Single and Dual Vector Gene Therapy with AAV9-PHP.B Rescues Hearing in Tmc1 Mutant Mice

Abstract

AAV-mediated gene therapy is a promising approach for treating genetic hearing loss. Replacement or editing of the Tmc1 gene, encoding hair cell mechanosensory ion channels, is effective for hearing restoration in mice with some limitations. Efficient rescue of outer hair cell function and lack of hearing recovery with later-stage treatment remain issues to be solved. Exogenous genes delivered with the adeno-associated virus (AAV)9-PHP.B capsid via the utricle transduce both inner and outer hair cells of the mouse cochlea with high efficacy. Here, we demonstrate that AAV9-PHP.B gene therapy can promote hair cell survival and successfully rescues hearing in three distinct mouse models of hearing loss. Tmc1 replacement with AAV9-PHP.B in a Tmc1 knockout mouse rescues hearing and promotes hair cell survival with equal efficacy in inner and outer hair cells. The same treatment in a recessive Tmc1 hearing-loss model, Baringo, partially recovers hearing even with later-stage treatment. Finally, dual delivery of Streptococcus pyogenes Cas9 (SpCas9) and guide RNA (gRNA) in separate AAV9-PHP.B vectors selectively disrupts a dominant Tmc1 allele and preserves hearing in Beethoven mice, a model of dominant, progressive hearing loss. Tmc1-targeted gene therapies using single or dual AAV9-PHP.B vectors offer potent and versatile approaches for treating dominant and recessive deafness.

Keywords: TMC1; auditory; gene therapy; genetic deafness; hair cell; hearing; hearing loss; inner ear; mechanotransduction; sensory transduction.

Graphical abstract

Figure 1

AAV2/9-PHP.B-Tmc1 Mediates Hearing Restoration in Tmc1^Δ/Δ Mice (A) Representative ABR waveform families recorded from mice at 4 weeks or 4 weeks postinjection for each indicated condition, using 11.3 kHz tone bursts at incrementally increasing sound-pressure levels. Thresholds were determined by the detection of peak 1 and are indicated by colored traces. Scale bar applies to all families. (B) Mean ABR (top) and DPOAE (bottom) thresholds of 4-week-old mice plotted as a function of stimulus frequency for wild-type (WT) C57BL/6 (black; n = 6), Tmc1^Δ/Δ uninjected controls (red; n = 5), and Tmc1^Δ/Δ mice injected with PHP.B-Tmc1 at P1 (blue; n = 28) with five best performers (green) tested at 4 weeks old. Lighter traces show individual responses. Average data shown as mean ± SEM. (C) Mean ABR (top) and DPOAE (bottom) thresholds plotted as a function of stimulus frequency for Tmc1^Δ/Δ mice injected with PHP.B-Tmc1 at P1, tested at 4, 8, and 12 weeks after injection (n = 6). Average data shown as mean ± SEM. (D) Mean ABR (top) and DPOAE (bottom) thresholds plotted as a function of stimulus frequency for Tmc1^Δ/Δ mice injected with PHP.B-Tmc1 at P7, tested at 4 weeks after injection (n = 19). Lighter traces show individual responses. Average data shown as mean ± SEM. (E) Representative 10× (top panels) and 63× (bottom panels) confocal images from middle-apical cochlear sections of indicated experimental conditions, immunostained against myosin VIIa at 12 weeks of age. Scale bars, 100 μm at 10× and 20 μm at 63×. (F) Mean cell counts of inner (left) and outer (right) hair cells (IHCs and OHCs, respectively) for WT C57BL/6 (n = 3), Tmc1^Δ/Δ uninjected controls (n = 5), and P1-injected Tmc1^Δ/Δ conditions (n = 6) at 12 weeks of age measured from 100 μm sections of the apex, middle, and basal cochlear turns. Individual samples are shown as a scatterplot (∗∗∗p ≤ 0.001; one-way ANOVA followed by Sidak’s multiple comparison test for respective measurements, i.e., Tmc1^Δ/Δ apex versus injected apex).

Figure 2

AAV9-PHP.B-Mediated Hearing Recovery in the Recessive Hearing-Loss Model, Baringo (A) Representative ABR waveform families recorded from mice at 4 weeks or 4 weeks postinjection for indicated conditions, using 11.3 kHz tone bursts at incrementally increasing sound-pressure levels. Thresholds were determined by the detection of peak 1 and are indicated by colored traces. Scale bar applies to all families. (B) Mean ABR (top) and DPOAE (bottom) thresholds of 4-week-old mice plotted as a function of stimulus frequency for Tmc1^Y182C/+ (black; n = 5), Baringo uninjected controls (red; n = 4), and Baringo mice injected with PHP.B-Tmc1 at P1 (blue; n = 8) with three best performers (green) tested at 4 weeks old. Lighter traces show individual responses. Average data shown as mean ± SEM. (C) Mean ABR (top) and DPOAE (bottom) thresholds plotted as a function of stimulus frequency for Baringo mice injected with PHP.B-Tmc1 at P7, tested at 4 weeks after injection (n = 13). Lighter traces show individual responses. Average data shown as mean ± SEM. (D) Representative 10× (left panels) and 63× (right panels) confocal images from middle-apical cochlear sections of indicated experimental conditions, immunostained against myosin VIIa at 4 weeks of age or 4 weeks postinjection. Scale bars, 100 μm at 10× and 20 μm at 63×. (E) Mean cell counts of IHCs (left) and OHCs (right) for Tmc1^Y182C/+ (n = 4), Baringo uninjected controls (n = 5), P1-injected Baringo (n = 9), and P7-injected Baringo (n = 9) conditions at 4 weeks of age, measured from 100 μm sections of the apex, middle, and basal cochlear turns. Individual samples are shown as a scatterplot (not significant [ns] ≥ 0.05, ∗∗∗p ≤ 0.001; one-way ANOVA followed by Sidak’s multiple comparison test for respective measurements, i.e., Baringo apex versus injected apex).

