|Home | About | Journals | Submit | Contact Us | Français|
Menopause is an important public health issue because of its association with a number of disorders. Androgens produced by residual ovarian tissue after menopause could impact the development of these disorders. It has been unclear, however, whether the postmenopausal ovary retains steroidogenic capacity. Thus, an ovary-intact mouse model for menopause that uses the occupational chemical 4-vinylcyclohexene diepoxide (VCD) was used to characterize the expression of steroidogenic genes in residual ovarian tissue of follicle-depleted mice. Female B6C3F1 mice (age, 28 days) were dosed daily for 20 days with either vehicle or VCD (160 mg kg−1 day−1) to induce ovarian failure. Ovaries were collected on Day 181 and analyzed for mRNA and protein. Acyclic aged mice were used as controls for natural ovarian senescence. Relative to cycling controls, expression of mRNA encoding steroidogenic acute regulatory protein (Star); cholesterol side-chain cleavage (Cyp11a1); 3beta-hydroxysteroid dehydrogenase (Hsd3b); 17alpha-hydroxylase (Cyp17a1); scavenger receptor class B, type 1 (Scarb1); low-density lipoprotein receptor (Ldlr); and luteinizing hormone receptor (Lhcgr) was enriched in VCD-treated ovaries. In acyclic aged ovaries, mRNA expression for only Cyp17a1 and Lhcgr was greater than that in controls. Compared to cycling controls, ovaries from VCD-treated and aged mice had similar levels of HSD3B, CYP17A1, and LHCGR protein. The pattern of protein immunofluorescence staining for HSD3B in follicle-depleted (VCD-treated) ovaries was homogeneous, whereas that for CYP17A1 was only seen in residual interstitial cells. Circulating levels of FSH and LH were increased, and androstenedione levels were detectable following follicle depletion in VCD-treated mice. These findings support the idea that residual ovarian tissue in VCD-treated mice retains androgenic capacity.
In mammals, the neonatal ovary contains its full complement of oocyte-containing primordial follicles that will be available throughout life. Because oocytes cannot be generated after birth, this set number of primordial follicles represents a finite source of germ cells for ovulation. Throughout life, ovarian follicles are continually lost by ovulation or a process of programmed cell death, known as atresia . Once all follicles are lost, the ovary is follicle-depleted, and ovarian failure (menopause in women) occurs.
Menopause is associated with an enhanced incidence of a number of disorders, including Alzheimer disease , cardiovascular disease , metabolic syndrome , osteoporosis , and ovarian cancer . In recent years, menopause has become a more important public health issue, because on average, women are now living approximately 30 years after ovarian failure, which means three decades of increased risk for these related disorders . As a result, understanding the biology of menopause and its associated disorders is receiving greater attention. An ongoing controversy exists regarding whether residual ovarian tissue in postmenopausal women is steroidogenic [8–11]. This issue is important to resolve, because in the face of declining 17β-estradiol, androgens produced by residual ovarian tissue could impact postmenopausal health in a positive or negative manner.
Previously, animal studies aimed at understanding the biology of the postmenopausal ovary have not been feasible because of the lack of appropriate models in which ovarian function has been lost but residual ovarian tissue remains. Recently, however, an ovary-intact mouse model of menopause has been developed using the occupational chemical 4-vinylcyclohexene diepoxide (VCD) [12–14]. Repeated daily dosing with VCD selectively destroys primordial and primary follicles in ovaries of mice and rats by accelerating the natural process of follicular atresia [15–18]. Because VCD does not target larger follicles, the animal continues to ovulate normally until no more follicles can be recruited. Thus, ovarian follicular depletion in the VCD-treated mouse is gradual. As with women undergoing perimenopause, VCD-treated mice show increased levels of FSH  as well as declining levels of estrogen and irregular estrous cycles  as they become follicle-depleted. Additionally, following ovarian failure, residual ovarian tissue is retained. Thus, preservation of residual ovarian tissue in the VCD-treated, follicle-depleted mouse makes this model ideal for studying the physiology of the postmenopausal ovary. Several menopause-related disorders have been modeled and studied using the VCD-treated mouse model of peri- and postmenopause [19–21]. Therefore, the VCD-treated mouse is relevant for studies related to both perimenopausal and postmenopausal stages .
A previous study  found that dispersed cells collected from residual ovarian tissue in VCD-treated mice were capable of producing androstenedione in vitro; however, how that related to cells from age-matched cycling controls was not investigated. Therefore, the present study was designed to characterize steroidogenic capacity in residual ovarian tissue from VCD-treated mice following ovarian failure by comparing ovarian mRNA and protein expression of steroidogenic enzymes between follicle-depleted and age-matched cycling control animals. In addition, circulating levels of androstenedione were monitored throughout the present study.
