Abstract
Bone morphogenetic proteins (BMPs) are known to play an indispensable role in preventing the precocious luteinization of granulosa cells within growing ovarian follicles. In this study, we found that the transcripts of BMP8 genes are enriched in the ovaries of humans and rodents. When analyzing transcriptomic datasets obtained from human mature granulosa cells, we further found that the BMP8 transcripts not only show the highest abundance among the searchable BMP-related ligands but also decrease significantly in women of advanced age or women with polycystic ovarian syndrome. The correlation between the BMP8 levels in granulosa cells and the decline in ovarian function in these subjects suggests that BMP8 protein may be involved in the regulation of granulosa cell function(s). Using a rat model, we demonstrated that human BMP8A protein activates the SMAD1/5/8 and the SMAD2/3 pathways simultaneously in both immature and mature granulosa cells. Furthermore, the expression of potential type I and type II receptors used by BMP8 in rat granulosa cells was characterized. We found that BMP8A treatment can significantly inhibit gonadotropin-induced progesterone production and steroidogenesis-related gene expression in granulosa cells. Pathway dissection using receptor inhibitors further revealed that such inhibitory effects occur specifically through the BMP8-activated SMAD1/5/8, but not SMAD2/3, pathway. Taken together, considering its abundance and possible functions in granulosa cells, we suggest that BMP8 may act as a novel luteinization inhibitor in growing follicles.
Introduction
Successful ovulation is governed by cooperation between gonadotropins and various local factors. During folliculogenesis, follicle-stimulating hormone (FSH) mediates follicular transition from antral follicles to preovulatory follicles. The evoking signal promotes not only proliferation but also estradiol production in granulosa cells. By the end of the follicular phase, the luteinizing hormone (LH) surge triggers ovulation, which is followed by the terminal differentiation of residual follicular cells into luteal cells that then produce progesterone (Richards et al. 1995). Of interest, while gonadotropins elicit their functions for follicle maturation and ovulation, the involvement of local factors, namely luteinization inhibitors, is required to selectively attenuate progesterone biosynthesis before ovulation in order to prevent premature luteinization (Shimasaki et al. 1999, Oktem & Urman 2010).
In mice, it has been shown that premature luteinization can block ovulation by trapping the enclosed oocytes within luteinizing granulosa cells (Pangas et al. 2006, Grasa et al. 2016). In humans, although it has not been clearly linked to a similar phenotype, premature luteinization is one of the common phenomena observed in older women with poor reproductive performance and has been hypothesized to be one of the causes of anovulation in women with polycystic ovarian syndrome (PCOS) (Willis et al. 1998, Wu et al. 2015). In women of advanced age, for example, their granulosa cells, which are collected from follicular fluid, tend to undergo earlier luteinization, as reflected by increases in the expression of LH receptor (LHCGR), CYP11A1 and progesterone receptor (PGR); this is consistent with the facts of declining numbers in retrieved oocytes and embryos and lower pregnancy rates in these individuals (Wu et al. 2015). Intriguingly, a study indicates that granulosa cells from follicles of PCOS patients express higher levels of LHCGR and CYP11A1 than those from size-matched control follicles of regularly cycling women; this suggests that granulosa cells in these selected PCOS subjects may undergo premature luteinization (Jakimiuk et al. 2001). In addition, in patients who undergo in vitro fertilization (IVF), premature luteinization has been associated not only with decreased oocyte recovery, poor oocyte maturation and a lower fertilization rate (Stanger & Yovich 1985, Lejeune et al. 1986, Silverberg et al. 1991), but also with a lower rate of pregnancy and a higher incidence of spontaneous abortion (Schoolcraft et al. 1991, Fanchin et al. 1993, Manzi et al. 1995). Therefore, preventing premature granulosa cell luteinization may greatly improve IVF outcomes.
The molecular mechanisms that prevent the precocious luteinization of granulosa cells are still poorly understood. However, previous studies have indicated that many members of the bone morphogenetic protein (BMP) subfamily have an indispensable role in guiding this event (Knight & Glister 2006). Indeed, several BMP members secreted by the ovary, including BMP2, BMP4, BMP7 and BMP15, have been tested individually in vitro and shown to prevent luteinization of human granulosa cells efficiently (Chang et al. 2013, 2015, Shi et al. 2010, 2011). However, whether other BMP candidates are involved in preventing premature luteinization and whether differential expression of BMPs can reflect the luteinization status of human mature granulosa cells remain unclear.
BMP8 is encoded by a pair of genes, BMP8A and BMP8B (Zhao & Hogan 1996). In this study, we initially found that the two BMP8 genes have a differential tissue expression profile in various tissues of humans and rats. Among the reproductive organs, the transcripts of both BMP8A and BMP8B were enriched in the ovary. We previously have demonstrated that both recombinant BMP8A protein and BMP8B protein are functional to activate canonical BMP signaling in different cells, including spermatogonia, P19 and 293T cells (Wu et al. 2017). As mentioned above, BMP signaling is indispensable for the prevention of the precocious luteinization of granulosa cells. In this study we therefore are particularly interested in the evaluation of the BMP8 roles in this event. To clarify the expression profiles of BMP8, the ovaries from superovulated rats and the transcriptomic datasets of granulosa cells derived from women at different ages or women with PCOS were used. To evaluate the effects of BMP8 protein on preventing luteinization, granulosa cells collected from a superovulated rat model were used. In addition, the underlying mechanisms whereby BMP8 protein is able to suppress the gonadotropin induction of progesterone production and steroidogenic enzyme expression were further characterized.
Materials and methods
Animals and ethics
Sprague–Dawley rats were obtained from the laboratory animal center (National Yang-Ming University, Taipei, Taiwan). In the superovulation model, immature female rats (26 days old) were intraperitoneally injected with 15 international units (IU) pregnant mare serum gonadotropin (PMSG) in the morning (09:00–10:00 h). Rats were then anesthetized and killed using CO2 at indicated time points. All animals were housed under a controlled humidity, temperature, and light regimen. Animal care and treatments were approved by the Institutional Animal Care and Use Committee of the National Yang-Ming University (Permit No. 1021229).
Hormones and reagents
L-15 Leibovitz medium, McCoy’s 5A medium, fetal bovine serum, and penicillin-streptomycin-glutamine were obtained from Invitrogen. PMSG, hCG and human FSH were obtained from Calbiochem. Human BMP4 (314-BP), activin A (338-AC) and human BMP8A (1073-BPC) proteins were purchased from R&D systems. SB431542, dorsomorphin, hyaluronidase, forskolin and other chemicals unless noted were purchased from Sigma-Aldrich.
Regarding generation of bioactive recombinant human BMP8A protein, we previously have demonstrated that BMP10 prodomain, known to facilitate the production and folding of the mature domain of some BMP molecules (Mazerbourg et al. 2005), greatly helps the production of functional BMP8A mature protein than BMP8A prodomain itself in the 293T expression system (Wu et al. 2017). Thus, this strategy was applied in this study. Briefly, recombinant DNAs that encode human BMP10 prodomain and human BMP8A mature domain or N-tagged His6-human BMP8A mature domain were fused and cloned into pcDNA3.1 vector. The 293T cell conditioned medium containing recombinant non-tagged BMP8A protein was collected and concentrated 20-fold by filtration with 3-kD molecular cutoff filters (Amicon Ultra, Millipore). The 293T cell conditioned medium containing His6-tagged BMP8A protein was used for recombinant protein purification by Ni Sepharose excel (GE Healthcare). Purified recombinant His6-tagged BMP8A protein was subjected to Western blotting and the band intensities specific to the mature domain were quantified using the ImageJ program for concentration conversion using a standard curve established by a serial dilution of a commercially available BMP8A mature domain (R&D Systems; 1073-BP). Any effect of the BMP10 prodomain on SMAD signaling was excluded based on the results in Supplementary Fig. 1E and F shown in our previous study (Wu et al. 2017).
