Development of an ER Action Indicator Mouse for the Study of ...

0013-7227/01/$03.00/0 Printed in U.S.A.

Endocrinology 142(11):4721– 4728 Copyright © 2001 by The Endocrine Society

Development of an ER Action Indicator Mouse for the Study of Estrogens, Selective ER Modulators (SERMs), and Xenobiotics SUSAN C. NAGEL, JENNIFER L. HAGELBARGER,

AND

DONALD P. MCDONNELL

Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, North Carolina 27710 We have developed a transgenic mouse that functions as a reporter of ER activity, termed ER action indicator (ERIN), by incorporating a transgene with an estrogen-responsive promoter (three copies of the vitellogenin estrogen response element with a minimal thymidine kinase promoter) linked to the reporter gene ␤-galactosidase. Evaluation of ER activity in female ERIN mice demonstrated estrogen-inducible expression of the reporter gene in the uterus, pituitary, and hypothalamus; established targets of estrogen action. Importantly, we also identified ER activity in a number of nonclassical estrogen target tissues, including kidney, liver, adrenal, and thyroid gland. ERIN provides a system to measure the same end point (transgene regulation) in different target tissues, permitting separation of the contributions of cell- and promoter-specific factors in determining ER pharmacology. In this regard we observed that on this specific promoter the

E

STROGENS FUNCTION BY binding to specific estrogen receptors (ERs) located within the nuclei of target cells. Upon binding ligand, the transcriptionally inactive ER undergoes an activating conformational change that facilitates the interaction of ER with specific DNA response elements within the regulatory regions of target genes (1). It was previously considered that ER ligands fell into two distinct classes: agonists and competitive antagonists. However, the discovery of selective ER modulators (SERMs), compounds whose agonist/antagonist activity can differ among cells, prompted a reevaluation of the pharmacological classification of ER ligands. This ongoing process has revealed that the classical models of ER action do not adequately describe the activity of the known ER ligands and that a more comprehensive analysis of the cellular factors that influence the agonist/antagonist activity of ER ligands is warranted. One of the most important advances in our understanding of ER pharmacology has been the discovery that different ER ligands induce different conformational changes in the receptor, and that this is a key determinant of the interaction between ER and specific comodulator proteins (1–3). Thus, the ability of ligand-bound ER to regulate target gene transcription is determined by the cell- and tissue-specific expression of both coactivator and corepressor proteins that impact the ER signaling pathway (for reviews, see Refs. 4 and Abbreviations: BPA, Bisphenol A; DES, diethylstilbestrol; E2, estradiol; EIA, enzyme immunoassay; ER, estrogen receptor; ERE, estrogen response element; ERIN, ER action indicator; ␤-gal, ␤-galactosidase; ICI, ICI 182,780; SERM, selective ER modulator; SV40, simian virus 40; TK, thymidine kinase.

pituitary gland was 25-fold more sensitive than the uterus to the estrogen diethylstilbestrol, implying the existence of cellspecific factors that influence ligand sensitivity. Our studies also identified considerable difference in the efficacy and potency of ER ligands in the uterus when ER transcriptional activity was assayed vs. uterine weight gain. Specifically, we observed that the environmental estrogen bisphenol A was a potent agonist in stimulating ER transcriptional activity, whereas it exhibited little uterotropic activity. In contrast to bisphenol A, tamoxifen significantly increased uterine weight, but minimally induced ER reporter activity in this tissue. Given the results of these studies, we believe that ERIN will be a useful model to evaluate ER ligand pharmacology and will assist in defining the cellular and molecular mechanisms that determine agonist and antagonist activity. (Endocrinology 142: 4721– 4728, 2001)

5). Adding to this complexity was the discovery of a second ER, ER␤ (6). It is likely, therefore, that tissue-specific expression of comodulators and ER subtypes will together play an important role in determining the tissue-specific actions of ER ligands. Clearly, identification of the ER comodulators in different target tissues will enable a more complete understanding of the pharmacology of the receptor and is likely to aid in the development of the next generation of SERMs. In most cases, however, the specific cell types, particularly outside of the reproductive tract, that permit tissues to respond to estrogen have not yet been identified. To address globally the issues of tissue specificity and sensitivity to both endogenous and xenobiotic estrogen, we have developed the ER action indicator (ERIN) transgenic mouse that functions as a reporter of ER activity. ERIN provides a model system that can be used to identify tissues and cells that contain functionally active ER, both ER␣ and ER␤, and to define their ability to respond to different ER ligands. This model system integrates the upstream requirements in ER action, including the receptor, ligand, and accessory comodulators. Activation of ER results in expression of the enzyme ␤-galactosidase (␤-gal), which allows for enzymatic amplification of the signal and histological localization of its activity. This model system can be used to identify novel estrogen target tissues and to define the elements that influence and regulate ligand specificity and sensitivity in the mouse. Importantly, by assessing the regulation of the same gene across many different tissues, the contribution of cell context (not target gene promoter) to the activity of a ligand can be evaluated. In this study we validated this model

4721

4722

Endocrinology, November 2001, 142(11):4721– 4728

system by identifying ER activity in a number of classical and nonclassical estrogen target tissues, and we used the system to study the tissue-specific activity of a select estrogen (diethylstilbestrol, DES), SERM (tamoxifen), and xenoestrogen (bisphenol A). Materials and Methods Biochemicals General laboratory chemicals, 17␤-E2, DES, and charcoal-stripped FCS were purchased from Sigma (St. Louis, MO), and ICI 182,780 (ICI) was obtained from Tocris Cookson, Inc. (Ballwin, MO). Chlorophenol red-␤-d-galactopyranoside, restriction enzymes, T4 DNA ligase, tissue culture media and supplements, lipofectin, and DH5␣ cells were purchased from Life Technologies, Inc. (Grand Island, NY), and luciferin was obtained from Promega Corp. (Madison, WI).

