Bone morphogenetic protein-15 inhibits follicle-stimulating hormone (FSH) action by suppressing FSH receptor expression.

We have recently reported that oocyte-derived bone morphogenetic protein-15 (BMP-15) can directly modulate follicle-stimulating hormone (FSH) action in rat granulosa cells. Here, we investigate underlying mechanisms of this BMP-15 effect. Treatment with BMP-15 alone exerted no significant effect on the basal expression of mRNAs encoding steroidogenic acute regulatory protein, P450 side chain cleavage enzyme, P450 aromatase, 3beta-hydroxysteroid dehydrogenase, luteinization hormone receptor, and inhibin/activin subunits. However, BMP-15 markedly inhibited the FSH-induced increases in these messages. In striking contrast, BMP-15 did not change the forskolin-induced levels of these transcripts. Thus, the inhibitory effect of BMP-15 on FSH action must be upstream of cAMP signaling. We next examined changes in FSH receptor mRNA expression. Interestingly, BMP-15 severely reduced the levels of FSH receptor mRNA in both basal and FSH-stimulated cells. To determine whether this effect was at the level of FSH function, we investigated the effect of BMP-15 on FSH bioactivity. Consistent with the mRNA data, BMP-15 inhibited the biological response of FSH, but not that of forskolin. Based on these results, we propose that BMP-15 is an important determinant of FSH action through its ability to inhibit FSH receptor expression. Because FSH plays an essential role in follicle growth and development, our findings could have new implications for understanding how oocyte growth factors contribute to folliculogenesis.

FSH 1 is an essential for normal folliculogenesis and female fertility (1,2). In the ovary, FSH interacts with its receptor on the granulosa cells (GCs) to initiate cytodifferentiation and proliferation that ultimately result in the development of pre-ovulatory follicles (3,4). The mechanism of FSH action involves the activation of specific genes in the GCs through the cAMPdependent protein kinase-A signaling pathway (5)(6)(7). Some of the physiologically important genes that are induced by FSH signaling include the P450 aromatase (P450arom), luteinizing hormone receptor (LH-R), steroidogenic acute regulatory protein (StAR), P450 side chain cleavage enzyme (P450scc), 3␤hydroxysteroid dehydrogenase (3␤-HSD), inhibin, and activin (5,8). The physiological importance of FSH action is demonstrated by the fact when restricted, the developing follicles die by apoptosis, and there are no ovulations (9,10). Therefore, it is important to define the mechanisms involved in FSH action.
Recently, attention has been focused on another oocyte growth factor involved in regulating FSH action: the new member of the transforming growth factor-␤ superfamily designated as bone morphogenetic protein-15 (BMP-15) or growth and differentiation factor-9B (GDF-9B) (29 -31). The cloning of the cDNAs in the mouse (29), human (29), and rat (32) has revealed that the primary structure of BMP-15 is most closely related to that of GDF-9 (29,32). To date the only cell type known to express BMP-15 is the mammalian oocyte (29 -31). Regarding function, we have recently reported that BMP-15 stimulates rat GC proliferation and selectively inhibits FSH-induced P 4 , but not E 2 , production (33). This is the only evidence available to date concerning the function of the BMP-15 molecule.
The potential physiological relevance of BMP-15 in vivo is beginning to emerge. A recent genetic study of a naturally occurring mutant in sheep termed Inverdale (FecX I ) has provided evidence for an essential role of BMP-15 in folliculogenesis and fertility (34). In the female homozygous FecX I mutants, follicular development arrests at the primary stage, resulting in infertility. This phenotype has been linked to a defect in the BMP-15 gene. By contrast, the heterozygotes exhibit increased ovulation rate and multiple pregnancies. Therefore, BMP-15 is associated with the mechanisms of infertility and super fertility in a dosage-sensitive manner (34). An important unanswered question is how BMP-15 regulates these physiological functions.
The present study provides the evidence that BMP-15 inhibits major FSH actions that are obligatory for follicle development and ovulation by virtue of its ability to suppress FSH * This work was supported in part by The Lalor Foundation, University of California, San Diego Academic Senate Grant RY 440M, and NICHD/National Institutes of Health through cooperative agreement U54HD12303 as part of Specialized Cooperative Centers Program in Reproduction Research. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Primary Cell Culture-Female Harlan Sprague-Dawley rats (23 days old) were implanted with silastic capsules containing 10 mg of diethylstilbestrol to increase GC number (3). After 4 days of diethylstilbestrol exposure, GCs were collected from the ovaries and cultured in serumfree McCoy's 5a medium supplemented with 2 mM L-glutamine and antibiotics as described previously (35). The animal protocols were approved by the University of California at San Diego Institutional Animal Care and Use Committee.