Cultures of primary granulosa cells
Immature granulosa cells were isolated from early antral follicles from immature rats primed with diethylstilbestrol for 3 days, whereas mature granulosa cells were isolated from large antral follicles from rats primed with PMSG for 2 days (Hung et al. 2012). In brief, the ovaries were punctured in L-15 Leibovitz medium with a 26 gauge needle to release granulosa cells. The cell suspensions were filtered through 40 μm cell strainer to remove large tissue clumps and oocytes. The purity of granulosa cells was confirmed based on the differential expression of Fshr and Lhcgr, which was determined by real-time PCR relative to Actb (cells before PMSG injection: Fshr, 0.0030; Lhcgr, 0.0022; cells at 48 h after PMSG injection: Fshr, 0.0045; Lhcgr, 0.0358) (Luo et al. 2004). The cells were cultured in McCoy’s 5A supplemented with penicillin, streptomycin, glutamine and androstenedione (10-7 M) and adjusted to appropriate cell number for subsequent experiments (Sun et al. 2010). For transcript quantification and progesterone measurement, 2 × 105 cells/mL were used. For immunoblotting against phosphorylated SMADs, 9 × 105 cells/mL were used.
cDNA preparation and real-time PCR quantification
For preparation of cDNA from granulosa cells, total RNA from each sample was extracted using the TRIzol reagent (Thermo Fisher Scientific) and then reverse-transcribed using the High-capacity cDNA RT kit (Applied Biosystems) with the oligo-dT primer. For subsequent quantitative real-time PCR, the Power SYBR Green Master Mix reagent (Applied Biosystems) was used. The primer pairs for each gene were listed as follows: rat Bmp8a forward, CATGGGGTTTGTGGGCTATC; rat Bmp8a reverse, GACTGAGGATGGCCAAGGAC (accession ID: NM_001109432.1; amplicon: 155 bp). Rat Bmp8b forward, TGTGGCAGAGGATTGGTTCA; rat Bmp8b reverse, CAGGGACTGTACTGCCCACA (accession ID: XM_002729526.5; amplicon: 145 bp). Rat Star forward, AGATGAAGTGCTAAGTAAGGTGGTG; rat Star reverse, CCAGTTCTTCATAGAGTCTGTCCAT (accession ID: NM_031558.3; amplicon: 98 bp). Rat Hsd3b forward, AGACCATCCTAGATGTCAATCTGAA; rat Hsd3b reverse, CAGGATGATCTTCTTGTAGGAGT (accession ID: NM_017265.4; amplicon: 134 bp). Rat Cyp11a1 forward, CCAAGTTCAACCTCATCCTGA; rat Cyp11a1 reverse, GTGTGACTGCAGCCTGCAAT (accession ID: NM_017286.3; amplicon: 127 bp). Rat Id1 forward, TGGACGAACAGCAGGTGAAC; rat Id1 reverse, ATCTCCACCTTGCTCACTTTGC (accession ID: NM_012797.2; amplicon: 112 bp). Rat Smad7 forward, CGATACAAAAACGGGAAGCAA; rat Smad7 reverse, ACACACTGACAACTGAAATGCTGAT (accession ID: NM_030858.1; amplicon: 100 bp). Rat Actb forward, CTCTGTGTGGATTGGTGGCTC; rat Actb reverse, CTGCTTGCTGATCCACATCTG (accession ID: NM_031144.3; amplicon: 70 bp). For Bmp8a and Bmp8b, single products were confirmed by melting curves, agarose gel electrophoresis and sequencing. The products of other genes have been verified previously (Luo et al. 2004).
Immunoblotting
To detect the existence of endogenous BMP8 protein, the ovaries collected from superovulated rats at indicated time points were homogenized in PBS supplemented with protease inhibitor (Roche, 04693132001) using a tissue grinder (Shineteh Inc) on ice and centrifuged at 12,000 g for 30 min, whereas the ovarian cells were lysed in reducing sample buffer directly. The derived supernatants from ovary homogenates were quantified by BCA protein assay (Pierce), a total of 20 μg protein was loaded per lane. To detect phosphorylated SMAD proteins in primary granulosa cells, the cells seeded overnight in 12-wells plates were changed to fresh serum-free medium and treated with indicated ligands for 1 h. The cells were lysed in reducing sample buffer. Protein samples were boiled to denature and analyzed by SDS-PAGE in 10–12% slab gels. Western blotting was performed using goat antibody against BMP8 (1:1000; R&D Systems, AF-1073), rabbit antibody recognizing phosphorylated SMAD1/5/8 (1:1000; Cell Signaling, 9511) or phosphorylated SMAD2/3 (1:1000; Cell Signaling, 9510). Detailed protocols were described previously (Wu et al. 2017).
Progesterone ELISA
To assess the production of progesterone, granulosa cells were pretreated with or without SB431542 or dorsomorphin for 30 min before being incubated with PBS control, FSH, hCG, and/or BMP8 for 48 h as indicated. The supernatant was collected and the amount of progesterone was measured by competitive ELISA. Briefly, 96-well microtiter plates pre-coated with progesterone-BSA (10 ng/well) were incubated with specific anti-progesterone antibody (Sigma, P1922) in the presence of the above supernatant or serially diluted progesterone-containing standards. The plates were then washed and incubated with horseradish peroxidase-conjugated secondary antibody, and the luminescent signals were read by adding the enhanced chemiluminescence substrate. The within-assay coefficient of variation and between-assay coefficient of variation were less than 5%. The sensitivity was less than 100 pg/mL (Vitt et al. 2000).
Characterization of transcriptomic datasets and statistical analyses
Normalized data from each human transcriptomic dataset provided by the submitters were retrieved from the Gene Expression Omnibus (GEO) repository. The group differences between two independent samples were examined by Mann–Whitney test. For other data analyses, Student’s t-test was used to compare differences between two groups, whereas one-way ANOVA followed by Bonferroni post test was used to compare differences among multiple groups. For progesterone assay, three independent experiments were performed and gave similar trends. Data are expressed as means ± s.d. for one of the experiments due to batch variation.
Results
The expression profiles of BMP8 in diverse mammalian tissues
BMP8 consists of a pair of genes, BMP8A and BMP8B. Importantly, the expression profiles of these two genes in diverse mammalian tissues have not yet been clarified. We therefore, as a first step, analyzed the expression profiles of these two BMP8 genes in humans and rodents in a whole tissue spectrum manner. Regarding the BMP8 profiles in humans, we browsed NCBI’s Gene Expression Omnibus (GEO) database and targeted a transcriptomic dataset that contains 42 normal human tissues (GSE14938) (She et al. 2009). By analyzing this dataset, we identified that the transcript level in the ovary of either BMP8A or BMP8B ranked third among 42 human tissue types analyzed (Fig. 1A and B). Specifically, their expression in the human ovary was much higher than in the testis.
