Processing of Integrin αv Subunit by Membrane Type 1 Matrix Metalloproteinase Stimulates Migration of Breast Carcinoma Cells on Vitronectin and Enhances Tyrosine Phosphorylation of Focal Adhesion Kinase*

Recently, we have shown that membrane type 1 matrix metalloproteinase (MT1-MMP) exhibits integrin convertase activity. Similar to furin-like proprotein convertases, MT1-MMP directly processes a single chain precursor of αvintegrin subunit (pro-αv) into the heavy and light α-chains connected by a disulfide bridge. To evaluate functionality of MT1-MMP-processed integrins, we examined breast carcinoma MCF7 cells co-expressing αvβ3 integrin with either the wild type or mutant MT1-MMP in a variety of migration and adhesion tests. Specific inhibitors of proprotein convertases and MMP were employed in our cell system to attenuate the individual pathways of pro-αv maturation. We present evidence that MT1-MMP cleavage of pro-αv in the cells did not affect RGD-ligand binding of the resulting αvβ3 integrin but enhanced outside-in signal transduction through a focal adhesion kinase pathway. Enhanced tyrosine phosphorylation of focal adhesion kinase in cells co-expressing MT1-MMP and αvβ3integrin contributed to efficient adhesion and, especially, migration of cells on vitronectin, a ligand of αvβ3integrin. These mechanisms underscore the significance of a coordinated interplay between MT1-MMP and αvβ3 integrin in tumor cells and identify downstream signaling pathways resulting from their interactions. Regulation of integrin maturation and functionality may be an important role of MT1-MMP in tumor cells.

Integrins are a family of ␣␤-heterodimeric cell adhesion molecules consisting of two type I transmembrane glycoprotein subunits. Eight integrin ␤-chains pair with a restricted number of 16 ␣-chains, giving rise to 22 different integrins with distinct expression patterns and ligand binding profiles (1,2). Integrins play central roles in cell adhesion, cell migration, and control of cell differentiation, proliferation, and apoptosis (3)(4)(5)(6). Among all known integrins, ␣ v ␤ 3 plays a unique functional role in tumor angiogenesis and metastasis (7,8).
Integrins mediate signal transduction through the cell membrane in both directions (9 -11). Binding of ligands to integrins transmits signals into the cell and results in cytoskeletal rearrangements, gene expression, and cellular differentiation (outside-in signaling). Signals from within the cell can also propagate through integrins and regulate integrin-ligand binding and cell adhesion (inside-out signaling) (12). The cytoplasmic domain of integrins plays a pivotal role in these bi-directional signaling processes (13).
It has been established that similar to certain other integrin types, the mature ␣ v integrin subunit is generated by posttranslational endoproteolytic cleavage of the pro-␣ v (14,15). The cleavage at the highly conserved pairs of basic amino acids is known to be a function of furin-like proprotein convertases from the subtilisin/kexin family. This cleavage converts a single ␣-chain precursor into the mature subunit consisting of an N-terminal heavy ␣-chain and a C-terminal light ␣-chain connected by a disulfide bridge (16). Adhesive function of integrins has been directly associated with outside-in signaling (17). Integrin ligation normally induces tyrosine phosphorylation through outside-in signaling and activation of cytoplasmic tyrosine kinases (12). The maturation of the ␣ v integrin subunit is likely to be essential to outside-in signaling (18,19).
Evidence emerges that in migrating cells integrins frequently collaborate with matrix metalloproteinases (MMPs) 1 (20 -28). The human MMP family is composed of at least 25 zinc-dependent endopeptidases expressed as inactive precursors, or zymogens, that require the proteolytic N-terminal processing to display the functional activity (29,30). MT1-MMP, the most common membrane type MMP, belongs to a membrane-type subfamily. MT1-MMP is distinguished by the presence of a transmembrane domain that anchors the molecule to the plasma membrane and a cytoplasmic domain that interacts with the intracellular milieu (31,32). MT1-MMP is a cell surface activator of pro-MMP-2 and has been implicated in collagen invasion (33) and turnover (34) as well as in the proteolytic processing of cell surface receptors (35,36). There is a strong correlation between the expression of MT1-MMP and tumorigenesis in many cancer types (33,37,38).
