The role of prolactin (PRL) in supporting the growth of human breast cancer is still unclear. The ability to grow primary breast cancer specimens in the soft agar clonogenic assay in the absence of serum gave us the opportunity to evaluate the growth-promoting effect of PRL and to compare it to that of estradiol in the same tumor samples. PRL was tested both at physiological concentrations (20 ng/ml) as well as in pharmacological amounts (200 ng/ml) comparable to circulating blood levels in hyperprolactinemic states. Estradiol was simultaneously tested in physiological amounts (10(-8)M). In 17 infiltrating ductal carcinomas, the lower dose of PRL stimulated colony formation to 126 +/- 5.2% (SE) of control, while the higher dose increased colony number to 159 +/- 10.4% of control. This latter effect was comparable to that observed with estradiol (159 +/- 8.5% of control). The effect of PRL was more pronounced in estrogen receptor-positive tumors. Nine of ten estrogen receptor-positive tumors were PRL sensitive, while three of seven estrogen receptor-negative tumors exhibited a clear response to PRL administration. PRL did not stimulate colony formation in a malignant cystosarcoma phylloides and in two benign lesions (fibroadenoma and fibrocystic disease). We conclude that, at least under the conditions of the soft agar clonogenic assay, PRL exerts a dose-dependent growth-promoting effect on human breast cancer. Such effect is comparable to that of estradiol when PRL is added in concentrations similar to circulating blood levels in hyperprolactinemic patients.