Androgenic Induction of Growth and Differentiation in the Rodent Uterus Involves the Modulation of Estrogen-Regulated Genetic Pathways

http://www.100md.com 内分泌学杂志 2005年第2期

     Departments of Molecular Endocrinology and Bone Biology (P.V.N., P.M., M.A.G., B.P., A.S., D.B.K., L.P.F., S.-i.H., W.J.R.), Laboratory Science and Investigative Technology (J.X.), and Biometrics Research (D.H.), Merck Research Labs, West Point Pennsylvania 19401; and Department of Medicine (D.T.), Division of Bone and Mineral Diseases, Washington University School of Medicine, Barnes-Jewish Hospital North Campus, St. Louis, Missouri 63110

    Address all correspondence and requests for reprints to: William J. Ray, Department of Molecular Endocrinology, Merck & Co. Inc., P.O. Box 4, WP26A1000, West Point Pennsylvania 19401. E-mail: james_ray@merck.com.

    Abstract

    The androgen receptor (AR) is expressed in the uterus; however, the role of AR in female reproductive physiology is poorly understood. Here we examined the effects of androgens on uterine growth and gene expression in adult ovariectomized rats. Nonaromatizable AR-selective agonists potently stimulate hypertrophy and induce significant myometrial expansion distinct from that induced by 17?-estradiol (E2). In the endometrium, androgens only modestly increase epithelial cell height and antagonize the trophic effects of E2. To identify underlying mechanisms, global changes in RNA levels 24 h after stimulation with E2 and 5-dihydrotestosterone (DHT) were compared. A total of 491 genes were differentially expressed after E2 treatment, including key regulators of tissue remodeling, cell signaling, metabolism, and gene expression. Of the 164 transcripts regulated by DHT, 86% were also affected by E2, including trophic genes like IGF-I and epithelial secretory genes such as uterocalin. In estrogen receptor (ER) knockout mice, DHT cannot induce uterine growth, suggesting a key role for ER. However, DHT appears not to activate ER directly because DHT induction of IGF-I is blocked by the AR antagonist bicalutamide, and multiple genes regulated directly by ER were not induced by DHT. The similarity between estrogens and androgens instead could reflect general trophic signaling in reproductive tissues because 93 of the 503 genes regulated in the uterus are similarly affected during prostate growth. Thus androgens regulate the trophic environment and architecture of the rodent uterus via a gene expression program that is overlapping but distinct from the estrogen response.

    Introduction

    FEMALES PRODUCE OVARIAN and adrenal androgens both before and after menopause (1, 2, 3). Testosterone, the major circulating androgen, exhibits cyclical serum concentrations during the menstrual cycle, peaking in most studies around midcycle (4, 5, 6, 7, 8, 9, 10, 11). During normal aging testosterone levels decline such that in postmenopausal women their concentration is about half of youthful peaks (12, 13). Thus, androgen replacement therapy, in which testosterone is administered to postmenopausal women, has been proposed to improve several age-related problems, such as reduced sexual function, cognitive decline, hot flashes, and depression, and could increase muscle mass and bone mineral density (14, 15, 16, 17, 18, 19, 20). Given the concerns associated with prolonged exposure to estrogen, androgen replacement therapy is a possible alternative for the management of these common postmenopausal problems. Furthermore, androgen receptor ligands are used to for the management of endometriosis (21), hirsutism (22), and in contraception [because many antiprogestins are testosterone derivatives and also have affinity for androgen receptor (AR)] (23, 24). Thus, it is important to understand the role of androgens in uterine physiology as a step toward safe and effective use of AR ligands in women.

    Androgen signal transduction is mediated by the AR, a steroid hormone nuclear receptor that functions primarily as a ligand-activated transcription factor (25). Testosterone diffuses into target cells and activates the receptor directly or can be converted by 5-reductases to 5-dihydrotestosterone (DHT), a more potent agonist. Once activated, AR forms homodimers and binds to the regulatory regions of target genes to either promote or inhibit transcription. In males, AR is expressed in reproductive organs such as the prostate, in which AR activity in the secretory epithelial cells and surrounding stromal smooth muscle cells is required for its development, maintenance, and function (26). Aberrant androgen signaling is central to the initiation and progression of prostate cancer, the second most common form of cancer in men in the United States (26, 27). In females, AR is also expressed in breast, ovaries, and uterus (28, 29), and several studies demonstrate functional consequences of AR activation or blockade (30, 31). But the role of androgens in regulating female reproductive physiology remains largely unknown. Furthermore, whereas there have been studies suggesting a role for AR in breast (32, 33) and ovarian (34) cancer, it is not clear how AR might influence tumorigenesis in these organs.

    The uterus is organized into the endometrium, which is comprised of secretory epithelial cells and surrounding mesenchymal cells, and an outer muscle layer, the myometrium. In humans AR is expressed throughout both the endometrium and myometrium, with less protein detectable in the secretory epithelial cells (35, 36). AR levels increase during the proliferative phase followed by a decline during the secretory phase (35, 36, 37, 38). In the rat uterus, AR is induced by estrogen and is most highly expressed in the myometrium (39). In contrast to the well-documented roles of estrogen receptors (ERs) and progesterone receptors (PRs), the contribution of AR to uterine physiology is poorly defined. Testosterone or synthetic derivatives have been shown to alter uterine weight and morphology (40, 41, 42, 43), and antiandrogens interfere with the proliferative effects of both androgens and estrogens (39, 41). In summary, AR is expressed in the uterus and manipulation of its activity produces physiological responses, but the mechanisms by which AR functions are unknown.

    To address this issue, we studied the effects of AR ligands on organ growth, morphology, and gene expression in mature ovariectomized (OVX) rats. We find that androgens increase uterine weight to the levels observed in intact rats. Gene expression studies show that DHT rapidly alters the expression of many genes, most of which are also regulated in a similar manner by estrogen. Thus, in the rat uterus, AR activation partially induces an estrogenic gene expression response that reconstitutes the trophic environment required for stromal and epithelial growth and differentiation.

    Materials and Methods

    Animals

    All animals were maintained in accordance with institutional Animal Care and Use guidelines and were individually housed with ad libitum access to food and water. Mice (Taconic, Germantown, NY) were obtained after OVX at 21 d and were treated at 7 wk of age for 3 wk with the indicated doses of 17?-estradiol (E2) or DHT from pellets (Innovative Research of America, Sarasota, FL). For the mibolerone experiments, Sprague Dawley female rats (Taconic) were ovariectomized at age 3 months and treated at age 7 months. Animals were randomized into eight groups (n = 10), with four groups receiving dosing three times weekly with E2 (Steraloids, Wilton, NH) in 3% benzyl alcohol in sesame oil (vehicle), and four vehicle alone for 2 wk by sc injection. Groups that received E2 during this phase continued to receive E2 throughout the remainder of the experiment. Three vehicle and three E2 groups were treated five times weekly for an additional 4 wk with sc injections of different doses of mibolerone 5 d/wk. For DHT, bicalutamide, and cyproterone acetate experiments (Steraloids), 7-month-old, 3 month post-OVX animals were treated by daily sc injection with propylene glycol as the vehicle. At necropsy, uteri were dissected at the level of the cervix, and either weighed wet and fixed in 4% paraformaldehyde in PBS for histology (n = 5) or frozen in liquid nitrogen for RNA analysis (n = 5). Tissue for histology was dehydrated and processed through paraffin. Four-micron sections of the right uterine horn, taken 1 cm above the uterine bifurcation were stained with hematoxylin and eosin. Histomorphometric analysis used the Bioquant semiautomated image analysis system (Bioquant Nova, version 4, R&M Biometrics, Nashville, TN). Data were analyzed by Fisher’s protected least significant difference (Statview, version 5.0).

    Uterine RNA and Northern blot analysis

    Total RNA was extracted using the TRIzol reagent (Invitrogen Life Technologies, Grand Island, NY). Ten micrograms of total RNA were separated by electrophoresis in denaturing agarose (1.8% formaldehyde and 1.2% agarose) and transferred by capillary blotting to a nylon membrane and cross-linked by UV irradiation. The blots were probed overnight with 50 ng cDNA fragment randomly labeled with 32P dCTP. The membranes were washed successively with 2x saline sodium citrate and 0.1% sodium dodecyl sulfate at room temperature for 20 min and twice with 1x saline sodium citrate and 01% sodium dodecyl sulfate at 65 C for 15 min and then exposed to X-OMAT film (Kodak, Rochester, NY).