Expression profiles of BMP8A and BMP8B in various mammalian tissues. In humans, the relative expression levels of BMP8A (A) and BMP8B (B) were compared across 42 human tissue samples. This data was extracted from a public microarray dataset (GSE14938) and only 20 tissues with the highest abundance levels for expression of the two BMP8 genes are shown. RNAs of each tissue were pooled from multiple donors for a microarray assay and the signal intensity was normalized against a pooled reference comprising about 20 normal adult tissue pools. No data variation was shown (She et al. 2009). In rats, the transcript levels of Bmp8a (C) and Bmp8b (D) in 22 rat tissue samples were evaluated by real-time PCR. Depending on the tissue size, tissues from four to six adult rats were pooled for RNA extraction. The ovaries and testes were collected from four rats. Data were normalized against Actb amounts are expressed as means ± s.d. from three independent quantification replicates.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Expression profiles of BMP8A and BMP8B in various mammalian tissues. In humans, the relative expression levels of BMP8A (A) and BMP8B (B) were compared across 42 human tissue samples. This data was extracted from a public microarray dataset (GSE14938) and only 20 tissues with the highest abundance levels for expression of the two BMP8 genes are shown. RNAs of each tissue were pooled from multiple donors for a microarray assay and the signal intensity was normalized against a pooled reference comprising about 20 normal adult tissue pools. No data variation was shown (She et al. 2009). In rats, the transcript levels of Bmp8a (C) and Bmp8b (D) in 22 rat tissue samples were evaluated by real-time PCR. Depending on the tissue size, tissues from four to six adult rats were pooled for RNA extraction. The ovaries and testes were collected from four rats. Data were normalized against Actb amounts are expressed as means ± s.d. from three independent quantification replicates.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Expression profiles of BMP8A and BMP8B in various mammalian tissues. In humans, the relative expression levels of BMP8A (A) and BMP8B (B) were compared across 42 human tissue samples. This data was extracted from a public microarray dataset (GSE14938) and only 20 tissues with the highest abundance levels for expression of the two BMP8 genes are shown. RNAs of each tissue were pooled from multiple donors for a microarray assay and the signal intensity was normalized against a pooled reference comprising about 20 normal adult tissue pools. No data variation was shown (She et al. 2009). In rats, the transcript levels of Bmp8a (C) and Bmp8b (D) in 22 rat tissue samples were evaluated by real-time PCR. Depending on the tissue size, tissues from four to six adult rats were pooled for RNA extraction. The ovaries and testes were collected from four rats. Data were normalized against Actb amounts are expressed as means ± s.d. from three independent quantification replicates.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
To further evaluate the expression profiles of these two Bmp8 genes in rodents, we performed real-time PCR quantification in 22 different types of rat tissues. Consistently, our results indicated that the transcript level of either Bmp8a or Bmp8b in the rat ovary was also relatively higher than in the testis (Fig. 1C and D), where BMP8 signaling has been reported to be essential for the maintenance of postnatal spermatogenesis in rodents (Zhao & Hogan 1996, Zhao et al. 1998, Wu et al. 2017).
Next, the amounts of BMP8 protein in the rodent gonads were assessed. The BMP8 antibody can only detect total BMP8 protein due to high sequence similarity between BMP8A and BMP8B protein. In mice, mature BMP8 protein can indeed be detected in the mature ovary and testis, with the ovary showing a higher abundance (Fig. 2A). In the rat model, the amounts of BMP8 protein in the pre-pubertal ovaries and the mature ovaries were similar (Fig. 2B). We next investigated the expression profiles of the two Bmp8 genes during folliculogenesis using a superovulated rat model. Quantitative real-time PCR analysis showed that the transcript levels of both Bmp8a and Bmp8b were attenuated after PMSG injection (Fig. 2C and D). Taken together, our findings suggest that the two BMP8 genes are expressed at relatively high levels in the ovary of diverse mammals and that such expression is regulated during folliculogenesis.
Evaluation of the protein amount and the RNA regulatory profile of BMP8 in the rodent ovary. (A and B) Detection of BMP8 protein. The ovaries and the testes from mature mice (8 weeks old) (A) or the ovaries from rats at 26 days and 8 weeks old (B) were homogenized. The derived supernatants (S) and pellets (P) (20 μg protein per lane) were subjected to immunoblotting using the antibody against BMP8. Blots are representative of three independent experiments. (C and D) The ovaries collected from immature rats at the indicated time points after PMSG injection and were used for real-time PCR quantification of the transcripts of Bmp8a (C) and Bmp8b (D). The ovaries from three to five rats were collected and pooled for each time point. Data were normalized against Actb amounts and are expressed as means ± s.d. from three independent quantification replicates. **P < 0.01; ***P < 0.001; one-way ANOVA, followed by Bonferroni post test.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Evaluation of the protein amount and the RNA regulatory profile of BMP8 in the rodent ovary. (A and B) Detection of BMP8 protein. The ovaries and the testes from mature mice (8 weeks old) (A) or the ovaries from rats at 26 days and 8 weeks old (B) were homogenized. The derived supernatants (S) and pellets (P) (20 μg protein per lane) were subjected to immunoblotting using the antibody against BMP8. Blots are representative of three independent experiments. (C and D) The ovaries collected from immature rats at the indicated time points after PMSG injection and were used for real-time PCR quantification of the transcripts of Bmp8a (C) and Bmp8b (D). The ovaries from three to five rats were collected and pooled for each time point. Data were normalized against Actb amounts and are expressed as means ± s.d. from three independent quantification replicates. **P < 0.01; ***P < 0.001; one-way ANOVA, followed by Bonferroni post test.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Evaluation of the protein amount and the RNA regulatory profile of BMP8 in the rodent ovary. (A and B) Detection of BMP8 protein. The ovaries and the testes from mature mice (8 weeks old) (A) or the ovaries from rats at 26 days and 8 weeks old (B) were homogenized. The derived supernatants (S) and pellets (P) (20 μg protein per lane) were subjected to immunoblotting using the antibody against BMP8. Blots are representative of three independent experiments. (C and D) The ovaries collected from immature rats at the indicated time points after PMSG injection and were used for real-time PCR quantification of the transcripts of Bmp8a (C) and Bmp8b (D). The ovaries from three to five rats were collected and pooled for each time point. Data were normalized against Actb amounts and are expressed as means ± s.d. from three independent quantification replicates. **P < 0.01; ***P < 0.001; one-way ANOVA, followed by Bonferroni post test.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Expression and regulation of the BMP8 genes in granulosa cells
As mentioned in the introduction, in this study we are particularly interested in characterizing the molecular candidates for their involvement in the prevention of precocious luteinization of granulosa cells, an event known to be guided by SMAD1/5/8 signaling (Knight & Glister 2006). By exploring public transcriptomic databases derived from human granulosa cells, we intriguingly found that BMP8 is potentially of importance to this event. For example, when analyzing a RNA-sequencing transcriptomic dataset derived from human mature granulosa cells (n = 12; GSE62093) (Yu et al. 2015), BMP8B exhibited the highest expressional abundance among the searchable ligands capable of activating SMAD1/5/8 signaling (Fig. 3A); unfortunately, BMP8A was not included in this dataset. In addition, as compared with young individuals (age 26 ± 2.2 years; n = 6), a significant decrease in the expression of BMP8B, as well as that of AMH, GDF7, BMP6 and BMP4, was observed in older women (age 40 ± 2.3 years; n = 6) (Fig. 3A); this suggests that there might be a decline in SMAD1/5/8 signaling in the mature granulosa cells as the age of woman increases.
