| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
From the Department of Reproductive and Vascular
Biology,*
The Medical School, University of Birmingham,
Edgbaston, Birmingham, United Kingdom; the Cardiovascular Medicine
Unit,
Imperial College School of Medicine,
London, United Kingdom; and the Division of
Cardiology,
Duke University Medical Center,
Durham, North Carolina
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The known biological responses of VEGF in endothelial cells are reported to be mediated by the activation of VEGF tyrosine kinase receptor-2 (VEGFR-2).1,2 Transfection of human VEGFR-1 and VEGFR-2 into porcine aortic endothelial (PAE) cells showed that human recombinant VEGF was able to stimulate chemotaxis and proliferation in VEGFR-2-transfected and not in VEGFR-1-transfected cells.3 Only a few functions of VEGF have been attributed to VEGFR-1, including stimulation of peripheral blood monocyte migration and tissue factor expression,4 nitric oxide (NO) release in trophoblasts,5 and up-regulation of matrix metalloproteinases in vascular smooth muscle cells.6 Placenta growth factor (PlGF) that binds to VEGFR-1 and not VEGFR-2 also stimulates monocyte migration.4
Knockout studies demonstrate that both VEGFR-1 and VEGFR-2 are essential for normal development of the embryonic vasculature.7,8 Mice lacking VEGFR-2 fail to develop a vasculature and have very few mature endothelial cells,7 whereas mice engineered to lack VEGFR-1 seem to have excess formation of endothelial cells that abnormally coalesce into disorganized tubules.8 More recently, Fong and colleagues9 showed that increased mesenchymal-hemangioblast transition is the primary defect in VEGFR-1 knock-out mice, whereas the formation of disorganized vascular channels is a secondary phenotype because of the overcrowding of the endothelial population. However, it is unclear how VEGFR-1 prevents overcrowding. As truncation of VEGFR-1 at the tyrosine kinase domain does not impair embryonic angiogenesis, this led to the suggestion that VEGFR-1 acts as an inert decoy by binding VEGF and thereby regulating the availability of VEGF for activation of VEGFR-2.10 However, this does not negate the involvement of VEGFR-1 signaling in adult endothelia. Indeed, there is now a considerable body of evidence that on the contrary supports this notion5,11,12 and the role of this receptor has been implicated in both physiological13 and pathological angiogenesis.10,14
Angiogenesis is initiated by vasodilatation, a NO-mediated process. Originally identified as endothelium-derived relaxing factor, NO has profound vasomotor regulatory effects on the vasculature.15 In addition to its potent vasodilatory function, NO inhibits platelet aggregation, leukocyte adherence, and smooth muscle proliferation and migration, supporting its role in the maintenance of vascular integrity. Synthesis of NO is under tight regulation of a family of NO synthase (NOS) isoenzymes that convert L-arginine to L-citrulline in the presence of molecular oxygen yielding free NO.16 Ziche and co-workers17 established the first line of evidence that NO can induce angiogenesis in vivo. We and others have shown that VEGF stimulates the release of NO from human umbilical vein endothelial cells (HUVECs)5 and from intact rabbit arterial strips.18 Further studies have shown that NO generated in response to VEGF induces angiogenesis both in vitro19 and in vivo.20 The most compelling evidence that NOS is a downstream effector of VEGF-mediated angiogenesis comes from studies in eNOS knockout mice in which VEGF fails to induce angiogenesis.21 However, it is not clear at what stage in the process of blood vessel formation that NO exerts its influence. Accordingly, we tested the hypothesis that VEGFR-1 negatively regulates VEGFR-2 action via NO that functions to inhibit endothelial cell growth and to promote vascular tube formation.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Recombinant VEGF165,
VEGF121, and VEGF:PlGF heterodimer were purchased
from Strathmann Biotech GmbH (Hanover, Germany). The soluble VEGFR-1
ectodomain (sVEGFR-1), a VEGF antagonist, was kindly provided by Dr. H.
Weich (GBF, Braunschweig, Germany). The neutralizing monoclonal
anti-VEGF antibody was a gift from Dr. N. Ferrara (Genetech, San
Francisco, CA). The neutralizing polyclonal anti-VEGFR-1 and
anti-VEGFR-2 antibodies were raised as previously
described5
and validated as shown in Figure 1
. Matrigel basement membrane matrix was
obtained from Becton Dickinson Labware (Bedford, MA). Polycarbonate
filters (8-µm pore size, polyvinylpyrrolidone-free) were obtained
from Nucleopore Corporation (Pleasanton, CA). Sodium nitroprusside,
N-(b-D-glucopyranosyl)-N2-acetyl-S-nitroso-D,L-penicillaminamide
(Glyco-SNAP-1), NG monomethyl
L-arginine (L-NNA), 8-bromo-cGMP, LY 83583, and
tyrosine kinase inhibitors were from Calbiochem Novabiochem Corporation
(La Jolla, CA). All cell culture reagents were obtained from Sigma
Chemical Co. Ltd. (Poole, Dorset, UK).
