Metformin is treatment of choice for the metabolic consequences seen in polycystic ovary syndrome for its insulin-sensitizing and androgen-lowering properties. Yet, the mechanism of action remains unclear. Two potential targets for metformin regulating steroid and glucose metabolism are AMP-activated protein kinase (AMPK) signaling and the complex I of the mitochondrial respiratory chain. Androgen biosynthesis requires steroid enzymes 17α-Hydroxylase/17,20 lyase (CYP17A1) and 3β-hydroxysteroid dehydrogenase type 2 (HSD3B2), which are overexpressed in ovarian cells of polycystic ovary syndrome women. Therefore, we aimed to understand how metformin modulates androgen production using NCI-H295R cells as an established model of steroidogenesis. Similar to in vivo situation, metformin inhibited androgen production in NCI cells by decreasing HSD3B2 expression and CYP17A1 and HSD3B2 activities. The effect of metformin on androgen production was dose dependent and subject to the presence of organic cation transporters, establishing an important role of organic cation transporters for metformin's action. Metformin did not affect AMPK, ERK1/2, or atypical protein kinase C signaling. By contrast, metformin inhibited complex I of the respiratory chain in mitochondria. Similar to metformin, direct inhibition of complex I by rotenone also inhibited HSD3B2 activity. In conclusion, metformin inhibits androgen production by mechanisms targeting HSD3B2 and CYP17-lyase. This regulation involves inhibition of mitochondrial complex I but appears to be independent of AMPK signaling.

CYP17A1 plays a pivotal role in the biosynthesis of androgens in the adrenals and the gonads. Although this enzyme catalyzes two different reactions on one single active site, its specific activities are regulated independently. Although the 17alpha-hydroxylase activity is rather constant and regulated by gene expression, the 17,20-lyase activity varies significantly with the amount of cofactors or by protein phosphorylation. cAMP increases CYP17A1 expression, P450c17 phosphorylation, and androgen production. However, the exact mechanism(s) and the specific regulators of CYP17A1 remain unknown. Therefore, we studied the regulation of adrenal androgen biosynthesis in human adrenal H295R cells focusing on CYP17A1. We analyzed androgen production and P450c17 activities in H295R cells grown under normal and serum-free conditions and/or after stimulation with 8-bromoadenosine-cAMP. H295R cells grown in starvation medium produced more androgens and had decreased HSD3B2 expression and activity but increased P450c17-17,20-lyase activity and serine phosphorylation. Although starvation increased serine phosphorylation of P450c17 specifically, cAMP stimulation enhanced threonine phosphorylation exclusively. Time-course experiments revealed that a short cAMP stimulation augmented threonine phosphorylation of P450c17 but did not increase 17,20-lyase activity. By contrast, long cAMP stimulation increased androgen production through increased P450c17 activities by enhancing CYP17A1 gene expression. We conclude that serum withdrawal shifts steroidogenesis of H295R cells towards androgen production, providing a suitable model for detailed studies of androgen regulation. In addition, our study shows that starvation and cAMP stimulation regulate P450c17 phosphorylation differentially and that an increase in P450c17 phosphorylation does not necessarily lead to enhanced enzyme activity and androgen production.

The CYP17A1 gene is the qualitative regulator of steroidogenesis. Depending on the presence or absence of CYP17 activities mineralocorticoids, glucocorticoids or adrenal androgens are produced. The expression of the CYP17A1 gene is tissue as well as species-specific. In contrast to humans, adrenals of rodents do not express the CYP17A1 gene and have therefore no P450c17 enzyme for cortisol production, but produce corticosterone. DNA methylation is involved in the tissue-specific silencing of the CYP17A1 gene in human placental JEG-3 cells. We investigated the role of DNA methylation for the tissuespecific expression of the CYP17A1 gene in rodents. Rats treated with the methyltransferase inhibitor 5-aza-deoxycytidine excreted the cortisol metabolite tetrahydrocortisol in their urine suggesting that treatment induced CYP17 expression and 17a-hydroxylase activity through demethylation. Accordingly, bisulfite modification experiments identified a methylated CpG island in the CYP17 promoter in DNA extracted from rat adrenals but not from testes. Both methyltransferase and histone deacetylase inhibitors induced the expression of the CYP17A1 gene in mouse adrenocortical Y1 cells which normally do not express CYP17, indicating that the expression of the mouse CYP17A1 gene is epigenetically controlled. The role of DNA methylation for CYP17 expression was further underlined by the finding that a reporter construct driven by the mouse K1041 bp CYP17 promoter was active in Y1 cells, thus excluding the lack of essential transcription factors for CYP17 expression in these adrenal cells.