X/00/$03.00/0 Vol. 85, No. 9 The Journal of Clinical Endocrinology & Metabolism Copyright 2000 by The Endocrine Society

Transcription

1 X/00/$03.00/0 Vol. 85, No. 9 The Journal of Clinical Endocrinology & Metabolism Printed in U.S.A. Copyright 2000 by The Endocrine Society Bone Morphogenetic Protein Inhibits Ovarian Androgen Production CHRISTINA A. DOOLEY, GEORGE R. ATTIA, WILLIAM E. RAINEY, D. RAYBURN MOORE, AND BRUCE R. CARR Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas ESTRADIOL FORMATION by developing follicles is critical for normal reproductive function in women. Estrogen biosynthesis relies on the intricate interaction of granulosa and thecal cells due to compartmentalization of 17 -hydroxylase (CYP17) in the theca and its absence in the granulosa. This allows follicular thecal cells to secrete and provide 19 carbon (C 19 ) steroids to the adjacent granulosa cells, for conversion to estrogens (1, 2). Thus, the regulation of C 19 steroid production in the thecal cell is a key component of estrogen biosynthesis. There is growing evidence that locally produced growth factors play an important role in thecal cell steroid production, either alone or in combination with gonadotropins (3). Among these growth factors is the transforming growth factor- (TGF- ) superfamily, which includes TGF-, mullerian-inhibiting substance, inhibin, activin, and the bone morphogenetic proteins (BMPs) (4 6). ABSTRACT Bone morphogenetic proteins (BMPs), members of the transforming growth factor superfamily, were recently shown to be expressed and to regulate steroidogenesis in rat ovarian tissue. The purpose of this study was to investigate the effect of BMP-4 on androgen production in a human ovarian theca-like tumor (HOTT) cell culture model. We have previously demonstrated the usefulness of these cells as a model for human thecal cells. HOTT cells respond to protein kinase A agonists by increased production of androstenedione and with an induction of steroid-metabolizing enzymes. In this investigation, HOTT cells were treated with forskolin or dibutyryl cyclic AMP (dbcamp) in the presence or absence of various concentrations of BMP-4. The accumulation of androstenedione, progesterone, and 17 -hydroxyprogesterone (17OHP) in the incubation medium was measured by RIA. The expression of 17 -hydroxylase (CYP17), 3 hydroxysteroid dehydrogenase (3 HSD), cholesterol side-chain cleavage (CYP11A1), and steroidogenic acute regulatory (StAR) protein was determined by protein immunoblotting analysis using specific rabbit polyclonal antibodies. We also examined the expression of BMP receptor subtypes in our HOTT cells using RT-PCR. In cells treated with medium alone, steroid accumulation and steroid enzyme expression was unchanged. In cells treated with BMP alone there was a modest decrease in androstenedione secretion. In the presence of forskolin, HOTT cell production of androstenedione, 17OHP, and progesterone increased by approximately 4.5-, 35-, and 3-fold, respectively. In contrast, BMP-4 decreased forskolin-stimulated HOTT cell secretion of androstenedione and 17OHP by 50% but increased progesterone production 3-fold above forskolin treatment alone. Forskolin treatment led to an increase in CYP17, CYP11A1, 3 HSD, and StAR protein expression. BMP-4 markedly inhibited forskolin stimulation of CYP17 expression but had little effect on 3 HSD, CYP11A1, or StAR protein levels. Similar results were observed with the camp analog dbcamp. In addition, BMP-4 inhibited basal and forskolin stimulation of CYP17 messenger RNA expression as determined by RNase protection assay. Other members of the transforming growth factor superfamily, including activin and inhibin, had minimal effect on androstenedione production in the absence of forskolin. In the presence of forskolin, activin inhibited androstenedione production by 80%. Activin also inhibited forskolin induction of CYP17 protein expression as determined by Western analysis. We identified the presence of messenger RNA for three BMP receptors (BMP-IA, BMP- IB, and BMP-II) in the HOTT cells model. In conclusion, BMP-4 inhibits HOTT cell expression of CYP17, leading to