Plasmids The thymidine kinase (TK)-luciferase reporter was a gift from Ligand Pharmaceuticals, Inc. (San Diego, CA). The 1⫻ estrogen response element (ERE)-TK-luciferase and 3⫻ERE-TATA-luciferase reporters were generated in the McDonnell laboratory by J. D. Norris and M. Huacani-Hamilton. The actin-luciferase, pW1Xb-simian virus 40 (SV40), and pXT-␤-gal3 plasmids were gifts from E. Linney (Duke University, Durham, NC).

Generation of transgene For subcloning, all DNA fragments were separated on agarose gels, excised, and purified using the Gene Clean II kit (no. 1001– 400, BIO 101, Vista, CA). lacZ was excised from pXT-␤-gal3 by BglII digestion and ligated into pW1Xb-SV40, which contains SV40 polyadenylase and intron sequences, digested with BglII and BamHI to generate pW1-lacZ. 3⫻ERE (three copies of the vitellogenin ERE, GATCCCGCAGGTCACAGTGACCTG) was excised from 3⫻ERE-tata-luciferase by BglII digestion and ligated into TK-luciferase digested with BamHI. 3⫻ERE-TK was excised from 3⫻ERE-TK-luciferase with BglII and HindIII and ligated into pW1-lacZ digested with BglII and HindIII. The final transgene, 3⫻ERE-TK-lacZ-SV40 polyadenylase and intron, was excised with NotI and SfiI for microinjections.

Cotransfection assays NIH-3T3 cells were maintained in DMEM. MCF-7 and HepG2 cells were maintained in MEM. All media were supplemented with 8% FCS (HyClone Laboratories, Inc., Logan, UT), 1 mm sodium pyruvate, and 0.1 mm nonessential amino acids. Cells were plated in 24-well plates (coated with gelatin for HepG2 cells) 24 – 48 h before transfection. DNA was introduced into cells by transfection using lipofectin. Triplicate transfections were performed using 3 ␮g total DNA. For standard transfections, 300 ng actin-luciferase (normalization vector), 2200 ng reporter, and 500 ng ER␣ expression vector (pRST7-hER) (7) were used for each triplicate. Cells were transfected for 5 h, at which time medium was removed and induced with hormone diluted in phenol red-free medium supplemented with 8% charcoal-stripped FCS (Sigma). Incubation with hormone continued for 24 or 48 h, after which cells were lysed and assayed for luciferase and ␤-gal activities.

Generation and identification of transgenics Transgenic mice were produced by the Duke University Transgenic Mouse Facility by microinjection of male pronuclei of zygotes produced from hybrid C57BL/6-SJL mice. Genomic DNA was isolated (DNeasy Tissue Kit, no. 69506, QIAGEN, Valencia, CA) from tail clips from 10-d-old pups, and the transgene was detected by PCR amplification of a 200-bp fragment using transgene-specific primers (CCGACTGCATCTGCGTGT and TAATACGACTCACTATAGGG) and control primers that amplified a 500-bp fragment of the TSH gene (TCCTCAAAGATGCTCATTAG and GTAACTCACTCATGCAAAGT).

Nagel et al. • ER Indicator Mice

Housing and mating All housing and procedures were approved by the Duke University animal care and use committee and were performed in accordance with federal, state, and local rules for the humane treatment of laboratory animals. Mice were housed in polystyrene cages with a 14-h light, 10-h dark cycle, with food (Purina Mouse Chow 5001, or for pregnant and nursing mice 5015, Ralston-Purina, St. Louis, MO) and water ad libitum. Transgenic founders were bred with C57BL/6 mice (The Jackson Laboratory, Bar Harbor, ME), and transgenic offspring were identified by PCR. Pups were weaned at 21 d and housed three to five per cage. All experimental animals for these studies were generated from the Founder 6 line. Founder 6 (C57BL/6-SJL hybrid) was bred with C57BL/6 mice, F1 offspring were crossed, and F2 transgenics were mated with C57BL/6 mice to generate experimental animals.

Ovariectomies and dosing Adult mice were ovariectomized after anesthetization with a ketamine and medetomidine cocktail. Ovariectomies were performed on females by an incision directly over the ovary through the skin and body wall. Ovaries were severed between the oviduct and uterus with cauterizing scissors. Incisions were closed with surgical stainless steel clips. Domitor was administered to reverse anesthesia. Mice were kept under heat lamps until sternal recumbency was regained. One week later, mice were given a single (unless otherwise noted) sc injection of test compound in tocopherol-stripped corn oil (ICN Biomedicals, Inc., Aurora, OH) as indicated in the text. Twenty to 24 h later (unless otherwise noted) mice were killed with CO2 asphyxiation, and organs were immediately removed, weighed, and either frozen on dry ice or placed in 1⫻ lysis buffer for immediate homogenization. We chose to use the estrogen DES for these studies to assure continuity with future studies in which compounds will be administered orally. DES is an orally active estrogen, whereas E2 is not.

