Thiazolidinediones but Not Metformin Directly Inhibit the Steroidogenic Enzymes P450c17 and 3β-Hydroxysteroid Dehydrogenase

Androgen biosynthesis requires 3β-hydroxysteroid dehydrogenase type II (3βHSDII) and the 17α-hydroxylase and 17,20-lyase activities of cytochrome P450c17. Thiazolidinedione and biguanide drugs, which are used to increase insulin sensitivity in type 2 diabetes, lower serum androgen concentrations in women with polycystic ovary syndrome. However, it is unclear whether this is secondary to increased insulin sensitivity or to direct effects on steroidogenesis. To investigate potential actions of these drugs on P450c17 and 3βHSDII, we used “humanized yeast” that express these steroidogenic enzymes in microsomal environments. The biguanide metformin had no effect on either enzyme, whereas the thiazolidinedione troglitazone inhibited 3βHSDII (KI = 25.4 ± 5.1 μm) and both activities of P450c17 (KI for 17α-hydroxylase, 8.4 ± 0.6 μm; KI for 17,20-lyase, 5.3 ± 0.7 μm). The action of troglitazone on P450c17 was competitive, but it was mainly a noncompetitive inhibitor of 3βHSDII. The thiazolidinediones rosiglitazone and pioglitazone exerted direct but weaker inhibitory effects on both P450c17 and 3βHSDII. These differential effects of the thiazolidinediones do not correlate with their effects on insulin sensitivity, suggesting that distinct regions of the thiazolidinedione molecule mediate these two actions. Thus, thiazolidinediones inhibit two key enzymes in human androgen synthesis contributing to their androgen-lowering effects, whereas metformin affects androgen synthesis indirectly, probably by lowering circulating insulin concentrations. Androgen biosynthesis requires 3β-hydroxysteroid dehydrogenase type II (3βHSDII) and the 17α-hydroxylase and 17,20-lyase activities of cytochrome P450c17. Thiazolidinedione and biguanide drugs, which are used to increase insulin sensitivity in type 2 diabetes, lower serum androgen concentrations in women with polycystic ovary syndrome. However, it is unclear whether this is secondary to increased insulin sensitivity or to direct effects on steroidogenesis. To investigate potential actions of these drugs on P450c17 and 3βHSDII, we used “humanized yeast” that express these steroidogenic enzymes in microsomal environments. The biguanide metformin had no effect on either enzyme, whereas the thiazolidinedione troglitazone inhibited 3βHSDII (KI = 25.4 ± 5.1 μm) and both activities of P450c17 (KI for 17α-hydroxylase, 8.4 ± 0.6 μm; KI for 17,20-lyase, 5.3 ± 0.7 μm). The action of troglitazone on P450c17 was competitive, but it was mainly a noncompetitive inhibitor of 3βHSDII. The thiazolidinediones rosiglitazone and pioglitazone exerted direct but weaker inhibitory effects on both P450c17 and 3βHSDII. These differential effects of the thiazolidinediones do not correlate with their effects on insulin sensitivity, suggesting that distinct regions of the thiazolidinedione molecule mediate these two actions. Thus, thiazolidinediones inhibit two key enzymes in human androgen synthesis contributing to their androgen-lowering effects, whereas metformin affects androgen synthesis indirectly, probably by lowering circulating insulin concentrations. 