CNG channel-related retinitis pigmentosa

Abstract

The genes CNGA1 and CNGB1 encode the alpha and beta subunits of the rod CNG channel, a ligand-gated cation channel whose activity is controlled by cyclic guanosine monophosphate (cGMP). Autosomal inherited mutations in either of the genes lead to a progressive rod-cone retinopathy known as retinitis pigmentosa (RP). The rod CNG channel is expressed in the plasma membrane of the outer segment and functions as a molecular switch that converts light-mediated changes in cGMP into a voltage and Ca²⁺ signal. Here, we will first review the molecular properties and physiological role of the rod CNG channel and then discuss the characteristics of CNG-related RP. Finally, we will summarize recent activities in the field of gene therapy aimed at developing therapies for CNG-related RP.

The rod CNG channel structure and molecular function

Cyclic nucleotide-gated (CNG) channels are ligand-gated cation channels whose activity can be controlled by cyclic nucleotides, such as cyclic adenosine monophosphate (cAMP) or cyclic guanosine monophosphate (cGMP). From a structural view, CNG channels belong to the superfamily of pore-loop cation channels (James & Zagotta, 2018, Yu, Yarov-Yarovoy, Gutman & Catterall, 2005). Other closely related members of this superfamily include the hyperpolarization-activated cyclic nucleotide-gated (HCN) channels (Biel, Wahl-Schott, Michalakis & Zong, 2009, Wahl-Schott & Biel, 2009) and the ether-à-go-go-type (KCNH) channels (Barros, de la Peña, Domínguez, Sierra & Pardo, 2020). Despite of their structural similarity, these channels differ significantly in their function, namely in their cation selectivity, voltage dependence, and cyclic nucleotide regulation (James & Zagotta, 2018). The CNG channel gene family in vertebrates consists of six members (CNGA1, CNGA2, CNGA3, CNGA4, CNGB1, CNGB3), four of which are linked to inherited retinal diseases such as retinitis pigmentosa (CNGA1 and CNGB1) or achromatopsia (CNGA3 and CNGB3).

Rods express a CNGA1/CNGB1 channel, whereas cones have a CNGA3/CNGB3 channel. The physiological function of the rod CNG channel is to confer a depolarizing Na⁺/Ca²⁺ inward current upon binding of cGMP to a C-terminal cyclic nucleotide binding domain (CNBD). The channel complex is heterotetrameric consisting of three alpha 1 (gene name: CNGA1) subunits and one beta 1 subunit (gene name: CNGB1) subunit (Xue, Han, Zeng & Jiang, 2022). CNGA1 encodes the approximately 78 kDa alpha subunit, which confers key channel properties and has been shown to form functional homotetrameric ion channels in heterologous expression systems (Kaupp & Seifert, 2002). CNGB1 encodes the approximately 240 kDa beta subunit, which confers specific biophysical properties to the rod channel (Kaupp & Seifert, 2002). Although the CNGB1 subunit has a similar structure to CNGA1, it cannot form functional ion channels on its own, but is important for the correct outer segment localization of the channel complex. A shorter CNGB1 splice variant, which lacks a large part of the cytosolic N-terminus, including the glutamic acid-rich protein (GARP) domain (Figure 1), is expressed in olfactory sensory neurons where it forms, together with CNGA2 and CNGA4, a cAMP-gated CNG channel involved in olfactory signal transduction (Boccaccio, Menini & Pifferi, 2021).

Figure 1.

Structure of the rod CNG channel. (A-B) Structures of individual subunits: (A) CNGA1 and (B) CNGB1. S1–S6, transmembrane segment 1–6; CNBD, cyclic nucleotide-binding domain; GARP, glutamic acid-rich domain. The structures of the cytosolic N termini of CNGA1 and CNGB1 could not be determined experimentally and are shown as dashed lines. (C) Top view of the heterotetrameric human rod CNGA1/CNGB1 channel complex showing the arrangement of the three CNGA1 and the one CNGB1 subunit within the complex. Structures in this figure were generated with the RSCB PDB 3D View tool (www.rcsb.org/3d-view/) based on PDB 7RHH.

