| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
research-article |
INSERM U-407, Faculté de Médecine Lyon-Sud, BP 12, 69921 Oullins Cedex, France
ABSTRACT
Bone morphogenetic protein (BMP) 15 and growth differentiation factor (GDF) 9 are oocyte-secreted growth factors that are critical local regulators of ovarian function and may be involved in preovulatory cumulus expansion. As cumulus expansion occurs in response to the ovulatory surge, the present study was designed: 1) to investigate whether GDF9 and BMP15 are regulated by gonadotropins in the mouse ovary; and 2) to visualize changes in both GDF9 and BMP15 immunostaining in response to gonadotropins. Immature 21-day-old mice were sequentially treated with recombinant human FSH (r-hFSH), 5 IU daily, at Days 21, 22, and 23 of life, then injected with 5 IU hCG at Day 24 of life. In response to r-hFSH, steady-state Bmp15 mRNA expression levels increased in both total ovaries and cumulus-oocyte complexes, whereas Gdf 9 mRNA levels did not. In addition, BMP15 protein levels increased in total ovaries. The GDF9 immunostaining was exclusively seen in growing oocytes in both control and gonadotropin-treated mice, whereas that of BMP15, which was also primarily seen in growing oocytes, exhibited important changes in response to gonadotropins. Strong BMP15 immunostaining was observed in the follicular fluid of atretic antral follicles after FSH treatment and in expanded, but not in compact, cumulus cells after hCG. The present results show for the first time that BMP15 levels increase during gonadotropin-induced follicular development, in parallel with oocyte maturation, and that this local factor is likely involved in cumulus expansion as previously suggested by studies in Bmp15-null mice.
cumulus cells, follicle-stimulating hormone, growth factors, oocyte development, ovary
Bone morphogenetic protein 15 (BMP15) and growth differentiation factor 9 (GDF9) are two members of the transforming growth factor (TGFB) superfamily [1] that are only produced by growing oocytes in mice [2, 3], rats [4, 5], and humans [4, 6]. Both BMP15 [7] and GDF9 [8] exert their biological effects by binding to the BMP receptor II (BMPRII), but, whereas the downstream actions of BMP15 are mediated by the BMP receptor 1B (BMPRIB) and the Smad1/5/8 proteins [7], those of GDF9 are mediated by the TGFB type-I receptor (also named ALK-5) and the Smad 2/3 proteins [9]. As these receptors are present at the granulosa cell surface, GDF9 and BMP15 constitute a unique system that contributed to the emergence of the new paradigm that the oocyte controls, at least partly, folliculogenesis [10].
In vitro experiments have shown in the mouse that recombinant GDF9 modulates various granulosa cell functions, including inhibition of expression of the LH receptor, stimulation of both proliferation and progesterone production even in the absence of FSH, and cumulus expansion [11]. The importance of Gdf 9 in mouse reproduction was demonstrated by the observation that in the Gdf 9-null mouse, folliculogenesis does not process beyond the primary stage [12].
In vitro experiments using recombinant BMP15 have shown that, in rats, BMP15 strongly stimulates granulosa cell proliferation [5] as well as kit ligand (KL) production [13], whereas it inhibits some FSH-induced steroidogenic functions, but not estradiol synthesis [5], via an inhibition of FSH receptor (FSHR) expression in granulosa cells [14]. In Bmp15-null mice, it has been suggested that the terminal cumulus-oocyte maturation is impaired, leading to a more or less impaired fertility, depending upon the background strain of mice [15]. In addition to the recent observation that BMP15 is involved in cumulus expansion [16], the loss of the BMPRIB, the type I receptor of the BMP15 [7], is associated in BmprIB-null mice to a defective cumulus expansion [17]. Thus, BMP15 plays an important role in female reproduction in mice, but is not as essential as GDF9.
Whereas FSH stimulates GDF9 production by hamster postnatal ovaries in vitro [18], the BMP15 protein is downregulated in both Fshr-null and Fshr+/– ovaries, indicating that FSHR and BMP15 have a direct relationship, at least at the level of translation of the latter [19]. However, a recent in vitro study in the mouse [20] has shown that, depending on the dose, FSH had either no significant effect or significantly decreased Bmp15 expression by oocyte-granulosa cell complexes from preantral follicles when compared with controls.
