ABSTRACT
In mammalian growing follicles, oocytes are arrested at the diplotene stage (which resembles the G2/M boundary in mitosis), while the granulosa cells (GCs) continue to proliferate during follicular development, reflecting a cell cycle asynchrony between oocytes and GCs. Hypoxanthine (Hx), a purine present in the follicular fluid, has been shown to induce oocytes meiotic arrest, although its role in GC proliferation remains ill-defined. Here, we demonstrate that Hx indiscriminately prevents G2-to-M phase transition in porcine GCs. However, oocyte-derived paracrine factors (ODPFs), particularly GDF9 and BMP15, maintain the proliferation of GCs, partly by activating the ERK1/2 signaling and enabling the G2/M transition that is suppressed by Hx. Interestingly, GCs with lower expression of GDF9/BMP15 receptors appear to be more sensitive to Hx-induced G2/M arrest and become easily detached from the follicular wall. Importantly, Hx-mediated inhibition of G2/M progression instigates GC apoptosis, which is ameliorated in the presence of GDF9 and/or BMP15. Therefore, our data indicate that the counterbalance of intrafollicular factors, particularly Hx and oocyte-derived GDF9/BMP15, fine-tunes the development of porcine follicles by regulating the cell cycle progression of GCs.
INTRODUCTION
As the basic functional unit within mammalian ovaries, each follicle comprises an oocyte surrounded by one or more layers of granulosa cells (GCs) (Gougeon, 1996). Some follicles at the antral stage are recruited and selected by pituitary gonadotropins to grow towards ovulation (Gougeon, 1996). During this phase, the GCs divide rapidly to form several layers. However, GCs usually stop proliferating and become apoptotic in most (99.9%) of the growing follicles, which are destined for the atretic degeneration (Baker, 1963). Unlike GCs, development of oocytes remains arrested at the diplotene stage of prophase I until the onset of ovulation (Jaffe and Egbert, 2017). Nevertheless, oocytes spontaneously resume meiosis when isolated from their follicles, indicating that the follicular niche hold the oocytes in a dormant state (Pincus and Enzmann, 1935).
Several decades ago, the purine catabolite hypoxanthine (Hx) was identified as the principal inhibitory component of follicular fluid that prevented meiotic resumption in mouse oocytes (Downs et al., 1985). Similar observations have since been made in numerous mammalian species, including rat, cow, monkey, hamster, rabbit and pig (Downs, 1997a). Hx is reported to maintain meiotic arrest by inhibiting phosphodiesterase and thus blocking the degradation of intracellular cAMP (Eppig and Downs, 1984; Mehlmann, 2005). Elevated levels of oocyte cAMP trigger PKA activation, which in turn deactivates CDC25B and activates Wee1/Myt1 kinases, resulting in phosphorylation of CDK1 and inactivation of maturation/metaphase-promoting factor (MPF, a complex of CDK1 and cyclin B), thereby impeding the G2/M transition beyond diplotene stage (Mehlmann, 2005). Heretofore, however, whether Hx exerts any influence on GCs proliferation is still unknown.
In vivo, the surge of luteinizing hormone (LH) from the pituitary gland allows both meiotic resumption of oocytes and ovulation (Mehlmann, 2005). The action of LH is mediated through binding to LH receptor (LHR), the expression of which in GCs is induced by follicle-stimulating hormone (FSH) (Erickson et al., 1979). It has been well established that FSH stimulates the growth and proliferation of GCs (Richards, 1979). The mitogenic response to FSH is partly achieved by its stimulation of estradiol production in GCs, and estradiol concomitantly induces expression of the GCs-specific FSH receptor (FSHR) (Simoni et al., 1997). In addition to gonadotropins, GCs proliferation is also influenced by multiple types of growth factors, including insulin-like growth factor 1 (IGF1), epidermal growth factor (EGF), transforming growth factor β (TGFβ) and fibroblast growth factor (FGF) (Vanderhyden et al., 1992). Either individually or together with other growth factors and/or gonadotropins, these intra-ovarian factors modulate follicular development via upregulation of the mitotic activity of GCs (Vanderhyden et al., 1992).
In recent years, it has become evident that oocytes modulate follicular development via oocyte-derived paracrine factors (ODPFs). These members of the TGFβ superfamily includes growth differentiation factor 9 (GDF9) and bone morphogenetic protein 15 (BMP15 or GDF9B) (Knight and Glister, 2006). Similar to other TGFβ ligands, GDF9 and BMP15 initiate their signaling pathway by binding type 2 (BMPR2) and type 1 (TGFBR1 and BMPR1B, respectively) receptors on the GCs membrane (Knight and Glister, 2006). There is already, in the follicle, a clear indication for some form of interaction between GDF9 and ERK1/2 (Elvin et al., 1999; Sasseville et al., 2010; Su et al., 2003). In cumulus cells, high levels of ERK1/2 activity have been detected following GDF9 treatment (Su et al., 2003). GDF9 also simulates ERK1/2-dependent cumulus expansion in the absence of FSH (Elvin et al., 1999; Su et al., 2003). Further investigations show a crosstalk between GDF9 and EGFR-ERK1/2 in mediating GCs survival and proliferation (Sasseville et al., 2010). Nevertheless, current evidence does not indicate a direct action of ERK1/2 on ODPF-induced cell cycle progression of GCs. Here, we have investigated whether and how Hx modulates GC proliferation, and have explored the mechanism required for maintaining GC mitotic activity during follicular development.
RESULTS
Hx in porcine follicular fluid arrests the cell cycle of ovarian GCs at G2/M
Hx is a major purine catabolite that exists in the porcine follicular fluid (Downs et al., 1985), but few researchers have accurately determined its physiological range of concentration. In this study, high-performance liquid chromatography (HPLC) was employed to analyze the Hx content in porcine follicular fluid (pFF) samples collected from 190 porcine ovarian follicles. As shown in Fig. 1A, Hx (peak 9) was identified by comparing the tR and UV (λmax 254 nm) of the reference. The HPLC-UV data were in agreement with earlier reports suggesting the presence of Hx in pFF (Downs et al., 1985). On the basis of frequency distribution, the levels of Hx observed in more than 94.12% of total follicles were between 0.25 mM and 1 mM (Fig. 1B), which might represent a physiological concentration range of Hx within pFF.
Follicular Hx is required for the induction of G2/M arrest in GCs retrieved from the pFF. (A) Upper panel: representative HPLC chromatograms of pFF collected from 190 porcine ovarian follicles. Hx, hypoxanthine. Lower panel: representative HPLC chromatograms of Hx standard solution (2 mM). (B) A frequency distribution of follicles over a range of Hx concentrations. (C-F) Antral follicles were dissected from porcine ovaries, and torn apart to release the pFF. GCs separated from the pFF were processed for flow cytometric detection of cell cycle. The pFF supernatant were collected for HPLC assay of Hx levels. The relationship between Hx concentrations and follicular size or GCs cell cycle was analyzed by linear regression in 68 follicles. (G-O) The relationship between Hx concentrations and GCs cell cycle was assessed using linear regression in porcine antral follicles with different sizes, including small follicles (1-3 mm; G-I), medium follicles (3-5 mm; J-L) and large follicles (5-8 mm; M-O). (P-R) Individual follicles dissected from porcine ovaries were torn apart to discharge the pFF. GCs isolated from the pFF were used for cell cycle determination with flow cytometry, while the pFF supernatant were collected for ELISA assay of estrogen levels. The relationship between estrogen concentrations and GCs cell cycle was analyzed by linear regression in 68 follicles.
Follicular Hx is required for the induction of G2/M arrest in GCs retrieved from the pFF. (A) Upper panel: representative HPLC chromatograms of pFF collected from 190 porcine ovarian follicles. Hx, hypoxanthine. Lower panel: representative HPLC chromatograms of Hx standard solution (2 mM). (B) A frequency distribution of follicles over a range of Hx concentrations. (C-F) Antral follicles were dissected from porcine ovaries, and torn apart to release the pFF. GCs separated from the pFF were processed for flow cytometric detection of cell cycle. The pFF supernatant were collected for HPLC assay of Hx levels. The relationship between Hx concentrations and follicular size or GCs cell cycle was analyzed by linear regression in 68 follicles. (G-O) The relationship between Hx concentrations and GCs cell cycle was assessed using linear regression in porcine antral follicles with different sizes, including small follicles (1-3 mm; G-I), medium follicles (3-5 mm; J-L) and large follicles (5-8 mm; M-O). (P-R) Individual follicles dissected from porcine ovaries were torn apart to discharge the pFF. GCs isolated from the pFF were used for cell cycle determination with flow cytometry, while the pFF supernatant were collected for ELISA assay of estrogen levels. The relationship between estrogen concentrations and GCs cell cycle was analyzed by linear regression in 68 follicles.