Figure 3

Coexpression of GFP and RFP by Dual Injection of AAV9-PHP.B Vectors Representative 10× confocal images (left and middle images) from middle-apical cochlear sections showing GFP (green) and RFP (red) coexpression in IHCs and OHCs. 63× magnification of 100 μm sections (right images) showing GFP and RFP expression in individual hair cells, merged, and stained against myosin VIIa (blue). Scale bars, 100 μm at 10× and 20 μm at 63×.

Figure 4

Screening of gRNAs for Selective Disruption of the Bth Allele Mediated by the SpCas9 Nuclease (A) Dual AAV vector constructs for expression of the gRNA and SpCas9 nuclease (tracrRNA, trans-activating CRISPR RNA; β-gl int, beta-globin intron; pA, 3′ UTR and polyadenylation signal; ITR, AAV2-inverted terminal repeats). (B) Design of gRNA sequences targeting the Bth mutation. Aligned sequences of the mouse Tmc1 and human TMC1 genes are shown in the lower boxes. The PAM sequence is depicted in blue and the nucleotides corresponding to the Bth mutation in red. For some gRNA sequences, a G (shown in green) was added to the 5′ end to facilitate the activity of the RNA polymerase. (C) Representative sequencing chromatograms of the Tmc1 gene in MEF cells. Asterisk (∗) indicates the position corresponding to the T-to-A Bth point mutation. Note the predominant T in this position in Bth/+ cells transfected with gRNA 15, which indicates selective SpCas9-mediated cleavage of the Bth allele. The arrowhead shows the SpCas9 cleavage site. (D) TIDER quantification of gRNA-induced cleavage efficacy based on the relative presence of the WT or Bth sequence in Bth/+ MEF cells transfected with each of the gRNA-expressing constructs. Note the significant selectivity for the Bth allele of gRNA 12 and 15 (ns ≥ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001; multiple t tests with Holm-Sidak method). Data shown as mean ± SD. (E) Representative sequencing chromatograms of the TMC1 gene in HAP TMC1+ and TMC1Bth cells. Note that indels (arrowhead) are present only in the SpCas9-expressing HAP TMC1Bth cells transfected with gRNA 16. (F) TIDE analysis of gRNA-induced cleavage efficacy based on indel frequency in HAP cells transfected with each of the gRNA-expressing constructs. Data shown as mean ± SD. Note the significant selectivity of gRNA 16 for the Bth allele (∗p ≤ 0.05; multiple t tests with Holm-Sidak method).

Figure 5

In Vivo Dual Vector System for SpCas9/gRNA Targeting of the Bth Allele (A) Representative ABR waveform families recorded from mice at 24 weeks for indicated conditions, using 11.3 kHz tone bursts at incrementally increasing sound-pressure levels. Thresholds were determined by the detection of peak 1 and are indicated by colored traces. Scale bar applies to all families. (B) Mean ABR (left)- and DPOAE (right)-threshold mice plotted as a function of stimulus frequency for Bth uninjected controls (red, n = 6 at 4 weeks; n = 3 at 6 weeks; n = 9 at 12 weeks; n = 9 at 24 weeks old) and Tmc1^Bth/+ mice dual injected with PHP.B-SpCas9 and PHP.B-gRNA15-GFP (blue, n = 9 at 4 weeks; n = 13 at 6 weeks; n = 9 at 12 weeks; n = 8 at 24 weeks old). Lighter traces show individual responses. Average data shown as mean ± SEM. (C) DPOAE thresholds at 11.3 kHz, measured in (B), plotted as a function of age from 4, 6, 12, and 24 weeks (for 4, 6, 12, and 24 weeks, respectively, WT, n = 8, 9, 9, 6; Tmc1^Bth/+, n = 6, 3, 9, 9; dual-injection SpCas9 + gRNA15-GFP, n = 9, 13, 9, 8) (∗p ≤ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001; multiple t tests with Holm-Sidak method). Average data shown as mean ± SEM.

Figure 6

PHP.B-SpCas9/gRNA Dual Vector Transduction Preserves Hair Cell Survival (A) Representative 63× confocal images of 100 μm sections from the apex, middle, and basal cochlear turns of WT C57BL/6 and uninjected Tmc1^Bth/+ mice (right two columns) or Tmc1^Bth/+ mice dual injected with PHP.B-SpCas9 and PHP.B-gRNA15-GFP immunostained against myosin VIIa at 24 weeks of age. Scale bar, 20 μm. (B) Mean cell counts of IHCs (left) and OHCs (right) for WT C57BL/6 (n = 3), Bth uninjected controls (n = 3), and Bth mice dual injected with PHP.B-SpCas9 and PHP.B-gRNA15-GFP (n = 6) at 24 weeks of age, measured from 100 μm sections of the apex, middle, and basal cochlear turns. Individual samples are shown as a scatterplot (∗∗p ≤ 0.01, one-way ANOVA; ∗p ≤ 0.01, Student’s t test). Data shown as mean ± SEM. (C) Percentage of surviving hair cells (left) and percentage of GFP-positive hair cells (right) at 24 weeks of age correlated to ABR thresholds measured with 8, 11.3, and 16 kHz tone bursts from Bth mice dual injected with PHP.B-SpCas9 and PHP.B-gRNA15-GFP (n = 5 mice; 15 measurements). Data fit with linear regression with r = −0.62 (p = 0.013) for % surviving hair cells and r = −0.59 (p = 0.020) for % GFP+ hair cells.