Direct Thiazolidinedione Action in the Human Ovary: Insulin ...

0021-972X/05/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 90(11):6099 – 6105 Copyright © 2005 by The Endocrine Society doi: 10.1210/jc.2005-0469

Direct Thiazolidinedione Action in the Human Ovary: Insulin-Independent and Insulin-Sensitizing Effects on Steroidogenesis and Insulin-Like Growth Factor Binding Protein-1 Production Donna Seto-Young, Maria Paliou, Jonathan Schlosser, Dimiter Avtanski, Alice Park, Parini Patel, Kevin Holcomb, Peter Chang, and Leonid Poretsky Division of Endocrinology, Departments of Medicine (D.S.-Y., M.P., J.S., D.A., P.P., L.P.) and Obstetrics and Gynecology (A.P., K.H., P.C.), Beth Israel Medical Center and Albert Einstein College of Medicine, New York, New York 10003; and Institute of Biology and Immunology of Reproduction (D.A.), Bulgarian Academy of Sciences, Sofia, Bulgaria, 1113 Context and Objective: Hyperinsulinemia contributes to the pathogenesis of ovarian dysfunction in insulin-resistant states, including polycystic ovary syndrome (PCOS). Peroxisome proliferator activated receptor-␥ (PPAR-␥) agonists [thiazolidinediones (TZDs)] ameliorate hyperandrogenism in polycystic ovary syndrome presumably because they reduce systemic hyperinsulinemia. Direct effects of TZDs in the ovary, however, cannot be excluded. We explored direct effects of TZDs in cultured human ovarian cells. Methods: Human ovarian cells, obtained from oophorectomy specimens, were cultured in the presence or absence of rosiglitazone or pioglitazone, insulin, and gonadotropins. Steroid hormone and IGFbinding protein-1 (IGFBP-1) concentrations were measured in conditioned tissue culture medium. Results: Rosiglitazone or pioglitazone stimulated progesterone production up to 156% (P ⬍ 0.001) and 131% (P ⬍ 0.001) of baseline, respectively. Pioglitazone but not rosiglitazone, inhibited baseline and FSH-stimulated estradiol production by 20% (P ⬍ 0.001) and 50%

I

NSULIN PARTICIPATES IN the regulation of ovarian steroidogenesis and folliculogenesis and plays an important role in the pathogenesis of ovarian dysfunction in patients with insulin-resistant states (1–5). It is thought that in these conditions hyperinsulinemia excessively stimulates ovarian androgen production, contributing to hyperandrogenism, anovulation, and infertility (1, 4, 5). On the other hand, insulin inhibits IGF-binding protein-1 (IGFBP-1) production, resulting in increased bioavailability of IGFs and enhancement of folliculogenesis (1). Ovarian hyperandrogenism, commonly observed in insulin-resistant states, can be ameliorated by insulin-sensitizing agents, including thiazolidinediones (TZDs) (6 –12). TZDs act as peroxisome proliferator activated receptor-␥ (PPAR-␥) agonists and include, among other agents, troglitazone, rosiglitazone, and pioglitazone. Recent studies suggested that,

First Published Online August 30, 2005 Abbreviations: FBS, Fetal bovine serum; IGFBP-1, IGF-binding protein-1; PCOS, polycystic ovary syndrome; PPAR-␥, peroxisome proliferator activated receptor-␥; TZD, thiazolidinedione. JCEM is published monthly by The Endocrine Society (http://www. endo-society.org), the foremost professional society serving the endocrine community.

(P ⬍ 0.001), respectively. Both rosiglitazone and pioglitazone abolished insulin-dependent stimulation of estradiol production in the presence of FSH. Rosiglitazone and pioglitazone inhibited testosterone production by 10% (P ⬍ 0.012) and 15% (P ⬍ 0.023), respectively, and abolished insulin-induced stimulation of testosterone production. In the absence of insulin, pioglitazone or rosiglitazone stimulated IGFBP-1 production up to 160% (P ⬍ 0.001) and 125% (P ⬍ 0.036) of baseline, respectively. Pioglitazone and rosiglitazone enhanced insulin-induced inhibition of IGFBP-1 production by 13% and 20%, respectively (P ⬍ 0.001). Conclusions: PPAR-␥ agonists directly stimulate progesterone and IGFBP-1 production, inhibit estradiol and testosterone production, abolish insulin-induced stimulation of testosterone production and insulin-dependent stimulation of estradiol production in the presence of FSH, and enhance insulin-induced inhibition of IGFBP-1 production in human ovarian cells. PPAR-␥ represents a novel system of ovarian regulation. (J Clin Endocrinol Metab 90: 6099 – 6105, 2005)

in addition to reducing circulating androgen levels (9 –12), TZDs can normalize (increase) circulating progesterone levels, which are commonly reduced in patients with polycystic ovary syndrome (PCOS) (13) and in a primate model of this syndrome (14). Although it is thought that the inhibitory effects of TZDs on ovarian androgen production are a result of the reduction of circulating insulin levels induced by TZDs, in several in vitro and in vivo studies in animal and human ovarian cells, TZDs have been reported to directly inhibit or increase activity of ovarian steroidogenic enzymes (15–20), progesterone (18), and estradiol (14, 19) production as well as androgen biosynthesis induced by LH and insulin (20). In this report, we explored whether PPAR-␥ agonists, rosiglitazone and pioglitazone, have direct effects on ovarian steroidogenesis and IGFBP-1 production. Patients and Methods All studies described in this report were approved by the Institutional Review Board at Beth Israel Medical Center. Patient’s consent, allowing the use of ovarian tissue in research studies, was obtained before surgery.

