Anti-angiogenic Therapy for Retinal Disease

Abstract

Recent breakthroughs in our understanding of the molecular pathophysiology of retinal vascular disease have allowed us to specifically target pathological angiogenesis while minimizing damage to the neurosensory retina. This is perhaps best exemplified by the development of therapies targeting the potent angiogenic growth factor and vascular permeability mediator, vascular endothelial growth factor (VEGF). Anti-VEGF therapies, initially introduced for the treatment of choroidal neovascularization in patients with age-related macular degeneration, have also had a dramatic impact on the management of retinal vascular disease and are currently an indispensable component for the treatment of macular edema in patients with diabetic eye disease and retinal vein occlusions. Emerging evidence supports expanding the use of therapies targeting VEGF for the treatment of retinal neovascularization in patients with diabetic retinopathy and retinopathy of prematurity. However, VEGF is among a growing list of angiogenic and vascular hyperpermeability factors that promote retinal vascular disease. Many of these mediators are expressed in response to stabilization of a single family of transcription factors, the hypoxia-inducible factors (HIFs), that regulate the expression of these angiogenic stimulators. Here we review the basic principles driving pathological angiogenesis and discuss the current state of retinal anti-angiogenic pharmacotherapy and as well as future directions.

Introduction

The development of the vascular system (i.e., vasculogenesis) occurs early during development and is an important foundational stage in organogenesis. In adults, the growth of new blood vessels (i.e., physiologic angiogenesis) is limited to wound healing and ovarian and cervical cycling (DiPietro 2016). Conversely, pathological angiogenesis is an essential component in disease, most notably in cancer, rheumatologic conditions including psoriasis, and ischemic disease (e.g., myocardial and cerebral infarction). The impact of pathological angiogenesis in disease is particularly evident for diseases of the retinal, where it can lead to dramatic and permanent loss of vision. Pathological angiogenesis occurs a list of settings including diabetic retinopathy, retinal vein occlusions, sickle cell retinopathy, hyperviscosity syndromes, pathologic myopia, age-related macular degeneration, retinal arterial occlusion, radiation retinopathy, ocular ischemic syndrome, retinopathy of prematurity, familial exudative vitreoretinopathy (FEVR), retinal vasculitis, uveitic and inflammatory conditions, tumors, retinal vascular tumors including hemangiomas, retinal degeneration, post-traumatic changes, macular dystrophies, angioid streaks, choroidal rupture, Eales disease, sarcoidosis, chronic retinal detachment, carotid-cavernous fistula, Coats’ disease, retinal artery macroaneurysms, and incontinentia pigmenti. Due to its easy accessibility and visualization, the retina is also an ideal model to study both vasculogenesis and angiogenesis: the retina is thin and transparent; retinal vascular development is stereotyped both spatially and temporally; and vascularization of the retina occurs post-natally in rodents, the most commonly used vertebrate animals to study ocular disease.

Retinal laser therapy was developed over 50 years ago with investigations by Kapany and colleagues in 1963 using the ruby laser (Kapany et al. 1963) and has had a profound impact on the treatment of numerous retinal neovascular disorders. Seminal early studies established laser as the first treatment for numerous eye conditions. The Diabetic Retinopathy Study (DRS) established panretinal laser photocoagulation as the first effective treatment for high risk proliferative diabetic retinopathy (The Diabetic Retinopathy Study Research Group 1981). And the Early Treatment Diabetic Retinopathy Study (ETDRS) study established laser as the first effective treatment for clinically significant diabetic macular edema (Early Treatment Diabetic Retinopathy Study Research Group 1991). The branch retinal vein occlusion study (BVOS) later demonstrated the efficacy of focal laser for the treatment of macular edema associated with branch vein occlusions (The Branch Vein Occlusion Study Group 1984). However, recent scientific discoveries have led to the development of new pharmacotherapies which have demonstrated significantly improved visual acuity outcomes with decreased side-effect profile. Indeed, the recent introduction of targeted pharmacologic anti-angiogenic therapies has transformed the care of patients with ocular neovascular diseases. This chapter will briefly describe the normal development of the retinal vasculature, the pathologic molecular events that lead to retinal and choroidal neovascularization, and currently available and possible future targets for anti-angiogenic therapy.

Embryology and Anatomy of the Retina

The development of the human eye has been described and refined over the past 90 years (Mann 1928). The outer layer of the optic cup evolves as a monolayer into the retinal pigment epithelium (RPE). The neurosensory retina develops from the inner layer of the neuroectodermal cells of the optic cup that begin to migrate in the second month of fetal development (Tripathi et al 1997). Vascular channels from the internal carotid artery develop in the mesenchyme around the optic vesicle late in the fourth week of gestation and the ophthalmic artery develops by the 6th week of gestation.

The most superficial retinal vascular layer first begins around 14 weeks gestation age as spindle-shaped undifferentiated mesenchymal cells arise from the hyaloid artery at the optic disc and develop into radial vessel extensions from the optic nerve head. Lumina develop behind the advancing edge of the mesenchymal cells. The veseels spread peripherally to reach the ora serrata just before term in centrifugal waves following neuronal differentiation with endothelial cells differentiating first followed by zonulae occludens and gap junctions (Penfold et al 1990). Deeper vascular layers arise sequentially by sprouting from the initial plexus (Michaelson 1948). The hyaloid system and tunica vasculosa lentis atrophy in the third trimester. The adult retinal vascular pattern occurs by remodeling through pruning and maturation soon after formation by 3 months after birth. The foveal pit develops in the first four years of life.

The normal human circulatory system is made of several cell types. Endothelial cells make up the inner wall of the blood vessel. Inter-endothelial cell adhesions and tight junctions comprise important blood-tissue barriers. Smooth muscle cells in arterioles and pericytes in capillaries surround the endothelial cells and provide vascular tone and regulation. Glial cells such as astrocytes and retinal Müller cells are closely apposed to vessels in neural tissue. Dendritiform cells such as microglia are antigen presenting cells associated with vessels in certain tissues.

The human retina has 4 interconnected planar vascular networks in neuronal layers in the nerve fiber layer, ganglion cell layer, and inner nuclear layer. The retinal vessels form in alignment with the radial orientation of ganglion cell axons (Figure 1). Importantly, the outer retina (the outer plexiform layer and photoreceptor layer) is avascular, and the macular pit (fovea) is also avascular at all times during human development (Figure 2). The absence of vessels in the outer retina and fovea is essential for human vision.

Figure 1. Schematic of the eye and blood flow.

A. Schematic of the eye demonstrating the cornea, anterior chamber (blue), vitreous cavity (red), and retina (black box). B. Retinal schematic diagram demonstrating juxtaposed retinal cells and the corresponding retinal and choroidal vasculature. The inner retina receives its blood supply from branches from the central retinal artery and has two distinct capillary plexi: the superficial capillary plexus within the ganglion cell layer and the deep capillary plexus within the inner nuclear layer. The outer retina, including the photoreceptors and retinal pigment epithelium, are supplied by the dense underlying vasculature of the choriocapillaris and choroid.

Figure 2. Optical Coherence Tomography Angiography (OCTA).

