Targeted Loss of Androgen Receptor Signaling in Murine Granulosa Cells of Preantral and Antral Follicles Causes Female Subfertility¹

Abstract

Ovarian granulosa cells display strong androgen receptor (AR) expression, suggesting a functional role for direct AR-mediated actions within developing mammalian follicles. By crossing AR-floxed and anti-Müllerian hormone (AMH)-Cre recombinase mice, we generated granulosa cell-specific androgen receptor knockout mice (GCARKO). Cre expression, assessed by lacZ activity, localized to 70%–100% of granulosa cells in most preantral to antral follicles, allowing for selected evaluation of granulosa cell AR-dependent actions during follicle development. Relative to wild-type (WT) females, GCARKO females were subfertile, producing a 24% reduction in the number of litters (P < 0.05) over 6 mo and an age-dependent decrease in total number of pups born, evident from 6 mo of age (P < 0.05). Follicle dynamics were altered in GCARKO ovaries at 3 mo of age, with a significant reduction in large preantral and small antral follicle numbers compared to WT ovaries (P < 0.05). Global premature follicle depletion was not observed, but increased follicular atresia was evident in GCARKO ovaries at 6 mo of age, with an 81% increase in unhealthy follicles and zona pellucida remnants (P < 0.01). Cumulus cell expansion was decreased (P < 0.01) and oocyte viability was diminished in GCARKO females, with a significant reduction in the percentage of oocytes fertilized after natural mating and, thus, in the rate of progression to the two-cell embryo stage (P < 0.05). In addition, compared with age-matched WT females, 6-mo-old GCARKO females exhibited significantly prolonged estrous cycles (P ≤ 0.05), suggesting altered hypothalamic-pituitary-gonadal feedback signaling. In conclusion, our findings revealed that selective loss of granulosa cell AR actions during preantral and antral stages of development leads to a premature reduction in female fecundity through reduced follicle health and oocyte viability.

androgen receptor, female reproduction, granulosa cell, ovarian function

Introduction

The androgen receptor (AR) gene, located on the X chromosome, has a proven role in female reproductive physiology [1, 2]; however, the specific mechanisms involved remain unclear. Expression of AR has been shown throughout the hypothalamic-pituitary-ovarian axis [3, 4], implying that AR-mediated actions have the potential to impact on all key processes involving the female reproductive tract. The consistent expression of AR in ovaries of many mammalian species (rat [3], pig [5], cow [6], primate [7], human [8]) and the influence of AR activity on key genes involved in follicle growth (follicle-stimulating hormone receptor and insulin-like growth factor 1 receptor [9]) strongly support a role for AR-mediated actions in follicle development. The oocyte, granulosa cells, theca cells, and cumulus cells all express AR, with expression predominantly in granulosa cells. Expression is highest in preantral/early antral follicles and gradually decreases as follicles mature [6, 10–13]. The role of the follicle as the principal source of estrogen biosynthesis, together with testosterone (T) being an obligate steroidogenic precursor of estradiol (E2) [14], means that the follicular cells are regularly exposed to high local concentrations of potent androgens. Despite the strong support of a functional role for local AR-mediated actions throughout follicle growth, pharmacological studies have provided conflicting findings, with androgens shown to inhibit and/or stimulate follicular development [15–18]. This ambiguity is mainly due to use of aromatizable androgens (T, androstenedione [A₄], dehydroepiandrosterone [DHEA]) that can be converted with a variable extent into corresponding estrogens, E2 and estrone (E1), which can exert estrogen receptor (ER)-mediated effects [1]. Furthermore, although the standard method to avoid this is to use the nonaromatizable alternative, dihydrotestosterone (DHT), this method is also flawed, as DHT is irreversibly metabolized to 3beta-diol, which interacts with ERβ, allowing it to manifest ER-mediated estrogen-like effects [19, 20]. Therefore, pharmacological approaches relying on endogenous steroids have difficulty in defining the precise AR or ER molecular mechanism(s) involved.

Although androgen-resistant females cannot be bred by natural mating due to sterility of the hemizygous AR knockout males, the development of genetic mouse models of androgen insensitivity created by functional inactivation of the AR in females using the Cre/LoxP system have confirmed a role for AR-mediated actions in female reproductive function (reviewed in [1, 2]). Global androgen receptor knockout (ARKO) females are subfertile with fewer pups per litter and exhibit increased follicular atresia and defective ovulation [9, 21–23]. However, the specific AR-mediated mechanisms leading to this subfertility remain to be fully identified.

Paracrine interactions between the oocyte and surrounding granulosa cells are critical for normal ovarian follicle development. Within the ovary, granulosa cells display strong AR expression and are exposed to potent endogenous AR agonists, such as T and DHT. Furthermore, expression levels of Kit ligand and bone morphogenetic protein 15 gene expression, both of which are involved in the oocyte-granulosa cell regulatory loop, are reduced in ARKO ovaries [22]. Therefore, granulosa cells may constitute a major site of AR-mediated action in the follicle. Recently, the Cre/LoxP system was used to target the homozygous deletion of AR protein in the granulosa cells [24]. This model used anti-Müllerian hormone receptor type 2 Cre (Amhr2Cre) gene promoter to direct transgenic Cre expression in granulosa cells. Females from this model exhibited nearly all of the fertility phenotypes associated with the global ARKO mice [24]. This led the authors to conclude that local AR signaling within the granulosa cells of the ovary is the primary site of androgen actions involved in regulating ovarian function, rather than possible AR-mediated actions in other sites such as the pituitary and/or hypothalamus. However, nonspecific expression of Amhr2Cre has been detected in other mouse models [24–26]. In particular the possibility of uterine [24, 25] and/or oocyte and theca cell [25] expression of Amhr2Cre raises the issue that nongranulosa cells may play a role in these findings.