an alteration of steroidogenic pathway resulting in reduced androstenedione accumulation and increased progesterone production. These effects of BMP-4 seem similar to those caused by activin, another member of the transforming growth factor superfamily of proteins. (J Clin Endocrinol Metab 85: , 2000) BMPs comprise one of the largest subgroups in the TGF- superfamily. Fifteen BMPs have been described, and seven BMPs (2, 3, 3b, 4, 6, 7, 15) have been localized in mammalian ovaries (7 13). Additionally, in situ hybridization has demonstrated BMP-4 and BMP-7 messenger RNA (mrna) in rat thecal cells (14). Attempts to define the molecular and biochemical mechanisms controlling human thecal cell steroidogenesis have been hampered by the difficulty in obtaining and maintaining sufficient numbers of human thecal cells in monolayer culture. Our laboratory has developed a human ovarian thecal-like tumor (HOTT) cell culture model that seems to act as an appropriate model system for thecal cell steroidogenesis (15, 16). These cells continue to produce C 19 steroids and express the enzymes involved in normal thecal cell steroidogeneisis. In the current study, the HOTT cell culture model system was used to investigate the effects of BMP-4 on androgen production. Received March 13, Revision received June 8, Accepted June 15, Address correspondence and requests for reprints to: Bruce R. Carr, M.D., Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, Texas Materials and Methods Cell isolation and culture The descriptions of the in vitro HOTT cell culture model have been reported previously (15 17). Briefly, a portion of the tumor was dis- 3331

2 3332 DOOLEY ET AL. JCE&M 2000 Vol. 85 No. 9 persed into single cells using constant gentle agitation in 0.025% trypsin in Dulbecco s Modified Eagle s (DME)/F-12 (Life Technologies, Inc., Grand Island, NY) and antibiotic (37 C, 30 min 8). After each time point the cell suspension was collected, pooled, pelleted, and resuspended in DME/F-12 medium containing 5% FBS to inactivate the trypsin. Aliquots of these cells were frozen and used as needed for the study. For monolayer culture, HOTT cells were thawed and maintained in DME/ F-12 medium supplemented with insulin (6.25 g/ml), transferrin (6.25 g/ml), selenious acid (6.25 ng/ml), BSA (1.25 mg/ml), and linoleic acid (5.35 g/ml) added in the form of 1% ITS plus (Collaborative Research, Waltham, MA), 2% Ultroser G (BioSepra, Villeneuve, France), and antibiotics. Cells were maintained and grown on 75-cm 2 flasks at 37 C under an atmosphere of 5% CO 2 /95% air. Cells were routinely subcultured using 0.05% trypsin and replated at a 1:4 split. All experiments described in this study were accomplished using cells in culture for 2 8 weeks. For the experiments, growth medium was removed and replaced with a low-serum medium (DME/F-12 medium containing antibiotics and 0.01% Ultroser G) for 24 h. Cells were then rinsed and experimentally treated in the same low-serum medium. At the end of the treatment period, the medium was removed and the cells and medium were stored at 20 C for subsequent assay. Experimental treatments Specifics of various treatments are detailed in the figure legends. Reagents included: forskolin (Sigma, St. Louis, MO), dbcamp (Sigma), and recombinant human BMP-4 (Research Diagnostics Inc., Flanders, NJ). Human recombinant inhibin-a and human recombinant activin-a were graciously supplied by Dr. A. F. Parlow, the scientific director of the National Hormone and Pituitary Program (Torrance, CA). Measurement of steroid content The steroid contents of culture medium were assayed after 36 h. This time was chosen as optimal for inhibition of androstenedione by BMP on forskolin-stimulated cells. Steroid contents was measured using standard RIA kits [androstenedione, 17OHP, and progesterone from Diagnostic Systems Laboratory (Webster, TX)]. The amount of steroid measured was expressed as picomoles steroid per milligram of cellular protein. Protein determination Cells were solubilized in Tris-HCl (50 mm/ph 7.4) containing NaCl (150 mm), deoxycholic acid (0.1%), EGTA (5 mm), MgCl 2 (0.5 mm), and phenylmethylsulfonylfluoride (0.2 mm) and stored frozen at 20 C. Subsequently, protein content was determined by bicinchonic acid protein assay, using the BCA assay kit (Pierce, Rockford, IL). Western analysis Using a Novex Xcell system (Novel Experimental Technology, San Diego, CA), one dimensional electrophoresis (200 V, 35 min) was performed in a 4 12% polyacrylamide pre-cast gel in a NuPAGE MES SDS running buffer [50 mm 2-(N-morpholino) ethane sulfonic acid, 50 mm Tris base, 3.5 mm SDS, and 1 mm EDTA (ph 7.2)] under reducing conditions. Proteins were transferred to nylon membrane (25 V, 1 h) in NuPAGE transfer buffer [25 mm Bicine, 25 mm Bis-tris, 1 mm EDTA, 0.05 mm chlorobutanol, and 20% methanol (ph 7.2)]. Immunoblotting was performed with rabbit polyclonal antibodies. Antibody sources were: human CYP17 antibody (M. Waterman, Vanderbilt University Medical Center, Nashville, TN), StAR antibody (D. Stucco, Texas Tech University Health Sciences Center, Lubbock, TX), human 3 HSD antibody (J. Ian Mason, University of Edinburgh, Edinburgh, Scotland), and human CYP11A1 (Bon-Chu Chung, Institute of Molecular Biology Taiwan, Taiwan, China). The primary antibody incubation was followed by incubation with an antirabbit Ig, and horseradish peroxidase-linked F (ab ) 2 fragment from donkey (Amersham Life Sciences, Buckinghamshire, England). Proteins were detected with enhanced chemiluminescence reagents (Amersham Life Sciences) on X-Omat Blue XB-1 Scientific Imaging Film (Kodak, Rochester, NY). RNA isolation Total RNA was isolated from a 15-mm follicle obtained from a patient undergoing oophrecetomy for a benign gynecological condition, as well as from HOTT cells that were maintained in low-serum medium for 48 h. RNA was isolated using Ultraspec RNA reagent according to the manufacturer s suggested protocol. Frozen follicle and HOTT cells were homogenized in Ultraspec RNA reagent. The homogenate was kept on ice for at least 5 min to permit the complete dissociation of nucleoprotein complexes. Chloroform (0.2 ml/1 ml Ultraspec RNA reagent) was added while vigorously shaking, incubated at 4 C for 5 min, and centrifuged at 12,000 g for 15 min at 4 C. The aqueous phase was carefully transferred to a fresh tube without disturbing the interphase. An equal volume of isopropanol was added, and the tube was stored at 4C for 1 h or longer. RNA precipitate was formed as a pellet at the bottom of the tube after centrifugation for 10 min at 12,000 g (4 C). The pellet was washed twice with 75% ethanol by vortexing and subsequent centrifugation for 5 min at 12,000 g (4 C) and briefly dried. The pellet was then dissolved in 50 ml diethylprocarbonate-treated water. RNase Protection Assay (RPA) A 295-bp fragment was obtained from CYP17 cdna using Sac1 and Xba 1. This fragment, containing the nucleotides between 6 and 300, was inserted in pbluescript KS vector and sequenced to confirm its identity and orientation. A 32 P-labeled probe was prepared from the linearized plasmid using Maxiscript T7/T3 polymerase kit (Ambion, Inc., Austin, TX). A 125-bp 32 P-labeled riboprobe for -actin was also prepared using Maxiscript T7/T3 polymerase kit and was used as internal control. Because -actin is more abundant than CYP17 mrna, the 32 P-labeled -actin probe was prepared at 100-fold lower specific activity. This ensured that the -actin probe could be used in molar excess to its target and a similar exposure time could be used for both probes. The labeled probes were purified on a 6% acrylamide gel. The labeled CYP17 and -actin probes were added to each of the RNA sample (10 g). RPA was performed according to the manufacturer-suggested protocol (Ambion). Following hybridization and RNase digestion the protected fragments were separated on a 6% polyacrylamide gel. The gel was placed in a phosophoimager caste and left overnight. A phosphoimager counter was used to measure the density of each band. RT-PCR The first-strand cdna synthesis from 4 g total RNA was catalyzed by superscript II RT using random hexamer primers according to the manufacturer s protocol. The 20-L of reaction mixture consisted of 4 g total RNA, extracted either from ovarian follicles or