RNA base editing therapy cures hearing loss induced by OTOF gene mutation

Abstract

Otoferlin (OTOF) gene mutations represent the primary cause of hearing impairment and deafness in auditory neuropathy. The c.2485C>T (p. Q829X) mutation variant is responsible for approximately 3% of recessive prelingual deafness cases within the Spanish population. Previous studies have used two recombinant AAV vectors to overexpress OTOF, albeit with limited efficacy. In this study, we introduce an enhanced mini-dCas13X RNA base editor (emxABE) delivered via an AAV9 variant, achieving nearly 100% transfection efficiency in inner hair cells. This approach is aimed at treating OTOF^Q829X, resulting in an approximately 80% adenosine-to-inosine conversion efficiency in humanized Otof^Q829X/Q829X mice. Following a single scala media injection of emxABE targeting OTOF^Q829X (emxABE-T) administered during the postnatal day 0-3 period in Otof^Q829X/Q829X mice, we observed OTOF expression restoration in nearly 100% of inner hair cells. Moreover, auditory function was significantly improved, reaching similar levels as in wild-type mice. This enhancement persisted for at least 7 months. We also investigated P5-P7 and P30 Otof^Q829X/Q829X mice, achieving auditory function restoration through round window injection of emxABE-T. These findings not only highlight an effective therapeutic strategy for potentially addressing OTOF^Q829X-induced hearing loss but also underscore emxABE as a versatile toolkit for treating other monogenic diseases characterized by premature termination codons.

Keywords: OTOF; RNA base editing; emxABE; gene therapy; hearing loss; otoferlin.

Graphical abstract

Figure 1

Screen of efficient RNA base editor and gRNAs for efficiently correcting hOTOF^Q829X mutation (A) Schematic illustration of RNA base editing mediated by single direct repeat (sDR) or dual direct repeat (dDR) elements. The RNA base editor consists of mini-dCas13X.1 protein and ADAR2dd protein from the RESCUE-S editor. The reporter contains a mCherry CDS with a stop codon at W148 (TGG>TAG). Red fluorescence (mCherry) can only be detected if A-to-I conversion is achieved. (B) Flow cytometry results of editors with different gRNA components. The higher the mCherry ratio, the higher the efficiency of A-to-I conversion. (C) Schematic illustration of RNA base editors with different subcellular localization sequences. NLS^SV40, NLS from simian virus 40; bpNLS^SV40, bipartite SV40 NLS. (D) A-to-I conversion rate of RNA base editors with different subcellular localization sequences at two endogenous sites in HEK293T cells. n = 3, except for 2xNLS^SV40 at SMAD4 site, where n = 2. (E) Schematic diagram of screening optimal gRNAs targeting hOTOF^Q829X mutation in HEK293T cells. The reporter construct contains an mCherry cassette fused with 2A peptide and part of hOTOF CDS with Q829X (CAG>TAG) mutation. The red C indicates mismatched base against A mutation in hOTOF transcript. pCMV, porcine cytomegalovirus. (F) A-to-I conversion efficiency of gRNAs containing different mismatch base positions. NT, nontargeting crRNA. n = 4. (G) Measurement of bystander A-to-I editing rate for multiple adenosines within targeting sequence of all gRNAs from deep sequencing. The targeting site number is 0. Others are sites with bystander editing, with -on the left and + on the right, which was omitted. Numbers represent the distance from the targeting site. The lower row is the corresponding amino acid, where X represents a stop codon. The top of the heatmap is the amino acid changes caused by bystander editing. The left side of > is the original amino acid; the right side of > is the amino acid after bystander editing. n = 1. (H) Transcriptome-wide off-target site numbers for mCherry (control), ADAR2dd^E488Q, emxABE_NT, or emxABE_OTOF, cotransfected with the OTOF^Q829X reporter plasmid, in HEK293T cells. All of the values are presented as mean ± SEM, n = 3, unless otherwise noted. ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001, 2-tailed unpaired 2-sample t test. See Table S3 for detailed data.

