Essential Roles of Androgen Signaling in Wolffian Duct Stabilization and Epididymal Cell Differentiation

Mice and their embryos

The mutant mice used in this study were AP2α-Cre (32), AR^flox (33), CAG-Cre (34), CAG-CreER^TM (35), and p63 mutants (23). For normal expression analyses, ICR (Crlj:CD1; Charles River Laboratories Japan, Kanagawa, Japan) mice were used. Experiments using laboratory mice were approved by the Committee on Animal Research at Kumamoto University. Embryos and pups for each experiment were collected from more than three independent pregnant females. Noon of the day of a vaginal plug was designated as E0.5. AR-null [AR^−/Y, AR knockout (KO)] male mutant mice were generated as previously described (36) (Supplemental Fig. 1A, published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org) using CAG-Cre mice (34) and loss of AR expression in epithelium and mesenchyme was confirmed (Supplemental Fig. 1B and data not shown). The tamoxifen (TM)-inducible Cre recombinase system removes the floxed sequence from the target genome upon exposure to TM (35). TM (Sigma, St. Louis, MO) was dissolved in sesame oil (Kanto Chemical, Tokyo, Japan) at a final concentration of 10 mg/ml. Four milligrams of TM per 40 g body weight was administrated (ip) to the pregnant mice. Under these conditions, no overt teratological effects or sexual disorders of the reproductive organs were observed (37). Control specimens included AR^+/Y (male) and AR^+/+ and AR^+/− control females. For analyses using AP2α-Cre and CAG-CreER^TM alleles, the control indicates Cre allele-negative mice.

Histology and immunohistochemistry

Hematoxylin and eosin staining was performed by standard procedures as previously described (38). Semithin sections were prepared and toluidine blue staining was performed as previously described (39). For immunohistochemistry, mouse embryos and epididymides were fixed in 4% paraformaldehyde, embedded in paraffin, and prepared for histological analyses by standard procedures. After deparaffinization, sections were treated for antigen retrieval (autoclave 121 C for 1 min in 10 mm citrate buffer at pH 6.0) and incubated with 3% H₂O₂ in PBS for 10 min to inactivate endogenous peroxidase before 1 h incubation with a blocking solution [1.5% fetal bovine serum (FBS) in PBS]. Anti-AR and anti-p63 polyclonal antibodies (Santa Cruz Biotechnology, Santa Cruz, CA), anti-K14 polyclonal antibody (Covance, Berkeley, CA), anti-AQP9 polyclonal antibody (Alpha Diagnostic, San Antonio, TX), or anti-Foxi1 polyclonal antibody (Abcam, Cambridge, UK) in PBS solution was added (in dilution of 1:200, 1:100, 1:1000, 1:400, and 1:100 for each antibody) to the slides and incubated for 1 h at room temperature. Negative controls were obtained without the primary antibodies. After washing with PBS, the sections were stained with the Vectastain ABC kit (Vector Laboratories, Burlingame, CA) or incubated with Alexa 546/488-labeled goat antirabbit/mouse IgG (Invitrogen, Carlsbad, CA) or horseradish peroxidase-labeled rabbit antigoat IgG (Zymed, South San Francisco, CA) for 1 h at room temperature. Where indicated, the sections were subsequently incubated with 3,3-diaminobenzidine tetrahydrochloride (DAB) containing H₂O₂.

Cell proliferation and apoptosis assays

Pregnant females were injected ip with 100 mg 5-bromo-2′-deoxyuridine (BrdU; Sigma) per 1 kg body weight. One hour after injection, embryos were collected. Terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling (TUNEL) assay for the detection of apoptotic cells was performed with the in situ apoptosis detection kit (TAKARA, Shiga, Japan). Statistical analyses were performed using Student's t test or Welch's t test followed by F-test (P < 0.05, considered significant).