We next investigated whether the abundance of Hx correlates with the follicular size. As shown in Fig. 1C, although there were variations in Hx content between individual follicles of the same diameter, the larger follicles tended to have higher levels of Hx. Actually, it has been reported that the oocyte maturation inhibitors are probably produced by the granulosa and/or theca cells (Tripathi et al., 2010). Predictably, the growing follicles with elevated number of GCs or theca cells would inevitably increase Hx production in pFF. Incidentally, our results might reflect a potential role of follicular somatic cells in Hx synthesis.
Using flow cytometry, we then evaluated the effects of Hx on the cell cycle progression of GCs released from pFF. As shown in Fig. 1D-F, Hx in each follicle did not significantly alter the proportion of GCs in G0/G1, but drove cells out of S and into the G2/M phase, indicating a possible role of Hx in triggering G2/M arrest of GCs. As the antral follicles could be divided into small follicles (1-3 mm), medium follicles (3-5 mm) and large follicles (5-8 mm), we thus further tested the correlation between Hx concentration and the GC cell cycle in follicles of different sizes. The results also showed a dose-dependent induction of G2/M arrest in GCs in response to the increasing level of Hx (Fig. 1G-O).
Considering that GCs in growing follicles could stimulate their own proliferation via secreting estrogen (Tonetta and diZerega, 1989), we asked whether the intrafollicular levels of this steroid hormone affects the GC cell cycle. As shown in Fig. 1P-R, no linear correlation was detected between estrogen level and the GC cell cycle (Fig. 1P-R), indicating that estrogen might be irrelevant to Hx-induced cell cycle arrest of GCs. Collectively, our in vivo findings raised the possibility that the induction of G2/M arrest by Hx might inhibit the proliferation activity of GCs during follicular development.
Hx-mediated G2/M arrest is associated with retardation of proliferation and apoptotic death in cultured GCs
To further clarify the effects of Hx on GC proliferation without the interference from other follicular factors, we thus applied Hx treatment in primary cultured GCs retrieved from porcine ovarian follicles (3-7 mm in diameter). Using the CCK-8 assay, we found that GCs proliferation activity was significantly decreased after 24 h of incubation in Hx-supplemented medium (Fig. S2A). As Hx has been reported to cause mammalian cell dysfunction through oxidative stress (Kim et al., 2017), we assessed whether the suppression of GCs proliferation was dependent on Hx-induced ROS production. As expected, the conversion of DCFH to DCF, which emits green fluorescence in ROS-positive cells, was markedly increased following Hx treatment (Fig. S2B). However, no definitive evidence for the involvement of ROS was observed in the proliferation retard induced by Hx exposure (Fig. S2C,D).
In recent years, a possible correlation has emerged between proliferation and autophagy (Cianfanelli and Cecconi, 2015). To investigate whether Hx induced autophagy in cultured GCs, protein lysates were collected for immunoblot analysis of autophagy marker proteins. As shown in Fig. S2E, the expression of ATG3 and MAP1LC3B-II was promoted in GCs pretreated with Hx at a concentration within the physiological range. In addition, the MAP1LC3B-II blots and GFP-MAP1LC3B puncta accumulated in cells with Hx and chloroquine (CQ) administration (Fig. S3A-C), indicating that the autophagic flux was enhanced upon Hx stimulation. However, the autophagy inhibitor 3-MA failed to inhibit Hx-induced retardation of proliferation in GCs (Fig. S3D,E). Therefore, Hx-mediated suppression of GCs proliferation might not be attributed to autophagy induction.
Given the pro-apoptotic role of Hx described previously (Kim et al., 2017), we next determined whether apoptosis might be responsible for the retardation of proliferation in GCs in response to Hx incubation. As shown in Figs S2F,G and S3F, the proportion of apoptotic GCs was remarkably increased following Hx treatment, along with an elevated expression of cleaved caspase 3. Notably, inhibition of apoptosis using Z-VAD-FMK (a pancaspase inhibitor) reversed the proliferation activity in cells exposed to Hx (Fig. S2H), suggesting that Hx-induced suppression of GC division might be mediated partly through the apoptotic pathway.
Although Hx has been shown to inhibit meiotic resumption of oocytes (Downs et al., 1985), there have been no reports about its influence on the cell cycle of cultured GCs. Here, the inhibitory effects of Hx on mitotic progression were investigated by monitoring PI-labeled GCs using flow cytometry. As shown in Fig. 2A-D, GCs subjected to Hx incubation became arrested in G2/M following the S-phase transition, while the proportion of cells in G0/G1 remained unchanged. These observations were consistent with the aforementioned results from in vivo experiments (Fig. 1). Because flow cytometry alone could not distinguish M from G2 phase, we also counted the mitotic cells at different stages (including prophase, metaphase, anaphase and telophase) by performing DAPI staining (Fig. 2G). As shown in Fig. 2E,F,H, Hx blocked the transition from G2 to M phase (Fig. 2E,F), thereby inhibiting GC division (Fig. 2H).
The suppression of GC proliferation is associated with Hx-induced G2/M arrest. (A) GCs with or without 24 h of Hx (1 mM) treatment were subjected to PI staining. The cell cycle distribution was then analyzed using flow cytometry. (B-D) The percentage of GCs in G0/G1 (B), S (C) or G2/M phase (D) was quantified by DNA content using the ModFit LT 3.2 program. (E-G) After culturing with 1 mM Hx for 24 h, the GC nuclei were labeled with DAPI, and representative images of cells in prophase, metaphase, anaphase and telophase were captured using a laser-scanning confocal microscope (G). Scale bar: 5 μm. Experiments were performed in triplicate, and the number of M-phase GCs was counted in five randomly selected fields in each well to quantify the percentage of mitotic cells (E) and the ratio of M/(G2+M) (F). (H) The total cell number was counted in GCs treated with or without Hx (1 mM) at 0 h and 24 h. The data are mean±s.e.m.; n=3. *P<0.05, ***P<0.001; NS, not significant (P>0.05).
The suppression of GC proliferation is associated with Hx-induced G2/M arrest. (A) GCs with or without 24 h of Hx (1 mM) treatment were subjected to PI staining. The cell cycle distribution was then analyzed using flow cytometry. (B-D) The percentage of GCs in G0/G1 (B), S (C) or G2/M phase (D) was quantified by DNA content using the ModFit LT 3.2 program. (E-G) After culturing with 1 mM Hx for 24 h, the GC nuclei were labeled with DAPI, and representative images of cells in prophase, metaphase, anaphase and telophase were captured using a laser-scanning confocal microscope (G). Scale bar: 5 μm. Experiments were performed in triplicate, and the number of M-phase GCs was counted in five randomly selected fields in each well to quantify the percentage of mitotic cells (E) and the ratio of M/(G2+M) (F). (H) The total cell number was counted in GCs treated with or without Hx (1 mM) at 0 h and 24 h. The data are mean±s.e.m.; n=3. *P<0.05, ***P<0.001; NS, not significant (P>0.05).
To further verify whether the G2/M arrest is dependent on Hx-induced apoptosis or autophagy, cell cycle was then examined in GCs treated with Z-VAD-FMK and 3-MA, respectively. As shown in Fig. S4A-D, the inhibition of apoptosis or autophagy did not seem to affect the cell cycle progression. In contrast, the increased percentage of cells in G2/M was accompanied by decreased proliferation activity and elevated apoptotic rates after prolonged treatment with Hx (Figs S4E-H, S5A,B). We also performed the above experiments in GCs cultured with FBS-free medium, but failed to detect cells in G2/M phase because a large population of GCs underwent apoptosis after 24 h of culture in serum-free conditions (Fig. S6). To better assess the cellular autophagic/apoptotic response to G2/M arrest during Hx exposure, we treated GCs with Ro-3306, a specific inhibitor that prevents CDK1 activation (Fig. S4I). According to our treatment procedure, the CDK1 inhibitor succeeded in arresting GCs at G2/M (Figs S4J,K, S5C), followed by enhanced activation of both autophagy (Fig. S4L-N) and apoptosis (Figs S4O and P, S5D), as well as a decline in mitotic capability (Fig. S4Q). Consistent with this, similar results were obtained when CDK1 was knocked down by small interfering RNAs (Fig. S4R-V). Thus, it was concluded that the suppression of G2/M transition and the resultant apoptosis might be responsible for Hx-induced retardation of proliferation in GCs.
