Ovarian Carcinoma Regulation of Matrix Metalloproteinase-2 and Membrane Type 1 Matrix Metalloproteinase through β1 Integrin

INTRODUCTION

Ovarian carcinomas often metastasize through cancer cell implantation into the collagen-rich peritoneal submesothelial layer, followed by localized invasion and proliferation. Previous work has demonstrated that ovarian epithelial carcinomas preferentially adhere to type I collagen, thereby facilitating submesothelial interactions during metastasis (1) . Normal ovarian epithelial cells do not express MMPs, 3 and acquisition of their expression by ovarian carcinomas may promote more aggressive invasion (2) . MMPs associated with ovarian carcinomas include MMP-2 (M_r 72,000 gelatinase A), MMP-9 (M_r 92,000 gelatinase B), and the cell membrane-anchored MMP, MT1-MMP (or MMP-14; Ref. 3 ). MMP-2 is secreted as a zymogen and is found in complex with the protease inhibitor, TIMP-2 (4, 5, 6) . Extracellular proteolytic processing to a mature enzyme is an important control point in the regulation of MMP-2 activity (7) . MT1-MMP activates pro-MMP-2, and it is hypothesized that the zymogen interacts with MT1-MMP via a mutual TIMP-2 binding partner, forming a ternary complex on the cell surface (8, 9, 10) . MT1-MMP activity can be controlled by TIMP-2 and may involve cell surface aggregation of the enzyme (10 , 11) . MT1-MMP is also expressed as a zymogen and possesses a putative cleavage site between its pro and catalytic domains for furin-like enzymes, suggesting that MT1-MMP may be activated intracellularly before reaching the cell surface (7 , 12 , 13) . The mechanisms that regulate MT1-MMP posttranslational processing and its importance to enzyme activity and stability, however, remain to be elucidated.

Integrins function as transmembrane linkers between the extracellular matrix and the actin cytoskeleton. Integrin-mediated signaling regulates a variety of cellular events, including cell growth, differentiation, migration, invasiveness, and gene expression (14, 15, 16, 17) . In cultured cells, integrin activation results in the accumulation of signaling and cytoskeletal associated molecules at cell matrix contact sites called focal contacts. It has been reported that integrin occupancy by monomeric ligand leads to accumulation of integrins at preexisting focal contacts but does not promote integrin signaling (18 , 19) . Aggregation of integrin molecules, however, results in tyrosine kinase-mediated events that lead to the focal accumulation of signaling molecules, including those of the src kinase family (19) . Studies addressing matrix regulation of MMP-2 have demonstrated that culturing a variety of cells, including carcinomas, in three-dimensional type I collagen can stimulate pro-MMP-2 activation and expression of MT1-MMP (20, 21, 22, 23, 24) . Experimental stimulation of integrins with soluble antibodies, however, does not reproduce the effects of multidimensional collagen on MMP-2 activation, instead promoting pro-MMP-2 expression rather than activation (25, 26, 27) .

The ovarian carcinoma cell line DOV 13 adheres preferentially to type I collagen via the α2β1 integrin and expresses MMP-2 and MT1-MMP (1 , 28, 29, 30) . In this study, we observed that DOV 13 cells exhibit pro-MMP-2 processing activity when cultured in a three-dimensional type I collagen lattice. β1 integrin has been reported to mediate type I collagen induction of pro-MMP-2 processing in normal fibroblasts (24) . To address the mechanistic requirements of β1 integrin signaling in this event, we stimulated DOV 13 cells with soluble β1 integrin antibodies or β1 integrin antibodies immobilized on latex beads to promote integrin aggregation. We report here that β1 integrin-mediated signaling affects the expression, processing, and activity of both MMP-2 and MT1-MMP and that the nature of this regulation is dependent on whether the integrin is divalently engaged by a soluble antibody or aggregated on the cell surface through immobilized antibodies. These in vitro observations extend the hypothesis that matrix-induced occupation or aggregation of β1 integrins can, in turn, differentially regulate production and activity of MMPs.

