Cell Physiology

Signaling mechanisms mediating vascular protective actions of vascular endothelial growth factor

Ian Zachary


Vascular endothelial growth factor (VEGF) is essential for angiogenesis in health and pathophysiology, and it is currently a major focus for drug targeting in the development of treatments for diverse human diseases. Recently, we proposed that VEGF could also play a role as a vascular protective factor in the adult vasculature and in disease. In this model, vascular protection is defined as a VEGF-induced enhancement of endothelial functions that mediate the inhibition of vascular smooth muscle cell proliferation, enhanced endothelial cell survival, suppression of thrombosis, and anti-inflammatory effects. A feature of this model is that protective effects of VEGF are essentially independent of angiogenesis or endothelial cell proliferation. VEGF-dependent cell survival and VEGF-induced synthesis of nitric oxide and prostacyclin are likely to be key mediators of a vascular protective effect. Vascular protection should help to improve insight into the underlying mechanisms of cardiovascular actions of VEGF and prove valuable for developing novel therapeutic approaches to cardiovascular disease.

  • angiogenesis
  • prostacyclin
  • nitric oxide
  • kinase domain receptor
  • apoptosis
  • endothelium

in embryogenesis, vascular endothelial growth factor (VEGF) is essential for endothelial cell differentiation (vasculogenesis) and for the sprouting of new capillaries from preexisting vessels (angiogenesis) (35, 97). VEGF is also thought to play a key role in postnatal angiogenesis in human pathophysiology including cancer, rheumatoid arthritis, ocular neovascularizing disorders, and cardiovascular disease (37,60). Unlike many diseases in which VEGF-driven neovascularization contributes to disease progression, in ischemic heart and peripheral vascular disease the problem is one of insufficient blood supply, and a recent development has been the use of VEGF to stimulate collateral artery formation in regions with a vascular deficit, an approach termed “therapeutic angiogenesis” (54, 60, 107).

Although the role of VEGF-induced angiogenesis in development and disease has been the main focus of research, VEGF regulates multiple biological functions in endothelial cells in vitro and in the adult vasculature in vivo. Because VEGF deficiency is lethal, relatively little is known about the biological role of VEGF in the normal adult vasculature. On the basis of studies of extravascular VEGF gene transfer in vivo and biological actions of VEGF in cultured endothelial cells, we have developed the concept that vascular protection may be an important mode of action for VEGF in the adult vasculature (103,132). Vascular protection is a mechanism through which VEGF can enhance antiproliferative, antithrombotic, and other protective functions of essentially intact endothelia independently of significant mitogenic or angiogenic effects. Furthermore, VEGF-mediated arterial protection may prove to be useful for the treatment of cardiovascular disease in particular clinical situations (132).

Elucidation of the mechanisms through which VEGF exerts its effects on the cardiovascular system is clearly an essential prerequisite for both understanding the complex biology of this molecule and realizing its therapeutic potential. This review considers the biologically functional signal transduction mechanisms underlying the actions of VEGF within the context of the vascular protection model. Accordingly, the main focus will be on the signaling pathways involved in vascular protective effects of VEGF and, particularly, cell survival and on the regulation of two key vasoactive mediators, nitric oxide (NO) and prostacyclin (PGI2).