RNA isolation and real-time RT-PCR analysis

HaCaT cells were maintained in DMEM (Invitrogen) medium supplemented with 10% FBS (Invitrogen) at 37 C in 5% CO₂. The cells were plated at 2.0 × 10⁵ cells/well in a six-well plate 2 d before the transfection. Transfection was performed in triplicate with mouse an AR expression vector (40) using Transfast (Promega, Madison, WI) according to the manufacturer's instructions. After 5 h incubation in DMEM with 10% charcoal/dextran-treated FBS (Hyclone Laboratories, Logan, UT), cells were treated with 10⁻⁸ m 5α-dihydrotestosterone (DHT, Sigma) or dimethylsulfoxide. Twenty-four hours after the treatment, cells were collected and total RNA was isolated by ISOGEN (Nippon Gene, Toyama, Japan) and reverse transcribed by Superscript III reverse transcriptase (Invitrogen). Real-time PCR was performed using 7500 real-time PCR system (Applied Biosystems, Foster City, CA) with SYBER Green master mix according to the manufacturer's instructions (41). The sets of primer sequences were described before (42, 43).

Plasmid DNA and luciferase assay

The conserved regions of the Trp63 locus were reported previously (44). Genomic sequences of the p63 androgen response region [PARR; mouse (mm9) July 2007 chr16:25780114-25780472, human (hg19) February 2009 chr3:189484233-189484627)] and keratinocyte-specific enhancer [KSE; mouse (mm9) July 2007 chr16:25843798-25844131, human (hg19) February 2009 chr3:189551827-189552160] of Trp63 loci in mouse or human, obtained from Ensemble (www.ensembl.org) were submitted for the analysis by rVISTA (genome.lbl.gov/vista). DNA fragments containing the PARR and KSE of the murine Trp63 locus, obtained from a genomic C57BL/6 mouse DNA by PCR, were inserted into the pGL3 basic vector with herpes simplex virus thymidine kinase (HSV-TK) promoter (Promega). Primers were reported previously (44). The mouse AR expression vector was reported previously (40). HaCaT cells were plated at 5.0 × 10⁴ cells/well in a 24-well plate in DMEM supplemented with 10% FBS 2 d before the transfection. Cells were transfected in triplicate with reporter plasmids and mouse AR expression vectors or pCS2+MT as control vector using Transfast (Promega) according to the manufacturer's instructions. Cells were incubated and treated with DHT in the same conditions as described above, and luciferase activity was measured by chemiluminescence using the dual-luciferase reporter assay system (Promega) 24 h after DHT treatment. The values were normalized against Renilla luciferase activity. More than three independent experiments were performed. Statistical analyses were performed using either the Student's t test or Welch's t test followed by the F test (P < 0.05, considered significant).

Epididymal grafting

p63-null embryos were generated as described previously (23). Epididymides from p63 KO and control embryos were collected at E18.5 and grafted under the kidney capsule of athymic male nude mice (25). The phenotypes of p63^+/+ and p63^+/− epididymides were essentially identical. Therefore, we used p63^+/+ and p63^+/− epididymides as controls. The results presented here were obtained analyzing six p63 KO and three control epididymal grafts.

Spatiotemporal AR expression during murine WD development

It is well established that androgen signaling through AR is necessary and sufficient for WD stabilization (5, 45–47). However, the precise events of WD masculinization that are controlled by androgen are unknown. To obtain further understanding, we first examined the expression of AR in mouse embryos from E12.5 to E18.5. At E12.5, the approximate onset of androgen production in the male gonad (48), we detected AR in the periductal mesenchyme of both male and female WDs, in agreement with previous reports (Fig. 1, A and I) (49). AR expression in the mesenchyme expanded from the cranial to the caudal direction during masculinization, from E14.5 to E18.5 (Fig. 1, B–D and F–H) (50). Low levels of AR protein are present in the cranial portion of the WD in both E14.5 male and female embryos during WD stabilization and regression, respectively (Fig. 1, B and J). In male embryos, AR expression gradually increased in both the epithelium and mesenchyme during the morphogenesis of the WD from E16.5 to E18.5 (Fig. 1, C, D, G, and H). No immunostaining was detected in the negative control samples (Fig. 1M, data not shown).

