Postnatal deletion of Myo7a leads to early-onset hearing loss. Targeted postnatal deletion of Myo7a was achieved by crossing Myo7a floxed mice (Myo7afl/fl) with Myo15-cre+/– mice (15, 27). Using auditory brainstem responses (ABRs) to assess auditory function, we found that hearing loss in Myo7afl/fl Myo15-cre+/– mice is progressive (Figure 1) and started several days following the induced downregulation of MYO7A from the IHCs (Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.182138DS1). ABR thresholds for clicks were significantly elevated in Myo7afl/fl Myo15-cre+/– compared with control mice at P25–P26 and P33–P35 (P < 0.0001 for both comparisons), but not at P20 (P > 0.9999, Tukey’s post hoc test, 1-way ANOVA; Figure 1A). Similar findings were also observed for pure-tone evoked ABRs (Figure 1B). Compared with controls, littermate Myo7afl/fl Myo15-cre+/– mice exhibited significantly elevated ABR thresholds only at the 2 older ages (P < 0.0001, for both comparisons; P20: P = 0.1164, Tukey’s post hoc test, 2-way ANOVA; Figure 1B). The amplitude and latency of ABR wave 1 at 12 kHz, which is the frequency region that more closely matches that used for the ex vivo experiments described below, became significantly affected by the lack of MYO7A at the 2 older age ranges tested (Supplemental Figure 2). These results show that, in the absence of MYO7A, mice progressively lose their hearing (Figure 1), which is linked to the loss of the mechanoelectrical transduction (Supplemental Figure 3).

Auditory brainstem response thresholds in control and Myo7afl/fl Myo15-cre+/– mice. (A) Average ABR thresholds for click stimuli recorded from control Myo7afl/fl (black) and Myo7afl/fl Myo15-cre+/– (red) male and female mice at P20, P25–P26, and P33–P35 The age group tested are shown on the x axis. *P < 0.0001, Tukey’s post hoc test (1-way ANOVA). (B) ABR thresholds for frequency-specific pure tone burst stimuli at 3, 6, 12, 18, 24, and 30 kHz recorded from controls (Myo7afl/fl) and littermate Myo7afl/fl Myo15-cre+/– mice. Significance found using 2-way ANOVA. The number of mice tested for each genotype are shown next to the data, while the dashed lines indicate the upper threshold limit selected for these recordings (95 dB). Data are shown as mean ± SD.

Time course of efferent synapse formation in IHCs from Myo7afl/fl Myo15-cre+/– mice. Axosomatic efferent cholinergic synapses with IHCs are normally present only during prehearing stages (21–23, 28) or in aging mice (24–26). In this condition, the release of acetylcholine (ACh) from the efferent terminals leads to the activation of α9α10 nicotinic ACh receptors (nAChRs) on the postsynaptic IHC membrane. Since α9α10nAChRs are permeable to Ca2+ and are closely colocalized with the small conductance Ca2+ activated K+ channels (SK2), the release of ACh from efferent terminals causes IHC hyperpolarization (21–23). The Myo7afl/fl Myo15-cre+/– mouse strain recapitulates the efferent rewiring of the aged cochlea in young adult IHCs (15).

We initially looked at changes in the expression of the postsynaptic SK2 channels in the IHCs located within 150 μm of the apical-coil sensory epithelium (9–12 kHz frequency region) using immunostaining (Figure 2, A–E). As expected, SK2 channels were expressed in all prehearing IHCs from both control Myo7afl/fl (P9, black symbols; Figure 2, A, D, and E) and WT C57BL/6N mice (P10, blue symbols; Figure 2, D and E) but were downregulated in adult mice (Figure 2, B–D). In Myo7afl/fl Myo15-cre+/– mice, both the percentage of IHCs expressing SK2 (Figure 2D) and the number of SK2 puncta per IHC (Figure 2E) were significantly increased compared with littermate controls (P < 0.0001 for both comparisons, 2-way ANOVA). By P49, all the IHCs in Myo7afl/fl Myo15-cre+/– mice reexpressed SK2 channels as in prehearing cells (Figure 2D), and the number of SK2 puncta per IHC was no longer significantly different than that of control prehearing IHCs (P = 0.1874). Similar to immature IHCs, the reexpressed SK2 puncta were juxtaposed to efferent terminals (ChAT immunoreactivity). Although a larger number of IHCs from P22 Myo7afl/fl Myo15-cre+/– mice seem to show SK2 puncta (P = 0.0057, Sidak’s post hoc test, 2-way ANOVA; Figure 2D), their number per IHC was not significantly different to that from control IHCs (P = 0.9936; Figure 2E). This indicates that SK2 channels start to be significantly upregulated in IHCs in Myo7afl/fl Myo15-cre+/– mice older than P22 and became more pronounced over the following days (Figure 2E).