Recent studies revealed structural details and insights into the activation mechanism of the rod CNG channel (Barret, Schertler, Benjamin Kaupp & Marino, 2022a, Barret, Schertler, Kaupp & Marino, 2022b, Xue et al., 2022, Xue, Han, Zeng, Wang & Jiang, 2021). Both subunits of the channel consist of six segments (S1-S6) of alpha helices spanning the plasma membrane (Figure 1). The channel core consists of a loop (P-loop) between S5 and S6, which together form the channel pore proper (Biel & Michalakis, 2009). The intracellular C-terminus bares the CNBD, which is coupled to S6 by the C-linker (Figure 1). The S4 has multiple positively charged residues, but its specific structure does not allow the necessary charge movements to control the opening and closing of the channel in a voltage-dependent manner, as is the case with conventional voltage-gated channels (Li, Zhou, Wang, Michailidis, Gong, Su, Li, Li & Yang, 2017). Therefore, the activation of the CNG channel is exclusively ligand controlled. Binding of cGMP to the CNBD induces a rotational movement of the entire C-terminus relative to the pore, which through corresponding rotation of the C-linker is transferred to the gating ring (James & Zagotta, 2018, Napolitano, Torre & Marchesi, 2021, Xue et al., 2022, Xue et al., 2021). This partially moves the channel gate upward, removing the confining forces that keep the gate constricted and opening the channel pore to allow ion permeation. Within the rod channel complex, the overall structures of the three alpha subunits and the single beta subunit align well and exhibit a similar domain arrangement, resulting in a quite symmetric pore in the closed state (Xue et al., 2022). However, during the opening of the channel, only the two CNGA1 subunits to the left and opposite of the CNGB1 subunit show the expected gating movements, while CNGB1 and the other CNGA1 subunit hardly move, resulting in an asymmetric opening of the pore (Xue et al., 2022). Closer examination of the structure of the rod CNG channel in the closed state revealed an asymmetric pore geometry, in which an arginine residue of CNGB1 S6 protrudes directly into and closes the ion conduction pathway of the pore (Barret et al., 2022a, Zheng, Hu, Li & Yang, 2022). Mutation of this arginine to a glycine led to a higher single channel conductance (Xue et al., 2022).

These studies have greatly improved our understanding of CNG channel structure and function, but there are still gaps in our knowledge, particularly with respect to the N- and C-terminal domains of CNGA1 and CNGB1. This is especially true for the N-terminal GARP domain of the CNGB1 subunit, which represents the most important structural difference between the CNGA1 and CNGB1 subunits and which could not be modeled so far for technical reasons (Barret et al., 2022a, Xue et al., 2022, Zheng et al., 2022). This is most likely due to the intrinsic disorder of the GARP domain, which will only assume a more constant structural identity when bound to an interaction partner. While the precise role of the GARP domain is not clear, recent findings indicate that the GARP region of the CNGB1 subunit contains two distinct functional regions: one being essential for proper subcellular targeting of the channel to rod outer segments, and one for connecting the CNG channel to the photoreceptor disc rim, likely mediated through an interaction with peripherin-2 (Pearring, Martinez-Marquez, Willer, Lieu, Salinas & Arshavsky, 2021).

The physiological function of the rod CNG channel in phototransduction

The rod photoreceptors of the retina mediate high-sensitivity vision in low light and do not contribute to high-acuity color vision in daylight conferred by the cone photoreceptors. While the basic function of the phototransduction cascades in rods and cones is similar and relies on the second messenger cGMP. Key proteins, including the CNG channel, differ and are encoded by distinct but functionally equivalent genes. In this review, we focus on the signal transduction in rods. The central second messenger cGMP controls activation of the CNG channel in the plasma membrane of outer segments. The cGMP concentration is kept high in the dark by the constant activity of the transmembrane guanylyl cyclases (retGC), keeping the CNG channels in an open conformation (Pugh, Duda, Sitaramayya & Sharma, 1997, Yang, Robinson, Xiong, Yau, Birch & Garbers, 1999). The CNG channels conduct a constant, mixed depolarizing Na⁺ and Ca²⁺ current. This so-called “dark current” keeps the rod photoreceptor in a depolarized state and leads to the activation of voltage-gated calcium channels (VGCC) at the synaptic terminals, which trigger glutamate release. Upon light-induced conformational change, rhodopsin, as a G protein-coupled receptor, releases its G protein transducin, which in turn binds to and activates the PDE6A/B phosphodiesterase that hydrolyzes cGMP (Chen, Getter, Salom, Wu, Quetschlich, Chorev, Palczewski & Robinson, 2022). PDE6A/B activity leads to a decrease in cGMP and closure of the rod CNG channel, which in turn hyperpolarizes the rod photoreceptor, inactivating the VGCC and reducing synaptic glutamate release. The changes in glutamate release are sensed by postsynaptic glutamate receptors in bipolar and horizontal cells, which transmit and/or shape the light signal to the retinal output neurons, the ganglion cells. The closure of CNG channels on light stimulation, together with the constant activity of Na⁺/Ca²⁺, K⁺ exchangers (NCKX1), which balance the concentration of Ca²⁺ in rod outer segments, leads to a decrease in intracellular Ca²⁺ concentration (Cervetto, Lagnado, Perry, Robinson & McNaughton, 1989, Hodgkin, McNaughton & Nunn, 1985, Kaupp & Seifert, 2002, Schnetkamp, 2004, Vinberg, Wang, De Maria, Zhao, Bassnett, Chen & Kefalov, 2017, Yau & Nakatani, 1984, Yau & Nakatani, 1985). The drop in Ca²⁺ facilitates recovery from the light response by modulating the activities of PDE6A/B and retGC (Ames, 2021, Chen et al., 2022, Gao, Eskici, Ramachandran, Poitevin, Seven, Panova, Skiniotis & Cerione, 2020, Irwin, Gupta, Gao, Cahill, Chu & Cote, 2019).

The pathobiology of rod CNG channel-related retinopathy

Pathogenic sequence variants in CNGA1 or CNGB1 impair the rod CNG channel function and/or expression and cause autosomal recessive RP (Bareil, Hamel, Delague, Arnaud, Demaille & Claustres, 2001, Dryja, Finn, Peng, McGee, Berson & Yau, 1995). RP is the most common inherited retinal disease and the leading cause of inherited blindness with an overall prevalence of approximately 1:4,000, affecting an estimated 1.5 to 3 million people worldwide (Hanany, Rivolta & Sharon, 2020, Hartong, Berson & Dryja, 2006). The disease is characterized by a primary loss of rod function causing night blindness. The loss of rod function is followed by a decline of photoreceptors due to progressive cell death, resulting in a gradual narrowing of the visual field. Additional changes such as abnormal migration and accumulation of pigment become visible as retinal atrophy progresses (Kalloniatis & Fletcher, 2004). As rod loss progresses, the structure and function of the primarily unaffected cones become impaired and visual acuity decreases. Due to the severity of symptoms such as night blindness, photophobia, visual field loss, and color and contrast vision deficits, which may already be evident in early adolescence, RP represents a significant loss of quality of life for those affected (Petrs-Silva & Linden, 2014). In many cases, RP even leads to complete loss of vision after the fourth or fifth decade of life. Based on the mode of inheritance, RP can be classified as either X-linked (XLRP), maternally/mitochondrially acquired, autosomal dominant (adRP), and autosomal recessive (arRP). Apart from those common inheritance patterns, rare non-mendelian mitochondrial and digenic inherited forms of RP exist (Kajiwara, Berson & Dryja, 1994, Mansergh, Millington-Ward, Kennan, Kiang, Humphries, Farrar, Humphries & Kenna, 1999). It is estimated that 50–60% of RP is inherited in an autosomal recessive manner (Hartong et al., 2006, Verbakel, van Huet, Boon, den Hollander, Collin, Klaver, Hoyng, Roepman & Klevering, 2018, Xu, Guan, Shen, Zhang, Xiao, Jiang, Li, Yang, Jia, Yin, Guo, Wang & Zhang, 2014) and mutations in more than 100 genes are associated with arRP (OMIM-268000) (see also https://web.sph.uth.edu/RetNet/). Two of them are CNGA1 (OMIM #123825) and CNGB1 (OMIM #600724).