Lack of androgen signaling results in prominent epithelial cell death and WD regression

Histological analyses revealed that the WD of AR KO males aberrantly regresses from cranial to caudal at E14.5, as is observed in control female embryos (Fig. 2, B and C). We analyzed cell proliferation and apoptosis in AR KO mice during this aberrant WD regression. Prominent apoptotic cell (TUNEL+) was observed in the WD epithelia of AR KO male and control female embryos compared with control male embryos at E14.5 (Fig. 2, D–F and J). In contrast, the percent of proliferating cells (BrdU+) in the cranial and middle WD epithelia was not different between control male and AR KO male WDs (Fig. 2, G, H, and K). The WD epithelia of control females and the caudal WD of AR KO males had fewer BrdU+ cells compared with control males (Fig. 2, I and K). No significant differences in the percent of BrdU+ mesenchymal cells were observed among these three groups (data not shown). These results indicate that AR regulates epithelial cell survival rather than proliferation during WD stabilization at E14.5.

Epithelially expressed AR is dispensable for WD stabilization during fetal development

To clarify whether epithelial AR prevents the apoptosis of WD epithelium, we conditionally ablated AR function in the epithelium using AP2α-Cre allele (32). Cre activity from this allele was detected predominantly in the WD epithelia; it was much less efficient in the mesenchyme (data not shown). In AP2α-Cre;AR^flox/Y mutant embryos, AR protein expression was undetectable in the majority of epithelial cells of the WD and the epididymis at E14.5 and E18.5 (Fig. 3, A–D). Despite the lack of AR in the epithelium, the WD persisted and coiled epididymides were present in mutant male embryos at E18.5 (Fig. 3, E and F). No significant differences were detected in the number of apoptotic or proliferating cells in mutant vs. control WD epithelia at E14.5 (Fig. 3, G and H). These results indicate that epithelial AR signaling is not required for survival of WD epithelia or stabilization, subsequent elongation, and coiling of the epididymides.

WD mesenchymal cells differentiate into smooth muscle actin- and smooth muscle myosin-positive cells during fetal stages (10). In other androgen target organs, such as prostate, it has been reported that mesenchymal smooth muscle differentiation is controlled by epithelial-derived signals (51). However, we detected expression of these differentiation markers in both control and AP2α-Cre;AR^flox/Y mutant epididymides at E18.5 (Fig. 3, I and J), indicating that WD mesenchymal cell differentiation occurs even in the absence of AR signaling to the WD epithelium.

Epithelially expressed AR contributes to postnatal epididymal cell differentiation

The adult epididymal epithelium consists of several types of cells that cooperate to provide the epididymal fluid microenvironment for sperm maturation (52–54). These epithelial cell types reportedly differentiate after birth (12, 55). Thus, we examined the requirement for epithelial AR in postnatal epididymal epithelial cell differentiation. AP2α-Cre;AR^flox/Y mutant epididymides were structurally regionalized into caput, corpus, and cauda, as seen in controls at postnatal day (P) 35 (Fig. 4, A and B). Histological analysis by toluidine blue staining revealed that control epididymides developed a typical pseudostratified epithelium containing principal, narrow, basal, and clear cells, whereas the mutant epididymal epithelia were composed of an irregular single layer of cuboidal cells (Fig. 4, C and D). AQP9 marks principal cells; its expression was detected in both the caput and cauda of control (Fig. 4, E and E′) but not mutant epididymides (Fig. 4, F and F′). p63 and K14 label basal cells, and expression of these markers was reduced in the caput region of the mutant epididymides relative to controls (Fig. 4, G–J). In cauda, the density of K14- and p63-expressing cells relative to other epithelial cells is rather higher in the mutant (Fig. 4, H′ and J′). No appreciable reduction of Foxi1 expression, a marker of narrow and clear cells, was observed in mutant epididymides (Fig. 4, K, K′, L, and L′). These results indicate that differentiation of caput and cauda principal and caput basal cells is dependent on androgen signaling to the epithelium, whereas that of narrow and clear cells is not.

