The administration route of gene therapy vectors could be the difference between a non-targeted and a targeted delivery. For eye diseases several routes can be used ranging from systemic or topical administration to subconjunctival, intravitreal and subretinal injections (Figure 1). Depending on the disease and the affected ocular tissue, a different delivery route can be chosen to allow a more precise delivery of the genetic therapy.

The topical administration route is used mostly in the form of eye droplets to treat diseases of the anterior section of the eye as pre-corneal factors (solution drainage, blinking, among others) and anatomical barriers prevent the accessibility of the carriers to the posterior section of the eye (Gaudana et al., 2010). Delivery of SLN-based vectors to treat corneal inflammation have been achieved in mice; however, translation to the human eye may be hindered by the thickness of the human cornea (Vicente-Pascual et al., 2020; Amador et al., 2022).

In the last decade, intravitreal injection became one of the most used techniques to deliver therapeutics to the retina in clinical practice. It can be used to target the inner retina, optic nerve, and lens, however it has some drawbacks, as it can cause retinal detachment, haemorrhage, cataracts, and endophthalmitis (Gaudana et al., 2010). Notwithstanding, the injection of anti-VEGF (vascular endothelial growth factor) agents is the standard approach for treating retinal diseases such as diabetic retinopathy (DR) and age-related macular degeneration (Lai et al., 2015).

Subretinal injection is a very specific delivery technique that can be used to target cell types unreachable by any other delivery method, such as RPE and photoreceptors (PR) cells (Venkatesh et al., 2013; Fayyaz et al., 2020). Similar to other injection methods, it has some limitations: only a small volume can be injected, and there is an associated risk of haemorrhage and retinal detachment (Pierce and Bennett, 2015).

Electroporation is a physical method that uses short electric pulses to direct and deliver genetic material to tissues in a reversible manner by temporarily increasing the cell membrane permeability. This technique has been successfully used in neonatal and adult mice to deliver naked plasmids and non-viral vectors (Melo and Blackshaw, 2011; Nickerson et al., 2012). In combination with the different injection routes, specific cells such as photoreceptors, RPE and retinal ganglion cells can be targeted. For instance, by combining a subretinal injection of genetic material with electroporation, photoreceptors or RPE can be targeted, depending on the direction in which the electric pulses are applied (Venkatesh et al., 2013). Although a powerful tool in animal models, electroporation has a limited use in human clinical practice. In the posterior segment of the eye, specifically the retina, electroporation requires a more invasive surgery for placement of the electrodes, as proximity to the target cells is needed to produce a localized and effective electric field. This creates a challenge to translate electroporation from preclinical studies in rodents to human clinical practice. That being said, there is an active clinical trial (NCT03308045) to assess safety and tolerability of electroporation in establishing the ciliary muscle as a biofactory of therapeutic proteins, in which non-infectious uveitis (NIU) patients are being electroporated in the ciliary muscle with a plasmid encoding a chimeric TNF-α soluble receptor (hTNFR-Is) (Bloquel et al., 2006).

One of the molecular strategies used in vector design to increase protein production and control the expression of therapeutic genes by plasmid DNA is the incorporation of specific promoters into the expression cassette. Promoters are genetic regions involved in transcription initiation and some are only active in specific cell types (Zheng and Baum, 2008). These DNA regulatory elements could be used to regulate the level of transgene expression and restrict it to a specific tissue, thus reducing off-target effects. Furthermore, several regulatory DNA elements (enhancers, silencers or insulators) can be used in addition to promoters to increase cell-type specificity (Nicoud et al., 2007; Toscano et al., 2011). In initial stages of gene therapy, transgenes were under the control of viral promoters such as cytomegalovirus (CMV) or simian virus 40 (SV40), which have demonstrated capacity to promote gene expression (Boshart et al., 1985; Schmidt et al., 1990; Makrides, 2003). However, in recent years, mammalian promoters have emerged, especially in cancer gene therapy, to direct therapeutic genes towards target cells (Yoon et al., 2018; Rama Ballesteros et al., 2020; Yan et al., 2020).

In the ophthalmic field, some groups have reported promising results using mammalian promoters to increase cell-specific gene expression, as summarized in Table 1. These promoters are usually evaluated and selected using reporter genes such as green fluorescent protein (GFP) or luciferase (Kachi et al., 2006; Wang et al., 2020) and ideally, the chosen promoters, should meet two criteria: they must regulate an endogenous gene that is expressed exclusively in target cells, and must promote robust expression of the reporter gene in those cells (Aguirre, 2017).