CNGA1 was among the first genes to be linked to arRP almost three decades ago (Dryja et al., 1995). Since then, nearly 50 probable pathogenic mutations in CNGA1 have been identified, with an overall reported prevalence ranging from 2 % of arRP cases in Spain to 5,1% in Japan and up to 7.6% in China (Chen, Zhao, Sheng, Li, Gao, Zhang, Kang, Pan, Liu, Jiang, Shi, Chen, Rong, Chen, Lai, Liu, Wang, Yuan, Liu, Vollrath, Pang & Zhao, 2013, Dryja et al., 1995, Hanany et al., 2020, Hartong et al., 2006, Jin, Qu, Hou, Xu, Meng, Pang & Yin, 2016, Katagiri, Akahori, Sergeev, Yoshitake, Ikeo, Furuno, Hayashi, Kondo, Ueno, Tsunoda, Shinoda, Kuniyoshi, Tsurusaki, Matsumoto, Tsuneoka & Iwata, 2014, Paloma, Martinez-Mir, Garcia-Sandoval, Ayuso, Vilageliu, Gonzalez-Duarte & Balcells, 2002). Most of the identified CNGA1 mutations cause deletions of key functional domains or are missense mutations that result in protein degradation or membrane trafficking impairment (Biel & Michalakis, 2009, Dryja et al., 1995, Kandaswamy, Zobel, John, Santhiya, Bogedein, Przemeck, Gailus-Durner, Fuchs, Biel, de Angelis, Graw, Michalakis & Amarie, 2022, Kaupp & Seifert, 2002, Mallouk, Ildefonse, Pages, Ragno & Bennett, 2002). There is no systematic study describing the natural history of CNGA1-RP. However, individual reports as well as the authors’ observation suggest that despite the early onset of clinical symptoms (e.g., night blindness from early childhood), the rate of progression of CNGA1-RP is low to moderate, providing a relatively large window of opportunity for therapeutic intervention to prevent vision loss (Petrs-Silva & Linden, 2014).

Pathogenic mutations in CNGB1 constitute 1–4% of arRP cases, again with the highest prevalence in Asian countries (Bareil et al., 2001, Ge, Bowles, Goetz, Scholl, Wang, Wang, Xu, Wang, Wang & Chen, 2015, Hanany et al., 2020, Hartong et al., 2006, Kondo, Qin, Mizota, Kondo, Hayashi, Hayashi, Oshima, Tahira & Hayashi, 2004, Nassisi, Smirnov, Solis Hernandez, Mohand-Said, Condroyer, Antonio, Kuhlewein, Kempf, Kohl, Wissinger, Nasser, Ragi, Wang, Sparrow, Greenstein, Michalakis, Mahroo, Ba-Abbad, Michaelides, Webster, Degli Esposti, Saffren, Capasso, Levin, Hauswirth, Dhaenens, Defoort-Dhellemmes, Tsang, Zrenner, Sahel, Petersen-Jones, Zeitz & Audo, 2021, Simpson, Clark, Alexander, Silvestri & Willoughby, 2011, Xu et al., 2014). In contrast to CNGA1-RP, most mutations in CNGB1 cause only minor deletions or single amino acid substitutions that affect rod CNG channel stability, transport, or channel function (Bareil et al., 2001, Becirovic, Nakova, Hammelmann, Hennel, Biel & Michalakis, 2010, Kondo et al., 2004, Michalakis, Zong, Becirovic, Hammelmann, Wein, Wanner & Biel, 2011). The largest retrospective natural history study involving 33 patients showed that CNGB1-RP is a rather slowly progressing subtype of RP similar to that observed in CNGA1-RP (Jackson, Dubis & Moosajee, 2022). To reach a more accurate picture on the natural history of CNGB1-RP and to determine clinical endpoints for future interventional trials, an international multicenter prospective natural history study (NCT04639635) is currently underway.

