Transforming Growth Factor-β Signaling Participates in the Maintenance of the Primordial Follicle Pool in the Mouse Ovary*

Background: Why only a few follicles are activated to enter the growing follicle pool each wave remains unclear. Results: TGF-β regulates oocyte growth through p70 S6 kinase 1/ribosomal protein S6 signaling. Conclusion: TGF-β participates in maintenance of the primordial follicle pool. Significance: Learning how TGF-β acts on primordial follicle growth. Physiologically, only a few primordial follicles are activated to enter the growing follicle pool each wave. Recent studies in knock-out mice show that early follicular activation depends on signaling from the tuberous sclerosis complex, the mammalian target of rapamycin complex 1 (mTORC1), phosphatase and tensin homolog deleted on chromosome 10, and phosphatidylinositol 3-kinase (PI3K) pathways. However, the manner in which these pathways are normally regulated, and whether or not TGF-β acts on them are poorly understood. So, this study aims to identify whether or not TGF-β acts on the process. Ovary organ culture experiments showed that the culture of 18.5 days post-coitus (dpc) ovaries with TGF-β1 reduced the total population of oocytes and activated follicles, accelerated oocyte growth was observed in ovaries treated with TGF-βR1 inhibitor 2-(5-chloro-2-fluorophenyl)pteridin-4-yl]pyridin-4-yl-amine (SD208) compared with control ovaries, the down-regulation of TGF-βR1 gene expression also activated early primordial follicle oocyte growth. We further showed that there was dramatically more proliferation of granulosa cells in SD208-treated ovaries and less proliferation in TGF-β1-treated ovaries. Western blot and morphological analyses indicated that TGF-β signaling manipulated primordial follicle growth through tuberous sclerosis complex/mTORC1 signaling in oocytes, and the mTORC1-specific inhibitor rapamycin could partially reverse the stimulated effect of SD208 on the oocyte growth and decreased the numbers of growing follicles. In conclusion, our results suggest that TGF-β signaling plays an important physiological role in the maintenance of the dormant pool of primordial follicles, which functions through activation of p70 S6 kinase 1 (S6K1)/ribosomal protein S6 (rpS6) signaling in mouse ovaries.

Reproductive productivity and lifespan in the female mouse are determined by follicle assembly and development. In fact, the activation amount from primordial follicles to growing follicles is more important to determine the exhausted velocity of primordial follicle pool. Naturally, primordial follicles are assembled from germ cell cysts undergoing programmed breakdown and pre-granulosa cells that invade germ cell cysts at ϳ19.5 days post-coitus (dpc) 4 in the mouse (1,2). The primordial follicle pool is established by 4 days post-parturition (dpp) and the number of primordial follicles reaches almost 80% by 6 dpp (2)(3)(4). The majority of primordial follicles in the ovary remain quiescent until they are activated to enter the growing follicle pool (5,6). The formation of primary follicles following the activation of primordial follicles is characterized by the change of granulosa cells from a squamous to a cuboidal morphology as well as an increase in oocyte size (7). Mature oocytes are generated from antral follicles, which arise as a result of primordial follicle development. Therefore, the activation of primordial follicles and the formation of growing follicles are of critical importance to female reproductive potential.
Recent studies have demonstrated that growth differentiation factor 9 (GDF9) and inhibin ␣ influence follicular development (8 -10). Additionally, the phosphatidylinositol 3-kinase (PI3K) and phosphatase and tensin homolog deleted on chromosome 10 (PTEN; PTEN/PI3K) and the tuberous sclerosis complex (TSC) and mammalian target of rapamycin complex 1 (mTORC1; TSC/mTORC1) signaling pathways have been found to participate in the regulation of follicular activation. Deletion of Forkhead box protein O3 (FOXO3a) or Pten in oocytes causes the excessive activation and depletion of primordial follicles, whereas Pten ablation also leads to increased FOXO3a phosphorylation and nuclear export (11)(12)(13)(14)(15). The PTEN/PI3K and TSC/mTORC1 pathways in oocytes regulate follicular dormancy and activation via p70 S6 kinase 1 (S6K1)/ * This work was supported by the National Basic Research Program Grants ribosomal protein S6 (rpS6) signaling through the phosphorylation of S6K1 at different threonine residues (16,17). The above mentioned signaling pathways are critical for cell proliferation and growth and their disruption may result in certain phenomena. It is difficult to determine whether these processes occur in vivo, for the activation only occurred in a few primordial follicles each wave, physiologically. So, it is important to identify the factor(s) that modulate these or other pathways during the regulation of early primordial follicular oocyte growth.
