Laser photocoagulation was the preferred treatment for the complications of diabetic retinopathy – diabetic macular edema (DME) and proliferative diabetic retinopathy (PDR) – for several decades with level I evidence emerging from the Early Treatment of Diabetic Retinopathy Study [45] and the Diabetic Retinopathy Study (DRS) [166–168]. Laser reduced the risk of severe vision loss (<5/200) in eyes with PDR and moderate vision loss (>15 letters) in eyes with DME by one half, but significant improvements in visual acuity (>15 letters) were unusual.

In an attempt to improve the treatment of macular edema, physicians began injecting triamcinolone acetate into the vitreous of the eyes with macular edema due to diabetic retinopathy and retinal vein occlusions [77, 91, 114] early in the twenty-first century. Corticosteroids improved macular edema dramatically but the high incidences of posterior subcapsular cataracts and elevated intraocular pressures limited their widespread adoption. Surgeons frequently performed pars plana vitrectomy for advanced complications of PDR (vitreous hemorrhage, traction retinal detachment, traction/rhegmatogenous retinal detachment) [169] and even for DME that was accompanied by vitreomacular traction [106]. Surgical outcomes for PDR were superior to the natural history of the disease [169], but vitrectomy only benefited a small number of patients with vision loss due to DME [60].

Effective pharmacotherapy for DME eventually came from a surprising source – tumor biology [24, 56]. The discovery of vascular endothelial growth factor (VEGF) and the early understanding of its role in neovascular age-related macular degeneration, macular edema due to retinal vein occlusions, DME, and PDR [5] set in motion a pharmacologic revolution in ophthalmology. The development of potent ocular pharmaceuticals created a paradigm shift in the treatment of center-involving diabetic macular edema and more recently in the treatment of PDR (Fig. 4.1).

Many manuscripts describing the treatment of DME with anti-VEGF drugs have been published, but most are flawed with small numbers of patients, short durations of treatment and follow-up, lack of control arms and randomization, and inadequate masking. Fortunately, several high-quality, randomized, controlled, double-masked, multicenter trials have been published, and these constitute the focus of this chapter.

Pathologic ocular neovascularization has long been recognized as the sequela to severe vascular injury. Michaelson (1948) postulated the existence of a soluble factor X that he believed was responsible for new blood vessel growth in conditions such as neovascular glaucoma (Fig. 4.2) [119]. Folkman (1971) noted that both primary and metastatic solid tumors in children were accompanied by leashes of blood vessels through which the tumor obtained oxygen and nutrients [61]. He postulated that solid tumors synthesized a pro-angiogenic substance that promoted blood vessel growth necessary to sustain tumors.

Vascular permeability factor (VPF) was discovered in 1983 [155], and Ferrara and Connolly independently discovered vascular endothelial growth factor (VEGF) [25, 56] 6 years later. Sequencing analysis showed that both investigators had identified the same molecule, which turned out to be the previously discovered VPF. This set in motion a multi-specialty, international research initiative that characterized VEGF, its receptors, and its downstream pathways.

VEGF is actually a group of closely related molecules that segregate into seven families: VEGF-A, VEGF-B, VEGF-C, VEGF-D, the orf virus encoded VEGF-E, VEGF-F, and placental growth factor (PlGF) (Table 4.1). Most of these families consist of several isoforms with similar binding sequences.

VEGF-A isoforms are the most important contributors to ocular angiogenesis. The 14 kb VEGF-A gene resides on chromosome 6p21.3 [174]. Splicing of eight exons and seven introns creates at least six major (VEGF₁₂₁, VEGF₁₄₅, VEGF₁₆₅, VEGF₁₈₃, VEGF₁₈₉, and VEGF₂₀₆) and several minor isoforms within the VEGF-A family [84, 170]. In addition, the split product VEGF₁₁₀ is produced when tissue-sequestered isoforms with 165 or more amino acids are cleaved by plasmin and matrix metalloproteinase-3. The shorter non-heparin-binding isoforms are biologically active but with only 10–20% the biological activity of VEGF_165. VEGF naturally exists as a homodimer, so it presents two receptor-binding sites to its environment [95, 104].

Each of the VEGF families controls a different type of angiogenesis (Table 4.1). VEGF-B stimulates coronary blood vessel growth, and VEGF-C and VEGF-D promote lymphangiogenesis by activating VEGF receptor (VEGFR) 3 [94]. The exact role of PlGF in angiogenesis is unclear, but its binding to VEGFR1 may force VEGF-A to preferentially bind to VEGFR2 [137], the receptor primarily responsible for ocular angiogenesis. PlGF may play an active role in the development of DR as it has been found in the retinas of diabetic rats [89, 99].