Female B6C3F1 mice (age, 21 days) were purchased from The Jackson Laboratory, housed in polycarbonate plastic cages, kept at 22 ± 2°C on 12L:12D cycles, and fed ad libitum. All animals were allowed to acclimate to the animal facilities for 1 wk before starting treatment. Two-year-old acyclic B6C3F1 mice (acyclic aged) were either provided by Dr. Loretta P. Mayer (Department of Biological Sciences, Northern Arizona University, Flagstaff, AZ) or aged in house. All experiments and methods were approved by the University of Arizona Institutional Animal Care and Use Committee and conformed to the Guide for the Care and Use of Experimental Animals.
On Postnatal Day 28, animals were randomly assigned to two groups: VCD (160 mg kg−1 day−1 i.p, n = 8; Sigma-Aldrich) and sesame oil (vehicle control, i.p., n = 8; Sigma-Aldrich). Animals were weighed and dosed daily for 20 days. Estrous cyclicity was monitored by daily vaginal cytology starting on Day 29 after the onset of dosing in the VCD-treated and cycling control mice and before death in 2-yr-old mice (to confirm acyclicity). Ovarian failure was assigned when mice showed 10 days or more of persistent diestrus . On average, ovarian failure occurred on Day 53.1 ± 1.47 after the onset of dosing (range, Days 46–68) in VCD-treated mice.
Animals were killed by CO2 inhalation on Day 181 after the onset of dosing (cycling controls, VCD-treated) or at 2 yr of age. It was not clear which would be the best stage of the estrous cycle in which to make a comparison with an acyclic or follicle-depleted ovary. Therefore, tissues from control animals were collected on the same day (age matched) as those from VCD-treated animals, and the cycle stage at the time of tissue collection was not assessed. Ovaries were excised, trimmed of fat, and weighed. Ovaries were either stored at −80°C in RNAlater (for RNA isolation, n = 4; Ambion), quick-frozen in liquid nitrogen (for protein isolation, n = 4), or fixed in 4% formalin (for immunostaining, n = 4) or Bouin fixative (for histological evaluation, n = 4). The contralateral ovary of each control mouse used in the mRNA analysis was prepared for morphological evaluation. Primordial, preantral, and antral follicles as well as corpora lutea were observed in ovaries of all control animals. Thus, all control ovaries showed evidence of cyclicity.
Total RNA was isolated from whole ovaries (n = 4 per group) using an RNeasy Mini kit (Qiagen) according to the manufacturer's protocol for tissue samples. Additionally, samples were incubated (15 min) with DNase (Qiagen) during the isolation to eliminate DNA contamination. Samples were concentrated using an RNeasy MinElute kit (Qiagen). The RNA concentration of each sample was determined at 260 nm using a Nanodrop ND1000 UV-Vis spectrophotometer (Nanodrop Technologies). Samples were amplified and converted to antisense RNA using an Amino Allyl MessageAmp kit (Ambion).
Amplified RNA samples (2 μg) were reverse transcribed with Moloney murine leukemia virus reverse-transcriptase enzyme (Invitrogen) using oligo(dT)18–20 primers (Invitrogen). Each cDNA sample was diluted with nuclease-free water to a final concentration of approximately 8 ng/μl. Real-time PCR experiments were carried out using a Rotor-Gene RG3000 real-time DNA detection system (Corbett Life Science). The PCR reactions were done in triplicate, and each contained 2 μl of diluted cDNA, 0.4 μl of 25 mM MgCl2, 0.35 μl of nuclease-free water, 0.25 μl of SYBR Green Dye, 5 μl of Quantitect SYBR Green PCR Master Mix (Qiagen), and 100 pmol (2 μl) of forward and reverse primers (Integrated DNA Technologies, Inc.) for a final volume of 10 μl. The program started with a hold step of 95°C for 15 min, followed by 45 cycles of 95°C for 15 sec, 58°C for 15 sec, and 72°C for 20 sec. The final step was a melt cycle, which consisted of 1°C increments from 72 to 99°C, spending 45 sec on the first step and 5 sec on each step thereafter. Primers were designed to the 3′ end of each gene using Primer3 (version 0.3.0 ), and their sequences are listed in Table 1. Primer specificity was assessed by matching primer sequences with that of the gene of interest using BLASTN 2.2.18+ [25, 26]. Each primer pair was tested by agarose gel electrophoresis to confirm that PCR amplification products were the correct size. The cycle threshold (Ct) numbers were obtained by setting threshold within the exponential phase of the reaction, and the expression of each gene of interest was determined by normalizing experimental Ct values to β-actin as described previously .
