| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Physiologie de la Reproduction et des Comportements, UMR 6175 INRA-CNRS-Université F. Rabelais de Tours Haras Nationaux, 37380 Nouzilly, France
ABSTRACT
Recently, the role of several elements of the bone morphogenetic protein (BMP) family has been studied in the ovary, some of them being crucial for ovarian function. In the present work, we have studied bone morphogenetic protein 5 (BMP5) expression and its biological role in the rat ovary. BMP5 is expressed by rat granulosa cells (GCs) and exerts specific biological effects on proliferation and steroidogenesis of these cells in an autocrine manner. These effects were shown to be associated with an increase in cyclin D2 protein level and a decrease in steroidogenic acute regulatory (StAR) protein expression in GCs in vitro. Ultimately, BMP5 actions were inhibited by follistatin. Overall, these data show that BMP5 is a novel element of the BMP family that might play a fully paracrine role in rodent ovarian folliculogenesis.
BMP5, follicle-stimulating hormone, granulosa cells, ovary, progesterone, rat
INTRODUCTION
During the past decade, a number of studies have identified several factors produced by the oocyte, granulosa cells (GCs), and/or theca cells that are able to play a role in ovarian folliculogenesis in a paracrine/autocrine manner. Among these factors, the bone morphogenetic protein (BMP) family has emerged as a central player in ovarian physiology and female fertility [1]. In particular, natural mutations in sheep, as well as gene-targeted invalidation in the mouse, revealed that BMP family plays a key role in ovarian function. For example, growth and differentiation factor-9 (GDF9), an oocyte-derived growth factor of the BMP family, has shown to be required for early ovarian folliculogenesis, enabling primary to secondary follicle transition [2, 3]. Moreover, homozygous natural mutant ewes carrying loss-of-function mutation in Gdf9 and Bmp15 gene sequence are infertile, follicular growth being arrested at the primary follicular stage, whereas heterozygous females exhibit increased ovulation rate [3, 4]. Ultimately, evidence that the Q249R substitution in the BMPR-IB sequence leads to high fecundity in ewes [5–7] recently provided further support that BMP signaling pathways play a key role in ovarian function.
A major concept in ovarian physiology is that FSH stimulates granulosa cells from growing follicles to produce increasing amounts of estradiol during terminal follicular phase. However, FSH-induced progesterone synthesis in vivo is inhibited until the preovulatory period, when it increases suddenly. In contrast to the steroidogenic properties of granulosa cells in vivo, granulosa cells from early antral follicles cultured in vitro in the presence of FSH secrete copious amounts of both estradiol and progesterone. These findings led to the hypothesis that there is a luteinizing inhibitor in the follicle in vivo that selectively inhibits FSH-dependent progesterone but not estradiol production by GCs.
In the rat ovary, the expression of Bmp3, Bmp3b, Bmp4, and Bmp7 mRNA has been observed in the theca cells [8], and Gdf9 and Bmp15 have been shown to be only expressed in the oocyte; Bmp6 in both oocyte and GCs; BmprIA and BmprIB in the oocyte, GCs, and theca cells; and BmprII in oocyte and GCs [8]. Moreover BMP4, –6, –7, and –15 were shown to inhibit FSH-induced progesterone production, associated with an inhibitory effect on Star, cytochrome P450 side chain cleavage (Cyp11a1), and 3 beta hydroxysteroid dehydrogenase (3 beta hsd) expression [9–11].
In the present work, we searched by using bioinformatic tools if other members of the BMP family are also expressed in the ovary. In this way, we have observed in UniGene (http://www.ncbi.nlm.nih.gov/UniGene) that ESTs of Bmp5 were highly represented in libraries from mouse egg and ovaries. BMP5 belongs to the 60A subgroup of BMPs, which also includes BMP6, BMP7, BMP8a, and BMP8b [12]. BMP5 is required for normal growth and patterning of skeletal structure [13], but nothing is known about its role in ovarian function. In this work, we have assessed for the first time BMP5 expression and function in the rat follicle. We report that Bmp5 is expressed by GCs and is able to both stimulate their proliferation rate and inhibit progesterone secretion.