    RT-PCR analysis

    Quantitative RT-PCR was performed using the two-step Taqman Gold RT-PCR kit and the Prism 7700 sequence detection system (Applied Biosystems, Foster City, CA) using 0.5 U of Amplitaq Gold polymerase. Then 75-ng aliquots of RNA were reverse transcribed and amplified using the following primers and Taqman probes: uterocalin, 5'-CCAGGGCAGGTGGTTCGTT-3', 5'-GGTAAAGCGGCTTTGTCTTTCTTT-3', and probe 5'-FAM-CTGGCAGCGAATGCGGTCCA-TAMRA-3'; IGF-I, 5'-CACAGCTCCGGAAGCAACA-3', 5'-CCATGGCTCCAGCATTCG-3', and probe 5'-FAM-CAATGCCCGTCTGTGGTGCCC-TAMRA-3'; oxytocin receptor, 5'-CATCACCTTCCGCTTCTATGG-3', 5'-AGGTAGGTGGAGGCGAACATG-3', and probe 5'-FAM-CCCGACCTGCTGTGTCGTCTGGT-TAMRA-3'; Nor-1, 5'-TGAAGGAAGTTGTGCGTACAGATAG-3', 5'-TCATCATACAGATCGGAGGAGATG-3', and probe 5'-FAM-CGTCTGCCTTCCAAACCAAAGAGCC-TAMRA-3'; and reverb-, 5'-CCACAGTGCTGGGAGAGTCA-3', 5'-GCTGAGCCGCCAAAAGGT-3', and probe 5'-FAM-ATGCCCGTCTACCCCCAACGACAA-TAMRA-3'. Other primer-probe sets were previously described (44). Data were normalized to glyceraldehyde-3-phosphate dehydrogenase or 18S. Average Ct values from duplicate PCR were normalized to average cycle threshold (Ct) values for glyceraldehyde-3-phosphate dehydrogenase from the same cDNA preparations (primers and probes from Applied Biosystems). The ratio of expression of each gene in experimental vs. control samples was calculated as 2–(mean Ct).

    Affymetrix microarray analysis

    Nine animals were used per group. RNA was collected from whole uteri as above and treated with DNase I as directed by the manufacturer (Genhunter, Nashville, TN), RNA was precipitated and concentration was determined by OD260. Three distinct pools of RNA, each from three animals, were generated by mixing equal amounts of RNA from three individual animals. These three RNA pools were each used to probe a single microarray chip, providing three nonredundant replicates per treatment. Hybridizations were performed as directed by Affymetrix (Santa Clara, CA) as described (44). Briefly, cDNA was synthesized from total RNA by Superscript II (Invitrogen Life Technologies, Carlsbad, CA) primed with oligo-dT coupled to the T7 promoter. Double-stranded cDNA was captured using Qiaquick columns (Qiagen, Santa Clarita, CA) and used as the template for in vitro transcription by T7 RNA polymerase (MEGAscript T7 kit, Ambion, Austin, TX) in the presence of biotin-uridine 5-triphosphate and biotin-CTP. These biotin-cRNA transcripts were collected on RNeasy columns (Qiagen) and then hydrolyzed to 50- to 100-nt fragments by 10 mM MgCl2 at 95 C for 35 min. Ten-microgram aliquots of these probes were hybridized at 45 C for 16 h to the RGU34A arrays (Affymetrix). After washing, signals were visualized with streptavidin-R-phycoerythrin using a fluorescence scanner (Molecular Dynamics, Sunnyvale, CA).

    Microarray analysis

    Pilot experiments confirmed the sequence of several microarray probes by mass spectrometry and verified their specificity toward their intended transcript by BLAST search. Protocols were optimized such that a high correlation (r2 = 0.91) between microarray and quantitative real-time RT-PCR (QRT-PCR) data were achieved (data not shown). Data were normalized to achieve identical median fluorescence intensity of each array. Then the normalized median fluorescent intensity for each probe set (probe sets consist of multiple oligonucleotides specific for a sequence) was compared between experiments using ANOVA. For a transcript to be considered differentially expressed, the hybridization signals must have exceeded the experimentally determined background value and then must have tested different from vehicle control (P < 0.01) by either DHT or E2 with an absolute difference of more than 1.5-fold. Accession numbers of probes were then used to search GenBank to confirm the identity of the gene. To allow for functional interpretation of the data, probe sets directed toward sequences that were not 90% homologous or greater to a known full-length gene were excluded.

    Functional clustering

    Each gene was analyzed using the primary literature at National Center for Biotechnology Information PubMed (www.ncbi.nlm.nih/gov/pubmed) to determine each gene product’s commonly accepted or best-known function. Genes were then subjectively assigned to various biochemical, metabolic, or cellular pathway-based analyses. Each functional category was then assigned to a general category (e.g. oxidative phosphorylation is included in general metabolism). The behavior of genes in each functional category was summarized by determining the percent of genes up-regulated by E2, the number regulated by E2 or DHT, and the percent of genes in the group that were regulated in the same direction by both treatments. Prostate data and analysis was previously described (44).

    Results

    AR activation stimulates uterine growth

    To determine the effects of AR activation on uterine physiology, we administered nonaromatizable ligands with high affinity for AR (mibolerone or DHT) to rats that had been rendered estrogen and androgen deficient by OVX after sexual maturation. OVX rats treated for 28 d with the AR agonist mibolerone showed significant gains in uterine weight with as little at 0.03 mg/kg five times weekly (Fig. 1A). The effects of mibolerone are dose dependent, and at the highest dose tested (0.3 mg/kg), induced uterine weight gains equivalent to that seen in sham-operated or estrogen-treated animals. In a second arm, animals were pretreated with E2 to restore the uterus and then given E2 plus mibolerone to assess antagonistic or synergistic effects (E2 was given by injection rather than using intact animals producing endogenous sex steroids to better control exposure and avoid secondary effects due to gonadotropin repression). Mibolerone plus E2 did not produce additive effects when coadministered (Fig. 1A). Similar uterotrophic effects are observed with the natural high-affinity AR ligand DHT, which causes gradual gains in uterine weight reaching a plateau after approximately 21 d of treatment (Fig. 1B). Cyproterone acetate, a steroidal antiandrogen, produces no change in uterine weight even after 24 d of treatment with 30 mg/kg·d, indicating that AR activation is required. Finally, the short-term effects of E2 and the endogenous AR ligand DHT were compared and reveal that within 24 h both treatments increase uterine weight, with E2 showing greater efficacy. Thus, AR-selective agonists induce uterine weight increases in OVX rats.

    FIG. 1. Effect of androgens on uterine weight. A, Rats (n = 10/group) ovariectomized at 3 months of age (OVX) or sham-operated (sham) were treated with the indicated dose of the AR agonist mibolerone (mib; numbers represent milligram per kilogram body weight per injection) or pretreated with E2 for 2 wk and continued to receive E2 for the 4 wk of mibolerone treatment. Intact uteri were collected and weighed after 42 d. Shown are mean wet weights ± SD. *, Different from OVX + vehicle (P < 0.05). B, Time course of androgen-induced uterine weight gains. OVX rats (n = 6/group) similar to the ones used in A were treated for the indicated number of days with 3 mg/kg·d DHT or vehicle or for 24 d with the AR antagonist cyproterone acetate (Cyp) at 30 mg/kg·d. Shown are mean wet weights ± SD. *, Different from OVX + vehicle (P < 0.05), ns, not significantly different from vehicle control. C, Wet weight of the samples used for microarray analysis (n = 9) ± SD. *, Different from OVX + vehicle (P < 0.05). Seven-month-old female rats 4 months after OVX (n = 9/group) were treated with a single injection of 3 mg/kg DHT, 0.025 mg/kg E2, or vehicle and uteri collected 24 h later.

    Cellular basis of androgenic uterine weight gains

    To determine the specific effects of AR activation on uterine morphology, the uteri treated with mibolerone and/or E2 in the 42-d treatment experiment (Fig. 1) were collected for histomorphometric analyses. Mibolerone-induced dose-dependent increases in total uterine area equal to that of a fully effective dose of E2 and mirroring the changes in uterine weight (Fig. 2A). Furthermore, mibolerone produced significant increases in endometrial and myometrial cross-sectional area, indicating that AR activation is trophic in both major compartments (Fig. 2, B and D). Essentially identical results were obtained with DHT (data not shown).

    FIG. 2. Effects of androgens and estrogens on the morphology of the uterus. Rats (n = 10) ovariectomized at 14 wk of age (OVX) or sham-operated (sham) were treated by daily sc injection with the indicated dose of the AR agonist mibolerone (mib; numbers represent micrograms per kilogram body weight) or pretreated with E2 for 2 wk and continued to receive E2 for the 4 wk of mibolerone treatment. Cross-sections through the uterine horn were analyzed by histomorphometry by a blinded observed. All values represent mean ± SD. A, Average cross-sectional area (upper left). B, Endometrial area (upper right). C, Epithelial cell height (lower left). D, Myometrial area (lower right) of all treatment groups.

    These data were examined more closely to determine the specific effects of androgens and their interactions with estrogens. In endometrial glands, mibolerone significantly increased the thickness of the epithelial layer by 23.3%, compared with vehicle control (Fig. 2C), but the cells remained relatively cuboidal (Fig. 3). E2 increased epithelial height by 2-fold, with the cells forming a pseudostratified layer (Fig. 3). This effect of E2 is blunted by even the lowest dose of mibolerone tested (Fig. 2C); the cells remain cuboidal and are similar in thickness to epithelial cells of animals treated with mibolerone alone (Fig. 3). Thus, androgen signaling inhibits a full estrogenic response in the endometrial epithelial cells. Similar antiestrogenic effects of androgens on endometrial cells have been reported in the porcine uterus (45).