Inhibition of MPF activity through the Hx-PKA-Wee1/Myt1 axis is required for the induction of G2/M arrest
Progression of cell cycle through the G2/M phase is precisely controlled by MPF (a cyclin B-CDK1 complex) and its interaction with multiple checkpoint regulators, including Myt1, Wee1 and CDC25B (Boutros et al., 2007), although their relative contributions to Hx-modulated G2/M transition in GCs are not well understood. Using qRT-PCR analysis, we found that both CDK1 and cyclin B1 were downregulated at the transcription level when GCs received Hx treatment (Fig. S7A,B). Correspondingly, impaired activity of the MPF complex was detected after Hx incubation (Fig. S7C). These findings prompted us to investigate the coordination between MPF and upstream regulators. As shown in Fig. S7D, Hx activates the Wee1/Myt1 kinases, leading to inactivation of CDK1, as suggested by phosphorylation at Thr14 and Tyr15. Concomitantly, inhibition of Cdc25B (the CDK1 activator) through phosphorylation at Ser323 occurred during Hx stimulation.
Several signaling cascades have been implicated in the regulation of mammalian oocyte meiotic resumption (Sun et al., 2009). To test whether these pathways might be involved in Hx-induced G2/M arrest in GCs, we performed western blot assay. As shown in Fig. S7E, cells subjected to Hx incubation exhibited significantly elevated levels of PRKACA, the catalytic subunit α of PKA. Conversely, Hx remarkably inhibited ERK1/2 activity, which was reflected by deceased phosphorylation at Thr202/Tyr204 (Fig. S7E). Blocking PKA with its inhibitor H89 reactivated ERK1/2, despite Hx administration (Fig. S7F), indicating that the suppression of ERK1/2 by Hx might be mediated through the PKA pathway. As a pro-apoptotic BH3-only protein, PUMA expression was also upregulated following Hx exposure (Fig. S7E). However, NF-κB p65, p53 and JNK presented no response to the Hx administration (Fig. S7E).
To further clarify whether PKA affected the G2/M transition upon Hx stimulation, we then examined the cell cycle of GCs in the presence of H89. As shown in Fig. S7G-J, the inhibition of PKA significantly reduced the proportion of Hx-treated GCs in G2/M. In addition, H89 increased the percentage of M phase cells in the G2/M population during Hx exposure, indicating a restoration of G2/M transition after silencing the PKA signaling (Fig. S7K). Using immunoblotting to detect the checkpoint factors required for G2/M progression, we further confirmed the role of PKA in Hx-induced G2/M arrest within GCs (Fig. S7L).
Hx fails to induce cell cycle arrest in those GCs that are strictly attached to the follicular wall
During the development of ovarian follicles, GCs proliferate rapidly to form multilayered structures. Given the fact that larger follicles usually contain higher levels of Hx, as mentioned in Fig. 1, we asked whether the proliferation of mural GCs might not be inhibited by Hx. To validate this speculation, the correlation between cell cycle and Hx concentration was examined in GCs collected by scraping the follicular wall. As shown in Fig. S8A-C, the G2/M, G0/G1 and S-phase population were not significantly changed by Hx. It therefore appears that certain cell cycle-promoting factors (CCPFs) might negate the presence of Hx in pFF.
Oocyte-derived factors GDF9 and BMP15 restore G2/M transition in Hx-treated GCs
To explore how GCs ‘escape’ from Hx-stimulated G2/M arrest, we examined the possible involvement of gonadotropins and growth factors, both of which have been reported to facilitate meiotic resumption in oocytes (Byskov et al., 1995). As shown in Fig. S10A-F, FSH did not significantly affect the process of G2/M transition or proliferation when GCs were exposed to Hx. In accordance with this, FSH failed to restore the activity of G2/M checkpoint factors during Hx incubation (Fig. S10G,H). Moreover, FSH did not significantly alter the G2/M distribution of GCs when PKA was inhibited (Fig. S11A,B). However, in GCs incubated with both Hx and a PKA inhibitor, more G2/M transition occurred following FSH treatment (Fig. S11C). Using flow cytometric analysis of PI-labeled GCs, we investigated the potent functions of growth factors in regulating G2/M transition. As shown in Fig. S12A-F, TGFβ markedly reduced the proportion of G2/M population in GCs suffering from Hx exposure, whereas EGF and IGF1 had no effect. Consistent results were obtained by determining the percentage of mitotic cells, the ratio of M/(G2+M), CDK1 phosphorylation status, proliferation activity and the apoptotic rates (Fig. S12G-P), further confirming the role of TGFβ in preventing GCs from Hx-induced G2/M arrest.
Oocyte-derived paracrine factors (ODPFs), including GDF9 and BMP15, are relevant members of the TGFβ superfamily (Knight and Glister, 2006). To test whether the physiological dose of TGFβ, GDF9 or BMP15 could overcome Hx-induced G2/M arrest in GCs, we determined the levels of these factors in pFF samples collected from 70 porcine ovarian follicles. As shown in Fig. 3A, the follicular concentration of TGFβ was around 1 ng/ml, while the average level of GDF9 or BMP15 was ∼10 ng/ml. We then treated cultured GCs with 1 ng/ml TGFβ, but this failed to restore G2/M progression or cell proliferation activity upon Hx exposure (Fig. 3B,C). In contrast, the decline in GCs proliferation activity caused by Hx was partly reduced after GDF9 treatment (Fig. 3D). Correspondingly, GDF9 and/or BMP15 significantly restored the proportion of mitotic cells in GCs with Hx incubation (Fig. 3E). Immunoblotting assays showed that Hx-induced expression of p-CDK1 (Tyr15), Wee1, p-Wee1 (ser642) and p-CDC25B (Ser323) were markedly decreased in the presence of GDF9 and/or BMP15 (Fig. 3H). Moreover, GDF9 and/or BMP15 blocked the induction of apoptosis in GCs that received Hx treatment (Fig. 3F,G). Based on these data, we propose that GDF9/BMP15-mediated G2/M transition might be essential for resuming GC proliferation upon Hx stimulation.
G2/M-arrested GCs resume proliferation activity in the presence of GDF9 and BMP15. (A) Seventy antral follicles (3-7 mm in diameter) were individually dissected from porcine ovaries, and torn apart to discharge the pFF for ELISA assay of TGFβ1, GDF9 or BMP15 levels. (B) GCs were cultured with Hx (1 mM) and/or TGFβ1 (1 ng/ml) for 24 h. The cell cycle distribution in G2/M phase was then analyzed using flow cytometry. Data are mean±s.e.m.; n=3. (C) The CCK-8 assay of cell proliferation activity in GCs with the same treatment as described above. Data are mean±s.e.m.; n=4. (D) Primary GCs were grown in medium containing Hx (1 mM) and/or GDF9 (10 ng/ml) for 24 h. Cell proliferation activity was then determined by CCK-8 assay. Data are mean±s.e.m.; n=4. (E) GCs incubated with Hx (1 mM) were cultured for 24 h in the presence or absence of GDF9 (10 ng/ml) and/or BMP15 (10 ng/ml), and stained with DAPI to visualize the M-phase nuclei. The bar charts showed the ratio of the M-phase population. (F,G) The detection of apoptotic signals in GCs with the same treatment as described above. The percent of apoptotic cells were quantified by flow cytometry. Data are mean±s.e.m.; n=3. (H) Western blotting showed expression levels of p-CDK1 (Thr14), p-CDK1 (Tyr15), p-Wee1 (Ser642), Wee1, p-Myt1 (Ser83), Myt1, p-CDC25B (Ser323) and CDC25B in GCs with the treatments indicated. TUBA1A served as the loading control. Data are mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001; NS, not significant (P>0.05).
