Endothelial cell adhesion on Vas

Document Sample
Endothelial cell adhesion on Vas Powered By Docstoc
					The FASEB Journal express article10.1096/fj.02-0691fje. Published online April 8, 2003.

Extracellular matrix-bound vascular endothelial growth
factor promotes endothelial cell adhesion, migration, and
survival through integrin ligation
Helen Hutchings, Nathalie Ortega, and Jean Plouët

GDR CNRS 1927 – Angiogenese and UMR CNRS 5089, Institut de Pharmacologie et Biologie
Structurale, Toulouse, France

Corresponding author: Jean Plouët, Institut de Pharmacologie et Biologie Structurale, UMR
CNRS 5089, 205 route de Narbonne, 31077 Toulouse cedex, France. E-mail:


Vascular endothelial growth factor (VEGF), a major factor mediating endothelial cell survival,
migration, and proliferation during angiogenesis, is expressed as five splice variants (121, 145,
165, 189, and 206 aminoacids) encoded by a single gene. Although the three shorter isoforms are
mainly diffusible, the two longer ones are sequestered in cell membranes after secretion.
However, their potential role as true components of the extracellular matrix has not been
investigated. We determined that endothelial cells could adhere and spread on VEGF189 and
VEGF165, but not on VEGF121. Adhesion was mediated by the α3β1 and αvβ3 integrins and
other αv integrins but not by the cognate VEGF receptors. Cells migrated on VEGF165 and
VEGF189 and displayed a stellate morphology with numerous lamellopodia and FAK staining but
no actin stress fibers. Tumstatin, an antiangiogenic peptide that interacts with the αvβ3 integrin,
could inhibit adhesion on VEGF, and this effect was potentiated by anti-αvβ3 blocking antibody.
Immobilized VEGF almost totally abolished endothelial cell apoptosis through interactions with
integrins. The inhibition of αvβ3 engagement with immobilized VEGF by tumstatin inhibited
most of its survival activity. We have thus determined a new VEGF receptor-independent role
for immobilized VEGF in supporting cell adhesion and survival through interactions with

Key words: angiogenesis • VEGF • tumstatin

        ngiogenesis, the formation of new blood vessels from the preexisting vasculature, is a
        critical process contributing to embryonic vessel development, wound healing, corpus
        luteum formation, and pathological processes such as tumor development, diabetic
retinopathies, and rheumatoid arthritis. The process of angiogenesis can be divided into several
overlapping phases. It begins with an increase of local blood vessel permeability and
extravasation of plasma proteins such as fibrinogen at the site where the new blood vessel is to
sprout. Degradation of the basement membrane by proteases of the metalloproteinase family
allows for proliferation and migration of endothelial cells toward the angiogenic stimulus.
Finally, during the maturation stage, endothelial cells differentiate, form a lumen and a new
basement membrane, and recruit pericytes to support the structure of the new vessel.
Vascular endothelial growth factor (VEGF) is a key regulator of both physiological and
pathological angiogenesis (1). VEGF was first described as an endothelial cell-specific mitogen
(2, 3), but it is now known to act on certain other cell types (4, 5). VEGF exists in five isoforms--
121, 145, 165, 189, and 206 amino acids--which are generated by alternative splicing of the pre-
mRNA, and these isoforms can be further matured by proteolytic cleavage. VEGF165, a 46-kD
homodimeric glycoprotein, is the most abundant and most intensely studied of the five isoforms.
Although the five isoforms can be cleaved by plasmin and generate a 110-amino acid factor,
which does not bind to heparin, only exon 6-containing isoforms are cleaved by urokinase (6).
However, little is known about the specific biological responses they might induce.

Cellular responses to VEGF165 are mediated by two high-affinity type III tyrosine kinase
receptors, KDR/Flk1 (7) and Flt (8), and two receptors of the semaphorin receptor family,
neuropilin-1 (9) and neuropilin-2 (10). Gene targeting studies have shown that VEGF, KDR, and
Flt1 are essential for fetal angiogenesis (11–13). The loss of even a single allele of the VEGF
gene is lethal, underlining the critical role of this factor. In VEGF null mice embryos, most steps
of early vascular development are impaired: endothelial cell differentiation, sprouting from
preexisting vessels, lumen formation, formation of vessel interconnections, and spatial
organization (11). The phenotype of VEGF null mice illustrates the multiple roles that VEGF
could have during vascular development: as a differentiation factor for angioblasts, as a mitogen
for endothelial cells, and as a mediator of cell-cell and cell-matrix interactions.

Little is known about VEGF’s role in mediating cell-matrix interactions. The extracellular matrix
(ECM) gluttonously sequesters VEGF soon after its secretion on heparan-sulfates (14) and
probably another binding site specific for exon 6 (6). Accordingly, systemic injections of VEGF
have very little effect because all the protein is absorbed by the ECM at the site of injection or in
the vascular wall. It has been shown recently that cells could adhere directly on cytokines
involved in angiogenesis. For instance, Carlson and coworkers demonstrated that cells could
adhere both on angiopoietin-1 (Ang-1) and angiopoietin-2 (Ang-2) but only spread and migrate
on Ang-1 (15). Several integrins were shown to be involved in adhesion on Ang-1 and Ang-2,
and the α5 integrin subunit had a particulary important role. Latent forms of transforming growth
factor β (TGF-β) have also been shown to be a substrate for cell adhesion, and this adhesion was,
as for the angiopoietins, mediated by integrins (16).

Here, we demonstrate that endothelial cells can adhere directly and spread on two immobilized
VEGF isoforms: VEGF165 and VEGF189. Endothelial cells clearly show a preference for
VEGF189 as an adhesion substrate because adhesion reaches higher levels and occurs at a faster
rate than on VEGF165. We propose the existence of additional cell binding sites for the
immobilized forms of VEGF, which include the α3β1 and αvβ3 integrins, and we have
determined that immobilized VEGF induces endothelial cell migration and survival through
direct αvβ3 ligation. Moreover, tumstatin, a fragment of collagen IV, reversed the effect of
VEGF on endothelial survival.


