Abstract
The mechanisms controlling the initiation and early stages of follicular growth are poorly understood. Our laboratory developed a serum-free culture system that supports spontaneous and wholesale activation of primordial follicles in pieces of cortex dissected from the ovaries of fetal calves and fetal baboons. However, very few follicles activated in vitro progressed to the secondary stage. To determine whether androgens can promote the primary to secondary follicle transition, pieces of fetal bovine ovarian cortex were cultured in serum-free medium in the absence or presence of testosterone (T, 10−7 and 10−6 M) or estradiol (E2, 10−6 M) for 10 days. Cortical pieces were then fixed and embedded in plastic for serial sectioning and morphometric analysis; fresh cortical pieces fixed on Day 0 served as uncultured controls. Freshly isolated cortical pieces contained mostly primordial follicles, whereas after 10 days in vitro, most primordial follicles had activated, differentiating into primary follicles as expected. Neither T nor E2 affected the number of primordial and primary follicles compared with controls (P > 0.05). However, T (10−7 and 10−6 M) increased the number of secondary follicles (P < 0.05), whereas E2 had no effect, suggesting that the effect of T was not due to conversion of T to E2. In the second experiment, the optimal concentration of T for preantral follicle growth was determined. A range of lower doses of T (10−10–10−7 M) increased the number of secondary follicles in cultured cortical pieces in a dose-dependent manner, with 10−7 M T being the most effective (P < 0.05). In the third experiment, addition of a specific androgen receptor blocker, flutamide, inhibited the stimulatory effects of T on the primary to secondary follicle transition (P < 0.05), suggesting a receptor-mediated action of T. Localization of androgen receptors by immunohistochemistry revealed immunostaining for the androgen receptor in ovarian stromal cells and increasing immunoreactivity in follicle cells as follicular development progressed from primordial and primary to secondary to antral follicles, suggesting the involvement of the androgen receptor in bovine folliculogenesis. In summary, our results show that T promotes the growth of bovine follicles activated in vitro and suggest that its stimulatory effect is mediated through androgen receptors in the stroma and/or follicular cells.
Introduction
Although the regulation of the later stages of ovarian follicular development has been studied extensively, the early stages of folliculogenesis, particularly the initiation of follicular growth (follicle activation) and the primary to secondary transition, remain poorly understood, especially in non-rodent species. Follicle formation occurs during fetal life in cattle and primates, including humans [1, 2]. Primordial follicles, which represent the earliest stage of follicular development, are composed of an oocyte that is surrounded by a single layer of flattened granulosa cells. During follicle activation, a primordial follicle enters the growing pool and follicular development progresses through the morphologically distinct stages of primary, secondary, antral, and preovulatory (Graafian) follicles. Understanding the mechanisms that control preantral follicular development is of practical as well as scientific interest, since the elucidation of this mechanism is a prerequisite for the use of primordial follicles to increase reproductive efficiency in domestic animals and endangered species and to ameliorate infertility in women exposed to gonadotoxic treatments. To date, only mouse follicles have been developed entirely in vitro to the point of meiotic and developmental competence of the oocyte [3, 4].
To study the regulation of preantral follicular development, we have developed a serum-free culture system that supports the activation of primordial follicles in pieces of ovarian cortex dissected from the ovaries of fetal calves or baboons obtained the last trimester of pregnancy [5, 6]. In vitro, most primordial follicles in pieces of the ovarian cortices of fetal or adult bovine ovaries or fetal baboon ovaries spontaneously activate within 2 days of culture and differentiate into primary follicles [5–7], and we have shown that insulin, but not IGF-I, is necessary for follicle activation and health [8]. However, after the onset of growth, only a few primary follicles progress to the secondary stage, even after 10–20 days in culture [6, 9, 10]. The reasons for this are not clear, although it is possible that follicles require more time in vitro to develop to the secondary stage. More likely, the culture conditions used in the current system are not optimal to support the primary to secondary follicle transition. One or more specific hormones and/or growth factors may be needed.
