A hop, skip, and a jump to evade USH2A deaf-blindness mutations

Main text

Over the past decade, many gene therapy approaches have been designed to restore auditory and visual function lost due to Usher syndrome (USH).¹ Among them, gene replacement therapy mediated by adeno-associated virus (AAV)-based vectors has been shown to be a safe and efficient strategy. However, this approach is not suitable for large genes, such as USH2A, which vastly exceed the cargo capacity of AAV vectors. In this issue of Molecular Therapy, Dulla et al.² provide evidence that synthetic antisense oligonucleotides (ASO or AON) can mediate exon-skipping, which may be a viable approach for a subset of USH2A mutations. With this approach, AONs bind specifically to the complementary exon 13 pre-mRNA sequence to induce exon skipping, thereby restoring translation of the rest of the USH2A sequence. At the protein level, this study validates that the shorter usherin Δexon 13 protein is expressed, functional, stable, and capable of restoring vision in several animal models of USH2A. The results of this work are the basis of an ongoing clinical trial (QR-421a phase 2/3; ClinicalTrials.gov: NCT03780257), which offers hope to patients who carry USH2A exon 13 mutations.

USH is a rare group of autosomal recessive disorders characterized by a progressive retinal degeneration (retinitis pigmentosa or RP), bilateral sensorineural hearing loss, and, in some cases, vestibular dysfunction. USH is clinically and genetically heterogenous, with 11 genes described so far that are associated with these pathologies.¹ Based on the severity and progression of this disease, three clinical forms have been described—USH1, USH2, and USH3—with USH1 being the most severe and USH3 the least severe form. USH2A is the most frequent form, representing over half of all USH cases. USH2A (OMIM: 608400) is one of the largest genes in the human genome, comprising 72 exons (15.6 kb coding sequence) that encode the usherin protein (5,202 amino acids, ∼570 kDa) (Figure 1A). To date, over 1,139 pathogenic and 474 likely pathogenic USH2A variants have been identified, with ∼35% of them localized to exon 13 (Figure 1C).^3,4 Among the mutations located within exon 13, the c.2299delG variant accounts for 31% of USH2A cases and represents the most common mutation in the USH2A gene. The c.2299delG mutation creates a frameshift in the coding sequence, resulting in a premature stop codon and the absence of functional usherin protein (Figure 1B). Usherin is expressed in photoreceptor cells of the retina and hair cells of the inner ear. While its role in the retina is poorly understood, the slow progression of the disease suggests that it is important for the maintenance of the photoreceptor cells rather than their initial development.⁵ Conversely, in the cochlea, usherin is essential for the development and maturation of hair bundles that form the sensory receptor organelle, which is essential for hair cell function.

Figure 1.

Exon skip strategy eludes a common mutation associated with USH2A

(A) Schematic representation of the USH2A protein. Predicted domains are illustrated that show the fibronectin and laminin repeats. The c.2299delG mutation falls within exon 13 and affects the short isoform encoded by exon 2-21 as well as the long isoform encoded by exon 2-71. (B) The common c.2299delG mutation leads to a stop codon and a non-functional protein. Use of antisense oligonucleotides (AONs) targeting this region led to an in-frame exon skip and production of a modified yet functional protein. (C) Minor allele frequency map of pathogenic and likely pathogenic USH2A variations juxtaposed with the USH2A gene map (vertical bars indicate coding exons). Scale bar, 5 E^-4%; from Deafness Variation Database.³

