Ocular angiogenesis can be physiological or pathological, with physiological ocular angiogenesis occurring primarily during embryonic development (reviewed by Gariano ). Ocular angiogenesis in adults is usually pathological and is a major cause of vision loss and blindness due to conditions such as choroidal neovascularization (CNV) related to age-related macular degeneration (AMD), diabetic retinopathy, neovascular glaucoma, corneal neovascularization, and retinopathy of prematurity. Each of these is discussed in at length elsewhere in this volume.
Research into mechanisms of physiological and pathological angiogenesis has led to a deeper understanding of molecular and cellular mechanisms involved in angiogenesis, with a principal focus of research efforts being the identification of molecules involved in promotion and inhibition of ocular neovascular disease. This chapter will focus on the molecules for which the role in ocular angiogenesis is best characterized, especially those that have led to the development of new drugs, and will present an overview of existing and developing therapies derived from this work.
The pathogenesis of ocular neovascularization involves a complex interaction between proangiogenic and antiangiogenic factors and molecules. There is also accumulating evidence supporting the inflammatory nature of both AMD and diabetic retinopathy. The depletion of monocytes inhibited pathologic (but not physiologic) retinal neovascularization in experimental models, strongly supporting a role for inflammation in ocular neovascular disease ( Figure 70.1 ). Certain haplotypes of factor H, a regulatory component of the complement cascade, are associated with an increased risk of developing AMD (reviewed by Donoso et al ). In addition, complement factors C3a and C5a and C5b–9 (the membrane attack complex, or MAC) were identified in drusen of patients with AMD. Extensive deposits of C3d and C5b–9 have also been identified in the choriocapillaris of human eyes with clinically evident diabetic retinopathy, but not in the vast majority of control eyes, and elevated levels of assorted complement factors have been found in the vitreous of patients undergoing surgery for proliferative diabetic retinopathy. Since complement is involved in opsonization, chemotaxis, and activation of leukocytes (reviewed by Rus et al ), the presence of complement in AMD and diabetic retinopathy lesions together with the co-localization of immune cells is evidence of active inflammation.
The final stage of angiogenesis involves stabilization of nascent vasculature by a process known as maturation, which involves selective pruning (remodeling) and recruitment of mural cells (pericytes and smooth-muscle cells). Leukocytes are believed to contribute actively to vascular pruning ( Figure 70.2 ) through Fas/FasL-mediated endothelial cell apoptosis. Maturation is tightly regulated by levels of vascular endothelial growth factor (VEGF), with new vessels becoming refractory to VEGF withdrawal over time. In other studies, maturation corresponded with the expression of angiopoietin 1 (Ang1) and platelet-derived growth factor-B (PDGF-B) and prevention of mural cell binding to the endothelium by PDGF-B blockade resulted in disorganized retinal vasculature; normalization was restored by the administration of Ang1.
The complexity of the pathogenesis of ocular neovascularization suggests numerous potential targets for intervention in the treatment of ocular neovascular diseases; yet this same complexity suggests that different interventions may be needed, either alone or in combination, to provide optimal benefits to all patients.
Endogenous promoters of angiogenesis
There are many molecules known to promote angiogenesis for which there is evidence supporting a role in the etiology of ocular neovascular disease ( Box 70.1 ).
Fibroblast growth factor-2
Platelet-derived growth factor-B
Tumor necrosis factor-α
Vascular endothelial growth factor
Vascular endothelial growth factor
A major research effort has identified VEGF as a master regulator in both physiologic and pathologic angiogenesis and a major contribution to ocular neovascular diseases. Elevated levels of VEGF have been shown to accompany the development of neovascularization in conditions such as retinal vein occlusion, neovascular glaucoma, retinopathy of prematurity, and proliferative diabetic retinopathy. VEGF has also been found to be overexpressed in the retinal pigment epithelium (RPE) in surgically excised CNV membranes of patients with AMD.
A number of approaches have demonstrated that VEGF is both necessary and sufficient for the development of ocular neovascularization. Experimentally induced ocular elevations of VEGF achieved by various means have led to pathological ocular neovascularization while inactivation of VEGF resulted in inhibition of ocular neovascularization. VEGF is produced by many cell types in the retina, including neurons, glia, and RPE cells. Hypoxia strongly induces the expression of VEGF through stabilization of hypoxia-inducible factor-1 alpha, a transcriptional regulator.
Alternative splicing of the VEGF gene results in six major isoforms. Evidence from rodent models suggests that VEGF 164/165 (VEGF 164 is the rodent equivalent of human VEGF 165 ) was preferentially upregulated in ischemia-induced pathological neovascularization and was significantly more potent at inducing inflammation.