|
HUVECs were isolated, characterized, and cultured as previously described.5 Experiments were performed on second or third passage HUVECs. PAE cells that were transfected with either human VEGFR-1 (PAEVEGFR-1) or VEGFR-2 (PAEVEGFR-2) were obtained from Dr. Johannes Waltenberger (Ulm, Germany). Scatchard analysis of receptor binding of VEGF to PAEVEGFR-1 and PAEVEGFR-2 showed that the binding and expression of these receptors are comparable to HUVEC.3 Peripheral blood monocytes were isolated from buffy coats using gradient centrifugation over Ficoll (histopaque 1077) and subsequent on plastic dishes.
Measurement of DNA Synthesis and Proliferation
Quiescent HUVECs or transfected PAEVEGFR-1 and PAEVEGFR-2 cells were incubated with test substances and assayed for DNA synthesis by measuring [3H]-thymidine incorporation5 and cell proliferation by counting the cells in a Coulter Counter (Coulter Electronics Ltd., Hialeah, FL).22
Measurement of NO Release
For concentration-dependence and time-course experiments, stimulations were initiated in confluent and quiescent HUVECs and PAE cells in serum-free culture medium. For the inhibition studies, cells were preincubated for 30 minutes with test substances as indicated, and stimulation then initiated with 50 ng/ml VEGF in a final volume of 0.5 ml at 37°C for a further 60 minutes. The reaction was terminated by removal of the supernatant that was centrifuged and immediately stored at -80°C for NO analysis within 1 week. Because increasing intracellular cGMP levels cannot be ascribed solely to NOS activation,23 we measured NO directly in the gas phase using a Sievers NOA 280 chemiluminescence analyzer (Analytix, Sunderland, UK) as previously described.5 Results are corrected for background levels of NO present in culture medium alone in all of the experiments excepts in low calcium buffer studies.
In Vitro Migration Assay
Chemotaxis of HUVECs and monocytes to VEGF was assessed using a modified Boyden’s chamber. Briefly, HUVECs (2 x 105) or monocytes (1.5 x 106/ml) were seeded in the upper chamber and their migration across a polycarbonate filter in response to 20 ng/ml of VEGF was investigated. The upper surface of the filter was scraped and filters were fixed and stained with Diff-Quik (Harleco, Gibbstown, NJ). Ten random fields at x200 magnification were counted and the results expressed as mean (±SEM) number per field.
Tube Formation
In vitro formation of tubular structures was studied on growth factor-reduced Matrigel diluted 1:1 in ice-cold Dulbecco’s modified Eagle’s medium, as previously described.24 PAE-transfected cells (5 x 104 cells/well) were seeded onto Matrigel-coated wells in culture medium containing 10% fetal calf serum (FCS). After cells had attached to the Matrigel (2 hours at 37°C), the FCS-containing media was removed and VEGF165 (25 ng/ml) or the vehicle alone in media containing 0.2% bovine serum albumin was added. Cells were pretreated with anti-VEGFR antibodies (200 ng/ml) or 8-bromo-cGMP or LY 83583 at 37°C for 30 minutes prior to stimulation with VEGF165. In parallel experiments, Trypan Blue exclusion showed that cell viability was >90%. Cells were observed with a Nikon inverted microscope and experimental results recorded with an Optimas image analysis software (Microscope Service, Surrey, UK).
In Vitro Co-Culture Angiogenesis Assay
In vitro angiogenesis was assessed as formation of capillary-like structures of HUVECs co-cultured with matrix-producing cells that had been UV irradiated before plating of primary HUVECs. The experimental procedure followed the manufacturer’s protocol provided with the In Vitro angiogenesis kit (TCS Biologicals, Buckingham, UK). Briefly, cells were stimulated with the test substances at day 3 and medium was replaced at day 5, 8, and 10. At day 13, the cells were fixed and HUVECs were stained using an anti-CD31 antibody (TCS Biologicals) according to the instructions provided with the kit. To measure the formation of the capillary network, the number of connections between three or more capillary-like structures was counted and expressed as the number of capillary connections per field. Furthermore, the average thickness of the tube or cell overcrowding, the total length of tubes per field, and the average length (distance) of tube between connections were quantified by image analysis at x4 magnification with a MicroImage analysis system (Cast Imaging srl, Venice, Italy) calibrated with an Olympus micrometer slide. Four (0.5 mm2) different fields were analyzed per well.
