New Therapies of Neovascular AMD—Beyond Anti-VEGFs

Neovascular age-related macular degeneration (nAMD) is one of the leading causes of blindness among the aging population. The current treatment options for nAMD include intravitreal injections of anti-vascular endothelial growth factor (anti-VEGF). However, standardized frequent administration of anti-VEGF injections only improves vision in approximately 30–40% of nAMD patients. Current therapies targeting nAMD pose a significant risk of retinal fibrosis and geographic atrophy (GA) development in nAMD patients. A need exists to develop new therapies to treat nAMD with effective and long-term anti-angiogenic effects. Recent research on nAMD has identified novel therapeutic targets and angiogenic signaling mechanisms involved in its pathogenesis. For example, tissue factor, human intravenous immune globulin, interferon-β signaling, cyclooxygenase-2 (COX-2) and cytochrome P450 monooxygenase lipid metabolites have been identified as key players in the development of angiogenesis in AMD disease models. Furthermore, novel therapies such as NACHT, LRR and PYD domains containing protein 3 (NLRP3) inflammasome inhibition, inhibitors of integrins and tissue factor are currently being tested at the level of clinical trials to treat nAMD. The aim of this review is to discuss the scope for alternative therapies proposed as anti-VEGFs for the treatment of nAMD.


Introduction
Age related macular degeneration (AMD) is the most common cause of irreversible blindness among the elderly population [1,2]. Current global prevalence of AMD stands at 170 million and with aging as a major risk factor, it is expected to increase to 288 million by the year 2040 [3]. AMD can be classified into early, intermediate, and advanced forms, depending on the severity of the symptoms [4,5]. Early and intermediate AMD, also referred to as dry AMD (non-exudative), is characterized by accumulation between the retinal pigment epithelium (RPE) and Bruch's membrane of a yellowish-lipid rich protein content known as drusen leading to functional loss of the retinal photoreceptors [4]. Drusen deposition is considered as the hallmark of AMD [4,5]. The advanced form of AMD is known as geographic atrophy (GA) and is characterized by the loss of RPE and choroid near the macular region leading to the loss of photoreceptors and central vision [4,5]. The severe form of AMD (exudative) presents with growth of abnormal blood vessels from the choroid extending into the avascular RPE and sub-retinal regions. This phenomenon is known as choroidal neo-vascularization (CNV) and the form of AMD with CNV is termed nAMD [4,5]. nAMD accounts for the majority of cases of blindness in AMD patients [1,3].
The formation of neovascularization in AMD is a complex process involving multiple signaling pathways mediated by vascular endothelial growth factor (VEGF), platelet-derived growth factor, fibroblast growth factor, transforming growth factor, the Wnt pathway, the NLRP3 inflammasome, mitogen-activated protein kinase (MAPK) signaling, interleukins and chemokines [6][7][8][9][10]. VEGF-A, and is expressed in the outer retina under normal conditions and the inner retina during hypoxia [38]. Previously, Sema3F was reported to be protective against subretinal vascularization in mouse models [39]. Sun et al. investigated the anti-angiogenic role of Sema3F in two different mouse models, a very low-density lipid-receptor knockout model of spontaneous subretinal neovascularization where the resultant lesions are comparable to retinal angiomatous proliferations in humans and laser-induced CNV mimicking human nAMD [39]. It was found that intravitreal injection of AAV2. Sema3F effectively inhibited subretinal neovascularization and CNV in both models ( Figure 1) [39]. Considering these data, Sema3F could be a potential target around which to design novel therapies for nAMD.

