Regulation of Tenascin-C, a Vascular Smooth Muscle Cell Survival Factor that Interacts with the α_vβ₃ Integrin to Promote Epidermal Growth Factor Receptor Phosphorylation and Growth

Vascular SMCs were isolated from pulmonary arteries of 300-g adult male Sprague-Dawley rats. Briefly, arteries were harvested from anaesthetized animals and placed in cold sterile PBS, pH 7.6. Endothelium was removed by gently scraping the lumenal surface with a scalpel blade, and the adventitia was also removed from the vessel. The medial layer was minced using scalpel blades and incubated at 37°C in medium 199 (M199; GIBCO BRL, Gaithersburg, MD) supplemented with 0.1% collagenase I (Sigma Chemical Co., St. Louis, MO.), and 0.1% BSA (Boehringer-Mannheim, Laval Canada) for 1 h with gentle rotation. Tissue was collected by centrifugation at 300 g, resuspended in M199/collagenase I solution, and incubated overnight at 37°C with gentle rotation. Smooth muscle cells were harvested by centrifugation and were routinely maintained in M199 containing 10% heat-inactivated FBS (Intergen, Purchase, NY), 10 U/ml penicillin G sodium, 10 mg/ml streptomycin sulfate, 0.25 mg/ml amphotericin B, and 0.1 mg/ml gentamicin sulfate (GIBCO BRL). Cells were passaged by trypsinization using 0.05% trypsin/EDTA (GIBCO BRL). Vascular SMCs were identified by their characteristic hills-and-valley morphology and immunohistochemical staining for α-smooth muscle actin.

Collagen gels were prepared based on methods that have been previously described (Jones and Rabinovitch., 1996). Briefly, 0.8 ml of a 3.1 mg/ ml solution of bovine dermal type I collagen (Vitrogen 100; Collagen Corp., CA), 0.1 ml of 0.1 M NaOH, and 0.1 ml of 10× PBS were mixed at 4°C for a final collagen concentration of 2.48 mg/ml. To determine the effect of TN-C on SMC behavior, neutralized collagen was supplemented with 15 μg/ml of human TN-C protein (Chemicon International, Temecula, CA). Fibrillogenesis was initiated overnight in a humid 5% CO₂ environment at 37°C, and the gels were rinsed extensively (three times for 3 h) with M199 containing 0.1% BSA before use. Confluent cultures of rat PA SMCs were serum starved in M199/0.1% BSA for 48 h. Smooth muscle cells were collected by trypsinization and centrifugation, and cell number was determined using an improved Neubauer hemocytometer (American Optical, Buffalo, NY). Cell pellets were resuspended in M199/0.1% BSA plus 0.5% serum, and a 1-ml aliquot of SMCs was seeded onto the surface of each gel. 6 h later, cells were rinsed in M199 and were thereafter cultured in M199/0.1% BSA.

We used collagen gels and heat-denatured collagen to determine the interrelationships between TN-C, MMP activity, β₃ integrins, and EGF-dependent SMC survival and growth. For these experiments, 2 × 10⁴ cells were plated on collagen gels as described above. For anti-α_vβ₃ antibody blocking studies, SMCs were plated on collagen gels in the presence of 15 μg/ml LM609 monoclonal antibody (a kind gift from Dr. David Cheresh, Scripps Research Institute, La Jolla, CA) or with 15 μg/ml of control IgG (DAKO Corp., Carpenteria, CA). For β₃ integrin studies on native and denatured collagen, SMCs were plated in the presence of 25 μg/ml anti-β3 monoclonal antibody (PharMingen, San Diego, CA) or with 25 μg/ml of control IgG (DAKO Corp.). For MMP inhibitor studies, 1–2 μm GM6001 (a kind gift from Dr. Stu Sweidler, Glycomed, Alameda, CA) in DMSO, or an equivalent volume of DMSO alone, was added to cultures 24 h after plating. GM6001 (N- [2 (R)  -  2  - ( hydroxamidocarbonylmethyl )  - 4 -  methylpentanoyl]  - l  - tryptophane methylamide) is a noncytotoxic synthetic inhibitor that is specific for MMPs. The hydroxamic group of GM6001 binds to the critical active site zinc atom present in MMPs. In addition, the isobutyl group and tryptophan side chain of GM6001 also binds to subsites on MMPs, which normally bind side chains of ECM proteins (Galardy, 1993; Boudreau et al., 1995a; Strauss et al., 1996). For assessment of the effects of heat denatured collagen, cells were plated on 50 μg of type I collagen/35-mm-diam dish that had been boiled for 30 min with 0.02 M acetic acid, and then air dried to the bottom of each dish. Before plating cells, collagen substrates were rinsed extensively, (three times for 1 h) with M199/0.1% BSA. To “float” the collagen gels, we detached them from the plastic substratum 24 h after plating using a spatula. 24 h after plating, cells were cultured in M199/ 0.1% BSA either with or without 50 ng/ml EGF (GIBCO BRL). The number of cells retained on collagen gels was determined 72 h after plating. The dose of EGF chosen to be stimulatory to SMC growth was determined in pilot studies and previously published by our group (Jones and Rabinovitch, 1996). All experiments were performed in triplicate. Attachment efficiencies were determined by counting the number of cells in the medium using a hemocytometer or Coulter counter. The number of attached cells (number of cells seeded − number of cells in the medium) was expressed as a percentage of the number of cells seeded. To assess the number of cells retained on collagen, gels were digested with 1 mg/ml collagenase type II (Sigma Chemical Co.) for 1 h at 37°C. Free cells were pelleted by centrifugation at 4°C for 10 min at 300 g and suspended in PBS containing 0.05% trypsin, and cell number was determined by counting aliquots in triplicate using a hemocytometer.