Human umbilical arterial endothelial cells (HUAEC) were isolated from dissected umbilical
arteria perfused with collagenase (Sigma, St. Louis, MO) to digest the basal membrane. HUAEC
were maintained in SFM (Life Sciences) supplemented with 20% heat-inactivated fetal calf
serum (FCS). Stock cultures received 1 ng/ml VEGF every other day. Human bone marrow
endothelial cells (HBMEC) were a kind gift from L. Pelletier and were cultured in the same
medium. Bovine retinal endothelial cells (BREC) and fetal aorta (FBAE) endothelial cells were
isolated from organs obtained from a local slaughterhouse as described previously (17). Cells
were maintained in Dulbecco’s modified Eagle’s medium (DMEM) glutamax (Life Sciences)
supplemented with 10% heat-inactivated newborn calf serum (NBCS), 100 µg/ml penicillin, and
100 µg/ml streptomycin at 37°C in 10% CO2. Rhodamine-conjugated phalloidin was purchased
from Molecular Probes (Eugene, OR). rhVEGF121, rhVEGF165, and rhVEGF189 were produced
using a baculoviral expression system and iodinated as described previously (6). rhVEGF110 was
produced by plasmin digestion of rhVEGF165 as described previously (18). Fibronectin, collagen
I, poly-lysine, heparin, and bovine serum albumin (BSA) were from Sigma. αvβ3 recombinant
protein was purchased from Chemicon (Temecula, CA). The RGD peptides and RAD control
peptides were also purchased from Sigma. The α3β1 integrin blocking peptide, p678
(FQGVLQNVRFVF), and its control peptide, p690 (FQGVLQNVAFVF) (19), as well as the
tumstatin peptide, (TMPFLFCNVNDVCNFASRNDYSYWL), were synthesized by Eurogentec
(Philadelphia, PA). The peptides corresponding to the VEGF amino acid sequences 119–134
(P2: RGKGKGQKRKRKKSRY) and 126–134 (P1: KRKRKKSRY) were synthesized by a
solid-phase technique. The anti-β1 (6S6), anti-αv subunit, and anti-αvβ5 (P1F6) blocking
antibodies were purchased from Chemicon, the αvβ3 blocking antibody was produced from the
HB-11029 hybridoma (ATCC), and the control monoclonal antibody was from Santa Cruz
Biotechnology (Santa Cruz, CA). The anti-β3 (RM13) antibody was purchased from Bioline
(Milan, Italy). The anti-FAK (sc 517) antibody was from Santa Cruz Biotechnology, and the
FITC conjugated anti-rabbit secondary antibody was from Sigma.

Cell adhesion assays

Ninety-six-well ELISA plates (Nunc, Rochester, NY) were coated with protein diluted in
carbonate buffer, 0.05 M, pH 9.6, overnight at 4°C. Nonspecific binding sites were blocked for 1
h at 37°C with 5 mg/ml BSA in carbonate buffer and washed twice with DMEM before
experiments. Cells were trypsinized, washed, and resuspended in 5 ml of DMEM with 10% FCS
in an untreated plastic tube and incubated for 1 h at 37°C with 10% CO2. Cells were then
pelleted and resuspended in DMEM and 0.2% BSA without serum, and the cell suspension was
treated for 20 min (37°C, 10% CO2) with the proteins, peptides, or antibodies used to modulate
adhesion. Cells (40,000 per well; three wells per condition) were distributed into the wells in a
volume of 100 µl DMEM and 0.2% BSA. Cells were allowed to adhere at 37°C, 10% CO2, for
the desired time. Wells were washed gently three times with DMEM to remove nonadherent,
cells and adherent cells were fixed with 1% glutaraldehyde for 20 min at room temperature.
Fixed cells were quantified by crystal violet incorporation (20): Cells were incubated with 0.1%
crystal violet (Sigma) diluted in 0.2 M borate buffer, pH 9.5, for 20 min at room temperature,
nonincorporated dye was eliminated by thoroughly washing the wells with large amounts of
water, and incorporated crystal violet dye was then solubilized by 100 µl of 10% acetic acid per
well. Optical density readings were performed at 595 nm. T1/2 is the time for which cell adhesion
has attained its half-maximal level.

Solid-phase αvβ3 binding assay

Ninety-six-well ELISA plates (Nunc) were coated with VEGF isoforms, BSA, or fibronectin at 4
µg/ml diluted in carbonate buffer, 0.05 M, pH 9.6, overnight at 4°C. Nonspecific binding sites
were blocked for 1 h at 37°C with 5 mg/ml BSA in carbonate buffer. Wells were washed and
incubated for 2 h with different concentrations of αvβ3 recombinant protein in PBS, 5 mg/ml
BSA, and 0.05% Tween 20. Binding of αvβ3 to VEGF isoforms was revealed by incubation
with an anti-β3 antibody (1/1000), followed by an horseradish peroxidase (HRP)-conjugated
anti-rabbit secondary antibody. Plates were developed with Sigma Fast OPD tablets, and optical
densities were read at 492 nm.


For all experiments, glass Labtek slides (Nunc) were coated with protein diluted in carbonate
buffer, 0.05 M, pH 9.6, overnight at 4°C. Nonspecific binding sites were blocked for 1 h at 37°C
with 5 mg/ml BSA in carbonate buffer. Cells were harvested by trypsinization, washed, and
resuspended in DMEM with 10% NBCS for bovine cells or FCS for human endothelial cells in
an untreated plastic tube and incubated for 1 h at 37°C with 10% CO2. Cells were then pelleted,
resuspended in DMEM and 2% BSA, and seeded at ~25% confluency in the Labtek wells. After
2 h at 37°C, cells were fixed with 3.7% paraformaldehyde and 15 mM sucrose in PBS for 20 min
at room temperature and permeabilized with 0.3% Triton-X100 in PBS for 5 min at room
temperature. Nonspecific binding sites were blocked by incubation with 5 mg/ml BSA for 2 h at
room temperature. For F-actin labeling, cells were incubated for 20 min at room temperature
with rhodamine-conjugated phalloidin. For anti-FAK labeling, cells were first incubated with an
anti-FAK antibody (1/100) for 1 h at 37°C and then with an anti-rabbit FITC-conjugated
secondary antibody (1/100) for 45 min at 37°C. The plastic wells of the Labtek slides were then
removed, and slides were mounted with Paramount (Dako, Glostrup, Denmark) and analyzed
with a Zeiss fluorescence microscope.

Migration assays

Cells were allowed to adhere for 2 h in plastic dishes coated with either 4 µg/ml VEGF165,
VEGF189, fibronectin, or BSA, as described previously. After this incubation period, dishes were
washed thoroughly with DMEM to eliminate nonadherent cells. Cell movement was monitored
by successive photography of cells at a given spot of the dish. Image overlay was used to
evaluate the migration patterns of individual cells.

Annexin V-fluorescein isothiocyanate assay

Annexin V, a calcium-dependent phospholipid binding protein with a high affinity for
phosphatidylserine was used to detect apoptosis. In brief, HUAEC were seeded in 12-well
culture plates previously coated with VEGF165, VEGF189, or collagen I (0.5 × 106 cells/well) in
DMEM and 2% FCS in the presence or absence of anti-αvβ3 antibody (1/3). Cells were left to
adhere for 2 h, and then 10 µg/ml tumstatin was added. Three hours later, floating and attached
cells were harvested and processed as described previously (21).

Statistical analysis

Data are mean ±SD. Statistical comparisons were made by ANOVA followed by an unpaired
Student’s t test. Data were considered significantly different if values of P<0.05 were observed.