Taken together, the available data suggest that, in the mouse, BMP15 might be involved in terminal follicular maturation, and especially at the time of cumulus expansion, although its regulation by gonadotropins in vivo appears somewhat confusing. Because in vivo cumulus expansion is induced by gonadotropins, the specific aims of the present study were: 1) to assess the ability of gonadotropins FSH and hCG to regulate expression of both Gdf 9 and Bmp15 genes in the mouse ovary and cumulus-oocyte complexes (COCs) in vivo; and 2) to describe the temporal relationships existing between GDF9 and BMP15 immunostaining and cumulus expansion at the time of terminal follicle maturation.
The treatment usually used to induce ovulation in immature rodents (i.e., one injection of 5 IU eCG, followed 48 h later by one injection of 5 IU hCG) was not appropriate for the present study, as eCG, is a mixture of FSH and LH. It was therefore impossible to differentiate which hormone was responsible for any observed effect. On the contrary to eCG, which possesses a long half-life (more than 48 h), recombinant human FSH (r-hFSH [Gonal-F]; Serono France, www.serono.fr) has a short half-life of several hours. It was therefore necessary to treat immature mice with FSH more than once. In rodents, 5 days are required for preantral follicles to reach the preovulatory size, and, because 21-day-old mice exhibit early antral follicles, it was decided to sustain the growth of these follicles with r-hFSH during 3 subsequent days. Consequently, 21-day-old OF1 mice (Charles River, www.criver.com) were treated by a daily s.c. injection of 5 IU r-hFSH on 3 consecutive days (Days 21, 22 and 23). On Day 24, they were injected i.m. with 5 IU hCG (Pergonal; Serono France). Histological analysis of ovaries showed that 48 h after hCG injection (n = 10) the number of corpora lutea was 8.3 ± 1.3 (mean ± SEM) per ovary. Treated mice were killed by cervical dislocation at Days 22 (24 h after r-hFSH injection [F24]), 23 (F48), 24 (F72), 25 (24 h after hCG injection [H24]), and 26 (H48) of age, and untreated control animals were killed at Days 21–26 of age. Ovaries were collected and carefully cleaned from surrounding fat tissue, ovarian bursa, and fallopian tubes. The freshly isolated ovaries were either fixed for 24 h in 4% buffered paraformaldehyde or Bouin solution, or directly snap-frozen. To study BMP15 and GDF9 in COCs, all visible follicles were punctured, and the COCs were pooled for each animal. Frozen ovaries and COCs were kept at –80°C for further biochemical studies. Investigations and animal care procedures were in accordance with the INSERM (French National Institute for Health and Medical Research) Animal Care Committee (decrees 2001–486 and 2001–464).
RNA isolation. Total RNA was extracted from whole ovaries with TRIzol reagent (Invitrogen, www.invitrogen.com), a monophasic solution of phenol and guanidine isothiocyanate. This reagent is an improvement on the single-step RNA isolation technique previously developed by Chomczynski and Sacchi [21]. For COCs, total RNA was extracted with Total Quick RNA reagent (Talent, www.spin.it./talent) following the mini RNA preparation approach, per the manufacturer's instructions. Briefly, it consisted of lysing cells and homogenizing the suspension. Then, RNAs were bound to a silica resin that was washed twice to eliminate proteins and almost all DNA contaminants. Finally, RNAs were eluted in diethyl pyrocarbonate-treated water with swirling at 37°C. Two sharp ethidium bromide-stained bands of 18s and 28s rRNA, without any obvious smear, were visualized on a 1% agarose gel, proving the quality of the RNA obtained in whole ovaries and COCs. Furthermore, no larger size band corresponding to any putative DNA contamination was evident. The amount of RNA was estimated by spectrophotometry at 260 nm.
Reverse Transcription. Single-stranded cDNAs were obtained from reverse transcription (RT) of 500 ng of total RNA. The reaction was carried out using random hexanucleotides as primers (5 µM) in the presence of deoxynucleotide triphosphates (250 µM; Invitrogen), dithiothreitol (10 mM; Invitrogen), and Moloney murine leukemia virus reverse transcriptase (10 µl; Invitrogen) in a final volume of 20 µl for 1 h at 37°C.
