Cellular invasion depends on cooperation between adhesive and proteolytic mechanisms. Evidence is provided that the matrix metalloproteinase MMP-2 can be localized in a proteolytically active form on the surface of invasive cells, based on its ability to bind directly integrin alpha v beta 3. MMP-2 and alpha v beta 3 were specifically colocalized on angiogenic blood vessels and melanoma cells in vivo. Expression of alpha v beta 3 on cultured melanoma cells enabled their binding to MMP-2 in a proteolytically active form, facilitating cell-mediated collagen degradation. In vitro, these proteins formed an SDS-stable complex that depended on the noncatalytic C-terminus of MMP-2, since a truncation mutant lost the ability to bind alpha v beta 3. These findings define a single cell-surface receptor that regulates both matrix degradation and motility, thereby facilitating directed cellular invasion.

The 72-kDa gelatinase/type IV collagenase (MMP-2) is a member of the matrix metalloproteinase (MMP) family of enzymes. This enzyme is known to cleave type IV collagen as well as degrade denatured collagens. However, native interstitial collagens are reportedly resistant to MMP-2 and are thought to be susceptible only to the interstitial collagenases MMP-1 and MMP-8. In this study we report that both human and chicken MMP-2, free of tissue inhibitors of metalloproteinases (TIMPs) are capable of cleaving soluble, triple helical type I collagen generating the 3/4- and 1/4-length collagen fragments characteristic of vertebrate interstitial collagenases. MMP-2 cleaves at the same Gly-Ile/Leu bond in the collagen alpha chains as interstitial collagenases with kcat and Km values similar to that of MMP-1. MMP-2 also is capable of degrading reconstituted type I collagen fibrils. The closely related 92-kDa gelatinase/type IV collagenase (MMP-9) is unable to cleave soluble or fibrillar collagen under identical conditions indicating that the specific collagenolytic activity of MMP-2 is not a general property of gelatinases. That MMP-2, a potent gelatinase, also can cleave fibrillar collagen provides an alternative to the proposal that two enzymes, an interstitial collagenase and a gelatinase, are required for the complete dissolution of stromal collagen during cellular invasion.

We have previously used a subtractive immunization (SI) approach to generate monoclonal antibodies (mAbs) against proteins preferentially expressed by the highly metastatic human epidermoid carcinoma cell line, M + HEp3. Here we report the immunopurification, identification and characterization of SIMA135/CDCP1 (subtractive immunization M + HEp3 associated 135 kDa protein/CUB domain containing protein 1) using one of these mAbs designated 41-2. Protein expression levels of SIMA135/CDCP1 correlated with the metastatic ability of variant HEp3 cell lines. Protein sequence analysis predicted a cell surface location and type I orientation of SIMA135/CDCP1, which was confirmed directly by immunocytochemistry. Analysis of deglycosylated cell lysates indicated that up to 40 kDa of the apparent molecular weight of SIMA135/CDCP1 is because of N-glycosylation. Western blot analysis using a antiphosphotyrosine antibody demonstrated that SIMA135/ CDCP1 from HEp3 cells is tyrosine phosphorylated. Selective inhibitor studies indicated that an Src kinase family member is involved in the tyrosine phosphorylation of the protein. In addition to high expression in M + HEp3 cells, the SIMA135/CDCP1 protein is expressed to varying levels in 13 other human tumor cell lines, manifesting only a weak correlation with the reported metastatic ability of these tumor cell lines. The protein is not detected in normal human fibroblasts and endothelial cells. Northern blot analysis indicated that SIMA135/ CDCP1 mRNA has a restricted expression pattern in normal human tissues with highest levels of expression in skeletal muscle and colon. Immunohistochemical analysis indicated apical and basal plasma membrane expression of SIMA135/CDCP1 in epithelial cells in normal colon. In colon tumor, SIMA135/CDCP1 expression appeared dysregulated showing extensive cell surface as well as cytoplasmic expression. Consistent with in vitro shedding experiments on HEp3 cells, SIMA135/CDCP1 was also detected within the lumen of normal and cancerous colon crypts, suggesting that protein shedding may occur in vivo. Thus, specific immunodetection followed by proteomic analysis allows for the identification and partial characterization of a heretofore uncharacterized human cell surface antigen.