Abundance and regulation of BMP8 in mammalian granulosa cells. (A) The transcript abundances of BMP-related ligands were analyzed in a RNA-sequencing transcriptomic dataset derived from mature granulosa cells harvested from older and young women (GSE62093 (Yu et al. 2015)). The normalized RNA-seq data (RPKM, reads per kilobase per million) is represented by a box-plot. Data are means ± s.d. *P < 0.05; **P < 0.01; Mann–Whitney test. (B) Changes in the expression of BMP-related ligands were compared in granulosa cells from normal (matched control subjects) and PCOS patients (GSE34526 (Kaur et al. 2012)). Data are means ± s.d. *P < 0.05; Mann–Whitney test. (C) Isolated granulosa cells collected from superovulated rats at indicated time points were lysed (three individual rats per group). The derived supernatant fractions (20 μg protein per lane) were subjected to immunoblotting using the antibody against BMP8 (left panel). The blots were assessed by densitometer using β-actin as a loading control. Data are shown as means ± s.d. (right panel). *P < 0.05; **P < 0.01. Data are means ± s.d. *P < 0.05; Student’s t test. n = 3 individual animals.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Abundance and regulation of BMP8 in mammalian granulosa cells. (A) The transcript abundances of BMP-related ligands were analyzed in a RNA-sequencing transcriptomic dataset derived from mature granulosa cells harvested from older and young women (GSE62093 (Yu et al. 2015)). The normalized RNA-seq data (RPKM, reads per kilobase per million) is represented by a box-plot. Data are means ± s.d. *P < 0.05; **P < 0.01; Mann–Whitney test. (B) Changes in the expression of BMP-related ligands were compared in granulosa cells from normal (matched control subjects) and PCOS patients (GSE34526 (Kaur et al. 2012)). Data are means ± s.d. *P < 0.05; Mann–Whitney test. (C) Isolated granulosa cells collected from superovulated rats at indicated time points were lysed (three individual rats per group). The derived supernatant fractions (20 μg protein per lane) were subjected to immunoblotting using the antibody against BMP8 (left panel). The blots were assessed by densitometer using β-actin as a loading control. Data are shown as means ± s.d. (right panel). *P < 0.05; **P < 0.01. Data are means ± s.d. *P < 0.05; Student’s t test. n = 3 individual animals.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Abundance and regulation of BMP8 in mammalian granulosa cells. (A) The transcript abundances of BMP-related ligands were analyzed in a RNA-sequencing transcriptomic dataset derived from mature granulosa cells harvested from older and young women (GSE62093 (Yu et al. 2015)). The normalized RNA-seq data (RPKM, reads per kilobase per million) is represented by a box-plot. Data are means ± s.d. *P < 0.05; **P < 0.01; Mann–Whitney test. (B) Changes in the expression of BMP-related ligands were compared in granulosa cells from normal (matched control subjects) and PCOS patients (GSE34526 (Kaur et al. 2012)). Data are means ± s.d. *P < 0.05; Mann–Whitney test. (C) Isolated granulosa cells collected from superovulated rats at indicated time points were lysed (three individual rats per group). The derived supernatant fractions (20 μg protein per lane) were subjected to immunoblotting using the antibody against BMP8 (left panel). The blots were assessed by densitometer using β-actin as a loading control. Data are shown as means ± s.d. (right panel). *P < 0.05; **P < 0.01. Data are means ± s.d. *P < 0.05; Student’s t test. n = 3 individual animals.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Moreover, using a transcriptomic dataset of granulosa cells collected from PCOS women (age less than 35 years; n = 7) and matched controls (age less than 35 years; n = 3) (GSE34526) (Kaur et al. 2012), we also observed that, among all searchable ligands capable of activating SMAD1/5/8 signaling, only the expression of BMP8B showed a significant suppression in women with PCOS (Fig. 3B). Taken together, the above findings derived from re-analysis of human transcriptomic data suggest that BMP8 transcripts are abundant in granulosa cells, where their levels correlated with the decline in ovarian function in these selected subjects. However, the functions of BMP8 protein in granulosa cells have never been explored.
Therefore, to investigate the above questions, we initially used a rat granulosa model based on the fact that high expression levels of the BMP8 genes in the mammalian ovaries seem to be common across species (Figs 1 and 2). Firstly, the protein amounts of BMP8 were evaluated in granulosa cells isolated from immature follicles (before PMSG injection) and mature follicles (48 h after PMSG injection) of superovulated rats (Fig. 3C). Consistent with the mRNA profiles assessed during the ovarian cycle (Fig. 2C and D), Western blotting results indicated that the protein amounts of BMP8 in granulosa cells encountered a significant drop after maturation, suggesting that BMP8 signaling will be reduced upon initiation of ovulation and luteinization.
BMP8 activates the SMAD1/5/8 and SMAD2/3 pathways simultaneously in rat granulosa cells
We previously have demonstrated that BMP8 can activate the SMAD1/5/8 and SMAD2/3 pathways simultaneously via different receptor complexes in spermatogonia (Wu et al. 2017). Here we further characterized BMP8 signaling in granulosa cells. Human BMP8A and BMP8B differ in only two amino acids in their mature domain and have been shown to activate similar downstream signaling (Zhao & Hogan 1996, Wu et al. 2017). Therefore, we chose only human BMP8A for subsequent experiments. Indeed, we found that the purified recombinant His6-tagged BMP8A protein not only induced the phosphorylation of SMAD1/5/8 but also the phosphorylation of SMAD2/3 in both immature (Fig. 4A and B) and mature rat granulosa cells (Fig. 4C and D). In addition to the detection of phosphorylated SMADs, induction of SMAD downstream genes was evaluated. Besides, bioactivities of pure human BMP8A protein obtained from a commercial source and the conditioned medium containing non-tagged recombinant BMP8A protein generated from our designed construct were also compared. As compared with the commercial BMP8A protein that showed negligible effect, both non-tagged and His6-tagged recombinant BMP8A proteins produced from transfected 293T cells can significantly induce the expression of Id1, an early response gene of SMAD1/5/8 signaling, and the expression of Smad7, an early response gene of SMAD2/3 signaling, in treated immature rat granulosa cells (Fig. 4E and F); this may further exclude the potential effect of His6 tag on affecting BMP8 bioactivity. Thus, we used purified His6-tagged recombinant BMP8A proteins, hereafter referred to simply as BMP8A, for subsequent treatments.
