Leptin impairs the synergistic stimulation by transforming growth factor-beta of follicle-stimulating hormone-dependent aromatase activity and messenger ribonucleic acid expression in rat ovarian granulosa cells.

determine whether leptin caused a general antagonism of the stimulatory modulators of FSH-dependent E 2 production. This study tested the effects of leptin on TGF- (cid:98) regulation of E 2 production by GC, and investigated intraovarian mechanisms whereby leptin suppresses E 2 synthesis in GC. was no effect of leptin on LH-dependent androgen production in the presence or absence of TGF- (cid:98) . Leptin (0.1–100 ng/ml) did not alter LH-dependent androsterone levels after 48 h in vitro (data not shown). These results showing a lack of an effect of leptin on TIC androgen production are consistent with the observation that the TIC did not express the signal-transducing isoform of the leptin receptor (OB-Rb).


INTRODUCTION
Leptin is a 16-kDa protein produced principally by adipocytes [1]. Leptin was initially identified because of its effects on appetite suppression and fat metabolism in rodents [2]. Additional studies demonstrated that exogenous leptin exerts regulatory functions within the rodent reproductive-endocrine axis by restoring fertility in the leptindeficient (Ob/Ob) mouse [3]. The positive effects on fertility appear to be related to increased secretion of LH and FSH, as a result of increased production of LHRH [4]. Collectively, these data indicate that leptin promotes fertility at the neuroendocrine level by increasing gonadotropin release. However, recent evidence demonstrates that direct ovarian effects of leptin cannot be ruled out [5][6][7].
Using mouse brain [8], pancreatic ␤ cells [9], and transfected cell lines [10], it has been demonstrated that leptin binds with high affinity to a cell-surface receptor (OB-R). OB-R is structurally similar to the class I cytokine receptor gp130 subunit [10]. Three major classes of leptin receptor isoforms have been identified in the rat brain: 1) the long, signal-transducing isoform, OB-Rb; 2) truncated isoforms, OB-Ra, OB-Rc, and OB-Rd; and 3) a soluble isoform, OB-Re [11]. The cellular distribution of leptin receptor isoforms within the rodent ovary has not been reported.
Upon receptor activation, an intricate signaling cascade that is mediated by one or more Janus kinase isoforms, and signal transducers and activators of transcription (STAT) proteins has been shown to mediate leptin bioactivity [10]. At present, the intracellular mechanisms that regulate leptin bioactivity within the ovary are unknown.
In the ovary, FSH is required for the recruitment of small antral follicles into the growing preovulatory cohort. As a consequence of FSH stimulation, the granulosa cells (GC) differentiate into estradiol-17␤ (E 2 )-producing cells, and E 2 is required for continued follicle viability. It has become apparent that a group of intraovarian growth factors and cytokines modulates the FSH-dependent growth and differentiation of GC [12]. Many of these regulatory factors have redundant actions in GC. For example, both insulin-like growth factor-I (IGF-I) and transforming growth factor ␤ (TGF-␤) augment FSH-dependent E 2 production. Hence, TGF-␤ appears to be a key supportive intraovarian factor because it promotes GC growth [13] and augments FSHdependent E 2 synthesis [14].
Leptin has been shown to exert direct inhibitory effects on ovarian GC. In cultures of rat [5], bovine [6], and human [7] GC, leptin suppressed the sensitizing effect of insulin and/or IGF-I on FSH-dependent E 2 production. Because E 2 production is essential for ovarian follicle growth and ovulation, and IGF-I appears to be an obligatory mediator of FSH-dependent follicle development [15], it appears that leptin can interfere with an important regulatory mechanism supporting follicle viability and ultimately ovulation. In light of the redundant modulation of FSH action by growth factors and cytokines, it was of interest to determine whether leptin caused a general antagonism of the stimulatory modulators of FSH-dependent E 2 production. This study tested the effects of leptin on TGF-␤ regulation of E 2 production by GC, and investigated intraovarian mechanisms whereby leptin suppresses E 2 synthesis in GC.