In another report, 2 we presented evidence that MT1-MMP is capable of processing the precursor of the ␣ v integrin subunit (pro-␣ v ), thereby functioning as an integrin convertase. Thus, in breast carcinoma MCF7 cells co-expressing ␣ v ␤ 3 integrin and MT1-MMP, the protease specifically cleaves pro-␣ v , generating a 115-kDa heavy ␣-chain with the truncated C terminus and a 25-kDa light ␣-chain commencing from the N-terminal Leu 892 . Consistent with these observations, our earlier findings (20,21) demonstrated that the cells co-expressing MT1-MMP and ␣ v ␤ 3 integrin were significantly more migratory relative to the cells transfected with MT1-MMP or ␣ v ␤ 3 integrin alone. This prompted us to investigate the migratory characteristics of transfected cells in detail.
Here we report that processing of pro-␣ v by MT1-MMP facilitated ␣ v ␤ 3 integrin-mediated adhesion and especially migration of cells on vitronectin, the extracellular matrix ligand of ␣ v ␤ 3 integrin. MT1-MMP cleavage of pro-␣ v did not affect RGD-ligand binding of ␣ v ␤ 3 integrin. However, the cells exhibiting both ␣ v ␤ 3 integrin and MT1-MMP were more efficient in outside-in signal transduction through a focal adhesion kinase (FAK) pathway relative to the cells expressing ␣ v ␤ 3 integrin alone. Our data suggest that regulation of integrin functionality may be an important role of MT1-MMP in migrating tumor cells.

MATERIALS AND METHODS
Antibodies and Reagents-Rabbit antibodies AB815 specific to a hinge region of MT1-MMP and ␣ v ␤ 3 integrin-specific mAb LM609 were from Chemicon International (Temecula, CA). Fab-9 was purified as described (39). Na 125 I was from Amersham Biosciences. Broad range hydroxamate inhibitor of MMP activity, AG3340 (40), was a kind gift of Peter Baciu (Allergan Pharmaceuticals, Irvine, CA). A furin inhibitor decanoyl-Arg-Val-Lys-Arg-chloromethyl ketone (dec-RVKR-cmk) was from Bachem (King of Prussia, PA). Protease inhibitors (PMSF, leupeptin, pepstatin, and aprotinin), sodium fluoride, sodium pyrophosphate, and sodium orthovanadate were from Sigma. Vitronectin purified from human plasma was a kind gift of Richard DiScipio (La Jolla Institute for Molecular Medicine, La Jolla, CA).
Immunoprecipitation and Western Blotting-Cells were plated in DMEM/FCS supplemented with the indicated concentrations of a hydroxamate inhibitor AG3340. After preincubation with the inhibitor for 12-72 h, cells were washed with DPBS and surface-biotinylated with sulfo-NHS-LC-biotin according to the manufacturer's instructions (Pierce). Following incubation for 60 min, cells were washed with icecold DPBS and lysed with 50 mM N-octyl-␤-D-glucopyranoside (Amresco, Solon, OH) in TBS supplemented with 1 mM CaCl 2 , 1 mM MgCl 2 , and protease inhibitor mixture containing 1 mM PMSF and 2 g/ml each of aprotinin, pepstatin, and leupeptin. The lysates were precleared with Pansorbin (Calbiochem). The samples of cell lysates each containing 1.0 mg of protein were mixed with 2-3 g of anti-integrin or anti-MT1-MMP antibodies and protein G-agarose beads. Following incubation for 14 h at 4°C, agarose beads were washed with TBS containing 1 mM CaCl 2 , 1 mM MgCl 2 , and protease inhibitor mixture (1 mM PMSF and 2 g/ml each of aprotinin, pepstatin, and leupeptin), then with DPBS supplemented with 0.05% Tween 20 (DPBS/Tween 20) and 0.5 M NaCl, and next with DPBS/Tween 20. Immune complexes were released by boiling the beads for 5 min in 2ϫ SDS sample buffer (20 l). After centrifugation, solubilized proteins were reduced and subjected to SDS-PAGE. Separated proteins were transferred to an Immobilon P membrane (Millipore, Bedford, MA). Following blocking with 1% casein in DPBS/Tween 20, the membrane was probed with ExtrAvidin-HRP (Sigma), and the bound HRP activity was visualized with the TMB/M substrate (Chemicon).