MATERIALS AND METHODS
Materials
Purified ovine FSH-20 (oFSH) (lot no. AFP-7028D, 4453 IU/mg, FSH activity = 175 times activity of oFSH-S1) was a gift from NIDDK, National Hormone Pituitary Program (Bethesda, MD). Recombinant human insulin-like growth factor-I (IGF1) was a gift from Dr. J. Zapf (University Hospital, Zurich). Recombinant human BMP5 and follistatin were obtained from R&D Systems Europe (Lille, France).
Isolation of cDNA Fragments of Mouse and Rat BMP5
The pGEMT-Bmp5 antisense and sense constructs used for in situ hybridization were generated by inserting the fragment of Bmp5 cDNA rat (600 bp) into pGEM-T vector (Promega) and selecting a clone with the appropriate antisense or sense orientation. The Bmp5 cDNA fragment was generated by RT-PCR from rat ovary mRNA using sense 5'-GGGTTGGCTTGTCTTTGATA-3' and antisense 5-TGGCATTTAGTTTGGTTGGA-3' primers.
In Situ Hybridization
Female adult rats (Janvier Laboratories, Genest St. Isle, France) at different sexual stages (estrus, diestrus, pre-estrus and postestrus, three females per sexual stage) were perfused with 4% (w/v) paraformaldehyde. The ovaries were then recovered, postfixed in PAF 4% overnight at 4°C, and put in PBS containing 15% of sucrose for 24 h at 4°C. The ovaries were then included in tissue Tek. Frozen ovaries were serially sectioned at a thickness of 10 µm with a cryostat to perform in situ hybridization experiments using 35S-labeled rat Bmp5 cRNA. For prehybridization, sections were incubated for 2 h at 50°C in the following prehybridization buffer: 50% formamide (v/v) (Merck), 0.6 M NaCl, 10 mM Tris, 1 mM EDTA, 1% SDS (Serva Biowhittaker), 10 mM dithiothreitol (DTT, Boehringer Mannheim), 250 µg/ml tRNA (Sigma), 2% Denhardt reagent (Eurogentec), and 100 µg/ml salmon sperm DNA. For hybridization, 35S-labeled antisense and sense cRNA probe were diluted in the prehybridization buffer with 10% dextran sulfate (w/v) and without salmon sperm DNA, denatured at 80°C for 3 min, and applied on sections (300000 cpm/50 µl) at 50°C overnight in a sealed humidified container. Nonspecifically bound RNA transcripts were removed by washing sections in Tris buffer (10 mM Tris, 0.5 M NaCl, pH 8) containing 20 µg/ml RNase A (Boehringer Mannheim) during 1 h at 37°C. Sections were then washed in 1) Tris buffer without RNase A for 30 min at 37°C; 2) in 50% formamide, 0.1% ß-mercaptoethanol (v/v), and single-strength SSC (single strength SCC = 150 mM sodium chloride, 15 mM sodium citrate, pH 7) for 30 min at 50°C; and 3) in 50% formamide, 0.1% ß-mercaptoethanol, and 0.1 single-strength SSC for 30 min at 37°C. Then sections were dehydrated and air-dried, and slides were dipped in Kodak NTB2 emulsion (Integra Bioscience) and exposed at 4°C for 3–6 wk in a desiccated dark box. Slides were developed and counterstained with hematoxylin. Specificity of hybridization was assessed by comparing signal obtained with the cRNA antisense probe and the corresponding cRNA sense probe.
Histological determination of follicular size and degree of atresia was performed on adjacent sections stained with feulgen. As previously described [14], follicles were judged normal or atretic using classical histological criteria (normal: frequent mitosis, no pyknosis; atretic: no mitosis, frequent pyknotic bodies in granulosa cells).
Isolation and Culture of Granulosa Cells
Twenty-two-day-old female wistar rats were injected with 1 mg of DES during 3 days, and animals were then slaughtered. Ovaries were removed and immersed for 15 min in isotonic solution containing fungizone and antibiotics (penicillin and streptomycin) immediately after slaughter, then placed in B2 medium prepared according to Menezo [15]. GCs were extracted, and cell suspensions were centrifuged at 1200 rpm for 7 min. Pellets were resuspended in culture medium (McCoy's 5a medium; Sigma), and GC were cultured according to the method described by Campbell et al. [16].
Cultures were performed in 96-well plates. Cells were seeded at 105 viable cells/well and cultured at 37°C in an humidified atmosphere with 5% CO2 in serum-free culture medium in the presence or absence of oFSH, IGF1, follistatin, and BMP5.
