of two antiprogestins (mifpristone and onapristone) on endometrial factors of ... tigestagen onapristone in rabbits: relationship to uterine estrogen receptors.
0021-972X/01/$03.00/0 The Journal of Clinical Endocrinology & Metabolism Copyright © 2001 by The Endocrine Society
Vol. 86, No. 6 Printed in U.S.A.
Progesterone Antagonists Increase Androgen Receptor Expression in the Rhesus Macaque and Human Endometrium* OV D. SLAYDEN, NIHAR R. NAYAK, KEVIN A. BURTON, KRISTOF CHWALISZ†, SHARON T. CAMERON, HILARY O. D. CRITCHLEY, DAVID T. BAIRD, AND ROBERT M. BRENNER Division of Reproductive Sciences (O.D.S., N.R.N., R.M.B.), Oregon Regional Primate Research Center, Beaverton, Oregon 97006; Center for Reproductive Biology (K.A.B., S.T.C., H.O.D.C., D.T.B.), University of Edinburgh, EH3 9ET Edinburgh, United Kingdom; and Schering AG (K.C.), D13342 Berlin, Germany ABSTRACT Antiprogestins (APs) inhibit estradiol (E2)-stimulated endometrial growth in women and nonhuman primates, but the mechanism of this “antiestrogenic” action is unknown. Here, we report that APs upregulate endometrial androgen receptor (AR) in both women and macaques, an effect that might play a role in the antiproliferative effects of APs on the primate endometrium. In addition, because there are discrepancies in the literature on the regulation and localization of AR in the primate endometrium, we used both in situ hybridization and immunocytochemistry to evaluate hormonal influences on endometrial AR in women and macaques. In ovariectomized macaques, the following treatments were given for 4 weeks each: E2 alone, E2 ⫹ progesterone (P), E2 ⫹ mifepristone (RU 486), and E2 ⫹ P ⫹ RU 486. In women, samples were obtained during the normal menstrual cycle and after treatment with either RU 486 for 30 days at 2 mg/day, or after a single oral administration of 200 mg RU 486 on cycle day LH ⫹ 2. In macaques, E2 significantly increased AR expression above vehicle controls; E2 ⫹ RU 486 increased binding further; E2 ⫹ P decreased AR binding; and E2 ⫹ P ⫹ RU 486 treatment caused an intermediate elevation in AR binding. In macaques treated with E2 alone, stromal AR staining was predominant, and P treatment
P
ROGESTERONE (P) ANTAGONISTS [antiprogestins (APs)], including mifepristone (RU 486; Roussel UCLAF, Romainville, France), ZK 137 316, and ZK 230 211 (ZK compounds; Schering AG), are compounds that bind to the P receptor and inhibit P-initiated gene transcription. It is now well established that chronic treatment of women (1–3) and nonhuman primates (4 – 6) with low doses of APs either during the menstrual cycle or with combined estrogen therapy results in inhibition of endometrial proliferation; in macaques there is especially severe atrophy of the endometrium (4). The mechanism of the apparent antiestrogenic effect of Received December 18, 2000. Revision received March 13, 2001. Accepted March 20, 2001. Address correspondence and requests for reprints to: Robert M. Brenner, Ph.D., Oregon Regional Primate Research Center, 505 N.W. 185th Avenue, Beaverton, Oregon 97006. * Supported by Grants DAMD15-96-C-6096 and RR-00163 and Wellbeing (UK). Part of this work was undertaken by R.M.B. during a 1997 sabbatical in Edinburgh under the auspices of a Burroughs Wellcome Fellowship. † Present address: TAP Pharmaceutical Products, Lake Forest, Illinois 60045.
suppressed that staining. E2 ⫹ RU 486 or E2 ⫹ P ⫹ RU 486 treatment produced a striking up-regulation of glandular epithelial AR staining and enhanced the stromal AR signal. In situ hybridization analyses confirmed the immunocytochemistry data. Similar induction of glandular AR staining and enhanced stromal AR staining were obtained in macaques treated with ZK 137 316 and ZK 230 211. During the natural cycle in women, stromal AR staining predominated and was greater in the proliferative than the late secretory phase. RU 486 treatment of women up-regulated glandular epithelial AR staining after either daily treatment for 30 days with 2 mg/day or after a single oral dose of 200 mg. In summary, endometrial AR was highest in the stroma during the human proliferative phase (or during E2 treatment in macaques) and lowest during the late secretory phase in women (or after E2 ⫹ P treatment in macaques). In both species, RU 486 induced AR expression in the glands and enhanced AR expression in stromal cells. Because androgens can antagonize E2 action, enhanced endometrial AR expression induced by APs could play a role in the antiproliferative, “antiestrogenic” effects of APs in primates. (J Clin Endocrinol Metab 86: 2668 –2679, 2001)
APs on epithelial cell proliferation and endometrial mass has not been fully explained because 1) APs do not bind to the estrogen receptor (ER), and 2) other well established actions of estrogen in the endometrium, including up-regulation of ER and P receptor (PR), are not blocked but rather enhanced by AP treatment (3–5, 7). There is substantial evidence that exogenous androgen can have inhibitory effects on female reproductive systems (8, 9), including induction of endometrial atrophy (10). A recent report suggests that elevated androgens in women with recurrent miscarriages may specifically antagonize estrogen action directly in the endometrium (11). In preliminary work we reported that androgens could block estrogen action in the oviduct and uterus of ovariectomized macaques (12). These suppressive effects of elevated or exogenously administered androgens on endometrial function suggested to us that the androgen receptor (AR) could play a role in the antagonistic effects of APs. ARs are present in the human and nonhuman primate endometrium, but the results of various studies of AR regulation and/or localization are contradictory. For instance,
2668
AR IN THE UTERUS
some immunocytochemistry (ICC) studies (13) report localization of AR in both endometrial stroma and endometrial glandular epithelium (13–15), others report absence of any staining in human uteri (16, 17), and some report specific staining predominantly in stroma (14). In one in situ hybridization (ISH) study, AR messenger RNA (mRNA) was localized to both stroma and glandular epithelium, but with a significantly stronger signal in the stromal compartment (18). Equally conflicting are reports of hormonal regulation of AR; some report up-regulation of AR mRNA when estradiol (E2) treatment was combined with either androgens or progestins, although androgens were the most effective (18), whereas others report that AR mRNA was up-regulated by estrogens and unaffected by androgens (19). Transsexual women treated with testosterone are reported to have elevated endometrial AR (17). None of the above studies have used both ICC and ISH on the same tissue specimens. Therefore, in the current work in macaques, we used multiple assays (ligand binding, ICC, and ISH) on the same set of endometria to ascertain the regulation and localization of AR during hormonally regulated cycles and to evaluate the effects of AP treatment on AR. With human endometria we used both ICC and ISH to evaluate the localization of endometrial AR during the normal cycle, and used ICC to evaluate changes in AR in women treated with RU 486. The results indicate that during the normal cycle in both women and macaques, stromal cells are the predominant cell type that express AR. Moreover, AR levels are enhanced by E2 and suppressed by P. Of great interest, AP treatment enhanced stromal AR and greatly increased AR expression in the glands. These data suggest that during normal cycles any androgen effects on the endometrium would be mediated by the stroma and that, during AP treatment, elevated AR could play some role in the endometrial suppressive effects induced by APs. Experimental subjects