Murine Matrigel Angiogenesis Assay
In vivo angiogenesis was assayed in mice as growth of blood vessels from subcutaneous tissue into a Matrigel plug containing the test sample, according to Passaniti and colleagues.25 Matrigel rapidly forms a solid gel at body temperature, trapping the factors to allow slow release and prolonged exposure to surrounding tissues. Matrigel (8.13 mg/ml) in liquid form at 4°C was mixed with 64 U/ml heparin and VEGF (40 ng/ml) or vehicle alone (saline containing 0.25% bovine serum albumin). In selected experiments, the neutralizing anti-VEGFR-1 or anti-VEGFR-2 antibodies were included in the Matrigel plug at a final concentration of 200 mg/ml. Matrigel was injected (0.5 ml) into the abdominal subcutaneous tissue of female C57 mice (6 to 8 weeks old) along the peritoneal mid-line. At day 6, mice were sacrificed and plugs were recovered and processed for histology. Typically, the overlying skin was removed, and gels were cut out by retaining the peritoneal lining for support, fixed in 10% buffered formalin, and embedded in paraffin. Sections (3 µm) were cut and stained with hematoxylin and eosin (H&E) and examined under a light microscope system (Nikon UK Limited Instrument Division, Kingston, UK). Vessel area and the total Matrigel area were planimetrically assessed from cross-sections of Matrigel plugs as described by Kibbey and colleagues.26 The mean area of H&E-stained vessels per field from 10 to 20 fields (x40 magnification) was evaluated using the computing integral area calculation of Lucia digital system (Nikon UK Limited). Blood vessels defined as those structures possessing a patent lumen and containing red blood cells were identified blind by an independent observer. Positive immunofluorescence staining for vWF, an endothelium-specific antigen, was shown to correlate with the presence of vessels detected by light microscopy.26 Results were expressed as percentage of the vessel area to the total Matrigel area.
Statistical Analysis
All data are expressed as mean ± SEM. Statistical analysis was performed by analysis of variance with Dunnett’s or Newmann-Keul’s multiple comparison test or by Mann-Whitney test where appropriate.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
VEGF165 stimulates chemotaxis in transfected
PAEVEGFR-2 and not
PAEVEGFR-1 cells demonstrating that VEGFR-2
activation is required for endothelial migration.3
In
contrast, VEGF stimulates monocyte chemotaxis via
VEGFR-1.4
To test the functional specificity of the
neutralizing anti-VEGF receptor antibodies, the chemotactic effect of
VEGF was evaluated on both HUVECs and monocytes. The anti-VEGFR-2
antibody (100 ng/ml) but not the anti-VEGFR-1 antibody significantly
inhibited the effect of VEGF165 (20 ng/ml) on
migration of HUVECs (Figure 1A
; P < 0.05). The VEGFR-2
agonist, VEGF121,27
induced a
chemotactic response comparable to VEGF. Basic fibroblast growth factor
(10 ng/ml)-induced HUVEC migration was unaffected by the anti-VEGFR
antibodies (data not shown). In contrast to HUVECs, preincubation of
monocytes (that only express VEGFR-1) with anti-VEGFR-1 antibody (50
ng/ml) but not with the anti-VEGFR-2 antibody (50 ng/ml) completely
inhibited the VEGF-mediated migration (Figure 1B
; P <
0.05), thus confirming the specificity of the two antibodies.
Effect of Blockade of VEGF Receptors on in Vivo Angiogenesis
The relative role of VEGF receptors in the process of angiogenesis
was investigated in vivo to determine whether the
subtype-selective blocking antibodies could replicate some of the
murine knockout phenotypes. The histological and morphometric analyses
of Matrigel plugs containing VEGF (40 ng/ml) revealed a diffuse
angiogenic process, characterized by canalized vessels and
microaneurysm-like structures (Figure 2, B and G
; P < 0.0006 versus control). The
blockade of VEGFR-2 reduced VEGF-induced angiogenic process as there
were small branched vessels (Figure 2; C, E, and G
; P
< 0.003 versus VEGF) suggesting impairment in endothelial
cell invasion of the Matrigel. In contrast, the blockade of VEGFR-1 by
anti-VEGFR-1 antibody caused the formation of large aneurysm-like
structures (Figure 2, D and G
; P < 0.03
versus VEGF). This angiogenic process covered >80% of the
Matrigel in the majority of mice treated with VEGF in the presence of
anti-VEGFR-1 antibody (Figure 2G)
. No angiogenesis was present in
control plugs containing the vehicle alone, heparin (64 U/ml) plus
vehicle (Figure 2A)
, or the antibodies alone (Figure 2F)
.