MATERIALS AND METHODS

Materials.

BSA, gelatin, cell culture reagents, human type I collagen, human type IV collagen, aprotinin, APMA, N-hydroxy-succinimidobiotin, anti-(rabbit IgG)-alkaline phosphatase and peroxidase conjugates, BCIP/NBT, 2.97-μm-diameter latex beads, ConA, cycloheximide, and actinomycin D were purchased from Sigma Chemical Co. (St. Louis, MO). Genistein and anti-TIMP-2 mAb clone T2-101 was purchased from Calbiochem (Cambridge, MA). Antihuman β1 integrin mAb clone 21C8, antihuman αvβ3 integrin mAb clone LM609, and antihuman MT1-MMP polyclonal antibody (AB815, hinge domain) were obtained from Chemicon (Temecula, CA). Nitrocellulose-1 (0.45 μm) and the antihuman β1 integrin mAb clone P4C10 were purchased from Life Technologies, Inc. (Grand Island, NY). Hydrobond-P:PVDF membrane and pro-MMP-2 ELISA kit were obtained from Amersham (Arlington Heights, IL). SuperSignal enhanced chemiluminescence reagents were purchased from Pierce (Rockford, IL). The peptide Asp-Gly-Glu-Ala-Gly-Ala-Gly (DGEAGAG) was acquired from Genosys (The Woodlands, TX). Purified TIMP-2 and the hydroxamic acid MMP inhibitor 3-(N-hydroxycarbamoyl)-(R)-isobutylpropionyl-l-trytophan methylamide (GM-6001-X) were gifts of Dr. Hideaki Nagase, University of Kansas Medical School (Kansas City, KS).

Cell Culture.

The ovarian carcinoma cell line DOV 13 was generously provided by Dr. Robert Bast, Jr. (M. D. Anderson Cancer Center, Houston, TX). Cell culture was maintained under standard conditions in 75-cm² cell culture flasks (30) . Cells were plated confluent (1.5 × 10⁵ cells/well) in 24-well cluster plates (Becton Dickinson, Bedford, MA) overnight and incubated 1 h in serum-free medium prior to addition of fresh serum-free medium. Cells were treated with soluble antibodies (12 μg/ml), ConA (20 μg/ml), or antibody-adsorbed latex beads (8–10 μg/ml; 0.2% beads by final volume) for the times indicated below. All final volumes were 500 μl/well. Cells were incubated with immobilized antibodies for up to 24 h because longer incubations resulted in extensive endocytosis of antibody-coated beads.

Thin-Layer Collagen.

Twenty-four-well cluster plate chambers were coated with 200 μl of 10 μg/ml type I collagen dissolved in 100 mm sodium carbonate (pH 9.6). Chambers were incubated for 1 h at 4°C and then washed with sterile PBS. Collagen-coated chambers were air-dried before plating of cells.

Three-Dimensional Collagen.

Type I collagen (3 mg/ml in 0.1% acetic acid) was treated with 2 mm 1,10-phenathrolene for 1 h at 4°C, followed by dialysis against PBS. Collagen was diluted to 1.5 mg/ml with cold MEM containing 20 mm Hepes (pH 7.4), added to 24-well plates (200 μl), and allowed to gel at 37°C for 30 min prior to addition of cells (1 × 10⁵). Cells were incubated for 48–72 h in serum-free medium at 37°C before collection of CM.

Latex Bead-immobilized Antibody Preparation.

Antibodies were passively adsorbed onto 2.97-μm-diameter latex beads using the following modifications of the procedure described by Miyamoto et al. (16) . Latex beads were incubated at a final suspension of 1% in 50 mm 4-morpholineethanesulfonic acid buffer (pH 6.1) with 75 μg/ml appropriate antibody overnight at 4°C with gentle agitation before being blocked with 10 mg/ml BSA for 75–90 min at room temperature. Following blocking, beads were centrifuged for 3 min at 3000 rpm and then washed twice by resuspending with 2 volumes of serum-free medium. Following washing, antibody-adsorbed beads were resuspended at a final concentration of 1% by volume. Protein concentration determination of unblocked beads with a bicinchoninic acid detection kit (Sigma) indicated 60–70% of immunoglobulins were adsorbed, resulting in final concentration of 8–10 μg/ml mAb bead suspension in culture wells.

