A Role for Androgens in Epithelial Proliferation and Formation of Glands in the Mouse Uterus

The endometrium consists of stromal and epithelial compartments (luminal and glandular) with distinct functions in the regulation of uterine homeostasis. Ovarian sex steroids, namely 17 (cid:2) estradiol (E 2 ) and progesterone (P 4 ), play essential roles in modulating uterine cell proliferation, stromal-epithelial cross-talk and differentiation in preparation for pregnancy. The impact of an-drogensonuterinefunctionremainspoorlyunderstood.Thecurrentstudyinvestigatedtheimpact of the non-aromatizable androgen dihydrotestosterone (DHT) on mouse endometrial function. Ovariectomized female mice were given a single subcutaneous injection (short treatment) or 7 daily injections (long treatment) of vehicle alone (5% ethanol, 0.4% methylcellulose) or vehicle with the addition of 0.2 mg DHT (n (cid:2) 8/group) and a single injection of Bromodeoxyuridine (BrdU) 2 hours prior to tissue recovery. Prl8, Prlr ) as well as a striking impact on the number of endometrial glands. This study has revealed a novel role for androgens in regulating uterine function with an impact on the glandular compartment of the endometrium. This previously unrecognised role for androgens has implications for our understanding of the role of androgens in regulation of endometrial function and fertility in women. ment group). Statistical analysis performed using GraphPad Prism 6.0. Student’s t test was used for comparison between VC and DHT groups, with statistical significance for P (cid:5) .05.

T he endometrium is a hormone-dependent multicellular tissue that consists of stromal and epithelial compartments. The stromal compartment contains mesenchymal fibroblasts, a well-developed vasculature and immune cells, while the epithelium is divided into a luminal and a glandular component, each with a distinct identity and role in reproductive physiology (1). Both compartments respond to hormonal cues in a coordinated spatial and temporal manner characterized by extensive cross-talk to regulate the development and homeostasis of the tissue (2)(3)(4).
While the impacts of estrogens on endometrial function have been extensively studied, androgens have only recently emerged as key players in the regulation of endometrial function. Analysis of circulating steroid hormone levels utilizing sensitive liquid chromatography-tandem mass spectrometry has revealed significant concentrations of androgens present throughout the menstrual cycle in women exceeding those of circulating estrogens (5). In-vestigation of circulating androgen levels in cycling female mice reported detectable concentrations of the circulating testosterone metabolite 5␣-dihydrotestosterone (DHT; 2.8 Ϯ 0.4 pg/ml) during all phases of the estrous cycle, with concentrations peaking during proestrus and metestrus (6). In aging female mice, there is a significant increase of circulating DHT, with a mean value of 12.8 Ϯ 0.6 pg/ml. To our knowledge there are no studies in the literature regarding the circulating levels of DHT during mouse pregnancy, however, there are several reports that androgen levels increase during pregnancy in women (reviewed in (7)). In women androgens appear to be 'Goldilocks' factors with both deficiency and excess contributing to pregnancy-and fertility-related disorders such as polycystic ovarian syndrome (8,9), endometriosis (10) and recurrent pregnancy loss (11).
The effects of androgens are mediated via the androgen receptor (AR), a ligand-activated transcription factor that binds testosterone (T) and its metabolite 5␣-dihydrotestosterone (DHT) with high affinity and specificity. In women, expression of AR in the endometrium is regulated both temporally and spatially during the normal menstrual cycle and detailed analysis of full-thickness uterine tissue sections identified endometrial stromal fibroblasts as a key targets for androgen action (12). Administration of antiprogestins in women and nonhuman primates induces upregulation of AR in the uterine epithelium, which may mediate the antiproliferative effects of these compounds (13). In the uteri of mice, quantitative in situ hybridization detected uniform labeling of Ar mRNA in all compartments, including the epithelium (14). A recent study reported nuclear stromal AR staining in the mouse uterus but AR was not detected in the luminal or glandular epithelium (15). A study using a transgenic mouse model of an androgen response element (ARE) construct fused to a luciferase reporter revealed strong AR activity in the uterus, with the antiandrogen bicalutamide inhibiting luciferase activity (16). Female global AR knockout (ARKO) mice are subfertile, with a neuroendocrine and ovarian phenotype accompanied by defects in uterine development (17).
Studies in female-to-male transsexuals have reported a negative impact of androgens on endometrial proliferation in response to long-term administration of testosterone, resulting in endometrial atrophy (18,19). Treatment of primary human endometrial stromal cells with DHT also results in a significant reduction in both proliferation and apoptosis (12). Androgen-estrogen cross-talk has been implicated in the initiation and progression of endometrial cancers, with androgens acting in an antagonistic manner to the proliferative effects of estrogens, restricting endometrial carcinogenesis (reviewed in (20)). Weihua et al reported that the proliferative effects of E 2 in the rat uterus were blocked by either the antiestrogen tamoxifen or the antiandrogen flutamide (21), suggesting a functional overlap between androgen and estrogen action in this rodent species. Microarray analysis of uteri recovered from ovariectomized mice after a single injection of DHT revealed time-dependent changes in the expression of genes involved in the regulation of cell division and organ development between 12 and 24 hours post injection (22).
Sex steroids have been shown to exert different effects in cell proliferation and apoptosis resulting from either acute or chronic stimulation (reviewed in (23,24)). In the 1970s, Lee and colleagues recorded three phases of epithelial cell proliferation within the mouse uterus in response to continuous oral treatment with estrogen, characterized by pronounced uterine epithelial cell proliferation on days 2 and 3, basal proliferation on days 4 and 5 and a second wave of proliferation on days 19 and 20 (25,26). Similar effects have been reported in prostate cancer cells in response to androgen treatment, characterized by an early proliferative response followed by a refractory period to androgen-induced proliferation (27). Taking into account this biphasic effect of androgens in controlling cell proliferation of target tissues, a short-term (24 hrs) and a long-term (7 days) treatment paradigm was selected to elucidate the impact of DHT on endometrial growth in a mouse model. Experiments were performed in ovariectomized mice to avoid the confounding impacts of endogenous ovarian hormones.
In the current study, we show that treatment with DHT has compartment-specific impacts on the immunoexpression of AR, cellular proliferation and expression of cellcycle regulators and epithelial growth factors resulting in an increase in the number of uterine glands. These striking findings identify a novel role for androgens in the regulation of endometrial homeostasis.