Transforming growth factor-␤ (TGF-␤) regulates a variety of biological processes in mammals by influencing cell proliferation, growth, differentiation, and apoptosis (18 -20). TGF-␤ binds to serine/threonine kinase receptor types I and II on the cell surface to form a complex that activates the Smad signaling pathway via a phosphorylation of Smad proteins. The phosphorylated Smad proteins then associate with the common Smad4 and translocate from the cytoplasm to the nucleus where they regulate target gene transcription (21)(22)(23). An increasing amount of research has demonstrated that TGF-␤ proteins play important roles in embryonic gonadal development and follicular development in mice. The expressions of TGF-␤1, TGF-␤2, TGF-␤R1, and TGF-␤R2 mRNA and proteins have been identified in the mouse ovary and previous immunolocalization experiments have detected Smads 2, 3, 4, and 6 in the mouse ovary (23)(24)(25). Although TGF-␤2 and TGF-␤3 knock-out mice die perinatally (26,27), TGF-␤1-null mice fail to display the normal surge of luteinising hormone and exhibit a 40% reduction in oocyte ovulation as well as impaired fertility (28). The treatment of cultured 0-day-old rat ovaries with TGF-␤1 for 2 days had no effect on follicle assembly or the number of oocytes but the primordial follicle pool size was reduced after 10 days (29). In another in vitro study, TGF-␤1 inhibits follicle development and progression in the rat ovary in the presence of folliclestimulating hormone (25). Although these studies provide evidence that TGF-␤1 may participate in the regulation of follicular development, its effect on oocyte growth in early follicular development remains unclear. Because nearly all TGF-␤ isoform-specific knock-out mice die perinatally and the role of TGF-␤ remains unclear, the present study, using a mouse fetal ovary culture system, is undertaken to investigate whether TGF-␤ is involved in early follicular development and whether it influences this process through PTEN/PI3K or TSC/ mTORC1 signaling or an alternative pathway.

EXPERIMENTAL PROCEDURES
Animals-Kunming white mice were purchased from the Laboratory Animal Centre of the Institute of Genetics (Beijing, China) and maintained in the University Animal Care Facility with free access to food and water under a light-dark cycle. Female mice (6 to 8 weeks old) were mated with adult male mice to induce pregnancy. Mice with a vaginal plug the next morning were considered to be at 0.5 dpc. All procedures were performed in accordance with institutional and national guidelines and regulations and were approved by the China Agricultural University Animal Care and Use Committee.
Fetal Ovary Culture and Chemicals-Ovaries were dissected from mice as described previously (30). Fetal ovaries were cultured in a 24-well culture plate at 37°C in 1 ml of Dulbecco's modified Eagle's medium/nutrient mixture F-12 (DMEM/F-12; 1:1, v/v; Invitrogen) containing 10% fetal bovine serum (FBS; Sigma), 10 g/ml of insulin (Sigma), and 10 g/ml of transferrin (Sigma) in an atmosphere of 5% CO 2 and 95% air. The medium was supplemented with penicillin and streptomycin to prevent bacterial contamination and was changed every other day.
Ovarian Follicle Counts-Ovarian germ cells and follicles were quantified according to a widely used approach. Ovaries were fixed, embedded in paraffin, and sectioned to a thickness of 5 m. Serial sections were stained with hematoxylin and every fifth section was analyzed for the presence of germ cells and follicles. The follicles were distinguished from each other as follows: primordial follicle (a single oocyte surrounded by several flattened pre-granulosa cells) and growing follicle (an enlarged oocyte surrounded by a mixture of squamous and cuboidal somatic cells or an enlarged oocyte surrounded by one layer of cuboidal granulosa cells). Finally, cumulative germ cell and follicle counts were multiplied by five because four-fifths of the ovary was not analyzed (32). Another method used for determining the number of oocytes involved counting the number of oocytes across two consecutive sections of the center of the ovary with the largest cross-section; these were then averaged. Previously, data obtained from this method of analysis using two mid-diameter cross-sections have been shown to be similar to those from analyses of compiled data from all serial sections (33). Only oocytes with a visible nucleus were counted, and 100 -300 oocytes were present in each cross-section.
TUNEL Staining-The degree of oocyte apoptosis was measured by terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate (TUNEL) assay using an In Situ Apoptosis Detection Kit (S7101; Millipore). Ovaries were fixed, embedded in paraffin, and sectioned to a thickness of 5 m. Paraffin-embedded slides were then treated according to the manufacturer's instructions in the apoptosis detection kit. Follicles positively stained by TUNEL had multiple breaks in their DNA and were deemed to be undergoing apoptosis at the time of fixation.