Samples were analyzed by real-time PCR using the LightCycler 3.5 instrument (Roche, www.roche-applied-science.com). Complementary DNAs (5 µl of RT mixture diluted 1:10) were amplified using an appropriate primer set (0.5 µM) and 10 µl of QuantiTect SYBR green Master Mix (Qiagen, www.qiagen.com), containing HotStartTaq DNA polymerase, MgCl2 (5 mM), diethylnitrophenyl thiophosphates, including 2'-deoxyuridine 5'-triphosphate, PCR buffer, and a double-stranded DNA-specific fluorescent dye. Amplification was carried out in a final volume of 20 µl, as follows: initial activation of HotStartTaq DNA polymerase at 95°C for 15 min; 50 cycles in three steps (denaturation at 95°C for 15 sec, annealing at specific AT [Table 1] for 20 sec, and elongation at 72°C for 20 sec). Amplification was followed by melting curve analysis to check PCR efficiency. Standard curves, for the quantification analysis, were generated using serial dilutions of each amplicon. The amounts of gene interest relative to the amount of beta-actin in the same sample were determined and analyzed using Real-Quant 1.0 software (Roche). The results were expressed as a normalized ratio. Primer pairs for Gdf 9 and Bmp15 were those used by Pakarainen and colleagues [22] and Thomas and colleagues [20], respectively, and were purchased from Invitrogen. Blast (URL: http://www.ncbi.nlm.nih.gov/blast) assured that Bmp15 and Gdf 9 primer sets did not cross-react with each other or with any other gene. To check the identity of the amplicons, PCR-amplified products were run on an agarose gel, and each single band was excised and directly sequenced (ABI Prism; 310 Genetic Analyzer; Applied Biosystems, www.appliedbiosystems.com). Sequences of oligonucleotide primer pairs used for real-time PCR experiments and experimental conditions are shown in Table 1.
|
Proteins were extracted by mechanical homogenization of ovaries in the presence of 200 µl ice-cold hypotonic buffer (25 mM Tris-HCl [pH 7.4], 1 mM EDTA, 1% protease inhibitor cocktail [Sigma, www.sigmaaldrich.com]). Protein concentrations were obtained by using a colorimetric Bradford method. Proteins (30 µg) were resolved on 12% SDS/polyacrylamide gels and electrophoretically transferred to nitrocellulose membranes (Hybond-C extra; Amersham) using 25 mM Tris and 185 mM glycine (pH 8.3) containing 20% methanol. The transfer was performed at a constant voltage of 100 V for 2 h. After transfer, the membranes were incubated in blocking buffer (Tris-buffered saline [TBS] buffer containing 10% fat-free dry milk) for 1 h at room temperature. The membranes were rinsed (3 x 10 min) with TBS/Tween 20 0.1% and incubated with the anti-BMP15 primary antibody [5] (dilution 1:500) overnight at 4°C. The membranes were rinsed with TBS/Tween 20 0.1% (3 x 10 min), and then incubated with horseradish peroxidase-labeled rabbit antibody (1:2000). The bound antibody was detected by chemiluminescence using a Covalab kit (Covalab, www.covalab.com) and Biomax MR-1 film (Eastman Kodak Co., www.kodak.com). The protein loading was checked by reprobing the blot with an anti-beta-actin antibody (1:1000; Sigma). Intensity of bands was estimated by using OptiQuant software (Packard). The molecular weight of the protein was determined using biotinylated protein markers (Covalab). This allowed validation of the antibody used.