The contribution of specific type I collagen remodeling in angiogenesis was studied in vivo using a quantitative chick embryo assay that measures new blood vessel growth into well-defined fibrillar collagen implants. In response to a combination of basic fibroblast growth factor (bFGF) and vascular endothelial growth factor (VEGF), a strong angiogenic response was observed, coincident with invasion into the collagen implants of activated fibroblasts, monocytes, heterophils, and endothelial cells. The angiogenic effect was highly dependent on matrix metalloproteinase (MMP) activity, because new vessel growth was inhibited by both a synthetic MMP inhibitor, BB3103, and a natural MMP inhibitor, TIMP-1. Multiple MMPs were detected in the angiogenic tissue including MMP-2, MMP-13, MMP-16, and a recently cloned MMP-9-like gelatinase. Using this assay system, wild-type collagen was compared to a unique collagenase-resistant collagen (r/r), with regard to the ability of the respective collagen implants to support cell invasion and angiogenesis. It was found that collagenase-resistant collagen constitutes a defective substratum for angiogenesis. In implants made with r/r collagen there was a substantial reduction in the number of endothelial cells and newly formed vessels. The presence of the r/r collagen, however, did not reduce the entry into the implants of other cell types, that is, activated fibroblasts and leukocytes. These results indicate that fibrillar collagen cleavage at collagenase-specific sites is a ratelimiting event in growth factor-stimulated angiogenesis in vivo. IntroductionAngiogenesis is the process by which the preexisting vascular tree gives rise to new blood vessels. This process is tightly regulated during development and occurs only under highly specialized circumstances in the normal adult animal. Unregulated angiogenesis may be a pivotal element of disease etiology, such as during tumor growth and atherosclerosis. 1,2 All forms of angiogenesis, however, are thought to share certain basic features, including migration and mitogenesis of endothelial cells, lumen formation, connection of new vascular segments with the preexisting circulation, and extensive remodeling of the extracellular matrix by proteases. The detailed mechanisms by which proteolytic enzymes, such as the matrix metalloproteinases (MMPs), mediate some of these events in vivo remain unclear. 3,4 The MMPs constitute a large family of zinc-dependent endopeptidases that have been strongly implicated in both normal angiogenesis and tumor vascularization. 2,4 These enzymes are characteristically regulated at multiple levels, including gene expression, spatial localization, zymogen activation, and inhibition by the tissue inhibitors of metalloproteinases (TIMPs). 5 In vitro remodeling of extracellular matrices by cultured endothelial cells has consistently been shown to rely on MMPs. 4 However, it is unclear which aspects of cell culture models for angiogenesis may be directly extrapolated to the animal. In vivo studies also have foun...

SummaryMany serine proteases play important regulatory roles in complex biological systems, but only a few have been linked directly with capillary morphogenesis and angiogenesis. Here we provide evidence that serine protease activities, independent of the plasminogen activation cascade, are required for microvascular endothelial cell reorganization and capillary morphogenesis in vitro. A homology cloning approach targeting conserved motifs present in all serine proteases, was used to identify candidate serine proteases involved in these processes, and revealed 5 genes (acrosin, testisin, neurosin, PSP and neurotrypsin), none of which had been associated previously with expression in endothelial cells. A subsequent gene-specific RT-PCR screen for 22 serine proteases confirmed expression of these 5 genes and identified 7 additional serine protease genes expressed by human endothelial cells, urokinase-type plasminogen activator, protein C, TMPRSS2, hepsin, matriptase/ MT-SP1, dipeptidylpeptidase IV, and seprase. Differences in serine protease gene expression between microvascular and human umbilical vein endothelial cells (HUVECs) were identified and several serine protease genes were found to be regulated by the nature of the substratum, ie. artificial basement membrane or fibrillar type I collagen. mRNA transcripts of several serine protease genes were associated with blood vessels in vivo by in situ hybridization of human tissue specimens. These data suggest a potential role for serine proteases, not previously associated with endothelium, in vascular function and angiogenesis.