Efferent synapses return to adult IHCs in Myo7afl/fl Myo15-cre+/– mice. (A–C) Maximum intensity projections of confocal Z stack images taken from the 9–12 kHz cochlear region in control Myo7afl/fl (P9 [A], P24 [B], P39 [C]) and littermate Myo7afl/fMyo15-cre+/– mice (P24 [B], P39 [C]). Cochleae were labeled with antibodies against SK2 (green), the presynaptic efferent marker ChAT (magenta), and the hair cell marker MYO7A (blue). The right panels in A–C show a single IHC rotated on the y, z plane, providing a lateral view of the IHCs, which show the juxtaposed SK2 puncta and ChAT labeling of the efferent synapses. Scale bars: 10 μm. (D and E) Percentage of IHCs that expressed SK2 puncta (D) and number of SK2 puncta per IHC (E) in 150 μm of the apical cochlea region at different age groups of both control Myo7afl/fl and Myo7afl/fl Myo15-cre+/–. At prehearing ages, P10 WT C57BL/6N mice were also used as a comparison with P9 Myo7afl/fl. Data are shown as mean ± SD. The number of mice used for each age group is indicated above the single data points/averages. *P < 0.05, Šídák post hoc test (2-way ANOVA).

Considering that the ACh-activated current can be elicited in the IHCs from Myo7afl/fl Myo15-cre+/– at P22 (15), and SK2 channels are significantly upregulated in IHCs after P22 (Figure 2), we investigated when the axo-somatic efferent contacts on the IHCs of Myo7afl/fl Myo15-cre+/– mice became active. The presence of functional axo-somatic efferent contacts was tested by applying an extracellular solution containing 40 mM KCl, instead of the normal 5.8 mM, onto the cochlear epithelium. High-K+ caused IHCs, which were voltage clamped at –84 mV, to respond with an inward sustained current owing to a positive shift in the K+ reversal potential from –81 mV (5.8 mM K+) to –31 mV (40 mM K+). In addition, high-K+ also depolarizes the efferent synaptic terminals causing the release of ACh-containing vesicles and the generation of transient inhibitory postsynaptic currents (IPSCs) superimposed on the KCl-induced sustained inward current in IHCs (21, 29). We found that, in Myo7afl/fl Myo15-cre+/– mice, 40 mM KCl elicited robust and reliable efferent synaptic currents in the large majority of IHCs starting from P25 (Figure 3, A–E). These currents were blocked by 1M strychnine (Figure 3D), a potent blocker of α9α10nAChRs (30), confirming the reexpression of these receptors in adult Myo7afl/fl Myo15-cre+/– IHCs. In Myo7afl/fl Myo15-cre+/– mice, we found that both the frequency and amplitude of the IPSCs were similar across the different age ranges (P = 0.5890, P = 0.4089, respectively; 1-way ANOVA; Figure 3, F and G). Both the frequency (4.8 ± 3.2 Hz, range 0.5–14.1 Hz, P24–P42) and amplitude (76 ± 46 pA, range 27–317 pA, P24–P42) of IPSCs from all 70 IHCs were comparable with those previous reported (15, 29). These results indicate that knocking out Myo7a causes the presynaptic efferent terminals to innervate the IHCs, or become functional, only a few days after the acquisition of the postsynaptic elements in the IHCs.