Additionally, many studies support a central role for androgen signaling in the regulation of female fertility [27–34]. Indeed, findings from reciprocal ovary transplantation between control and global ARKO females provide strong evidence of a role for neuroendocrine and ovarian AR-mediated actions in maintaining female fertility [28]. Therefore, to determine the role of AR specifically in granulosa cells of developing follicles, we have developed a novel transgenic granulosa cell-specific ARKO (GCARKO) female mouse model using granulosa cell-specific transgenic Cre directed via the anti-Müllerian hormone gene promoter (AMH-Cre). Detailed studies in rodents have shown that AMH expression is highest in granulosa cells of preantral and antral follicles and gradually diminishes in the final stages of follicle growth [35, 36]. Hence, we predicted that our AMH-Cre-mediated GCARKO model would provide a unique in vivo platform to determine the importance of AR actions in granulosa cells during preantral and antral stages of follicle development, without the confounding effects of ectopic AR disruption in other ovarian cells or reproductive tissues. Our findings show that classical genomic AR-mediated signaling within the granulosa cells of preantral- and antral-stage follicles is vital for normal female fertility.

Materials and Methods

Mice

Mice were maintained under standard housing conditions (ad libitum access to food and water in a temperature- and humidity-controlled, 12L:12D environment) at the ANZAC Research Institute. All procedures were performed under ketamine/xylazine anesthesia. All procedures were approved by the Sydney South West Area Health Service Animal Welfare Committee within National Health and Medical Research Council guidelines for animal experimentation.

Generation of GCARKO Mice

To generate GCARKO mice, transgenic AMH-Cre mice (Cre shown to be expressed in granulosa cells; The Jackson Laboratory, Bar Harbor, ME) [37] were mated with AR-floxed mice in which exon 3 is flanked by loxP sites [38] as previously described [39]. Due to the X-linked Ar gene, females heterozygous or homozygous for AR floxed were bred with hemizygous AMH-Cre^(+/–);Ar^wt/Y males. To establish AMH-Cre females homozygous for AR floxed (Ar^flox/flox), breeding strategies used AMH-Cre^(+/-);Ar^wt/ARfloxed or AMH-Cre^(+/-);Ar^flox/flox females mated with non-AMH-Cre Ar^flox/Y males. Global female homozygous ARKO mice were generated by crossing AR-floxed mice with CMV-Cre as previously described [23].

DNA and RNA Extraction, Genotyping, and RT-PCR

Genomic DNA isolated from toe clip or tail biopsy, as previously described [40], was used as a template for PCR genotyping to detect rearrangements in the mouse Ar gene, as previously described [23]. Forward PCR primers upstream of the first LoxP site within mouse AR exon 3 (AREx3-F, CTTCTCTCAGGGAAACAGAAGT) and within NEO cassette (ARNeo-F, TAGATCTCTCGTGGGATCATTG) were used with a common reverse primer located within intron 3 (AR-R, GGGAGACACAGGATAGGAAATT). Two product sizes were obtained, 613 bp for intact Ar and 289 bp for AR-floxed Ar (Fig. 1A). Mice containing the CMV- or AMH-Cre gene were detected using primers and PCR conditions as described [41]. Global ARKO males and females were distinguished using primers and PCR conditions for the mouse Y chromosome sry gene as previously described [23, 38].

Fig. 1

Characterization of GCARKO mice. A) Representation of PCR genotyping by which all mice were genotyped using genomic DNA. The intact AR exon 3 product was 613 bp, and the floxed AR exon 3 product was 289 bp. Cre had a product of 213 bp. B) Representation of RT-PCR AR analyses using cDNA extracted from WT, ARKO, and GCARKO ovary, uterus, brain, and pituitary and WT and GCARKO granulosa cells. The intact AR exon 3 product is 288 bp, and the excised AR exon 3 product is 171 bp. Mouse β-actin, used as an internal control, had a product of 431 bp. C) AR immunohistochemistry in WT and GCARKO ovaries. Sections were counterstained with Harris hematoxylin. Original magnification ×2.

