Aberrant BMP15/HIF-1α/SCF signaling pathway in human granulosa cells is involved in the PCOS related abnormal follicular development

Abstract Aims To investigate the regulatory mechanism of SCF expression in human GCs of PCOS related follicles. Materials and Methods SCF, BMP15 and HIF-1α were evaluated in human serums, follicular fluids (FFs) and GCs, which were collected from 69 PCOS patients and 74 normal ovulatory patients. KGN cell line was used in this study. Results Our results showed that the rate of MII oocyte and 2PN fertilization was lower in PCOS group, though PCOS patients retrieved much more oocytes. The level of BMP15 in FF and the level of SCF in serum and FF were also lower in PCOS patients. We found a weakened expression of HIF-1α and SCF in GCs from PCOS patients when compared with the non-PCOS patients. The expression of HIF-1α and SCF was significantly increased in KGN cells after treating cells with rhBMP15, however, this promotion effects of BMP15 on HIF-1α and SCF expression were obviously abolished by co-treatment with BMP-I receptor inhibitor (DM). Moreover, knock down of HIF-1α expression in KGN cells significantly reduced the expression of SCF in human GCs, in spite of activating BMP15 signaling pathway. Conclusions The present study suggest that BMP15 could induce SCF expression by up-regulating HIF-1α expression in human GCs, the aberrance of this signaling pathway might be involved in the PCOS related abnormal follicular development.


Introduction
Polycystic ovary syndrome (PCOS) is one of the most common reproductive endocrine disorders in women of childbearing age [1]. It is diagnosed if two of the three following characteristics are present: clinical or biochemical hyperandrogenism, oligo or amenorrhea, and polycystic ovaries [2]. Patients with PCOS often suffer from obesity, abdominal distribution of body fat, type 2 diabetes mellitus, cardiovascular diseases, and even gynecological cancers, which is detrimental to their heath and quality of life [3][4][5][6]. Besides, the major problem of PCOS is anovulatory infertility resulting from abnormal follicular development [7,8]. However, its etiology and pathogenesis remain unknown.
Normally, ovarian follicle develops from primordial phase into mature phase, which is tightly regulated by hormones, protein factors and cell-cell interactions [9,10]. Therefore, the interactions between inhibitory signals and stimulatory factors ensure the ovarian follicular develop normally. Stem cell factor (SCF) is a granulosa-derived cytokine growth factor, which stimulates the c-kit receptor expressed by oocytes [11,12]. Recent studies have demonstrated the important role of SCF in the initiation and the maintenance of folliculogenesis [13]. Exogenous added SCF increased the diameter of oocytes from primordial to early primary follicles in the mouse, and this promotion was inhibited by SCF-neutralizing antibody [14]. Furthermore, Tan et al have showed a positive and statistically significant relationship between SCF level and oocyte developmental potential in human follicles [15]. These studies suggest that SCF might be a critical factor in regulation of follicular development. It has been known that PCOS patients are often accompanied with abnormal follicular development and poor quality oocytes [16]. Remarkably, our previous studies have found a reduced SCF expression in serum, follicular fluid (FF) and granulosa cells (GCs) in PCOS patients, indicating that decreased SCF expression might be associated with PCOS related abnormal follicular development [17]. Nevertheless, the reasons of weakened SCF expression and it related regulatory mechanisms in PCOS are still unclear.
The communication between GCs and oocyte appears to be crucial for the regulation of early follicular development and the promotion of the secretion of oocyte factors, which in turn stimulates the proliferation and differentiation of the surrounding GCs [18]. Bone morphogenetic protein (BMP) 15 is one of the important oocyte-derived factors [19], which belongs to the member of the TGF-β superfamily [20]. It has been demonstrated that BMP15 can bind with BMP receptor expressed on the GCs membrane, and subsequently stimulate the downstream SMAD signaling pathway leading to regulation of gene transcription [20]. Several studies have found the positive relationship between BMP15 and follicular development [21,22]. Increased BMP15 in GCs is associated with oocyte maturation, fertilization, and embryo quality in humans [23]. Whereas, knock down of BMP15 expression by miR-378 significantly disturbed the oocyte-cumulus interaction and resulted in oocyte maturation arrest in mouse [24]. It has been reported that a reduced BMP15 expression has been found in follicles of PCOS ovarian tissues, and this reduction of BMP15 might not only influence the follicular development, but also impact on the formation of zona pellucida structure, finally causing subfertility or infertility in PCOS patients [25,26]. Moreover, a recent study has reported that BMP15 may positively regulate SCF gene expression in mouse GCs [27]. Together with our previous findings that patients with PCOS showed a decreased SCF expression in serum, FF and GCs [17], we speculate that the reduction of SCF expression in PCOS is likely to be attributed to the weakened BMP15 expression. Nonetheless, the molecular regulatory mechanism between BMP15 and SCF expression in PCOS has yet to be investigated.
More recently, the hypoxia inducible factor-1α (HIF-1α) has been showed to promote the follicular development in mouse ovary [28]. HIF-1α, the master orchestrator of cellular adaptation to hypoxia environment, serves as a key adjuster of the cellular adaptive response to both physiological and pathological low oxygen conditions [29]. Activation of HIF-1α regulates many downstream genes, such as VEGF, STAR and endothelin (ET)-2 [30][31][32], which plays an important function during the process of follicular development [33]. In mammalian ovarian GCs, HIF-1α could trigger cell autophagy and glucose uptake and facilitate follicular development [34], as well as promote the proliferation and the steroid production of GCs [35]. Interestingly, recent studies have identified HIF-1α as a candidate transcriptional regulator of SCF in multiple kinds of cells, such as pancreatic ductal adenocarcinoma cell, epithelial breast cancer cell and bone marrow stem cell [36]. Therefore, we hypothesize that whether HIF-1α could be a pivotal element in the program of BMP15 regulating SCF expression in human GCs during the follicular development.
In the present study, we firstly detected the expression of BMP15, HIF-1α and SCF in patients with or without PCOS undergoing in vitro fertilization (IVF) treatment, respectively. Subsequently, we investigated the regulatory mechanism between BMP15, HIF-1α and SCF by using granulosa cells collected from patients with or without PCOS, and to further confirmed our findings by using human granulosa cell line (KGN).