|
The relative role of VEGF receptors in the process of angiogenesis
was investigated in vitro using the angiogenesis kit assay
to determine the underlying mechanism. Quantitative analysis is shown
for the average thickness of the tube or cell overcrowding (Figure 3A)
, the total length of tubes per field
(Figure 3B)
, the number of capillary connections per field (Figure 3C)
,
and the average length (distance) of tube between connections (Figure 3D)
. The basal formation of capillary-like structures was increased
after stimulation with VEGF165 (20 ng/ml). There
was a significant increase in tube thickness (Figure 3A
;
P < 0.05, n = 4), the total length of
tubes per field (Figure 3B
; P < 0.04,
n = 4), number of capillary connections (Figure 3C
;
P < 0.05, n = 4), and the average
length of tube between connections (Figure 3D
; P <
0.003, n = 4). Anti-VEGFR-1 antibody inhibited
VEGF-induced total tube length by threefold (Figure 3B
;
P < 0.0001, n = 4), capillary
connections by a dramatic sevenfold (Figure 3C
; P <
0.0001, n = 4), and the average length of tube between
connections were also reduced (Figure 3D
; P < 0.0001,
n = 4), whereas tube thickness was increased by almost
fourfold (Figure 3, A and G
; P < 0.04,
n = 4) compared with VEGF alone. HUVECs remained in
isolated islands, proliferated, and accumulated in large aggregates in
the presence of anti-VEGFR-1 (Figure 3G)
. In contrast, blockade of
VEGFR-2 by anti-VEGFR-2 antibody (100 ng/ml) did not cause a
significant increase in VEGF-induced tube thickness (Figure 3A)
but did
inhibit both total tube length (Figure 3B
; P < 0.002,
n = 4) and tube length between connections (Figure 3D
;
P < 0.0001, n = 4). Capillary
connections per field were also inhibited by the anti-VEGFR-2 antibody
(100 ng/ml) when compared with 20 ng/ml VEGF165
(Figure 3C
; P < 0.0004, n = 4) but
under basal conditions, anti-VEGFR-2 antibody alone did not inhibit
capillary connections (Figure 3C)
. Interestingly, the anti-VEGFR-1
antibody did not significantly inhibit tube length between connections
under basal conditions whereas anti-VEGFR-2 caused a marked reduction
in tube length between connections (Figure 3D
; P <
0.0001, n = 4). In the presence of
VEGF165 and anti-VEGFR-2 antibody, short tubes
were detectable that failed to form capillary connections (Figure 3H)
.
When the NO donor glyco-SNAP (10-5 mol/L) was
added to cells stimulated with VEGF165 in the
presence of anti-VEGFR-1 antibody, the ability of HUVECs to form
capillary-like structures was partially restored as determined by the
increase in total tube length (Figure 3B
; P < 0.02,
n = 4) and capillary connections (Figure 3C
;
P < 0.0001, n = 4) and decrease in
tube thickness (Figure 3A
; P < 0.0001,
n = 4). Basal capillary network was inhibited by
soluble VEGFR-1 that blocks VEGF activity indicating that VEGF is
involved in the process under basal conditions. Addition of
8-bromo-cGMP alone induced a diffuse capillary network (Figure 3, B and F)
comparable to VEGF (Figure 3E)
.
|
PAEVEGFR-1 cells, plated on a Matrigel
substratum, elongated and connected in basal conditions (Figure 4A)
. When stimulated with
VEGF165 (25 ng/ml), cells organized into complete
tubular structures (Figure 4B)
. The blockade of VEGF with the
neutralizing anti-VEGF antibody (50 ng/ml) inhibited all these
morphological rearrangements (data not shown) suggesting a specific
response. Preincubation with anti-VEGFR-2 antibody (200 ng/ml) had no
effect on tube formation (Figure 4C)
, whereas anti-VEGFR-1 antibody
(200 ng/ml) inhibited both the basal and the VEGF-mediated tubular
network (Figure 4D)
. Similar results were obtained with HUVECs (data
not shown). cGMP inhibitor LY83583 (2.5 µmol/L) completely inhibited
the PAEVEGFR-1 cell organization (Figure 4, E and F)
supporting the role of NO released via VEGFR-1 in endothelial cell
differentiation. In contrast to PAEVEGFR-1 cells,
the PAEVEGFR-2 cells were unable to establish a
network of tubular-like structures on Matrigel under basal conditions
or when stimulated with VEGF165 (25 ng/ml)
(Figure 5, A and B)
. No significant
change from basal or VEGF was observed after the addition of
anti-VEGFR-2 (Figure 5C)
or anti-VEGFR-1 (Figure 5D)
antibodies.