Tyrosine Phosphorylation of FAK-Cells were plated for 48 h in DMEM/FCS with or without dec-RVKR-cmk (100 M), a furin inhibitor. Cells were serum-starved for 24 h with or without the inhibitor, detached with minimum trypsinization, extensively washed with ice-cold 1% BSA/DMEM, and resuspended in the prewarmed 1% BSA/DMEM supplemented with 20 mM HEPES, pH 7.2. Tissue culture dishes (150 mm in diameter) were coated overnight at 4°C with the indicated concentrations of vitronectin and then washed with PBS and blocked with 1% BSA. Cells were allowed to adhere for 1 h to vitronectin-coated surfaces. Adherent cells were washed with ice-cold DPBS supplemented with 1 mM CaCl 2 and 1 mM MgCl 2 and lysed with 1% Triton X-100 in TBS containing 1.0 mM EDTA, 50 mM sodium fluoride, 10 mM sodium pyrophosphate, 1 mM sodium orthovanadate, and inhibitor mixture (1 mM PMSF and 2 g/ml each of aprotinin, pepstatin, and leupeptin). After incubation for 1 h on ice, cell lysates were centrifuged at 14,000 rpm for 20 min at 4°C. The samples of cell lysates each containing 1.0 mg of protein were mixed with anti-FAK rabbit antibodies sc-557 (Santa Cruz Biotechnology, Santa Cruz, CA) and protein G-agarose. Following incubation for 14 h at 4°C, agarose beads were washed with TBS containing 1.0 mM EDTA, 50 mM sodium fluoride, 10 mM sodium pyrophosphate, 1 mM sodium orthovanadate, and inhibitor mixture (1 mM PMSF and 2 g/ml each of aprotinin, pepstatin, and leupeptin), then with TBS supplemented with 0.05% Tween 20 and 0.5 M NaCl, and next with TBS supplemented with 0.05% Tween 20. Immune complexes were released by boiling the beads for 5 min in 2ϫ SDS sample buffer (20 l). After centrifugation, solubilized proteins were reduced and subjected to SDS-PAGE. Separated proteins were transferred to a membrane. After blocking with 1% BSA, the membrane was incubated with 2 g/ml anti-FAK rabbit sc-557 antibodies or anti-phosphotyrosine murine mAb 4G10 (Upstate Biotechnology, Lake Placid, NY), followed by incubation with HRP-conjugated secondary antibodies (Sigma) and the TMB/M substrate. To estimate the phosphotyrosine (Tyr(P)) content in the immunoprecipitated FAK samples, the intensity of the Tyr(P) and FAK bands was quantified using an AlphaEase camera and AlphaImager software (Alpha Innotech, San Leandro, CA).
Fab-9 Binding to ␣ v ␤ 3 -expressing Cells-Fab-9 binding was performed as reported previously (43). Cells (3 ϫ 10 6 /ml) were incubated with increasing concentrations of 125 I-Fab-9 at 14°C for 70 min. Nonspecific binding was measured in the presence of 5 mM EDTA. Following incubation, free 125 I-Fab-9 was separated from cell-bound ligand by centrifugation through a layer of 20% sucrose in TBS at 14,000 rpm for 3 min. The radioactivity in pellets was determined in a gamma counter. The K D values of the ␣ v ␤ 3 ⅐Fab-9 complex were calculated by Scatchard analysis and by fitting the data with a single binding site curve using GraphPad Prism software (GraphPad Software, San Diego, CA).