Gelatin Zymography.

Gelatinase activities in CM were determined using SDS-PAGE zymography as described previously (30) . SDS-polyacrylamide gels were prepared with 9% acrylamide and 0.1% gelatin, and samples were electrophoresed without reduction. All samples were analyzed in the absence of APMA activation unless otherwise indicated. Control samples were activated by preincubation with APMA (1.5 mm) for 1 h at 37°C prior to electrophoresis. Following removal of SDS through a 1-h incubation in 2.5% Triton X-100, the gels were incubated in 20 mm glycine, 10 mm CaCl₂, and 1 μm ZnCl₂ (pH 8.3) at 37°C for 24 h prior to staining for protein. The pro-MMP-2 protein levels in DOV 13 CM were determined by ELISA according to manufacturer’s protocol (Amersham).

Inhibitors.

GM-6001-X, actinomycin D, and genistein were dissolved in DMSO prior to use. Inhibitors were added to confluent culture wells 30 min prior to addition of immobilized antibodies. Cells were found to be >95% viable by exclusion of trypan blue dye at the highest concentration of all inhibitors tested. The highest concentration of genistein used in these experiments was slightly above the reported IC₅₀ of src tyrosine kinases but well below the IC₅₀ of nontyrosine kinases (31) .

Gelatin Hydrolysis.

Biotin-labeled gelatin was prepared by mixing 20 mg of type IV gelatin with 1 mg of N-hydroxy-succinimidobiotin in 10 ml of PBS and incubating on a rotator for 1 h at room temperature. Free biotin was removed through extensive dialysis against 20 mm Hepes (pH 7.4). Biotinylated gelatin was added directly to culture wells at a final concentration of 500 ng/ml. Following incubation, aliquots of CM were resolved by 7% SDS-PAGE electrophoresis under nonreducing conditions and transferred to PVDF membranes. Membranes were blocked with 3% BSA in Tris-buffered saline (pH 7.5) with 0.1% Tween 20 (TBST). Biotin-tagged gelatin fragments were detected with streptavidin peroxidase polymer (200 ng/ml in 3% BSA) and enhanced chemiluminescence (SuperSignal; Pierce).

TIMP-2 Analysis.

CM was diluted in TBS and added to nitrocellulose using a slot blot apparatus (Bio-Rad). Nitrocellulose blots were washed in TBST, blocked with 3% BSA/TBST, and incubated at room temperature with 2 μg/ml anti-TIMP-2 mAb (clone T2-101) in 3% BSA/TBST followed by anti-(mouse-IgG)-alkaline phosphatase (1:5000 in 3% BSA/TBST) and BCIP/NBT detection.

MT1-MMP Immunoblots.

Following incubation of cells under various conditions, cells were lysed with a lysis buffer [50 mm Tris, 150 mm NaCl, 1 mm CaCl₂, 1 mm MgCl₂, 1% NP40, 10 ng/ml benzamadine, 1 mm phenylmethylsulfonyl fluoride, and 1 μm aprotinin (pH 7.5)], and protein concentration of lysates was analyzed using the bicinchoninic acid detection kit (Sigma). Cell lysate (5–15 μg) was electrophoresed on 9% SDS-polyacrylamide gels under reducing conditions, transferred to PVDF membrane, and blocked with 3% BSA in TBST (32 , 33) . Membranes were incubated 1 h at room temperature in a 1:4000 dilution of antihuman MT1-MMP polyclonal antibody in 3% BSA in TBST. Immunoreactive bands were visualized with an anti-(rabbit-IgG)-alkaline phosphatase (1:5000 in 3% BSA/TBST) and BCIP/NBT detection or anti-(rabbit-IgG)-peroxidase (1:15,000 in 3% BSA/TBST) and enhanced chemiluminescence. Recovery of cells from collagen gels for Western analysis required treatment with bacterial collagenase. In control experiments, we observed that bacterial collagenase very efficiently removed MT1-MMP from the surface of DOV 13 cells, so that it was impossible to assess MT1-MMP processing in this system.