Patients
From September 2018 to January 2019, we enrolled 69 patients with PCOS (PCOS group) and 74 patients with normal ovulatory cycles (non-PCOS group) who under IVF therapy in Reproductive Medicine Center of Jiangxi Provincial Maternal and Child Health Hospital. The study was a single-center study. All patients were in good physical and mental condition. PCOS patients were diagnosed based on the Rotterdam criteria [2]. Patients with a history of the following procedures or disorders were excluded: ovarian surgery, radiotherapy or chemotherapy, premature ovarian failure, ovarian dysfunction, hyperprolactinemia, thyroid dysfunction, or ovulation induction within 3 months.

Controlled ovarian stimulation
Ovarian stimulation was performed with the use of a prolonged protocol. Briefly, standard full dose of gonadotropinreleasing hormone agonist (3.75 mg, GnRH-a, Ipsen, France) was used on the second day of menstrual cycle for down regulation. Pituitary down regulation (Endometrial thickness ≤5 mm, serum FSH <5 mIU/mL, LH <5 mIU/mL, E2 < 50 pg/mL) was confirmed with transvaginal ultrasound and endocrine examination after 30 days. Then, according to the patient's age, body mass index, serum basal FSH levels, LH levels, estradiol levels and antral follicle count, initial doses of 75-112.5 IU/d of recombinant human FSH (MerckSerono, German) were used. The time and dose of recombinant human FSH was adjusted according to ovarian response as monitored by serum estradiol levels and vaginal ultrasound. When the dominant follicle was ≥19 mm in diameter or at least 2 follicles were ≥18 mm in diameter, recombinant human FSH was stopped and a single injection of 6000-8000 IU of hCG (Merck-Serono, German) was administered. Oocyte retrieval was performed 36-40 h later under transvaginal ultrasound guidance.

Samples collection
Serums (n = 143) and FFs (n = 161) samples were collected during oocyte retrieval. For FF samples, the follicles visualized by ultrasound were aspirated individually without flushing at the day of oocyte retrieval. The FFs were pooled for each patient, and both were used to test the concentrations of BMP15 and SCF. Any follicular aspirate that was not clear or showed gross contamination with blood was discarded. The FF samples were immediately centrifuged at 3000 rpm for 10 min. Supernatants were aspirated, divided into aliquots, and frozen at −80 °C for future analysis.

ELISA for BMP15 and SCF measurements
Concentrations of BMP15 and SCF in FFs and serums were determined with a commercially available enzyme-linked immunosorbent assay (ELISA) kit, respectively. All of the procedures were performed according to the manufacturers' instructions.