Reestablishment of efferent synapses onto the IHCs of Myo7afl/fl Myo15-cre+/– mice. (A–D) Inward membrane currents recorded from IHCs of Myo7afl/fl (P24 [A]; P32 [C]) and littermate Myo7afl/fl Myo15-cre+/– KO (P24 [B]; P25 [D]) mice during the extracellular perfusion of 40 mM KCl (holding potential: –84 mV). The size of the slow-activating and sustained inward current, which is independent from the efferent system activation, varied among cells and was larger in Myo7afl/fl Myo15-cre+/– (365 ± 230 pA, n = 129) compared with Myo7afl/fl mice (165 ± 153 pA, n = 108) due to the smaller K+ currents active at the holding potential of –84 mV (15). Note that the superimposed inhibitory synaptic currents (IPSCs) were only evoked in Myo7afl/fl Myo15-cre+/– (see also expanded view in D), which were blocked by the selective α9α10nAChR blocker strychnine. (E) Percentage of IHCs responding to 40 mM KCl with IPSCs as a function of age. Numbers above the data represent the IHCs showing IPSCs versus total IHCs tested. (F and G) Average frequency (F) and amplitude (G) of the IPSCs as a function of age. Number of mice in E–G for each age group from left to right are: Myo7afl/fl: n = 7, 10, 14, 26, 17; Myo7afl/fl Myo15-cre+/–: n = 7, 7, 9, 16, 17. Data are shown as mean ± SD.

Afferent ribbon synapses are retained during reinnervation of IHCs by the efferent system. In the adult cochlea, the LOC efferent neurons form axo-dendritic synapses with the SGN terminals that contact the IHCs, the majority of which being predominantly located on the modiolar side of the cell (31). Considering that the modiolar SGNs are believed to be more vulnerable to damage and are largely reduced during aging (32, 33), it has been suggested that the disappearance of their physiological target could prompt the unconnected efferent terminals to establish direct axo-somatic contacts with IHCs (24, 26, 34). Therefore, we tested this hypothesis in Myo7afl/fl Myo15-cre+/– mice by evaluating the number of afferent synapses using antibodies against the presynaptic ribbon protein RIBEYE (CtBP2) and the postsynaptic AMPA-type glutamate receptor GluR2 (32, 35). We found that both CtBP2 and GluR2 puncta were present in the pre- and postsynaptic sites at each age tested (P22–P50) in both control and Myo7afl/fl Myo15-cre+/– mice (Figure 4, A–C). The number of CtBP2 and GluR2 puncta, as well as the number of colocalized puncta (indicative of functional afferent synapses) in 150 μm of the apical cochlear region (9–12 kHz), did not significantly differ between both genotypes over the age range investigated (CtPB2, P = 0.6638; GluR2, P = 0.6020; percentage of colocalization, P = 0.7501, 2-way ANOVA; Figure 4C). Overall, these data indicate that the reestablishment of the axo-somatic efferent synapse in Myo7afl/fl Myo15-cre+/– mice, which occurs as early as P24, was not due to the loss of afferent fibers.

Ribbon synapse number is not affected in Myo7afl/fl Myo15-cre+/– mice. (A and B) Maximum intensity projections of confocal Z stack images taken from the 9–12 kHz cochlear region in Myo7afl/fl and Myo7afl/fl Myo15-cre+/– mice at P22 (A) and P50 (B). IHCs were labeled with antibodies against the ribbon synapse marker CtBP2 (magenta), the postsynaptic marker GluR2 (green), and the cell marker MYO7A (blue). The right panels in A and B show a higher magnification of the synaptic region within the boxed IHCs depicted in the left panels. Colocalization between the pre- (CtBP2) and postsynaptic (GluR2) markers is highlighted in white. Scale bars: 10 μm (left), 5 μm (right).(C) Number of CtBP2 puncta (top panel) and GluR2 puncta (middle panel) per IHC; bottom panel shows the percentage of colocalized CtBP2-GluR2 puncta, at each age group tested in Myo7afl/fl (black) and Myo7afl/fl Myo15-cre+/– (red) mice. Significance was found using 2-way ANOVA. Data are shown as mean ± SD. The number of mice is indicated above the data groups and numbers of IHCs is shown in parentheses.