Figure 2

In vivo correction of Otof^Q829X mutation and restoration of OTOF protein expression with emxABE-T delivered by AAV-HGHC at P0–P3 Otof^Q829X/Q829X mice (A) Diagrammatic workflow of the in vivo studies with emxABE-T by AAV-HGHC via SM injection at P0–P3 Otof^Q829X/Q829X mice (hereafter marked Homo). IF, immunofluorescence staining. (B) Deep-sequencing results in the cochlea of Otof ^Q829X/Q829X mice injected with vehicle or low and medium doses of emxABE-T after 4 weeks of administration (vehicle, n = 3; low dose, n = 5; medium dose, n = 5). Targeting site number. Others are sites with bystander editing, with -on the left and + on the right. Numbers represent the distance from the targeting site. The lower row is the corresponding amino acid, and X represents a stop codon. The top of the heatmap is the amino acid changes caused by bystander editing. The left side of > is the original amino acid, the right side of > is the amino acid after bystander editing. On the right is the statistical analysis of editing efficiency at the only on-target site. (C) The confocal immunofluorescence results of the entire basilar membrane of mice after 4 weeks of virus injection. MYO7A⁺ indicates the HC marker. n = 1. (D) High-magnification views of IHCs immunolabeled for synaptic ribbons (CTBP2) of middle cochlear turns after 30 weeks of virus injection. n = 3. Scale bar: 15 μm. HA tag indicates the expression of emxABE. (E) The proportion of successful recovery of OTOF protein expression in IHC cells infected with the virus (HA⁺). n = 3. (F) Synaptic ribbon numbers quantified from IHCs in apical, middle, and basal cochlear turns after 30 weeks of administration (WT, n = 16 IHCs; Homo+vehicle, n = 15 IHCs; Homo+emxABE-T, n = 15 IHCs from 3 mice). HC, hair cell. Data are presented as mean ± SEM. Statistical tests are 2-way ANOVA with Bonferroni correction for multiple comparisons. ∗p < 0.05; ∗∗∗∗p < 0.0001; ns, not statistically significant. Unless otherwise specified, Homo+ emxABE-T indicates the result of medium dose administration. See Table S4 for detailed data.

Figure 3

Recovery of hearing function of Otof^Q829X with AAV-emxABE-T injection at P0–P3 (A) A representative graph of the Click ABR waveform recorded after 4 weeks of ear administration. The green trace represents the threshold of the mouse ears, with no threshold for the group of Homo+vehicle. (B) ABR thresholds of mice for tone-burst stimuli are recorded 4 weeks after therapeutic injection of low, medium, and high doses (WT, n = 10; Homo+vehicle, n = 7; low dose, n = 6; medium dose, n = 8; high dose, n = 9). (C and D) ABR wave I latency (C) and amplitude (D) at 8,16, and 32 kHz after 4 weeks of administration. (Latency: WT, n = 7; Homo+vehicle, n = 9; Homo+emxABE-T, n = 8. Amplitude: WT, n = 6; Homo+vehicle, n = 5; Homo+emxABE-T, n = 8.) (E) ABR thresholds of mice for tone-burst stimuli are recorded 12 weeks after therapeutic injection under medium and high dose (WT, n = 3; Homo+vehicle, n = 5; medium dose, n = 8; high dose, n = 6). Data are presented as mean ± SEM. Statistical tests were 2-way ANOVA with Bonferroni correction for multiple comparisons. ∗∗p < 0.01; ∗∗∗∗p < 0.0001. Homo, Otof^Q829X/Q829X mice. Unless otherwise specified, Homo+emxABE-T indicates the result of medium-dose administration. See Table S5 for detailed data.

Figure 4

In vivo therapy of Otof^Q829X mice with emxABE-T delivery at P30 by RW injection (A) Diagrammatic workflow of the in vivo studies with emxABE-T delivered at P30 by AAV-HGHC in Otof^Q829X/Q829X mice (hereafter marked Homo). (B) Deep-sequencing results after injection 4 weeks in the cochlea of Otof^Q829X/Q829X mice at P30 (vehicle, n = 4; emxABE-T, n = 5). (C) Confocal of IHCs immunolabeled for otoferlin, IHC (MYO7A) and AAV-emxABE-T (HA) of middle cochlear turns after dosing 6 weeks. Scale bars: 50 μm. (D) The proportion of successful recovery of OTOF protein expression in IHC cells infected with the virus (HA⁺). n = 3. (E) The OTOF recovery efficiency in all IHCs (MYO7A⁺) are calculated by OTOF⁺/MYO7A⁺. n = 3. (F) ABR thresholds of mice for tone-burst stimuli are recorded 4 weeks after therapeutic injection (WT + vehicle contralateral ear, the contralateral ear from WT + vehicle group, n = 6; WT + vehicle, n = 6; Homo+vehicle, n = 5; Homo+emxABE-T, n = 14). Data are presented as mean ± SEM. Statistical tests were 2-way ANOVA with Bonferroni correction for multiple comparisons. See Table S6 for detailed data.