6099

6100

J Clin Endocrinol Metab, November 2005, 90(11):6099 – 6105

Seto-Young et al. • Thiazolidinedione Action in the Human Ovary

Patients

IGFBP-1 production

Ovarian tissue samples were collected from 25 women at the time of indicated oophorectomy. All patients were premenopausal or perimenopausal. The patient age ranged from 25–54 yr. No patient was using hormonal contraception or other hormonal therapy at the time of oophorectomy. Reasons for oophorectomy included follicular or dermoid cysts and benign uterine or ovarian neoplasms. Only unaffected ovarian tissue was used for the cell culture. For each experiment, the ovarian tissue from a single patient was used. The experiments were repeated with ovarian tissue obtained from four to 10 patients.

The cells were incubated in tissue culture medium as described in the experiments assessing progesterone production, except that, during the final 24-h incubation, the medium was supplemented with 0.1% (rather than 1%) FBS. A lower concentration of FBS was used because FBS, while increasing cell survival, inhibits IGFBP-1 synthesis (23, 24). The medium was supplemented with 0, 10, 102, 103, or 104 ng/ml insulin in the presence of 50 ␮m rosiglitazone or pioglitazone. IGFBP-1 concentration in the conditioned medium was determined by immunoradiometric assay (Diagnostic Systems Laboratories).

Cell culture system

Materials

The cell culture used in these studies was developed and previously described by us (21). Briefly, the unaffected ovarian tissue was obtained from oophorectomy specimens after dissection, which separated cysts or benign tumors from the tissue used to establish the culture. Tissue fragments were placed in the McCoy’s 5A tissue culture medium supplemented with 10% fetal bovine serum (FBS) (Gemini Bio-Products, Woodland, CA), 100 ␮g/ml penicillin, 100 ␮g/ml streptomycin, and 250 ng/ml amphotericin. The capsule was removed, and the tissue was cut into small pieces (⬃0.5 cm2 ⫻ 0.1 cm) and then homogenized in a glass homogenizer. The tissue was then plated in 60 ⫻ 15-mm tissue culture dishes. Sterile Vaseline was placed around the tissue fragments, and the fragments were covered with a 24 ⫻ 30-mm glass cover slide. Four milliliters of the culture medium were added to the culture dishes. The ovarian tissue was cultured without passaging (usually for 2–3 wk) until the cells were confluent. The cells were then trypsinized and subcultured until adequate cell numbers required for the experiments were available. Enzymatic and immunocytochemistry studies demonstrated the presence of stroma, theca, and granulosa cells in this tissue culture system (21). Optimal concentrations of rosiglitazone or pioglitazone for a maximal effect were determined in dose-response experiments for each of the examined steroid hormones and for IGFBP-1. Approximately 5 ⫻ 104 cells per well were used in each experiment; cell protein concentration was approximately 0.1 mg per well (22).

Dehydroepiandrosterone, pregnenolone, progesterone, and testosterone were obtained from Sigma Chemical Co. (St. Louis, MO). LH and FSH were obtained from Calbiochem (San Diego, CA). Insulin was obtained from Roche Applied Science (Indianapolis, IN). Rosiglitazone was obtained from Cayman Chemical (Ann Arbor, MI), and pioglitazone was obtained from Takeda Pharmaceuticals America, Inc. (Lincolnshire, IL).

Progesterone production The cells were incubated in McCoy’s 5A medium supplemented with 10% FBS, 100 ␮g/ml penicillin, 100 ␮g/ml streptomycin, and 250 ng/ml amphotericin for 48 h and then in 1% FBS medium for another 24 h. The cells were then incubated in 1% FBS medium containing pregnenolone (30 ␮m) as substrate in the presence of 0, 10, 102, 103, or 104 ng/ml insulin and in the presence of 0 or 30 ␮m rosiglitazone or pioglitazone. The tissue culture medium was removed after 24 h incubation. Progesterone concentration in the conditioned medium was measured using RIA kits from Diagnostic Systems Laboratories, Inc., Webster, TX.

Estradiol production The cells were incubated for 48 h in McCoy’s 5A medium supplemented with 10% FBS, 100 ␮g/ml penicillin, 100 ␮g/ml streptomycin, and 250 ng/ml amphotericin and then in 1% FBS medium for 24 h. The cells were then incubated in 1% FBS medium containing testosterone (30 ␮m) as substrate in the presence of 0, 10, 102, 103, or 104 ng/ml insulin and in the presence of rosiglitazone or pioglitazone (30 ␮m). The tissue culture medium was removed after 24 h incubation. Estradiol concentration in the conditioned medium was measured using RIA kits from Diagnostic Systems Laboratories.