VEGF (VEGF-A) is a distant relative of platelet-derived growth factor (PDGF) and is a member of a family of related growth factors that now includes VEGF-B, -C, -D, and -E and placenta growth factor (PlGF) (2, 35, 61, 88, 92). Alternative splicing of human VEGF mRNA from a single gene containing eight exons gives rise to at least five different isoforms of 121, 145, 165, 189 and 206 amino acid residues (35, 93, 97). VEGF121, VEGF145, and VEGF165 are secreted and form dimeric proteins which, except for VEGF121, bind heparin to differing degrees. Human VEGF165 is typically expressed as a 46-kDa homodimer of 23-kDa monomers and is the most abundant and, in in vitro studies, the most biologically active form (35,92). All isoforms possess a signal sequence, but only VEGF121, VEGF145, and VEGF165 are readily diffusible and have been unambiguously demonstrated to have biological effects in endothelial cells. Exons 1–5 encode the core regions essential for binding to the receptors VEGF receptor 2 (VEGFR2) and VEGFR1 (see vegf receptors). Exon 6 encodes a region rich in basic amino acid residues and is a putative heparin-binding domain. Exon 7 is implicated in binding to both heparin and neuropilin-1 (NP-1). Exon 6 is absent in VEGF121 and VEGF165, whereas exon 7 is absent from VEGF121and VEGF145. The central role of VEGF in embryonic blood vessel development has been confirmed by the finding that targeted inactivation of only a single allele of the VEGF gene in mice causes a lethal impairment of endothelial cell differentiation, development of the primitive vascular plexus, and angiogenesis (16). The biological role of other VEGFs is less clear, but VEGF-C is strongly implicated in lymph angiogenesis (92) and promotes angiogenesis in vivo (130).


Two distinct receptor tyrosine kinases (RTKs) have been identified for VEGF, VEGFR1 (Flt-1) and VEGFR2 (KDR/Flk-1), which share ∼44% amino acid homology with each other (see reviews, Refs.84 and 92). A third receptor, VEGFR3 (Flt-4), binds VEGF-C and -D and does not bind VEGF-A (2, 62). PlGF and VEGF-B bind with high affinity only to VEGFR1 and do not bind to VEGFR2 (15, 20, 84, 92). VEGF/PlGF heterodimers do, however, have biological activity in vivo and exhibit high-affinity binding to VEGFR2 (15). VEGF-E binds with high affinity to VEGFR2. The specificities of members of the VEGF family for their receptors are illustrated in 1. The three VEGF receptors are structurally related to the PDGF family of RTKs (class III) and have a similar domain structure characterized by cytoplasmic regions with an insert sequence within the catalytic domain, a single hydrophobic transmembrane domain, and seven immunoglobulin-like domains in the extracellular regions (Fig. 1). The reported affinities of VEGF for VEGFR1 and VEGFR2 are, respectively, 16–114 pM and 400–1,000 pM (15, 20, 29, 78, 79, 94, 115, 122).

Fig. 1.

Vascular endothelial growth factor (VEGF) ligands and receptors. VEGF tyrosine kinase receptors are related to the platelet-derived growth factor (PDGF) subfamily of receptor protein tyrosine kinases (RTKs; class III) and possess an extracellular domain (ECD) containing 7 immunoglobulin-like loops (red ovals), a single hydrophobic membrane-spanning domain (TM; white boxes), and a large cytoplasmic domain comprising a single catalytic domain (Cat; pink boxes) containing all the conserved motifs found in other RTKs and that is interrupted by a non-catalytic region, called the kinase insert (KI; yellow boxes), within the catalytic domain. The extracellular domain of VEGFR1 is also independently expressed as a soluble protein. VEGFR3 undergoes proteolytic processing to yield disulfide-linked 120- and 75-kDa polypeptides. VEGF binds with high affinity to both VEGFR2 (KDR/Flk-1) and VEGFR1 (Flt-1) receptors. Placenta growth factor (PlGF) and VEGF-B exhibit high-affinity binding to VEGFR1 only. VEGF-C and -D are VEGF-related factors that bind to a related receptor, Flt-4 (VEGFR3), and also to VEGFR2. Neuropilin-1 (NP-1) is a novel non-RTK receptor for VEGF165. VEGFR1 and VEGFR2 are present in vascular endothelium, and VEGFR1 is also uniquely expressed on monocytes, while VEGFR3 is preferentially expressed on lymphatic endothelium. NP-1 is a non-tyrosine kinase receptor expressed in some tumor cells and in endothelial cells (EC). The structure of NP-1 comprises an extracellular region with MAM (or C), a and b domains, a transmembrane region, and a short cytoplasmic domain (Cyt).