After embedding in paraffin, 5-µm sections of Bouin-fixed ovaries were mounted on Polysin-coated glass slides (Menzel-Glaser, www.menzel.de). Sections were dewaxed in xylene, rehydrated in a graded series of ethanol solutions, treated for 20 min at 98°C in pH 6 citric buffer, rinsed in osmosed water (2 x 5 min), washed (2 x 5 min) in PBS-Tween 20 0.1%, incubated 10 min at 37°C in the peroxidase blocking reagent (DakoCytomation, www.dakocytomation.fr), and washed (2 x 5 min) in PBS-Tween 20 0.1%. The primary antibody (GDF9, 1:800 dilution; BMP15, 1:200 dilution) was then added and the sections incubated overnight at 4°C. The GDF9 and BMP15 antibodies were provided by Aaron Hsueh (Stanford University, Stanford, CA)[23] and Shunichi Shimasaki (University of California, San Diego, CA) [5], respectively. The GDF9 antibody is specific for the mature form of GDF9 protein [23], whereas the BMP15 antibody recognizes both the proprotein and the mature BMP15 protein, and does not cross-react with the rat GDF9 [5]. After incubation with the primary antibody, the sections were rinsed, washed (2 x 5 min) in PBS-Tween 20 0.1%, and then incubated for 30 min at 37°C in the presence of the secondary antibody attached to a peroxidase-conjugated polymer backbone (Envision+ kit; DakoCytomation). After incubation with the secondary antibody, the sections were rinsed, washed (2 x 5 min) in PBS-Tween 20 0.1%, incubated 3 min at room temperature with 3-amino-9-ethylcarbazole (DakoCytomation), which generated a red color at the site of peroxidase activity, rinsed, and then washed (2 x 5 min) in osmosed water. Sections were counterstained with Mayer's hematoxylin (DakoCytomation), rinsed in 10 g/L lithium chloride, and mounted with Faramount mounting medium (DakoCytomation). Control sections received either buffer or rabbit serum, diluted appropriately, in place of the primary antibody. For each control or treatment day, one ovary from five mice was used, and two to four sections per ovary were examined. The intensity of the immunostaining was graded as weak, moderate, or strong.
Stages of Follicle Development and Morphometric Studies
Follicles were ranged into the following groups: primordial, in which the oocyte was surrounded by some flattened granulosa cells; primary, in which an enlarged oocyte was surrounded by a single layer of cuboidal granulosa cells; preantral follicles; and antral follicles. Preovulatory follicles were the largest (
350 µm) healthy follicles. Follicles were considered as atretic when they displayed more than 10 pycnotic granulosa cells in the largest cross-section. Follicles in late atresia were, the most often, located in the central part of the ovary; their oocytes either displayed meiotic resumption or were fragmented.
Construction of BMP15 Riboprobes
Riboprobes for in situ hybridization studies were prepared by using a pT7T3 plasmid containing a 541 bp-long Bmp15 cDNA sequence from the IMAGE clone BX529299. The plasmid (German Human Genome Project Resource Center [RZPD]; URL: http:// www.rzpd.de) was linearized by Xho1 or Stu1 restriction enzymes. This plamid exhibited no obvious sequence homology with other known TGFB superfamily members, including Gdf 9.
In vitro transcription was performed using this plasmid as DNA template, and a digoxigenin (DIG) RNA labeling kit (Roche Diagnostics), containing T7 and T3 polymerase and DIG-labeled UTP, to synthesize DIG-labeled antisense (positive) and sense (negative control) riboprobes.
Paraffin sections (5 µm) of 4% buffered paraformaldehyde-fixed ovaries were mounted on Polysin-coated glass slides. The sections were dewaxed in xylene, rehydrated in graded ethanol and PBS, and permeabilized 10 min at room temperature with proteinase K (DakoCytomation) 1/25 in 50 mM Tris (pH 7.5). Sections were prehybridized for 60 min at 55°C in a humidified chamber with hybridization buffer containing 50% deionized formamide, 2.5% dextran sulfate (Sigma), 4x sodium chloride/sodium citrate (SSC), 5x Denhardt solution, 250 µg/ml herring sperm DNA (Promega) and 250 µg/ml yeast tRNA (Roche Diagnostics). Slides were incubated overnight with sense or antisense DIG-labeled riboprobes diluted in hybridization buffer in a humid chamber at 55°C. Concentrations of the RNA probes in the hybridization buffer were 0.5–1.0 µg/ml. After hybridization, the sections were washed at 37°C in 2x SSC and 1x SSC for 30 min each, treated with 20 µg/ml ribonuclease A (Roche Diagnostics) for 30 min, then washed for 60 min in 50% formamide 0.5 x SSC at 37°C.
Immunological detection was then performed according to the manufacturer's instructions (Roche Diagnostics) using an alkaline phosphatase-conjugated anti-DIG antiserum (1:500 dilution). The labeled probes were visualized with nitroblue tetrazolium chloride/5 bromo-4-chloro-3-indolyl phosphate (Roche Diagnostics) as substrates for 60 min in the dark. Sections were counterstained with Mayer hematoxylin and mounted with Faramount mounting medium. The brown-red precipitates seen in Figure 4, indicate the presence of the mouse Bmp15 mRNA.