Reagents and Supplies
Recombinant murine leptin (carrier-free) and recombinant human transforming growth factor ␤1 (TGF-␤; lyophilized with BSA as carrier) were purchased from R&D Sys-tems (Minneapolis, MN). Human recombinant FSH and ovine LH were supplied by the National Hormone and Pituitary Program of the NIDDK, NICHD, and USDA (Rockville, MD). McCoy's 5a medium (M5a, serum-free) and Medium 199 were purchased from GIBCO-BRL (Grand Island, NY). Culture plates were purchased from Falcon (Lincoln Park, NJ). [1␤-3 H(N)]Androstenedione (A 4 ; 21.5 Ci/ mmol) was obtained from Dupont NEN (Boston, MA). The E 2 RIA kit was obtained from Diagnostic Products Corporation (Los Angeles, CA). The estrone (E 1 ) RIA kit was obtained from Diagnostic Systems Laboratories (Webster, TX). Unless otherwise specified, all assay reagents were purchased from Sigma (St. Louis, MO).

GC Culture
All procedures using live animals were approved by the CSMC Institutional Animal Care and Use Committee. Immature (26-day-old) Sprague-Dawley rats (Harlan Industries, Indianapolis, IN) were killed via CO 2 inhalation followed by cervical dislocation. Ovaries were removed and placed in ice-cold Medium 199 supplemented with 0.1% BSA. Ovaries were cleaned of bursa and other extraneous tissues, and GC were collected from the surrounding medium after follicle puncture [16]. GC were centrifuged (250 ϫ g) and resuspended in a known volume of M5a, supplemented with penicillin (100 U/ml), streptomycin sulfate (100 g/ml), and L-glutamine (2 mM). GC number and viability were determined by trypan blue exclusion using a hemacytometer.
Aliquots containing 50 000-60 000 viable GC were placed in 96-well culture plates. GC were incubated in a final volume of 200 l M5a/well containing 0.1 M A 4 at 37ЊC in a humidified atmosphere containing 5% CO 2 in air. Control GC were incubated without additional hormones. Designated GC were challenged with FSH (0.001-1.0 IU/ ml) with or without TGF-␤ (10 ng/ml). Separate cultures were treated with FSH (0.001-1.0 IU/ml) plus leptin (10 ng/ml) with and without TGF-␤ (10 ng/ml). The leptin concentration was chosen on the basis of the reported K d (0.7 nM) for leptin binding [8] and previous studies by our lab [5], as well as serum leptin concentrations in the human [17]. Cultures were terminated at 48 h, and the conditioned media were collected and frozen at Ϫ20ЊC pending RIAs to measure E 2 and E 1 content. RIAs were conducted according to the manufacturers' protocols.

Theca-Interstitial Cell (TIC) Culture
To measure the effect of leptin on androgen production, purified populations of TIC were obtained from the enzymatically dispersed ovaries of 26-day-old hypophysectomized rats as previously described [18]. TIC viability was determined using trypan blue exclusion. TIC were incubated in 96-well plates (approximately 4 ϫ 10 5 viable TIC/ well) in a final volume of 200 l. TIC were either cultured in M5a alone (control), leptin (0.1, 1.0, 10, and 100 ng/ml), or LH (0.03-10 ng/ml) to induce steroidogenic differentiation of the cells. Designated TIC were challenged with LH in the presence of TGF-␤ (10 ng/ml), with and without leptin (10 ng/ml). TIC were incubated for 48 h at 37ЊC in a humidified atmosphere containing 5% CO 2 in air. Cultures were terminated at 48 h, and media were removed and frozen at Ϫ20ЊC until analyzed for androsterone content by RIA [19]. The cells were frozen at Ϫ80ЊC pending extraction of RNA.
Leptin Receptor (OB-R) Isoform mRNA Expression GC and TIC were harvested after the 48-h incubation period described above. For all reverse transcription (RT)polymerase chain reaction (PCR) reactions, total RNA, DNA, and protein were extracted from the cells using the Tri Reagent method, according to the manufacturer's protocol (Molecular Research Center, Inc., Cincinnati, OH). Four replicate wells were pooled from GC and TIC cultures, and RT was performed as described [20]. Previous studies have shown that the multiple isoforms of the leptin receptor present in rat hypothalamus can be grouped into three classes: short forms with truncated intracellular domains (OB-Ra, OB-Rc, OB-Rd), the full-length signaltransducing isoform (OB-Rb), and a soluble isoform lacking the transmembrane and intracellular domains (OB-Re) [11]. Therefore, as a positive control, RNA was extracted from fresh hypothalamus tissue harvested from intact 26day-old female Sprague-Dawley rats. All samples were amplified using oligonucleotide primers (synthesized by GIB-CO-BRL) previously shown to amplify the leptin receptor isoforms OB-Ra, OB-Rb, and OB-Re [11]. After 35 cycles of PCR (94ЊC, 1 min; 52ЊC, 1 min; 72ЊC, 1.5 min), the amplification products were separated on a 2% agarose gel and visualized with ethidium bromide.