Atretic follicles are characterized by a low estradiol/androgen ratio [11–13], and androgens are generally considered to be inhibitory to antral follicular development (reviewed in [14]). For example, androgens inhibit the induction of LH receptors by FSH in cultured rat granulosa cells [15]. Treatment of hypophysectomized, estrogen-treated rats with androgen increases the apoptosis of granulosa cells in a subpopulation of early antral and late preantral follicles [16], and follicular atresia is increased by androgen administration in vivo [17]. However, a recent study has shown that androgen, but not estradiol, suppresses apoptosis in human ovarian tissue in vitro [18], and other recent studies have provided evidence that androgens stimulate the growth of mouse late preantral follicles in vitro via an androgen receptor-mediated mechanism [19, 20]. Interestingly, Vendola et al. [21, 22] observed a progressive increase in the number of growing preantral and small antral follicles when they treated rhesus monkeys with testosterone or the non-aromatizable androgen DHT for 3 to 10 days. Since the results of these studies with mice and primates suggest that androgens promote the growth of preantral follicles, we used our serum-free culture system with the bovine ovarian cortex to test the hypothesis that testosterone stimulates early bovine follicular development, especially the primary to secondary follicle transition.
Materials and Methods
Animals and Preparation of Ovarian Cortical Pieces
Ovaries were collected from late term bovine fetuses (5–8 mo gestation, estimated by crown-rump length [23]) at a local slaughterhouse (Wyalusing, PA). Fetal ovaries were quartered and placed in Leibovitz L-15 medium that was supplemented with 1% fetal bovine serum, 50 IU/ml penicillin and 50 μg/ml streptomycin (Invitrogen, Carlsbad, CA) for transport to the laboratory at ambient temperature (20–22°C), as previously described [5]. At the laboratory, ovarian cortical pieces, rich in primordial follicles, were dissected from the medullary tissue and cut into ~0.5 to 1 mm3 pieces.
Culture of Ovarian Cortical Pieces
Four ovarian cortical pieces from each fetus were fixed immediately (day 0) as uncultured controls. The other cortical pieces were placed on uncoated culture well inserts (two pieces/well; two wells/treatment; Millicell-CM, 0.4 μm pore size; Millipore Corporation, Bedford, MA) in the wells of 24-well Costar culture plates (Corning Inc., Corning, NY) with 300 μl serum-free Waymouth medium MB 752/1 (Invitrogen) that was supplemented with 25 mg/L pyruvic acid (Sigma Chemical Co., St. Louis, MO), 50 IU/ml penicillin, 50 μg/ml streptomycin (Invitrogen), and ITS+ (6.25 μg insulin, 6.25 μg transferrin, 6.25 ng selenous acid, 1.25 mg BSA, 5.35 μg/ml linoleic acid; Collaborative Biomedical Products, Becton Dickinson Labware, Bedford, MA). Cortical pieces were cultured at 38.5°C in a humidified incubator with 5% CO2 and 95% air, and 200 μl of culture medium was replaced with fresh medium every other day. The cultures were terminated on day 10 of culture.
In the first experiment, cortical pieces were isolated and cultured as described above, in defined medium without or with testosterone (10−7 and 10−6 M) or estradiol (10−6 M), to test the effects of these hormones on primordial follicle activation and subsequent growth and to determine whether the effects of testosterone are due to its conversion to estradiol. The second experiment was conducted to determine the optimal concentration of testosterone for preantral follicle growth. Cortical pieces were cultured in defined medium without or with graded concentrations of testosterone (10−10, 10−9, 10−8, and 10−7 M). The aim of the third experiment was to examine whether the effects of testosterone on preantral follicle growth are mediated by binding to the androgen receptor. Cortical pieces were cultured in defined medium with testosterone (10−7 M) alone or in combination with graded doses of flutamide (10−6 and 10−5 M; Sigma), which is a specific androgen receptor antagonist.
Testosterone, estradiol, and flutamide were dissolved initially in 100% ethanol to final stock concentrations of 2.38 M, 2.4 M, and 0.1 M, respectively, and then diluted to the desired concentrations in culture medium. The final ethanol concentration in the cultures did not exceed 0.042% (v/v). Each experiment was replicated with 2 or 3 fetuses that were obtained on separate occasions.