17α-hydroxypregnenolone dehydroepiandrosterone 3β-hydroxysteroid dehydrogenase type II 17α-hydroxyprogesterone polycystic ovary syndrome peroxisome proliferator-activated receptor γ oxidoreductase The first and rate-limiting step in the biosynthesis of all steroid hormones is the conversion of cholesterol to pregnenolone by the mitochondrial cholesterol side chain cleavage enzyme, P450scc, which is thus the quantitative regulator of steroidogenesis (1Miller W.L. Endocr. Rev. 1988; 9: 295-318Crossref PubMed Scopus (1187) Google Scholar). The qualitative regulator is microsomal P450c17, which sequentially catalyzes both 17α-hydroxylase and 17,20-lyase activities (2Nakajin S. Shively J.E. Yuan P.M. Hall P.F. Biochemistry. 1981; 20: 4037-4042Crossref PubMed Scopus (257) Google Scholar, 3Zuber M.X. Simpson E.R. Waterman M.R. Science. 1986; 234: 1258-1261Crossref PubMed Scopus (350) Google Scholar, 4Lin D. Harikrishna J.A. Moore C.C. Jones K.L. Miller W.L. J. Biol. Chem. 1991; 266: 15992-15998Abstract Full Text PDF PubMed Google Scholar). In the absence of either activity of P450c17, adrenal pregnenolone is directed toward the biosynthesis of mineralocorticoids. When only 17α-hydroxylase activity is present, the resulting 17α-hydroxypregnenolone (17-Preg)1 is converted to cortisol; when both the 17α-hydroxylase and 17,20-lyase activities are present, the C19 steroid dehydroepiandrosterone (DHEA) is produced. DHEA then can be converted to androstenedione by 3β-hydroxysteroid dehydrogenase type II (3βHSDII), and androstenedione is converted to testosterone and estradiol by isozymes of 17β-hydroxysteroid dehydrogenase and by aromatase (P450aro) (5Penning T.M. Endocr. Rev. 1997; 18: 281-305Crossref PubMed Scopus (391) Google Scholar,6Simpson E.R. Mahendroo M.S. Means G.D. Kilgore M.W. Hinshelwood M.M. Graham-Lorence S. Amarneh B. Ito Y. Fisher C.R. Michael M.D. Mendelson C.R. Bulun S.C. Endocr. Rev. 1994; 15: 342-355Crossref PubMed Scopus (1060) Google Scholar). The biosynthesis of all sex steroids proceeds through DHEA because human P450c17 does not convert 17α-hydroxyprogesterone (17OHP) to androstenedione (4Lin D. Harikrishna J.A. Moore C.C. Jones K.L. Miller W.L. J. Biol. Chem. 1991; 266: 15992-15998Abstract Full Text PDF PubMed Google Scholar, 7Lin D. Black S.M. Nagahama Y. Miller W.L. Endocrinology. 1993; 132: 2498-2506Crossref PubMed Google Scholar, 8Auchus R.J. Lee T.C. Miller W.L. J. Biol. Chem. 1998; 273: 3158-3165Abstract Full Text Full Text PDF PubMed Scopus (476) Google Scholar). Thus P450c17 and 3βHSDII are key enzymes required for the synthesis of all androgens (Fig.1 A). Polycystic ovary syndrome (PCOS) is the most frequent cause of female infertility, affecting ∼5–10% of women of reproductive age (9Ehrmann D.A. Barnes R.B. Rosenfield R.L. Endocr. Rev. 1995; 16: 322-353Crossref PubMed Scopus (507) Google Scholar, 10Dunaif A. Endocr. Rev. 1997; 18: 774-800Crossref PubMed Scopus (2269) Google Scholar). The insulin-sensitizing drugs metformin and troglitazone decrease hyperandrogenemia and increase fertility in women with PCOS (11Velazquez E.M. Mendoza S. Hamer T. Sosa F. Glueck C.J. Metabolism. 1994; 43: 647-654Abstract Full Text PDF PubMed Scopus (787) Google Scholar, 12Nestler J.E. Jakubowicz D.J. N. Engl. J. Med. 1996; 335: 617-623Crossref PubMed Scopus (688) Google Scholar, 13Dunaif A. Scott D. Finegood D. Quintana B. Whitcomb R. J. Clin. Endocrinol. Metab. 1996; 81: 3299-3306Crossref PubMed Scopus (549) Google Scholar, 14Ehrmann D.A. Schneider D.J. Sobel B.E. Cavaghan M.K. Imperial J. Rosenfield R.L. Polonsky K.S. J. Clin. Endocrinol. Metab. 