Measurement of Aromatase Cytochrome P450 (P450 arom ) mRNA
In order to determine the effect of leptin on P450 arom mRNA expression, RNA was extracted from GC cultures as described above. P450 arom mRNA was measured using semiquantitative RT-PCR. Primers (sense: 5Ј-ACT GTG CCT GTC AGT GCC AT-3Ј; antisense: 3Ј-GAC CAG AAT AAG CTT ACC A-5Ј) were synthesized in our lab (using an Applied Biosystems model 391 DNA synthesizer, Foster City, CA) and were designed to amplify a 426-base pair (bp) segment of the rat P450 arom cDNA [21]. To control for variations in individual PCR reactions, a mutant control P450 arom cDNA fragment was synthesized by site-directed mutagenesis [22]. In the P450 arom cDNA, a C was substituted for a T at base 320 to introduce an MspI restriction site. The resultant mutant cDNA can be amplified by the P450 arom primers but can be distinguished from the amplified wild-type P450 arom cDNA by digestion with MspI. The control cDNA (1 pg) was included in each PCR reaction (25 cycles: 94ЊC, 1 min; 55ЊC, 2 min; 72ЊC, 1 min), and all samples from each experiment were amplified at the same time in the presence of [ 32 P]dCTP. The amplification products were separated on a 2% agarose gel and visualized with ethidium bromide. The individual bands were cut from the gel and counted in a ␤-spectrometer. P450 arom mRNA values were normalized to ␤-actin mRNA levels measured [23] in the same samples to account for procedural variability and differences in cell number.

P450 arom Activity
P450 arom activity was estimated by measuring the production of 3 H 2 O from [1␤-3 H]-A 4 [24]. GC (5 ϫ 10 5 viable GC/well, 1 ml final volume) were incubated in 6-well plates without hormones (control), with FSH alone (0.1 IU/ml), with FSH plus TGF-␤ (10 ng/ml), and with a combination of FSH, TGF-␤, and leptin (10 ng/ml). After 48 h, fresh hormones were added to the appropriate wells and [1␤-3 H]-A 4 (2 ϫ 10 6 cpm, 0.1 M) was added to all wells. After a 4-h incubation at 37ЊC, the media were removed and the amount of 3 H 2 O produced was measured [24]. Briefly, trichloroacetic acid (TCA) was added, and the precipitated proteins were removed by centrifugation at 1700 ϫ g for 15 min. The supernatants were collected, and 1 ml of H 2 Osaturated chloroform was added. The reactions were vigorously shaken for 5 sec; then the aqueous phase was aspirated from each tube and mixed with an ice-cold 5% charcoal, 0.5% Dextran T-70 solution to remove the unreacted substrate. The mixtures were centrifuged (1700 ϫ g, 15 min). The supernatants were collected, scintillation fluid was added, and then the supernatants with scintillation fluid added were counted in a ␤-spectrometer. To control for variations in cell numbers, GC were scraped from the wells and protein levels were measured using the Bradford method [25].

Statistical Analyses
Treatments were administered in quadruplicate, and each experiment was repeated a minimum of three times. Mean values from independent experiments were statistically analyzed by unpaired t-test, and multiple comparisons were performed using one-way ANOVA followed by Tukey's test. Values were determined to be significant when P Յ 0.05.

