Cancer Research Infection and Cancer: Biology, Therapeutics, and Prevention Cancer Health Disparities Conference 2009
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Nisato, R. E.
Right arrow Articles by Pepper, M. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Nisato, R. E.
Right arrow Articles by Pepper, M. S.
[Cancer Research 65, 9377-9387, October 15, 2005]
© 2005 American Association for Cancer Research

Cell and Tumor Biology

Dissecting the Role of Matrix Metalloproteinases (MMP) and Integrin {alpha}vß3 in Angiogenesis In vitro: Absence of Hemopexin C Domain Bioactivity, but Membrane-Type 1-MMP and {alpha}vß3 Are Critical

Riccardo E. Nisato1, Ghamartaj Hosseini1, Christian Sirrenberg2, Georgina S. Butler4, Thomas Crabbe5, Andrew J.P. Docherty5, Matthias Wiesner3, Gillian Murphy6, Christopher M. Overall4, Simon L. Goodman2 and Michael S. Pepper1,7

1 Department of Morphology, University Medical Center, Geneva, Switzerland; 2 Oncology Research and 3 Medicinal Chemistry, Merck KGaA, Darmstadt, Germany; 4 Centre for Blood Research, Department of Oral Biological and Medical Sciences, and Department of Biochemistry and Molecular Biology, University of British Columbia, Vancouver, British Columbia, Canada; 5 Celltech, Slough, Berkshire, United Kingdom; 6 School of Biological Sciences, University of East Anglia, Norwich, United Kingdom; and 7 NetCare Molecular Medicine Institute, Unitas Hospital, and Department of Immunology, Faculty of Health Sciences, University of Pretoria, Pretoria, South Africa

Requests for reprints: Michael S. Pepper, NetCare Molecular Medicine Institute, Unitas Hospital, Clifton Avenue, 0140 Lyttleton, Pretoria, South Africa. Phone: 27-12-677-8504; Fax: 27-12-677-8505; E-mail: mpepper{at}doctors.netcare.co.za.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Matrix metalloproteinase (MMP)-2 and its hemopexin C domain autolytic fragment (also called PEX) have been proposed to be crucial for angiogenesis. Here, we have investigated the dependency of in vitro angiogenesis on MMP-mediated extracellular proteolysis and integrin {alpha}vß3–mediated cell adhesion in a three-dimensional collagen I model. The hydroxamate-based synthetic inhibitors BB94, CT1399, and CT1847 inhibited endothelial cell invasion, as did neutralizing anti–membrane-type 1-MMP (MT1-MMP) antibodies and tissue inhibitor of MMP (TIMP)-2 and TIMP-3 but not TIMP-1. This confirmed the pivotal importance of MT1-MMP over other MMPs in this model. Invasion was also inhibited by a nonpeptidic antagonist of integrin {alpha}vß3, EMD 361276. Although PEX strongly inhibited pro-MMP-2 activation, when contaminating lipopolysaccharide was neutralized, PEX neither affected angiogenesis nor bound integrin {alpha}vß3. Moreover, no specific binding of pro-MMP-2 to integrin {alpha}vß3 was found, whereas only one out of four independently prepared enzymatically active MMP-2 preparations could bind integrin {alpha}vß3, and this in a PEX-independent manner. Likewise, integrin {alpha}vß3–expressing cells did not bind MMP-2-coated surfaces. Hence, these findings show that endothelial cell invasion of collagen I gels is MT1-MMP and {alpha}vß3- dependent but MMP-2 independent and does not support a role for PEX in {alpha}vß3 integrin binding or in modulating angiogenesis in this system.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Angiogenesis is critical for the growth of normal tissues and tumors, and depends on precisely controlled interactions between endothelial cells and the extracellular matrix (ECM; ref. 1). On the one hand, the ECM is remodeled to allow endothelial cell migration and capillary morphogenesis, and on the other hand, the ECM recognized by endothelial cells allows motility and provides signals important for cell survival.

Proteinases of the matrix metalloproteinase (MMP) and serine proteinase families direct ECM remodeling, and the integrin family of cell surface receptors mediates adhesive interactions between endothelial cells and the ECM (2). MMPs are a family of zinc-dependent endopeptidases that can be divided into two distinct groups, the secreted MMPs and the membrane-type MMPs (MT-MMP). Of six MT-MMPs, four are anchored to the plasma membrane via a transmembrane domain, whereas, for the other two, anchoring is mediated by glycosylphosphatidylinositol (3). MMPs are secreted and may be stored in extracellular depots as precursor zymogens (4). These are activated by specific proteases, many of which are localized at the cell membrane and help to spatially define regions of matrix breakdown, so-called invadopodial structures, at the tips of sprouting capillaries. MMPs are in turn modulated by inhibitors, the tissue inhibitors of MMPs (TIMP; ref. 2). For example, pro-MMP-2, a potent gelatinase and type IV collagenase, is activated by the cell membrane enzyme MT1-MMP, and is retained on MT1-MMP in an active ternary complex with TIMP-2 (3). MMP-2 binds MT1-MMP via a TIMP-2 bridge at the junction of hemopexin modules III and IV of the MMP-2 hemopexin C domain. In addition to its role in activating MMP-2, MT1-MMP cleaves native type I collagen on binding its hemopexin C domain (5) and many bioactive mediators (6) and also has weak gelatinase activity. Gelatinases are believed to be involved in angiogenesis, for example, by cutting a pathway for invading endothelial cells and by releasing ECM-bound growth factors and cytokines. ECM cleavage can expose novel sequences in collagens, which are then recognized by integrins, notably {alpha}vß3 (7).

Integrins are heterodimeric cell surface receptors composed of noncovalently associated transmembrane glycoproteins, {alpha} and ß, which connect adhesive proteins in the ECM to the cytoskeleton. Excluding splice variants, at present, 16 different {alpha} subunits and 8 different ß subunits have been identified, which associate to form >20 receptors each recognizing its own spectrum of extracellular ligands (8). Integrins not only mediate attachment of cells to their substratum, but also are involved in intracellular signal transduction. Integrin {alpha}vß3 is a receptor for proteins bearing an exposed Arg-Gly-Asp (RGD) tripeptide; these proteins include vitronectin, fibronectin, fibrinogen, thrombospondin, osteopontin, von Willebrand factor, and some degraded laminins and collagens (9). {alpha}vß3 has been shown to be important during angiogenesis. In vivo, {alpha}vß3 is not widely expressed. It seems to be most prominent on cytokine-activated endothelial cells during angiogenesis in a wide variety of settings and is also expressed by smooth muscle cells in postangioplasty restenosis, in atherosclerotic plaques, in healing arterial wounds, and in osteoclasts (10). A significant body of experimental evidence has shown that {alpha}vß3 antagonists (e.g., antibodies and cyclic RGD peptides) can inhibit angiogenesis during development, on wound healing, and in growing tumors (11, 12). The relevance of {alpha}vß3 to angiogenesis and its potential as a therapeutic target has, therefore, been established.

It has been reported that integrin {alpha}vß3 binds MMP-2 via its hemopexin C domain (13, 14) and colocalizes with degraded collagen type I (15). More recently, MT1-MMP was reported to physically interact with and to process {alpha}vß3, increasing outside-in signaling via {alpha}vß3 (16). In this way, MMP-2 and MT1-MMP (17, 18) activity may be colocalized where {alpha}vß3 ligands freshly generated by proteolysis can have a rapid effect: at the {alpha}vß3 receptor. MMPs are susceptible to autolysis and the autolytically generated hemopexin C domain of MMP-2 has been reported to have biological activities independent of its role in the full-length enzyme, which is to bind TIMP-2 (19) and substrates (20). The reports showing that isolated hemopexin C domain of MMP-2 blocks angiogenesis in model systems and inhibits the interaction of MMP-2 with integrin {alpha}vß3 (13, 14) have generated considerable controversy (21). Moreover, confirmatory reports are lacking. Expression of PEX in neonatal mouse retina parallels the development and stabilization of retinal blood vessels (14). In view of the important biological and therapeutic implications of MMP-2/{alpha}vß3/MT1-MMP/TIMP-2 interactions, we dissected the roles of these molecules in angiogenesis using a well-defined three-dimensional in vitro model and focused specifically on the role of the hemopexin C domain of MMP-2.