Animals and Treatments
Female C57BL/6J mice were purchased from Charles Rivers Laboratories and allowed to acclimate for a week with ad libitum access to food and water. Mice were maintained in accordance with UK legal requirements under licensed approval from the UK Home Office. Eight-to ten-week old female mice were ovariectomized (ovx) by dorsal bilateral ovariectomy and allowed to recover for 7 days prior to treatment to deplete endogenous sexsteroid hormones. Surgery was performed under isoflurane anesthesia followed by a postoperative analgesic, buprenorphine (0.1 mg/kg) for pain management. Mice were randomly assigned into four treatment groups (n ϭ 8/group, 32 in total). To explore the short-term and long-term impacts of androgens mice received either one (24 hrs group) or seven (7 days group) daily subcu-taneous injections of DHT (0.2 mg/mouse) or vehicle control (5% ethanol-0.4% methylcellulose). Two hours prior to being sacrificed, mice received an intra-peritoneal injection of 0.1 ml BrdU (25 mg/ml in PBS).
To measure the serum concentration of DHT after treatments, blood was collected via cardiac puncture and exsanguination from mice treated with VC (5% ethanol-0.4% methylcellulose) or DHT (0.2 mg/mouse) for 24 hrs (n ϭ 4/group). Serum DHT concentrations were assayed using a commercially available mouse DHT ELISA kit (Amsbio, BlueGene; Cat. Number E03D0001 -Assay sensitivity: 1 pg/ml). Samples were assayed in duplicate and a standard curve was generated using a four parameter logistic curve fit (4-PL); sample DHT concentrations were interpolated from the standard curve. For analysis of tissues, one uterine horn was fixed in 4% neutral buffered formalin (NBF) for paraffin embedding, sectioning and histology while the other horn was kept in RNA Save at -80°C.

RNA Extraction and Reverse Transcription
Total RNA was extracted from homogenized mouse uterine tissue (20 mg/sample) using TriReagent and chloroform followed by RNeasy Mini Kit (Qiagen, UK) column isolation and elution as per the manufacturer's instructions. RNA concentration and quality was measured using a Nanodrop® ND-1000 spectrophotometer (Nanodrop Technologies, USA) and the samples were standardized to 100ng/l in RNAse-free water. Reverse transcription was performed using SuperScript VILO cDNA Synthesis Kit and Master Mix (Invitrogen, UK) as per the manufacturer's instructions using a thermal cycler programmed at: 25°C for 10 minutes, 42°C for 60 minutes and 52°C for 5 minutes. Two negative controls (no enzyme control and no RNA control) and one positive control (mouse RNA positive control) were included for each set of RNA samples.

