum binding proteins (i.e., sex hormone binding globulin) were not present during the in vitro assay. DHT concentrations were compared to [free T] since [free T] but not [free DHT] during these critical windows of human development were identified in the literature, and responses to T and DHT in the AR-mediated luciferase reporter gene assay were similar (Kemppainen et al., 1999Go). At stages overlapping the critical window for male urogenital tract differentiation, the [free T] measurements in cord blood range from ~0.05–0.5 nM (Table 1Go; Abramovich and Rowe, 1973Go; Diez d’Aux and Murphy, 1974). Within the critical window for external genitalia differentiation, the [free T] measurements in cord blood ranged from ~0.1–0.5 nM (Table 1Go; Diez d’Aux and Murphy, 1974). During Tanner stage I, a stage just before puberty onset that may correspond to the critical window for pubertal development, the [free T] measurements ranged from ~0.001–0.2 nM. Thus, three of the four tested DHT concentrations, 0.001, 0.01, and 0.1 nM, fell within the expected physiological [free T] range during some androgen-sensitive developmental intervals while 1 nM DHT was above the in vivo [free T] during human development.

At 0.01 and 0.1 nM DHT, the model is most consistent with the data and is the most robust since the greatest number of data points and the most complete coverage of three-way dose combinations were tested (Fig. 3Go). The shapes of the response surfaces were similar indicating that the relationships among the three components were similar at these two DHT concentrations. [M2] exhibited a significant inhibitory effect on the response that is consistent with the mechanism of action of M2 as an androgen antagonist. As predicted by the mode of action for vinclozolin, the inhibitory effect of M2 decreased as [AR] increased at these two DHT concentrations. The additivity tests for [M2] and [AR] indicated that their relationship was significant and antagonistic for the response since the null hypothesis of additivity was rejected. This finding is consistent with the biology of receptor binding, i.e., as [AR] increases, the luciferase response would be expected to increase and, as [M2] increases, the luciferase response would be expected to decrease. As expected, at 0.1 nM and 0.01 nM DHT the response contours of the contour plots (Figs. 3F and 3IGo) suggest a consistently antagonistic relationship between AR and M2 across all dose combinations. At 0.01 nM DHT, [AR] had a significant inhibitory effect on the response, which is inconsistent with the fact that DHT has a higher binding affinity for AR than M2, i.e., within the [AR] range tested, as [AR] increases a greater proportion of DHT-AR compared to M2-AR complexes would be expected to form. Comparable in vivo [AR] during human development was not identified in the literature. Further data on [AR] and distribution during development in humans and rodents are needed to understand comparable in vivo AR levels and their effect on the response to vinclozolin exposure.

The models at 0.001 and 1 nM DHT were less reliable than at either 0.01 or 0.1 nM DHT since they were based on fewer data points and less coverage of the arithmetic dose intervals. The statistics did not reveal significant patterns about the relationship between [M2] and [AR] at 0.001 and 1 nM DHT. The response-surface shapes at 0.001 and 1 nM DHT were significantly different from those at 0.1 nM DHT, but the results of M2*AR were not significant (1 nM DHT) or were borderline significant (0.001 nM DHT). The differences in response-surface shapes may be due to constraints of the in vitro system, and therefore, may not inform the in vivo developmental scenario. For example, at 1 nM DHT, the decreased luciferase response may reflect a limiting concentration of transcription cofactors (i.e., titration of transcription cofactors at saturating [DHT]; "squelching," E.M. Wilson, unpublished results). The DHT level at 0.001 nM is often too low to achieve a robust transcriptional response in the assay (data not shown). A high [AR] was found to be inhibitory and may also be due to titration of transcription cofactors by the AR. Alternatively, changes in the M2*AR relationship to the response at the highest and lowest DHT levels may indicate that at lower DHT levels and in the presence of excess M2 and limited DHT, M2 may antagonize DHT-AR more effectively. Certain EDCs have been observed to behave as an antagonist at relatively high concentrations and as an agonist at very high concentrations (Kemppainen and Wilson, 1996Go; Maness et al., 1998Go; Wong et al., 1995Go).