Western Blotting-Total protein from the ovaries was extracted with a MEM-R kit (Pierce) according to the manufacturer's protocol and protein concentrations were measured using a bicinchoninic acid procedure (CellChip; BJ Biotechnology Co., Ltd., Beijing, China). Proteins were separated with 10% SDS-PAGE and then electrophoretically transferred onto pieces of Protran nitrocellulose membrane (Schleicher and Schuell, Dassel, Germany). After the transfer, membranes were incubated for 1 h at room temperature in 5% bovine serum albumin (BSA) in Tris-buffered saline with Tween (TBST; 20 mM Tris-HCl, 150 mM NaCl, and 0.1% Tween 20, pH 7.6). The relative quantity of each gene was calculated by the 2 Ϫ⌬⌬Ct method as described previously (34), and normalized to the endogenous ␤-actin reference gene.
RNA Interference (RNAi)-To assure that siRNAs would be transfected into the inner cells of fetal ovaries, 0.5 l of the siRNAs were first injected into isolated fetal ovaries at 18.5 dpc using glass pipettes with a stereomicroscope. After the ovaries were full of liquid, electrotransfection was accomplished by applying three 5-ms long quasi-square pulses at a pulse-field strength of up to 40 V/cm. TGF-␤R1 siRNA was purchased from Sigma. Control siRNA contained a scrambled siRNA sequence that did not lead to the specific degradation of any known mouse mRNA. The ovaries then were cultured for 48 h to test the transfection efficiency of mRNA levels using realtime PCR, for 72 h to test protein levels using Western blotting, or for 7 days for histological examination and follicle counting.
Statistical Analysis-Each experiment was repeated at least three times and the values are the mean Ϯ S.E. Data were analyzed using an analysis of variance with the StatView software (SAS Institute, Inc., Cary, NC). Values of p Ͻ 0.05 were regarded as statistically significant.

The Expression of TGF-␤1 and TGF-␤R1 in the Fetal Mouse
Ovary-To better understand the actions of TGF-␤ during the early stage of mouse ovary development, immunohistochemical and Western blot analyses were employed to determine the cellular localization and expression of TGF-␤1 and TGF-␤R1. The cytoplasm of germ cells with the cyst structure in 18.5-dpc mouse ovaries exhibited positive staining for TGF-␤1 and TGF-␤R1, whereas somatic cells exhibited negative staining (Fig. 1, A and D). Both TGF-␤1 and TGF-␤R1 continued to be expressed in the oocytes of primordial and primary follicles in 1 and 4 dpp mouse ovaries as well as the cuboidal granulosa cells of primary follicles, but not in the flattened pre-granulosa cells of primordial follicles ( Fig. 1, B, C, E, and F). Western blotting results revealed that the TGF-␤1 protein was expressed at low levels from 17.5 to 19.5 dpc, gradually increased from 1 to 4 dpp, and slightly decreased at 7 dpp compared with 4 dpp levels in fetal mouse ovaries. TGF-␤R1 protein levels were similar to those of TGF-␤1 in that they declined to a nadir at 7 dpp relative to 4 dpp (Fig. 1G).
The Attenuation of TGF-␤ Signaling Activates Early Primordial Follicle Growth-To investigate the function of TGF-␤ during early follicular development, ovaries at 18.5 dpc were cultured without treatment (as a control), with 10 ng/ml of TGF-␤1, or with 1 M SD208 (an inhibitor of TGF-␤R1) for 7 days. After culture, ovaries were fixed, sectioned, and the shape and the numbers of germ cells and growing follicles were evaluated. Following 7 days of in vitro culture (equivalent to 6 dpp), ovaries were chiefly composed of primordial follicles with some activated follicles, which included transient and primary follicles (Fig. 2, A-I). The growth of primordial follicle oocytes in TGF-␤1-treated ovaries was inhibited (Fig. 2, D-F), whereas, the growth of oocytes in SD208-treated ovaries was faster (Fig.  2, G-I), and the activated oocyte diameter was larger compared with control ovaries (Fig. 2, A-C). Moreover, the number of germ cells and growing follicles in TGF-␤1-treated ovaries (3,421 Ϯ 427 and 128 Ϯ 27, respectively) were significantly reduced compared with control ovaries (5,146 Ϯ 495 and 313 Ϯ 43, respectively). SD208-treated ovaries contained a similar number of germ cells (4,970 Ϯ 413) but more activated follicles (763 Ϯ 58) than the control group (Fig. 2J). Furthermore, we cultured 3-dpp mouse ovaries for 5 days without treatment (as a control), with 10 ng/ml of TGF-␤1, or with 1 M SD208. As shown in supplemental Fig. S1, the ovarian morphology results revealed that the growth of primordial follicle oocytes in TGF-␤1-treated ovaries was significantly inhibited, whereas the growth of oocytes in SD208-treated ovaries was apparently accelerated. The activated oocyte diameter was larger in SD208-treated ovaries and smaller in TGF-␤1-treated ovaries compared with control ovaries (supplemental Fig. S1, A-F).
