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

In addition to pituitary gonadotropins and paracrine factors, ovarian follicle development is also modulated by oocyte factors capable of stimulating granulosa cell proliferation but suppressing their differentiation. The nature of these oocyte factors is unclear. Because growth differentiation factor-9 (GDF-9) enhanced preantral follicle growth and was detected in the oocytes of early antral and preovulatory follicles, we hypothesized that this oocyte hormone could regulate the proliferation and differentiation of granulosa cells from these advanced follicles. Treatment with recombinant GDF-9, but not FSH, stimulated thymidine incorporation into cultured granulosa cells from both early antral and preovulatory follicles, accompanied by increases in granulosa cell number. Although GDF-9 treatment alone stimulated basal steroidogenesis in granulosa cells, cotreatment with GDF-9 suppressed FSH-stimulated progesterone and estradiol production. In addition, GDF-9 cotreatment attentuated FSH-induced LH receptor formation. The inhibitory effects of GDF-9 on FSH-induced granulosa cell differentiation were accompanied by decreases in the FSH-induced cAMP production. These data suggested that GDF-9 is a proliferation factor for granulosa cells from early antral and preovulatory follicles but suppresses FSH-induced differentiation of the same cells. Thus, oocyte-derived GDF-9 could account, at least partially, for the oocyte factor(s) previously reported to control cumulus and granulosa cell differentiation.     

Paracrine actions of growth differentiation factor-9 in the mammalian ovary.

Although the transforming growth factor-beta (TGF-beta) superfamily is the largest family of secreted growth factors, surprisingly few downstream target genes in their signaling pathways have been identified. Likewise, the identities of oocyte-derived secreted factors, which regulate important oocyte-somatic cell interactions, remain largely unknown. For example, oocytes are known to secrete paracrine growth factor(s) which are necessary for cumulus expansion, induction of hyaluronic acid synthesis, and suppression of LH receptor (LHR) mRNA synthesis. Our previous studies demonstrated that absence of the TGF-beta family member, growth differentiation factor-9 (GDF-9), blocks ovarian folliculogenesis at the primary follicle stage leading to infertility. In the present study, we demonstrate that mouse GDF-9 protein is expressed in all oocytes beginning at the type 3a follicle stage including antral follicles. To explore the biological functions of GDF-9 in the later stages of folliculogenesis and cumulus expansion, we produced mature, glycosylated, recombinant mouse GDF-9 using a Chinese hamster ovary cell expression system. A granulosa cell culture system was established to determine the role of GDF-9 in the regulation of several key ovarian gene products using semiquantitative RT-PCR. We find that recombinant GDF-9 induces hyaluronan synthase 2 (HAS2), cyclooxygenase 2 (COX-2), and steroidogenic acute regulator protein (StAR) mRNA synthesis but suppresses urokinase plasminogen activator (uPA) and LHR mRNA synthesis. Consistent with the induction of StAR mRNA by GDF-9, recombinant GDF-9 increases granulosa cell progesterone synthesis in the absence of FSH. Since induction of HAS2 and suppression of the protease uPA in cumulus cells are key events in the production of the hyaluronic acid-rich extracellular matrix which is produced during cumulus expansion, we determined whether GDF-9 could mimic this process. Using oocytectomized cumulus cell-oocyte complexes, we show that recombinant GDF-9 induces cumulus expansion in vitro. These studies demonstrate that GDF-9 can bind to receptors on granulosa cells to regulate the expression of a number of gene products. Thus, in addition to playing a critical function as a growth and differentiation factor during early folliculogenesis, GDF-9 functions as an oocyte-secreted paracrine factor to regulate several key granulosa cell enzymes involved in cumulus expansion and maintenance of an optimal oocyte microenvironment, processes which are essential for normal ovulation, fertilization, and female reproduction.     

PTX3 plays a key role in the organization of the cumulus oophorus extracellular matrix and in in vivo fertilization

PTX3 is a prototypic long pentraxin that plays a non-redundant role in innate immunity against selected pathogens and in female fertility. Here, we report that the infertility of Ptx3(-/-) mice is associated with severe abnormalities of the cumulus oophorus and failure of in vivo, but not in vitro, oocyte fertilization. PTX3 is produced by mouse cumulus cells during cumulus expansion and localizes in the matrix. PTX3 is expressed in the human cumulus oophorus as well. Cumuli from Ptx3(-/-) mice synthesize normal amounts of hyaluronan (HA), but are unable to organize it in a stable matrix. Exogenous PTX3 restores a normal cumulus phenotype. Incorporation in the matrix of inter-alpha-trypsin inhibitor is normal in Ptx3(-/-) cumuli. PTX3 does not interact directly with HA, but it binds the cumulus matrix hyaladherin tumor necrosis factor alpha-induced protein 6 (TNFAIP6, also known as TSG6) and thereby may form multimolecular complexes that can cross-link HA chains. Thus, PTX3 is a structural constituent of the cumulus oophorus extracellular matrix essential for female fertility.     