Cell Adhesion Assay-Cell adhesion was performed in the wells of a high binding 96-well plate (Corning Glass) precoated overnight at 4°C with vitronectin or Fab-9 at the indicated concentrations. Plates were blocked for 1 h at 37°C with 1% BSA/DMEM supplemented with 10 mM HEPES, pH 7.2. Cells were incubated overnight in DMEM/FCS, detached with enzyme-free buffer (Specialty Media, Lavalette, NJ), washed, and resuspended in with 1% BSA/DMEM supplemented with 10 mM HEPES, pH 7.2. Cells were plated at 5 ϫ 10 4 , 0.1 ml in 1% BSA/DMEM supplemented with 10 mM HEPES, pH 7.2, and allowed to attach for 1 h at 37°C. Where indicated, the increasing concentrations of cyclo(Arg-Gly-Asp-D-Phe-Val) peptide (cRGD; Peptide Institute, Osaka, Japan) were used to inhibit cell adhesion. Adherent cells were fixed and stained with Crystal Violet, and the incorporated dye was extracted and quantified (21).
Cell Migration Assay-The directional migration of cells in Transwells (Costar, Cambridge, MA) was analyzed under serum-free conditions as described (21). The undersurface of a 6.5-mm insert membrane with an 8-m pore size was coated at 4°C overnight with 2 g/ml vitronectin and blocked with 1% BSA. Cells cultured overnight in DMEM/FCS were detached with enzyme-free buffer. Cells (7.5 ϫ 10 4 ) were plated in 0.15 ml AIM-V medium (Invitrogen) per insert. The outer chamber was filled with 0.6 ml of AIM-V medium. Following incubation for 48 h, cells that migrated to the undersurface of the membrane were detached with trypsin/EDTA and counted. Where indicated, cells were first incubated in the presence of 10 M AG3340 and/or 100 M furin inhibitor dec-RVKR-cmk for 24 -72 h and then plated into Transwells with the same inhibitor(s) added to both outer and inner chambers.

RESULTS
Two Pathways of Integrin Maturation-In another study, 2 we reported the novel mechanisms involving MT1-MMP in the limited proteolytic cleavage of pro-␣ v in MCF7 breast carci-noma cells. Our findings demonstrated that MT1-MMP functions as an integrin convertase. Similar to furin-like proprotein convertases, MT1-MMP converts a single chain pro-␣ v into the processed heavy and light ␣-chains connected by a disulfide bridge. To corroborate these findings and to confirm that the processing of pro-␣ v by MT1-MMP is not restricted to a single cell type, we examined the cell line of the glioma U251 cells that overexpressed MT1-MMP (42). Empty vector construct served as control.
To attenuate the effects of MT1-MMP on pro-␣ v , we inhibited the furin pathway of pro-␣ v maturation by incubating glioma and breast carcinoma cells with 100 M dec-RVKR-cmk, a furin inhibitor (44,45). Consistent with our previous report (41), activation of MT1-MMP is not affected by inhibition of furin in breast carcinoma and glioma cells. Following incubation with the inhibitor for 48 h, cells were surface-biotinylated, and the cell lysates were immunoprecipitated with anti-␣ v ␤ 3 mAb LM609. The inhibitor suppressed the furin-dependent cleavage of pro-␣ v and increased the relative levels of the 150-kDa pro-␣ v chain in control cells. Concomitantly, the levels of the 125-kDa mature heavy ␣-chain decreased in both control cell types, MCF7 and U251, co-incubated with the inhibitor (Fig. 1). In contrast, the inhibitor had no significant effect on the processing of pro-␣ v in both breast carcinoma and glioma cells coexpressing ␣ v ␤ 3 integrin and MT1-MMP. In these cells the furin inhibitor failed to block pro-␣ v maturation. These findings extend our observations 2 and suggest that the MT1-MMP-dependent pathway of pro-␣ v maturation is likely to exist in many cancer cell types, thereby representing a general phenomenon.