Testosterone production The cells were incubated for 48 h in McCoy’s 5A medium supplemented with 10% FBS, 100 ␮g/ml penicillin, 100 ␮g/ml streptomycin, and 250 ng/ml amphotericin and then in 1% FBS medium with 10 ng/ml insulin for 24 h. The cells were then incubated in 1% FBS medium containing dehydroepiandrosterone (30 ␮m) as substrate in the presence of 0, 10, 102, 103, or 104 ng/ml insulin and in the presence of 0 or 20 ␮m rosiglitazone or pioglitazone. The tissue culture medium was removed after 24 h incubation. Testosterone concentration in the conditioned medium was measured using RIA kits from Diagnostic Systems Laboratories.

Statistical analysis All experiments were carried out in quadruplicate and repeated four to 10 times. Two-way ANOVA was used to compare mean values according to insulin concentrations in the presence or absence of rosiglitazone or pioglitazone. The statistical interactions between the sets of data obtained with or without rosiglitazone or pioglitazone were examined. Pairwise Bonferroni-adjusted contrasts were used to determine statistical significance. Because absolute concentration of steroid hormones and IGFBP-1 in the tissue culture medium at baseline differed among the samples from different patients, the results were normalized to the each sample’s baseline level, which was considered 100%.

Results

The concentration of progesterone in the conditioned tissue culture medium at baseline ranged from 300 – 600 nmol/ liter, of estradiol from 1.1– 4.8 nmol/liter, of testosterone from 70 – 420 nmol/liter, and of IGFBP-1 from 0.05– 0.31 nmol/liter. Progesterone production

Insulin alone stimulated progesterone production up to 125% of baseline (P ⬍ 0.004) (Fig. 1). Rosiglitazone (30 ␮m) alone stimulated progesterone production up to 156% (P ⬍ 0.001), and pioglitazone (30 ␮m) alone stimulated progesterone production up to 131% of baseline (P ⬍ 0.001). Adding insulin to either rosiglitazone or pioglitazone produced no significant additional effect on progesterone production. There was no statistically significant difference between rosiglitazone or pioglitazone effects on progesterone production. Estradiol production

Insulin alone stimulated estradiol production up to 122% of baseline (P ⬍ 0.049) (Fig. 2A). In the absence of insulin, pioglitazone (30 ␮m) inhibited estradiol production by 15% (P ⬍ 0.003), whereas rosiglitazone had no effect. The stimulatory effect of insulin on estradiol production was reduced by 20% (P ⬍ 0.001) in the presence of pioglitazone, whereas rosiglitazone, once again, had no effect. Pioglitazone inhibited FSH-dependent estradiol production by 50% (P ⬍ 0.001) (Fig. 2B), whereas rosiglitazone had no effect. In the presence of FSH (100 ng/ml), insulin-induced estradiol production was stimulated up to 160% of baseline (P ⬍ 0.001). Rosigli-



6101

FIG. 1. TZD-induced stimulation of progesterone production in human ovarian cells. ␣, P ⬍ 0.004 (n ⫽ 10) compared with control (in the absence of insulin, rosiglitazone, or pioglitazone); ␤, P ⬍ 0.001 (n ⫽ 4) compared with absence of insulin, rosiglitazone, or pioglitazone; ␥, P ⬍ 0.001 (n ⫽ 4) compared with absence of insulin, rosiglitazone, or pioglitazone. To convert ng/ml insulin to pmol/liter, multiply by 167.

tazone and pioglitazone completely abolished insulindependent stimulation of estradiol production in the presence of FSH. The inhibitory effect of pioglitazone (30 ␮m) on FSH/insulin-stimulated estradiol production was more potent than that of rosiglitazone (P ⬍ 0.001).

pletely abolished insulin-induced stimulation of testosterone production (Fig. 3). The inhibitory effect of pioglitazone on basal and insulin-induced testosterone production was more potent than that of rosiglitazone (P ⬍ 0.05). IGFBP-1 production

Testosterone production

Insulin alone stimulated testosterone production up to 131% of baseline (P ⬍ 0.001) (Fig. 3). Rosiglitazone (20 ␮m) alone inhibited testosterone production by up to 10% (P ⬍ 0.012), and pioglitazone (20 ␮m) alone inhibited testosterone production by up to 15% (P ⬍ 0.023). Rosiglitazone and pioglitazone com-

FIG. 2. TZD-induced inhibition of estradiol production in the absence of FSH (A) and in the presence of 100 ng/ml FSH (B). ␣, P ⬍ 0.049 (n ⫽ 9) compared with the absence of insulin, rosiglitazone, or pioglitazone; ␤, P ⬍ 0.001 (n ⫽ 6) in the presence of pioglitazone compared with absence of insulin or rosiglitazone; ␥, P ⬍ 0.001 (n ⫽ 4) compared with absence of pioglitazone or rosiglitazone in the presence of insulin; ␦, P ⬍ 0.001 (n ⫽ 7) compared with absence of insulin, rosiglitazone, or pioglitazone; ⑀, P ⬍ 0.001 (n ⫽ 4) compared with the absence of rosiglitazone or pioglitazone; ␲, P ⬍ 0.001 (n ⫽ 4) compared with absence of rosiglitazone or pioglitazone. To convert ng/ml insulin to pmol/liter, multiply by 167.