Knockdown of Myo7a in IHCs alone is sufficient to induce efferent reinnervation. Having established that the efferent reinnervation of IHCs in Myo7afl/fl Myo15-cre+/– mice was not associated with a loss of SGNs, we tested whether it was a secondary effect caused by lack of MYO7A in the outer hair cells (OHCs). This is because adult OHCs, which are innervated by MOC efferent fibers, are known to be more susceptible than IHCs to cochlear insult, such as loud noise or aging (36, 37). If OHCs are lost or degenerate, then the unconnected MOC fibers could divert toward the IHCs. To address this hypothesis, we crossed the Myo7afl/fl with Otof-cre mice (Figure 5, A–G), which express cre-recombinase under the otoferlin promoter that primarily targets IHCs (38, 39). Similar to Myo7afl/fl Myo15-cre+/–, we found that Myo7afl/fl Otof-cre+/– mice had normal hearing thresholds at least up to P19 but rapidly increased with age such that, by P40, they were almost completely deaf (Figure 5, A and B). We verified the overall function of OHCs from Otof-cre mice by measuring distortion product otoacoustic emissions (DPOAEs), which are a product of cochlear amplification caused by OHC electromotility during acoustic stimulation of their hair bundles. We found that DPOAE thresholds were indistinguishable between the 2 genotypes at P39–P44 (Supplemental Figure 4), indicating that OHCs were functional. Although Myo7afl/fl Otof-cre+/– mice do not downregulate MYO7A in 100% of the IHCs (Figure 5, C–E), they offer the advantage of allowing comparisons within the same cochlea between cells with and without MYO7A. A characteristic of adult IHCs is the expression of large conductance calcium-activated potassium channels (BK) at their neck region, which carry the rapid-activating outward K+ current IK,f (40, 41). We found that IHCs from Myo7afl/fl Otof-cre+/– mice that still expressed the BK channels did not show SK2 channels, which are normally present only in immature IHCs (22, 23). However, several IHCs from Myo7afl/fl Otof-cre+/– mice lacked BK channels and instead reexpressed SK2 channels (Figure 5, C–E), which is a sign of the reestablishment of the efferent postsynaptic specialization. The presence of functional axo-somatic efferent synapses in Myo7afl/fl Otof-cre+/– mice was supported by the presence of synaptic currents in response to the application of 40 mM extracellular KCl (Figure 5, F and G). IPSCs recorded from IHCs of P42–P50 Myo7afl/fl Otof-cre+/– mice had an average frequency of 2.3 ± 1.2 Hz (range 1.2–4.3.1 Hz, 7 IHCs, 4 mice) and amplitude of 47 ± 12 pA (range 32–71 pA). These values were not significantly different from those measured in Myo7afl/fl Myo15-cre+/– mice of a comparable age range (P36–P42 from Figure 3, F and G: 3.9 ± 3.5 Hz, P = 0.4744; 71 ± 33 pA, P = 0.0606; 25 IHCs; Mann-Whitney U test). Overall, the above findings indicate that the targeted deletion of Myo7a in IHCs alone is sufficient to reestablish axo-somatic efferent reinnervation in the adult cochlea.

Conditional KO of Myo7a only in the IHCs is sufficient to induce the efferent reinnervation. (A) Average ABR thresholds for click stimuli recorded from Myo7afl/fl (black) and Myo7afl/fl Otof-cre+/– (red) mice at P19, P23, and P36–P40. Two-tailed t test was used. (B) ABR thresholds for frequency-specific pure tone burst stimuli from 3 to 30 kHz recorded from both genotypes at P19 (left), P23 (middle), and P36–P40 (right). Significance was found using 2-way ANOVA. Number of mice tested for each genotype is shown next to the data. The dashed lines indicate the upper threshold limit for these recordings (95 dB). (C and D) Maximum intensity projections of confocal Z stack images taken from the 9–12 kHz apical region of the cochlea in Myo7afl/fl (C) and Myo7afl/fl Otof-cre+/– mice (D) at P23 (top panels) and P39 (bottom panels). IHCs were labeled with antibodies against SK2 (green), BK (magenta), and the IHC marker MYO7A (cyan). BK channels are expressed in mature IHCs. Arrows point to SK2 puncta; arrowhead indicates BK puncta. Scale bars: 10 μm. (E) Percentage of IHCs that expressed MYO7A (left) and SK2 (right) puncta in 150 μm of the apical cochlea region at P19 or P37–P39 in Myo7afl/fl (black) and Myo7afl/fl Otof-cre+/– mice (red). *P < 0.0001. NS indicates P = 0.0645 (Tukey’s post hoc test, 1-way ANOVA). The number of mice used for each genotype is shown above the data. (F and G) Voltage-clamp recordings obtained from IHCs in Myo7afl/fl (F) and Myo7afl/fl Otof-cre+/– (G) mice at P42 during the extracellular application of 40 mM KCl. IPSCs were only evoked in Myo7afl/fl Otof-cre+/– IHCs. Recordings were made from 3 Myo7afl/fl and 4 Myo7afl/fl Otof-cre+/– mice. Data are shown as mean ± SD.