␤-Gal enzyme immunoassay (EIA) Immulon 4 flat-bottom microtiter plates (no. 3855, Dynex Technologies, Chantilly, VA) were used to capture 1.5 ␮g/well rabbit anti Escherichia coli ␤-gal antibody (AB1211, Chemicon International, Inc., Temecula, CA) in PBS with 0.05% NaN3 overnight at 4 C or for 2 h at 37 C. Plates were washed with water twice. Nonspecific binding was blocked with 200 ␮l/well blocking buffer (borate-buffered saline, 0.05% Tween, 0.25% BSA, 1 mm EDTA, and 0.05% NaN3) for 1 h or more at room temperature, and the plates were washed twice with water immediately before assay. Animals were killed by CO2 asphyxiation, tissues were removed immediately, approximately 50 mg tissue were placed in 500 ␮l 1⫻ lysis buffer [25 mm Tris-phosphate (pH 7.8), 2 mm dithiothreitol, 2 mm trans-1-2-diaminocyclohexane-N,N,N,N⬘tetraacetic acid, 10% glycerol, and 0.5% Triton X-100], minced with scissors, and kept on ice. All tissues were homogenized (PT1200, Brinkmann Instruments, Inc., Westbury, NY) for 10 sec at maximum speed. Insoluble material was pelleted by centrifugation at 4 C at 15,000 ⫻ g for 10 min. ␤-Gal capture was performed by incubating 140 ␮l homogenate overnight at 4 C in antibody-coated plates. For ␤-gal assays, plates were washed four times by submersion in water, then incubated with 200 ␮l 0.67 mg/ml chlorophenol red-␤-d-galactopyranoside, and the absorbance at 575 nm was measured at intervals (10 min and 3, 6, and 18 h). Wild-type mice were routinely included in the EIA analysis, and it was found that nontransgenic mice do not express any proteins that crossreact with the anti-␤-gal antibody. Thus, absorbance readings from wild-type mice were the same as background. For protein assays, tissue homogenates were diluted and performed in microplates using the Coomassie Plus Protein Assay Reagent Kit (no. 23236, Pierce Chemical Co., Rockford, IL).

Statistics Data were log-transformed to reduce the variance heterogeneity, and ANOVA was conducted on log-transformed data using StatView (SAS Institute, Inc., Cary, NC). Planned comparisons were conducted using Fisher’s planned least significant difference test when the overall ANOVA was statistically significant.

Nagel et al. • ER Indicator Mice

Endocrinology, November 2001, 142(11):4721– 4728 4723

Results Selection of the ER reporter for introduction into mice

Generation and identification of transgenic mice containing an estrogen-responsive reporter

To select an estrogen-responsive promoter for transgene production, we created and characterized seven different reporters containing one to seven copies of the vitellogenin ERE (1⫻ 7⫻ERE) linked to a minimal TK promoter. Interestingly, the presence of more than three EREs did not increase, and in some cases decreased, reporter activity when assayed in transiently transfected MCF-7 human breast cancer cells (Fig. 1A). In the course of the analysis we observed that all of the ERE-TK reporters had a much higher absolute level of induction than the 3⫻ERE-tata promoter (data not shown). After systematic evaluation of these estrogenresponsive promoters in MCF-7 cells, HepG2 human hepatocarcinoma cells, and NIH-3T3 mouse fibroblast cells with transfected ER (data not shown), we selected the 3⫻ERETK-lacZ for transgenic production (Fig. 1B). This promoter has the highest absolute level of estrogen-induced expression in cell-based transfection assays, a selection criteria that had been used successfully in the creation of the retinoic acid receptor response element-lacZ transgenic mouse (8).

DNA microinjections and creation of potential founder mice were performed by the Duke University Transgenic Mouse Facility. Initially, 7 founders that had incorporated the transgene were identified by PCR amplification and Southern blot analysis. Founders (C57BL/6-SJL hybrids) were subsequently bred with C57BL/6 mice. Transgenic offspring were analyzed for expression of the reporter transgene, among which 2 founder lines (F6 and F7) expressed the transgene broadly in target tissues. However, F7 was extremely subfertile; therefore, 2 more rounds of microinjections were performed, and 10 additional founders were identified. Four of these founder lines showed broad transgene expression, although their characterization are not described in this report. For the following studies, Founder 6 was bred with C57BL/6 mice, F1 offspring were crossed, and F2 transgenics were mated with C57BL/6 mice to generate experimental animals.

FIG. 1. A, 3XERE-Tk-lacZ showed the highest absolute level of estrogen-induced activity in cultured cells. MCF-7 cells were seeded in 24-well plates in phenol red-free medium for 2 d, transfected for 5 h, and induced with hormone for 18 h. Graphed is one assay performed in triplicate; it is representative of three separate assays. Error bars represent the SEM. B, Schematic of the ERIN transgene and the PCR primers used to detect transgenic animals. The final transgene (4.4kb) consists of three copies of the vitellogenin ERE (3⫻ERE) and a minimal TK promoter linked to the lacZ gene and including SV40 polyadenylation sequences and an intron from the small t antigen. The locations of primers used for PCR detection and the sizes of amplified fragments are shown below the diagram.