Assessment of Follicular Activation and Growth
Follicular activation and growth were assessed by histological morphometry, as previously described [5]. Briefly, after 0 or 10 days of culture, pieces of ovarian cortex were fixed and embedded in LR White plastic (EMS, Fort Washington, PA) and 2-μm serial sections were cut with a glass knife. Every other set of 10 consecutive sections from each piece of ovarian cortex was mounted on gelatin-coated slides and stained with toluidine blue. Sections were examined under an inverted microscope. To avoid counting or measuring the same follicle twice, only one section in each set of 10 consecutive sections was used and only follicles with the germinal vesicle present in that section were counted and measured. Follicles were first classified according to their stage of development as primordial (oocyte surrounded by one layer of flattened pre-granulosa cells), primary (oocyte surrounded by a single layer of cuboidal granulosa cells) or secondary (oocyte surrounded by two or more layers of cuboidal granulosa cells). Follicles were further classified as either healthy (intact germinal vesicle and nucleolus; oocyte with no more than three cytoplasmic vacuoles) or atretic (in early stages, oocyte with more than three cytoplasmic vacuoles and slight condensation of chromatin; in later stages, fragmentation of the oocyte cytoplasm and/or nucleus and heavy chromatin condensation) using criteria defined previously [5]. The microscopic image was projected on a video monitor. The diameters of individual healthy follicles and their oocytes were measured by a computer-driven image analysis program (NIH Image; NIH, Bethesda, MD). Each follicle and the enclosed oocyte were measured in two dimensions and the two measurements were averaged.
Immunohistochemistry for Androgen Receptors
Fetal bovine ovaries collected from fetuses (n = 13; 127 to 244 days of gestation, length of gestation = ~281 days) and antral follicles isolated from adult bovine ovaries (n = 3) were fixed in Bouin solution. All tissue samples were then embedded in paraffin, and 5-μm sections were cut, mounted on slides coated with poly-L-lysine (Poly-prep slides; Sigma), and dried overnight at 37°C and then for 15 min at 50°C. After deparaffinization in xylene and rehydration, the sections were microwaved for 20 min at 700 W to enhance antigen retrieval, with the evaporated water being replaced every 5 min. Endogenous peroxidases were blocked by incubating the sections with 3% H2O2 for 15 min. To prevent nonspecific reactions, sections were incubated with 1% bovine serum albumin in phosphate-buffered saline for 1 h and then with non-immune goat serum for 30 min at room temperature. A polyclonal rabbit anti-human androgen receptor antiserum (PG-21; Upstate, Lake Placid, NY) was used for immunohistochemical detection of androgen receptors. This antiserum was chosen because the supplier has reported it as being bovine-reactive and non-cross-reactive with estrogen or progesterone receptors. Each section was incubated with 3 μg/ml of primary antibody overnight at 4°C. The biotinylated secondary antibody (goat anti-rabbit IgG; Zymed Laboratories, San Francisco, CA) was then applied for 1 h at room temperature. All incubations were carried out in a humidified chamber. The location of the antigen/antibody/enzyme complex was detected and visualized using the Histostain-SP kit (Zymed Laboratories). Diaminobenzidine (DAB) chromogen substrate, which was added for 1 min, produced brown staining. Negative controls, which were included in each run, consisted of sections adjacent to the sections being tested but with the primary antibody replaced with non-immune rabbit IgG. Sections were counterstained with light green. For each ovary or follicle, sections were analyzed by immunohistochemistry on two or three different occasions with similar results.
To examine the specificity of the anti-human antiserum for bovine tissues, tissues reported to be positive (prostate, seminal vesicle, and heart) or negative (spleen) for androgen receptors in humans and mice [24] were collected from adult cattle at the abattoir. Each tissue was processed as described above for fetal ovaries.
Statistical Analysis
The mean numbers of total and healthy primordial, primary, and secondary follicles per section and the mean diameters of the healthy follicles and oocytes were calculated for each replicate (culture well) for each treatment. The data were log-transformed (base 10) before statistical analysis when the Hartley test indicated heterogeneity of variance. However, for ease of comprehension, the means ± SEMs of the non-transformed data are presented in the figures. Differences among the means were evaluated by two-way ANOVA, with treatment and fetus as the two factors, and statistical significance was assigned at P < 0.05. No significant effects of fetus or significant interactions between treatment and fetus were detected. When ANOVA indicated a significant effect of treatment, differences among the treatment means were tested using the Duncan multiple range test.