Differences in both Gdf 9 and Bmp15 mRNA levels and BMP15 protein levels between control and treatment groups were determined by one-way ANOVA followed by the post hoc Scheffé test (Statview Software). Differences were considered to be significant at P < 0.05.
Effects of r-hFSH and hCG Treatments on Steady-State Gdf 9 and Bmp15 mRNA Expression Levels in Mouse Ovaries
Although the normalized Gdf 9/Actb mRNA ratio was higher in COCs than in total ovarian homogenates (around 2.5-fold), the Gdf 9 mRNA levels were not significantly different between control animals and gonadotropin-treated mice in both total ovaries (Fig. 1A) and COCs (Fig. 2A).
|
|
In control mice, no significant changes in the steady-state Bmp15 mRNA expression levels were observed in either whole ovaries (Fig. 1B) or in COCs (Fig. 2B). Whereas these levels were unchanged at F24, they were significantly (P < 0.001) higher than the control values at F48 in both total ovaries (Fig. 1B) and COCs (Fig. 2B). At F72, the Bmp15 mRNA levels were significantly (P < 0.001) higher than those at F48 in both total ovaries (Fig. 1B) and COCs (Fig. 2B), and reached their highest value. At H24, the Bmp15 mRNA levels in both total ovaries (Fig. 1B) and COCs (Fig. 2B) significantly dropped (P < 0.001) to reach those observed at F48, and at H48, these levels dropped again; however, although they were significantly lower (P < 0.001) than those observed at H24, they remained significantly higher (P < 0.001) than F24 and control values. Similar to the situation for Gdf 9, the normalized Bmp15/Actb mRNA ratio was higher in COCs than in total ovaries.
Effects of r-hFSH and hCG Treatments on BMP15 Protein Levels in Mouse Ovaries
Mouse kidney proteins were subjected to SDS-PAGE immunoblotting analysis using the anti-recombinant human BMP15 [5] as negative controls. Because no band was obtained, it can be assumed that this antibody does not cross-react with proteins present in the kidney, especially BMP6 that is present in both kidney tissues [24] and oocytes [25], and, consequently, specifically binds both the BMP15 proprotein and mature protein. In control mice, from Day 21 to Day 26 there were no significant changes in either the BMP15 proprotein or the mature BMP15 protein levels in total ovaries (Fig. 3). Gonadotropin treatment did not induce any changes in the BMP15 proprotein levels when compared with controls. At F48 and F72, the mature BMP15 protein levels were significantly higher (P < 0.01) than those in controls and F24 samples, but were not significantly different from each other. The treatment with hCG did not induce any further changes in mature BMP15 protein levels when compared with the F72 value (Fig. 3).
|
In Situ Hybridization for Bmp15
Hybridization with the sense Bmp15 probe did not show any background signal (Fig. 4B) when compared with the use of the antisense Bmp15 probe (Fig. 4A). The Bmp15 transcript was present only in growing oocytes from both control and treated mice. Whatever the stage of follicle development, the immunostaining was strong in oocytes from healthy follicles, including oocytes in preovulatory follicles showing an expanded cumulus (Fig. 4C). Interestingly, oocytes from follicles in late atresia did not show any Bmp15 signal (Fig. 4D). Neither r-hFSH nor hCG induced any visible change in either the localization or the intensity of Bmp15 mRNA staining in growing oocytes.
|
Localization of GDF9 and BMP15 Proteins—Effects of r-hFSH and hCG Treatments
The replacement of the primary antibody by the antibody diluent or rabbit normal serum led to the absence of any immunostaining in all ovarian tissues (data not shown).
In immature control mice, a moderate to strong immunostaining for both GDF9 and BMP15 was detected in growing oocytes from the secondary follicular stage; the immunostaining was weak in primary oocytes and absent in primordial oocytes (Fig. 5, A and B). After FSH treatment, no changes were observed for the ovarian GDF9 immunostaining (Fig. 5C), but strong changes were observed for BMP15 in atretic antral follicles. Although the BMP15 immunostaining was unchanged in healthy follicles in which the oocyte was the only positive cell with a regular staining, the follicular fluid of almost all atretic antral follicles was strongly stained (Figs. 5D and 6). In addition, roughly half of the oocytes in these atretic follicles displayed a heterogeneous BMP15 staining with clusters of strongly stained spots (Fig. 7), whereas 78% of oocytes in late atretic follicles displayed this characteristic. After hCG injection, the GDF9 immunostaining remained moderate to strong in the preovulatory oocyte, whereas the expanded cumulus cells were not stained (Fig. 5E). In preovulatory follicles showing an expanded cumulus, the extracellular matrix of inner granulosa and cumulus cells was strongly stained in the presence of the BMP15 antibody (Fig. 5F), whereas, in preovulatory follicles that did not respond to hCG and still possessed a compact cumulus, the cumulus was not stained (Fig. 5G).