HOTT cells, 125 ng random hexamers, 50 mm Mg Cl 2, 0.5 mm dntp, 10 mm dithiothreitol, and 200 U Superscript II RT in 20 mm Tris-HCl buffer (ph 8.4). Two microliters of the first-strand cdna reaction was used for PCR reaction to amplify the three different subtypes of BMP receptors. Primers used for PCR reaction are listed in Table 1. The PCR condition was 94 C for 3 min to denature the RNA/cDNA hybrid, then 30 cycles for 94 C for 1 min, 55 C for 1 min, and 72 C for 1 min. PCR products were examined on a 1% agarose gel. Statistical analysis Statistical comparison of means of three or more samples was accomplished by ANOVA with Newman-Keuls post hoc testing. Significance was accepted at the 0.05 level of P. TABLE 1. The expected size of PCR products and the primer sequences used for detection of BMP receptors subtypes BMPR-IA (1401 bp) Sense Antisense BMPR-IB (634 bp) Sense Antisense BMPR-II (694 bp) Sense Antisense GCATAACTAATGGACATTGCT TAGAGTTTCTCCGATGG GCAGCACAGACGGATATTGT TTTCATGCCTCATCAACACT ACGGGAGAGAAGACGAGCCT CTAGATCAAGAGAGGGTTCG

3 BMP INHIBITS OVARIAN ANDROGEN PRODUCTION 3333 Results Steroid production in HOTT cells Medium accumulation of androstenedione, 17 hydroxyprogesterone (17OHP), and progesterone was evaluated after 36 h of treatment. Treatment groups consisted of control (basal), forskolin (10 m), BMP-4 (50 ng/ml), and forskolin (10 m) plus BMP-4 (50 ng/ml). In cells treated with control medium or BMP-4, steroid accumulation was unchanged. In the presence of forskolin, HOTT cell production of androstenedione, 17OHP, and progesterone increased by approximately 4.5-, 35-, and 3-fold, respectively (Fig. 1, A, B, and C, respectively). In contrast, BMP-4 decreased forskolin-stimulated HOTT cell secretion of androstenedione and 17OHP by 50% but increased progesterone 3-fold above forskolin treatment alone (P 0.001). To examine the concentration-dependent effect of BMP-4, cells were treated for 36 h with control medium, forskolin (10 m), BMP-4 (50 ng/ml), or forskolin (10 m) plus BMP-4 in increasing doses (1, 3, 10, 30, and 50 ng/ml). Medium accumulation of androstenedione was evaluated after 36 h of treatment. BMP-4 caused a concentration-dependent inhibition of forskolin-stimulated androstenedione production for all concentrations greater than 3 ng/ml (P 0.001) (Fig. 2). Comparison of BMP-4, inhibin, and activin effects on HOTT cell steroid production Cells were treated for 36 h with control media, forskolin (10 m), BMP-4 (50 ng/ml), inhibin-a (50 ng/ml), activin-a (50 ng/ml), or forskolin (10 m) plus BMP-4 (50 ng/ml), inhibin-a (50 ng/ml), or activin-a (50 ng/ml) (Fig. 3). Androstenedione levels were evaluated using RIA. In cells treated with medium only, BMP-4 modestly inhibited steroid accumulation. In the presence of forskolin, HOTT cell production of androstenedione, increased by approximately 4-fold and was not affected by the presence of inhibin. In contrast, BMP-4 and activin significantly decreased forskolin-stimulated HOTT cell secretion of androstenedione (BMP-4, P 0.001; activin-a, P 0.001). FIG. 1. Effect of BMP-4 with and without forskolin on steroid production in HOTT cells. Cells were treated for 36 h with medium alone, forskolin (10 M), BMP-4 (50 ng/ml), or forskolin (F) (10 M) plus BMP-4 (50 ng/ml). A, Androstenedione production. B, 17OHP production. C, Progesterone production. All are expressed as a percentage of steroid production in control (medium only treatment). The basal values: A, pmol/mg protein; B, pmol/mg protein; C, pmol/mg protein. Each data point represents SE of replicate dishes (n 3). Data are from a representative experiment that was run three times. *, P 0.001, compared to forskolin alone. Protein immunoblotting analysis for steroid-metabolizing enzymes We sought to elucidate the effect of BMP-4 on HOTT cell expression of immuno-detectable steroidogenic enzymes using Western analysis. Cells were treated for 36 h with medium only, forskolin (10 m), BMP-4 (50 ng/ml), or forskolin (10 m) plus BMP-4 (50 ng/ml). Additionally, cells were treated for 36 h with dbcamp (1 mm) or dbcamp (1 mm) plus BMP-4 (50 ng/ml). Western analysis was performed as described