Growth Differentiation Factor-9 Induces Smad2 ... - Oxford Journals

0013-7227/03/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 88(2):755–762 Copyright © 2003 by The Endocrine Society doi: 10.1210/jc.2002-021317

Growth Differentiation Factor-9 Induces Smad2 Activation and Inhibin B Production in Cultured Human Granulosa-Luteal Cells ¨ MA ¨ RA ¨ INEN, JANNE KOSKIMIES, NOORA KAIVO-OJA, JONAS BONDESTAM, MEERIT KA URSULA VITT, MARK CRANFIELD, KAISA VUOJOLAINEN, JANNE P. KALLIO, VESA M. OLKKONEN, MASARU HAYASHI, ARISTIDIS MOUSTAKAS, NIGEL P. GROOME, PETER TEN DIJKE, AARON J. W. HSUEH, AND OLLI RITVOS Program for Developmental and Reproductive Biology, Biomedicum Helsinki, and Departments of Bacteriology and Immunology, Haartman Institute (N.K.-O., J.B., M.K., J.K., K.V., J.P.K., O.R.), University of Helsinki, 00014 Helsinki, Finland; Division of Reproductive Biology, Department of Gynecology and Obstetrics, Stanford University (U.V., M.H., A.J.W.H.), Palo Alto, California 94305; School of Biological and Molecular Sciences, Oxford Brookes University (M.C., N.P.G.), Headington, Oxford, United Kingdom OX3 0BP; Department of Molecular Medicine, National Public Health Institute, Biomedicum Helsinki (V.M.O.), 00251 Helsinki, Finland; Ludwig Institute for Cancer Research, Uppsala Branch (A.M.), SE-752 37 Uppsala, Sweden; and Department of Cellular Biochemistry, Netherlands Cancer Institute (P.t.D.), 1066 CX Amsterdam, The Netherlands The TGF␤ family member growth differentiation factor-9 (GDF-9) is an oocyte-derived factor that is essential for mammalian ovarian folliculogenesis. GDF-9 mRNAs have been shown to be expressed in the human ovarian follicle from the primary follicle stage onward, and recombinant GDF-9 has been shown to promote human ovarian follicle growth in vitro. In this study with primary cultures of human granulosa-luteal (hGL) cells, we investigated whether recombinant GDF-9 activates components of the Smad signaling pathways known to be differentially activated by TGF␤ and the bone morphogenetic proteins (BMPs). As with TGF␤, GDF-9 treatment caused the phosphorylation of endogenous 53-kDa proteins detected in Western blots with antiphospho-Smad2 antibodies (␣PS2). However, unlike BMP-2, GDF-9 did not activate the phosphorylation of antiphospho-Smad1 antibody (␣PS1)-immunoreactive proteins in hGL cells. Infection of hGL cells with an adenovirus expressing Smad2 (Ad-Smad2) confirmed that GDF-9 activates specifically phosphorylation of the Smad2 protein.

Infection of hGL cells with Ad-Smad7, which expresses the inhibitory Smad7 protein, suppressed the levels of both GDF9-induced endogenous and adenoviral ␣PS2-reactive proteins. Furthermore, GDF-9 increased the steady state levels of inhibin ␤B-subunit mRNAs in hGL cells and strongly stimulated the secretion of dimeric inhibin B. Again, Ad-Smad7 blocked GDF-9-stimulated inhibin B production in a concentration-dependent manner. We identify here for the first time distinct molecular components of the GDF-9 signaling pathway in the human ovary. Our data suggest that GDF-9 mediates its effect through the pathway commonly activated by TGF␤ and activin, but not that activated by many BMPs. The results are also consistent with the suggestion that in addition to endocrine control of inhibin production by gonadotropins, a local paracrine control of inhibin production is likely to occur via oocyte-derived factors in the human ovary. (J Clin Endocrinol Metab 88: 755–762, 2003)

G

ROWTH DIFFERENTIATION factor-9 (GDF-9) was first identified in the early 1990s as a member of the TGF␤ family (1). GDF-9 is expressed mainly in the oocytes of ovarian follicles (2), although some testicular and hypophyseal expression has been detected (3). Gene ablation studies in mice show that the essential role of GDF-9 is restricted to the regulation of ovarian function. Female mice deficient in GDF-9 are infertile and fail to demonstrate any normal follicles beyond the primary one-layer follicle stage, whereas GDF-9-deficient male mice are fertile (4). Recombinant GDF-9 has been shown to mimic many of the actions the oocytes exert on follicular somatic cells (reviewed in Refs. 5

and 6). Recently, we and others have shown that GDF-9 together with its closely related homolog, GDF-9B/bone morphogenetic protein-15 (BMP-15) (7–9), are expressed in the human ovary from the primary follicular stage onward (10, 11). Under in vitro conditions, recombinant GDF-9 promotes effectively early follicular growth in humans and rats (12, 13) and also in vivo in the rat model (14). In the polycystic ovary syndrome (PCOS), excessive formation of small follicles occurs in the ovary, and in the most severe form of PCOS follicles fail to develop beyond the small antral follicle stage. GDF-9 mRNAs were recently shown to be more weakly expressed in the ovaries of PCOS patients than in normal ovaries, whereas GDF-9B/BMP-15 levels were unchanged (11). These studies suggest that a balanced expression of GDF-9 and GDF-9B/BMP-15 might be important for normal ovarian folliculogenesis. Although previous studies have provided important new information on GDF-9-regulated cellular events, the receptor systems and downstream signaling pathways used by GDF-9

Abbreviations: ActR-II, Activin receptor type II; ALK, activin receptor-like kinase; BMP-2, bone morphogenetic protein-2; BMPR, bone morphogenetic protein receptor; FCS, fetal calf serum; GDF-9, growth differentiation factor-9; hGL, human granulosa-luteal; MOI, multiplicity of infection; PCOS, polycystic ovary syndrome; ␣PS1, antiphospho-Smad1 antibody; ␣PS2, antiphospho-Smad2 antibody; R-Smad, receptor-regulated Smad protein.