Tissue Factor Inhibition
Tissue factor (TF) is a transmembrane receptor for plasma coagulation factor VII [40,41]. Studies have reported that TF is one of the key mediators in pathological neovascularization and thrombosis [42]. Under normal physiological conditions TF is not expressed by cells, however vascular endothelial cells, monocytes and macrophage express TF in response to inflammation [43]. Increased expression of TF has been observed in the RPE of nAMD patients compared to non-AMD retinas [42,44]. Intravitreal injection of anti-TF monoclonal antibody contributed to the reduction of CNV in a mouse model [45]. With this evidence, TF has been identified as a novel target with which to treat nAMD through the development of hI-con1. hI-con1, a synthetic molecule coupled with factor VII conjugated to the Fc region of an antibody, selectively binds to TF and destroys pathological vessels [46]. hI-con1 is being tested in a multi-center phase II clinical trial, with results pending (ClinicalTrials.gov identifier: NCT02358889) [46].
Despite the reported protective role of the ω3, ω6-LCPUFAs (DHA, EPA), in reducing nAMD in animal models, results from the Age-Related Eye Disease Study 2 (AREDS2) a multi-center, randomized, double masked phase III human clinical trial, report conflicting results [52]. AREDS2 tested whether the addition of ω3, ω6-LCPUFAs to the original AREDS formulation containing vitamin C, vitamin E, beta-carotene, zinc, and copper would improve the clinical outcome in AMD patients [52]. Results from AREDS2 concluded that the use of ω3, ω6-LCPUFAs did not reduce the risk of AMD progression [52]. On the other hand, a randomized prospective study, Nutritional AMD Treatment 2 (NAT2), has been reported, simultaneous to AREDS2. NAT2 investigated the relationship between ω3, ω6-LCPUFAs in AMD patients [53,54]. NAT2 reported that the incidence and the time-to-occurrence of CNV was not significantly different between the DHA and the placebo group [53,54]. However, a further detailed analysis of NAT2 patient data revealed that individuals with high levels of red blood cell membrane (RBCM) EPA + DHA (good marker for LCPUFAs) are at reduced risk to develop nAMD, suggesting that RBCM EPA + DHA can be a clinically useful biomarker of AMD [54]. Clearly, there are few similarities and differences between the AREDS2 and NAT2 studies (discussed in detailed [25]) in terms of the number of participating patients and the variety of treatment groups that may account for reported differences.

Interleukin-33 Therapy
Interleukin-33 (IL-33), a pro-inflammatory cytokine, is a member of the type-2 IL-1 family [55]. Once activated, IL-33 signals via its receptor, ST2 and the IL-1R accessory protein [56]. In humans, IL-33 is expressed in epithelial cells, endothelial cells and fibroblasts, and in rodents, its expression has been detected in RPE, the inner retina, and the choroid, along with the lymph nodes, spleen and central nervous system (CNS) [55,57,58]. In mouse experimental autoimmune uveitis, systemic administration of IL-33 attenuated the disease severity [58]. Similarly, Theodoropoulou et al. reported a protective role of IL-33 in a laser-induced CNV mouse model [59]. Intravitreal injection of recombinant IL-33 inhibited the development of CNV in mice via inhibition of ST2-expressing fibroblasts and endothelial cells, but did not alter the levels of VEGF [59]. This study discovered a novel mechanism involved in attenuating CNV independent of VEGF signaling, suggesting that recombinant IL-33 therapy could serve as an alternative treatment for nAMD patients who do not respond to conventional anti-VEGF treatments.

Targeted Intraceptor Nanoparticle Therapy
Targeted intraceptor nanoparticle therapy is a three-component system that consists of (1) plasmids expressing Flt23k intraceptors, (2) PLGA-biodegradable nanoparticles, and (3) the tripeptide adhesion motif Arg-Gly-Asp (RGD) [60]. Flt23k intraceptors are composed of VEGF-binding domains 2-3 of Flt, a high affinity VEGF receptor, while RGD facilitates the selective localization of nanoparticles to CNV after intravenous injection [60]. The Flt23k component inhibited CNV, and the RGD component suppressed fibrosis in mice and primates [60]. Although this is an anti-VEGF strategy to inhibit CNV, it has advantages over the conventional treatment for nAMD, (intravitreal injections) that is associated with pain, retinal detachment, and scarring [60]. Targeted intraceptor nanoparticle therapy is administered intravenously and can cross the blood-retinal barrier, which is generally a major obstacle for intravenous delivery. Furthermore, this therapy is nontoxic in vivo, suggesting that this could provide a means for alternative drug delivery route to treat nAMD [60,61].