Targeted disruption of VEGFR1 and R2 in mice prevents normal vascularization and embryonic development, but the two knockouts have distinctive phenotypes. VEGFR2-deficient mice produce neither differentiated endothelial cells nor organized blood vessels and also possess no hematopoietic precursors, suggesting that this receptor is essential for development of both endothelial and hematopoietic precursors (97, 104). In contrast, the VEGFR1 knockout mice possess mature, differentiated endothelial cells but have large, disorganized vessels (38). Surprisingly, the primary defect in VEGFR1 knockout mice appears to be overproduction of endothelial progenitor cells (39), rather than vascular disorganization per se, consistent with a negative regulatory role for VEGFR1 (see vegf receptor signaling).

VEGFR3 expression starts during E8 in developing blood vessels but subsequently is largely confined to the lymphatic vasculature, consistent with a specific role of this receptor in lymph angiogenesis. Disruption of the VEGFR3 gene in mice did not prevent vasculogenesis or angiogenesis but caused defects in normal vascular development, leading to fluid accumulation from leaky vessels by E9.5 (33).

Neuropilin-1 (NP-1) was recently identified as a new receptor for VEGF (108) (Fig. 1). NP-1 is a non-tyrosine kinase transmembrane receptor with a short cytoplasmic tail and a large extracellular domain (92) (Fig. 1), earlier identified as a receptor for the semaphorin/collapsin family of polypeptides implicated in axonal guidance (66). Previous work had pointed to a role for NP-1 in cardiovascular development. Overexpression of NP-1 in mice results in diverse vascular abnormalities, including excess capillaries and blood vessels, and in malformation of the heart (65), while NP-1 knockout mice display impaired neural vascularization, defects in the aorta and other large blood vessels, and aberrant yolk sac vascularization (63). In the human fetal heart, NP-1 is expressed in the endocardium, coronary vessels, myocardial capillaries, and epicardial blood vessels and is coexpressed with VEGFR1 and VEGFR2 in the endocardium and myocardial capillaries, but only with VEGFR1 in coronary vessels (90).

VEGF165, but not VEGF121, was shown to bind to NP-1 with an affinity similar to that for VEGFR2 (dissociation constant ∼0.3 nM), and coexpression of NP-1 with VEGFR2 increased binding of VEGF165 fourfold (108). Because NP-1 has a short cytoplasmic tail with no known signaling function and is able to bind a variety of semaphorins with equal affinity but with each having different biological activities, NP-1 by itself may not be a functional receptor but may act as a “docking” coreceptor for VEGFR2. A major NP-1 binding site for VEGF165 has been mapped to the domain encoded by exons 7 and 8 in VEGF165(108). NP-1 has also been shown to bind VEGF-B and -E and the PlGF-2 splice variant (92). The role of NP-1 in the functions of VEGF in vitro and in vivo is not yet known, and the mechanisms by which it acts as a coreceptor have not been clarified.


Evidence that VEGF has a vascular protective effect in the adult vasculature came from the finding that perivascular VEGF gene transfer inhibits neointima formation in a non-endothelial injury rabbit carotid artery model (10, 71, 72, 110). With the use of the collar as a gene delivery reservoir, extravascular VEGF gene transfer was found to strikingly inhibit neointimal smooth muscle cell (SMC) hyperplasia (72) in the absence of angiogenesis. The endothelial nitric oxide synthase (eNOS) inhibitor nitro-l-arginine methyl ester (l-NAME) prevented VEGF-mediated inhibition of neointima formation, suggesting that the NO pathway is involved (72).

Several investigators have established that VEGF stimulates endothelial production of NO and PGI2, intercellular mediators predicted to have a vascular protective effect (58, 68, 69, 72,82, 103, 117, 119, 128, 132). Though the role of these factors in vasodilatation is well known, NO and PGI2 have other potentially vascular protective effects, including antiproliferative effects in SMCs, anti-platelet actions, and, in the case of NO, inhibition of leukocyte interactions with endothelium (Fig.2). In addition, both NO and PGI2 may also mediate angiogenic and permeability-increasing effects of VEGF.

Fig. 2.