In the absence of insulin, rosiglitazone (50 ␮m) stimulated IGFBP-1 production up to 160% of baseline (P ⬍ 0.001) and pioglitazone (50 ␮m) up to 125% of baseline (P ⬍ 0.036) (Fig. 4A). The stimulatory effect of rosiglitazone on IGFBP-1 production in the absence of insulin was more potent than that of pioglitazone (P ⬍ 0.024). Insulin alone inhibited IGFBP-1 pro-

6102



FIG. 3. TZD-induced inhibition of testosterone production. ␣, P ⬍ 0.001 (n ⫽ 9) compared with absence of insulin, rosiglitazone, or pioglitazone; ␤, P ⬍ 0.012 (n ⫽ 4) compared with the absence of insulin, rosiglitazone, or pioglitazone; ␥, P ⬍ 0.023 (n ⫽ 9) compared with the absence of insulin, rosiglitazone, or pioglitazone; ␦, P ⬍ 0.001 (n ⫽ 9) compared with the presence of insulin, in the absence of rosiglitazone or pioglitazone; ⑀, P ⬍ 0.001 (n ⫽ 4) compared with the presence of insulin, in the absence of rosiglitazone or pioglitazone. To convert ng/ml insulin to pmol/liter, multiply by 167.

duction by 42% (P ⬍ 0.001) (Fig. 4B). In the presence of insulin, pioglitazone enhanced insulin-induced inhibition of IGFBP-1 production by 13% (P ⬍ 0.001, compared to insulin alone) and rosiglitazone by 20% (P ⬍ 0.001, compared to insulin alone) (Fig. 4B). There was no statistically significant difference between the effects of pioglitazone or rosiglitazone on insulininduced inhibition of IGFBP-1 production.

Discussion

Traditionally, gonadotropins have been viewed as exclusive regulators of ovarian function. More recently, it has been demonstrated that insulin and IGFs, their receptors, IGFBPs, and IGFBP-proteases participate in the regulation of ovarian function (1–5, 25–27). We examined

FIG. 4. TZD effect on IGFBP-1 production. A, In the absence of insulin; data are normalized to the control level of IGFBP-1 (the absence of insulin, rosiglitazone, or pioglitazone). B, In the absence of (0 ng/ml insulin) or presence of (1–103 ng/ml) insulin; data are normalized to the baseline IGFBP-1 production for each experimental group. ␣, P ⬍ 0.001 (n ⫽ 10) compared with the absence of insulin, rosiglitazone, or pioglitazone; ␤, P ⬍ 0.036 (n ⫽ 10) compared with the absence of insulin, rosiglitazone, or pioglitazone; ␥, P ⬍ 0.001 (n ⫽ 10) compared with the absence of insulin, rosiglitazone, or pioglitazone; ␦, P ⬍ 0.002 (n ⫽ 10) compared with the presence of insulin in the absence of rosiglitazone or pioglitazone; ⑀, P ⬍ 0.029 (n ⫽ 10) compared with the presence of insulin in the absence of rosiglitazone or pioglitazone. To convert ng/ml insulin to pmol/liter, multiply by 167.


whether PPAR-␥ may represent yet another system of ovarian regulation. In this report, we demonstrated that PPAR-␥ agonists rosiglitazone and pioglitazone directly regulate ovarian steroidogenesis. Both rosiglitazone and pioglitazone enhanced progesterone production and inhibited testosterone synthesis. Pioglitazone inhibited basal and FSH-stimulated estradiol production whereas rosiglitazone had no significant effect. However, both PPAR-␥ agonists completely abolished stimulation of estradiol production induced synergistically by FSH and insulin as well as insulininduced stimulation of testosterone production. In the absence of insulin, both pioglitazone and rosiglitazone stimulated IGFBP-1 production. Both TZDs, however, enhanced insulin-induced inhibition of IGFBP-1 production. The cell system, used in this study, has been described by us in detail and is derived from oophorectomy specimens (21). Our cell culture system contains cells from three types of ovarian steroidogenic compartments: granulosa, theca, and stroma. In this kind of system, it is difficult to be certain whether an observed effect is a result of a primary action on the cells representing one compartment, because primary effects on one cellular compartment may produce secondary effects in another compartment. For example, if theca cells are stimulated by gonadotropins or insulin to increase production of androgens (primary effect), the increased amount of androgens may be used as a substrate for aromatization by granulosa cells, resulting in increased synthesis of estrogens (secondary effect). Experiments with isolated cell cultures derived from a single steroidogenic cell compartment (i.e. isolated granulosa, stroma, or theca cultures) are necessary to clarify these mechanisms. However, our cell culture system, which contains all steroidogenic ovarian compartments, reflects normal ovarian physiology in vivo, where all cellular components are present simultaneously and interact with each other. Oophorectomy specimens were used as a source for our cell culture system. Although we used unaffected ovarian tissue, our cells may have been influenced by pathological processes for which oophorectomies were performed (28). In each of our experiments, however, a specimen from a single patient was used for both control and experimental cell samples. The experiments were repeated with cells from four to 10 patients, and although the absolute amount of steroid hormones and IGFBP-1 produced by the cells varied among the patients, similar effects were observed in tissue samples obtained from different patients. Furthermore, in some patients, the pathology for which oophorectomy was performed involved only the uterus. The concentrations of insulin that we used ranged from physiological (1–10 ng/ml), to moderately supraphysiological (10 –100 ng/ml), to significantly supraphysiological (up to 104 ng/ml). The significance of studying physiological concentrations of insulin is quite clear, because insulin participates in the regulation of normal ovarian function, including steroidogenesis and ovulation (1). Moderately supraphysiological circulating concentrations of insulin are observed in patients with PCOS (1), whereas significantly supraphysiological concentrations are present in patients with the syndromes of extreme insulin resistance and acanthosis nigricans (1– 4). Thus, our data regarding the response