Here, we report that both MT1-MMP and {alpha}vß3 were critical for in vitro angiogenesis. However, and to our surprise, angiogenesis occurred independently of the activation status of MMP-2. Furthermore, we found no evidence for high-affinity binding of MMP-2 or its hemopexin C domain to {alpha}vß3.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Reagents. The synthetic MMP inhibitor BB94 and inactive isomer BB1268 were provided by Dr. P. Brown (British Biotechnology, Oxford, United Kingdom). The synthetic MMP inhibitors CT1399 and CT1847 and recombinant TIMP-1, TIMP-2, and TIMP-3 have been described previously (22–24). The different preparations of hemopexin C domain as well as pro- and active MMP-2 are indicated in Supplementary Tables S2 and S3, respectively. Pro-MMP-2 was activated by treatment with 4-aminophenylmercuric acetate (APMA). Anti-human MT1-MMP-neutralizing monoclonal antibodies (mAb) LEM-1/58 and LEM-2/15 were a gift from Dr. A.G. Arroyos (Immunology Department, Hospital de la Princesa, Madrid, Spain; ref. 17). Recombinant human vascular endothelial growth factor (VEGF)-A (165–amino acid isoform), fibroblast growth factor-2 (FGF-2), and transforming growth factor-ß1 (TGF-ß1) were from Peprotech, Inc. (Rocky Hill, NJ), Dr. P. Sarmientos (Farmitalia Carlo Erba, Milan, Italy), and R&D Systems Europe (Oxon, United Kingdom), respectively. Concavalin A, phenylmethylsulfonyl fluoride (PMSF), pepstatin A, and phorbol 12-myristate 13-acetate (PMA) were from Sigma Chemical Co. (St. Louis, MO). Polymyxin B sulfate was from Calbiochem-Novabiochem Corp. (La Jolla, CA). Native and recombinant human integrins {alpha}vß3, {alpha}vß5, and {alpha}vß6, native gpIIbIIIa, plasma fibronectin and vitronectin were prepared as described previously (25–27). For binding assays, collagen type IV (human placenta) and collagen type I (rat tail) were purchased from Sigma Chemical and Serva (Heidelberg, Germany), respectively. Integrin {alpha}vß3 inhibitory cyclic peptide EMD 66203 and control EMD 69601 were provided by Dr. A. Jonczyk (Merck KGaA, Darmstadt, Germany) and are described elsewhere (28). The synthesis of the racemic mixture of {alpha}vß3-specific inhibitor EMD 361276 and the noninhibitory control EMD 362619 have been described.8 Monoclonal anti-biotin peroxidase-conjugated antibody, NHS-biotin, and TMB peroxidase substrate were from Sigma Chemical, Pierce Chemicals (Rockford, IL), and Bio-Rad Laboratories (Hercules, CA), respectively. All other reagents were from Merck KGaA unless otherwise stated.

Cell culture. Adrenal cortex–derived bovine microvascular endothelial (BME) cells (provided by Drs. M.B. Furie and S.C. Silverstein, Columbia University, New York, NY; ref. 29) were grown in {alpha}-modified MEM (Life Technologies, Gaithersburg, MD) supplemented with 2% to 15% heat-inactivated donor calf serum (DCS; Life Technologies) and antibiotics. Bovine aortic endothelial (BAE) generated as described previously (30), were cultured in low-glucose DMEM (Life Technologies) supplemented with 10% DCS and antibiotics.

In vitro angiogenesis in a three-dimensional collagen gel model. In vitro angiogenesis was assayed as described previously (31). BME and BAE cells were seeded onto rat type I collagen gels at 1.0 x 105 cells per well. On reaching confluence, DCS was reduced to 2%, and cells were stimulated with FGF-2 (10 ng/mL) or VEGF-A (30 ng/mL) alone or in combination in the presence or absence of {alpha}vß3- inhibitors, MMP-inhibitors, TIMP-2, or hemopexin C domain. Where indicated, polymyxin B sulfate was added at 10 µg/mL. Media and reagents were renewed after 2 days, and after a further 2 to 3 days, cultures were fixed with 2.5% glutaraldehyde in 100 mmol/L sodium cacodylate buffer (pH 7.4) and photographed using a Nikon (Tokyo, Japan) Diaphot TMD inverted photomicroscope. Invasion was quantitated as described previously (30). Results are shown as the mean ± SEM total additive sprout length (in µm) from three randomly selected fields per experiment; at least three experiments were done per condition.

Gelatin zymography. MMP-2 activity was analyzed using gelatin zymography (32). Confluent BME cell monolayers on three-dimensional type I collagen gels were stimulated in medium supplemented with 2% DCS with the reagents indicated in the corresponding figure. After 2 days, conditioned media were collected and prepared as described previously (28). For cell extracts, confluent monolayers on collagen gels were washed and gently detached from the sides of the dishes and then mixed with lysis buffer [1% Triton X-100, 150 mmol/L NaCl, 5 mmol/L EDTA, 1 mmol/L PMSF, 1 µg/mL pepstatin A, 50 mmol/L Tris-HCl (pH 8)]. Lysates were sonicated and centrifuged (10 minutes, 15,800 x g). Protein (50 µg) from cell lysates or conditioned media (30 µL) were processed as described previously (28). Zymograms were photographed, photographs were scanned using a Scan Jet II CX scanner, and band intensities were evaluated with ImageQuant analysis software.

Binding assays. Detailed methodologies for binding of pro-MMP-2, active MMP-2, and integrin {alpha}vß3 with other proteins and MMP-2-{alpha}vß3 inhibition assays can be found in Supplementary Data.

Northern blot assays. Northern blot assays were done as described previously (28). Confluent monolayers of BME cells on collagen gels were washed and gently detached from the dishes. Medium was removed by placing the gels onto sterilized filters (Schleicher & Schüll AG, Feldbach, Switzerland) and applying mild suction. Total cellular RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA). 32P-dUTP cRNA probes were generated from human MMP-2 and MT1-MMP cDNAs provided by Drs. W.G. Stetler-Stevenson (Bethesda, MD), M. Seiki (Tokyo, Japan), and K.H. Scheit (Göttingen, Germany). As an internal control, the filters were hybridized simultaneously with a 32P-labeled human P0 ribosomal phosphoprotein cRNA probe (33). Autoradiograms were scanned and analyzed using ImageQuant software 3.3 (Molecular Dynamics, Sunnyvale, CA) and bands were normalized relative to P0. One representative autoradiogram is shown and quantitative data represent the mean ± SEM from three independent experiments.