Interestingly the addition of 8-bromo-cGMP stimulated the formation of
a network in PAEVEGFR-2 cells in the absence
(Figure 5E)
or in the presence of 25 ng/ml
VEGF165 (Figure 5F)
. Wild-type
PAEWT seeded on Matrigel did not display the
ability to form tubular structures both under basal conditions and when
stimulated with VEGF165 (data not shown). Cell
viability after stimulation with cGMP or LY 83583 was >90% as
assessed by Trypan Blue exclusion.
|
|
To investigate the mechanism, by which VEGFR-1 promotes capillary
morphogenesis, we examined the effect of VEGF on NO production in
HUVECs. Addition of VEGF165 to quiescent HUVECs
resulted in a dose- and time-dependent release of NO (Figure 6A)
. At the highest concentration tested
(50 ng/ml), VEGF165 caused a 258.66 ±
14.75% increase in NO release above control (Figure 6A
;
P < 0.001, n = 3). Interestingly, VEGF
had no significant effect on NO release within the concentration range
of 0.5 to 5 ng/ml that is known to promote maximal endothelial cell
proliferation.1
This response was specific to VEGF,
because the neutralizing monoclonal anti-VEGF antibody28
inhibited VEGF165-stimulated NO release by
89.7 ± 1.6% (Figure 6B
; P < 0.01,
n = 3). Several experimental approaches were undertaken
to identify the receptor subtype responsible for VEGF-mediated NO
release. To investigate the role of VEGFR-2 in VEGF-mediated NO
release, cells were stimulated with increasing concentrations of
VEGF121 and VEGF:PlGF heterodimer that bind with
high affinity to VEGFR-2.27
Neither
VEGF121 or VEGF:PlGF stimulated a significant
amount of NO release as compared with VEGF165
(Figure 6A)
. To further investigate the role of the VEGF receptors,
quiescent HUVECs were exposed to neutralizing anti-VEGFR-1 or
anti-VEGFR-2 antibodies (250 ng/ml) for 30 minutes before stimulation
with 50 ng/ml VEGF165 for 60 minutes (Figure 6C)
.
VEGF165-stimulated NO release was completely
inhibited by anti-VEGFR-1 antibody (P < 0.001,
n = 3; Figure 6C
) whereas the anti-VEGFR-2 antibody had
no effect (Figure 6C)
. Specific activation of eNOS was confirmed by the
attenuation of VEGF165-stimulated NO release by
L-NNA (Figure 6D)
as well as by the complete inhibition obtained by the
removal of extracellular calcium (P < 0.001,
n = 3; Figure 6E
). Both genistin (30 µmol/L) and
lavendustin A (25 µmol/L) inhibited 50 ng/ml VEGF-induced NO release
(P < 0.05; data not shown) and PlGF-1 but not
PlGF-2 stimulated NO release (P < 0.01,
n = 3; Figure 6F
).
|
To conclusively demonstrate that VEGF effect on NO
release was mediated via VEGFR-1, we used PAE cells transfected
with human VEGFR-1 or VEGFR-2. VEGF165 stimulated
NO release in a concentration-dependent manner from
PAEVEGFR-1 cells (Figure 7A)
. The effect of VEGF on NO release was
more potent and efficacious in the PAEVEGFR-1
cells compared with HUVECs. PAEVEGFR-1 cells
displayed an increased sensitivity to VEGF165 in
mediating NO release compared with HUVECs as VEGF induced a
twofold increase in NO production at low concentrations (1 and 5
ng/ml), ineffective in HUVECs (Figure 7A)
. In contrast,
VEGF165 failed to stimulate NO release from
PAEVEGFR-2 cells (Figure 7A)
. Basal NO levels in
PAEVEGFR-2 cells, however, were elevated compared
with PAEVEGFR-1. The time period during
which VEGF165 caused changes in eNOS activity in
PAEVEGFR-1 cells is shown in Figure 7B
. VEGF (10
ng/ml) stimulated an increase in eNOS activity that was significant at
20 minutes (P < 0.016, n = 3),
and peaked at 60 minutes to reach a maximum of 190 ± 3.8% more
than basal levels (P < 0.001, n
= 3; Figure 7B
). The addition of 50 ng/ml of sVEGFR-1 (VEGF antagonist)
after 35 minutes of stimulation with VEGF165 (10
ng/ml) completely inhibited VEGF-evoked NO release within 35 minutes
(P < 0.001, n = 3; Figure 7B
)
indicating that VEGF-mediated NO release is sustained during this
period. Likewise, the addition of 50 ng/ml of monoclonal
anti-VEGF antibody completely inhibited
VEGF165-evoked NO release (data not shown).
Activation of eNOS by VEGF was confirmed by the complete attenuation of
VEGF165-mediated NO release by the removal of
extracellular calcium (P < 0.001,
n = 3; Figure 7C
). Premixing
VEGF165 with excess (1 µg/ml) heparin
sulfate proteoglycan that saturates the heparin-binding
sites, significantly reduced the VEGF-dependent release of NO
(78 ± 11% reduction, P < 0.001,
n = 4). Heparin sulfate proteoglycan alone had no
effect. The results in HUVECs and transfected PAE cells conclusively
demonstrate that activation of VEGFR-1 mediates
VEGF165-stimulated NO release in endothelial
cells.
|
Addition of VEGF165 (10 ng/ml) to quiescent
HUVECs for 48 hours caused an increase in cell proliferation by
58.44 ± 11.91% compared with basal (1% FCS) cell growth (Figure 8A)
. The response was specific to VEGF as
VEGF165-mediated endothelial cell proliferation
was completely inhibited in HUVECs preincubated with 50 ng/ml of
neutralizing anti-VEGF monoclonal antibody (P <
0.001, n = 3) or 50 ng/ml of sVEGFR-1
(P < 0.001, n = 3; Figure 8A
).