Human GC collection
Follicular fluids (n = 231) were centrifuged at 2000 rpm for 5 min. The cells were resuspended with DMEM/F12 medium and transferred to a 50% (volume fraction) Percoll gradient (Sigma-Aldrich, USA); they were centrifuged at 4000 rpm for 20 min to purify human GCs. After washing and recentrifugation, sheets of human GCs were digested with hyaluronidase at a 1:1 ratio for 2 min to separate them. The GCs were removed using a pipette and washed with phosphate buffered saline (PBS). The cells were stored at −80 °C for future analysis.

KGN cell culture
Human granulosa cell line (KGN) was kindly provided by Prof. Qiao Jie (Department of Obstetrics and Gynecology, Peking University Third Hospital, Beijing). Cells were plated in 6-well at a density of 1 × 10 5 cells per well in DMEM/F12 (Life Technologies, USA) containing 10% fetal bovine serum (FBS, Gibco, USA) and incubated at 37 °C in a humidified atmosphere of 95% air and 5% CO 2 . Subsequently, cells were treated with BMP15 (100 ng/mL) alone or in combination with Dorsomorphin (1 μM) alone for 24 h.

Small interfering RNA (siRNA) transfection
HIF-1α siRNA and control sequences were obtained from RiboBio (ShenZhen, China). Transient transfection of KGN cells with HIF-1α siRNA and control sequences was performed with Lipofectamine 2000 in accordance with the manufacturer's protocol but with slight modifications. After 48 h, the cells were treated with BMP15 (100 ng/mL) for 1 h, then real-time PCR and western blotting assays were performed.

Western blot
Total proteins were extracted from cells using the RIPA lysis buffer containing protease inhibitors (Applygen, China) and phosphatase inhibitors (Sigma, USA). The protein concentrations were determined by NanoDrop 2000c spectrophotometer using BCA protein assay kit (Applygen, China). After loading equal amount of protein samples, SDS-PAGE (12% sodium dodecyl sulfate polyacrylamide gel electrophoresis) was performed. The proteins were then transferred to a PVDF membrane (Merck-Millipore, USA). After blocking with Tris buffered saline containing 0.05% Tween-20 (TBST) and 5% nonfat dry milk or 5% BSA for 1 h, the membrane was incubated with corresponding antibodies at 4 °C overnight, washed in TBST, followed by incubation with the corresponding horseradish peroxidase-conjugated secondary antibodies for 1 h. Visualization of the proteins was detected with ECL chemiluminescence. Beta-actin was used as a loading control. The intensity values were assessed and analyzed with Image J software.

RNA extraction and real-time PCR
Total RNA was extracted from embryos with RNeasy kit (Qiagen, China) according to the manufacturers' instructions. Reverse transcription reactions were performed using Super cDNA First-Strand Synthesis Kit (CWBiotech, China), and the quantification of total RNA was performed on a NanoDrop spectrophotometer. Real-time PCR was performed in an ABI 7500 real-time PCR system (Applied Biosystems, USA) using Ultra SYBR Mixture with ROX (CWBiotech, China). The following primers were used: SCF (Forward: CAGAGTCAGTGTCACAA AACCATT, Reverse: TTGGCCTTCCTATTACTGCTACTG); HIF-1α (Forward: GTCTGAGGGGACAGGAGGAT, Reverse: CTCCTCAGGTGGCTTGTCAG); GAPDH (Forward: AGAAG GCTGGGGCTCATTTG, Reverse: AGGGGCCATCCACAG TCTTC). The reactions were incubated at 95 °C for 10 min, followed by 40 cycles at 95 °C for 15 s and at 60 °C for 1 min. All reverse transcription reactions included no-template controls; and all PCR reactions were run in triplicate. Relative gene expression was determined using comparative CT (2-ΔΔCt) method.

Oocyte and embryo assessment
Oocyte was regarded as being at the metaphase II (MII) stage when the first polar body was observed in the oocyte cytoplasm. The oocyte maturation rate refers to the number of MII oocytes divided by the total number of all retrieved oocytes. Oocyte fertilization was assessed 16-20 h after conventional adding sperms. Normal fertilization was confirmed when two pronuclei (2PN) were found in the cytoplasm. The normal fertilization rate refers to the number of fertilized oocytes divided by the total number of all retrieved oocytes. Embryo cleavage was examined 40-44 h after adding sperms. An embryo with a good quality should consist of 7-9 blastomeres with a uniform size, and the fragment proportion should be less than 20% at day 3. The good quality embryo rate refers to the number of good quality embryos divided by the total number of all embryos. The assessment was made in a blinded manner by two embryologists.