Characterization of the 3XERE-TK-lacZ-containing mice

Previous studies in animals and in vitro have shown that it is difficult to select a single time point for estrogen exposure that is suitable for all target organs. Thus, a preliminary time-course study was performed. Adult female ERIN mice were ovariectomized, and 1 wk later they were given a single sc injection of corn oil or 5 ␮g/kg DES (a synthetic estrogen) 20, 46, or 70 h before collection. We developed an EIA specific for the bacterial ␤-gal protein (coded for by the lacZ gene used for transgene production) that was required to avoid the confounding influence of endogenous enzymes that also hydrolyze the ␤-gal substrate. Using this assay, tissues were collected, homogenized, and assayed for ␤-gal activity. Robust estrogen-induced expression of the transgene (ERIN activity) was detected in the uterus, pituitary, hypothalamus, and kidney (Fig. 2). The pituitary and uterus showed significant induction of ERIN activity at all three time points tested, whereas the ␤-gal signal was significantly reduced in the kidney and hypothalamus after 46 and 20 h, respectively. The optimal expression of ERIN activity in the four organs occurred at 20 h postinjection (Fig. 2), and consequently, this time point was used for all of the experiments presented in this study. To verify that the increased ERIN activity was due to estrogen induction and not to an estrogen-stimulated increase in cell number, ␤-gal activity was normalized to both protein and DNA content in a representative experiment. Interestingly, the fold induction in the pituitary was not changed when normalized to DNA content whereas the fold induction in the uterus was slightly increased. As the two types of normalization yielded similar results, the data were normalized to protein content for all of the experiments in this study. ER activity was detected in a broad range of tissues

It is now well established that estrogen can manifest biological activity in tissues not generally considered to be involved in reproduction. The identification of a second ER together with clinical data that have defined a role for es-

4724

Endocrinology, November 2001, 142(11):4721– 4728

Nagel et al. • ER Indicator Mice

FIG. 2. Demonstration of estrogeninducible transgene activity. Female ERIN mice were ovariectomized at 10 wk. One week later a single sc injection of corn oil control (䡺) was given 20 h before collection or a singe dose of 5 ␮g/kg DES (f) was given 70, 46, or 20 h before collection. Tissues were harvested and homogenized for EIA (see Materials and Methods). ␤-Gal activity was normalized to protein content. n ⱖ 3/treatment group. Error bars represent the SEM. *, P ⬍ 0.05 relative to corn oil control.

FIG. 3. Identification of ER activity in a broad range of tissues. Female ERIN mice were ovariectomized at 16 wk. One week later a single sc injection of corn oil control (䡺) or 5 ␮g/kg DES (f) was given 22 h before collection. The graph represents the average of three separate experiments. Data are presented as the fold induction over the oil control. For each experiment, DES was divided by the average oil control and then averaged across experiments (control, n ⫽ 11; DES, n ⫽ 9). Error bars represent the SEM. *, P ⬍ 0.05; for tissues that did not reach this level of significance the P value is given above the bars. Mamm, Mammary gland tissue; hypoth., hypothalamus.

trogen in a broad range of tissues suggest that additional targets of estrogen remain to be identified. We believe that one of the major uses of ERIN mice will be the identification and cataloging of tissues in which estrogen can manifest activity. Thus, the first comprehensive study performed with the ERIN mouse was designed to identify tissues containing estrogen-induced ERIN activity. Transgenic females (ovariectomized) were analyzed for expression of the reporter

transgene after exposure to corn oil control or DES (5 ␮g/kg BW). A summary of expression from three independent experiments is shown in Fig. 3. Evaluation of ERIN activity in DES-treated female mice revealed that tissues could be classified into five groups based on ER activity. ERIN activity was strongly induced (⬃4-fold) in the pituitary, uterus, and kidney; however, the absolute levels of expression varied greatly (pituitary ⬎ uterus ⬎⬎ kidney; compare scale in Fig. 2). A moderate level of estrogen-induced ER activity (⬃2-fold) with a high level of basal ER activity was detected in the liver, hypothalamus (and other areas of the brain), and adrenal. A moderate level of estrogen-induced ER activity (⬃2-fold) with a low level of basal ER activity was detected in the thyroid, fat, mammary gland, and muscle. Several tissues displayed a significant basal level of ERIN activity that was not enhanced further with estrogen treatment, including the heart, thymus, and intestine (data not shown). Finally, there was no detectable activity in the spleen or lung. The apparent ligand independent ERIN activity observed in some tissues can be inhibited by an antiestrogen

We were concerned that the ERIN activity observed in some tissues was not ER dependent. To confirm that the basal reporter activity was indeed due to ER activation, we treated ovariectomized mice with the pure antiestrogen ICI and reevaluated reporter expression in the ERIN mice. These studies revealed that ICI (25 mg/kg) inhibited basal ER activity in the intestine, fat, liver, thyroid, and uterus by 50% or more (P ⬍ 0.05; Fig. 4). In addition, there was a trend toward inhibition of the basal level of activity present in the heart, kidney, adrenal, pituitary, and hypothalamus, but

Nagel et al. • ER Indicator Mice

FIG. 4. ER-dependent basal activity demonstrated by anti-estrogen treatment. Female ERIN mice were ovariectomized at 7 wk. One week later a single sc injection of corn oil control (䡺) or 25 mg/kg ICI (f) was given 22 h before collection. The graph represents the average of three separate experiments. Data are normalized to oil control. For each experiment, ICI was divided by the average oil control and then averaged across experiments (control, n ⫽ 8; ICI, n ⫽ 6). Error bars represent the SEM. *, P ⬍ 0.05.

basal ER activity in thymus and mammary gland was not affected (data not shown). As these experiments were conducted in mice lacking the primary source of estrogen (ovary), the variable level of basal ER activity may be due to local aromatization of adrenal androgens, as the adrenal is a source of estrogen precursors that can be used as substrates to synthesize estrogen in other tissues, or to phytoestrogens in the feed. This suggests that the ERIN mouse may be sensitive enough to detect very low levels of estrogen exposure. Differential tissue sensitivity to estrogen can be demonstrated in ERIN mice

