Vascular Endothelial Growth Factor Inhibits Anchorage-Disruption-Induced Apoptosis in Microvessel Endothelial Cells by Inducing Scaffold Formation

EXPERIMENTAL CELL RESEARCH ARTICLE NO.

233, 340 –349 (1997)

EX973583

Vascular Permeability Factor/Vascular Endothelial Growth Factor Inhibits Anchorage-Disruption-Induced Apoptosis in Microvessel Endothelial Cells by Inducing Scaffold Formation Yoshifumi Watanabe*,†,1 and Harold F. Dvorak* *Department of Pathology, Beth Israel Hospital and Harvard Medical School, 99 Brookline Avenue, Research North, Boston, Massachusetts 02215; and †Department of Biomolecular Engineering, Tokyo Institute of Technology, 4259 Nagatsuda, Midori-ku, Yokohama 226, Japan

Survival and proliferation of endothelial cells requires both growth factors and an appropriate extracellular matrix to which cells can attach. In the absence of either, endothelial cells rapidly undergo apoptosis. Thus, when human microvascular endothelial cells (HDMEC) are plated on a hydrophobic surface such as untreated polystyrene, they rapidly undergo apoptosis and die. The present study demonstrates that vascular permeability factor/vascular endothelial growth factor (VPF/VEGF), an endothelial cell-selective cytokine, inhibits apoptosis of HDMEC cultured on untreated polystyrene and induces these cells to adhere, spread, and proliferate. VPF/VEGF-induced HDMEC adhesion was time-dependent, required de novo protein synthesis, and was inhibited by a soluble RGD peptide but not by an inhibitor of collagen synthesis. Under the conditions of these experiments, VPF/VEGF downregulated expression of collagen IV and fibronectin but did not change collagen I mRNA levels. VPF/VEGF-induced HDMEC adhesion was inhibited by antibodies to avb5 and vitronectin but not by antibodies to avb3. Other endothelial growth factors and cytokines such as bFGF, HGF, and TGFb did not reproduce the VPF/VEGF effect. We suggest that VPF/VEGF induces endothelial cells to deposit a scaffolding (likely involving vitronectin) that allows them to attach to and proliferate on an otherwise nonsupportive surface (hydrophobic polystyrene) and in this manner serves as both a survival factor and a growth factor. q 1997 Academic Press

INTRODUCTION

Angiogenesis is required for a wide variety of physiological and pathological processes, including embryogenesis, organ development, ovulation, corpus luteum 1 To whom correspondence and reprints should be addressed. Fax: 81-45-924-5815.

340

0014-4827/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.

AID ECR 3583

/

6i21$$$141

formation, and wound repair. The formation and regression of new vessels are regulated processes in which the growth and death of endothelial cells play critical roles. These events are orchestrated by the presence (or the absence) of specific combinations of hormones, growth factors, and extracellular matrix (ECM). Like hormones and growth factors, ECM plays an important role in the regulation of cell growth, differentiation, and behavior [1 –5]. In fact, endothelial cell adhesion to the matrix is an absolute requirement for survival and proliferation in response to growth factors. Without attachment to an appropriate matrix, endothelial cells undergo apoptosis [6, 7]. Some investigators have designated this type of apoptosis ‘‘anoikis’’ in epithelial cells [8]. This mechanism is thought to be crucial for maintaining tissue organization, thereby preventing ectopic cell growth of the type occurring in tumor-associated angiogenesis and retinal neovascularization, which is a major cause of blindness in the United States [9]. Stimuli, such as injury, inflammation, and vasoactive mediators, lead to extravasation and extravascular deposition of various plasma matrix proteins, which in turn stimulate angiogenesis [10 –12]. In these examples, formation of new vessels requires endothelial cell binding to these matrix proteins and the use of deposited matrix as a provisional scaffolding. Acting in concert with growth factors, newly deposited ECM is thought to inhibit the programmed cell death (PCD) of endothelial cells but the mechanism by which this occurs is not clearly understood. Vascular permeability factor/vascular endothelial cell growth factor (VPF/VEGF) is an endothelium-specific growth factor that plays a pivotal role in angiogenesis [13 –19]. VPF/VEGF acts selectively on vascular endothelial cells in vivo to increase microvascular permeability to circulating macromolecules, resulting in deposition of several plasma proteins around the vessels [20, 21]. VPF/VEGF also acts as a survival factor in vivo for the pathological neovascularization found in retinopathy [22]. However, the mode of action of VPF/

