Because cells within a colon carcinoma in vivo are most likely exposed to circulating human serum plas minogen and since the cell lines in our study did not incorporate thymidine or divide in serumfree medium, fetal calf serum in RPMI was routinely used as the culture medium.Under these conditions, traces of Benzocaine plasmin activity in the medium may explain both the presence of in creased levels of active cellassociated uPA and the enhanced ability of PAI to inhibit cancer cellassociated uPA in our fibrin zymography studies.Despite the presence of small traces of plasmin in the culture medium, the addition of exogenous plasminogen produced markedly increased degradation and susceptibility to inactivation by recombinant PAI, indicating that only miniscule traces of plasmin in plasma are required for such a process to operate.Expression or secretion of active cellassociated uPA could M.Direct hydrolysis by plasmili of matrix components allowing collagen and elastin to be more thoroughly de graded by other proteases have been reported.The addition of exogenous plasmin to basement membrane produced a loss of laminin and fibronectin antigen staining while collagen type IV antigens were retained. Although it is obvious that the presence of cell surface plasmin is not sufficient for complete basement membrane degradation, the expression of cell surface plasmin alone will result in a basement membrane with only the type collagen IV framework of the basement membrane intact in the immediate environment surrounding the cell.In turn, the collagen framework may be degraded by cancer cell collagenases, previously activated by cell surface plasmin.In this fashion, malignant cancer cells may cause localized ex tracellular matrix proteolysis in a very focused manner by the expression of cell surface plasmin, activated through the action ofuPA.In this study, purified uPA, tumor cell secreted uPA, and cancer cell homogenate uPA were susceptible to inhibition by the specific inhibitor recombinant PAI, sometimes at approx imately a: molar ratio.In contrast, orders of magnitude higher molar amounts of PAI in culture medium were re quired to inhibit cancer cell degradation of basement membrane extracellular matrix.Similar concentrations were required to inhibit degradation of smooth mus cle extracellular matrix by HT fibrosarcoma cells in the presence of serum. Differences in the susceptibility of membranebound serine proteases to inhibitors compared to that of secreted proteases, similar to those found in this study, have been reported previously. In a recent work, the formation of active cellassociated plasmin by uPA, in the presence of exogenous plasminogen, was inhibited by antiuPA active siteblocking antibodies, exogenous PAI, or exogenous PAI. The membrane localization of macrophage uPA has been suggested to protect it from inhibitors, while secreted uPA is markedly more sensi tive to inhibition. PAI secreted from endotoxinstimulated macrophages inhibited cellassociated uPA by only, while secreted uPA was inhibited by, reacts rapidly and with: stoichiometry with nativerecombinant uPA under dilute solution conditions.Unfortunately the same type of cysteine kinetic analysis is not appropriate for cellassociated or receptor bound uPA.Although these studies imply that increased protease expression is a significant factor in the formation of a more invasive cell type, the loss of genes for particular endog enous antiproteases may be equally responsible for the devel opment of the invasive phenotype.