Animal models of CNGA1-RP and CNGB1-RP

Various animal models for CNGA1-RP and CNGB1-RP have been described so far (Ahonen, Arumilli & Lohi, 2013, Hüttl, Michalakis, Seeliger, Luo, Acar, Geiger, Hudl, Mader, Haverkamp, Moser, Pfeifer, Gerstner, Yau & Biel, 2005, Liu, Wang, Xiao, Li, Ruan, Luo, Wan, Wang & Sun, 2021, Wiik, Ropstad, Ekesten, Karlstam, Wade & Lingaas, 2015, Winkler, Ekenstedt, Occelli, Frattaroli, Bartoe, Venta & Petersen-Jones, 2013, Winkler, Occelli & Petersen-Jones, 2020, Zhang, Molday, Molday, Sarfare, Woodruff, Fain, Kraft & Pittler, 2009). Among those, three animal models have been reported for CNGA1-RP. First, a naturally occurring dog model with progressive retinal atrophy identified in a Shetland sheepdog breed (Wiik et al., 2015) that has yet to be phenotypically characterized. More recently, two mouse models have been described and characterized (Kandaswamy et al., 2022, Liu et al., 2021). One is a mouse line with a the CRISPR/Cas9-induced 65-bp frame-shift deletion in exon 2 of Cnga1 (Liu et al., 2021). Although not yet experimentally verified at the protein level, the genomic deletion is expected to result in a premature stop codon and loss of most of the Cnga1 protein starting at amino acid 27. Homozygous mice show a progressive retinal degeneration and loss of photoreceptors and a greatly reduced scotopic electroretinogram (ERG) response (Liu et al., 2021). Another CNGA1-RP mouse model is a Cnga1 mutant generated by N-ethyl-N-nitrosourea (ENU) treatment (Kandaswamy et al., 2022). These mice have a c.1526 A>G mutation in Cnga1 that results in the replacement of a single tyrosine (Y509) in the b3 strand of the CNBD of the Cnga1 protein with a cysteine. This Y509C mutation destabilizes the Cnga1 protein, resulting in complete loss of Cnga1 and Cngb1 proteins, presumably due to protein degradation. The Cnga1 mutant mice lack rod-specific ERG responses and show progressive retinal degeneration, initially affecting only rods and, from the sixth month of life, also cone photoreceptors. At the age of 1 year, ERG responses are no longer measurable (Kandaswamy et al., 2022). Interestingly, a mutation affecting the corresponding tyrosine residue in the human CNGA1 protein (Y513) was identified in patients with CNGA1-arRP (Jin et al., 2016, Kandaswamy et al., 2022). Among the various animal models, the Cngb1 animal models have been best characterized to date (Blank, Goldmann, Koch, Amann, Schon, Bonin, Pang, Prinz, Burnet, Wagner, Biel & Michalakis, 2017, Chen, Woodruff, Wang, Concepcion, Tranchina & Fain, 2010, DeRamus, Stacks, Zhang, Huisingh, McGwin & Pittler, 2017, Hüttl et al., 2005, Petersen-Jones, Occelli, Winkler, Lee, Sparrow, Tsukikawa, Boye, Chiodo, Capasso, Becirovic, Schon, Seeliger, Levin, Michalakis, Hauswirth & Tsang, 2018, Winkler et al., 2013, Zhang et al., 2009, Zhang, Rubin, Fineberg, Huisingh, McGwin, Pittler & Kraft, 2012). The first described model is a knockout mouse that carries a deletion of exon 26 of the Cngb1 gene (Hüttl et al., 2005). The knockout results in a truncated transcript that lacks the sequence encoding the pore-forming S6 and all down-stream domains (Hüttl et al., 2005). These Cngb1-X26 knockout mice lack Cngb1 protein, which in turn leads to degradation of the remaining Cnga1 subunit. The data obtained in this mouse model provided the first evidence that Cnga1 protein requires Cngb1 for proper expression (Hüttl et al., 2005). As a consequence of the loss of rod CNG channel, Cngb1-X26 knockout mice show decreased rod light responses and diminished scotopic ERG responses. Similar to Cnga1-deficient mice, the functional impairment is accompanied by progressive degeneration of the rods and secondary degeneration of the primarily unaffected cones. The rate of progression of degeneration was also similar, with first signs of cone degeneration at 6 months of age, when about half of the rods have been lost. At the age of 1 year, only one to two rows of photoreceptors were still present in the retina (Hüttl et al., 2005). Just as in CNGA1-RP and CNGB1-RP patients, the phenotypes of Cnga1- and Cngb1-deficient mice were very similar in disease manifestation and progression. Based on the observations in the animal models, this is likely due to the fact that loss of CNGA1 or CNGB1 leads to destabilization and secondary degradation of the remaining CNG channel subunit. Thus, the cause of the RP phenotype in both cases is most likely a complete loss of the rod CNG channel complex and its functions.

In addition to murine models, a well characterized dog model of CNGB1-RP exists (Petersen-Jones et al., 2018, Winkler et al., 2013). The dog model carries a c.2387delA;2389_2390insAGCTAC mutation in Cngb1, which was first described in a Papillon dog breed with markedly reduced or absent rod function and slowly progressive retinal degeneration (Winkler et al., 2013). The identified c.2387delA;2389_2390insAGCTAC mutation leads to premature termination of the Cngb1 protein at a similar position as in the Cngb1-X26 mice (Hüttl et al., 2005, Winkler et al., 2013, Winkler et al., 2020). Just like CNGB1-RP patients and Cngb1-X26 mice, Cngb1 mutant dogs show early loss of normal rod function with a slow rod degeneration followed by a secondary cone degeneration and loss in cone function (Petersen-Jones et al., 2018). Interestingly, careful evaluation of the scotopic ERG responses showed that the young affected dogs have a residual but very reduced and desensitized rod response, likely due to residual function of homotetrameric CNGA1 channels that still reach the outer rod segments (Petersen-Jones, Pasmanter, Occelli, Querubin & Winkler, 2022). The availability of this well characterized large animal model for CNGB1-RP with eyes similar in size and structure to the human eye greatly facilitates translational programs to develop future gene therapies (Winkler et al., 2020).

Besides the Cngb1-X26 mice, other mouse models with distinct deletions were generated (Chen et al., 2010, Zhang et al., 2009). One model carries a genetic deletion in exon 1 of Cngb1 (Cngb1-X1 mice) that eliminates all transcripts of the Cngb1 locus, including those encoding the cytosolic glutamic acid-rich proteins (GARPs) (Zhang et al., 2009). Cngb1-X1 mice showed a similar functional phenotype to Cngb1-X26 mice, whereas the structure of the rod outer segment was more affected, suggesting an important role of GARP proteins in rod disc morphogenesis and outer segment integrity. Several studies have examined the function of GARPs and found a role in shaping rod outer segment morphology, transport, and channel function (Ardell, Bedsole, Schoborg & Pittler, 2000, Ba-Abbad, Holder, Robson, Neveu, Waseem, Arno & Webster, 2019, Chakraborty, Conley, DeRamus, Pittler & Naash, 2015, DeRamus et al., 2017, Michalakis et al., 2011, Pearring et al., 2021, Ritter, Khattree, Tam, Moritz, Schmitz & Goldberg, 2011). For an overview of the structural properties of GARPs, its distribution and association with other proteins see (Goldberg, Moritz & Williams, 2016). However, the function of soluble GARP proteins, as well as the corresponding channel-attached GARP domain, which are inherently unfolded (Batra-Safferling, Abarca-Heidemann, Korschen, Tziatzios, Stoldt, Budyak, Willbold, Schwalbe, Klein-Seetharaman & Kaupp, 2006) remain enigmatic.