Human endometrial samples were obtained from three different patient groups. Institutional ethical approval was obtained for each of the study situations described, and all women provided written informed consent. The first group consisted of fertile women (n ⫽ 52) with regular menstrual cycles lasting between 25 and 35 days who had not been using hormonal preparations for the preceding 3 months. Samples were obtained at the time of hysterectomy for benign indications. Subjects with significant uterine pathology (for example, fibroids) were excluded. Endometrial tissue was collected from women during the proliferative phase (n ⫽ 15), early secretory phase (n ⫽ 5), midsecretory phase (n ⫽ 5), and late secretory phase (n ⫽ 7) of the menstrual cycle. After the uterus was removed, a full thickness biopsy of the endometrium was obtained, extending from the lumen to the muscular myometrial layer that included superficial and basal endometrium as well as myometrium. The stage of the menstrual cycle was consistent with the patients reported last menstrual period and histological dating using the criteria of Noyes et al. (20). The biopsy was fixed overnight in 4% paraformaldehyde and then embedded in paraffin for AR ICC analysis. Some samples were also frozen in Tissue Tek
2669
II OCT (Van Waters and Rogers International, South Plainfield, NJ), as described for macaque tissues, and shipped on dry ice to the Oregon Regional Primate Research Center for ISH of AR. The second group consisted of six healthy women with regular menstrual cycles (25–35 days). The effects of daily low-dose RU 486 on endometrial maturation and proliferation have already been described in this group of women (3). No subject had used hormonal contraception in the preceding 3 months. The subjects were studied in a control and a treatment cycle. Two placebo capsules were taken daily from the onset of menses of the first cycle (control). These were then replaced with identical capsules of RU 486 (two 1-mg capsules), commencing on the first day of menses of the second (treatment) cycle and daily for 30 days. Endometrial biopsies were collected with a Pipelle endometrial sampling device (Prodimed, Neuilly-en-Thelle, France) in the control cycle, 7–11 days after the plasma LH surge and on the corresponding day of treatment with RU 486 (days 19 –28). In the previous report on samples obtained from these women (3) we noted that RU 486 treatment suppressed mitotic activity in the endometrial glands and elevated ER and progestin receptor levels, but AR levels were not evaluated. Paraffinembedded tissue from these studies was available for the current work. The third group consisted of an additional 10 healthy women with regular menses (25–35 days), all of whom had participated in a study evaluating the effects of postovulatory administration of APs on the endometrium (21). Again, no subject had used hormonal contraception in the preceding 3 months. Endometrial biopsies from this group of women were available for evaluation of AR expression. Each subject had been studied over a control cycle and a treatment cycle. In the treatment cycle the woman received 200 mg oral RU 486 on the second day after the onset of the LH surge in urine (LH ⫹ 2). Endometrial biopsies were collected with a Pipelle endometrial sampler, either 4 or 6 days after the LH surge in the control cycle and on the corresponding day of the treatment cycle, that is, 2 or 4 days after oral administration of RU 486. All endometrial biopsies were fixed and embedded in paraffin by standard procedures. Experimental animals
The Division of Animal Resources at the Oregon Regional Primate Research Center provided all animal care following the NIH guidelines for use of nonhuman primates. Thirtyone adult rhesus macaques (Macaca mulatta) were ovariectomized and treated sequentially with 3-cm Silastic capsules (0.34 cm inner diameter, 0.46 cm outer diameter; Dow Corning Corp., Midland, MI) containing E2 for 14 days, and then E2 ⫹ P (6-cm Silastic capsules) to induce artificial cycles as described previously (6). Removal of the P implant at the end of the artificial cycle induces menstruation and begins the next cycle. E2 and P were purchased from either Steraloids Inc. (Wilton, NH) or Sigma (St. Louis, MO). Analysis of serum collected during these artificial cycles confirmed implant delivery of normal levels of E2 and P for the primate cycle (7, 22).
2670
SLAYDEN ET AL.
JCE & M • 2001 Vol. 86 • No. 6
FIG. 1. Representative saturation binding studies for AR in endometria of E2- and E2 ⫹ RU 486-treated macaques. A, Nuclear and cytosolic extracts of E2-treated animals. The insets in A and B show data plotted by the method of Scatchard. Kd ⫽ 1.4 –2.4 nM. B, Nuclear and cytosolic extracts of E2 ⫹ RU 486-treated animals. In each case, saturation of binding was achieved and the majority of specific binding was found in the cytosolic fraction.
E2, P, and RU 486 treatments. Artificially cycled macaques were treated with steroids as previously published (7). Briefly, 14 animals received implants of E2 for 2 weeks, to create a proliferative endometrium. The animals were then treated for an additional 2 weeks with the following four treatments: E2 only (n ⫽ 3); E2 ⫹ P ⫹ vehicle (n ⫽ 3); E2 ⫹ P ⫹ RU 486 (n ⫽ 4); and E2 ⫹ RU486 (n ⫽ 4). To create these treatments the E2 implant was left in place, and the animals received either an implant of P and/or injection with RU 486 (1 mg/kg in ethanol) im daily. Four additional animals received no hormones for 4 – 8 weeks (spay). Reproductive tract tissues were collected by midventral laparotomy at the end of treatment. The uterus was separated from the oviducts and cervix, quartered along the longitudinal axis, and crosssections (2 mm thick) from two uterine quarters were cut
with a razor blade and prepared for AR ICC and ISH. The samples of fresh tissue for ICC were microwave stabilized and frozen in liquid propane, as described previously (23). The samples for ISH were similarly frozen without microwave treatment. The remaining quarter was separated from the myometrium with fine scissors and analyzed by AR binding assay and, in some cases, by a sucrose gradient shift assay (24). E2 ⫹ ZK treatments. Beginning at the end of one artificial menstrual cycle (P implant withdrawal), three rhesus macaques were treated for 28 days with the following three treatments: no hormones (hormone withdrawal; spay; n ⫽ 3); E2 only (n ⫽ 4); E2 ⫹ 0.1 mg/kg ZK 137 316 (n ⫽ 3); and E2 ⫹ 0.032 mg/kg ZK 230 211 (n ⫽ 3). ZK 137 316 and ZK
AR IN THE UTERUS
FIG. 2. Endometrial total AR levels (mean ⫾ SE) measured by [3H]R1881 binding in spay, E2-, E2 ⫹ RU 486-, E2 ⫹ P-, and E2 ⫹ P ⫹ RU 486-treated macaques. Bars with different superscripts are significantly different (P ⬍ 0.05).