    FIG. 3. Structure of the androgen-treated uterus. All photomicrographs are of cross-sections of the right uterine horn, 1 cm above the uterine bifurcation. Tissue was stained with hematoxylin and eosin. Photomicrographs in each horizontal row are representative of a treatment group: A–C, vehicle; D–F, 0.3 mg/kg·d mibolerone; G–I, 0.01 mg/kg·d E2; J–L, 0.01 mg/kg·d E2 and 0.3 mg/kg·d mibolerone. The first column (A, D, G, and J) are low-magnification images of the uterine horn cross-section to show the global size changes with treatment. Photomicrographs in the second column (B, E, H, and K) are higher-magnification views of the endothelial cell layer (bracket) and stroma (S) of the endometrium. Micrographs in the third column show the effect of treatment on the thickness of the myometrial layers (M and bracket) of the uterus.

    In contrast, E2 plus mibolerone effects in the myometrium were additive because myometrial area was significantly greater when the lowest dose of mibolerone was combined with E2 than when either of these treatments was given alone (Fig. 2B). The endometrial stroma appears less compact with increasing mibolerone dose, whereas the stromal matrix appears dense with E2 alone (Fig. 3). The appearance of the stroma in the mibolerone-treated samples is consistent with edema and/or perhaps unique effects on cell-matrix interactions (see below). E2 appears to partially reduce the edematous appearance of the endometrial stroma seen with mibolerone alone (Fig. 3). Thus, unlike with epithelial cells in which mibolerone antagonized the effect of E2, E2 did not antagonize the effects of mibolerone on myometrial area and only partially ameliorated the effects on cellular density. These data show that AR activation leads to gains in uterine weight due to stimulation of both the myometrium and endometrium and suggest that androgens and estrogens have distinct effects in the rat uterus.

    Comparative genomics analysis of uterine hypertrophy

    The organ weight and histology data imply that AR promotes the expression of genes that control the trophic environment and architecture of the uterus. Because both AR and ER predominantly function as transcription factors regulating the expression of target genes in hormone-sensitive cells, large-scale gene expression experiments provide an ideal means to compare the functional consequences of their activation. Thus, the global gene expression changes produced by E2 and DHT were assessed using oligonucleotide microarrays. OVX rats were treated with a single dose of vehicle, E2, or DHT, and uteri were collected 24 h later; 24 h was chosen for comparison because most genes directly regulated by AR or ER should be induced within this time frame, whereas secondary effects on gene expression resulting from dynamic changes in cell proportion or other factors would be minimized. For each treatment group, three independent groups of animals were used to generate samples, allowing for statistical analyses of the data. RNA prepared from each of the nine samples (three controls, three E2-treated, and three DHT-treated) was hybridized to a 7000-feature oligonucleotide microarray.

    Transcripts were selected for further analyses if the following criteria were met. First, the median hybridization intensity had to be above background, differ from vehicle significantly (P < 0.01) in either the DHT- or E2-treated samples, and be at least 1.5-fold different from control. Second, each gene product must encode for a known protein for which functional information is available. This latter requirement excludes uncharacterized genes and expressed sequence tags but allows the microarray data to be interpreted in the context of uterine physiology by functional clustering. A total of 503 transcripts met these criteria and are listed in Supplemental Table 1, published on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org.

    Confirmation of microarray data

    Several safeguards are employed to ensure the overall validity of the gene expression data as described in Materials and Methods. To assess the reproducibility of these data, multiple genes were investigated in separate experiments by two independent techniques. First, using QRT-PCR, seven genes were examined in a replicate experiment and performed as in the microarray, providing 14 replication data points (Fig. 4A). All selected genes regulated significantly in the first experiment were also significantly regulated in the same direction in the replicate experiment. Importantly, with the exception of uterocalin, nonsignificant differences from the microarray did not repeat in the second experiment (uterocalin P = 0.018 in the microarray, and thus was greater than 0.01 and considered nonsignificant). Two genes were then chosen for further validation by Northern blot: IGF-I and uterocalin/lipocalin-2 (Fig 4B). These genes were chosen because IGF-I codes for a stromal-derived factor that plays a major role in the regulation of uterine growth (46, 47), whereas uterocalin encodes a secreted protein synthesized by uterine epithelial cells (48) and thus represents gene activation in that compartment. Northern blot data confirm that both IGF-I and uterocalin transcripts were significantly induced by DHT. Thus, the microarray experiment successfully identified authentic changes in gene expression in the uterus. Because every gene cannot be confirmed independently, it remains possible that some regulations identified by microarray analysis were produced by type I error (estimated to be fewer than five genes with the P < 0.01 criteria used here) or cross-reactivity of the probes.

    FIG. 4. Validation of microarray results. A, Gene expression values in a replication experiment using independent animals for seven transcripts scored as regulated in the microarray (MA) experiment. DHT-MA and E2-MA values are the expression of each gene relative to vehicle in the microarray experiment for DHT and E2, respectively, whereas DHT-QRTPCR and E2-QRTPCR are the values measured in the replication experiment by QRT-PCR. Asterisks indicate no difference from vehicle (P > 0.05). B, Northern blot analysis of IGF-I and uterocalin/lipocalin-2 RNA levels in total RNA prepared from the uteri of animals treated for 24 h with DHT or E2. Bottom panel is 28S and 18S RNA as a loading control.

    DHT-responsive genes are largely a subset of E2-responsive genes

    We then compared the effects of AR and ER activation on gene expression on the uterus. E2 produces profound changes in the gene expression profile of uterine tissue, with nearly all of the 503 regulated transcripts responding. Interestingly, the majority of transcripts significantly regulated by DHT are included within this set of estrogen-responsive genes. Of the transcripts regulated by DHT, 85.9% were altered in qualitatively the same manner by E2 (Fig. 5A). Only 12 of the 503 genes were regulated by DHT but not E2, and only 11 showed induction by one hormone but repression by the other. Scatter plots comparing the magnitude of gene expression responses reveals that the DHT and E2 gene signatures are significantly correlated (Fig. 5B, r = 0.57, P < 0.001). However the DHT response tended to be quantitatively less than the E2 response, as seen by the deviance of the best-fit correlation line from unity at 45 degrees, which would represent identical magnitudes of regulation by both treatments. On average, genes responded to DHT treatment 26.4% as much as when the animals were given E2, and few transcripts were altered by DHT to the same extent as with E2 (Fig. 5C). Thus, AR induces a gene expression response that is largely a subset of the estrogenic program but typically does not evoke the same magnitude of change in transcript abundance.

    FIG. 5. Comparison of AR- and ER-mediated gene expression programs in the uterus. A, Venn diagram showing the distribution of transcripts expressed at significantly (P < 0.01) different levels from vehicle controls 24 h after a single dose of 3 mg/kg DHT or 0.1 mg/kg E2 (n = 503 genes). B, Scatter plot analysis of magnitude of gene regulation by E2 (y-axis) vs. DHT (x-axis). Each point is a unique gene transcript. Solid line is best-fit regression, dashed line is 45 degrees. C, Histogram showing the number of transcripts regulated by DHT subdivided by the magnitude of change relative to E2 (in which the effect of E2 is set for each transcript’s relative abundance to 100%).

    Comparative analysis of DHT and E2

    We then sought to determine the specific similarities and differences between estrogens and androgens. To do so, the 503 genes regulated by E2 or DHT were then subjected to functional clustering (44). In this method, genes are scrutinized individually for their most probable function in the context of uterine physiology. Primary research publications, reviews, and gene product summaries in public databases are used to generate a summary of each gene product’s best-known function. Based on this manual data-mining exercise, genes are clustered into pathways or functional classes. This analysis is in contrast to other clustering methods that rely exclusively on numerical expression data to group genes. As shown in Fig. 6, the effects of E2 on gene expression could be divided into four major groups: 1) protein synthesis, maturation, degradation, and secretion; 2) intracellular signaling and signal transduction; 3) tissue growth and remodeling; and 4) metabolism and metabolite transport. Within each group, genes are subdivided into more specific functional clusters, and the response of these subclasses of genes to DHT relative to E2 is graphically depicted. Functionally similar groups of genes varied by the percent up-regulated and the proportion of the class significantly regulated by DHT (the greater the amount of gray in the bar adjacent to each group’s name, the more similarity between E2 and DHT). Each major group is discussed briefly below.

    FIG. 6. Functional clustering of the AR- and ER-mediated gene expression programs in the uterus. The 503 genes regulated by either DHT and/or E2 after 24 h were analyzed by functional clustering. Genes with putatively related functions are grouped, and then these groups are assigned to one of four categories. For each grouping, shown is the number of genes (N), the percent that were up-regulated by E2 and/or DHT (%up, subtract this number from 100% for the percentage of genes that were down-regulated), and the scale bar to the right shows the relative proportion of transcripts in the class regulated in a qualitatively similar fashion by DHT and E2 (gray) vs. E2 alone (white).