Oocyte control of metabolic cooperativity between oocytes and companion granulosa cells: energy metabolism

Intercellular communication between oocytes and granulosa cells is essential for normal follicular differentiation and oocyte development. Subtraction hybridization was used to identify genes more highly expressed in cumulus cells than in mural granulosa cells of mouse antral follicles. This screen identified six genes involved in glycolysis: Eno1, Pkm2, Tpi, Aldoa, Ldh1, and Pfkp. When oocytes were microsurgically removed from cumulus cell-oocyte complexes, the isolated cumulus cells exhibited decreased expression levels of genes encoding glycolytic enzymes, glycolysis and activity of the tricarboxylic acid (TCA) cycle. These decreases were prevented by culturing the cumulus cells with paracrine factors secreted by fully grown oocytes. Paracrine factors from fully grown oocytes exhibited greater ability than those from growing oocytes to promote expression of genes encoding glycolytic enzymes and glycolysis in the granulosa cells of preantral follicles. However, neither fully grown nor growing oocytes secreted paracrine factors affecting activity of the TCA cycle. These results indicate that oocytes regulate glycolysis and the TCA cycle in granulosa cells in a manner specific to the population of granulosa cells and to the stage of growth and development of the oocyte. Oocytes control glycolysis in granulosa cells by regulating expression levels of genes encoding glycolytic enzymes. Therefore, mouse oocytes control the intercellular metabolic cooperativity between cumulus cells and oocytes needed for energy production by granulosa cells and required for oocyte and follicular development.     

Transient invasive migration in mouse cumulus oocyte complexes induced at ovulation by luteinizing hormone

Ovulation, the release of the oocyte from the ovarian follicle, is initiated by the luteinizing hormone surge. It is clear that highly controlled degradation of the follicle and ovarian wall is required for passage of the oocyte and accompanying cumulus cells from the follicle, but the mechanism has not yet been elucidated. Here we show that cumulus oocyte complexes (COCs) adopt transient adhesive, migratory, and matrix-invading capacities at the time of ovulation. We characterized cell adhesion, migration, and invasion in preovulatory and postovulatory mouse COCs collected over a time course post-human chorionic gonadotropin (hCG) administration. Adhesion of dispersed cumulus cells and intact COCs to extracellular matrix proteins present in the ovarian wall (collagens, laminin, and fibronectin) increased significantly after hCG treatment and declined immediately after ovulation. Cumulus cell migration was low in unexpanded, equine chorionic gonadotropin-only treated COCs, but increased 4, 8, and 10 h post-hCG, reaching a peak at 12 h post-hCG that coincided with ovulation. The ability of cumulus cells to migrate was rapidly diminished in COCs isolated from the oviduct within 2 h postovulation. Cell migration was cumulus cell specific and was not observed in granulosa cells. Invasion through three-dimensional collagen I and matrigel barriers by preovulatory expanded COCs was equivalent to that of a known invasive breast cancer cell line (MB-231). Cumulatively, these results demonstrate that cumulus cells in the expanded COC transition to an adhesive, motile, and invasive phenotype in the periovulatory period that may be required for successful release of the oocyte from the ovary at ovulation.     

Local roles of TGF-beta superfamily members in the control of ovarian follicle development