Disruption of IHC exocytosis does not lead to efferent reinnervation. Similar to our findings in Myo7a-deficient mice, the efferent reinnervation of IHCs in the aged cochlea occurs at a time when the MET current is reduced compared with that of young adult mice (25). A reduction in the IHC MET current could decrease or even abolish the activity of some of the afferent terminals, even if they are still physically intact. The lack of activity in the target afferent fiber could trigger the LOC fibers to rewire onto the IHCs. To selectively silence the activity of the afferent fibers, we used 2 conditional KO mice to delete otoferlin in the IHCs (Figure 6), which is crucial for exocytosis at their ribbon synapses (42). This was achieved by crossing Otoffl/fl mice with either Myo15-cre or the tamoxifen inducible Vglut3-cre (43). The vesicular glutamate transporter 3 (VGLU3) is essential for glutamate release at IHC ribbon synapse since it is required to transport and repackage glutamate into synaptic vesicles (44, 45). While otoferlin expression was completely absent in the IHCs of Otoffl/fl Myo15-cre+/– mice at P16 (Supplemental Figure 5), that in Otoffl/flVglut3-cre+/– was only reduced by 35% ± 4 % (n = 4) at 4 weeks after the injection of tamoxifen. However, the latter strain allowed us to make side-by-side comparisons of IHCs with and without otoferlin.

Abolishing IHC exocytosis does not trigger efferent reinnervation. (A) Average peak Ca2+ current-voltage (ICa-Vm) curves from control Otoffl/fl (black) and Otoffl/fl Myo15-cre+/– (red) mice between P24 and P32. Recordings were obtained in response to 50 ms voltage steps from –81 mV in 10 mV increments. (B) ICa-Vm curves from control Otoffl/fl and Otoffl/flVglut3-cre+/– mice between P48 and P65. Recording conditions are as in A. (C and D) Exocytosis was recorded from both Otoffl/fl Myo15-cre+/– (C) and Otoffl/flVglut3-cre+/– (D) mice and their respective controls (Otoffl/fl) mice from the same age-range stated in A and B, respectively. ΔCm was elicited by applying 50 ms voltage steps to –11 mV (holding potential: –81 mV) between 2 ms and 0.6 seconds (interstep interval: at least 11 seconds) using 1.3 mM extracellular Ca2+ and at 35°C–37°C. Data are shown as mean ± SD. Significance was found using 2-way ANOVA. Number of IHCs is shown next to the data. Number of mice: Otoffl/fl and Otoffl/fl Myo15-cre+/– (n = 5 [A], 4 [C]); Otoffl/fl (n = 3 [B], 2 [D]) and Otoffl/flVglut3-cre+/– (n = 3 [B], 3 [D]). (E and F) Maximum intensity projections of confocal Z stacks of IHCs taken from the apical cochlea region of Otoffl/fl and Otoffl/fl Myo15-cre+/– (E) and Otoffl/flVglut3-cre+/– (F) mice at P50. IHCs were labeled with antibodies against the efferent presynaptic terminal marker ChAT (magenta), the postsynaptic efferent marker SK2 (green), and otoferlin (cyan). Right images show a side view of an IHC from the left images. SK2 puncta were absent in the IHCs of both strains. Three mice were used for each genotype and age group. Scale bars: 10 μm. (G) Voltage-clamp recordings obtained from IHCs held at –84 mV in P66 Otoffl/fl and Otoffl/flVglut3-cre+/– mice during the extracellular application of 40 mM KCl. Different from Myo7afl/fl Myo15-cre+/– mice (Figure 3), IHCs responded to KCl with slow inward sustained currents (Otoffl/fl: 343 ± 126 pA, 3 IHCs from 2 mice; Otoffl/flVglut3-cre+/–: 355 ± 98 pA, 5 IHCs from 4 mice; P = 0.8745, 2-tailed t test) without the superimposed IPSCs in both genotypes.

