Progesterone is required for the full proliferative activity of the breasts and may be directly or indirectly involved in either stimulating or inhibiting breast cancer. To determine whether the effects on breast cancer are attributable to progesterone metabolites, we compared the capacity of nontumorous and tumorous breast tissue to convert progesterone and then tested the effects of these metabolites on breast cell proliferation and anchorage. Tissues from the operated breasts of six patients with infiltrating duct carcinomas were incubated with [14C]progesterone for 2, 4, and 8 h, and the metabolites were identified and quantified. The identified metabolites (equal to >95% of recovered radioactivity) can be divided into those that retain the double bond of progesterone in the carbon-4 position of ring A (4-pregnenes) and those that are 5alpha-reduced (5alpha-pregnanes). The results show that tumorous breast tissue has elevated 5alpha-reductase activity, which results in significantly higher total levels of 5alpha-pregnanes, especially 5alpha-pregnane-3,20-dione (5alphaP), whereas normal (nontumorous) breast tissue produces more 4-pregnenes, especially 3alpha-hydroxy-4-pregnen-20-one (3alphaHP). 5alphaP and 3alphaHP are each one enzymatic step removed from progesterone, resulting from the action of either 5alpha-reductase or 3alpha-hydroxysteroid oxidoreductase (3alpha-HSO), respectively. The ratio of 5alpha-pregnanes:4-pregnenes is >5-fold greater and the ratio of 5alphaP:3alphaHP is nearly 30-fold greater in tumorous than nontumorous breast tissue incubates. In vitro studies with three breast cell lines (MCF-7, MCF-10A, and ZR-75-1) show that 3alphaHP dose dependently inhibits, whereas 5alphaP significantly stimulates, proliferation. Additional studies with MCF-7 and MCF-10A cells indicate that each of the 4-pregnenes isolated from breast tissue suppresses, whereas each respective 5alpha-reduced product stimulates, cell proliferation. Studies of cell anchorage were conducted using MCF-7 cells and various concentrations of 5alphaP or 3alphaHP. The number of cells attached to the substrate was significantly (P<0.05) decreased by treatment with > or =30 nM 5alphaP and increased by treatment with > or =50 nM 3alphaHP. Conversely, the number of cells detached from the substrate after partial trypsin exposure was significantly increased by treatment with > or =40 nM 5alphaP and decreased by treatment with > or =30 nM 3alphaHP. The results suggest that a change in in situ progesterone metabolism, resulting in an increased 5alpha-pregnane:4-pregnene (especially 5alphaP:3alphaHP) ratio, may promote breast cancer by promoting increased cell proliferation and detachment, whereas increases in 4-pregnenes may retard these tumorigenic processes. These studies suggest that endogenous progesterone metabolites may provide a new hormonal basis for breast cancer.