Quantitative RT-PCR Analysis
Quantitative PCR was performed using SuperMix with Premixed ROX dye (Invitrogen, UK), primer sets designed using the Roche Universal Probe Library Assay Design Centre (sequences in Table 1) purchased from Eurofins (MWG Operon) and probes from the Roche Universal Probe Library Mouse Set (Roche Applied Science, UK). Samples were assayed in duplicate and run on an ABI 7900 HT Fast Real Time PCR machine using the following cycling conditions: 95°C for 10 minutes then 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. Primer amplification efficiency was validated and analysis was performed using the relative standard curve method. Data were normalized to 18S and fold-change is expressed as the ratio of expression of each gene of interest in the treated groups vs the average of the VC groups (one for the 24 hrs and one for the 7 days time point respectively). Statistical analysis was performed using GraphPad Prism 6.0. Data are presented as mean-/ϩSEM. Student's t test was used for comparison between VC and DHT groups. Criterion for significance was P Ͻ .05. Primer pair and probe information is provided in Supplemental Table 1.

Immunohistochemistry and Immunofluorescence
Samples were fixed in 4% NBF overnight at room temperature, transferred in 70% ethanol and the tissues were processed for paraffin wax embedding. Transverse sections of 5 m thickness were cut using a microtome (Leica RM2135) and the slides were incubated overnight at 55°C. For histology, the tissues were dewaxed, rehydrated and antigen retrieval was performed using citrate (0.01M citrate, pH 6.0) in a decloaking chamber followed by several blocking steps with serum from the species the secondary antibody was raised in, to avoid nonspecific binding and hydrogen peroxide was used to block endogenous peroxidase activity. Sections were incubated with the primary antibody diluted in serum blocking solution overnight at 4°C. Negative controls were included, in which the primary antibody was omitted to identify any nonspecific binding by the secondary antibody. The following day, detection was performed by using either a conjugated secondary antibody specific for the primary antibody or a polymer-based detection system (ImmPress, Vector Labs, UK). For chromogenic immunohistochemistry, DAB (Vector, UK) was used as a chromogen, while for immunofluorescence the Tyramide Signal Amplification system (Perkin Elmer, UK) was used as per the manufacturer's instructions. Between incubations, washes were performed with TBS-Tween. Hematoxylin was used as a counterstain for immunohistochemistry and DAPI as a counterstain for immunofluorescence. Antibody information is provided in Table 1.

Image Acquisition and Stereology
For chromogenic immunohistochemistry images were acquired with a Provis AX70 microscope (Olympus Optical, London, UK) fitted with a Canon DS6031 camera (Canon Europe, Amsterdam) while fluorescent images were acquired with a LSM 710 confocal microscope (Carl Zeiss, UK) and processed with the ZEN software (Carl Zeiss). Tissue surface area was measured using a Zeiss AXIO Imager A1 (Carl Zeiss, UK) while quantification of BrdU-positive epithelial cells and of the number of glands was performed by blinded manual counting by two independent investigators. For the calculation of cell density, sections were counterstained with DAPI and acquired images were analyzed using the image analysis software Fiji (28). In all cases, a minimum of two nonserial sections were used (n ϭ 4 -8/treat-

Trophic effects of androgens in the mouse uterus
In order to elucidate the actions of DHT in mouse uterine physiology, ovx mice were treated with either vehicle or DHT under a short-or long-term treatment regime (24 hrs and 7 days respectively). ELISA analysis of serum from mice injected with vehicle control or DHT 24 hrs previously had a mean DHT concentration of 180 pg/ml in the DHT group, while DHT concentrations in the control group were below the level of detection (Ͻ1 pg/ml) (Supplemental Figure 1). These results appear consistent with those of Zhang et al who previously demonstrated that ovx C57BL/6 female mice injected subcutaneously with 0.1 mg/mouse DHT, exhibit a peak of circulating DHT 1 hour postinjection (ϳ14 ng/ml) which rapidly declines over the following 24 hrs (29).
The short-term treatment with DHT (24 hrs) induced a significant increase in uterine weight compared to vehicle which was further increased when treatment was extended to 7 days ( Figure 1A, P Ͻ .001 at 7 days). The effect of DHT on uterine weight was accompanied by significant increase of total uterine surface area ( Figure 1B), endometrial area ( Figure 1C) and myometrial area ( Figure 1D) at both time points. Analysis of cell density suggested the change in uterine size was due to an increase in cell numbers accompanied by a significant reduction in cell density at both time points ( Figure 1E, F).