The Down-regulation of TGF-␤R1 Expression Activates Early Primordial Follicle Growth-To further investigate the action of TGF-␤, a siRNA-mediated knockdown of TGF-␤R1 gene expression was employed to attenuate TGF-␤ signaling in the fetal mouse ovary. Ovaries at 18.5 dpc were transfected with scrambled or TGF-␤R1 siRNA in vitro to evaluate transfection efficiency and ovarian histology. Real-time PCR and Western blot analyses revealed an obvious decrease in TGF-␤R1 gene and protein expressions following transfection with TGF-␤R1 siRNA (Fig. 3A). Both scrambled and TGF-␤R1 siRNA-treated ovaries contained mostly primordial follicles. However, the growth of oocytes in primordial follicles was more accelerated (Fig. 3, E and F) and the number of growing follicles was greater in TGF-␤R1 siRNA-treated ovaries (Fig. 3B) compared with scrambled siRNA-treated ovaries (Fig. 3, C and D). Similar to SD208 treatment, TGF-␤R1 siRNA-treated ovaries displayed enlarged growing follicles relative to control ovaries.
TGF-␤1 Partially Reverses the Effect of SD208 on Primordial Follicle Growth-Because rapid oocyte growth was observed in ovaries treated with SD208, it was investigated whether TGF-␤1 could partially reverse the effects of SD208 on oocyte growth. Ovaries at 14.5 dpc were cultured in medium containing 4% fetal bovine serum (FBS) for 4 days (equivalent to 18.5 dpc) to make them adhere to the dish and then cultured in fresh medium without treatment as a control (Fig. 4E), with 10 ng/ml of TGF-␤1, 1 M SD208 (Fig. 4F), 1 M SD208 plus 10 ng/ml of TGF-␤1, 1 M SD208 plus 20 ng/ml of TGF-␤1, 1 M SD208 plus 50 ng/ml of TGF-␤1, or 1 M SD208 plus 100 ng/ml of TGF-␤1 (Fig. 4G) for 7 days. Additionally, 18.5-dpc ovaries were directly cultured for 7 days as described above with the different treatments. Following culture, ovaries were fixed and sectioned (Fig. 4, A-D) and revealed a higher number of activated follicles in SD208-treated ovaries (Fig. 4, C and F). Although there were no significant differences in the number of germ cells among the ovaries (data not shown), TGF-␤1 partially reduced the number of growing follicles at high doses in SD208-treated ovaries (Fig. 4H).
Effects of TGF-␤1 and SD208 Treatments on Granulosa Cell Proliferation and Oocyte Apoptosis-To determine the cause of rapid oocyte growth, it was investigated whether the proliferation of granulosa cells and the apoptosis of oocytes were affected in 18.5-dpc ovaries following treatment with 10 ng/ml of TGF-␤1 or 1 M SD208. There was dramatically less proliferation of granulosa cells in TGF-␤1-treated ovaries and significantly more proliferation in SD208-treated ovaries after 7 days of culture as shown by incorporation assays (Fig. 5, A-C) and immunohistochemical staining for proliferating cell nuclear antigen (PCNA) (Fig. 5, D-F). In the follicles of these ovaries, less granulosa cells were BrdU-or PCNA-positive in TGF-␤1treated ovaries (Fig. 5, B and E) and significantly more granulosa cells were BrdU-or PCNA-positive in SD208-treated ova- ries (Fig. 5, C and F) compared with control ovaries (Fig. 5, A  and D). Terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate (TUNEL) assays for apoptosis were performed on ovarian sections after 5 days of culture and revealed more TUNEL-positive oocytes in TGF-␤1-treated ovaries (Fig. 5, G-I). Furthermore, the number of apoptotic oocytes per section was greater in TGF-␤1-treated ovaries compared with control and SD208-treated ovaries (Fig. 5K). Western blot results showed that the expression of cleaved caspase-3 (the ultimate apoptosis effector) was significantly elevated in TGF-␤1-treated ovaries but not altered in SD208treated ovaries compared with control ovaries (Fig. 5J).