Confirmation of the genotype of GCARKO, ARKO, and wild-type (WT) female mice was confirmed by RT-PCR on the RNAs extracted from ovary, uterus, brain, and pituitary of 12-wk-old mice (n = 3 per genotype). Reverse-transcription was performed using SuperScript III First-Strand Synthesis System (Invitrogen); subsequent PCR was carried out and products analyzed by electrophoresis on a 1.5% agarose gel [23]. RNA was also extracted from granulosa cells of WT and GCARKO ovaries. Females were treated with 10 IU pregnant mare serum gonadotropin (Folligon; Intervet), and 46 h later ovaries (ovaries pooled from three females per genotype) were removed and transferred to a 100-mm cell culture dish containing Leibovitz L-15 medium (Invitrogen) supplemented with 0.3% bovine serum albumin (BSA; Sigma), 50 U/ml penicillin, and 50 μg/ml streptomycin. Antral follicles were punctured with 27-gauge needles, the granulosa cells were collected and spun down, the supernatant was removed, and the final cell pellet was frozen at −80°C until RNA extraction. Total RNA was extracted using RNeasy Micro Kit (Qiagen). The 5′-GGA CAG TAC CAG GGA CCA T-3′ and 5′-CCA AGT TTC TTC AGC TTA CGA-3′ primers identified intact and excised exon 3 Ar. The conditions for the PCR were as follows: 94°C for 2 min, followed by 35 cycles of 94°C for 5 sec, 60°C for 15 sec, 72°C for 30 sec, and final extension at 72°C for 2 min. Two product sizes were obtained, 288 bp for intact Ar and 171 bp for Cre-mediated exon 3-excised Ar. Primers specific for mouse β-actin were used as an internal control (product size 431 bp).

Specimen Collection

WT and GCARKO dissected ovaries, oviduct, and uteri were weighed, fixed in 4% paraformaldehyde at 4°C overnight, and stored in 70% ethanol before histological processing. One ovary from each mouse was randomly selected and processed through graded alcohols into glycol methacrylate resin (Technovit 7100; Heraeus Kulzer). Ovaries were serially sectioned at 20 μm, stained with periodic acid-Schiff, and counterstained with hematoxylin. Oviducts were dehydrated in a graded series of ethanols, cleared in xylene, and embedded in paraffin. Sections (5 μm) were stained with hematoxylin and eosin and analyzed using a light microscope.

X-Gal Staining and LacZ Activity Quantification

Functional recombinase activity of transgenic Cre in granulosa cells was analyzed using transgenic AMH-Cre mice crossed with R26R (ROSA) reporter mice [42]. Expression of functional Cre recombinase by lacZ activity in tissues was detected by whole-mount galactosidase staining of dissected ovaries. Briefly, organs were prefixed in 2% paraformaldehyde for 1 h at room temperature, then subjected to 5-bromo-galactopyranoside (X-gal; 1 mg/ml) staining for 4 h at 37°C, followed by rinsing with PBS and postfixation overnight in 4% paraformaldehyde at 4°C. Stained ovaries were photographed and then paraffin embedded and sectioned at 5 μm before counterstaining with Nuclear Fast Red Stain.

To quantify LacZ activity, 9 or 10 follicles from three different ovaries were randomly selected per developmental stage, and the total percentage of small preantral, large preantral, small antral, and large antral follicles to exhibit lacZ staining was calculated. Then the average percentage of granulosa cells/follicles stained positively for LacZ was calculated by dividing the total number of positively stained granulosa cells by the total number of granulosa cells for each follicle and then working out the average for each developmental stage.

Androgen Receptor Immunohistochemistry

Androgen receptor immunohistochemistry was carried out as previously described [23]. Ovaries from 3-mo-old mice (n = 3 per genotype) were collected, fixed in 4% paraformaldehyde, dehydrated in alcohols, cleared in xylene, embedded in paraffin, and serially sectioned (5 μm). Ovarian sections were subjected to microwave (600 watt) antigen retrieval in 10 mM citrate buffer (pH 6.0) for 15 min. Endogenous peroxidase activity was blocked by incubating sections in 2% hydrogen peroxide for 5 min. To block nonspecific binding, sections were incubated in 1% BSA for 1 h at room temperature followed by goat serum for 30 min. Sections were then incubated with an anti-AR antibody (N-20; sc-816; Santa Cruz Biotechnology Inc.) diluted 1:50 in 1% BSA/PBS for 1 h at 37°C. Antibody binding was visualized using the Vectastain Elite ABC Kit followed by color development using liquid DAB+ (Dako) according to the manufacturer's instructions. Sections were briefly counterstained with Harris hematoxylin, dehydrated through ethanol series and xylene, and mounted.

Assessment of Estrous Cycle and Fertility

Estrous cycle stage was determined in virgin females at 3 and 6 mo of age (3 mo: WT [n = 7], GCARKO [n = 7]; 6 mo: WT [n = 5], GCARKO [n = 5]) by light microscope analysis of vaginal epithelial cell smears collected daily (1000 h) in 20 μl sterile PBS that were transferred to glass slides, air dried, and stained with 0.05% Trypan Blue for microscopy.

To estimate natural fertility, 6- to 8-wk-old females (WT, n = 6; GCARKO, n = 7) were continuously mated with an individual fertile male (at least 8 wk old) for a 6-mo period. Cages were monitored daily and the number of pups and litters recorded.

Classification and Enumeration of Follicles

The follicle classification system used was based on the system used by Myers et al. [43]. Briefly, follicles classified as small preantral follicles contained an oocyte with one and a half to two layers of cuboidal granulosa cells, large preantral follicles were classified by having an oocyte surrounded by more than two and up to five layers of cuboidal granulosa cells, small antral follicles contained an oocyte surrounded with more than five layers of cuboidal granulosa cells and/or one or two small areas of follicular fluid, large antral follicles contained a single large antral cavity, and preovulatory follicles possessed a single large antrum and an oocyte surrounded by cumulus cells at the end of a stalk of mural granulosa cells. Corpus lutea were identified by morphological properties consistent with luteinized follicles and by being visible throughout several serial sections. Zona pellucida remnants (ZPRs) (representing end-stage atretic follicles) were recognized as collapsed ZPR structures stained brightly pink with periodic acid-Schiff, as previously described [43].