Statistical analysis
The software package SPSS 17.0 (SPSS Inc., Chicago, IL, USA) was used for all data analysis. In general, results among experimental groups were analyzed by student's t-test or one-way ANOVA. For all tests, p-value < 0.05 was considered statistically significant.

General characteristics
The general characteristics of PCOS patients and non-PCOS patients were summarized in Table 1.

Comparison of the IVF outcomes between PCOS patients and non-PCOS patients
It has been known that patients suffered from PCOS were often accompanied with aberrant follicles and unsatisfactory fertilization rate [16]. Similarly, our results also showed a worse IVF outcome in PCOS patients than that in non-PCOS patients. As shown in  > 0.05). These data suggested high rate of immature oocyte and low rate of fertilization in PCOS patients.

Concentration of BMP15 and SCF in serums and FFs collected from PCOS patients and non-PCOS patients
Several studies have reported that BMP15 and SCF are associated with abnormal follicular development in PCOS [25,26]. Thus, we analyzed the concentration of BMP15 and SCF in serums and FFs collected from PCOS patients and non-PCOS patients, respectively. The mean BMP15 concentration were 9.01 ± 2.38 ng/ mL (PCOS) vs 8.94 ± 2.52 ng/mL (non-PCOS) in serum and 0.89 ± 0.29 ng/mL (PCOS) vs 1.74 ± 0.51 ng/mL (non-PCOS) in FF, respectively (Table 3). Furthermore, The mean SCF concentration were 1.57 ± 0.11 ng/mL (PCOS) vs 2.15 ± 0.17 ng/mL (non-PCOS) in serum and 0.31 ± 0.03 ng/mL (PCOS) vs 0.67 ± 0.07 ng/mL (non-PCOS) in FF, respectively (Table 3). Compared with the non-PCOS patients, the concentration of BMP15 in FF and the concentration of SCF in serum and FF from PCOS patients were significantly lower (p < 0.05) ( Table 3).
No statistically significant correlation was observed for the concentration of BMP15 in the serum (p > 0.05) ( Table 3).

Expression of HIF-1α and SCF in GCs collected from non-PCOS patients and PCOS patients
Previous studies have found the regulatory relationship between HIF-1α and SCF in multiple types of tumor cells [37]. Next, we investigated the expression of HIF-1α and SCF in GCs from these two groups. As shown in Figure 1A, the level of HIF-1α and SCF protein was obviously weaken in GCs from PCOS patients when compared with the non-PCOS patients. Furthermore, real time-PCR also showed a decreased HIF-1α and SCF mRNA in GCs from PCOS group those from non-PCOS patients (p < 0.05) ( Figure 1B). These results indicated that the expression of HIF-1α and SCF was reduced in GCs from PCOS patients.

Recombinant human BMP15 increases SCF mRNA and protein by up-regulating HIF-1α expression in human granulosa cells
It has been reported that BMP15 could up-regulate SCF expression in mouse GCs [27]. Together with our above findings, we speculated that HIF-1α might be a key factor in the program of BMP15 regulating SCF expression in human GCs.  note: The data were expressed as mean ± Sd. abbreviations: BMP15, bone morphogenetic protein 15; ScF, stem cell factor; FF, follicular fluid; PcoS, polycystic ovary syndrome; non-PcoS patients, normal ovulatory cycles patients. Therefore, we firstly treated KGN cells with rhBMP15 to investigate the expression of HIF-1α and SCF. Our results showed that the expression of HIF-1α and SCF was significantly increased after treating with recombinant human BMP15 ( Figure 2). However, this promotion effects of BMP15 on HIF-1α and SCF expression were abolished by co-treatment with BMP type I receptor inhibitors (Dorsomorphin, DM) ( Figure 2), implying that BMP15 might promote SCF expression via regulating HIF-1α factor. To further confirm our hypothesis, we then knocked down the HIF-1α expression in KGN cells by transfecting cells with si-HIF-1α RNA (Supplement Figure 1), and found that knock down of HIF-1α significantly reduced the expression of SCF in human GCs, in spite of activating BMP15 signaling pathway ( Figure 3). These data supported that BMP15 increase SCF expression mainly through up-regulating HIF-1α in human GCs.