RNA Extraction and Analysis by Quantitative Competitive RT-PCR-GCs (2 ϫ 10 6 viable cells) were cultured in a six-well plate with 2 ml of McCoy's 5a medium containing one or a combination of the following: 10 ng/ml FSH, 100 ng/ml BMP-15, 10 M forskolin. Onehundred nM androstenedione was added to the culture medium as indicated. After 48 h culture, total RNA was extracted by guanidium acid-isothiocyanate-phenol-chlorform methods using TriZOL® (Life Technologies Inc.), quantified by measuring absorbance at 260 nm, and stored at Ϫ80°C until assay. Oligonucleotides used for RT-PCR were custom-ordered from Life Technologies, Inc. PCR primer pairs were selected from different exons of the corresponding genes to discriminate PCR products that might arise from possible chromosome DNA contaminants. Specifically, they are derived from the cDNA clones at the following nucleotide numbers: 1651-1670 and 1751-1770 for StAR (36) (44). The steady state levels of mRNA encoding StAR, P450scc, P450arom, and FSH-R (referred to "target" mRNA) were analyzed by quantitative competitive RT-PCR. Prior to PCR the internal control DNAs (ϳ340 base pairs) having target-specific primer pairs were generated by PCR as reported previously (45). The extracted RNA (500 ng) was subjected to a RT reaction using a First-Strand cDNA Synthesis System (Life Technologies, Inc.) with random hexamer (2 ng/l), reverse transcriptase (200 units), and deoxynucleotide triphosphate (dNTP; 0.5 mM) at 42°C for 50 min, 70°C for 10 min. The resultant simple strand cDNA was resuspended in 50 l of water for competitive PCR. The linear portion of the relationship between target cDNAs and internal control DNAs was determined individually for all targets. For this, a fixed amount (20 ng) of cDNA derived from GCs treated with FSH was mixed with a series of 10-fold dilutions of the internal control DNA, and the target and the internal control DNAs were coamplified by PCR using a specific primer set for the individual target (Fig. 1). Competitive PCR was performed using MgCl 2 (1.5 mM), dNTP (0.2 mM), and 2.5 units of Platinum Taq DNA polymerase (Life Technologies, Inc.) under the following conditions: 35 cycles of denatur-ation at 94°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 30 s. Aliquots of PCR products were electrophoresed on 2% agarose gels, visualized after ethidium bromide staining, photographed, and scanned. The relative integrated density of each band was digitized by multiplying the absorbance of the surface area (NIH image 1.62). Finally, the ratios of the densitometric readings of the amplified target cDNA and internal control DNA were plotted on the ordinate against the serial 10-fold dilutions of internal control DNA on the abscissa (Fig.  1). After establishing the working ranges in which linear relationship existed, the following concentrations of the internal control DNA were selected: 10 Ϫ9 pM for StAR, 10 Ϫ7 pM for P450scc, 10 Ϫ7 pM for P450arom, 10 Ϫ4 pM for L19, 10 Ϫ9 pM for FSH-R. For competitive PCR reactions, 2 l each of these internal control samples were added to 20 ng of each cDNA sample. The resultant PCR products were quantified by densitometric scanning as described above. Control analysis was performed using L19 ribosomal protein primers as well. For the other targets such as LH-R, inhibin subunits, 3␤-HSD, and 20␣-hydroxysteroid dehydrogenase (20␣-HSD), semiquantitative PCR analysis was performed without internal control under the same condition, and the level of their PCR products were compared with that of L19 target.
Bioactvity Study-GCs (10 5 viable cells) were cultured in a 96-well plate with 200 l of McCoy's 5a medium containing 100 nM androstenedione together with either 10 ng/ml FSH, 100 ng/ml BMP-15, and/or 10 M forskolin. After 48 h, the levels of P 4 and E 2 in the media were measured by radioimmunoassay (33).
Statistical Analysis-All results shown are mean Ϯ S.E. of at least three separate experiments, with triplicate determinations for each treatment. Differences between groups were analyzed for statistical significance using analysis of variance (StatView 5.0 software, Abacus Concept, Inc., Berkely, CA). p values Ͻ 0.01 were accepted as statistically significant.

RESULTS
In previous experiments (33) we found that BMP-15 inhibits FSH-induced P 4 , but not E 2 , production by cultured rat GCs. To test the hypothesis that BMP-15 may be specifically involved in modulating the expression of steroidogenic genes in response to FSH stimulation, we analyzed mRNA levels for StAR, P450scc, and P450arom by quantitative competitive RT-PCR.