    Protein synthesis, maturation, degradation, and secretion

    E2 induces rapid growth and secretory activity in the uterus; accordingly more than 100 genes regulated by E2 encode proteins that function in the various steps of protein metabolism (Supplemental Table 1). In terms of mRNA synthesis, 28 transcription factor-encoding genes are differentially expressed relative to controls 24 h after treatment. These include well-known targets of ER activation, such as the PR, and less well-studied genes with potentially novel roles in uterine physiology such as Deaf1/suppressin and Cited2, which were repressed by both hormones. There is nearly uniform induction of the polypeptide synthesis and maturation pathways: genes involved in translation, folding, glycosylation, and protein trafficking were generally induced by E2. Whereas DHT regulated many of these same genes similarly, albeit to a generally lesser degree, there is a notable difference in the DHT-mediated expression of genes involved in the processing of immature polypeptides (Fig. 7). For example, two genes involved in the maturation of peptide hormones were regulated by E2 but unaffected by DHT: prolyl endopeptidase, which was induced, and peptidylglycine -amidating monooxygenase, which was repressed nearly 6-fold. Given the functional significance of peptide hormones such as oxytocin (49) and adrenomedullin (50) in uterine physiology, the diminished effect of DHT on genes regulating the processing of peptide hormones could be an important difference between androgenic and estrogenic uterine stimulation.

    FIG. 7. Selected protein processing and modification genes regulated by DHT and/or E2 in the uterus. Shown are the mean (n = 3) normalized fluorescent intensity hybridization values for the RNAs of each of the listed genes normalized to vehicle (± SE) for genes whose protein products are involved in polypeptide proteolysis and maturation. Asterisk denotes P < 0.05 relative to vehicle.

    Intracellular signaling and signal transduction

    Intercellular signaling, both in the form of paracrine signaling from stromal to epithelial cells and autocrine signaling, mediates much of the growth and function of uterine cells (e.g. Refs.51 , 52). In agreement with this concept, E2 altered the expression of more than 100 genes involved in cell-cell communication and signal transduction (Supplemental Table 1). Closer examination reveals potentially important similarities and differences between DHT and E2, specifically in terms of genes involved in peptide hormone and growth factor signaling. In the peptide hormone category, RNAs for the receptors for galanin (galanin type 2 receptor), the C-type natriuretic peptide (natriuretic peptide receptor type 2), and the peptide hormones (prepronociceptin, oxytocin, and the diazepam binding inhibitor) were all regulated significantly (although by different magnitudes) by both DHT and E2. In contrast, adrenomedullin and the opioid peptide proenkephalin were regulated by DHT and E2 alone, respectively (Fig. 8A). Each of these peptide hormone signaling genes are expressed in the myometrium and are generally involved in regulating the contractility of smooth muscle cells (53, 54, 55). Thus, these gene regulations perhaps provide insight into the effects of androgens in the rodent myometrium.

    FIG. 8. Selected intercellular signaling genes regulated by DHT and/or E2 in the uterus. Shown are the mean (n = 3) normalized fluorescent intensity hybridization values for the RNAs of each of the listed genes normalized to vehicle (± SE) for genes whose protein products are involved in peptide hormone signaling (A), growth factor signaling (B), and IGF-I and wnt signaling (C). Asterisk denotes P < 0.05 relative to vehicle.

    In contrast to genes involved in peptide hormone signaling, E2-sensitive growth factors and their receptors were generally not significantly regulated by DHT (Fig. 8B). These genes include fibroblast growth factor (FGF) receptors-1 and -2b, TGF?3, and the nerve growth factor receptor TrkA. The repression of these progrowth and -differentiation genes by E2 might reflect complex kinetics; for example TGF?3 is transiently induced and then repressed relative to baseline after 6 h in mice treated with estrogen (56). The relative lack of effect of DHT on the TGF?3 and FGF receptors could reflect the reduced ability of androgens to stimulate the endometrium (Fig. 2) because these systems are thought to regulate maturation in that compartment (57). The role of nerve growth factor in the uterus has not been extensively studied, but the more than 7-fold induction of nerve growth factor receptor RNA specifically by E2 suggests that it might play an unappreciated role.

    Although DHT did not significantly alter the expression of all growth factors, it did have potentially important effects on the expression of genes involved in IGF-I and wnt signaling. As shown above, IGF-I was highly induced by both DHT and E2, as was IGF-I binding protein (IGFBP)-2 (Fig. 8C), which modulates IGF-I signaling (58). Genes for the inhibitory IGFBP-3 were significantly repressed by both hormones, thereby producing a pattern of gene expression suggesting elevated intracrine IGF-I tone. Because IGF-I is involved in the development of sex steroid-dependent secretory organs in general, this observation is likely to be important for understanding the effects of DHT in the uterus. Likewise both DHT and E2 repressed the RNAs for the wnt receptor frizzled-1 and the soluble decoy wnt receptor secreted frizzled-related protein-4. Wnt signaling, particularly by wnt-7a, is critical for uterine development (59) and appears from these and other data to be regulated by both estrogens (60, 61) and androgens (Fig. 8C).

    Tissue growth and remodeling

    E2 and DHT produce extensive changes in the morphology of the uterus (Fig. 3); and approximately 100 genes that function in tissue growth and remodeling were regulated by these hormones. Several genes associated with mitosis and apoptosis were regulated similarly by DHT and E2, suggesting that both hormones regulate cell division and survival. Three genes that control the cell cycle were regulated by both DHT and E2: retinoblastoma-1 (Rb1), Mcmd6, and Btg1 (Fig. 9A). Rb1 is an inhibitor of S-phase progression previously shown to be repressed by E2 (62), whereas Mcmd6 is a positive factor required for DNA replication (63) and was induced by both. Consistent with this prodivision pattern, BTG-1 RNA, which encodes an antiproliferative protein in the TIS21/PC3/BTG1/TOB family of cell cycle repressors (64), was also repressed by both hormones. Together, these data suggest proliferation in the uterus by androgens and estrogens might involve regulation of Rb1, Mcmd6, and Btg1.

    FIG. 9. Selected proliferation and tissue remodeling genes regulated by DHT and/or E2 in the uterus. Shown are the mean (n = 3) normalized fluorescent intensity hybridization values for the RNAs of each of the listed genes normalized to vehicle (±SE) for genes whose protein products are involved in cell cycle regulation (A) and collagen matrix synthesis (B). Asterisk denotes P < 0.05 relative to vehicle. NGF, Nerve growth factor.

    Consistent with the differing morphologies of the uterus after long-term E2 and DHT treatment, these hormones have overlapping but distinct effects on genes involved in the collagen matrix (Fig. 9B). Types I, III, and V collagen genes have previously been shown to be regulated by E2 in the immature rat uterus (65). In these adult animals, E2 induced the expression of Col1a1 and Col5a2 RNA, whereas DHT induced Col1a1, Col3a1, and both Col5a1 and Col5a2 RNAs. Collagen gene regulation by E2 was accompanied by reductions in the relative abundance of the collagen maturation cofactor lysyl oxidase and matrix protein-encoding RNAs for decorin and biglycan, whereas DHT had no effect. Interestingly, the RNA for matrix gla protein, which is expressed in uterine smooth muscle (66), was repressed by E2 but induced by DHT. These differences in expression of collagen matrix-associated genes could be important for understanding the effects of androgens on uterine architecture.

    Metabolism and metabolite transport

    Estrogens control the metabolic state of cells throughout the mammalian uterus, influencing the production of numerous classes of biomolecules (e.g. Refs.67 , 68). The microarray data provide insight into the gene expression changes that underlie metabolic activation of uterine cells by estrogens and androgens. Genes involved in the regulation of lipid metabolism, energy utilization, amino acid and nucleotide synthesis and catabolism, metabolite transport, and several other metabolic processes were differentially expressed 24 h after E2 treatment (Fig. 6). DHT affected a subset of these genes in a manner generally consistent with the effects of E2. For example, cholesterol synthesis genes, some of which have been shown to be estrogen induced (e.g. Ref.69), were uniformly up-regulated by E2 within 24 h (Fig. 6). Several of these genes were also significantly induced by DHT and all trended in the same direction as E2 (Fig. 10). These data confirm and extend the concept that estrogens rapidly modulate cholesterol metabolism in the uterus at the level of gene expression and demonstrate that DHT has a similar (albeit quantitatively lesser) effect.

    FIG. 10. Selected metabolism genes regulated by DHT and/or E2 in the uterus. Shown are the mean (n = 3) normalized fluorescent intensity hybridization values for the RNAs of each of the listed genes normalized to vehicle (±SE) for genes whose protein products are involved in cholesterol synthesis and metabolism. Asterisk denotes P < 0.05 relative to vehicle.