BMP8A activates both the SMAD1/5/8 and SMAD2/3 pathways in rat granulosa cells. Granulosa cells from immature rats primed with PMSG for 0 h (A and B) or 48 h (C and D) were treated with PBS control, BMP4 (1 nM), activin A (1 nM), or His6-tagged BMP8A (1 or 3 nM) for 1 h. The cell lysates were then subjected to immunoblotting using antibody against phosphorylated SMAD1/5/8 (A and C) or phosphorylated SMAD2/3 (B and D). The signals of phosphorylated-SMAD were quantified and then normalized against β-actin signals. Real-time qPCR analysis of the transcripts of Id1 (E) and Smad7 (F), the SMAD target genes, in immature rat granulosa cells treated with indicated proteins for 1 h. Data are expressed as means ± s.e.m. n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001 compared to no-treatment control; one-way ANOVA, followed by Bonferroni post test. n = 3 independent biological replicates.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
BMP8A activates both the SMAD1/5/8 and SMAD2/3 pathways in rat granulosa cells. Granulosa cells from immature rats primed with PMSG for 0 h (A and B) or 48 h (C and D) were treated with PBS control, BMP4 (1 nM), activin A (1 nM), or His6-tagged BMP8A (1 or 3 nM) for 1 h. The cell lysates were then subjected to immunoblotting using antibody against phosphorylated SMAD1/5/8 (A and C) or phosphorylated SMAD2/3 (B and D). The signals of phosphorylated-SMAD were quantified and then normalized against β-actin signals. Real-time qPCR analysis of the transcripts of Id1 (E) and Smad7 (F), the SMAD target genes, in immature rat granulosa cells treated with indicated proteins for 1 h. Data are expressed as means ± s.e.m. n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001 compared to no-treatment control; one-way ANOVA, followed by Bonferroni post test. n = 3 independent biological replicates.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
BMP8A activates both the SMAD1/5/8 and SMAD2/3 pathways in rat granulosa cells. Granulosa cells from immature rats primed with PMSG for 0 h (A and B) or 48 h (C and D) were treated with PBS control, BMP4 (1 nM), activin A (1 nM), or His6-tagged BMP8A (1 or 3 nM) for 1 h. The cell lysates were then subjected to immunoblotting using antibody against phosphorylated SMAD1/5/8 (A and C) or phosphorylated SMAD2/3 (B and D). The signals of phosphorylated-SMAD were quantified and then normalized against β-actin signals. Real-time qPCR analysis of the transcripts of Id1 (E) and Smad7 (F), the SMAD target genes, in immature rat granulosa cells treated with indicated proteins for 1 h. Data are expressed as means ± s.e.m. n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001 compared to no-treatment control; one-way ANOVA, followed by Bonferroni post test. n = 3 independent biological replicates.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Our previous study has characterized the possible combinations of receptor complexes for BMP8 protein using overexpression and knockdown strategies in 293T cells (Wu et al. 2017). Specifically, in 293T cells, BMP8 can activate SMAD1/5/8 signaling potentially through receptor complexes formed by the type I receptor ALK3 or ALK6 and the type II receptor BMPR2 or ACVR2A, whereas it is also able to activate SMAD2/3 signaling potentially through receptor complexes formed by the type I receptor ALK4 or ALK5 and the type II receptor ACVR2A, ACVR2B or TGFBR2. Based on the above findings, the expression levels of these receptor candidate genes were quantified in granulosa cells in order to evaluate which can form the possible receptor complexes for BMP8. The results indicated that Alk3, Alk5 and Alk6, which are the type I receptor genes, and Acvr2a and Bmpr2, which are the type II receptor genes, are stably expressed and relatively abundant in both immature (Fig. 5A) and mature granulosa cells (Fig. 5B). In granulosa cells, promiscuous interactions between the type I and the type II receptors have been found and this can lead to crosstalk of SMAD pathways. For example, GDF9 can mediate SMAD2/3 signaling via the BMPR2/ALK5 complex (Mazerbourg et al. 2004). Taken together, these findings suggest that BMP8 can act on granulosa cells as an autocrine factor. It activates SMAD1/5/8 signaling and SMAD2/3 signaling simultaneously through different receptor complexes potentially formed by the type I receptor ALK3, ALK5 or ALK6 and the type II receptor BMPR2 or ACVR2A (Fig. 5C). However, the exact receptor combination responsible for BMP8-mediated SMAD signaling needs more studies.
The transcript levels of the type I and type II receptors linked to BMP8 signaling. The transcript levels of various type I receptors (Alk3, Alk4, Alk5 and Alk6) and various type II receptors (Acvr2a, Acvr2b, Bmpr2 and Tgfbr2) were evaluated by real-time PCR in granulosa cells from immature rats primed with PMSG for 0 h (A) or 48 h (B). Data were normalized against Actb amounts and are expressed as means ± s.e.m. n = 3 biological replicates. (C) Proposed model of BMP8-mediated signal transduction in rat granulosa cells.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
The transcript levels of the type I and type II receptors linked to BMP8 signaling. The transcript levels of various type I receptors (Alk3, Alk4, Alk5 and Alk6) and various type II receptors (Acvr2a, Acvr2b, Bmpr2 and Tgfbr2) were evaluated by real-time PCR in granulosa cells from immature rats primed with PMSG for 0 h (A) or 48 h (B). Data were normalized against Actb amounts and are expressed as means ± s.e.m. n = 3 biological replicates. (C) Proposed model of BMP8-mediated signal transduction in rat granulosa cells.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
The transcript levels of the type I and type II receptors linked to BMP8 signaling. The transcript levels of various type I receptors (Alk3, Alk4, Alk5 and Alk6) and various type II receptors (Acvr2a, Acvr2b, Bmpr2 and Tgfbr2) were evaluated by real-time PCR in granulosa cells from immature rats primed with PMSG for 0 h (A) or 48 h (B). Data were normalized against Actb amounts and are expressed as means ± s.e.m. n = 3 biological replicates. (C) Proposed model of BMP8-mediated signal transduction in rat granulosa cells.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
BMP8 inhibits gonadotropin-induced progesterone production in granulosa cells via the SMAD1/5/8 pathway
BMPs are known to inhibit premature luteinization of granulosa cells by suppressing gonadotropin-driven progesterone synthesis (Knight & Glister 2006, Shimasaki et al. 1999). We here demonstrated that, among various BMP molecules, BMP8 is abundant in granulosa cells (Fig. 3). Therefore, whether and how BMP8 affects progesterone production in granulosa cells needed to be further characterized. Similar to the BMP4 control, we were able to demonstrate that addition of BMP8A significantly dampened gonadotropin-induced progesterone production in a dose-dependent manner in both immature (Fig. 6A) and mature (Fig. 6B) rat granulosa cells.