Endothelial cell adhesion to VEGF165 and VEGF189

To determine whether endothelial cells could adhere to VEGF, HUAEC were plated on plastic
wells that had been coated with BSA, fibronectin, VEGF121, VEGF165, or the full-length or
cleaved forms of VEGF189. As shown in Figure 1A, HUAEC adhered well to the VEGF189
isoforms (cleaved or full-length) compared with BSA within 20 min after plating. Cells adhered
moderately well to VEGF165 (60% of maximum adhesion reached on VEGF189) and not at all to
VEGF121. The amount of cells adhered to VEGF189 was similar to that on fibronectin, which is
one of the best adhesion substrates for endothelial cells. Differences were shown not to be due to
variable absorption rates between the VEGF isoforms, becauses we determined that 12% and 9%
of VEGF165 and VEGF189, respectively, bound to plastic. VEGF121 adhered as well as the other
isoforms to plastic and inhibited the binding of iodinated VEGF165 and VEGF189 with a similar
efficacy (IC50 = 7 µg/ml) when it was coincubated with 4 µg/ml of either long isoform (data not
shown). This decrease of the amount of longer isoforms immobilized on plastic achieved by
coincubation with VEGF121 reduced HUAEC adhesion in a dose-response manner (IC50 =
8µg/ml for VEGF165 and 15 µg/ml for VEGF189).

HUAEC adhesion on VEGF was comparable, on a molar basis, to that on other ECM proteins
(data not shown). Maximal adhesion levels on VEGF165 and VEGF189 were attained at a coating
concentration of 100 nM (parallel experiments showed that maximal absorption of VEGF on the
plastic was not achieved at this concentration and therefore that the amount of VEGF bound to
the plastic was not a limiting step for cell adhesion).

Microscopic observation of the cells during adhesion assays revealed that HUAEC adhered
extremely rapidly to VEGF189. We therefore decided to study adhesion in a time-dependent
manner. As expected, HUAEC adhered much faster to VEGF189 or to fibronectin than to
VEGF165 (Fig. 1B) (T1/2 = 7.5 min).

In a second set of experiments, we evaluated the involvement of heparan-sulfate proteoglycans in
cell adhesion on VEGF. Soluble heparin (at concentrations >50 nM) inhibited HUAEC adhesion
to VEGF189; maximal inhibition represented ~50% total adhesion (Fig. 2A). On VEGF165,
heparin increased rather than inhibited cell adhesion. Heparin had no effect on adhesion to

Because HUAECs did not adhere to VEGF121 (which is encoded by exons 1–5 and 8), we
investigated whether peptides derived from exon 6 of VEGF could inhibit cell adhesion. High
concentrations of peptides P1 and P2 (20 µg/ml) significantly delayed HUAEC adhesion to
VEGF (Fig. 2B). The longer peptide, P2, was a better inhibitor than P1. Neither P1 nor P2 had
any effect on HUAEC adhesion to fibronectin. The inhibitory activity of these two peptides
seems to be mostly due to their basic charge, because they both inhibit adhesion to VEGF165,
which does not contain the exon 6 from which they are derived. Strong concentrations of these
peptides were needed to obtain any effect; this could be due to a difference in the conformations
of soluble and immobilized VEGF domains.

Integrin blocking agents inhibit adhesion

We next determined which cell surface binding sites were involved in adhesion to the two VEGF
isoforms. Adhesion assays showed that pretreatment of HUAEC with soluble VEGF165 or
VEGF189 had no significative effect on cell adhesion (Fig. 3). Because soluble VEGF should
induce internalization of known VEGF receptors or at least block access to them, we conclude
that the molecules that interact with the immobilized VEGF and mediate cell adhesion differ
from known VEGF receptors.

To test whether or not integrins may serve as receptors for immobilized VEGF, adhesion
experiments were performed in the presence of the calcium chelator EDTA, a universal inhibitor
of integrins; RGD peptides, which inhibit integrins binding this sequence; and p678, a specific
α3β1 blocking peptide. Blocking peptides for RGD integrins and α3β1 were used at maximal
concentrations effective for blocking adhesion--500 µM (15, 22) and 20 µM (19), respectively.
As shown in Figure 3, addition of these inhibitors revealed different adhesion profiles for
HUAEC on VEGF165 and VEGF189. Whereas EDTA inhibits all adhesion on both VEGF
isoforms, the α3β1 blocking peptide, p678, completely blocks adhesion on VEGF165 but only
~60% of that on VEGF189. RGD peptides had little effect alone, but when used together with
p678, all adhesion was inhibited on VEGF189. Time-course experiments with HUAEC showed
that adhesion was not completely abolished on VEGF165 or VEGF189 by addition of p678 or
RGD peptides (Fig. 4). However, all adhesion was inhibited on both VEGF isoforms by addition
of a combination of p678 and RGD peptides. These results demonstrate that HUAEC adhere to
VEGF165 and VEGF189 uniquely via integrins: the α3β1 integrin and one or more of the RGD
sequence binding integrins. Similar results were obtained with aortic endothelial cells and two
capillary-derived endothelial strains cultured from bone marrow or retina (Table 1). These
findings underline that the observations were not a particular feature of artery endothelial cells.

Next, we used blocking antibodies against specific integrins. Figure 5 shows time-course
adhesion experiments in the presence of antibodies directed against the β1 integrin subunit, the
αv integrin subunit, and the αvβ3 and αvβ5 integrins. β1 and αvβ3 blocking antibodies both
partially inhibit adhesion on VEGF165. The αv blocking antibody strongly inhibits adhesion on
VEGF165 (from 100% inhibition after 7.5 min to ∼65% inhibition after 40 min). Adhesion is
dependent on integrins β1 and αvβ3 on VEGF165. Because a combination of these antibodies
inhibits practically all adhesion (Fig. 5A), the small remaining fraction of adhesion seems to be
due to another αv-subunit integrin (but not αvβ5, data not shown). On VEGF189, initial adhesion
can be significantly blocked with the anti-αvβ3 antibody (Fig. 5B). The β1-subunit antibody
only very slightly inhibits adhesion but enhances inhibition by the αvβ3 and αv subunit blocking
antibodies. About 50% of adhesion on VEGF189 could not be attributed to any of the tested
integrins and must depend on other RGD integrins that do not possess a αv subunit.
αvβ3 binds directly to immobilized VEGF

In a solid-phase binding assay, we investigated whether recombinant αvβ3 could bind directly to
VEGF isoforms absorbed on plastic. As shown in Figure 6, we observed that αvβ3 bound to
immobilized VEGF165 and VEGF189 isoforms but not to VEGF121. αvβ3 bound to VEGF165 and
VEGF189 as strongly as it did to fibronectin, a well-characterized αvβ3 ligand. Control
experiments with a nonspecific primary antibody showed optical densities comparable to
background signaling on BSA (data not shown). These results demonstrate that VEGF121 is not a
ligand for αvβ3 and therefore confirms that the interaction of αvβ3 and VEGF occurs through
the basic domain of VEGF corresponding to exons 6 and 7.

Endothelial cells spread on VEGF but display no actin stress fibers

Endothelial cell adhesion to ECM proteins usually induces reorganization of the actin
cytoskeleton with a great number of filaments of polymerized actin that attach the cells to their
substrate via focal adhesions at the abluminal membrane. We examined the cell morphology
induced by VEGF. As seen in Figure 7A, HUAEC displayed a classical well-organized actin
cytoskeleton with many actin stress fibers on fibronectin. On VEGF165 and VEGF189, cells spread
well compared with cells attached to poly-lysine but presented a very different morphology.
Cells adhering to VEGF display a stellate morphology with numerous lamellopodia (HUAEC)
and filipodia. Although polymerized actin could be discerned in these membrane protrusions, no
actin stress fibers were observed and the filamentous actin appeared to be organized into large
bundles in the cytoplasm.