Targeting MyD88 Pathway and DICER 1
GA is an advanced form of AMD for which there is no current effective treatment [9]. GA can occur over time in patients after repeated administration of anti-VEGFs [62][63][64][65]. Hence, it is important to understand the molecular mechanism involved in the pathogenesis of GA to identify novel therapeutic targets. RPE degeneration leading to loss of photoreceptor function is commonly seen in the patients with GA [9,66]. This RPE degeneration is associated with the accumulation of Alu RNA which was previously shown to cause RPE cell death (Figure 1) [66]. Alu RNA elements are non-coding genomic sequences that constitute almost 11% of the human genome [67]. They belong to a class of retroelements termed SINEs (short interspersed elements) and are primate-specific [67,68]. Tarallo et al. recently discovered that Alu RNA accumulates in GA patients due to a deficiency of the enzyme DICER 1 that functions to cleave Alu RNA [8]. The accumulated Alu RNA activated the NLRP3 inflammasome ( Figure 1) and triggered IL-18-dependent MyD88 signaling in the RPE [8]. The NLRP3 inflammasome is an intracellular NOD-like receptor that operates in innate immunity [69]. Upon activation, it cleaves pro-IL-1β and pro-IL-18 into their biologically active forms via caspase-1 [69]. Pharmacological inhibition of the NLRP3 inflammasome, MyD88, or IL-18 in mouse models and human RPE cell cultures prevented RPE cell death resulting from DICER 1 deficiency [66]. Furthermore, activation of caspase-11 (caspase-4 in humans) in mice has been implicated in the pathogenesis of GA [70]. This activation was mediated by cyclic GMP-AMP synthase (cGAS) leading to IFN-β production and gasdermin D-dependent IL-18 secretion (Figure 1) [70]. Elevated levels of gasdermin D, IFN-β, caspase-4, and cGAS have also been observed in the RPE of human eyes with GA [70]. Discovery of these series of events from DICER 1 deficiency in RPE to cGAS is a breakthrough in understanding the pathogenesis of GA and opens new platforms for novel therapies to treat GA [66,70].

Interferon-β Therapy
Immune cells such as microglia and mononuclear phagocytes play important roles in angiogenesis [71,72]. Microglia, the resident macrophages in the retina, are attracted to the choroid and RPE during CNV [14]. Inhibition of monocyte (precursors for macrophages) migration into the retina reduced CNV in a laser-induced mouse model, suggesting that microglia and monocyte derived macrophages may be pro-angiogenic [71,73]. Targeting the signaling pathways involving macrophage migration could be of therapeutic benefit in neovascular diseases such as AMD [74]. Interferon-beta (IFN-β), via interferon-alpha/beta receptor (IFNAR) signaling, has been identified as a critical pathway in regulating autoimmunity and monocyte/microglia influx in the CNS [75,76]. Luckoff et al. investigated the role of IFN-β and its receptor IFNAR in a laser-induced CNV mouse model and reported that IFN-β and IFNAR play a pivotal role in retinal microglia/macrophage activation and infiltration [77]. IFNAR knock-out mice presented with increased microglia/macrophage activation and CNV [77]. Intraperitoneal injection of recombinant human IFN-β 1a to CNV-induced wild-type mice significantly attenuated CNV formation, vascular leakage, and microglia/macrophage infiltration, suggesting that systemic IFN-β therapy could be a promising treatment option for nAMD patients [77]. Since IFN-β therapy is a well-established treatment option for multiple sclerosis and autoimmune encephalomyelitis, it could have a great potential for treating neovascular diseases such as AMD [78][79][80]. Systemic treatment could reduce the complications resulting from intravitreal injections. However, IFN-β therapy is associated with side-effects such as flu-like symptoms, increased spasticity, menstrual disorders, muscle pain, and headache [81,82]. These-side effects should be taken into consideration in optimizing and designing IFN-β therapies to treat nAMD.

Intravenous Injection of Immune Globulin
Intravenous immune globulin (IVIg) is pooled plasma from thousands of healthy donors with a diverse antibody repertoire and specificity [83]. IVIg has been approved by the FDA for the treatment of primary immunodeficiency diseases [84,85]. The first record of IVIg use dates to the year 1981 for the treatment of thrombocytopenic purpura in children [86]. Bogdanovich et al. reported that human IgG1 is a potent anti-angiogenic factor acting via Fc-mediated signaling through the FcγR1 receptor, a high-affinity receptor for IgG1 [87]. Based on these facts, Yasuma et al. tested the anti-angiogenic properties of IVIg, which is composed of approximately 60% IgG1 in five different humanized mouse models of angiogenesis [88]. Intravenous and intravitreal administration of IVIg in nAMD mice suppressed angiogenesis effectively and attenuated macrophage infiltration, a key factor in promoting angiogenesis [88]. Most importantly, IVIg inhibited neovascularization via the activation of the FcγR1 receptor, a VEGF-independent pathway [88]. As intravenous administration of IVIg effectively suppressed CNV as effectively as intravitreal injections, this could provide an alternative mode of treatment to repeated intravitreal injection of anti-VEGFs in nAMD patients [88].

Inhibitors of Integrins
Integrins are transmembrane proteins that bind to extracellular matrix proteins such as laminin, fibronectin and collagens [89]. Integrins are localized on the surface of RPE and mediate the process of phagocytosis of the outer segment particles of the photoreceptors by RPE [90,91]. Members of the integrin family α5β1, αvβ3, and αvβ5 are expressed during CNV and their antagonists could possibly have a therapeutic role in inhibiting CNV in AMD patients (Figure 1) [92][93][94].
Integrin α5β1 is a fibronectin receptor which is linked to endothelial cell migration and proliferation. ATN-161 is a small peptide inhibitor of Integrin α5β1 [95]. Intravitreal injection of ATN-161 after laser photocoagulation inhibited CNV leakage and neovascularization in rats [95]. Optical coherence tomography and histological examination indicated that treatment with ATN-161 significantly reduced the size of laser-induced lesions in rats [95]. JSM6427 is also an inhibitor of integrin α5β1. Intravitreal injection of JSM6427 significantly attenuated neovascularization in laser-induced CNV in mice [96]. A phase I clinical trial (ClinicalTrials.gov Identifier: NCT00536016) evaluated the pharmacological efficacy and safety of JSM6427 intravitreal injections in 36 patients. The study ended in 2009 with promising results however, to date no further studies have been undertaken to investigate JSM6427.
Volociximab is a monoclonal antibody (Ophthotech Corporation, Princeton, NY, USA) that inhibits the binding of fibronectin to integrin α5β [97]. A phase I clinical trial (ClinicalTrials.gov identifier: NCT00782093) which evaluated the safety of Volociximab as intravitreal injection in combination with ranibizumab, reported positive results [97].
ALG-1001 is a synthetic oligopeptide (Allegro Ophthalmics, San Juan Capistrano, CA, USA) that attenuates α5β1, αvβ3, and αvβ5 integrin mediated blood vessel growth [92]. Phase I and II b clinical trials (ClinicalTrials.gov Identifier: NCT01749891, NCT02348918) reported that ALG-1001 was safe to administer via intravitreal injection, and that it improved the visual acuity of both nAMD and diabetic macular edema patients [92].
Tenascin-C is an extracellular glycoprotein which is mainly expressed in the CNS during developmental stages and its levels are upregulated under inflammatory conditions where it is localized within the CNV membranes of AMD patients ( Figure 1) [98,99]. Tenascin-C promotes retinal neovascularization in proliferative diabetic retinopathy patients [100]. Kobayashi et al. reported that tenascin-C co-localizes with integrin αvβ3 in the CNV membranes of AMD patients and laser-induced CNV mice [101]. Furthermore, tenascin-C promoted CNV in mice by binding to integrin αvβ3. Tenascin-C knock out and tenascin-C messenger RNA (mRNA)-silenced (intravitreal injection of siRNA) mice showed a significant reduction in CNV formation, suggesting that tenascin-C-mediated integrin αvβ3 could be a potential target for nAMD treatment [101].