Ovaries were collected at the beginning of the diestrus stage for follicle enumeration. Follicles were counted on all serial sections throughout each ovary using an Olympus microscope with Stereo Investigator software (MicroBrightField). Total numbers of growing follicles per ovary at different developmental stages for each of the genotypes (three or more ovaries counted per genotype per time point) was determined as previously described [23, 43]. Follicles were classified as unhealthy if they contained a degenerate oocyte and/or >10% of the granulosa cells were pyknotic in appearance as previously described [23, 44]. For all histological analyses, repetitive counting of follicles was avoided by only counting follicles containing an oocyte with a visible nucleolus. To avoid bias, all ovaries were analyzed without knowledge of genotype.

In Vivo Expansion of the Cumulus

Cumulus expansion in vivo was assessed by morphological examination of postovulatory oocyte cumulus complexes (OCC) before fertilization. Females (3 mo old, five or more per genotype) were treated with 10 IU pregnant mare serum gonadotropin and 48 h later with 10 IU human chorionic gonadotropin (hCG). OCC were recovered from the oviducts of superovulated mice 16 h post-hCG treatment. Degree of expansion was evaluated morphologically according to a subjective scoring system: 0 (indicating no detectable response) to 4 (indicating the maximum degree of expansion), as previously described [45]. The experiments were independently repeated at least three times.

Embryo Culture

As a reduction in cumulative pups per female was evident after 4 mo of breeding (6 mo of age), embryo viability was assessed in younger females. Four-mo-old WT and GCARKO females (six or more per genotype) were naturally mated with an individual fertile WT male. Cages were monitored daily, and on the identification of a copulatory plug, mice were euthanized, and OCC and embryos were recovered from the oviducts, counted, and placed in M2 medium (Sigma). Fertilization was identified by the extrusion of the second polar body before embryos were cultured overnight at 37°C with 5% CO₂ in EmbryoMax KSOM with 1/2 Amino Acids medium (Millipore) to assess the number of embryos reaching cleavage to the two-cell stage.

Quantitative Real-Time RT-PCR

Total RNA was extracted from whole ovaries at the diestrus stage (five per genotype) using Tri Reagent (Sigma) according to the manufacturer's protocol. Reverse-transcription was performed with total RNA using SuperScript III First-Strand Synthesis System (Invitrogen). Quantitative real-time PCR analysis of ovarian cDNAs was performed on a Corbett RotorGene 6000 (Corbett Research) using the SensiMix SYBR Kit (Bioline) as recommended and previously described [23, 28]. A standard curve was generated for each gene from five serial dilutions of purified PCR product from the same primers designed for quantitative PCR using Wizard DNA Clean-Up System (Promega). Standards (dilutions used for each gene were 10⁻²–10⁻⁶) were assigned an arbitrary value, and mean relative mRNA expression of samples determined in duplicate were standardized to mouse Rpl19 levels (housekeeping gene), as previously described [23, 28]. No template controls, substituting water for cDNA, and a negative reverse transcription were included in each run. Gene expression was detected with the following primer pairs and annealing temperatures: Fshr (ACACAACTGTGCATTCAACGG and GACTTGTTGCAAATTGGATGA, 55°C) to amplify a region spanning exons 7–10 [46] present in all Fshr mRNA variants found in mouse ovary [47], Kitl (TCATGGTGCACCGTATCCTA and CCTTGGCATGTTCTTCCACT, 55°C) to detect transcripts encoding both membrane and soluble Kit-L isoforms, and Igfr1 (GCGGCCCTCCTTCCTGGAGA and GGGAGGCTGAGGAGGCCGAA, 65°C). There were no significant differences in Rpl19 mRNA levels between treatment groups (P > 0.05). Reaction steps were 15 min at 95°C, followed by 40 cycles of denaturation at 95°C for 30 sec, annealing at 55°C for 30 sec, and extension at 72°C for 30 sec.

Hormone Assays

Blood was collected by cardiac exsanguination under ketamine/xylazine anesthesia, rested at room temperature for 20 min, and centrifuged at 5000 rpm for 5 min, and then collected serum was stored at −20°C. Serum samples (3 mo: WT [n = 10], GCARKO [n = 8]; 6 mo: WT [n = 7], GCARKO [n = 6]) for hormone assay were obtained at diestrus as determined by light microscopic analysis of vaginal epithelial cell smears collected daily. Mouse serum luteinizing hormone (mLH) was determined using a species-specific immunofluorometric assay, as previously described [23, 48], with modification. The capture antibody used was the anti-LH antibody (5303 SPRN-1; Medix Biochemica), and the detection antibody was the anti-LH antibody (MAb 518B7; supplied by Dr. J Roser, Dept of Animal Science, UC Davis) [49] directly labeled with a Europium chelate using the DE-LFIA Eu-labeling kit (Perkin Elmer) as per suppliers methodology. The mLH detection limit was 0.02 ng/ml, the quantification limit was 0.05 ng/ml, and the within-assay QC was 6.8% at low (0.25 ng/ml), 4.7% at mid (0.49 ng/ml), and 7.4% at high (1.18 ng/ml) range. Mouse serum follicle-stimulating hormone (FSH) was determined using a species-specific immunofluorometric assay, as previously described [23, 50]. All immunoassays were performed in a single batch.