MT1-MMP Does Not Affect RGD Ligand Binding of ␣ v ␤ 3 Integrin-To study the functional implications of our findings, we analyzed whether the RGD ligand binding of ␣ v ␤ 3 integrin was affected in cells expressing MT1-MMP. For these purposes, the ligand binding ability of ␣ v ␤ 3 integrin expressed in ␤ 3 /zeo, ␤ 3 /MT-WT, and ␤ 3 /MT-E240A cells was assessed using a recombinant Fab fragment, Fab-9, designed as an RGD-ligand highly specific for ␤ 3 integrins (39,46). In another study, 2 we demonstrated that similar to anti-␣ v ␤ 3 mAb LM609 the chimeric ligand efficiently precipitated MT1-MMP-processed species of ␣ v ␤ 3 integrin.
To quantitatively confirm these observations further, we evaluated the ligand binding efficiency of 125 I-labeled Fab-9 to ␣ v ␤ 3 integrin expressed in ␤ 3 /zeo, ␤ 3 /MT-WT, and ␤ 3 /MT-E240A cells. Binding of ␣ v ␤ 3 integrin by Fab-9 was specific and cation-dependent, as addition of 5 mM EDTA completely abrogated Fab-9 binding (data not shown). Similar K D values of the Fab-9⅐␣ v ␤ 3 interactions were demonstrated for these cell types (Table I). There was no significant difference in the total number of Fab-9-binding sites/cell in ␤ 3 /zeo, ␤ 3 /MT-WT, and ␤ 3 /MT-E240A cells.
Furthermore, ␤ 3 /zeo, ␤ 3 /MT-WT, and ␤ 3 /MT-E240A cells were similarly efficient in short term adhesion to increasing concentrations of Fab-9 ( Fig. 2A). To elaborate further on the effect of MT1-MMP on RGD ligand binding capacity of ␣ v ␤ 3 integrin, we examined the attachment of ␤ 3 /zeo, ␤ 3 /MT-WT, and ␤ 3 /MT-E240A cells on plastic precoated with Fab-9 in the presence of increasing concentrations of the soluble cRGD peptide. Our observations indicate that the peptide was similarly efficient in inhibiting adhesion of all these cell lines to Fab-9 (Fig. 2B). Overall, these findings indicate that the MT1-MMPmediated processing of integrin ␣ v subunit does not significantly affect the RGD ligand binding characteristics of the resulting ␣ v ␤ 3 integrin.
Adhesion and Migration of Cells on Vitronectin-To evaluate the possible relevance of the MT1-MMP-␣ v ␤ 3 interactions on cell motility, we analyzed cell adhesion and migration on vitronectin, a natural ligand of ␣ v ␤ 3 integrin. First, we assessed the ability of cells to adhere to increasing concentrations of vitronectin. Consistent with our earlier reports (21), our current studies indicated that in a short term assay, ␤ 3 /MT-WT cells were more adhesive to vitronectin relative to ␤ 3 /zeo or ␤ 3 /MT-E240A cells (Fig. 3A). This difference in adhesion efficiency of the cells was especially pronounced at the relatively low concentrations of the ligand. However, all three cell lines were highly similar in their response to increasing concentrations of the cRGD peptide that competitively inhibited vitronectin-mediated attachment of cells (Fig. 3B). According to these observations, co-expression of MT1-MMP and ␣ v ␤ 3 integrin promoted cell adhesion to vitronectin without affecting the RGD binding ability of the integrin.