05-22-97 18:08:26

eca

VPF/VEGF INHIBITS APOPTOSIS IN EC

VEGF as an endothelial cell survival factor is poorly understood, aside from its known role as a growth factor. In this study we tested the possibility that VPF/ VEGF exerts its function as a survival factor by inducing the deposition of an extracellular matrix scaffold that permits endothelial cell attachment to an unfavorable (hydrophobic) surface that otherwise led to PCD. In contrast to VPF/VEGF, another growth factor, bFGF, did not serve as an endothelial cell survival factor under these conditions. We used early passage microvascular endothelial cells for these studies because new blood vessels arise from microvessels in vivo and because endothelial cells derived from large vessels differ in a number of respects from microvascular endothelium. EXPERIMENTAL PROCEDURES Reagents. Recombinant human vascular cell growth factor (VEGF165), recombinant human basic fibroblast growth factor (bFGF), recombinant human hepatocyte growth factor (HGF), and purified human transforming growth factor b (TGFb) were purchased from R&D Systems (Minneapolis, MN). Plasmids containing human fibronectin and human collagen type I (a1) or IV (a1) inserts were obtained from American Type Culture Collection (ATCC) (Rockville, MD). Anti-avb3 integrin monoclonal antibody (LM609) [23] was purchased from Chemicon International, Inc. (Temecula, CA). Anti-human vitronectin rabbit antiserum [24] was purchased from GIBCO BRL (Grand Island, NY). Anti-human integrin avb5 (P1F6) [23], anti-human integrin b1 (P4C10) [23] monoclonal antibody ascites, cell attachment inhibitory RGD peptide (GRGDTP) [25], and other reagents were purchased from Sigma Chemical Co. (St. Louis, MO). Cell culture. Human microvessel endothelial cells (HDMEC) derived from neonatal foreskins were established as described previously [26] and were kindly provided by Dr. Michael Detmar. HDMEC were cultured on collagen-coated dishes in endothelial cell basal medium (EBM; Clonetics, San Diego, CA) with 20% fetal bovine serum (GIBCO BRL), 50 mM dibutyryl cyclic AMP, 1 mg/ml hydrocortisone acetate, and antibiotics (Sigma). Since the composition of EBM has not been disclosed, we used Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal calf serum for our experiments to eliminate the influences of unknown factors present in EBM. HDMEC at passage 7 to 10 were used. They were plated on nontreated 96well hydrophobic polystyrene plates (Nunc, Roskilde, Denmark) at a density of 1.0 –1.5 1 104 cells/well in the presence or the absence of various additives (see below). At predetermined times, cell number and viability were measured by cell counting with trypan blue exclusion or by using the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay as described below. Cell proliferation was measured by pulse labeling with [ 3H]thymidine (10 mCi/ml) for 4 h. MTT assay. Cell viability was evaluated by the MTT assay [27]. Briefly, MTT was added to cells plated in a 96-well flat-bottomed plate at a final concentration of 500 mg/ml, and the cells were incubated for 4 h at 377C. One hundred microliters of acidic isopropyl alcohol was then added to each well, and the solution was vigorously mixed to solubilize the precipitated dye. The absorbance of each well at 550 nm was measured using a microplate reader, Thermo max (Molecular Devices, Sunnyvale CA). DNA isolation and agarose gel electrophoresis. DNA was isolated according to the method described by Sambrook et al. [28] with minor modification. Briefly, cells were incubated with lysis buffer (10 mg/ ml proteinase K (Sigma), 10 mM Tris– HCl, 150 mM NaCl, 1 mM EDTA, 1% SDS) for 15 h at 377C. Chromosomal DNA was obtained

AID ECR 3583

/

6i21$$$142

05-22-97 18:08:26

341

by phenol:chloroform (1:1) extraction and ethanol precipitation. The samples in TE solution (10 mM Tris– HCl, pH 8.0; 1 mM EDTA) with 1 mg/ml RNase were incubated for 1 h at 377C. The same amount of DNA from each sample was subjected to electrophoresis in a 1.0% agarose gel containing 0.1 mg/ml ethidium bromide. Cell adhesion assay. Trypsinized HDMEC were washed twice and then plated on a nontreated 96-well hydrophobic plate at a density of 3 1 104 cells/well. After incubation for varying times, cells were washed three times with PBS containing 3% BSA and then fixed with 4% paraformaldehyde for 10 min at room temperature. Adherent cells were stained with 0.02% crystal violet for 10 min at room temperature. Cells were then washed with PBS and solubilized with 0.2% SDS. The absorbance of the solution at 570 nm was quantified using a microplate reader. RNA isolation and Northern analyses. Total RNA was extracted with TRI zol (GIBCO BRL), subjected to electrophoresis, and then transferred to a Hybond nylon membrane (Amersham, Arlington Heights, IL). 32P-radiolabeled cDNA probes were prepared using a random primer synthesis kit (GIBCO BRL) that employed purified cDNA inserts as templates. Following hybridization, the blots were washed several times at room temperature in 11 SSC plus 0.1% SDS, followed by 0.11 SSC plus 0.1% SDS, and then exposed to Xray film. Extracellular matrix analysis. Extraction of the extracellular matrix was performed as described [29]. Briefly, cell cultures in 10-mm dishes were fed serum-free DMEM supplemented or not with VPF/ VEGF. After 12 h, the medium was replaced with methionine-free MEM containing 100 mCi/ml [35S]methionine and the corresponding concentration of VPF/VEGF. After 4 h of labeling, the medium was removed and the cultures were washed with cold PBS. The cells were extracted in 3 ml of urea-containing buffer (1 M urea, 1 mM dithiothreitol, 10 mM Tris –HCl, pH 7.4, 10 mM EDTA, 2 mM PMSF). The cells were scraped and vortexed in cold buffer for 5 min and