Differential tissue and end-point sensitivities to a variety of ER ligands have been previously described. The ERIN mouse, in which the same simple promoter and reporter is examined in different tissues, afforded us the chance to examine whether differential hormone sensitivity was an inherent property of cells or was a consequence of the target gene promoter. For this study we administered doses of DES ranging from 5 to 5000 ng/kg to ovariectomized females and measured ERIN reporter activity 20 h later. Interestingly, DES was a more potent agonist of ER activity in the pituitary than in the uterus and kidney (Fig. 5). The EC50 of DES stimulation of ER activation in the pituitary (20 ng/kg) was 25-fold lower than that required for ER activation in the uterus and kidney (EC50 ⬃500 ng/kg). It will be important in the future to microdissect these tissues and determine the relative sensitivities of different cell types to estrogen exposure. However, these data suggest that cellular context is a principle determinant of differential estrogen responsiveness in different tissues. Uterine wet weight assays do not faithfully assess the estrogenic activity of ER ligands

With the development of the ERIN mouse and, more specifically, a model system to track estrogen action without the

Endocrinology, November 2001, 142(11):4721– 4728 4725

FIG. 5. Differential tissue sensitivity to estrogen revealed by DES dose response in ERIN mice. Female ERIN mice were ovariectomized at 7 wk. One week later a single sc injection of corn oil control or DES was given 22 h before collection. The graph represents the average of three separate experiments. For each experiment, treatment group was divided by the average oil control and then averaged across experiments. Data are presented as a percentage of maximum induction for each organ (mean of normalized treatment ⫼ mean of normalized DES at maximum effective dose ⫻ 100; control, n ⫽ 11; 0.005 ␮g/kg DES, n ⫽ 7; 0.158 ␮g/kg DES, n ⫽ 8; 5 ␮g/kg DES, n ⫽ 5).

influence of target gene promoter context, we wanted to evaluate the ability of this model system to study xenoestrogens. One of the most widely used assays to measure the estrogenic activity of a compound is to assess its ability to increase uterine wet weight in immature or ovariectomized mice (or rats). With the goal of evaluating this assay for the purpose of studying xenobiotic estrogen, we compared the ability of the environmental estrogen bisphenol A (BPA) and the SERM tamoxifen to increase uterine wet weight and stimulate ER transcriptional activity simultaneously in ERIN mice. BPA has previously been reported to have very weak uterotropic activity (9 –13), lower than would be predicted from its affinity for ER␣, the dominant ER expressed in the uterus. However, when BPA was administered to ERIN mice, it induced ER transcriptional activity in the uterus at 25 mg/kg (P ⬍ 0.01) and 0.8 mg/kg (P ⬍ 0.01), and there was a trend toward stimulation at only 25 ␮g/kg (P ⫽ 0.052). At 25 mg/kg (the maximum dose given), BPA stimulated ER transcriptional activity to 60% of the maximum activity stimulated by 5 ␮g/kg DES. Contrary to the strong response in ER transcriptional activity, when uterine wet weight was measured in the same ERIN females, BPA showed very weak uterotropic activity, stimulating uterine weight gain to only 18% of that induced by DES (Fig. 6). In fact, BPA has previously been demonstrated to be a partial agonist for this end point, stimulating uterine weight gain to only 20% of that induced by E2 (10). To further characterize the selective ER-modulated activities in the uterus, we extended the analysis to the SERM tamoxifen. Tamoxifen has been characterized as having more ER agonist activity in the mouse than in humans and, in fact, is a very strong agonist for stimulation of uterine weight gain. In ERIN mice exposed to either DES or tamoxifen for

4726

Endocrinology, November 2001, 142(11):4721– 4728

FIG. 6. BPA stimulation of ER activity in the uterus is more sensitive than stimulation of uterine weight gain in ERIN mice. Female ERIN mice were ovariectomized at 7 wk. One week later a single sc injection of corn oil control or DES was given 22 h before collection. The graph represents the average of three separate experiments. For each experiment, treatment group was divided by the average oil control and then averaged across experiments. Data are presented as a percentage of DES maximum induction of ER activity or uterine weight gain (mean normalized treatment ⫼ mean normalized 5 ␮g/kg DES ⫻ 100; n ⫽ 11 control; n ⫽ 7, 6, and 6 for 25, 791, and 25,000 ␮g/kg BPA, respectively). Error bars represent the SEM. *, P ⬍ 0.01 (except where noted) and represents the difference between treatment and oil control (0%; not shown).

20 h, tamoxifen (25 mg/kg) stimulated uterine weight gain to 70% of that induced by DES (Fig. 7). Surprisingly, however, tamoxifen was found to stimulate very little ERIN activity in the same mice. It appears that BPA and tamoxifen stimulate estrogenic responses selectively in the uterus (Fig. 7). BPA is a relatively potent agonist for ER transcriptional activation in the uterus, but is a weak agonist in stimulating uterine wet weight gain. In contrast to BPA, tamoxifen is a weak agonist for ER activation, but a potent agonist for uterine wet weight gain. Thus, although BPA would probably be scored as a compound with insignificant estrogenic activity based on its performance in the uterotropic assay, it is, in fact, a relatively potent activator of ER transcription activity in this tissue. These data have important implications for environmental estrogen assessment and for the evaluation of SERMs, where lack of uterotropic activity has been taken to indicate inactivity for all end points. Discussion

The aim of the ERIN project was to develop a transgenic mouse model that could be used to identify target tissues that contain ER that are active and to aid in the evaluation of the contributions of the cellular environment, rather than promoter context, in determining the pharmacological activities of estrogen. We have validated this model system by demonstrating estrogen-inducible activity in several estrogen target tissues and have identified ER activity in several nonclassical target tissues, the significance of which seems likely to be important in understanding estrogen and SERM action.

