Adhesion-dependent control of matrix metalloproteinase-2 activation in human capillary endothelial cells.

The growth and regression of capillary blood vessels during angiogenesis is greatly influenced by changes in the activity of matrix metalloproteinases (MMPs), which selectively degrade extracellular matrix (ECM) and thereby modulate capillary endothelial cell shape, growth and viability. However, changes in cell-ECM binding and cell spreading have also been reported to alter MMP secretion and activation. Studies were carried out to determine whether changes in integrin binding or cell shape feed back to alter MMP-2 processing in human capillary endothelial (HCE) cells. Catalytic processing of proMMP-2 to active MMP-2 progressively decreased when HCE cells were cultured on dishes coated with increasing densities of fibronectin (FN), which promote both integrin binding and cell spreading. Conversely, the highest levels of active MMP-2 were detected in round cells cultured on low FN. When measured 24 hours after plating, this increase in active MMP-2 was accompanied by a concomitant rise in mRNA and protein levels for the membrane-type 1 MMP (MT1-MMP), which catalyzes the cleavage of proMMP-2. To determine whether proMMP-2 processing was controlled directly by integrin binding or indirectly by associated changes in cell shape, round cells on low FN were allowed to bind to microbeads (4.5 microm diameter) coated with a synthetic RGD peptide or FN; these induce local integrin receptor clustering without altering cell shape. ProMMP-2 activation was significantly decreased within minutes after bead binding in these round cells, prior to any detectable changes in expression of MT1-MMP, whereas binding of beads coated with control ligands for other transmembrane receptors had no effect. This inhibitory effect was mimicked by microbeads coated with activating antibodies against alphaVbeta3 and beta1 integrins, suggesting a direct role for these cell-surface ECM receptors in modulating proMMP-2 activation. Similar inhibition of proMMP-2 processing by integrin binding, independent of cell spreading, was demonstrated in cells that were cultured on small, microfabricated adhesive islands that prevented cell spreading while presenting a high FN density directly beneath the cell. Interestingly, when spread cells were induced to round up from within by disrupting their actin cytoskeleton using cytochalasin D, proMMP-2 processing did not change at early times; however, increases in MT1-MMP mRNA levels and MMP-2 activation could be detected by 18 hours. Taken together, these results suggest the existence of two phases of MMP-2 regulation in HCE cells when they adhere to ECM: (1) a quick response, in which integrin clustering alone is sufficient to rapidly inhibit processing of proMMP-2 and (2) a slower response, in which subsequent cell spreading and changes in the actin cytoskeleton feed back to decrease expression of MT1-MMP mRNA and, thereby, further suppress cellular proteolytic activity.