Conditioned medium was collected from cells maintained on attached or floating collagen gels for 4, 8, 12, and 24 h in M199 with 50 ng/ml EGF. For each culture and at every time point, SMCs were rinsed gently three times with culture medium and then reincubated with 1 ml of fresh medium. At the designated time points, conditioned medium was harvested from the cultures, and equal volumes of medium were lyophilized overnight. Conditioned medium was mixed 3:1 (vol/vol) with 4× sample buffer (10% SDS, 25% glycerol in 150 mM Tris-Cl, pH 6.8) and separated on nonreducing 10% polyacrylamide gels that contained 0.1% gelatin (Sigma Chemical Co.). After electrophoresis, gels were soaked in 2.5% Triton X-100 to remove SDS and incubated in substrate buffer (50 mM Tris-Cl, pH 7.5, 5 mM CaCl₂) for 18 h at 37°C. The gels were then stained with Coomassie blue R250 (Bio-Rad Labs, Hercules, CA). Gelatinases appear as a clear zone on a blue background. To demonstrate the ability of GM6001 to inhibit MMPs, we incubated gelatin zymograms in substrate buffer containing 2 μM of this inhibitor. This treatment abolished MMP activity. For detection of the pro-form of MMP-2, 20 μg of total protein contained in conditioned medium harvested from SMCs cultured on attached and floating collagen for 24 h in serum-free medium (SFM) was subjected to SDS-PAGE electrophoresis. Proteins were electrophoretically transferred to nylon membranes, blocked for 1 h at 37°C in wash buffer (10 mM Tris, pH 7.5, 10 mM NaCl, and 0.1% Tween-20), containing 0.1% BSA and 1% normal goat serum (Sigma Chemical Co.), and then with an anti–MMP-2 rabbit polyclonal antisera (a kind gift from Drs. M. Silverman and M. Ailenburg, Department of Medicine, University of Toronto, Toronto, Canada) diluted 1:200 in wash buffer for 30 min at 37°C. To detect MMP-2, membranes were then incubated with a horseradish peroxidase–conjugated goat anti–rabbit secondary antibody (GIBCO BRL) diluted 1:3,000 in wash buffer, rinsed in wash buffer three times for 15 min, visualized by enhanced chemiluminescence (ECL kit; Amershan Corp., Arlington Heights, IL), and exposed to film (X-Omat; Kodak, Rochester, NY).

For immunofluorescent studies, serum starved SMCs were cultured on glass coverslips (12 mm, No.1 thickness) coated with 2.48 mg/ml of neutralized type I collagen (+/− 15 μg/ml TN) in M199/0.1% BSA. For actin, EGF-Rs, and tyrosine-phosphorylated proteins, cells were exposed to 50 ng/ml EGF in M199/0.1% BSA for 30 min or subjected to control treatment with M199/0.1% BSA alone. To determine the effects of blocking the β₃ integrin on EGF-R distribution, SMCs were preincubated with 25 μg/ml anti–rat β₃ integrin monoclonal antibody (CD61; PharMingen) before plating on collagen- and TN-C–enriched substrates. At the appropriate time point, cells were rinsed in PBS and then fixed. For detection of F-actin, SMCs were fixed with extraction buffer (60 mM Pipes, 25 mM Hepes, 10 mM EDTA, 2 mM MgCl₂, pH 6.9) supplemented with 0.2% Triton X-100 for 5 min at 4°C. Smooth muscle cells were then washed three more times for 5 min in extraction buffer without Triton X-100. Fixed cells were incubated with 2.5 U of rhodamine-phalloidin (Molecular Probes, OR) in PBS for 20 min at 37°C before rinsing briefly with PBS. To detect EGF-Rs, tyrosine-phosphorylated proteins, and vinculin, SMCs were fixed in 2% paraformaldehyde (Sigma Chemical Co.) for 10 min at room temperature and rinsed in PBS containing 0.1 M glycine for 30 min. For tyrosine-phosphorylated proteins and vinculin, SMCs were permeabilized with 0.2% Triton X-100 in PBS for 5 min and washed three more times. The fixed cells were incubated for 1 h at 37°C in wash buffer (PBS/ 1% BSA), supplemented with 10% normal goat serum for EGF-Rs, and then incubated for 1 h at 37°C with anti-EGF-R antisera (RK-2, diluted 1:100; a kind gift from Dr. Irit Lax, New York University Medical Center, New York), antiphosphotyrosine monoclonal antibody (clone 4G10; 5 μg/ ml; Upstate Biotechnology Inc., Lake Placid, NY), antivinculin monoclonal antibody (diluted 1:50; Sigma Chemical Co.), or appropriate preimmune sera and IgG controls in wash buffer. After being washed three times, coverslips were incubated for 1 h at 37°C either with fluorescein-conjugated goat anti–rabbit antibody (diluted 1:100; Sigma Chemical Co.) or with goat anti–mouse antibody (diluted 1:100; Sigma Chemical Co.). For double immunofluorescent staining experiments, all cells were fixed in 2% paraformaldehyde. For vinculin and EGF-Rs, cells were permeabilized in methanol for 2 min and were then air dried for 1 h at room temperature. For vinculin, tyrosine-phosphorylated proteins, and F-actin, fixed cells were washed in PBS containing 0.2% Triton X-100. All cells were then washed with PBS/0.1 M glycine and were sequentially incubated with appropriate antisera, antibodies, or rhodamine-phalloidin as described above. After washing, all coverslips were mounted onto glass slides using Antifade reagent (Molecular Probes). Observations and photomicrographs were obtained with a fluorescent microscope (Olympus Corp., Lake Success, NY) using epifluorescence.