1997; 82: 2108-2116Crossref PubMed Scopus (494) Google Scholar). These agents might decrease circulating androgens indirectly by lowering insulin levels, which may secondarily inhibit steroidogenesis by an unknown mechanism. Alternatively, metformin and troglitazone also might inhibit steroidogenic enzymes directly (12Nestler J.E. Jakubowicz D.J. N. Engl. J. Med. 1996; 335: 617-623Crossref PubMed Scopus (688) Google Scholar). Metformin and troglitazone are members of two fundamentally different drug families (Fig. 1 B). Metformin, a biguanide, decreases hepatic gluconeogenesis and enhances peripheral glucose uptake, either secondarily by alleviating glucose toxicity (15DeFronzo R.A. Goodman A.M. N. Engl. J. Med. 1995; 333: 541-549Crossref PubMed Scopus (1186) Google Scholar) or possibly by inhibiting complex 1 of the mitochondrial respiratory chain (16El-Mir M.Y. Nogueira V. Fontaine E. Averet N. Rigoulet M. Leverve X. J. Biol. Chem. 2000; 275: 223-228Abstract Full Text Full Text PDF PubMed Scopus (1074) Google Scholar, 17Owen M.R. Doran E. Halestrap A.P. Biochem. J. 2000; 348: 607-614Crossref PubMed Scopus (1635) Google Scholar). Troglitazone, a thiazolidinedione drug, decreases both hepatic gluconeogenesis and peripheral insulin resistance, probably through its action as a ligand for the nuclear peroxisome proliferator-activated receptor γ (PPARγ) (18Spiegelman B.M. Diabetes. 1998; 47: 507-514Crossref PubMed Scopus (1640) Google Scholar, 19Kersten S. Desvergne B. Wahli W. Nature. 2000; 405: 421-442Crossref PubMed Scopus (1665) Google Scholar, 20Olefsky J.M. J. Clin. Invest. 2000; 106: 467-472Crossref PubMed Scopus (508) Google Scholar). To determine whether metformin or troglitazone inhibits either P450c17 or 3βHSDII, and if so by what mechanism, we examined the actions of these two drugs using our humanized yeast expression system (8Auchus R.J. Lee T.C. Miller W.L. J. Biol. Chem. 1998; 273: 3158-3165Abstract Full Text Full Text PDF PubMed Scopus (476) Google Scholar, 21Lee T.C. Miller W.L. Auchus R.J. J. Clin. Endocrinol. Metab. 1999; 84: 2104-2110Crossref PubMed Scopus (69) Google Scholar), in which individual human steroidogenic enzymes are examined in native microsomal environments without interfering factors that are often present in whole-cell experiments. Reagent-grade metformin was purchased from Sigma and dissolved in water; reagent-grade troglitazone, a generous gift of Dr. Andrea Dunaif (Harvard University), was dissolved in 100% Me2SO. Tablets of troglitazone, rosiglitazone, and pioglitazone were purchased as Rezulin (Parke-Davis), Avandia (Lilly), and Actos (Smith Kline Beecham), respectively, and were dissolved in 100% Me2SO and centrifuged to remove insoluble tablet material. The chemical equivalence and purity of the troglitazone solutions prepared from reagent-grade powder and tablets were confirmed by liquid chromatography/mass spectrometry. Reagent-grade metformin and troglitazone were used in all experiments except for the comparison of troglitazone to rosiglitazone and pioglitazone in which troglitazone prepared from tablets was used to control for any variations in preparation. Saccharomyces cerevisiae strain W303B (22Pompon D. Louerat B. Bronine A. Urban P. Methods Enzymol. 1996; 272: 51-64Crossref PubMed Google Scholar) was transformed by the lithium acetate procedure (23Gietz D. St. Jean A. Woods R.A. Schiestl R.H. Nucleic Acids Res. 1992; 20: 1425Crossref PubMed Scopus (2895) Google Scholar) with the yeast expression vector V10 (22Pompon D. Louerat B. Bronine A. Urban P. Methods Enzymol. 1996; 272: 51-64Crossref PubMed Google Scholar) containing the cDNA sequences for human P450c17 or 3βHSDII as described (8Auchus R.J. Lee T.C. Miller W.L. J. Biol. Chem. 1998; 273: 3158-3165Abstract Full Text Full Text PDF PubMed Scopus (476) Google Scholar, 21Lee T.C. Miller W.L. Auchus R.J. J. Clin. Endocrinol. Metab. 