Total protein was isolated from whole ovaries of Day 181 cycling controls; VCD-treated, follicle-depleted; and 2-yr-old acyclic aged mice (n = 2–3 per group). Ovaries were homogenized in cold lysis buffer (1% Triton-X-100, 50 mM Hepes, 150 mM NaCl, 10% glycerol, 50 mM NaF, 2 mM EDTA, and 0.1% SDS) containing protease inhibitor cocktail (100 μl/ml lysis buffer; Sigma) using a variable-speed Tissue Tearor homogenizer (Biospec Products, Inc.) at medium speed for intervals of 5 sec. Protein samples (10–20 μg) were separated by SDS-PAGE using 9–10% bis-acrylamide gels (5 min at 70 V, then 1 h at 115 V for 10% gels or at 135 V for 9% gels; Bio-Rad Laboratories) and transferred to nitrocellulose membranes (Protran BA85; 1 h at 100 V). Blotted samples were blocked for 1 h in 5% milk (5 g of dry milk in 100 ml of TTBS [0.5 M NaCl, 20 mM Tris, and 0.15% Tween-20]) at 4°C. Blots were then incubated with anti-HSD3B (1:1000 dilution; Santa Cruz Biotechnologies, Inc.), anti-CYP17A1 (1:200 dilution; Santa Cruz Biotechnologies), anti-LHCGR (30 μg/ml; a gift from J. Wimalansena, University of Tennessee, Knoxville, TN ), or anti-PPIB (1 μg/ml; Cyclophilin B (Peptidylpropyl isomerase B); Abcam, Inc.) in 5% milk for 1 h at room temperature. Blots were washed three times in TTBS (10 min/wash) and horse radish peroxidase-conjugated secondary antibody in 5% milk (donkey anti-goat IgG, 1:10000 dilution [Santa Cruz Biotechnologies]; goat anti-rabbit IgG, 1:5000 dilution [Pierce]; goat anti-mouse IgG, 1:1000 dilution [Pierce]) was added for 1 h at room temperature. Blots were washed three times in TTBS (10 min/each), followed by an additional wash with TBS (10 min/wash), and then detected using ECL Plus Western detection reagents (GE Healthcare) according to manufacturer's instructions. Densitometry of the appropriate bands was performed using the Gel Analysis feature in ImageJ software (version 1.40g; http://rsb.info.nih.gov/ij) .
Ovaries were fixed in 4% formalin for 4 h, transferred to 70% ethanol, and prepared for immunostaining as described by Cannady et al.  with modifications. Briefly, fixed ovaries were embedded in paraffin, and every 10th and 11th section through the ovary was prepared and deparaffinized (n = 8 sections/ovary and 3 ovaries/treatment). Sections were either baked with antigen retrieval (anti-HSD3B) or air-dried without antigen retrieval (anti-CYP17A1). Tissue was blocked for 5 min with 5% BSA (Gamma Biological) and then incubated with anti-CYP17A1 (1:50 dilution) or anti-HSD3B (1:100 dilution) overnight at 4°C. Specificity of the antibodies was confirmed by observing the expected bands on Western blots and performing staining with antibodies that were preabsorbed with their respective blocking peptides (fourfold strength, overnight at 4°C; Santa Cruz Biotechnologies, Inc.). Slides were incubated with secondary biotinylated antibody (rabbit anti-goat, 1:75 dilution; Vector Laboratories) for 1 h. This was followed by incubation with Cy5-streptavidin (1 h, 1:100 dilution; Jackson Laboratories) and staining with YOYO-1 (genomic DNA, 10 min, 1:50 dilution; Molecular Probes). Immunofluorescence was visualized using a Zeiss LSM 510 NLO-Meta confocal microscope with an argon and helium-neon laser projected through the tissue into a photomultiplier at I = 488 and 633 nm for YOYO-1 (green) and CY-5 (red), respectively.
Ovaries were trimmed and placed in 4% formalin (4 h), transferred to 70% ethanol, and paraffin-embedded. Each ovary was sectioned (thickness, 4–5 μm), mounted, and stained with hematoxylin and eosin. Overlapping images of the slides (magnification ×10) were taken using an Olympus IX-70 inverted microscope connected to a Nikon camera and then stitched together using Adobe Photoshop CS software (version 8.0; Adobe Systems, Inc.).