The results suggest that the response to vinclozolin is altered by changes in androgen and AR concentration, both of which are known to vary during development. If the relationship between DHT and AR concentrations and their effect on M2 inhibition pertain in vivo, susceptibility to vinclozolin exposure could be affected by developmental stage as well as individual differences in AR or DHT concentrations. The model predicts that susceptibility to an equivalent vinclozolin exposure would be expected to be greatest when DHT and AR concentrations are relatively low but sufficiently high to induce a transcriptional response (i.e., in the absence of AR-dependent transcription, inhibition cannot be observed). However, the free androgen concentration cannot be the only factor affecting sensitivity and response to an androgen antagonist. In the human, the free androgen concentration is lower before puberty than it is during male sexual differentiation in utero (Table 1Go) in the rat. It was this period of male sexual development in the fetal rat that was shown to be most susceptible to androgen antagonist effects. Significant effects on anogenital distance at birth, the most sensitive endpoint of sexual differentiation in the developing rat embryo, were observed after treatment at doses as low as 3.125 mg/kg/day (Gray et al., 1999Go). In contrast, effects on the most sensitive endpoint of puberty, the timing of preputial separation, were not observed at 10 or 30 mg/kg/day but were observed at 100 mg/kg/day (Monosson et al., 1999Go).

Decreased sensitivity to antiandrogens prior to puberty versus the time of sexual differentiation in the developing male fetus may result from differences in feedback regulation at the hypothalamic-pituitary-gonadal axis. In the prepubertal and adult male, androgen feedback inhibits luteinizing hormone (LH) production by the anterior pituitary. Exposure to an antiandrogen such as vinclozolin would inhibit androgen-induced downregulation of LH. The subsequent LH-induced rise in testosterone synthesis by the Leydig cells of the testis leads to elevated androgen levels, which compete with antiandrogens for binding to the AR (Monosson et al., 1999Go). The compensatory increase in testosterone thereby minimizes the sensitivity of the prepubertal rat to the effects of antiandrogen exposure. In contrast, in the fetal rat and human male, the developing hypothalamic-pituitary-gonadal axis feedback regulation is not fully functional (Forest, 1985Go; Huhtaniemi et al., 1981aGo,bGo, 1995Go; Quigley, 2002Go; Warren et al., 1987Go). LH production by the fetal male pituitary is suppressed by placental estrogen and fetal testicular androgen production is stimulated in the human by human chorionic gonadotropin (hCG) from the placenta. Because fetal androgens do not regulate LH production, there may be no compensatory rise in testosterone production when challenged with an antiandrogen. The absence of hypothalamic-pituitary-gonadal feedback regulation in the fetus may contribute to the increased sensitivity to antiandrogen exposure during male sexual differentiation.

Developmental stage information on the concentration of components other than androgen and androgen receptor concentrations is needed to further understand the effect of developmental stage on response to an antiandrogen. One other important component is sex hormone binding globulin (SHBG), the mean serum concentration of which decreases from six months to 14 years of age, concomitant with an increase in non-SHBG-bound T (Belgorosky and Rivarola, 1987). While [T] is relatively low during the prepubertal stage, [SHBG] is also low, presumably leading to an increase in unbound T. Additionally, serum SHBG decreased from prepuberty (Tanner stage I) to puberty onset (Tanner stage II), suggesting that androgen-dependent developmental events correspond to a relatively high unbound androgen concentration that is available to bind to AR (Kim et al., 1999Go). Data on SHBG concentrations throughout development would be useful to define its effect on the response to an antiandrogen.

Statistical modeling is useful in risk assessment, because responses can be predicted at untested dose combinations and because all concentration combinations cannot be tested. The modeling approach applied in this paper can be used to predict the effect of developmental stage on the response to exposure to an EDC or to developmental toxic agents in general. A similar modeling approach has been used for a number of mixture studies (Charles et al., 2002Go; Weller et al., 1999Go; Gennings et al., in press). In this study, the "mixture" studied was the combination of the chemical exposure (M2) and two developmental stage components (DHT and AR). The use of an in vitro assay allowed for the manipulation of the concentrations of these two developmental stage components and consequently the effect of different dose combinations, reflecting different developmental stages, on the response to the chemical exposure could be determined by the modeling approach. Determining if the relationships among the developmental stage components and developmental toxic agent exposure change at different dose combinations is an approach that can be applied to a number of scenarios. Understanding the impact of developmental stage and its components is critical to risk assessment for children and susceptible populations. Additionally, the results of this analysis are useful to cumulative risk assessment efforts, especially for chemicals such as DHT and M2, which have the same mode of action but have opposite effects.