G2/M-arrested GCs resume proliferation activity in the presence of GDF9 and BMP15. (A) Seventy antral follicles (3-7 mm in diameter) were individually dissected from porcine ovaries, and torn apart to discharge the pFF for ELISA assay of TGFβ1, GDF9 or BMP15 levels. (B) GCs were cultured with Hx (1 mM) and/or TGFβ1 (1 ng/ml) for 24 h. The cell cycle distribution in G2/M phase was then analyzed using flow cytometry. Data are mean±s.e.m.; n=3. (C) The CCK-8 assay of cell proliferation activity in GCs with the same treatment as described above. Data are mean±s.e.m.; n=4. (D) Primary GCs were grown in medium containing Hx (1 mM) and/or GDF9 (10 ng/ml) for 24 h. Cell proliferation activity was then determined by CCK-8 assay. Data are mean±s.e.m.; n=4. (E) GCs incubated with Hx (1 mM) were cultured for 24 h in the presence or absence of GDF9 (10 ng/ml) and/or BMP15 (10 ng/ml), and stained with DAPI to visualize the M-phase nuclei. The bar charts showed the ratio of the M-phase population. (F,G) The detection of apoptotic signals in GCs with the same treatment as described above. The percent of apoptotic cells were quantified by flow cytometry. Data are mean±s.e.m.; n=3. (H) Western blotting showed expression levels of p-CDK1 (Thr14), p-CDK1 (Tyr15), p-Wee1 (Ser642), Wee1, p-Myt1 (Ser83), Myt1, p-CDC25B (Ser323) and CDC25B in GCs with the treatments indicated. TUBA1A served as the loading control. Data are mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001; NS, not significant (P>0.05).
We next examined whether the levels of GDF9, BMP15, TGFβ, E2 or FSH in pFF might influence the GC cell cycle in vivo. As shown in Fig. 4A-F and Fig. S9A-E, there is no evidence for a significant correlation between the cell cycle distribution of mural GCs and GDF9, BMP15, TGF-β, E2 or FSH content in pFF. In accordance with this, Hx alone also failed to alter the GC cell cycle in ovarian follicles (Fig. S8A-C). We thus asked whether Hx might coordinate with GDF9, BMP15, TGFβ, E2 or FSH in the regulation of mitotic progression in GCs. To test this possibility, we determined the correlation between GCs G2/M distribution and pFF concentration of GDF9, BMP15, TGFβ, E2 or FSH in follicles with similar Hx levels and follicular sizes. The results showed a dose-dependent suppression of G2/M arrest in GCs by the increasing levels of GDF9 and BMP15, rather than TGFβ, E2 or FSH (Fig. 4G,H and Fig. S9F-J). On the other hand, when we fixed the pFF content of GDF9 or BMP15 and follicular sizes, a significant positive correlation was observed between Hx levels and G2/M distribution in follicular GCs (Fig. 4I,J and Fig. S9K-N). Therefore, the in vivo findings indicated a possible counteraction among Hx, GDF9 and BMP15 in controlling G2/M transition of GCs.
Hx coordinates with GDF9 and BMP15 to control G2/M transition of follicular GCs in vivo. (A-F) Seventy antral follicles were individually dissected from porcine ovaries, and torn apart to discharge the pFF for ELISA analysis of GDF9 and BMP15 levels. Meanwhile, the mGCs were isolated by scraping the follicular wall. The relationship between levels of GDF9 or BMP15 and the mGC cell cycle was analyzed by linear regression. (G,H) The relationship between concentrations of GDF9 or BMP15 and the GC G2/M ratio was analyzed by linear regression in seven follicles with similar Hx levels (0.6-0.7 mM) and follicular sizes (3-5 mm). (I,J) The relationship between concentrations of Hx and GCs G2/M ratio was analyzed by linear regression in seven follicles with similar levels of GDF9 (8.5-9.5 ng/ml) (I) or BMP15 (9-10 ng/ml) (J) and similar follicular sizes (3-5 mm).
Hx coordinates with GDF9 and BMP15 to control G2/M transition of follicular GCs in vivo. (A-F) Seventy antral follicles were individually dissected from porcine ovaries, and torn apart to discharge the pFF for ELISA analysis of GDF9 and BMP15 levels. Meanwhile, the mGCs were isolated by scraping the follicular wall. The relationship between levels of GDF9 or BMP15 and the mGC cell cycle was analyzed by linear regression. (G,H) The relationship between concentrations of GDF9 or BMP15 and the GC G2/M ratio was analyzed by linear regression in seven follicles with similar Hx levels (0.6-0.7 mM) and follicular sizes (3-5 mm). (I,J) The relationship between concentrations of Hx and GCs G2/M ratio was analyzed by linear regression in seven follicles with similar levels of GDF9 (8.5-9.5 ng/ml) (I) or BMP15 (9-10 ng/ml) (J) and similar follicular sizes (3-5 mm).
GDF9/BMP15 receptor-mediated activation of ERK1/2 restores G2/M progression during Hx exposure
Growing evidence indicates some form of crucial interaction between ODPFs and ERK1/2 (Elvin et al., 1999; Sasseville et al., 2010; Su et al., 2003). This prompted us to test whether ERK1/2 correlates with GDF9/BMP15-induced mitotic progression of GCs during Hx exposure. As shown in Fig. S13A, GDF9/BMP15 remarkably inhibited Hx-induced dephosphorylation of ERK1/2 without altering total ERK1/2 protein levels. Using SCH772984, a specific ERK1/2 inhibitor, we then blocked ERK1/2 activity in cultured GCs (Fig. S13B). As shown in Fig. S13C,D, SCH772984 completely abolished the cell cycle progression from G2/M into mitosis. Notably, GDF9/BMP15 did not restore the proportion of cells in M-phase when cells were pretreated with the ERK1/2 inhibitor, indicating that the activation of ERK1/2 by GDF9/BMP15 might contribute to the G2/M transition in GCs. This assumption was further confirmed by monitoring the expression of several checkpoint proteins required for G2/M progression. As shown in Fig. S13G, GDF9/BMP15 failed to suppress Hx-triggered expression of p-CDK1 (Tyr15), Wee1, p-Wee1 (ser642) and p-CDC25B (Ser323) in the presence of an ERK1/2 inhibitor. Additionally, the flow cytometric data from annexin V-FITC staining indicated that GDF9/BMP15 might protect against Hx-induced apoptotic GCs death through the ERK1/2 pathway (Fig. S13E,F). By performing the above experiments in GCs treated with SCH772984 alone, we further confirmed that the ERK1/2 inhibitor itself did not affect G2-to-M progression of GCs, although an increased apoptosis rate was detected (Fig. S13H,I). GDF9 and BMP15 are believed to signal through type II (BMPR2) and type I (TGFBR1 and BMPR1B, respectively) receptors (Knight and Glister, 2006). To further elucidate the mechanism of cell cycle regulation by GDF9/BMP15, we next determined whether blocking their receptors might impede the G2/M transition. As shown in Fig. S14A-D, Repsox (a TGFBR1 inhibitor) and K02288 (a BMPR1B inhibitor) significantly reduced the percentage of mitotic cells and the ratio of M/(G2+M) in GCs receiving GDF9/BMP15 treatment. Accordingly, GDF9/BMP15 revealed no inhibitory effect on Hx-induced inactivation of MPF when GCs were pretreated with Repsox or K02288 (Fig. S14E,F). These data therefore suggest that the suppression of TGFBR1/BMPR1B restores the G2/M arrest elicited by Hx. We then explored whether TGFBR1/BMPR1B affected ERK1/2 activity. It was found that antagonizing TGFBR1/BMPR1B with Repsox/K02288 abrogated GDF9/BMP15-induced ERK1/2 phosphorylation during Hx exposure (Fig. S14G,H). Therefore, the ERK1/2 pathway might act downstream of receptor-mediated GDF9/BMP15 signaling in GCs. Moreover, by performing annexin V-FITC staining under the same conditions, we further confirmed that TGFBR1/BMPR1B is essential for GDF9/BMP15-induced GCs protection against Hx-derived apoptosis (Fig. S14I,J). Taken together, our data suggest that the induction of MPF by ERK1/2 activation through the interaction between GDF9/BMP15 and their receptors might prevent G2/M arrest and thus repress apoptosis in Hx-treated GCs.