Members of the transforming growth factor-beta (TGF-beta) superfamily have wide-ranging influences on many tissue and organ systems including the ovary. Two recently discovered TGF-beta superfamily members, growth/differentiation factor-9 (GDF-9) and bone morphogenetic protein-15 (BMP-15; also designated as GDF-9B) are expressed in an oocyte-specific manner from a very early stage and play a key role in promoting follicle growth beyond the primary stage. Follicle growth to the small antral stage does not require gonadotrophins but appears to be driven by local autocrine/paracrine signals from both somatic cell types (granulosa and theca) and from the oocyte. TGF-beta superfamily members expressed by follicular cells and implicated in this phase of follicle development include TGF-beta, activin, GDF-9/9B and several BMPs. Acquisition of follicle-stimulating hormone (FSH) responsiveness is a pre-requisite for growth beyond the small antral stage and evidence indicates an autocrine role for granulosa-derived activin in promoting granulosa cell proliferation, FSH receptor expression and aromatase activity. Indeed, some of the effects of FSH on granulosa cells may be mediated by endogenous activin. At the same time, activin may act on theca cells to attenuate luteinizing hormone (LH)-dependent androgen production in small to medium-size antral follicles. Dominant follicle selection appears to depend on differential FSH sensitivity amongst a growing cohort of small antral follicles. Activin may contribute to this selection process by sensitizing those follicles with the highest "activin tone" to FSH. Production of inhibin, like oestradiol, increases in selected dominant follicles, in an FSH- and insulin-like growth factor-dependent manner and may exert a paracrine action on theca cells to upregulate LH-induced secretion of androgen, an essential requirement for further oestradiol secretion by the pre-ovulatory follicle. Like activin, BMP-4 and -7 (mostly from theca), and BMP-6 (mostly from oocyte), can enhance oestradiol and inhibin secretion by bovine granulosa cells while suppressing progesterone secretion; this suggests a functional role in delaying follicle luteinization and/or atresia. Follistatin, on the other hand, may favor luteinization and/or atresia by bio-neutralizing intrafollicular activin and BMPs. Activin receptors are expressed by the oocyte and activin may have a further intrafollicular role in the terminal stages of follicle differentiation to promote oocyte maturation and developmental competence. In a reciprocal manner, oocyte-derived GDF-9/9B may act on the surrounding cumulus granulosa cells to attenuate oestradiol output and promote progesterone and hyaluronic acid production, mucification and cumulus expansion.     

Synergistic roles of BMP15 and GDF9 in the development and function of the oocyte-cumulus cell complex in mice: genetic evidence for an oocyte-granulosa cell regulatory loop

Bone morphogenetic protein 15 (BMP15) and growth differentiation factor 9 (GDF9) are oocyte-specific growth factors that appear to play key roles in granulosa cell development and fertility in most mammalian species. We have evaluated the role(s) of these paracrine factors in the development and function of both the cumulus cells and oocytes by assessing cumulus expansion, oocyte maturation, fertilization, and preimplantation embryogenesis in Gdf9+/-Bmp15-/- [hereafter, double mutant (DM)] mice. We found that cumulus expansion, as well as the expression of hyaluronon synthase 2 (Has2) mRNA was impaired in DM oocyte-cumulus cell complexes. This aberrant cumulus expansion was not remedied by coculture with normal wild-type (WT) oocytes, indicating that the development and/or differentiation of cumulus cells in the DM, up to the stage of the preovulatory luteinizing hormone (LH) surge, is impaired. In addition, DM oocytes failed to enable FSH to induce cumulus expansion in WT oocytectomized (OOX) cumulus. Moreover, LH-induced oocyte meiotic resumption was significantly delayed in vivo, and this delayed resumption of meiosis was correlated with the reduced activation of mitogen-activated protein kinase (MAPK) in the cumulus cells, thus suggesting that GDF9 and BMP15 also regulate the function of cumulus cells after the preovulatory LH surge. Although spontaneous in vitro oocyte maturation occurred normally, oocyte fertilization and preimplantation embryogenesis were significantly altered in the DM, suggesting that the full complement of both GDF9 and BMP15 are essential for the development and function of oocytes. Because receptors for GDF9 and BMP15 have not yet been identified in mouse oocytes, the effects of the mutations in the Bmp15 and Gdf9 genes on oocyte development and functions must be produced indirectly by first affecting the granulosa cells and then the oocyte. Therefore, this study provides further evidence for the existence and functioning of an oocyte-granulosa cell regulatory loop.     

Molecular characterization of the follicle defects in the growth differentiation factor 9-deficient ovary.