A. OCT (Cirrus HD-OCT, Zeiss AngioPlex, Carl Zeiss Meditec, Inc., Dublin, CA) demonstrating calculations of the retinal thickness (purple hashed lines) and location of blood flow (red). B. OCTA of a normal human retinal vasculature demonstrating the foveal avascular zone and surrounding vascular plexus. (Image courtesy of Gregory Hoffmeyer, Carl Zeiss Meditec, Dublin, CA)

Vasculogenesis

There are two primary modes of blood vessel assembly: vasculogenesis and angiogenesis. Vasculogenesis is the de novo assembly of vessels from endothelial precursor cells (angioblasts) and mesenchymal precursors that form clusters (blood islands). The cells aggregate into tubes that fuse, form tube-like structures, interconnect into cords, and slowly lumenize. Vasculogenesis forms the initial embryonic vessels and begins in gastrulation but plays a more limited role in adult vessel growth (McLeod et al. 1987; Hughes et al. 2000; Chan-Ling et al. 2004, McLeod et al. 2006).

In the developing human fetal retina, mesenchymal precursors precede the ingrowth of retinal blood vessels (Ashton 1970, Chan-Ling et al., 2004). Astrocytes are noted to be the majority of dividing cells in developing blood vessels in the retina (Sandercoe et al. 1999). Hemangioblasts serve as the precursor to angioblasts, which proliferate and migrate peripherally as a precursor to endothelial cells. Both astrocyte precursor cells and angioblasts have been described in developing fetal human retina (Chu et al. 2001; Chan-Ling et al. 2004). Once believed to happen exclusively in fetuses, new blood vessel formation due to angioblast differentiation has recently been demonstrated in pathological angiogenesis in adults (Jiang et al 2005). In this setting, angioblasts stored in the bone marrow of adults are recruited to tissue to initiate vasculogenesis.

Vascular guidance and patterning during vessel growth occurs through attractant and repellant cues remarkably similar to axons in the developing nervous system. This requires a complex interplay between different cell types, including endothelial cells, pericytes, and periendothelial cells. Ganglion cells promote astrocyte spread in a placental-derived growth factor (PDGF)-dependent manner. Astrocytes appear to serve as a template for vessels by secreting vascular endothelial growth factor (VEGF) and through the expression of cadherins to promote cellular adhesion. Vessels, in turn, downregulate expression of VEGF by astrocytes and induce astrocytic maturation. Early genetic and local tissue factors determine arterial versus venous fate of a plexus (Gariano 2003).

Angiogenesis

Angiogenesis is the growth of new vessels from pre-existing vessels by a sprouting or branching process. Angiogenesis elaborates on the vascular plexus begun with vasculogenesis and accounts for most vascular growth in adults, including pathologic vessel growth in disease. In stark contract with normal retinal vasculogenesis, pathologic retinal angiogenesis creates chaotically oriented and physiologically deficient vessels that do not conform to neuronal histology. These vessels do not respect the normally avascular outer retina and macular pit, and can contribute to profound vision loss.

There are several steps in angiogenesis, including destabilization, vasodilation, hyperpermeability, cell migration, and cell proliferation. Each of these steps is regulated by the coordinated expression and/or activation of secreted mediators, cell-surface receptors, intracellular signaling molecules, and nuclear transcription factors. Circulating progenitor cells and hematopoietic stem cells from the bone marrow can contribute to angiogenesis, providing cellular building blocks for endothelial cells and secreting angiogenic molecules or differentiating into precursor perivascular cells. Angiogenesis is regulated by numerous physiologic properties including tissue oxygen tension, extracellular matrix composition, intrinsic genetic programs of endothelial cells, and a complex milieu of pro and anti-angiogenic cells and molecules.

Vasodilation occurs in response to chemical signals (e.g., nitric oxide) as well as hyper-permeability factors (e.g., VEGF). Angiogenic growth factors (including VEGF) activate receptors on endothelial cells of pre-existing blood vessels. The vascular endothelial cells release proteases, including matrix metalloproteases (MMPs) that degrade the basement membrane, allowing vascular endothelial cells to escape from the parent vessel (Rodrigues et al. 2013), while plasminogen activator promotes the disruption of the endothelial cell junctions (Yao and Tsirka 2011). Migrating endothelial cells proliferate in the surrounding extracellular matrix towards the angiogenic stimulus using integrins and other adhesion molecules. Periendothelial cells and pericytes are then recruited to promote vessel maturity (Birbrair et al. 2014).

Additional Methods of New Vessel Development

Additional mechanisms of formation of blood vessels have also been described but play a more limited role in the retina. Intussusception involves the formation of transvascular tissue pillars and includes intussusceptive microvascular growth to expand a capillary network, intussusceptive arborization to form feeding vessels from capillaries, and intussusceptive branching remodeling through pillars arising close to a bifurcation or through pillars arising at some distance from the bifurcation point. Intussesception has been described in numerous organs including the eye, as well as lung, intestine, and kidney (Djonov et al. 2002; Burri et al. 2004). Tube formation and remodeling, vascular mimicry, and biochemical translocation and expansion have also been described. Another mechanism, hamo-vasculogenesis, has been described in which blood vessel and blood cells differentiate from a common precursor, the haemangioblast. This is thought to play a role in human choroidal vascular development from 6 until 22 weeks gestation and involves endothelial cells or angioblasts and erythroblasts in the choriocapillaris. Fenestrations, smooth muscle actin, extracellular basal lamina, and pericytes develop later (Lutty et al. 2010).

Hypoxia and the Retinal Vasculature

In 1948, Michaelson used dye-perfusion techniques to evaluate the embryonic and perinatal retinal vasculature. He found that capillaries grow extensively around venules and less around arteries. Michaelson posited a “Factor X”, an oxygen-sensitive molecule, which controlled retinal vascular development through a concentration gradient (Michaelson 1948). Indeed, decreased oxygenation due to relative non-perfusion or an increase in metabolic demand is thought to drive the development of blood vessels in normal development (Yu et al 2001).

The important role of oxygen in retinal vascular development has been further refined since the observations of Michaelson. The superficial vascular plexus in the developing retina follows the metabolic demand of neuronal development and physiologic hypoxia (Chan-Ling 1995). Lowering the percentage of inhaled oxygen in animals reduces the rate and density of the retinal vasculature (Phelps 1990; Li et al. 2008). Hypoxia also plays an essential role in many retinal vascular diseases (Kaur et al. 2008), providing a link between physiologic and pathologic vascular processes in the eye. Indeed, the oxygen-dependent angiogenic stimulators expressed during normal vascular development are also expressed (albeit at much higher levels) in retinal vascular disease (Gariano et al. 2005).

Hypoxia Inducible Factor (HIF)

A group of transcriptional activators, the hypoxia-inducible factor (HIF)s, have recently emerged as the master regulators of these hypoxia-regulated angiogenic stimulators (Semenza 2012) and have helped illuminate the overlap between physiologic and pathologic vascular processes in the eye. HIFs are heterodimeric proteins composed of an exquisitely oxygen-sensitive α subunit and a ubiquitously-expressed β subunit (Semenza 2007). HIF-1α was the first HIF α subunit isoform to be identified (Wang et al. 1995) and its role has been extensively studied in retinal vascular disease (Scholz et al. 2013). Two other isoforms, HIF-2α and HIF-3α have since been reported; while HIF-2α is closely related to HIF-1α and activates hypoxia-inducible gene transcription, HIF-3α is distantly related and may antagonize HIF-1α and HIF-2α (Ratcliffe 2007).

In normoxic conditions, HIF-1α is hydroxylated and ubiquited by the VHL E3 ubiquitin ligase, labeling it for degradation in the proteosome (Maxwell et al. 1999). In hypoxic conditions, the VHL hydroxylase is inhibited. Thus, in normal conditions, HIF-1α is produced but constantly degraded. Conversely, under conditions of hypoxia (e.g., in ischemic retinal disease) HIF-1α is no longer targeted for degradation, and can promote the expression of pro-angiogenic gene products, including VEGF. Indeed, VEGF expression is temporally and spatially correlated with HIF-1α expression and retinal vascular growth (Kurihara et al. 2014). HIF-1α regulates expression of the angiogenic template by astrocytes (Nakamura-Ishizu et al. 2012) and is essential for retinal vascular development (Caprara et al. 2011).

The precise molecular regulation of HIF-1α has been elegantly demonstrated. Under standard tissue culture conditions (20% O₂), proline residues 402 and 564 on the HIF-1α subunit are hydroxylated by a family of HIF prolyl hydroxylases (PHDs) (Kaelin 2005). Hydroxylated HIF-1α binds to the von Hippel-Lindau (VHL) tumor suppressor protein, which ubiquitinates HIF-1α and targets it for degradation by the proteasome (Hubbi et al. 2014; Mole et al. 2002). An additional level of regulation is provided by an asparaginyl hydroxylase, factor inhibiting HIF-1 (FIH-1) (Lando et al. 2002; Mahon et al. 2001). FIH-1 hydroxylates asparagine residue 803 on HIF-1α and prevents binding of the transcriptional co-activator, p300, to HIF-1α, thereby inhibiting its transcriptional activity.

Under hypoxic conditions (<5% O₂), the ability of PHDs and FIH to hydroxylate HIF-1α is impaired (Figure 3). In the absence of hydroxylation, VHL does not bind to HIF-1α to trigger its degradation, whereas p300 binds to HIF-1α to enhance its transcriptional activity (Maxwell 2005; Schofield et al. 2004; Manalo et al. 2005). The resulting increased amount of active HIF-1α protein localizes to the nucleus and binds to HIF-1β forming a heterodimer (HIF-1) that is capable of binding to the DNA of specific (hypoxia-inducible) genes, and inducing broad changes in gene expression that mediate acclimation of cells, tissues, and the organism to conditions of low oxygen tension (Semenza 2012). Indeed, HIF-1 targets include numerous genes that play essential adaptive roles by promoting angiogenesis to increase O₂ delivery, regulating the metabolic shift from oxidative phosphorylation to glycolysis and lactic acid production to decrease O₂ demand, protecting cells from acidosis, and influencing adaptive survival mechanisms (Semenza 2012). These genes work together to collectively promote the survival of cells exposed to hypoxia.

Figure 3. Schematic of HIF-1α stabilization of the ischemic retina.

A. Ultrawide field fluorescein angiogram (Optos 200Tx, Optos plc, Dunfermline, Scotland, U.K.) of a normal retina. B. Schematic of HIF-1α in normoxia demonstrating degradation involving von Hippel-Lindau (VHL) protein. HIF-1α is constitutively transcribed. In normoxia, HIF-1α is hydroxylated on proline and acetylated on lysine. This complex is preferentially bound by VHL, ubiquinated, and degraded in the proteasome. C. Ultrawide field fluorescein angiogram demonstrating extensive peripheral capillary non-perfusion (star) and microaneurysms along with regions of neovascularization (red arrows) adjacent to regions of capillary non-perfusion (yellow arrows). D. Schematic of HIF-1α in hypoxia demonstrating HIF-induced transcription of hypoxia-inducible genes. In hypoxia, p300 is not inhibited by FIH (factor inhibiting HIF-1α) and interacts with the COOH-terminal transactivation domain (C-TAD). This results in HIF-1α stabilization against degradation through NH₂ -terminal transactivation domain (N-TAD), resulting in nuclear localization, heterodimerization of HIF-1α and HIF-1β, and binding of hypoxia response elements (HRE). This results in the transcription of hypoxia-inducible genes, including vascular endothelial growth factor (VEGF) and angiopoietin-2 (Ang2).

Numerous ischemic retinopathies exist where abnormal perfusion, permeability, and vasculature creates a mismatch between tissue metabolic demand and supply. These ischemic retinopathies involving retinal neovascularization include diabetic retinopathy, central and branch retinal vein occlusions, sickle cell retinopathy, hyperviscosity syndromes, radiation retinopathy, retinal tumors including retinoblastoma and uveal melanoma, ocular ischemic syndrome, chronic retinal detachment, and retinopathy of prematurity. Proliferative diabetic retinopathy has been show to result in an increase in numerous pro-angiogenic factors in the vitreous, including inflammatory cytokines, chemokines, extracellular matrix adhesion molecules, complement, polyamines, vasoactive peptides, and inflammatory cells (Gariano et al. 2005). Retinal oxygen saturation has been shown to decrease in patients with diabetes without retinopathy (Beach et al. 1999) and shown to increase adjacent to panretinal photocoagulation laser scars (Stefánsson et al. 1992). Retinal oxygen partial pressure has been shown to be reduced in cats with diabetes of 6–8 years duration (Linsenmeier et al. 1988). In addition to reduced oxygen, inflammation and aberrant wound-healing play an important role in diabetic retinopathy with microglial and glial cell activation resulting in the release of cytokines and chemokines (Amin et al. 1997; Krady et al. 2005).

In age-related macular degeneration (AMD), the development of choroidal neovascularization (CNV) appears to be a consequence of the expression of the same angiogenic mediators that drive the development of retinal neovascularization in ischemic retinal disease (Figg et al. 2008). However, it is postulated that oxidative stress instead of – or in addition to – hypoxia may stimulate accumulation of HIF-1α in AMD (Park et al. 2015). CNV occurs in conditions including age-related macular degeneration, pathologic myopia, uveitic and inflammatory conditions, choroiditis, presumed ocular histoplasmosis (POHS), punctate inner choroidopathy (PIC), multifocal choroiditis, angioid streaks, ocular trauma, vitelliform dystrophies, optic disc drusen, and idiopathic CNV. CNV occurs when the RPE or Bruch’s membrane is compromised. Bruch’s membrane is a 5-layered extracellular matrix that separates the RPE from the choriocapillaris and is composed of the basement membrane of the RPE, inner collagenous zone, central band of elastic fibers, outer collagenous zone, and basement membrane of the choriocapillaris. AMD can develop CNV of the choroid in addition to retinal angiomatous proliferation (RAP) originating from the deep capillary bed of the retina.

Identifying Therapeutic Targets for the Treatment of Ocular Neovascular Disease

There are numerous considerations when identifying an ideal target for treatment of ocular neovascular disease. One would consider a target whose expression is found only in the disease state (or markedly elevated in the diseased state). Low levels of the target in normal tissue would minimize potential side-effects of medications that inhibit that target. In addition, one would prefer a secreted factor or transmembrane receptor that would be readily accessible to a binding molecule given the difficulty in achieving adequate intracellular or intranuclear concentrations of some medications. An ideal therapeutic agent would be administered locally to the eye in an easily accessible form, such as topical instillation, subconjunctival depot, or intravitreal. Administration would preferably be non-invasive, so that administration would involve minimal risk to patients. And the duration of action would ideally be long, so as to minimize the number of re-treatments patients would need for these chronic diseases.

Vascular Endothelial Growth Factor (VEGF)

No single molecule has captured the attention of the retinal community as much as the angiogenic and vascular permeability stimulator, VEGF. The VEGF family consists of VEGF-A, -B, -C, and -D, and placental-induced growth factor (PIGF) 1 and 2. VEGF is a hypoxia-inducible cytokine that is necessary for retinal vascularization and is spatially and temporally expressed in developing retinal blood vessels (Stone et al. 1995; Provis et al. 1997; Ozaki et al. 2000). VEGF plays numerous roles in the retina, including angiogenesis promotion, increasing vascular permeability, stimulating endothelial cell migration and proliferation, and increasing endothelial expression of plasminogen activator and interstitial collagenase. Mice expressing a single VEGF isoform demonstrate unique retinal vascular appearance, suggesting that each isoform acts within a limited spatial location to create the vasculature. Mice expressing VEGF164 appear normal, whereas mice expressing VEGF188 have aborted arterioles and normal veins (Stalmans et al 2002). VEGF120, a more diffusible version, results in reduced vessel branching. VEGF also plays an important role in vascular pruning in the developing retina. Relative hyperoxia surrounding arteries reduces VEGF production and its ability to maintain endothelial cell survival (Alon et al. 1995; Claxton et al. 2003). VEGF has been shown to be elevated in the eye in patients with proliferative diabetic retinopathy and retinal vein occlusions (Adamis et al. 1994; Aiello et al. 1994).

The VEGF receptors, VEGFR-1 or Flt-1, VEGFR-2 or Flk-1, VEGFR-3 or Flt-4, and neuropilin 1 and 2, are expressed in a temporally and spatially distinct pattern that also allows for vascular patterning (Shih et al. 2003; Favier et al. 2006; Gariano et al. 2006). Signals by a given receptor can elicit different responses at different sites. VEGFR-2 activation causes migration and filopodia in endothelial tip cells and proliferation in more proximal stalk cells (Gerhardt et al 2003). Several inducing agents exist for VEGF, including TGF-α, TGF-β, PDGF, and hypoxia. VEGF also further refines vascular development through its interactions with other angiogenic factors, including placental growth factors, angiotensin, angiopoietin, and pigment epithelium-derived factor (Dawson et al 1999; Feeney et al. 2003; Sarlos et al 2003). Another potential use of anti-VEGF treatment is in the “normalization” of vasculature, reducing hypoxia and creating a window for chemotherapy (Datta et al. 2015).

Several therapies currently exist utilizing VEGF (table 1). The first approved therapy for age-related macular degeneration (AMD) in December 2004 was pegaptanib sodium (Macugen, EyeTech Pharmaceuticals/Pfizer), an RNA aptamer directed against VEGF-165. Pegaptanib was the first aptamer approved for use in humans (Ng et al. 2006). Aptamers are oligonucleotide ligands with high-affinity binding to molecular targets. Pegaptanib was demonstrated to be safe and effective in several prospective, randomized, multicenter, double-masked, sham-controlled studies (Gragoudas et al. 2004; Chakravarthy et al. 2006; D’Amico et al. 2006).

Table 1.

Current VEGF/KDR (kinase-insert-domain-containing receptor) anti-angiogenic targets in clinical trials

Drug	Drug target	Disease target	Clinicaltrials.gov ID	Phase	Number patients	Relevant clinical study
Pazopanib	VEGFR tyrosine kinase inhibitor	AMD	NCT00612456 NCT01134055	2, Completed 2009     2b	70 510	Danis R, et al. Br J Ophthalmol 2014;98:172–178
TG100801	VEGF tyrosine kinase inhibitor	AMD, PDR	NCT00414999	1, Completed	44
TG100801	Same	AMD	NCT00509548	2, Terminated early	Terminated after 7 due to corneal deposits
OPT-302	VEGFR-3	AMD	NCT02543229	1	Enrolling 50
Regorafenib	Receptor tyrosine kinase VEGFR2-TIE2	AMD	NCT02222207	2	52
rAAV.sFlt-1	Flt1	AMD	NCT01494805	1 and 2	40	Rakoczy EP, et al. Lancet. 2015;386(10011):2395–403.
PAN-90806	Anti-VEGFR	AMD	NCT02022540	1	50
Squalamine lactate	VEGF	AMD RVO PDR	NCT01678963 NCT02614937 NCT01769183	2     1 and 2     2	142   20     6
Abicipar Pegol (AGN-150998)	Anti-VEGF	AMD DME	NCT01397409 NCT02181517 NCT02181504 NCT02462928 NCT02186119	2     2     2     3     2	271   25   25 900, R 151
PF582	Anti-VEGF	AMD	NCT02121353	1 and 2	25
EYE001	Anti-VEGF Pegylated Aptamer	AMD	NCT00021736	2 and 3	540
AL-39324	Inhibit VEGFR and PDGFR	AMD	NCT00992563	2	35
LHA510	VEGFA-inhibitor	AMD	NCT02076919 NCT02355028	1     2	135
X-82	VEGF and PDGF inhibitor	AMD	NCT02348359 NCT01674569	2, R     1 and 2	132   35
ESBA1008	Anti-VEGF	AMD	NCT01304693	1 and 2	376
BCD-021	Anti-VEGF	AMD	NCT02450981	1	10
LMG324	Anti-VEGF	AMD	NCT02398500	1 and 2	25, Terminated
TK001	Anti-VEGF	AMD	NCT02613559	1	27,R
ORA102		AMD	NCT00745511	1 and 2	96
RG7716	Anti-VEGFA	AMD	NCT02484690	2	271
AGN211745	VEGFR-1 siRNA	AMD	NCT00395057	2	138, Terminated
PTK787 (Vatalanib)	Tyrosine kinase inhibitor targets VEGFR	AMD	NCT00138632	1 and 2	50

Abbreviations: R= recruiting; AMD = neovascular age-related macular degeneration; DME = diabetic macular edema; PDR = proliferative diabetic retinopathy; RVO = retinal vein occlusion

Bevacizumab (Avastin, Genentech/Roche) was approved by the United States Food and Drug Administration (FDA) in 2004 as treatment for metastatic colon cancer and is a whole humanized from mouse antibody to VEGF-A. It was described for use initially systemically as an intravenous treatment (Michels et al. 2005) and subsequently via off-label intravitreal administration for AMD (Avery et al. 2006; Rich et al. 2006). Prospective randomized controlled trials have proven the efficacy of bevacizumab in treating AMD (Sacu et al. 2009; Tufail et al. 2010), macular edema from CRVO (Epstein et al. 2012), and diabetic macular edema (DME) in the BOLT trial (Rajendram et al. 2012).

Shortly after bevacizumab was approved, the same company Genentech developed another monoclonal antibody to VEGF-A, ranibizumab or Lucentis (Ferrara et al. 2006). Ranibizumab is a humanized mouse Fab fragment to an epitope of VEGF (AS82-91) from within the receptor binding domain (AS8-109) of VEGF165 and was approved by the FDA in June 2006 for treatment of AMD, June 2010 for treatment of macular edema from retinal vein occlusions (RVO), August 2012 for treatment of diabetic macular edema (DME), and February 2015 for the treatment of diabetic retinopathy. Like bevacizumab, ranibizumab has neutralizing activity on all VEGF isoforms. Prospective randomized controlled trials have proven the efficacy of ranibizumab in treating classic CNV in the ANCHOR trial (Brown et al 2009), occult CNV in the MARINA trial (Rosenfeld et al. 2006), DME in RISE, RIDE, and RESTORE (Mitchell et al. 2011; Nguyen et al. 2012), BRVO in BRAVO (Campochiaro et al. 2010), and CRVO in CRUISE (Brown et al. 2010).

Several important differences exist between ranibizumab and bevacizumab: (1) ranibizumab (48 kDa) contains only the Fab fragment of the antibody, whereas bevacizumab (149 kDa) contains both the Fab and Fc fragments; (2) ranibizumab has six corresponding amino acids which differ from bevacizumab; (3) ranibizumab has one binding site for VEGF whereas bevacizumab has two; (4) ranibizumab is produced in prokaryotic Escherichia coli and therefore is not glycosylated, whereas bevacizumab is produced in a eukaryotic cell line (CHO cells) and is N-glycosylated in its Fc region; and (5) ranibizumab costs over $2000 per injection whereas bevacizumab costs approximately $60 per injection (Krispel et al. 2013).

However, it is unclear how much these differences between ranibizumab and bevacizumab translate into a clinically meaningful difference. The smaller size of ranibizumab was deliberate to enhance diffusion from the vitreous cavity into the retina and choroid (Ferrara et al. 2006). However, subsequent studies suggest that the predicted enhanced diffusion may not translate into a clinically meaningful therapeutic advantage (Martin et al. 2012). Affinity maturation of ranibizumab was also deliberate to increase the binding affinity to VEGF and increase the biologic activity compared to bevacizumab (Ferrara et al. 2006). While initial studies using a monovalent Fab-12 were significantly lower than what was later demonstrated for ranibizumab (Chen et al. 1999), later evaluation demonstrated that ranibizumab and bevacizumab had a similar affinity for VEGF-A165 given bevacizumab’s bivalent nature and dissociation constant (Papadopoulos et al. 2012). A possible disadvantage of the Fc fragment is that bevacizumab may be more stable systemically than is ranibizumab and thus have higher systemic levels (Miki et al. 2009). However, it is unclear whether this results in any clinically significant systemic effects.

Aflibercept (Eylea, Regeneron Pharmaceuticals Inc) is a recombinant fusion protein with the extracellular binding portions of VEGFR-1 and VEGFR-2 fused to the Fc portion of the human IgG1. Aflibercept received FDA approval in November 2011 for AMD, September 2012 for RVO, July 2014 for DME, and March 2015 for diabetic retinopathy. Aflibercept has been proven in large, randomized clinical trials for treating AMD in VIEW1 and 2 (Heier et al. 2012), CRVO in GALILEO (Korobelnik et al. 2014), BRVO in VIBRANT (Campochiaro et al. 2015), and DME in DA VINCI (Do et al. 2012).

Several studies have evaluated the relative efficacies of these bevacizumab and ranibizumab for treatment of AMD and found no difference in visual acuity or complications (Subramanian et al. 2010; Biswas et al. 2011; Chakravarthy et al. 2013; Kodjikian et al. 2013; Krebs et al. 2013). The CATT study found monthly treatments of bevacizumab not inferior to monthly ranibizumab, although ranibizumab had a greater mean decrease in central retinal thickness and bevacizumab had more serious systemic adverse events (CATT 2011). A Cochrane Review including 12 randomized controlled trials including 5496 patients for AMD found no difference in mean visual acuity outcomes between ranibizumab and bevacizumab, no difference in adverse events, and a statistically significant but clinically insignificant increase in reduction in central retinal thickness (−14 μm) with ranibizumab compared to bevacizumab (Solomon et al. 2014). Some disagreement also exists whether a difference in safety of bevacizumab and ranibizumab exists (Martin et al. 2012). Some retina specialists state that ranibizumab has been well studied in many randomized clinical trials with more long-term findings when compared with bevacizumab, although the superiority of ranibizumab over bevacizumab has not been proven. This remains a hot topic of debate among retina specialists given the significant difference in price between the medications, and both drugs are actively used to treat VEGF driven retinopathies currently.

A comparison of intravitreal bevacizumab, ranibizumab, and aflibercept for DME found that for visual acuity of 20/40 or better there was on average no difference in the improvement in visual acuity between the three treatments. For patients with visual acuity of 20/50 or worse, aflibercept resulted in more mean improvement in visual acuity (Diabetic Retinopathy Clinical Research Network 2015).

Further modifications in VEGF therapy have occurred through the development of new anti-VEGF molecules. Conbercept (KH902; Chengdu Kanghong Biotech Co., Ltd., Sichuan, China) consists of the VEGF binding domains of the human VEGFR-1 and VEGFR-2 combined with the Fc portion of the human immunoglobulin G1. It binds the VEGF-A along with VEGF-B and placental growth factor (Wang et al. 2013). A randomized, double-masked, multicenter, phase 2 clinical trial, of conbercept, AURORA, demonstrated the drug to be safe and efficacious with as needed dosing (Li X et al. 2014).

Recent concerns have been raised regarding possible increase in geographic atrophy with anti-VEGF administration, particularly ranibizumab (Grunwald et al. 2014 & 2015). In addition, repeat administration of intravitreal anti-VEGF results in sustained elevation of intraocular in some patients (Bakri et al. 2008; Pershing et al. 2013; Bakri et al. 2014; Dedania & Bakri 2015), suggesting the importance of VEGF to the trabecular meshwork.

Other Secreted Growth Factors Currently Under Study

In addition to VEGF, numerous other secreted growth factors that contribute to the angiogenesis cycle exist and are currently under investigation as potential therapeutic targets. Current anti-angiogenic targets are summarized in table 2. An important proangiogenic secreted molecule implicated in retinal neovascular disease is fibroblast derived growth factor (FGF) (Presta et al. 2005). FGF are a family that play an important role in angiogenesis, wound healing, and embryonic development. FGF plays a role in the migration and proliferation of endothelial cells along with the proliferation of smooth muscle cells and fibroblasts. Basic FGF (bFGF), a secreted cytokine, regulates angiogenesis through induction of VEGF expression via the FGFR1/c-Src/p38/NF-κB (nuclear factor κB) signaling pathway, triggering angiogenesis of endothelial progenitor cells (Tzeng et al. 2015).

Table 2.

Current non-VEGF targets in clinical trials

Drug	Drug target	Disease target	Clinicaltrials.gov ID	Phase	Number patients
Fractalkine (FKN)	CX3C chemokine	PDR	NCT00728598	1, completed 1998	30
OC-10X	tubulin inhibitor	PDR, AMD	NCT01869933	1	10
Fovista (E10030)	Anti-PDGF-B	AMD	NCT02591914 NCT02214628 NCT01089517 NCT01944839	1   2   2  3,R	25 100 449 622
Palomid 529	TORC1/2inhibitor of mTOR	AMD	NCT01033721 NCT01271270	1   1	13     5
AdGVPEDF.11D	PEDF	AMD	NCT00109499	1
Sirolimus	mTOR IL-2	AMD	NCT00766337	2	62
Everolimus (RAD001)	mTOR	AMD	NCT00857259	2	16, terminated
PF-04523655	Block RTP801 mTOR	DME AMD	NCT01445899 NCT00713518	2   2	258 152
ARC1905	Inhibit complement C5 aptamer	AMD	NCT00709527	1	60
iSONEP	S1P lipid	AMD	NCT01414153	2	158
ASP8232	Vascular adhesion protein-1 inhibitor	DME	NCT02302079	2	96
hI-con1	Factor VIIa inhibitor	AMD	NCT02358889	2	88
Tocilizumab	Antibody IL-6	DME	NCT02511067	2	66
iCo-007	inhibitor C-raf	DME	NCT01565148	2	208
Pf-04634817	CCR2/5 antagonist	DME	NCT01994291	2	212
ATG003 (mecamylamine)	Tubulin depolymerize, tight junction disruption	AMD	NCT00607750	2	60
RO6867461		DME AMD	NCT02699450 NCT01941082	2   1	150, R
Luminate (ALG-1001)	αVβ3 and αVβ5 integrin inhibitor	DME	NCT02348918	2	150
Apremilast (CC-10004)	Inhibit PDE4 (breaks down cAMP)	Behcet’s	NCT00866359 NCT02307513	2   3	111 204
Efalizumab	Bind CD11a	DME	NCT00676559	1	0, Withdrawn
Zimura	Anti-C5 Aptamer)	Polypoidal	NCT02397954	2	5
AKB-9778	Tie-2 activator	DME	NCT02050828	2	144
AL-78898A	Complement C3 inhibitor	AMD	NCT01157065	2	99
Volociximab	α5β1 integrin antagonist	AMD	NCT00782093	1	63
JSM6427	α5β1 integrin antagonist	AMD	NCT00536016	1	36

Abbreviations: R= recruiting; AMD = neovascular age-related macular degeneration; DME = diabetic macular edema; PDR = proliferative diabetic retinopathy; RVO = retinal vein occlusion

Platelet derived growth factor (PDGF) is critical for pericyte recruitment and maturation of neovascular vessels (Alvarez et al. 2006). Therapies targeting PDGF may destabilize neovacular tissue thereby enhancing current anti-VEGF approaches. This interplay among angiogenic factors is also observed between FGF and PDGF. FGF-2 works with PDGF-BB to upregulate expression of PDGFR-α and -β in newly formed blood vessels (Cao et al. 2003). Thus therapies targeting PDGF and FGF may prove to complement one another as well as current therapies targeting VEGF.

Placental-induced growth factor (PIGF) is a member of the VEGF family and plays an important role in several steps of vasculogenesis and angiogenesis through its receptor VEGFR-1 (Flt-1). An anti-Flt-1 antibody suppressed neovascularization in tumors and ischemic retina, and angiogenesis and inflammatory joint destruction in arthritis (Luttun et al. 2002). PIGF plays a role in endothelial cell migration and survival, recruitment of smooth muscle cells, and differentiation and activation of monocytes. PIGF mRNA expression was reduced by hypoxia in mice and elevated with anti-VEGF therapy (Zhou et al. 2014).

Hepatocyte growth factor (HGF) is a mesenchyme-derived pleiotropic factor and acts upon vascular endothelial cells to regulate cell growth, cell motility, and morphogenesis through epithelial mesenchymal interactions (Morishita et al. 2004). NK4, the N-terminal hairpin and subsequent four kringle domains of HGF, acts as the competitive antagonist for HGF and has been demonstrated to inhibit HGF-induced ERK1/2 (p44/42 mitogen-activated protein kinase) activation to prevent angiogenesis and tumor growth in mice (Kuba et al. 2000). HGF has also been implicated in papillary thyroid cancer angiogenesis (Scarpino et al. 2003).

Pigment epithelium-derived factor (PEDF) plays an important role in endothelial cell migration and proliferation, apoptosis through the p38 MAPK pathway or FAS/FASL, and mediates angiogenesis through effects on VEGFR-1 and -2 (Longeras et al 2012). An N-terminal 34-amino acid peptide (PEDF-34) has been demonstrated to have antiangiogenic properties along with anti-vasculogenic properties.

Pro-angiogenic Signaling Molecules as Therapeutic Targets

The VEGFR is a receptor tyrosine kinase and one of the earliest and most studied kinases targeted by synthetic small molecules for oncologic applications and may prove to be effective for the treatment of ocular neovascular disease. There are currently 7 approved small molecules inhibitors targeting the VEGFR: sorafenib (Nexavar, Bayer), sunitinib (Sutent, Pfizer), pazopanib (Votrient, GlaxoSmithKline), axitinib (Inlyta, Pfizer), regorafenib (Stivarga, Bayer), nintedanib (Ofev, Boehringer Ingelheim), and lenvatinib (Lenvima, Eisai Inc.). Both sorafenib and regorafenib are multitarget protein kinase inhibitors. Besides inhibition of VEGFR and B-raf, sorafenib has also been shown to be a potent low-nanomolar inhibitor of p38a (Simard et al. 2009). In addition, vandetanib (Caprelsa, AstraZeneca) is a multiple kinase inhibitor against EGFR, VEGFR, and RET (Knowles et al. 2006). Cabozantinib (Cometriq, Exelixis) is a dual MET and VEGFR2 inhibitor.

Numerous drugs targeting other signaling pathways that promote pathological angiogenesis have been developed or are being investigated for oncologic applications. Extension of these therapies for the treatment of ocular neovascular disease may follow. Trebananib (AMG 386) is a peptide-Fc fusion protein, or peptibody, reported to neutralize the interaction between angiopoietins (Ang1/2) and their Tie2 receptors, which has been shown to be promising in ovarian cancer (Marchetti et al. 2015). Nintedanib has activity against platelet derived growth factor receptor (PDGFR) and fibroblast growth factor receptor (FGFR) in addition to VEGFR, thus offering a broader spectrum of anti-angiogenic activity than simply VEGFR. Nintedanib has been shown to be effective in lung cancer (Hilberg et al. 2008; Reck et al. 2014). Linifanib (ABT-869) is a tyrosine-kinase inhibitor that selectively targets VEGFR and PDGFR and has low off-target inhibitory activity and anti-angiogenic activity (Aversa et al. 2015).

The Wnt signaling pathways are a group of signal transduction pathways that pass extracellular signals through cell surface receptors to intracellular. The canonical Wnt signaling pathway involves β-catenin. Wnt signaling pathways are activated by the binding of a Wnt-protein to a Frizzled receptor, which activates Dishevelled. Wnt controls cell fate specification, cell proliferation, and cell migration. A glial-derived extracellular ligand, Norrin, can stimulate Wnt signaling and acts on the transmembrane receptor, Frizzled4, a coreceptor, Lrp5, and an auxiliary membrane protein, Tspan12, on the surface of developing endothelial cells. The resulting signal controls a transcriptional program that regulates endothelial growth and maturation (Ye et al. 2010). Therapies targeting Wnt signaling are also under development for the treatment of cancer; these, too, may prove effective as an anti-angiogenic approach for the treatment of ocular neovascular disease.

Nestin is an important molecule expressed in the cell soma of diving neural progenitor cells and their leading processes. After this, it follows vascular branches, suggesting vasculogenesis along microglia migrating routes sustains its angiogenic potential (Lee et al. 2012). Endothelial nitric oxide synthases occurs in the nuclei of endothelial vascular cells in vasoformative cells, suggesting it may also play a role in vasculogenesis (McLeod et al. 2012). Neuropilins are transmembrane glycoproteins that are receptors for VEGF and are essential for blood vessel development and assists with the separation of veins and arteries (Fantin et al. 2011). In addition to VEGF, neuropilins bind to semaphorins, molecules critical for axon guidance which have recently been implicated also in retinal neovascular disease (Cerani et al. 2013).

Akt-mediated phosphorylation of Girdin, an actin-binding protein, promotes VEGF-dependent migration of endothelial cells and tube formation. Exogenously delivered adenovirus harbouring Girdin short interfering RNA markedly inhibited VEGF-mediated angiogenesis (Kitamura et al 2008). Mammalian target of rapamycin (mTOR), a key mediator of PI3K/Akt/mTOR signaling pathway, has recently emerged as a compelling molecular target in glioblastoma. The mTOR is a member of serine/threonine protein kinase family that functions as a central controller of growth, proliferation, metabolism and angiogenesis, but its signaling is dysregulated in various human diseases especially in certain solid tumors including the glioblastoma (Cui et al. 2015). mTOR has been implicated in the regulation of HIF-1 translation and may play a critical role in regulating HIF-directed angiogenesis in retinal neovascular disease.

The small GTPase RhoA and its downstream effectors, ROCK1 and ROCK2, are important mediators in a number of angiogenic processes, including EC migration, survival, and cell permeability, and suggest that Rho/ROCK inhibition may prove useful for the treatment of angiogenesis-related disorders (Bryan et al 2010).

α_vβ₃ integrin has been reported as a promising therapeutic target for angiogenesis. GOPPP, a novel antagonist of α_vβ₃ integrin, has been shown to inhibit the pro-angiogenic effects of vitronectin on HUVECs, including adhesion, proliferation, and migration, and inhibit ERK1/2 and Akt phosphorylation. HIF-1α and VEGF were also inhibited by GOPPP in mice (Li YJ et al. 2014). Proliferative diabetic retinopathy has been shown to recruit bone-marrow-derived CD133+ endothelial progenitor cells and CD14+ monocytes to assist (Abu El-Asrar et al. 2011). Other integrins under active investigation include α_vβ₅ and α₅β₁ in current clinical trials for AMD and diabetic macular edema.

MicroRNAs (miRNA) are small non-coding RNA molecules (containing approximately 22 nucleotides) that functions in RNA silencing and post-transcriptional regulation of gene expression. A single miRNA can regulate the expression of hundreds of gene products and can have broad effects on cell function. MiRNA-132 as a highly upregulated miRNA in a human embryonic stem cell model of vasculogenesis and was found to be highly expressed in the endothelium of human tumors. Anti-miRNA-132 has been shown to effectively inhibit neovascularization (Anand et al. 2010). However, whether therapies targeting microRNAs will be specific enough for the treatment of human disease remains unknown.

Emerging Therapeutic Targets

Future anti-angiogenic approaches may include therapies upstream of the angiogenic secreted factors [e.g., HIF-1 (Subhani et al. 2016)] or novel angiogenic secreted molecules [e.g., angiopoietin-like 4, or ANGPTL4 (Xin et al. 2013; Sodhi et al. 2015)] or signaling pathways regulated by these molecules. However, recent focus has been directed at the molecules that regulate other steps in pathological angiogenesis (table 3). Appreciation for the role of circulating progenitor cells and hematopoietic stem cells from the bone marrow to pathological angiogenesis (Liekens et al. 2010) has exposed the SDF1/CXCR4 axis as a novel therapeutic target for ocular neovascular disease. The extracellular matrix plays an essential role in retinal neovascularization. MMPs are necessary to degrade the basement membrane to allow migrating vascular endothelial cells to escape from the parent vessel (Rodrigues et al. 2013). Although therapies targeting MMPs have not proven successful, other molecules that regulate extracellular matrix protein production have emerged as potential targets including transforming growth factor (TGF)-α and -β, endoglin, epidermal growth factor (EGF), inflammatory cytokines including IL-6 and 8, TNF-alpha and beta inhibitors, a5b1 integrin receptor inhibitor, and α_vβ₃ and α_vβ₅. Another critical step in pathological angiogenesis is the promotion of endothelial cell survival through both intrinsic (e.g., VEGF) and extrinsic mechanisms. The latter is ensured through stabilization of endothelial cells through recruitment of pericytes, smooth muscle cells, and deposition of extracellular matrix proteins in addition to focal adhesion kinase (FAK) (Roy-Luzarraga et al. 2016). In addition, endothelial cell transdifferentiation has exposed additional targets including ID1 and ID3 (Ruzinova et al. 2003).

Table 3.

Future anti-angiogenic targets to consider

Drug target	Relevant literature/studies
HIF-1/2	Subhani S, et al. Angiogenesis. 2016 Jul;19(3):257–73.
Angiopoietin-like 4 (ANGPTL4)	Babapoor-Farrokhran S, et al. Proc Natl Acad Sci U S A. 2015;112(23):E3030-E3039. Xin X, et al. Proc Natl Acad Sci U S A. 2013 Sep 3;110(36):E3425–34. Kwon SH, et al. Ophthalmology. 2015 May;122(5):968–75.
Stromal-derived factor-1 (SDF-1) and its receptor CXCR4	Ghanem I, et al. Am J Transl Res. 2014 Jul 18;6(4):340–52.
Metabolic gene products	Treps L, et al. Pharmacol Rev. 2016 Jul;68(3):872–87.
Matrix metalloproteinases (MMP)	Sampieri CL, et al. J Cancer Res Ther. 2013 Jul-Sep;9(3):356–63. Zhang X, et al. Int J Oncol. 2016 May;48(5):1783–93. Chang JH, et al. Surv Ophthalmol. 2016 Jul-Aug;61(4):478–97.
Inflammation and inflammatory cytokines	de Oliveira Dias JR, et al. Br J Ophthalmol. 2011 Dec;95(12):1631–7. Schor AM, Schor SL. Eye (Lond). 2010 Mar;24(3):450–8.
Reactive oxygen species (ROS)	Wilkinson-Berka JL, et al. Clin Sci (Lond). 2013 May;124(10):597–615.
Stem cell and endothelial progenitor therapy	Nazari H, et al. Prog Retin Eye Res. 2015 Sep;48:1–39.
Integrin antagonists	Salehi-Had H, et al. PLoS One. 2011 Apr 29;6(4):e18864. Varner JA, et al. Important Adv Oncol. 1996:69–87. Tolentino MJ. Curr Mol Med. 2009 Nov;9(8):973–81.
Methionine aminopeptidase	Ma AC, et al. Blood. 2011 Nov 17;118(20):5448–57. Mauriz JL, et al. Curr Drug Targets. 2010 Nov;11(11):1439–57.

Vascular Permeability Targets

There is significant overlap between pathological angiogenesis (and retinal neovascularization) and vascular hyper-permeability (and macular edema). Indeed, one of the first steps in pathological angiogenesis is the induction of vascular permeability. It is therefore not surprising that many of the therapeutic targets for retinal neovascularization have proven (or may prove) effective for the treatment of macular edema. This is most notable for therapies targeting VEGF. Nonetheless, macular edema is a major cause of vision loss in numerous disorders, including uveitis, pars planitis, post-operative, and retinal degenerations such as retinitis pigmentosa, in which retinal neovascularization is less common, and there are several mediators that promote vascular permeability independent of their ability to promote angiogenesis.

Macular edema results from a breakdown of the blood-retinal barrier (BRB), which is composed of an inner and outer component (Cunha-Vaz 1976). The inner component is formed by tight functions and adherens junction between vascular endothelial cells. Pericytes and perivascular astrocytes mediate the inner component also and the paucity of intraendothelial cell vesicles. The outer component is established by tight junctions between RPE cells that prevents fluid from choroidal vessels entering the retina (Vinores 1995; Rizzolo 1997; Vinores et al. 1999).

In diabetes, an upregulation of trans-endothelial vesicular transport and increased membrane permeability correlates with BRB breakdown. The players in vascular permeability are identical to those in pathological angiogenesis, including retinal glial cells, vascular endothelial cells, and neutrophils, which together promote breakdown of adherens and tight junctions and increase vesicular transport. And the secreted factors observed in neovascularization, including VEGF, TNF-α, and IL-1β, are elevated also in macular edema (Cuff et al. 1996; Luna et al. 1997). Similarly, ANGPTL4, an emerging target in pathological angiogenesis and retinal neovascularization (Babapoor-Farrokhran et al. 2015) may also play an important role in vascular permeability and macular edema (Xin et al. 2013). The downstream signaling molecules are also similar between retinal neovascularization and diabetic macular edema. As such, src kinase inhibitors (Doukas et al. 2008) and tyrosine kinase inhibitors such as pazopanib have been shown to reduce diabetic macular edema in animal models (Thakur et al. 2011). Protein kinase C (PKC) inhibitors such as ruboxistaurin showed initial promise in reducing diabetic macular edema (Gálvez 2011; Aiello et al. 2011), although some phase 3 studies have shown mixed results (Sheetz et al. 2013).

Molecular Imaging and Theranostics

Currently available ophthalmic imaging modalities are capable of imaging retinal pathologic anatomic changes, including hemorrhages, exudates, and edema, with unprecedented resolution. However, anatomical abnormalities are the end-product of complex molecular processes. Subclinical molecular changes occur before retinal disease can be detected by current anatomy-driven imaging instrumentations. Thus, the field of molecular imaging is emerging as an important factor to assist with early disease detection, improved treatment monitoring, and improved understanding of retinal pathophysiology (Xie et al. 2012; Capozzi et al. 2013; Evans et al. 2014). Several recent studies have reported feasibility of molecular imaging in detecting retinal ganglion cells (RGCs), RPE cells, vascular endothelial cells, VEGFR, and leukocytes.

Detection of apoptosing retinal cells (DARC) has been described for the single-cell detection of RGC apoptosis (Cordeiro et al. 2004; Coxon et al. 2011). In DARC, Annexin V was intravitreally injected to specifically bind the apoptosis biomarker phosphatidylserine (PS), and then it was detected by ophthalmic fluorescence imaging instrumentation. A peptide-based fluorescent probe (TcapQ) sensitive to active caspases such as caspase 3 involved in RGC apoptosis has also been described in vivo to quantify apoptotic RGCs (Barnett et al. 2009). Another promising opportunity for molecular imaging of RGCs lies in imaging RGC dysfunction before cell death. Several imaging probe such as reactive oxygen (ROS) (Dickinson et al. 2011), mitochondria permeability transition pore (Vrabec et al. 2003), and E glutamate (Okubo et al. 2010) hold great promise in improving the molecular imaging of RGCs. RPE-related molecular targets that have been evaluated include ROS (King et al. 2004), β-amyloid, esterified cholesterol and carbohydrate moieties in drusen (Hageman et al. 2001).

Vascular endothelial cells have emerged as critical surface biomarkers of neovascularization and as potential candidates for development of targeted contrast agents for ophthalmic imaging of retinal and choroidal neovascularization. The C-C chemokine receptor 3 (CCR3) is a promising biomarker of choroidal neovascularization (CNV), as demonstrated by CCR3 expression on choroidal neovascular endothelial cells in human CNV specimens (Takeda et al. 2009). Other promising biomarkers include targeting proliferating endothelial, endoglin (Grisanti et al. 2004), and integrin α_vβ₃ (Li et al. 2010; Peiris et al. 2012). VEGFR-2 expression has been demonstrated with molecular imaging to be elevated in retinal capillaries in diabetes (Sun et al. 2014). OCT has been used with anti-mouse CD45 coated gold nanorods to visualize leukocytes at sites of laser-induced retinal injury (Sen et al. 2016). Photoacoustic imaging can also be used to perform molecular imaging (de la Zerda et al. 2010; Hu et al. 2015). Molecular imaging is now further evolving into the development of theranostics, agents that combine diagnostic imaging with targeted therapy (Ding et al. 2013; Yan et al. 2016). While currently in its infancy, molecular imaging of the retina is rapidly developing and will play a critical role in early disease diagnosis and treatment monitoring in the near future.

Drug Delivery Pathways

Treatment with anti-VEGF is often transient and must be repeated with rapid regrowth of neovascularization on cessation of treatment. In addition to intravitreal injections which are the most common routes of administration, numerous additional pathways of delivery are being investigated. Recent developments in drug delivery have led to investigations on longer-lasting agents through biodegradable implants, gene delivery, biodegradable polymers, suprachoroidal delivery, microspheres, contact lenses with sustained delivery, punctal plugs with sustained delivery, nanoparticles, liposomes, and gels (Querques et al. 2015). ForSight Vision has developed a refillable, nonbiodegradable port delivery system that is surgically implanted in the pars plana beneath the conjunctiva through a 3.2 mm scleral incision. A phase 1 study was conducted in Latvia included 20 treatment-naïve patients with neovascular AMD (Rubio 2014). The primary endpoint of the study was at 1 year. There were 4 serious adverse effects, including 1 case of endophthalmitis, 2 cases of persistent vitreous hemorrhage, and 1 case of traumatic cataract. The visual acuity gain for all patients at month 12 was 10 letters, and the average number of refills for the full 20-patient cohort was 4.8. A phase 2 study is being planned.

Topical therapy has also been developed, pazopanib (GlaxoSmithKline, Brentford, UK), a multi-tyrosine kinase inhibitor having an effect on VEGFR-1, VEGFR-2, VEGFR-3, PDGFR-α, PDGFR-β. A study evaluating 70 patients with AMD treated for 28 days found a decrease in central retinal thickness and an increase in visual acuity in only a subset of patients (Danis et al. 2014). Bevacizumab therapy was initially investigated in AMD as a systemic treatment given intravenously, and oral therapies are also being investigated.

Another strategy of sustained delivery is through gene delivery through adeno-associated viral vectors (AAV). Gene therapy trials have been completed for Leber’s congenital amaurosis and have shown safety and promising results (Maguire et al. 2008). AAV2-sFLT01 (Genzyme, Cambridge, MA, USA) allows delivery of VEGFR1 (sFlt1) and has demonstrated sustained expression in animals (Pechan et al. 2009; Lukason et al. 2011). Similarly, AVA-101 (rAAV.sFLT-1 recombinant adeno-associated virus) is currently undergoing a phase I study (ClinicalTrials.gov identifier: NCT01494805).

Additional routes of administration which can be considered include transcleral, subretinal, and high-velocity (needleless) delivery (Todorich et al. 2014), siRNA, and cell-based therapy. Single chain antibody fragments are being developed which allow for sustained delivery due to their small size. A phase I study of the safety and tolerability of a single chain antibody fragment ESBA1008 (Alcon, Fort Worth, Tex., USA) versus ranibizumab has been completed in 194 patients in a prospective, randomized, multicenter trial. Encapsulated cell technology involves implants which can continuously produce recombinant therapy without needing to be refilled. NT-503 (Neurotech Pharmaceuticals, RI, USA), an intraocular implant delivering anti-VEGF, is being tested in a phase I prospective, multicenter study (Querques et al. 2015). In addition to pharmacologic therapies, laser photocoagulation can be used along with photodynamic therapy, transpupillary thermotherapy, and radiation therapy as a possible treatment by itself or in conjunction with pharmacotherapy.

Conclusion

Anti-angiogenic pharmacologic therapy has become an indispensable component in the management of diseases of the vitreous and retina. Anti-angiogenic pharmacotherapy has revolutionalized the treatment of many ocular conditions, including diabetic retinopathy, retinal vein occlusions, pathologic myopia, choroidal neovascularization, and age-related macular degeneration. Our improving understanding of the pathophysiology of embryology, angiogenesis, and vasculogenesis has allowed us to increasingly target a select tissue while minimizing the side effects of therapy. Selective therapeutic approaches will achieve desirable biochemical, cellular, and tissue effects while excluding unwanted damage, and will improve our understanding of the function and pathology of posterior segment diseases. Continuous innovations in pharmacotherapy and progress in understanding of retinal pathophysiology make us believe that improvements in the treatment of retinal diseases using anti-angiogenic therapy will continue for many years to come.