FIG. 1. Viability of HDMEC on noncoated hydrophobic plastic. Cells were plated on a collagen-coated or nontreated 96-well hydrophobic plastic plate at 1 1 104 cells/well in DMEM containing 10% serum. At sequential intervals, cells were harvested from the supernatant and attached cells by trypsinization. Cell viability was assessed by trypan blue dye exclusion. Data are expressed as means { SD of counts in triplicate from one representative experiment of four. (h) Collagen-coated, (l) noncoated, (m) noncoated with VPF/ VEGF (50 ng/ml), (s) noncoated with bFGF (50 ng/ml).

eca

342

WATANABE AND DVORAK

FIG. 3. DNA degradation of endothelial cells. Cells were cultured on noncoated hydrophobic dishes for 24 h with or without VPF/VEGF or bFGF. They were then lysed and cellular DNA was subjected to gel electrophoresis. Lane 1, DNA from cells cultured on a collagencoated dish. Lane 2, DNA from cells on a noncoated hydrophobic dish. Lane 3, DNA from cells cultured on a noncoated hydrophobic dish with VPF/VEGF (0.5 ng/ml). Lane 4, DNA from cells cultured on a noncoated hydrophobic dish with VPF/VEGF (5 ng/ml). Lane 5, DNA from cells cultured on a noncoated hydrophobic dish with VPF/ VEGF (50 ng/ml). Lane 6, DNA from cells cultured on a noncoated hydrophobic dish with bFGF (50 ng/ml). FIG. 2. Effect of cytokines on endothelial cell viability on a hydrophobic substratum. Cells were plated (1.5 1 104 cells/well) on a 96-well noncoated hydrophobic plate and cultured in DMEM with 10% serum in the presence of VPF/VEGF (m), bFGF (s), and HGF (j) at various concentrations. At 48 h, MTT assay was performed as described under Experimental Procedures.

then pelleted at 13,000g. The supernatants were subjected to SDS– PAGE (4 –15% gradient). After drying, the gel was exposed to X-ray film for 7 days at room temperature.

RESULTS

VPF/VEGF Inhibits Anchorage-Disruption-Dependent Apoptosis of Endothelial Cells The present study employed untreated polystyrene plastic as the cell culture substratum. Due to its highly hydrophobic nature, this material inhibits cell adhesion and is commonly used for growing cells in suspension culture. When many types of adherent cells are plated on this substratum, they fail to attach and undergo apoptosis. As shown in Fig. 1, HDMEC also underwent apoptosis when they were cultured on untreated polystyrene. Similar results have been reported when endothelial cells were cultured on poly (HEMA)coated or BSA-coated tissue culture plates [6, 7]. Untreated polystyrene plastic was used in preference to artificially coated substrates to eliminate possible side effects that might result from uneven coating, etc.

When VPF/VEGF was added to cultures of HDMEC on untreated polystyrene, cells not only survived but they also proliferated (Fig. 1). In contrast, another potent endothelial cell growth factor, bFGF, did not inhibit anchorage-disruption-dependent apoptosis (Fig. 1). When HDMEC were cultured on collagen-coated wells in DMEM –serum, cell number decreased over time (Fig. 1). This probably reflected growth factor deprivation (survival and growth of endothelial cells requires growth factors [30] as well as a matrix for cell attachment). Indeed, when VPF/VEGF was added to HDMEC cultured on collagen, or when HDMEC were cultured on collagen in growth factor-containing EBM medium, cells not only survived but proliferated (data not shown). Another assay of cell viability (the MTT assay) was used to demonstrate that the survival effect of VPF/ VEGF was dose-dependent (Fig. 2). In contrast, two other growth factors, bFGF and HGF, did not show a survival effect for endothelial cells (Fig. 2). As shown in Fig. 3, HDMEC cultured for 24 h on untreated polystyrene exhibited extensive DNA degradation, a hallmark of apoptosis. DNA degradation was inhibited by addition of VPF/VEGF to such cultures in a dose-dependent fashion. HDMEC cultured on collagen-coated wells in DMEM–serum showed some evidence of DNA degradation (Fig. 3), presumably due to lack of necessary growth factors [30].

FIG. 4. Comparison of morphological characteristics of microvascular endothelial cells cultured on a noncoated hydrophobic substratum under various conditions. Cells were plated (3 1 104 cells/well) and cultured on 96-well collagen-coated or noncoated hydrophobic wells in the absence or the presence of VPF/VEGF or bFGF for 36 h. Morphological changes in these cells were photographed by phase-contrast microscopy. (A) Cells cultured on a collagen-coated well. (B) Cells cultured on a noncoated well. (C) Cells cultured on a noncoated well with VPF/VEGF (50 ng/ml). (D) Cells cultured on a noncoated well with bFGF (50 ng/ml). Original magnifications: (left panels) 140; 1220 (right panels).