Platelet-derived growth factor-B
The PDGF group of dimeric proteins is composed of combinations of four different polypeptide chains (PDGF A–D) with a cellular distribution that includes fibroblasts, vascular smooth-muscle cells, endothelial cells, RPE cells, and macrophages. PDGFs interact with two related tyrosine kinases, PDGF receptor (PDGFR)-α and PDGFR-β, leading to receptor dimerization and autophosphorylation (reviewed by Heldin and Westermark ).
During angiogenesis, the homodimeric PDGF-B has been found to play a particularly important role in the recruitment of PDGFR-β-expressing mural cells (pericytes and vascular smooth-muscle cells) to the developing vasculature ( Figure 70.3 ). Jo et al employed three different murine models to study the contributions of signaling induced by VEGF and PDGF-B in ocular neovascularization. Physiologic development of neonatal retinal vasculature was significantly inhibited by blockade of signaling induced by PDGF-B but not by VEGF164; simultaneous blockade of both factors led to additional reductions. In contrast, blocking PDGF-B had little effect on developing or established neovascularization in a model of CNV, whereas VEGF inhibition significantly reduced the growth of new vessels; the most potent inhibition was again observed when both factors were inhibited. Finally, PDGF-B blockade of established corneal neovascularization between days 10 and 20 postinjury led to detachment of mural cells from corneal neovessels ( Figure 70.4A ) while PDGR-B blockade immediately following corneal injury did not significantly reduce neovascularization; in contrast, blocking VEGF led to a significant inhibition in the growth of new vessels. In established vessels, the combination led to the regression of pathological vessels. When both factors were blocked there was a significantly greater reduction than with VEGF inhibition alone ( Figure 70.4B ). In these models, inhibition of PDGF-B signaling led to pericyte depletion in retinal and corneal vessels, but not in quiescent adult limbal vessels, suggesting that therapies which block both VEGF and PDGF-B are more likely to achieve regression of both established and developing ocular neovascular lesions.
Fibroblast growth factor 2
Experimental models have not clearly defined the role of fibroblast growth factor 2 (FGF2; also known as basic FGF) in ocular neovascular disease. While studies have demonstrated the presence of FGF2 in surgically removed CNV membranes, other studies suggest that FGF overexpression is insufficient in itself to provoke CNV in the absence of an additional stimulus such as cell injury.
Tumor necrosis factor-α
The role of tumor necrosis factor-α (TNF-α) in ocular neovascularization is not fully understood; it may contribute to angiogenesis indirectly by promoting leukostasis in neovascular tissues and by inducing expression factors such as VEGF, Ang1, and Ang2. Although intravitreal administration of infliximab, an anti-TNF-α monoclonal antibody, reduced the formation of laser-induced CNV in a rat model, findings with ischemia-induced retinopathy models in knockout mice were inconclusive. In mice lacking TNF-α expression there was no reduction in neovascularization with respect to wild-type mice, while mice lacking TNF-receptor p55 had a reduction in ischemia-induced neovascularization. These apparent contradictions in results may reflect differences in experimental methodology.
Angiopoietins 1 and 2
Ang1 and Ang2 are factors that act as ligands for Tie2, a receptor tyrosine kinase. Ang1 binds to and induces the phosphorylation of Tie2 whereas Ang2 usually behaves as an antagonist of Ang1 and Tie2 (reviewed by Eklund and Olsen ). Both Ang1 and Ang2 have been found to co-localize with VEGF in neovascular proliferative membranes of patient eyes.
Ang1 is produced by vascular smooth-muscle cells. Experimentally induced elevations in Ang1 in rodents caused reductions in retinal vascular leukocyte adhesion, endothelial cell damage, and blood–retinal barrier breakdown in a diabetic retinopathy model, suppressed the development of CNV following laser wounding, and inhibited VEGF-mediated breakdown of the blood–retinal barrier in response to ischemia ; however, Ang1 had no effect on established neovascularization.
Ang 2, which is produced by endothelial cells and is prominently expressed at sites of vascular remodeling, is believed to serve primarily as an antagonist to Ang1/Tie2 during angiogenesis. Studies suggest that Ang2 functions as a promoter of angiogenesis, mainly in combination with VEGF, and induces vascular regression when VEGF levels are low. Thus, the bulk of evidence suggests that Ang1 acts largely to inhibit the development of neovascularization whereas Ang2 acts to destabilize the vascular endothelium, making it more responsive to factors such as VEGF ( Figure 70.5 ).