Results
Effects of Testosterone on Growth of Bovine Follicles In Vitro
Experiment 1. Effects of testosterone and estradiol on preantral follicular development
The purpose of the first experiment was to test the hypothesis that testosterone, but not estradiol, promotes the primary to secondary follicle transition. On day 0, cortical pieces contained mostly primordial follicles, whereas after 10 days in vitro, most of the primordial follicles had activated and differentiated into primary follicles, as expected based on previous studies (Fig. 1, A–D; [5]). Neither testosterone (10−7 and 10−6 M) nor estradiol (10−6 M) affected the total number or the number of healthy primordial or primary follicles compared with control cultures (P > 0.05; Fig. 1, A–D). However, both doses of testosterone increased the total number and the number of healthy secondary follicles by 3- to 4-fold (P < 0.05; Fig. 1, E and F). In contrast, estradiol had no effect (P > 0.05; Fig. 1, E and F). The diameters of the healthy primary follicles and of the few remaining primordial follicles were slightly larger after 10 days of culture than the diameters of the follicles in freshly isolated tissues (P < 0.05; Fig. 2, A and C), whereas oocyte diameter was not affected by culture (P > 0.05; Fig. 2, B and D). Neither testosterone nor estradiol had any effect on the diameters of healthy primordial, primary or secondary follicles or their oocytes (Fig. 2, A–F).
Effects of testosterone (T) and estradiol (E2) on types and numbers of follicles (mean ± SEM) in bovine cortical pieces cultured for 10 days (open bars). Black bars show the data for freshly isolated cortical pieces (Day 0). CON, Day 10 controls. Within each panel, means (bars) with no common letters are significantly different (P < 0.05; n = 6 cortical pieces, two from each of three fetuses; 36–43 sections examined per treatment).
Effects of testosterone (T) and estradiol (E2) on types and numbers of follicles (mean ± SEM) in bovine cortical pieces cultured for 10 days (open bars). Black bars show the data for freshly isolated cortical pieces (Day 0). CON, Day 10 controls. Within each panel, means (bars) with no common letters are significantly different (P < 0.05; n = 6 cortical pieces, two from each of three fetuses; 36–43 sections examined per treatment).
Effects of testosterone (T) and estradiol (E2) on the mean diameters (in μm ± SEM) of healthy primordial, primary, and secondary follicles and their oocytes in bovine cortical pieces cultured for 10 days (open bars). Black bars show the data for freshly isolated cortical pieces (Day 0). CON, Day 10 controls. Within each panel, means (bars) with no common letters are significantly different (P < 0.05; n = 6 cortical pieces, two from each of three fetuses; 36–43 sections examined per treatment).
Effects of testosterone (T) and estradiol (E2) on the mean diameters (in μm ± SEM) of healthy primordial, primary, and secondary follicles and their oocytes in bovine cortical pieces cultured for 10 days (open bars). Black bars show the data for freshly isolated cortical pieces (Day 0). CON, Day 10 controls. Within each panel, means (bars) with no common letters are significantly different (P < 0.05; n = 6 cortical pieces, two from each of three fetuses; 36–43 sections examined per treatment).
Experiment 2. Effects of graded doses of testosterone on preantral follicular development
Since the numbers of secondary follicles were similar in cortical pieces treated with 10−7 and 10−6 M testosterone in Experiment 1 (Fig. 1, E and F), in the second experiment, cortical pieces were cultured with a range of lower concentrations of testosterone, to determine the optimal concentration. After 10 days in culture, there were no differences in the numbers of total or healthy primordial or primary follicles between cortical pieces cultured without or with testosterone (Fig. 3, A–D). However, there was a dose-dependent increase in the numbers of total and healthy secondary follicles in cortical pieces after culture with 10−10–10−7 M testosterone, with the two highest concentrations (10−8 and 10−7 M) significantly stimulating of the primary to secondary follicle transition (Fig. 3, E and F).