in Materials and Methods. The results are presented in Fig. 4. In cells treated with medium alone or BMP-4, steroidogenic enzyme expression was unchanged. Forskolin treatment increased CYP17, CYP11A1, and StAR protein expression. In contrast, BMP-4 markedly inhibited forskolinstimulated CYP17 expression but had little effect on 3 HSD, CYP11A1, or StAR protein. Similar results were observed with the camp analog dbcamp. Moreover, we compared the effects of BMP-4, inhibin, and activin on steroidogenic enzymes in forskolin-stimulated HOTT cells. Cells were treated for 36 h with and without forskolin (10 m) or forskolin (10 m) plus BMP-4 (50 ng/ml, inhibin-a (50 ng/ml, or activin-a (50 ng/ml). Western analysis was performed as described in Materials and Methods, and the results are shown in Fig. 5. In cells treated with medium alone, BMP-4, inhibin-a, or activin-a, steroid enzyme expression was unchanged. Forskolin treatment increased CYP17, CYP11A1, and StAR protein expression. Inhibin showed no inhibition of fors-

4 3334 DOOLEY ET AL. JCE&M 2000 Vol. 85 No. 9 FIG. 2. Concentration-dependent effect of BMP-4 on forskolin stimulation of androstenedione production in HOTT cells. Cells were treated for 36 h with medium alone, forskolin (10 M), BMP-4 (50 ng/ml), or forskolin (F) (10 M) plus BMP-4 in increasing doses (1 50 ng/ml). Data are expressed as a percentage of control steroid production (medium only treatment). Each data point represents the mean SE of replicate dishes (n 3). The basal value is pmol/mg protein. Data are for a representative experiment that was run twice. *P and **P 0.025, compared with forskolin alone. FIG. 3. Effect of BMP-4, inhibin-a, and activin-a with and without forskolin treatment on androstenedione production in HOTT cells. Cells were treated for 36 h with medium only, forskolin (10 M), BMP-4 (50 ng/ml), inhibin-a (50 ng/ml), activin-a (50 ng/ml), or forskolin (F) (10 M) plus BMP-4 (50 ng/ml) inhibin-a (50 ng/ml), and activin-a (50 ng/ml). Each data point represents the mean SE of replicate dishes (n 3). Basal value is pmol/mg protein. Data are from a representative experiment that was run three times. *P compared to forskolin alone. kolin-stimulated enzymes. In contrast, BMP-4 markedly inhibited forskolin-stimulated CYP17 expression but had no effect on 3 HSD, CYP11A1, or StAR protein. Activin also markedly inhibited forskolin-stimulated CYP17 expression. Additionally, activin mildly inhibited forskolinstimulated StAR expression. FIG. 4. Effect of BMP-4 with and without forskolin treatment on steroidogenic enzyme expression in HOTT cells. Cells were treated for 36 h with medium only, forskolin (10 M), BMP-4 (50 ng/ml) or forskolin (10 M) plus BMP-4 (50 ng/ml). Additionally, cells were treated for 36 h with dbcamp (1 mm) or dbcamp (1 mm) plus BMP-4 (50 ng/ml). Western analysis was performed as described in Materials and Methods with 20 g cellular protein per lane. Lanes corresponding to each experimental condition are listed above the figure. Molecular weight markers are indicated with a dash. RNase Protection Assay for CYP17 transcripts To better define the mechanism for BMP-4 inhibition of CYP17 protein expression, we sought to evaluate CYP17 mrna using a RPA for CYP17 as discussed in Materials and Methods. Cells were incubated for 24 h in basal media, BMP-4 (10 ng/ml), forskolin (10 m) or forskolin (10 m) plus BMP-4 (10 ng/ml). The results of the experiment are shown in Fig. 6, A and B. Compared with basal levels, BMP-4 alone inhibited CYP17 mrna expression by 60%. Forskolin treatment caused a 20-fold increase in CYP17 mrna levels, but BMP-4 inhibited forskolin stimulated CYP17 mrna expression by 40%. These data suggest BMP-4 action on CYP17 protein expression occurs through alterations in mrna levels. Expression of BMP receptors using RT-PCR To better define the molecular action of BMP-4 on steroidogensis, we examined for the presence of BMP receptors subtypes in HOTT cells maintained in low serum medium as described in Materials and Methods. As seen in Fig. 7, the expression of BMP-IA, BMP-IB, and BMP-II was observed in a human ovarian follicle and in HOTT cells.

