Ovarian cancer is the most lethal gynecological cancer affecting women. Hormone-based therapies are variably successful in treating ovarian cancer, but the reasoning behind these therapies is paradoxical. Clinical reagents such as tamoxifen are considered to inhibit or reverse tumor growth by competitive inhibition of the estrogen receptor (ER); however, high-dose estrogen is as clinically effective as tamoxifen, and it is unlikely that estrogen is acting by blocking ER activity; however, it may be activating a unique function of the ER that is nonmitogenic. For poorly defined reasons, 90% of ovarian cancers derive from the ovarian surface epithelium (OSE). In vivo the ER-positive OSE is exposed to high estrogen levels, reaching micromolar concentrations in dominant ovarian follicles. Using cultured rhesus OSE cells in vitro, we show that these levels of estradiol (1 mug/ml; approximately 3 mum) block the actions of serum growth factors, activate the G(1) phase retinoblastoma checkpoint, and induce p21, an inhibitor of kinases that normally inactivate the retinoblastoma checkpoint. We also show that estradiol increases p53 levels, which may contribute to p21 induction. Supporting the hypothesis that clinical selective ER modulators activate this novel ER function, we find that micromolar doses of tamoxifen and the "pure antiestrogen" ICI 182,780 elicit the same effects as estradiol. We propose that, in the context of proliferation, these data clarify some paradoxical aspects of hormone-based therapy and suggest that a fuller understanding of normal ER function is necessary to improve therapeutic strategies that target the ER.