Blood was collected by retro-orbital puncture under anesthesia (1.25% Avertin) on Days 35, 63, 91, 119, 152, and 181 after the onset of dosing or at time of death for acyclic aged mice. To obtain serum, samples were kept on ice for 1 h and then centrifuged at 14000 rpm for 15 min at 4°C. Samples were stored at −20°C for subsequent analysis. Serum FSH and LH levels were measured by RIA according to instructions with kits from the National Hormone and Pituitary Distribution Program. Circulating levels of androstenedione were determined using a Direct Androstenedione RIA kit (Coat-A-Count; Diagnostic Products Corp.). The sensitivity and interassay coefficients of variation were as follows: FSH, 1.0 ng/ml and 0.922%, respectively; LH, 0.5 ng/ml and 3.7%, respectively; and androstenedione, 70 pg/ml and 5.7%, respectively. The results of all RIA assays were calculated by four-parameter logistic analysis using AssayZap (BioSoft).
The FSH, LH, and real-time PCR data were compared using t-tests. Levels of circulating androstenedione throughout the present study were compared between cycling controls and VCD-treated animals using Mann-Whitney nonparametric tests. Levels of androstenedione within each treatment group were compared over time using ANOVA followed by a Fisher protected-least-significant-difference post hoc test. Western blot data were analyzed using a Kruskal-Wallis nonparametric test. Tests were conducted using StatView for Windows (version 5.0; SAS Institute, Inc.), with the significance level set at P < 0.05.
The expression of several genes encoding steroidogenic enzymes was evaluated in residual ovarian tissue from VCD-treated, follicle-depleted mice and acyclic aged mice (age, 2 yr) and compared to that in ovaries from cycling controls (Fig. 1). Relative to age-matched cycling controls, increased (P < 0.05) abundance of mRNAs encoding Star (VCD/control, 2.15 ± 0.30), Cyp11a1 (VCD/control, 2.48 ± 0.25), Hsd3b1 (VCD/control, 2.15 ± 0.18), and Cyp17a1 (VCD/control, 147.71 ± 0.14) was observed in ovaries from VCD-treated animals. Because mRNA encoding Cyp17a1 (v1) was substantially greater in follicle-depleted ovaries compared with cycling controls, two additional primer pairs directed at distinct regions of the gene were used (v2 and v3) (Table 1). Cyp17a1 mRNA was still more highly enriched using the other primer sets (VCD/control, 116.4 ± 0.14 and 237.8 ± 0.67; P < 0.05). To compare the VCD-treated, follicle-depleted ovary to ovaries that had undergone natural senescence, expression of mRNA encoding those enzymes also was determined using ovaries from acyclic aged mice. Relative to cycling controls, no difference (P > 0.05) was found in mRNA levels of Star (aged/control, 0.74 ± 0.37), Cyp11a1 (aged/control, 1.17 ± 0.21), and Hsd3b1 (aged/control, 0.98 ± 0.29) in ovaries of acyclic aged mice. Similar to follicle-depleted (VCD-treated) ovaries, ovaries from acyclic aged mice were enriched (P < 0.05) in mRNA for Cyp17a1 (aged/control, 17.59 ± 0.72). Follicle-depleted ovaries from VCD-treated mice had higher (P < 0.05) expression of all genes tested relative to those of acyclic aged mice.
Expression of HSD3B and CYP17A1 protein was measured by Western blot analysis in ovaries of VCD-treated, follicle-depleted mice; cycling controls; and acyclic aged mice (Fig. 2). Densitometric analysis of the bands from each sample was normalized to peptidylprolyl isomerase B, also known as cyclophilin B, and showed similar (P > 0.05) expression of both HSD3B and CYP17A1 protein in VCD-treated and acyclic aged mice as compared to cycling controls. Interestingly, two of the three follicle-depleted, VCD-treated samples showed a second band slightly shifted above the 57-kDa molecular weight expected for CYP17A1 protein. This second band was not seen in cycling controls or acyclic aged ovaries.