To investigate the relative roles of two high-affinity VEGF receptor
subtypes on endothelial cell growth, VEGF-induced proliferation was
studied in the absence or presence of anti-VEGFR1 or anti-VEGFR-2
antibodies in HUVECs. In the presence of the neutralizing anti-VEGFR-1
antibody (30 ng/ml) alone, HUVEC proliferation increased by 51.36
± 2.09% more than basal levels (P < 0.01,
n = 3; Figure 8B
) and was comparable with that achieved
with 1 ng/ml VEGF165. This increase in cell
proliferation was inhibited in the presence of anti-VEGF antibody
(P < 0.01, n = 3; Figure 8B
)
indicating endogenous VEGF action. Addition of 1 ng/ml
VEGF165 to HUVECs incubated with the anti-VEGFR-1
antibody lead to a weak increase in cell proliferation
(P < 0.05, n = 3; Figure 8B
).
In contrast, the proliferative response of endothelial cells to the
higher concentrations of VEGF (10 to 50 ng/ml) was significantly
enhanced by the addition of the anti-VEGFR-1 antibody (Figure 8C)
. The
anti-VEGFR-2 antibody completely inhibited
VEGF165-mediated endothelial cell proliferation
(P < 0.001, n = 3; Figure 8B
).
Cell viability, assessed by Trypan Blue exclusion was >95% in all
experiments.
|
To investigate the role of VEGFR-1-mediated NO production on HUVEC
proliferation, quiescent subconfluent HUVECs were incubated
with the anti-VEGFR-1 antibody (30 ng/ml) in the presence of NO donors.
The increase in DNA synthesis observed in the presence of anti-VEGFR-1
antibody was significantly attenuated by the addition of glyco-SNAP-1
(10-4 to 10-6 mol/L)
(Figure 9A)
. However, under basal
conditions, glyco-SNAP at 10-6 to
10-8 mol/L significantly increased endothelial
cell proliferation as assessed by cell number compared with control,
whereas at 10-4 mol/L it had no effect (Figure 9B)
. In contrast, VEGF165 (10 ng/ml)-induced
proliferation was significantly attenuated by glyco-SNAP-1 at the doses
of 10-4 to 10-6 mol/L
(Figure 9B)
. Significant levels of NO, comparable with that obtained
with 50 ng/ml VEGF, were produced when glyco-SNAP-1 was added to HUVECs
at concentrations more than 10-5 mol/L (data not
shown). This result is consistent with the concentration dependency of
glyco-SNAP inhibition of endothelial cell proliferation. As the
addition of exogenous NO suppressed
VEGF165-mediated endothelial cell proliferation,
it is proposed that inhibition of endogenous NO would promote cell
growth. Indeed, incubation of HUVECs with increasing concentrations of
L-NNA potentiated VEGF165-stimulated, at the dose
of 10-4 mol/L L-NNA, endothelial cell
proliferation by 49 ± 14% (P < 0.05,
n = 3) as compared to stimulation with
VEGF165 alone (Figure 9C)
. In addition,
VEGF165-mediated DNA synthesis was completely
blocked by the addition of 10-4 mol/L
8-bromo-cGMP (P < 0.02, n = 3;
Figure 9D
).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
VEGFR-2 is absolutely critical for the earliest stages of vasculogenesis as blood islands, endothelial cell, and major blood vessels fail to develop in appreciable numbers in embryos lacking VEGFR-2.7 The reduction of VEGF-induced angiogenesis in the murine Matrigel assay in the presence of the anti-VEGFR-2 antibody confirms that this receptor is essential for angiogenesis. In the VEGFR-1 knockout mouse embryos the vessels do form, but are disorganized.8 The blockade of the VEGFR-1 with anti-VEGFR-1-neutralizing antibodies prevented vessel stabilization and promoted formation of large aneurysm-like structures in the murine Matrigel angiogenesis assay further supporting a regulatory role for VEGFR-1 in vessel formation.
In two important studies, Fong and co-workers8,9 demonstrated that the primary role of VEGFR-1 during vasculogenesis is to limit the differentiation of hemangioblasts into endothelial cells, it was unclear however how this was prevented by VEGFR-1. Increased hemangioblast commitment in the absence of VEGFR-1 resulted in increased numbers of endothelial cells that caused an overcrowded microvascular environment and markedly disorganized vasculature. The absence of VEGFR-1 resulted in mouse embryos and in cystic embryoid bodies derived from embryonic stem cells in vitro in large, disorganized vascular channels filled with aggregates of hemangioblasts and endothelial cells. These findings are strikingly similar to those observed in both our in vitro and in vivo angiogenesis assays. Inhibition of VEGFR-1 in vivo resulted in the formation of massive blood vessels lined with aggregates of endothelial cells and unchecked proliferation of cultured endothelial cells led to the collection of large cellular aggregates.
