Systems biology of vascular endothelial growth factors

Abstract

Several cytokine families have roles in the development, maintenance, and remodeling of the microcirculation. Of these, the vascular endothelial growth factor (VEGF) family is one of the best studied and one of the most complex. Five VEGF ligand genes and five cell-surface receptor genes are known in the human, and each of these may be transcribed as multiple splice isoforms to generate an extensive family of proteins, many of which are subject to further proteolytic processing. Using the VEGF family as an example, we describe the current knowledge of growth-factor expression, processing, and transport in vivo. Experimental studies and computational simulations are being used to measure and predict the activity of these molecules, and we describe avenues of research that seek to fill the remaining gaps in our understanding of VEGF family behavior.

Figure 1. Complexity of expression regulation and control for VEGF family ligands and receptors

Many cytokines, mechanical factors and microenvironmental cues influence the expression level of each of the five ligand and five receptor genes. The transcript of each gene can be alternatively spliced to produce mRNAs of different lengths, or even in the case of VEGF₁₆₅b, mRNAs of similar lengths that contain alternative exons. The mRNAs is then translated into proteins; full-length receptor proteins insert into the membrane; truncated receptor proteins (denoted ‘s’ for soluble) are secreted; and ligand proteins dimerize (with cysteine-cysteine linkages) before secretion. The full extent of ligand heterodimerization is not known, but one example is shown here. Each secreted ligand has unique structure that confers on it the ability to bind a subset of the receptors. (Figure 2). Because the two receptor binding sites of the dimer are not located on each monomer, but rather are formed by the interface between monomers, it is difficult to predict the receptor-binding profile of the heterodimers. We use the systems biology notation of Kitano [77, 111]: genes are rectangles, mRNA rhomboids, and proteins are rounded rectangles. Homodimers are stacked shapes. Activation is represented by open circles, translocations by open triangles.

Figure 2. VEGF ligands have different binding profiles to cell surface VEGF receptors on the cell surface and to proteoglycans in the matrix

The isoforms of VEGF that bind to the matrix can be proteolytically released by one of two methods: first, the protease may cleave the matrix itself, leaving the ligand intact and diffusible, possibly with a small proteolytic product of the matrix attached; second, the protease can cleave VEGF itself at one of two sites, to form VEGF₁₁₀ or VEGF₁₁₃ (which are still dimers), and two cleavage products of the heparin-binding domain (e.g., VEGF_111–165 also known as VEGF₅₅) that may or may not remain attached to the matrix. Neuropilin-1 and VEGFR2 can be coupled together by VEGF₁₆₅ but not by VEGF₁₂₁. This may confer a signaling advantage on the longer isoform. BM: basement membrane; ECM, extracellular matrix.

Figure 3. Three heterodimerizing receptors lead to nine distinct signaling pathway-initiating receptor states

Each homodimer transmits a single set of signals due to its phosphorylation profile; heterodimers produce two sets of signals, neither identical to the homodimers.

Figure 4. Tissue-level view of the trafficking of VEGF and its receptors

Each tissue is a multicellular system. It is typically thought that VEGF ligands are secreted by parenchymal cells and bind to endothelial cells, but in fact VEGF expression by endothelial cells and VEGFR expression by parenchymal and stromal cells has also been noted. The relative importance of each of these ligand and receptor pools in maintaining homeostasis, or in generating pathological neovascularization, is not known.

Figure 5. Dimeric structure of VEGF causes homo- and hetero-dimerization of VEGF receptors

A, VEGF ligands are secreted as a dimer (blue and yellow monomers), covalently linked by two cysteine-cysteine linkages from CYS51 on each monomer to CYS60 on the other. The receptor-binding surfaces are present across the junction between the monomers, at each pole. The VEGFR2 binding surface is shown in green and VEGFR1 in red. Structural data from [106]. B, The VEGF dimers bind two VEGF receptors, one at each pole (in this case, VEGFR1). Structural data from [164]. C, The heparin-binding domain of VEGF₁₆₅, also known as VEGF₅₅. The residues involved in Neuropilin binding are labeled in orange. Structural data from [147]. D, Domain b1/b2 of Neuropilin-1 in complex with a peptide containing the Neuropilin-binding site of VEGF₁₆₅ (the last five amino acids, VEGF_161–165). Structural data from [158]. Figures generated using Swiss-PDB Viewer.

Figure 6. Whole organism-view of the transport and communication of VEGF

VEGF is secreted by parenchymal, stromal and perivascular cells in the various tissues of the body. The expression level is different for each tissue and cell type, but all the tissues communicate – to a greater or lesser extent depending on the permeability – through the blood. Microvessels of the brain and retina, for example, are tight and these tissues may be almost closed VEGF microenvironments. Liver and kidney, however, with fenestrated vessels, likely exchange VEGF freely with the blood. Tumors, if present, have leaky vessels also. Plasma VEGF may be cleared, or sequestered by the formed elements in the blood, possibly for release at a later time. The extent to which VEGF returns from the blood to tissues, e.g. from tumors to normal tissue, is not known, but global VEGF communication throughout the body is possible.

Figure 7. Pro- and anti-angiogenic drugs that impact the VEGF-VEGFR system

In recent years, several drugs have been approved by the FDA for the inhibition of VEGF signaling and angiogenesis in diseases of hypervascularization, notably cancer and age-related macular degeneration. Most belong to two main classes: molecules that sequester the VEGF ligands, and tyrosine kinase inhibitors that block downstream VEGFR signaling. However, there are other possibilities, e.g. blocking receptor binding and blocking of co-receptor coupling. Despite many clinical trials, there are still no approved treatments for diseases of hypovascularization (e.g. coronary artery disease, peripheral arterial disease) that impact the VEGF system directly. Exercise is the current standard of care for peripheral arterial disease (PAD) patients, and it does increase both VEGF secretion and VEGF receptor expression, but as yet direct manipulation has not been successful. Figure adapted from [43].

Figure 8. Computational modeling predictions of therapeutic strategies based on the VEGF system

Computational models of VEGF transport in vitro and in vivo have been created for the study of cancer, exercise and peripheral arterial disease. A, Exercise training increases VEGF signaling and VEGF gradients in extensor digitorum longus for rat model of peripheral arterial disease. These figures show the heterogeneity of VEGF concentration (left) and also VEGF gradients. The white box outlines the area expanded in the right hand figure. Note that vessels close together can experience significantly different VEGF gradients (high, *; low, †). Data adapted from [66, 91] B, Exercise increases both VEGF and VEGF receptor expression. A compartmental model of the human vastus lateralis muscle was used to predict the effect of multiple-gene therapy to mimic this physiological response. While VEGF increase alone increases the signaling of both the pro-angiogenic VEGFR2 and anti-angiogenic VEGFR1 receptors (denoted *), combination therapies allow synergistic changes in signaling in both receptors. Data adapted from [97, 98]. C, The microenvironment of a tissue can determine its response to therapeutics. Six therapeutic strategies are compared by their predicted ability to inhibit VEGFR2 signaling of VEGFR2 for 48 hours: Inhibition of Neuropilin expression; peptide blocking of VEGF-Neuropilin binding; antibody blocking VEGFR-Neuropilin coupling; antibody sequestering VEGF; antibody sequestering VEGFR2; and tyrosine kinase inhibitor blocking VEGFR activation. Some drugs are predicted to be effective for tissues many different receptor population profiles; some drugs are effective in only selected environments. Data adapted from [95] and unpublished data.