RESULTS AND DISCUSSION

Three-Dimensional Collagen Culture Stimulation of pro-MMP-2 Processing.

Previous work has shown that a three-dimensional collagen matrix can stimulate pro-MMP-2 activation in a variety of cell lines (20, 21, 22, 23, 24) . Because type I collagen is a prevalent matrix encountered by invasive ovarian carcinomas, the effect of DOV 13 interaction with a collagen matrix on pro-MMP-2 processing was determined. Equal numbers of DOV 13 cells were cultured on tissue culture plastic wells coated with a thin layer of type I collagen or on three-dimensional collagen lattices, and the collected CM was analyzed by gelatin zymography (Fig. 1A) ⇓ . Culturing on three-dimensional collagen stimulated processing of pro-MMP-2 to a M_r 62,000 form (Fig. 1A ⇓ , Lanes 5 and 6), comparable in electrophoretic migration to the APMA-activated species (Fig. 1A ⇓ , Lane 1). In time course experiments, maximal induction of pro-MMP-2 processing by three-dimensional collagen was observed after 72 h. Incubation of DOV 13 cells in three-dimensional collagen with the MMP inhibitor GM-6001-X at 20 μm effectively blocked pro-MMP-2 processing activity, suggesting that peri/extracellular MMP activity is required (data not shown). Culturing DOV 13 cells on a thin layer of type I collagen (Fig. 1A ⇓ , Lanes 3 and 4) or on other matrix proteins, including type IV collagen, laminin, fibronectin, vitronectin (1) , and Matrigel (data not shown), had no effect on pro-MMP-2 processing, suggesting that this event is specific to β1 integrin-mediated collagen binding. Monomeric occupation of β1 integrins with soluble DGEAGAG peptide (500 μm; Ref. 34 ) was also insufficient to promote pro-MMP-2 processing. Together, these results suggest that type I collagen can influence pro-MMP-2 processing by DOV 13 cells and that an intact three-dimensional collagen is required for this event.

Effects of β1 Integrin Divalent Ligation or Aggregation on pro-MMP-2 Processing and TIMP-2 Expression.

Although requirements for integrin signaling are not well characterized, studies have shown that integrin binding events that result in receptor occupancy, receptor aggregation, or both (occupancy and aggregation) mediate distinct subsets of cellular events that are differentiated based on the physical nature of the specific ligand-receptor interaction (16 , 19) . A recent study using confocal microscopy provided direct evidence that cellular interaction with a three-dimensional network of type I collagen fibrils results in aggregation of β1 integrins (35) . The observation that collagen stimulation of pro-MMP-2 processing requires an intact three-dimensional collagen matrix (Fig. 1A) ⇓ suggests that β1 integrin aggregation may be involved in pro-MMP-2 regulation. To address this hypothesis, DOV 13 carcinoma cells were treated with soluble β1 integrin antibodies to stimulate β1 integrin signaling or β1 integrin antibodies immobilized on latex beads to promote additional multivalent integrin aggregation. Culturing DOV 13 cells with soluble mAb clone 21C8, an adhesion-stimulating antibody that has been reported to activate p21^ras in lymphoblasts (36) , had no effect on pro-MMP-2 levels (Fig. 1B ⇓ , Lane 2). Addition of mAb clone P4C10, an adhesion-blocking antibody (37) , resulted in a modest but reproducible increase in pro-MMP-2 levels (Fig. 1B ⇓ , Lane 3), which was confirmed to be 1.5-fold by an ELISA specific for pro-MMP-2 (112 and 172 ng/ml in IgG and P4C10 mAb-treated CM, respectively). In contrast to divalent ligation, aggregation of β1 integrins with either antibody clone promoted pro-MMP-2 processing to the M_r 62,000 form (Fig. 1B ⇓ , Lanes 6 and 7), with processing detected as early as 6 h (data not shown). Aggregation of αVβ3 integrins had no effect on pro-MMP-2 processing. Unlike P4C10 antibodies, 21C8 antibodies do not mimic ligand binding, suggesting that aggregation of integrins in the absence of occupation is sufficient to stimulate pro-MMP-2 processing. Together, these results support the hypothesis that β1 integrin aggregation by three-dimensional collagen may function as a mechanism for matrix-induced pro-MMP-2 processing. In support of these findings, it has recently been reported that culturing HT1080 fibrosarcoma cells on β1 integrin antibodies immobilized on tissue culture plastic increased pro-MMP-2 processing (38) . It is interesting to speculate that proteinase processing and/or activity is controlled by the structural integrity of the collagen matrix, such that production of matrix-degrading proteinases by ovarian carcinoma cells results from integrin clustering by intact matrix and is subsequently down-regulated following proteolysis of matrix into fragments incapable of promoting integrin aggregation (39) . In control experiments, treatment of DOV 13 cells with ConA, a stimulator of MMP-2 activation through MT1-MMP, resulted in significant pro-MMP-2 processing to the M_r 62,000 form (Fig. 1B ⇓ , Lane 8; Ref. 40 ).