Immunoprecipitation. It is technically challenging to perform an immunoprecipitation from cells grown on a thick collagen gel; we thus opted to perform this experiment with cells grown on gelatin. Confluent BME cell monolayers were treated for 48 hours with the reagents indicated in the corresponding figure. Cells were washed with PBS, scraped, and centrifuged (340 x g, 10 minutes, 4°C). Pellets were resuspended in 100 mmol/L Tris (pH 7.4), 20 mmol/L EDTA, 1 µg/mL pepstatin A, 1 µg/mL antipain, 1 mmol/L benzamidine, 200 Kallikrein IU/mL Trasylol, 2 mmol/L PMSF, and 1 mmol/L diisopropyl fluorophosphate. The homogenates were centrifuged (900 x g, 10 minutes), and supernatants were further centrifuged (100,000 x g, 1 hour). Pellets were resuspended in PBS containing 0.5% Triton X-100, 1 mmol/L PMSF, and 200 Kallikrein IU/mL Trasylol and frozen. Protein was estimated in the BCA assay. Plasma membrane fractions from HT1080 cells were used as a positive control. Cell membrane fractions (40 µg) in 250 µL of 50 mmol/L Tris-HCl (pH 8.1), 150 mmol/L NaCl, 1 mg/mL bovine serum albumin (BSA), and 1% Triton X-100 were cleared by incubation with protein G-Sepharose beads (Pharmacia Biotech, Uppsala, Sweden; 1 hour, 4°C, with agitation). Samples were then centrifuged and supernatants were incubated with anti-MT1-MMP mAb (1 µg, 2 hours; clone 113-5B7, Chemicon, Temecula, CA). Protein G-Sepharose was added to the samples which were then agitated (2 hours, 4°C). Samples were washed twice with TBS and twice with 50 mmol/L Tris-HCl (pH 6.8) and resuspended in 2% w/v SDS and boiled (95°C, 3 minutes). Samples were electrophoresed under nonreducing conditions and transferred to polyvinylidene difluoride membranes (Bio-Rad Laboratories). The blots were blocked with 5% dry milk in TBS and then incubated with anti-MT1-MMP (113-5B7; 3 µg/mL; 16 hours, 4°C) followed by incubation (1 hour) with horseradish peroxidase–conjugated mouse IgG. Immunoreactive bands were visualized by enhanced chemiluminescence (Amersham Pharmacia Biotech, Little Chalfont, United Kingdom) according to the manufacturer's instructions. The results shown are representative of two independent experiments.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Integrins {alpha}vß3 and {alpha}vß5 are required for angiogenesis in vitro. RGD-containing inhibitors of {alpha}vß3 and {alpha}vß5 integrins can inhibit FGF-2- and VEGF-A-driven endothelial cell invasion of three-dimensional collagen type I gels (28). However, it has been suggested that such inhibitors might function by entering cells and acting at sites other than integrins (34). We therefore investigated whether a nonpeptidic inhibitor could also affect endothelial cell invasion and tube formation in a three-dimensional collagen model. EMD 361276 is an indol-3-yl derivative that has an IC50 of 6 nmol/L on {alpha}vß3 versus 700 nmol/L on {alpha}vß5 at the receptor level (400 nmol/L versus 20 µmol/L at the cellular level), activity similar to the peptide inhibitor EMD 66203 (35). EMD 361276 is thus selective for {alpha}vß3 with little effect on {alpha}vß5. The activities of the compounds are summarized in Supplementary Data. The effect of these compounds in the three-dimensional collagen assay is summarized in Fig. 1. Unstimulated BME cells did not invade the gel (Fig. 1A). When stimulated by FGF-2, they invaded the collagen gel and formed tubes (Fig. 1B); this was reduced by the {alpha}vß3-selective inhibitor EMD 361276 (Fig. 1C) but not by its enantiomer EMD 362619. No signs of cytotoxicity or morphologic changes were seen in control monolayers incubated with the compounds up to 50 µmol/L (data not shown). To quantitate anti-invasive activity, cells were exposed to FGF-2 or VEGF-A in the absence or presence of serially diluted inhibitors. Invasion was quantitated and IC50s were determined. Contrary to inactive EMD 362619, EMD 361276 reduced BAE and BME cell invasion in a dose-dependent manner (Fig. 1D and E) and was more effective in inhibiting invasion into fibrin than collagen.



View larger version (66K):
[in this window]
[in a new window]
  		Figure 1. Nonpeptidic EMD 361276 but not its enantiomer inhibits angiogenesis in vitro. Confluent monolayers of BME cells on three-dimensional collagen gels were treated with 10 ng/mL FGF-2 for 5 days. Phase-contrast views of a control untreated monolayer (A) and cells treated with basic FGF-2 alone (B) or in combination with 50 µmol/L EMD 361276 (C). Bar, 100 µm. EMD 361276 but not its enantiomer EMD 362619 inhibits FGF-2 (10 ng/mL)–induced BME cell collagen gel invasion in a dose-dependent manner (D). IC50s (E) were estimated from the dose-response curves plotted in (D) as well as additional BAE and BME cell invasion inhibition profiles for which the graphics are not shown. The Origin 3.5 analysis program (Microcal Software, Inc., Northampton, MA) was used to fit a sigmoidal curve to the data. Residual cell invasion in the presence of 50 µmol/L EMDs is expressed as a percentage of cell invasion induced by FGF-2 or VEGF-A alone (100%). Statistical analysis was done as described in Materials and Methods. Results are from at least three experiments per condition (three fields per well per experiment). Points, mean percentage of total cell invasion with respect to cell treatment with cytokine alone (100%); bars, SEM.

 
Neither matrix metalloproteinase-2 nor its hemopexin C domain is involved in angiogenesis in three-dimensional collagen gels. Exogenous hemopexin C domain, which competes with endogenous pro-MMP-2 and reduces its activation (14, 19, 36), was added to the system to determine whether the active form of MMP-2 was directly involved in BME cell collagen gel invasion. The hemopexin C domain partially inhibited invasion induced by FGF-2 plus VEGF-A (Fig. 2E). However, because the hemopexin C domain was expressed in Escherichia coli and because the preparation presumably contained lipopolysaccharide (LPS), polymixin B sulfate was also included in the assay to neutralize the effect of LPS. Surprisingly, in vitro angiogenesis was not significantly altered by the hemopexin C domain in the presence of polymixin B sulfate (Fig. 2F and G). Polymyxin B sulfate alone was added to BME cells (Fig. 2A) treated with FGF-2 plus VEGF-A (Fig. 2D); polymyxin B sulfate itself had no effect on cell invasion. Cell invasion induced by FGF-2 and VEGF-A combined in the presence of increasing concentrations of the hemopexin C domain of MMP-2 (and polymyxin B) is shown in Fig. 2G; hemopexin C domain had no effect up to 2 µmol/L. A similar lack of effect on cell invasion was observed with all the hemopexin preparations, including the {Delta}MMP-2 inactive mutant.



View larger version (96K):
[in this window]
[in a new window]
  		Figure 2. Lack of effect of hemopexin C domain of MMP-2 on angiogenesis in vitro. Confluent monolayers of BME cells on three-dimensional collagen gels were treated with FGF-2 (10 ng/mL) and VEGF-A (30 ng/mL) in combination for 4 days. Phase-contrast views of control untreated monolayer (A and B) and cells treated with FGF-2 + VEGF-A (C and D), which results in the formation of cell cords within the collagen gel (the plane of focus is beneath the surface monolayer); partial inhibition of invasion in the presence of hemopexin C domain (from Dr. C.M. Overall) in the absence (E) but not in the presence (F) of polymyxin B. These images are representative fields from one of at least three experiments per condition. Bar, 75 µm. Quantitation of total additive sprout length from BME cell monolayers on three-dimensional collagen gels stimulated with FGF-2 + VEGF-A in the presence of 10 µg/mL polymyxin B sulfate and increasing concentrations of hemopexin C domain (G).

 
Recombinant human matrix metalloproteinase-2 hemopexin C domain inhibits cytokine-induced matrix metalloproteinase-2 activation in bovine microvascular endothelial cells cultured on three-dimensional type I collagen gels. We next investigated whether hemopexin C domain preparations could block the activation or activity of MMP-2 produced by our cells. We observed by zymography that several cytokines and reagents induced MMP-2 activation (Fig. 3A) in either cell extracts or conditioned media. Hemopexin C domain inhibited FGF-2 plus VEGF-A–induced activation of pro-MMP-2 in a dose-dependent manner (Fig. 3B) in BME cells cultured on three-dimensional collagen gels. Generation of active MMP-2 at the cell surface was inhibited by ~90% (Fig. 3B, right) by 2 µmol/L hemopexin C domain, and generation of this form in conditioned media was also inhibited by ~2-fold (Fig. 3B, left). Similar inhibition was obtained with the hemopexin C domain and the {Delta}MMP-2 inactive mutant.



View larger version (35K):
[in this window]
[in a new window]
  		Figure 3. Effect of hemopexin C domain of MMP-2 on cytokine-induced MMP-2 activation in BME cells cultured on three-dimensional type I collagen gels. Confluent monolayers of BME cells were treated for 48 hours in the presence of FGF-2 (10 ng/mL), VEGF-A (30 ng/mL), TGF-ß1 (1 ng/mL), PMA (20 ng/mL), or concavalin A (Con-A; 50 µg/mL). Conditioned medium from the MCF-7 cell line (known to produce MMP-2) was used as a positive control. Different cytokines induce MMP-2 active forms (A). Confluent monolayers were treated with hemopexin C domain (from Dr. C.M. Overall; 10 nmol/L-2 µmol/L) in the presence of FGF-2 + VEGF-A and polymyxin B sulfate (10 µg/mL) for 48 hours. Hemopexin C domain decreased MMP-2 activity as assessed by gelatin zymography in either conditioned medium (left) or cell extract (right; B, top). MMP-2 standard is shown on the left of the zymograms. Zymograms are representative of at least three experiments. Semiquantitative representation of the active form of MMP-2 (B, bottom). Columns, mean percentage (from at least three independent experiments) of active MMP-2 compared with the total amount of MMP-2 (latent + active) for each condition (bottom; left, conditioned medium; right, cell extracts); bars, SEM.

 
Specificity of pro–matrix metalloproteinase-2 or active matrix metalloproteinase-2 binding to integrin {alpha}vß3. We investigated whether we could measure binding interactions within the MMP-2/{alpha}vß3/hemopexin C domain triptych. We found that pro-MMP-2 did bind {alpha}vß3 but only in the absence of carrier protein (BSA; Fig. 4A). Indeed, different carrier proteins (casein, IgG, and ovalbumin) all efficiently prevented the interaction of pro-MMP-2 with {alpha}vß3. Neither vitronectin binding to {alpha}vß3 nor pro-MMP-2 binding to collagen was affected by carrier proteins. In the absence of carrier protein, pro-MMP-2 bound strongly to {alpha}vß3, {alpha}vß5, and gpIIbIIIa (data not shown). Hence, these results exclude any specific interaction between pro-MMP-2 and {alpha}vß3.



View larger version (31K):
[in this window]
[in a new window]
  		Figure 4. Specificity of pro-MMP-2 or active MMP-2 and hemopexin C domain for integrin {alpha}vß3. The binding of biotinylated pro-MMP-2 or active MMP-2 and vitronectin to immobilized proteins was detected using an anti-biotin peroxidase-conjugated antibody. Biotinylated pro-MMP2 or vitronectin was allowed to bind to increasing concentrations of {alpha}vß3 in the absence (A, left) or presence (A, right) of 0.5% BSA. Biotinylated pro-MMP-2 binding viability was confirmed by pro-MMP-2 binding to type IV collagen (A, right). Pro-MMP-2 was activated with APMA during the indicated time course and then biotinylated and exposed to immobilized BSA, {alpha}vß3, and collagen I (B, left). Increasing concentrations of biotinylated preactivated MMP-2 (Chemicon) were exposed to {alpha}vß3, {alpha}vß5, or BSA-coated surfaces; biotinylated vitronectin binding to integrin {alpha}vß3 (VN to {alpha}vß3) was used as a positive control (B, right). Biotinylated preactivated MMP-2 (Chemicon) was allowed to bind to immobilized {alpha}vß3 or BSA in the presence of 0.5% BSA and increasing concentrations of either EMD 66203 or control inactive EMD 69601 (C, left). Preactivated MMP-2 (Chemicon) was allowed to bind to {alpha}vß3 in the presence of 0.5% BSA and either glutathione S-transferase (GST)-PEX 111198 or GST-PEX 131198 or GST and EMD 66203 (C, right). Experiments were done in duplicate and repeated four times. Columns, mean of four independent experiments; bars, SD.

 
The ability of activated MMP-2 to bind {alpha}vß3 was source dependent. Only preactivated MMP-2 (Chemicon) selectively bound {alpha}vß3 but not {alpha}vß5 in the presence of a carrier protein (Fig. 4B, right), whereas pro-MMP-2 (from Dr. G. Murphy), which we activated with APMA, failed to bind {alpha}vß3 but could bind collagen I (Fig. 4B, left), an interaction that has been well characterized (5). Similarly, neither APMA-activated pro-MMP-2 (Chemicon) or active MMP-2 (Oncogene) could specifically bind {alpha}vß3.

These discrepancies in binding seem not to be related to levels of MMP-2 activity because zymograms done with the different MMP-2 preparations revealed similar levels of activation (data not shown). Furthermore, M21 cells expressing {alpha}vß3 did not bind to surfaces coated with increasing amounts of active MMP-2 but bound strongly to vitronectin-coated surfaces via {alpha}vß3 (data not shown).

The interaction between active matrix metalloproteinase-2 and {alpha}vß3 is Arg-Gly-Asp dependent but hemopexin C domain- independent. The binding of active MMP-2 to {alpha}vß3 was prevented by the cyclic RGD peptide EMD 66203 but not by a control peptide, EMD 69601, which does not contain RGD (Fig. 4C, left and right). However, in accord with the lack of effect of the hemopexin C domain in cell invasion tests (above), hemopexin C domain was unable to prevent the interaction of active MMP-2 with {alpha}vß3.

Matrix metalloproteinases involved in angiogenesis in vitro. Because hemopexin C domain inhibited MMP-2 activity but not cytokine-induced tubular morphogenesis in collagen gels, it seemed unlikely that the proteolytic activity of MMP-2 was crucial for mediating tube formation in our model. We next investigated which MMPs were involved in cell invasion in collagen gels by using both hydroxamate-based inhibitors (CT1399 and CT1847) and naturally occurring TIMPs (1–3). We found that in vitro angiogenesis was completely inhibited by CT1399 at 100 nmol/L, whereas 10 µmol/L CT1847 was required for the same inhibition. FGF-2- and FGF-2 plus VEGF-A–induced invasion were inhibited in a dose-dependent manner by BB94, CT1399, and CT1847 (Fig. 5A). BB1268, an inactive isomer of BB94, did not significantly alter invasion (76 ± 11%; Fig. 5A). Thus, the ability of the specific inhibitor CT1399 to block invasion at a concentration two orders of magnitude lower than the general MMP inhibitors (BB94 and CT1847) when related to the inhibition constants of the inhibitors for the MMPs (see Supplementary Data) implicated gelatinase MMP-9 and MT1-MMP in BME collagen gel invasion. Furthermore, TIMP-2 and TIMP-3, but not TIMP-1, significantly inhibited FGF-2 plus VEGF-A–induced BME cell invasion (Fig. 5B; data not shown). These results suggest that MT1-MMP is a key enzyme mediating cell invasion in our model because TIMP-2 and TIMP-3 inhibit MT1-MMP, whereas TIMP-1 is a very poor MT1-MMP inhibitor (36, 37).



View larger version (18K):
[in this window]
[in a new window]
  		Figure 5. Effect of MMP inhibitors on angiogenesis in vitro. Confluent monolayers of BME cells on three-dimensional collagen gels were treated for 3 to 5 days with the indicated cytokines (10 ng/mL FGF-2 and 30 ng/mL VEGF-A) in the absence or presence of different MMP inhibitors. Synthetic MMP inhibitors BB94, CT1399, and CT1847 but not a BB94-related isomer (BB1268) inhibit BME cell invasion of three-dimensional collagen gels (A). TIMP-2 and TIMP-3 exhibit a robust inhibition of BME cell invasion, whereas TIMP-1 has little or no effect (B). Results are mean ± SEM total additive sprout length (in µm; A) or percentage ± SEM of cell invasion induced by FGF-2 + VEGF-A alone (100%; B) from three randomly selected photographic fields per experiment; at least three experiments were done per condition. Note that the concentration range for CT1399 differs from that of the other inhibitors.

 
Cytokines modulate membrane-type 1-matrix metalloproteinase levels. We next investigated the expression of MT1-MMP in cytokine-stimulated BME cells invading collagen gels and found that mRNA levels were significantly increased by FGF-2, FGF-2 plus VEGF-A, and PMA, whereas VEGF-A alone induced a modest increase (Fig. 6A). Protein levels of MT1-MMP were significantly increased by FGF-2 plus VEGF-A, only slightly by FGF-2, and not at all by VEGF-A (Fig. 6B). Because MT1-MMP activity is needed for BME cell tubulogenesis driven by FGF-2, VEGF-A, or FGF-2 plus VEGF-A, and MT1-MMP levels are only modestly enhanced by these cytokines, if at all, perhaps MT1-MMP activity might be regulated by specific relocalization to motility and/or collagen-degrading sites (invadopodia). This has been shown previously for migrating cells (17). MMP-2 mRNA levels were unaltered or only slightly increased by cytokines and did not correlate with increased MMP-2 activity at least in cells treated with FGF-2 plus VEGF-A (Fig. 3; data not shown).



View larger version (33K):
[in this window]
[in a new window]
  		Figure 6. Effect of cytokines, PMA, and concavalin A on MT1-MMP and MMP-2 expression by BME cells cultured on three-dimensional collagen gels. Confluent BME cell monolayers were treated with the aforementioned reagents for 24 or 48 hours, after which total cellular RNA was extracted and analyzed (5-10 µg/lane) by Northern blot (A). Filters were hybridized simultaneously with 32P-labeled cRNA probes for human MT1-MMP and bovine P0 ribosomal phosphoprotein (used as an internal control for determination of the amount of RNA loaded; top). Autoradiograms were scanned and bands were quantitated using a densitometric scanner. Values were normalized relative to the P0 signal and an arbitrary value of 1 was assigned to controls (unstimulated cultures). Results after 24 hours of stimulation are mean ± SEM from at least two independent experiments (bottom). Western blot analysis of MT1-MMP (B). BME cells were treated for 48 hours with the aforementioned reagents. Plasma membrane fractions (40 µg) were then immunoprecipitated and analyzed by Western blot using the same antibody; plasma membrane fractions from HT1080 cells were used as a positive control.

 
Membrane-type 1-matrix metalloproteinase is the master enzyme mediating bovine microvascular endothelial cell tubulogenesis. As a final test for the role of MT1-MMP in BME cell tubular morphogenesis, we used the MT1-MMP–neutralizing antibodies LEM-1 and LEM-2. These antibodies inhibited cell invasion in a dose-dependent manner (Fig. 7) leading to an almost complete block of cytokine-induced invasion at a concentration of 10 µg/mL (Fig. 7D). No morphologic changes or toxicity were observed on unstimulated monolayers even in the presence of 10 µg/mL antibody (Fig. 7B; data not shown). Interestingly, these antibodies prevented cytokine-induced tube formation, but cells were clearly still activated as illustrated by their elongated, spindle-shaped morphology and greater cell density when compared with controls (Fig. 7D). Taken together, these data unequivocally exclude MMP-2, and any binding interaction it might have with integrin {alpha}vß3, and emphasize the role of MT1-MMP in mediating cytokine-induced bovine endothelial cell invasion in type I collagen gels.



View larger version (74K):
[in this window]
[in a new window]
  		Figure 7. Specific MT1-MMP-neutralizing antibodies inhibit cytokine-induced BME cell invasion of three-dimensional collagen gels. Confluent monolayers were left untreated or were stimulated with 10 ng/mL FGF-2 (4 days) or 30 ng/mL VEGF-A (4 days) alone or in combination (2 days) with increasing concentrations of mAbs LEM-1 and LEM-2. Phase-contrast views of control untreated monolayers (A) and control monolayer cultured in the presence of 10 µg/mL LEM-1 for 7 days (B), treated with FGF-2 + VEGF-A (C), and treated with FGF-2 + VEGF-A in the presence of 3 µg/mL LEM-1 antibody (D). Bar, 100 µm. Quantitation of LEM-1 (E, top) and LEM-2 (F, bottom) dose-response inhibition. Results are mean ± SEM total additive sprout length (in µm) from three randomly selected photographic fields per experiment; three experiments were done per condition.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
The objectives of this study were to identify the differential importance of the proteolytic members of the collagenolytic MMP-2/MT1-MMP axis (6) in our in vitro angiogenesis model and the relationship between MMP-2 and integrin {alpha}vß3 in this system. In the three-dimensional culture system we used here, we found that for in vitro angiogenesis in type I collagen, (a) MT1-MMP and integrin {alpha}vß3 function are necessary; (b) MMP-2 is both secreted and activated during angiogenesis, but its enzymatic activity is not required for in vitro tubulogenesis; (c) there is no interaction between MMP-2 and {alpha}vß3 of significance to cytokine-driven endothelial morphogenesis; and (d) that contrary to previous published reports, when contaminating LPS is neutralized, the hemopexin C domain of MMP-2 does not modulate angiogenesis. Our culture system mimics physiologic angiogenesis: quiescent endothelial cells are activated by physiologically relevant stimulators; endothelial cells proliferate and undergo morphologic and biochemical differentiation characteristic of angiogenesis in vivo; and the system is sensitive to reagents that modulate angiogenesis in vivo. For example, growth factors like FGF-2 and VEGF-A stimulate in vitro angiogenesis (38), and inhibitors of {alpha}vß3 integrins block the process (28).

Our data implicate the proteinase MT1-MMP in in vitro angiogenesis and deemphasize the roles of MMP-1, MMP-2, MMP-3, MMP-7, and MMP-9. We examined the specificity of a range of MMP inhibitors and compared their activities in vitro. Angiogenesis was blocked by synthetic mixed inhibitors of MT1-MMP, MMP-2 and MMP-9, and by naturally occurring inhibitors of MT1-MMP (i.e., TIMP-2 and TIMP-3 but not by TIMP-1). TIMP-1 has poor inhibitory activity toward MT-MMPs (36, 37). Furthermore, in vitro angiogenesis was efficiently inhibited by reagents that specifically inhibit MT1-MMP (17). Pro-MMP-2 is activated by the cell membrane enzyme MT1-MMP and is retained on MT1-MMP in an active ternary complex with TIMP-2 (3, 36). Indeed, TIMP-2 binds the active site of MT1-MMP and the C-terminal hemopexin-like domain of pro-MMP-2 via its N- and C-termini, respectively. In our model, we found that MMP-2 was secreted and activated, as reported for rat endothelial cells in a collagen gel model, in a process that involves MT1-MMP, because both TIMP-2 and hemopexin C domain inhibit it (39). As reported by Brooks et al. (14), we observed that the MMP-2 hemopexin C domain blocks MMP-2 activation, which indirectly but selectively blocks MMP-2 gelatinase activity. However, neither a variety of hemopexin C domain preparations from separate laboratories nor a broad range MMP inhibitor (i.e., TIMP-1, which includes MMP-2 but not MT1-MMP in its spectrum) had an effect in our model. This implicates MT1-MMP, but not MMP-2, in FGF-2 plus VEGF-A–induced in vitro angiogenesis. Similarly, human umbilical vein endothelial cell tubulogenesis in three-dimensional fibrin gels was MT-MMP dependent but MMP-2 independent based on its insensitivity to TIMP-1, to recombinant hemopexin C domain, and to a function-blocking antibody to MMP-2 (40). Partial inhibition of PMA-induced tubulogenesis in collagen gels was also reported by Galvez et al. (17) using these antibodies. Because TIMP-2 or TIMP-3 and MT1-MMP function-blocking antibodies almost totally blocked BME invasion in collagen gels, but TIMP-1 or hemopexin C domain did not, despite essentially complete inhibition of MMP-2 activity, MT1-MMP is identified as the key player in matrix degradation, whereas direct MMP-2 activity is excluded, at least in this system.

Our findings are supported by data emanating from genetically modified mice. MMP-2-null mice develop with no apparent abnormal phenotype, whereas a mutation in the MMP-2 gene shows relative abnormal osteogenesis (41, 42). Thus, it seems that MMP-2 does not mediate angiogenesis during development. In adults, MMP-2-null mice display reduced angiogenesis in a tumor model and in a corneal micropocket assay (43, 44). In contrast, MT1-MMP-null mice have an apparently normal phenotype at birth but succumb after a few weeks postpartum. These mice are 30% to 40% smaller than wild-type mice and display skeletal defects and reduced vasculogenesis and angiogenesis; the latter are believed to be responsible for the abnormal skeletal development (45). In an in vitro model of angiogenesis using murine aortic vessel implants or endothelial cells from mice deficient in MMP-2, MMP-9, or MT1-MMP, it was shown that only MT1-MMP was required for collagenolytic activity, endothelial cell invasion, and generation of new vessels (46).

MMPs and {alpha}vß3 were involved in angiogenesis in our model, and MMP-2 was generated and activated. The interaction of MMP-2 with {alpha}vß3 has been targeted to prevent angiogenesis, and both hemopexin C domain and low molecular weight inhibitors of the {alpha}vß3-MMP-2 interaction have been reported to be antiangiogenic (14, 47). In biochemical assays, MMP-2 has been shown to bind {alpha}vß3 only in the absence of carrier proteins, including albumin, a protein ubiquitously present in vivo. By contrast, binding of {alpha}vß3 to its ligand vitronectin and binding of MMP-2 to its ligands collagen type IV and I were independent of carrier protein. Interaction of {alpha}vß3 with MMP-2 was not enhanced by MMP2-activation, nor was it related to the means of MMP-2 detection: direct labeling, antibody detection, and detection of enzyme activity all gave similar results. Furthermore, cells expressing {alpha}vß3 could not use MMP-2-coated surfaces either to attach or to spread (data not shown). A commercial source of activated MMP-2 (Chemicon), when directly labeled, did however bind {alpha}vß3 as reported previously (14). This binding was blocked both by a competitive {alpha}vß3 integrin inhibitor and by fibronectin (data not shown), a ligand of {alpha}vß3. Similarly, vitronectin was also reported to inhibit the integrin-MMP-2 interaction (48). Interestingly, this was the commercial source (Chemicon) originally used to define the only reported MMP-2-{alpha}vß3 interaction (14). MMP-2 is purified from cell culture supernatants using gelatin affinity chromatography, a technique also used to purify fibronectin (49). Thus, other {alpha}vß3-binding ligands might have been present in the MMP-2 preparation, whose direct labeling and binding led to the original observation of its direct interaction with {alpha}vß3. This hypothesis is further supported by the fact that, in our assays, the interaction of active MMP-2 with {alpha}vß3 was wholly RGD dependent: MMP-2 does not contain the contiguous RGD motives whose obligate interaction with {alpha}vß3 has been defined at the atomic level (50). However, it will not be possible to further investigate the nature of a putative contaminant in this commercial preparation because Chemicon no longer supplies the reagent. An alternative hypothesis is that the putative interaction between MMP-2 and {alpha}vß3 via hemopexin C domain may need an unknown linker protein bearing RGD motives.

Angiogenesis involves {alpha}v integrin binding to ECM targets like proteolysed or denatured collagen and fibrin (9). Inhibitors of {alpha}v integrin can block angiogenesis in vitro and in vivo and are in clinical trials as antiangiogenic cancer therapeutics. Integrin {alpha}vß3 binding to fibrinogen was blocked by EMD 361276 but not by its enantiomer; similarly, EMD 361276 was able to drastically decrease cell invasion in fibrin. In line with previous results with cyclic-RGD peptides (28), EMD 361276 could significantly decrease cell invasion in type I collagen gels, although the primary endothelial collagen receptors, integrins {alpha}2ß1 and {alpha}1ß1, are insensitive to it.9 What then is the source of ligands for {alpha}vß3? First, deposition of serum components, such as fibronectin or vitronectin, may provide ligands for {alpha}vß3. Second, ECM components, such as collagen IV and laminins, produced by the endothelial cells undergoing tubulogenesis as reported in a similar three-dimensional collagen gel model of tubulogenesis using human umbilical vascular endothelial cells (51). Thus, ECM components produced by the cells and/or deposited from serum may provide ligands for the {alpha}vß3 integrin used during in vitro angiogenesis.

Alternatively, proteolytic activity and partial digestion of collagens would be a prerequisite for the cells to disrupt their basal lamina, to invade the matrix, and to organize into tube-like structures. Collagenolytic activity may result in the unmasking of cryptic binding sites for {alpha}vß3 near the surface of invading cells. Recent reports have shown that {alpha}vß3 and MT1-MMP physically interact and colocalize to motility-associated structures and that {alpha}vß3 regulates MT1-MMP internalization and activity in migrating endothelial cells (17, 18). Furthermore, MT1-MMP was reported to process pro-{alpha}v into the mature {alpha}v subunit leading to enhanced focal adhesion kinase phosphorylation, whereas an MT1-MMP-{alpha}vß3 integrin interaction cooperatively increased breast carcinoma cell locomotion (16). Taken together with our study, the evidence suggests that {alpha}vß3 may directly interact with deposited or degraded components of the matrix. {alpha}vß3 could be regulated by MT1-MMP; alternatively, {alpha}vß3 could regulate MT1-MMP activity/localization during endothelial cell migration and tube formation.

In summary, the direct role of MMP-2 in in vitro angiogenesis and its interaction with {alpha}vß3 integrin seems to be minor compared with its interaction with either of its proteolytic substrates or with MT1-MMP. We cannot exclude that in vivo (a) MMP-2 collaborates with MT1-MMP in degrading ECMs more complex than the ones used in this study and affects a late resolution phase of angiogenesis; (b) that the low-affinity interactions of MMP-2 with {alpha}vß3 we report are physiologically important; or (c) that other subtleties are in play. For example, we have used a monoculture, which excludes possible effects of nonendothelial cells like pericytes, macrophages, and fibroblasts. Nor does our model replicate the in vivo situation in that MMP-2 may release or activate matrix-associated growth factors. Hemopexin C domain, by blocking MT1-MMP activation of MMP-2, might in this way inhibit angiogenesis. This could also explain the increase in hemopexin C domain seen during late phases of neonatal murine retinal vascularization. The presence of MMP-2 and its state of activation thus seem to be irrelevant in our in vitro angiogenic model, which we found is dominated by a requirement for MT1-MMP activity.


   Acknowledgments
 
Grant support: Swiss National Science Foundation grant 3100.064037.00 (M.S. Pepper) and Medical Research Council and Wellcome Trust UK (G. Murphy).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank C. Di Sanza and M. Quayzin for expert technical assistance, Drs. M.B. Furie and S.C. Silverstein for providing the BME cells, Dr. D. Cheresh for PEX and MMP-2, Dr. A.G. Arroyo for anti-MT1-MMP antibodies, Dr. A. Jonczyk for EMD 66203 and EMD 69601, Dr. P. Sarmientos for recombinant human FGF-2, Drs. W.G. Stetler-Stevenson and M. Seiki for the human gelatinase A and MT1-MMP cDNAs, respectively, and Dr. P. Brown for BB94 and BB1268.


   Footnotes
 
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

8 S.L. Goodman et al. Indol-3-yl derivatives. Patent WO 01/58893, August 16, 2001. Back

9 S.L. Goodman unpublished. Back

Received 5/ 2/05. Revised 8/ 9/05. Accepted 8/15/05.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Mignatti P, Rifkin DB. Plasminogen activators and matrix metalloproteinases in angiogenesis. Enzyme Protein 1996;49:117–37.[Medline]
  2. Pepper MS. Extracellular proteolysis and angiogenesis. Thromb Haemost 2001;86:346–55.[Medline]
  3. Visse R, Nagase H. Matrix metalloproteinases and tissue inhibitors of metalloproteinases: structure, function, and biochemistry. Circ Res 2003;92:827–39.[Abstract/Free Full Text]
  4. Steffensen B, Bigg HF, Overall CM. The involvement of the fibronectin type II-like modules of human gelatinase A in cell surface localization and activation. J Biol Chem 1998;273:20622–8.[Abstract/Free Full Text]
  5. Tam EM, Moore TR, Butler GS, Overall CM. Characterization of the distinct collagen binding, helicase and cleavage mechanisms of matrix metalloproteinase 2 and 14 (gelatinase A and MT1-MMP): the differential roles of the MMP hemopexin C domains and the MMP-2 fibronectin type II modules in collagen triple helicase activities. J Biol Chem 2004;279:43336–44.[Abstract/Free Full Text]
  6. Tam EM, Morrison CJ, Wu YI, Stack MS, Overall CM. Membrane protease proteomics: isotope-coded affinity tag MS identification of undescribed MT1-matrix metalloproteinase substrates. Proc Natl Acad Sci U S A 2004;101:6917–22.[Abstract/Free Full Text]
  7. Montgomery AM, Reisfeld RA, Cheresh DA. Integrin {alpha}vß3 rescues melanoma cells from apoptosis in three-dimensional dermal collagen. Proc Natl Acad Sci U S A 1994;91:8856–60.[Abstract/Free Full Text]
  8. Hynes RO. Integrins: bidirectional, allosteric signaling machines. Cell 2002;110:673–87.[CrossRef][Medline]
  9. Hodivala-Dilke KM, Reynolds AR, Reynolds LE. Integrins in angiogenesis: multitalented molecules in a balancing act. Cell Tissue Res 2003;314:131–44.[CrossRef][Medline]
  10. Byzova TV, Rabbani R, D'Souza SE, Plow EF. Role of integrin {alpha}(v)ß3 in vascular biology. Thromb Haemost 1998;80:726–34.[Medline]
  11. Clark RA, Tonnesen MG, Gailit J, Cheresh DA. Transient functional expression of {alpha}vß3 on vascular cells during wound repair. Am J Pathol 1996;148:1407–21.[Abstract]
  12. Drake CJ, Cheresh DA, Little CD. An antagonist of integrin {alpha}vß3 prevents maturation of blood vessels during embryonic neovascularization. J Cell Sci 1995;108:2655–61.[Abstract]
  13. Brooks PC, Stromblad S, Sanders LC, et al. Localization of matrix metalloproteinase MMP-2 to the surface of invasive cells by interaction with integrin {alpha}vß3. Cell 1996;85:683–93.[CrossRef][Medline]
  14. Brooks PC, Silletti S, von Schalscha TL, Friedlander M, Cheresh DA. Disruption of angiogenesis by PEX, a noncatalytic metalloproteinase fragment with integrin binding activity. Cell 1998;92:391–400.[CrossRef][Medline]
  15. Petitclerc E, Stromblad S, von Schalscha TL, et al. Integrin {alpha}(v)ß3 promotes M21 melanoma growth in human skin by regulating tumor cell survival. Cancer Res 1999;59:2724–30.[Abstract/Free Full Text]
  16. Deryugina EI, Ratnikov BI, Postnova TI, Rozanov DV, Strongin AY. Processing of integrin {alpha}(v) subunit by membrane type 1 matrix metalloproteinase stimulates migration of breast carcinoma cells on vitronectin and enhances tyrosine phosphorylation of focal adhesion kinase. J Biol Chem 2002;277:9749–56.[Abstract/Free Full Text]
  17. Galvez BG, Matias-Roman S, Albar JP, Sanchez-Madrid F, Arroyo AG. Membrane type 1-matrix metalloproteinase is activated during migration of human endothelial cells and modulates endothelial motility and matrix remodeling. J Biol Chem 2001;276:37491–500.[Abstract/Free Full Text]
  18. Galvez BG, Matias-Roman S, Yanez-Mo M, Sanchez-Madrid F, Arroyo AG. ECM regulates MT1-MMP localization with ß1 or {alpha}vß3 integrins at distinct cell compartments modulating its internalization and activity on human endothelial cells. J Cell Biol 2002;159:509–21.[Abstract/Free Full Text]
  19. Overall CM, Wrana JL, Sodek J. Independent regulation of collagenase, 72-kDa progelatinase, and metalloendoproteinase inhibitor expression in human fibroblasts by transforming growth factor-ß. J Biol Chem 1989;264:1860–9.[Abstract/Free Full Text]
  20. McQuibban GA, Gong JH, Tam EM, et al. Inflammation dampened by gelatinase A cleavage of monocyte chemoattractant protein-3. Science 2000;289:1202–6.[Abstract/Free Full Text]
  21. Overall CM, Tam E, McQuibban GA, et al. Domain interactions in the gelatinase A.TIMP-2.MT1-MMP activation complex. The ectodomain of the 44-kDa form of membrane type-1 matrix metalloproteinase does not modulate gelatinase A activation. J Biol Chem 2000;275:39497–506.[Abstract/Free Full Text]
  22. Pender SL, Tickle SP, Docherty AJ, et al. A major role for matrix metalloproteinases in T cell injury in the gut. J Immunol 1997;158:1582–90.[Abstract]
  23. Murphy G, Willenbrock F. Tissue inhibitors of matrix metalloendopeptidases. Methods Enzymol 1995;248:496–510.[Medline]
  24. Apte SS, Olsen BR, Murphy G. The gene structure of tissue inhibitor of metalloproteinases (TIMP)-3 and its inhibitory activities define the distinct TIMP gene family. J Biol Chem 1995;270:14313–8.[Abstract/Free Full Text]
  25. Kraft S, Diefenbach B, Mehta R, et al. Definition of an unexpected ligand recognition motif for {alpha}vß6 integrin. J Biol Chem 1999;274:1979–85.[Abstract/Free Full Text]
  26. Mehta RJ, Diefenbach B, Brown A, et al. Transmembrane-truncated {alpha}vß3 integrin retains high affinity for ligand binding: evidence for an "inside-out" suppressor? Biochem J 1998;330:861–9.
  27. Mitjans F, Sander D, Adan J, et al. An anti-{alpha}v-integrin antibody that blocks integrin function inhibits the development of a human melanoma in nude mice. J Cell Sci 1995;108:2825–38.[Abstract]
  28. Nisato RE, Tille JC, Jonczyk A, Goodman SL, Pepper MS. {alpha}vß3 and {alpha}vß5 integrin antagonists inhibit angiogenesis in vitro. Angiogenesis 2003;6:105–19.[CrossRef][Medline]
  29. Furie MB, Cramer EB, Naprstek BL, Silverstein SC. Cultured endothelial cell monolayers that restrict the transendothelial passage of macromolecules and electrical current. J Cell Biol 1984;98:1033–41.[Abstract/Free Full Text]
  30. Pepper MS, Ferrara N, Orci L, Montesano R. Potent synergism between vascular endothelial growth factor and basic fibroblast growth factor in the induction of angiogenesis in vitro. Biochem Biophys Res Commun 1992;189:824–31.[CrossRef][Medline]
  31. Montesano R, Orci L. Tumor-promoting phorbol esters induce angiogenesis in vitro. Cell 1985;42:469–77.[CrossRef][Medline]
  32. Taraboletti G, Sonzogni L, Vergani V, et al. Posttranscriptional stimulation of endothelial cell matrix metalloproteinases 2 and 1 by endothelioma cells. Exp Cell Res 2000;258:384–94.[CrossRef][Medline]
  33. Mandriota SJ, Pepper MS. Regulation of angiopoietin-2 mRNA levels in bovine microvascular endothelial cells by cytokines and hypoxia. Circ Res 1998;83:852–9.[Abstract/Free Full Text]
  34. Buckley CD, Pilling D, Henriquez NV, et al. RGD peptides induce apoptosis by direct caspase-3 activation. Nature 1999;397:534–9.[CrossRef][Medline]
  35. Gurrath M, Muller G, Kessler H, Aumailley M, Timpl R. Conformation/activity studies of rationally designed potent anti-adhesive RGD peptides. Eur J Biochem 1992;210:911–21.[Medline]
  36. Morrison CJ, Butler GS, Bigg HF, et al. Cellular activation of MMP-2 (gelatinase A) by MT2-MMP occurs via a TIMP-2-independent pathway. J Biol Chem 2001;276:47402–10.[Abstract/Free Full Text]
  37. Will H, Atkinson SJ, Butler GS, Smith B, Murphy G. The soluble catalytic domain of membrane type 1 matrix metalloproteinase cleaves the propeptide of progelatinase A and initiates autoproteolytic activation. Regulation by TIMP-2 and TIMP-3. J Biol Chem 1996;271:17119–23.[Abstract/Free Full Text]
  38. Tille JC, Wang X, Lipson KE, et al. Vascular endothelial growth factor (VEGF) receptor-2 signaling mediates VEGF-C({delta}N{delta}C)- and VEGF-A-induced angiogenesis in vitro. Exp Cell Res 2003;285:286–98.[CrossRef][Medline]
  39. Haas TL, Davis SJ, Madri JA. Three-dimensional type I collagen lattices induce coordinate expression of matrix metalloproteinases MT1-MMP and MMP-2 in microvascular endothelial cells. J Biol Chem 1998;273:3604–10.[Abstract/Free Full Text]
  40. Lafleur MA, Handsley MM, Knauper V, Murphy G, Edwards DR. Endothelial tubulogenesis within fibrin gels specifically requires the activity of membrane-type-matrix metalloproteinases (MT-MMPs). J Cell Sci 2002;115:3427–38.[Abstract/Free Full Text]
  41. Itoh T, Ikeda T, Gomi H, et al. Unaltered secretion of ß-amyloid precursor protein in gelatinase A (matrix metalloproteinase 2)-deficient mice. J Biol Chem 1997;272:22389–92.[Abstract/Free Full Text]
  42. Martignetti JA, Aqeel AA, Sewairi WA, et al. Mutation of the matrix metalloproteinase 2 gene (MMP2) causes a multicentric osteolysis and arthritis syndrome. Nat Genet 2001;28:261–5.[CrossRef][Medline]
  43. Itoh T, Tanioka M, Yoshida H, et al. Reduced angiogenesis and tumor progression in gelatinase A-deficient mice. Cancer Res 1998;58:1048–51.[Abstract/Free Full Text]
  44. Kato T, Kure T, Chang JH, et al. Diminished corneal angiogenesis in gelatinase A-deficient mice. FEBS Lett 2001;508:187–90.[CrossRef][Medline]
  45. Zhou Z, Apte SS, Soininen R, et al. Impaired endochondral ossification and angiogenesis in mice deficient in membrane-type matrix metalloproteinase I. Proc Natl Acad Sci U S A 2000;97:4052–7.[Abstract/Free Full Text]
  46. Chun TH, Sabeh F, Ota I, et al. MT1-MMP-dependent neovessel formation within the confines of the three-dimensional extracellular matrix. J Cell Biol 2004;167:757–67.[Abstract/Free Full Text]
  47. Silletti S, Kessler T, Goldberg J, Boger DL, Cheresh DA. Disruption of matrix metalloproteinase 2 binding to integrin {alpha}vß3 by an organic molecule inhibits angiogenesis and tumor growth in vivo. Proc Natl Acad Sci U S A 2001;98:119–24.[Abstract/Free Full Text]
  48. Deryugina EI, Ratnikov B, Monosov E, et al. MT1-MMP initiates activation of pro-MMP-2 and integrin {alpha}vß3 promotes maturation of MMP-2 in breast carcinoma cells. Exp Cell Res 2001;263:209–23.[CrossRef][Medline]
  49. Engvall E, Ruoslahti E. Binding of soluble form of fibroblast surface protein, fibronectin, to collagen. Int J Cancer 1977;20:1–5.[Medline]
  50. Xiong JP, Stehle T, Zhang R, et al. Crystal structure of the extracellular segment of integrin {alpha}Vß3 in complex with an Arg-Gly-Asp ligand. Science 2002;296:151–5.[Abstract/Free Full Text]
  51. Bell SE, Mavila A, Salazar R, et al. Differential gene expression during capillary morphogenesis in 3D collagen matrices: regulated expression of genes involved in basement membrane matrix assembly, cell cycle progression, cellular differentiation and G-protein signaling. J Cell Sci 2001;114:2755–73.



This article has been cited by other articles:


Home page
 INT J LOW EXTREM WOUNDSHome page
M. S. Agren and M. Werthen
The Extracellular Matrix in Wound Healing: A Closer Look at Therapeutics for Chronic Wounds
International Journal of Lower Extremity Wounds, June 1, 2007; 6(2): 82 - 97.
[Abstract] [PDF]

Home page
 J. Leukoc. Biol.Home page
H. Piccard, P. E. Van den Steen, and G. Opdenakker
Hemopexin domains as multifunctional liganding modules in matrix metalloproteinases and other proteins
J. Leukoc. Biol., April 1, 2007; 81(4): 870 - 892.
[Abstract] [Full Text] [PDF]

Home page
 J. Biol. Chem.Home page
C. J. Morrison and C. M. Overall
TIMP Independence of Matrix Metalloproteinase (MMP)-2 Activation by Membrane Type 2 (MT2)-MMP Is Determined by Contributions of Both the MT2-MMP Catalytic and Hemopexin C Domains
J. Biol. Chem., September 8, 2006; 281(36): 26528 - 26539.
[Abstract] [Full Text] [PDF]

This Article