Distribution of HSD3B and CYP17A1 protein was morphologically evaluated in ovaries of VCD-treated, follicle-depleted animals and cycling control animals by immunofluorescence and confocal microscopy (Fig. 3). HSD3B protein was present in theca cells, interstitium, and corpora lutea (Fig. 3A) as well as in preovulatory follicles (data not shown). CYP17A1 protein was present in theca and interstitial cells in cycling control ovaries (Fig. 3D) but not in corpora lutea (data not shown). In follicle-depleted ovaries, staining for HSD3B (Fig. 3B) was relatively homogeneous throughout residual ovarian tissue, whereas staining for CYP17A1 protein was more localized to scattered individual cells (Fig. 3E). No staining occurred when either primary antibody was preabsorbed with antigen (Fig. 3, C and F). Follicle depletion and ovarian atrophy are apparent in ovaries from VCD-treated mice as well as acyclic aged mice when compared with cycling controls (Fig. 3, H, I, and G, respectively).
Expression of mRNA encoding receptors involved in ovarian steroidogenesis also was assessed (Fig. 4). Compared to cycling controls, the follicle-depleted (VCD-treated) ovaries contained greater levels (P < 0.05) of Scarb1 (Sr-b1; VCD/control, 3.35 ± 0.25) and Ldlr (VCD/control, 5.31 ± 0.11). Relative to cycling controls, no difference (P > 0.05) was found in the expression of Scarb1 (aged/control, 1.07 ± 0.31) and Ldlr (aged/control, 1.43 ± 0.30) in ovaries of acyclic aged mice. Relative to acyclic aged mice, follicle-depleted ovaries from VCD-treated mice had higher (P < 0.05) expression of Scarb1 and a trend for increased levels of Ldlr mRNA (P = 0.06). Expression of mRNAs encoding the oocyte-specific genes Zar1 (zygote arrest 1) and Zp3 (zona pellucida glycoprotein 3), used as negative controls, was undetectable in all follicle-depleted ovaries (data not shown).
Compared to cycling controls, follicle-depleted, VCD-treated and acyclic aged ovaries contained greater (P < 0.05) levels of Lhcgr (VCD/control, 21.31 ± 0.66; aged/control, 6.41 ± 0.26). Relative to VCD-treated, follicle-depleted ovaries, the ovaries from acyclic aged mice showed a nonsignificant trend for lower (P = 0.06) Lhcgr mRNA levels. As determined by Western blotting, similar to HSD3B and CYP17A1, there was no difference (P > 0.05) in ovarian LHCGR protein between cycling controls, VCD-treated, and acyclic aged mice.
Circulating levels of FSH were measured (Fig. 5A). Relative to cycling controls, VCD-treated mice had increased (37.9 ± 8.8 vs. 11.8 ± 2.6 ng/ml, P < 0.05) levels of FSH. However, FSH was not different (P > 0.05) between aged (19.4 ± 4.3 ng/ml) and control (11.8 ± 2.6 ng/ml) or VCD-treated (37.9 ± 8.8 ng/ml) mice. Even so, a nonsignificant (P = 0.07) trend was found for lower FSH levels in aged relative to VCD-treated mice.
Circulating levels of LH also were determined (Fig. 5B). No difference (P > 0.05) was found between the three groups. A trend was found, however, for increased levels of LH in VCD-treated mice (8.5 ± 2.0 ng/ml) as compared with cycling controls (4.6 ± 0.09 ng/ml).
Levels of androstenedione were determined in the circulation of cycling controls and VCD-treated mice throughout the present study (Fig. 5C). Compared to age-matched cycling controls, circulating levels of androstenedione were lower (P < 0.05) in VCD-treated, follicle-depleted mice on Days 63 and 119 after the onset of dosing. At the other time points throughout the present study, however, no differences were found between groups. Compared to baseline (Day 35), circulating levels of androstenedione in cycling controls did not vary between Days 35 and 152 but were reduced (P < 0.05) on Day 181 (age, 7 mo). Conversely, compared to baseline (Day 35), circulating androstenedione was reduced (P < 0.05) between Days 119 and 181 in follicle-depleted animals.
During the biosynthesis of ovarian steroids, STAR mobilizes cholesterol into mitochondria, where it is converted to pregnenolone by CYP11A1. Pregnenolone is further metabolized by either CYP17A1 or HSD3B to produce dehydroepiandrosterone (DHEA) or progesterone, respectively. In turn, DHEA can be further converted by HSD3B to androstenedione or by HSD17B to androstenediol. Androstenedione can be used as a substrate for synthesis of estrone by CYP19 (aromatase) or, like androstenediol, for synthesis of testosterone by HSD17B .