6103

of cultured ovarian cells to insulin are relevant for both physiological and pathological conditions in the ovary. The concentrations of rosiglitazone and pioglitazone that we used were chosen in dose-response experiments based on the optimal cell responses. Circulating concentrations of pioglitazone in patients receiving this compound are approximately 5 ␮m (29), whereas for rosiglitazone, these concentrations range from 0.25–2.5 ␮m (30). Target tissue concentrations of these TZDs, however, are 10 –20 times higher than circulating concentrations (Dr. Ronald Law, Takeda Pharmaceuticals, personal communication), thus approximating concentrations used in our studies. Furthermore, concentrations of rosiglitazone or pioglitazone used in other reported in vitro studies ranged from 10 –25 ␮m (18, 31) and are therefore comparable to the concentrations used by us (20 –50 ␮m). At the concentrations that we examined, pioglitazone appeared to be more potent than rosiglitazone in its ability to inhibit estradiol or testosterone production, whereas rosiglitazone appeared to be more potent than pioglitazone in its ability to stimulate insulin-independent IGFBP-1 production. Pioglitazone and rosiglitazone exhibited similar potency in their effects on progesterone and on insulin-induced inhibition of IGFBP-1 production. Differences in the potency of different TZDs are commonly observed when these agents are used in vivo and may be due to their different binding characteristics in relation to PPAR-␥ (32) or, possibly, due to different signaling mechanisms used by TZDs for their specific effects. In our studies, TZDs had a clear stimulatory effect on progesterone production. It appears that under our experimental conditions, TZDs alone were capable of maximizing progesterone response. Once this response was maximized, insulin had no additional effect, although it had a potent independent effect on progesterone production in the absence of TZDs. The stimulatory effect of TZDs on progesterone production in our cell system is in agreement with in vivo studies, which demonstrated normalization of the reduced circulating progesterone levels in patients with PCOS by TZDs (13). We demonstrated that pioglitazone inhibited basal and FSH-stimulated estradiol production, whereas both rosiglitazone and pioglitazone abolished insulin-dependent stimulation of estradiol production in the presence of FSH. These findings may be either a result of the direct effect of TZDs on estradiol production or due to TZD-induced inhibition of testosterone production, because testosterone serves as a substrate for aromatase-dependent estradiol production. Additionally, TZDs may have direct inhibitory effect on aromatase (16). In our system, we would not expect testosterone concentration in the tissue culture medium to increase, even if aromatase activity is inhibited, because in the studies of estradiol production, we used a high concentration of testosterone as substrate (30 ␮m), whereas the concentration of estradiol as the end product was approximately 6 nm. Under these conditions, a small rise in the testosterone concentration in the conditioned tissue culture medium because of the inhibition of aromatase would not be detected. In in vivo experiments (19), in agreement with our studies, pioglitazone inhibited FSH-stimulated estradiol production.

6104


In the same study, pioglitazone also prolonged the estradiol peak observed in the course of the ovulatory cycle, an effect that could not be examined in our in vitro system. As we will discuss below, in vitro and in vivo TZD effects can be quite different because of the differences between the mechanisms involved in the direct TZD action and their systemic insulinsensitizing activity. TZDs in our cell culture system had a significant direct inhibitory effect on testosterone production, in agreement with in vivo observations in patients with PCOS (9 –11). Our studies suggest, however, that inhibition of testosterone production observed in PCOS patients treated with TZDs may be caused not only by a reduction of hyperinsulinemia (which occurs because of systemic insulin-sensitizing activity of TZDs) but also because of the direct action of TZDs in the ovary, specifically because of their direct inhibitory effect on testosterone synthesis. In an interesting display of the differences between insulin-independent and insulin-sensitizing effects of TZDs, pioglitazone and rosiglitazone stimulated IGFBP-1 production by ovarian cells in the absence of insulin but enhanced insulin-induced inhibition of IGFBP-1 production. The latter effect could presumably increase bioavailable concentration of IGFs, thus enhancing the role of IGF in the development of ovulatory follicles (1, 33). In PCOS patients treated with rosiglitazone, plasma levels of IGFBP-1 were reported to increase (34). This observation is unlikely to be a result of a direct stimulatory effect of TZDs on IGFBP-1 production in the ovary but rather due to the systemic insulin-sensitizing action of TZDs: with a reduction of hyperinsulinemia, circulating levels of IGFBP-1 will be expected to rise because IGFBP-1 synthesis in the liver is under inhibitory control of insulin (1, 35). The opposite (stimulation or inhibition) direct effects of TZDs on specific steroidogenic pathways and the differences between their insulin-dependent vs. insulin-independent effects may occur because of variable signaling pathways used by TZDs for their specific effects. Similarly, a variety of signaling mechanisms has been demonstrated to exist for the ovarian effects of insulin. For example, phosphatidylinositol-3 kinase (36) or MAPK (37) activation is not necessary for the insulin effects on steroidogenesis, indicating that classical insulin-signaling pathways, used for glucose transport and for mitogenic activity, differ from those involved in insulin effects on steroidogenesis. Furthermore, unlike steroidogenic effects of insulin, IGF-I effects on steroidogenesis,