5 BMP INHIBITS OVARIAN ANDROGEN PRODUCTION 3335 FIG. 5. Effect of BMP-4, inhibin-a, and activin-a with and without forskolin on steroidogenic enzyme expression in HOTT cells. Cells were treated for 36 h with medium only, forskolin (10 M), BMP-4 (50 ng/ml), inhibin-a (50 ng/ml), activin-a (50 ng/ml), or forskolin (10 M) plus BMP-4 (50 ng/ml) inhibin-a (50 ng/ml) or activin-a (50 ng/ml). Western analysis was performed as described in Materials and Methods with 20 g cellular protein per lane. Lanes corresponding to each experimental condition are listed above the figure. Molecular weight markers are indicated with a dash. Discussion Recruitment, selection, growth and atresia of ovarian follicles are not only regulated by endocrine hormones but also by locally produced growth factors that can act in an autocrine or paracrine manner (3). The ovary seems to produce a wide range of growth factors that influence not only growth but modulate steroid hormone biosynthesis in both granulosa and theca cells. Growth factor influences on thecal cell androgen synthesis directly impacts estradiol production by influencing substrate availability for granulosa cells. Members of the TGF superfamily seem to play such a role in the regulation of ovarian steroid hormone production (5, 6, 18 23). Here, we examine the effects of a recently discovered member of the TGF superfamily, BMP. BMPs are multifunctional cytokines that regulate cellular proliferation, differentiation, and apoptosis of various cell types, including osteoblasts and chondroblasts (7 13). These functions are related to various biological functions in vivo (e.g. formation of bone and cartilage, embryogenesis, and organogenesis). There are at least 15 structurally related BMPs, all of which have structural similarities that place them in the TGF superfamily of proteins. Recently, the expression of BMPs was demonstrated in rat ovarian tissues, and predominantly in thecal cells (14). Thecal cells expressed both BMP-4 and BMP-7 mrna as determined by in situ FIG. 6. Effect of BMP-4 and forskolin on CYP17 mrna expression in HOTT cells. A, RPA for CYP17. Cells were treated without (basal) or with BMP-4 (10 ng/ml), forskolin (10 M) or BMP-4 (10 ng/ml) plus forskolin (10 M) for 24 h. RNA was then extracted, and the RPA was performed as described in Materials and Methods. Yeast was used as a negative control, and -actin was used to standardize results from each lane. B, The data from panel A is expressed as a percentage of CYP17 mrna expression compared to basal (control) cells. hybridization. BMP type IA, IB, and II receptors were also studied and shown to be predominantly expressed in the granulosa cells (14). Granulosa cell steroidogenesis was also influenced by BMP-4 and BMP-7 treatment. In addition, human granulosa cells obtained from in vitro fertilization cycles seem to express BMP-3, and its expression was regulated negatively by human chorionic gonadotropin (hcg) (11). The production of BMPs in granulosa cells suggests the possibility of an autocrine role on granulosa cells or a paracrine effect on thecal cells. The purpose of this study was to investigate the role of one of the BMP family members, BMP-4, on the regulation of steroidogenesis as well as evaluating protein and mrna expression of the principal enzymes involved in human ovarian thecal cell steroidogenesis. The difficulty in obtaining human thecal cell in sufficient quantities has slowed progress directed at defining the mechanism regulating of steroidogenesis. Our laboratory has isolated cells from ovarian tumors and place them in monolayer cell culture (15, 16). These HOTT cells have retained many of the characteristics of normal human thecal cells maintained in primary culture. These include the production of C 19 steroids and expression of steroid-metabolizing enzymes, each of which is under the control of camp. In addition, we have demonstrated that the HOTT cell model responds to TGF, activin, and inhibin with similar responses to those seen in primary cultures of human thecal cells (5, 6, 21, 22).