    Involvement of ER in androgen activity

    The above data demonstrate that within 24 h AR regulates the expression of genes that are generally also regulated by estrogens, presumably mostly via ER (70). One explanation is that DHT or AR modulates ER activity. To determine the role of ER and ER? in androgen activity, DHT and E2 were given to OVX ER and ER? homozygous knockout mice (71) for 3 wk, and uterine weight was measured. As in rats, the uteri in wild-type mice were stimulated by both hormones (Fig. 11). Ablation of ER? had no effect on either hormone. In contrast, ER–/– animals did not respond to either hormone. We then tested the possibility that AR expression is simply absent in ER–/– uteri. QRT-PCR analysis confirms that the level of AR RNA is similar among wild-type, ER–/–, and ER?–/– mice. Interestingly, in the ER–/– animals only, AR RNA was significantly induced by DHT (Fig. 11B), demonstrating that DHT was active in the uterus but was unable to promote uterine growth. Furthermore, ER might have a role in regulating androgen sensitivity by repressing AR RNA levels in response to androgens because this induction was not observed in the other strains. These data indicate ER is required for the androgen sensitivity of the murine uterus.

    FIG. 11. Effect of DHT on uterine hypertrophy in ER-deficient mice. A, Wild-type (WT), ER–/– mice (ERKO), and ER?–/– mice were ovariectomized after sexual maturation and 3 wk later treated for 3 wk with 3 mg/kg·d DHT, 100 μg/kg·d E2, or blank pellets. Uteri were collected and weighed; shown are mean wet weight ± SD. *, Different from respective control (P < 0.05). B, QRT-PCR analysis of uterine AR RNA content relative to cyclophilin RNA and normalized to wild-type vehicle value. BERKO, ER?–/– mice.

    We then reasoned that if AR modulates ER directly, then genes known to be specific direct targets of ER activity in the uterus should be activated by DHT. The PR (72, 73), calbindin D-9k (74), cellular retinoic acid binding protein II (75), C3 (76), and oxytocin (77) are induced by ER binding directly to regulatory regions of the gene. Whereas all five of these genes were significantly induced by E2 within 24 h (with calbindin D-9k and C3 more than 30-fold), none were significantly induced by DHT, and calbindin D-9k was actually significantly repressed (Fig. 12). QRT-PCR experiments 4, 7, 10, and 17 d after DHT treatment revealed that at no time point were these genes induced by DHT (data not shown). To demonstrate a role for AR directly, we treated OVX rats for 4 d with DHT with or without the nonsteroidal, highly selective AR antagonist bicalutamide. Bicalutamide completely blocked the effect of DHT on IGF-I RNA induction (Fig. 12). Thus, in the OVX rat, DHT does not lead to the induction of several direct targets of ER, and an AR antagonist blocks its effects on IGF-I.

    FIG. 12. ER target genes after DHT treatment. A–E, Microarray data showing change in relative abundance of RNAs that are known direct targets of ER activity (see text for references). E2 induced all genes (*, Different from vehicle, P < 0.05), whereas DHT had no significant effect. CRABP2, cellular retinoic acid binding protein II. F, IGF-I induction in rat uteri. OVX rats were treated for 4 d with 3 mg/kg·d DHT and/or 30 mg/kg·d bicalutamide (Bic). Shown is the abundance of IGF-I RNA relative to cyclophilin RNA and normalized to vehicle value.

    A common gene expression motif in uterus and prostate

    We then asked whether the effects of androgens in the uterus are tissue specific or rather whether the gene expression changes downstream of AR are similar to that observed in the prostate, in which DHT is the primary trophic hormone. In a previous study (44), we identified 234 genes that respond to DHT within 24 h in the ventral prostate glands of castrated male rats. In both the prostate and uterus experiments, the same dose and duration of DHT was used, and the experimental design and analysis were similar.

    When the results of the prostate and uterus gene expression studies are compared, 93 genes are significantly regulated in both tissues (Supplemental Table 2, published on The Endocrine Society’s Journals Online web site at http://endo. endojournals.org.). Twenty-eight of the 165 genes regulated by DHT in the uterus were also significantly regulated at the same time point in the prostate. In contrast, almost all genes (90 of the 93) regulated in both the prostate and uterus were regulated by E2 in uterus and by DHT in prostate. Regression analysis shows more similarity between the effects of E2 in the uterus (as indicated by proximity of regression line to the 45-degree unity line) to DHT in the prostate than there is between DHT in the two organs (Fig. 13A). Analysis of the common genes shows that many function in protein synthesis, metabolism, and degradation, whereas a subset encode growth and differentiation regulators (Fig. 13B). The latter category includes components of the IGF-I signaling system (Fig. 13C). In the prostate, as in the uterus, IGF-I is a key component of the stromal-epithelial interaction and a regulator of cell growth and survival (78). Also included are regulators of transcription and the cell cycle [Rb1, early growth response protein-1 (egr-1), c-jun, and inhibitor of differentiation-2 (Id-2), Fig. 13D]. Examination of the reported functions of these genes suggests that they have important roles in the early stages of growth in both tissues. Rb1 is a cell cycle inhibitor and tumor suppressor, whereas egr-1 regulates apoptosis and survival and is transiently induced by estrogen in uterine cells entering the cell cycle (79). c-jun is a component of the activator protein-1 complex implicated in many processes in both tissues, whereas Id-2 encodes a basic helix-loop-helix inhibitory protein that functions in differentiation of breast epithelial cells (80). Finally, there is very similar regulation of developmental patterning genes, notably jagged-1, the wnt decoy receptor secreted frizzled-related protein-4, and the homeobox (Hox) gene and p53 regulator Hox A5/Hox 1.3 (81) (Fig. 13E). Jagged-1 protein is a ligand for the notch family of receptors, which are involved in uterine development in Caenorhabditis elegans (82) and also appear to play a role in prostate growth (44). Hox 1.3 is involved in many developmental processes, including epithelial-mesenchymal patterning in the gut (83), and is involved in breast cancer tumorigenesis (81). In summary, these observations suggest that there is a group of genes generally involved in growth and differentiation in sex steroid-dependent tissues that respond to proliferative or survival signals in general rather than to the specific hormone. DHT regulates a subset of these genes to a limited extent in the uterus.

    FIG. 13. Identification of a shared gene expression set in prostate and uterus. A, Scatter plots comparing gene expression changes in the uterus and prostate of the 93 genes that were regulated by DHT in the prostate (44 ) and uterus by either DHT and/or E2 after 24 h treatment in castrated or OVX rats, respectively. Dashed line represents 45 degrees (equal values on both axes). B, Functional categorization of the genes regulated in both prostate and uterus at 24 h. Right panels, Mean (n = 3) normalized fluorescent intensities, relative to vehicle ± SD, from the microarray experiments comparing the effects of DHT and E2 on gene expression in the uterus and prostate. C, IGF-I signaling genes, IGFBP3, and IGF-I receptor (GFIR). D, Growth and differentiation regulators Rb1, egr-1, c-jun, and Id-2. E, Developmental pathway genes jagged-1, secreted frizzled-related protein 4 (Sfrp4), and Hox A5/Hox 1.3.

    Discussion

    Androgens are important regulators of somatic growth, sexual differentiation, and higher cognitive and behavioral functions in males, but their role in females is less clear. However, androgens are present in females throughout adult life (1, 2, 3) and AR is expressed in many tissues, including those in the female reproductive tract (28, 29). Here we sought to define the role of AR in the uterus.

    The data presented here demonstrate that androgens promote the growth and differentiation of the rodent uterus in a manner similar to, yet distinct from, estrogens. This interpretation is based on three observations. The first is that AR ligands potently induce gains in uterine weight in adult OVX rats and can fully restore the uterus to the size seen in hormone-replete animals (Fig. 1, A and B). Androgens, like estrogens, have rapid effects, with weight gains significant after 24 h (Fig. 1C) but unlike estrogens may take several weeks to reach full efficacy (Fig. 1B). The antiandrogen cyproterone acetate does not produce any change in weight, showing that AR activation and not simply binding to the receptor is required. Given these data and that antiandrogens inhibit the effect of estrogen on proliferation in rats (39) and nonhuman primates (84), it will be important to determine what role endogenous androgens have in uterine trophism in vivo. Recently AR knockout mice have been described (85, 86), and these were developed to permit tissue-selective gene deletion. It will be interesting to determine the effects of AR deletion specifically in the uterus.

    Second, we examined the effects of androgens using histomorphometric analyses. Mibolerone produced a total uterine area no different from that observed with E2 alone or in intact animals (Fig. 2). However, androgens differentially affect the two major compartments of the uterus. Androgens stimulate the rodent myometrium to a greater extent than observed with E2, whereas the endometrial cells were less responsive (Fig. 2). Interestingly, the effects of AR on the myometrium were not blocked by E2, whereas mibolerone antagonized the full restoration of the epithelial cell height by estrogen. Thus, although broadly similar in their abilities to stimulate uterine growth, important differences between the two hormones are evident. These observations suggest specific effects of androgens and argue against an interpretation of the data that rely on mibolerone or DHT simply mimicking estrogen. Histologically the myometrium appears different after mibolerone treatment than it does with E2 treatment, with extended interstitial space, suggesting either some degree of edema and/or altered formation or adherence to extracellular matrix components (Fig. 3). The cellular basis of this finding, once uncovered, might provide important additional clues to the specific function of AR.