1999; 84: 2104-2110Crossref PubMed Scopus (69) Google Scholar). For P450c17 transformation, we cotransformed the yeast cultures with the vector pYcDE2 (24Hadfield C. Cashmore A.M. Meacock P.A. Gene ( Amst. ). 1986; 45: 149-158Crossref PubMed Scopus (60) Google Scholar) expressing human P450 oxidoreductase cDNA (8Auchus R.J. Lee T.C. Miller W.L. J. Biol. Chem. 1998; 273: 3158-3165Abstract Full Text Full Text PDF PubMed Scopus (476) Google Scholar). Yeast was grown, microsomes were prepared, and the microsomal P450 content was measured as described (8Auchus R.J. Lee T.C. Miller W.L. J. Biol. Chem. 1998; 273: 3158-3165Abstract Full Text Full Text PDF PubMed Scopus (476) Google Scholar). Yeast microsomes were incubated in the presence and absence of various concentrations of thiazolidinediones or metformin added in 4 μl of Me2SO or water to 196 μl of 50 mm potassium phosphate buffer (pH 7.4). Enzymatic assays were performed with 0.5–4 μm progesterone or 17α-hydroxypregnenolone for P450c17 activities and 1–20 μm pregnenolone for 3βHSDII activity. Each reaction also contained 20,000 cpm of 14Cpregnenolone (55.4 mCi/mmol) (Amersham Pharmacia Biotech) for analysis of 3βHSDII activity, 20,000 cpm of 14Cprogesterone (55.4 mCi/mmol, PerkinElmer Life Sciences) for 17α-hydroxylase activity, or 50,000 cpm of 3H17-Preg (21.1 Ci/mmol, PerkinElmer Life Sciences) for 17,20-lyase activity. Catalysis was initiated by adding 1 mm NADPH for P450c17 or 1 mm NAD+ for 3βHSDII activity, and assays were done in the linear time range of the enzymatic reaction. Assays of 17,20-lyase activity were performed with and without addition of purified recombinant human cytochrome b5 (PanVera, Madison, WI) in 10-fold molar excess to the total P450 content of the microsomes. Steroids were extracted from the reaction mixtures with 400 μl of ethyl acetate/isooctane (1:1) concentrated by evaporation under continuous nitrogen flow and assayed by thin layer chromatography on phosPE SIL G/UV silica gel plates (Whatman) using 3:1 chloroform/ethyl acetate as the solvent system (8Auchus R.J. Lee T.C. Miller W.L. J. Biol. Chem. 1998; 273: 3158-3165Abstract Full Text Full Text PDF PubMed Scopus (476) Google Scholar, 25Geller D.H. Auchus R.J. Miller W.L. Mol. Endocrinol. 1999; 13: 167-175Crossref PubMed Scopus (141) Google Scholar). The radiolabeled steroids were quantified by phosphorimaging analysis on a Storm 860 PhosphorImager (Molecular Dynamics, Sunnyvale, CA). All assays were performed in triplicate, and data are presented as means ± S.D. Kinetic behavior was approximated as a Michaelis-Menten system, and data were plotted as described by Lineweaver and Burk and by Dixon and Webb (26Dixon M. Webb E.C. Enzymes. 3rd Ed. Academic Press, New York1979Google Scholar).KI values were calculated from the equation for competitive inhibition, Vapp =Vmax × S/(Km (1 + I/KI) + S), whereas data from noncompetitive-competitive inhibition were approximated by the formula for generic inhibition, Vapp =Vmax/(1 + I/KI) at each S (27Cornish-Bowden A. Analysis of Enzyme Kinetic Data. Oxford University Press, Oxford1995Google Scholar). Data fitting was carried out by LEONORA version 1.0 for analysis of steady-state enzyme kinetics (27Cornish-Bowden A. Analysis of Enzyme Kinetic Data. Oxford University Press, Oxford1995Google Scholar). COS-1 monkey kidney cells were grown as monolayers in 10-cm Petri dishes in 10 ml of Dulbecco's modified Eagle's medium-H21 containing 4.5 g/liter glucose, 10% fetal bovine serum, 100 IU/ml penicillin, and 0.1 mg/ml streptomycin. For transient transfection, COS-1 cells were grown to 80% confluence and split into six-well plates 24 h prior to transfection. Plasmid pcDNA3 containing the wild-type human P450c17 cDNA (28Chung B.C. Picado-Leonard J. Haniu M. Bienkowski M. Hall P.F. Shively J.E. Miller W.L. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 407-411Crossref PubMed Scopus (413) Google Scholar) was transfected into COS-1 cells at ∼60% confluence using the calcium phosphate/DNA coprecipitation method. Thirty-six hours after transfection, cells were incubated with 2 ml of fresh Dulbecco's modified Eagle's medium-H21 containing 20,000 cpm 14Cprogesterone and 0–100 μmtroglitazone; each troglitazone concentration was added in 10 μl of Me2SO to triplicate wells. The steroids then were extracted with 8 ml of ethyl acetate/isooctane (1:1), concentrated, and analyzed as described above for the microsomal assays. Transfection efficiency was monitored by cotransfecting COS-1 cells with the pRL-CMV plasmid (Promega, Madison, WI) containing the Renilla luciferase gene driven by the cytomegalovirus promoter. After removing the medium, the cells were lysed and assessed for luciferase activity by the dual luciferase reporter assay system (Promega). Both 17α-hydroxylase and 17,20-lyase activities require the interaction of P450c17 with its electron donor P450 oxidoreductase (OR); 17,20-lyase activity is enhanced further by allosteric interaction of the complex with cytochrome b5 (8Auchus R.J. Lee T.C. Miller W.L. J. Biol. Chem. 1998; 273: 3158-3165Abstract Full Text Full Text PDF PubMed Scopus (476) Google Scholar). from yeast expressing human P450c17 cDNA and the conversion of pregnenolone to and of progesterone to but the 17,20-lyase reaction is 100 with as the with D. Black S.M. Nagahama Y. Miller W.L. Endocrinology. 1993; 132: 2498-2506Crossref PubMed Google Scholar, 8Auchus R.J. Lee T.C. Miller W.L. J. Biol. Chem. 1998; 273: 3158-3165Abstract Full Text Full Text PDF PubMed Scopus (476) Google Scholar). Thus, 17α-hydroxylase activity was measured as the conversion of progesterone to because this is not and 17,20-lyase activity was measured as the conversion of to DHEA in the presence of added (8Auchus R.J. Lee T.C. Miller W.L. J. Biol. Chem. 1998; 273: 3158-3165Abstract Full Text Full Text PDF PubMed Scopus (476) Google Scholar, 21Lee T.C. Miller W.L. Auchus R.J. J. Clin. Endocrinol. Metab. 1999; 84: 2104-2110Crossref PubMed Scopus (69) Google Scholar). of metformin to 100 μm not inhibit either activity of P450c17, but troglitazone inhibited both activities of P450c17 (Fig. a concentration of 1 μm troglitazone inhibited 17α-hydroxylase activity with an of and at 1 troglitazone inhibited 17,20-lyase activity with an of These 1 concentrations are above the of μm for both the 17α-hydroxylase reaction and the 17,20-lyase reaction in this system (8Auchus R.J. Lee T.C. Miller W.L. J. Biol. Chem. 1998; 273: 3158-3165Abstract Full Text Full Text PDF PubMed Scopus (476) Google Scholar). The presence or absence of cytochrome not the of the reaction not Thus, the inhibitory effect of troglitazone on 17,20-lyase activity does not on interaction with Kinetic analysis using both the plotted and the Dixon plotted inhibitor a competitive of inhibition values of 8.4 ± 0.6 μm for 17α-hydroxylase activity and of 5.3 ± 0.7 μm for 17,20-lyase activity (Fig. Thus, troglitazone inhibits both activities of P450c17 by the analysis of the of inhibition