Effects of graded doses (10−10 – 10−7 M) of testosterone (T) on the types and numbers of follicles (mean ± SEM) in bovine cortical pieces cultured for 10 days (open bars). Black bars show the data for freshly isolated cortical pieces (Day 0). CON, Day 10 controls; ND, not detected. Within each panel, means (bars) with no common letters are significantly different (P < 0.05; n = 6 cortical pieces, two from each of three fetuses; 28–34 sections examined per treatment).
Effects of graded doses (10−10 – 10−7 M) of testosterone (T) on the types and numbers of follicles (mean ± SEM) in bovine cortical pieces cultured for 10 days (open bars). Black bars show the data for freshly isolated cortical pieces (Day 0). CON, Day 10 controls; ND, not detected. Within each panel, means (bars) with no common letters are significantly different (P < 0.05; n = 6 cortical pieces, two from each of three fetuses; 28–34 sections examined per treatment).
Experiment 3. Effects of flutamide on preantral follicular development
The specific androgen receptor-blocking agent flutamide was used to determine whether the stimulatory effects of testosterone on early preantral follicle growth are mediated by the binding of testosterone to the androgen receptor. As in Experiments 1 and 2, testosterone increased the numbers of total and healthy secondary follicles (P < 0.05; Fig. 4, E and F). The combination of testosterone and flutamide had no effect on the numbers of total or healthy primordial or primary follicles (Fig. 4, A–D). In contrast, cotreatment with flutamide and testosterone decreased the numbers of total and healthy secondary follicles in a dose-dependent manner, compared with testosterone alone, to levels that were similar to those observed in cortical pieces cultured in control medium (Fig. 4, E and F). Flutamide alone had no effect on the numbers of follicles in any developmental category (Fig. 4).
Inhibitory effect of flutamide (F), an androgen receptor antagonist, on the testosterone-induced (T; 10−7 M) increase in the number of secondary follicles (mean ± SEM) in bovine cortical pieces cultured for 10 days (open bars). Black bars show the data for freshly isolated cortical pieces (Day 0). CON, Day 10 controls. Within each panel, means (bars) with no common letters are significantly different (P < 0.05; n = 4 cortical pieces, two from each of two fetuses; 18–21 sections examined per treatment).
Inhibitory effect of flutamide (F), an androgen receptor antagonist, on the testosterone-induced (T; 10−7 M) increase in the number of secondary follicles (mean ± SEM) in bovine cortical pieces cultured for 10 days (open bars). Black bars show the data for freshly isolated cortical pieces (Day 0). CON, Day 10 controls. Within each panel, means (bars) with no common letters are significantly different (P < 0.05; n = 4 cortical pieces, two from each of two fetuses; 18–21 sections examined per treatment).
Immunohistochemical Localization of Androgen Receptors
Since the experiments described above provided evidence for a receptor-mediated effect of testosterone on the primary to secondary follicle transition, we localized the androgen receptors in bovine fetal ovaries (n = 13) and large antral follicles (n = 3; 5–10 mm) isolated from three adult ovaries. Androgen receptors were observed in different types of ovarian cells and the staining intensities differed for follicles at different stages of development (Fig. 5). There was widespread and strong staining for the androgen receptor in cells of the ovarian stroma (Fig. 5A). Compared with the negative controls (Fig. 5, B and D), the granulosa cells of the primordial and primary follicles showed no immunoreactivity for the androgen receptor (Fig. 5A), whereas those of the secondary follicles exhibited weak androgen receptor staining (Fig. 5C). The emerging presumptive theca cells around the secondary follicles displayed moderate reactivites (Fig. 5C). In addition, the oocytes of some preantral follicles were stained weakly (Fig. 5, A and C). At the antral stage, strong staining for androgen receptors was typically present in both granulosa and theca cells of healthy follicles (Fig. 5E), whereas atretic follicles had moderate androgen receptor staining in the theca but almost no immunoreactivity in degenerating granulosa cells (Fig. 5F). The specificity of the antiserum was tested with three tissues (prostate, seminal vesicle, and heart) that have been reported to be positive and one tissue (spleen) that has been reported to be negative for androgen receptors in mice and humans [24]. The expected specificity was observed with the bovine tissues, with staining observed in the prostate, heart, and seminal vesicle (Fig. 5, G–I) but not in the spleen (Fig. 5J).