The model presented here suggests that developmental stage, as defined by androgen and androgen receptor concentrations can affect the response to an antiandrogen and that the relationship between [AR] and [M2] is similar and antagonistic in the 0.01–0.1 nM DHT range. Information on additional components of developmental stage could be incorporated into a future model to predict the response to an EDC at different developmental stages and the effect of each component on the response. Further expansion of the model could include information reflecting individual sex and racial/ethnic differences in circulating androgen levels during development and in adulthood to predict the response to an EDC.


    ACKNOWLEDGMENTS
 
This work was supported by an internal grant from The National Center for Environmental Assessment, Office of Research and Development at the U.S. Environmental Protection Agency. We gratefully acknowledge Lester Yuan for making the contour and response surface plots in Figure 3Go, and we thank anonymous reviewers for their comments.


    NOTES
 
The views expressed in this paper are those of the authors and do not necessarily reflect the views or policies of the U.S. Environmental Protection Agency. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.

1 To whom correspondence should be addressed. Fax: (202) 565-0078. E-mail: euling.susan{at}epa.gov. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Abramovich, D. R., and Rowe, P. (1973). Foetal plasma testosterone levels at mid-pregnancy and at term: Relationship to fetal sex. J. Endocrinol. 56, 621–622.[ISI][Medline]

Belgorsky, A., and Rivarola, M. A. (1987). Changes in serum sex hormone-binding globulin and in serum non-sex hormone-binding testosterone during prepuberty in boys. J. Steroid Biochem. 27, 291–295.[ISI][Medline]

Bentvelsen, F. M., Brinkmann, A. O., van der Schoot, P., van der Linden, J. E., van der Kwast, T. H., Boersma, W. J., Shroder, F., and Nijman, J. M. (1995). Developmental pattern and regulation by androgens of androgen receptor expression in the urogenital tract of the rat. Mol. Cell. Endocrinol. 113, 245–253.[ISI][Medline]

Berenbaum, M. C. (1985). The expected effect of a combination of agents: The general solution. J. Theor. Biol. 114, 413–431.[ISI][Medline]

Carter, W. H., Jr., Gennings, C., Staniswalis, J. G., Campbell, E. D., and White, K. L., Jr. (1988). A statistical approach to the construction and analysis of isobolograms. J. Am. Coll. Toxicol. 7, 963–973.[ISI]

Charles, G. D., Gennings, C., Zacharewski, T. R., Gollapudi, B. B., and Carney, E. W. (2002). Assessment of interactions of diverse ternary mixtures in an estrogen receptor-{alpha} reporter assay. Toxicol. Appl. Pharmacol. 180, 11–21.[ISI][Medline]

Choong, C. S., Quigley, C. A., French, F. S., and Wilson, E. M. (1996). A novel missense mutation in the amino-terminal domain of the human androgen receptor gene in a family with partial androgen insensitivity syndrome causes reduced efficiency of protein translation. J. Clin. Invest. 98, 1423–1431.[Abstract/Free Full Text]

Cooke, P. S., Young, P., and Cunha, G. R. (1991). Androgen receptor expression in developing male reproductive organs. Endocrinology 128, 2867–2873.[Abstract]

Diez d’Aux, R. C., and Pearson Murphy, B. E. (1974). Androgens in the human fetus. J. Steroid Biochem. 5, 207–210.[ISI][Medline]

Euling, S. Y., and Kimmel, C. A. (2001). Developmental-stage sensitivity and mode-of-action information for androgen agonists and antagonists. Sci. Total Environ. 274, 103–113.[ISI][Medline]

Fisher, D. A., Ed. (1998). Endocrinology test selection and interpretation. Quest Diagnostics Manual, 2nd ed. Quest Diagnostics, Inc., San Juan Capistrano, CA.