BMP8A suppresses gonadotropin-induced progesterone synthesis in rat granulosa cells. Granulosa cells from immature rats primed with PMSG for 0 h (A) or 48 h (B) were treated with FSH (0.1 nM) or hCG (0.1 nM) in combination with either BMP4 (a positive control; 1 nM) or BMP8A (1 nM or 3 nM). The levels of progesterone in the medium were determined after 48 h incubation. Data are expressed as means ± s.d. **P < 0.01; ***P < 0.001; one-way ANOVA, followed by Bonferroni post test. Three experiments were performed and gave similar results.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
BMP8A suppresses gonadotropin-induced progesterone synthesis in rat granulosa cells. Granulosa cells from immature rats primed with PMSG for 0 h (A) or 48 h (B) were treated with FSH (0.1 nM) or hCG (0.1 nM) in combination with either BMP4 (a positive control; 1 nM) or BMP8A (1 nM or 3 nM). The levels of progesterone in the medium were determined after 48 h incubation. Data are expressed as means ± s.d. **P < 0.01; ***P < 0.001; one-way ANOVA, followed by Bonferroni post test. Three experiments were performed and gave similar results.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
BMP8A suppresses gonadotropin-induced progesterone synthesis in rat granulosa cells. Granulosa cells from immature rats primed with PMSG for 0 h (A) or 48 h (B) were treated with FSH (0.1 nM) or hCG (0.1 nM) in combination with either BMP4 (a positive control; 1 nM) or BMP8A (1 nM or 3 nM). The levels of progesterone in the medium were determined after 48 h incubation. Data are expressed as means ± s.d. **P < 0.01; ***P < 0.001; one-way ANOVA, followed by Bonferroni post test. Three experiments were performed and gave similar results.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Interestingly, in both immature (Fig. 7A) and mature (Fig. 7B) granulosa cells, the suppression of gonadotropin-driven progesterone production by BMP8A can be reversed by adding dorsomorphin, an inhibitor of the SMAD1/5/8 pathway, but not by adding SB431542, an inhibitor of the SMAD2/3 pathway, while in the controls these SMAD pathway inhibitors had no effect on gonadotropin-driven progesterone production themselves. Furthermore, changes in the transcriptional levels of steroidogenesis-related genes, including Star, Cyp11a1 and Hsd3b, were evaluated. In immature granulosa cells, it seems that only the FSH-elevated level of Cyp11a1, but not that of Star or Hsd3b, was significantly reduced by BMP8A (Fig. 7C, D and E). In contrast, in mature granulosa cells, the levels of Star, Cyp11a1 and Hsd3b induced by hCG can all be suppressed by BMP8A (Fig. 7F, G and H). Specifically, the above suppressive effects of BMP8A can be reversed by the presence of dorsomorphin but not of SB431542 (Fig. 7C, D, E, F, G and H), suggesting that the suppression of gonadotropin-driven progesterone synthesis and steroidogenesis-related genes by BMP8 is mediated via the SMAD1/5/8 pathway only.
Effects of BMP8A on gonadotropin-induced steroidogenesis-related genes in rat granulosa cells. Granulosa cells from immature rats primed with PMSG for 0 h (A, C, D and E) or 48 h (B, F, G and H) were treated with FSH (0.1 nM) or hCG (0.1 nM) in combination with dorsomorphin (D; 10 μM), SB431542 (S; 10 μM) and/or BMP8A (3 nM) as indicated. (A and B) The levels of progesterone in the medium were determined after 48 h incubation. Data are expressed as means ± s.d. Three experiments were performed and gave similar results. Changes in the transcripts of steroidogenesis-related genes, including Star (C and F), Cyp11a1 (D and G) and Hsd3b (E and H), were then evaluated by real-time PCR. Data were normalized against Actb amounts and are expressed as means ± s.e.m. ***P < 0.001; one-way ANOVA, followed by Bonferroni post test. n = 3 biological replicates.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Effects of BMP8A on gonadotropin-induced steroidogenesis-related genes in rat granulosa cells. Granulosa cells from immature rats primed with PMSG for 0 h (A, C, D and E) or 48 h (B, F, G and H) were treated with FSH (0.1 nM) or hCG (0.1 nM) in combination with dorsomorphin (D; 10 μM), SB431542 (S; 10 μM) and/or BMP8A (3 nM) as indicated. (A and B) The levels of progesterone in the medium were determined after 48 h incubation. Data are expressed as means ± s.d. Three experiments were performed and gave similar results. Changes in the transcripts of steroidogenesis-related genes, including Star (C and F), Cyp11a1 (D and G) and Hsd3b (E and H), were then evaluated by real-time PCR. Data were normalized against Actb amounts and are expressed as means ± s.e.m. ***P < 0.001; one-way ANOVA, followed by Bonferroni post test. n = 3 biological replicates.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Effects of BMP8A on gonadotropin-induced steroidogenesis-related genes in rat granulosa cells. Granulosa cells from immature rats primed with PMSG for 0 h (A, C, D and E) or 48 h (B, F, G and H) were treated with FSH (0.1 nM) or hCG (0.1 nM) in combination with dorsomorphin (D; 10 μM), SB431542 (S; 10 μM) and/or BMP8A (3 nM) as indicated. (A and B) The levels of progesterone in the medium were determined after 48 h incubation. Data are expressed as means ± s.d. Three experiments were performed and gave similar results. Changes in the transcripts of steroidogenesis-related genes, including Star (C and F), Cyp11a1 (D and G) and Hsd3b (E and H), were then evaluated by real-time PCR. Data were normalized against Actb amounts and are expressed as means ± s.e.m. ***P < 0.001; one-way ANOVA, followed by Bonferroni post test. n = 3 biological replicates.
Citation: Reproduction 159, 3; 10.1530/REP-19-0305
Discussion
Using a model of rat granulosa cells, we demonstrated BMP8 is able to inhibit gonadotropin-stimulated progesterone production as well as the expression of various steroidogenesis-related genes, including Star, Cyp11a1 and Hsd3b, via activation of SMAD1/5/8 signaling. Interestingly, among the ligands capable of activating the SMAD1/5/8 pathway, we found that BMP8 is relatively abundant in granulosa cells, where both the transcript levels and protein amounts of BMP8 decrease in parallel with follicle maturation. Taken together, our findings suggest that BMP8 signaling can counteract the activities of gonadotropins in granulosa cells and thus potentially prevents their premature luteinization. In other words, regulation of BMP8 expression would be critical for proper follicle development and for the timing control of subsequent ovulation. Thus, BMP8 may act as a novel luteinization inhibitor in the rodent ovary.
Although some physiological differences exist between the primate and rodent ovaries, the expression patterns of BMP-related ligands seem to be conserved in terms of evolution across species. For example, it has been shown that the expression profiles of several BMPs, such as BMP2, BMP6, BMP7 and BMP15, in the human ovaries are virtually identical with those in the rodent ovary (Yoshino et al. 2011). Moreover, most of these factors also play similar and critical roles during ovarian follicle development in both primates and rodents (Knight & Glister 2006). Therefore, our findings in rodents create a further interest in characterizing the expression profiles and physiological roles of these two BMP8 genes in the human ovary. Although these have not yet been fully characterized, in this study we have preliminarily found that BMP8 exhibited similar expressional characteristics in the ovaries of humans and rodents. For example, expressions of both the BMP8A and BMP8B genes were enriched in the ovaries of humans and rats (Fig. 1). In terms of granulosa cells, we also identified that BMP8B is likely to be the most abundant candidate among all searchable BMP-related ligands in human mature granulosa cells (Fig. 3A). Women of advanced age or women with PCOS often suffer from premature luteinization (Willis et al. 1998, Wu et al. 2015). Interestingly, our findings also indicated that the level of BMP8B in granulosa cells from these women is relatively low compared with that in granulosa cells from the matched normal controls (Fig. 3B). Taken together, these results provide a hint that BMP8 may also act as a luteinization inhibitor in the human ovary to control the timing of ovulation; if this is true, then the levels of the BMP8 genes in residual granulosa cells may reflect the quality of oocyte retrieved from the same follicle.
It has been reported that deletion of either Bmp8a or Bmp8b in mice can cause defects in postnatal spermatogenesis (Zhao et al. 1996, 1998). Although we found that the ovary exhibits higher levels of transcripts and protein amounts of BMP8 than the testis in rodents (Fig. 1), no phenotypic change regarding the ovary or other female reproductive tracts has been implicated in the above knockout mice. Many BMP-related and TGFβ-related ligands have been shown to be expressed and interact in a complex manner in the ovary (Knight & Glister 2006). Also, both Bmp8a and Bmp8b are expressed in the ovary. Therefore, it is likely that compensation by other ligands limits deleterious effects of functional loss of BMP8A or BMP8B protein in the knockout mouse models.