The common receptor-binding sequence of all VEGF-A isoforms resides in the amino acid region 82–93, whereas the heparin-binding sequence is found in the 111–165 amino acid region of longer isoforms. Molecular charge determines isoform diffusibility with the acidic VEGF₁₂₁ isoform being freely diffusible whereas the basic VEGF₂₀₆ isoform is nearly totally bound to interstitial matrix. Fifty to seventy percent of the electrically neutral VEGF₁₆₅ is bound to the interstitial matrix, and heparin-binding sequences enable isoforms of 165 amino acids and longer to bind to interstitial matrix heparin moieties. Sequestration of these longer molecules provides a readily available reservoir of VEGF when injury or inflammation induces a protease-mediated breakdown of the interstitium. The heparin-binding domain also enables VEGF₁₆₅ to bind to the transmembrane neuropilin-1 co-receptor.

Diffusible VEGF binds to the extracellular binding domains of three transmembrane receptors: VEGFR1 (flt-1), VEGFR2 (flk-1), and VEGFR3 (flt-4). Each of these receptors is composed of three major parts: the seven extracellular immunoglobulin-like VEGF-binding domains, the single transmembrane region, and the intracellular tyrosine kinase moieties [28, 165]. VEGFR1 binds VEGF-A, VEGF-B, and placental growth factor; VEGFR2 binds VEGF-A and VEGF-C; and VEGFR3 binds VEGF-C, VEGF-D, VEGF-E, and VEGF-F. VEGFR-1 binds VEGF-A with a much higher affinity than does VEGFR2. Binding of VEGF by VEGFR1 and VEGFR2 occurs at the second and third extracellular domains. Alternative splicing of the RNA transcript for the binding domains of the VEGF receptors creates soluble VEGFR2 (sVEGFR2) and VEGFR1 (sVEGFR1), which inhibit vascular development by promoting vascular maturation and maintaining corneal avascularity [137].

The role of VEGFR1 in angiogenesis is unclear because its downstream activity varies with both the age of the individual and the cell type. Activation of VEGFR1 during embryological development stimulates angiogenesis, but during adulthood it dampens VEGF activity by acting as a decoy receptor.

VEGF-A activity occurs when either naturally occurring VEGF₁₂₁ or VEGF_165, or the fibrin-split product VEGF₁₁₀, binds to VEGFR2 (Fig. 4.3). Each of two VEGFR2 receptors binds to the VEGF dimer causing conformational changes in the 3-dimensional structures of the receptor molecules and activating their intracytoplasmic tyrosine kinase moieties (Fig. 4.2). This phosphorylates several intracellular enzymes that activate downstream pathways, which lead to VEGF’s potent effects (Fig. 4.3) [21, 48, 79, 164].

Hypoxic environments created by processes such as tumor growth and occlusive vasculopathies upregulate VEGF [42]. Tissue hypoxia reduces hydroxylation of the cell’s “oxygen sensor” HIF-1α, preventing it from binding to the von-Hippel Lindau factor, and undergoing ubiquination and subsequent destruction within proteosomes. Stable HIF-1α dimerizes with HIF-1β to form a complex that enters the cell nucleus where it activates the promoter region of the VEGF gene and also upregulates both VEGFR1 and VEGFR2 [71, 172] synthesis (Fig. 4.4).

VEGF is upregulated during starvation and by several metabolic factors, chemokines, and cytokines. A variety of metabolic regulators including ROS (reactive oxygen species such as superoxide (O₂ ⁻)) increase the synthesis of both VEGF and its receptors [173]. Several growth factors including epidermal growth factor, tumor necrosis factor (TNF)-α, transforming growth factor (TGF)-β, basic fibroblast growth factor, interleukin-6, insulin-like growth factor-1, keratinocyte growth factor, and platelet-derived growth factor upregulate VEGF and may work in conjunction with tissue hypoxia [128]. Oncologic mutations that stimulate RAS, a family of ubiquitous intracellular GTP-ases, upregulate VEGF [78]. Mechanical forces such as sheer stress and stretch that occur with vitreomacular traction promote VEGF production [46].

VEGF has been detected in both the neuroretina and pigment epithelium [97], and it can be produced by several cell types including pericytes, glia, vascular endothelial cells, invading leukocytes, and retinal pigment epithelium [2, 134]. Some VEGF isoforms modulate the synthesis of others by affecting HIF-1α stability [183], and VEGF inhibitors probably function via a similar mechanism.