As shown in Fig. 2, treating GCs for 48 h with a saturating dose of BMP-15 had no effect on the steady state mRNA levels for StAR, P450scc, P450arom, and a housekeeping gene L19. Treatment with FSH markedly increased the mRNA levels for StAR, P450scc, and P450arom, whereas L19 mRNA levels were unchanged. BMP-15 inhibited FSH-stimulated StAR and P450scc mRNA levels by 60 and 70%, respectively, but had only a weak effect on P450arom mRNA levels in the presence of androstenedione. L19 mRNA levels showed no difference among four different treatments. Taken into consideration that StAR and P450scc are rate-limiting factors of P 4 synthesis in GCs (46), the suppression of these gene expressions seems to be a major cause of the selective suppression of FSH-induced P 4 production.
We next investigated the effect of BMP-15 on forskolin-stimulated GC responses. As shown in Fig. 3, forskolin (10 M) FIG. 1. Linear range analysis for quantitative competitive RT-PCR. The linear ranges for the ratios of coamplified PCR products were determined using cDNA reverse-transcribed from RNA of FSH-treated GCs and competitive internal control DNA for StAR, P450scc, P450arom, and L19 measured at 10-fold serial dilutions. A, PCR products were gel-electrophoresed and stained by ethidium bromide. The concentrations of the internal control DNA in each lane should refer to the number on the abscissa of B. B, the ratios of PCR products (target/internal control; ordinate) are plotted against the internal control DNA (serial 10-fold dilutions; abscissa). stimulated the levels of StAR, P450scc, and P450arom mRNAs similar to that seen with FSH. However, in striking contrast to the FSH results, BMP-15 failed to change the forskolin-induced mRNA levels of these steroidogenic factors. L19 mRNA levels had no change by the forskolin treatment regardless of the presence of BMP-15 among four groups. These results suggest that the inhibitory effect of BMP-15 on FSH-induced StAR and P450scc mRNA expression is mediated by a pre-cAMP signaling event.
The failure of BMP-15 to inhibit FSH-induced P450arom mRNA expression is paradoxical. Given that androgens can up-regulate FSH-induced P450arom gene activity (47,48), we wondered whether the androstenedione in the culture medium could explain this paradox. As seen in Fig. 4, FSH stimulated P450arom mRNA expression in the absence of androstenedione and importantly, this increase was significantly inhibited by BMP-15. As expected (see Fig. 3), the forskolin-induced P450arom expression was not changed by BMP-15 in the absence of androstenedione. The steady state mRNA levels of StAR, P450scc, and L19 were unaffected by androstenedione (data not shown). Thus, it is clear that in the absence of androstenedione, P450arom mRNA induced by FSH was also inhibited by BMP-15 at a site upstream of the cAMP signaling event.
To investigate whether the BMP effect involves changes in FSH-R expression, we measured the mRNA levels for FSH-R by quantitative competitive PCR. First, we established a linear relationship between the expression of target FSH-R and internal control (Fig. 5A) and then selected 10 Ϫ9 pM concentration of internal control to perform the competitive reactions. Control untreated cells expressed FSH-R mRNA spontaneously. Interestingly, the level of FSH-R mRNA in control cells was markedly reduced (80%) following treatment with BMP-15 (Fig. 5B). FSH treatment increased FSH-R mRNA levels ϳ2fold, and this increase was completely abolished by cotreatment with BMP-15, actually decreasing the FSH-R mRNA to that seen with BMP-15 alone. Thus, both basal and FSHinduced FSH-R mRNA expression is negatively regulated by BMP-15.
To determine whether BMP-15 modulates other FSH-dependent functions in GCs, we examined the possible role of BMP-15 in the regulation of LH-R, the inhibin subunits (␣, ␤A, and ␤B), 3␤-HSD, and 20␣-HSD. As seen in Fig. 6, FSH induced marked increases in the levels of mRNA for each of these end points, except for 20␣-HSD. Although BMP-15 alone had no effect on the basal levels of these mRNAs, it totally abolished the stimulatory effects seen with FSH (Fig. 6). BMP-15, either alone or together with FSH, did not change the basal levels of 20␣-HSD mRNA. As illustrated in Fig. 7, forskolin induced the expression of high levels of the mRNAs encoding LH-R, inhibin ␣, ␤A and ␤B subunits, and 3␤-HSD. Consistent with our earlier results with StAR, P450scc and P450arom, cotreatment with BMP-15 failed to alter the forskolin-induced expression of these mRNAs (Fig. 7).
To determine whether this BMP-15 inhibition was at the level of FSH function, we tested the effect of BMP-15 on FSH bioactivity, specifically, P 4 and E 2 production in the GCs. In agreement with our recent study (33), FSH-induced P 4 , but not E 2 , production was suppressed by BMP-15 (Fig. 8). In contrast, the forskolin-induced steroidogenic responses were unaffected by BMP-15 (Fig. 8). These results fit our hypothesis that BMP-15 inhibits FSH action by suppressing functional FSH-R in GCs.