Because spreading of cells on ECM components requires integrin clustering and subsequent
activation of downstream signaling molecules, we investigated whether focal adhesion kinase
(FAK) localized at focal adhesions in cells adhering to VEGF. FAK immunofluorescence
experiments revealed that cells adhering to both VEGF165 and VEGF189 showed typical FAK
localization at focal adhesions (Fig. 7B). On VEGF, FAK clustering was mostly concentrated at
the edges of the cells, whereas FAK clusters were observed all over the abluminal membrane of
cells on fibronectin.

VEGF supports cell migration

The ability of cells to migrate on VEGF165 and VEGF189 was tested next. After plating, cells
were observed over a 4-h period, and successive photography was used to analyze movement.
VEGF189 supported strong HUAEC migration in the absence of any other ECM components or
serum. This migration was comparable to that seen on fibronectin (Fig. 8). Migration on
VEGF189 could not be attributed to its basic charge because endothelial cells plated on poly-
lysine, a highly basic substrate, did not migrate at all. Migration of cells on VEGF165 was
significantly lower than on VEGF189 over the same time period but was however statistically
significant compared with cells on poly-lysine.
Tumstatin inhibits the αvβ3-dependent effects of immobilized VEGF on adhesion and

Because tumstatin is known to interact with αvβ3 and induce endothelial cell apoptosis, we
tested its activity on cells plated on VEGF165 or VEGF189. Tumstatin inhibited HUAEC adhesion
to both VEGF isoforms in a dose-dependent manner (Fig. 9). This effect was enhanced when
cells were treated simultaneously with an anti-αvβ3 blocking antibody.

Next, we compared the survival rates of HUAEC adhered to VEGF or to collagen I. Cells were
plated on VEGF165, VEGF189, or collagen I and treated with either an anti-αvβ3 blocking
antibody, tumstatin, or both. Apoptosis levels were analyzed by FACS after Annexin-V
isothiocyanate and propidium iodide staining. Table 2 shows that VEGF165 and VEGF189 protect
HUAEC from apoptosis at a level similar to that observed in the presence of collagen I.
Tumstatin and an anti-αvβ3 antibody induced a synergistic effect on apoptosis of HUAEC
adhering to either VEGF isoform, whereas they had no combined action on HUAEC seeded on
collagen I. This sugests that ligation of αvβ3 to immobilized VEGF generates a survival signal
that opposes the proapoptotic action of tumstatin. However, all the survival effecst of VEGF
could not be accounted for by αvβ3 because none of the treatments could induce more than 60%
of apoptosis, thus demonstrating that other integrins involved in adhesion could also be involved
in survival signals.


Our results show for the first time that endothelial cells adhere directly to immobilized VEGF
isoforms VEGF165 and VEGF189 and that this adhesion is mediated by integrins. We find that the
basic moieties of VEGF trigger adhesion mechanisms, because the anionic form did not promote
cell adhesion, whereas VEGF165 and VEGF189 supported endothelial cell spreading, migration,
and survival.

Cell surface VEGF binding sites have been identified as two tyrosine kinase receptors (namely,
VEGFR1 and VEGFR2), two semaphorin receptors (neuropilin-1 and neuropilin-2), and
heparan-sulfate proteoglycans. It is unlikely that adhesion on VEGF is mediated by its tyrosine
kinase receptors because the binding domains of VEGF to its receptors are present in VEGF121,
which cannot mediate adhesion. VEGF121 does not interact directly with αvβ3 either, one of the
integrins largely responsible for adhesion on VEGF165 and VEGF189. Additional evidence comes
from the fact that no difference was observed between adhesion on VEGF189, which does not
bind to VEGFR2, and its cleaved counterpart uPA-VEGF189, which does (6). These results are
again strengthened by the fact that peptides derived from exon 6, which does not bind either
VEGFR1 or VEGFR2, partially inhibit cell adhesion. Addition of soluble heparin partially
inhibited cell adhesion on VEGF189, but not on VEGF165, which suggests that nonspecific
electrostatic interactions between the anionic charges of heparin and the basic sequence encoded
by exon 6 partially mask VEGF189 binding sites for endothelial cells. On VEGF165, soluble
heparin slightly enhanced cell adhesion, which suggests that as for it’s binding to VEGFR2,
soluble heparin might help to give recombinant VEGF165 a more active conformation and thus
enhance adhesion. VEGF145 and VEGF206 that contain exons 6a or 6b and 7 bind strongly to the
ECM, and therefore they could be expected to act as substrates for cell adhesion as well.
However, cell pretreatment with chlorate, which inhibits the sulfatation of proteoglycans and
reduces their anionic charge, had no effect on adhesion (data not shown). These results contrast
with the observation that heparan-sulfate proteoglycans could serve as coreceptors for cell
adhesion on Cyr61 mediated by the α6β1 integrin (23).

We observed an important difference in the way that endothelial cells adhere to VEGF165 and
VEGF189, which suggests that neuropilins, which interact with the basic sequence of VEGF
encoded by exon 7 (9), are not involved in endothelial cell adhesion to VEGF. Little is known
about the putative distinct functions of VEGF189 compared with VEGF165. VEGF189, in contrast
to VEGF165, is rapidly sequestered by heparan-sulfate proteoglycans of the ECM (14) and
another as yet unidentified binding site (6).

We show here that all cell adhesion on VEGF165 can be attributed to the α3β1 and αvβ3
integrins and another αv subunit containing integrin (but not αvβ5). On VEGF189, the same
integrins are involved (α3β1, αvβ3, and another αv integrin); however, ~50% of adhesion is
mediated by a yet unidentified receptor. If the results shown in Figure 4 are to be taken into
account, this receptor should be an RGD integrin, because a combination of α3β1 blocking
peptides and RGD peptides inhibit all adhesion on VEGF189. The only known RGD integrin that
does not contain either a β1 subunit or a αv subunit is αIIbβ3. However, it could be possible that
the RGD peptides might nonspecifically block other nonintegrin cell surface receptors.

Several hypotheses could explain the differences (maximal levels and rapidity) between adhesion
on VEGF165 and VEGF189. First, a simple explanation could be that α3β1 and αvβ3 integrins
bind with a greater affinity to VEGF189 than to VEGF165. It has been shown that in some cases,
integrins must be activated to be able to induce a functional response, so it is plausible that
contact with immobilized VEGF189 could activate this integrin much faster than immobilized
VEGF165. It has been shown that soluble VEGF can induce activation of several integrins (24). In
our case, immobilized VEGF could act in the same way and the efficiency of integrin activation
could depend on the VEGF isoform. Another hypothesis is that the yet unidentified receptors for
VEGF189 are responsible for the observed differences. This would be a very interesting finding
because it would mean that VEGF189 possesses a specific cell surface receptor that could mediate
isoform-specific responses (such as cell migration).