Tyrosine-phosphorylated proteins were analyzed by immunoblotting whole cell lysates, SMC membrane–enriched protein fractions, or EGF-R immunoprecipitates with a horseradish peroxidase–conjugated antiphosphotyrosine (RC20H; Transduction Labs, Lexington, KY). 4 × 10⁶ SMCs cultured for 24 h on 10-cm-diam collagen gels (+/− TN-C) were treated with EGF (50 ng/ml) for 10 and 30 min or were further incubated in M199/ 0.1% BSA alone. Tenascin-C protein expression on collagen gels was assessed by immunoblotting equal numbers (2 × 10⁴) of cell lysates with an anti–chicken TN-C rabbit polyclonal antisera (pK7; gift from Dr. M. Schachner, Swiss Federal Institute of Technology, Zurich, Switzerland). To determine the effect of the β₃ integrin on TN-C–dependent EGF-R phosphorylation, SMCs were cultured on TN-C–supplemented collagen gels with 25 μg/ml of anti–rat β₃ integrin monoclonal antibody (CD61; PharMingen) or with control IgG (DAKO Corp.).

To enrich for membrane proteins, SMCs were harvested at the appropriate time points in ice-cold buffer containing 50 mM Hepes, 1 mM benzamidine, 1 mM PMSF, 1 μg/ml aprotinin, 5 μg/ml leupeptin, 2 mM NaF, 2 mM Na₃VO₄, 1 mM MgCl₂, and 0.25 M sucrose, pH 7.4 at 4°C. Cells were homogenized with a Dounce B glass homogenizer (60 strokes), and intact cells and nuclei were sedimented by centrifugation at 3,000 g for 5 min at 4°C. The resulting supernatants were centrifuged at 100,000 g for 1 h at 4°C, and the membrane-enriched pellet was solubilized by sonication in homogenization buffer supplemented with 1% Triton X-100. Protein content was determined using the Bio-Rad protein assay according to the manufacturer's instructions. For EGF-R immunoprecipitation experiments, 15 μg of membrane-enriched protein, resuspended in 0.5 ml RIPA lysis buffer (25 mM Tris, pH 7.2, 0.1% SDS, 1% Triton X-100, 1% sodium deoxycholate, 150 mM NaCl, and 1 mM EDTA) containing 1 mM benzamidine, 1 mM PMSF, 1 μg/ml aprotinin, 5 μg/ml leupeptin, 2 mM NaF, and 2 mM Na₃VO₄ was precleared with normal rabbit IgG and protein A–Sepharose (Sigma Chemical Co.) for 1 h at 4°C. Supernatants were then incubated for 1 h at 4°C with the rabbit polyclonal anti–EGF-R antibody diluted 1:400 in RIPA buffer and precipitated with protein A–Sepharose. Pellets were washed three times with ice-cold RIPA buffer and once in 50 mM Tris, pH 7.4. Membrane fractions (15 μg) and EGF-R immunoprecipitates were boiled in reducing sample buffer containing 0.2 M Tris-HCl, pH 8.8, 18% glycerol, 4% SDS, 0.01% bromophenol blue, and 10% β-mercaptoethanol. All samples were separated in 7% SDS–polyacrylamide gels. Proteins were transferred to Immobilon polyvinyl fluoride membranes (Millipore Corp., Milford, MA), and these were blocked for 1 h at 37°C in wash buffer (10 mM Tris, pH 7.5, 10 mM NaCl, and 0.1% Tween-20), supplemented with 0.1% BSA. To detect tyrosine-phosphorylated proteins, blots were incubated for 30 min with horseradish peroxidase–conjugated antiphosphotyrosine antibody, diluted 1:2,500 in wash buffer. To detect TN-C protein, membranes were sequentially incubated with anti–TN-C antisera and a horseradish peroxidase–conjugated goat anti–rabbit secondary antibody (GIBCO BRL) diluted 1:5,000 in wash buffer. Thereafter, all membranes were rinsed in wash buffer three times for 15 min, and tyrosine-phosphorylated and TN-C proteins were visualized by enhanced chemiluminescence (Amersham ECL kit) before exposure to film (X-Omat; Kodak).

To cluster β₃ integrins, SMCs were first cultured on collagen gels and were then washed gently with medium and incubated with 150 μg/ml anti-β₃ monoclonal antibody (PharMingen) diluted in M199/0.1% BSA for 60 min. The SMCs were then rinsed with medium and incubated either with 20 μg/ml of goat anti–mouse IgG F(ab′)₂ fragment (ICN Biomedicals, Inc., Costa Mesa, CA) or with medium alone for 60 min at 37°C. For tyrosine-phosphorylation studies, SMCs were rinsed in medium and treated with 50 ng/ml EGF for 10 or 30 min. Control cultures were maintained in M199/0.1% BSA. Epidermal growth factor receptor immunoprecipitates were analyzed by immunoblotting with antiphosphotyrosine antibodies as described above.