6 3336 DOOLEY ET AL. JCE&M 2000 Vol. 85 No. 9 FIG. 7. Expression on BMP receptors RNA in HOTT cells. RT-PCR was performed to demonstrate the presence of the three different subtypes of BMP receptors in HOTT cells. A, BMP receptor IA (BMPRIA). B, BMP receptor IB (BMPRIB). C, BMP receptor II (BMP- RII). Follicular RNA was used as positive control for all three reactions. RT was used as a negative control in which the RT reaction was performed in the absence of Superscript II enzyme. RNA was performed in the absence of RNA in the RT reaction. Molecular weight markers are noted by the arrows on the right. Treatment of HOTT cells with BMP-4 caused a concentration-dependent inhibition of forskolin-stimulated androstenedione production. This effect was also observed on the stimulation of androstenedione by dbcamp, an analog of camp. The effect was not due to an overall inhibition of steroid hormone biosynthesis because the level of progesterone production actually increased following BMP-4 treatment. This effect (decreased androgen production and increased progestin formation) is similar to that we reported previously for activin treatment of HOTT cells (6). In vivo such a shift in thecal cell androgen production is observed at ovulation when the cells start to luteinize. Thus, one interpretation of the effects of BMP and activin would be that they act to promote a luteinized phenotype or alternatively that they act as a negative regulator of androgen production to prevent excessive androstenedione production by thecal cells. BMPs transduce signals through binding to specific BMP receptors that have been grouped as type IA, type IB, and type II BMP receptors, each of which has serine/threonine kinase activity (7 10). The BMPs can also bind and activate acitivin receptors (24 26). Therefore, the effects on thecal cell androgen production could result from BMP-4 activation of specific BMP receptors or through activin receptors. The observation that activin and BMP-4 have the similar effects on enzyme expression and androgen production support this hypothesis. However, we observed that the HOTT cells model expresses receptors for BMP; therefore, we suggest that the effect of BMP could be mediated through its BMP receptors or to a lesser degree through activin receptors. The production of thecal cell androstenedione relies on the activities of CYP11A1, 17 -hydroxylase; CYP17, and 3 HSD (27). In addition, the production of androstenedione can be regulated by the expression of StAR protein that is needed for cholesterol to enter the mitochondria (28). As mentioned above, BMP-4 inhibited androstenedione production but increased the production of progesterone, suggesting that the enzymes involved in progesterone biosynthesis are not effected. Indeed, Western analysis revealed that BMP-4 exhibited specific effects on CYP17 without an inhibition of StAR, CYP11A1, or 3 HSD. Human CYP17 and 3 HSD represent key regulatory enzymes at branchpoints in the production of C 21 steroids (progesterone and 17 -OHP) vs. the production of C 19 steroids (dehydroepiandrosterone and androstenedione). CYP17, through its 17 -hydroxylase activity, will readily bind and metabolize pregnenolone or progesterone to 17 -hydroxypregnenolone or 17 -OHP, respectively. By virtue of its broad substrate specificity 3 HSD will complete with CYP17 for the metabolism of pregnenolone and 17 hydroxypregnenolone (29, 30). Therefore, the relative expression of CYP17 and 3 HSD will have direct effects on the amount of androstenedione produced by thecal tissue. Activin inhibited CYP17 in a similar manner and had little effect on StAR, CYP11A1, or 3 HSD. Taken together, these data suggest that activin and BMP-4 decrease thecal cell androgen production by modifying the ratio of CYP17 to 3 HSD within the cell. This results in the production of more progesterone and less androstenedione. It could be hypothesized that such an effect could inhibit overall estrogen production in vivo by decreasing production of C 19 steroids in the thecal cells. Such an effect could be considered luteotropic because the production of estrogen does decrease following ovulation as the thecal cells luteinize. We have previously observed similar effects using activin, another member of the TGF family of proteins (6). Activin also lead to an increase in progesterone production and an inhibition of androstenedione production. Activin and BMP-4 may well share signaling pathways and, therefore, act through a similar manner to regulate thecal cell steroidogenesis.

