Estrogens and androgens are well known to exert opposing effects in several tissues. In this study, we explored the possibility that there might be a direct antiandrogenic effect of estradiol on human breast cancer cells (MCF-7). Since the biological activity of androgens is mediated by specific androgen receptors, and because the abundance of androgen receptors in target tissues is thought to be rate limiting for androgen action, we examined whether estradiol regulates the quantity of androgen receptors in MCF-7 cells. Cells treated with 2.6 nM estradiol exhibited markedly lower levels of cytoplasmic androgen receptors, measured by [3H]5 alpha-dihydrotestosterone ([3H]DHT) binding, compared to levels in cells receiving ethanol vehicle alone. The effect was time dependent, and 6-day treatment of cells with estradiol resulted in an 80% reduction in [3H]DHT binding. Occupancy of androgen receptors by estradiol did not account for this difference. Cytosol competition studies demonstrated that the androgen receptor in MCF-7 cells possesses an approximately 125-fold lower affinity for estradiol than for DHT. In addition, tamoxifen, a nonsteroidal estrogen antagonist, blocked the estradiol effects on [3H]DHT binding. These latter studies support the hypothesis that this estradiol action is mediated by the estrogen receptor. Equilibrium binding studies indicated that the observed decrease in [3H]DHT binding after estradiol treatment was due to an absolute decrease in the number of cytoplasmic androgen receptors per cell. The estradiol-mediated reduction in androgen receptor content was dose dependent; a 50% reduction in androgen receptor number was observed after 6 days of treatment with 2.6 X 10(-11) M estradiol. Additional experiments revealed that MCF-7 cells exhibited a time-dependent increase in androgen receptor content when estradiol was withdrawn; continued estradiol treatment prevented this rise in receptor content. Moreover, androgen receptor levels began to decrease from the point when the ethanol vehicle added to the medium was replaced with 2.6 nM estradiol. In summary, estradiol treatment caused a reduction in androgen receptors, and estradiol withdrawal lead to a rise in androgen receptors. We believe that these results provide a mechanism whereby estradiol may directly antagonize androgen action. Conversely, estradiol withdrawal may potentiate androgen action by allowing androgen receptor levels to rise. This hypothesis may help explain the basis of the estrogen/androgen ratio as a predictor of sex steroid response.