In these studies, we also used immunofluorescence to demonstrate the clustering of β₃ integrins. Accordingly, SMCs were rinsed in PBS and fixed in acetone at −20°C for 1 min. Cells were rehydrated in PBS, blocked in 1% PBSA for 1 h at 37°C, and incubated with a fluorescein-conjugated goat anti–mouse antibody diluted 1:100 in 1% PBSA for 30 min at 37°C, before being washed and visualized by epifluorescence. In these experiments, EGF-R distribution was also examined by immunofluorescence as described above.

For immunoprecipitation of radiolabeled EGF-R and TN-C proteins, 2.5 × 10⁴ serum-starved SMCs were cultured for 24 h on collagen gels in 35-mm-diam tissue culture dishes. To determine the effect of TN-C on EGF-R synthesis, collagen gels were supplemented with 15 μg/ml of TN-C. 24 h after plating, SMCs were rinsed twice and incubated in cysteine/methionine-free medium (ICN Biomedicals, Inc.) supplemented with 0.1% BSA for 3 h and then pulse-labeled with 100 μCi/ml [³⁵S]methionine/cysteine translabel (ICN Biomedicals, Inc.) for 18 h. To determine the effects of MMP inhibition on TN-C protein synthesis, cells were treated with 1- or 2-μm GM6001 in DMSO or with DMSO alone. Tissue culture medium was collected in RIPA buffer and cleared by centrifugation. Total counts per minute were normalized by TCA precipitation and equal counts per minute (5 × 10⁵ cpm/sample) were immunoprecipitated with 5 μl rabbit polyclonal antisera raised against rat EGF-Rs (RK-2) or human TN-C (GIBCO BRL). Immune complexes were precipitated with 25 μl protein A–Sepharose (Sigma Chemical Co.), washed in 1× RIPA buffer, 50 mM Tris, pH 7.4, and resuspended in 40 μl 2× SDS-PAGE sample buffer. Proteins were separated on 6% SDS–polyacrylamide gels, fixed, dried, and exposed to film (X-Omat; Kodak) for 3 d.

Internucleosomal DNA fragmentation is a characteristic of apoptosis and was used to assess programmed cell death in SMCs cultured on attached and floating type I collagen gels (+/− TN-C). 48 h after floating, cells were rinsed in PBS and incubated on ice with DNA lysis buffer (10 mM Tris, pH 8, 100 mM NaCl, 2 mM EDTA, and 0.5% SDS) for 5 min with rotation. Lysates were incubated with 1 μg/ml proteinase-K (Boehringer-Mannheim Corp.) at 37°C for 24 h with rotation. Genomic DNA was purified by sequential extraction with phenol/chloroform and was precipitated overnight in ethanol at −20°C. DNA pellets were treated with 10 μg/ml RNase A for 1 h at room temperature, and DNA was then quantified based on optical density. 10 μg of genomic DNA was analyzed by electrophoresis through 1.2% agarose gels and visualized with ethidium bromide.

All statistical assessments were compared by one-way analysis of variance and Student Newman Keuls post hoc analysis. A P value of <0.05 was considered statistically significant. Mean values ±SEM are given in the figures.

When chick embryo fibroblasts are cultured on attached type I collagen gels, they express TN-C protein, whereas culture on floating collagen leads to suppression of TN-C expression (Chiquet-Ehrismann et al., 1994). Using this culture method as a model system, we first compared SMC morphology, TN-C production, and MMP activity and expression on attached and floating collagen gels. Within the first few hours of releasing attached collagen gels into the culture medium, SMCs lost their elongated morphology and became rounded (Fig. 1,A). Immunoprecipitation of radiolabeled TN-C protein from SMC conditioned medium after 24 h (Fig. 1 B) showed that synthesis of two TN-C protein isoforms, of apparent molecular masses 220 and 180 kD, was suppressed in SMCs cultured on floating collagen when compared to those cultured on attached collagen gels. No significant differences between the total number of TCA precipitable counts were noted between cells cultured on attached (1.10 × 10⁶ cpm/culture, SEM ± 9.6 × 10⁴, P < 0.05) versus floating collagen (1.13 × 10⁶ cpm/culture, SEM ± 1.59 × 10⁴, P < 0.05), indicating that despite suppression of TN-C protein synthesis, general protein synthesis at this time point was unaffected.

The production and activity of MMPs under these conditions was then examined by gelatin substrate zymography. Two proteins with approximate molecular masses of 68 and 57 kD were detected in conditioned medium harvested from SMCs cultured on attached and floating collagen (Fig. 1,C). Densitometric analysis of these gelatinases showed that, relative to attached collagen, the activities of the 68- and 57-kD gelatinase decreased with time on floating collagen (Fig. 1,D). Under the denaturing and nonreducing conditions used for zymography, the pro-form of MMP-2 has an apparent molecular mass of 68 kD, whereas the active form appears at ∼57 kD (Ailenburg et al., 1993). Thus, our data indicate that the gelatinases detected by zymography may represent MMP-2. Their identity as MMPs was substantiated by the abolishment of gelatinase activity after incubation of zymograms with EDTA or GM6001 (data not shown) and Western immunoblot analysis, which showed that expression of the pro-form of MMP-2 (which migrates at 72 kD under denaturing and reducing conditions) is suppressed on floating collagen (Fig. 1 E).