230 211 were provided by Schering AG (Berlin, Germany) and administered by im injection in a nonirritating vehicle that consisted of 37.5% HBSS (Life Technologies, Inc., Grand Island, NY), 37.5% 1,2-propanediol, and 25% ethanol (Aaper, Shelbyville, KY). At the end of treatment, the reproductive tracts were collected by laparotomy, and samples of uterus were frozen for AR ICC and ISH as described above. Materials and Methods AR binding assay Except where indicated, reagents for this and all other assays were purchased from the Sigma. Radioligand binding assays followed procedures similar to those previously validated for ER and PR (6). Fresh pieces of endometrium were weighed and homogenized in 10 vol (wt/ vol) TEDGM buffer [10 mm Tris, 1.5 mm EDTA, 1 mm dithiothreitol (DTT), 10% glycerol (vol/vol), and 25 mm NaMoO4 (pH 7.4)] on ice with precooled Duall tissue grinders, size 22 (Kontes Glass Co., Vineland, NJ) and then centrifuged at 1000 ⫻ g for 10 min at 4 C to separate the cytosols and crude nuclear pellets. The nuclear pellets were washed three times with 1 mL TEDGM buffer. The washed nuclear pellets were extracted with a high-salt buffer [10 mm Tris, 1.5 mm EDTA, 1 mm DTT, and 0.8 m potassium chloride (pH 7.4)] for 25 min on ice. An aliquot of the nuclear extract was frozen for DNA analysis (25), and the cytosols and nuclear extracts were centrifuged at 100,000 ⫻ g for 1 h at 4 C. Aliquots (100 L) of the postultracentrifugation cytosols or nuclear extracts were mixed with 50 L titrated methyltrienolone ([3H]-R1881; NEN Life Science Products, Boston, MA; 80 –90 Ci/mmol) or the same quantity of [3H]-R1881 ⫹ 200-fold excess of radioinert R1881. All assay tubes also contained 10 m final concentration of triamcinolone acetonide. For saturation studies and Scatchard analysis, a series of concentrations of [3H]-R1881 from 0.2–20 nm were used. The cytosols and nuclear extracts were incubated with steroid for 20 h at 4 C, and then the bound steroid was separated from free with the aid of Sephadex LH-20 minicolumns, as described previously (6). Specific binding was determined by subtraction of nonspecific binding ([3H]-R1881 ⫹ 200-fold radioinert R1881) from total binding ([3H]-R1881 only) and is expressed as femtomoles per micrograms DNA in the nuclear extract.
2671
FIG. 3. Sucrose density gradients showing shift of AR by the specific anti-AR antibody F-39.4. Nuclear extracts prepared from macaque endometrium incubated with [3H]-1881 were mixed with either monoclonal anti-AR (F-39.4) antibody, monoclonal AT, or no antibody and analyzed on 5–20% sucrose gradients. When mixed with AT, endometrial or oviductal extracts gave a peak ⬃4S. Aliquots of the same endometrial extracts, when mixed with F-39.4, gave a peak at ⬃7S. The positions of aldolase (8S) and BSA (4.5S) marker proteins are indicated by arrows.
Sucrose gradient analysis Endometrial slices (⬃1 mm thick) from an ovariectomized, 28-day E2-treated macaque were incubated for 1 h at 37 C in Trowells medium containing 10 nm [3H] R1881. The slices were washed three times in Trowells medium at 4 C, blotted dry, and then frozen in liquid nitrogen. AR was extracted and analyzed on sucrose gradients as described previously (24), with the following modifications. The samples of tissue were homogenized on ice in 1:10 (w:v) TEDGM buffer and centrifuged at 100,000 ⫻ g for 1 h as described above for binding assays. Aliquots of resulting cytosol were incubated in duplicate for 20 h at 4 C with either the AR-specific mouse monoclonal antibody (F39.4; BioGenex Laboratories, Inc., San Ramon, CA) (26), or an irrelevant antibody [antitimothy pollen (AT); courtesy of Dr. Arthur Malley, Oregon Regional Primate Research Center], or no antibody. The incubate was layered on top of linear 5–20% sucrose gradients and then centrifuged at 205,600 ⫻ g overnight at 4 C. Twenty-six fractions (25 drops each) were then collected from the bottom of each tube and analyzed by liquid scintillation counting. Protein solutions (BSA, 4.5S, and aldolase, 8S) were layered over replicate gradients as sedimentation markers, and the fractions were read in a spectrophotometer at 280 nm.
ICC ICC for AR in frozen sections of macaque was conducted with monoclonal antihuman AR antibody F39.4. Briefly, cryostat sections (5 m) were thaw-mounted on SuperFrost Plus slides (Fisher Scientific, Pittsburgh, PA), placed on ice, and microwaved for 2 sec. The microwavetreated sections were lightly fixed [0.2% picric acid and 2% paraformaldehyde in phosphate-buffered saline (PBS)] for 10 min; and then immersed twice for 2 min in 85% ethanol ⫹ 1.5% polyvinylpyrollidone at 4 C, rinsed in PBS, and then immersed twice in 0.37% glycine in PBS⫹ polyvinylpyrollidone. After further rinsing the slides were treated for 20 min at 4 C with a nonspecific serum (horse) and then incubated overnight at 4 C with F39.4. As a control for nonspecific reaction some sections were incubated with a nonspecific antibody directed against an antigen of AT. After various trials, F39.4 was used at a 1:45 dilution of the BioGenex Laboratories, Inc. stock solution, and AT was used at 10 ng/mL. Following incubation the slides were rinsed with 0.1% gel-
2672
SLAYDEN ET AL.