The G2/M transition blocked by Hx is resumed in GCs co-cultured with oocytes
Because both GDF9 and BMP15 are oocyte-secreted factors, we wonder whether the oocyte itself could induce GCs mitotic progression upon Hx stimulation. As shown in Fig. 5A,B, the proportion of GCs in G2/M phase after Hx exposure was markedly reduced when oocytes were added to the culture. Accordingly, the GCs co-cultured with oocytes exhibited relatively higher percentage of mitotic population, as well as an elevation in the ratio of M/(G2+M) during Hx incubation (Fig. 5C,D). Consistent results were obtained in cumulus cells, a specialized type of GC. As expected, Hx inhibited the spontaneous GVBD of oocytes in cumulus-oocyte complexes (COCs) (Fig. 5E,F). However, the cumulus counterpart showed no definitive evidence of cell cycle arrest in spite of Hx treatment (Fig. 5G-J), indicating a possible role of oocytes in preserving GCs mitotic activity, which is suppressed by Hx exposure. In addition, the inhibitory effects of Hx on cumulus expansion were removed in COCs treated with GDF9/BMP15, while GDF9/BMP15 failed to restore GVBD in denuded oocyte during HX incubation (Fig. S15). We also examined whether oocytes might influence the production of GDF9 and BMP15 in the culture medium. As shown in Fig. 5K,L, the levels of GDF9 and BMP15 were significantly increased in the presence of oocytes. Collectively, these data suggest the possibility that oocytes, through secreting GDF9/BMP15, partly restore the cell cycle progression from G2 to M phase in Hx-treated GCs.
Oocytes restore the G2/M transition in Hx-treated GCs. (A-D) Primary GCs were either cultured alone and treated with 1 mM Hx, or co-cultured with denuded oocytes at a density of 100 oocytes/ml and treated with 1 mM Hx. (A) 24 h later, GCs were stained with PI, and the cell cycle distribution was analyzed using flow cytometry. (B) The percentage of GCs in G2/M phase was quantified using the ModFit LT 3.2 program. (C,D) The GCs subjected to DAPI staining were processed for calculating the percentage of mitotic cells (C) and the ratio of M/(G2+M) (D). (E) COCs were cultured with or without Hx (1 mM) for 12 h or 24 h. The morphological changes of the COCs were observed at a magnification of 50× under a surgical dissecting microscope. Scale bar: 80 μm. (F) After 24 h of Hx (1 mM) incubation, the denuded oocytes isolated from the COCs were stained with Hoechst 33342 to assess the GVBD rates. (G-J) COCs grown with or without Hx (1 mM) for 24 h were digested with hyaluronidase. The cumulus GCs were collected for the flow cytometric determination of cell cycle distribution (G). The percentage of GCs in G0/G1 (H), S (I) or G2/M phase (J) was quantified by DNA content using the ModFit LT 3.2 program. (K,L) Primary GCs were either cultured alone and treated with 1 mM Hx, or co-cultured with denuded oocytes at a density of 100 oocytes/ml and treated with 1 mM Hx. 24 h later, the culture medium was collected for ELISA analysis of GDF9/BMP15 levels. Data are mean±s.e.m.; n=3. **P<0.01, ***P<0.001; NS, not significant (P>0.05).
Oocytes restore the G2/M transition in Hx-treated GCs. (A-D) Primary GCs were either cultured alone and treated with 1 mM Hx, or co-cultured with denuded oocytes at a density of 100 oocytes/ml and treated with 1 mM Hx. (A) 24 h later, GCs were stained with PI, and the cell cycle distribution was analyzed using flow cytometry. (B) The percentage of GCs in G2/M phase was quantified using the ModFit LT 3.2 program. (C,D) The GCs subjected to DAPI staining were processed for calculating the percentage of mitotic cells (C) and the ratio of M/(G2+M) (D). (E) COCs were cultured with or without Hx (1 mM) for 12 h or 24 h. The morphological changes of the COCs were observed at a magnification of 50× under a surgical dissecting microscope. Scale bar: 80 μm. (F) After 24 h of Hx (1 mM) incubation, the denuded oocytes isolated from the COCs were stained with Hoechst 33342 to assess the GVBD rates. (G-J) COCs grown with or without Hx (1 mM) for 24 h were digested with hyaluronidase. The cumulus GCs were collected for the flow cytometric determination of cell cycle distribution (G). The percentage of GCs in G0/G1 (H), S (I) or G2/M phase (J) was quantified by DNA content using the ModFit LT 3.2 program. (K,L) Primary GCs were either cultured alone and treated with 1 mM Hx, or co-cultured with denuded oocytes at a density of 100 oocytes/ml and treated with 1 mM Hx. 24 h later, the culture medium was collected for ELISA analysis of GDF9/BMP15 levels. Data are mean±s.e.m.; n=3. **P<0.01, ***P<0.001; NS, not significant (P>0.05).
The GCs with lower expression of GDF9/BMP15 receptors are more sensitive to Hx-induced G2/M arrest and apoptosis, and might become detached from the follicular wall
The aforementioned results showed that, in GCs collected from pFF, the fraction of G2/M arrested cells was positively correlated with Hx content within the follicles (Fig. 1). In contrast, Hx did not affect cell cycle progression of the GCs distributed in the follicular wall (Fig. S8). Considering the role of GDF9/BMP15 in promoting mitotic progression of GCs upon Hx stimulation, we suspected that different sensitivity to GDF9/BMP15 might exist between detached GCs (dGCs) and mural GCs (mGCs). To test this assumption, the expression of GDF9/BMP15 receptors was determined in dGCs and mGCs collected from 120 porcine ovarian follicles. Data from all paired dGCs and mGCs from each follicle sample was shown in Fig. 6. In general, the expressions of both TGFBR1 and BMPR1B in mGCs were higher than in dGCs (Fig. 6A,B). In addition, in situ hybridization with specific RNA probes showed that mRNAs of TGFBR1 and BMPR1B were expressed predominantly by mGCs and cumulus cells, while lower levels of expression were detected in some dGCs (Fig. 6C). Similar results were obtained by qRT-PCR, which also showed higher expression of GDF9/BMP15 receptors in cumulus and mural GCs than that in dGCs (Fig. 6D). Correspondingly, the ratio of G2/M distribution and apoptosis clearly paralleled TGFBR1/BMPR1B expression in dGCs and mGCs (Fig. 6E-H). Using western blotting to detect the checkpoint factors required for G2/M progression, we further confirmed a probable causal relationship between levels of GDF9/BMP15 receptors and the occurrence of G2/M arrest in porcine ovarian GCs (Fig. 6I).
The decreased expression of GDF9/BMP15 receptors in GCs is responsible for the induction of G2/M arrest, apoptosis and cell shedding. (A,B) 120 antral follicles (3-7 mm in diameter) were individually dissected from porcine ovaries and torn apart to release the detached GCs (dGCs). Meanwhile, the mGCs were isolated by scraping the follicular wall. dGCs and mGCs were then processed for determining the mRNA levels of TGFBR1 or BMPR1B using qRT-PCR. (C) In situ hybridization showing localization of TGFBR1 and BMPR1B mRNA expression in porcine ovaries. Localization of specific mRNAs was monitored by DIG-UTP-labeled RNA probes (green). The cell nuclei were counterstained with DAPI (blue). Scale bars: 500 μm. d, dGCs; m, mGCs. Areas outlined in red are enlarged in the two lower panels. Scale bars: 50 μm. (D) The comparison of the mRNA levels of TGFBR1 or BMPR1B in mGCs, dGCs, cumulus cells and oocytes using qRT-PCR. (E,F) dGCs and mGCs collected from each of three follicles around 4 mm in diameter were labeled with PI for cell cycle analysis. The proportion of GCs in G2/M phase was quantified with flow cytometry. (G,H) The detection of apoptotic signals in dGCs and mGCs. The percentages of apoptotic cells were quantified by flow cytometry. (I) Proteins were extracted from three paired dGCs or mGCs harvested as described above, and western blotting was performed to detect changes in p-CDK1 (Thr14), p-CDK1 (Tyr15), p-Wee1 (Ser642) and p-CDC25B (Ser323). TUBA1A served as the control for loading. Data are mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001.