Cellular mechanisms of VEGF-mediated vascular protection. VEGF production in arteries may be increased by gene transfer, or endogenous production may be upregulated in vascular smooth muscle cells (VSMC) by hypoxia, growth factors (basic fibroblast growth factor and PDGF-BB), and cytokines. Intimal thickening could reduce oxygen tension and lead to increased expression of regulatory factors in medial VSMC in vivo, leading to increased VEGF production. VEGF is most likely to act through receptors (VEGFR2 and possibly VEGFR1) in the endothelium to increase production of nitric oxide (NO) and prostacyclin (PGI2) and augment intracellular endothelial cell survival signaling. In VSMC, NO activates soluble guanylate cyclase (GC), leading to increased cGMP synthesis, and PGI2 binds to specific G protein-coupled prostacyclin receptors (IP), which activate adenylyl cyclase (AC) and increase cAMP synthesis. Both intracellular cyclic nucleotides mediate antiproliferative and vasodilatory effects in VSMC. NO and PGI2 are predicted to have other biological consequences: decreased platelet aggregation, thrombosis, and, in the case of NO, inhibition of leukocyte adhesion. The combined effect of these biological actions is vascular protection.

Antimitogenic effects of NO and PGI2 on SMCs have been demonstrated in vitro and in vivo acting via production of the intracellular messengers cGMP and cAMP, respectively (6, 7, 23,44, 73, 121). Though short-term PGI2 administration failed to inhibit restenosis after balloon injury (8, 47), gene transfer of PGI synthase was shown to accelerate reendothelialization and reduce neointima formation after balloon injury (86). eNOS gene transfer also reduces neointimal hyperplasia in balloon injury models of restenosis (120,121). Inhibition of neointimal SMC hyperplasia following VEGF delivery in the rabbit collared carotid artery or balloon denudation and stent implantation may be mediated in part through the antimitogenic effects of these two intercellular mediators.

An ability of VEGF to inhibit platelet aggregation (Fig. 2) and hence exert an antithrombotic effect is predictable from increased NO and PGI2 production (101, 129). Though vascular VEGF delivery markedly reduces mural thrombus formation after balloon injury-induced intimal thickening (60, 132), there has so far been no direct demonstration of an antithrombotic effect of VEGF mediated by NO and PGI2. In fact, current findings suggest that VEGF stimulates prothrombotic as well as antithrombotic pathways. Thus VEGF increases the expression and activation of the fibrinolytic proteases urokinase and tissue-type plasminogen activator (91) but also induces expression of the prothrombotic components plasminogen activator inhibitor, von Willebrand factor (vWF), and tissue factor (11, 20, 83, 91, 128). However, VEGF only appears to increase surface expression of active tissue factor on endothelial cells in cooperation with tumor necrosis factor-α (14). Other findings may point toward a role for vWF and tissue factor in angiogenic functions of VEGF. Mice embryos deficient in tissue factor have an impaired pattern of extraembryonic angiogenesis (17, 106), and vWF increases endothelial cell adhesion (28). It is noteworthy that VEGF is released by platelets, that its synthesis is increased by thrombopoietin in megakaryocytic cell lines, and that increased levels of VEGF are found at the site of hemostatic plugs in humans (81, 124, 127). Whether VEGF regulates platelet function and/or thrombosis remains an elusive but potentially important aspect of the vascular protective action of VEGF that needs to be addressed.

VEGF may also cause an impairment of leukocyte interactions with the endothelium (Fig. 2) via the ability of endogenous NO synthesis to inhibit leukocyte rolling and adhesion and upregulation of intercellular adhesion molecule-1 and vascular cell adhesion molecule-1 (27, 70). Striking support for this hypothesis has come from the finding that systemic administration of VEGF attenuates thrombin- or l-NAME-induced leukocyte adhesion to the mesenteric endothelium in rats. VEGF was unable to inhibit leukocyte interactions with the endothelium in eNOS-deficient mice, indicating a requirement of endothelial NO generation for mediation of this action of VEGF (100). Because adhesion molecule expression and leukocyte adhesion are important triggers during the early stages of atherosclerosis, VEGF-induced NO synthesis may have an anti-inflammatory effect with the potential to protect against proatherogenic factors.