PlGF and VEGF-B binding to VEGFR1 activates signaling pathways implicated in monocyte chemotaxis [36] and inflammatory-related angiogenesis. Though PlGF appears to be a critical mediator of inflammatory angiogenesis, it does not appear to directly influence embryologic or physiologic adult angiogenesis [59]. Tyrosine kinase activity of VEGFR1 is relatively weak but several downstream signaling molecules including phospholipase C, the p85 subunit of PI-3 kinase, growth factor receptor bound protein, SHP-2, and NcK are associated with VEGFR1 phosphorylation sites. These pathways result in eNOS activation and calcium influx in vascular endothelial cells.

VEGFR2 activation is responsible for the full spectrum of VEGF-A-mediated responses within human endothelial cells (survival, proliferation, migration, and formation of a vascular tube) by stimulating Raf-MEP-ERK and MAPK pathways [164]. These pathways phosphorylate several endothelial cell proteins including phospholipase-3, PI-3 kinase, RAS GTP-ase activating protein and the Src family [48, 79]. They stimulate cell migration via FAK and paxillin, PI-3 kinase, and MAPK and prolong survival via PI-3 kinase and Akt/PKB [21].

VEGF-A isoforms also bind to the heparin-binding domains of neuropilin-1 and neuropilin-2, transmembrane proteins with small, non-catalytic tails that present VEGF₁₆₅ favorably to VEGFR2. The neuropilin-modulated binding of VEGF₁₆₅ increases its biological activity by tenfold over that of VEGF₁₂₁, which cannot interact with neuropilin [63].

VEGF-A upregulates gene expression that promotes the division, migration, and survival of vascular endothelial cells, regulates vascular dilation and permeability, and stabilizes immature blood vessels (Fig. 4.5) [55]. VEGF-A enhances vascular permeability 50,000 times more than does histamine, increases nitric oxide production, and stimulates the synthesis of several molecules that are important to angiogenesis, such as matrix metalloproteinases, which break down the interstitial matrix to allow advancement of proliferating endothelial cells [175]. VEGF₁₆₅ induces endothelial cell growth by activating the Raf-MEK-Erk pathway [52]. VEGF degrades the blood retinal barrier (BRB) in venules but not capillaries [126] by causing a calcium-mediated increase in hydraulic conductivity [14], increase in endothelial cell fenestrations [147], and opening of intercellular tight junctions by phosphorylating several junctional proteins [67]. Increased permeability allows plasma to extravasate and form an interstitial fibrin gel into which endothelial cells can proliferate and migrate [44]. VEGF-A increases hexose transport to meet the metabolic demands of neovascularization [138]. VEGF-A produces a dose-dependent vasodilation in vitro via nitric oxide synthesis [102] and retinal vasodilation in animal models and human retinal vein occlusions [51, 171].

VEGF-A promotes the survival of rat endothelial cells in vitro and in vivo, while VEGF blockade causes extensive apoptotic changes in the vasculature of neonatal but not adult mice [73]. Newly formed but not mature vessels are VEGF dependent [16, 181], and a complete vestment of pericytes is key to establishing VEGF independence [16]. Nonetheless, VEGF-A plays a role in maintaining the choriocapillaris in adults [53]. VEGF-A administration to hypoxic vascular endothelial cells upregulates phosphatidylinositol (PI)-3 kinase-Akt pathways and the expression of antiapoptotic proteins Bl-2 and A1 and prevents apoptosis and regression of the vasculature [6, 72]. The effects of VEGF on the retinal pigment epithelium remain controversial as some investigators have noted breakdown of the outer blood-retinal barrier whereas others have not [1, 74, 85, 122].

VEGF-A promotes monocyte chemotaxis and colony formation by granulocyte-macrophage progenitor cells. VEGF-A upregulates the synthesis of intercellular adhesion molecule (ICAM)-1 and vascular cellular adhesion molecule (VCAM)-1, which allows natural killer (NT) cells to adhere to vascular endothelial cells.

Exogenous administration of VEGF into primate eyes induces changes that resemble diabetic retinopathy [171]. VEGF has been successfully blocked in animals with soluble decoy receptors [93], tyrosine kinase inhibitors, and antibodies directed against the VEGF dimer or its receptors [101]. In animal models, blocking VEGF pathways appears to be more effective than blocking other angiogenic pathways [52]. Drugs that block VEGF work primarily by reestablishing the blood-retinal barrier, “normalizing” the circulation, and preventing further new blood vessel growth.

Healthy human eyes have vitreous VEGF concentrations of 8.8 +/− 9.9 ng/ml [5], whereas patients with DME and other retinal vascular conditions have elevated intraocular VEGF concentrations [5, 80]. Following pars plana vitrectomy, the intravitreal half-life of VEGF in rabbit eyes decreases by 75%.

Currently available drugs that inhibit the actions of vascular endothelial growth factor are listed in Table 4.2. Included are the major binding and pharmacokinetic properties of each drug.