Furthermore, we assessed the migration efficiency of cells in Transwell chambers with the membrane undersurface precoated with vitronectin. As additional control, in these experiments we used neo/MT-WT cells that are deficient in ␣ v ␤ 3 integrin but express the wild type MT1-MMP. To determine the levels of MT1-MMP and ␣ v ␤ 3 integrin (Fig. 4A)   and ␤ 3 /MT-WT cells and largely existed as the 42-and 39-kDa stable autolytic forms. Because catalytically inactive MT-E240A is incapable of self-proteolysis (41), the individual 60-kDa full-length MT1-MMP form has been found in ␤ 3 /MT-E240A cells.
In agreement with our earlier observations, expression of MT1-MMP alone in neo/MT-WT cells failed to stimulate cell locomotion (20,21). In turn, co-expression of the wild type MT1-MMP with ␣ v ␤ 3 integrin in ␤ 3 /MT-WT cells strongly promoted cell migration on vitronectin. Concomitantly with the essential role of MT1-MMP activity in the pro-␣ v maturation, co-expression of the catalytically inactive MT-E240A construct with ␣ v ␤ 3 integrin in ␤ 3 /MT-E240A cells did not stimulate cell locomotion. In agreement, vitronectin-mediated migration of ␤ 3 /zeo cells and ␤ 3 /MT-E240A both expressing no proteolytically competent MT1-MMP was similarly low (Fig. 4B).
To support these data further, we preincubated cells for 24 -72 h with AG3340 (10 M) to inhibit MT1-MMP activity and, consequently, the MT1-MMP-dependent pathway of pro-␣ v maturation. AG3340 had no effect both on the pattern of ␣ v ␤ 3 integrin in ␤ 3 /zeo cells and migration of these cells (data not shown). As shown in Fig. 5A (lower panel), this treatment completely blocked the cellular MT1-MMP activity and, correspondingly, self-proteolysis of MT1-MMP. Accordingly, the fulllength MT1-MMP was the major species observed in the cells co-incubated with the inhibitor. This correlated with the inhibition of the MT1-MMP-dependent pathway of pro-␣ v maturation. Agreeably, preincubation of cells with AG3340 caused an increase in the levels of pro-␣ v and the 125-kDa heavy ␣-chain levels and a reciprocal decrease in the amounts of the 115-kDa heavy ␣-chain. In about 60 -70 h AG3340 restored the ␣ v pattern in ␤ 3 /MT-WT cells to that observed in the MT1-MMPnegative ␤ 3 /zeo cells (Fig. 5A, upper panel; Fig. 1A).
To elucidate whether inhibition of furin-and MT1-MMP-dependent pathways of pro-␣ v maturation affects cell mobility, we evaluated migration of cells pretreated with the furin inhibitor, the hydroxamate inhibitor, or both. As shown in Fig. 5B, preincubation of ␤ 3/ MT-WT cells with the hydroxamate inhibitor decreased in a time-dependent manner their migration efficiency to the levels characteristic to ␤ 3 /zeo and ␤ 3 /MT-E240A cells.
In contrast, preincubation for 48 h of ␤ 3 /MT-WT cells but not ␤ 3 /zeo cells with dec-RVKR-cmk (100 M), a furin inhibitor, strongly enhanced cell migration (Fig. 6). Co-addition of AG3340 reversed the stimulatory effects of the furin inhibitor on migration efficiency of ␤ 3 /MT-WT cells. Both inhibitors, individually or jointly, failed to modulate significantly migration of ␤ 3 /zeo cells (Fig. 6). Together, these results suggest that ␣ v ␤ 3 integrin processed by MT1-MMP is likely to be more efficient in vitronectin-mediated cell migration relative to the integrin processed by the conventional furin-like proprotein convertases.

MT1-MMP Stimulates Outside-in Signaling Mediated by
␣ v ␤ 3 Integrin-Increased migration of ␤ 3 /MT-WT cells could be associated with altered outside-in signaling through phosphorylation of the proteins downstream of the integrin-ligand complexes. In this regard, phosphorylation of FAK correlates with the maturation of integrins (47,48). Therefore, the processing of the ␣ v -chain by MT1-MMP could affect the levels of FAK phosphorylation induced by ␣ v ␤ 3 integrin ligation.