Gene therapy approaches for the treatment of CNG channel-linked RP

The improved understanding of rod CNG channel biology, the availability of suitable animal models, and efficient vectors are now enabling the development of gene therapy treatments for CNGA1-RP and CNGB1-RP.

As in other autosomal recessive retinopathies, a valid therapeutic strategy appears to be the use of adeno-associated virus (AAV) vectors to introduce a normal copy of CNGA1 or CNGB1 into the affected rod cells to render them functional and halt retinal degeneration (Figure 2).

Figure 2.

Gene supplementation therapy with adeno-associated virus (AAV) vectors for the treatment of CNG channel-related RP. Rod photoreceptors that lack the CNG channel function are unable to respond to light stimuli. The concept of gene supplementation therapy uses AAV-based vectors to deliver a functional copy of the gene that is mutated in patients. When mutated, CNG channel subunits are instable and degraded, which destabilizes the remaining endogenous CNG channel subunit. Expression of the missing CNG channel subunit rescues the endogenous from degradation and restores a functional CNG channel in the rod outer segments. In most cases the vectors are delivered via single injection under the retina (subretinal injection).

So far, there are no gene therapy approaches for CNGA1-linked RP. For CNGB1-linked RP, successful preclinical proof-of-concept studies for AAV-based gene supplementation have been established with species-matched vectors in both the Cngb1-X26 mouse model (Koch, Sothilingam, Garcia Garrido, Tanimoto, Becirovic, Koch, Seide, Beck, Seeliger, Biel, Muhlfriedel & Michalakis, 2012) and the Cngb1 mutant dog model (Petersen-Jones et al., 2018). Due to the genome size limitation (~4.7 kb) of AAV vectors and the relatively large size of the Cngb1 cDNA (~4 kb), the two studies had to choose small promoter and polyadenylation sequences. In both studies, AAV-mediated expression of Cngb1 restored rod CNG channel localization, improved rod-mediated retinal function, preserved rod structure, prevented secondary cone degeneration, and improved rod-mediated vision (Koch et al., 2012, Petersen-Jones et al., 2018). Based on these promising feasibility studies, translational studies with a vector optimized for human application (AAV5-RHO-CNGB1) could be initiated (Wagner, Zobel, Gerhardt, O’Riordan, Frederick, Petersen-Jones, Biel & Michalakis, 2021). AAV5-RHO-CNGB1 harbors a short human rhodopsin promoter for rod-specific expression of the human CNGB1 gene. Single subretinal injection of AAV5-RHO-CNGB1 in 4-week-old Cngb1-X26 mice, resulted in efficient expression of the human CNGB1 protein in mouse rods, which restored expression of the endogenous mouse Cngb1 protein (Wagner et al., 2021). This led to a dose-dependent improvement in rod-driven ERG responses and preservation of retinal structure (Wagner et al., 2021). Translational studies in Cngb1 mutant dogs are currently underway with the expectation of initiating clinical trials for the treatment of CNGB1-RP.

Conclusions and outlook

The rod CNG channelopathies are rare inherited RP-type retinopathies with an estimated total number of more than 100,000 affected patients (Hanany et al., 2020). Several well-characterized animal models have helped improve our understanding of the underlying pathobiology and are important for preclinical testing of new therapies. Gene therapy programs based on AAV vectors targeting CNGA1-RP or CNGB1-RP are currently in preclinical development and are expected to enter the clinical phase in the coming years.