The clear preference that endothelial cells show for adhesion to VEGF189 compared with
VEGF165 is not translated by different cell morphology. In contrast to cells plated on fibronectin,
which display prominent stress fibers, cells on VEGF possess numerous lamellopodia and
filopodia but no actin stress fibers. The cell shape on VEGF can be compared with that seen by
Waltenberger and coworkers who observed the appearence of lamellopodia and cytoskeletal
reorganization when soluble VEGF was added to endothelial cultures (25). However, the
morphology of cells attached to VEGF and those treated by soluble VEGF is quite distinct
because soluble VEGF did not cause stress fibers to disappear. Actin dynamics appear to be
coordinated in time and space by intracellular signaling pathways, and it has been shown that
two separate signaling pathways acting on actin reorganization can be initiated by integrins (26,
27). Engagement of integrins with the ECM leads to activation of Cdc42, which subsequently
leads to activation of Rac, and together, these GTPases mediate cell spreading and formation of
lamellopodia and filopodia. Integrins can also activate Rho independently of Rac, and this leads
to formation of stress fibers. After observation of endothelial morphology on VEGF, one could
imagine that only one of these signaling pathways has been activated--the Cdc42-Rac pathway,
leading to formation of lamellopodia and filopodia. The functionality of this morphology is yet
to be determined, but the absence of stress fibers does not weaken the cells because they survive
on VEGF189 in serum-deprived conditions just as long as their counterparts plated on fibronectin.

Cells on VEGF may not display actin stress fibers, but they show signs of active integrin
signaling. FAK, a major integrin signaling molecule, is localized at focal adhesions in cells
adhering to VEGF. Integrin clustering and recruitment of several adaptator molecules and
kinases is known to be a key step in the process of cell spreading and migration. Our results
show the presence of focal adhesions in cells adhering to VEGF, and it is therefore possible that
integrins clustered at these focal adhesions induce active signal transduction. Note that focal
adhesions are found mostly near the edges of cells on VEGF, whereas they can be observed
throughout cells on fibronectin. This particularity could be related to the cell morphology
observed on VEGF, where actin is not organized into stress fibers. This underlines the fact that
the role of VEGF immobilized in the ECM is probably not to induce the formation of an actin
stress fiber network but to provide the cell with a specific growth factor-like stimulus (e.g.,
migration or antiapoptosis).

Although cell morphology on VEGF165 and VEGF189 is very similar, a functional distinction
appears when cells are tested for their capacity to migrate on a surface coated with the two
isoforms. After 4–6 h in the absence of serum, cells migrate on VEGF189 as much as on
fibronectin. However, although migration on VEGF165 is significant compared with that on poly-
lysine, it is much lower than on VEGF189. This experiment describes a distinct role for VEGF189,
which does not involve, at least in the initial stages, known VEGF receptors. This VEGF189-
specific activity could be due to release of FGF2, because we have already shown that soluble
VEGF189 added to corneal endothelial cells that do not express VEGF receptors induces cell
proliferation and migration through FGF2 release from the cell membranes (28).

To our knowledge, this is the first work demonstrating that VEGF can be used by endothelial
cells as an adhesion substrate but not the first that shows that cytokines can support cell
adhesion. Carlson and coworkers recently showed that angiopoietins Ang-1 and Ang-2, two
others factors involved in angiogenesis, could serve as substrates for cell adhesion (15). Latent
forms of TGF-β can also support cell adhesion via integrins (16, 29), and factor XIIIa can
support microvascular endothelial cell adhesion (30).

In the last part of our work, we observed the action of tumstatin on cells attaching to
immobilized VEGF. Both immobilized VEGF and tumstatin interact directly with the αvβ3
integrin: VEGF promotes endothelial cell adhesion, whereas tumstatin induces apoptosis (31,
32). In our experiments, tumstatin inhibits adhesion to VEGF (both VEGF165 and VEGF189) in a
dose-dependent fashion. This inhibition is potentiated by an anti-αvβ3 blocking antibody, but
both treatments cannot fully reverse the survival activity of VEGF, thus demonstrating that other
integrins such as α3β1 involved in adhesion may also play a role in apoptosis protection.
However, tumstatin and anti-αvβ3 blocking antibodies do not have a combined effect on
inducing apoptosis in cells attached on collagen I, which is not a substrate for αvβ3. This
suggests that immobilized VEGF and tumstatin compete for signaling through αvβ3. Tumstatin
and VEGF do not interact directly (data not shown) and probably do not share the same binding
site on αvβ3: Tumstatin has been shown to bind to the β3 subunit, whereas VEGF seems to bind
the classical RGD binding site (even if no RGD sequence is present in VEGF). This could mean
that both molecules might bind αvβ3 simultaneously, implying that the αvβ3 integrin is the
major regulation point leading to either survival or death of the endothelial cell (see Fig. 10).

The role of αvβ3 in angiogenesis is still controversial. Pioneer work has focused attention on the
proangiogenic role of αvβ3 (33), showing that disruption of αvβ3 anchorage with the ECM
promoted apoptosis, whereas other work has shown that disruption of either the αv or β3 genes
leads to enhanced angiogenesis (34). However, these demonstrations address functional
relationships between αvβ3 and soluble VEGF, which in turn activates VEGF receptors. The
recent demonstration that ligation of αvβ3 with tumstatin induces apoptosis specifically in
endothelial cells allowed the postulation that it might bind simultaneously to other substrates that
have not been yet identified. We propose that this missing link for αvβ3-dependent survival is
immobilized VEGF. Therefore, the balance between the amounts of matrix-bound VEGF and
tumstatin could be an essential mechanism used by the organism to control the quiescence of
endothelial cells or the fate of angiogenic endothelial cells (death or proliferation).

The finding that VEGF serves as an adhesive protein in the ECM helps to explain many other
previous results. It is well known that VEGF is not a factor that circulates well in the blood
stream; it is rapidly sequestered by the proteoglycans of the ECM (14, 35). This results in VEGF
being present and available only close to its production site. One of the hypotheses concerning
the role of VEGF trapped in the matrix is that it constitutes a natural reserve of growth factor that
can be freed by proteolytic digestion when necessary. Here, we have determined another
biological function for these sequestered molecules of VEGF, especially for the 189-amino acid
isoform. Immobilized VEGF189 remains functional and could serve in vivo as a provisional
matrix and participate in directing migrating endothelial cells at the site of angiogenesis. This
correlates with results from Pat D'Amore’s team, which showed that the basic ECM binding
VEGF isoforms are necessary in retinal vascular development to allow vessels to branch out
from the center of the retina (36). It has been shown that VEGF189 expression is up-regulated
more than other VEGF isoforms in esophageal xenografts (37) and by some types of tumor cells,
for example, in human colon and in non-small-cell lung cancer (38). In these cases, VEGF189 in
the tumor ECM or associated with the cell surface could be a support for endothelial cell
migration into the tumor mass. It has also been reported that VEGF189, and not VEGF165,
expression is up-regulated in osteoartritic cartilage (39), where uncontrolled angiogenesis
contributes to the pathology.