JCE & M • 2001 Vol. 86 • No. 6
FIG. 4. AR immunostaining in the endometrium of macaques. Only weak staining for AR was observed in the spayed animals (a and d). Treatment with E2 increased AR stromal immunostaining (b and e), which was strongest in the functionalis zone. Treatment with E2 ⫹ P decreased stromal AR staining (c and f). Treatment with RU 486 (g and j), ZK 137 316 (h and k), and ZK 230 211 (i and l) resulted in increased epithelial and stromal AR staining. No staining was observed in the vascular endothelium or vascular muscle of the spiral arteries (m). Myometrial staining was always strong (n), and no staining was observed after staining with AT (negative control). Gl, Glands; S, stroma. Original magnification, ⫻312.
atin, 0.075% BRIJ 35 Solution (Sigma, St. Louis, MO) in PBS (4 C), reincubated with nonspecific serum, and then with a biotinylated second antibody for 30 min (25 C). The biotinylated antibody complexes were then reacted with an avidin-biotin peroxidase kit (Vector Laboratories, Inc., Burlingame, CA). The slides were then treated with 0.05% osmium tetroxide for 1 min, washed, dehydrated, and coverslipped. Quantitative analysis of the abundance of AR-positive cells in the macaque endometrium was done on six random fields from the functionalis and basalis zones captured through an Optronics DEI-750TD color CCD camera (Optronics, Goleta, CA) at ⫻312 original magnification and displayed on a Sony Trinitron HR color monitor (Sony, Tokyo, Japan). Approximately 1000 cells were analyzed in each zone per animal.
The total number of epithelial cells and stromal cells per field and the number of epithelial cells and stromal cells that were clearly AR positive were counted by a trained observer (Kunie Mah). The percentage of AR-positive epithelial and stromal cells was then calculated, and differences in the percentage of positive cells among the animals in each treatment group were analyzed by ANOVA. For ICC of AR in human tissues, procedures followed those previously published for ER (3), except that ICC was conducted with monoclonal antibody F39.4, briefly as follows. Paraffin sections (5 m) were dewaxed in Histoclear (National Diagonistics, Atlanta, GA), rehydrated through a series of alcohols, and washed with PBS. The slides were then subjected to pressure cooker antigen retrieval (27) in 0.01 m sodium
AR IN THE UTERUS
2673
FIG. 5. Mean (⫾ SE) percentage of endometrial cells staining positive for AR by ICC. Means with different superscripts are significantly different (P ⬍ 0.05).
citrate buffer at pH6 for 5 min. Endogenous peroxidase activity was quenched by immersion in 3% hydrogen peroxide (Merck, Poole, UK) in distilled water for 10 min at room temperature. Nonspecific binding of the primary antibody was blocked by incubating the sections for 20 –30 min at room temperature in nonimmune horse serum (Vectastain; Vector Laboratories, Inc., Peterborough, UK). Slides were then incubated with either F39.4 overnight at 4 C at a 1:480 dilution in PBS/BSA gel or similarly with a control mouse IgG antibody at 1:600 dilution in PBS/BSA gel. Following a wash in PBS with Tween 20, the slides were incubated in biotinylated horse antimouse secondary antibody (Vectastain) in normal horse serum for 60 min at room temperature, reacted with the avidin-biotin peroxidase complex (Vectastain Elite) for 60 min at room temperature and visualized with substrate and chromagen 3,3⬘-diaminobenzidine (Vector Laboratories, Inc.). Negative controls were performed by replacing the primary antibody with mouse IgG at a matched concentration. Quantification of immunohistochemical observations was performed by objective image analysis using Openlab software (Improvision Inc., Lexington, MA). Twelve random digital fields were examined from the functionalis and basalis regions. The AR content of each section was expressed as the mean percentage of chromogen-positive cells from the total number of cells. Further analysis divided the tissue samples into menstrual cycle stages for temporal analysis using an ANOVA.
taining the appropriate concentration of the antisense probe (5 ⫻ 106 cpm/mL). After hybridization all the slides were treated with RNase A at 37 C for 30 min, rinsed in a descending series of SSC, and then washed in 0.1⫻ SSC at 65 C (high stringency) for 30 min. Sections were dehydrated in alcohol, vacuum dried, coated with NTB2 autoradiographic emulsion (Eastman Kodak Co., Rochester, NY), stored at 4 C for 2 weeks, developed in aqueous d-19 (Eastman Kodak Co.), lightly counterstained with hematoxylin, dehydrated in alcohol, cleared with xylene, and coverslipped with Permount (Fisher Scientific). Silver grains were counted over stroma and glandular epithelium in sections hybridized with radiolabeled probe. The counts were made with Optimas (Media Cybernetics, Silver Spring, MD) on images captured at ⫻250 original magnification. In RNase-treated slides, regardless of hormone treatment or state, grains over the tissues were identical to those over the glass slide. Therefore, a region of the slide away from the section was counted as background, and these counts were subtracted from each epithelial and stromal field counted. The abundance of silver grains in the field was expressed as the number of silver grains per nucleus (stromal cell and glandular cell, respectively). These counts were then expressed as a percentage of the maximum signal in all the sections analyzed.
ISH
Numeric data from binding assays, ISH grain counts, and percentage of AR-positive cells in human ICC preparations were analyzed by oneway ANOVA, followed by Fisher’s protected least significant difference test (30).
ISH of frozen sections of macaque and human endometrium was conducted with a riboprobe labeled with [35S]UTP (NEN Life Science Products) derived from a rhesus monkey-specific 330-bp AR complementary DNA (cDNA) (Ref. 28; GenBank accession number AF092930). The nucleotide sequence of this AR cDNA has 99% homology with the corresponding AR human sequence (28). Techniques for ISH with 35Slabeled probes were previously published (29). Briefly, 10-m thick frozen sections were mounted on Super Frost Plus slides (Fisher Scientific) and fixed in 4% paraformaldehyde in PBS for 10 min at 4 C. The tissue sections were rinsed in 2⫻ SSC, acetylated with 0.25% acetic anhydride in 0.1 m triethanolamine (pH 8.0) for 10 min, and then airdried. At this point, at least one slide per tissue group was treated with RNase A [20 mg/mL RNase A, 0.5 m NaCl, 0.01 m Tris, and 1 mm EDTA (pH 8.0)] as a negative control. All the slides were prehybridized for 1 h at 42 C in 10 mm DTT, 0.3 m NaCl, 20 mm Tris (pH 8.0), 5 mm EDTA, 1⫻ Denhardt solution, 10% dextran sulfate, and 50% formamide. Then, sections were incubated at 55 C overnight in the same solution con-
Statistical analysis
Results AR expression in the rhesus macaque
Binding kinetics. Figure 1 shows a representative saturation study of AR binding with [3H]-R1881 done on endometrial cytosolic and nuclear extracts from E2-treated and E2 ⫹ RU 486-treated animals. Scatchard transformation of these data showed linear plots with an estimated Kd of 1.4 –2.4 nm from both treatments, indicating that RU 486 had no effect on the binding affinity of [3H]-R1881 to the AR. Regardless of hormone treatment, there was consistently
2674
SLAYDEN ET AL.