We first established whether exocytosis in the IHCs from the above mouse lines was abolished (Figure 6, A–D). IHC exocytosis was estimated by measuring increases in cell membrane capacitance (Cm) following depolarizing voltage steps that activate the calcium current (ICa). Cm is generally interpreted as a sign of neurotransmitter release from presynaptic cells (46–48). We found that the size of ICa, which was elicited by applying voltage steps in 10 mV increments from –81 mV to more depolarized potentials, was not significantly different between controls Otoffl/fl and either of the KO mice (Otoffl/fl Myo15-cre+/–, P = 0.2932; Otoffl/flVglut3-cre+/–, P = 0.101; 2-way ANOVA; Figure 6, A and B, respectively). However, the rate of neurotransmitter release in IHCs, which was studied by measuring Cm at the peak ICa (–11 mV) in response to depolarizing voltage steps from 2 ms to 0.6 seconds in duration, was almost completely abolished in both Otof-deficient mice (P < 0.0001 for both comparisons, 2-way ANOVA; Figure 6, C and D). Despite the absence of exocytosis in both conditional Otof-deficient mice, IHCs did not express SK2 channels (Figure 6, E and F) nor did they show any efferent-mediated synaptic currents (Figure 6G), indicating that abolishing exocytosis and thus the afferent activity did not trigger efferent reinnervation of IHCs in the adult cochlea.

Nature of the newly formed efferent synapses in IHCs lacking MYO7A. Recent data from aged mice have suggested that LOC fibers are the most likely candidate to reinnervate the IHCs (25). To confirm this observation, we performed immunostaining experiments using the anti-ATP1A3 antibody, which labels the MOC efferent neurons (49) known to contact the OHCs in the adult cochlea (19, 50). Since the anti-ATP1A3 antibody has also been shown to label the afferent fibers, MOC fibers are identified as those that also showed expression of the efferent marker ChAT. Using this approach, we confirmed that the efferent synapses on the OHCs from both Myo7afl/fl and Myo7afl/fl Myo15-cre+/– mice were MOC fibers since they showed both ATP1A3 and ChAT labeling (Supplemental Figure 6, A–F). However, the efferent synapses on IHCs from Myo7afl/fl Myo15-cre+/– mice were not labeled by the ATP1A3 antibody (Supplemental Figure 6F), suggesting they were LOC terminals. We also found that in IHCs, ChAT+ terminals were found juxtaposed to the efferent postsynaptic SK2 channel puncta only in Myo7afl/fl Myo15-cre+/– mice (Supplemental Figure 6, G and H), further supporting the reexpression of the postsynaptic machinery.

AAV-mediated Myo7a rescue in posthearing mice partially restores IHC and hearing function in Myo7a-deficient mice. We next assessed the plasticity of the adult sensory epithelium by testing whether the AAV-mediated in vivo replacement of Myo7a in posthearing mice could restore the function of the MET apparatus and revert the efferent innervation back to its normal adult-like configuration. To do this, we used a dual AAV approach to deliver Myo7a cDNA into the cochlear perilymphatic space in vivo via the round window membrane (RWM) of P13–P15 control and Myo7a-deficient mice.

The hearing function of the treated mice was assessed using ABRs. We found that for click ABR stimuli, noninjected WT mice had thresholds around 45 dB sound pressure level (SPL), while noninjected Myo7afl/fl Myo15-cre+/– mice had no detectable thresholds even at levels up to 120 dB SPL (Figure 7A). Most AAV9-Myo7a–injected Myo7afl/fl Myo15-cre+/– mice had significantly lower thresholds compared with noninjected Myo7afl/fl Myo15-cre+/– mice, although their thresholds remained significantly elevated compared with controls (P < 0.0001, Tukey’s post hoc test, 1-way ANOVA; Figure 7A). For pure tone ABRs, we found a similar trend as for the clicks, with an average recovery of about 30 dB across all frequencies tested in the AAV-injected mice (between 3 kHz and 24 kHz) (Figure 7A). We then analyzed the amplitude and latency of the ABR wave 1 at 12 kHz and at 2 sound intensities (95 and 110 dB SPL). We found that the amplitude and latency of wave 1 was reemerging in Myo7afl/fl Myo15-cre+/– mice transduced with AAV9-Myo7a for sound intensities above 90 dB (Figure 7, B–E). These results are encouraging considering that dual-AAV9-Myo7a does not transduce 100% of adult IHCs (Supplemental Figure 7). Injection into the cochlea of either AAV9-Myo7A-Nterm or AAV9-Myo7a-Cterm alone failed to produce MYO7A in hair cells (Supplemental Figure 8).