Discussion
The clinical outcomes of PCOS and non-PCOS patients in our study, in which PCOS patients showed a high rate of immature oocyte and low rate of fertilization, further confirmed the the previous studies that PCOS was often accompanied with aberrant follicular development [16]. But the cleavage rate and high-quality embryo rate between these two groups was no statistical  difference. The reasons for no statistical difference between these two groups were likely that we calculated the cleavage rate and the high-quality embryo rate with 2PN as the denominator. It is well known that SCF activates c-kit receptor expressed on oocyte and plays a key role during the development of oocyte/ follicles [12,38,39]. Our previous studies have found a weakened SCF expression in FFs and GCs from PCOS patients, suggesting that the reduction of SCF was closely associated with PCOS related abnormal follicular development. However, the reasons why SCF was reduced in PCOS related abnormal follicles are still unclear.
Studies have demonstrated that BMP15 could promote SCF expression in rat antral GCs [40], furthermore, decreased BMP15 was found to be associated with PCOS related abnormal follicles [25,26]. Therefore, we firstly compared the concentration of BMP15 and SCF in FFs and serums between the PCOS patients and non-PCOS patients. In line with the previous study [17], the levels of SCF in FFs and serums were significantly lower in PCOS patients. Moreover, we also found a significantly weakened BMP15 concentration in PCOS patients' FFs, but there was no statistically difference of BMP15 concentration in serum, which could be due to the differences in cells or organs producing BMP15. These findings suggest that the weakened BMP15 level might be an answer of that why SCF is decreased in PCOS related abnormal follicles. Nonetheless, little is known about the molecular regulatory mechanism between BMP15 and SCF in PCOS.
It has been proven that HIF-1α serves as a candidate transcriptional regulator of SCF in many kinds of cells [41], more importantly, HIF-1α expression was significantly increased during the follicular growth and development of postnatal rats [42] and facilitated the proliferation and steroid production of GCs [43]. These studies indicated a key role of HIF-1α in follicular development and reminded us that HIF-1α might also be a transcriptional regulator of SCF in human GCs, which is similar to other kinds of cells. Therefore, we then detected the expression of HIF-1α and SCF in GCs from patients with and without PCOS. The results showed an obvious reduction of HIF-1α and SCF in PCOS patients, implying that HIF-1α is closely related to PCOS abnormal follicles and is positively associated with SCF expression in human GCs. Moreover, in the light of previous studies that the expression of HIF-1α could be regulated by MEK/ERK signaling pathway [44], which is an important signaling pathway downstream of BMP receptor in granulosa cells [45]. All these results encourage us to raise a possible that HIF-1α might be correlated to the program of BMP15 regulating SCF expression in human GCs during the PCOS related abnormal follicular development.
To verify our hypothesis, we subsequently investigated the regulatory relationship between BMP15, HIF-1α and SCF by using KGN cell line. Firstly, we test the effects of BMP15 on HIF-1α expression in KGN cell line, and found that exogenous added BMP15 statistically significantly increased HIF-1α protein as well as mRNA levels. Furthermore, the promotion effect of BMP15 on HIF-1α was remarkably weakened after adding BMP type I receptor inhibitors (Dorsomorphin, DM), indicating that HIF-1α expression is positively regulated by BMP15 in human GCs. Additionally, we also found a similar promotion function of BMP15 on SCF expression in human GCs, which was in line with the previous studies [40]. To further confirmed our findings in detail, we then knocked down the HIF-1α mRNA expression in KGN cells by transfecting cells with si-HIF-1α RNA. As we expected that the promotion effect of BMP15 on SCF expression was obviously abolished after knocking down HIF-1α mRNA expression. These observations strongly supported that HIF-1α might be a major transcriptional regulator involved in SCF up-regulation in response to BMP15 in human GCs.

Conclusion
In summary, the results of this study showed a significantly reduction of BMP15, HIF-1α and SCF in PCOS related abnormal follicles, and demonstrated that BMP15 could induce SCF expression by up-regulating HIF-1α expression in human GCs. The aberrance of this signaling pathway might be involved in the PCOS related abnormal follicular development.
Next, we plan to increase the number of patients and animal experiments to further confirmed our findings.

Author's note
The study was approved by the Clinical Ethical Committee of Jiangxi Provincial Maternal and Child Health Hospital, and informed consents from patients were obtained before the initiation of the study.