VEGF was reported to stimulate NO release from both cultured HUVECs5,19 and intact arterial strips.18 Blockade of VEGFR-2 did not affect VEGF-stimulated NO release in HUVECs. In addition, neither VEGF121 nor the VEGF:PlGF heterodimer that specifically binds to VEGFR-227,29 were able to induce NO release demonstrating that VEGFR-2 activation is not required for NO release in HUVECs. These results are further supported by the ability of PAEVEGFR-1-transfected cells to release NO in response to increasing concentrations of VEGF. Although VEGF up-regulates eNOS expression in PAEVEGFR-2,30 this does not equate to NOS activity. Indeed, it was recently stated that expression of eNOS in bovine endothelial cells does not correlate with activity (Professor Ron R. Magness, University of Wisconsin-Madison, personal communication; and data presented at FASEB Summer Conference, Copper Mountain, CO on June 2000). In our hands, we noted that basal levels of NO were higher in PAEVEGFR-2 cells compared with PAEVEGFR-1, but when VEGF was added, a significant increase in NO release was only observed in PAEVEGFR-1 cells. As sVEGFR-1 inhibited the NO release when added to cells that were previously stimulated for 35 minutes with VEGF, this increase was sustained. These findings however are in contrast to those of Kroll and Waltenberger31 that reported VEGF caused a transient increase in cGMP levels at 2 minutes in PAEVEGFR-2 cells, but at subsequent time points VEGF had no effect. This is surprising as VEGF-mediated NO release in endothelial cells is sustained.5,19,20
Murohara and colleagues32 reported that VEGF stimulates permeability via NO and prostacyclin through VEGFR-2 based on their observation that PlGF that binds to VEGFR-1 but not VEGFR-2, failed to increase vascular permeability. The negative effect of PlGF often used to conclude a lack of biological response via VEGFR-1 is misleading. Our findings clearly demonstrate that PlGF-1 that lacks a heparin-binding domain is able to stimulate NO release in HUVECs, whereas PIGF-2 has no effect on NO production. The apparent lower potency of PlGF-2 may be explained by the fact that it contains a heparin-binding domain.33 We suggest that the use of PlGF as a VEGFR-1-selective ligand to assess VEGF receptor function must involve the use of both isoforms. It is also important to remember that agonists that bind to a common receptor do not necessarily transduce the same signals as exemplified by angiopoietin-1 and angiopoietin-2 that bind to Tie-2.34 Moreover, the fact that PlGF homodimers that bind to VEGFR-1 and PlGF:VEGF heterodimers that bind to VEGFR-2 induced tyrosine phosphorylation of a 38-kd protein in HUVECs whereas VEGF homodimers did not29 supports the view that PlGF and VEGF can exert different biological actions through the same receptor. Thus an absence of a response to PlGF should not be misconstrued as a lack of VEGFR-1 activity.
The specific blockade of VEGFR-1 with a neutralizing anti-VEGFR-1 antibody potentiating HUVEC proliferation and sustaining the maximal mitogenic effect of VEGF suggests that VEGFR-1 transduces a regulatory signal to control VEGFR-2-dependent endothelial cell proliferation. Similar results were reported in trophoblasts5 and in PAE cells transfected with VEGF receptor chimeras.35 However, the underlying mechanism involved in regulating VEGF-mediated endothelial cell proliferation is still unclear, and the role of NO in this process remains controversial. NO may stimulate proliferation of endothelial cells at low concentrations but inhibit their proliferation at higher concentrations by a ribonucleotide-dependent mechanism.36 In the present study, VEGF-stimulated proliferation was attenuated by glyco-SNAP-1 as well as by 8-bromo-cGMP indicating that activation of cGMP by NO controls VEGF-mediated endothelial cell proliferation. Furthermore, the inhibitory effect of the anti-VEGFR-1 antibody on endothelial cell proliferation was bypassed by treatment of cells with NO donor glyco-SNAP-1. Our data are supported by a recent study showing a role for NO in inhibiting endothelial proliferation and promoting in vitro capillary organization induced by basic fibroblast growth factor.37