which are mediated by type I IGF receptors, are MAPK dependent (37). The latter observation is important because in supraphysiological concentrations, insulin up-regulates and activates type I IGF receptors in the ovary (1, 38). In summary, in analogy with insulin and IGFs, TZDs may employ multiple postreceptor signaling pathways, thus explaining a variety of their effects in the ovary. The multiplicity of TZD effects in the ovary can be best understood in terms of their direct effects (both insulinindependent and insulin-sensitizing) vs. the indirect effects, which are caused by the systemic insulin-sensitizing action of the TZDs leading to a reduction in hyperinsulinemia (Table 1). Direct effects can be demonstrated in vitro and may be present in vivo, although in some cases, they may be minimized or even abolished by the systemic insulin-sensitizing TZD action. For example, TZDs increase IGFBP-1 production by ovarian cells (direct effect demonstrated in vitro) but enhance insulin-induced inhibition of IGFBP-1 production (direct insulin-sensitizing effect, also demonstrated in vitro). In vivo, because IGFBP-1 production in the liver is under inhibitory control of insulin (1, 35) and because circulating insulin levels fall under the influence of TZDs, TZDs produce an increase in circulating IGFBP-1 levels (34). Another example of the interaction between the direct and indirect TZD effects would be their influence on the circulating testosterone levels. The direct effect of TZDs leads to a reduction of both insulin-independent and insulin-dependent testosterone production in the ovary. In vivo, in addition, circulatory free testosterone levels are reduced because of systemic insulin-sensitizing action of TZDs: since the production of SHBG in the liver is under the inhibitory control of insulin (1, 39), reduction of hyperinsulinemia leads to a rise of SHBG and a fall in circulating free testosterone levels (1, 8). The multiplicity of TZD effects, combined with the potential multiplicity of signaling mechanisms employed by TZDs, provide for a complex novel system of ovarian regulation by PPAR-␥. We conclude that, in cultured human ovarian cells, TZDs 1) enhance progesterone production, 2) inhibit estradiol and testosterone production, 3) completely abolish insulin-induced stimulation of testosterone production, 4) completely abolish insulin-induced stimulation of estradiol production in the presence of FSH, 5) stimulate insulin-independent IGFBP-1 production, and 6) enhance insulin-induced inhi-

TABLE 1. Effects of TZDs related to ovarian function Direct: can be observed in vitro; may be present in vivo

Insulin-independent 1 Progesterone 2 Testosterone 2 Estradiol 1 IGFBP-1 (in the absence of insulin) Insulin-sensitizing (enhanced insulin effect) 2 IGFBP-1 production 1 Estradiol production (in vivo, in a setting of high-dose insulin infusion) 1, Increased; 2, decreased. Please see the text for references.

Indirect: observed in vivo; are due to systemic insulin-sensitizing action and reduction of hyperinsulinemia

2 1 1 2

Testosterone IGFBP-1 SHBG Free testosterone



bition of IGFBP-1 production. Thus, PPAR-␥ represents a novel regulatory system in the human ovary.

18.

Acknowledgments

19.

Received March 3, 2005. Accepted August 23, 2005. Address all correspondence and requests for reprints to: Leonid Poretsky, M.D., or Donna Seto-Young, Ph.D., Division of Endocrinology, Beth Israel Medical Center, 317 East 17th Street, Fierman Hall, Seventh Floor, New York, New York 10003. E-mail: [email protected] or [email protected]. This work was supported in part by Gerald J. Friedman and by the Thanks to Scandinavia Foundations, the Singer/Hellman award at Beth Israel Medical Center, Pharmacia Corp., Chinese American Medical Society and Independent Practice Association, and National Institutes of Health Grant RO1-MH60563.