|
|
|
The present study shows that the treatment of immature mice with r-hFSH and hCG, effectively induced follicular growth and ovulation, as evidenced by normal numbers of corpora lutea per ovary at H48. It also shows, for the first time, that in the mouse, BMP15 and GDF9, although both produced by the oocyte, do not react similarly to in vivo treatment with gonadotropins.
The steady-state Gdf 9 mRNA expression levels were higher in COCs than in total ovaries, likely because of exclusive expression of Gdf 9 in the oocyte, but were not altered by FSH and hCG in both total ovaries and COCs.
Similar to Gdf 9, exclusive expression of Bmp15 in the oocyte can explain the higher levels of its mRNA in COCs than in total ovaries. However, in contrast to Gdf 9, the Bmp15 mRNA levels started to increase above the control levels from F48 in both total ovaries and COCs, reached their highest levels at F72, and then decreased after hCG; in addition, the levels of BMP15 mature protein also started to increase at F48. These results agree with those of Thomas and colleagues [20] concerning the absence of FSH effect on in vitro production of GDF9 by preantral COCs from immature mice. Nevertheless, they appear to be in conflict with the observation that FSH, at 5 ng/ml, negatively regulates Bmp15 mRNA expression in the same preantral COCs [20], a discrepancy that could be explained by differences in doses of hormones, stage of follicular development, and in vitro conditions. Interestingly, it has been recently reported that mRNAs of BmprIB and BmprII, as well as those of Smads 1 and 5, were upregulated by FSH (10 ng/ml) in a human cell-line model [26]. The present observation that Bmp15 mRNA levels increased in gonadotropin-treated immature mice suggests that the complete BMP15 system could be positively regulated by FSH, as previously shown for the TGFbeta2 system [27].
As BMP15 is exclusively produced by the oocyte, the question of whether FSH directly or indirectly regulates Bmp15 expression and production remains to be answered. A direct effect of FSH on the oocyte cannot be excluded, as the presence of FSHR on the oocyte cell surface was reported in mouse [28], porcine [29], and human oocytes [29, 30]. However, the delay between the first injection of FSH and the observation of its effect on Bmp15 suggests that the effect of FSH could be indirect. In 21-d-old mice, preantral follicles contain partly grown oocytes and constitute the largest part of the follicular population. On the one hand, Joyce and colleagues [31] have observed that mouse oocytes at this stage of development decrease Kl mRNA levels in FSH-treated preantral granulosa cells. On the other hand, it was reported that KL downregulates Bmp15 expression in the rat [13], and that FSH-stimulated KL results in suppression of BMP15 in the mouse [20]. Consequently, it is tempting to suggest that a decreased production of KL could have been subsequently followed by the increase of both Bmp15 mRNA and BMP15 mature protein levels in total ovaries and COCs at F48. At F72, follicles have reached the preovulatory stage and their oocytes are fully grown. At this stage of maturation, oocytes reduce Kl1 and Kl2 expression in cumulus cells [31], which could be followed by a subsequent increase of Bmp15 expression, and may explain the highest Bmp15 mRNA levels in total ovaries and COCs at F72. Whether, these high levels result from a gonadotropic effect, or are subsequent to FSH-induced follicle growth and oocyte cytoplasmic and/or nuclear maturation, remains, however, to be determined.