To exclude any possibility that MT1-MMP proteolysis of cell receptors other than ␣ v ␤ 3 integrin might cause stimulation of tyrosine phosphorylation of FAK, we compared FAK activation in cells expressing either ␣ v ␤ 3 integrin (␤ 3 /zeo) or MT1-MMP alone (neo/MT-WT) with that in the cells co-expressing ␣ v ␤ 3 integrin and MT1-MMP (␤ 3 /MT-WT). Expression of MT1-MMP alone was incapable of stimulating integrin-specific outside-in signaling associated with FAK activation, and the observed levels of tyrosine phosphorylation of FAK in neo/MT-WT cells were significantly lower as compared with those in ␤ 3 /zeo and especially ␤ 3 /MT-WT cells (Fig. 7A).
Furthermore, ␤ 3 /zeo, ␤ 3 /MT-WT, and ␤ 3 /MT-E240A cells were preincubated with or without dec-RVKR-cmk and then allowed to adhere to vitronectin. After incubation for 1 h, attached cells were lysed, and FAK was immunoprecipitated from the total cell lysates followed by Western blotting with anti-Tyr(P) mAb or anti-FAK antibody. In agreement with our hypothesis, ␤ 3 /MT-WT cells exhibited higher levels of tyrosine phosphorylation of FAK relative to those observed in ␤ 3 /zeo or ␤ 3 /MT-E240A cells (Fig. 7B). As expected, incubation with dec-RVKR-cmk decreased the levels of tyrosine phosphorylation of FAK in ␤ 3 /zeo cells. The content of Tyr(P) in FAK increased in ␤ 3 /MT-WT cells treated with the inhibitor (Fig. 7B). This pattern of phosphorylation indicated that cells co-expressing ␣ v ␤ 3 integrin with MT1-MMP could be more efficient in stimulating signal transduction through the FAK pathway.
To confirm further that MT1-MMP-dependent proteolysis of ␣ v ␤ 3 integrin could affect outside-in signaling through FAK activation, we analyzed tyrosine phosphorylation of FAK in ␤ 3 /zeo and ␤ 3 /MT-WT cells plated on vitronectin coated on plastic at increasing concentrations. In both cell types, FAK was phosphorylated in a dose-dependent manner (Fig. 7C). However, the content of Tyr(P) was significantly higher in ␤ 3 /MT-WT cells relative to that found in ␤ 3 /zeo cells. Given the MT1-MMP-dependent processing of ␣ v ␤ 3 integrin in ␤ 3 /MT-WT cells, these results provide further evidence that MT1-MMP is involved in the processes underlying the increased adhesion and migration of breast carcinoma cells on vitronectin. DISCUSSION Emerging evidence indicates that membrane-tethered MMPs such as MT1-MMP are directly involved in endoproteolytic modifications of cell surface receptors including CD44, tissue transglutaminase, and ␣ v ␤ 3 integrin (20,21,35,36). changing extracellular matrix environment. Our work supplements the existing knowledge and identifies a novel functional link between MT1-MMP and ␣ v ␤ 3 integrin in tumor cells. Our observations demonstrate for the first time that MT1-MMP exhibits integrin convertase activity and that MT1-MMP is able to specifically cleave pro-␣ v in tumor cells. In breast carcinoma MCF7 cells, the MT1-MMP cleavage appears to occur at two distinct sites localized within a loop between the disulfidebonded Cys 852 and Cys 904 of pro-␣ v . These cleavages generate a 115-kDa heavy ␣-chain and a light ␣Ϫchain commencing from the N-terminal Leu 892 . 2 In MT1-MMP-transfected glioma cells, which naturally express high levels of ␣ v ␤ 3 integrin (49,50), the proteinase is also capable of processing pro-␣ v . Thus, we suggest that the processing of ␣ v subunit of integrin by MT1-MMP may occur in cells of various tissue/organ origin.
