Skip to main page content
U.S. flag

An official website of the United States government

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Review
. 2014 Feb;25(1):1-19.
doi: 10.1016/j.cytogfr.2013.11.002. Epub 2013 Nov 27.

Extracellular regulation of VEGF: isoforms, proteolysis, and vascular patterning

Affiliations
Review

Extracellular regulation of VEGF: isoforms, proteolysis, and vascular patterning

Prakash Vempati et al. Cytokine Growth Factor Rev. 2014 Feb.

Abstract

The regulation of vascular endothelial growth factor A (VEGF) is critical to neovascularization in numerous tissues under physiological and pathological conditions. VEGF has multiple isoforms, created by alternative splicing or proteolytic cleavage, and characterized by different receptor-binding and matrix-binding properties. These isoforms are known to give rise to a spectrum of angiogenesis patterns marked by differences in branching, which has functional implications for tissues. In this review, we detail the extensive extracellular regulation of VEGF and the ability of VEGF to dictate the vascular phenotype. We explore the role of VEGF-releasing proteases and soluble carrier molecules on VEGF activity. While proteases such as MMP9 can 'release' matrix-bound VEGF and promote angiogenesis, for example as a key step in carcinogenesis, proteases can also suppress VEGF's angiogenic effects. We explore what dictates pro- or anti-angiogenic behavior. We also seek to understand the phenomenon of VEGF gradient formation. Strong VEGF gradients are thought to be due to decreased rates of diffusion from reversible matrix binding, however theoretical studies show that this scenario cannot give rise to lasting VEGF gradients in vivo. We propose that gradients are formed through degradation of sequestered VEGF. Finally, we review how different aspects of the VEGF signal, such as its concentration, gradient, matrix-binding, and NRP1-binding can differentially affect angiogenesis. We explore how this allows VEGF to regulate the formation of vascular networks across a spectrum of high to low branching densities, and from normal to pathological angiogenesis. A better understanding of the control of angiogenesis is necessary to improve upon limitations of current angiogenic therapies.

Keywords: Angiogenesis; Computational model; Extracellular matrix; Gradient; Mathematical model; Microenvironment; Protease; Receptor; Systems biology.