Each combination of treatment was tested in triplicate in three to five independent cultures. Cultures of GCs were performed for 96 h. Culture media were partially changed at 48 h, by replacing 180 µl of the total (250 µl) with prewarmed medium. The spent medium between 48 and 96 h was stored at –20°C before progesterone assay.
All procedures were approved by the Agricultural and Scientific Research Government committees in accordance with the guidelines for the care and use of Agricultural Animals in Agricultural Research and Teaching (approval A37801).
Assessment of Progesterone Production
Progesterone amounts (ng/50000 cells) in the culture media conditioned between 48 and 96 h from each experiment were measured by radioimmunoassay as previously described [17].
Western Blot Analysis
GC whole-cell extracts were obtained by resuspension in lysis buffer (10 mM Tris pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 0.5% Igepal) containing several protease inhibitors (2 mM PMSF, 10 mg/ml leupeptin, 10 mg/ml aprotinin) and phosphatase inhibitors (100 mM sodium fluoride, 10 mM sodium pyrophosphate, 2 mM sodium orthovanadate; Sigma). Lysates were centrifuged at 15000 g for 30 min at 4°C, and the protein concentration in the supernatants was determined by a colorimetric assay (kit BC Assay; Uptima Interchim). The protein samples (30 µg) were submitted to electrophoresis using SDS-PAGE in 12% polyacrylamide gels and transferred to nitrocellulose membranes (Schleicher & Schuell). For steroidogenic protein detection, GCs were harvested after 48 h of treatment with FSH and/or BMP5 in serum-free condition. The same samples were run and transferred on two independent membranes to allow individual detection of CYP11A1 and dual detection on the same membrane of steroidogenic acute regulatory (StAR) protein and tubulin used as a loading control. Rabbit polyclonal anti-CYP11A1 antibody was purchased from Chemicon (Euromedex), and mouse StAR was kindly provided by Dr. D.B. Hales (Chicago, IL). Tubulin was revealed using a mouse monoclonal antitubulin antibody (1:1000; Calbiochem). For cycline D2 detection (1:1000; Santa Cruz), GCs were seeded in a 12-well plate with 3% fetal ovine serum and harvested after 48 h of culture in the presence or absence of BMP5. Incubation of membranes with rabbit polyclonal or mouse monoclonal primary antibodies were followed by incubation with peroxidase conjugated anti-rabbit (1:20000; Interchim) or anti-mouse IgG (1:20000; Bio-Rad), respectively. Finally, the protein bands were detected using ECL detection reagents (Amersham-Pharmacia Biotech).
Cell Proliferation
Proliferation was assessed by measuring [3H] thymidine incorporation (du Pont de Nemours). GCs were seeded in tissue culture chambers mounted on glass microslides (lab-Tek; Nunc) in the presence or absence of BMP5 (50 ng/ml). After 48 h of culture, cells were incubated for 2 h at 37°C in the presence of [3H] thymidine in modified B2 without thymine and fixed with Bohm-Sprenger fixative (15% formaldehyde, 5% acetic acid, 80% methanol) for 10 min at 20°C. Afterward, cells were stained with feulgen (Merck) and slides were dipped in NTB2 emulsion (Kodak), air-dried, and exposed for 6 days at 4°C. Autoradiographs were then developed and the labeling index (LI; percentage of thymidine-labeled cells) was assessed by counting labeled and unlabeled cells in randomly chosen microscopic fields. For each culture well, estimation of LI was performed on about 1000 cells.
Data Analysis
All experimental data are presented as mean ± SEM. The effects of hormones on progesterone and estradiol secretion were analyzed using two-way ANOVA to appreciate the hormone effect as well as the culture effect. Paired t-test was used to appreciate treatment effect on Western blot analysis. Post hoc comparisons were performed with the Scheffé and Newman-Keuls tests. For all analysis, differences with P > 0.05 were considered as not significant.
RESULTS
Expression of BMP5 mRNA in Ovary
By systematically screening most of the elements of the BMP family in UniGene, we found that a high number of Bmp5 ESTs were present in mouse egg and ovary libraries. We then decided to make an overall study on the expression and the role of BMP5 in the rat ovary, in vitro culture of GCs being easier in this species than in the mouse.