The decreased expression of GDF9/BMP15 receptors in GCs is responsible for the induction of G2/M arrest, apoptosis and cell shedding. (A,B) 120 antral follicles (3-7 mm in diameter) were individually dissected from porcine ovaries and torn apart to release the detached GCs (dGCs). Meanwhile, the mGCs were isolated by scraping the follicular wall. dGCs and mGCs were then processed for determining the mRNA levels of TGFBR1 or BMPR1B using qRT-PCR. (C) In situ hybridization showing localization of TGFBR1 and BMPR1B mRNA expression in porcine ovaries. Localization of specific mRNAs was monitored by DIG-UTP-labeled RNA probes (green). The cell nuclei were counterstained with DAPI (blue). Scale bars: 500 μm. d, dGCs; m, mGCs. Areas outlined in red are enlarged in the two lower panels. Scale bars: 50 μm. (D) The comparison of the mRNA levels of TGFBR1 or BMPR1B in mGCs, dGCs, cumulus cells and oocytes using qRT-PCR. (E,F) dGCs and mGCs collected from each of three follicles around 4 mm in diameter were labeled with PI for cell cycle analysis. The proportion of GCs in G2/M phase was quantified with flow cytometry. (G,H) The detection of apoptotic signals in dGCs and mGCs. The percentages of apoptotic cells were quantified by flow cytometry. (I) Proteins were extracted from three paired dGCs or mGCs harvested as described above, and western blotting was performed to detect changes in p-CDK1 (Thr14), p-CDK1 (Tyr15), p-Wee1 (Ser642) and p-CDC25B (Ser323). TUBA1A served as the control for loading. Data are mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001.
DISCUSSION
The development of ovarian follicles constitutes the foundation of female reproduction. The meiotic inhibition of oocytes and the mitotic proliferation of GCs are two basic aspects that must be established to ensure normal follicular development (Senbon et al., 2003). However, it remained unclear how these two processes were coordinated in the same follicular environment. As reported, the primary oocytes are halted in meiotic prophase I until ovulation, because the presence of inhibitory factors (such as Hx and natriuretic peptide) in follicular fluid prevents MPF from being activated (Wigglesworth et al., 2013). Unlike oocytes, GCs proliferate rapidly during follicular growth, reflecting a seemingly uninterrupted mitotic capability. The results described here, however, suggest that Hx within pFF indiscriminately represses the MPF of ovarian GCs, leading to G2/M cell cycle arrest. Interestingly, although the resumption of meiosis by the oocyte itself is blocked by Hx, our data show that the oocyte, via secreting ODPFs, including GDF9 and BMP15, prevents GCs from undergoing G2/M arrest, thus facilitating the mitotic progression of GCs upon Hx exposure. These findings might provide compelling evidence to explain why a cell cycle asynchrony exists between GCs and oocytes during follicular development (Fig. S16).
Hx inhibits the mitotic progression of GCs by repressing MPF activity through the PKA-Wee1/Myt1 axis
It has been well documented that the follicular purine Hx, which maintains meiotic arrest in cultured oocytes, is the principal inhibitor of oocyte maturation in vivo (Downs et al., 1985), but few researchers have investigated its role in GCs proliferation. As shown in this study, together with the results from others (Downs et al., 1985), porcine ovarian follicular fluid contains high concentrations of Hx. Specifically, Hx blocked GCs from entering M phase, hence triggering a retardation of proliferation. To our knowledge, the present study first describes a negative role for Hx in regulating GC proliferation by inhibiting the G2/M transition.
Throughout the duration of oocyte meiotic arrest, PKA is constitutively activated due to the accumulation of cAMP resulting from the suppression of cAMP phosphodiesterase (PDE) by Hx or cGMP that diffuses into the oocyte from GCs via gap junctions (Tripathi et al., 2010). Shortly before ovulation, the disconnection of communication between GCs and oocyte blocks the transfer of Hx and cGMP, leading to the reactivation of PDE that reduces cAMP level by hydrolysis, and thus removes PKA-mediated inhibition of MPF activity and meiotic resumption (Tripathi et al., 2010). However, there is no definite evidence about the PKA response when ovarian GCs experience Hx exposure. Here, we have found that Hx incubation improved PKA activation, as suggested by elevated levels of PRKACA, the catalytic subunit α of PKA. The increased PKA activity inactivates CDC25B phosphatase and ERK1/2, but activates Wee1/Myt1 kinase, which results in phosphorylation of CDK1 and inactivation of MPF, thereby impeding the G2/M transition in GCs. These findings are consistent with those observed in oocytes (Tripathi et al., 2010), suggesting that a common mechanism for modulating meiotic/mitotic progression through the PKA pathway might be shared by oocytes and GCs. Nevertheless, given that Hx has been reported to be converted into purine nucleotides by the purine salvage pathway at least in cumulus cells (Downs, 1997b), the possibility of alternative pathways for Hx in regulating GCs cell cycle remains to be investigated in the future.
ODPFs act through receptor-mediated activation of ERK1/2 to restore the G2/M transition in GCs with Hx exposure
Although Hx inhibits oocyte maturation, it exerts no apparent influence on follicular development (Senbon and Miyano, 2002), as indicated by uninterrupted proliferation of GCs (Robker and Richards, 1998). Consistent with this, our current results show that the increased level of Hx in pFF does not prevent follicles from growing larger. Actually, Hx failed to induce G2/M arrest in the GCs that were tightly anchored to the follicular wall, implying the existence of specific follicular factors that might overcome the inhibitory effects of Hx. To explore the possible protection factors against the action of Hx, we examined the involvement of gonadotropins and growth factors, both of which have been reported to facilitate meiotic resumption in oocytes. No significant results emerged easily, until we tested the effects of TGFβ, which efficiently restored the G2/M transition in GCs with Hx exposure. However, the physiological dose of TGFβ within ovarian follicles failed to restore G2/M progression or cell proliferation activity upon Hx exposure. As reported, oocyte-derived paracrine factors (ODPFs), including GDF9 and BMP15 (also known as GDF9B), are relevant members of the TGFβ superfamily that are abundant in the follicular fluid (Knight and Glister, 2006). It has become evident that GDF9/BMP15 regulates follicular development, e.g. by inhibiting the precocious differentiation of GCs (Knight and Glister, 2006). However, it remains unclear whether GDF9/BMP15 could maintain the mitotic activity of GCs during Hx exposure. In this study, we observed that both GDF9 and BMP15 relieved Hx-induced G2/M arrest in GCs, which was associated with a restoration of proliferation capability. By co-culturing GCs with the oocytes, we further confirmed that oocytes, by secreting GDF9/BMP15, propelled the cell cycle progression from G2 to M phase in Hx-treated GCs. On this basis, we further demonstrate that GDF9/BMP15 regulates G2/M transition of GCs through receptor-mediated ERK1/2 signaling in GCs. These findings might provide new evidence for the bidirectional communication between oocytes and GCs.
ODPFs might be required for gonadotropin-stimulated GCs proliferation
As the ovarian follicles grow in response to the gonadotropin stimulation, the rate of GCs proliferation is exponentially increased (Robker and Richards, 1998). The accelerated GCs division might be at least in part due to the effects of FSH signaling on promoting the cell cycle progression (Rao et al., 1978). However, our results showed that FSH failed to restore G2/M transition or proliferation in Hx-incubated GCs (Fig. S10). Although FSH has been reported to signal via activation of PKA (Hunzicker-Dunn and Maizels, 2006), the present study revealed no evident influence of FSH on PKA-dependent G2/M distribution in GCs treated with or without Hx. Notably, FSH appeared to reduce the proportion of Hx-treated GCs in G1 phase. Accordingly, when G2/M transition was resumed after blocking the PKA pathway, more GCs entered M phase in the presence of FSH. Indeed, evidence has emerged regarding the role of FSH in regulating GC proliferation by promoting the G1/S transition (Robker and Richards, 1998). Considering the effects of GDF9/BMP15 on removing G2/M arrest as shown in this study, it is therefore imaginable that FSH-induced GCs proliferation in vivo might actually be achieved via the coordination of FSH with other ODPFs.
The G2-to-M phase switch mechanism that is orchestrated via the counterbalancing actions of Hx and ODPFs might be relevant to the occurrence of GC apoptosis and detachment
It remains inconclusive whether there is a direct link between the cell cycle and apoptosis. Here, we showed that the induction of apoptotic GCs death is a result of prolonged G2/M arrest. In atretic follicles, apoptotic GCs are detached from the basal lamina, and become suspended in the follicular fluid (Yang and Rajamahendran, 2000). Our present data suggest a dose-dependent induction of G2/M arrest in detached GCs (dGCs) by the elevated concentration of Hx within pFF. Importantly, Hx-triggered G2/M arrest markedly increased the apoptosis rate in both cultured GCs and dGCs. Therefore, it is probable that Hx, through prohibiting the cell cycle transition in GCs, might clandestinely elicit the atretic degeneration of porcine follicles.