DISCUSSION
In a recent study we showed that BMP-15 stimulates proliferation of rat GCs and negatively regulates FSH-induced P 4 , but not E 2 , production in vitro (33). The objective of the present study was to begin to explore the molecular basis by which BMP-15 modulates FSH action. The major new findings of the present study are as follows. First BMP-15 reduced the steady state levels of mRNA induced by FSH, including StAR, P450scc, 3␤-HSD, LH-R, inhibin/activin subunits, and FSH-R. These observations suggest that BMP-15 can act broadly to inhibit the expression of a large battery of genes induced by FSH. Accordingly, BMP-15 can be considered a negative regulator of the major actions of FSH in the rat ovary. Second, in striking contrast to the FSH action, BMP-15 did not affect the stimulation of this battery of mRNAs in response to forskolin. Therefore, one can conclude that BMP-15 exerts its inhibitory effect on FSH action upstream of cAMP signaling. Third, BMP-15 markedly decreased both the basal and FSH-induced increases in FSH-R mRNA. And fourth, FSH-induced, but not forskolin-induced, steroidogenesis was suppressed by BMP-15, a finding that is consistent with the decrease in FSH-R message. Based on these findings, we propose that the negative regulation of FSH-R expression is the primary cause of BMP-15 inhibition of FSH action. It is important to note, however, that because we did not measure FSH protein levels directly, we cannot exclude the possibility that BMP-15 might exert its effects by a mechanism other than reducing the number of FSH-R.
The finding that BMP-15 is a negative regulator of FSH-R expression together with our earlier report that BMP-15 stimulates DNA synthesis (33) establishes the concept that oocytederived BMP-15 is functionally involved in regulating two major aspects of GC development, namely stimulating proliferation and inhibiting FSH-dependent cytodifferentiation. This suggests a role for BMP-15 in coordinating GC proliferation with FSH-dependent cytodifferentiation during follicle growth and development. Our observation that FSH stimulates the expression of FSH-R mRNA is consistent with the previous reports demonstrating that FSH itself increases FSH-R gene transcription and cell surface FSH receptors (49,50). Our results clearly demonstrated that BMP-15 completely blocks this FSH effect. Therefore, we propose that BMP-15 is a determinant of FSH action by exerting inhibitory effects on FSH-R expression.
It was found that BMP-15 does not inhibit the amplifying effect of androstenedione on FSH-stimulated P450arom mRNA expression. But, it did suppress the ability of FSH to induce P450arom in the absence of androstenedione. This is consistent with the failure of BMP-15 to inhibit FSH-induced E 2 production in the presence of added androstenedione. The potential significance of this finding is unknown. In this regard, however, a fundamental principle in ovarian physiology is that FSH stimulates the expression of P450arom and E 2 production in the dominant follicle that contains a high concentration of androstenedione in the follicular fluid. Therefore, the finding that BMP-15 does not inhibit FSH-induced P450arom mRNA and E 2 production in the presence of androstenedione may be physiologically relevant.
Genetic and physiological studies in Inverdale (FecX I ) sheep have identified a point mutation in the BMP-15 gene that has profound effects on follicle development and ovulation quota (34). Homozygous FecX I mutant females are infertile because follicle growth is arrested at the primary preantral stage. Consequently there are no ovulations in these animals. This provides compelling evidence for a requirement of bioactive BMP-15 in follicle cell proliferation and differentiation in sheep (51). Surprisingly, the heterozygous FecX I mutants exhibit increased ovulation rate. In this connection, several interesting abnormal features of the ovaries of the heterozygotes have been identified: (i) there are more healthy estrogenic follicles, ii) the number of GCs in these developing follicles is significantly smaller, iii) these GCs have a higher mean LH responsiveness at smaller follicle stages, and iv) the corpora lutea are smaller (52). The fact that the plasma FSH and LH levels in the heterozygotes are normal would imply that the mechanisms responsible for these unusual features in the heterozygote reside in the ovary. Our observation that BMP-15 inhibits FSH-R expression could explain the cause of the abnormal phenotype of the Inverdale ewe. In the heterozygotes, we propose that reduced levels of BMP-15 result in higher levels of FSH-R in the GCs, which in turn lead to more developing healthy estrogenic follicles with more LH-R. The end result of this sequence of events would be increased ovulation rate. At the opposite extreme, follicle development in the homozygotes is arrested at the primary follicle (FSH-independent) stage. In these animals we propose that the absence of BMP-15 results in the cessation of GC proliferation, which in turn leads to arrested follicle development.
In summary, the conclusion emerging from these findings is that BMP-15 can inhibit some of the most important events in the process of FSH-dependent GC cytodifferentiation, including the expression of the FSH-R itself. Consequently the oocyte can play an important role in determining GC proliferation and FSH sensitivity in developing follicles.