Previous studies have shown that interstitial collagenases, including MMP-2, are capable of degrading fibrillar native type I collagen to generate collagen fragments (Aimes and Quigley, 1995). At a functional level, this leads to exposure of cryptic Arg-Gly-Asp (RGD) sites within the native type I collagen molecule that bind the α_vβ₃ integrin receptor to promote cell survival (Davis, 1992; Montgomery et al., 1994). This integrin effect may also be achieved by culturing cells on heat-denatured collagen (Montgomery et al., 1994). Since increased SMC TN-C expression correlates with increased MMP expression and activity, we hypothesized that degradation of native type I collagen by MMPs may upregulate SMC TN-C expression through β₃ integrin ligation.

To first establish whether MMPs regulate TN-C expression, we cultured SMCs on attached type I collagen gels in the presence of a specific MMP inhibitor, GM6001 (Galardy, 1993; Strauss et al., 1996). Immunoprecipitation of TN-C from radiolabeled cell lysates showed that inhibition of MMP activity results in suppression of TN-C protein synthesis (Fig. 2 A). In addition, the ability of this inhibitor to suppress TN-C protein synthesis is not due to cytotoxicity since there were no significant differences between the total number of TCA precipitable counts in radiolabeled control (3.18 × 10⁶ cpm/culture, SEM ± 2.12 × 10⁵, P < 0.05) and GM6001-treated cultures (3.75 × 10⁶ cpm/culture, SEM ± 1.63 × 10⁵, P < 0.05) at the time point examined.

We next assessed TN-C protein expression on native versus heat-denatured collagen. Western immunoblot analysis showed that TN-C protein expression is upregulated on denatured collagen (Fig. 2,B), and that this substrate obviated the need for MMPs (Fig. 2,C). To address whether TN-C protein expression is regulated by β₃ integrins, we used a function-blocking anti-β₃ antibody. Western immunoblots showed decreased TN-C protein expression in the presence of the anti-β₃ antibody when compared with IgG-treated cultures on both native (Fig. 2,D) and proteolyzed collagen. (Fig. 2 E).

Since MMPs regulate SMC TN-C protein synthesis on native type I collagen, and exogenous TN-C cooperates with EGF to facilitate SMC growth on this substrate (Jones and Rabinovitch, 1996), we next determined the functional consequences of suppressing MMPs on EGF-dependent SMC morphology, growth, and survival, in the presence and absence of exogenously added TN-C. Inhibition of MMPs with GM6001 in SMCs cultured on attached collagen resulted in cellular rounding (data not shown) and a significant decline in SMC number, whereas inclusion of exogenous TN-C in the collagen gels prevented this decrease (Fig. 3,A) and restored their characteristic elongated morphology (data not shown). Similarly, on floating collagen, in which MMPs and TN-C expression are suppressed, SMC survival is diminished (Fig. 3,B) via apoptosis as determined by oligonucleosomal DNA fragmentation assays (Fig. 3,C). In contrast, SMC numbers were unaffected (Fig. 3,B), and apoptosis was suppressed in cells cultured on floating collagen gels supplemented with exogenous TN-C protein (Fig. 3 C).

Tenascin-C has the capacity to bind and interact with multiple cell surface receptors, including the α_vβ₃ integrin (Prieto et al., 1993; Sriramarao et al., 1993). To determine whether this integrin mediates TN-C/EGF–dependent SMC growth, we evaluated the effect of a functional-blocking anti-α_vβ₃ integrin monoclonal antibody (LM609) on cell morphology and proliferation of SMC cultured on collagen and TN-C–enriched substrates in the presence of EGF. While LM609 had no effect on the stellate SMC morphology observed on collagen, the more elongated SMC morphology induced by TN-C was attenuated by this antibody. Instead, SMCs failed to spread and remained rounded (Fig. 4,A). LM609 had no significant effect on SMC attachment to TN-C–enriched substrates (Fig. 4,B) or on cell number after culture of SMC on collagen in the presence of EGF (Fig. 4,C). In contrast, TN-C–dependent growth in response to EGF was inhibited by LM609 (Fig. 4 C).

The effects of TN-C–α_vβ₃ integrin interactions on SMC shape and EGF-dependent proliferation provided us with the rationale for investigating the organization of the filamentous actin cytoskeleton and the distribution of EGF-Rs and tyrosine-phosphorylated proteins. Accordingly, we used immunofluorescence microscopy to evaluate the distribution of F-actin in SMCs cultured on collagen gels supplemented with exogenous TN-C, and we determined whether this pattern was further altered by EGF (Fig. 5 A). Rhodamine-phalloidin staining revealed a longitudinal distribution pattern of F-actin stress fibers in SMCs cultured on collagen, which became more cortical after addition of EGF. On TN-C–supplemented collagen gels, however, high-intensity F-actin staining was observed at the cell periphery, and this pattern intensified after addition of EGF.