    Next we compared the 24-h effects of DHT and E2 on gene expression. E2, mostly via the ER (70), is a powerful regulator of gene expression in the rodent uterus, and even with relatively strict selection criteria, approximately 500 transcripts showed differential expression in E2-treated animals. DHT was also active, causing altered expression of 164 RNAs within 24 h. Because these data were shown to be generally reliable (Fig. 4), they provide an overview of the effects of androgens and estrogen at a molecular level. Furthermore, comparison of the effects of DHT and E2 reveal that there is much similarity. Of the 164 genes whose expression was affected by DHT, 86% of these were regulated in a similar fashion by E2 (although the effects of E2 were typically more pronounced) (Fig. 5). Using functional clustering (44), each gene was grouped into categories or physiological pathways based on published information. This analysis allows for a comparison between the consequences of AR and ER activation at the level of specific genetic pathways.

    As shown in Fig. 6, DHT effects relative to E2 varied by the presumed function of the genes’ products. In some categories, such as the various steps of protein synthesis, regulators of the cell cycle, or cholesterol metabolism, the effects of DHT and E2 are quite similar and differ mostly by magnitude of effect. Thus, it seems likely that androgens influence these systems in a manner similar to estrogens. Likewise, the data suggest the possibility that androgens are similar to estrogens in that they both act as positive effectors of the stromal-epithelial interaction. The stromal-epithelial interaction is a complex trophic relationship between mesenchymal cells, which produce paracrine factors such as IGF-I, and epithelial cells, which require these factors for survival, growth, and differentiation. Estrogens are trophic in the uterus, in part due to stimulation of these factors, particularly IGF-I (52). DHT also elevates the IGF-I RNA content of the rat uterus severalfold (Fig. 4), and both androgens and estrogens regulate multiple genes in the IGF-I signaling system (Fig. 8C). Thus, a major role of androgens in the rodent uterus could be to induce IGF-I expression and responsiveness, which could perhaps lead indirectly to modulation of estrogen signaling. Signaling by the wnt family of secreted proteins, particularly wnt-7a (59), governs multiple aspects of uterine physiology, and the data here reveal that both AR and ER alter the expression of genes involved in this signal transduction cascade (Fig. 8C). Other potentially important similarities include regulated expression of several peptide hormones or their receptors, such as oxytocin and prepronociceptin. Interestingly, most of these peptide hormone systems are thought to be expressed within the myometrium (e.g. Refs.87, 88, 89), the region of the uterus most affected by DHT (Fig. 2). Oxytocin critically regulates many aspects of uterine myometrial physiology (90); the local induction of its RNA by androgens has not to our knowledge been described. Finally, the induction of certain secreted molecules such as uterocalin/lipocalin-2, which is synthesized specifically in the epithelial cells in the uterus (91), suggests that AR, like ER, induces secretory activity in these cells.

    Other findings highlight differences between the hormones and hint at specific roles for androgens, particularly in the endometrial compartment. One example is adrenomedullin, a peptide hormone that is expressed in the rodent and primate endometrium (88, 92, 93) and whose RNA was suppressed by DHT but unaffected by E2 (Fig. 8A). Adrenomedullin promotes angiogenesis in the endometrium (93) and inhibits spontaneous contractions (94), and thus its regulation by androgens could be functionally important. Other genes that were differentially expressed include the one encoding the calcium binding protein, calbindin D-9k, which was induced over 30-fold by E2 but repressed 2.8-fold by DHT (Supplemental Table 1). Finally, the lack of androgenic regulation of a number of growth factor signaling genes that act in the endometrium (Fig. 8B) further suggests mechanistic bases for the distinct effects of androgens in the endometrium. These findings could shed light on the mechanisms by which androgens are beneficial in the management of endometriosis (94). Furthermore, they suggest that androgens might act to limit normal estrogenic proliferation in vivo, as has been suggested in the porcine uterus (45).

    It will be important to unravel the molecular mechanisms by which AR promotes the expression of genes that are largely estrogen responsive. Experiments in ER–/– mice (Fig. 11) indicate that ER activity is required for DHT-mediated trophism. We interpret these data to mean either that DHT treatment activates ER (directly or secondarily through growth factor signaling) or that ER is required during uterine development to promote androgen sensitivity. AR RNA itself appears unaffected by the absence of ER, and DHT induced the expression of AR in ER–/– uteri, suggesting that the receptor is present (Fig. 11B). Furthermore, five validated ER direct target genes were induced by E2 but not AR, whereas an AR antagonist blocks the induction of IGF-I (Fig. 12). These data agree with previous studies showing that both ER and AR antagonists blunt the trophic effects of androgens (39, 41). It will be important to unravel the mechanism by which ER participates in androgenic activity in uterine cells, which may be complex because ER can potentially be activated by growth factor signaling cascades.

    One possibility is that the similarity between DHT and E2 in the uterus reflects the fact that both hormones are trophic in this organ, and that most of the similarly regulated genes reflect the myriad of changes that occur during the initiation of growth by any mechanism. If so, then one might expect significant overlap between gene expression changes in the uterus and prostate, in which DHT is the dominant trophic hormone. When expression data were compared between these organs as they undergo rapid growth, there was some similarity in response, with 28 genes being regulated by DHT in both organs (Fig. 13) These include known direct targets of AR such as ornithine decarboxylase (Ref.95 and Supplemental Table 2). More striking, however, was the similarity between E2 effects in the uterus and DHT effects in the prostate. Of the 93 genes common between uterus and prostate, 90 were regulated by both E2 in the uterus and DHT in the prostate. This group of genes includes many that are likely downstream effectors of tissue growth and metabolic activation, including those involved in protein synthesis and cholesterol metabolism. We suggest that these genes are likely induced in a general manner in tissues undergoing growth and remodeling. However, the similarity between key regulators of prostate and uterine physiology bears further consideration. RNAs for proteins such as the activator protein-1 component c-jun, the cell cycle inhibitor and tumor suppressor Rb1, the notch ligand Jagged1, and other modulators of growth and differentiation were regulated similarly or nearly identically by DHT in the prostate and E2 in the uterus (Fig. 13). These data suggest that ER in the uterus and AR in the prostate elicit mostly tissue-specific effects because the majority of genes regulated within 24 h by DHT in the prostate were unaffected by E2 in the uterus and vice versa. In addition to these specific effects, we propose that a subset of genes respond generally to proliferative signals during the growth and development of sex steroid-dependent secretory organs. Because both ER and AR exert many of their effects via common growth factors such as IGF-I and epidermal growth factor (96, 97), perhaps this putative common gene program results from signal transduction downstream from key signaling molecules rather than direct activation of transcription by steroid hormone receptors.

    In summary, we conclude that AR functions in the rat uterus as a modulator of both myometrial and endometrial growth. It achieves this effect by inducing a gene expression program that shares significant overlap with that produced by ER. It is not clear to what extent the data from rodents can be generalized to humans. Because AR is present in the primate endometrium in muscle, stromal, and epithelial cells and varies with the reproductive cycle (35, 98, 99), it is likely to perform some function there. To our knowledge there are no reports indicating uterine hypertrophy in postmenopausal women receiving androgens. Thus, it remains an important challenge to determine the function of uterine AR in women.

    Acknowledgments

    We thank Greg Seedor and Christine Weiss for technical assistance.

    References

    Burger HG 2002 Androgen production in women. Fertil Steril 77(Suppl 4):S3–S5

    Rittmaster RS 1995 Clinical relevance of testosterone and dihydrotestosterone metabolism in women. Am J Med 98:17S–21S

    Burger HG 1996 The endocrinology of the menopause. Maturitas 23:129–136

    Abraham GE 1974 Ovarian and adrenal contribution to peripheral androgens during the menstrual cycle. J Clin Endocrinol Metab 39:340–346

    Rao AJ, Kotagi SG 1982 Serum testosterone levels during the menstrual cycle and early pregnancy in the bonnet monkey (Macaca radiata). Endocrinol Jpn 29:271–275

    Epstein MT, McNeilly AS, Murray MA, Hockaday TD 1975 Plasma testosterone and prolactin in the menstrual cycle. Clin Endocrinol (Oxf) 4:531–535

    Guerrero R, Aso T, Brenner PF, Cekan Z, Landgren BM, Hagenfeldt K, Diczfalusy E 1976 Studies on the pattern of circulating steroids in the normal menstrual cycle. I. Simultaneous assays of progesterone, pregnenolone, dehydroepiandrosterone, testosterone, dihydrotestosterone, androstenedione, oestradiol and oestrone. Acta Endocrinol (Copenh) 81:133–149

    Vermeulen A, Verdonck L 1976 Plasma androgen levels during the menstrual cycle. Am J Obstet Gynecol 125:491–494

    Stahl F, Dorner G, Rohde W, Schott G 1976 Total and free testosterone and total and free 17?-oestradiol in normally menstruating women. Endokrinologie 68:112–114