Treatment with DHT induces epithelial AR expression in the mouse uterus
Positive immunoexpression of AR was detected in the stromal compartment at both 24 hrs and 7 days in both vehicle and DHT-treated mice. DHT treatment induced expression of AR in the luminal and glandular epithelium at 24 hrs ( Figure 2B) that was absent in vehicle groups at both time points (Figure 2A, C). In mice treated with DHT for 7 days, AR was detected in the stroma and the glandular epithelium but not the luminal epithelium ( Figure  2D). Real-time PCR showed a significant decrease of Ar mRNA after 7 days of DHT treatment compared to vehicle ( Figure 2E). Investigation of the expression pattern of ER␣ revealed no difference between vehicle and DHT treatedgroups at both time points (Supplemental Figure 2).

Treatment with DHT induces epithelial proliferation in the mouse uterus
Immunofluorescence was performed with an antibody targeting the proliferation marker MKi67 that detects all cells in interphase or undergoing mitosis. Staining identified extensive proliferation of luminal and glandular epithelial cells after treatment with DHT at 24 hrs compared to vehicle ( Figure 3A

DHT regulates genes necessary for proliferation, cell-cycle regulation and stromal-epithelial crosstalk in the mouse uterus
Real-time PCR analysis identified dynamic changes in the expression of several genes associated with cell-cycle regulation and stromal-epithelial cross-talk following treatment with DHT.
Wee1 expression was significantly upregulated by DHT after 24 hrs, followed by a significant downregulation at 7 days ( Figure 5A). Ccnd1 expression was significantly increased by DHT at 24 hrs followed by a significant decrease after 7 days ( Figure 5B). Rb1 expression was significantly downregulated by DHT at both time points ( Figure 5C).
Wnt signaling, involved in stromal-epithelial cross-talk in the uterus, was significantly affected by treatment with DHT. Wnt4 expression was significantly downregulated after treatment with DHT for 24 hrs and 7 days ( Figure  5D). Wnt5a mRNA concentration was significantly increased by DHT at both time points (Figure 5E), while press.endocrine.org/journal/endo 5 Wnt7a expression was downregulated by DHT only at 7 days ( Figure 5F). Cdh1 which is expressed in the uterine epithelium was significantly upregulated by DHT at both 24 hrs and 7 days ( Figure 5G), while expression of the adherens junction component vinculin (Vcl) involved in intercellular adhesion and signaling displayed a significant upregulation by DHT at 24 hrs, with no significant change observed between vehicle and DHT at 7 days ( Figure 5H). Expression of the putatively androgen-regulated epithelial growth factor Igf1 was significantly upregulated by DHT at both time points compared to vehicle treatments, reaching an average 10-fold increase at 7 days ( Figure 5I). Prl8, whose role has been reported in gland formation in the mouse uterus, was significantly upregulated more than 20-fold at both time points ( Figure 5J), with a similar pattern observed for it receptor (Prlr, Figure 5J). Furthermore, expression of the epithelial growth factors Fgf7 ( Figure 7L) and Hgf ( Figure 7M) was significantly altered after treatment with DHT. The transcription factor Foxa2, which is exclusively expressed in uterine glands exhibited a trend for an increase after DHT treatment at 24 hrs (P ϭ .14) but the changes at both time points didn't reach statistical significance ( Figure 5K).

A role for androgens in regulating gland formation in the mouse uterus
Histological examination of uterine sections at 7 days suggested more glands with DHT treatment. FOXA2 is a transcription factor expressed exclusively in the glands in the mouse uterus and FOXA2 immunohistochemistry was performed to quantify the number of glands between the treatment groups ( Figure 6A, B). A significant increase in the total number of glands was observed with DHT compared to vehicle treatment verifying the qualitative observations ( Figure 6C). To compare the effect of DHT with that with E 2 in glandular expansion, ovx mice treated with an E 2 pellet for a week were used as a comparator. Staining for FOXA2 and quantification of glands demonstrated fewer glands in the E 2 group compared to the DHT-treated group accompanied by a weaker staining pattern (Supplemental Figure 3A, D). Real-time PCR identified a significant reduction in Foxa2 expression levels between E 2 and DHT-treated groups (Supplemental Figure 3).