Nagel et al. • ER Indicator Mice

FIG. 7. Estrogen-stimulated uterine weight gain does not reflect ER transcriptional activity in the uterus. Stimulation of uterine weight gain vs. ER transcriptional activity by DES (5 ␮g/kg), BPA (25 mg/kg), and tamoxifen (TAM; 25 mg/kg). Data are presented as a percentage of DES maximum induction of uterine weight gain or ERIN activity (mean normalized treatment ⫼ mean normalized 5 ␮g/kg DES ⫻ 100). The graph represents the average of three separate experiments. For each experiment, treatment group ⫼ average oil control and then averaged across experiments. For BPA experiments, n ⫽ 11 control, n ⫽ 5 DES, n ⫽ 6 for BPA; for TAM experiments, n ⫽ 12 control, n ⫽ 11 DES, and n ⫽ 5 TAM. Error bars represent the SEM. *, P ⬍ 0.01, Difference between treatment and oil control (0%; not shown).

Although there are other potentially important ER-mediated actions that function through pathways besides direct binding to ERE, e.g. interactions with AP-1 (14, 15), we chose to focus only on those that function through the classical ERE pathway. It is important to note that ERIN is a model system that will complement, not replace, established models of ER action that have been validated for various aspects of ER signaling. We have used the ERIN mouse to examine the sensitivity of the same end point in different tissues to DES administration. In ovariectomized female mice, we found the pituitary to have 25-fold greater sensitivity to estrogen than the uterus in the same mouse at the identical time. These studies revealed that cell context is a major determinant of end-point responsiveness to estrogen. Definition of the molecular mechanisms that enable the pituitary to respond to low concentrations of estrogen is a primary goal of our continued research, although several potential mechanisms for differential responsiveness are already apparent. For instance, in transient transfection assays, both the PR A isoform and ER␤ have been shown to transrepress the transcriptional activity of ER␣ (16, 17), and this may contribute to the decreased sensitivity of the uterus relative to the pituitary that we observed in ERIN mice. In addition, thyroid hormone-activated TR has been shown to potentiate estrogen activity in some tissues (18, 19), whereas in other tissues, TR␣1 inhibits ER activity (20). Thus, isolation of the specific pituitary cells that display a heightened response to estrogen and an evaluation of the relative expression levels of the receptors will be important. Finally, it is likely that tissue-specific expression of ER comodulatory proteins will significantly impact ER signaling (21, 22). The ERIN mouse will clearly be a useful model that can be used to define the mechanisms in the mouse responsible for the differential tissue sensitivity of the

Nagel et al. • ER Indicator Mice

pituitary and uterus and other targets not considered in our study. In addition to well characterized estrogen target tissues, we observed ER transcriptional activity in a number of nonclassical tissues, including liver, kidney, thyroid, adipose tissue, and adrenal glands. The level of transgene expression varied greatly between tissues and may reflect in part the expression level of ER and/or the number of ER-containing cells in different tissues. For example, the pituitary, uterus, and hypothalamus express ER in a large number of cells and, not surprisingly, have a much higher level of estrogeninduced activity in ERIN mice. The kidney has been shown to express both ER␣ and ER␤ (23), although the number of cells expressing ER is lower than that in the more classic estrogen target tissues. Importantly, however, all of the tissues we found to have ER activity have previously been shown to express ER, albeit often at low levels. ERIN can be used in future studies to specifically isolate the ER-containing cells and determine the role of ER in these tissues. An important new aspect in estrogen physiology is the potential ER-selective actions of xenoestrogens, a research area that has been complicated by contradictory reports of estrogenicity of several of these xenobiotics. For example, BPA, a widely used synthetic compound with demonstrated estrogenic activity, is used as a component of epoxy resins found in the lining of metal food cans (24), as a monomer in the manufacture of polycarbonate plastics (25), and in dental sealants (26). BPA has a relatively weak affinity for ER (27, 28) and exhibits low activity in the classic uterine weight gain assay (9 –13). However, this xenoestrogen can display significant estrogenic activity, as evidenced 1) in mice where prenatal exposure results in increased prostate weight in adult males and earlier puberty in female mice (27, 29 –31), and 2) in Fischer 344 rats where its administration leads to the development of hyperprolactinemia (32). The differential, and seemingly paradoxical, actions of BPA may be due to selective ER modulatory activities of this ligand, which result in species-, life stage-, and tissue-selective ER activity (33, 34). In addition, BPA is a full ER agonist in many cellbased assays and a partial agonist in others (25, 34), and Diel et al. (35) recently reported that BPA can induce estrogenresponsive endogenous genes in the uterus at lower concentrations than stimulation of uterine weight gain. It is likely that the ER conformation induced by BPA results in recruitment of cell- and tissue-specific factors responsible for the differential responses to this ligand (36). By definition, therefore, BPA can be classified as a SERM. Our results have important implications for the screening of environmental estrogen. Currently, there are several in vitro assays that adequately measure the ability of xenoestrogens to interact with and/or activate ER, including binding, transcriptional activation, and proliferation assays (37). Clearly, however, species- and tissue-selective ER-mediated responses will occur in the animal. In this study the potency and efficacy of BPA in stimulating uterotropic activity were very low; however, the potency and efficacy of BPA in stimulating ER transcriptional activity in the uterus are equal to or greater than its potency in vitro. Specifically, BPA stimulated ER activity in the uterus at 0.8 mg/kg (efficacy, 40% of DES maximum; P ⬍ 0.05) and tended to stimulate at 25