Another hypothesis as to the function of VEGF in the ECM could be that the interaction between
immobilized VEGF and integrins may not be strictly necessary for cell adhesion but might
influence VEGF receptor activity. It has previously been shown that αvβ3 interacts and
modulates the activity of VEGFR2 (40). Because it seems that VEGF interacts with VEGF
receptors and integrins via distinct sites, it is possible that matrix-bound VEGF could cocluster
integrins and receptors and thus enhance signal transduction.


We wish to thank Laurent Pelletier for valuable discussions concerning this work. This work was
supported by a grant from La Ligue contre le Cancer (Equipe Labelisée) and the Région Midi-
Pyrénées. H.H. is a recipient of a Ph.D. studentship from the Fondation pour la Recherche
Médicale. We also thank Françoise Viala for great help with graphics.


1.    Ferrara, N., Davis-Smyth, T. (1997) The biology of vascular endothelial growth factor.
Endocr. Rev. 18, 4–25

2.     Plouet, J., Schilling, J., Gospodarowicz, D. (1989) Isolation and characterization of a
newly identified endothelial cell mitogen produced by AtT-20 cells. EMBO J. 8, 3801–3806

3.    Ferrara, N., Henzel, W. J. (1989) Pituitary follicular cells secrete a novel heparin-binding
growth factor specific for vascular endothelial cells. Biochem. Biophys. Res. Commun. 161, 851–

4.     Barleon, B., Sozzani, S., Zhou, D., Weich, H. A., Mantovani, A., Marme, D. (1996)
Migration of human monocytes in response to vascular endothelial growth factor (VEGF) is
mediated via the VEGF receptor flt-1. Blood 87, 3336–3343

5.     Gabrilovich, D., Ishida, T., Oyama, T., Ran, S., Kravtsov, V., Nadaf, S., Carbone, D. P.
(1998) Vascular endothelial growth factor inhibits the development of dendritic cells and
dramatically affects the differentiation of multiple hematopoietic lineages in vivo. Blood 92,

6.     Plouet, J., Moro, F., Bertagnolli, S., Coldeboeuf, N., Mazarguil, H., Clamens, S., Bayard,
F. (1997) Extracellular cleavage of the vascular endothelial growth factor 189- amino acid form
by urokinase is required for its mitogenic effect. J. Biol. Chem. 272, 13390–13396

7.     Terman, B. I., Dougher-Vermazen, M., Carrion, M. E., Dimitrov, D., Armellino, D. C.,
Gospodarowicz, D., Bohlen, P. (1992) Identification of the KDR tyrosine kinase as a receptor for
vascular endothelial cell growth factor. Biochem. Biophys. Res. Commun. 187, 1579–1586

8.    de Vries, C., Escobedo, J. A., Ueno, H., Houck, K., Ferrara, N., Williams, L. T. (1992)
The fms-like tyrosine kinase, a receptor for vascular endothelial growth factor. Science 255,

9.     Soker, S., Takashima, S., Miao, H. Q., Neufeld, G., Klagsbrun, M. (1998) Neuropilin-1 is
expressed by endothelial and tumor cells as an isoform- specific receptor for vascular endothelial
growth factor. Cell 92, 735–745

10.    Gluzman-Poltorak, Z., Cohen, T., Herzog, Y., Neufeld, G. (2000) Neuropilin-2 is a
receptor for the vascular endothelial growth factor (VEGF) forms VEGF-145 and VEGF-165. J.
Biol. Chem. 275, 29922

11.    Carmeliet, P., Ferreira, V., Breier, G., Pollefeyt, S., Kieckens, L., Gertsenstein, M.,
Fahrig, M., Vandenhoeck, A., Harpal, K., Eberhardt, C., et al. (1996) Abnormal blood vessel
development and lethality in embryos lacking a single VEGF allele. Nature 380, 435–439
12.    Shalaby, F., Rossant, J., Yamaguchi, T. P., Gertsenstein, M., Wu, X. F., Breitman, M. L.,
Schuh, A. C. (1995) Failure of blood-island formation and vasculogenesis in Flk-1-deficient
mice. Nature 376, 62–66

13.    Fong, G. H., Rossant, J., Gertsenstein, M., Breitman, M. L. (1995) Role of the Flt-1
receptor tyrosine kinase in regulating the assembly of vascular endothelium. Nature 376, 66–70

14.     Houck, K. A., Leung, D. W., Rowland, A. M., Winer, J., Ferrara, N. (1992) Dual
regulation of vascular endothelial growth factor bioavailability by genetic and proteolytic
mechanisms. J. Biol. Chem. 267, 26031–26037

15.     Carlson, T. R., Feng, Y., Maisonpierre, P. C., Mrksich, M., Morla, A. O. (2001) Direct
cell adhesion to the angiopoietins mediated by integrins. J. Biol. Chem. 276, 26516–26525

16.     Munger, J. S., Harpel, J. G., Giancotti, F. G., Rifkin, D. B. (1998) Interactions between
growth factors and integrins: latent forms of transforming growth factor-beta are ligands for the
integrin alphavbeta1. Mol. Biol. Cell 9, 2627–2638

17.     Bensaid, M., Malecaze, F., Prats, H., Bayard, F., Tauber, J. P. (1989) Autocrine
regulation of bovine retinal capillary endothelial cell (BREC) proliferation by BREC-derived
basic fibroblast growth factor. Exp. Eye Res. 48, 801–813

18.      Keyt, B. A., Berleau, L. T., Nguyen, H. V., Chen, H., Heinsohn, H., Vandlen, R., Ferrara,
N. (1996) The carboxyl-terminal domain (111-165) of vascular endothelial growth factor is
critical for its mitogenic potency. J. Biol. Chem. 271, 7788–7795

19.    Krutzsch, H. C., Choe, B. J., Sipes, J. M., Guo, N., Roberts, D. D. (1999) Identification of
an alpha(3)beta(1) integrin recognition sequence in thrombospondin-1. J. Biol. Chem. 274,

20.     Kueng, W., Silber, E., Eppenberger, U. (1989) Quantification of cells cultured on 96-well
plates. Anal. Biochem. 182, 16–19

21.     Moore, A., Donahue, C. J., Bauer, K. D., Mather, J. P. (1998) Simultaneous measurement
of cell cycle and apoptotic cell death. Methods Cell Biol. 57, 265–278

22.     Rusnati, M., Tanghetti, E., Dell'Era, P., Gualandris, A., and Presta, M. (1997)
alphavbeta3 integrin mediates the cell-adhesive capacity and biological activity of basic
fibroblast growth factor (FGF-2) in cultured endothelial cells. Mol. Biol. Cell 8, 2449–2461.