Alterations of Growth Dynamics and the Activation Status of Akt, FOXO3a, S6K1, and rpS6 in TGF-␤1and SD208-treated Ovaries-To investigate the molecular mechanisms underlying the accelerated enlargement of oocytes, the changes in growth factors known to be involved in oocyte growth and development in treated ovaries were analyzed by real-time PCR after 7 days of culture. Levels of Gdf9 mRNA in control ovaries were higher than in TGF-␤1-treated ovaries but lower relative to SD208-treated ovaries (Fig. 6A). Levels of Inhibin ␣ were significantly decreased in SD208-treated ovaries compared with control or TGF-␤1-treated ovaries (Fig. 6B). Next, whether this phenomenon involved the activation of PTEN/PI3K or S6K1-rpS6 signaling was also investigated. Ovaries at 18.5 dpc were cultured alone or with 10 ng/ml of TGF-␤1 or 1 M SD208 for 5, 7, or 9 days, respectively. Western blot analyses were performed to detect the phosphorylated forms of protein kinase B (Akt), FOXO3a, S6K1, and rpS6. The results showed that the phosphorylation levels of Akt at its serine 473 (p-Akt, Ser-473) and FOXO3a at its threonine 32 (p-Foxo3a, Thr-32) in SD208treated ovaries were similar to control and TGF-␤1-treated FIGURE 2. Phenotypes of ovary cultures treated with TGF-␤1 and SD208. Ovaries at 18.5 dpc were cultured for 7 days without treatment (as a control), with 10 ng/ml of TGF-␤1, or with 1 M SD208. Following culture, ovaries were fixed, sectioned, and the shape and number of total germ cells and growing follicles were evaluated. All cultured ovaries exhibited mostly primordial follicles and some activated follicles. However, the primordial follicle oocytes in TGF-␤1treated ovaries seemed to grow more slowly (D-F), whereas those in the SD208-treated ovaries appeared to grow faster (G-I, arrowheads) compared with control ovaries (A-C). The numbers of total germ cells and growing follicles were quantified (J); germ cell data were obtained from at least six ovaries. The ovaries (Fig. 6, C and E). However, the phosphorylation of S6K1 at threonine 389 (p-S6K1, Thr-389) and the phosphorylation of rpS6 at its serine 235 and 236 (p-rpS6, Ser-235/Ser-236) and at its serine 240 and 244 (p-rpS6, Ser-240/Ser-244) were elevated in SD208-treated ovaries but reduced in TGF-␤1-treated ovaries compared with control ovaries (Fig. 6, G, I, and J).
Rapamycin Partially Prevents the Accelerated Primordial Follicle Oocyte Growth in SD208-treated Ovaries-The phosphorylation of S6K1 (Thr-389) and rpS6 (Ser-235/Ser-236, Ser-240/Ser-244) were elevated in SD208-treated ovaries but reduced in TGF-␤1-treated ovaries, and the phosphorylation levels of Akt (Ser-473) and FOXO3a (Thr-32) were not altered in TGF-␤1-and SD208-treated ovaries (Fig. 6). To provide evidence that it is not PI3K-Akt signaling but the elevated mTORC1 activity that drives the accelerated oocyte growth in SD208-treated ovaries, we cultured 3-dpp mouse ovaries for 5 days without treatment (as a control), with 1 M SD208, 25 M LY294002, 50 nM rapamycin, 1 M SD208 plus 25 M LY294002, or with 1 M SD208 plus 50 nM rapamycin. Following culture, ovaries were fixed, sectioned, and the morphological analyses and the average numbers of growing follicles were evaluated. Clusters of primordial follicles and some activated follicles were seen in all cultured ovaries, whereas more primordial follicles were activated into growing follicles with enlarged oocytes in SD208-treated ovaries (Fig. 8, B and BЈ) as compared with control ovaries (Fig. 8, A and AЈ). The ovaries treated only with LY294002 (Fig. 8, C and CЈ) or rapamycin (Fig. 8, D and DЈ) seemed to show fewer numbers of growing follicles compared FIGURE 4. TGF-␤1 partially reverses the effect of SD208 on primordial follicle oocyte growth. Ovaries at 14.5 dpc were cultured in medium containing 4% FBS for 4 days (equivalent to 18.5 dpc) to make them adhere to the dish and then cultured in fresh medium without treatment as a control (E), with 10 ng/ml of TGF-␤1, 1 M SD208 (F), 1 M SD208 plus 10 ng/ml of TGF-␤1, 1 M SD208 plus 20 ng/ml of TGF-␤1, 1 M SD208 plus 50 