p63 is necessary for the basal cell differentiation in the epididymis

Previous studies have suggested that p63 plays critical roles for basal cell differentiation; furthermore, basal cells are essential for proper differentiation of other cell types (23, 24, 28). We questioned whether the impaired epithelial cell differentiation marker expression in the postnatal caput epididymides of AP2α-Cre;AR^flox/Y mice was a consequence of reduced p63 expression. Hence, we examined p63 KO mouse epididymides. Because p63 KO mice die shortly after birth (23, 24), it was necessary to perform this experiment on transplanted tissue: epididymides from mutant and control embryos were dissected at E18.5 and grafted into the renal capsule of adult male nude mice. p63 KO-derived epididymides developed ductal structures comparable with those derived from controls, but the p63 KO epithelia consisted of a single layer of columnar cells, which is devoid of basal cell, at 1 month after the transplant (Fig. 5, A and B). Consistent with this finding, expression of K14 in p63 KO epididymides was markedly reduced relative to control epididymides (Fig. 5, C and D). Notably, AQP9 and Foxi1 were detected in both control and p63 KO epididymal epithelia, indicating proper differentiation of principal, clear, and narrow cells (Fig. 5, E–H). These results suggest that p63 is an essential factor for the development of epididymal basal cells and that basal cells are not required to direct or support the differentiation of principal or narrow/clear cells.

Androgen signaling possibly regulates the expression of p63

To explore the potential cell-autonomous regulation of basal cell differentiation by AR, we analyzed colocalization of p63 with AR in epididymides of which we induced mosaic AR expression. We used CAG-CreER^TM;AR^flox/Y mice with induction of Cre by TM treatment at E13.5, which results in mosaic recombination of AR throughout the embryos from approximately E14.0 to E15.0 (data not shown). The benefit of this experimental system is that the genetic mosaicism for AR loss of function permits us to examine the correlation between the presence of AR and p63 expression at the individual cell level. As for the possibility of mesenchymal AR deletion, the frequency of the recombination in the mesenchyme was rather low in this mutant (Supplemental Fig. 2A). In CAG-CreER^TM;AR^flox/Y mice, epididymal epithelial cells that retain AR immune reactivity were capable of expressing p63, but few AR negative cells expressed p63 (Fig. 6A and Supplemental Fig. 2B). These data suggest that AR is required, but not sufficient, for p63 expression in a cell-autonomous manner.

These results led us to question whether p63 could be a target of AR transcriptional regulation. p63 immune reactivity was barely detectable in the fetal epididymides, but its expression gradually increases after birth and was detectable throughout the epididymal epithelium at P5 (Fig. 6B). The HaCaT cell line is a spontaneously immortalized human keratinocyte line that expresses p63 (both TAp63 and ΔNp63) (56). Although this cell line does not express endogenous AR, it has been reported that transfection of exogenous AR gene into HaCaT cells could install AR transcriptional activity (57). Real-time RT-PCR analysis revealed that 5α-dihydrotestosterone (DHT) treatment promoted the transcription of ΔNp63 in AR-transfected HaCaT cells (Fig. 6C). We searched for AREs (13) in the evolutionally conserved region of the mouse p63 locus (Fig. 6D) and found two predicted AREs in a previously published candidate enhancer (44). We tentatively assigned this enhancer candidate region as the PARR. To assess whether PARR confers androgen responsiveness to a reporter construct, we transiently transfected HaCaT cells with a PARR/luciferase reporter construct with cotransfected AR expression vector and examined the transcriptional response to exogenous DHT. In AR-cotransfected cells, DHT treatment significantly activated the PARR/luciferase reporter (Fig. 6E).

To further evaluate the transactivating function of AR on PARR, we introduced a mutation into each of the putative AREs in PARR (Fig. 6F). Nucleotide substitutions at the ARE1 (AGAACA to AgttgA) resulted in a significant reduction in the transactivation of the mutant PARR in response to DHT, whereas mutation of another putative ARE [ARE2 (TGTTTTAAGTAGTAAATGTTGC to TtaTTTAAcTctTcAcTcgaGC)] did not (Fig. 6F). The combined mutation of both sites did not result in any further decrement in androgen response than observed with mutation of ARE1 (data not shown). The orthologous conserved region of the human Trp63 gene also contains putative AREs, and this region also conferred transcriptional response to DHT treatment in vitro (data not shown). These results suggest the possibility that AR positively regulates Trp63 expression via ARE1 in PARR.