To assess whether pro-MMP-2 processing correlated with increased gelatinase activity, we incubated cultured cells with biotin-labeled gelatin, and degradation was monitored by Western blotting with streptavidin. Aggregation of β1 integrins resulted in an increase in gelatinase activity (Fig. 1C ⇓ , Lanes 1 and 2), which was blocked by the MMP inhibitor GM-6001-X (Fig. 1C ⇓ , Lane 3), providing evidence that pro-MMP-2 processing resulted in pro-MMP-2 activation. In control experiments, treatment of DOV 13 cells with ConA resulted in robust gelatinase activity that was blocked by GM-6001-X (Fig. 1C ⇓ , Lanes 5 and 6). Divalent ligation of β1 integrins with soluble antibodies did not promote changes in gelatinase activity.

To assess the effects β1 integrin ligation or aggregation on TIMP-2 expression, we analyzed TIMP-2 levels in CM by slot blotting. Divalent ligation of β1 integrins with soluble P4C10 antibodies significantly increased the level of TIMP-2 in CM, whereas β1 integrin aggregation or divalent ligation with 21C8 antibodies appeared consistent with IgG control (Fig. 2) ⇓ . In agreement with previous work, treatment of cells with ConA resulted in a depression of TIMP-2 levels in CM (40) . Although TIMP-2 is required for MT1-MMP-mediated pro-MMP-2 activation (10, 11, 12) , excess TIMP-2 blocks cell surface MT1-MMP activity. These data indicate that down-regulation of TIMP-2 expression may facilitate the induction of pro-MMP-2 activation by ConA through alleviating TIMP-2 inhibition of MT1-MMP on the cell surface. At the same time, TIMP-2 levels were not significantly altered by β1 integrin aggregation, suggesting that integrin stimulation of pro-MMP-2 activation may work through other mechanisms and/or sufficient inhibitor-free MT1-MMP exists on the cell surface.

Inhibition of β1 Integrin Induced pro-MMP-2 Activation.

To obtain preliminary information on the mechanisms by which aggregation of β1 integrins regulate pro-MMP-2 activation, we used inhibitors of translation, transcription, and tyrosine kinase activity. Preincubation with as little as 10 ng/ml of the protein translation inhibitor cycloheximide prior to treatment with immobilized 21C8 antibody-blocked integrin stimulation of pro-MMP-2 activation (Fig. 3A ⇓ , Lanes 6–9). Incubation with 200 ng/ml transcription-blocking agent actinomycin D was required before a slight decrease in activation was observed (Fig. 3B ⇓ , compare Lanes 6 and 9), and activation was never completely abolished, even at the highest concentration tested (500 ng/ml; data not shown). These observation are in agreement with previous work that demonstrated that stimulation of pro-MMP-2 activation in MDA-MB-231 human breast cancer involves protein synthesis and is only partially dependent on gene transcription (41) .