Given that high-affinity EGF-Rs bind directly to actin (den Hartigh et al., 1992), we then assessed whether TN-C–dependent alterations in the organization of the F-actin cytoskeleton may be accompanied by EGF-R clustering, a prerequisite for EGF-dependent signaling, including ligand-dependent tyrosine phosphorylation (Heldin, 1995). Immunofluorescent studies revealed a diffuse pattern of EGF-R staining and tyrosine phosphorylation in SMCs cultured on collagen alone (Fig. 5,A). By contrast, in SMCs cultured on TN-C–enriched collagen gels, prominent EGF-R clustering was observed, especially at the cell periphery. After addition of EGF to collagen gels, a modest increase in both EGF-R clustering and tyrosine phosphorylation was apparent, whereas in TN-C–treated cultures, EGF-R clustering remained pronounced and high levels of tyrosine phosphorylation were evident within the nucleus and at a relatively lower level at the cell periphery. Furthermore, double immunostaining experiments with an antivinculin antibody established that TN-C–treated SMCs contained more focal adhesion complexes, which were larger than those produced on type I collagen alone (data not shown). Most strikingly, exogenous TN-C promoted accumulation of F-actin, EGF-Rs, and tyrosine-phosphorylated proteins within these focal adhesion sites (Fig. 5 B).

Having established that culturing SMCs on exogenous TN-C promotes EGF-R clustering, we next determined the effects of TN-C on EGF-R protein synthesis and tyrosine phosphorylation of this receptor. Immunoprecipitation of radiolabeled whole cell lysates with an anti–EGF-R antibody showed that the levels of receptor protein synthesis were identical on collagen and TN-C–supplemented gels (Fig. 6,A). Consistent with our immunofluorescent studies, Western immunoblotting of membrane-enriched fractions with an antiphosphotyrosine antibody revealed that there was increased tyrosine phosphorylation in the presence of exogenous TN-C (Fig. 6,B). In addition, on both type I collagen and TN-C–supplemented cultures, EGF treatment resulted in a transient increase in protein tyrosine phosphorylation that was enhanced by TN-C. Of particular interest was a 170-kD tyrosine-phosphorylated species that corresponds in molecular mass to the EGF-R. Also noted was an increase in an ∼125-kD species, which may represent focal adhesion kinase (Fig. 6,B). To specifically determine whether TN-C promotes EGF-R activation, EGF-Rs were immunoprecipitated from SMC membrane fractions, and their activation status was assessed using Western immunoblotting with an antiphosphotyrosine antibody. This confirmed that EGF-R phosphorylation occurs by 10 min after addition of EGF, and that this response is enhanced in TN-C–supplemented as opposed to SMC cultures on collagen gels alone (Fig. 6 C).

We next examined the effects of blocking TN-C–β₃ integrin SMC interactions on EGF-R clustering and tyrosine phosphorylation. Smooth muscle cells were preincubated with an anti–β₃ integrin antibody before plating on TN- C–supplemented collagen gels. As with the α_vβ₃ functional-blocking antibody experiments, the anti–β₃ integrin antibody prevented cell spreading on TN-C (data not shown). We therefore used the anti–β₃ integrin antibody in subsequent studies (although similar results would be expected with the α_vβ₃ antibody, LM609) and showed inhibition of TN-C–dependent EGF-R clustering (Fig. 7,A). Immunoprecipitation of SMC membrane–enriched fractions followed by Western immunoblotting with an antiphosphotyrosine antibody showed that upon addition of EGF, EGF-R activation on TN-C–enriched collagen substrates was abrogated in cells pretreated with the function-blocking anti–β₃ integrin antibody (Fig. 7 B).

To determine whether clustering of β₃ integrins on SMC surfaces, in the absence of exogenous TN-C, promotes EGF-R clustering and ligand-dependent phosphorylation of EGF-Rs, anti–β₃ integrin antibody cross-linking studies were performed. Smooth muscle cells cultured on type I collagen gels were treated with either a combination of anti–β₃ integrin and secondary goat anti–mouse IgG F(ab′)₂ antibodies or with primary mouse monoclonal anti–β₃ integrin antibody alone. Immunofluorescent microscopy revealed extensive β₃ integrin clusters on SMC surfaces treated with primary and secondary antibodies, whereas a more diffuse pattern of β₃ integrins was apparent in cultures treated with primary antibody alone (Fig. 8,A). Immunofluorescent microscopy with an anti–EGF-R antibody demonstrated that cross-linking the β₃ integrin receptor leads to clustering of EGF-Rs, which appeared diffuse in SMCs treated with primary antibody alone (Fig. 8 A).

To further assess the effect of β₃ integrin–dependent EGF-R clustering on tyrosine phosphorylation of EGF-Rs, EGF was added to SMCs, which had been pretreated either with a combination of mouse anti–β₃ integrin and secondary goat anti–mouse IgG F(ab′)₂ antibodies, or with anti–β₃ integrin antibody alone. Immunoprecipitation of EGF-Rs from SMC membrane fractions followed by Western immunoblotting with an antiphosphotyrosine antibody showed that a significant increase in EGF-dependent activation of EGF-Rs resulted from cross-linking β₃ integrin receptors (Fig. 8, B and C).