of P450c17 activities by troglitazone of 17α-hydroxylase activity and 17,20-lyase activity and Dixon of 17α-hydroxylase activity in the absence and presence of troglitazone μm) and with and 4 and for the of 17α-hydroxylase and 17,20-lyase activity, were carried out with yeast microsomes of total human P450c17 and human P450 Each data the ± S.D. of triplicate The inhibitory effect of troglitazone was both in the yeast system and in cells expressing P450c17. inhibited 17α-hydroxylase activity in transfected COS-1 cells in a with an of μm which is to the of μm in yeast microsomes. The two of human are J. Biochem. PubMed Scopus Google Scholar, E. Y. N. M. Y. C. V. J. F. Mol. Endocrinol. 1991; PubMed Scopus Google Scholar). 3βHSDII is the only in human and E. Y. N. M. Y. C. V. J. F. Mol. Endocrinol. 1991; PubMed Scopus Google Scholar), we examined the potential inhibitory effects of metformin and troglitazone on this assayed activity in microsomes from yeast transformed with an expression vector for human 3βHSDII. all for and the values and when analyzed in this yeast expression system T.C. Miller W.L. Auchus R.J. J. Clin. Endocrinol. Metab. 1999; 84: 2104-2110Crossref PubMed Scopus (69) Google Scholar), we measured the conversion of pregnenolone to which the of conversion of DHEA to Metformin had no effect on 3βHSDII activity, but troglitazone inhibited 3βHSDII activity in a with an of 24 μm To determine the by which troglitazone inhibited 3βHSDII activity we performed and Dixon In to the competitive inhibition of P450c17, that troglitazone inhibited 3βHSDII activity by a noncompetitive with of 25.4 ± 5.1 μm A). noncompetitive enzymatic a linear in Dixon analysis (26Dixon M. Webb E.C. Enzymes. 3rd Ed. Academic Press, New York1979Google Scholar). However, of a linear Dixon of 3βHSDII inhibition by troglitazone a (Fig. B). that are of inhibition at inhibitor concentrations noncompetitive and at inhibitor concentrations competitive behavior might be by the of two of inhibitor to molecule of enzyme (26Dixon M. Webb E.C. Enzymes. 3rd Ed. Academic Press, New York1979Google Scholar). Thus, the by which troglitazone inhibits 3βHSDII activity is different from the by which it inhibits analysis of the of inhibition of 3βHSDII activity by troglitazone and Dixon of 3βHSDII activity. were carried out with yeast microsomes total expressing human 3βHSDII in the absence and presence of troglitazone μm) with and Each data the ± S.D. of triplicate To determine whether the inhibitory effects of troglitazone are to this drug or are a of thiazolidinedione we the effects of troglitazone on P450c17 and 3βHSDII activities to of two rosiglitazone and pioglitazone inhibited both the 17α-hydroxylase = μm) and the 17,20-lyase = μm) activities of P450c17 and also inhibited 3βHSDII activity = μm) but to a troglitazone = and had inhibitory effects on all enzymatic activities with values of of all the inhibition of the thiazolidinediones on the enzymatic activities is in Thus, the potential for direct inhibition of both P450c17 and 3βHSDII to be a of thiazolidinedione drugs, but the variations in their inhibitory effects that in the thiazolidinedione effects on their actions as of steroidogenic of 1 for enzyme in COS-1 cells = ± 0.7 ± 0.6 ± 5.1 of 1 1 μm for enzyme in COS-1 cells = in a drugs increase insulin sensitivity and decrease hyperandrogenemia in women with PCOS (11Velazquez E.M. Mendoza S. Hamer T. Sosa F. Glueck C.J. Metabolism. 