It remains problematic whether VEGFR1 is able to transduce a signal in endothelial cells. Some studies have reported activation of phospholipase C (PLC)-γ in VEGFR1-expressing cells (74,102) and Ca2+ mobilization in trophoblast cells (3). VEGFR1 also interacts with the p85 subunit of phosphatidylinositol 3′-kinase (PI 3-kinase) in a yeast two-hybrid system (26), but so far this has not been associated with a biological activity in VEGFR1-expressing cells. In general, signaling mediated by VEGFR1 appears to be weak and largely dissociated from biological effects such as proliferation and migration in endothelial cells. VEGFR1 is, however, essential for normal vascular development.

Intriguingly, mice expressing only the extracellular domain of VEGFR1 and lacking the kinase domain develop normally (53), suggesting that signaling through the kinase domain of VEGFR1 is dispensable for endothelial cell differentiation or angiogenesis. Together with the phenotype of VEGFR1 knockout mice discussed earlier in vegf receptors, other findings support the conclusion that VEGFR1 functions as a negative regulator of VEGFR2. A soluble form of the extracellular VEGFR1 domain occurs naturally, and overexpression of this form, but not of an artificial soluble VEGFR2, inhibits VEGF-induced migration and proliferation of human microvascular endothelial cells and human umbilical vein endothelial cells (HUVECs) by forming an inactive complex with VEGF and the full-length VEGFR2 (64, 98). A study of chimeras made between the extracellular domain of the receptor for colony-stimulating factor 1 (CSF-1) and the kinase domains of either VEGFR1 or VEGFR2 found that activation of the R1 kinase domain chimera with CSF-1 suppressed ERK activation and proliferation mediated via the R2 kinase chimera (95). Analysis of domain swapping between VEGFR1 and VEGFR2 indicates that a short motif (ANGG) unique to the intracellular juxtamembrane VEGFR1 domain can suppress VEGFR2-mediated signaling and cell migration (48).

VEGFR1 does appear to mediate VEGF- and PlGF-induced tissue factor expression and chemotaxis in monocytes, a cell type that expresses only VEGFR1 (19, 20). Moreover, migration of macrophages was suppressed in VEGFR1 kinase domain-deficient mice, suggesting that signaling through this receptor may play a role in regulation of monocyte/macrophage functions in vivo (53). VEGFR1 is also implicated in VEGF stimulation of urokinase and plasminogen activator inhibitor 1 (87) and of metalloproteinase expression in SMCs (123).

Like other class III RTKs, VEGFR2 is presumably activated through ligand-stimulated receptor dimerization and transphosphorylation (autophosphorylation) of tyrosine residues in the cytoplasmic kinase domain. Four major autophosphorylation sites have been identified in VEGFR2: Y951, Y996, Y1054, and Y1059 (32). Y951 and Y996 are in the kinase insert region, and Y1054 and Y1059 are in the kinase domain. Like other RTKs, VEGF has been reported to stimulate tyrosine phosphorylation of several SH2 domain-containing signaling proteins (1, 51, 67). Overall, most functional endothelial VEGF cell signaling described to date is either known to be mediated via VEGFR2 or strongly suspected to involve primarily VEGFR2 on the basis of ligand specificity. Almost all of what follows therefore relates primarily to VEGFR2-mediated signaling.


A fundamental cellular mechanism by which VEGF can exert a vascular protective effect and augment endothelial function is the promotion and maintenance of endothelial cell survival (Fig. 2). VEGF was first shown to act as a survival factor for retinal endothelial cells (4). VEGF has been reported to inhibit HUVEC apoptosis by activating the antiapoptotic kinase Akt/PKB via a PI 3′-kinase-dependent pathway (46, 116) (Fig.3). As discussed below, an important link has recently been established between the Akt pathway and Ca2+-independent NO generation. Longer term effects of VEGF on cell survival may be mediated through the upregulation of the antiapoptotic proteins Bcl-2 and A1 (45).

