Plasminogen activator inhibitor-1 (PAI-1) is a predictor of poor outcome in cancer. An explanation for this paradoxical role of PAI-1 has been its documented pro-angiogenic activity. The effect of PAI-1 on tumor cells however has not been fully explored. Here we have examined the effect of PAI-1 knockdown (KD) on the survival of four human cancer cell lines (high PAI-1 expression: HT-1080 fibrosarcoma and MDA-MB-231 breast cancer cells; low PAI-1 expression: A549 lung adenocarcinoma and HCT-116 colon carcinoma cells) in vitro and in vivo. We first demonstrated a statistically significant decrease in cell survival and a statistically significant increase in apoptosis in all four cell lines upon PAI-1 downregulation by small interfering RNA (siRNA) or in the presence of PAI-039 (Tiplaxtinin), a small molecule inhibitor of PAI-1 (P < 0.05). Increased levels of cell-associated urokinase plasminogen activator (uPA) and plasmin were also detected upon PAI-1 downregulation, and spontaneous apoptosis was blocked by a caspase-8 inhibitor, Fas/FasL neutralizing antibodies, and plasmin inhibitors, indicating that PAI-1 protects tumor cells from Fas/FasL-mediated extrinsic apoptosis involving plasmin. Tumor cells then were transduced with scramble (sc) short hairpin RNA (shRNA) or PAI-1 shRNA lentivirus and were implanted into PAI-1 wildtype (WT) or knockout (KO) mice, generating four experimental groups (group 1: sc shRNA tumor cells in WT mice; group 2: PAI-1 shRNA tumor cells in WT mice; group 3: sc shRNA tumor cells in KO mice; group 4: PAI-1 shRNA tumor cells in KO mice). In the HT-1080 model, we observed a statistically significant decrease in tumor growth, tumor take, bromodeoxyuridine incorporation (P = 0.0036), and microvessel density (P < 0.001), associated with a statistically significant increase in apoptosis (P = 0.0006) only in tumors from group 4 mice, but not from group 2 or 3 mice, when compared to tumors from group 1 mice. A similar statistically significant inhibition in tumor growth was also observed in tumors from group 4 mice when compared with tumors from group 1 mice in A549 and HCT-116 models. In all tumor-bearing group 4 mice, low levels of human PAI-1 (5-10 ng/mL) were detected indicating an incomplete inhibition of tumor-derived PAI-1 by shRNA. Five out of 15 group 4 mice in the HT-1080 model and 2 out of 8 group 4 mice in the MDA-MB-231 model never developed tumors, and in those mice human PAI-1 was never detected in the plasma, suggesting that PAI-1 may be necessary for tumor growth. In conclusion, PAI-1 exerts a protective effect against Fas/FasL-mediated extrinsic apoptosis in tumor cells. Downregulation of PAI-1 in both tumor and host cells is necessary for a significant inhibitory activity on tumorigenesis through a dual effect on tumor cell and endothelial cell apoptosis. Our data confirm PAI-1 as a therapeutic target in cancer and support further investigation of PAI-1 inhibitors in cancer therapy.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 2462. doi:1538-7445.AM2012-2462