In the studies by Fong and colleagues,8,9 the primary abnormality in VEGFR-1 null mice was found to be one of differentiation, not increased proliferation of endothelial cell precursors. In contrast, our results clearly indicate a role for VEGFR-1 in the modulation of VEGFR-2-mediated proliferation through the generation of NO. A possible explanation for this difference is that our studies examined the effects of blocking VEGFR-1 during physiological angiogenesis, whereas the observations of Fong and colleagues8,9 were during vascular development. The similar phenotypes suggest a common role for VEGFR-1 in the modulation of endothelial cell differentiation during both embryonic angiogenesis and adult angiogenesis, one that might involve VEGFR-1-mediated NO production because eNOS is expressed as early as day 4 in the embryo during murine development.38 Our data demonstrate that VEGFR-1 signal generates NO that counteracts VEGFR-2-mediated endothelial cell proliferation to allow differentiation to occur. Recently, Hiratsuka and co-workers14 proposed that VEGFR-1 acts as a positive regulator of pathological angiogenesis when levels of PlGF are elevated because murine Lewis lung carcinoma cells overexpressing PlGF-2 grew faster in wild-type mice than in VEGFR-1 tyrosine domain-deficient mice. We have previously shown that addition of PlGF-1 or PlGF-2 to cultured trophoblasts causes tyrosine phosphorylation of VEGFR-2 that is inhibited by anti-VEGF or anti-VEGFR-2 antibodies22 indicating that VEGFR-1, the signaling receptor for PlGF, can modulate VEGFR-2. If a similar mechanism exists in endothelial cells then the observation that PlGF promotes pathological angiogenesis and tumor growth directly by VEGFR-114 could be explained as indirect VEGFR-2 activation. VEGF signaling is very similar in both trophoblast and endothelial cells. We have previously shown that VEGF can induce mitogen-activated protein (MAP) kinase extracellular signal-regulated kinase 1 and 2 (ERK1/2) in trophoblasts39 that was later confirmed in endothelial cells.40
This function of VEGFR-1 may be masked in an in vivo environment leading to the suggestion that VEGFR-1 acts as a VEGF-sink.10 Indeed, Cunningham and colleagues41 demonstrated an interaction between VEGFR-1 and the p85 subunit of the phosphatidylinositol 3-kinase. Moreover, members of the Src family showed an increased level of phosphorylation after VEGF stimulation in PAEVEGFR-1 and not in PAEVEGFR-2 cells.3 Furthermore, VEGFR-1 kinase was shown to display a morphogenic activity as fibroblasts expressing the activated cytoplasmic domain of the VEGFR-1 spontaneously form tubular structures on matrix.42 Moreover, using the novel in vitro co-culture angiogenesis assay, we show that both VEGFR-1 and VEGFR-2 are required for angiogenesis, but it seems that VEGFR-1 may promote vascular connections within the capillary network (branching angiogenesis) whereas VEGFR-2 predominantly promotes tube length between branches. This suggests that endothelial cells stimulated to proliferate and migrate via VEGFR-2 and are unable to organize themselves into vascular structures without VEGFR-1. Indirect evidence that supports this comes from the JunB null mice where the failure to establish proper vascular interactions with the maternal circulation is associated with down-regulation of VEGFR-1. In these mice, fetal blood vessels could only be located in the chorio-allantoic plate and they were not able to penetrate or to branch into the labyrinth trophoblasts.43
In summary, our studies show that VEGFR-1 activation releases NO and
that NO is the cross-talking second messenger that inhibits
VEGFR-2-mediated endothelial cell proliferation and induces these cells
to differentiate into capillary-like structures. A model for VEGF
action during physiological or pathological angiogenesis through the
regulatory pathways mediated by VEGFR-1 and VEGFR-2 is proposed (Figure 10)
. Although the activation of VEGFR-2
may be sufficient for embryonic angiogenesis, activation of
VEGFR-1-mediated NO release may be required to promote the development
of the intricate architecture of the vascular network (branching
angiogenesis) during menstruation, luteal development, and wound
healing. It is therefore likely that altered expression or activity of
VEGFR-1 may cause vascular abnormalities. There is a reciprocal
relation between VEGF and NO in the endogenous regulation of
endothelial integrity after arterial injury44
and in eNOS
knockout mice VEGF fails to induce angiogenesis.21
The
present study, together with earlier reports on
trophoblasts9
and tumor epithelial cells,45
provides clear evidence that NO via VEGFR-1 activation leads to
vascular morphogenesis by arresting VEGF-induced proliferation and
initiating a program of cell differentiation. Thus NO seems to be a
molecular switch for endothelial cell differentiation and that
induction of eNOS may have therapeutic activity in vascular
insufficiencies whereas NOS inhibitors may limit tumor growth.
|
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by a British Heart Foundation grant (no. RG/98/0003) and Wellcome Trust grants.
Accepted for publication May 31, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. Kusmartsev, E. Eruslanov, H. Kubler, T. Tseng, Y. Sakai, Z. Su, S. Kaliberov, A. Heiser, C. Rosser, P. Dahm, et al. Oxidative Stress Regulates Expression of VEGFR1 in Myeloid Cells: Link to Tumor-Induced Immune Suppression in Renal Cell Carcinoma J. Immunol., July 1, 2008; 181(1): 346 - 353. [Abstract] [Full Text] [PDF]
|