By in situ hybridization on ovarian sections (Fig. 1), Bmp5 mRNA was found to be mainly expressed in granulosa and cumulus cells of early (EA) to large antral (LA) follicles (1, 2, 3, 4, 9, 10, 11, 12). Note that Bmp5 expression is observed in GCs of both healthy (3, 4) and atretic follicles (5, 6). A low but significant expression was also detected in oocyte of small preantral follicles. No expression was detected in theca cells. No clear difference of expression was observed between diestrus, pre-estrus, estrus, or postestrus stages (data not shown).
|
BMP5 Action on GC Proliferation
We then examined whether BMP5 was able to regulate GC proliferation (Fig. 2). Treatment of rat GCs with BMP5 (50 ng/ml) stimulated by 2-fold the level of the thymidine incorporation (Fig. 2A) (P < 0.01). This stimulation was associated with a 1.5-fold increase in the cyclin D2 protein accumulation (Fig. 2B) (P < 0.05).
|
BMP5 Action on GC Steroidogenesis
We next investigated the effect of BMP5 on basal and FSH-induced progesterone (P4) and estradiol (E2) production using primary rat GCs cultured in serum-free medium. As expected, FSH alone (5 ng/ml) increased P4 and E2 production (P < 0.001). Cotreatment of GCs with FSH (5 ng/ml) and BMP5 caused a marked dose-dependent inhibition of the FSH-induced P4 production (60% with 50 ng/ml of BMP5) (Fig. 3A). Moreover, BMP5 alone affected basal level of P4 production in a dose-dependent manner as well (65% with 50ng/ml of BMP5 P < 0.001) (Fig. 3A). As expected, treatment with FSH (0–10 ng/ml) dose-dependently increased progesterone production by GCs, and BMP5 (50 ng/ml) inhibitory action was also observed in the presence of the higher concentration of FSH (10 ng/ml) (Fig. 3B).
|
As previously shown, IGF1 dose-dependently enhanced FSH action on progesterone secretion (Fig. 3C). Addition of BMP5 (50 ng/ml) totally abolished IGF1 action at 10 ng/ml and reduced by 75% IGF1 action at 100 ng/ml (P < 0.001). BMP5 also inhibited FSH action by 60% and 55% in the presence of 10 and 50 ng/ml of IGF1, respectively (P < 0.001) (Fig. 3C). BMP5 had no effect on basal and FSH-induced estradiol production (Fig. 4).
|
Effect of BMP5 on Steroidogenic Enzymes: Underlying Mechanisms
To understand the mechanism by which BMP5 inhibited basal and FSH-induced progesterone secretion, we studied the effects of BMP5 on forskolin- and IBMx-induced P4 production by GCs in comparison with FSH effects (Fig. 5). As expected, addition of IBMx or forskolin enhanced P4 production by GCs. BMP5 inhibited both forskolin- and IBMx-induced P4 level, similarly to the effects observed in the presence of FSH. This suggests that BMP5 exerts, at least partly, an inhibitory action on cAMP action.
|
To further elucidate the molecular mechanisms that underlie the action of BMP5 on GCs, we have assessed the expression of proteins implicated in progesterone synthesis including the StAR and the CYP11A1. As shown in Figure 6, treatment of GCs with BMP5 (50 ng/ml) during 48 h inhibited basal StAR and CYP11A1 accumulation and tended to inhibit FSH-induced StAR accumulation (P = 0.23). These results indicate that recombinant BMP5 exhibits specific biological activities on rat primary GCs.
|
Regulation of BMP5 Activity
Ultimately, we have investigated the ability of follistatin, a well-known activin antagonist expressed in the ovary [18–20], to regulate BMP5 activity on GCs (Fig. 7). Addition of follistatin on GCs significantly reduced the effect of BMP5 on progesterone production by rat GCs in vitro in a dose-dependent manner (P < 0.001).
|
DISCUSSION
There is accumulating evidence supporting a crucial role of BMPs in ovarian function. In the present study, we showed that Bmp5 is expressed in rat GCs of antral follicles. Moreover, in vitro experiments showed that BMP5 is able to inhibit basal and FSH-induced P4 production without affecting E2 production, suggesting that BMP5 preferentially inhibits biochemical pathways that lead to P4 production. This BMP5-inhibitory effect was associated with a marked decrease in StAR and CYP11A1 expression in basal and in FSH-stimulated conditions.