JCE & M • 2001 Vol. 86 • No. 6
FIG. 6. ISH of AR in the macaque endometrium. ISH confirmed immunolocalization of AR to the endometrial stroma of macaques treated with E2 or E2 ⫹ P (a–f). Treatment with E2 ⫹ RU 486, E2 ⫹ ZK 137 316, and E2 ⫹ ZK 230 211 in the functionalis (g–i) as well as the basalis (j–l) all caused increases in the AR mRNA signal in the epithelium as well as the stroma. E2 ⫹ ZK 230 211 caused the least increase in the AR mRNA signal. The inset (m) shows negative (RNase) control. Gl, Gland; E, epithelium; S, stroma. Original magnification, ⫻500.
10-fold more specific binding of [3H]-R1881 in the cytosolic than the nuclear fraction, an indication that the AR complex was loosely bound to nuclear components regardless of treatment. Binding assays. Figure 2 compares specific [3H]-R1881 binding in endometrial extracts from untreated spay, E2, E2 ⫹ P, E ⫹ RU 486, and E2 ⫹ P ⫹ RU 468-treated macaques. Because nuclear AR quantities in the high-salt nuclear extracts made up only a small percentage of the total AR, the nuclear and cytosolic values were combined. Binding to the endometrium of the untreated spay animals was minimal, and treatment with E2 alone resulted in significantly more AR binding
(P ⬍ 0.05). Sequential treatment with E2 ⫹ P lowered binding compared with E2 alone (P ⬍ 0.05). However, sequential treatment with E2 ⫹ RU 486 increased AR binding (P ⬍ 0.05) above the level induced by E2 alone, and E2 ⫹ P ⫹ RU 486 elevated binding to a level above that found with E2⫹P and intermediate between that induced by E2 ⫹ RU 486 and E2 alone. Gradient shift assays. Sucrose gradient analysis was used to validate that the monoclonal AR-specific antibody F-39.4 binds to the macaque endometrial AR. A representative experiment is shown in Fig. 3. When nuclear extracts prepared from endometrium were incubated at 37 C with [3H]-R1881
AR IN THE UTERUS
2675
FIG. 7. Quantification of the AR mRNA levels by grain counts in ISH preparations. Values represent (mean ⫾ SE) percentage of maximum signal. Means with different superscripts are statistically different (P ⬍ 0.05). The results support the micrographs shown in Fig. 6.
⫹ F-39.4 and then analyzed by ultracentrifugation on sucrose gradients, the labeled antibody-receptor complex sedimented as a 7S moiety. However, when either the irrelevant monoclonal antibody AT or no antibody was mixed with aliquots of the same endometrial extract, the peak of radioactivity was seen at slightly less than 4S. This indicates that F-39.4 binds with high affinity to endometrial AR to form a heavy F-39.4-receptor-[3H]-R1881 complex that migrates further in the gradient than the lighter, uncomplexed receptor. These data validate the ability of the F-39.4 antibody to recognize macaque endometrial AR. ICC. Figure 4 presents photomicrographs of AR immunostaining in the macaque; the percentage of positive cells staining for AR in each treatment group is presented in Fig. 5. In untreated ovariectomized animals (spay), specific AR staining was just barely detectable in some of the endometrial stromal cells in the functionalis and basalis zones (Fig. 4, a and d). Treatment with E2 for 28 days resulted in increased specific AR nuclear immunostaining in the stromal cells, but not in the glandular epithelial cells. This increase in stromal AR was more striking in the endometrial functionalis (Fig. 4b), but also occurred in the basalis zone (Fig. 4e) and included a significant increase in the number of stromal cells staining (Fig. 5). After sequential E2, followed by E2 ⫹ P treatment (Fig. 4c), AR staining intensity was reduced, and significantly fewer stromal cells were positive for AR in the functionalis zone. This P-mediated reduction was less evident in the basalis zone (Fig. 4f). Staining remained absent from the glandular epithelium. Combined treatment with E2 ⫹ RU 486 (Fig. 4, g and j) or E2 ⫹ ZK 137 316 (Fig. 4, h and k) resulted in a remarkable appearance of AR staining in the glandular epithelium of both the functionalis and basalis zones, although it was
most striking in the functionalis zone. Stromal staining was also strong in the RU 486-treated animals, but the number of cells staining was similar to endometrium after E2 treatment. Similar effects were seen after combined E2 ⫹ P⫹ RU 486 (data not shown), indicating that RU 486 blocked the down-regulatory action of P and led to an increase in epithelial AR. Endometria from animals treated with E2 ⫹ ZK 230 211 also showed an increase in both epithelial and stromal staining over those treated with E2 alone, but this increase was not as marked as with RU 486 or ZK 137 316. The statistical significance of the changes indicated above are displayed in Fig. 5, which summarizes the percentage of immunopositive cells in each cellular compartment for each treatment. Regardless of treatment, endothelial cells and vascular muscle cells of the spiral arteries (Fig. 4m) and endothelial cells of veins were negative for AR. Myocytes in the macaque myometrium stained strongly for AR (Fig. 4n), but there was no apparent effect of treatment on myometrial staining. There was no staining with an irrelevant antibody (Fig. 4o). ISH. Overall, the localization and quantification of AR mRNA by ISH confirmed the changes in AR protein revealed by ICC. In spayed macaques a basal, low level of AR mRNA was detected in the stroma of the functionalis and basalis zones, but the glands showed only near-background levels (Fig. 6, a and d) and the silver grain counts (Fig. 7) over the glands were less than 5% of the maximum signal. Treatment with E2 for 28 days significantly (P ⬍ 0.05; Fig. 7) increased AR mRNA in the stromal cells, but not the glands (Fig. 6, b and e) in both the functionalis and basalis zones. E2 ⫹ P treatment (Fig. 6, c and f) reduced the abundance of AR transcript by approximately 40% (Fig 7,
2676
SLAYDEN ET AL.
JCE & M • 2001 Vol. 86 • No. 6
FIG. 8. ICC and ISH of AR in the human endometrium. Top, ICC of AR in the proliferative (a and d), early secretory (b and e), and late secretory (c and f) of the natural menstrual cycle. All AR staining was evident only in the stroma; staining intensity and number of immunopositive cells were least in the late secretory phase. The inset shows absence of staining when the control IgG was used as first antibody; all nuclear staining is due to hematoxylin. Original magnification, ⫻20. Middle, ISH of AR in the early proliferative phase. A comparison of low (⫻25) magnification bright field (g) and dark field (h) optics shows that all the AR mRNA signal is in the stroma. A higher magnification (⫻500; i) of the ISH clearly shows that the glands are negative. Bottom, ICC of AR in controls vs. RU 486-treated human endometria. Compared with controls (j), treatment with RU 486 increased AR staining in the glandular epithelium (k and l) and also enhanced stromal staining (⫻312).