755

756

J Clin Endocrinol Metab, February 2003, 88(2):755–762

have not been determined. Other TGF␤ family members are known to mediate their signals across the plasma membrane by binding to and activating distinct combinations of heteromeric complexes of specific type I and type II serine/ threonine kinase receptors (reviewed in Ref. 15). Activated type I receptors initiate intracellular signaling through the phosphorylation of specific receptor-regulated Smad proteins (R-Smads). TGF␤ and activin can induce the phosphorylation of Smad2 and Smad3, whereas the phosphorylation of Smad1, Smad5, and Smad8 is induced by the BMPs. The activated Smads form heteromeric complexes with a common partner Smad4 and translocate into the nucleus where they, together with certain transcription factors and coregulators, modulate target gene expression. The inhibitory Smads, Smad6 and Smad7, antagonize the R-Smads either by competing with R-Smads for binding to type I receptors (Smad7) (16) or by preventing complex formation between activated R-Smads and Smad4 (Smad6) (17). Smad7 acts as a common inhibitor for R-Smads, whereas Smad6 appears to act more specifically as an inhibitor for BMP signaling (reviewed in Ref. 15). We recently reported that activin, TGF␤, and BMP-2 regulate inhibin ␤B-subunit mRNA expression and inhibin B production in human granulosa-luteal (hGL) cells (18 –20) and that the mRNAs for Smad proteins are expressed in these cells (20). Here we report how GDF-9 affects Smad-mediated signaling pathways and inhibin B production in hGL cells. Materials and Methods Reagents and supplies Recombinant BMP-2 and TGF␤1 were purchased from R&D Systems, Inc. (Minneapolis, MN). Fetal calf serum (FCS) was purchased from Euroclone Ltd. (Devon, UK). DMEM and Ham’s F-12 were purchased from Life Technologies, Inc. (Gaithersburg, MD). Anti-Flag M2 monoclonal antibody and puromycin were purchased from Sigma-Aldrich (St. Louis, MO). Smad1 and Smad2 activation was detected with rabbit antiphospho-Smad1 (␣PS1) and antiphospho-Smad2 (␣PS2) antibodies as previously described (21, 22). Heparin (Fragmin) was purchased from Pharmacia & Upjohn, Inc. (Stockholm, Sweden). BSA was purchased from Roche Molecular Biochemicals (Mannheim, Germany). Hybond C and Hybond N membranes were purchased from Amersham Pharmacia Biotech (Little Chalfont, UK). Fugene 6 was purchased from Roche Molecular Biochemicals (Indianapolis, IN). Peroxidase-conjugated AffiniPure goat antirabbit IgG and rabbit antimouse IgG were purchased from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA).

Kaivo-oja et al. • GDF-9 Induces Smad2 Activation and Inhibin B Production

Generation of monoclonal antibodies for GDF-9 A synthetic peptide to annotated amino acids 420 – 450 of the Cterminal part of human GDF-9 (GenBank accession no. NP 005251) of sequence VPAKYSPLSVLTIEPDGSIAYKEYEDMIATKC was made using F-moc chemistry, coupled to tuberculin through the cysteine thiol using a heterobifunctional agent, and used to immunize female BALB/c mice by standard methods. After an initial immunization and two boosts at monthly intervals, the sera of the mice collected by tail bleed were tested against ELISA wells coated directly with the immunizing peptide. High responding mice were boosted iv, and 4 d later the spleens were removed and used for fusion to SP2/0 splenocytes by standard methods. Positive hybridomas were tested in Western blot against recombinant GDF-9. Monoclonal antibody 37 was purified by protein A chromatography using a high salt protocol. The antibody has minimal crossreaction with preparations of recombinant GDF-9B.

Cell culture and stimulations hGL cells were obtained with permission from women undergoing in vitro fertilization treatments. Granulosa cells from one to six patients were pooled and extracted as described previously (24). For Smad activation experiments, 200,000 –300,000 hGL cells/well were plated on 6-well plates, and for inhibin B assays, 30,000 – 40,000 cells/well were plated on 24-well plates. The HepG2 cells were plated on 6-well plates at a density of 500,000 cells/well and were cultured overnight for control Smad activation experiments. Both cells were cultured in DMEM supplemented with 10% FCS at 37 C in a humidified atmosphere of 5% CO2 in air. hGL cells were stimulated with 3–300 ng/ml GDF-9, 10 –25 ng/ml TGF␤, or 25 ng/ml BMP-2, depending on the experiment, 2–3 d after plating. When recombinant adenoviruses were used, the cells were stimulated with the ligands 24 h postinfection. For Smad activation experiments, the cells were incubated in low serum medium containing 0.5% FCS for 3 h before stimulations. Stimulations for Smad activation were made in fresh low serum medium. For inhibin B assays the cells were washed once with PBS (pH 7.4) and stimulated in DMEM containing 2% FCS (the inclusion of FCS is essential for maintaining good cell viability up to 96 h of culture).

Recombinant adenoviruses Generation of the recombinant Smad adenoviruses used in this study has been previously described (25), and their use in hGL cultures has been recently optimized (26). These serotype 5 adenoviruses express Smad proteins that contain a Flag tag in their N terminus. All viruses were amplified and titrated in transcomplemental 293T cells and purified with cesium chloride gradient ultracentrifugation as described previously (27). The purified viruses were stored at –70 C with 10% glycerol in PBS. Before adenovirus infections, hGL cells were cultured for 2–3 d. The hGL cells were infected by incubating them with the virus(es) at 37 C in serum-free DMEM supplemented with l-glutamine and antibiotics for 1 h, and DMEM containing 2% FCS was added on top to stop the infection. The cells were then incubated for 24 h before continuing the experiments. The multiplicity of infection (MOI) used in all experiments varied from 0.1–30. The expression of adenovirally produced Smads was detected with monoclonal anti-Flag M2 antibody.