In response to FSH, the levels of the mature BMP15 protein significantly increased at F48; however, contrary to Bmp15 mRNA levels, those of the mature BMP15 protein did not exhibit any further changes. Although the levels of the mature BMP15 protein increased, those of the BMP15 proprotein did not exhibit any significant changes. Consequently, the possibility cannot be excluded that, in addition to its effect on steady-state Bmp15 mRNA levels, FSH upregulates the enzyme that cleaves the BMP15 proprotein, leading to the observed results. The BMP15 antibody used in the present study is an anti-recombinant human BMP15 tagged with glutathione S-transferase [5]. Despite the fact that the homology between the mouse and human BMP15 is low (76%) when compared with that of GDF9 (96%) [3], this antibody specifically binds the rat BMP15 [5], and, as the homology between the rat and mouse BMP15 is 93% [3, 32], it appears likely that this antibody specifically binds the mouse BMP15. In addition, negative controls performed in the present study have shown that this antibody specifically binds both mouse BMP15 proprotein and mature protein. Our results are in conflict with those by Yoshino and colleagues [16], who detected neither the BMP15 proprotein in mouse oocyte lysates nor the mature protein before 9 h after hCG administration to eCG-primed immature mice. This discrepancy might be attributable to either an inability of their antibody to bind BMP15 proprotein or the use of different treatments: r-hFSH vs. eCG (LH + FSH).
In control mice, both GDF9 and BMP15 immunostaining were restricted to growing oocytes, as previously described [2, 3, 19]. In response to gonadotropins, neither changes of the GDF9 immunostaining, whatever quality and stage of follicle development, nor of the BMP15 immunostaining in healthy, nonovulatory follicles were observed. However, localization of BMP15 was dramatically altered in atretic antral follicles. When compared with control animals, the follicular fluid of atretic follicles was strongly stained as early as F24. Presence of BMP15 within the antral cavity cannot be related to degenerative alteration of the oocyte membrane permeability, as it was not observed in control ovaries in which numerous atretic follicles were observed. In mouse antral follicles, aromatase, which converts testosterone into estradiol, is expressed in mural granulosa cells [33]. During follicular atresia, aromatizing capacity of granulosa cells is lost [34] and antral atretic follicles mainly produce testosterone. Interestingly, testosterone can increase Kl mRNA in mouse mural granulosa cells, an effect that is enhanced by FSH [31]. Consequently, increased production of KL after FSH treatment could be responsible for the observed arrest of Bmp15 expression [13], as shown by the present in situ hybridization experiments showing that oocytes in atretic follicles no longer synthesize any Bmp15 mRNA. In the present immunohistochemical study, most oocytes in atretic follicles were rather less stained than oocytes from healthy follicles, and, above all, contained small clusters of positively BMP15-stained spots, whereas the staining was homogeneous in all healthy oocytes. Whether this follicular fluid staining can be related to arrest of BMP15 production and escape within the follicular fluid remains unknown.
Furthermore, at H24, a surprising BMP15 immunostaining was observed in preovulatory follicles. In those showing an expanded cumulus, the extracellular matrix of the cumulus and inner granulosa cells was strongly stained, whereas in preovulatory follicles with a compact cumulus, no BMP15 immunostaining was seen in cumulus cells. Interestingly, after hCG injection, expanded cumuli of preovulatory follicles exhibited no GDF9 immunostaining. Clearly, the present study shows that after hCG injection, BMP15 was released from the oocyte to the cumulus extracellular matrix at the time of cumulus expansion. The drop of Bmp15 mRNA at H24 and then at H48 is likely due to the disappearance of mature oocytes that ovulated in response to hCG.
It has been previously demonstrated that, just before ovulation and in response to the ovulatory surge, mouse oocytes produce a soluble cumulus expansion-enabling factor (CEEF) that is absolutely required for cumulus expansion [35]. To date, the exact identity of the CEEF remains unknown. GDF9 was hypothesized to be the CEEF [36, 37]; however, in vitro experiments have suggested that GDF9 alone is not the oocyte-secreted factor promoting cumulus expansion [38]. The results of the present study, showing that there are neither changes in steady-state Gdf 9 mRNA expression and protein levels nor GDF9 release at the time of cumulus expansion, would rather argue in favor of this latter view. In contrast to Gdf 9, Bmp15 mRNA and mature BMP15 protein levels increased in parallel with FSH-induced follicular growth and oocyte maturation, a BMP15 immunostaining being detected in the extracellular matrix of expanded, but not in compact, cumuli after hCG. Are these observations evidence that BMP15 might be the CEEF? In BmprIB-null mice, defective cumulus expansion was observed [17]. As BMP15 is an oocyte-secreted factor that binds to BMPRIB [25], it is tempting to suggest that BMP15 may be constitutive of the CEEF. Recombinant hBMP15 does not result in in vitro expansion of oocytectomized cumulus complexes [11], but BMP15 stimulates in vitro mouse cumulus expansion, an effect that requires epidermal growth factor receptor signaling [16].