Growth differentiation factor-9 (GDF-9), a secreted member of the transforming growth factor-beta superfamily, is expressed at high levels in the mammalian oocyte beginning at the type 3a primary follicle stage. We have previously demonstrated that GDF-9-deficient female mice are infertile because of an early block in folliculogenesis at the type 3b primary follicle stage. To address the molecular defects that result from the absence of GDF-9, we have analyzed the expression of several important ovarian marker genes. The major findings of our studies are as follows: 1) There are no detectable signals around GDF-9-deficient follicles for several theca cell layer markers [i.e. 17alpha-hydroxylase, LH receptor (LHR), and c-kit, the receptor for kit ligand]. This demonstrates that in the absence of GDF-9, the follicles are incompetent to emit a signal that recruits theca cell precursors to surround the follicle; 2) The primary follicles of GDF-9-deficient mice demonstrate an up-regulation of kit ligand and inhibin-alpha. This suggests that these two important secreted growth factors, expressed in the granulosa cells, may be directly regulated in a paracrine fashion by GDF-9. Up-regulation of kit ligand, via signaling through c-kit on the oocyte, may be directly involved in the increased size of GDF-9-deficient oocytes and the eventual demise of the oocyte; 3) After loss of the oocyte, the cells of the GDF-9-deficient follicles remain in a steroidogenic cluster that histologically resembles small corpora lutea. However, at the molecular level, these cells are positive for both luteal markers (e.g. LHR and P-450 side chain cleavage) and nonluteal markers (e.g. inhibin alpha and P-450 aromatase). This demonstrates that initially the presence of the oocyte prevents the expression of luteinized markers, but that the absence of GDF-9 at an early timepoint alters the differentiation program of the granulosa cells; and 4) As demonstrated by staining with either proliferating cell nuclear antigen (PCNA) or Ki-67 and TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling) labeling, the granulosa cells of GDF-9-deficient type 3b primary follicles fail to proliferate but also fail to undergo cell death. This suggests that granulosa cells of type 3b follicles require GDF-9 for continued growth and also to become competent to undergo apoptosis, possibly through a differentiation event Thus, these studies have enlightened us as to the paracrine roles of GDF-9 as well as the normal steps of granulosa cell and theca cell growth and differentiation within ovarian follicles.     

Qualitative and quantitative changes in protein synthesis of bovine follicular cells during the preovulatory period.

To understand the mechanisms governing oocyte maturation better, the effects of the gonadotropin surge were studied on follicular cells of bovine preovulatory follicles. For this purpose, qualitative and quantitative changes in protein synthesis by both granulosa cells and cumulus cells were compared relative to the luteinizing hormone (LH) surge and the resumption of meiosis in the oocyte. Follicular cells were collected at different times before and up to 25 hr after the LH surge. For each individual preovulatory follicle, granulosa and cumulus cells were incubated separately for 3 hr with 3H-methionine or with 35S-methionine. Newly synthesized cytosolic proteins from granulosa and cumulus cells and proteins secreted into the medium were analyzed by polyacrylamide gel electrophoresis. The radioactivity was measured by liquid scintillation counting after slicing of the gels or revealed by fluorography. Three major peaks of the newly synthesized proteins, with molecular weights of 76, 56, and 30 kDa, were studied throughout the preovulatory period. After the LH surge, the overall level of protein synthesis increased in granulosa cells. In addition, the pattern of cytosolic proteins in granulosa cells changed, and, in particular, the relative synthesis of the 30 kDa peak decreased. These changes in cytosolic protein synthesis may be due to the action of LH since they could be reproduced in vitro in LH-stimulated granulosa cells. A predominant peak of 56 kDa was secreted by granulosa cells throughout the experimental period. No significant change was observed in proteins synthesized by cumulus cells under the same experimental conditions. The amounts of radioactivity incorporated into the three major proteins secreted by granulosa cells, however, were correlated significantly with the amounts of radioactivity incorporated by similar proteins synthesized by cumulus cells. These results indicate that cumulus cells respond differently from granulosa cells to the gonadotropin surge but not in an independent manner.     

Synergistic roles of bone morphogenetic protein 15 and growth differentiation factor 9 in ovarian function

Knockout mouse technology has been used over the last decade to define the essential roles of ovarian-expressed genes and uncover genetic interactions. In particular, we have used this technology to study the function of multiple members of the transforming growth factor-beta superfamily including inhibins, activins, and growth differentiation factor 9 (GDF-9 or Gdf9). Knockout mice lacking GDF-9 are infertile due to a block in folliculogenesis at the primary follicle stage. In addition, recombinant GDF-9 regulates multiple cumulus granulosa cell functions in the periovulatory period including hyaluronic acid synthesis and cumulus expansion. We have also cloned an oocyte-specific homolog of GDF-9 from mice and humans, which is termed bone morphogenetic protein 15 (BMP-15 or Bmp15). To define the function of BMP-15 in mice, we generated embryonic stem cells and knockout mice, which have a null mutation in this X-linked gene. Male chimeric and Bmp15 null mice are normal and fertile. In contrast to Bmp15 null males and Gdf9 knockout females, Bmp15 null females (Bmp15(-/-)) are subfertile and usually have minimal ovarian histopathological defects, but demonstrate decreased ovulation and fertilization rates. To further decipher possible direct or indirect genetic interactions between GDF-9 and BMP-15, we have generated double mutant mice lacking one or both alleles of these related homologs. Double homozygote females (Bmp15(-/-)Gdf9(-/-)) display oocyte loss and cysts and resemble Gdf9(-/-) mutants. In contrast, Bmp15(-/-)Gdf9(+/-) female mice have more severe fertility defects than Bmp15(-/-) females, which appear to be due to abnormalities in ovarian folliculogenesis, cumulus cell physiology, and fertilization. Thus, the dosage of intact Bmp15 and Gdf9 alleles directly influences the destiny of the oocyte during folliculogenesis and in the periovulatory period. These studies have important implications for human fertility control and the maintenance of fertility and normal ovarian physiology.     