BMP8 proteins are uniquely marked by an additional 8th cysteine in the mature domains of the TGFβ superfamily that normally contain seven conserved cysteines (Ozkaynak et al. 1992). Therefore, certain folding machinery in specific cell types and/or conditions may be required for the generation of functional BMP8 mature domains. Using prodomain shuffling, we previously have found that BMP10 prodomain may overcome the above issue and can successfully facilitate the production of bioactive BMP8A protein in 293T cells (Wu et al. 2017). In this study we further demonstrated that purified BMP8A protein can activate both the SMAD1/5/8 pathway and the SMAD2/3 pathway in immature and PMSG-primed mature rat granulosa cells (Fig. 4). Of interest, gene depletion models using mice have indicated that both sets of SMAD signaling in granulosa cells are involved in the control of follicle ovulation and luteinization. For example, female mice with granulosa cell-specific deficiency in Smad4 exhibit prematurely luteinized follicles with trapped oocytes and eventually become infertile by 4-6 months of age (Pangas et al. 2006). In contrast to this, although individual disruption of either SMAD1/5/8 signaling or SMAD2/3 signaling in mouse granulosa cells can lead to disorders of multiple ovarian processes or even to development of metastatic granulosa cell tumors, none of each can totally phenocopy the Smad4-conditional knockout mice (Li et al. 2008, Pangas et al. 2008). Taken together, these findings favor the hypothesis that the SMAD1/5/8 and SMAD2/3 pathways may mediate different activities in granulosa cells, while at the same time these two SMAD pathways do also work synergistically in order to control the correct timing of folliculogenesis and luteinization (Pangas 2012). Consistently, our data also indicated that the two SMAD pathways activated by BMP8 in granulosa cells may function in different ways. We showed that blockage of the SMAD1/5/8 pathway, but not blockage of the SMAD2/3 pathway, activated by BMP8A can restore gonadotropin-induced progesterone production in granulosa cells, suggesting that BMP8-driven SMAD2/3 signaling is not involved in the control of steroidogenesis. On the other hand, activation of the SMAD2/3 pathway by either GDF9 or TGFβ has been shown to have pronounced effects on promoting proliferation and survival of granulosa cells (Dorrington et al. 1988, Sasseville et al. 2010). Therefore, it is worthwhile to further explore whether BMP8 protein can affect other activities of granulosa cells, such as proliferation, differentiation and/or their subsequent survival after luteinization, via activation of the SMAD2/3 pathway.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by grants from Ministry of Science and Technology in Taiwan (MOST 103-2314-B-010-006-MY3; MOST 107-2314-B-010-016-MY3).
Author contribution statement
F-J W and C-W L designed the research. F-J W and Y-W W did the research and data analysis. F-J W, Y-W W and C-W L wrote the manuscript. All authors read and approved the final manuscript.
References
Chang HM, Cheng JC, Fang L, Qiu X, Klausen C, Taylor EL & Leung PC 2015 Recombinant BMP4 and BMP7 downregulate PENTRAXIN 3 in human granulosa cells. Journal of Clinical Endocrinology and Metabolism E365–E374. (https://doi.org/10.1210/jc.2014-2496)
Chang HM, Cheng JC, Klausen C & Leung PC 2013 BMP15 suppresses progesterone production by down-regulating StAR via ALK3 in human granulosa cells. Molecular Endocrinology 2093–2104. (https://doi.org/10.1210/me.2013-1233)
Dorrington J, Chuma AV & Bendell JJ 1988 Transforming growth factor beta and follicle-stimulating hormone promote rat granulosa cell proliferation. Endocrinology 353–359. (https://doi.org/10.1210/endo-123-1-353)
Fanchin R, de Ziegler D, Taieb J, Hazout A & Frydman R 1993 Premature elevation of plasma progesterone alters pregnancy rates of in vitro fertilization and embryo transfer. Fertility and Sterility 1090–1094. (https://doi.org/10.1016/s0015-0282(16)55933-0)
Grasa P, Sheikh S, Krzys N, Millar K, Janjua S, Nawaggi P & Williams SA 2016 Dysregulation of follicle development in a mouse model of premature ovarian insufficiency. Reproduction 591–601. (https://doi.org/10.1530/REP-16-0091)
Hung WT, Wu FJ, Wang CJ & Luo CW 2012 DAN (NBL1) specifically antagonizes BMP2 and BMP4 and modulates the actions of GDF9, BMP2, and BMP4 in the rat ovary. Biology of Reproduction 158, 1–158, 9. (https://doi.org/10.1095/biolreprod.111.096172)
Jakimiuk AJ, Weitsman SR, Navab A & Magoffin DA 2001 Luteinizing hormone receptor, steroidogenesis acute regulatory protein, and steroidogenic enzyme messenger ribonucleic acids are overexpressed in thecal and granulosa cells from polycystic ovaries. Journal of Clinical Endocrinology and Metabolism 1318–1323. (https://doi.org/10.1210/jcem.86.3.7318)
Kaur S, Archer KJ, Devi MG, Kriplani A, Strauss JF, 3rd & Singh R 2012 Differential gene expression in granulosa cells from polycystic ovary syndrome patients with and without insulin resistance: identification of susceptibility gene sets through network analysis. Journal of Clinical Endocrinology and Metabolism E2016–E2021. (https://doi.org/10.1210/jc.2011-3441)
Knight PG & Glister C 2006 TGF-beta superfamily members and ovarian follicle development. Reproduction 191–206. (https://doi.org/10.1530/rep.1.01074)
Lejeune B, Degueldre M, Camus M, Vekemans M, Opsomer L & Leroy F 1986 In vitro fertilization and embryo transfer as related to endogenous luteinizing hormone rise or human chorionic gonadotropin administration. Fertility and Sterility 377–383. (https://doi.org/10.1016/s0015-0282(16)49220-4)
Li Q, Pangas SA, Jorgez CJ, Graff JM, Weinstein M & Matzuk MM 2008 Redundant roles of SMAD2 and SMAD3 in ovarian granulosa cells in vivo. Molecular and Cellular Biology 7001–7011. (https://doi.org/10.1128/MCB.00732-08)
Luo CW, Kawamura K, Klein C & Hsueh AJ 2004 Paracrine regulation of ovarian granulosa cell differentiation by stanniocalcin (STC) 1: mediation through specific STC1 receptors. Molecular Endocrinology 2085–2096. (https://doi.org/10.1210/me.2004-0066)
Manzi DL, Dumez S, Scott LB & Nulsen JC 1995 Selective use of leuprolide acetate in women undergoing superovulation with intrauterine insemination results in significant improvement in pregnancy outcome. Fertility and Sterility 866–873. (https://doi.org/10.1016/s0015-0282(16)57495-0)
Mazerbourg S, Klein C, Roh J, Kaivo-Oja N, Mottershead DG, Korchynskyi O, Ritvos O & Hsueh AJ 2004 Growth differentiation factor-9 signaling is mediated by the type I receptor, activin receptor-like kinase 5. Molecular Endocrinology 653–665. (https://doi.org/10.1210/me.2003-0393)
Mazerbourg S, Sangkuhl K, Luo CW, Sudo S, Klein C & Hsueh AJ 2005 Identification of receptors and signaling pathways for orphan bone morphogenetic protein/growth differentiation factor ligands based on genomic analyses. Journal of Biological Chemistry 32122–32132. (https://doi.org/10.1074/jbc.M504629200)
Oktem O & Urman B 2010 Understanding follicle growth in vivo. Human Reproduction 2944–2954. (https://doi.org/10.1093/humrep/deq275)
Ozkaynak E, Schnegelsberg PN, Jin DF, Clifford GM, Warren FD, Drier EA & Oppermann H 1992 Osteogenic protein-2. A new member of the transforming growth factor-beta superfamily expressed early in embryogenesis. Journal of Biological Chemistry 25220–25227.