ng/ml of TGF-␤1, or 1 M SD208 plus 100 ng/ml of TGF-␤1 (G) for 7 days. Ovaries at 18.5 dpc were directly cultured with the same treatments as described above for 7 days. After culture, ovaries were fixed, sectioned, and the shape (A-D) and the number of growing follicles (H) was determined. Data were obtained from at least six ovaries. Arrows indicate growing follicles. with control ovaries. In SD208-treated ovaries that had also been treated with LY294002 (Fig. 8, E and EЈ) or rapamycin (Fig.  8, F and FЈ), the ovaries showed decreased numbers of growing follicles compared with SD208-treated ovaries but still had more enlarged growing follicles as compared with control ovaries. Quantification of follicle numbers showed that treatment with LY294002 or rapamycin in SD208-treated ovaries exhibited lower numbers of activated follicles than in only SD208treated ovaries but higher numbers of growing follicles than in control ovaries (Fig. 8G). It had been reported that activation of S6K1-rpS6 in oocytes was dependent on both PI3K and mTORC1 signaling (17,(35)(36). To check the effectiveness of LY294002 and rapamycin in suppressing PI3K or mTORC1 signaling, we measured the phosphorylation levels of Akt and rpS6 in the ovaries. We found that ovaries treated with LY294002 or SD208 plus LY294002 largely suppressed levels of p-Akt (Ser-473) and p-rpS6 (Ser-240/Ser-244), whereas treatment with rapamycin or SD208 plus rapamycin only suppressed the level of p-rpS6 (Ser-240/Ser-244), but not the level of p-Akt (Ser-473), treatment with SD208 only elevated the level of p-rpS6 (Ser-240/Ser-244), but did not alter the level of p-Akt (Ser-473) (Fig. 8H).

DISCUSSION
Physiologically, to avoid quickly exhausting of primordial follicle pool and maintain the reproductive life span in mammals, the activation of primordial follicles is limited in a few amounts for each wave. The current findings indicate that TGF-␤ signaling participates in the regulation of early primordial follicle growth. Furthermore, Western blot and immunohistochemical analyses indicate that TGF-␤ regulate primordial follicle growth phase via activation of rpS6 in oocyte, but may not via activation of Akt or FOXO3a.
Previous studies have shown that TGF-␤1 immunostaining in 26-day-old mice is strong in primordial oocytes and moderate in growing oocytes, whereas TGF-␤R1 immunostaining is of the same intensity in primordial and growing follicle oocytes in 23-day-old mice (24). This indicates that TGF-␤1 immunostaining becomes weaker during the transition from primordial follicle oocytes to growing oocytes. Meanwhile, our results FIGURE 5. Proliferation and apoptosis in TGF-␤1-and SD208-treated ovaries. Ovaries at 18.5 dpc were cultured alone as a control, with 10 ng/ml of TGF-␤1, or with 1 M SD208 for the staining of PCNA, a BrdU incorporation assay, and the TUNEL assay. Incorporation of BrdU into the granulosa cells was decreased in TGF-␤1-treated ovaries and increased in SD208-treated ovaries compared with control ovaries (A-C, arrows: BrdU positive granulosa cells) after 7 days of culture. Immunostaining of PCNA expression in ovarian sections from TGF-␤1-and SD208-treated ovaries (D-F, arrows, PCNA negative granulosa cells; arrowheads, PCNA positive granulosa cells) after 7 days of culture. TUNEL assays for apoptosis were performed in control, TGF-␤1-, and SD208-treated ovaries (G-I, arrows, TUNEL positive oocytes) after 5 days of culture and the number of apoptotic germ cells was analyzed in the largest cross-section of all treated ovaries (K). Western blot was performed using anti-cleaved caspase-3 and anti-GAPDH antibodies after 5 days of culture (J). Data were obtained from at least six ovaries. Scale bars: 40 m (A-I). Bars with different letters are significantly different (p Ͻ 0.05).
revealed that the protein levels of TGF-␤1 and TGF-␤R1 in ovaries sharply decreased between 4 and 7 dpp, which is the first time that primordial follicle oocytes are recruited into the growing follicle pool. This suggests that TGF-␤ is a potential regulator of primordial follicle growth.