From the hematoxylin and eosin-stained sections of whole ovaries on Day 181, evidence of ovarian cyclicity was still obvious in age-matched cycling controls, but follicle depletion and ovarian atrophy had occurred in VCD-treated animals. This also was observed in ovaries from acyclic aged mice. Whereas the whole ovary from cycling control animals contains a heterogeneous mixture of functional compartments, the follicle-depleted (postmenopausal) ovary lacks ovarian follicles and corpora lutea. Instead, it is composed mainly of ovarian stroma and secondary interstitial cells. Secondary interstitial cells originate from the theca interna of atretic follicles that do not undergo apoptosis at the time of oocyte and granulosa cell death [32, 33]; however, the onset of apoptosis in secondary interstitial cells occurs later and at a slower rate than in granulosa cells [34–37]. Thus, relative to cycling control ovaries, the follicle-depleted ovary appears to be highly enriched in certain genes, but this reflects a relative absence of nonsteroidogenic compartments. In the present study, mRNA encoding Star, Cyp11a1, Hsd3b, and Cyp17a1 was highly enriched in VCD-treated, follicle-depleted ovaries compared to ovaries from age-matched cycling controls. These findings are in agreement with those of previous studies using human ovaries, in which expression of STAR, CYP11A1, HSD3B, and CYP17A1 mRNAs was detected in postmenopausal ovarian stroma [10, 38]. Additionally, mRNA encoding STAR, CYP11A1, and HSD3B has been detected when measured in cultured human ovarian stromal cells . To compare steroidogenic capacity in the follicle-depleted ovary from VCD-treated mice to that in ovaries from animals that had undergone natural senescence, the expression of mRNA encoding Star, Cyp11a1, Hsd3b1, and Cyp17a1 was evaluated. No significant difference in Star, Cyp11a1, and Hsd3b1 mRNA levels was found between acyclic aged mice and younger cycling controls. Expression of Cyp17a1 was enriched in ovaries from VCD-treated, follicle-depleted mice as well as in those from acyclic mice relative to cycling controls. Furthermore, Cyp17a1 mRNA was lower (P < 0.05) in acyclic aged ovaries relative to ovaries from VCD-treated mice.
It has been shown that LH and FSH, via a cAMP-signaling cascade, stimulate the expression of steroidogenic enzymes in ovarian cells [39, 40]. Because both VCD-treated and acyclic follicle-depleted ovaries are enriched in residual ovarian tissue, the differences in mRNA expression might be the result of increased levels of LH and FSH that accompany ovarian failure . To evaluate this possibility, circulating levels of LH and FSH were determined. Compared to cycling controls, VCD-treated, follicle-depleted mice had significantly higher levels of FSH and a nonsignificant trend for higher LH in circulation. This finding (Day 181) is in contrast to animals evaluated at an earlier time point (Day 120) following the onset of VCD dosing, in which circulating LH was significantly greater than in age-matched controls . This might relate to declining gonadotropin production as time after ovarian failure progresses. Interestingly, levels of both gonadotropins in acyclic aged mice were not different from those in young cycling controls. Past studies have shown that aged postmenopausal women (40 yr after the onset of menopause) have a reduced response of LH and FSH compared to that of younger postmenopausal women (~10 yr after the onset of menopause) following acute injection of GnRH . Also, aged postmenopausal women (≥15 yr after the onset of menopause) have significantly lower levels of both gonadotropins compared to women within 2–5 yr after the onset of menopause . This effect of aging could explain the differences observed between VCD-treated, follicle-depleted mice and acyclic aged mice. Therefore, the ovary from a VCD-treated mouse (following ovarian failure) may be a more relevant model for the early postmenopausal woman than naturally senescent animals.
Determining expression of Lhcgr in ovaries from VCD-treated, follicle-depleted mice and acyclic aged mice also was of interest. Relative to cycling controls, ovaries from VCD-treated, follicle-depleted mice and acyclic aged mice were enriched in mRNA encoding Lhcgr. As with genes encoding steroidogenic enzymes, Lhcgr mRNA also was lower in ovaries from acyclic aged mice than in ovaries from VCD-treated mice. A previous study  reported LHCGR protein in follicle-depleted ovaries of VCD-treated mice as well as an increase in androstenedione production in response to LH in cultures of dispersed cells collected from follicle-depleted ovaries at single time points (Days 127–120, respectively) earlier than the time point evaluated in the present study (Day 181). In the present study, we observed expression of LHCGR protein in both VCD-treated, follicle-depleted ovaries and acyclic aged ovaries. The observations in VCD-treated and acyclic aged mice agree with those of studies demonstrating expression of LHCGR mRNA in human postmenopausal residual ovarian tissue . Thus, as with human ovaries, the follicle-depleted mouse ovary has the potential to respond to LH. Because the expression of mRNA for both Cyp17a1 and Hsd3b is regulated by the binding of LH to its receptor, elevated expression of mRNAs encoding steroidogenic enzymes in the VCD-treated, follicle-depleted ovary may result from the action of LH on LH receptor in residual steroidogenic cells.