7 BMP INHIBITS OVARIAN ANDROGEN PRODUCTION 3337 References 1. Short RV Steroids in the follicular fluid and the corpus luteum of the mare: a two-cell type theory of ovarian steroid synthesis. J Endocrinol. 24: McNatty KP, Makris A, DeGrazia C, Osathanondh R, Ryan K The production of progesterone, androgens, and estrogens by granulosa cells, theca tissue and stroma tissue from human ovaries in vitro. J Clin Endocrinol Metab. 49: Ruutiainen K, Adashi EY Intraovarian factors in hyperandrogenism. Semin Reprod Endocrinol. 11: Koppa SD TGF- : what s in a name? Drug Ther. 24: Carr BR, McGee EA, Sawetawan C, Clyne CD, Rainey WE The effect of transforming growth factor- on steroidogenesis and expression of key steroidogenic enzymes with a human ovarian thecal-like tumor cell model. Am J Obstet Gynecol. 174: Sawetawan C, Carr BR, McGee E, Bird IM, Hone TL, Rainey WE Inhibin and activin differentially regulate androgen production and 17 -hydroxylase expression in human ovarian thecal-like tumor cells. J Endocrinol. 148: Hogan BL Bone morphogenetic proteins: multifunctional regulators of vertebrate development. Genes Dev. 10: Wozney JM, Rosen V Bone morphogenetic protein and bone morphogenetic protein gene family in bone formation and repair. Clin Orthop. 346: Dube JL, Wang P, Elvin J, Lyons KM, Celeste AJ, Matzuk MM The bone morphogenetic protein 15 gene is X-linked and expressed in oocytes. Mol Endocrinol. 12: Leong LM, Brickell PM Molecules in focus. Bone morphogenetic protein-4. Int J Biochem Cell Biol. 28: Jaatinen R, Rosen V, Tuuri T, Ritvos O Identification of ovarian granulosa cells as a novel site of expression for bone morphogenetic protein-3 (BMP-3/osteogenin) and regulation of BMP-3 messenger ribonucleic acids by chorionic gonadotropin in cultured human granulosa-luteal cells. J Clin Endocrinol Metab. 81: Ring CJ, Cho KWY Insights from model systems. Specificity in transforming growth factor- signaling pathways. Am J Hum Genet. 64: Shimizu K, Yoshikawa H, Matsui M, Masuhara K, Takaoka K Periosteal and intratumorous bone formation in a thymic nude mice by Chinese hamster ovary tumors expressing murine bone morphogenetic protein-4. Clin Orthop Relat Res. 300: Shimasaki S, Zachow RJ, Li D, et al A functional bone morphogenetic protein system in the ovary. Proc Natl Acad Sci USA. 96: Rainey WE, Sawetawan C, McGee EA, Bird IM, Word RA, Carr BR Human ovarian tumor cells: a potential model for theca cell steroidogenesis. J Clin Endocrinol Metab. 80: Carr BR, McGee EA, Sawetawan C, Rainey WE Development of a human thecal tumor cell model: regulation of steroidogenesis and enzyme expression. Proceedings of the Conference on Polycystic Ovarian Syndrome. Serono Symp (USA) Sawetawan C, Rainey WE, Word RA, Carr BR Immunohistochemical and biochemical analysis of a human Sertoli-Leydig cell tumor: autonomous steroid production characteristic of ovarian theca cells. J Soc Gynecol Invest. 2: Chegini N, Flanders KC Presence of transforming growth factor- and their selective localization in human ovarian tissue of various reproductive stages. Endocrinology. 130: Chegini N, Williams RS Immunocytochemical localization of transforming growth factors (TGF s ) TGF- and TGF- in human ovarian tissues. J Clin Endocrinol Metab. 74: Magoffin DA, Gancedo B, Erickson GF Transforming growth factor- promotes differentiation of ovarian thecal-interstitial cells but inhibits androgen production. Endocrinology. 125: Hillier SG, Yong EL, Illingworth PJ, Baird DT, Schwall RH, Mason AJ Effect of recombinant inhibin on androgen synthesis in cultured human thecal cells. Mol Cell Endocrinol. 157:R1 R Hillier SG, Yong EL, Illingworth PJ, Baird DT, Schwall RH, Mason AJ Effect of recombinant activin on androgen synthesis in cultured human thecal cells. J Clin Endocrinol Metab. 72: McAllister JM, Byrd W, Simpson ER The effects of growth factors and phorbol esters on steroid biosynthesis in isolated human theca interna and granulosa-lutein cells in long term culture. Endocrinology. 79: Peng C, Ohno T, Koh LY Chen VT, Leung PC Human ovary and placenta express messenger RNA for multiple activin receptors. Life Sci. 64: Miyazono K Signal transduction by bone morphogenetic protein receptors: function roles of Smad proteins. Bone. 25: Yamashita H, Miyazono K Bone morphogenetic protein (BMP) receptors and transduction. Nippon Rinsho. 57: Carr BR Disorders of the ovaries and female reproductive tract. In: Wilson JD, Foster DW, Kronenbert HM, Larsen PR, eds. Williams Textbook of Endocrinology, ed 9. Philadelphia: WB Saunders Co; Orly J, Stocco DM The role of the steroidogenic acute regulatory (StAR) protein in reproductive tissues. Horm Metabol Res. 31: Barnes HJ, Arlotto MP, and Waterman MR Expression and enzymatic activity of recombinant cytochrome P hydroxylase in Escherichia coli. Proc Natl Acad Sci USA. 88: Brock BJ, Waterman MR Biochemical differences between rat and human cytochrome P450c17 support the different steroidogenic needs of these two species. Biochemistry. 38: