Inhibitors of Ocular Neovascularization: Title and subTitle BreakPromises and Potential Problems

Molecular medicine offers promise for the prevention of vision loss caused by ocular neovascularization in diabetic retinopathy and exudative age-related macular degeneration (ARMD). During the past decade, significant advances have been made in angiogenesis research, such that the understanding about new vessel formation in disease has increased considerably. This knowledge has led to the development of numerous inhibitors of angiogenesis. Among a host of novel therapeutics for ocular neovascularization, 2 inhibitors of the angiogenic agent vascular endothelial growth factor (VEGF)—pegaptanib sodium and ranibizumab—are poised for imminent clinical application. However, the need for repeated intraocular injection of these agents and the potential for local and systemic adverse effects may pose hurdles for these emerging therapies.

The proliferative retinopathies, principally diabetic retinopathy and exudative ARMD, are leading causes of vision loss worldwide, and their prevalence is projected to increase.^{1 - 2} Central to the pathogenesis of both disorders are increased vascular permeability, leading to retinal edema and subretinal fluid accumulation, and the proliferation of new vessels that are prone to hemorrhage. The established therapy for retinal neovascularization in diabetic retinopathy, laser photocoagulation, may be effective in delaying the progression of the disease but lacks specificity and is associated with retinal destruction, causing impaired visual function.³ Moreover, retinopathy can progress despite the best available treatment.⁴ The management of choroidal neovascularization in ARMD has been boosted by the advent of photodynamic therapy, which involves laser ablation of choroidal new vessels with the aid of a photosensitizer administered systemically via intravenous injection. However, photodynamic therapy is only helpful in a subset of neovascular lesions, and repeated treatments are often required.⁵ While photodynamic therapy is often effective in ablating established pathological vessels, it does not prevent new vessel formation. Accordingly, there is a need for treatments that selectively target the molecular mediators of ocular neovascularization.

Vascular endothelial growth factor, a central mediator of the complex cascade of angiogenesis and a potent permeability factor, is an attractive target for novel therapeutics. Vascular endothelial growth factor is a peptide growth factor, and alternative messenger RNA splicing gives rise to at least 6 isoforms, of which VEGF₁₆₅ is the major pathogenic species.⁶ Vascular endothelial growth factor is the ligand for 2 membrane-bound tyrosine kinase receptors, VEGFR-1 and VEGFR-2. Most of the proangiogenic functions of VEGF are mediated by VEGFR-2. Ligand binding triggers VEGFR-2 dimerization and transphosphorylation with subsequent activation of an intracellular tyrosine kinase domain (Figure). The ensuing intracellular signaling axis results in vascular endothelial cell proliferation, migration, and survival.

Vascular endothelial growth factor has been identified in neovascular membranes in both diabetic retinopathy and ARMD, and intraocular levels of the factor correlate with the severity of neovascularization in diabetic retinopathy.^{7 - 8} In the developing human retina, as well as in animal models of proliferative retinopathy, VEGF expression has been temporally, spatially, and quantitatively associated with new vessel formation.^{9 - 10} Therapeutic antagonism of VEGF in these models results in significant inhibition of both retinal and choroidal neovascularization, as well as a reduction in vascular permeability.^{11 - 13}

Several therapeutic strategies are under development to inhibit the activities of VEGF in the proliferative retinopathies (Figure). Approaches involve the sequestration and neutralization of VEGF or the blockade of VEGFR-2. Examples include a VEGF-neutralizing oligonucleotide aptamer (pegaptanib), a humanized anti-VEGF monoclonal antibody fragment (ranibizumab), a receptor analogue (sFlt-1), and a receptor-immunoglobulin fusion protein.^{12 ,14 - 15} Other strategies are the inhibition of the tyrosine kinase signaling cascade or the degradation of VEGF messenger RNA using small interfering RNAs.^{16 - 17} Of these approaches, pegaptanib and ranibizumab have shown promise in clinical trials in patients with exudative ARMD.

Pegaptanib is an RNA oligonucleotide of 28 bases in length with extremely high affinity, in the picomolar range, for the human VEGF₁₆₅ peptide.¹⁸ It binds to VEGF₁₆₅ by a combination of charge and shape complementarity, sequestering it and preventing VEGF receptor activation. The aptamer has demonstrated significant inhibition of vascular permeability and retinal neovascularization in animal models.¹⁹ In a randomized, double-masked, placebo-controlled multicenter phase 3 clinical trial,¹⁸ 1208 patients with exudative ARMD were randomized to receive either intravitreal pegaptanib (0.3 mg, 1.0 mg, or 3.0 mg) or a sham subconjunctival injection every 6 weeks for 48 weeks prior to rerandomization at 54 weeks. At 54 weeks, each patient group demonstrated a progressive loss of vision, but the extent of this loss differed among groups. Treatment with pegaptanib (0.3-mg dose) was associated with a 15% benefit over sham in terms of the primary efficacy end point—the loss of less than 15 letters of visual acuity, as measured with the Early Treatment of Diabetic Retinopathy Study (ETDRS) chart, at 2 m.¹⁸ Paradoxically, the high-dose pegaptanib treatment group (3 mg) was not significantly different from the control group for the efficacy end point. While most trial participants continued to lose vision, fewer patients in the 0.3-mg treatment group experienced severe vision loss (≥30 letters ETDRS) than in the control group (10% vs 22%, P<.001).¹⁸ Pegaptanib was approved by the US Food and Drug Administration for the treatment of exudative ARMD in December 2004.²⁰ At a price of $995 per injection,^{21 - 22} the annual drug cost per patient is approximately $8600, assuming a 6-week dosing schedule.

In contrast to pegaptanib, ranibizumab is a recombinant humanized monoclonal antibody fragment with specificity for all isoforms of human VEGF.²³ Ranibizumab demonstrates high affinity for human VEGF and exerts its neutralizing effect by inhibiting the VEGF-receptor interaction. Unlike the larger whole antibody, ranibizumab can penetrate the internal limiting membrane and reach the subretinal space following intravitreal injection in rhesus monkeys.^{12 ,24} In nonhuman primate studies of laser-induced choroidal neovascularization, intravitreal injection reduced the incidence of new vessel formation as well as leakage from established vessels.¹²

In a phase 2 trial of the antibody fragment in human ARMD, patients were randomized to receive either usual care or intravitreal injections of 300 μg or 500 μg of ranibizumab every 28 days for 4 doses.²⁵ Ranibizumab treatment was associated with few adverse events aside from reversible ocular inflammation. Visual acuity, as measured with the ETDRS chart, improved 8.5 (3.3) and 12.8 (3.4) (mean [SD]) letters from baseline in the group treated with the 300-μg dose at days 98 and 210, respectively. Patients in the usual care group demonstrated a decrease of 3.0 (5.6) letters at day 98, but improved to a gain of 7.3 (6.6) letters at day 210 following crossover to ranibizumab treatment. Ranibizumab is currently under study in phase 3 clinical trials.

A major limitation of both treatments is the need for repeated intraocular injection. Intravitreal injection is invasive, with the potential for blinding sequelae such as endophthalmitis and retinal detachment. In clinical trials, administration of pegaptanib was associated with an overall risk of endophthalmitis of 0.16% per dose, or 1.3% per patient per year.¹⁸ The incidence of endophthalmitis was significantly reduced following the introduction of more stringent infection control measures at the time of administration. The rates of retinal detachment and traumatic cataract per injection were 0.08% and 0.07%, respectively.¹⁸ While the incidence of serious complications of intraocular injection is low, cumulative risk exposure may be significant for patients requiring serial treatments over many years. Although patients with advanced disease may tolerate repeated injections, given the hope of improved visual function, it is unlikely that individuals with early proliferative diabetic retinopathy or ARMD would be as likely to do so. Accordingly, attempts are being made to formulate alternative delivery vehicles for these drugs.²⁶

In health, the eye is relatively sequestered from the systemic circulation by the tight blood-ocular barrier; however, breakdown of this barrier is common in neovascular eye disease.^{27 - 28} Thus, while intraocular injection of an anti-VEGF therapeutic may provide relative selectivity for VEGF in the eye, systemic exposure is inevitable. This notion is supported by pharmacokinetic findings in trials of pegaptanib and ranibizumab. In rhesus monkeys, peak plasma levels of approximately 0.4 μg/mL were achieved following bilateral intravitreal injections of 0.5 mg of pegaptanib, and mean levels were in excess of 3 ng/mL 28 days later.²⁹ In humans, mean plasma levels of the aptamer were approximately 80 ng/mL following a single intravitreal injection of 3 mg of pegaptanib (10 times the recommended dose for intraocular injection) and its plasma half-life was 10 (4) (mean [SD]) days.³⁰ Similarly, in cynomolgus monkeys, peak serum ranibizumab levels were 150 ng/mL following bilateral intravitreal injections of the drug (500 μg) and the serum half-life was 3.5 days.³¹ While it is difficult to extrapolate these values directly to humans, it is likely that serum levels of ranibizumab will be lower owing to different kinetics of release from the ocular compartment and the larger volume of distribution. To put these figures into context, plasma VEGF levels in the healthy human adult are typically less than 100 pg/mL, 2 orders of magnitude lower than the observed mean drug levels.³² While VEGF concentrations at sites of active angiogenesis are substantially higher, the effects of chronic low-level VEGF antagonism are not well characterized.

Although VEGF antagonists appear to be well tolerated in the short term, a growing body of evidence, from animal and in vitro experiments, hints at the potential for serious systemic adverse effects. In addition to playing a role in pathological neovascularization, VEGF is required for normal wound healing, bone growth, cyclic endometrial development, and placental vascularization. Preclinical studies of an antivascular endothelial growth factor antibody (rhuMAbVEGF)—a recombinant humanized whole antibody closely related to ranibizumab—in young adult cynomolgus monkeys demonstrated physeal dysplasia following biweekly intravenous doses of the antibody, as low as 2 mg/kg.³³ At higher doses (10 mg/kg), uterine and ovarian weights were reduced and corpora lutea were absent, indicating impaired reproductive function. Partial restoration of these changes was noted 4 weeks after the cessation of treatment. Vascular endothelial growth factor plays other vital roles, such as the formation of collateral vessels critical to the viability of ischemic limbs and myocardium.³⁴

Because individuals with diabetic retinopathy and ARMD may be at increased risk of cardiovascular and peripheral vascular disease, the implications of long-term systemic inhibition of VEGF could be profound.^{35 - 37} These concerns are compounded by the recent discovery of a doubling in the incidence of serious thromboembolic events in patients with colon cancer receiving intravenous anti-VEGF monoclonal antibodies in combination with 5-fluorouracil, relative to those receiving standard chemotherapy.³⁸ It is estimated that the risk of such events in patients treated with this agent may be as high as 5%. While this population is unique, both in terms of the level of systemic exposure to the anti-VEGF therapeutic and in terms of comorbid illness, these findings warrant concern for patients receiving chronic anti-VEGF therapy for non–life-threatening ocular disease.

The 54-week safety data from the pegaptanib trial are reassuring: cardiovascular events and all-cause mortality were comparable for the pegaptanib-treated and sham-injected groups. Because individuals deemed to be at high risk of cardiac and cerebrovascular events were excluded from trial participation, it remains to be seen whether similar results can be attained in a real-world population. Trials of ranibizumab are still under way. In view of the theoretical potential for adverse cardiovascular events, the US Food and Drug Administration has endorsed the need for postmarketing surveillance to determine the long-term safety of pegaptanib. While concerns have been voiced about the rigor of postmarketing surveillance, triggered largely by events surrounding the withdrawal of the cyclooxygenase 2 inhibitor rofecoxib,³⁹ there is a clear need for greater vigilance and timely reporting of adverse safety outcomes of novel therapeutics.

It appears that VEGF has important roles in neuronal function.⁴⁰ Vascular endothelial growth factor receptors are widely expressed in the brain and spinal cord, and mice deficient in VEGF have a phenotype analogous to the neurodegenerative disorder amyotrophic lateral sclerosis. In addition, the motor neuron degeneration observed in a well-established mouse model of amyotrophic lateral sclerosis (superoxide dismutase with Gly93Al1 substitution [SOD1^G93A] mice) can be significantly delayed by the induction of VEGF expression.⁴¹ Vascular endothelial growth factor has recently been implicated in the proliferation of neuronal stem cells in the murine hippocampus and is thought to play an important role in memory and learning.⁴² The factor serves a protective role in acute neuronal ischemia and stimulates the proliferation of a wide range of neuronal and glial cell types in vitro and in vivo.⁴⁰ Vascular endothelial growth factor expression has been detected in all classes of neurons and glial cells in the disease-free human retina.⁴³ It has been postulated that the basal expression of VEGF by the neural retina, which may approach 15 to 20 pg/mg of protein, may serve a role in maintaining retinal vascular homeostasis; however, the potential for an autocrine or paracrine neuroprotective role remains.⁴⁴ In vitro experiments suggest that VEGF plays a role in photoreceptor differentiation and may contribute to photoreceptor survival.⁴⁵ Granted that photoreceptor degeneration is a key pathological event in ARMD and neuronal ischemia is central to diabetic retinopathy, it remains to be seen whether VEGF antagonists will accelerate these processes in the long-term.

Drugs targeting VEGF offer promise as sight-saving therapies for individuals with advanced diabetic retinopathy and exudative ARMD, for whom other therapeutic interventions are limited. However, enthusiasm for these agents must be tempered by recognition of the potential for significant local and systemic adverse sequelae. While an intraocular excess of VEGF can contribute to disease, the factor also serves numerous essential functions in extraocular tissues. It follows that the ability to selectively localize an anti-VEGF therapeutic to the intraocular environment may be critical to its clinical success. Moreover, as VEGF may serve roles in the maintenance of the neural retina and functional retinal vessels, greater therapeutic precision could be afforded by targeting the downstream effectors of VEGF-signaling that are specific to angiogenesis. Further advances in the understanding of the molecular bases of pathological angiogenesis will lead to the design of therapeutic agents with the potential to ease the burden of neovascular eye disease. Such developments should go hand in hand with therapies targeting the broader spectrum of pathological events in these disorders.