BMP5 inhibited forskolin- as well as IBMx-induced P4 secretion, supporting the hypothesis that BMP5 acts downstream of FSH-R, at least partly at the level of phosphodiesterase and cyclase adenylate. Similar results have been previously obtained by studying BMP4 inhibitory effects on P4 secretion by ovine granulosa cells, suggesting a common mechanism of action of BMP ligands on granulosa cells [21].
We have also observed that the BMP5-enhanced proliferation of GCs was associated with an increase in cyclin D2 expression. Interestingly, cyclin D2-deficient females are sterile, ovarian GCs being unable to proliferate in response to FSH [22]. Of note, in the rat GCs, activin alone had no effect on cyclin D2 accumulation but increased the FSH-stimulated cyclin D2 accumulation [23]. So, one possible hypothesis is that at least part of the BMP5 effect on GC proliferation is mediated by this increase in cyclin D2 expression.
Finally, we showed that BMP5-inhibitory actions on progesterone secretion could be regulated by follistatin, initially described for its binding affinity for activin [18]. Interestingly, the inhibition of BMP5 action by follistatin was observed here in a molar ratio similar to that previously observed for the inhibition of activin action on human GCs [24]. Of note, follistatin was also shown to inhibit the dendrite-promoting activity of BMP5 in a concentration-dependent manner [25].
What can we hypothesize about the physiological relevance of BMP5 in ovarian function? BMP5 belongs to the 60A subfamily of the BMP system, which contains BMP6, BMP7, BMP8A, and BMP8B, all of these elements presenting more than 70% of identity of sequence (for review, see [12]). All these factors have already been described to be implicated in female (BMP6, BMP7) and male reproduction (BMP8A, BMP8B) [10, 11, 26, 27]. Despite their expression in the mouse ovary (BMP5, our unpublished data; BMP6 [28]), mice exhibiting natural or experimental loss of function mutation in BMP5 and BMP6, respectively, are viable and have no apparent ovarian abnormalities [13, 28]. This suggests that, in contrast to GDF9, these BMPs are dispensable for normal ovarian function, their absence being likely compensated by other members of the BMP subfamily. Alternatively, one could hypothesize that invalidation of one of these genes leads to subtle alteration of ovarian function, without clear consequence on fertility, invalidation of both genes being required to lead to a clear ovarian phenotype. In particular, in the mouse, the loss of one copy of the Gdf9 gene or the loss of two copies of Bmp15 have only subtle consequences on ovarian follicle population and female fertility, whereas Gdf9+/–/Bmp15–/– female mice are sterile. Similarly, Bmp7–/– mice die early after birth [29], whereas the same targeted invalidation leads to an early embryonic mortality in a Bmp5-invalidated genetic background. These data suggest that both factors may be able to, in part, functionally compensate each other at early stages of embryonic and fetal development [30]. Similarly, Bmp6 null mice present sternum defect, this phenotype being slightly exacerbated in Bmp5/6 double mutant animals [28].
Overall, our data have shown that BMP5 is expressed in rat GCs, stimulate their proliferation, and inhibit their P4 secretion. Moreover, BMP5 action is inhibited by follistatin.
ACKNOWLEDGMENTS
We thank Claude Cahier, Jean Claude Braguer, and Eric Jean-Pierre for expert animal care. We acknowledge Michele Peloille for sequencing and Danielle Monniaux and Stephane Fabre for helpful discussion.
FOOTNOTES
1 Supported by Institut National de la Recherche Agronomique. A.P. was supported by a fellowship from Institut National de la Recherche Agronomique and Région Centre. 
2 Correspondence: Philippe Monget, Physiologie de la Reproduction et des Comportements, UMR 6175 INRA-CNRS-Université F. Rabelais de Tours Haras Nationaux, 37380 Nouzilly, France. FAX: 33 2 47 42 77 43; monget{at}tours.inra.fr
Received: 21 April 2005.
First decision: 23 May 2005.
Accepted: 27 July 2005.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  S Elis, J Dupont, I Couty, L Persani, M Govoroun, E Blesbois, F Batellier, and P Monget Expression and biological effects of bone morphogenetic protein-15 in the hen ovary J. Endocrinol., September 1, 2007; 194(3): 485 - 497. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. G Knight and C. Glister TGF-{beta} superfamily members and ovarian follicle development. Reproduction, August 1, 2006; 132(2): 191 - 206. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