Leptin Receptor Isoform mRNA Expression in Rat GC and TIC
We examined the cell-specific expression of OB-Ra, OB-Rb, and OB-Re in the immature rat ovary. In GC, the mRNA for OB-R isoforms OB-Ra (Fig. 1, lane 1) and OB-Rb (Fig. 1, lane 2) were expressed. In contrast, TIC expressed only the OB-Ra mRNA (Fig. 1, lane 1). OB-Re mRNA was not detected in either GC or TIC (Fig. 1, lane  3). We tested the direct effect of leptin on androgen production in TIC. As shown in Figure 2, there was no effect of leptin on LH-dependent androgen production in the presence or absence of TGF-␤. Leptin (0.1-100 ng/ml) did not alter LH-dependent androsterone levels after 48 h in vitro (data not shown). These results showing a lack of an effect of leptin on TIC androgen production are consistent with the observation that the TIC did not express the signaltransducing isoform of the leptin receptor (OB-Rb).

Effect of Leptin on GC Estrogen Production
FSH stimulated E 2 production in GC (Fig. 3). The maximal stimulatory effect of FSH was detected in the presence of 0.1 and 1.0 IU/ml FSH. In the presence of 0.1 and 1.0 IU/ml FSH, TGF-␤ augmented FSH-dependent E 2 accumulation by 2.7-and 1.45-fold, respectively (Fig. 3). Leptin did not significantly alter basal or FSH-dependent E 2 production, whereas leptin did impair the synergistic effect of TGF-␤ on FSH-stimulated E 2 synthesis (Fig. 3).
In order to determine whether there was a selective effect of leptin on the conversion of E 1 to E 2 (17␤-hydroxysteroid dehydrogenase activity), E 1 levels in GC-conditioned media were measured. As expected, FSH induced a dose-dependent increase in E 1 accumulation, and leptin did not alter the FSH effect (Fig. 4). In the presence of FSH (0.01, 0.1, and 1.0 IU/ml), TGF-␤ significantly augmented E 1 accumulation. When leptin was added to FSH-and TGF-␤-stimulated cells, E 1 levels were diminished at the 2 highest concentrations of FSH tested (0.1 and 1.0 IU/ml) (Fig.  4).

Effects of Leptin on P450 arom mRNA Levels and Activity
In order to understand the mechanism of leptin interference with TGF-␤ stimulation of estrogen production, the effect of leptin on P450 arom mRNA expression was examined. In the absence of FSH, P450 arom mRNA expression was not stimulated above control levels by TGF-␤, leptin, or TGF-␤ plus leptin (Fig. 5). FSH (0.01 and 0.1 IU/ml) induced an increase in P450 arom mRNA above control levels that was not significantly altered by leptin. TGF-␤ augmented the stimulatory effect of FSH (0.1 IU/ml) on P450 arom mRNA levels (Fig. 5). When GC were treated with leptin in the presence of FSH plus TGF-␤, P450 arom mRNA levels were reduced to levels equivalent to FSHstimulated levels.
We next measured the effect of leptin on P450 arom activity. As shown in Figure 6, P450 arom activity was increased in FSH-treated GC compared to untreated control cells. In the presence of TGF-␤, FSH-stimulated P450 arom was augmented 4-fold. Addition of leptin reduced the up-regulation in FSH-dependent P450 arom activity by 27% (Fig. 6).