A number of studies have described the induction of TN-C in remodeling vascular tissues, where it has been functionally linked to SMC proliferation (Hedin et al., 1991; Hahn et al., 1995; Chen et al., 1996; Jones and Rabinovitch, 1996; Jones et al., 1997). However, remarkably little was known about the nature of the factors that determine its expression in these tissues or the mechanism by which TN-C supports SMC growth. In this study, we demonstrate that MMPs positively regulate TN-C protein expression and that the mechanism likely involves exposure of cryptic β₃ integrin–binding sites in type I collagen. We also show that TN-C rescues SMCs from apoptosis and promotes EGF-dependent SMC survival through interactions with the α_vβ₃ integrin receptor, which modulates cell shape and EGF-dependent tyrosine phosphorylation. These studies therefore provide new evidence to show how remodeling of the extracellular microenvironment by ECM-degrading proteinases may influence integrin growth factor receptor signaling functions and thus cell behavior. Fig. 9 shows a hypothetical model for the regulation and function of TN-C in SMCs based on our findings.

Recent studies by Chiquet-Ehrismann et al. (1994) demonstrated that chick embryo fibroblasts cultured on attached collagen gels express TN-C protein, whereas in floating cultures, TN-C synthesis is preferentially suppressed. This regulation occurs at the transcriptional level and was mapped to a putative “ECM response element” in the TN-C gene promoter that is distinct from the region important for induction by serum. However, the reason for this differential TN-C production on attached and floating collagen and the functional consequences of suppressing TN-C expression were not explored.

It is well established that specific cell–ECM interactions provide critical cues that regulate proliferation, migration, differentiation, and cell viability (Jones et al., 1993; Boudreau et al., 1995b), and that these functions may depend upon proteolysis of ECM and cell surface molecules (Fujii and Imamaura, 1995; Brooks et al., 1996). Since MMPs and TN-C are often expressed at the same site within normal (Sympson et al., 1994; Jones et al., 1995) and injured remodeling tissues (Latijnhouwers et al., 1996; Vaalamo et al., 1996), including blood vessels (Bendeck et al., 1994; Strauss et al., 1996; Jones, P.L., J. Crack, and M. Rabinovitch. 1996. Mol. Biol. Cell. 7:418a), we examined the possibility that MMPs regulate SMC growth and viability through the induction of TN-C. Accordingly, we established a role for MMPs in regulating TN-C expression and SMC survival by showing that reducing MMP activity and expression on floating collagen gels, or with a specific MMP inhibitor, limits TN-C protein synthesis and SMC growth and survival.

Matrix metalloproteinase–mediated alterations in cell– ECM interactions have profound effects on receptor-mediated intracellular signaling and subsequent cell behavior (for review see Bausbaum and Werb, 1996). For example, MMP-2 exposes cryptic RGD ligands in native collagen, which may bind α_vβ₃ integrins to promote cell survival (Montgomery et al., 1994), an effect which can be mimicked by culturing cells on heat-denatured collagen (Davis, 1992; Montgomery et al., 1994). In the present study, we demonstrate that on native type I collagen, MMPs promote TN-C protein expression via β₃ integrins, and using denatured collagen, we show that this substrate supports TN-C expression in a manner that still requires β₃ integrins but is independent of MMPs. Thus, on attached type I collagen gels, activated MMPs may generate two RGD-containing ligands, i.e., denatured collagen and TN-C, and in this way, serve a dual function in cell survival. In addition, recent studies indicate that MMP-2 directly interacts with the α_vβ₃ integrin receptor to form a functional complex (Brooks et al., 1996). Since the appearance of TN-C expression within remodeling tissues is restricted in a precise spatial manner (Chiquet-Ehrismann et al., 1995), it is also tempting to speculate that induction of TN-C by MMPs serves to localize this ECM component with its cognate adhesion receptors, which include the α_vβ₃ integrin (Prieto et al., 1993; Sriramarao et al., 1993).

Considerable evidence demonstrates that regulation of cell shape and tissue organization by the ECM profoundly influences gene expression, differentiation, and growth behavior (Folkman and Moscona, 1978; Mooney et al., 1995; Ingber, 1993; Roskelley et al., 1994; Boudreau et al., 1996; Chen, 1997; Weaver et al., 1997). Our results show that suppression of endogenous TN-C occurs within a rounded SMC population, in which EGF-dependent growth and survival are limited, whereas addition of exogenous TN-C to these cultures is associated with the restoration of an elongated cell shape, the reestablishment of an EGF-dependent growth response, and the prevention of apoptosis. Also, using functional-blocking anti–β₃ integrin antibodies, we show that EGF-R signaling is attenuated in SMCs that fail to spread on TN-C–enriched collagen substrates. These data strongly indicate that TN-C–dependent cell shape alterations may impact cell growth and survival functions in SMCs. Consistent with this idea, suppression of TN-C expression in mammary epithelial cells is associated with cellular rounding (Jones et al., 1995) and repression of growth-related genes including c-myc (Boudreau et al., 1996). Furthermore, we have shown that SMCs plated on nonadhesive substrates remain rounded and are unable to proliferate in response to EGF, whereas on tissue culture plastic, SMC spread and are able to grow (Jones, P.L., and M. Rabinovitch, unpublished observations).