Endocrinology, November 2001, 142(11):4721– 4728 4727

␮g/kg (20% of DES maximum; P ⫽ 0.052). This is the lowest reported estrogenic response of BPA in adult animals. The estrogen-stimulated uterine wet weight assay has been the gold standard for determining estrogenicity in vivo. However, based on our results and others (12), this assay appears to be a relatively insensitive end point, and is likely to underestimate the number of compounds with estrogenic potential in vivo. For example, based on results of uterotropic assays in mice, BPA would be considered an extremely weak partial ER agonist, whereas SERMs such as tamoxifen would be rather potent agonists; conclusions that are not supported by the results of this study. Although the Founder 6 used in these studies showed expression in all tissues reported to be estrogen responsive, it showed relatively limited expression in two organs, the mammary gland and ovary. However, we have recently identified a new founder that in preliminary experiments showed strong expression of the transgene in both the mammary gland and ovary and broadly in other tissues. Due to the insertion site of the transgene into areas of chromatin that are variably inactive in different tissues, it is possible that there will not be a single founder line that possesses the ability to respond strongly to ER ligands in all tissues. However, we have demonstrated estrogen-dependent ER activity in a number of classic estrogen target tissues in ERIN mice in addition to tissues with less characterized estrogenic responses. ERIN will provide an excellent in vivo model system 1) to characterize the pharmacology of ER␣- and ER␤-selective ligands, 2) to detect local sources of estrogen production, 3) to identify cell types within nonclassical estrogen target tissues containing active ER, and 4) to aid in the characterization of environmental estrogen, which, like BPA, will undoubtedly manifest tissue-specific estrogenic responses. We believe that ERIN is an important model system and anticipate that it will gain widespread use in the field to study different aspects of estrogen signaling. Acknowledgments We are grateful to E. Linney for his suggestions for this project and for contribution of plasmids. We thank members of the McDonnell laboratory, specifically M. Jansen, C.-Y. Chang, and G. Sathya, for critical review of the manuscript; J. D. Norris and M. Huacani-Hamilton for plasmids; H. Cui and V. Clack for technical assistance; T. Martelon for technical and administrative assistance; and M. Huacani-Hamilton, A. Wijayaratne-Fernando, J. O’Campo, and F. Folger for assistance with animal collections. We are also grateful to W. V. Welshons, F. S. vom Saal, and D. Sheehan for suggestions for the manuscript. Received May 22, 2001. Accepted July 16, 2001. Address all correspondence and requests for reprints to: Donald P. McDonnell, Ph.D., Box 3813, Duke University Medical Center, Durham, North Carolina 27710. E-mail: [email protected]. This work was supported by NIH Grants DK-07012-23 (to S.C.N.) and DK-48807 (to D.P.M.).

References 1. Beekman JM, Allan GF, Tsai SY, Tsai M-J, O’Malley BW 1993 Transcriptional activation by the estrogen receptor requires a conformational change in the ligand binding domain. Mol Endocrinol 7:1266 –1274 2. Paige LA, Christensen DJ, Gron H, Norris JD, Gottlin EB, Padilla KM, Chang C-Y, Ballas LM, Hamilton PT, McDonnell DP 1999 Estrogen receptor (ER) modulators each induce distinct conformational changes in ER␣ and ER␤. Proc Natl Acad Sci USA 96:3999 – 4004 3. McDonnell DP, Clemm DL, Hermann T, Goldman ME, Pike JW 1995 Anal-

4728

4. 5. 6. 7.

8. 9. 10.

11. 12. 13. 14.

15. 16. 17.

18. 19. 20.

Endocrinology, November 2001, 142(11):4721– 4728

ysis of estrogen receptor function in vitro reveals three distinct classes of antiestrogens. Mol Endocrinol 9:659 – 668 McKenna NJ, Lanz RB, O’Malley BW 1999 Nuclear receptor coregulators: Cellular and molecular biology. Endocr Rev 20:321–344 Robyr D, Wolffe AP, Wahli W 2000 Nuclear hormone receptor coregulators in action: diversity for shared tasks. Mol Endocrinol 14:329 –347 Kuiper GGJM, Enmark E, Pelto-Huikko M, Nilsson S, Gustafsson J-Å 1996 Cloning of a novel estrogen receptor expressed in rat prostate and ovary. Proc Natl Acad Sci USA 93:5925–5930 Tzukerman MT, Ety A, Santiso-Mere D, Danielian P, Parker MG, Stein RB, Pike JW, McDonnell DP 1994 Human estrogen receptor transactivational capacity is determined by both cellular and promoter context and mediated by two functionally distinct intramolecular regions. Mol Endocrinol 8:21–30 Balkan W, Colbert M, Bock C, Linney E 1992 Transgenic indicator mice for studying activated retinoic acid receptors during development. Proc Natl Acad Sci USA 89:3347–3351 Yamasaki K, Sawaki M, Takatsuki M 2000 Immature rat uterotrophic assay of bisphenol A. Environ Health Perspect 108:1147–1150 Papaconstantinou AD, Umbreit TH, Fisher BR, Goering PL, Lappas NT, Brown KM 2000 Bisphenol A-induced increase in uterine weight and alterations in uterine morphology in ovariectomized B6C3F1 mice: role of the estrogen receptor. Toxicol Sci 56:332–339 Tinwell H, Joiner R, Pate I, Soames A, Foster J, Ashby J 2000 Uterotrophic activity of bisphenol A in the immature mouse. Regul Toxicol Pharmacol 32:118 –126 Markey CM, Michaelson CL, Veson EC, Sonnenschein C, Soto AM 2001 The mouse uterotrophic assay: a reevaluation of its validity in assessing the estrogenicity of bisphenol A. Environ Health Perspect 109:55– 60 Ashby J, Tinwell H 1998 Uterotrophic activity of bisphenol A in the immature rat. Environ Health Perspect 106:719 –720 Webb P, Nguyen P, Valentine C, Lopez GN, Kwok GR, McInerney E, Katzenellenbogen BS, Enmark E, Gustafsson J-Å, Nilsson S, Kushner PJ 1999 The estrogen receptor enhances AP-1 activity by two distinct mechanisms with different requirements for receptor transactivation functions. Mol Endocrinol 13:1672–1685 Paech K, Webb P, Kuiper GG, Nilsson S, Gustafsson J-Å, Kushner PJ, Scanlan TS 1997 Differential ligand activation of estrogen receptors ER␣ and ER␤ at AP1 sites. Science 277:1508 –1510 Giangrande PH, McDonnell DP 1999 The A and B isoforms of the human progesterone receptor: two functionally different transcription factors encoded by a single gene. Recent Prog Horm Res 54:291–313 Hall JM, McDonnell DP 1999 The estrogen receptor ␤-isoform (ER␤) of the human estrogen receptor modulates ER␣ transcriptional activity and is a key regulator of the cellular response to estrogens and antiestrogens. Endocrinology 140:5566 –5578 Shao ZM, Sheikh MS, Rishi AK, Dawson MI, Li XS, Wilber JF, Feng P, Fontana JA 1995 Thyroid hormone enhancement of estradiol stimulation of breast carcinoma proliferation. Exp Cell Res 218:1– 8 Rabelo EM, Tata JR 1993 Thyroid hormone potentiates estrogen activation of vitellogenin genes and autoinduction of estrogen receptor in adult Xenopus hepatocytes. Mol Cell Endocrinol 96:37– 44 Zhu YS, Yen PM, Chin WW, Pfaff DW 1996 Estrogen and thyroid hormone interaction on regulation of gene expression. Proc Natl Acad Sci USA 93: 12587–12592