Tyrosine kinase activity is required for the recruitment of signaling molecules to focal complexes and complex-mediated signal transduction (19) . To evaluate whether β1 integrin-stimulated pro-MMP-2 activation involves tyrosine kinase signaling, DOV 13 cells were treated with the tyrosine kinase inhibitor genistein (30) . Preincubation of cells with 10 μg/ml of genistein was sufficient to decrease pro-MMP-2 processing, with 20 μg/ml genistein blocking the majority of activation (Fig. 3C ⇓ , Lanes 6–9). Similar results were obtained with immobilized P4C10 antibody (data not shown). These results indicate that tyrosine kinase signaling is involved in β1 integrin stimulated pro-MMP-2 activation and are consistent with hypothesis that tyrosine kinase activity is required for integrin-aggregated signaling.

Effects of β1 Integrin Divalent Ligation or Aggregation on MT1-MMP.

MT1-MMP has been demonstrated to be a cell surface activator of pro-MMP-2, catalyzing conversion of the M_r 66,000 zymogen to the active form (8) . To determine whether induction of pro-MMP-2 activation reflects changes in MT1-MMP expression, we analyzed cell lysates by immunoblotting with an antibody to the MT1-MMP hinge domain. Immunoblots of DOV 13 cell extracts demonstrated an immunoreactive band of M_r 66,000, comparable to the reported weight of pro-MT1-MMP (Fig. 4 ⇓ , Lanes 1–7). Aggregation of β1 integrins resulted in the accumulation of a M_r 43,000 processed form of MT1-MMP (Fig. 4 ⇓ , Lanes 5 and 6). In contrast, soluble β1 integrin antibodies, which do not initiate cellular activation of pro-MMP-2, failed to stimulate pro-MT1-MMP processing (Fig. 4 ⇓ , Lanes 2 and 3). ConA stimulation of DOV 13 cells resulted in accumulation of the M_r 43,000 processed form as well as a M_r 55,000 intermediate form of MT1-MMP in cell extracts (Fig. 4 ⇓ , Lane 7). It has been reported that phorbol 12-myristate 13-acetate stimulation of HT1080 fibrosarcoma cells induces processing of MT1-MMP to M_r 60,000 and 43,000 forms that correlate with pro-MMP-2 processing activity (38 , 42 , 43) . The generation of a M_r 43,000 processed form of MT1-MMP in both fibrosarcoma and ovarian carcinoma cell lines supports the hypothesis that this form may be involved in a general mechanism of MT1-MMP regulation. Analysis of CM for soluble forms of MT1-MMP failed to detect any fragment of MT1-MMP containing the hinge domain recognized by the MT1-MMP polyclonal antibody.

Effect of Protease Inhibitors on MMP Processing.

To address whether pro-MMP-2 activation requires extracellular serine or metalloproteinase activity, DOV 13 cells were treated with various concentrations of the serine protease inhibitor aprotinin or the MMP inhibitor GM-6001-X. β1 integrin-mediated pro-MMP-2 activation was abolished by addition of 100 nm GM-6001-X to culture wells (Fig. 5A ⇓ , Lanes 2–4), whereas 5 μg/ml aprotinin had no effect (data not shown). In control cells, addition of 100 nm GM-6001-X decreased ∼50% of ConA-driven pro-MMP-2 activation, whereas 1 μm GM-6001-X blocked all processing activity (Fig. 5A ⇓ , Lanes 5–7). Aprotinin had no effect on ConA stimulated pro-MMP-2 processing (data not shown). These results are consistent with the conclusion that β1 integrin regulation of pro-MMP-2 activation involves MT1-MMP or another MMP.