Fig. 3.

VEGF survival signaling. VEGF-dependent endothelial cell survival is mediated in part via phosphatidylinositol 3-kinase (PI3K)-mediated activation of the antiapoptotic kinase Akt. Increased tyrosine phosphorylation of focal adhesion kinase (FAK), which has been strongly implicated in cell survival signaling, is a point of convergence for diverse endothelial cell survival stimuli, including VEGF, matrix-integrin interactions, and fluid shear stress (FSS). FAK may also be important for endothelial cell migration and, hence, angiogenesis. Interactions between the integrin αvβ3 and VEGFR2 may also play a role in survival functions of VEGF. Longer term antiapoptotic effects of VEGF may involve upregulation of antiapoptotic proteins such as Bcl-2 and A1.

Another mechanism through which VEGF could maintain survival signals in endothelial cells is tyrosine phosphorylation of focal adhesion kinase (FAK) (see review, Ref. 133) (see Fig. 3). VEGF increases tyrosine phosphorylation and focal adhesion association of FAK and the FAK-associated protein paxillin (1, 99). VEGF also stimulates tyrosine phosphorylation of the FAK-related tyrosine kinase Pyk2 in a bone marrow endothelial cell line (76). FAK is critical for maintaining survival signals in several adherent cell types, and in endothelial cells FAK tyrosine dephosphorylation and caspase-mediated proteolytic cleavage are early responses to apoptogenic stimuli (40, 59, 75, 77, 118). While these findings suggest that VEGF-dependent survival signaling is relayed in part through increased FAK tyrosine phosphorylation, it is important to note that FAK is strongly implicated in cell migration and, specifically, in VEGF-induced changes in actin cytoskeletal organization (1, 99). FAK may therefore also play a role in the chemotactic response to VEGF, which in turn has implications for angiogenic signaling.

VEGF-dependent survival and migration may also involve interactions between integrins and VEGF receptors (Fig. 3). VEGFR2 associates selectively with αvβ3 (13), and VEGF mitogenicity and receptor activity were enhanced by endothelial adhesion to the αvβ3 ligand vitronectin (109). However, β3-null mice exhibited no defects in retinal neovascularization, a physiological angiogenic context in which αvβ3 has been strongly implicated (55). This latter finding suggests that, at least in the case of some angiogenic (and presumably survival) functions of VEGF, interactions between VEGFR2 and αvβ3 may play a redundant role.

Short-term NO production induced by VEGF probably involves activation of the constitutive eNOS isoform. This may occur in part by VEGF-induced Ca2+ mobilization (24, 25) in common with other activators of eNOS (Fig.4). Another mechanism for VEGF-dependent NOS activation may be through activation of the heat shock protein Hsp 90 or an Hsp 90-associated protein (42). Activation of Hsp 90 seems to increase its affinity for and association with eNOS to stimulate eNOS activity (42). The signaling pathways mediating VEGF-induced NO production may be still more complex. Activation of c-Src was reported to mediate VEGF signaling through PLC-γ, leading to inositol 1,4,5-trisphosphate formation and Ca2+ mobilization (52). An important mechanism through which VEGF could induce prolonged NO production is increased expression of eNOS mRNA and protein (9, 57, 105). The mechanism mediating VEGF-stimulated eNOS expression is unclear, but the effect appears to occur via activation of VEGFR2 and was inhibited by selective protein kinase C (PKC) inhibitors, suggestive of a role for PKC in VEGF-dependent gene expression (105). A novel insight into how VEGF can promote sustained generation of NO has come from the discovery that Akt activation by VEGF and shear stress mediates phosphorylation of eNOS at serine 1179 to cause Ca2+-independent NO generation (22, 30, 36, 41,43) (Fig. 4).

Fig. 4.

