Search

Metformin is commonly used to treat women with polycystic ovary syndrome, but its precise mechanism of action is unclear, and it even appears to have direct ovarian effects. At the cellular level, it may act either via an insulin-dependent pathway or an independent pathway by activating AMP-activated protein kinase (AMPK). In the ovary, metformin directly decreased estradiol and progesterone production by human granulosa cells, and inhibition of progesterone production by metformin in rat granulosa cells caused an increase in phosphorylated AMPK (pAMPK). We investigated whether metformin activates AMPK in the human ovary by looking for changes in phosphorylation of AMPK and its downstream target acetyl CoA carboxylase (ACC). mRNA and protein for α1 and α2 AMPK subunits were present in all human ovarian tissue. Neither 100 nm nor 2 mm of metformin affected subunit expression. After 1 or 4 h, neither dose of metformin increased pAMPK or pACC, although after 1 h, the addition of insulin significantly enhanced pAMPK, whereas insulin alone had no effect on phosphorylation of either AMPK or ACC. The addition of compound C, an inhibitor of AMPK, negated the effect of metformin in the presence of insulin on pAMPK. This effect on AMPK was not due to a change in the ADP/ATP ratio measured by HPLC. In summary, the presence of insulin was required to cause a metformin-induced increase in pAMPK in these human ovarian cells. Although previous data suggest that metformin may act via an insulin-independent pathway, our results therefore imply that insulin may be required to initiate an effect.

Metformin treatment, now widely prescribed in polycystic ovary syndrome, is aimed at correcting the associated insulin resistance, but it has also been shown to directly inhibit ovarian steroidogenesis. The mechanisms, however, by which metformin inhibits estradiol production in human granulosa cells remains unknown. Granulosa luteal cells were incubated with metformin, insulin, or combined metformin and insulin treatment, and aromatase mRNA expression was quantified using real-time RT-PCR. Enzyme activity was assessed by the conversion of (3)H-androstenedione to estrone and estradiol. Metformin's effect on the expression of specific untranslated first exon aromatase promoters was analyzed using semiquantitative PCR. The involvement of MAPK kinase (MEK)/ERK pathway was investigated by immunoblotting for aromatase, phosphorylated, and total ERK-1,2 from cells cultured as above with/without the MEK inhibitor PD98059. Metformin significantly inhibited basal and insulin-stimulated aromatase mRNA expression, with parallel results from the aromatase activity assay and protein assessment. This suppression was via down-regulation of aromatase promoter II, I.3, and 1.4 expression and was reversed by the addition of PD98059. Involvement of the ERK signaling pathway was demonstrated by the significant increase in phosphorylated ERK-1,2 with the combined metformin and insulin treatment. We have shown for the first time in human granulosa cells that metformin signficantly attenuated basal and insulin-stimulated P450 aromatase mRNA expression and activity, via silencing of key promoters. This occurred by activation of MEK/ERK pathway, which negatively regulated aromatase production. This is an important consideration given metformin's widespread use in polycystic ovary syndrome and may further support a possible therapeutic indication in estrogen-dependent breast tumors.