Expression of mouse and rat recombinant GDF-9 At the initial stages of this study we used tagged and untagged recombinant rat GDF-9 that have been previously described (12). A cell line expressing fully processed mouse GDF-9 was also developed and used as an abundant source of bioactive recombinant GDF-9 protein. Mouse GDF-9 full-length cDNA (7) was subcloned into pEFIRES-P expression vector (23) and transfected into a human embryonic kidney cell line, HEK-293T, by Fugene 6 transfection reagent. Cells expressing high levels of the recombinant protein were selected with increasing concentrations of puromycin in DMEM supplemented with 10% FCS, 2 mmol/liter l-glutamine, 100 IU/ml penicillin, 100 ␮g/ml streptomycin, and 0.25 ␮g/ml amphotericin B. The recombinant protein was produced into serum-free harvesting medium (DMEM/Ham’s F-12, 1:1) supplemented with l-glutamine and antibiotics, 0.01% (vol/vol) BSA, and 100 ␮g/ml heparin. Levels of recombinant mouse GDF-9 were estimated in immunoblots using the purified N-tagged rat GDF-9 as a standard (12).

Smad activation experiments and Western blot analysis The cells were cultured for 48 –72 h before endogenous Smad activation experiments or adenoviral infections. Infected cells were stimulated 24 h after infections with the stimulant concentrations mentioned above. Control cells were treated with conditioned medium. The cells were incubated in low serum medium (0.5% FCS) for 3 h before stimulations. Stimulations were performed in fresh low serum medium with stimulation times varying from 15 min to 3 h, and cells were thereafter recovered in ice-cold PBS and subsequently lysed in reducing SDSPAGE sample buffer. The samples were sonicated for 10 sec (amplitude, 5 ␮m) as described previously (28), run in 10% SDS-PAGE gels, and blotted onto a Hybond C nitrocellulose membrane. The membranes were blocked for 1 h in 2.5% (wt/vol) nonfat dried milk in Tris-buffered saline/Nonidet P-40 [20 mm Tris-HCl (pH 7.5), 150 mm NaCl, and 0.1% Nonidet P-40] and incubated with the primary antibody in 2.5% milk in


J Clin Endocrinol Metab, February 2003, 88(2):755–762 757

Tris-buffered saline/Nonidet P-40 overnight at 4 C (dilution, 1:15 000 for ␣PS1 and ␣PS2). After washing, the membranes were incubated with the secondary antibody, a peroxidase-conjugated anti-IgG (Jackson ImmunoResearch Laboratories, Inc.; 1:15,000), for 1 h. Immunoreactive proteins were detected using enhanced chemiluminescence reagents (Amersham Pharmacia Biotech). The adenoviral Smad proteins were detected with an ␣-Flag M2 monoclonal antibody according to the manufacturer’s protocol.

RNA isolation and Northern blotting Cytoplasmic RNAs from cultured hGL cells were extracted with the modified Nonidet P-40 lysis procedure (24, 29). RNA was quantitated by absorbance measurement at 260 nm. For Northern blots, 10 –20 ␮g of the RNAs were size-fractionated in 1.5% agarose gels. Before transfers, even loading of the gels was checked based on RNA fluorescence at 254 nm, after which the RNAs were transferred to Hybond N nylon membranes.

Preparation and labeling of cDNA probes and filter hybridizations For the detection of inhibin ␣- and ␤B-subunit mRNAs by Northern blot hybridization, double- and single-stranded cDNA probes were prepared as previously described (24, 29). A human cyclophilin cDNA was used as a loading control for the experiment (29). Northern blot hybridizations were performed for 16 h at 42 C, and the filters were washed three times for 20 min each time with 0.1–1⫻ standard saline citrate/1% sodium dodecyl sulfate at 50 C. Thereafter, filters were exposed to a Fujifilm Ip-Reader Bio-Imaging Analyzer Bas 1500 (Fuji Photo Film Co., Ltd., Tokyo, Japan).

Inhibin B ELISA hGL cells were cultured on 24-well plates in DMEM supplemented with 10% FCS for 48 h at 37 C in 5% CO2 in air before infections and stimulations. The infections were made as described above, with MOI values ranging from 0.1–30. The cells were carefully washed with PBS before adding the stimulants. The stimulations were made in DMEM supplemented with 2% FCS 24 h postinfection with the stimulant concentration mentioned above. The spent media were harvested after 24 –96 h of incubation, and dimeric inhibin B concentrations were quantified from the medium using an inhibin B ELISA from Serotec (Oxford, UK) together with a signal amplification kit (Life Technologies, Inc.).

Results Recombinant GDF-9 induces ␣PS2-immunoreactive proteins in hGL cells

We studied first whether either the activin/TGF␤ or the BMP-activated Smad pathway would be activated by GDF-9 in hGL cells. We used both recombinant rat and mouse GDF-9 in these experiments. Figure 1A shows that a high affinity monoclonal antibody raised against human GDF-9, mAb-37, readily detects both rat and mouse recombinant GDF-9 in Western blots analysis of the culture medium of GDF-9-expressing 293T cells. The mouse GDF-9 expressed in 293T cells appears to be fully processed, whereas the rat GDF-9 is a mixture of the processed mature region and the unprocessed protein precursor (12). However, untagged rat and mouse GDF-9 are equally potent stimulants when similar amounts of processed mature region proteins are present in the 293T-conditioned medium. The activation of different Smads was followed by Western blot analyses using ␣PS1 and ␣PS2 antibodies, which detect Smad1 and Smad2 proteins, respectively, that are C-terminally phosphorylated by distinct ligand-activated type I Ser/Thr kinase receptors. These antibodies are specific to the phosphorylated forms of the Smad proteins and do not recognize the unphosphory-