Immunohistochemical staining for androgen receptors in different types of bovine cells. A–F) Typical results for the 13 fetal ovaries and three large antral follicles from three adult ovaries examined. A and C) Ovarian stromal cells (S) exhibit staining for androgen receptors. Staining for androgen receptors is not detected in the granulosa cells of primordial (PrF; arrowhead) and primary (PF; small arrow) follicles, whereas secondary (SF; large arrow) follicles show weak immunoreactivity. Emerging presumptive thecal cells (T) around the secondary follicle display moderate staining for androgen receptors. Staining is also observed in the oocytes of some follicles. B and D) Negative controls for A and C; staining is absent. E) Very strong staining for androgen receptors is present in both the granulosa (G) and thecal cells of healthy antral follicles. F) Atretic follicles have moderate androgen receptor staining in the theca, while degenerating granulosa cells show almost no immunoreactivity. G–I) Staining is detected in the prostate, heart, and seminal vesicle, respectively. J) No specific staining is detected in the spleen. K) Negative control for J. Bars = 100 μm; H–K are at the same magnification as G.
Immunohistochemical staining for androgen receptors in different types of bovine cells. A–F) Typical results for the 13 fetal ovaries and three large antral follicles from three adult ovaries examined. A and C) Ovarian stromal cells (S) exhibit staining for androgen receptors. Staining for androgen receptors is not detected in the granulosa cells of primordial (PrF; arrowhead) and primary (PF; small arrow) follicles, whereas secondary (SF; large arrow) follicles show weak immunoreactivity. Emerging presumptive thecal cells (T) around the secondary follicle display moderate staining for androgen receptors. Staining is also observed in the oocytes of some follicles. B and D) Negative controls for A and C; staining is absent. E) Very strong staining for androgen receptors is present in both the granulosa (G) and thecal cells of healthy antral follicles. F) Atretic follicles have moderate androgen receptor staining in the theca, while degenerating granulosa cells show almost no immunoreactivity. G–I) Staining is detected in the prostate, heart, and seminal vesicle, respectively. J) No specific staining is detected in the spleen. K) Negative control for J. Bars = 100 μm; H–K are at the same magnification as G.
Discussion
Together, the results provide the first evidence for a role for testosterone in early follicular development in cattle, particularly in the primary to secondary transition. When newborn rodent ovaries are cultured, follicles develop spontaneously to the secondary stage in vitro, but this has previously not been the case in cultured cortical pieces from bovine or primate ovaries, with very few follicles progressing to the secondary stage. The effects of testosterone on the primary to secondary follicle transition were dose-dependent, were not due to aromatization to estradiol, and were inhibited by an antagonist to the androgen receptor. Taken together with the findings on the patterns of expression of androgen receptors in bovine ovaries, these results strongly suggest a role for androgens in follicle development in vivo, as well as in vitro. In addition to their economic importance, cattle provide an excellent animal model for human follicular development. Therefore, the current evidence that a critical transition in early follicular development is promoted by testosterone, which is the first factor reported to stimulate this transition in cattle, is a significant step towards the long-term goal of complete in vitro growth of follicles from species of practical importance.
Our previous and current studies have shown that although the vast majority of primordial follicles are activated in cultures of cortical pieces isolated from fetal bovine ovaries, only a few activated follicles develop to the secondary stage in our serum-free culture system. These results suggest that the conditions required to induce follicle activation in vitro may not be adequate to support further growth to the secondary and small preantral stages. We hypothesize that the addition of one or more specific hormones and/or growth factors or longer culture time is needed for the primary to secondary follicle transition in the cortical culture system. In a previous study, when we cultured bovine cortical pieces with combinations of graded doses of ITS+ and fetal bovine serum (FBS), a small but significant increase in the number of secondary follicles was observed when half-strength ITS+ was combined with 5% FBS [10]. The results of the present study show that the addition of testosterone to the culture medium has a dose-dependent and very robust stimulatory effect on the number of secondary follicles. The dose of testosterone (10−7 M; 28.8 ng/ml) that was most effective in the present study is in the physiological range [13]. These findings imply that testosterone stimulates the primary to secondary follicle transition in cattle. The increases in secondary follicle numbers were not accompanied by significant decreases in the primary or primordial follicle numbers. However, the total number of follicles was not affected by androgen treatment (data not shown), which suggests that testosterone does not cause the formation of new follicles. It is possible that the loss of follicles from the primary pool was simply too small numerically to be detected statistically.