Forest, M. G. (1985). Sexual maturation of the hypothalamus: Pathophysiological aspects and clinical implications. Acta Neurochir. (Wien) 75, 23–42.[ISI]

Forest, M. G., Cathiard, A. M., and Bertrand, J. A. (1973). Total and unbound testosterone levels in the newborn and in normal and hypogonadal children: Use of a sensitive radioimmunoassay for testosterone. J. Clin. Endocrinol. Metab. 36, 1132–1142.[ISI][Medline]

Gennings, C. (2000). On testing for drug/chemical interactions: Definitions and inference. J. Biopharm. Stat. 10, 457–467.[Medline]

Gennings, C., Charles, G., Gollapudi, B., Zacharewski, T., Carney, E. (in press). Analysis of resulting data from estrogen receptor reporter gene assays. J. Agric. Biol. Environ. Stat.

George, F. W., and Wilson, J. D. (1994). Sex determination and differentiation. In The Physiology of Reproduction, 2nd ed. (E. Knobil and J.D. Neill, Eds.), pp. 3–28. Raven Press, New York.

Gray, L. E., Jr., Ostby, J. S., and Kelce, W. R. (1994). Developmental effects of an environmental antiandrogen: The fungicide vinclozolin alters sex differentiation of the male rat. Toxicol. Appl. Pharmacol. 129, 46–52.[ISI][Medline]

Gray, L. E., Jr., Ostby, J., Monosson, E., and Kelce, W. R. (1999). Environmental antiandrogens: Low doses of the fungicide vinclozolin alter sexual differentiation in the male rat. Toxicol. Ind. Health 15, 48–64.[ISI][Medline]

He, B., Bowen, N. T., Minges, J. T., and Wilson, E. M. (2001). Androgen-induced NH2- and carboxyl-terminal interactions inhibits p160 coactivator recruitment by activation function 2. J. Biol. Chem. 276, 42293–42301.[Abstract/Free Full Text]

He, B., Kemppainen, J. A., Voegel, J. J., Gronemeyer, H., and Wilson, E. M. (1999). Activation function 2 in the human androgen receptor ligand-binding domain mediates interdomain communication with the NH2-terminal domain. J. Biol. Chem. 274, 37219–37225.[Abstract/Free Full Text]

Huhtaniemi, I. (1995). Molecular aspects of the ontogeny of the pituitary-gonadal axis. Reprod. Fertil. Dev. 7, 1025–1035.[ISI][Medline]

Huhtaniemi, I. T., Katikineni, M., and Catt, K. J. (1981a). Regulation of luteinizing-hormone-receptors and steroidogenesis in the neonatal rat testis. Endocrinology 109, 588–595.[ISI][Medline]

Huhtaniemi, I. T., Katikineni, M., Chan, V., and Catt, K. J. (1981b). Gonadotropin-induced positive regulation of testicular luteinizing hormone receptors. Endocrinology 108, 58–65.[ISI][Medline]

Jost, A. (1953). Problems in fetal endocrinology: The gonadal and hypophyseal hormones. Recent Prog. Horm. Res. 8, 379–418.[ISI]

Jost, A. (1972). A new look at the mechanism controlling sexual differentiation in mammals. Johns Hopkins Med. J. 130, 38–53.[ISI][Medline]

Kelce, W. R., Gray, L. E., and Wilson, E. M. (1998). Antiandrogens as environmental endocrine disruptors. Reprod. Fertil. Dev. 10, 105–111.[ISI][Medline]

Kelce, W. R., Lambright, C. R., Gray, L. E., Jr. and Roberts, K. P. (1997). Vinclozolin and p, p'-DDE alter androgen-dependent gene expression: In vivo confirmation of an androgen receptor-mediated mechanism. Toxicol. Appl. Pharmacol. 142, 192–200.[ISI][Medline]

Kelce, W. R., Monosson, E., Gamcsik, M. P., Laws, S. C., and Gray, Jr. L. E. (1994). Environmental hormone disruptors: Evidence that vinclozolin developmental toxicity is mediated by antiandrogenic metabolites. Toxicol. Appl. Pharmacol. 126, 276–285.[ISI][Medline]

Kemppainen, J. A., Langley, E., Wong, C. I., Bobseine, K., Kelce, W. R., and Wilson, E. M. (1999). Distinguishing androgen receptor agonists and antagonists: Distinct mechanisms of activation by medroxyprogesterone acetate and dihydrotestosterone. Mol. Endocrinol. 13, 440–454.[Abstract/Free Full Text]

Kemppainen, J. A., and Wilson, E. M. (1996). Agonist and antagonist activities of hydroxyflutamide and Casodex relate to androgen receptor stabilization. Urology 48, 157–63.[ISI][Medline]

Kim, M. R., Gupta, M. K., Travers, S. H., Rogers, D. G., Van Lente, F., and Faiman, C. (1999). Serum prostate-specific antigen, sex hormone binding globulin, and free androgen index as markers of pubertal development in boys. Clin. Endocrinol. 50, 203–210.[ISI][Medline]

Loewe, S. (1953). The problem of synergism and antagonism of combined drugs. Arzneim. Forsh. 3, 285–290.