Mechanisms mediating VEGF-induced NO and PGI2 synthesis. Short-term NO production induced by VEGF is mediated via increased cytosolic Ca2+, resulting from activation of phospholipase C (PLC)-γ and subsequent generation of inositol 1,4,5-trisphosphate (IP3). c-Src has been implicated in signaling upstream of PLC-γ. Activation of Akt leads to phosphorylation and activation of endothelial NO synthase (eNOS-P), providing a mechanism for sustained Ca2+-independent NO synthesis. PLC-γ-mediated production of diacylglycerol (DAG) leads to activation of PKC, and this pathway plays an important role in mediating VEGF-induced activation of extracellular signal-regulated kinases (ERKs). In turn, ERK activation mediates cytosolic phospholipase A2(cPLA2)-mediated PGI2 synthesis. Increased cytosolic Ca2+ also stimulates the cellular release of PGI2. Protein kinase C (PKC) is also reported to mediate VEGF-induced upregulation of eNOS, another mechanism leading to long-term NO generation. AA, arachidonic acid; COX-1, cyclooxygenase-1.

In addition to the protective functions of VEGF-induced NO production discussed above, NO is also strongly implicated in VEGF-induced angiogenesis (134) and permeability (82), NO regulates focal adhesion integrity and FAK tyrosine phosphorylation in endothelial cells (50) (Fig. 3), and NO production plays a permissive role in VEGF-induced endothelial cell migration (85).

VEGF-induced PGI2 production is mediated via extracellular signal-regulated kinase (ERK) 1/2-dependent activation of cytosolic phospholipase A2 (128), and our recent findings indicate that both ERK activation and PGI2generation are mediated through PKC (49) (Fig. 4). Similar to NO, PGI2 production has also been shown to mediate other biological functions of VEGF, including increased vascular permeability (82).

A novel aspect of VEGF receptor signaling concerns the mechanism of ERK activation. The paradigmatic pathway through which protein tyrosine kinase receptors activate ERKs involves tyrosine phosphorylation and receptor association of the adapter protein Grb-2, subsequent stimulation of the guanine nucleotide exchange protein SOS, and activation of Ras, which in turn activates Raf-1 and the distal ERK cascade. VEGF stimulates tyrosine phosphorylation of Shc and promotes the formation of a complex between Shc and Grb-2 in porcine aortic endothelial cells overexpressing VEGFR2 (67). However, neither tyrosine phosphorylation of Shc nor VEGF-stimulated association between Shc and Grb-2 have been shown to lead to activation of Ras or the ERK pathway. Moreover, VEGF induces Ras-independent activation of the Raf-MEK-ERK pathway in sinusoidal endothelial cells (31,114) and PKC-mediated ERK activation in HUVECs (49). There have been two reports that NO may mediate VEGF-induced Raf-1/ERK activation (56, 89), but the mechanism involved is not clear. However, our own findings suggest that VEGF-induced ERK activation and PGI2 production in HUVECs are unaffected by inhibitors of eNOS, indicating that the VEGF signaling pathways leading to NO and PGI2 generation bifurcate upstream of ERK (49). Together, these findings suggest that VEGFR2 may be unique among RTKs in activating the ERK cascade via a PKC-dependent, Ras-independent pathway.

Activation of the ERK cascade and activation of PLC-γ, leading to generation of diacylglycerol and inositol 1,4,5-trisphosphate and subsequent PKC activation and Ca2+ mobilization, are also strongly implicated in VEGF mitogenic signaling (113, 126,131). The involvement of PKC in the mitogenic effects of VEGF is indicated by the finding that antisense oligonucleotides to PKC-α and PKC-ζ isoforms block VEGF-induced endothelial cell proliferation (126).

An aspect of VEGF signal transduction that remains largely obscure concerns the mechanism(s) responsible for VEGF-induced effects on vascular permeability (21, 35). As noted above, NO and PGI2 production have been implicated in VEGF-induced permeability changes, but their possible mechanism of action is not known. VEGF increases phosphorylation of VE-cadherin (34), the major component of endothelial intercellular adherens junctions, and of the tight junction proteins occludin and zonula occludens-1 (5). Phosphorylation of junctional components may be a mechanism through which cell-cell adhesions are weakened, leading to increased permeability, though this has not yet been established for VEGF.