DISCUSSION
Several isoforms of the leptin receptor (OB-R) have been identified in rodents [11]. This report demonstrates that the signal-transducing isoform of the leptin receptor (OB-Rb) [26] was expressed in leptin-sensitive rat GC. Interestingly, the pattern of expression of OB-R isoforms was cell-specific within the immature rat ovary. In GC, OB-Rb and OB-Ra mRNAs were detected, whereas, in LH-treated TIC obtained from hypophysectomized rats, only isoform OB-Ra mRNA was expressed. The demonstration that the GC responded to leptin and the TIC did not provides further evidence that leptin bioactivity in the immature rat ovary is mediated by the long OB-Rb isoform and not by the short OB-Ra isoform. These data are consistent with reports that OB-Ra may bind leptin but is not involved in leptin signaling [26]. Whether OB-Ra can bind leptin in the ovary has not been determined. In human theca, the long form of the OB-R is expressed [7,27], and the theca cells respond to leptin [7]. The difference in OB-R expression between rat and human theca may be due to the fact that the rat theca were obtained from immature hypophysectomized animals. Alternatively, there may be species differences. For example, in the bovine ovary, 125 I-labeled leptin binding was detected in theca cells [28], whereas, in the human ovary (GC and theca cells) both long and short isoforms of the leptin receptor may be expressed [7,27]. The precise intraovarian mechanisms of leptin action are unknown, but a point of convergence between the FSH/TGF-␤ and leptin signaling pathways appears to occur. Further studies will be required in order to determine how leptin affects FSH/ TGF-␤ signaling cascades and how these changes regulate P450 arom mRNA expression and P450 arom enzyme activity in GC.
Several lines of evidence support the conclusion that TGF-␤ is an important intraovarian regulator that potentiates FSH action in GC. First, TGF-␤ mRNAs are expressed by GC [29] and TIC [30], and TGF-␤ is secreted by TIC [31]. Second, in rat GC, TGF-␤ augments the FSH-stimulated cAMP-dependent second messenger pathway by increasing FSH-induced cAMP levels [32] and synthesis of both cAMP-dependent protein kinase RII␤ subunit mRNA and protein [33]. Third, TGF-␤ stimulates GC growth [13] and steroidogenesis [14] in vitro. The present study expands this body of knowledge by demonstrating an up-regulatory effect of TGF-␤ on FSH-stimulated P450 arom mRNA expression and P450 arom activity in rat GC.
In previous reports by our laboratory and others, leptin was shown to reduce the synergistic effect of IGF-I and/or insulin on FSH-dependent E 2 synthesis in rat [5], bovine [6], and human [7] GC in vitro. These observations raised the question whether the direct effect of leptin in GC was specific to IGF-I or if leptin exerted a generalized effect on multiple positive modulators of FSH action. This report demonstrated that leptin blocks the positive modulatory effects of TGF-␤ on FSH-dependent estrogen (E 1 and E 2 ) production. Hence, the actions of TGF-␤ and IGF-I, two well-characterized stimulatory modulators of GC function, are impaired by leptin. These data support the concept that the intracellular signaling pathways mediating TGF-␤ and IGF-I enhancement of P450 arom gene transcription in GC converge at a common point. Leptin appears to block P450 arom gene transcription distal to the point of convergence.
In the present study, leptin blocked the TGF-␤-dependent increase in FSH-stimulated estrogen synthesis. It appears that there was a greater effect of leptin on estrogen production than on aromatase activity. Whether or not the apparent difference is important is unclear. The concentrations of estrogens measured in conditioned medium reflect the net metabolism of A 4 over a 48-h period of time by a combination of steroidogenic enzymes. E 2 synthesis from A 4 requires the aromatase and 17␤-hydroxysteroid dehydrogenase enzymes and may be influenced by other enzymes such as 5␣-reductase that can metabolize the A 4 substrate or others that could metabolize the estrogen products. Unlike RIA, the aromatase enzyme assay specifically measures aromatase activity. The results of these experiments indicate that leptin treatment may affect not only aromatase activity but also the activities of other steroidogenic enzymes in granulosa cells.
Leptin mRNA has been detected in human GC [34], and immunoreactive leptin has been found in human follicular fluid [7,27]. Together the aforementioned show a potential intraovarian leptin system, replete with ligand and one or more OB-R isoforms. Importantly, leptin production by human GC has yet to be demonstrated, and there is no difference between circulating and follicular fluid concentrations of leptin [7,27], indicating that intraovarian leptin is likely to be of endocrine origin. It appears that the physiologic role of leptin in the ovary may be limited to conditions of obesity. For example, the circulating concentrations of leptin in lean women are too low to alter ovarian function significantly [7]. In contrast, the levels observed in obese women are sufficient to interfere with the sensitizing actions of IGF-I and TGF-␤ on FSH-dependent E 2 production. Such an effect could inhibit fertility because sensitization of small antral follicles to FSH by intraovarian factors (i.e., IGF-I and TGF-␤) is thought to be important for selection of dominant follicles [35]. Disruption of E 2 production during follicle growth could cause follicle atresia. Thus, by counteracting the effect of TGF-␤ in GC, leptin may interfere with an essential support mechanism (e.g., augmentation of E 2 production) that promotes follicle growth and maturation. Such a mechanism can help to explain how weight loss in obese women can improve their fertility [36].