Cochlear implants or hearing aids are the standard of care for USH2A hearing impairment, but no biological treatments are available that alter the progression of the disease, which leads to vision loss in USH2A patients. Owing to the large size of USH2A, classic gene replacement is not feasible. To palliate this difficulty, Dulla et al.² explored an AON-mediated exon-skipping approach, previously developed to treat Duchenne muscular dystrophy (DMD) disorder.^6,7 The objective of their strategy was to hop over out-of-frame mutations in exon 13 and thereby restore in-frame production of a shorter yet functional protein (Figure 1B). Interestingly, usherin shares several similarities with dystrophin, as they are encoded by two of the largest genes of the human genome and both are very large proteins with multiple repeat domains. In addition, exon 13 of USH2A is an in-frame exon that encodes three of ten rod-like laminin-epidermal growth factor (EGF)-like modules (LE). The investigators hypothesized that jumping over the mutated exon would lead to production of a shorter but functional pseudo-protein (Figure 1B). Removing portions of a large protein while preserving function is not trivial. Interestingly, in silico analysis performed by Dulla et al.² predicted similar structures for full-length and shorter pseudo-usherin proteins.² Recently, Pendse et al.⁸ demonstrated the potential efficacy of a pseudo-USH2A protein in vitro and in vivo in auditory and visual tissues using a novel Ush2a^Δexon12 mouse model (mouse exon 12 is homologous of human exon 13).

Development of therapeutic strategies for USH2A have been hampered by the lack of valid animal models. Similarities between zebrafish and human USH2A genes, proteins, and subcellular localization of usherin in the retina renders zebrafish a useful model for assessing therapeutic strategies and outcomes. As such, Dulla et al.² used a ush2a^rmc1 zebrafish null model previously shown to have retinal defects.⁹ Injection of a combination of morpholino AONs into the yolk of one or two cell-stage mutants successfully led to recovery of wild-type subcellular localization of the protein and restoration of function in the mutant ush2a^rmc1 zebrafish. One important consideration for any gene therapy is how much protein is required to rescue the phenotype. Not enough, and recovery may not occur; too much, and toxicity may follow. The advantage of the AON approach is that there is no concern of overexpression of the protein since the endogenous promoters should remain undisturbed. However, depending on the bioavailability and efficiency of the AON, it is possible that too little protein expression would be recovered to rescue function. Interestingly, in an Ush2a^KO/Δexon12 mouse model, where one allele is knocked out, Pendse et al.⁸ demonstrated normal hearing and visual function, suggesting that mouse Δ exon 12 usherin is not only functional, but also that expression of 50% or less of this protein is sufficient to recover function. Here, Dulla et al.² report that production of as little as ∼20% of zebrafish Δ exon 13 usherin is sufficient to restore visual impairment in ush2a^rmc1 nulls.²

Base on the ability of the Δ exon 13 usherin transcripts to restore visual function in zebrafish, the authors developed another set of AONs (QR-421a) capable of specifically targeting and restoring the exon-skipped reading frame in wild-type mouse retina and human tissues (WEBI-rb1 and induced pluripotent stem cell [iPSC]-derived photoreceptor progenitor cells). The high efficiency and limited inflammatory effect of this approach suggested AON-mediated exon skipping may be a robust approach for clinical application. The effect of QR-421a in RNA therapy is currently being investigated in a phase 1/2 clinical trial (ClinicalTrials.gov: NCT03780257). An interim report from March 2021 for the phase 1/2 dose escalation trial revealed no adverse effects, with two of eight participants demonstrating improvement in functional and structural measures. These results are encouraging, and they corroborate those observed by Dulla et al.² Will there be a need for repeated injections? QR-421 includes a fully phosphorothiated backbone that provides significant resistance to nuclease degradation as well as a 2′ O-methoxyethyl (2′-MOE) alkyl substituent that improves safety and efficacy. However, a single injection may not be sufficient to maintain adequate Δ exon 13 usherin transcript levels in humans. To bypass this issue, multiple injections may be required to maintain stable production of Δ exon13 usherin over time. Alternative delivery systems may be considered, such as nanoparticles, which could potentially protect AONs from degradation and extend their therapeutic durability. Finally, studies aimed at clarifying the turnover of usherin protein in retinal tissue will also help define the therapeutic window of AON treatment required to provide long-term protection of visual function in USH2A patients.

This promising therapeutic approach is relevant not only for patients carrying mutations in exon 13 of USH2A, but also for other in-frame exon mutations in this gene. To further evaluate the therapeutic potential of this promising strategy, it will be interesting to adapt this approach to other USH genes or other forms of genetic deaf-blindness.