Our studies comparing SMC growth and survival on attached and floating collagen gels add to previous work that suggests that mechanical factors are able to modulate growth factor–dependent functions, by relating this property to alterations in TN-C expression, and concomitant changes in cell shape and growth factor receptor activation. For example, fibroblasts in mechanically stressed collagen continue to proliferate in response to growth factors (Nakagawa et al., 1989), whereas cells in mechanically relaxed matrices become arrested in G₀ and are less responsive to addition of serum or purified growth factors (Nishiyama et al., 1991). Mechanistically, autophosphorylation of PDGF and EGF-Rs in fibroblasts is decreased in mechanically relaxed collagen (Lin and Grinnell, 1993), whereas β₁ integrin–dependent autophosphorylation of platelet derived growth factor (PDGF β) receptors is induced in fibroblasts in response to the application of external tension (Sundberg and Rubin, 1996). It is therefore possible that mechanical induction of TN-C and/or other ECM components might be important in the activation of a more extensive repertoire of growth factor receptors. Indeed, we have previously shown that TN-C cooperates with another soluble growth factor, bFGF, to facilitate SMC proliferation (Jones and Rabinovitch, 1996). Also, the cooperative effect between TN-C, EGF, and bFGF does not appear to be restricted to a single cell type, or cell surface receptor. Chung et al. (1996) showed that cell surface annexin II, which also acts as a TN-C receptor (Chung and Erickson, 1995), mediates TN-C/bFGF–dependent endothelial cell growth, whereas End et al. (1992) showed that TN-C and EGF stimulate proliferation of Swiss fibroblasts, NIH-3T3 cells and SMCs isolated from spontaneously hypertensive rats. Our present studies therefore confirm the idea that TN-C cooperates with soluble growth factors to promote cell proliferation and extend these studies by providing a mechanism for these cooperative effects that is based on ECM-dependent cell shape changes and cross-talk between integrins and growth factor signaling pathways.

Binding of EGF to its receptor leads to dimerization, which is essential to the subsequent activation of intrinsic tyrosine kinase activity (Schlessinger, 1988). In turn, EGF-R activation catalyzes the tyrosine phosphorylation of several protein substrates, including the receptor itself (Hunter and Cooper, 1988). In the present study, we show for the first time that the α_vβ₃ integrin mediates TN-C/EGF–dependent growth, and that this interaction leads to cell shape changes that are underscored by a redistribution of filamentous actin, clustering of EGF-Rs, and accumulation of tyrosine-phosphorylated proteins at the cell periphery within focal adhesion contacts, as well as increased activation of EGF-Rs after addition of EGF. Moreover, these TN-C–dependent effects on EGF-R clustering and tyrosine phosphorylation could be mimicked by cross-linking SMC β₃ integrins. It is therefore possible that other ECM components that bind β₃ integrins, e.g., vitronectin and proteolysed collagen, which have also been implicated in vascular pathobiology (Montgomery et al., 1994; Brooks et al., 1996), could function in a manner similar to TN-C. Indeed, a recent report shows that denatured monomeric collagen stimulates increased focal adhesion contact formation and PDGF-BB–dependent SMC proliferation (Koyama et al., 1996), whereas polymerized collagen inhibits growth by upregulating cyclin-dependent kinase inhibitors including p27^kip1 and p21^Cip1/Waf1.

Most integrins do not possess intrinsic kinase activity, and their ability to generate a signal after ligation to ECM appears to be dependent upon the formation of specialized structures comprised of multiple components, including clustered integrins, actin, growth factor receptor signaling molecules, and cytoskeletal-associated proteins such as vinculin and focal adhesion kinase (Miyamoto et al., 1995a; Plopper et al., 1995; Burridge and Chrzanowska-Woodnicka, 1996). In addition, a number of studies have provided evidence in favor of a direct role for the actin cytoskeleton in EGF-R signaling. For example, the EGF-R has been shown to be an actin-binding protein (den Hartigh et al., 1992), and from a functional standpoint, this association leads to the accumulation and activation of EGF-R substrates at the binding site (Gronowski and Bertics, 1993; Diakonova et al., 1995). However, the functional consequences of these types of interactions and the identity of cell surface receptors involved had not been explored. Our results strongly indicate that ligation of substrate-bound TN-C to SMC α_vβ₃ integrins leads to the formation of specialized cytoskeletal structures enriched with tyrosine-phosphorylated proteins that together favor the recruitment and clustering of actin-associated high-affinity EGF-Rs. Addition of EGF to these cells, which are now both physically and biochemically primed, would allow the generation of an optimal and ECM-specific mitogenic response.

In conclusion, our results represent a model for neointimal formation, whereby induction of MMPs upregulates TN-C expression, which in turn facilitates growth factor– dependent SMC proliferation. Based on the present studies, and our work that has linked increased TN-C expression and SMC proliferation in vivo (Jones and Rabinovitch, 1996; Jones et al., 1997), we suggest that TN-C and its receptors may be prime therapeutic targets for inhibiting SMC growth that is associated with vascular disease.

We would like to thank Drs. D. Cheresh, I. Lax, M. Silverman and M. Ailenburg, and Glycomed Inc. (GM6001 inhibitor) for providing reagents. We are also indebted to Claire Coulber, Joan Jowlabar, and Susy Taylor for their help in preparing this manuscript.

This study was supported by program grant T2229 from the Heart and Stroke Foundation of Canada.

ECM

extracellular matrix

EGF-R

epidermal growth factor receptor

MMP

matrix metalloproteinase

SFM

serum-free medium

SMC

smooth muscle cell

TN-C

tenascin-C