Fresh brain-tumor samples were obtained at operation and analyzed for their content of tissue type plasminogen activator (tPA) using an activity assay (gel chromatography zymogram) and an enzyme-linked immunospecific assay. The specimens were taken from 23 glioblastomas, 35 metastatic tumors, 42 meningiomas, 16 low-grade gliomas, and seven acoustic neurinomas; seven specimens were from normal brain. A strong correlation was found between the results of the two assays (r = 0.77, p less than 0.0001). There was a threefold difference in the tPA content of the benign tumors as compared to malignant tumors (p = 0.0006), the latter having less tPA. Histologically benign meningiomas contained higher tPA than malignant meningiomas (p = 0.01); however, the difference between low-grade gliomas and high-grade gliomas was less evident. Overall regression analysis data have shown an inverse relationship between the tissue content in tPA and the presence and degree of tumor necrosis and peritumoral brain edema (p = 0.004 and p = 0.0004, respectively). This finding was most consistent in the glioblastoma group where the correlation coefficient values were r = 0.53 and r = -0.55, respectively. There was no significant correlation between the tissue tPA content and the age and sex, steroid use, or plasma tPA of the patients or the duration of symptoms. In summary, this is the first demonstration of tPA in a large series of human brain tumors and in normal brain. The differences observed have clear biological significance and, although the source of tPA in tumor tissue is still unknown, a relative reduction in tPA in tumor tissue may play an integral role in the development of tissue necrosis and tissue edema. The lack of tPA in tumor necrosis was not due to tissue destruction and cell death since urokinase was readily detectable in that tissue.