In addition, we found that the G2/M progression of GCs might also depend on their sensitivity to GDF9/BMP15, as reflected by the differential expression of GDF9/BMP15 receptors between mGCs and dGCs. It is possible that the inaccessibility of GDF9/BMP15 receptors to their ligands, and the resultant G2/M arrest, is responsible for the dissociation of GCs from the follicular wall. As the atretic process is characterized by GC detachment, we speculate that the varied expression of GDF9/BMP15 receptors in GCs might contribute to the selective follicular atresia.
In mammalian ovaries, some follicles prefer to grow towards ovulation, but most of them undergo atretic degeneration during development (Baker, 1963). Unfortunately, the exact mechanisms for selective atresia have not hitherto been elucidated. Even worse, follicles do not show any significant morphological alterations before the initiation of atresia; it is thus difficult to determine which follicle will be destroyed in future. As shown in the present study, there were variations in GCs cell cycle between individual follicles of the same diameter. Importantly, cell cycle arrest is closely related to the occurrence of GCs apoptosis and detachment. On the other hand, we found that the cell cycle, mitotic activity and viability of GCs might be fine-tuned by the counterbalance of intrafollicular factors, particularly HX and oocyte-derived GDF9/BMP15. Indicatively, these data might provide novel insights into the mechanisms required for selective atresia in mammalian follicles. Moreover, it remains unclear whether oocytes are relevant to the initiation of atresia in antral follicles. Our results demonstrate that the secretory factors of oocytes could prevent GCs apoptosis by removing the inhibitory effects of Hx on GCs mitosis; it thus raises the possibility that the secretory dysfunction of oocytes might initiate the atretic process of mammalian follicles.
MATERIALS AND METHODS
Reagents and antibodies
3-Methyladenine (3-MA; S2767), H89 (S1582), Ro-3306 (S7747), SCH772984 (S7101), Z-VAD-FMK (S7023), Repsox (S7223) and K02288 (S7359) were purchased from Selleck Chemicals. Hx, Recombinant human EGF (SRP3027) and anti-MAP1LC3B (L7543) were bought from Sigma-Aldrich. Antibodies against Myt1 (4282), phospho-Myt1 Ser83 (4281), cyclin B1 (12231), CDC25B (9525), ERK1/2 (4695), phospho-ERK1/2 Thr202/Tyr204 (4370), NF-κB p65 (8242), PCNA (2586), TUBA1A (2125), ATG3 (3415), ATG7 (8558), PUMA (4976), phospho-p53 Ser15 (9286), SAPK/JNK (9252) and phospho-SAPK/JNK Thr183/Tyr185 (4668) were obtained from Cell Signaling Technology. Anti-TGFBR1 (AF3025), anti-BMPR1B (MAB505) and recombinant human BMP15 (5096-BM-005) were from R&D Systems. Antibodies against phospho-CDK1 Thr14 (D155338), phospho-CDK1 Tyr15 (D155115), CDK1 (D160158), PRKACA (D163374), Wee1 (D163186), phospho-Wee1 Ser642 (D151274), phospho-CDC25B Ser323 (D155350) and BECN1 (D199499) were purchased from BBI Life sciences. HRP-conjugated goat anti-mouse IgG H&L (ab6789), HRP-conjugated goat anti-rabbit IgG H&L (ab6721) (ab6789) and anti-SQSTM1/p62 (ab56416) were obtained from Abcam. Anti-caspase 3 antibody (19677-1-AP) and anti-GAPDH antibody (10494-1-AP) were bought from Proteintech. FSH and LH were purchased from Ningbo Second Hormone Factory (Zhejiang, China). Recombinant human IGF1 (10598-HNAY1) and recombinant human TGFβ1 (10804-HNAC) were purchased from Sino Biological. GDF9 (APA427Hu01) was obtained from Cloud-Clone.
Sample collection
Porcine ovaries were collected from mature sows at a local slaughterhouse and transferred to the laboratory as soon as possible. The pFF was then extracted with a syringe from the antral follicles on the surface of ovaries. After centrifugation (1000 g, 5 min), GCs were isolated for primary cell culture. The COCs within the follicular extracts were retrieved by mouth pipetting under a surgical dissecting microscope (Olympus). The oocytes were separated from the COCs by digestion with hyaluronidase. In some experiments, individual follicles were dissected from the ovaries with small forceps. After measuring the size, follicles were torn apart to release the suspended GCs (referred to as dGCs in this study), which were collected by centrifugation (1000 g, 5 min). Meanwhile, the mural GCs (mGCs) were obtained by scraping the follicular wall. dGCs or mGCs were immediately processed for the determination of cell cycle distribution and apoptosis rate, or restored for qRT-PCR and immunoblotting analysis. The pFF supernatant were harvested for detecting the concentration of Hx or estrogen. The detailed process for the collection and separation of mGCs, dGCs, cumulus cells and oocytes is illustrated with a schematic diagram in Fig. S1.
Cell culture and treatments
Porcine ovarian GCs were collected from antral follicles (3-7 mm in diameter) using a syringe, washed with PBS (Gibco) and then cultured in DMEM/F-12 (1:1) medium (Life Technologies) supplemented with 10% fetal bovine serum (FBS; Sigma) and 100 units/ml penicillin plus 100 μg/ml streptomycin (Gibco) for 2 days at 37°C with 5% CO2. For drug administration, GCs were cultured with Hx (0, 1, 1.5 or 4 mM) alone or combined with the treatment of chloroquine (50 μM), 3-MA (250 μM), Z-VAD-FMK (50 μM), Ro-3306 (10 μM), H89 (5 μM), FSH (7.5 IU/ml), LH (15 IU/ml), EGF (10 ng/ml), IGF1 (20 nM), TGFβ1 (10 ng/ml), GDF9 (10 ng/ml), BMP15 (10 ng/ml), SCH772984 (1 μM), Repsox (2 μM) or K02288 (2 μM) for 0, 2, 4, 6, 12, 24 or 48 h as indicated. In some experiments, GCs were co-cultured with the oocytes at a density of 100 oocytes/ml, and treated with 1 mM Hx for 24 h. Culture of COCs was carried out in IVF medium [75.75% M199 (Sigma), 1% L-GlutaMAX (Gibco), 10% pFF, 10% FBS (Sigma), 1.5% PMSG (Ningbo Second Hormone Factory), 1.5% HCG (Ningbo Second Hormone Factory), 0.25% Ggentamicin (Shanxi Kelong Veterinary Drug)] with or without Hx (1 mM). 24 h later, cumulus cells were isolated from the COCs by digestion with hyaluronidase, and collected for the determination of cell cycle distribution. The denuded oocytes were fixed with 4% PFA (KeyGEN), and the percentage of GVBD was assessed by nuclear morphology using Hoechst 33342 staining.
Detection of germinal vesicle breakdown (GVBD)
To assess germinal vesicle breakdown (GVBD), the denuded oocytes were fixed for 2 h in 4% PFA (KeyGEN), and stained for 10 min at room temperature in 10 μg/ml of Hoechst 33342. The stained oocytes were then placed on glass slides and compressed with coverslips. The mounted oocytes were observed under a fluorescence microscope (Olympus). Oocytes were first examined under bright-field illumination to visualize morphology of nucleoli and nuclear envelope, and then observed with fluorescence optics to display Hoechst staining excited by ultraviolet light emitted from a mercury lamp.
High-performance liquid chromatography (HPLC)
The concentration of Hx in pFF was measured via HPLC. An aliquot (10 μl) of each pFF sample dissolved in homeopathic alcohol (49.5-50.5% ethanol) was injected into a reversed phase column Kromasil-C18 (Chrom Matrix) and processed for chromatographic fractionation at 35°C with a binary mobile phase (methanol-water, 10:90). The flow rate was set at 1 ml/min. Peaks were monitored at 254 nm through an ultraviolet visible detector (Thermo). The follicular Hx levels were quantified using calibration curves of the reference Hx compound (Sigma).