References 1. Poretsky L, Cataldo N, Rosenwaks Z, Guidice L 1999 The insulin-related ovarian regulatory system in health and disease. Endocr Rev 20:535–582 2. Kahn CR, Flier JS, Bar RS, Archer JA, Gorden P, Martin MM, Roth J 1976 The syndromes of insulin resistance and acanthosis nigricans: insulin-receptor disorders in man. N Engl J Med 294:739 –745 3. Flier JS, Kahn CR, Roth J, Bar RS 1975 Antibodies that impair insulin receptor binding in an unusual diabetic syndrome with severe insulin resistance. Science 190:63– 65 4. Taylor SI, Dons RF, Hernandez E, Roth J, Gorden P 1982 Insulin resistance associated with androgen excess in women with autoantibodies to the insulin receptor. Ann Intern Med 97:851– 855 5. Barbieri RL, Smith S, Ryan KJ 1988 The role of hyperinsulinemia in the pathogenesis of ovarian hyperandrogenism. Fertil Steril 50:197–212 6. Lord JM, Flight IHK, Norman RJ 2004 Insulin-sensitizing drugs (metformin, troglitazone, rosiglitazone, pioglitazone, d-chiro-inositol) for polycystic ovary syndrome. Cochrane Database Systemat Rev 4:82–119 7. Guido M, Romualdi D, Suriano R, Giuliani M, Costantini B, Apa R, Lanzone A 2004 Effect of pioglitazone treatment on the adrenal androgen response to corticotrophin in obese patients with polycystic ovary syndrome. Hum Reprod 19:534 –539 8. Ghazeeri G, Kutteh WH, Bryer-Ash M, Haas D, Ke RW 2003 Effect of rosiglitazone on spontaneous and clomiphene citrate-induced ovulation in women with polycystic ovary syndrome. Fertil Steril 79:562–566 9. Brettenthaler N, DeGeyter C, Huber PR, Keller U 2004 Effect of the insulin sensitizer pioglitazone on insulin resistance, hyperandrogenism, and ovulatory dysfunction in women with polycystic ovary syndrome. J Clin Endocrinol Metab 89:3835–3840 10. Sepilian V, Nagamani M 2005 Effects of rosiglitazone in obese women with polycystic ovary syndrome and severe insulin resistance. J Clin Endocrinol Metab 90:60 – 65 11. Azziz R, Ehrmann D, Legro RS, Whitcomb RW, Hanley R, Fereshetian AG, O’Keefe M, Ghazzi MN 2001 Troglitazone improves ovulation and hirsutism in the polycystic ovary syndrome: a multi-center, double blind, placebo-controlled trial. J Clin Endocrinol Metab 86:1626 –1632 12. Bloomgarden Z, Futterweit W, Poretsky L 2001 Use of insulin sensitizing agents in patients with polycystic ovary syndrome. Endocr Pract 7:279 –286 13. Cataldo NA, McLaughlin TL, Abbasi F, Basina M, Fechner PY, Reaven GM, Giudice LC 2004 Prediction of ovulatory response to rosiglitazone in insulinresistant women with polycystic ovary syndrome. Program of the 86th Annual Meeting of The Endocrine Society, New Orleans, LA, 2004, p 118 (Abstract OR31-3) 14. Abbott DH, Zhou R, Bruns CM, Bird IM, Kemnitz JW, Turk JM, Dumesic DA 2004 Pioglitazone normalizes menstrual cycle duration and improves insulin sensitivity in a nonhuman primate model for polycystic ovary syndrome. Program of the 86th Annual Meeting of The Endocrine Society, New Orleans, LA, 2004, p 90 (Abstract OR14-3) 15. Gasic S, Nagamani M, Green A, Urban RJ 2001 Troglitazone is a completive inhibitor of 3-␤-hydroxysteroid dehydrogenase enzyme in the ovary. Am J Obstet Gynecol 184:575–579 16. Mu Y-M, Yanase T, Nishi Y, Waseda N, Oda T, Tanaka A, Takayanagi R, Nawata H 2000 Insulin sensitizer, troglitazone directly inhibits aromatase activity in human ovarian granulosa cells. Biochem Biophys Res Commun 271:710 –713 17. Lovekamp-Swan T, Jetten AM, Davis BJ 2003 Dual activation of PPAR-␣ and

20.

21.

22. 23. 24.

25. 26.

27. 28. 29. 30. 31.

32.

33. 34. 35. 36. 37.

38.

39.