VEGF increases the expression of diverse genes and proteins (see review, Ref. 92), but the role of VEGF-induced gene expression in either angiogenesis or vasculogenesis is still poorly understood and is not considered in detail in the present review.


An outstanding question is whether there is a physiological role in the adult vasculature for VEGF-mediated vascular protection. The abilities of VEGF to induce NO and PGI2 production, increase endothelial integrity and survival, and inhibit intimal SMC proliferation make it a particularly attractive candidate for an endogenous vascular protective factor. SMCs produce VEGF in response to hypoxia, growth factors, and cytokines (12, 111, 112) (Fig. 2). Intimal thickening and plaque formation are associated with increased production of growth factors and cytokines and may cause reduced oxygen tension in medial SMCs by increasing the diffusion distance of oxygen from the lumen. The atherosclerotic milieu may therefore promote endogenous VEGF synthesis, and, in agreement with this hypothesis, VEGF expression has been demonstrated in atherosclerotic lesions (18, 96). Reduced expression or impaired function of VEGF would in turn be predicted to attenuate endothelial antiproliferative and antithrombotic functions and, hence, encourage SMC proliferation and promote atherogenesis. Conditional VEGF or VEGF receptor knockouts in animal organisms may be required to determine whether VEGF has a maintenance and/or protective role in adult blood vessels. Studies of the effects of VEGF in the apolipoprotein E-deficient mouse model of atherosclerosis may also be useful for evaluating the antiatherogenic potential of VEGF.

The use of VEGF as a therapeutic angiogenic cytokine in human heart disease has attracted much interest, but recently the problems associated with proangiogenic therapy have been highlighted (see review, Ref. 132). The use of targeted low-level VEGF delivery to augment vascular protective effects locally in the absence of angiogenesis may be a way in which the therapeutic potential of VEGF can be harnessed more effectively and possibly more safely. However, the discussion of the implications of VEGF-mediated biological actions for thrombosis highlighted the difficulty of integrating these diverse effects into the vascular protection model. It is also likely that the context, in terms of pathophysiology, tissue type, and the cytokine milieu, will be crucial for determining the overall outcome of VEGF action. In turn, this suggests that VEGF could have deleterious as well as beneficial consequences for the cardiovascular system depending on the site of action, the specific type of disease (e.g., bypass graft, angioplasty) being targeted, and the presence of other cooperating cytokines. These considerations are particularly relevant for VEGF therapy. Thus VEGF delivered locally to the site of anastomosis in a bypass graft may reduce the risk of stenosis, while VEGF within an existing atherosclerotic plaque could have the contradictory effects of enhancing endothelium-dependent protective functions on one hand and inducing neovascularization on the other hand. It flows from this observation that both the careful selection of the pathophysiological context in which VEGF is delivered to patients and the need for targeted delivery are likely to be crucial for ensuring successful VEGF therapy.


I. Zachary is supported by the British Heart Foundation.


  • Address for reprint requests and other correspondence: I. Zachary, Dept. of Medicine, Univ. College London, 5 Univ. St., London WC1E 6JJ, United Kingdom.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 81.
  81. 82.
  82. 83.
  83. 84.
  84. 85.
  85. 86.
  86. 87.
  87. 88.
  88. 89.
  89. 90.
  90. 91.
  91. 92.
  92. 93.
  93. 94.
  94. 95.
  95. 96.
  96. 97.
  97. 98.
  98. 99.
  99. 100.
  100. 101.
  101. 102.
  102. 103.
  103. 104.
  104. 105.
  105. 106.
  106. 107.
  107. 108.
  108. 109.
  109. 110.
  110. 111.
  111. 112.
  112. 113.
  113. 114.
  114. 115.
  115. 116.
  116. 117.
  117. 118.
  118. 119.
  119. 120.
  120. 121.
  121. 122.
  122. 123.
  123. 124.
  124. 126.
  125. 127.
  126. 128.
  127. 129.
  128. 130.
  129. 131.
  130. 132.
  131. 133.
  132. 134.
View Abstract