    Apter D, Viinikka L, Vihko R 1978 Hormonal pattern of adolescent menstrual cycles. J Clin Endocrinol Metab 47:944–954

    Smith KD, Rodriguez-Rigau LJ, Tcholakian RK, Steinberger E 1979 The relation between plasma testosterone levels and the lengths of phases of the menstrual cycle. Fertil Steril 32:403–407

    Longcope C, Franz C, Morello C, Baker R, Johnston Jr CC 1986 Steroid and gonadotropin levels in women during the peri-menopausal years. Maturitas 8:189–196

    Bancroft J, Cawood EH 1996 Androgens and the menopause; a study of 40–60-year-old women. Clin Endocrinol (Oxf) 45:577–587

    Burd ID, Bachmann GA 2001 Androgen replacement in menopause. Curr Womens Health Rep 1:202–205

    Braunstein GD 2002 Androgen insufficiency in women: summary of critical issues. Fertil Steril 77(Suppl 4):S94–S99

    Rohr UD 2002 The impact of testosterone imbalance on depression and women’s health. Maturitas 41(Suppl 1):S25–S46

    Riggs BL, Khosla S, Melton 3rd LJ 2002 Sex steroids and the construction and conservation of the adult skeleton. Endocr Rev 23:279–302

    Guay A, Davis SR 2002 Testosterone insufficiency in women: fact or fiction? World J Urol 20:106–110

    Greene RA, Dixon W 2002 The role of reproductive hormones in maintaining cognition. Obstet Gynecol Clin North Am 29:437–453

    Khosla S, Bilezikian JP 2003 The role of estrogens in men and androgens in women. Endocrinol Metab Clin North Am 32:195–218

    Redmond GP 1998 Androgens and women’s health. Int J Fertil Womens Med 43:91–97

    Carmina E 2002 Anti-androgens for the treatment of hirsutism. Expert Opin Investig Drugs 11:357–363

    Schneider HP 2003 Androgens and antiandrogens. Ann NY Acad Sci 997:292–306

    Raudrant D, Rabe T 2003 Progestogens with antiandrogenic properties. Drugs 63:463–492

    Lee HJ, Chang C 2003 Recent advances in androgen receptor action. Cell Mol Life Sci 60:1613–1622

    Abate-Shen C, Shen MM 2000 Molecular genetics of prostate cancer. Genes Dev 14:2410–2434[Free Full Text]

    Huang H, Tindall DJ 2002 The role of the androgen receptor in prostate cancer. Crit Rev Eukaryot Gene Expr 12:193–207

    Horie K, Takakura K, Fujiwara H, Suginami H, Liao S, Mori T 1992 Immunohistochemical localization of androgen receptor in the human ovary throughout the menstrual cycle in relation to oestrogen and progesterone receptor expression. Hum Reprod 7:184–190

    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

    Dimitrakakis C, Zhou J, Wang J, Belanger A, LaBrie F, Cheng C, Powell D, Bondy C 2003 A physiologic role for testosterone in limiting estrogenic stimulation of the breast. Menopause 10:292–298

    Murray AA, Gosden RG, Allison V, Spears N 1998 Effect of androgens on the development of mouse follicles growing in vitro. J Reprod Fertil 113:27–33

    Moinfar F, Okcu M, Tsybrovskyy O, Regitnig P, Lax SF, Weybora W, Ratschek M, Tavassoli FA, Denk H 2003 Androgen receptors frequently are expressed in breast carcinomas: potential relevance to new therapeutic strategies. Cancer 98:703–711

    Brys M 2000 Androgens and androgen receptor: do they play a role in breast cancer? Med Sci Monit 6:433–438

    Wang PH, Chang C 2004 Androgens and ovarian cancers. Eur J Gynaecol Oncol 25:157–163

    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

    Mertens HJ, Heineman MJ, Theunissen PH, de Jong FH, Evers JL 2001 Androgen, estrogen and progesterone receptor expression in the human uterus during the menstrual cycle. Eur J Obstet Gynecol Reprod Biol 98:58–65

    Ito K, Suzuki T, Akahira J, Moriya T, Kaneko C, Utsunomiya H, Yaegashi N, Okamura K, Sasano H 2002 Expression of androgen receptor and 5-reductases in the human normal endometrium and its disorders. Int J Cancer 99:652–657

    Vermeulen-Meiners C, Poortman J, Nabuurs M, Thijssen JH 1988 The endogenous concentration and subcellular distribution of androgens in normal human premenopausal endometrium, myometrium and vagina. Gynecol Endocrinol 2:121–130

    Weihua Z, Ekman J, Almkvist A, Saji S, Wang L, Warner M, Gustafsson JA 2002 Involvement of androgen receptor in 17?-estradiol-induced cell proliferation in rat uterus. Biol Reprod 67:616–623

    Schmidt WN, Katzenellenbogen BS 1979 Androgen-uterine interactions: an assessment of androgen interaction with the testosterone- and estrogen-receptor systems and stimulation of uterine growth and progesterone-receptor synthesis. Mol Cell Endocrinol 15:91–108

    Papaconstantinou AD, Umbreit TH, Goering PL, Brown KM 2002 Effects of 17-methyltestosterone on uterine morphology and heat shock protein expression are mediated through estrogen and androgen receptors. J Steroid Biochem Mol Biol 82:305–314

    Kuhl H, Born HJ, Schneider R, Schneider M, Taubert HD 1985 Dose- and time-dependent effects of a dimeric ethynodiol-testosterone depot-preparation in female rat. Endocrinol Exp 19:105–115

    Skarda J 2002 Sensitivity and specificity of bioassay of estrogenicity on mammary gland and uterus of female mice. Physiol Res 51:407–412

    Nantermet PV, Xu J, Yu Y, Hodor P, Holder D, Adamski S, Gentile MA, Kimmel DB, Harada S, Gerhold D, Freedman LP, Ray WJ 2004 Identification of genetic pathways activated by the androgen receptor during the induction of proliferation in the ventral prostate gland. J Biol Chem 279:1310–1322

    Cardenas H, Pope WF 2004 Attenuation of estrogenic effects by dihydrotestosterone in the pig uterus is associated with downregulation of the estrogen receptors. Biol Reprod 70:297–302

    Sato T, Wang G, Hardy MP, Kurita T, Cunha GR, Cooke PS 2002 Role of systemic and local IGF-I in the effects of estrogen on growth and epithelial proliferation of mouse uterus. Endocrinology 143:2673–2679

    Adesanya OO, Zhou J, Samathanam C, Powell-Braxton L, Bondy CA 1999 Insulin-like growth factor 1 is required for G2 progression in the estradiol-induced mitotic cycle. Proc Natl Acad Sci USA 96:3287–3291

    Liu Q, Ryon J, Nilsen-Hamilton M 1997 Uterocalin: a mouse acute phase protein expressed in the uterus around birth. Mol Reprod Dev 46:507–514

    Shemesh M 2001 Actions of gonadotrophins on the uterus. Reproduction 121:835–842

    Nikitenko LL, MacKenzie IZ, Rees MC, Bicknell R 2000 Adrenomedullin is an autocrine regulator of endothelial growth in human endometrium. Mol Hum Reprod 6:811–819

    Bigsby RM 2002 Control of growth and differentiation of the endometrium: the role of tissue interactions. Ann NY Acad Sci 955:110–117;discussion 118:396–406

    Rutanen EM 1998 Insulin-like growth factors in endometrial function. Gynecol Endocrinol 12:399–406

    Niiro N, Nishimura J, Hirano K, Nakano H, Kanaide H 1998 Mechanisms of galanin-induced contraction in the rat myometrium. Br J Pharmacol 124:1623–1632

    Reis AM, Jankowski M, Mukaddam-Daher S, Tremblay J, Dam TV, Gutkowska J 1997 Regulation of the natriuretic peptide system in rat uterus during the estrous cycle. J Endocrinol 153:345–355

    Wahlstrom T, Laatikainen T, Salminen K, Leppaluoto J 1985 Immunoreactive ?-endorphin is demonstrable in the secretory but not in the proliferative endometrium. Life Sci 36:987–990

    Takahashi T, Eitzman B, Bossert NL, Walmer D, Sparrow K, Flanders KC, McLachlan J, Nelson KG 1994 Transforming growth factors ?1, ?2, and ?3 messenger RNA and protein expression in mouse uterus and vagina during estrogen-induced growth: a comparison to other estrogen-regulated genes. Cell Growth Differ 5:919–935

    Giudice LC 1994 Growth factors and growth modulators in human uterine endometrium: their potential relevance to reproductive medicine. Fertil Steril 61:1–17

    Wolf E, Lahm H, Wu M, Wanke R, Hoeflich A 2000 Effects of IGFBP-2 overexpression in vitro and in vivo. Pediatr Nephrol 14:572–578

    Miller C, Sassoon DA 1998 Wnt-7a maintains appropriate uterine patterning during the development of the mouse female reproductive tract. Development 125:3201–3211