The finding that testosterone promotes the primary to secondary transition in bovine cortical pieces in vitro is consistent with and extends the results of previous in vitro and in vivo studies with mice and primates. Androgens, but not estradiol, have been shown to increase the in vitro growth rates of medium-sized [20] and large [19] preantral follicles isolated from mouse ovaries. Treatment of rhesus monkeys with testosterone or the non-aromatizable androgen 5α-dihydrotestosterone (DHT) for 3, 5, or 10 days increased the number of growing preantral and small antral follicles [21, 22], and it has been proposed that elevated androgen production by thecal-interstitial cells in women with polycystic ovarian syndrome stimulates an abnormally large cohort of developing follicles [25]. In contrast, the addition of testosterone (10−8 M) to cultures of human ovarian cortex had no effect on the percentage of follicles that reached secondary stage [26], but the dose used in that study was 10-fold lower than the dose that was maximally effective in bovine cortical cultures in the current study.
Thus, the results of the experiments with mice and rhesus monkeys support a stimulatory role for androgens in late preantral follicular development, and the results of the current study further suggest that the stimulatory effects of androgens commence at the late primary stage to facilitate the transition to the secondary stage. In contrast, the reported effects of androgens on the development of antral follicles are much more diverse. Although there is evidence to suggest that androgens are atretogenic for antral follicles (reviewed in [14]), there is also evidence that androgens increase the expression of FSH receptors and enhance the positive effects of FSH on follicular growth, granulosa cell proliferation, and aromatase activity [27–30]. Although the levels of androgens are higher and the estrogen to androgen ratios are lower in the follicular fluid of atretic versus healthy large antral follicles in a number of species, this may simply reflect a decrease in follicular aromatizing capacity as the granulosa cells degenerate. Westergaard et al. [31] have reported that a high level of androgen in human follicular fluid is not by itself a marker of follicular atresia, but reflects the stage of follicular development. They found that in large antral follicles (> 6 mm in diameter), the intrafollicular hormonal milieu was dominated by androgens when the follicles were atretic, as opposed to estrogen predominance when the follicles were healthy, whereas small antral follicles (healthy as well as atretic) were always androgenic. In addition, the mode of androgenic modulation of FSH action on marmoset granulosa cells in vitro switches from stimulatory to inhibitory as follicular development advances [32]. Therefore, it appears that the effects of androgens on the development of antral follicles are very complex and stage-dependent, as opposed to their largely stimulatory effects on preantral follicles.
Androgens can exert direct effects mediated by androgen receptors, indirect effects mediated by their aromatization to estrogens, or both types of effect. In the current study, the number of secondary follicles increased in cortical pieces treated with testosterone whereas estradiol had no effect, which provides evidence that the stimulatory effect of testosterone is not due to conversion of testosterone to estradiol. These findings are consistent with the results of studies with primates and mice showing that treatment with the non-aromatizable androgen DHT had effects identical to testosterone in stimulating preantral follicle growth in vivo or in vitro [20–22]. Moreover, exogenous estrogen administration inhibited the growth of both preantral and medium-sized antral follicles in cynomolgus monkeys [33]. In the current study, coculture with testosterone and the specific androgen receptor antagonist flutamide inhibited the increase in secondary follicles stimulated by testosterone alone, which suggests that the stimulatory effect of testosterone on the primary to secondary follicle transition is a specific, receptor-mediated action, consistent with the effects of androgen receptor antagonists on androgen-stimulated in vitro growth of large and medium-sized secondary follicles isolated from mice [19, 20].
