Cathepsin G--induced release of PAI-1 in the culture medium of endothelial cells: a new thrombogenic role for polymorphonuclear leukocytes?

Activated polymorphonuclear leukocytes (PMNs) may affect the integrity of blood vessels by endothelial cell injury. We investigated the effects of cathepsin G purified from human neutrophils on the fibrinolytic potential of cultured human umbilical vein endothelial cells (HUVECs). Cathepsin G (5 and 10 micrograms/ml) induced marked intercellular gap formation after 1 hour of treatment, whereas 1 microgram/ml did not, even after 6 hours incubation. In contrast, plasminogen activator inhibitor-1 (PAI-1) antigen levels, measured by a double antibody enzyme-linked immunosorbent assay, were significantly increased in culture media (CM) on cathepsin G (1 microgram/ml) treatment after 15 minutes (5.1 +/- 1.2 ng/ml vs 2.6 +/- 0.6 ng/ml for controls, p < 0.01) and 6 hours of incubation (69.6 +/- 17.5 ng/ml vs 40.0 +/- 9.0 ng/ml for controls, p < 0.01). Likewise, PAI activity, measured by reverse fibrin autography, increased on cell treatment with cathepsin G. Preincubation of cathepsin G with eglin C (10 micrograms/ml) almost completely abolished the increase in both PAI antigen and activity levels induced by cathepsin G. Cycloheximide, a protein synthesis inhibitor, did not block cathepsin G-induced PAI-1 release. PAI-1 mRNA levels were not affected by HUVEC treatment with cathepsin G (1 microgram/ml for 15 minutes), even after 24 hours. In the extracellular matrix (ECM) PAI-1 antigen levels decreased to 77% and 40% of controls, respectively, after 15 minutes and 6 hours of cathepsin G (1 micrograms/ml) treatment. Reverse fibrin autography also demonstrated a dose-dependent reduction of PAI activity in the ECM on 6 hours of cell treatment with 1 or 5 micrograms/ml cathepsin G. Moreover, ECM prepared from confluent HUVECs released PAI-1 in supernatants on 1 micrograms/ml cathepsin G incubation in a cell-free system. Tissue-type plasminogen activator (t-PA) activity was strongly depressed on cathepsin G treatment, both in CM from HUVECs or in a cell-free system. Finally, PAI-1 was also released from cathepsin G-stimulated platelets in a dose-dependent manner. In summary, our results support a potentially thrombogenic role of cathepsin G, which could impair the fibrinolytic potential of the endothelium. These data give a new insight into the mechanisms by which activated PMNs may promote thrombus formation. On the other hand, the decrease of PAI-1 in ECM could favor penetration and migration of inflammatory or tumor cells through the subendothelial layers.