Much of the research in this area focuses on targeting photoreceptors, as a vast number of human retinal degenerations are primarily associated with dysfunction and death of these cells. Regarding photoreceptor-targeted gene expression, opsin promoters have been extensively studied. The rhodopsin promoter has been shown to drive strong gene expression with viral (Flannery et al., 1997; Matsuda and Cepko, 2004; Nicoud et al., 2007) and non-viral (Zack et al., 1991; Apaolaza et al., 2016; Wang et al., 2016) vectors in photoreceptors, when administered both by subretinal and intravitreal administration. It was also reported that this promoter was able to direct the expression of genes carried by PEGylated liposome-based nanoparticles specifically to the retina after intravenous administration (Zhang et al., 2003). In addition, promoters of other genes such as rhodopsin kinase (GRK1), RP1 axonemal microtubule associated (RP1) and cGMP phosphodiesterase (PDE) have been demonstrated to induce expression of reporter genes in photoreceptor cells (Nicoud et al., 2007; Beltran et al., 2017). The use of other opsin promoters to direct expression to photoreceptors using non-viral vectors was also reported by Apaolaza et al. The authors used a solid lipid nanoparticle carrying the retinoschisin (RS1) gene under the control of a murine opsin promoter and administered it by intravitreal administration in Rs1-deficient mice. Interestingly, the expression driven by the opsin promoter was confined to photoreceptors and the improvement observed in the retina was higher using this promoter than the ubiquitous CMV promoter, demonstrating the relevance of a cell-specific strategy (Apaolaza et al., 2016).

In some retinal pathologies, an even more specific targeting to limit the transgene expression to just one type of photoreceptors is required. For example, in achromatopsia it is important to achieve cone-specific expression (Ye et al., 2016) and some studies have shown this possibility by using a cone opsin promoter (Wang et al., 2016; Ye et al., 2016; Khabou et al., 2018). When studying cell-specific promoters it is important to be cautious when extrapolating and applying results obtained in healthy wild-type retinas to mutants and between different species. The expression level of a regulated gene in normal animals may be altered in mutants (Aguirre et al., 2014). Aguirre et al. and Beltran et al for example, have shown that even though the GRK1 promoter directed gene expression exclusively to rods in normal canine retinas, in mutated ones this vector drove expression also in cones (Beltran et al., 2010; Aguirre, 2017). The variability of promoter specificity between species was demonstrated by Ye et al. when they tried to express transgenes carried by AAV vectors and regulated by cone-specific promoters in different animal models. Indeed, they showed that the IRBP/GNAT2 promoter induced a strong expression restricted to photoreceptors in mice and canines, while in primate cones it was ineffective (Ye et al., 2016). A possible explanation for this divergence is that the DNA sequence of the promoter may not have some regulatory elements needed for expression in primate cones.

The implication of the retinal pigment epithelium (RPE) genetic defects in several retinal degenerative diseases, such as retinitis pigmentosa, age-related macular degeneration or Stargardt disease suggest that RPE could be an interesting target to gene therapies (Dougherty et al., 2008; Sparrow et al., 2010). Thus, in recent years, many promoters have been evaluated for their ability to drive specific expression in these cells. Kachi et al. studied two mammalian promoters of RPE cells, a promoter fragment of the vitelliform macular dystrophy 2 gene (VMD2) and another of Rpe65 gene and verified an increased and sustained expression for at least 56 and 112 days, respectively. This suggests that in addition to achieving cell-specific expression, it is also possible to obtain sustained transgene expression with mammalian promoters (Kachi et al., 2006). Interestingly, VMD2 and Rpe65 promoters have been the preferential promoters used to reach RPE specific expression; despite the promising results with the Rpe65 promoter (Kachi et al., 2009, 2006), VMD2 was shown to have higher specificity (Wang et al., 2019). This promoter drove RPE specific transduction of the reporter gene in either viral (Kachi et al., 2009; Wang et al., 2019; Wu et al., 2021) or non-viral (Esumi et al., 2004; Kachi et al., 2006; Wang et al., 2016) vectors.

In the retina field, most retinal degenerations initially affect photoreceptors and underlying RPE, and thus many therapeutic studies have been focused on targeting these cells; however, there are retinal disorders involving retinal ganglion cells, Müller cells, and bipolar cells that could benefit from targeted gene therapies as well (Hellström and Harvey, 2011; Devoldere et al., 2019a; Hulliger et al., 2020).

Retinal ganglion cells (RGCs) are substantially affected in patients with glaucoma, and their degeneration has also been reported in other diseases such as diabetic retinopathy, retinitis pigmentosa, or even neurodegenerative diseases. Intravitreal injection seems to be the most suitable administration route when targeting RGCs, with Wang et al. testing liposome-protamine-DNA (LPD) nanoparticles to deliver genes to RGCs through this administration method. They showed that it is possible to achieve RGC-specific expression of a reporter gene under the control of the thymocyte antigen promoter (Thy-1.2) (Wang et al., 2016). Other promoters from the gamma-synuclein (SNCG) and neurofilament heavy chain (Nefh) genes were tested, and both promoters, when conveyed in AAV vectors, were shown to successfully induce gene expression limited to retinal ganglion cells (Wang et al., 2016; Chaffiol et al., 2017; Hanlon et al., 2017).

Currently, there is limited research on the specific targeting of Müller cells; however, the use of glial promoters has been shown to be a promising approach. Greenberg et al., for example, have shown that lentiviral vectors with the Cluster of Differentiation 44 (CD44), glial fibrillary acidic protein (GFAP) or vimentin promoters, led to targeted expression in Müller cells after subretinal injection, with CD44 achieving the highest expression level (Greenberg et al., 2007). The GFAP promoter has been the most studied in this context and several other groups have verified its effectiveness (Kuzmanovic et al., 2003; Dorrell et al., 2009; Aartsen et al., 2010; Prentice et al., 2011). In prior studies the Mlc1 (megalencephalic leukoencephalopathy with the subcortical cysts 1) and RLBP1 (Retinaldehyde-binding protein 1) promoters have also shown to be of interest as specific tools to obtain targeted expression on Müller cells (Pellissier et al., 2014; Danjo et al., 2022). Furthermore, encouraging results were obtained using a construct carrying the cellular retinaldehyde binding protein (CRALBP) promoter upstream of the gene that was introduced by subretinal injection and electroporation into the retina (Matsuda and Cepko, 2004).

Since much of the early-stage processing of visual inputs occurs in retinal bipolar cells, their genetic modifications may be of value in various ophthalmic conditions. To reach these cells located in the inner nuclear layer of the retina, the promoter of glutamate metabotropic receptor 6 (Grm6), a gene expressed in these cells, has been studied. This promoter sequence was cloned into lentiviral vectors and shown to direct expression to bipolar cells when subretinal (Cronin et al., 2014) and intravitreal (Lu et al., 2016) administration were used. Cepko’s group has also obtained encouraging results in these cells by using a non-viral strategy combined with the calcium binding protein 5 (CABP5) promoter (Matsuda and Cepko, 2004).

Most recently, biotechnological advances created synthetic promoters capable of spatial and temporal control of transgene expression. This new approach has been used to enhance specificity for retinal cells. Jüttner et al., for example, have developed a library of 230 AAVs, each with a different synthetic promoter and showed that 14% of those drove successful targeting in mice (Jüttner et al., 2019). In their collection of promoters, it is possible to find some capable of targeting human cones (ProA7) or ganglion cells (ProC29), for example, which could be used to express optogenetic tools (Jüttner et al., 2019). For RPE cells, Johari et al presented a synthetic promoter that exhibited higher expression levels than VMD2 gene promoter in vitro (Johari et al., 2021).

Patients with retinal diseases would benefit greatly from the implementation of cell-specific promoters in gene therapy. The combination between different strategies could be relevant for developing and translating retinal gene therapies into clinics. In theory, if the use of a specific promoter is coupled with one of the previously mentioned targeting strategies, the overall efficacy of the therapy could be improved while reducing undesirable gene expression.

To increase the targeting capability of a non-viral vector, a combined strategy may be required to obtain an efficient and precise gene delivery to the retina. This can be achieved by combining a specific promoter or a targeting peptide with an adequate delivery route for non-viral vectors. Rajala et al. modified liposome-protamine-DNA (LPD) polyplexes with nuclear localization signal (NLS) and trans-activator of transcription (TAT) peptides, in combination with the CMV promoter and subretinal administration (Rajala et al., 2014). The authors have shown that when injected with these modified LPD polyplexes the RPE layer of BALB/c mice showed a strong green fluorescent protein (GFP) expression which was not observed in other retinal layers, thus confirming that RPE can be targeted. In another study, using a mouse model of retinal disease—the RPE65 KO mice—the expression of RPE65 LPD polyplexes-derived was able to reduce cone cell death and partially recover the function of photoreceptors, as observed in the electroretinogram (ERG) analysis (Rajala et al., 2014). In a follow-up and extended study, LPD polyplexes were combined with promoters from either mouse rhodopsin (rod photoreceptors), human red/green opsin (cone photoreceptors), thymocyte antigen (retinal ganglion cells), or vitelliform macular dystrophy (RPE cells). The injected eyes of the BALB/c mice presented a cell specific expression of GFP for the rod and cone photoreceptors, as well as ganglion and RPE cells. Additionally, LPD polyplexes did not affect the morphology of the retina and no inflammation was observed (Wang et al., 2016).