Observations from experiments in which 18.5-dpc ovaries were cultured for 7 days demonstrate that TGF-␤1 reduces the total population of primordial and growing follicles. This result was consistent with previous rat ovary culture experiments that demonstrated that the total number of oocytes was significantly reduced in TGF-␤1-treated ovaries after 10 days of culture (29). Our results demonstrated that the higher expression of cleaved caspase-3 in TGF-␤1-treated ovaries may explain why the total number of oocytes was reduced and the negative effects of TGF-␤1 on the development of primordial and growing follicles may reflect the fact that a sufficient amount of TGF-␤1 is produced endogenously during this process. Thus, the specific TGF-␤1 receptor inhibitor SD208 was used to block the function of endogenous TGF-␤1. The results showed that SD208 could promote rapid oocyte growth and significantly increase the diameter of primordial follicle oocyte. These results give the suggestion that, naturally, the activation of some primordial follicles may be a program process that needs the decreasing of TGF-␤1.
In TGF-␤1-null female mice, follicles are present across the full range of developmental stages and there is no difference in the proportion of various types of follicles (28). An accelerated enlargement of primordial follicle oocytes does not occur in the ovaries, but in our study, the diameter of primordial follicle oocyte in ovaries cultured with SD208 was increased. One possible explanation for this is that the TGF-␤R1 inhibitor SD208 could bind competitively to TGF-␤R1 and effectively block the binding of all TGF-␤ ligands to TGF-␤R1, not just of the binding of TGF-␤1. Additionally, because TGF-␤ ligands share 98 -100% of their identity and function, they are indistinguishable in most bioassays (37). Thus, TGF-␤2 and TGF-␤3 may play compensatory roles during ovarian development in TGF-␤1 knock-out mice. Rapid oocyte growth was not observed in TGF-␤1 knock-out mice but was detected in this experiment because TGF-␤ function was blocked. Alternatively, it is possible that other factors secreted by granulosa cells in preantral and antral follicles in the TGF-␤1 knock-out mouse ovary play a role in limiting the excessive growth of oocytes. The conditional knock-out of TGF-␤R1 in the female repro- ductive tract with anti-Müllerian hormone receptor type II (AMHR2) promoter-driven Cre recombinase produced minimal ovarian defects (38) and primordial follicles and oocytes appear to be AMHR2 mRNA-negative in the rat ovary (39). These observations do not contradict the present results because this set of experiments focused only on primordial and primary follicles. At this stage, TGF-␤R1 is solely expressed in oocytes, particularly oocytes of the primordial follicles.
PTEN/PI3K and TSC/mTORC1 signaling are known to be involved in oocyte growth and development in a coordinated manner. The deletion of Foxo3a or Pten in oocytes results in the excessive activation and depletion of primordial follicles and the ablation of Pten leads to increased FOXO3a phosphorylation and nuclear export (13)(14)(15). PTEN and TSC suppress the phosphorylation of S6K1 at different threonine residues resulting in an inhibition of S6K1/rpS6 activation in oocytes, which helps to preserve the quiescence of primordial follicles (16,17). The accelerated follicular activation in Pten-and Tsc-deleted mice occurs via the enhanced activation of S6K1/rpS6 signaling (16,17). The disruption of these signaling pathways can lead to an accelerated depletion of primordial follicles and premature ovarian failure. The two above mentioned signaling pathways play critical roles in cell proliferation and growth but the manner in which these pathways are normally regulated, especially in oocytes, and whether or not TGF-␤1 acts on them is not known. TGF-␤ activates the PI3K/Akt pathway in various cells (40,41), which results in the phosphorylation and inactivation of three FOXO proteins: FOXO1/FKHR, FOXO3a/FKHRL1, and FOXO4/AFX (42). TGF-␤1 can induce the phosphorylation of Akt and FOXO3a (43). Additionally, the phosphorylation of S6K1 is induced in epithelial cells that undergo epithelial-mesenchymal transition in response to TGF-␤ (44). TGF-␤ also activates the mTORC1 pathway in fibroblasts via a PI3K-Akt-TSC2-dependent mechanism (45). This study demonstrated that the levels of p-Akt and p-FOXO3a were similar in TGF-␤1-and SD208-treated ovaries, indicating that TGF-␤ FIGURE 7. Immunohistochemistry of p-rpS6 (Ser-235/Ser-236) in TGF-␤1-and SD208-treated ovaries. Ovaries at 18.5 dpc were cultured alone (as a control), with 10 ng/ml of TGF-␤1, or 1 M SD208 for 5, 7, or 9 days, respectively, and then collected for immunohistochemistry of p-rpS6 (Ser-235/Ser-236). In 1-dpp (A, arrows) and 4-dpp (B, arrows) mouse ovaries, p-rpS6 (Ser-235/Ser-236) was only expressed in the oocytes of growing follicles but not in the oocytes of primordial follicles. In 18.5-dpc ovaries after 5, 7, or 9 days of culture, respectively, a greater number of growing follicle oocytes exhibited positive staining for p-rpS6 (Ser-235/Ser-236) in SD208-treated ovaries and fewer in TGF-␤1-treated ovaries compared with control ovaries (C-K, arrows). More rapid growth of primordial follicles and increased staining of p-rpS6 (Ser-235/Ser-236) in growing follicle oocytes were observed in SD208-treated ovaries (E, H, and K, arrows), whereas slower growth of primordial follicles and decreased staining were observed in TGF-␤1-treated ovaries (D, G, and J, arrows) compared with control ovaries (C, F, and I, arrows). Scale bars: 40 m (A-K). The experiments were repeated at least three times and representative results are shown.