COX-2 Inhibitors
Cyclooxygenases (COXs) are a group of enzymes that are involved in inflammatory immune responses required for the conversion of arachidonic acid to prostaglandins [102]. Out of the three COX isoforms (COX-1, COX-2, and COX-3), COX-2 mediates inflammation and is induced by pathological stimuli [102]. In mice, COX-2 involvement has been implicated in CNV and subretinal fibrosis of the retina (Figure 1). Subretinal fibrosis was associated with the upregulation of transforming growth factor-β2 (TGF-β2) by COX-2 in AMD ( Figure 1) [103,104]. Zhang et al. reported that the expression of COX-2 in CNV, and inhibition of COX-2 using NS-398 a COX-2-selective antagonist significantly attenuated CNV and subretinal fibrosis via the inhibition of macrophage infiltration, TGF-β2, and VEGF [105].

Inhibition of CCR3
CCR3 (also known as CD193) is a cell surface chemokine receptor that is expressed by eosinophils, Th2 cells, basophils, and mast cells [106]. CCR3 expression by choroidal endothelial cells was recently reported in CNV membranes excised from nAMD patients [107]. Inhibition of CCR3 using intravitreal injection of anti-CCR3 antibody, a small molecular CCR3 antagonist, or by using CCR3 knockout mice significantly attenuated the formation of CNV in mice [107]. Furthermore, a comparison of CCR3 neutralization versus anti-VEGF treatments in mice reported that CCR3 inhibition was superior to anti-VEGF treatment. [107]. On the contrary, results from a recent study that investigated the broad inhibitory outcome of CCR3 resulted in retinal cell death in laser-induced CNV mice [108]. Intravitreal delivery of CCR3 inhibitor in laser-induced CNV mice led to the apoptosis of Muller cells and reduced the survival of retinal cells surrounding CNV lesions [108]. Inhibition of CCR3 can be considered as a potential target for the treatment of nAMD; however, further studies examining the safety of CCR3 inhibitors and targeting of choroidal endothelial cell CCR3 inhibition might provide greater insight.

Targeting Other Signaling Pathways with Involvement of VEGF
Several other pathways and novel agents that involve VEGF signaling were recently implicated in the pathogenesis of nAMD. Such mechanisms/agents include but are not limited to the complement pathway, BMP9/ALK1 signaling, erythropoietin signaling, long non-coding RNAs, STAT3 activation, neuropilin 1, platelet activating factor, mTOR signaling, and Yes-associated protein (YAP) inhibitors. Some of these pathways act in parallel (BMP9/ALK1 signaling), downstream (neuropilin 1) or upstream of VEGF signaling (complement pathway, BMP9/ALK1 signaling, platelet-activating factor receptor, STAT3 activation, TGF-β signaling). Since the primary aim of this review is to focus on the pathways independent to VEGF, other signaling pathways that inhibit VEGF directly or indirectly with possible therapeutic benefits in nAMD are not discussed in detail, but are listed in form of table (Table 1), briefly discussing the involvement of key molecules, signaling pathways, and their mechanisms of action.

Concluding Remarks
The current review highlights a broad range of non-VEGF pathway targets that offer potential new treatments for nAMD. Intravitreal injection of anti-VEGFs has been the standard treatment of nAMD for many years. Despite the success of anti-VEGFs, there is no improvement in vision for one-third of nAMD patients, and the long-term use of anti-VEGF therapy is associated with adverse events such as the development of GA and retinal fibrosis, to name a few. Therefore, a need exists to develop improved strategies that can reduce or eliminate ocular injections and improve clinical outcomes. Recent research has identified many molecular targets other than anti-VEGFs, as well as alternative drug delivery routes, which are currently being tested at the level of clinical trials, offering potential new avenues for treating nAMD. It could also be postulated that employing multiple targeted approaches to treat nAMD could yield better results than single pathway targeting, especially for simultaneous treatment of nAMD and subretinal fibrosis.