We have demonstrated previously that new blood vessel formation induced by angiogenic growth factors in onplants placed on the chorioallantoic membrane (CAM) of the chick embryos is critically dependent on the cleavage of fibrillar collagen by a previously unidentified interstitial collagenase. In the present study we have used a quantitative CAM angiogenesis system to search for and functionally characterize host avian collagenases responsible for the collagen remodeling associated with angiogenesis. Among the matrix metalloproteinases (MMPs) identified in the CAM onplant tissue, the chicken MMP-13 (chMMP-13) was the only enzyme whose induction and expression coincided with the onset of angiogenesis and blood vessel formation. The chMMP-13 cDNA has been cloned and recombinantly expressed. The chMMP-13 protein has been purified, characterized in vitro, and examined in situ in the CAM. MMP-13-positive cells appear in the CAM shortly after angiogenic stimulation and then accumulate in the collagen onplant tissue. Morphologically, the chMMP-13-containing cells appear as hematopoietic cells of monocyte/macrophage lineage. In vitro, the chMMP-13 proenzyme is rapidly and efficiently activated through the urokinase plasminogen activator/plasminogen/plasmin cascade into a collagenase capable of cleaving native but not the (r/r) mutant collagenase-resistant collagen. Surprisingly, nanogram levels of purified chMMP-13 elicit an angiogenic response in the CAM onplants comparable with that induced by the angiogenic growth factors. The chMMP-13-mediated response was efficiently blocked by select protease inhibitors indicating that plasmin-activated chMMP-13 can function as an angiogenic factor in vivo. Altogether, the results of this study extend the physiological role of MMP-13, previously associated with cartilage/ bone resorption, to the collagen remodeling involved in the angiogenic cascade.

The glycine-rich loop, one of the most important motifs in the conserved protein kinase catalytic core, embraces the entire nucleotide, is very mobile, and is exquisitely sensitive to what occupies the active site cleft. Of the three conserved glycines [G(50)TG(52)SFG(55) in cAMP-dependent protein kinase (cAPK)], Gly(52) is the most important for catalysis because it allows the backbone amide of Ser(53) at the tip of the loop to hydrogen bond to the gamma-phosphate of ATP [Grant, B. D. et al. (1998) Biochemistry 37, 7708]. The structural model of the catalytic subunit:ATP:PKI((5)(-)(24)) (heat-stable protein kinase inhibitor) ternary complex in the closed conformation suggests that Ser(53) also might be essential for stabilization of the peptide substrate-enzyme complex via a hydrogen bond between the P-site carbonyl in PKI and the Ser(53) side-chain hydroxyl [Bossemeyer, D. et al. (1993) EMBO J. 12, 849]. To address the importance of the Ser(53) side chain in catalysis, inhibition, and P-site specificity, Ser(53) was replaced with threonine, glycine, and proline. Removal of the side chain (i.e., mutation to glycine) had no effect on the steady-state phosphorylation of a peptide substrate (LRRASLG) or on the interaction with physiological inhibitors, including the type-I and -II regulatory subunits and PKI. However, this mutation did affect the P-site specificity; the glycine mutant can more readily phosphorylate a P-site threonine in a peptide substrate (5-6-fold better than wild-type). The proline mutant is compromised catalytically with altered k(cat) and K(m) for both peptide and ATP and with altered sensitivity to both regulatory subunits and PKI. Steric constraints as well as restricted flexibility could account for these effects. These combined results demonstrate that while the backbone amide of Ser(53) may be required for efficient catalysis, the side chain is not.

A variety of invertebrates possess plasma lectins with sialic acid recognition capabilities. One of the best studied of these lectins is limulin, which is a member of the pentraxin family of proteins and is found in the plasma of the American horseshoe crab, Limulus polyphemus. We find that limulin is one of several sialic acid-binding lectins of Limulus plasma and is present at a much lower abundance than Limulus C-reactive protein, the other plasma pentraxin. Limulin was purified by sequential affinity chromatography on phosphorylethanolamine-agarose, which isolates the pentraxins and separates limulin from the other sialic acid-binding lectins of the plasma, followed by fetuin-Sepharose, which binds limulin and separates it from Limulus C-reactive protein, the most abundant pentraxin of the plasma. We show here that limulin is the mediator of the Ca+2-dependent hemolytic activity found in the plasma of Limulus. Plasma that was depleted in the pentraxins by passage over phosphorylethanolamine-agarose or was depleted in the sialic acid-binding lectins by passage over fetuin-Sepharose lacked hemolytic activity. Purified limulin was hemolytic at concentrations of 3-5 nM. The other sialic acid-binding lectins of Limulus plasma and Limulus C-reactive protein were nonhemolytic. Foreign cell cytolysis by limulin represents a novel function for a plasma lectin and is the first documented function for limulin.