Statistical Analysis

Statistical analysis was performed using NCSS 2007 software (NCSS Statistical Software). Unless otherwise stated, all results were expressed as mean ± SEM. Data that were not normally distributed were transformed using a log transformation. Statistical differences were tested by t-test or ANOVA with post hoc test using a Fisher Least Significant Differences Multiple-Comparison Test. The percentage of females to produce seven litters was analyzed by Fisher Exact Test, while the time each breeder took to produce their 30th pup and sixth litter was analyzed with the use of Kaplan-Meier analysis using logrank test to compare groups. All parametric tests were confirmed by nonparametric equivalent tests. P ≤ 0.05 was considered statistically significant.

Results

Verification of GCARKO Female Mice

To confirm the tissue-specific excision of exon 3, RNA extracted from the ovary, granulosa cells, uterus, brain, and pituitary was subjected to RT-PCR. Global ARKO mice exhibited only Ar with excised exon 3 and WT exhibited only Ar with intact exon 3. GCARKO uteri, brain, and pituitaries showed only Ar with intact exon 3, while ovaries exhibited Ar with both intact and excised exon 3, consistent with only some cells (granulosa cells) within the ovary expressing Cre (Fig. 1B). Granulosa cells exhibited Ar with both intact and excised exon 3, indicating that while the deletion of functional AR was specific to granulosa cells, the deletion was not present in all granulosa cells. AR immunohistochemistry showed the normal pattern of AR protein expression in follicles of both genotypes (Fig. 1C), confirming that the in-frame deletion of exon 3 did not disrupt truncated AR protein production.

Expression of functional Cre recombinase by lacZ activity in the ovary was detected by whole mount-galactosidase staining of dissected ovaries (Fig. 2A). LacZ staining revealed specific and strong Cre expression in granulosa cells, with no staining in the oocyte or theca cells, consistent with the endogenous AMH expression pattern [35, 36, 51]. Staining was present in large preantral and antral follicles (Fig. 2B), but was at lower or undetectable levels in granulosa cells of primordial, primary, and small preantral follicles. None of the small preantral follicles assessed exhibited LacZ staining, while large preantral, small antral, and large antral follicles exhibited on average 100% ± 0%, 91.6% ± 5.6% and 69.7% ± 10.6% of granulosa cells per follicle stained positively for LacZ, respectively (Fig. 2C).

Fig. 2

LacZ activity in GCARKO ovaries. A) LacZ expression after Cre recombination detected by X-Gal staining of WT and GCARKO ovaries. GC, granulosa cells; LA, large antral; LP, large preantral; O, oocyte; SA, small antral; SP, small preantral; TC, theca cells. Arrow shows staining in granulosa cells. B and C) Black bars represent GCARKO. B) Percentage of follicles at each developmental stage exhibiting LacZ staining. n = 9–10 follicles per developmental stage. C) Average percentage of LacZ-stained granulosa cells per follicle at each developmental stage. n = 9–10 follicles per developmental stage. Data are the mean ± SEM.

Fertility of GCARKO Females

GCARKO females were subfertile. Although litter size did not differ between genotypes (GCARKO, 8.4 ± 0.4; WT, 9.2 ± 0.8; Fig. 3A), there was an age-dependent reduction in cumulative pups per month, evident after 4 mo of breeding (6 mo of age) (P < 0.05; Fig. 3B). GCARKO females exhibited a 32% reduction in the number of cumulative pups produced over the 6-mo breeding trial period (P < 0.05). GCARKO female breeders took longer than WT to produce 30 pups (median GCARKO = 124 days, WT = 88 days, P < 0.05; Fig. 3C). GCARKO females produced significantly fewer litters per female compared to WT females (GCARKO, 5.3 ± 0.5; WT, 7 ± 0.0; P < 0.05; Fig. 3D). 100% of WT breeders produced seven litters during the 6-mo breeding trial, while less than 20% of GCARKO breeders produced seven litters (P < 0.01; Fig. 3E). Furthermore, GCARKO females took longer to produce their sixth litter, with a median time of 22 wk, compared to WT breeders, with a median time of 18.9 wk (P < 0.05; Fig. 3F).

Fig. 3

Fertility. A, D, and E) White bars, WT; black bars, GCARKO. B, C, and F) White squares, WT; black circles, GCARKO. In B, dotted line with gray triangles indicates historical global ARKO data [23]. A) Average pups per litter. B) Average cumulative number of pups per month over a 6-mo period of continual mating (ANOVA, P < 0.05). C) Percentage of females to have produced their 30th pup and median time in days to 30th pup (Kaplan-Meier analysis, P < 0.05). D) Average litters per female (ANOVA, P < 0.05). E) Percentage of females to have produced seven litters (Fisher Exact Test, P < 0.01). F) Percentage of females to have produced their sixth litter and median time in days to sixth litter (Kaplan-Meier analysis, P < 0.05). Six or more per genotype. Data are the mean ± SEM. Different lowercase letters denote statistically significant differences.