AID ECR 3583

/

6i21$$$142

05-22-97 18:08:26

eca

VPF/VEGF INHIBITS APOPTOSIS IN EC

AID ECR 3583

/

6i21$$3583

05-22-97 18:08:26

eca

343

344

WATANABE AND DVORAK

In addition to its survival effect, VPF/VEGF also induced significant morphological changes in cultured HDMEC (Fig. 4). As anticipated, cells plated on hydrophobic plastic with or without added bFGF failed to attach, formed cell aggregates, and died (Figs. 4B and 4D). HDMEC plated on collagen-coated wells in DMEM–serum attached and formed monolayer cell islands (Fig. 4A); however, in the absence of added growth factors many cells detached and fragmented. In contrast, cells plated on hydrophobic plastic in DMEM–serum with added VPF/VEGF attached to the substratum and proliferated extensively (Fig. 4C). Of particular interest, viable cells were not confined to a monolayer and exhibited a considerable amount of overlapping (Fig. 4C, right panel). Cellular overlapping of VPF/VEGF-treated cells should not be considered as evidence of cell transformation; we have shown elsewhere that VPF/VEGF does not induce transformation of HDMEC [31]. Rather, it is likely that endothelial cells preferred attachment to each other rather than to the hydrophobic substratum. VPF/VEGF Induces Endothelial Cell Adhesion to and Spreading on a Hydrophobic Substratum Based on the experiments just described, we hypothesized that VPF/VEGF inhibited anchorage-disruptioninduced apoptosis by inducing HDMEC to adhere to a hydrophobic plastic surface. To test this hypothesis, we performed adhesion assays (Fig. 5). First, we demonstrated that the extent of HDMEC adhesion on polystyrene plastic was dependent on the concentration of added VPF/VEGF (Fig. 5A). Second, we demonstrated that VPF/VEGF, but not bFGF or HGF, induced endothelial cell adhesion (Fig. 5B). Third, we demonstrated that adhesion of HDMEC in the presence of added VPF/ VEGF was time-dependent (Fig. 5C); thus, significant adhesion was detected by 4 h after cell plating and reached plateau levels at about 12 h. Finally, we showed that cycloheximide, an inhibitor of protein synthesis, inhibited VPF/VEGF’s ability to promote HDMEC adhesion to polystyrene. Thus, whereas HDMEC adhesion to collagen was unaffected by cycloheximide, adherence of HDMEC to polystyrene in the presence of VPF/VEGF was strikingly inhibited by increasing concentrations of cycloheximide (Fig. 5D). These data provide evidence that VPF/VEGF exerts its effect on promoting HDMEC attachment to polystyrene by inducing the synthesis of one or more proteins that are necessary for cell attachment. To determine whether VPF/VEGF induced HDMEC spreading on polystyrene as well as cell attachment, we stained attached cells for F-actin. In the absence of added VPF/VEGF, very few cells attached and such few cells as were adherent at 24 h after plating were poorly spread and exhibited poorly developed actin filaments

AID ECR 3583

/

6i21$$$143

05-22-97 18:08:26

(Fig. 6A). In contrast, many more HDMEC attached when cells were cultured on noncoated polystyrene with added VPF/VEGF; moreover, these cells exhibited considerable cell spreading and had developed wellformed actin cables (Fig. 6B). Regulation of VPF/VEGF-Induced Attachment and Proliferation of HDMEC on Hydrophobic Plastic We next investigated factors that might modify the ability of VPF/VEGF to cause HDMEC to attach to hydrophobic plastic. As shown in Fig. 7A, a soluble RGD sequence-containing peptide inhibited VPF/ VEGF-induced cell attachment in a dose-dependent fashion. In addition to facilitating cell attachment and spreading, VPF/VEGF also enhanced [3 H]thymidine incorporation of HDMEC cultured on polystyrene and such incorporation was inhibited by GRGDTP (Fig. 7B). In contrast, a relatively specific inhibitor of collagen synthesis, cis-hydroxyproline [32], had only a modest inhibitory effect on [3H]thymidine incorporation. Taken together, these results suggest that both HDMEC attachment to and proliferation on untreated polystyrene wells are dependent on an RGD-containing protein but not on collagen synthesis. Northern analysis gave results consistent with this interpretation (Fig. 8). Thus, expression of collagen type IV and fibronectin was actually downregulated when VPF/VEGF was added to cultures of HDMEC on polystyrene and expression of collagen type I was unchanged. Levels of these proteins on the surface of HDMEC were evaluated by immunofluorescence at 24 h after these various treatments and were found to be unchanged (data not shown). We next attempted to identify the molecule(s) responsible for mediating HDMEC attachment to polystyrene by VPF/VEGF. We found that antibodies to the avb5 integrin and to vitronectin strikingly inhibited the stimulatory effect of VPF/VEGF on [3H]thymidine incorporation by HDMEC cultured on noncoated polystyrene (Fig. 9). In contrast, antibodies to b1 or to avb3 integrins had little inhibitory effect. We also showed that another growth factor (bFGF) and a cytokine that stimulates matrix synthesis (TGFb) could not duplicate the effect of VPF/VEGF in stimulating HDMEC proliferation (Fig. 9). Finally, we investigated the extracellular matrix deposited on untreated polystyrene to determine whether VPF/VEGF altered HDMEC synthesis of extracellular matrix proteins. As shown in Fig. 10, VPF/VEGF strikingly upregulated the synthesis of a 180-kDa protein. This protein has not yet been identified. DISCUSSION