The roles of androgen signaling on epithelial cell proliferation and apoptosis during WD stabilization

Studies of WD masculinization are expected to reveal molecular mechanisms that underlie the role of androgen on determining the fate of bipotential embryonic structures during sexual development. Most current knowledge as to androgen effects on the WD was obtained by comparing the response of male and female embryos to androgen or antiandrogen treatment. These previous studies showed prominent cell death and reduced cell proliferation in both the epithelium and mesenchyme during normal and induced WD regression (58, 59). However, treatment with antiandrogenic compounds may not block AR signaling completely and could also have other effects that are remote to androgen signaling.

To elucidate the roles of AR signaling definitively, we studied the cell proliferation and apoptosis in the WD of AR KO mice at E14.5. We detected abundant apoptotic cells in the WD epithelia but not in the mesenchyme in both AR KO male and control female of E14.5 embryos. Surprisingly, the proportion of proliferating cells in the caudal and middle regions of the epididymal epithelia was indistinguishable between AR KO male and control male embryos, whereas it was significantly reduced in the control female embryos (Fig. 2K). These results indicate that the regression of WD in AR KO males and control females do not occur by identical cellular mechanisms, consistent with a previous study that showed that the regression of the WD was delayed by 2 d in AR KO male embryos relative to control female embryos (9). AR KO male mice possess differentiated testes (36). These observations suggest that there may be additional testicular factors that promote WD masculinization or ovarian factors that promote regression of the WD (36). Some lumicrine testicular growth factors have been reported that are secreted by the testis into the duct and influence ductal epithelial cells in the adult caput epididymis (60–62). Whether such factors exist in the embryonic testis and influence the embryonic WD requires further analysis.

In the current study, androgen signaling regulates epithelial cell survival, rather than proliferation, during early WD stabilization (E14.5 in mice). However, epithelial cell proliferation in the caudal WD is suppressed in AR KO males, as seen in control females, suggesting that cell proliferation in this region is androgen dependent. Moreover, Welsh et al. (9) reported that apoptotic cells were present in the epithelium of the WD in control females but not in AR KO males at E15.5 and E16.5, when the ductal elongation and coiling normally occur. WD epithelial cell death and proliferation may be regulated by not only androgen signaling but also male/female-specific factor(s) in a temporospatial manner.

In previous studies, AR was not detected in the WD epithelium until E15.5, at the onset of ductal morphogenesis (49, 63, 64). We detected AR expression in the epithelium at E14.5, when the WD regression is already prominent in female embryos in the current study. This observation prompted us to analyze the function of epithelial AR during WD stabilization. We examined epithelial AR function during WD masculinization in AP2α-Cre;AR^flox/Y mutants. These mutants displayed WD stabilization, elongation, and coiling, suggesting that AR expressed in the WD epithelium is not required for these morphogenetic events. It further indicates that mesenchymal-epithelial interactions via mesenchymal AR are necessary for fetal stage WD masculinization. The current study is also supported by previous tissue recombination experiments (16). Recent molecular analyses of reproductive organ development indicate the presence of putative effector(s) that can potentially interact with hormonal signaling (65). Epidermal growth factor, keratinocyte growth factor, and inhibin-βA are paracrine factors expressed in the mesenchyme and shown to maintain epithelial proliferation during WD masculinization (63, 66, 67). However, the putative androgen effector(s) that inhibit epithelial cell death through mesenchymal AR have not been elucidated. Further experiments focusing on the mechanisms of epithelial cell survival may provide new information about effector(s) of androgen signaling.

