Age-related macular degeneration is a complex, multifactorial disease with aging, environmental/lifestyle, and genetic/epigenetic factors playing a role in its pathogenesis [17]. The complexity of this disease is also underlined in that each of these elements has several variants that may interplay both within each group of factors and with factors from a different group. For instance, AMD is not a monogenic disease, and as shown in genome-wide association studies, several genetic loci may be involved in its pathogenesis, containing genes of the following three main pathways: the complement pathway, lipid metabolism, and extracellular matrix remodeling. Variants of these genes may interact and their effect can be modulated by environmental/lifestyle AMD risk factors, including an unhealthy diet [18]. The most consistently reported loci that are associated with AMD are the rs1061170 (Tyr402His/p.Y402H) single nucleotide polymorphism variant in the complement factor H (CHF) and the age-related maculopathy susceptibility 2 and high-temperature requirement A serine peptidase 1 (ARMS2/HTRA1) [19,20]. Therefore, many mechanisms may be involved in AMD pathogenesis which, along with the limited possibility of studying live human eyes, results in limited treatment options for AMD.

Advanced AMD happens in two clinically distinct forms, dry and wet, and individuals affected by either form may suffer from gaps or dark spots in their vision (Figure 1). Although wet AMD is responsible for a minority of all AMD cases, it accounts for about 90% of sight loss related to AMD and, somewhat paradoxically, only wet AMD is treatable [21]. The causal relationship, if any, between these two forms of AMD is not clear, and some studies suggest that they might be considered as two distinct diseases [22]. We showed that wet AMD might correlate with mortality in a 12-year prospective case–control study, but the question of whether wet AMD might be an independent risk factor for death is still open [23].

Drusen, small, light, solid objects containing proteins and lipids, are present in both forms of AMD but their presence in dry AMD is much more common than in wet AMD (approximately 90:10%). They can be divided into hard and soft drusen [24]. Dry AMD is characterized by a gradual deterioration of the retina following from the death of retinal cells that are not renewed [25]. Drusen and edema may be visible in a microscopic picture of the affected eye.

Choroidal neovascularization (CVN), presenting the growth of new blood vessels derived from the choroid through a break in the Bruch’s membrane, is a hallmark of wet AMD [26]. In wet AMD the retinal pigment epithelium (RPE) is not broken, as it is in dry AMD, but RPE may be detached [27]. Furthermore, the functionality of the RPE cells is changed as they overproduce VEGFA and other angiogenic factors that are essential in CNV. Fresh blood vessels formed in CNV are fragile and release their content into the retina, resulting in fibrosis, with subsequent formation of disciform scar that can result in sight loss unless treated [28]. The angiogenic effect of VEGFA in CNV is mainly mediated by the PLCγ-PKC-MAPK pathway initiated from the 1175-PY site on VEGFA receptor 2 (VEGFR-2) located on the exterior of endothelium [29].

At present, the treatment of patients with antibodies against VEGFA and its receptor is the only accepted therapy in wet AMD, whose efficacy has been consistently confirmed and has become a routine procedure in ophthalmic clinics [30]. The anti-VEGFA treatment significantly improves the outcomes of wet AMD, although cases of improving visual acuity are rare; the treatment stops the progression of the disease and ameliorates damage in the retina [3,31]. However, the treatment per se is troublesome, as it involves repeated intravitreal injections, every 8 to 12 weeks after three-monthly loading doses, with a relatively expensive formulation of monoclonal antibodies against VEGFA and its receptor.

Many formulations of VEGFA antibodies are present in the market, including the following that are FDA approved: aflibercept (Eylea), brolucizumab (Beovu), ranibizumab (Lucentis), and faricimab-svoa (Vabysmo); bevacizumab (Avastin) is used off-label as it is primarily designated as an anti-cancer drug [32]. Also, biosimilars to anti-VEGF have been recently approved by FDA (https://www.reviewofophthalmology.com/article/an-update-on-the-antivegf-biosimilar-pipeline; accessed on 15 January 2024).

In summary, the introduction of anti-VEGFA treatment revolutionized the therapy for wet AMD. Although the treatment does not cure the disease, it may protect against sight loss and is in general safe for repeated applications over a long period of time. However, the necessity of lifetime intravitreal injections and the relatively high cost of anti-VEGF drugs are a serious burden for patients.

The eye seems to be predisposed as a target for gene therapy due to its relatively small size, with a compartmentalized structure and immune-privileged status [16]. These eye characteristics lower the risk of systemic exposure. Moreover, advanced non-invasive methods of imaging in the eye, including optical coherence tomography, fundoscopy, angiography, and two-photon microscopy, assist in the real-time monitoring of the progress of the gene therapy procedures and their safety [33]. Another advantageous feature of the eye for gene therapy is that changes in a single gene may be associated with various clinical states. For example, homozygous mutation in the “historic” RPE65 gene may result in either Leber’s congenital amaurosis 2 or rare forms of retinitis pigmentosa (RP) [34].

The development of gene therapy has brought a better understanding of the biology of viral vectors, as they are basic vectors used in eye gene therapy [35]. In general, virus vectors can be divided into integrating and non-integrating with the host genome. Due to safety concerns, non-integrating viral vectors may be the current strategies and those used in at least the near future for transgene delivery in gene therapy [36].

Among many non-integrating viral vector types, adeno-associated viral (AAV) vectors have many features of key significance for eye gene therapy [37].

Adeno-associated viruses consist of two parts, an icosahedral protein capsid and a single-stranded DNA (ssDNA) genome [38,39,40]. The AAV genome has 4.8 kb of ssDNA and two T-shaped inverted terminal repeats (ITRs), on either end of the genome [41] (Figure 2). The terminal repeats flank two open reading frames, Rep and Cap, that are transcribed and translated to produce the virus life cycle proteins Rep78, Rep68, Rep52, Rep40, and VP1, VP2, and VP3 capsid proteins resulting from the use of alternate promoters and alternate splicing.

In the transfer plasmid construct, which is to be packaged by the AAV virus, the transgene is placed between ITRs, and Rep and Cap are attached in trans (Figure 2).

AAV presents various serotypes that differ in the capsid structure, immunogenicity, and cellular tropism; the most suitable serotypes can be chosen to accommodate the eye environment [42]. The parental adeno-associated virus requires a helper virus to replicate to exert its pathogenic action in humans [43]. Moreover, AAV vectors were shown to elicit a limited immune response due to the restricted ability of the parent virus to infect antigen-presenting cells [44].

The virus needs the Rep protein to replicate and the open reading frame for encoding this protein can be removed by gene editing, depriving the virus of the ability to replicate and integrate [43]. If this open reading frame is not edited, the virus integrates in a site-specific fashion with the human genome on chromosome 19 [45]. Therefore, the AAV vector offers the possibility to integrate or not with the host genome by the manipulation with the Rep locus.

The AAV vectors have some limitations [46]. They cannot package fragment DNA longer than 4.5 kb, their long-term expression is limited to non-dividing cells and, despite their limited ability to elicit immune response, there is a high incidence of pre-existing immunity in humans [47].

If the Rep locus is retained, one may expect a long-term, controlled expression of target DNA in chromosome 19 and accumulation of the product of that expression. Therefore, if the target DNA could encode a product that would stimulate the eye to continuously produce an antagonist of VEGFA, this would replace multiple injections with anti-VEGFA antibodies. This is the general concept of the exploitation of AAV vectors in gene therapy of wet AMD. This is a definite change of the general strategy of gene therapy which, in its standard form, aims to replace a faulty or absent gene with a gene that displays normal expression. Here, nothing is replaced, but the strategy with the AAV vector carries an added value for target cells. Moreover, the recent strategies of gene therapies with the AAV vectors have created the possibility of therapeutic gene manipulations with minimal concerns for inflammatory and immune reactions and with maximal efficacy [16].

The above idea is just an idea, and many details need to be elaborated before its accomplishment. The main problems to solve are the delivery route and sustained expression of the VEGFA antagonist.

In general, three routes of the delivery of a transgene are considered in ocular gene therapy delivery methods (Figure 3). The delivery to the suprachoroidal and subretinal spaces is characterized by targeted access and wide transduction of the retinal cells, low exposure to the vitreous and anterior segments, and compartmentalized delivery of the AAV vector as observed in preclinical studies [48,49,50]. Subretinal delivery poses a low risk of immune response and inflammation [51]. Intravitreal injections are characterized by limited transduction of retinal cells, a wide exposure of the vitreous and anterior segments, and a high risk of immune/inflammatory reactions [52,53].

Subretinal introduction of the RPE65 gene in patients with Leber’s congenital dystrophy was a fundamental piece of work for gene therapy for AMD. The normal RPE65 gene encodes the RPE65 protein that transforms all-trans-retinol into 11-cis-retinol in the retinal cycle, a reaction that is crucial for the conversion of visual information into electric pulses [54]. Transduction of RPE cells with the transgene results in the production of the functional RPE65 protein. Therefore, experiments showed that AAV vectors may be used to provide a stable expression of the target DNA in the retina, opening a perspective for the “one-and-done” in-office treatment of retinal diseases requiring a continuous production of a therapeutic protein. However, the question remains whether a subretinal injection is an optimal route of administration of the target DNA.

It appears likely that the past, present, and future of gene therapy in the eye is not limited to the use of AAV-based vectors as delivery systems, and the use of several other types of viruses, as well as non-viral vectors, has been attempted. More detailed analysis can be found elsewhere, e.g., in the review of Rodrigues et al. [47].