In summary, the present study shows that, in immature mice in vivo, gonadotropin-induced follicular growth and ovulation was not accompanied by changes in Gdf 9 expression and GDF9 production. In contrast to GDF9, expression of Bmp15 and production of BMP15 increased in oocytes as they matured in response to gonadotropins. In addition, the present observation, that the BMP15 protein is released from the oocyte to the cumulus extracellular matrix at the time of cumulus expansion, constitutes additional evidence to support the view that BMP15 is strongly involved in preovulatory cumulus expansion.
ACKNOWLEDGMENTS
We thank Professors Aaron Hsueh and Shunichi Shimasaki for their kind gift of GDF9 and BMP15 antibodies, Catherine Deschildre, Nicole Riu, Pierre Contard, and Aurélien Delangle for their technical assistance, and Anne Florin-McLeer for her critical reading of the manuscript.
FOOTNOTES
Correspondence: 2 Alain Gougeon, INSERM U-407, Faculté de Médecine Lyon-Sud, Centre hospitalier Lyon-Sud, 165 chemin du Grand Revoyet, 69310 Pierre Bénite, France. FAX: 33 04 26 23 59 16; e-mail: gougeon{at}lyon.inserm.fr
Received: 13 July 2006.
First decision: 4 August 2006.
Accepted: 21 August 2006.
REFERENCES
-hydroxylase/C17–20 lyase and aromatase cytochrome P450 in the human ovary during the menstrual cycle. J Endocrinol 1992; 135:589–595[Abstract]This article has been cited by other articles:
|
|
 |
|
  J. E. Swain and T. B. Pool ART failure: oocyte contributions to unsuccessful fertilization Hum. Reprod. Update, September 1, 2008; 14(5): 431 - 446. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J C Sadeu, T Adriaenssens, and J Smitz Expression of growth differentiation factor 9, bone morphogenetic protein 15, and anti-Mullerian hormone in cultured mouse primary follicles Reproduction, August 1, 2008; 136(2): 195 - 203. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. E. McMahon, O. Hashimoto, P. L. Mellon, and S. Shimasaki Oocyte-Specific Overexpression of Mouse Bone Morphogenetic Protein-15 Leads to Accelerated Folliculogenesis and an Early Onset of Acyclicity in Transgenic Mice Endocrinology, June 1, 2008; 149(6): 2807 - 2815. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. B. Gilchrist, M. Lane, and J. G. Thompson Oocyte-secreted factors: regulators of cumulus cell function and oocyte quality Hum. Reprod. Update, March 1, 2008; 14(2): 159 - 177. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. X. Yeo, R. B. Gilchrist, J. G. Thompson, and M. Lane Exogenous growth differentiation factor 9 in oocyte maturation media enhances subsequent embryo development and fetal viability in mice Hum. Reprod., January 1, 2008; 23(1): 67 - 73. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y.-Q. Su, K. Sugiura, K. Wigglesworth, M. J. O'Brien, J. P. Affourtit, S. A. Pangas, M. M. Matzuk, and J. J. Eppig Oocyte regulation of metabolic cooperativity between mouse cumulus cells and oocytes: BMP15 and GDF9 control cholesterol biosynthesis in cumulus cells Development, January 1, 2008; 135(1): 111 - 121. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Sugiura, Y.-Q. Su, F. J. Diaz, S. A. Pangas, S. Sharma, K. Wigglesworth, M. J. O'Brien, M. M. Matzuk, S. Shimasaki, and J. J. Eppig Oocyte-derived BMP15 and FGFs cooperate to promote glycolysis in cumulus cells Development, July 15, 2007; 134(14): 2593 - 2603. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. A. Dragovic, L. J. Ritter, S. J. Schulz, F. Amato, J. G. Thompson, D. T. Armstrong, and R. B. Gilchrist Oocyte-Secreted Factor Activation of SMAD 2/3 Signaling Enables Initiation of Mouse Cumulus Cell Expansion Biol Reprod, May 1, 2007; 76(5): 848 - 857. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