    Bui TD, Zhang L, Rees MC, Bicknell R, Harris AL 1997 Expression and hormone regulation of Wnt2, 3, 4, 5a, 7a, 7b and 10b in normal human endometrium and endometrial carcinoma. Br J Cancer 75:1131–1136

    Parr BA, McMahon AP 1998 Sexually dimorphic development of the mammalian reproductive tract requires Wnt-7a. Nature 395:707–710

    Green AR, Parrott EL, Butterworth M, Jones PS, Greaves P, White IN 2001 Comparisons of the effects of tamoxifen, toremifene and raloxifene on enzyme induction and gene expression in the ovariectomised rat uterus. J Endocrinol 170:555–564

    Labib K, Tercero JA, Diffley JF 2000 Uninterrupted MCM2–7 function required for DNA replication fork progression. Science 288:1643–1647

    Rouault JP, Rimokh R, Tessa C, Paranhos G, Ffrench M, Duret L, Garoccio M, Germain D, Samarut J, Magaud JP 1992 BTG1, a member of a new family of antiproliferative genes. EMBO J 11:1663–1670

    Frankel FR, Hsu CY, Myers JC, Lin E, Lyttle CR, Komm B, Mohn K 1988 Regulation of 2(I), 1(III), and 2(V) collagen mRNAs by estradiol in the immature rat uterus. DNA 7:347–354

    Luo G, D’Souza R, Hogue D, Karsenty G 1995 The matrix Gla protein gene is a marker of the chondrogenesis cell lineage during mouse development. J Bone Miner Res 10:325–334

    Stacey K, Beasley B, Wilce PA, Martin L 1991 Effects of female sex hormones on lipid metabolism in the uterine epithelium of the mouse. Int J Biochem 23:371–376

    Gorski J, Morgan MS 1967 Estrogen effects on uterine metabolism: reversal by inhibitors of protein synthesis. Biochim Biophys Acta 149:282–287

    Khan SA, Moulton BC 1990 Regulation of 3-hydroxyl-3-methylglutaryl-coenzyme A reductase activity in rat uterine tissues. J Steroid Biochem 35:70170–70174

    Das SK, Taylor JA, Korach KS, Paria BC, Dey SK, Lubahn DB 1997 Estrogenic responses in estrogen receptor- deficient mice reveal a distinct estrogen signaling pathway. Proc Natl Acad Sci USA 94:12786–127891

    Dupont S, Krust A, Gansmuller A, Dierich A, Chambon P, Mark M 2000 Effect of single and compound knockouts of estrogen receptors (ER) and ? (ER?) on mouse reproductive phenotypes. Development 127:4277–4291

    Savouret JF, Bailly A, Misrahi M, Rauch C, Redeuilh G, Chauchereau A, Milgrom E 1991 Characterization of the hormone responsive element involved in the regulation of the progesterone receptor gene. EMBO J 10:1875–1883

    Kraus WL, Montano MM, Katzenellenbogen BS 1993 Cloning of the rat progesterone receptor gene 5'-region and identification of two functionally distinct promoters. Mol Endocrinol 7:1603–1616

    L’Horset F, Blin C, Colnot S, Lambert M, Thomasset M, Perret C 1994 Calbindin-D9k gene expression in the uterus: study of the two messenger ribonucleic acid species and analysis of an imperfect estrogen-responsive element. Endocrinology 134:11–18

    Li XH, Ong DE 2003 Cellular retinoic acid-binding protein II gene expression is directly induced by estrogen, but not retinoic acid, in rat uterus. J Biol Chem 278:35819–35825

    Fan JD, Wagner BL, McDonnell DP 1996 Identification of the sequences within the human complement 3 promoter required for estrogen responsiveness provides insight into the mechanism of tamoxifen mixed agonist activity. Mol Endocrinol 10:1605–1616

    Bale TL, Dorsa DM 1997 Cloning, novel promoter sequence, and estrogen regulation of a rat oxytocin receptor gene. Endocrinology 138:1151–1158

    DiGiovanni J, Kiguchi K, Frijhoff A, Wilker E, Bol DK, Beltran L, Moats S, Ramirez A, Jorcano J, Conti C 2000 Deregulated expression of insulin-like growth factor 1 in prostate epithelium leads to neoplasia in transgenic mice. Proc Natl Acad Sci USA 97:3455–3460

    Cicatiello L, Sica V, Bresciani F, Weisz A 1993 Identification of a specific pattern of "immediate-early" gene activation induced by estrogen during mitogenic stimulation of rat uterine cells. Receptor 3:17–30

    Parrinello S, Lin CQ, Murata K, Itahana Y, Singh J, Krtolica A, Campisi J, Desprez PY 2001 Id-1, ITF-2, and Id-2 comprise a network of helix-loop-helix proteins that regulate mammary epithelial cell proliferation, differentiation, and apoptosis. J Biol Chem 276:39213–39219

    Raman V, Martensen SA, Reisman D, Evron E, Odenwald WF, Jaffee E, Marks J, Sukumar S 2000 Compromised HOXA5 function can limit p53 expression in human breast tumours. Nature 405:974–978

    Cinar HN, Sweet KL, Hosemann KE, Earley K, Newman AP 2001 The SEL-12 presenilin mediates induction of the Caenorhabditis elegans uterine pi cell fate. Dev Biol 237:173–182

    Aubin J, Dery U, Lemieux M, Chailler P, Jeannotte L 2002 Stomach regional specification requires Hoxa5-driven mesenchymal-epithelial signaling. Development 129:4075–4087

    Slayden OD, Brenner RM 2003 Flutamide counteracts the antiproliferative effects of antiprogestins in the primate endometrium. J Clin Endocrinol Metab 88:946–949

    Yeh S, Tsai MY, Xu Q, Mu XM, Lardy H, Huang KE, Lin H, Yeh SD, Altuwaijri S, Zhou X, Xing L, Boyce BF, Hung MC, Zhang S, Gan L, Chang C, Hung MC 2002 Generation and characterization of androgen receptor knockout (ARKO) mice: an in vivo model for the study of androgen functions in selective tissues. Proc Natl Acad Sci USA 99:13498–134503

    Sato T, Matsumoto T, Yamada T, Watanabe T, Kawano H, Kato S 2003 Late onset of obesity in male androgen receptor-deficient (AR KO) mice. Biochem Biophys Res Commun 300:167–171

    Vrontakis M, Schroedter I, Leite V, Friesen HG 1993 Estrogen regulation and localization of galanin gene expression in the rat uterus. Biol Reprod 49:1245–1250

    Laoag-Fernandez JB, Otani T, Maruo T 2000 Adrenomedullin expression in the human endometrium. Endocrine 12:15–19

    Dos Reis AM, Fujio N, Dam TV, Mukaddam-Daher S, Jankowski M, Tremblay J, Gutkowska J 1995 Characterization and distribution of natriuretic peptide receptors in the rat uterus. Endocrinology 136:4247–4253

    Gimpl G, Fahrenholz F 2001 The oxytocin receptor system: structure, function, and regulation. Physiol Rev 81:629–683

    Ryon J, Bendickson L, Nilsen-Hamilton M 2002 High expression in involuting reproductive tissues of uterocalin/24p3, a lipocalin and acute phase protein. Biochem J 367:271–277

    Michishita M, Minegishi T, Abe K, Kangawa K, Kojima M, Ibuki Y 1999 Expression of adrenomedullin in the endometrium of the human uterus. Obstet Gynecol 93:66–70

    Cameron VA, Fleming AM 1998 Novel sites of adrenomedullin gene expression in mouse and rat tissues. Endocrinology 139:2253–2264

    Yanagita T, Yamamoto R, Sugano T, Kobayashi H, Uezono Y, Yokoo H, Shiraishi S, Minami SI, Wada A 2000 Adrenomedullin inhibits spontaneous and bradykinin-induced but not oxytocin- or prostaglandin F(2)-induced periodic contraction of rat uterus. Br J Pharmacol 130:1727–1730

    Bai G, Kasper S, Matusik RJ, Rennie PS, Moshier JA, Krongrad A 1998 Androgen regulation of the human ornithine decarboxylase promoter in prostate cancer cells. J Androl 19:127–135

    Nelson KG, Takahashi T, Bossert NL, Walmer DK, McLachlan JA 1991 Epidermal growth factor replaces estrogen in the stimulation of female genital-tract growth and differentiation. Proc Natl Acad Sci USA 88:21–25

    Sciavolino PJ, Abate-Shen C 1998 Molecular biology of prostate development and prostate cancer. Ann Med 30:357–368

    Pelletier G 2000 Localization of androgen and estrogen receptors in rat and primate tissues. Histol Histopathol 15:1261–1270

    Adesanya-Famuyiwa OO, Zhou J, Wu G, Bondy C 1999 Localization and sex steroid regulation of androgen receptor gene expression in rhesus monkey uterus. Obstet Gynecol 93:265–270(Pascale V. Nantermet, Pat)

    http://www.100md.com/html/DirDu/2006/09/10/16/84/17.htm