To evaluate whether MT1-MMP processing involves extracellular serine or metalloproteinase activity, we analyzed lysates of DOV 13 cells treated with aprotinin or GM-6001-X under various conditions by immunoblotting. Aprotinin did not affect MT1-MMP processing in cell extracts of β1 integrin aggregated cells, whereas GM-6001-X depressed MT1-MMP processing to the M_r 43,000 form (Fig. 5B ⇓ , Lanes 2–4). In ConA-treated cells, aprotinin had no effect on MT1-MMP processing, whereas GM-6001-X resulted in loss of the M_r 43,000 form of MT1-MMP and accumulation of the M_r 55,000 form of the enzyme in cell extracts (Fig. 5 ⇓ , Lanes 5–7). Addition of exogenous TIMP-2 to ConA-stimulated cells also blocked pro-MMP-2 activation (Fig. 6A) ⇓ and MT1-MMP processing to the M_r 43,000 form with an accumulation of the M_r 55,000 intermediate (Fig. 6B) ⇓ . These data suggest that the M_r 43,000 form is a product of MMP-dependent M_r 55,000 MT1-MMP proteolysis on the cell surface and is linked to pro-MMP-2 activation.

To address the time course of MT1-MMP processing in relation to pro-MMP-2 activation, we collected cell lysates at various times after ConA stimulation and immunoblotted for MT1-MMP (Fig. 6B) ⇓ , whereas CM was analyzed for pro-MMP-2 activation by gelatin zymography (Fig. 6A) ⇓ . Presence of the M_r 55,000 form of MT1-MMP strongly correlated with pro-MMP-2 activation, whereas significant changes in the amount of M_r 43,000 MT1-MMP were only observed in lysates collected at the 24-h time point. These data support the hypothesis of Lehti et al. (38) , which suggests that an intermediate processed specie of MT1-MMP is the active pro-MMP-2 processing enzyme on the cell surface, whereas the M_r 43,000 species is a byproduct of MMP dependent proteolysis and likely to be inactive (43) . This work indicates that the M_r 55,000 processed form of MT1-MMP is the active enzyme on the surface of DOV 13 carcinoma cells and may be the functional equivalent to the M_r 60,000 intermediate expressed by HT1080 fibrosarcoma cells. Furthermore, the presence of M_r 43,000 MT1-MMP in β1 integrin aggregated DOV13 extracts supports the conclusion that β1 integrin stimulated MMP-2 activation proceeds via the M_r 55,000 MT1 intermediate.

On the basis of the data summarized above, we propose the following model to illustrate the functional link between α2β1-mediated adhesion and MMP proteolysis in ovarian carcinoma cells (Fig. 7) ⇓ . Ovarian carcinoma engagement of an intact collagen matrix promotes aggregation of α2β1 integrin receptors (Fig. 7A ⇓ , [a]), resulting in the accumulation of active M_r 55,000 MT1-MMP through tyrosine kinase and translation-dependent events. M_r 55,000 MT1-MMP activates pro-MMP-2, resulting in enhanced cell surface MMP activity (Fig. 7A ⇓ , [b]). As pericellular MMP activity increases, it is down-regulated by two mechanisms. MT1-MMP and MMP-2 have been reported to degrade type I collagen and gelatin, respectively (4 , 5 , 44 , 45) . MMP-mediated proteolysis of the collagen-rich matrix decreases β1 integrin aggregation (Fig. 7B ⇓ , [c]), thereby disrupting signaling pathways required for induction of MMP activation (Fig. 7B ⇓ , [d]). Enhanced proteolytic activity also leads to MMP-dependent processing of M_r 55,000 MT1-MMP to the inactive M_r 43,000 form (43) , resulting in decreased pro-MMP-2 processing potential and MT1-MMP mediated matrix degradation (Fig. 7B ⇓ , [e]). As increased surface MMP activity correlates with enhanced invasive behavior of ovarian carcinoma cells (29) , our data suggest that matrix regulation of MMP-2 and MT1-MMP processing may function as a biochemical mechanism for the control of ovarian carcinoma invasion.