The roles of epithelial AR signaling for the cell differentiation in the epididymis

The trial to analyze the function of epithelial AR for the postnatal male reproductive organs has been demonstrated using conditional AR KO system. Loss of epithelial AR expression caused impaired secretory function in the prostate and epididymis resulted in male infertility (18, 68). In these systems, loss of AR function is prominent from P7, and thickening of mesenchyme is noted as a developmental defect in the epididymides. In the current study, the epithelia in AP2α-Cre;AR^flox/Y epididymides showed presumably more severe phenotypes also in the epithelia, cuboidal single-layered epithelium, than those of postnatal deletion of AR. These observations may indicate that epithelial AR contributes to epithelial cell differentiation in early postnatal or embryonic stages. The previous report also showed increment of the density of basal cells, other than the secretory dysfunction, in the epithelial AR-deficient prostate. It has been speculated that epithelial AR regulates basal cell proliferation directly or indirectly (68). It is noteworthy that mutant cauda epididymis did not display decrement of p63 in the current study, but instead it showed higher density of the p63+ basal cells compared with the control (Fig. 4, J and J′). In that aspect, cauda epididymis might share some common epithelial AR-dependent regulatory mechanism with those of prostate to regulate basal cell proliferation.

p63 plays crucial roles in basal cell differentiation in skin, vagina, trachea, and prostate (23–25, 27). We further investigated p63 function in the developing epididymis using p63 KO mice and assessed epididymal epithelial cell differentiation. The single-layered epithelia of p63-deficient epididymides lacked basal cells, but expression of molecular markers for principal, narrow, and clear cells were intact. These results imply that p63 is essential for the basal cell differentiation in epididymis and that basal cells are not required for the differentiation of principal and narrow/clear cells. A previous study of cell morphology in the rat epididymis also suggests that basal cells may not be a stem cell population but rather that principal and basal cells originate from a common precursor (12). AP2α-Cre;AR^flox/Y mutant epididymides showed impaired basal and principal cell differentiation. In addition, not all AR+ epithelial cells express p63 in CAG-CreER^TM;AR^fox/Y mutant epididymides, indicating that AR expression is necessary (but not sufficient) for p63 expression cell autonomously. Accordingly, principal and basal cells may originate from a common precursor and epithelial AR may contribute to their establishment and/or maturation. Detailed analyses of epididymal epithelial cell lineages and the regulation of their behavior and fate are required to gain insight into the function of epithelial AR on principal and basal cell differentiation.

Possible regulation of p63 expression by AR

Functional analysis of an evolutionally conserved genomic region of mouse p63 revealed that there are three candidate cis-acting regulatory elements (44). One of the elements located 160 kb downstream of the first p63 exon has been reported to function as a KSE (44). This KSE did not confer a response to androgen treatment in our experimental conditions (data not shown). In contrast, the other element located 96 kb downstream of the p63 transcriptional initiation site contains an ARE sequence (ARE1) and conferred androgen responsiveness (Fig. 6F); hence, the element is termed as PARR. The corresponding conserved noncoding region in the human TRP63 locus also responded to androgen in our system (data not shown). Our in vitro studies suggest that androgen signaling regulates p63 expression through tissue and/or stage specific regulatory element(s). Consistent with this, AR binding sites in cell lines and adult mouse epididymis were globally mapped recently (69–71), and these AR occupancy data support stage- and tissue-specific AR functions that are likely influenced by chromatin conformation, coregulator recruitment, and cellular environment. Applying these methods to developing androgen target tissues may shed light on the developmental role of AR and its cooperating factors.

The purpose of the current study was to obtain a better understanding of the cellular processes and molecular mechanisms underlying androgen actions during WD/epididymal development. Although the roles of androgen have been studied based on the (anti) androgen treatment or analyses using spontaneous Tfm mutants in rodents, few studies have been done with tissue- or stage-specific AR KO mice. In summary, we reported that epithelial apoptosis is a major cellular process repressed by androgen signaling during WD stabilization. Furthermore, epithelial AR is not required for survival of the WD epithelia but is essential for subsequent epididymal epithelial cell differentiation. In the absence of epithelial AR, we did not detect any increment of epithelial apoptosis; however, the mature epithelia were devoid of basal and principal cells. Of note is the fact that p63, which plays crucial roles in basal cell differentiation, is not expressed in the AP2α-Cre;AR^flox/Y mutant caput epididymis. Our in vitro experiments suggest that p63 is a candidate target of AR signaling. To our knowledge, this could be the first genetic study analyzing the epithelial AR function in epididymal cell differentiation.