23.    Chen, N., Chen, C. C., Lau, L. F. (2000) Adhesion of human skin fibroblasts to Cyr61 is
mediated through integrin alpha 6beta 1 and cell surface heparan sulfate proteoglycans. J. Biol.
Chem. 275, 24953–24961

24.     Byzova, T. V., Goldman, C. K., Pampori, N., Thomas, K. A., Bett, A., Shattil, S. J., Plow,
E. F. (2000) A mechanism for modulation of cellular responses to VEGF: activation of the
integrins. Mol. Cell 6, 851–860
25.    Waltenberger, J., Claesson-Welsh, L., Siegbahn, A., Shibuya, M., Heldin, C. H. (1994)
Different signal transduction properties of KDR and Flt1, two receptors for vascular endothelial
growth factor. J. Biol. Chem. 269, 26988–26995

26.   Price, L. S., Leng, J., Schwartz, M. A., Bokoch, G. M. (1998) Activation of Rac and
Cdc42 by integrins mediates cell spreading. Mol. Biol. Cell 9, 1863–1871

27.    Clark, E. A., King, W. G., Brugge, J. S., Symons, M., Hynes, R. O. (1998) Integrin-
mediated signals regulated by members of the rho family of GTPases. J. Cell Biol. 142, 573–586

28.     Jonca, F., Ortega, N., Gleizes, P. E., Bertrand, N., Plouet, J. (1997) Cell release of
bioactive fibroblast growth factor 2 by exon 6-encoded sequence of vascular endothelial growth
factor. J. Biol. Chem. 272, 24203–24209

29.     Munger, J. S., Huang, X., Kawakatsu, H., Griffiths, M. J., Dalton, S. L., Wu, J., Pittet, J.
F., Kaminski, N., Garat, C., Matthay, M. A., et al. (1999) The integrin alpha v beta 6 binds and
activates latent TGF beta 1: a mechanism for regulating pulmonary inflammation and fibrosis.
Cell 96, 319–328

30.    Dallabrida, S. M., Falls, L. A., Farrell, D. H. (2000) Factor XIIIa supports microvascular
endothelial cell adhesion and inhibits capillary tube formation in fibrin. Blood 95, 2586–2592

31.    Maeshima, Y., Yerramalla, U. L., Dhanabal, M., Holthaus, K. A., Barbashov, S.,
Kharbanda, S., Reimer, C., Manfredi, M., Dickerson, W. M., Kalluri, R. (2001) Extracellular
matrix-derived peptide binds to alpha(v)beta(3) integrin and inhibits angiogenesis. J. Biol. Chem.
276, 31959–31968

32.     Maeshima, Y., Sudhakar, A., Lively, J. C., Ueki, K., Kharbanda, S., Kahn, C. R.,
Sonenberg, N., Hynes, R. O., Kalluri, R. (2002) Tumstatin, an endothelial cell-specific inhibitor
of protein synthesis. Science 295, 140–143

33.    Brooks, P. C., Montgomery, A. M., Rosenfeld, M., Reisfeld, R. A., Hu, T., Klier, G.,
Cheresh, D. A. (1994) Integrin alpha v beta 3 antagonists promote tumor regression by inducing
apoptosis of angiogenic blood vessels. Cell 79, 1157–1164

34.    Reynolds, L. E., Wyder, L., Lively, J. C., Taverna, D., Robinson, S. D., Huang, X.,
Sheppard, D., Hynes, R. O., Hodivala-Dilke, K. M. (2002) Enhanced pathological angiogenesis
in mice lacking beta3 integrin or beta3 and beta5 integrins. Nat. Med. 8, 27–34

35.    Park, J. E., Keller, G. A., Ferrara, N. (1993) The vascular endothelial growth factor
(VEGF) isoforms: differential deposition into the subepithelial extracellular matrix and
bioactivity of extracellular matrix-bound VEGF. Mol. Biol. Cell 4, 1317–1326

36.    Stalmans, I., Ng, Y. S., Rohan, R., Fruttiger, M., Bouche, A., Yuce, A., Fujisawa, H.,
Hermans, B., Shani, M., Jansen, S., et al. (2002) Arteriolar and venular patterning in retinas of
mice selectively expressing VEGF isoforms. J. Clin. Invest. 109, 327–336
37.    Okamoto, K., Oshika, Y., Fukushima, Y., Ohnishi, Y., Tokunaga, T., Tomii, Y., Kijima,
H., Yamazaki, H., Ueyama, Y., Tamaoki, N., et al. (1999) Xenografts of human solid tumors
frequently express cellular-associated isoform of vascular endothelial growth factor (VEGF) 189.
Oncol. Rep. 6, 1201–1204

38.     Yuan, A., Yu, C. J., Kuo, S. H., Chen, W. J., Lin, F. Y., Luh, K. T., Yang, P. C., Lee, Y.
C. (2001) Vascular endothelial growth factor 189 mRNA isoform expression specifically
correlates with tumor angiogenesis, patient survival, and postoperative relapse in non-small-cell
lung cancer. J. Clin. Oncol. 19, 432–441

39.    Pufe, T., Petersen, W., Tillmann, B., Mentlein, R. (2001) The splice variants VEGF121
and VEGF189 of the angiogenic peptide vascular endothelial growth factor are expressed in
osteoarthritic cartilage. Arthritis Rheum. 44, 1082–1088

40.    Soldi, R., Mitola, S., Strasly, M., Defilippi, P., Tarone, G., Bussolino, F. (1999) Role of
alphavbeta3 integrin in the activation of vascular endothelial growth factor receptor-2. EMBO J.
18, 882–892

                                               Received August 5, 2002; accepted February 14, 2003.
Table 1

Adhesion of four endothelial cell strains on VEGF isoformsa

                                                                Cell type
                                                        (optical density, 595 nm)

Adhesion substrate             HUAEC                FBAE                  BREC                  HBMEC

BSA                              0.08                0.02                  0.10                   0.08
VEGF121                          0.05                0.03                  0.11                   0.08
VEGF165                          0.42               0.224                  0.41                   0.43
    + heparin                    0.46                 nd                   0.32                    nd
    + p678                       0.08               0.067                  0.06                   0.14
    + RGD                        0.37                0.09                  0.40                   0.32
    + p678 + RGD                 0.11                0.02                  0.04                   0.09
    + soluble VEGF165            0.33                0.16                  0.44                    nd
VEGF189                          0.92                0.38                  0.51                   0.66
    + heparin                    0.44                 nd                   0.25                    nd
    + p678                       0.44                0.21                  0.42                   0.48
    + RGD                        0.75                0.20                  0.49                   0.55
    + p678 + RGD                 0.16                0.11                  0.31                   0.27
    + soluble VEGF189            0.87                0.19                  0.51                    nd
 ELISA plates were coated with concentrations of 4 µg/ml of all proteins. Cells were trypsinized, washed, and
resuspended in DMEM with 10% FCS and incubated for 1 h at 37°C. Cells were then pelleted and resuspended
in DMEM without serum in the presence of heparin (50 nM), p678 (20 µM), RGD peptides (500 µM), or
soluble VEGF (200 nM) for an additioanl 20 min. Cells were then distributed (40,000 cells per well) and
allowed to adhere for 20 min. Nonadherent cells were washed away, and the amount of cell adhesion was
quantified by crystal violet incorporation. All conditions were performed in triplicate, and results are
representative of at least two experiments. Nd, not determined. Values are given as optical density readings at
595 nm.
Table 2