In this study we demonstrated that VPF/VEGF rescues microvascular endothelial cells from anchorage-

eca

VPF/VEGF INHIBITS APOPTOSIS IN EC

345

FIG. 5. Analysis of VPF/VEGF-induced endothelial cell adhesion to a non-ionic substratum. Cells were cultured on a 96-well noncoated hydrophobic plate at a density of 3 1 104 cells/well for varying times with serum. Adherent cells were then quantified as described under Experimental Procedures. (A) Dose-dependency of VPF/VEGF on HDMEC adhesion to a hydrophobic substratum at 12 h. (B) Effects of various endothelial growth factors (bFGF; 50 ng/ml, HGF; 50 ng/ml or VPF/VEGF; 50 ng/ml) on HDMEC adhesion to hydrophobic plastic at 12 h. (C) Kinetics of VPF/VEGF-induced HDMEC adhesion to hydrophobic plastic. (D) Inhibition of VPF/VEGF-induced cell adhesion by cycloheximide (CHX) at 12 h. (j) Cells cultured on collagen-coated wells. (s) Cells cultured on noncoated polystyrene wells with VPF/ VEGF (50 ng/ml).

disruption-dependent apoptosis when they are plated on uncoated (hydrophobic) polystyrene. This substratum inhibits cell attachment and is commonly used for culturing cells in suspension. The approach we have taken differs from that employed in many previous studies of cell adhesion. When cells are cultured on adherent substrata such as collagen or fibronectin, they attach rapidly and firmly by way of integrin receptors. As a consequence, any effects of extracellular matrix

AID ECR 3583

/

6i21$$$143

05-22-97 18:08:26

produced by the attaching cells are masked and can be easily overlooked. In this study we chose a hydrophobic substratum that does not ordinarily permit cell attachment. Therefore, any cell attachment that occurred required activity on the part of the attaching cells. We demonstrated that an endothelial selective cytokine, VPF/VEGF, stimulated HDMEC to attach, spread, and proliferate on a highly unfavorable substrate, uncoated polystyrene. Several other cytokines and growth factors

eca

346

WATANABE AND DVORAK

FIG. 6. Cell adhesion, spreading, and F-actin organization of VPF/VEGF-treated endothelial cells on a hydrophobic substratum. Cells were plated on noncoated hydrophobic plastic for 24 h with or without VPF/VEGF (50 ng/ml). Cells were then fixed with 0.25% glutaraldehyde and stained with Hoechst 33342 (0.01 mM) and TRITC– phalloidin. (A) Cells cultured without VPF/VEGF. (B) Cells cultured with VPF/ VEGF. Magnification: 1170.

did not have this effect. The VPF/VEGF effect was timedependent, developing over the course of at least several hours and achieving plateau levels at 12 h. De novo protein synthesis was required and the effect was inhibited by a soluble RGD-containing peptide and by antibodies to vitronectin or to avb5 integrin. Antibodies to b1 or to avb3 integrins were noninhibitory. The effect was masked when cells were cultured on tissue culturetreated (commercially available, ion-charged) plastic or on collagen-, fibronectin-, vitronectin-, or laminincoated plates (data not shown). Other cytokines/growth factors that stimulate endothelial cell growth (bFGF)

AID ECR 3583

/

6i21$$$143

05-22-97 18:08:26

or stimulate matrix synthesis (TGFb) could not mimic the VPF/VEGF effect. HGF, a cytokine that induces carcinoma cell spreading with rapid (õ1 h) kinetics and independent of new protein synthesis [33], was unable to induce HDMEC to attach to untreated polystyrene. The mechanisms responsible for VPF/VEGF-induced endothelial cell adhesion remain unclear. However, several possibilities exist, including: (1) VPF/VEGF induces cell adhesion by increasing the number or functional capability of ECM receptors, presumably integrins, (2) VPF/VEGF stimulates endothelial cells to synthesize ECM components, or (3) VPF/VEGF stimu-

eca

VPF/VEGF INHIBITS APOPTOSIS IN EC

347

FIG. 8. Northern analysis of extracellular matrix mRNA expression in endothelial cells. Cells (2 1 106 cells/dish) were cultured on noncoated hydrophobic dishes with or without VPF/VEGF (50 ng/ ml) for 12 h. Northern analysis of total RNA (20 mg/lane) isolated from these cells was then performed separately and sequentially with cDNA probes for collagen I, collagen IV, or fibronectin.

microvascular endothelium [34 –36], did not enhance HDMEC adhesion to uncoated polystyrene. Also, although VPF/VEGF-induced HDMEC adhesion and [3 H]thymidine incorporation on uncoated polystyrene

FIG. 7. Parallel inhibition of VPF/VEGF-induced endothelial cell adhesion to and endothelial cell proliferation on a hydrophobic substratum. (A) Cells were plated on a noncoated hydrophobic plate for 15 h with or without VPF/VEGF (50 ng/ml) in the presence of an RGD peptide (GRGDTP) at various concentrations. Adhesion assay was then performed as described. VPF/VEGF-induced cell adhesion was inhibited by the RGD peptide. (B) Cells were plated on a noncoated hydrophobic 96-well plate (1 1 104 cells/well) with or without VPF/VEGF (50 ng/ml) in the presence of an RGD peptide (GRGDTP) (j) or a collagen synthesis inhibitor, cis-hyolroxyproline (s), at various concentrations, cultured for 1 day, and then pulsed with 1 mCi of [3H]thymidine for 4 h.

lates endothelial cells to synthesize proteins that bind serum matrix proteins (e.g., vitronectin (Vn)) either to themselves or to the polystyrene matrix. We have little evidence for the first possibility. bFGF, which increases several types of integrin expression, including avb3 in

AID ECR 3583

/

6i21$$$143

05-22-97 18:08:26

FIG. 9. Effect of various antibodies on VPF/VEGF-induced endothelial cell proliferation on a hydrophobic substratum. Cells were cultured on a hydrophobic 96-well plate (1 1 104 cells/well) for 48 h with bFGF (50 ng/ml), TGFb (50 ng/ml), or VPF/VEGF (50 ng/ml), alone or in the presence of normal rat IgG (10 mg/ml), anti-human integrin avb3 (LM609) (10 mg/ml) antibody, anti-human avb5 (P1F6) (1:500) antibody, anti-human b1 (P4C10) (1:500) ascites, or antihuman vitronectin antiserum (1:500), and then pulsed with 1 mCi of [3H]thymidine for 4 h.

eca

348

WATANABE AND DVORAK

FIG. 10. Analysis of newly synthesized extracellular matrix in endothelial cells. Cells were cultured for 12 h with or without VPF/ VEGF (10 ng/ml). After cells were labeled with [35 S]methionine for 4 h, extracellular matrix was extracted with urea buffer as described under Experimental Procedures. Arrowhead indicates a significantly increased Ç180-kDa band expressed by VPF/VEGF-treated cells.

were strongly inhibited by antibodies to avb5 integrin (Fig. 9), VPF/VEGF did not upregulate b5 expression (data not shown). The second possibility has some merit in that VPF/ VEGF-induced cell proliferation on polystyrene was clearly Vn-dependent (Fig. 9). However, although others have reported that endothelial cells produce Vn [37], we were not able to detect Vn synthesis in our cultures. The third possibility also has merit. VPF/ VEGF induced a striking increase in the synthesis of a 180-kDa protein by HDMEC cultured on untreated polystyrene (Fig. 10). The nature and function of this protein are not known. However, since VPF/VEGF-induced cell attachment and proliferation were apparently Vn-dependent (Fig. 9), this molecule could have a role in binding serum Vn to endothelial cells or to the polystyrene substratum, thus facilitating the attachment of endothelial cells to polystyrene. Further study, especially characterization of the overexpressed 180-kDa protein, will be required to elucidate these mechanisms further. The dependency of VPF/VEGF-induced EC adhesion to inappropriate substrata on Vn, one of the major adhesive proteins in blood plasma, could be very important. Vn is primarily synthesized in the liver [38] and circulates in the plasma. VPF/VEGF induces extravasation of this protein in microvessels by increasing microvascular permeability [20, 21]. We suggest that VPF/VEGF induces not only Vn extravasation but also the construction of a Vn scaffold which allows endothelial cells to attach, spread, and proliferate. We also suggest that these processes are of general importance in angiogenesis induced by VPF/VEGF.

AID ECR 3583

/

6i21$$$143

05-22-97 18:08:26

VPF/VEGF induced HDMEC to downregulate their expression of fibronectin and collagen IV mRNAs. At first, this might seem to contradict our suggestion that VPF/VEGF promotes endothelial cell adhesion by inducing scaffold formation. However, Hitraya et al. also reported that growth factors such as endothelial cell growth factor (ECGF) and FGFs downregulated collagens I and IV and fibronectin production in human microvessel endothelial cells [39]. Thus, both their data and the present findings suggest that such downregulation is commonly induced by growth factors that induce endothelial cell proliferation. The lack of inhibition of VPF/VEGF-induced cell adhesion by antibodies against integrin b1 is consistent with the hypothesis that collagens and fibronectin are not involved in VPF/VEGFdependent induction of cell adhesion. Thus, one possible mechanism of angiogenesis in vivo involves the dependency of endothelial cell survival induced by VPF/VEGF on avb5 integrin. Several studies have demonstrated that avb3 integrin plays a critical role in angiogenesis in vivo [40– 43]. However, Friedlander et al. reported that bFGF-induced angiogenesis requires avb3 integrin, whereas, in the case of VPF/ VEGF-induced angiogenesis, avb5 integrin is predominant in vivo [44]. The present in vitro data are consistent with these reports and may explain the different pathways of angiogenesis induced by bFGF or VPF/ VEGF in vivo. It is interesting that VPF/VEGF-treated HDMEC tend to utilize avb5 rather than avb3 integrin. VPF/VEGF induces the production of plasminogen activator inhibitor-1 (PAI-1) in endothelial cells [45]. PAI-1 binds to receptors in the proximity of the determinant region of Vn [46]. Formation of a PAI-1/Vn complex is thought to sterically inhibit the determinant region of Vn for avb3 integrin because Stefansson et al. recently reported in smooth muscle cells that PAI1 inhibits the binding of Vn to avb3, whereas other sites for other integrin receptors remain intact [47]. It is unlikely that the PAI-1/Vn complex induced by VPF/ VEGF in HDMEC directly caused cell adhesion because bFGF is a more potent inducer of PAI-1 than VPF/VEGF [45], and bFGF did not significantly induce HDMEC adhesion to hydrophobic polystyrene. In conclusion, VPF/VEGF rescues HDMEC from apoptosis when they are cultured on an otherwise inappropriate substratum by inducing the construction of a scaffold that permits cell attachment, spreading, and proliferation. This function may be important for VPF/ VEGF-induced angiogenesis in vivo. The authors thank Dr. Michael Detmar for help with the culture of human microvascular endothelial cells. This work was supported by USPHS NIH Grants CA-50453 and HL-54465 and by salary support from the BIH Pathology Foundation, Inc.

REFERENCES 1. Boudreau, N., Sympson, C. J., Werb, Z., and Bissell, M. J. (1995) Science 267, 891– 893.

eca

VPF/VEGF INHIBITS APOPTOSIS IN EC 2. 3.

Zhu, X., Ohtsubo, M., Boehmer, R. M., Roberts, J. M., and Assoian, R. K. (1996) J. Cell Biol. 133, 391–403. Fang, F., Orend, G., Watanabe, N., Hunter, T., and Ruoslahti, E. (1996) Science 271, 499–502.

4.

Montesano, R., Orci, L., and Vassalli, P. (1983) J. Cell Biol. 97, 1648 –1652.

5.

Kubota, Y., Kleinman, H. K., Martin, G. R., and Lawley, T. J. (1988) J. Cell Biol. 107, 1589 –1598. Meredith, J. E., Jr., Fazeli, B., and Schwartz, M. A. (1993) Mol. Biol. Cell 4, 953–961. Re, F., Zanetti, A., Sironi, M., Polentarutti, N., Lanfrancone, L., Dejana, E., and Colotta, F. (1994) J. Cell Biol. 127, 537– 546. Frisch, S. M., and Francis, H. (1994) J. Cell Biol. 124, 619– 626.

6. 7.

8. 9.

Anonymous (1993) A report of the national advisory eye council. Natl. Inst. Health, Bethesda, Maryland, DHHS Pub. 10. Clark, R. A. (1990) J. Invest. Dermatol. 94, 128S–134S. 11. Gordon, S. R. (1988) J. Histochem. Cytochem. 36, 409– 416. 12. Brown, L. F., van de Water, L., Harvey, V. S., and Dvorak, H. F. (1988) Am. J. Pathol. 130, 455–465. 13. Ferrara, N., Carver-Moore, K., Chen, H., Dowd, M., Lu, L., O’Shea, K. S., Powell-Braxton, L., Hillan, K. J., and Moore, M. W. (1996) Nature 380, 439–442. 14. Thomas, K. A. (1996) J. Biol. Chem. 271, 603– 606. 15. Carmeliet, P., Ferreira, V., Breier, G., Pollefeyt, S., Kieckens, L., Gertsenstein, M., Fahrig, M., Vandenhoeck, A., Harpal, K., Eberhardt, C., Declercq, C., Pawling, J., Moons, L., Collen, D., Risau, W., and Nagy, A. (1996) Nature 380, 435– 439. 16. Shalaby, F., Rossant, J., Yamaguchi, T. P., Gertsenstein, M., Wu, X., Breitman, M. L., and Schuh, A. C. (1995) Nature 376, 62–66. 17. Leung, D. W., Cachianes, G., Kuang, W., Goeddel, D. V., and Ferrara, N. (1989) Science 246, 1306 – 1312. 18. Pierce, E. A., Avery, R. L., Foley, E. D., Aiello, L. P., and Smith, L. E. H. (1995) Proc. Natl. Acad. Sci. USA 92, 905– 909. 19. Shweiki, D., Neeman, M., Itin, A., and Keshet, E. (1995) Proc. Natl. Acad. Sci. USA 92, 768– 772. 20. Senger, D. R., Galli, S. J., Dvorak, A. M., Perruzzi, C. A., Harvey, V. S., and Dvorak, H. F. (1983) Science 219, 983–985. 21. Dvorak, H. F., Orenstein, N. S., Carvalho, A. C., Churchill, W. H., Dvorak, A. M., Galli, S. J., Feder, J., Bitzer, A. M., Rypysc, J., and Giovinco, P. (1979) J. Immunol. 122, 166–174. 22. Alon, T., Hemo, I., Itin, A., Pe’er, J., Stone, J., and Keshet, E. (1995) Nature Med. 1, 1024 – 1028. 23. Wayner, E. A., Orlando, R. A., and Cheresh, D. A. (1991) J. Cell Biol. 113, 919–929. 24. Hayman, E. G., Pierschbacher, M. D., Oehgren, Y., and Ruoslahti, E. (1983) Proc. Natl. Acad. Sci. USA 80, 4003 –4007.