Loewe, S., and Muischnek, H. (1926). Uber kombinationswirkunger: I. Mitteilung: Hiltsmittelder fragstellung. Naunyn-Schmiedebergs. Arch. Pharmacol. 114, 313–326.

Maness, S. C., McDonnell, D. P., and Gaido, K. W. (1998). Inhibition of androgen receptor-dependent transcriptional activity by DDT isomers and methoxychlor in HepG2 human hepatoma cells. Toxicol. Appl. Phamacol. 151, 135–142.[ISI][Medline]

Marshall, W. A., and Tanner, J. M. (1970). Variations in the pattern of pubertal changes in boys. Arch. Dis. Child 45, 13–23.[ISI][Medline]

McKenna, N. J., Lanz, R. B., and O’Malley, B. W. (1999). Nuclear receptor coregulators: Cellular and molecular biology. Endocr. Rev. 20, 321–344.[Abstract/Free Full Text]

Monosson, E., Kelce, W. R., Lambright, C., Ostby, J., and Gray, L. E., Jr. (1999). Peripubertal exposure to the antiandrogenic fungicide, vinclozolin, delays puberty, inhibits the development of androgen-dependent tissues, and alters androgen receptor function in the male rat. Toxicol. Ind. Health 15, 65–79.[ISI][Medline]

Pasqualini, J. R., and Kincl, F. A., Eds. (1985). Hormones and the Fetus. Pergamon Press, New York.

Quigley, C. A. (2002). Editorial: The postnatal gonadotropin and sex steroid surge: Insights from the androgen insensitivity syndrome. J. Clin. Endocrinol. Metab. 87, 24–28.[Abstract/Free Full Text]

SAS Institute (1999). SAS Campus Drive, Cary, NC 27513.

Seber, G. A. F., and Wild, C. J. (1989). Nonlinear Regression. Wiley, New York.

Tomlin, C. D. S., Ed. (1997). The Pesticide Manual, 11th ed. British Crop Protection Council, Farnham, Surrey, UK.

U.S. EPA (2001). Vinclozolin; notice of use cancellations. U.S. Environmental Protection Agency. Fed. Reg. 66(163), 44134–44136.

Warren, D. W., Huhtaniemi, I. T., Dufau, M. L., and Catt, K. J. (1987). Regulation of LH receptors and steroidogenesis in the fetal rat testis in vivo. Acta Endocrinol. (Copenh.) 115, 189–195.[ISI][Medline]

Weller, E., Long, N., Smith, A., Williams, P., Ravi, S., Gill, J., Henessey, R., Skornik, W., Brain, J., Kimmel, C., Kimmel, G., Holmes, L., and Ryan, L. (1999). Dose-rate effects of ethylene oxide exposure on developmental toxicity. Toxicol. Sci. 50, 259–270.[Abstract]

Wilson, E. M., and French, F. S. (1976). Binding properties of androgen receptors. Evidence for identical receptor in rat testis, epididymis, and prostate. J. Biol. Chem. 251, 5620–5629.[Abstract]

Wolf, C. J., LeBlanc, G. A., Ostby, J. S., and Gray, L. E., Jr. (2000). Characterization of the period of sensitivity of fetal male sexual development to vinclozolin. Toxicol. Sci. 55, 152–161.[Abstract/Free Full Text]

Wong, C., Kelce, W. R., Sar, M., and Wilson, E. M. (1995). Androgen receptor antagonist versus agonist activities of the fungicide vinclozolin relative to hydroxyflutamide. J. Biol. Chem. 270, 19998–20003.[Abstract/Free Full Text]





This Article
Abstract
FREE Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (3)
Disclaimer
Request Permissions
Google Scholar
Articles by Euling, S. Y.
Articles by Kimmel, C. A.
PubMed
PubMed Citation
Articles by Euling, S. Y.
Articles by Kimmel, C. A.