Cooperation between integrin avß3 and tumstatin to induce apoptosis of HUAEC plated on VEGF isoformsa

                                                                  (apoptosis, %)
                                      BSA               VEGF165               VEGF189              Col I

Blank                                 100                  22.8                    31.6             27.7
IgG Ctrl                              —                    27.8                    35.8             28.4
Anti-αvβ3                              —                   39.9                    34.2             26.9
Tumstatin                             —                    34.9                    49.9             36.0
Anti-αvβ3 + Tumstatin                  —                   61.3                    62.3             33.9
 HUAEC cells were trypsinized and plated in wells coated with either BSA, VEGF165, VEGF189, or collagen I
(Col I) in the presence or absence of anti-αvβ3. Cells were left to adhere for 2 h and were then treated with 10
µg/ml tumstatin. Annexin V-fluorescein isothiocyanate staining and propidium iodide (PI) incorporation were
performed after 3 h treatment. FACS analysis was done to quantitate the percentage of cells undergoing
apoptosis (annexin V-positive, PI-negative cells). Values are expressed as percentages of apoptotic cells.
Results of one out of three similar experiments are shown.
Fig. 1

Figure 1. HUAEC adhesion on VEGF. A) Adhesion of HUAEC to fibronectin, VEGF121, VEGF165, VEGF189, and uPA-
VEGF189. Coating concentrations were 4 µg/ml for all proteins. Cells were allowed to adhere for 20 min at 37°C.
B) Time-course adhesion of HUAEC to VEGF165, VEGF189, and fibronectin.
Fig. 2

Figure 2. Effect of heparin and exon 6-derived peptides on endothelial cell adhesion to VEGF isoforms.
A) HUAEC were pretreated with various concentrations of heparin before adhesion in wells coated with either VEGF165,
VEGF189, or fibronectin. B) Time-course HUAEC adhesion to VEGF165, VEGF189, or fibronectin in the presence of 20
µg/ml P1 and P2 peptides. i) Adhesion on VEGF165, ii) adhesion on VEGF189, and iii) adhesion on fibronectin.
Fig. 3

Figure 3. Effect of integrin blocking agents on HUAEC adhesion to VEGF. HUAEC adhesion was observed on
VEGF165 (A), VEGF189 (B), and fibronectin (C). Ninty-six-well ELISA plates were coated with concentrations of 4 µg/ml
for all proteins. Cells were trypsinized, washed, and resuspended in DMEM with 10% FCS and incubated for 1 h at 37°C.
Cells were then pelleted and resuspended in DMEM without serum in the presence of EDTA (10 mM), RGD peptides
(500 µM), RAD peptides (500 µM), p678 (20 µM), p690 (20 µM), or soluble VEGF (200 nM) for an additional 20 min.
Cells were then distributed and allowed to adhere for 20 min. All conditions were performed in triplicate. Shown values
are the mean ±SD for each condition. (*P<0.05, **P<0.005, and ***P<0.0005 compared with control nontreated cells).
Fig. 4

Figure 4. Time-course adhesion of HUAEC on VEGF isoforms VEGF165 and VEGF189. A) Time-dependent adhesion
on VEGF165. B) Time-dependent adhesion on VEGF189. Before distribution in wells coated with either substrate, cells
were treated by RGD peptides (500 µM) or p678 (20 µM) for 30 mon. Cells were allowed to adhere for up to 60 min. All
conditions were performed in triplicate.
Fig. 5

Figure 5. Time-course HUAEC adhesion on and VEGF165 and VEGF189 in the presence of anti-integrin antibodies.
Before distribution in wells coated with either substrate, cells were treated by anti-β1 6S6 (2 µg/ml), anti-αvβ3 (1/3), or
anti-αv (1/10) antibodies for 20 min. Cells were then allowed to adhere for various lengths of time. All conditions were
performed in triplicate. Shown values are the mean ±SD for each condition.
Fig. 6

Figure 6. Solid-phase αvβ3 binding assay. VEGF 165, VEGF189, fibronectin, or BSA was immobilized on plastic. αvβ3
recombinant protein was incubated with the immobilized protein, and interaction between the two proteins was revealed
by the ELISA technique with an anti-β3 antibody followed by a HRP-conjugated secondary antibody. Three
concentrations of αvβ3 recombinant protein were used to observe a dose response. All conditions were performed in
triplicate. Shown values are the mean ±SD for each condition.
Fig. 7

Figure 7. HUAEC and BREC actin morphology on VEGF165 and VEGF189. Glass Labtek slides precoated with
VEGF165, VEGF189, fibronectin, or poly-lysine at 4 µg/ml. A) F-actin labeling. B) FAK staining. Arrows point to FAK
localizing to focal adhesions.
Fig. 8

Figure 8. BREC and HUAEC migrate on VEGF. Cells were seeded at ∼25% confluency on plastic bacteriological
dishes coated with 4 µg/ml of VEGF165, VEGF189, fibronectin, or poly-lysine as previously described. Cells were allowed
to adhere for 2 h at 37°C. Nonadherent cells were removed, and cell movement was monitored by successive photography
of cells at a given spot of the dish. Image overlay was used to evaluate the migration patterns of individual cells. A) Mean
values ±SD of the trajectories of individual BREC (n=10) over a 5-h migration period. B) Mean values ±SD of the
trajectories of individual HUAEC (n=10) over a 5-h migration period. Distances in arbitrary units (AU). Results are
representative of three separate experiments. ***P<0.0005 compared with migration distance on VEGF189.
Fig. 9

Figure 9. HUAEC adhesion on and VEGF165 and VEGF189 in the presence of tumstatin and anti-αvβ3 integrin
antibodies. Ninety-six-well ELISA plates were coated with either VEGF165 or VEGF189 at 4 µg/ml. Cells were
trypsinized, washed, and resuspended in DMEM with 10% FCS and incubated for 1 h at 37°C. Cells were then pelleted
and resuspended in DMEM without serum in the presence of either anti-αvβ3 (1/3), tumstatin, or both for an additional 20
min. Cells were then distributed at 40,000 cells per well and allowed to adhere for 20 min. Nonadherent cells were washed
away, and the amount of cell adhesion was quantified by crystal violet as described in Materials and Methods. All
conditions were performed in triplicate. Shown values are the mean ±SD for each condition. (*P<0.05, **P<0.005, and
***P<0.0005 compared with control nontreated cells).
Fig. 10

Figure 10. Schematic representation of the opposite effects of tumstatin and matrix-bound VEGF on endothelial
cell survival. Endothelial cells adhere to matrix-bound VEGF by integrins such as αvβ3. A) Cells attached to
immobilized VEGF survive, whereas floating cells do not. B) If the αvβ3 integrin is engaged with matrix-bound VEGF,
tumstatin can induce a certain amount of endothelial cell apoptosis. C) If the interaction between matrix-bound VEGF and
αvβ3 is blocked by an anti-αvβ3 antibody, tumstatin-induced apoptosis is significantly increased.