Auditory brainstem response thresholds partially recover in Myo7afl/fl Myo15-cre+/– mice injected with AAV9-Myo7a. (A) Average ABR thresholds for click (left) and frequency-specific pure tone burst (right) stimuli recorded from Myo7afl/fl (black), Myo7afl/fl Myo15-cre+/– (red), and Myo7afl/fl Myo15-cre+/–AAV9-Myo7a (blue) mice at P48–P54. The number of mice used is shown near the data points. The dashed lines indicate the upper threshold limit used for this experiment (120 dB SPL). For click thresholds, statistical comparison is from 1-way ANOVA (Tukey’s post hoc test: P < 0.0001 between control and Myo7afl/fl Myo15-cre+/– without or with surgery; P = 0.0001 between Myo7afl/fl Myo15-cre+/– without and with surgery). For pure tone thresholds, statistical comparison is from 2-way ANOVA (Tukey’s post hoc test: P < 0.0001 for all comparisons). (B) Average ABR waveform responses at 12 kHz and using 2 stimulus intensities (95 and 110 dB SPL) relative to threshold from the 3 different mouse lines described in A. Continuous lines represent mean values, and shaded areas represent the SD. P1 and N1 indicate the positive and negative peaks of wave 1. (C) Same average ABR waveform responses as shown in B but only for the noninjected Myo7afl/fl Myo15-cre+/– and Myo7afl/fl Myo15-cre+/–AAV9-Myo7a mice to better highlight the size of wave 1. (D and E) Average amplitude (D) and latency (E) of wave 1 (from P1 to N1) for 3 conditions described in A. Please note that, due to wave 1 amplitude being zero in Myo7afl/fl Myo15-cre+/– mice (red), the latency could not be measured in E. Data in A, B, D, and E are shown as mean ± SD.

As mentioned above (Figure 5, C and D), BK channels are a distinctive characteristic of adult IHCs and are expressed around their neck region (Figure 8, A and D). While IHCs from Myo7afl/fl Myo15-cre+/– mice were almost completely devoid of BK channels (Figure 8, B and D), those from Myo7afl/fl Myo15-cre+/– mice transduced with AAV9-Myo7a reexpress BK channels (Figure 8, C and D). The reexpression of BK channels in the IHCs of Myo7afl/fl Myo15-cre+/– mice transduced with AAV9-Myo7a was confirmed electrophysiologically, since the size of IK,f was significantly upregulated compared with that recorded in the IHCs of Myo7afl/fl Myo15-cre+/– mice (P < 0.0001, Tukey’s post hoc test, 1-way ANOVA) but comparable with that of control Myo7afl/fl cells (P = 0.6246) (Figure 8, E and F). The presence of a large IK,f was not due to the injection of AAV9-Myo7a in P13 mice preventing the observed downregulation of BK channels in the IHCs of Myo7afl/fl Myo15-cre+/– mice (15), since IK,f was still very small or absent at P27 (Supplemental Figure 9, A and B). We then investigated whether adult IHCs transduced with AAV9-Myo7a returned to their normal functional state by also downregulating the efferent postsynaptic SK2 channels and their ability to respond to the efferent system (Figure 8, G–L). We found that IHCs from Myo7afl/fl Myo15-cre+/– mice exhibited a robust reexpression of SK2 channels (Figure 8, H and J), as described in Figure 2. However, SK2 puncta were almost completely absent in IHCs transduced with AAV9-Myo7a (Figure 8, I and J), which is similar to control IHCs (Figure 8, G and J). We also found that the extracellular application of 40 mM KCl onto the IHCs of Myo7afl/fl Myo15-cre+/– mice transduced with AAV9-Myo7a elicited inward currents with no or very few superimposed IPSCs, the frequency of which was significantly reduced compared with that recorded in IHCs from Myo7afl/fl Myo15-cre+/– mice (P = 0.0009, Tukey’s post hoc test, 1-way ANOVA). However, IHC responses to high-K+ were not significantly different from those recorded from control mice (P = 0.2918). As for the BK channels (Supplemental Figure 9, A and B), the downregulation of the efferent activity was not due to the transduction of AAV9-Myo7a at P13 preventing the normal changes seen in IHCs from Myo7afl/fl Myo15-cre+/– mice (Figure 2 and Figure 3), since SK2 channels and IPSCs were present in P22–P27 IHCs (Supplemental Figure 9, C–F).