6105

PPAR-␥ by mono-(2-ethylhexyl) phthalate in rat ovarian granulosa cells. Mol Cell Endocrinol 28:133–141 Gasic S, Bodenburg Y, Nagamani M, Green A, Urban RJ 1998 Troglitazone inhibits progesterone production in porcine granulosa cells. Endocrinology 139:4962– 4966 Coffler MS, Patel K, Dahan MH, Yoo RY, Malcom PJ, Chang RJ 2003 Enhanced granulosa cell responsiveness to follicle-stimulating hormone during insulin infusion in women with polycystic ovary syndrome treated with pioglitazone. J Clin Endocrinol Metab 88:5624 –5631 Veldhuis JD, Zhang G, Garmey JC 2002 Troglitazone, an insulin-sensitizing thiazolidinedione, represses combined stimulation by LH and insulin of de novo androgen biosynthesis by thecal cells in vitro. J Clin Endocrinol Metab 87:1129 –1133 Seto-Young D, Leonardi O, Park A, Holcomb K, Salehi M, Chang P, Yih M, Rosenwaks Z, Poretsky L, Hormonally active non-transformed human ovarian cell culture from oophorectomy specimens: methods of development and initial characterization. Horm Res, in press Lowry OH, Rosebrough NJ, Farr AL, Randall RJ 1951 Protein measurement with the folin phenol reagent. J Biol Chem 193:265–275 Jousse C, Bruhat A, Ferrara M, Fafournoux P 1998 Physiological concentration of amino acids regulates insulin-like growth factor binding protein-1 expression. Biochem J 334:147–153 Thissen JP, Pucilowska JB, Underwood LE 1994 Differential regulation of insulin-like growth factor I (IGF-I) and IGF binding protein-1 messenger ribonucleic acids by amino acid availability and growth hormone in rat hepatocyte primary culture. Endocrinology 134:1570 –1576 Poretsky L, Grigorescu F, Moses AC, Flier JS 1985 Distribution and characterization of the insulin and IGF-I receptors in the human ovary. J Clin Endocrinol Metab 61:728 –734 El-Roeiy A, Chen X, Roberts V, Shimasaki S, Ling N, Le-Roith D, Roberts CT, Yen SC 1994 Expression of the genes encoding the insulin-like growth factors (IGF-I and -II), the IGF and insulin receptors, and IGF-binding proteins1– 6 and the localization of their gene products in normal and polycystic ovary syndrome ovaries. J Clin Endocrinol Metab 78:1488 –1496 Adashi EY 1994 Growth factors and ovarian function: the IGF-I paradigm. Horm Res 42:44 – 48 Aiman J, Nalick RH, Jacobs A, Porter JC, Edman CD, Vellios F, MacDonald PC 1977 The origin of androgen and estrogen in a virilized postmenopausal woman with bilateral benign cystic teratomas. Obstet Gynecol 49:695–704 Zhong WZ, Williams MG 1996 Simultaneous quantitation of pioglitazone and its metabolites in human serum by liquid chromatography and solid phase extraction. J Pharm Biomed Anal 14:465– 473 Roa NVSM, Biju B, Mullangi R, Nuggehally RS 2003 Simple method for the determination of rosiglitazone in human plasma using a commercially available internal standard. Biomed Chromatogr 17:417– 420 Goetze S, Kintscher U, Kim S, Meehan WP, Kaneshiro K, Collins AR, Fleck E, Hsueh WA, Law R 2001 Peroxisome proliferator-activated receptor-␥ ligands inhibit nuclear but not cytosolic extracellular signal-regulated kinase/ mitogen-activated protein kinase-regulated steps in vascular smooth muscle cell migration. J Cardiovasc Pharmacol 38:909 –921 Young PW, Buckle DR, Cantello BC, Chapman H, Clapham JC, Coyle PH, High D, Hindley RM, Holder JC, Kallender H, Latter AJ, Lawrie KW, Mossakowska D, Murphy GJ, Roxbee Cox L, Smith SA 1998 Identification of high-affinity binding sites for the insulin sensitizer rosiglitazone (BRL-49653) in rodent and human adipocytes using a radioiodinated ligand for peroxisomal proliferator-activated receptor-␥. J Pharmacol Exp Ther 284:751–759 Hammond JM, Mondschein JS, Samaras SE, Canning SF 1991 The ovarian insulin-like growth factors, a local amplification mechanism for steroidogenesis and hormone action. J Steroid Biochem Mol Biol 40:411– 416 Beilli SH, Graffigna NM, Oneto A, Otero P, Schurman L, Levell OA 2004 Effect of rosiglitazone on insulin resistance, growth factors, and reproductive disturbances in women with polycystic ovary syndrome. Fertil Steril 81:624 – 629 Plymate SR, Matej LA, Jones RE, Friedl KE 1988 Inhibition of sex hormonebinding globulin production in the human hepatoma (Hep G2) cell line by insulin and prolactin. J Clin Endocrinol Metab 7:460 – 464 Poretsky L, Seto-Young D, Shrestha A, Dhillon S, Mirjany N, Liu H-C, Yi MC, Rosenwaks Z 2001 Phosphatidyl-inositol-3 kinase-independent insulin action pathway(s) in the human ovary. J Clin Endocrinol Metab 86:3115–3119 Seto-Young D, Zajac J, Liu H-C, Rosenwaks Z, Poretsky L 2003 The role of mitogen activated protein kinase (MAPK) in insulin and IGF-I signaling cascades for progesterone and IGFBP-1 production in human granulosa cells. J Clin Endocrinol Metab 88:3385–3391 Poretsky L, Glover B, Laumas V, Kalin M, Dunaif A 1988 The effects of experimental hyperinsulinemia on steroid secretion, ovarian [125I]insulin binding and ovarian [125I]insulin-like growth factor I binding in the rat. Endocrinology 122:581–585 Nestler JE 1991 Sex hormone-binding globulin: a marker for hyperinsulinemia and/or insulin resistance. J Clin Endocrinol Metab 76:273–274

JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.