The high prevalence of prostatic carcinoma (PRCA) and the limited therapeutic possibilities provide a strong stimulus for exploring new approaches in experimental research that ultimately may lead to improved therapy. Indeed, methods for assessing carcinoma prognosis, such as clinical staging (clinical examination, ultrasound, and plasmatic levels of prostatic acid phosphatase and prostate specific antigen) and histopathological grading according to the Gleason score, usually fail to provide consistent predictive information regarding the clinical outcome of single tumors. Increased plasminogen activator (PA) activities have been associated with high-grade malignancies and with the potential for invasion/metastasis in many tumors. Urokinase-type plasminogen activator (uPA) is present in prostatic secretion, and an increased uPA activity has been noted in human prostatic cell lines with metastatic behavior. Unfortunately, any study of uPA production or gene regulation in primary tumors is complicated by the inherent mixture of host stromal cells, infiltrating macrophages, and subpopulations of tumor cells that may have variable metastatic capacity and ability to synthesize uPA. In short-term tissue culture of prostatic samples, it is possible to grow in vitro cancer prostatic epithelial cells and thus exclude the presence of contaminant cells. We have shown elsewhere that the levels of a type IV collagenase, 92-kDa matrix metalloproteinase, a protease involved in tumor progression and invasion, are increased in PRCA primary cell cultures if compared with benign prostatic hyperplasia (BPH) cell cultures (C. Festuccia et al., manuscript in preparation). Activation of matrix metalloproteinases also can be correlated with uPA expression; therefore we studied the expression of uPA in serum-free culture media of primary cultures of PRCA or BPH tissue samples.(ABSTRACT TRUNCATED AT 250 WORDS)