Chronic treatment with ethinyl estradiol, levonorgestrel or cetrorelix decreases brain neurosteroid levels and selectively alters GABAA receptor plasticity

Oral contraceptives are among the most widely prescribed drugs, especially in young healthy women. 10 to 40% of oral contraceptive users develop mild to moderate depressive symptoms, anxiety or other mood disorders(Epperson et al., 1999). We have shown that long-term administration of ethinyl estradiol and levonorgestrel, two of the most widely used steroids in the oral contraceptive pills, decreased neurosteroid concentrations in rat cerebral cortex and plasma. Oral contraceptive administration for 3 months also reduced the serum concentrations of neurosteroids in women. The decrease in neurosteroid levels was accompanied by altered -aminobutyric acid type A (GABAA) receptor plasticity, with increased expression of the 2 subunit, and by anxiety-like behavior in the elevated plus-maze test(Follesa et al., 2002). We hypothesized that these neurochemical changes may contribute to the development of mood disorders observed in some oral contraceptive users. To better clarify the role of the estrogen and of the progestin component of the oral contraceptive pill on neurosteroid levels and GABAA receptor plasticity, we evaluated the effects of long-term treatment with ethinyl estradiol alone or levonorgestrel alone on brain concentrations of (3,5)-3-hydroxypregnan-20-one (allopregnanolone) and its precursor progesterone. We further evaluated their effects on GABAA receptor plasticity. Female rats were injected subcutaneously with ethynil estradiol, levonorgestrel or a mixture of these two steroids once a day for six weeks. Administration of ethinyl estradiol decreased cerebral cortical levels of allopregnanolone and progesterone by 76 and 72%, respectively. Likewise, administration of levonorgestrel decreased cerebral cortical levels of allopregnanolone and progesterone by 75 and 68%, respectively. As expected, administration of the ethinyl estradiol-levonorgestrel combination decreased cerebral cortical levels of neurosteroids. These effects were associated to changes in GABAA receptor plasticity. As expected, administration of the steroid combination increased the 2 subunit peptide in the cerebral cortex by 30%. Administration of levonorgestrel also increased the 2 subunit peptide in the cerebral cortex by 38%. In contrast, administration of ethinyl estradiol failed to alter the 2 subunit peptide in the region examined. We also examined the  subunit peptide in the cerebral cortex but none of these treatments altered its expression. These results suggest that changes in allopregnanolone levels are not always associated with GABAA receptor plasticity. To further evaluate the role of allopregnanolone on GABAA receptor plasticity we used other animal models associated with a decrease in neurosteroid concentrations. For instance, inhibition of hypothalamic-pituitary-gonadal axis with Cetrorelix (CET)(Halmos et al., 1996), a gonadotropin-releasing hormone antagonist, (0.100 mg s.c. once a day for 12 days) decreased cerebral cortical levels of allopregnanolone and progesterone by 44 and 49%, respectively. However, administration of CET did not change the expression of 2 and  subunits peptides in the cerebral cortex. Finally, we also examined the effect of ovariectomy in adult female rats. Ovariectomy decreased cerebral cortical levels of allopregnanolone and progesterone by 42 and 59%, respectively while the levels of the GABAA receptor subunits were not altered. These results suggest that changes in GABAA receptor plasticity observed following administration of oral contraceptives are mediated by the progestinic component of the oral contraceptive pill. Because both ethynil estradiol and levonorgestrel decrease allopregnanolone levels in a similar manner, these results suggest that changes in allopregnanolone levels are not associated with GABAA receptor plasticity.