P ⬍ 0.05). The highest levels of AR mRNA in both stroma and glands were detected by ISH after either combined E2 ⫹ RU 486 or E2 ⫹ ZK 137 316 (Fig. 6, g-k; Fig. 7). Moreover, as seen with ICC, there were significant increases in grain
counts over glands after AP treatment (Fig 7; P ⬍ 0.05) in both the functionalis and basalis zones. Treatment with ZK 230 211 also increased AR mRNA but not as greatly as RU 486 or ZK 137 316.
AR IN THE UTERUS AR expression in the human endometrium
Expression of AR during the menstrual cycle. In all stages of the cycle, AR staining was localized predominantly in the endometrial stroma with no (or barely detectable) staining in the glands (Fig. 8, a–f). The abundance of AR-positive cells was similar in the functionalis and basalis stroma throughout the entire proliferative, early secretory and midsecretory phases of the cycle, but there was a clear decrease in such staining in the late secretory phase. Quantitation of the frequency of AR-positive cells showed that the decrease in the late secretory phase was highly significant (Fig. 9; P ⬍ 0.001). We also performed ISH of AR in human endometrium to determine whether the same cells that expressed the AR protein also expressed AR mRNA. A representative sample from the early proliferative phase is depicted at low magnification with bright field (Fig. 8g) and dark field (Fig. 8h) illumination and at high magnification bright field (Fig. 8i). These images confirm the ICC data and show that AR mRNA was predominantly localized in the endometrial stroma and not the glandular epithelium. As with ICC, ISH of late secretory phase endometrium showed a decline in the intensity of the stromal cell signal in the late secretory phase (data not shown). Elevation of endometrial AR immunopositive staining by RU 486 treatment. We evaluated the ICC preparations of endometria from six control patients and six women treated with 2 mg/ day RU 486 for 30 days as described in Materials and Methods. In the control specimens there was essentially no staining of the glandular or surface epithelium, but distinct staining of the stroma (Fig. 8j). After 30 days of RU 486, there were distinct and notable increases in the AR staining of the glands and surface epithelium plus some enhancement of stromal AR staining (Fig. 8k). Furthermore, we evaluated the ICC preparations from five placebo controls and five patients
FIG. 9. Quantification of AR-positive cells in human endometrium. Data are expressed as mean (⫾SE) percentage of cells staining positive for AR during the menstrual cycle (P, Proliferative; ES, early secretory; MS, midsecretory; LS, late secretory). The percentage of AR immunopositive stromal cells was significantly lower (P ⬍ 0.05) during the late secretory phase compared with other phases of the cycle.
2677
treated with a single dose of 200 mg RU 486 on day LH ⫹ 2 and sampled on days LH ⫹ 4 and LH ⫹ 6. The controls, who were also sampled on days LH ⫹ 4 and LH ⫹ 6, had essentially no AR staining in the glands or surface epithelium, but had definite stromal AR staining (Fig. 8j). We note here that Fig. 8j is representative of the controls from both the 2-mg and the 200-mg studies. On both days LH ⫹ 4 and LH ⫹ 6 after the 200 mg RU 486dose there were clear increases in glandular and surface epithelial staining and some increase in stromal AR staining (Fig. 8l). These findings indicate that both chronic and acute RU 486 treatment can moderately enhance stromal AR staining and greatly increase glandular and surface epithelial staining in the human endometrium. Discussion Hormonal regulation of AR expression
In both women and macaques, expression of AR during the cycle seems regulated by changes in systemic levels of E2 and P. Results obtained by three independent means (binding assays, ICC, and ISH) revealed that AR mRNA and protein levels were minimal in ovariectomized macaques and were increased significantly by 28 days of E2 treatment. This increase under E2 was inhibited when P was added to the E2 treatment for the last 14 days of the 28-day period. In women, an analogous regulation was shown, as the percentage of AR immunopositive stromal cells was significantly higher in the proliferative than the late secretory phase of the cycle. In macaques, our binding, ICC, and ISH studies further showed that treatment with RU 486 significantly increased AR levels, and our ICC studies revealed that this increase in AR was especially notable in the glandular epithelium in both women and macaques. Our binding studies also revealed that the endometrial AR is functional, in the sense that it could bind radiolabeled ligand, and that the F-39.4 antibody was specific, in that it could bind and shift radiolabeled endometrial AR on a sucrose gradient. However, whereas all specific immunostaining for AR in women and macaques was localized to the nuclei of AR-positive cells, greater than 90% of the total AR was found in the cytosolic fraction after homogenization and low-salt extraction of macaque endometria. This indicates that AR is nuclear in situ but is only loosely bound to chromatin and is easily extracted regardless of hormone treatment. Because the physiological role of androgens and endometrial AR during the menstrual cycle is unclear, additional studies are needed to assess whether the affinity of AR for nuclear components differs under different hormonal conditions and, in particular, whether androgens can affect the affinity of primate endometrial AR for chromatin. In macaques, systemic androgen levels are higher during the follicular than the luteal phase of the cycle, and there is a definite periovulatory surge (31). In women, systemic androgen concentrations show less variation during the cycle but show a similar periovulatory surge (32). Currently, the role of endogenous androgens in the endometrium during normal menstrual cycles is not clear, but these data suggest that any AR-mediated effects of normal levels of endogenous
2678
SLAYDEN ET AL.