signaling manipulates early primordial follicle growth not through activation of PI3K-Akt signaling. Although treatment with LY294002 in SD208-treated ovaries can partially prevent the accelerated oocyte growth, it is because the PI3K-specific inhibitor LY294002 itself can result in an inhibition of downstream S6K1-rpS6 activation through negative regulation of PI3K-Akt signaling, which helps to preserve the quiescence of primordial follicles and suppress follicular activation. On the contrary, the phosphorylation levels of S6K1 at Thr-389 and rpS6 at Ser-235/Ser-236 and Ser-240/Ser-244 were elevated in SD208-treated ovaries but reduced in TGF-␤1-treated ovaries.
In other words, TGF-␤ signaling may regulate primordial follicle growth through activation of S6K1-rpS6, a downstream signaling of TSC/mTORC1 known to be involved in oocyte growth. It is known that mTORC1 is sensitive to rapamycin inhibition (46), and only suppresses the phosphorylation of S6K1 and rpS6, but not that of Akt. So we checked whether rapamycin could reverse the stimulated effect of SD208 on oocyte growth. Our results showed that rapamycin partially prevented the accelerated oocyte growth and decreased the numbers of growing follicles with enlarged oocytes. Western blot results also demonstrated that SD208 plus rapamycin only FIGURE 8. Rapamycin partially prevents the accelerated primordial follicle oocyte growth in SD208-treated ovaries. A-F, the accelerated primordial follicle oocyte growth in SD208-treated ovaries can be reversed by treatment with the PI3K-specific inhibitor LY294002 (LY, 25 M) or the mTORC1-specific inhibitor rapamycin (Rap, 50 nM). Ovaries at 3 dpp were cultured for 5 days without treatment (as a control), with 1 M SD208, 25 M LY294002, 50 nM rapamycin, 1 M SD208 plus 25 M LY294002, or with 1 M SD208 plus 50 nM rapamycin. Following culture, ovaries were fixed, sectioned, and the morphological analyses and average numbers of growing follicles were evaluated. Clusters of primordial follicles were seen in all cultured ovaries, whereas more primordial follicles were activated into growing follicles in SD208-treated ovaries (B and BЈ, arrows) as compared with control ovaries (A and AЈ). In SD208-treated ovaries that had also been treated with LY294002 or rapamycin, the ovaries showed decreased numbers of growing follicles compared with SD208-treated ovaries but still had more enlarged growing follicles as compared with control ovaries (E, EЈ, F, and FЈ, arrows). Average numbers of growing follicles were quantified (G). H, 3-dpp mouse ovaries cultured with the same treatments as described above were collected to evaluate the rpS6 and Akt phosphorylation levels by Western blotting. Ovaries treated with LY294002 or SD208 plus LY294002 largely suppressed levels of p-Akt (Ser-473) and p-rpS6 (Ser-240/Ser-244), whereas treatment with rapamycin or SD208 plus rapamycin only suppressed the level of p-rpS6 (Ser-240/Ser-244), but not the level of p-Akt (Ser-473), treatment with SD208 only elevated the level of p-rpS6 (Ser-240/Ser-244), but did not altere the level of p-Akt (Ser-473). Levels of total Akt, rpS6, and ␤-actin were used as internal controls. suppressed the level of p-rpS6 (Ser-240/Ser-244), but not the level of p-Akt (Ser-473). Our immunohistochemical staining also indicated that only those oocytes that began to grow from primordial follicles showed positive staining of p-rpS6.
In summary, in this study we have shown that TGF-␤ signaling plays an important physiological role in maintenance of the dormant pool of primordial follicles, which functions through activation of S6K1-rpS6 in mouse ovaries. We believe that these observations may provide a new resource for studying the developmental and physiological role of TGF-␤ during mammalian reproductive biology and have broad physiological and clinical implications for a better understanding of ovarian physiology and pathology.