ELISA assay
Levels of estrogen, GDF9, BMP15 and TGFβ1 were determined using the Porcine Estradiol (E2) ELISA Kit (ColorfulGene), Porcine Growth Differentiation Factor 9 (GDF9) ELISA Kit (ColorfulGene), Porcine Bone Morphogenetic Protein 15 (BMP15) ELISA Kit (ColorfulGene) and Porcine transforming growth factor-β1 ELISA Kit (ColorfulGene) according to the manufacturer's instructions. Briefly, pFF samples or standards resolved in the dilution buffer were added to the microelisa stripplate pre-coated with an antibody specific to E2, GDF9, BMP15 or TGFβ1, followed by 30 min of incubation at 37°C with a HRP-conjugated antibody against E2, GDF9, BMP15 or TGFβ1. After washing, the substrate solution was added to trigger the chromogenic reaction. Only those wells that contain E2, GDF9, BMP15 or TGFβ1 and HRP-conjugated antibody will appear blue in color and then turn yellow after the addition of the stop solution. The absorbance was measured at 450 nm using a TECAN microplate reader. The concentration of E2, GDF9, BMP15 or TGFβ1 in the pFF was calculated from the standard curve.
Measurement of cell proliferation
The proliferation activity of GCs was measured using Dojindo Cell Counting Kit-8 (CCK-8; Dojindo Laboratories). The experimental procedures were carried out following the manufacturer's directions. Briefly, GCs were seeded in 24-well plates and cultured for 2 days. After the indicated treatments, 50 μl of tetrazolium substrate was added to each well and incubated at 37°C in a humid atmosphere containing 5% CO2 for 4 h. The optical density (OD) at 450 nm was then measured with a microplate reader (TECAN). The cell proliferation activity was calculated by the following formula: proliferation activity (%)=[OD (experiment)−OD (blank)]/[OD (control)−OD (blank)]×100. Cell proliferation was also evaluated by comparing the total number of cells before and after Hx treatment. Briefly, GCs seeded in 24-well dishes were treated with or without Hx, the cell numbers were counted with hemocytometer (QIUJING) at 0 h and 24 h.
Detection of ROS production
ROS levels were determined by measuring the oxidative conversion of nonfluorescent DCFH-DA to highly fluorescent DCF. The experiments were performed using the Intracellular ROS Fluorescence Determination Kit (GENMED) according to the manufacturer's instructions. Fluorescence intensity was monitored using a fluorescence microplate reader (TECAN).
RNA interference
The siRNA directed against CDK1 and the scrambled control siRNA were obtained from GenePharma (Shanghai, China). siRNA transfection was performed using Lipofectamine 3000 reagent (Invitrogen, L3000015) according to the manufacturer's instructions. The siRNA sequences were provided in Table S1.
Confocal imaging of autophagic puncta
GCs grown in 12-well plates were transfected with the GFP-MAP1LC3B expression vector (Addgene 22418). After 24 h, cells were subjected to the indicated treatments. The formation of autophagosomes represented by the intracellular GFP-MAP1LC3B dots was visualized using a confocal fluorescence microscope (Zeiss).
Flow cytometry
GCs were digested with 0.25% trypsin (Gibco), rinsed twice with ice-cold PBS, fixed in 70% ethanol for at least 2 h. After washing, cells were incubated with the PI/RNase A staining solution (KeyGEN) for 30-60 min in the dark at 4°C. The cell cycle was then analyzed using a BD Accuri C6 flow cytometer (Becton Dickinson). Apoptosis was measured by flow cytometry using the annexin V FITC/PI Kit (Vazyme Biotech) according to the manufacturer's instructions.
CDK1-cyclin B activity assay
The fluorescence resonance energy transfer (FRET) quantitative detection kit (GENMED) was used for determining cellular CDK1-cyclin B activity. This fluorescent strategy is based on CDK1-mediated phosphorylation of Ser/Thr in target sequence EDAN-Ser/Thr-Pro-X-Lys/Arg-DABCYL, thus preventing aminopeptidase-catalyzed release of EDAN-Ser/Thr, which generates strong fluorescence when detached from DABCYL. The alterations of fluorescence intensity (with an excitation at 340 nm) were analyzed to quantify the specific activity of CDK1-cyclin B.
Staining of M-phase GCs
To calculate the percentage of M-phase cells, GCs were treated as indicated and exposed to DAPI staining. Briefly, cells were rinsed with PBS, and fixed using 4% PFA for 1 h. The nuclei were then labeled by culturing cells with 5 μg/ml DAPI (KeyGEN) at room temperature for 15 min. After being washed twice with PBS, the number of mitotic cells at different stages (including prophase, metaphase, anaphase and telophase) were counted under a fluorescence microscope (Olympus).
Quantitative RT-PCR (qRT-PCR)
Total RNA was isolated with TRIZOL reagent (Invitrogen), and reverse transcribed into cDNA using the PrimeScript RT Master Mix (Takara) according to the manufacturer's instructions. The qRT-PCR was conducted using AceQ qPCR SYBR Green Master Mix (Vazyme) and gene-specific primers (see Table S2 for primer sequences) in a StepOnePlus Real-Time PCR System (Applied Biosystems). Expression data were normalized to the amount of Tuba1a expressed. Melting curves were analyzed to verify amplification specificity.
Fluorescent in situ hybridization
Paraformaldehyde (4%)-fixed porcine ovaries were embedded in paraffin, sectioned with a microtome onto glass slides, deparaffinized and rehydrated before digestion with 1 μg/ml Proteinase K (Sigma) for 25 min. After 15 min of 3% H2O2 (Sigma) incubation to quench intracellular peroxidase activity, ovarian sections were blocked with 1% BSA, incubated with the indicated DIG-UTP-labeled RNA probes (see Table S3 for probe sequences), followed by sequential detection using HRP-labeled anti-DIG antibody (Boehringer, Mannheim) and FITC-conjugated tyramide amplification. The cell nuclei were counterstained with DAPI. Fluorescent images were captured under a confocal fluorescence microscope (Zeiss).
Western blot
Equal amounts of proteins extracted from GCs using radioimmune precipitation assay buffer (Beyotime) were fractionated using 12% Express Plus PAGE gel (Genscript) and transferred to a polyvinylidene difluoride membrane (Millipore). Nonspecific binding sites were blocked with TBST (Solarbio) containing 5% bovine serum albumin. The membranes were then treated with primary antibodies diluted (1:1000) in blocking solution overnight at 4°C, and incubated with a horseradish peroxidase-conjugated secondary antibody (1:2000) at room temperature for 2 h. The immune-reactive bands were visualized by a WesternBright ECL HRP substrate kit (Advansta) according to the manufacturer's instructions. The relative expression of target proteins was normalized to TUBA1A as the control for loading.
Statistical analysis
Data are expressed as means±s.e.m. Statistical analysis was performed using the SPSS version 16.0 software (SPSS). Differences between two groups were assessed using an unpaired Student t-test, and between multiple groups using one-way ANOVA. Values of P<0.05 were considered significant. All experiments were repeated at least three times.
Footnotes
Author contributions
Conceptualization: M.S., H.L., C.L.; Methodology: Jilong Zhou, J.L., J.T., W.W., M.S., H.L.; Software: M.S., C.L.; Validation: M.S., C.L.; Formal analysis: C.L., M.S.; Investigation: C.L., X.M., S.L., W.L., X.Z., W.Y., C.D., Z.L., Jiaqi Zhou, M.S.; Resources: M.S., H.L.; Data curation: C.L., M.S.; Writing - original draft: M.S.; Writing - review & editing: M.S, C.L..; Visualization: M.S., C.L.; Supervision: M.S., H.L.; Project administration: M.S., H.L.; Funding acquisition: M.S., H.L.
Funding
This work was supported by the National Natural Science Foundation of China (31972571, 31630072, 31972564 and 31601939), by a project funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions (DKQB201903), by the Fundamental Research Funds for the Central Universities (KJQN201705), by the Natural Science Foundation of Jiangsu Province (BK20150664) and by the National Major Project for Breeding of Transgenic Pigs (2016ZX08006001-003).
Peer review history
The peer review history is available online at https://dev.biologists.org/lookup/doi/10.1242/dev.184838.reviewer-comments.pdf
References
Competing interests
The authors declare no competing or financial interests.