androgens would be most prominent in the proliferative and periovulatory periods. Cellular localization of AR
Our data, which are the first to be based on both ICC and ISH analysis of the same specimens in both women and macaques, show that during the menstrual cycle, in both women and macaques, endometrial AR was primarily stromal in distribution with only an occasional cell in the glandular epithelium showing either positive staining or a clear hybridization signal for AR. These findings are most similar to those reported by Mertens et al. (14) for the human endometrium and Adesanya et al. (18) for the rhesus macaque endometrium. However, unlike Adesanya et al. (18), who reported an up-regulation of AR mRNA by E2, but no effect of P on E2 action, we found that P significantly suppressed the elevations in both AR binding and AR expression induced by E2 treatment. At present, we have no explanation for these different findings, except that different quantitative methodologies were used in the different laboratories. We found no substantial, specific AR staining or AR hybridization in the endometrial glandular epithelium, vascular endothelium, or smooth muscle of the spiral arteries under normal cyclic conditions. Consequently, any genomic effects of androgens on the endometrium during the cycle would be effected through the AR in stromal cells. Treatment with APs in both macaques and women not only enhanced stromal AR expression but induced striking increases in AR in the endometrial glands. Thus, after AP treatment, endometrial androgens could have direct glandular as well as indirect stromal effects. The APs used in this study differ in potency; both ZK 230 211 and ZK 137 316 are considerably more potent than RU 486. In addition, they may act through the AR in different ways. For example, in castrated rats, RU 486 is an antiandrogen, ZK 230 211 is a mixed androgen/antiandrogen, and ZK 137 316 is a weak androgen (K.C., data not shown). However, the fact that all three APs elevate AR suggests that this specific effect is mediated by the PR. The precise mechanism through which AP treatment elevates AR is not clear, but our previous studies (4, 7) and those of others (33) have shown that AP treatment, either during the menstrual cycle, or with combined estrogen therapy, also leads to elevations of the two other main uterine steroid receptors, ER and PR. Our current view is that the underlying mechanism may paradoxically involve an enhancement of E2 action at the molecular level, for the following reasons. Recent evidence indicates that unoccupied PR (in the presence of E2 and absence of P) can downmodulate the effects of E2. For example, in PR knockout mice, E2 action on the uterus and oviducts results in hyperplasia (34) whereas in wild-type mice, similar E2 treatment normally has a far less dramatic effect on uterine proliferation. This suggests that the unoccupied PR in wild-type mice can have a “braking” effect on E2 action. A similar role for unoccupied PR in the rabbit endometrium (35) and the rat brain (36) has been described. Because treatment with an AP would effectively occupy PR and remove it from play, any braking
JCE & M • 2001 Vol. 86 • No. 6
effect of PR on E2 action would be removed. As a result, E2 action at the molecular level would be enhanced, and levels of ER, PR, and AR would increase, because these are all molecules whose expression is increased by E2 (22). How an enhanced effect of E2 action at the molecular level could lead to mitotic inhibition and endometrial atrophy at the cellular and organ level is a puzzling issue that has so far resisted analysis. Our new data on the higher levels of AR in APtreated tissues suggests that AR could play a role in this inhibitory response. The increase in AR could lead to increased binding of endometrial androgens, which might antagonize the effects of estrogens on endometrial growth. For example, testosterone and danazol have been shown to inhibit human endometrial cell proliferation in tissue culture when the medium contained phenol red in amounts adequate to provide an estrogenic stimulus (37). These authors suggested that the antiendometriotic effects of danazol may involve direct antiestrogenic effects in the endometriotic lesions. Androstenedione can inhibit human endometrial cell growth and secretory activity in vitro (38), and high plasma androgen levels have been associated with adverse reproductive outcome, including recurrent miscarriage (39, 40). Hyperandrogenism is associated with lowered endometrial glycodelin secretion during the secretory phase (11), an effect the authors presumed was due to androgen antagonism of the ability of estrogen to prime the endometrium during the proliferative phase. In preliminary work with ovariectomized macaques, we noted that systemic treatment with dihydrotestosterone blocked the stimulatory effects of E2 on ER and PR levels, oviductal ciliogenesis, and endometrial mitosis (12). These various findings give credence to the possibility that elevated AR might mediate antiestrogenic effects of androgens in the endometrium during AP treatment. However, the endometrial atrophy induced by APs involves not only a specific blockade of glandular cell mitosis, but a severe compaction of the stroma that contributes to the overall thinning of the endometrium, an increase in apoptosis, and degenerative changes in the macaque spiral arteries. The latter effect may greatly reduce endometrial blood flow and contribute to the overall endometrial atrophy observed after AP treatment (41). Whether androgens could mediate all these varied effects of APs or whether APs interact in concert with other factors remains to be established. We suspect that several factors are involved. In summary, ARs are expressed predominantly in the human and nonhuman primate endometrial stroma throughout the menstrual cycle. Under normal cyclic conditions, stromal AR levels seem to be up-regulated by E2 and suppressed by P. The effects of endogenous androgens on the endometrium during the normal cycle remain to be elucidated, but it seems clear that any such AR-mediated effects would occur through the stroma. After AP treatment, AR expression is enhanced in the endometrial stroma and induced in the glandular epithelium, and these elevations could mediate androgen-dependent antiestrogenic effects. Whether androgens do, in fact, contribute to the endometrial suppression observed during AP treatment is a matter for further research.
AR IN THE UTERUS Acknowledgments We thank Kunie Mah, Xiao Jing Nie, and Teresa Henderson for technical assistance and Angela Adler for word processing.
References 1. Batista MC, Cartledge TP, Zellmer AW, et al. 1992 Delayed endometrial maturation induced by daily administration of the antiprogestin RU 486: a potential new contraceptive strategy. Am J Obstet Gynecol. 167:60 – 65. 2. Marions L, Gemzell Danielsson K, Swahn M-L, Bygdeman M. 1998 Contraceptive efficacy of low doses of mifepristone. Fertil Steril. 70:813– 816. 3. Cameron ST, Critchley HOD, Thong KJ, Buckley CH, Williams AR, Baird DT. 1996 Effects of daily low dose mifepristone on endometrial maturation and proliferation. Hum Reprod. 11:2518 –2526. 4. Slayden OD, Zelinski-Wooten MB, Chwalisz K, Stouffer RL, Brenner RM. 1998 Chronic treatment of cycling rhesus monkeys with low doses of the antiprogestin ZK 137 316: morphometric assessment of the uterus and oviduct. Hum Reprod. 13:269 –277. 5. Wolf JP, Hsiu JG, Anderson TL, Ulmann A, Baulieu EE, Hodgen GD. 1989 Noncompetitive antiestrogenic effect of RU 486 in blocking the estrogenstimulated luteinizing hormone surge and the proliferative action of estradiol on endometrium in castrate monkeys. Fertil Steril. 52:1055–1060. 6. Slayden OD, Hirst JJ, Brenner RM. 1993 Estrogen action in the reproductive tract of rhesus monkeys during antiprogestin treatment. Endocrinology. 132:1845–1856. 7. Slayden OD, Brenner RM. 1994 RU 486 action after estrogen priming in the endometrium and oviducts of rhesus monkeys (Macaca mulatta). J Clin Endocrinol Metab. 78:440 – 448. 8. Miller N, Bedard YC, Cooter NB, Shaul DL. 1986 Histological changes in the genital tract in transsexual women following androgen therapy. Histopathology. 10:661– 669. 9. Futterweit W, Deligdisch L. 1986 Histopathological effects of exogenously administered testosterone in 19 female to male transsexuals. J Clin Endocrinol Metab. 62:16 –21. 10. Loeser AA. 1938 The action of testosterone propionate on the uterus and breast. Lancet. 12:373–374. 11. Okon MA, Laird SM, Tuckerman EM, Li T-C. 1998 Serum androgen levels in women who have recurrent miscarriages and their correlation with markers of endometrial function. Fertil Steril. 69:682– 690. 12. Hirst JJ, Slayden OD, West NB, Brenner RM. Androgens suppress estrogen and progestin receptors in the endometrium and oviduct of rhesus monkeys. Proc 39th Meeting of the Society for Gynecological Investigation, San Antonio, TX, 1992; p 309. 13. Horie K, Takakura K, Imai K, Liao S, Mori T. 1992 Immunohistochemical localization of androgen receptor in the human endometrium, decidua, placenta and pathological conditions of the endometrium. Hum Reprod. 7:1461–1466. 14. Mertens HJ, Heineman MJ, Koudstaal J, Theunissen P, Evers JL. 1996 Androgen receptor content in human endometrium. Eur J Obstet Gynecol Reprod Biol. 70:11–13. 15. Kimura N, Mizokami A. 1994 Immunocytochemical localization of androgen receptor with polyclonal antibody in paraffin-embedded human tissues: authors’ response. J Histochem Cytochem. 42:126 (Letter). 16. Ruizeveld de Winter JA, Trapman J, Vermey M, Mulder E, Zegers ND, van der Kwast TH. 1991 Androgen receptor expression in human tissues: an immunohistochemical study. J Histochem Cytochem. 39:927–936. 17. Chadha S, Pache TD, Brinkmann AO, van der Kwast TH. 1994 Androgen receptor expression in human ovarian and uterine tissue of long term androgen-treated transsexual women. Hum Pathol. 25:1198 –1204. 18. Adesanya OO, Zhou J, Wu G, Bondy C. 1999 Location and sex steroid regulation of androgen receptor gene espression in rhesus monkey uterus. Obstet Gynecol. 93:265–270. 19. Fujimoto J, Nishigaki M, Hori M, Ichigo S, Itoh T, Tamaya T. 1994 The effect of estrogen and androgen on androgen receptors and mRNA levels in uterine
20. 21. 22. 23. 24. 25. 26. 27. 28.