Both HSD3B and CYP17A1 protein were detected in ovaries from VCD-treated and acyclic aged mice. Constitutive expression of HSD3B has been reported in the interstitial compartment and theca interna of rat ovaries irrespective of the stage of estrous cycle or aging . Also, HSD3B expression has been demonstrated in theca interna of human ovarian follicles . In the present study, the distribution of HSD3B protein was found to be relatively homogeneous in residual ovarian tissue of follicle-depleted ovaries, whereas in cycling controls, staining was specifically observed in theca and interstitial cells, preovulatory follicles, and corpora lutea. When compared to cycling controls by Western blot analysis, HSD3B protein levels in follicle-depleted ovaries (VCD-treated and acyclic aged) were similar. Conversely, staining for CYP17A1 protein in the follicle-depleted ovary was more scattered and only present in residual interstitial cells. In cycling controls, staining for CYP17A1 also was localized to theca and interstitial cells, but in contrast to HSD3B, this was not observed in corpora lutea. Additionally, like HSD3B protein, CYP17A1 protein levels as determined by Western blot analysis were similar in VCD-treated and acyclic aged ovaries when compared to those in ovaries from cycling control animals. Interestingly, samples from two of three VCD-treated ovaries showed an additional higher-molecular-weight band suggestive of a phosphorylated form. Previous studies have provided evidence for posttranslational regulation of CYP17A1 by phosphorylation [46–48]. Specifically, it has been shown that 17,20-lyase activity of CYP17A1, which is responsible for the production of DHEA and androstenedione, is favorably regulated by serine phosphorylation in cultured human adrenocortical carcinoma cells . Therefore, it could be hypothesized that these larger forms represent phosphorylated CYP17A1 with greater 17,20-lyase activity and, thus, greater androgenic capacity. Previous studies have reported expression of CYP17A1 as specifically localized in steroidogenically active theca cells in rat, human, porcine, ovine, and bovine cycling ovaries [49–51]. Immunofluorescent staining for CYP17A1 protein appeared to be in theca cell-containing follicles in cycling controls and in scattered cells in the VCD-treated, follicle-depleted ovary. Follicle-depleted ovaries are devoid of follicle-associated theca cells; rather, they contain dispersed secondary interstitial cells derived from atretic follicles, which are the likely site of that staining.
Substantially greater levels of mRNA encoding Cyp17a1 were measured in follicle-depleted (VCD-treated and acyclic aged) ovaries relative to controls. To rule out a cross-reactivity of Cyp17a1 primers with another gene, two other sets of primers directed at distinct regions of the gene were designed and used for real-time PCR analysis. Levels of Cyp17a1 mRNA also were as greatly enhanced in follicle-depleted ovaries using the other primers. An apparent disconnect exists between Hsd3b, Cyp17a, and Lhcgr mRNA expression and their respective protein levels. These findings support the idea that in residual ovarian tissue, mRNA for these genes is highly enriched, whereas protein is not. Regardless of differences in levels of expression, these findings provide evidence that the follicle-depleted mouse ovary is composed mainly of stroma and secondary interstitial cells that express enzymes required for androgen synthesis. Thus, the observations reported here demonstrate that like human ovaries, the VCD-treated, follicle-depleted ovary retains the capacity to synthesize androgens.
The initial step in ovarian steroidogenesis is the internalization of cholesterol from circulation. It has been shown previously that both low- and high-density lipoprotein receptors facilitate ovarian androgen biosynthesis [52, 53]. In addition, ovarian granulosa cells as well as luteal cells derive cholesterol from high-density lipoprotein via the receptor SCARB1 [54–57]. In the present study, abundance of mRNA encoding Ldlr and Scarb1 was enhanced in residual ovarian tissue from VCD-treated, follicle-depleted mice compared to cycling controls. Conversely, relative to cycling controls, acyclic aged ovaries showed no difference in the expression of Scarb1 and a nonsignificant trend (P = 0.06) for higher levels of Ldlr. Increased expression of Scarb1 mRNA is consistent with previous findings, in which immunostaining of the follicle-depleted ovary was positive for SCARB1 protein . Taken together, these findings support the notion that steroidogenic cells in the follicle-depleted ovary have the capacity to obtain cholesterol from lipoprotein particles to serve as substrates for steroidogenesis.
