Characterization of gelatinases linked to extracellular matrix invasion in ovarian adenocarcinoma: purification of matrix metalloproteinase 2.

Substantial evidence indicates that proteolytic degradation of the extracellular matrix is necessary for invasion and metastasis by cancer cells. Our previous work has demonstrated elevated secretion by cultured ovarian adenocarcinoma cells of two gelatinolytic metalloproteinases, a 72-kDa enzyme resembling matrix metalloproteinase 2 (MMP-2) and a 92-kDa enzyme resembling MMP-9 (Moser et al, Int. J. Cancer 56, 552-559, 1994). To assess the potential in vivo relevance of these enzymes, we have examined ovarian carcinoma ascites using gelatin substrate zymography. MMP species identical to those secreted from several well-characterized ovarian adenocarcinoma cell lines were found in the majority of ascites: MMP-2-like gelatinase (23 of 23 cases) and MMP-9-like gelatinase (18 of 23 cases), suggesting a prevalence of these species in the ovarian carcinoma microenvironment and their availability for tumor-associated proteolysis. The contribution of these proteinases to ovarian cancer invasion was further demonstrated by experiments measuring tumor cell-mediated proteolysis of native endothelial cell extracellular matrix (ECM) and tumor cell invasion of reconstituted basement membrane (Matrigel). These data showed that secretion of type IV collagenase activity by a series of independently isolated ovarian adenocarcinoma cell lines correlated well with the ability of these cells to proteolyze the ECM and invade the basement membrane. Furthermore, we have identified and characterized an ovarian carcinoma-associated gelatinase, the 72-kDa MMP found in conditioned media of the DOV 13 cell line, as MMP-2. This enzyme was identical to the previously described MMP-2 from other sources by Western blot, amino terminal sequence, and substrate specificity. Additionally, a large portion of the MMP-2 activity found in DOV 13 conditioned media is active without organomercurial treatment, suggesting that ovarian cancer cells have an endogenous activator of the zymogen. Together, these data suggest that ECM proteolysis mediated by tumor-associated proteinases plays an important role in the invasion and/or metastasis of ovarian carcinoma.

Comparison of plasminogen binding and activation on extracellular matrices produced by vascular smooth muscle and endothelial cells.

Plasminogen is the zymogen form of the serine proteinase plasmin. Although plasmin functions primarily as a fibrinolytic enzyme, recent evidence from numerous laboratories indicates that plasmin is also active in extracellular-matrix (ECM) proteolysis. The role of plasmin in ECM degradation suggests that activation of plasminogen may be regulated by interaction with components of the ECM. In the current study, we have investigated binding and kinetic interactions between plasminogen, plasminogen activators and ECM synthesized by either vascular smooth muscle cells (SMCECM) or endothelial cells (ECECM). We report binding of plasminogen, tissue-type plasminogen activator (t-PA) and urinary-type plasminogen activator (u-PA) to intact SMCECM with concentrations of ligand yielding half-maximal binding (B50) of 34, 5 and 15 nM, respectively. ECECM bound only plasminogen and t-PA, with B50 values of 32 nM and 10 nM, respectively. The initial rate of t-PA-catalyzed plasminogen activation was enhanced 41-fold in the presence of SMCECM and 27-fold on ECECM. In contrast, u-PA-catalyzed activation on SMCECM and ECECM was increased only 1.5-fold or 3-fold, respectively. These data suggest that the ECM may provide an alternative surface for assembly and regulation of plasminogen activation.

Human alpha 1-proteinase inhibitor binds to extracellular matrix in vitro.

alpha 1-Proteinase inhibitor (alpha 1-PI) is the major endogenous inhibitor of human leukocyte elastase (HLE). We have employed two different methods to quantitate the binding of alpha 1-PI to extracellular matrix (ECM), composed of 51% glycoproteins and proteoglycans, 37% types I and III collagen, and 12% elastin, derived from rat heart smooth muscle cells. alpha 1-PI is tightly bound to ECM via a saturable adsorption process; the bound protein fails to dissociate from the matrix after repeated washing. Binding of alpha 1-PI is unaffected by the prior removal of ECM glycoproteins with trypsin. Binding to ECM is not decreased in the presence of high salt but is decreased at low pH. A 40-fold excess of unlabeled alpha 1-PI displaces only 50% of [125I]alpha 1-PI prebound to ECM. A 30% decrease in the levels of alpha 1-PI bound to ECM is observed after DTT washes of ECM preincubated with alpha 1-PI or when alpha 1-PI is modified with iodoacetamide prior to incubation with ECM, implying that a fraction of bound alpha 1-PI is covalently linked to ECM via disulfide bond formation. Moreover, high molecular weight complexes between [125I]alpha 1-PI and ECM components can be visualized by SDS-PAGE under nonreducing conditions but disappear upon reduction. Approximately 50% of the total alpha 1-PI bound covalently or noncovalently to ECM retains the ability to inhibit HLE-mediated ECM proteolysis. alpha 1-PI-HLE complexes bound to ECM can be visualized by SDS-PAGE following the addition of HLE to ECM that was pretreated with [125I]alpha 1-PI. alpha 1-PI from normal plasma or serum also binds to ECM with retention of immunoreactivity and partial retention of inhibitory activity. However, ECM pretreated with alpha 1-PI-deficient serum retains no HLE-inhibitory activity.

Inhibition of human leucocyte elastase by ursolic acid. Evidence for a binding site for pentacyclic triterpenes.

Several pentacyclic triterpenoid metabolites of plant origin are inhibitors of hydrolysis of both synthetic peptide substrates and elastin by human leucocyte elastase (HLE). Ursolic acid, the most potent of these compounds, has an inhibition constant of 4-6 microM for hydrolysis of peptide substrates in phosphate-buffered saline. With tripeptide and tetrapeptide substrates, the inhibition is purely competitive, whereas with a shorter dipeptide substrate the inhibition is non-competitive, suggesting that ursolic acid interacts with subsite S3 of the extended substrate-binding domain in HLE, but not with subsites S1 and S2. The carboxy group at position 28 in the pentacyclic-ring system of the triterpenes contributes to binding to HLE, since replacement of this group with a hydroxy group, as in uvaol, the alcohol analogue of ursolic acid, reduces the potency of inhibition. The inhibitory potency of ursolic acid is also reduced by addition of 1 M-NaCl, further supporting a postulated electrostatic interaction between the negative charge on the triterpene and a positively charged residue on the enzyme, which we assign to the side chain of Arg-217, located in the vicinity of subsites S4 and S5 in HLE. These observations are consistent with a binding site for ursolic acid which extends from S3 towards S4 and S5 on the enzyme. Other triterpenes, including oleanolic acid, erythrodiol, hederagenin and 18 beta-glycyrrhetic acid, can also interact with this binding site. On the basis of these results we conclude that the extended substrate-binding domain of HLE can accommodate a variety of hydrophobic ligands, including not only such molecules as fatty acids [Ashe & Zimmerman (1977) Biochem. Biophys. Res. Commun. 75, 194-199; Cook & Ternai (1988) Biol. Chem. Hoppe-Seyler 369, 629-637], but also polycyclic molecules such as the pentacyclic triterpenoids.