PubMed Disclaimer

Figures

Figure 1
Figure 1. Properties of VEGF isoforms and proteolytic cleavage sites
A, The acidity of the individual amino acids (pI) for human (black) and murine (red) VEGF shows the basic residues responsible for the heparin-binding domains of exons 6 and 7. Murine VEGF contains a deletion of Gly-8 found in human VEGF and is frame-shifted for comparison. The overall sequence identity between murine and human VEGF189 orthologs is 89%. B, Exon structure of the predominant VEGF isoforms in humans, scaled to the 189 amino acids shown in panel A. Note that VEGF165/164 replace the last residue in exon 5, Lys, with Asp. Exon 1 is not present in processed VEGF, it is removed by signal peptidase. Of the anti-angiogenic VEGFxxxb isoforms, which use exon 8b instead of 8a, VEGF165b is most common. C, Disulfide bonding structure (black and purple lines) (223) and known proteolytic cleavage sites for the serine proteases (plasmin, green; uPA, blue) and MMPs (red). The uPA cleavage site has not been specifically mapped but is thought to reside in the C-terminal portion of exon 6a. Exon 7 is linked to the first amino acid in exon 8 in all isoforms except VEGF121. Processing of exon 7 by the MMPs results in a single fragment due to linkage by disulfide bonds, whereas complete processing by plasmin may yield a second two-amino-acid Arg147-Lys148 (in VEGF189) fragment. D, VEGF receptors (VEGFR1, R2, R3) bind to the bivalent dimeric ligand in the region encoded by exons 3 and 4 (blue). Neuropilins (NRP) and heparan sulfates (HS) bind VEGF in the region encoded by exons 6 and 7 (yellow/purple). E, VEGF isoforms differentially bind to Neuropilin-1, VEGFR2, and heparin/heparan sulfate. Cleavage by plasmin (P) and uPA (U) can activate VEGFR2 binding in VEGF189 and decreases NRP1 and HSPG binding.
Figure 2
Figure 2. Relative expression level of the VEGF isoforms in normal tissues and tumors
This ternary diagram depicts relative isoform expression based on mRNA RT-PCR of cell culture, or of embryonic, adult, and tumor tissues (see Supplemental Table S5). The location of each point on this ternary diagram denotes the relative expression of VEGF121, of VEGF165, and of the exon 6-containing isoforms grouped together; VEGF145 typically has low levels of expression and VEGF183 seems to be functionally equivalent to VEGF189 (62). At vertices, all VEGF expression is comprised only of that isoform, while the midpoint of the opposite edge would indicate 0% expression of that isoform and 50% expression each of other isoforms. Each point represents specific organs from individual studies; points with the same color denote the same organ of origin; circles: tumors, squares: normal embryonic or adult tissue. The dashed line represents a line of equal VEGF121:VEGF189 expression, as discussed in Section 5.2; the VEGF164-only and VEGF120/188 mice would be expected to fall on this line. Note that most tumors fall below this line (i.e. high VEGF120/121 expression).
Figure 3
Figure 3. Mechanisms of proteolytic regulation of extracellular VEGF
Proteases play a key role in determining the fate of VEGF and its detection by endothelial cells. Proteases can degrade soluble VEGF inhibitors (e.g. CTGF, sVEGFR1) (A) or release matrix-sequestered VEGF (B); both allow free VEGF to escape inactive states and bind to endothelial cell receptors (C). VEGF gradients are altered by release of matrix-sequestered VEGF. In tumor angiogenesis, cancer cells, endothelial cells and inflammatory cells can all contribute to proteolytic activity. VEGF164 has two heparin-binding domains; cleavage of either domain results in a VEGF164/113 intermediate (B), with lower overall affinity for the ECM. Subsequent cleavage of the second domain results in freely diffusing VEGF113. Some MMPs can cleave VEGF bound to heparin/HSPGs (e.g. MMP3) while others cannot (e.g. MMP9). Heparanase activity on cell-surface HSPGs can lead to upregulation of MMP9 leading to cleavage of both GAG chains and core protein (D) and enhanced signaling at the cell surface.
Figure 4
Figure 4. In vivo patterning and biological effects of VEGF isoforms
Heparin-binding affinity and proteolytic susceptibility modulate the VEGF patterning in tissues and the subsequent vascular phenotype. A, Angiogenesis in tumor xenografts expressing VEGF120 or VEGF113 only (left); VEGF164 only (middle); VEGF188 or the non-cleavable VEGF164Δ108–118 only (right). Note differences in spatial distribution with increase in cell-surface associated VEGF (green outlines) in the absence of proteases or in presence of VEGF188/189 (16, 17, 111). Heavier VEGF isoforms result in networks with greater capillary density, lower caliber, and increased pericyte coverage. B, Hindbrain VEGF distribution and angiogenesis in mice secreting (i) VEGF120 only or (ii) wildtype VEGF, which is predominantly VEGF164 (15). VEGF localizes closer to the source in wildtype relative to VEGF120-secreting hindbrain; note that the two VEGF gradients will intersect if superimposed. C, Postnatal murine retina (14) showed diffuse VEGF distribution (green dots) with a lack of association of VEGF to astrocytes (yellow), plus greater intercellular VEGF in VEGF120-only mice (left) compared to wildtype mice or VEGF164-only mice (right). Note the enlarged vessels and apparent increase in total VEGF staining in VEGF120-secreting retina.
Figure 5
Figure 5. Model-predicted effects of degradation and sequestration on VEGF patterning
A, Isoform-specific VEGF patterning as seen in vivo relies on isoform-specific differences in both degradation and sequestration (19). The graphs show the spatial distribution of the total VEGF (soluble + sequestered) for three VEGF isoforms (representing increasing matrix binding affinities and/or degradation). All conditions have identical secretion rates. Graphs are scaled to maximum concentration of the lowest affinity isoform (VEGF121 – solid line, identical in each case), except the inset in the middle graph which shows each distribution normalized to its maximum concentration, to show relative steepness. Arrows indicate effect of increasing degradation and/or sequestration. B, In vivo, different rates of uptake of VEGF isoforms by surrounding tissues may account for experimentally observed gradients. We hypothesize that VEGF165 (left) shows greater pericellular accumulation and localization because of dual effects of greater sequestration by HSPGs or cell-surface receptors, and greater degradation (loss from the system) by the resultant cellular internalization. Secretion of VEGF121 (right) has lower binding to cells and lower degradation, and thus increased levels in solution and dispersed spatial gradients. Internalization (arrows) may be due to interstitial cells as well as the endothelium, because both cell types can express VEGF receptors.

Similar articles

Cited by

References

    1. Koch S, Claesson-Welsh L. Signal transduction by vascular endothelial growth factor receptors. Cold Spring Harb Perspect Med. 2012;2:a006502. - PMC - PubMed
    1. Mac Gabhann F, Popel AS. Systems biology of vascular endothelial growth factors. Microcirculation. 2008;15:715–738. - PMC - PubMed
    1. Carmeliet P, Jain RK. Molecular mechanisms and clinical applications of angiogenesis. Nature. 2011;473:298–307. - PMC - PubMed
    1. Egginton S. In vivo shear stress response. Biochem Soc Trans. 2011;39:1633–1638. - PubMed
    1. Chung AS, Ferrara N. Developmental and pathological angiogenesis. Annu Rev Cell Dev Biol. 2011;27:563–584. - PubMed

Publication types

MeSH terms