29.
30. 31. 32. 33. 34. 35.
36. 37. 38. 39. 40. 41.
2679
leiomyoma, myometrium and endometrium of human subjects. J Steroid Biochem Mol Biol. 50:137–143. Noyes RW, Hertig AT, Rock J. 1950 Dating the endometrial biopsy. Fertil Steril. 1:3–25. Cameron ST, Critchley HOD, Buckley CH, Kelly RW, Baird DT. 1997 Effect of two antiprogestins (mifpristone and onapristone) on endometrial factors of potential importance for implantation. Fertil Steril. 67:1046 –1053. Brenner RM, Slayden OD. 1994 Cyclic changes in the primate oviduct and endometrium. In: Knobil E, Neill JD, eds. The physiology of reproduction. New York: Raven Press; 541–569. Slayden OD, Koji T, Brenner RM. 1995 Microwave stabilization enhances immunocytochemical detection of estrogen receptor in frozen sections of macaque oviduct. Endocrinology. 136:4012– 4021. Choate JVA, Slayden OD, Resko JA. 1998 Immunocytochemical localization of androgen receptors in brains of developing and adult male rhesus monkeys. Endocrine. 8:51– 60. Burton K. 1956 A study of the conditions and mechanism of the diphenylamine reaction for the colorimetric estimation of deoxyribonucleic acid. Biochem J. 62:315–323. Ruizeveld de Winter JA, van Weerden WM, Faber PW, et al. 1992 Regulation of androgen receptor expression in the human heterotransplantable prostate carcinoma PC-82. Endocrinology. 131:3045–3050. Norton AJ, Jordan S, Yeomans P. 1994 Brief, high-temperature heat denaturation (pressure cooking): a simple and effective method of antigen retrieval for routinely processed tissues. J Pathol. 173:371–379. Abdelgadir SE, Roselli CE, Choate JVA, Resko JA. 1999 Androgen receptor messenger ribonucleic acid in brains and pituitaries of male rhesus monkeys: studies of distribution, hormonal control, and relationship to luteinizing hormone secretion. Biol Reprod. 60:1251–1256. Nayak NR, Critchley HO, Slayden OD, et al. 2000 Progesterone withdrawl up-regulates vascular endothelial growth factor receptor type-2 in the superficial zone stroma of the human and macaque endometrium: potential relevance to menstruation. J Clin Endocrinol Metab. 85:3442–3452. Petersen RG. 1985 Design and analysis of experiments, ed 66. New York: Marcel Dekker, Inc. Hess DL, Resko JA. 1973 The effects of progesterone on the patterns of testosterone and estradiol concentrations in the systemic plasma of the female rhesus monkey during the intermenstrual period. Endocrinology. 92:446 – 453. Genazzani AR, Magrini G, Facchinetti F, et al. 1977 Behavior and origin of plasma androgens throughout the menstrual cycle. In: Martini L, Motta M, eds. Androgens and antiandrogens. New York: Raven Press; 247–261. Neulen J, Williams RF, Hodgen GD. 1990 RU 486 (Mifepristone): induction of dose dependent elevations of estradiol receptor in endometrium from ovariectomized monkeys. J Clin Endocrinol Metab. 71:1074 –1075. Lydon JP, DeMayo FJ, Funk CR, et al. 1995 Mice lacking progesterone receptor exhibit pleiotropic reproductive abnormalities. Genes Dev. 9:2266 –2278. Chwalisz K, Hegele-Hartung C, Fritzemeier K-H, Beier HM, Elger W. 1991 Inhibition of the estradiol-mediated endometrial gland formation by the antigestagen onapristone in rabbits: relationship to uterine estrogen receptors. Endocrinology. 129:312–322. Chappell PE, Levine JE. 2000 Stimulation of gonadotropin-releasing hormone surges by estrogen.1. Role of hypothalamic progesterone receptors. Endocrinology. 141:1477–1485. Rose GL, Dowsett M, Mudge JE, White JO, Jeffcoate SL. 1988 The inhibitory effects of danazol, danazol metabolites, gestrinone, and testosterone on the growth of human endometrial cells in vitro. Fertil Steril. 49:224 –228. Tuckerman EM, Okon MA, Li T-C, Laird SM. 2000 Do androgens have a direct effect on endometrial function? An in vitro study. Fertil Steril. 74:771–779. Bussen S, Sutterlin M, Steck T. 1999 Endocrine abnormalities during the follicular phase in women with recurrent spontaneous abortion. Hum Reprod. 14:18 –20. Tulppala M, Stenman U-H, Cacciatore B, Ylikorkala O. 1993 Polycystic ovaries and levels of gonadotropins and androgens in recurrent miscarriage: prospective study in 50 women. Br J Obstet Gynaecol. 100:348 –352. Reinsch RC, Murphy AA, Morales AJ, Yen SSC. 1994 The effects of RU 486 and leuprolide acetate on uterine artery blood flow in the fibroid uterus: a prospective, randomized study. Am J Obstet Gynecol. 170:1623–1628.