Corneal Neovascularisation and Anti-VEGF Therapy

Abstract

Cornea vascularisation is a significant cause of ocular morbidity. Disease or injury often triggers the development of new blood vessels in the cornea, compromising its clarity and impairing vision. Common causes of corneal neovascularisation include infections, chemical burns, and local and systemic inflammatory disorders. Topical corticosteroid eye drops remain the standard therapy; however, extended use of corticosteroids has been known to cause side-effects including cataracts and raised intraocular pressure. As such, an alternative therapy has been actively sought. Vascular endothelial growth factor (VEGF) is a major angiogenic factor implicated in neovascularisation. The success of anti-VEGF agents in managing leaking blood vessels in neovascular age-related macular degeneration provides an opportunity to explore its use in the treatment of corneal neovascularisation. The therapeutic potential of anti-VEGF agents has been evaluated in experimental models of corneal neovascularisation and clinical trials with variable results. Here, we review the study results and discuss the development of new strategies that may improve treatment outcomes for corneal neovascularisation.

1. Introduction

Preventing neovascularisation in the normally transparent cornea is essential to maintain vision [1]. Aberrant neovascularisation in the cornea can be triggered by conditions such as infection, injury, or systemic and local inflammation, and also compromises the survival of a cornea transplant (Figure 1). To manage neovascularisation in corneal diseases, topical application of a broad-spectrum corticosteroid remains the standard therapy, however extended use of steroids has been known to cause side-effects such as cataracts and glaucoma [2]. As such, other approaches including cauterisation, anti-metabolite treatment, and laser are also utilised [3]. The cornea has been known to maintain its transparency by an active process termed corneal angiogenic and lymphangiogenic privilege [4]. This process involves the generation of endogenous angiogenic inhibitors to counteract the production of angiogenic factors. Stimuli such as injury, infections, and diseases can disrupt this balance and consequently promote corneal neovascularisation [5]. As a result, inhibitors targeting the production or activation of the factors involved have been investigated (reviewed by Drzyzga et al. [6]). They include interleukins, matrix metalloproteinases, and transforming growth factor-β (TGF-β). Among these factors, the family of vascular endothelial growth factors (VEGFs) is one of the major angiogenic factors implicated in neovascularisation in the cornea [7] and retina [8].

High expression of VEGF and its receptors have been detected by immunohistochemistry and ELISA in inflamed and vascularised human corneal specimens [9]. In addition, a significantly higher concentration of VEGF is found in tears from patients with corneal neovascularisation compared to controls [10]. The approval of anti-VEGF therapy for the management of choroidal and retinal neovascularisation [11] provides additional therapeutic opportunities for the treatment of corneal neovascularisation, with the hope that targeted anti-VEGF therapy can effectively suppress the growth of blood vessels in diseased corneas or corneal transplants. The therapeutic potential of anti-VEGF agents has been evaluated in pre-clinical animal models of corneal neovascularisation or clinical trials [2], with mixed outcomes. Variations in dosing regimen, route of administration, and results have limited the ability to draw any conclusions on the efficacy of the treatment. Furthermore, its limited efficacy in established neovascularization and its off-label use have limited its uptake in the routine clinical setting [12]. As the crucial role of VEGF in corneal neovascularisation has been comprehensively reviewed by other authors [13,14], in this article we will discuss the results from preclinical and clinical evaluation of anti-VEGF agents and ways to refine treatment efficacy to improve treatment outcomes.

2. Efficacy of Anti-VEGF Agents in the Treatment of Experimental Corneal Neovascularisation

Since the approval of intravitreal anti-VEGF injections for the treatment of choroidal and retinal neovascularisation in 2006, different anti-VEGF formulations have been developed to improve the efficacy [11]. Among these anti-VEGF agents, the anti-angiogenic effects of pegaptanib (Macugen^®), bevacizumab (Avastin^®), ranibizumab (Lucentis^®), aflibercept (Eylea^®), and conbercept (Lumitin^®) have also been evaluated in experimental models of corneal neovascularisation and small clinical trials [14,15]. These anti-VEGF agents differ in terms of the molecular size and specific mechanisms of action. Pegaptanib is an RNA aptamer that binds to the predominant isoform of VEGF-A, VEGF165 [16,17]. Both bevacizumab and ranibizumab are neutralising antibodies of all isoforms of VEGF-A, but ranibizumab has a smaller molecular size (48kDa vs.149kDa) to improve tissue penetration [18]. Aflibercept is a soluble decoy receptor of both VEGF receptor 1 (VEGFR1) and VEGFR2; hence, aflibercept binds to the ligands of these receptors, including VEGF-A, VEGF-B, and placental growth factor, for prolonged effect [19,20]. Conbercept works like aflibercept but it has an additional binding affinity to VEGF-C for a broader anti-VEGF spectrum [21]. As corneal, choroidal, and retinal neovascularisation share a similar causative factor in VEGF [8,22], it is likely that anti-VEGF treatment is also effective in controlling corneal neovascularisation. This idea has been explored in pre-clinical animal models of corneal neovascularisation followed by clinical studies.

In experimental models of corneal neovascularisation, the anti-angiogenic effects of pegaptanib [23], ranibizumab [15,23], bevacizumab [15,23], aflibercept [24], and conbercept [25,26] have been demonstrated following either a single subconjunctival injection or repeated topical administration. It should be noted that the anti-VEGF agents were given prior to the onset of angiogenesis in all of these studies, hence demonstrating their effect in preventing the formation of new blood vessels. Conversely, in the clinical setting, patients often present following the development of neovascularisation [2], so the inhibitory effects of anti-VEGF agents on established blood vessels have also been evaluated in experimental models of corneal neovascularisation. Unfortunately, the study outcomes are not consistent and may be affected by different routes of administration (subconjunctival versus topical) and the method of inducing corneal neovascularisation (e.g., chemical injury using silver nitrate versus an alkali burn). Yoeruek et al. [27] demonstrated that daily topical bevacizumab 25 mg/mL solution given immediately after an alkali corneal burn markedly reduced vessel growth; however, it did not cause any anti-angiogenic effect in already vascularised corneas. In contrast, Hashemian et al. [28] reported that a single subconjunctival injection of bevacizumab (1, 5 or 25 mg/mL) when given at either day 0 or 7 post-injury prevented and reduced angiogenesis in silver nitrate-cauterised corneas, and the anti-angiogenic effect was superior to the corticosteroid betamethasone 6 mg/mL. Silver nitrate cauterisation is known to induce a superficial injury on the epithelium while an alkali burn tends to cause a deeper injury to the stroma [12]. In addition, established neovascularisation in rabbit eyes exposed to a different alkylating agent, namely sulfur mustard, responded to two subconjunctival injections of bevacizumab 5 mg given twice a week over a 3-week period [29]. Furthermore, a single injection of ziv-aflibercept 2 mg had a longer anti-angiogenic effect than the two doses of bevacizumab in the same experimental model [29]. As such, there seem to be discrepancies in the reported anti-angiogenic activities among research groups. Due to the growing number of pre-clinical research on bevacizumab, Papathanassiou et al. [30] performed a meta-analysis on eighteen publications dated from 2007 to 2011 on experimental models in mice, rats, and rabbits to identify an overall trend in the anti-angiogenic activity. The analysis covered a wide spectrum of pre-clinical animal models of corneal neovascularisation triggered by chemical eye burns, corneal transplantation, suture-induced injury, or limbal injury [30]. The authors concluded that bevacizumab given either in eye drops or via injection (i.e., intravitreal, subconjunctival, or intracameral) effectively causes a reduction in neovascularisation [30], but the effective dosing regime has not been established.

3. Efficacy of Anti-VEGF Agents in Clinical Studies

The off-label use of bevacizumab is the most investigated anti-VEGF agent in clinical studies. Early clinical studies trialled the anti-angiogenic effect of anti-VEGF agent in patients showing no or minimal response to prior steroid therapy [31]. Among these trials, DeStafeno and Kim [31] assessed topical bevacizumab in two clinical cases with superficial and deep stromal neovascularisation that were unresponsive to topical steroids after several months. Both patients were given bevacizumab (1%, four times a day) for twenty-five days and the affected eyes displayed a significant reduction in corneal neovascularisation with no adverse ocular events [31]. Since no final visual acuity was reported by DeStafeno and Kim [31], it is unclear if the reduction in neovascularisation improved vision. It is worth noting that the effect of prior topical steroid therapy would have been negligible as the two patients had a 6-month washout prior to anti-VEGF drops [31]. In contrast, Kim et al. [32] trialled topical bevacizumab (1.25%, twice a day) in ten eyes from seven patients where some cases remained on concurrent topical steroid treatment. Although bevacizumab effectively reduced neovascularisation in seven of the ten treated eyes within the first month, adverse effects appeared from the second month, including stromal thinning in one treated eye and an epithelial defect and erosions in six treated eyes [32]. The authors speculated that the adverse events may be attributable to a disruption in VEGF-dependent wound healing [32]. These adverse events have also been reported in isolated cases from several preclinical animal studies [30].

With the increasing number of clinical studies assessing the efficacy of anti-VEGF agents, the variations in data reporting and dosing regimen make it difficult to draw conclusions on their efficacy and safety. For instance, Papathanassiou et al. selected seven out of thirteen published clinical studies over the period of 2007–2010 to perform a meta-analysis based on the following exclusion criteria: missing variance or quantification of neovascularisation even though eye photos were provided [30]. From the meta-analysis outcome, bevacizumab effectively controlled progressive corneal neovascularisation associated with ocular disorders such as pterygium, lipid keratopathy, failed transplant, inflammation, and infection [30,33]. However, there was no consensus on treatment regime in the dose or route of administration. Between 2009 and 2022, several randomised controlled trials and non-randomised trials were conducted to assess topical or subconjunctival treatment of bevacizumab, ranibizumab, or aflibercept on corneal neovascularisation in patients undergoing high-risk corneal transplant [33]. By performing a meta-analysis on these studies, Lai et al. [33] concluded that anti-VEGF agents cause a reduction in corneal neovascularisation and are well tolerated. However, it is still uncertain whether anti-VEGF treatment improves visual acuity [34] or survival in cornea transplantation [35,36]. Recently, Sun et al. demonstrated that a single subconjunctival injection of conbercept (1 mg) suppressed existing neovascularisation in ten patients with chemical burns, blepharokeratoconjunctivitis, neurotrophic keratopathy, and viral infections, without adverse effects on visual acuity or safety issues at one month post-injection [37]. Overall, the findings from pre-clinical and clinical studies support an anti-angiogenic effect of anti-VEGF (primarily bevacizumab) agents. More robust data from larger clinical studies would ascertain the benefit of anti-VEGF agents on improving affected visual acuity or success rate following corneal transplantation.

4. Developing New Strategies to Overcome the Challenges and Limitations of Anti-VEGF Treatment

It is well known that VEGF is a crucial growth factor in maintaining normal ocular function. For example, ablation of VEGF signalling in retinal pigment epithelial cells induces a loss of the adjacent choriocapillaris and death of cone photoreceptors which are required for central and colour vision [38]. Furthermore, gene deletion of VEGF-A or its associated receptors (VEGFR1 and VEGFR2) is lethal in mouse embryos due to multiple defects in the formation of vascular structures [39]. VEGF is a crucial growth factor that is required not only during ocular development but also in maintaining mature and healthy blood vessels. Thus, inhibiting VEGF signalling could have detrimental effects. Although the cornea is an avascular structure, a healthy limbal vascular plexus is needed to maintain the integrity of the cornea. As such, applications of anti-VEGF agents for controlling corneal neovascularisation should proceed with caution.

4.1. Gene-Based Approach to Long-Term Inhibition of VEGF

Clustered Regularly Interspaced Short Palindromic Repeat (CRISPR) gene editing technology is made up of two components: a guide RNA and the protein nuclease Cas 9. The guide RNA-Cas9 complex can be designed to target specific DNA sequences, and modified to perform various tasks such as cutting DNA, modifying single bases (base editing) or sections of DNA (Prime editing), or modifying the transcription activity at the site (e.g., CRISPRa transcription activator; CRISPRi transcription inhibitor). Using various tools in the CRISPR toolbox, VEGF-A expression can be modified by a single subconjunctival injection of an adeno-associated virus or AAV-SpCas9/sgRNA-VEGFA system that causes a complete suppression of VEGF-A protein and neovascularisation in mouse sutured corneas [40]. Interestingly, the complete inhibition of VEGF-A has not caused any adverse effects on ocular function measured by electroretinography or eye morphology [40], supporting the therapeutic use of a CRISPA/Cas9 system in suppressing VEGF-A expression.

Another gene-based approach is to use a recombinant AAV vector as a vehicle to deliver a transgene to mediate long-term gene expression either in vitro or in vivo. Su et al. [41] developed recombinant AAV8-KH902 (rAAV-KH902) to mediate gene expression of KH902 (conbercept). A single intrastromal injection of rAAV8-KH902 was found to sustain the gene expression of KH902 for three months, and was able to prevent the growth of existing corneal neovascularisation from alkali burns or suture injury [41]. The prolonged suppression of the VEGF-A gene also translated to extended anti-angiogenic activity on corneal neovascularisation in comparison to a single injection of conbercept (i.e., 12 weeks versus 2 weeks) [41]. Gene therapy may be an effective strategy to prolong the efficacy of anti-VEGF therapy and can be safely achieved by the available gene manipulation technology [40,42].

4.2. Sustained-Release Formulation for Progressive Inhibition of VEGF

The development of sustained-release formulations aims to induce a progressive inhibition of VEGF. This approach not only addresses the safety concerns, but it would also prolong the efficacy of anti-VEGF. This can be achieved with the use of biomaterials such as nanoparticles. Biomaterials have become an important tool to facilitate the progressive release of pharmacological or molecular agents in the different compartments of the eye [43]. Moreover, multiple agents can be packaged into a single nanoparticle material to target multiple signalling pathways, with the hope to improve efficacy and minimise adverse effects.

4.3. Combined Therapy or Multiple Targeting to Improve and Prolong Efficacy

Since inflammation is also a major driver of corneal neovascularisation, targeting two separate pathways simultaneously would be expected to produce superior efficacy. As a result, several recent pre-clinical studies have synthesised a dual inhibitor by encapsulating both bevacizumab and the corticosteroid dexamethasone in a hydrogel [44,45]. The study outcome confirmed that combining bevacizumab with dexamethasone significantly improved the efficacy compared with a single bevacizumab treatment in an alkali burn model in rats [44,45]. Shi et al. [46] also demonstrated the synergistic effect of dual inhibition by combining bevacizumab with the non-steroidal anti-inflammatory agent diclofenac in a rabbit model of corneal neovascularisation. Similarly, Zhou et al. [47] incorporated bevacizumab with infliximab in a thermosensitive polymer solution for subconjunctival injection. Infliximab is a monoclonal antibody to the inflammatory cytokine tumour necrosis factor-α (TNF-α) and infliximab eye drops have been shown to reduce corneal neovascularisation in alkali-burnt rabbit corneas [48]. Combined bevacizumab/infliximab therapy produced a significantly greater reduction in both neovascularisation and inflammation than anti-VEGF monotherapy in rabbits following alkali burn injury [47]. The level of antibodies was also sustained for twelve weeks after a single injection of bevacizumab/infliximab gel [47].

Another approach is to target the signalling effector that is common to the growth factors implicated in corneal neovascularisation, for example, VEGF and platelet-derived growth factor (PDGF) [49]. Tyrosine kinases are the signalling receptors for both growth factors. Sunitinib is a tyrosine kinase inhibitor that targets both VEGF and PDGF [50] and is currently approved for cancer therapy [51]. Sunitinib has been trialled in a preclinical animal model with suture-induced corneal neovascularisation and the inhibitory effect of sunitinib eye drops was found to be superior to topical bevacizumab [52,53]. By incorporating sunitinib into a sustained-release formulation with poly (D,L-lactic-co-glycolic acid) (PLGA) microspheres, the inhibitory effect was more effective than sunitinib alone in the suture-induced corneal vascularisation model, with a sustained release of the drug for 30 days in vitro [54].

Faricimab is a bispecific antibody that neutralises the activity of both angiopoietin-2 and VEGF-A [55]. Recent phase 3 clinical studies in neovascular age-related macular degeneration have demonstrated the effectiveness of faricimab in maintaining best-corrected vision acuity when given at an interval of 16 weeks, with results comparable to an 8-weekly injection of the single anti-VEGF agent aflibercept [55]. Since angiopoietin 2 has also been implicated in corneal neovascularisation [56], it remains to be investigated if faricimab could be effective in limiting corneal neovascularisation. The success of combined therapy in experimental models thus opens up the opportunity of creating more efficacious and safer combined therapeutics with biomaterials for the treatment of corneal neovascularisation.

4.4. Simultaneous Targeting of Oxidative Stress and VEGF

Adding antioxidant treatment to anti-VEGF therapy should have synergistic efficacy for multiple reasons. Oxidative stress is caused by an overproduction of reactive oxygen species (ROS) and is a causative factor in many pathological conditions such as cancer, cardiovascular diseases [57], and ocular disease [58]. It is known that ROS products such as superoxide and hydrogen peroxide can amplify the responses of VEGF. VEGF increases the generation of ROS by activating the ROS-inducing enzyme NADPH oxidase (NOX), and, in turn, NOX can trigger VEGF signalling via up-regulation of VEGF expression and activation of its receptor. As such, NOX activation amplifies the VEGF response through a positive feedback loop. Deleting the sources of ROS can limit corneal neovascularisation in mice with alkali burn injury [59,60,61,62]. Furthermore, mice that are deficient in NOX isoforms tend to show less retinal vascularisation in comparison to wild-type mice when they are subjected to physiological stressors such as hypoxia [63]. Inhibition of ROS can be achieved by antioxidants and small molecules targeting ROS-inducing enzymes. Using nanotechnology, Zhu et al. [64] bioengineered an eye drop formulation containing an antioxidant epigallocatechin gallate (EGCG) and an siRNA targeting VEGF, and the combined formulation demonstrated inhibitory effects in in vitro angiogenic assays and a pre-clinical animal model of suture-induced corneal neovascularisation. Since there is no comparison to monotherapy with either ECCG or VEGF-siRNA, it is uncertain if the combined therapy has a superior effect; however, the study outcome highlights the feasibility of conjugating antioxidants to a gene-targeting molecule using nanotechnology.

ECCG belongs to a large class of plant-derived compounds or flavonoids that have health benefits such as anti-inflammatory, anti-microbial and antioxidant effects. The use of flavonoids in combined therapy with anti-VEGF would conceivably produce favourable outcomes. There is evidence supporting the benefits of flavonoids in ocular diseases; for example, flavonoid treatment with an intravitreal injection of quercetin has been shown to prevent apoptosis of retinal ganglion cells in an experimental model of glaucoma [65]. Flavonoids are likely to synergise the inhibitory effects that anti-VEGF has on corneal neovascularisation since they have been shown to inhibit VEGF/basic fibroblast growth factor (bFGF)-induced angiogenesis [66] and transforming growth factor (TGF-β)-induced fibrosis [67,68]. Moreover, a meta-analysis has also shown that flavonoid supplementation can minimise toxicity associated with nanomaterials [69]. As such, flavonoids are ideal candidates for incorporating with anti-VEGF agents. Figure 2 summarises the targets of current anti-VEGF agents and how advanced technology would improve and prolong the inhibition of VEGF responses.

5. Future Directions and Conclusions

The findings from the two meta-analyses, which included over 20 clinical studies, demonstrate that a single anti-VEGF treatment for corneal neovascularisation is relatively efficacious and safe [30,33]. However, many of these studies used placebo treatment in the control group, whereas in clinical practice, topical corticosteroid treatment is the mainstay in treating corneal neovascularisation. Therefore, studies comparing anti-VEGF treatment and the current standard of care is required. Studies should also address the treatment of both new and established corneal vascularisation. In addition, safety and pharmacokinetic studies should be performed to elucidate if the newly developed sustained-release formulations, combined therapy, and gene-based therapy have any adverse effects, in particular any effects of potential systemic exposure of VEGF from sustained-release formulations.

Since the causes of corneal neovascularisation cover a wide spectrum including infections, injury, and local and systemic inflammatory disorders [70], a single therapy or a generalised dosing regimen may also not be suitable, and treatment of new versus established neovacularisation may differ. However, a targeted approach would only be possible if patients presented during the early phase of the disease and if ophthalmologists have reliable historical data and tools to design an individualised treatment plan, but this is not always feasible. Recently, a deep learning-based model has been developed to provide an efficient method to analysis slit-lamp photographs [71]. This artificial intelligence (AI) technology would be useful for building up patient databases to better understand historical data and the outcomes of treatment plans, which would help in the advancement of future treatment strategies.