25. Gehlsen, K. R., Argraves, W. S., Pierschbacher, M. D., and Ruoslahti, E. (1988) J. Cell Biol. 106, 925 –930. 26. Detmar, M., Yeo, K. T., Nagy, J. A., Water, L. V. D., Brown, L. F., Berse, B., Elicker, B. M., Ledbetter, S., and Dvorak, H. F. (1995) J. Invest. Dermatol. 105, 44 –50. 27. Mosmann, T. (1983) J. Immunol. Methods 65, 55–63. 28. Sambrook, J., Fritsch, E. F., and Maniantis, T. (1989) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 29. Ignotz, R. A., and Massague, J. (1986) J. Biol. Chem. 261, 4337 – 4345. 30. Araki, S., Shimada, Y., Kaji, K., and Hayashi, H. (1990) Biochem. Biophys. Res. Commun. 168, 1194 –1200. 31. Watanabe, Y., Lee, S. W., Detmar, M., Ajioka, I., and Dvorak, H. F. (1997) Oncogene, in press. 32. Daly, J. M., Steiger, E., Prockop, D. J., and Dudrick, S. J. (1973) J. Surg. Res. 14, 551–555. 33. Matsumoto, K., Matsumoto, K., Nakamura, T., and Kramer, R. H. (1994) J. Biol. Chem. 269, 31807 –31813. 34. Swerlick, R. A., Brown, E. J., Xu, Y., Lee, K. H., Manos, S., and Lawley, T. J. (1992) J. Invest. Dermatol. 99, 715– 722. 35. Enenstein, J., Waleh, N. S., and Kramer, R. H. (1992) Exp. Cell Res. 203, 499–503. 36. Klein, S., Giancotti, F. G., Presta, M., Albelda, S. M., Buck, C. A., and Rifkin, D. B. (1993) Mol. Biol. Cell 4, 973– 982. 37. Berge, V., Johnson, E., Hogasen, K., and Hetland, G. (1992) Scand. J. Immunol. 36, 119–123. 38. Seiffert, D., Crain, K., Wagner, N. V., and Loskutoff, D. J. (1994) J. Biol. Chem. 269, 19836– 19842. 39. Hitraya, E. G., Tan, E. M. L., Rudnicka, L., and Jimenez, S. A. (1995) Lab. Invest. 73, 393–402. 40. Brooks, P. C., Montgomery, A. M. P., Rosenfeld, M., Reisfeld, R. A., Hu, T., Klier, G., and Cheresh, D. A. (1994) Cell 79, 1157 – 1164. 41. Drake, C. J., Cheresh, D. A., and Little, C. D. (1995) J. Cell Sci. 108, 2655 –2661. 42. Brooks, P. C., Stroemblad, S., Klemke, R., Visscher, D., Sarkar, F. H., and Cheresh, D. A. (1995) J. Clin. Invest. 96, 1815 –1822. 43. Brooks, P. C., Clark, R. A. F., and Cheresh, D. A. (1994) Science 264, 569– 571. 44. Friedlander, M., Brooks, P. C., Shaffer, R. W., Kincaid, C. M., Varner, J. A., and Cheresh, D. A. (1995) Science 270, 1500 – 1502. 45. Pepper, M. S., Ferrara, N., Orci, L., and Montesano, R. (1991) Biochem. Biophys. Res. Commun. 181, 902–906. 46. Ciambrone, G. J., and McKeown-Longo, P. J. (1990) J. Cell Biol. 111, 2183 –2195. 47. Stefansson, S., and Lawrence, D. A. (1996) Nature 383, 441– 443.

Received October 19, 1996 Revised version received January 31, 1997

AID ECR 3583

/

6i21$$$144

05-22-97 18:08:26

349

eca