Cyclic guanosine 5'-monophosphate-dependent protein kinase II is induced by luteinizing hormone and progesterone receptor-dependent mechanisms in granulosa cells and cumulus oocyte complexes of ovulating follicles

Cyclic GMP (cGMP)-dependent protein kinase II (Prkg2, cGK II) was identified as a potential target of the progesterone receptor (Nr3c3) in the mouse ovary based on microarray analyses. To document this further, the expression patterns of cGK II and other components of the cGMP signaling pathway were analyzed during follicular development and ovulation using the pregnant mare serum gonadotropin (PMSG)-human chorionic gonadotropin (hCG)-primed immature mice. Levels of cGK II mRNA were low in ovaries of immature mice, increased 4-fold in response to pregnant mare serum gonadotropin and 5-fold more within 12 h after hCG, the time of ovulation. In situ hybridization localized cGK II mRNA to granulosa cells and cumulus oocyte complexes of periovulatory follicles. In progesterone receptor (PR) null mice, cGK II mRNA was reduced significantly at 12 h after hCG in contrast to heterozygous littermates. In primary granulosa cell cultures, cGK II mRNA was induced by phorbol 12-myristate 13-acetate enhanced by adenoviral expression of PR-A and blocked by RU486 and trilostane. PR-A in the absence of phorbol 12-myristate 13-acetate was insufficient to induce cGK II. Expression of cGK I (Prkg1) was restricted to the residual tissue and not regulated by hormones. Guanylate cyclase-A (Npr1; GC-A) mRNA expression increased 6-fold by 4 h after hCG treatment in contrast to pregnant mare serum gonadotropin alone and was localized to granulosa cells of preovulatory follicles. Collectively, these data show for the first time that cGK II (not cGK I) and GC-A are selectively induced in granulosa cells of preovulatory follicles by LH- and PR-dependent mechanisms, thereby providing a pathway for cGMP function during ovulation.     

Gene expression profiles of cumulus cell oocyte complexes during ovulation reveal cumulus cells express neuronal and immune-related genes: does this expand their role in the ovulation process?

Ovulation is a complex process initiated by the preovulatory LH surge, characterized by cumulus oocyte complex (COC) expansion and completed by the release of a mature oocyte. Although many ovarian genes that impact ovulation have been identified, we hypothesized that genes selectively expressed in COCs would be overlooked by approaches using whole ovary or granulosa cell samples. RNA isolated from COCs collected from preovulatory follicles of equine chorionic gonadotropin (CG) primed mice and at selected times after human CG treatment was subjected to microarray analyses and results confirmed by RT-PCR analyses, Western blotting, and immunofluorescent studies. A remarkable number of genes were up-regulated in COCs including Areg, Ereg, and Btc. Several genes selectively expressed in cumulus cells compared with granulosa cells were related to neuronal (Mbp, Tnc, Nts) or immune (Alcam, Pdcd1, Cd34, Cd52, and Cxcr4) cell function. In addition to Sfrp2, other members of the Wnt/Fzd family (Sfrp4, Fdz1 and Fdz2) were expressed in COCs. Thus, there is a cumulus cell-specific, terminal differentiation process. Furthermore, immunofluorescent analyses documented that cumulus cells are highly mitotic for 4-8 h after human CG and then cease dividing in association with reduced levels of Ccnd2 mRNA. Other down-regulated genes included: Cyp19a1, Fshr, Inhb, and the oocyte factors Zp1-3 and Gja4. In summary, the vast number of matrix, neuronal, and especially immune cell-related genes identified by the gene- profiling data of COCs constitutes strong and novel evidence that cumulus cells possess a repertoire of immune functions that could be far greater than simply mediating an inflammatory-like response.