FIG. 1. GDF-9 induces endogenous ␣PS2-immunoreactive proteins in cultured hGL cells. Recombinant GDF9 proteins were expressed in 293T cells and detected with anti-GDF-9 monoclonal antibodies raised against hGDF-9 peptide conjugates as described in Materials and Methods. hGL cells were first cultured for 2–3 d and thereafter stimulated with various GDF-9 concentrations for the indicated time periods and analyzed by ␣PS2 or ␣PS1 Western blotting. A, Western blot analysis of the titration of recombinant mouse GDF-9 against purified N-terminally tagged recombinant rat GDF-9. One and 3 ng N-terminally tagged recombinant rat GDF-9 were loaded on SDSPAGE gels together with (from left to right) 1:10, 1:5, 1:2, and 1:1 dilutions of mGDF-9, and Western blotting was performed using mAb 37. The untagged mouse GDF-9 mature region monomer is indicated in the Western blot by arrowheads (the 14 amino acid longer tagged rat GDF-9 migrates more slowly in the neighboring lanes). B, ␣PS2 Western blot analysis of hGL cells stimulated for increasing lengths of time with 150 ng/ml recombinant GDF-9 (top panel) or treated with increasing doses of GDF-9 (10, 20, 30, 60, 100, and 150 ng/ml) for 75 min (lower panel). C, ␣PS2 (top panel) and ␣PS1 (lower panel) Western blot analysis of hGL cells stimulated with conditioned control medium, TGF␤ (25 ng/ml), BMP-2 (25 ng/ml), or GDF-9 (150 ng/ml) for 75 min. The phosphorylation of Smad2 and Smad1 was detected in Western blots with ␣PS2 and ␣PS1 antibodies, respectively. The specific 53-kDa ␣PS1- and ␣PS2-reactive bands are shown by arrows.

lated proteins (21, 22). With GDF-9 stimulation, ␣PS2-reactive bands first appeared after 30 min of stimulation, and the strongest effect was detected at 75 min (Fig. 1B, top panel). In

758



control samples harvested at the same time points, no ␣PS2 immunoreactivity was detected (data not shown). GDF-9 induced the appearance of ␣PS2-immunoreactive proteins in a concentration-dependent manner, with maximal effects seen with 150 ng/ml GDF-9 (Fig. 1B, lower panel). The upper panel of Fig. 1C shows that, like TGF␤, GDF-9 induced the appearance of ␣PS2-immunoreactive protein. The lower panel of Fig. 1C shows that neither GDF-9 nor TGF␤ induced the appearance of ␣PS1-immunoreactive proteins, whereas the positive control BMP-2 did, as previously described (26). The ␣PS2- and ␣PS1-immunoreactive bands correspond to the expected molecular masses (⬃3 kDa) of the phosphorylated forms of Smad2 and Smad1, respectively. Immunofluorescence staining of ␣PS2-reactive proteins showed that in control cells phosphorylation of the putative Smad2 is practically absent without ligand stimulation, whereas cells stimulated for 60 min with GDF-9 or TGF␤ show nuclear staining, but a part of the ␣PS2 immunoreactivity remains in the cytoplasm (data not shown). Ad-Smad7 inhibits GDF-9-induced phosphorylation of both endogenous and adenovirally expressed Smad2 proteins in hGL cells

To verify that GDF-9 is actually able to induce phosphorylation of the Smad2 protein, we introduced an exogenous Smad2 gene into hGL cells through adenovirus infection. Adenovirus-mediated transduction is a very effective method for introducing genes of interest into primary hGL cells, which are otherwise very difficult to transfect. Close to 100% transduction efficiencies in hGL cells can be reached, as detected by green fluorescent protein expressing adenoviruses (26). The adenovirally expressed Smad2 protein (AdSmad2) contains an N-terminal Flag tag that can be readily detected using an ␣-Flag antibody in Western blots and can be distinguished from the endogenous protein because of its slightly slower mobility in electrophoresis. We infected hGL cells with Smad2 adenoviruses at various MOI values, stimulated the cells 24 h postinfection with recombinant GDF-9, and analyzed the samples by ␣PS2 Western blotting. Recombinant GDF-9 activated the phosphorylation of both endogenous and exogenous Smad2 proteins (Fig. 2A). At MOI values of 5 (five viruses per cell) or lower, ␣PS2-reactive bands were rarely seen in unstimulated cells, whereas with Smad2 viruses given at MOI values of 10 –100, spontaneous activation of Smad2 phosphorylation was clearly detected. Receptor-regulated Smad proteins overexpressed excessively from expression plasmids have been previously shown to be spontaneously activated in other types of mammalian cells (30), and here Ad-Smad2 MOI values of 3–5 were chosen for most experiments to avoid excessive Smad2 protein expression. To study whether inhibitory Smads could block the GDF-9-induced Smad2 phosphorylation, we introduced increasing amounts of Ad-Smad6 or Ad-Smad7 (MOI ranging from 0.3–30) into hGL cells simultaneously with a constant amount of Ad-Smad2 (MOI 5). The GDF-9-induced phosphorylation of both the endogenous and the adenovirally expressed Smad2 proteins was gradually reduced close to the control level in cells infected with Ad-Smad7 (Fig. 2B),

FIG. 2. Ad-Smad7 suppresses the GDF-9-induced endogenous and Ad-Smad2-expressed ␣PS2-reactive protein levels in cultured hGL cells. A, Effect of GDF-9 on ␣PS2 immunoreactivity in noninfected hGL cells or cells infected with Ad-Smad2 at MOI of 1 or 10 virus concentrations. B, Effect of GDF-9 on ␣PS2 immunoreactivity in noninfected hGL cells or cells infected with Ad-Smad2 at MOI 5 in combination with Ad-Smad7 at MOI of 0.3, 1, 3, 10, and 30 virus concentrations. C, Effect of GDF-9 on ␣PS2 immunoreactivity in noninfected hGL cells or cells infected with Ad-Smad2 at MOI 5 in combination with Ad-Smad6 at MOI of 0.3, 1, 3, 10, or 30 virus concentrations. hGL cells were first cultured for 2–3 d, infected there after for 24 h with Ad-Smad2 and/or Ad-Smad7 or Ad-Smad6, and subsequently stimulated with 150 ng/ml GDF-9 for 75 min. Cell lysates were analyzed with ␣PS2 and ␣-Flag-M2 Western blotting. Dashes and arrows indicate the migration of endogenous and AdSmad2-expressed ␣PS2-immunoreactive bands. The bands representing adenovirus-expressed Smad2, Smad7, and Smad6 Flagtagged proteins are also indicated in the ␣-Flag-M2 Western blots.

whereas Ad-Smad6 was very ineffective in suppressing Smad2 phosphorylation (Fig. 2C). GDF-9 induces the expression of inhibin ␤B subunit mRNAs and dimeric inhibin B proteins

Figure 3A shows Northern blot analyses indicating that GDF-9 increased inhibin ␤B-subunit mRNA steady state levels after 8 h of stimulation, whereas inhibin ␣-subunit or ␤A-subunit (data not shown) transcript levels were hardly affected. Subsequently, the concentrations of secreted dimeric inhibin B proteins were measured from the spent media of untreated and GDF-9-stimulated cells using a specific inhibin B ELISA. GDF-9 stimulated the production of



inhibin B after 24 h, and the maximal effects were seen after 72 h (Fig. 3B). GDF-9 induced the production of inhibin B in a concentration-dependent manner, and maximal stimulations were seen at 300 ng/ml GDF-9 (Fig. 3C). As Ad-Smad7 suppressed GDF-9-induced Smad2 phosphorylation, we determined whether overexpression of this inhibitory Smad would affect GDF-9-stimulated inhibin B production. Increasing amounts of Ad-Smad7 (MOI values of 0.3–30) suppressed in a concentration-dependent manner GDF-9induced dimeric inhibin B production in hGL cells (Fig. 4, A and C), whereas Ad-Smad7 alone did not affect basal inhibin B levels (Fig. 4, B and D). No cytopathic effects could be detected at these MOI values. Discussion

FIG. 3. GDF-9 induces the expression of inhibin ␤B-subunit mRNAs and the production of inhibin B dimers in cultured hGL cells. A, Northern blot analysis of the effect of GDF-9 on inhibin subunit mRNA levels in hGL cells. Cells were first cultured for 3 d and subsequently stimulated for 2, 8, or 24 h with 100 ng/ml GDF-9. RNAs were extracted, and Northern blots were prepared and hybridized with 32P-labeled inhibin subunit and cyclophilin cDNA probes as described in Materials and Methods. Dashes and arrows indicate the migration of 28S and 18S rRNAs and the specific inhibin ␤B- and ␣-subunit and cyclophilin transcripts. The time course (B) and concentration dependence (C) of the effect of GDF-9 on secreted inhibin B proteins in cultured hGL cells are shown. Cells were treated with 150 ng/ml GDF-9 for 24, 48, 72, and 96 h or with 3–300 ng/ml GDF-9 for 72 h, and the spent media were harvested for measurement of inhibin B concentrations with an inhibin B ELISA. Triplicate samples were used (B and C), and the mean ⫾ SEM are shown. Results are displayed relative to the value for control cultures treated with conditioned medium (1.0). In C, two independent experiments are shown in parallel.

GDF-9 mRNAs and polypeptides are expressed in oocytes (10, 11), and it is likely that this factor regulates somatic cells in a paracrine manner. Although major effects of GDF-9 on rodent folliculogenesis have been described (4, 12), this study describes specific effects of recombinant GDF-9 on isolated human granulosa cells, including effects on the inhibin system. Hreinsson et al. (13) recently showed that early human follicular growth is stimulated by GDF-9, and Yamamoto et al. (31) reported that steroidogenesis is regulated by GDF-9 in human granulosa and thecal cell cultures. Although in vivo experiments concerning GDF-9 effects in primates have not yet been performed, our recent evidence from immunization experiments carried out on sheep indicate that both GDF-9 and GDF-9B/BMP15 are essential for follicular development in this monoovulatory species (32). Thus, GDF-9 has important biological effects in larger mammals and seems to play an important role in regulating ovarian function in various mammalian species. How the cellular signal of GDF-9 is mediated to the target cells, however, has been unclear. This study together with our other recent signaling studies (26, 33) have begun to unveil the GDF-9 signaling system with somewhat surprising results. Our present study focused on the postreceptor effects of GDF-9 by determining whether the Smad protein system used by many members of the TGF␤ family would be activated by this factor. Recent data obtained with rat granulosa cells identified the BMP receptor type II (BMPRII) as a component of the GDF-9 receptor system (33), and consequently, a BMP-like effect of GDF-9 on the Smad system might have been expected. However, our present data clearly indicate that GDF-9 activates Smad2 phosphorylation in hGL cells and thus acts in a very similar manner as TGF␤ and activin. The phosphorylation of Smad2 occurs as early as 30 min after GDF-9 stimulation, and a strong effect persists for at least 3 h. A similar time course of the effect of TGF␤ on Smad2 phosphorylation has been observed in other types of cells in culture (34). The effect of GDF-9 was concentration dependent, and 100 –300 ng/ml doses were needed for maximal effects, suggesting that GDF-9 acts over a similar concentration range as activin (18, 26). Whereas BMP-2-activated endogenous Smad1 phosphorylation, we did not observe any effect of GDF-9 or TGF␤ on Smad1 phosphorylation in hGL cells. Introduction of exogenous Flag-tagged Smad2 proteins to hGL cells via recombinant adenoviruses directly con-

760



FIG. 4. Ad-Smad7 suppresses GDF-9stimulated inhibin B production in cultured hGL cells. hGL cells were first cultured for 3 d, infected thereafter for 24 h with Ad-Smad7 at increasing MOI values (0.1, 0.3, 1, 3, 10, and 30), and treated with 150 ng/ml GDF-9 (A and C) or with control conditioned medium (B and D). The spent media were harvested 72 h after stimulation, and inhibin B levels were measured using an inhibin B ELISA. Triplicate samples were used, and the mean ⫾ SEM are shown. Results are displayed relative to the value for control cultures treated with conditioned medium (1.0) and are representative of experiments performed at least three times.

firmed that Smad2 is activated in the hGL cell context after GDF-9 stimulation. The hGL cells cannot be transfected easily with the use of common transfection methods, such as lipofection (our unpublished data), but recombinant adenoviruses have proven to be very efficient in expressing exogenous gene products in human and rat primary granulosa cells (26, 35). Adenoviral infection allowed us to effectively overexpress Smad2 proteins, and coinfection studies with respective viruses permitted simultaneous expression of exogenous Smad2 and the inhibitory Smad6 and Smad7 proteins in various combinations. The results suggest that Smad7 could completely quench GDF-9-induced Smad2 phosphorylation, whereas the BMP-Smad pathway inhibitor Smad6 had only weak effects. Thus, it seems that GDF-9 is better functionally classified as a TGF␤ or activin-like factor rather than a BMP-like factor. However, GDF-9 seems to show a more restricted target cell specificity than TGF␤ and activin, as the HepG2 hepatoma cells used in our experiments as positive controls for Smad2 activation were responsive to activin and TGF␤, but not to GDF-9 (data not shown). As a further example of differential target cell response to GDF-9 compared with activin, our previous experiments have shown that GDF-9 does not induce mesoderm in the early Xenopus laevis embryo model, which is very responsive to activin (10). Taken together, GDF-9 causes activation of the Smad2 pathway in a TGF␤/activin-like manner, but in a more target cell-specific fashion, whereas it is not able to activate the BMP-Smad pathway in the ovary. Our recent

parallel studies with rat granulosa cells have also confirmed that GDF-9 activates endogenous Smad2, but not Smad1, proteins in rodent granulosa cells (36). Other recent rodent studies have indicated that Smad2 and Smad3 are expressed in the ovary at follicular stages that are probably regulated by GDF-9 (37–39). We and others have also shown the presence of the activin/TGF␤ Smads in the human ovary (20, 40). All of these recent studies indicate that the GDF-9-regulated Smads identified in this study are found in the mammalian ovary in GDF-9 target cells. After demonstrating a stimulatory effect of GDF-9 on Smad2 phosphorylation, we determined how GDF-9 affects the inhibin system in hGL cells and whether the Smad system would be involved. Interestingly, Lanuza et al. (41) reported recently that denuded oocytes produce substances that stimulate inhibin A and B production in rat granulosa cells, and TGF␤ and activin were reported to have similar effects. The oocyte factor GDF-9 was found in this study to stimulate inhibin B production in hGL cells in a similar way as activin and TGF␤ did in previous studies (18, 19, 26, 42). In the present study GDF-9 induced a specific increase in ␤Bsubunit transcript levels in hGL cells. This subunit is apparently induced in these cells by GDF-9, activin (18), and TGF␤ (19), as well as BMP-2 (20). As the inhibin ␣-subunit is available in abundance, selective stimulation of ␤B-subunit expression seems to be sufficient to cause a clear increase in the amount of dimeric inhibin B produced by hGL cells after treatment with these factors. We have recently found that


strong overexpression of Smad1 or Smad2 by recombinant adenoviruses stimulates inhibin B production in hGL cells, and activation of either pathway thus seems to be sufficient for the induction of inhibin B production in hGL cells (26). GDF-9 is likely to stimulate inhibin B production via the Smad2 pathway, and our present results, showing that GDF9-stimulated inhibin B production is suppressed by Smad7 overexpression, would favor this hypothesis. The results do not, however, rule out the possibility that Smad-independent pathways could also be involved (43) in the cellular effect of GDF-9, but it seems clear that at least part of the effects of GDF-9 would be mediated through specific Smads. Interestingly, Su et al. (44) recently showed that some of the effects of GDF-9 on cumulus expansion in mouse ovary could be mediated through mitogen-activated kinases, possibly in a Smad-independent manner. Thus, it is anticipated that the effects of GDF-9 in the target cells are mediated via several parallel pathways, and cross-talk with the gonadotropinactivated pathways at various stages of folliculogenesis is likely to result in more complexity in the cellular response. To be able to study in detail the downstream signaling of GDF-9, the next challenge will be identification of the various components of the GDF-9 receptor complex. One of the type II receptors involved seems to be BMPRII, which is essential in mediating the effects of GDF-9 in rodent granulosa cells (33). We have previously shown that of the known type II Ser/Thr kinase receptors, BMPRII, TGF␤ receptor II, activin receptor type II (ActR-II) and ActR-IIB are expressed in hGL cells as well as ActR-I [activin receptor-like kinase (ALK) 2], ActR-IB (ALK4), BMPRIA (ALK3), and T␤R-I (ALK5) of the type I receptors (18 –20). BMPRIB (ALK6) is known to be expressed in sheep granulosa cells (45), and therefore, it is likely to be expressed in human cells as well. Further studies are in progress to identify which of these ALKs are involved in GDF-9 signaling and to determine what other possible components are needed in the receptor complex mediating the cellular effect of this ligand that appears to play a pivotal role in the regulation of cell growth and differentiation in the mammalian ovary. Acknowledgments The skillful technical assistance of Ms. Marjo Rissanen, Ms. Anita Saarinen, and Mr. Samu Myllymaa is kindly acknowledged. Dr. Steve Hobbs is thanked for the pEFIRES-P vector and helpful discussions. Drs. David Mottershead and Kenneth McNatty are acknowledged for their comments about the manuscript. Received August 16, 2002. Accepted November 3, 2002. Address all correspondence and requests for reprints to: Dr. Olli Ritvos, Biomedicum Helsinki, Room C502b, P.O. Box 63, Haartmaninkatu 8, University of Helsinki, 00014 Helsinki, Finland. E-mail: [email protected]. The work of the Ritvos laboratory was supported by grants from the Academy of Finland, the Finnish National Technology Agency, The Juselius Foundation, The Jalmari and Rauha Ahokas Foundation, the Novo Nordisk Foundation, the Finska La¨ karesa¨ llskapet, and Helsinki University Central Hospital Funds. The Ritvos and Groome laboratories were also supported by joint grants from the European Commission and The Juselius Foundation. The work of P.t.D. was supported by the Netherlands Institute for Earth and Life Sciences (ALW 809.67.024). The work of A.J.W.H.’s laboratory was supported by NIH Grant HD-31398. N.K.-o. and J.B. are recipients of Ph.D. fellowships from the Helsinki Biomedical Graduate School.


References 1. McPherron AC, Lee SJ 1993 GDF-3 and GDF-9: two new members of the transforming growth factor-␤ superfamily containing a novel pattern of cysteines. J Biol Chem 268:3444 –3449 2. McGrath SA, Esquela AF, Lee SJ 1995 Oocyte specific expression of growth differentiation factor-9. Mol Endocrinol 9:131–136 3. Fitzpatrick SL, Sindoni DM, Shughrue PJ, Lane MV, Merchenthaler IJ, Frail DE 1998 Expression of growth differentiation factor-9 messenger ribonucleic acid in ovarian and nonovarian rodent and human tissues. Endocrinology 139:2571–2578 4. Dong J, Albertini DF, Nishimori K, Kumar R, Lu N, Matzuk MM 1996 Growth differentiation factor-9 is required during early ovarian folliculogenesis. Nature 383:531–535 5. Eppig JJ 2001 Oocyte control of ovarian follicular development and function in mammals. Reproduction 122:829 –938 6. Erickson GF, Shimasaki S 2000 The role of the oocyte in folliculogenesis. Trends Endocrinol Metab 11:193–198 7. Laitinen M, Vuojolainen K, Jaatinen R, Ketola I, Aaltonen J, Lehtonen E, Heikinheimo M, Ritvos O 1998 A novel growth differentiation factor-9 (GDF-9) related factor is co-expressed with GDF-9 in mouse oocytes during folliculogenesis. Mech Dev 78:135–140 8. Dube JL, Wang P, Elvin J, Lyons KM, Celeste AJ, Matzuk MM 1998 The bone morphogenetic protein 15 gene is X-linked and expressed in oocytes. Mol Endocrinol 12:1809 –1817 9. Galloway SM, McNatty KP, Cambridge LM, Laitinen MPE, Juengel JL, Jokiranta S, McLaren RJ, Luiro K, Dodds KG, Montgomery GW, Beattie A, Davis GH, Ritvos O 2000 Mutations in an oocyte-derived growth factor gene (BMP15) cause increased ovulation rate and infertililty in a dosage-sensitive manner. Nat Genet 25:279 –283 10. Aaltonen J, Laitinen MP, Vuojolainen K, Jaatinen R, Horelli-Kuitunen N, Seppa¨ L, Louhio H, Tuuri T, Sjo¨berg J, Bu¨tzow R, Hovatta O, Dale L, Ritvos O 1999 Human growth differentiation factor-9 and its novel homolog GDF-9B are expressed in oocytes during early folliculogenesis. J Clin Endocrinol Metab 84:2744 –2750 11. Teixeira Filho FL, Baracat EC, Lee TH, Suh CS, Matsui M, Chang RJ, Shimasaki S, Erickson GF 2002 Aberrant expression of growth differentiation factor-9 in oocytes of women with polycystic ovary syndrome. J Clin Endocrinol Metab 87:1337–1344 12. Hayashi M, McGee EA, Min G, Klein C, Rose UM, van Duin M, Hsueh AJW 1999 Recombinant growth and differentiation factor-9 (GDF-9) enhances growth and differentiation of cultured early ovarian follicles. Endocrinology 140:1236 –1244 13. Hreinsson JG, Scott JE, Rasmussen C, Swahn ML, Hsueh AJW, Hovatta O 2002 Growth differentiation factor-9 promotes the growth, development and survival of human ovarian follicles in organ culture. J Clin Endocrinol Metab 87:316 –321 14. Vitt UA, McGee EA, Hayashi M, Hsueh AJ 2000 In vivo treatment with GDF-9 stimulates primordial and primary follicle progression and theca cell marker CYP17 in ovaries of immature rats. Endocrinology 141:3814 –3820 15. Massague´ J, Chen YG 2000 Controlling TGF-␤ signaling. Genes Dev 14: 627– 644 16. Miyazono K, ten Dijke P, Heldin CH 2000 TGF-␤ signaling by Smad proteins. Adv Immunol 75:115–157 17. Ishisaki A, Yamamoto K, Hashimoto S, Nakao A, Tamaki K, Nonaka K, ten Dijke P, Sugino H, Nishihara T 1999 Differential inhibition of Smad6 and Smad7 on bone morphogenetic protein- and activin-mediated growth arrest and apoptosis in B cells. J Biol Chem 274:13637–13642 18. Era¨maa M, Hilden K, Tuuri T, Ritvos O 1995 Regulation of inhibin/activin subunit messenger ribonucleic acids (mRNAs) by activin A and expression of activin receptor mRNAs in cultured human granulosa-luteal cells. Endocrinology 136:4382– 4389 19. Era¨maa M, Ritvos O 1996 Transforming growth factor-␤1 and -␤2 induce inhibin and activin ␤B-subunit messenger ribonucleic acid levels in cultured human granulosa-luteal cells. Fertil Steril 65:954 –960 20. Jaatinen R, Bondestam J, Raivio T, Hilden K, Dunkel L, Groome N, Ritvos O 2002 Activation of the bone morphogenetic protein signaling pathway induces inhibin ␤B-subunit messenger ribonucleic acid and secreted inhibin B levels in cultured human granulosa-luteal cells. J Clin Endocrinol Metab 87: 1254 –1261 21. Persson U, Izumi H, Souchelnytskyi S, Itoh S, Grimsby S, Engstrom U, Heldin CH, Funa K, ten Dijke P 1998 The L45 loop in type I receptors for TGF-␤ family members is a critical determinant in specifying Smad isoform activation. FEBS Lett 434:83– 87 22. Piek E, Westermark U, Kastemar M, Heldin C-H, van Zoelen EJ, Nister M, ten Dijke P 1999 Expression of transforming growth factor (TGF)-␤ receptors in glioblastoma cell lines with distinct responses to TGF-␤1. Int J Cancer 80:756 –763 23. Hobbs S, Jitrapakdee S, Wallace JC 1998 Development of a bicistronic vector driven by the human polypeptide chain elongation factor 1␣ promoter for creation of stable mammalian cell lines that express very high levels of recombinant proteins. Biochem Biophys Res Commun 252:368 –372

762 J Clin Endocrinol Metab, February 2003, 88(2):755–762


24. Era¨ maa M, Heikinheimo K, Tuuri T, Hilden K, Ritvos O 1993 Inhibin/activin subunit mRNA expression in human granulosa-luteal cells. Mol Cell Endocrinol 92:R15–R20 25. Fujii M, Takeda K, Imamura T, Aoki H, Sampath TK, Enomoto S, Kawabata M, Kato M, Ichijo H, Miyazono K 1999 Roles of bone morphogenetic protein type I receptors and smad proteins in osteoblast and chondroblast differentiation. Mol Biol Cell 10:3801–3813 26. Bondestam J, Kaivo-oja N, Kallio J, Hyden-Granskog C, Groome N, Fujii M, Moustakas A, Jalanko A, ten Dijke P, Ritvos O 2002 Engagement of activin and bone morphogenetic protein signaling pathway Smad proteins in the induction of inhibin B production in cultured human granulosa-luteal cells. Mol Cell Endocrinol 195:79 – 88 27. He TC, Zhou S, da Costa LT, Yu J, Kinzler KW, Vogelstein B 1998 A simplified system for generating recombinant adenoviruses. Proc Natl Acad Sci USA 95:2509 –2514 28. Kumar A, Novoselov V, Celeste AJ, Wolfman NM, ten Dijke P, Kuehn MR 2000 Nodal signaling uses activin and transforming growth factor-␤ receptorregulated smads. J Biol Chem 276:656 – 661 29. Laitinen M, Rutanen EM, Ritvos O 1995 Expression of c-kit ligand messenger ribonucleic acids in human ovaries and regulation of their steady-state levels by gonadotropins in cultured granulosa-luteal cells. Endocrinology 136:4407– 4414 30. Zhang Y, Feng X, We R, Derynck R 1996 Receptor-associated mad homologues synergize as effectors the TGF␤-response. Nature 383:168 –172 31. Yamamoto N, Christenson LK, McAllister JM, Strauss III JF 2002 Growth differentiation factor-9 inhibits 3⬘5⬘-adenosine monophosphate-stimulated steroidogenesis in human granulosa and theca cells. J Clin Endocrinol Metab 87:2849 –2856 32. Juengel JL, Hudson NL, Heath DA, Smith P, Reader KR, Lawrence SB, O’Connell AR, Laitinen MPE, Cranfield M, Groome NP, Ritvos O, McNatty KP 2002 Growth differentiation factor 9 and bone morphogenetic protein 15 are essential for ovarian follicular development in sheep. Biol Reprod 67:1777– 1789 33. Vitt UA, Mazerbourg S, Klein C, Hsueh AJW 2002 Bone morphogenetic protein receptor type II is a receptor for growth differentiation factor-9. Biol Reprod 67:473– 480 34. Lo R, Massague J 1999 Ubiquitin-dependent degradation of TGF-␤-activated Smad2. Nat Cell Biol 1:472– 478

35. Somers JP, DeLoia JA, Zeleznik AJ 1999 Adenovirus-directed expression of a nonphosphorylatable mutant of CREB (cAMP response element-binding protein) adversely affects the survival, but not the differentiation, of rat granulosa cells. Mol Endocrinol 13:1364 –1372 36. Roh J-S, Bondestam J, Mazerbourg S, Kaivo-oja N, Groome N, Ritvos O, Hsueh AJW 2003 Growth differentiation factor-9 (GDF-9) stimulates inhibin production and activates Smad2 in cultured rat granulosa cells. Endocrinology 144:172–178 37. Xu J, Oakley J, McGee EA 2002 Stage-specific expression of smad2 and smad3 during folliculogenesis. Biol Reprod 66:1571–1578 38. Tomic D, Brodie SG, Deng C, Hickey RJ, Babus JK, Malkas LH, Flaws JA 2002 Smad 3 may regulate follicular growth in the mouse ovary. Biol Reprod 66:917–923 39. Kano K, Notani A, Nam SY, Fujisawa M, Kurohmaru M, Hayashi Y 1999 Cloning and studies of the mouse cDNA encoding Smad3. J Vet Med Sci 61:213–219 40. Pangas SA, Rademaker AW, Fishman DA, Woodruff TK 2002 Localization of the activin signal transduction components in normal human ovarian follicles: implications for autocrine and paracrine signaling in the ovary. J Clin Endocrinol Metab 87:2644 –2657 41. Lanuza GM, Groome NP, Baranao JL, Campo S 1999 Dimeric inhibin A and B production are differentially regulated by hormones and local factors in rat granulosa cells. Endocrinology 140:2549 –2554 42. Va¨ nttinen T, Liu J, Hyden-Granskog C, Parviainen M, Penttila¨ I, Voutilainen R 2000 Regulation of immunoreactive inhibin A and B secretion in cultured human granulosa-luteal cells by gonadotropins, activin A and insulin-like growth factor type-1 receptor. J Endocrinol 167:289 –294 43. Yu L, Hebert MC, Zhang YE 2002 TGF-␤ receptor-activated p38 MAP kinase mediates Smad-independent TGF-␤ responses. EMBO J 21:3749 –3759 44. Su YQ, Wigglesworth K, Pendola FL, O’Brien MJ, Eppig JJ 2002 Mitogenactivated protein kinase activity in cumulus cells is essential for gonadotropininduced oocyte meiotic resumption and cumulus expansion in the mouse. Endocrinology 143:2221–2232 45. Mulsant P, Lecerf F, Fabre S, Schibler L, Monget P, Lanneluc I, Pisselet C, Riquet J, Monniaux D, Callebaut I, Cribiu E, Thimonier J, Teyssier J, Bodin L, Cognie Y, Chitour N, Elsen JM 2001 Mutation in bone morphogenetic protein receptor-IB is associated with increased ovulation rate in Booroola Merino ewes. Proc Natl Acad Sci USA 98:5104 –5109