1994; 43: 647-654Abstract Full Text PDF PubMed Scopus (787) Google Scholar, 12Nestler J.E. Jakubowicz D.J. N. Engl. J. Med. 1996; 335: 617-623Crossref PubMed Scopus (688) Google Scholar, 13Dunaif A. Scott D. Finegood D. Quintana B. Whitcomb R. J. Clin. Endocrinol. Metab. 1996; 81: 3299-3306Crossref PubMed Scopus (549) Google Scholar, 14Ehrmann D.A. Schneider D.J. Sobel B.E. Cavaghan M.K. Imperial J. Rosenfield R.L. Polonsky K.S. J. Clin. Endocrinol. Metab. 1997; 82: 2108-2116Crossref PubMed Scopus (494) Google Scholar), and it that metformin decreases P450c17 activity in PCOS (12Nestler J.E. Jakubowicz D.J. N. Engl. J. Med. 1996; 335: 617-623Crossref PubMed Scopus (688) Google Scholar, J.E. Jakubowicz D.J. J. Clin. Endocrinol. Metab. 1997; 82: PubMed Google A. T. A. V. A. 2000; 15: PubMed Scopus Google Scholar), but this on serum values on direct In our microsomal system we no for a direct inhibition of P450c17 or 3βHSDII by in concentrations the concentrations of μm J. M. Goodman A.M. J. Clin. 1996; PubMed Scopus Google Scholar). Thus the inhibitory effect of metformin on androgen synthesis to be only indirectly, most by lowering circulating insulin concentrations. lowering insulin concentrations without drug by and decreases circulating androgen concentrations in women with PCOS J.E. Jakubowicz D.J. J. Clin. Endocrinol. Metab. 1997; 82: PubMed Google Scholar, D. A. F. V. S. Clin. Endocrinol. 1992; PubMed Scopus Google Scholar, S. N. M. J. Clin. Endocrinol. Metab. 1995; PubMed Google Scholar). However, we that troglitazone, in to a direct inhibitory effect on the enzymatic activities of P450c17 and 3βHSDII, thus directly interfering with androgen effect probably to the androgen-lowering effect of troglitazone in women with PCOS because enzymatic inhibition at concentrations. concentrations of troglitazone in serum are μm J. Clin. 1999; PubMed Scopus Google Scholar). the concentrations are μm troglitazone inhibit of P450c17 activity and of 3βHSDII activity. troglitazone concentrations not we values in both the assay and in whole-cell suggesting that troglitazone concentrations may be to circulating concentrations. that troglitazone exerted competitive inhibition of P450c17 and a mainly noncompetitive inhibition of 3βHSDII. Thus, the interaction of troglitazone with each enzyme is and pioglitazone also inhibited P450c17 and 3βHSDII with values their serum concentrations of μm Y. Y. Y. M. S. 1997; 47: Google D.A. A.M. Miller M. Metab. 2000; Google Scholar). Thus, the direct inhibition of both P450c17 and 3βHSDII to be a of thiazolidinedione drugs, only troglitazone does so at concentrations in the potential actions of rosiglitazone and pioglitazone in PCOS not it is that all thiazolidinediones can decrease androgen by circulating insulin concentrations. The differential inhibitory effects of thiazolidinedione drugs on the key androgen enzymes P450c17 and 3βHSDII do not correlate with their differential to the to and is only a whereas rosiglitazone and pioglitazone to with and insulin-sensitizing effects J.M. Moore T.M. J. Biol. Chem. 1995; Full Text Full Text PDF PubMed Scopus Google Scholar, J. P. C. R. M.D. Endocrinology. 1996; PubMed Scopus Google Scholar, T.M. J.E. D.J. Moore J.M. J. Med. Chem. 1996; PubMed Scopus Google Scholar, S.C. X. R. T. Diabetes. 2000; PubMed Scopus Google Scholar). Thus, distinct and different regions of the thiazolidinedione drugs might be in and inhibition of P450c17 and 3βHSDII, of in thiazolidinedione drugs may drug at inhibitory effects on steroidogenesis as as at insulin-sensitizing W. M. and V. for and Dr. for with the liquid chromatography/mass