Injection of AAV9-Myo7a rescues the normal efferent wiring of the adult IHCs in Myo7afl/fl Myo15-cre+/– mice. (A–C) Maximum intensity projections of confocal Z stack images taken from the 9–12 kHz cochlear region in noninjected control Myo7afl/fl (A), Myo7afl/fl Myo15-cre+/– (B), and Myo7afl/fl Myo15-cre+/– mice injected with AAV9-Myo7a (C) at P49–P54. AAV-Myo7a was injected between P13 and P15. Cochleae were labeled with antibodies against BK (magenta) and the IHC markers MYO7A (cyan) or MYO6 (gray). BK was almost complete absence in Myo7afl/fl Myo15-cre+/– mice (B). Right images show a single IHC from the left panels rotated on the y,z plane, providing a lateral view of the IHC and allowing visualization of the juxtaposed MYO7A/MYO6 and the BK puncta. (D) Percentage of IHCs expressing the BK channels over 150 μm range. Statistical comparisons (post hoc test, 1-way ANOVA): Myo7afl/fl vs. Myo7afl/fl Myo15-cre+/–, P < 0.0001; Myo7afl/fl vs. Myo7afl/fl Myo15-cre+/–AAV9-Myo7a, P = 0.0068; Myo7afl/fl Myo15-cre+/– vs. Myo7afl/fl Myo15-cre+/–AAV9-Myo7a, P = 0.0016. Number of mice for A–C is shown in D. (E) Example of outward K+ current responses from IHCs of P37 Myo7afl/fl, P37 Myo7afl/fl Myo15-cre+/– and P36 Myo7afl/fl Myo15-cre+/– mice injected with AAV9-Myo7a. Currents were elicited by using 10 mV depolarizing voltage steps from –84 mV to the various test potentials shown by some of the traces. BK current IK,f is indicated with an arrow. (F) Size of the isolated IK,f measured at –25 mV and at 1 ms from the onset in the 3 experimental conditions shown in E. The number of IHCs is shown above the data. Number of mice from left to right: 15, 16, 3. (G–I) Images obtained as described in A–C for the same 3 mouse lines (P49–P54) and using antibodies against SK2 (magenta) and the IHC marker MYO7A (cyan) and otoferlin (gray). (J) Percentage of IHCs expressing the SK2 channels and number of SK2 puncta over 150 μm of the apical cochlear region. Number of mice used in G–I is shown in J. (K) Examples of inward membrane currents recorded from IHCs of P36 Myo7afl/fl Myo15-cre+/– mice injected with AAV9-Myo7a. Recording protocol is as described in Figure 3. Note that 1 IHC (left) only shows the inward current, while in the others (right) 40 mM K+ also elicited a few IPSCs. (L) Average frequency of the IPSCs recorded from IHCs of 17 Myo7afl/fl, 17 Myo7afl/fl Myo15-cre+/– and 3 Myo7afl/fl Myo15-cre+/– mice injected with AAV9-Myo7a. One-way ANOVA was used. Data are shown as mean ± SD.

The functional recovery of Myo7a-deficient IHCs following the transduction of AAV9-Myo7a was further validated on Myo7afl/fl Otof-cre+/– mice (Supplemental Figure 10, A–G). The in vivo delivery of AAV9-Myo7a to P13 Myo7afl/fl Otof-cre+/– mice led to the partial recovery of ABR thresholds (P < 0.0001, 2-way ANOVA; Supplemental Figure 10A). As shown for Myo7afl/fl Myo15-cre+, IHCs from Myo7afl/fl Otof-cre+/– mice downregulate the BK channels, reexpress SK2 channels, and show IPSC responses to efferent activation (Figure 5). However, as seen for Myo7afl/fl Myo15-cre+/– mice (Figure 8), IHCs from Myo7afl/fl Otof-cre+/– mice transduced with AAV9-Myo7a reexpressed BK channels (Supplemental Figure 10, B and C) and no longer showed IPSCs (Supplemental Figure 10, D–G).

The above results indicate that the reestablishment of MYO7A expression, and thus partial ABR recovery, was sufficient for the IHCs to establish MET current, downregulate the efferent postsynaptic machinery, and no longer respond to efferent stimulation.