Nagel et al. • ER Indicator Mice

21. Xu J, Qiu Y, DeMayo FJ, Tsai SY, Tsai M-J, O’Malley BW 1998 Partial hormone resistance in mice with disruption of the steroid receptor coactivator-1 (SRC-1) gene. Science 279:1922–1925 22. Xu J, Liao L, Ning G, Yoshida-Komiya H, Deng C, O’Malley BW 2000 The steroid receptor coactivator SRC-3 (p/CIP/RAC3/AIB1/ACTR/TRAM-1) is required for normal growth, puberty, female reproductive function, and mammary gland development. Proc Natl Acad Sci USA 97:6379 – 6384 23. Potier M, Elliot SJ, Tack I, Lenz O, Striker GE, Striker LJ, Karl M 2001 Expression and regulation of estrogen receptors in mesangial cells: influence on matrix metalloproteinase-9. J Am Soc Nephrol 12:241–251 24. Brotons JA, Olea-Serrano MF, Villalobos M, Pedraza V, Olea N 1995 Xenoestrogens released from lacquer coatings in food cans. Environ Health Perspect 103:608 – 612 25. Krishnan AV, Stathis P, Permuth SF, Tokes L, Feldman D 1993 Bisphenol-A: an estrogenic substance is released from polycarbonate flasks during autoclaving. Endocrinology 132:2279 –2286 26. Olea N, Pulgar R, Perez P, Olea-Serrano F, Rivas A, Novillo-Fertrell A, Pedraza V, Soto AM, Sonnenschein C 1996 Etrogenicity of resin-based composites and sealants used in dentistry. Environ Health Perspect 104:298 –305 27. Nagel SC, vom Saal FS, Thayer KA, Dhar M, 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 28. 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 Bio Med 217:300 –309 29. Howdeshell KL, Hotchkiss AK, Thayer KA, Vandenbergh JG, vom Saal FS 1999 Exposure to bisphenol A advances puberty. Nature 401:763–764 30. Gupta C 2000 The role of estrogen receptor, androgen receptor and growth factors in diethylstilbestrol-induced programming of prostate differentiation. Urol Res 28:223–229 31. Gupta C 2000 Reproductive malformation of the male offspring following maternal exposure to estrogenic chemicals. Proc Soc Exp Biol Med 224:61– 68 32. Steinmetz R, Brown NG, Allen DL, Bigsby RM, Ben-Jonathan N 1997 The environmental estrogen bisphenol A stimulates prolactin release in vitro and in vivo. Endocrinology 138:1780 –1786 33. Bergeron RM, Thompson TB, Leonard LS, Pluta L, Gaido KW 1999 Estrogenicity of bisphenol A in a human endometrial carcinoma cell line. Mol Cell Endocrinol 150:179 –187 34. Gould JC, Leonard LS, Maness SC, Wagner BL, Conner K, Zacharewski T, Safe S, McDonnell DP, Gaido KW 1998 Bisphenol A interacts with the estrogen receptor ␣ in a distinct manner from estradiol. Mol Cell Endocrinol 142:203–214 35. Diel P, Schulz T, Smolnikar K, Strunck E, Vollmer G, Michna H 2000 Ability of xeno- and phytoestrogens to modulate expression of estrogen- sensitive genes in rat uterus: estrogenicity profiles and uterotropic activity. J Steroid Biochem Mol Biol 73:1–10 36. Routledge EJ, White R, Parker MG, Sumpter JP 2000 Differential effects of xenoestrogens on coactivator recruitment by estrogen receptor (ER) ␣ and ER␤. J Biol Chem 275:35986 –35993 37. Andersen HR, Andersson AM, Arnold SF, et al. 1999 Comparison of shortterm estrogenicity tests for identification of hormone-disrupting chemicals. Environ Health Perspect 107(Suppl 1):89 –108