Estrous Cycle Analysis in GCARKO Females

All 3- and 6-mo-old WT and GCARKO female mice displayed estrous cyclicity. There was no difference in cycle length (average cycle length in days: GCARKO, 5.6 ± 0.6; WT, 4.9 ± 0.4; Fig. 4A) at 3 mo of age. However, at 6 mo of age GCARKO females had significantly longer estrous cycles compared to WT females (average cycle length in days: GCARKO, 6.6 ± 1.9; WT, 4.6 ± 0.3; P ≤ 0.05; Fig. 4, A and B).

Fig. 4

Estrus cycling and hormone levels. A, C, and D) White bars, WT; black bars, GCARKO. B) White squares, WT; black circles, GCARKO. A) Average cycle length (ANOVA, P ≤ 0.05). n = 5–7 per genotype. Different lowercase letters denote statistically significant differences. B) Estrous cycle pattern in representative WT and GCARKO females. P, proestrus; E, estrus; M, metestrus; D, diestrus. C) LH serum hormone levels at the diestrus stage in WT and GCARKO mice at 3 and 6 mo of age. Data are the mean ± SEM. n = 6–10 per genotype. D) FSH serum hormone levels at the diestrus stage in WT and GCARKO mice at 3 and 6 mo of age. Data are the mean ± SEM. n = 6–10 per genotype.

Effect of GCAR Deficiency on Serum FSH and LH Levels

At 6 mo of age there was a non-significant trend towards an increase in serum LH levels in GCARKO females compared to WT females (GCARKO, 0.46 ± 0.1 ng/ml; WT, 0.28 ± 0.03 ng/ml; P = 0.09; Fig. 4C). Levels of serum LH at 3 mo of age (Fig. 4C) and FSH at 3 or 6 mo of age (Fig. 4D) at diestrus were not significantly different between the genotypes.

Phenotype of GCARKO Females

There was no overt effect of genotype on the macroscopic appearance of the ovaries and uterus (Fig. 5A). Analysis of ovary (Fig. 5B), uterine (data not shown), and body weights (data not shown) revealed a significant effect of age (P < 0.01), but no effect according to genotype and no interaction (six or more per genotype per age group). There was no overt effect of genotype on the microscopic appearance of oviducts. WT and GCARKO isthmus morphology was normal (data not shown), and the ampulla of both WT and GCARKO females exhibited long, longitudinal mucosal folds displaying secretory cells, as expect at diestrus (Fig. 5C).

Fig. 5

Ovary and uterus phenotype, ovary weight, and ovarian follicle and corpora lutea numbers. A) Representation of WT and GCARKO female reproductive tracts. Original magnification ×0.67. B and D) White bars, WT; black bars, GCARKO. B) WT and GCARKO ovary weights at 3 and 6 mo of age. Data are the mean ± SEM. n = 6–14 per genotype. C) Histological sections of WT and GCARKO oviducts (original magnification ×40). Both WT and GCARKO ampulla exhibit long longitudinal mucosal folds displaying secretory cells (cell protrusions shown by arrow). D) Average number of growing follicles and corpus lutea per ovary at the diestrus stage at 3 and 6 mo of age (ANOVA, P < 0.05). Numbers represent total counts from serially sectioned ovaries. n ≥ 3 per genotype per age. Different lowercase letters denote statistically significant differences. Histological sections of WT and GCARKO ovaries and OCC in WT (i) and GCARKO (ii) preovulatory follicles. Original magnification ×2 and ×20 (i and ii).

Effect of GCAR Deficiency on Ovarian Histology and Follicle Health

Histological sections of ovaries showed that the full range of expected stages of follicle development were present, with overall normal morphology regardless of genotype (Fig. 5D). Cumulus granulosa cells maintained close contact with the oocyte throughout follicle development (Fig. 5D, i and ii). Quantitative analysis showed that at 3 mo of age, large preantral and small antral follicle numbers, at the diestrus stage of the estrous cycle, were significantly reduced by 36% and 57%, respectively (P < 0.05), in GCARKO female compared with WT mice (Fig. 5D). There was no significant difference in follicle population numbers between genotypes in ovaries from 6-mo-old mice (Fig. 5D). Corpus lutea numbers did not differ according to genotype or age (Fig. 5D). The number of morphologically unhealthy follicles and zona pellucida remnants within ovaries did not differ between genotypes at 3 mo of age. However, at 6 mo of age GCARKO ovaries had a significant increase in numbers of unhealthy follicles and ZPR counts (GCARKO, 1640.7 ± 881.7; WT, 907.7 ± 115.9; P < 0.01; Fig. 6A).

Fig. 6

Ovarian health. A) Average number of unhealthy follicles and ZPRs per ovary at 3 and 6 mo of age at the diestrus stage (ANOVA, P < 0.01). n ≥ 3 per genotype per age. Different lowercase letters denote statistically significant differences. Histological ovarian sections of WT and GCARKO ovaries are shown below. Arrowheads, ZPR. Original magnification ×10. B) Real-time RT-PCR analysis of the expression levels of Kitl, Igfr1, and Fshr. n = 5 per genotype. Data are the mean ± SEM.

Effect of GCAR Deficiency on Ovarian Gene Expression

Real-time RT-PCR on ovaries collected at 3 mo of age at diestrus revealed that expression levels of Kitl, Igfr1, or Fshr did not differ according to genotype (Fig. 6B).