MT1-MMP-mediated processing of pro-␣ v does not affect RGD-ligand binding of ␣ v ␤ 3 integrin. However, cells co-expressing integrin ␣ v ␤ 3 and MT1-MMP are more efficient relative to the cells expressing the integrin alone in stimulating outside-in signal transduction through the FAK pathway.
Adhesive function of integrins has been associated with the transduction of biochemical signals into the interior of the cell (17). Integrin ligation normally induces tyrosine phosphorylation through outside-in signaling and activation of cytoplasmic tyrosine kinases and, specifically, FAK (11,12). In many cell types, FAK, a known mediator of cell-matrix signaling events, appears to be the initial protein that becomes tyrosine-phosphorylated in response to integrin-mediated adhesion. FAK specifically modulates integrin-mediated cell migration (51). Thus, in the absence of endoproteolytic processing of pro-␣ v , the signaling function of ␣ v ␤ 5 integrin was impaired, and integrinmediated cell adhesion was suppressed (18). Because FAK is tyrosine-phosphorylated in almost all migratory cell types (47), it appears that activation of FAK is a prerequisite for efficient cell migration. It has been suggested that the highly conserved NPXY motif of the C-terminal segment of the ␤ 3 cytoplasmic domain is involved in tyrosine phosphorylation of FAK (13,52).
Expression of MT1-MMP alone in the cells deficient in ␣ v ␤ 3 integrin fails to stimulate tyrosine phosphorylation of FAK. This finding supports our suggestion that co-expression of ␣ v ␤ 3 integrin and MT1-MMP in cells is a prerequisite for the effects observed in our studies. Co-expression of MT1-MMP and ␣ v ␤ 3 integrin promoted cell adhesion and migration on vitronectin. These data are consistent with the recent observations that the increased strength of the ␣ v ␤ 3 -vitronectin interactions requires phosphorylation of the ␤ 3 subunit, intracellular signaling, and the binding of cytoskeletal proteins to cytoplasmic domains of the ␤ 3 subunit (53). In contrast to pro-␣ v processing, the activation pathway of MT1-MMP in breast carcinoma cells does not necessary involve furin (41,54). Concomitantly, the furin inhibitor specifically enhanced processing of pro-␣ v by MT1-MMP. This remarkably facilitated outside-in signal transduction via a FAK pathway and increased motility of cells coexpressing ␣ v ␤ 3 integrin and MT1-MMP.
We suggest that MT1-MMP may serve to activate pro-␣ v in processes and tissues characterized by MT1-MMP induction, e.g. aggressive tumor cells known to overexpress this protease activity. Because it has been demonstrated that in migrating cells MT1-MMP predominantly localize the invasive cell front and cell protrusions (55)(56)(57), we suggest that MT1-MMP works in concert with the integrins. To this end, the additional portions of ␣ v integrins generated from pro-␣ v via the MT1-MMPdependent pathway would co-localize the same cellular sites with the protease in migrating cells, specifically the invasive cell front and the invadopodia.
Evidently, there are more fully functional ␣ v ␤ 3 integrin molecules in MT1-MMP-expressing cells. Furthermore, our immunoprecipitation (Fig. 1A) and cell migration data (Fig. 6) indicate that MT1-MMP-processed species of ␣ v ␤ 3 integrin may be more effective in transmitting outside-in signals and, therefore, impart the cell with the elevated migratory capacity. This phenomenon is especially evident in the presence of the furin inhibitor when the furin-dependent maturation of the integrin is suppressed.
There is an attractive hypothesis that the mechanisms involving MT1-MMP may in a timely and spatially sensitive manner supply mature integrins from the cellular pro-␣ v pool, which has escaped the furin convertase-dependent maturation. The fine level of control exerted by MT1-MMP may allow aggressive migratory tumor cells to adjust their receptor profile to the extracellular matrix environment and, potentially, adds another layer of complexity to the control of integrin functionality.