Discussion
In this study, we have demonstrated that androgens induce epithelial proliferation resulting in an increased numbers of endometrial glands in the adult mouse uterus. Uterine glands are essential for uterine receptivity and implantation and secrete factors which are necessary for the survival and development of the conceptus. Our results suggest that homeostatic balance of androgens and the local steroidogenic environment could be important physiological drivers of gland formation during endometrial remodelling in early pregnancy. Animal models in which uterine gland development is disrupted, exhibit compromised fertility and altered expression of decidualization factors (30,31), highlighting the potential for prospective therapeutic targeting of androgen signaling in the uterus for fertilityassociated disorders (32). Although a trophic effect of androgens in the rodent uterus has previously been reported, we identify for the first time the molecular mechanisms involved and selective effects of androgens on distinct tissue compartments within the uterus. Notably, we describe for the first time that DHT promotes gland formation in the adult mouse uterus through directly stimulating glandular epithelial proliferation and by driving expression of regulators of cell cycle and stromal-epithelial interactions. The novel data generated in the current study provide new insights into the role of androgens in the regulation of endometrial function.
In the uterus, sex steroid action occurs via direct steroid hormone receptor signaling in target cells and indirect paracrine effects on neighboring cellular compartments. Such examples of intercellular paracrine signaling include the stromal-derived secretion of IGF1 and WNT5␣ binding on epithelial receptors controlling proliferation and gland formation respectively. Thus, regulation of the complex stromal-epithelial interactions resulting from androgen action can be investigated using in vivo model systems, rendering rodents invaluable models. To investigate the impact of androgen treatment in the steroid-depleted mouse uterus, we treated ovx mice with DHT at a supraphysiological dose (0.2 mg/mouse) previously shown to elicit uterine responses (29,33). Although the DHT concentrations generated in our experimental conditions are above the physiological levels detected in cycling female mice (6), this dose is consistent with those used in several previous studies investigating the direct action of DHT in the rodent uterus (22,34) and adipose tissue (29,35).
In this study, DHT induced a time-dependent increase in uterine weight in agreement with previous reports in rats treated with DHT or the synthetic androgen mibolerone, where a dose-and time-dependent increase in uterine weight after androgen treatment was detected in that species (34). In the present study, increases in total uterine area, endometrial area and myometrial area accompanied by a significant decrease in cell density, possibly due to water imbibition, were detected, demonstrating the trophic effects of DHT in both uterine compartments in the mouse. Although trophic effects are also induced by estrogens in the rodent uterus, in the current study uterine morphology after DHT treatment lacked the characteristic distended phenotype, edema and thinness of the endometrial layer induced by E 2 (36,37) consistent with a unique effect of androgens in regulating uterine physiology that is distinct from that of estrogens. Early studies using in vitro cell-free systems and in vivo rodent models have demonstrated interaction between testosterone and DHT with ER, leading to nuclear translocation of ER, while cotreatment with antiestrogens but not antiandrogens blocked this effect (38,39). However, a later study demonstrated that the in vivo doses of DHT required to elicit ER cytoplasmic depletion and nuclear translocation in the rat uterus are very high (5-10 mg), suggesting that the uterine trophic effects of DHT observed in our study are unlikely to be mediated by ER (40). In addition, ovariectomized female rats cotreated with DHT and the AR antagonist bicalutamide exhibited a complete suppression of DHT-induced expression of Igf1, a growth factor whose upregulation is both ER␣-and ARcontrolled (34). This finding together with the fact that putative ER␣ target-genes are not induced after treatment with DHT in the mouse uterus (34) suggest more of a functional interaction between AR and ER␣ than a physical one. Notably, in a study describing the development of a uterine glandular epithelial specific ARKO mouse model (ugeARKO), uterine parameters (uterine weight and surface area) of adult ugeARKO female mice were within the normal range but androgen-induced uterine growth in ovx mice was impaired (41) consistent with the results in the present study which describe androgenic induction of uterine growth via epithelial AR.