Effect of GCAR Deficiency on Hyperstimulated and Natural Ovulation, Cumulus Expansion, Fertilization, and Early Embryo Development

The number of oocytes collected from the oviducts of 3-mo-old superovulated females (GCARKO, 21.3 ± 6.1; WT, 21.3 ± 4.9) did not differ significantly between genotypes (data not shown). The degree of cumulus expansion observed in the collected OCC was significantly decrease in GCARKO OCC compared to WT OCC (GCARKO, 3.2 ± 0.1; WT, 3.5 ± 0.1; P < 0.01; Fig. 7A). The number of oocytes collected from the oviducts of 4-mo-old naturally mated females (GCARKO, 10.4 ± 1.3; WT, 9 ± 1.1) did not differ significantly between genotypes (Fig. 7B). Viability of eggs (proportion of ovulated eggs to be successfully fertilized; GCARKO, 38.6 % ± 17.5; WT, 94.5% ± 3.6) and embryos (proportion undergoing cleavage to the two-cell stage; GCARKO, 35.7% ± 17.5; WT, 89.5% ± 5.0) were significantly reduced in GCARKO female mice compared with WT (P < 0.05; Fig. 7C).

Fig. 7

Cumulus cell expansion, natural ovulation, fertilization, and early embryo development. In all graphs: white bars, WT; black bars, GCARKO. A) Average cumulus expansion index of WT and GCARKO OCC recovered from the oviducts of superovulated mice 16 h post-hCG treatment (t-test, P < 0.01). n = 79–113 per genotype. Black arrowhead, OCC with maximal cumulus expansion; white arrowhead, GCARKO OCC with reduced cumulus expansion. Original magnification ×6. B) Average number of naturally ovulated oocytes collected from WT and GCARKO mice. n = 6 per genotype. C) Percentage of oocytes to be fertilized after natural mating and progress to the two-cell stage of early embryo development after culture overnight (ANOVA, P < 0.05). n = 6 per genotype. Data are the mean ± SEM. Different lowercase letters denote statistically significant differences.

Discussion

A vital role for AR-mediated actions in regulating ovarian function and thus female fertility has been identified (reviewed in [1, 2]). However, the specific site of action(s) and AR-mediated mechanisms involved remain to be fully identified. The current study investigated the importance of classical genomic AR signaling pathways within the granulosa cells in female fertility. We developed a GCARKO female mouse model using AMH-Cre mice mated with our AR-floxed mouse model, in which exon 3 of the Ar gene is floxed [23], to determined the specific role of AR in granulosa cells, independent of possible contributions to the observed phenotype by the loss of AR in the uterus, oocyte, or theca cells. ROSA staining revealed that AMH-Cre was not present in oocytes or theca cells of GCARKO ovaries, with staining restricted to granulosa cells. Furthermore, PCR analysis confirmed that Ar with the excised exon 3 was only present in the ovary and granulosa cells, thus avoiding potential confounding effects from Ar inactivation in the uterus, pituitary, or brain. Consistent with AMH expression patterns [35, 36], Cre expression was highest in preantral and antral follicles, with expression decreasing in the final stages of follicle growth. Within follicles not all granulosa cells exhibited AR with the excised exon 3, hence our model, using AMH-Cre, may have underestimated the role of granulosa cell AR actions in female fertility and ovarian function. However, observed effects were not contributed to by the loss of AR function in other tissues, which cannot be ruled out by using Amhr2Cre [24, 25].

In the present study we show that granulosa cell classical AR-mediated actions play an important role in maintaining female fertility. GCARKO females are subfertile with fewer litters and pups produced over the 6-mo breeding trial, which is in agreement with a recent study [24]. Furthermore, GCARKO females exhibited an age-dependent reduction in cumulative pups per female, indicating that alterations in granulosa cell AR signaling have long-term detrimental effects on reproductive function.

We identified no difference in corpus lutea numbers or natural ovulation rates between our GCARKO and WT controls, which is in contrast to our [23, 28] and other [9, 22] global ARKO models and a recent GCARKO mouse model [24]. This finding implies that the defective ovulation in the other models is due to either loss of AR-mediated actions in the final stages of follicle development, as AR expression is normal at this stage in our GCARKO model, or that ovulation must be at least in part regulated by AR signaling in tissues other than the granulosa cells. Other studies have implicated central defects in androgen signaling as playing an important role in maintaining female fertility [27–34]. Previous findings in global ARKO females of delayed production of first litter [23], longer estrous cycles [9, 28], and reduced naturally ovulated oocyte numbers being overcome by gonadotropin hyperstimulation [23] suggested that abnormal AR signaling may lead to defects in hypothalamic-pituitary regulation of ovulation. Indeed, an extraovarian defect was confirmed by reciprocal ovary transplantation studies in which ARKO or control ovaries transplanted into ovariectomized control hosts displayed normal estrous cycles and fertility, whereas transplantation of control ovaries into ovariectomized ARKO hosts resulted in abnormal estrous cycles and reduced fertility [28].