Circulating levels of androstenedione were lower than those in controls on Days 63 and 119 in VCD-treated mice but were not different (P > 0.05) at the other time points. The reason for the differences seen on Days 63 and 119 are not clear, but they might relate to differences in stage of the estrous cycle for individual control animals on those days. At any rate, throughout the time course of the present study, androgen production was similar between the two groups. Recently, it has been reported that at a single time point, circulating androstenedione levels were not different between cycling controls and VCD-treated mice 2 mo after ovarian failure (age, 3 mo), whereas androstenedione was undetectable in age-matched mice that had been ovariectomized at the time of VCD-induced ovarian failure . To our knowledge, no previous study has compared circulating androstenedione levels over the time course of ovarian failure in VCD-treated mice with age-matched controls. Relative to baseline measures (Day 35; age, 2 mo) in the present study, a decline (P < 0.05) was observed in circulating androstenedione in cycling control animals on Day 181 (age, 7 mo). Although not completely understood, the decline in androstenedione in cycling control animals, as seen in aging cycling women  and aged postmenopausal women [59, 60], could be the result of an age-related decline in ovarian function, even though in the present study, the ovaries still contained follicles and corpora lutea as evidence of continued ovulations. In VCD-treated mice, the decrease in androstenedione began earlier (Day 119; age, 5 mo). However, the decline did not precede ovarian failure (Day 53; age, 2.5 mo), and levels remained detectable throughout the present study. It has been reported that the rodent adrenal gland does not produce androgens . Therefore, collectively, these findings support the idea that residual ovarian tissue in the VCD-treated mouse is the likely source of androgen production and continues to be substantial following ovarian failure.
A previous study demonstrated that dispersed cells from residual ovarian tissue in VCD-treated mice (2 mo following ovarian failure; age, 3 mo) were capable of synthesizing androstenedione in vitro and that this could be stimulated by LH . Additionally, at that time, circulating androstenedione levels in those animals were approximately 30% lower relative to those in cycling controls. Those findings suggested that residual ovarian tissue in the VCD-treated mouse is androgenic. However, androstenedione production in vitro was not compared with that in dispersed ovarian cells from cycling controls and circulating androstenedione levels were not measured over the time course of the study. Therefore, it is unknown whether androstenedione production fluctuates and differs from that in controls throughout a longer period of time, both with impending ovarian failure and following ovarian failure. Thus, the present study provides a more thorough characterization and more convincing support for androgenic capacity in residual ovarian tissue following ovarian failure.
In summary, the present study has provided evidence that the follicle-depleted ovary of VCD-treated mice expresses the enzymatic and receptor machinery necessary for synthesizing androgens de novo. These results favor the idea that the follicle-depleted ovary has an active role in the production of androgens following ovarian failure. This is supported by the fact that circulating androstenedione levels were not consistently different from those of age-matched, follicle-intact, cycling controls. Although steroidogenic capacity in the follicle-depleted, VCD-treated ovary does not completely correspond to that in naturally senescent acyclic mice, it more closely approximates previously reported characteristics of the ovaries of women in their early postmenopausal years than it does those of the acyclic aged mouse. In addition, the present results agree with those of several previous studies performed in human tissues, thus supporting use of the VCD-treated, follicle-depleted mouse as a highly relevant animal model for studying ovarian physiology of menopause. Furthermore, these studies also support the VCD-treated mouse model as being potentially useful for studies of the ovarian stroma separate from the major steroidogenic compartments in the ovary. Because menopause has become an important public health issue, studies aimed at understanding the physiology of the postmenopausal ovary in the context of its potential contributions to the prevention or progression of related disorders would benefit greatly from using the VCD-treated, follicle-depleted mouse.
The authors wish to thank Dr. Aileen Keating, Nivedita Sen, and Dr. Janet Funk for technical training. The authors also thank Andrea Grantham (Histology Service Laboratory, Department of Cell Biology and Anatomy, University of Arizona, Tucson, AZ) for histological processing of samples and Doug Cromey for help with the acquisition of immunofluorescent images and figure preparation (SWEHSC grant ES06694). Ovaries from aged mice in the mRNA and gonadotropin measurements were kindly donated by Dr. Loretta P. Mayer (Northern Arizona University, Flagstaff, AZ).
1Supported by NIH R01-AG21948, R01-ES06694, Center Grant ES09246 to P.B.H. and by American Physiological Society (APS) Porter Physiology Fellowship to Z.R.