Since androgens promote growth of the rodent uterus, we investigated expression of proliferation markers in the tissue to elucidate compartment-specific effects. Assessment by immunofluorecence of the proliferation marker MKi67 which detects all actively proliferating cells (cells not in G 0 of the cell-cycle) revealed extensive proliferation within the uterine epithelium that was associated press.endocrine.org/journal/endo with a significant increase in concentrations of Mki67 mRNA after 24 hours that was also accompanied by nuclear translocation of cyclin D1 in the epithelium (data not shown). The proliferative effect of DHT was confirmed by assessing incorporation of BrdU, which revealed extensive proliferation in the epithelial compartment but not in the stroma after treatment with DHT. Notably, increased proliferation was only detected after 24 hrs of treatment consistent with DHT inducing rapid changes in the tissue. Our findings are further complemented by Terada et al who reported that treatment of ovx E 2 -primed mice with DHT significantly decreased the apoptotic index of the uterine epithelium, thus arguing for an inhibitory effect of DHT on cell death on the mouse uterine epithelium (42). Furthermore, the effects of DHT partially mimic those seen after treatment with E 2 where maximal epithelial proliferation rate is observed after 21 hrs of E 2 treatment followed by a decline in proliferation and increased apoptosis after repeated E 2 injections (43). Treatment of ovx mice with DHT resulted in a striking increase in AR immunoreactivity in the uterine stroma accompanied by induced expression of AR in glandular and luminal epithelial cells. Previous studies have demonstrated that while androgens decrease AR transcription, they positively modulate translation efficiency and stability of the AR protein, thus increasing AR protein levels (44 -47). These findings are in agreement with our data where we detected a significant decrease of Ar mRNA following treatment with DHT for 7 days accompanied by an increase of AR protein levels. As already mentioned, previous studies have reported that androgenic effects in the mouse uterus can be mediated via ER␣ signaling (21,34), with extensive cross-talk between the two receptors (40,48). However no change was observed in the expression pattern of ER␣ after DHT treatment, suggesting that if any functional interaction occurs between AR and ER␣ following DHT treatment it is not due to altered ER␣ levels.
Taking together the uterine epithelial proliferation following DHT treatment with the known role of androgens in controlling cell-cycle progression in the prostate (49) and the mouse uterus (22), transcript expression of cell-cycle regulators was analyzed by real-time PCR. The multifunctional cell-cycle regulator Wee1 controls chromatin integrity and entry into mitosis by regulating the transcription of histones and forming active complexes with factors that control cell-cycle progression (reviewed in (50)). DHT induced upregulation of Wee1 at 24 hrs, consis- tent with the actively proliferating epithelial compartment at that time point, while its expression after 7 days of DHT treatment was significantly reduced compared to vehicle control suggesting a negative feedback mechanism restoring tissue proliferation to basal levels. The same expression pattern was observed for Ccnd1 (cyclin D1), whose levels and nuclear localization have been shown to be regulated by E 2 in mouse uterine epithelial cells (51). Expression of the cell-cycle inhibitor and tumor suppressor Rb1 was significantly downregulated by DHT at both time points consistent with increased proliferation and growth of the uterus detected in response to DHT. Since the changes seen in uterine size reflect changes in all cellular compartments (including the stroma) further studies are required to investigate whether changes in stromal proliferation occur in response to DHT across other time points post androgen administration. The results of the current study extend previous transcriptional analyses of samples generated from DHT-treated isolated human endometrial stromal cells which demonstrate AR-dependent repression of cell cycle regulators (52) and DHT-dependent decreases in both proliferation and apoptosis (12). Taken together these results suggest androgens may impact on the stromal compartment to inhibit stromal cell apoptosis, rather than promoting proliferation, in addition to the direct effects on the uterine epithelial proliferation described in the current study. Wnt signaling is a major pathway involved in stromalepithelial interactions and organ patterning during neonatal and postnatal uterine development. Wnt4 and Wnt5␣ are expressed by stromal fibroblasts and Wnt7␣ by epithelial cells with paracrine and autocrine signaling controlling cell fate and differentiation of the uterus (53). We observed significant changes in the expression pattern of these factors, with Wnt4 and Wnt7a being significantly downregulated and Wnt5a significantly upregulated after DHT treatment. Studies with mice have demonstrated that all three Wnts are required for gland formation in the uterus and knockout mice for these Wnts have compromised