Despite the support for a central role for AR actions in the control of female fertility, our GCARKO model demonstrates that within the ovary granulosa cells also provide an important site for AR actions. GCARKO females exhibit defective follicle development with a significant reduction in large preantral and small antral follicle populations in 3-mo-old GCARKO ovaries. This finding is supported by several studies implicating AR actions as playing a stimulatory role in normal follicle development [18, 52–54], in particular in the preantral and antral stages [18, 53]. The temporal reduction in the number of large preantral and small antral follicles suggests that a loss of AR signaling in granulosa cells may lead to ovaries undergoing a premature decline in follicle populations that then remain constant; however, the reason for this remains unknown. This reduction in follicle populations supports a previously proposed role for AR in the regulation of follicle numbers and follicle atresia [9, 22, 55–57]. However, our GCARKO females did not exhibit global accelerated follicle depletion, which is in agreement with our global ARKO model [23], but in contrast to another global ARKO model [22] and the previous GCARKO Amhr2cre model [24]. Our female global ARKO model and the present GCARKO model exhibit an in-frame excision of Ar exon 3 and thus retain a minimally truncated but nonfunctional AR protein [23]. However, female global ARKO models with major AR deletions, such as the exon 2 deletion used in the GCARKO Amhr2cre model, exhibit a more severe phenotype with the loss of AR protein, causing accelerated follicle depletion in some or all ARKO females [9, 22]. Thus, the more severe phenotype of premature ovarian failure may be due to secondary consequences of the loss of AR protein in these models, for example, the loss of ligand-independent protein-protein interactions, such as AR heterodimerization or co-regulator binding [1, 2].

Although global accelerated follicle depletion was not observed, a role for AR in regulating granulosa cell survival and hence protecting the follicle from undergoing follicular atresia is supported by our finding of a significant increase in unhealthy follicles and zona pellucida remnants within GCARKO ovaries at 6 mo of age. AR actions have previously been reported as playing an important survival role within the ovary, with abnormal levels of AR expression in estrogen receptor beta knockout mouse ovaries associated with high levels of follicular atresia, but this effect can be reversed by the addition of an AR antagonist [58]. The age-dependent increase in follicle atresia observed in our model may be contributed to by the observed trend toward increased LH levels in 6-mo-old GCARKO females, as it has been proposed that LH levels beyond a certain elevated threshold suppress granulosa cell proliferation and initiate follicle atresia [59]. Moreover, the elevated LH levels coupled with the lengthened estrous cycles observed in our and other [24] 6-mo-old GCARKO females imply that the loss of granulosa cell AR function within the ovary may alter hypothalamic-pituitary-gonadal feedback signaling.

Observed effects of androgens on oocyte maturation and embryonic development [60–62], along with the persistent high AR expression in the OCC and granulosa cells lying in close proximity to the OCC [63], are consistent with a paracrine role between oocytes and surrounding granulosa cells in the maintenance of oocyte health and embryo development. In the current study, GCARKO females exhibited significantly reduced cumulus expansion and oocyte/embryo viability, manifested as decreased rates of fertilization and thus progression to the two-cell stage. KIT ligand, which is involved in the oocyte-granulosa cell regulatory loop, has reduced expression in ARKO ovaries [22]. However, this reduction observed in global ARKO females was not apparent in our GCARKO females. Although altered timing of mating cannot be ruled out, oviduct environment appeared normal as there was no difference between in WT and GCARKO ampulla and isthmus morphology. Thus, these findings imply that granulosa cell AR actions play an important role in regulating cumulus expansion and the ability of the oocyte to be fertilized and undergo cleavage, which may be due to reduced oocyte quality.

In agreement with global ARKO females [9, 22–23], our GCARKO females are subfertile and exhibit altered follicle populations and reduced follicle health, confirming a vital role for AR signaling in granulosa cells in the regulation of female fertility and ovarian follicle development. However, in comparison to global ARKO models [9, 22–23], the current GCARKO model did not exhibit reduced ovulation rates, implying that the presence of AR during the final stage of follicle development plays a key role in the regulation of ovulation or that ovulation is regulated by AR signaling in nongranulosa cells. Global ARKO females exhibit a disassociation of cumulus cells from the oocyte during ovulation [9] and a reduction in gene expression of hyaluronan synthase 2 and tumor necrosis factor-α-stimulated gene 6, which are both required for cumulus expansion [9]; however, oocyte/embryo viability was normal [23]. Defective cumulus expansion and reduced oocyte/embryo viability were observed in our GCARKO model, confirming a specific role for granulosa cell AR actions in the maintenance of oocyte heath. In the present GCARKO females, the incomplete inactivation of AR in all granulosa cells, and/or the possibility that the granulosa cells that retain their AR may compensate for the lack of AR signaling in the granulosa cells expressing Cre, cannot be ruled out as the cause of the less severe phenotype observed when compared with global ARKO females. However, despite this, these findings confirm granulosa cells as an important site for AR signaling involved in the regulation of female fertility.

In summary, we demonstrate that a loss of classical genomic AR signaling in the granulosa cells of preantral and antral follicles leads to subfertility, defective follicle dynamics without altered ovulation rates or follicle depletion, increased follicle atresia, and reduced embryo viability. We reveal that preantral and antral follicle granulosa cells constitute a crucial site for AR-mediated actions involved in maintaining follicle and embryo survival and, ultimately, optimal female fertility.

Acknowledgment

The authors thank Jeff Zajac (University of Melbourne) for access to the AR floxed mice and Jenny Spaliviero, Li Ann Ooi, and Lucy Yang for technical support.

Author notes

¹

Supported by Australian National Health and Medical Research Council grant 632678 and the Australasian Menopause Society.