Little was known about the expression of androgen receptors or their mRNA in the bovine ovary; to our knowledge, there has been only one previous study. Hampton et al. [34] examined the expression of androgen receptor mRNA in bovine ovaries by in situ hybridization and found that although expression is absent in bovine primordial follicles, it is present in the granulosa cells of some primary follicles and in all the secondary and early antral follicles examined. In the present study, immunohistochemistry for the androgen receptor revealed an absence of immunoreactivity in the primordial and primary follicles, while staining was observed in granulosa cells as follicular development progressed to the secondary stage, with very intense staining observed in healthy antral follicles. However, we also detected strong immunoreactivity in thecal cells, beginning at the early secondary stage and increasing in intensity in the antral follicles, and in widespread stromal cells. Although Hampton et al. [34] detected expression of the mRNA in about 40% of primary follicles, we did not detect immunoreactivity in the primary follicles, which indicates that the actions of androgen on primary follicles may be indirect via receptor-mediated effects on stromal cells or direct via small numbers of receptors on granulosa cells of the primary follicles. Consistent with our results, studies with sheep [35] and rhesus monkeys [36] have shown little or no androgen receptor mRNA in primary follicles, while abundant expression has been detected in granulosa cells of healthy secondary to small antral follicles, with the theca interna and stromal cells also expressing mRNAs for androgen receptors. In contrast, specific immunoreactivity for the androgen receptor was minimal or not detected in the thecal and stromal cells of rat and marmoset ovaries [37, 38], although another group reported positive reactions in these cell types in human and marmoset ovarian tissue [39]. Immunostaining for androgen receptors increases as follicular development progresses from primordial and primary to secondary preantral follicles to antral follicles in canine [40], non-human primate [41], and human [42] ovaries, similar to our results for cattle. Moreover, in agreement with previous studies on other species [41, 42], both thecal and granulosa cells of healthy bovine antral follicles and thecal cells of antral follicles in advanced stages of atresia stained strongly for androgen receptors, and staining was observed in some oocytes. The strong expression in antral follicles of protein or mRNA for androgen receptors becomes attenuated as follicles develop to the preovulatory stage in rats, rhesus monkeys, and marmosets [36–38].
Folliculogenesis is estimated to take several months in large mammals, such as humans and cattle [43, 44]. Even in rodents, complete follicular development in adult ovaries is estimated to take about 60 days, with the early stages proceeding considerably more slowly than the later stages [45]. Although it is not known how long the primary-secondary follicle transition takes in bovine or primate ovaries in vivo, one can argue that simply extending the culture period may be more effective for promoting the development of secondary follicles than the addition of specific growth factors/hormones, but there were still very few follicles at the secondary stage in baboon cortical pieces after 20 days in culture [6]. In the present study, testosterone dramatically increased the number of secondary follicles in bovine cortical pieces cultured for 10 days, which suggests that the addition of specific hormones and/or growth factors is more effective than longer culture time for promoting the primary to secondary follicle transition. Most of the secondary follicles observed in our study were at the early secondary stage and testosterone did not promote further progression of these follicles to later stages even after more extended culture (20 days, Yang and Fortune, unpublished observations). This suggests that additional growth factors, such as bone morphogenetic protein-15 (BMP-15) and/or growth and differentiation factor-9 (GDF-9), which have been reported to be essential for normal follicular development beyond the primary stage in other species [46, 47], may be necessary for the subsequent development of bovine preantral follicles in vitro.
In summary, our results show that testosterone exerts a receptor-dependent stimulatory effect on the bovine primary to secondary follicle transition in vitro. In newborn rodent ovaries, this transition occurs spontaneously in vitro and can be enhanced by several hormones/growth factors (reviewed in [48]). In contrast, we have added hormones and growth factors to the culture medium in an attempt to promote the development of secondary follicles in cultured bovine ovaries, without success [49]. Therefore, the effects of testosterone reported in the current study represent both a first step towards understanding the signals that allow primary follicles of large mammals to progress to the secondary stage and a step towards the long-term practical goal of growing follicles activated in vitro to the stage of competence for meiotic maturation and fertilization in vitro.
Acknowledgments
The authors thank Taylor Packing Co. (Wyalusing, PA) for the donation of bovine ovaries and Dr. P.J. Bridges for reading a draft of the manuscript.