fertility (54 -56). In addition, administration of either progesterone or the synthetic estrogen diethylstilbestrol (DES) during an early postnatal time window inhibits endometrial formation of glands (adenogenesis) in the mouse uterus (3,57). Similarly, cadherin 1 (Cdh1) has been to shown to be vital for endometrial differentiation of the uterine epithelium, while loss of Cdh1 in the uterus leads to absence of glands, a disorganized luminal epithelium and upregulation of Wnt4a with concomitant downregulation of Wnt5a in the neonatal mouse uterus (58). In our study, treatment with DHT induced a significant upregulation of Cdh1 at both 24 hrs and 7 days accompanied by downregulation of Wnt4␣ and upregulation of Wnt5␣, suggesting that androgens in the mouse uterus positively regulate formation of glands. press.endocrine.org/journal/endo 9 Trophic effects in tissues are often accompanied by tissue remodelling and that led us to investigate the expres-sion of vinculin (Vcl), an adherens junction component involved in cell-cell adhesion, whose expression is andro- gen-regulated in the prostate (59). Vcl was significantly upregulated after 24 hrs of DHT treatment and by 7 days its expression was the same as the vehicle, indicating an initial phase of alteration in cell adhesion properties followed by restoration to basal levels. In addition, this is the first evidence of androgenic regulation of Vcl in the uterus.
To corroborate the increased epithelial proliferation after treatment with DHT, expression of epithelial growth factors was assessed. IGF1 is predominantly produced by mouse endometrial stromal cells following E 2 treatment and signals to the epithelium to proliferate (51) and previous studies have reported that the Igf1 promoter contains androgen response elements (60). In the current study, a significant increase in the expression of the epithelial growth factor Igf1 was detected after 7 days of DHT treatment. Notably, Nantermet et al reported that DHT did not induce uterine hypertrophy in mice lacking a functional estrogen receptor (estrogen receptor ␣ knockout mice; ERKO), suggesting that ER␣ is required to mediate DHT responses (34). However, Klotz et al have previously reported that ER␣ expression is necessary to mediate the proliferative effect of IGF1 in the mouse uterus (61). Thus, taken together with the data in the current study these results are consistent with DHT also mediating indirect effects on proliferation in the uterus through stromal-epithelial cross-talk following induction of the IGF signaling pathway. The stromal-derived epithelial growth factors FGF7 and HGF both mediate stromal-epithelial interactions in the uterus by promoting epithelial proliferation with important roles in epithelial morphogenesis (62). While progesterone administration in the neonatal mouse uterus inhibits uterine gland formation accompanied by a significant reduction in the expression of Fgf7 and a significant increase in the expression of Hgf (62), in our study, treatment with DHT induced the opposite effect by significantly increasing Fgf7 mRNA levels and decreasing Hgf mRNA levels.
The most striking result observed was the impact of DHT on uterine glands. In control uteri, AR expression was not detected in glandular epithelial cells, however DHT treatment induced AR expression and increased glandular epithelial proliferation. Quantification of FOXA2-positive glands demonstrated more uterine glands in the DHT-treated mouse uterus, an effect which was unique to DHT, as treatment with E 2 did not produce this effect. FOXA2 is expressed exclusively in the uterine glandular epithelium of mice and Foxa2 uterine knockout mice fail to develop glands (30). In the current study, DHT induced changes in expression of a number of factors that promote adenogenesis in adult mice. Although expression of Foxa2 mRNA in whole tissue homogenates was not significantly different between control and DHT treated mice, treatment with E 2 significantly decreased Foxa2 mRNA concentrations highlighting the importance of androgen-estrogen balance in the regulation of this cellular compartment in the uterus. Taken together these data are consistent with DHT regulating gland formation in the adult mouse uterus by inducing a transcriptional program that results in de novo adenogenesis.
In the present study, treatment of mice with DHT revealed androgenspecific effects on endometrial architecture characterized by a rapid increase in proliferation of epithelial cells lining the glands and significant DHT-dependent changes in expression of genes associated with cell-cycle regulation and stromal-epithelial cross-talk. These results suggest that any endocrine pathology that resulted in significant deviation from normal intra-tissue concentrations of androgens could have a long term impact on the abundance of glandu- press.endocrine.org/journal/endo lar epithelial cells, with consequent impacts on glandular secretions and the epithelial-stromal cell cross-talk required to support establishment of pregnancy in women.