Pangas SA 2012 Bone morphogenetic protein signaling transcription factor (SMAD) function in granulosa cells. Molecular and Cellular Endocrinology 40–47. (https://doi.org/10.1016/j.mce.2011.06.021)
Pangas SA, Li X, Robertson EJ & Matzuk MM 2006 Premature luteinization and cumulus cell defects in ovarian-specific Smad4 knockout mice. Molecular Endocrinology 1406–1422. (https://doi.org/10.1210/me.2005-0462)
Pangas SA, Li X, Umans L, Zwijsen A, Huylebroeck D, Gutierrez C, Wang D, Martin JF, Jamin SP, Behringer RR et al. 2008 Conditional deletion of Smad1 and Smad5 in somatic cells of male and female gonads leads to metastatic tumor development in mice. Molecular and Cellular Biology 248–257. (https://doi.org/10.1128/MCB.01404-07)
Richards JS, Fitzpatrick SL, Clemens JW, Morris JK, Alliston T & Sirois J 1995 Ovarian cell differentiation: a cascade of multiple hormones, cellular signals, and regulated genes. Recent Progress in Hormone Research 223–254. (https://doi.org/10.1016/b978-0-12-571150-0.50014-7)
Sasseville M, Ritter LJ, Nguyen TM, Liu F, Mottershead DG, Russell DL & Gilchrist RB 2010 Growth differentiation factor 9 signaling requires ERK1/2 activity in mouse granulosa and cumulus cells. Journal of Cell Science 3166–3176. (https://doi.org/10.1242/jcs.063834)
Schoolcraft W, Sinton E, Schlenker T, Huynh D, Hamilton F & Meldrum DR 1991 Lower pregnancy rate with premature luteinization during pituitary suppression with leuprolide acetate. Fertility and Sterility 563–566. (https://doi.org/10.1016/S0015-0282(16)54186-7)
She X, Rohl CA, Castle JC, Kulkarni AV, Johnson JM & Chen R 2009 Definition, conservation and epigenetics of housekeeping and tissue-enriched genes. BMC Genomics 269. (https://doi.org/10.1186/1471-2164-10-269)
Shi J, Yoshino O, Osuga Y, Koga K, Hirota Y, Nose E, Nishii O, Yano T & Taketani Y 2011 Bone morphogenetic protein-2 (BMP-2) increases gene expression of FSH receptor and aromatase and decreases gene expression of LH receptor and StAR in human granulosa cells. American Journal of Reproductive Immunology 421–427. (https://doi.org/10.1111/j.1600-0897.2010.00917.x)
Shi J, Yoshino O, Osuga Y, Nishii O, Yano T & Taketani Y 2010 Bone morphogenetic protein 7 (BMP-7) increases the expression of follicle-stimulating hormone (FSH) receptor in human granulosa cells. Fertility and Sterility 1273–1279. (https://doi.org/10.1016/j.fertnstert.2008.11.014)
Shimasaki S, Zachow RJ, Li D, Kim H, Iemura S, Ueno N, Sampath K, Chang RJ & Erickson GF 1999 A functional bone morphogenetic protein system in the ovary. PNAS 7282–7287. (https://doi.org/10.1073/pnas.96.13.7282)
Silverberg KM, Burns WN, Olive DL, Riehl RM & Schenken RS 1991 Serum progesterone levels predict success of in vitro fertilization/embryo transfer in patients stimulated with leuprolide acetate and human menopausal gonadotropins. Journal of Clinical Endocrinology and Metabolism 797–803. (https://doi.org/10.1210/jcem-73-4-797)
Stanger JD & Yovich JL 1985 Reduced in-vitro fertilization of human oocytes from patients with raised basal luteinizing hormone levels during the follicular phase. British Journal of Obstetrics and Gynaecology 385–393. (https://doi.org/10.1111/j.1471-0528.1985.tb01113.x)
Sun SC, Hsu PJ, Wu FJ, Li SH, Lu CH & Luo CW 2010 Thyrostimulin, but not thyroid-stimulating hormone (TSH), acts as a paracrine regulator to activate the TSH receptor in mammalian ovary. Journal of Biological Chemistry 3758–3765. (https://doi.org/10.1074/jbc.M109.066266)
Vitt UA, Hayashi M, Klein C & Hsueh AJ 2000 Growth differentiation factor-9 stimulates proliferation but suppresses the follicle-stimulating hormone-induced differentiation of cultured granulosa cells from small antral and preovulatory rat follicles. Biology of Reproduction 370–377. (https://doi.org/10.1095/biolreprod62.2.370)
Willis DS, Watson H, Mason HD, Galea R, Brincat M & Franks S 1998 Premature response to luteinizing hormone of granulosa cells from anovulatory women with polycystic ovary syndrome: relevance to mechanism of anovulation. Journal of Clinical Endocrinology and Metabolism 3984–3991. (https://doi.org/10.1210/jcem.83.11.5232)
Wu FJ, Lin TY, Sung LY, Chang WF, Wu PC & Luo CW 2017 BMP8A sustains spermatogenesis by activating both SMAD1/5/8 and SMAD2/3 in spermatogonia. Science Signaling 10 eaal1910. (https://doi.org/10.1126/scisignal.aal1910)
Wu YG, Barad DH, Kushnir VA, Lazzaroni E, Wang Q, Albertini DF & Gleicher N 2015 Aging-related premature luteinization of granulosa cells is avoided by early oocyte retrieval. Journal of Endocrinology 167–180. (https://doi.org/10.1530/JOE-15-0246)
Yoshino O, Shi J, Osuga Y, Harada M, Nishii O, Yano T & Taketani Y 2011 The function of bone morphogenetic proteins in the human ovary. Reproductive Medicine and Biology 1–7. (https://doi.org/10.1007/s12522-010-0072-3)
Yu B, Russanova VR, Gravina S, Hartley S, Mullikin JC, Ignezweski A, Graham J, Segars JH, DeCherney AH & Howard BH 2015 DNA methylome and transcriptome sequencing in human ovarian granulosa cells links age-related changes in gene expression to gene body methylation and 3'-end GC density. Oncotarget 3627–3643. (https://doi.org/10.18632/oncotarget.2875)
Zhao GQ, Deng K, Labosky PA, Liaw L & Hogan BL 1996 The gene encoding bone morphogenetic protein 8B is required for the initiation and maintenance of spermatogenesis in the mouse. Genes and Development 1657–1669. (https://doi.org/10.1101/gad.10.13.1657)
Zhao GQ & Hogan BL 1996 Evidence that mouse Bmp8a (Op2) and Bmp8b are duplicated genes that play a role in spermatogenesis and placental development. Mechanisms of Development 159–168. (https://doi.org/10.1016/0925-4773(96)00543-6)
Zhao GQ, Liaw L & Hogan BL 1998 Bone morphogenetic protein 8A plays a role in the maintenance of spermatogenesis and the integrity of the epididymis. Development 1103–1112.
