Utricles and saccules were collected separately from five WT mice at postnatal time points: P2, P14, P28, P60. RNA extraction and purification utilized the RNeasy Mini Kit (Qiagen). Total RNA was measured by spectrophotometer (Nanodrop, ND100, Thermo Fisher Scientific) and reverse transcribed to cDNA using iScript cDNA Synthesis Kit (Bio-Rad). Quantitative PCR reactions were then performed with the TaqMan Gene Expression Assays with intron spanning Tmc1 (5′-CATCTGCAGCCAACTTTGGTGTGT-3′ and 5′-AGAGGTAGCCGGAAATTCAGCCAT-3′) and Tmc2 (5′-AGATCTTTGCGTTCCTTGCCAACC-3′ and 5′-GATCTTCTTTCGCAGCTGGGCATT-3′). TaqMan probes were labeled with FAM reporter dye (Applied Biosystems). Cycle threshold (Ct) values of triplicate reactions were measured for each sample. Tmc1 and Tmc2 plasmid DNA of known concentrations and lengths (1:10, 1:100, 1:1000, 1:10000, 1:100000, 1:1000000) were used to validate Tmc1 and Tmc2 primer efficiency and to estimate particle number across all time points.

Ear capsules were removed and fixed in 1.5 mL 4% PFA/PBS under RNAse-free conditions, rocking at 4°C. Fixative was replaced with 100% MeOH and stored at −20°C until further use. Tissues were dissected in MeOH and then rehydrated through a graded methanol (25, 50, 75%) nuclease-free PBS series @ room temperature (RT). Tissues were washed in nuclease free PBS for 10 min at 25°C in 150 μL, then immersed ears in 150 μL of 100 μg/mL Proteinase K solution and incubated for 25 min @RT on a nutator. Tissues were post-fixed with 150 μL 4% PFA for 20 min @ RT on a nutator and washed in PBS/0.05% tween thrice in 150 μL for 5 min at 25°C. Tissues were then incubated in pre-hybridization buffer for 30 min at 37°C followed by incubation in 200 μL of probe hybridization buffer for 5 min. Finally, a pre-hybridization solution was prepared by adding 3.2 pmol of the desired probe, 16 nm in working solution into 200 μL of solution and embryos were incubated overnight at 37°C on a nutator in a pre-cleaned container which retains moisture.

For assessment of FISH following AAV injection at P1, mice with variable injection efficiency and VsEP thresholds were collected by P60 by the same process described above. Proteinase K concentration was increased to 125 μg/mL working solution and FISH was conducted otherwise similarly. Maculas were imaged at 7 μm depth with a range defined by the apex of phalloidin stained stereovilli containing Tmc1 puncta and basally to by Gfi1 signal. Striolar and extrastriolar utricular regions were defined by the presence or absence of Ocm signal, and by the distance from the lateral edge of the sensory domain. ROIs were restricted to hair cells expressing both Gfi1 and phalloidin signals, and all Tmc1 signal was normalized to the amount of Gfi1 present in each tissue.

P3 vestibular tissue was rapidly dissected following decapitation and placed under tungsten pins affixed to a glass coverslip using Sylgard. Tissue was submerged into 5 mL culture media containing 5 μM FM1-43 dye for 10–30 s with gentle agitation, transferred to a new container of 5 mL of culture media, washed three times, and placed in a slide holder for imaging. P60 mice were injected intraperitoneally with 5 mg/g mouse weight of FM1-43/FX fixable form as previously described (17). Tissue was collected after ~24 h and fixed for 1 h at 25°C in 4% PFA/PBS protected from light. Tissue was then transferred into 120 mM EDTA pH 7.4 for 48–72 h protected from light. Finally, tissue was transferred to PBS, dissected, and incubated in 1:1000 Phalloidin 405 Invitrogen Cat#A30104, Lot# 2403691.

All tissue was fixed in 4% PFA/PBS for 1 h at room temperature before being removed and replaced with PBS for storage at 4°C. Tissue was later dissected in PBS and freeze/thaw permeabilized in 30% sucrose for up to 30 min on dry ice. Tissue was then blocked and permeabilized further using 5% donkey serum (DS), 1% bovine serum albumin (BSA), and 0.3% triton X-100 in PBS for 30 min. Solution was then removed, and tissue was rinsed three times with PBS/0.05% tween 20. Primary antibodies were prepared in a solution of 5% DS, 1% BSA, 0.3% triton in PBS. The following antibodies were used: Mouse anti-CtBP2 BD Transduction Labs Cat#612044 (1:200), Mouse anti βIII tubulin (Tuj1), Biolegend Cat #801202 Lot#B249869 (1:200), Swant Rabbit anti-Oncomodulin (1:200) Cat#OMG4, Rabbit anti-Myo7a Proteus Biosciences Cat#6790 (1:200), Guinea-pig anti-calretinin Swant Cat# CRgp7 (1:200). Invitrogen Phalloidin 405 (1:1000), Cat#A30104, Lot#2403691, Goat anti-mouse IgG1(y1) 2 mg/mL Alexa 633, Thermo Fisher Cat#A21126, Lot#73B1-1 (1:1000), Goat anti-Rabbit IgG Alexa 488 2 mg/mL, Thermo Fisher Cat#A11008, Lot#2179202 (1:1000), Donkey anti-Mouse Alexa 488 Thermo Fisher Ref#A21202, Lot#45966A (1:1000), Donkey anti-Guinea Pig Alexa 647 Jackson Labs Ref#A30104, Lot#2403691 (1:1000).

Myo7a labeled tissues were imaged at 63X and tiled across all regions. Absolute total hair cell counts were completed using the FIJI (ImageJ) cell counter plugin across utricles, saccules, and cristas by three independent analysts (ER, JL, and MM). Quantification compared means ± standard error of the mean (SEM).

Quantification of CtBP2+ puncta was done with images of two striolar regions and two extrastriolar regions acquired from each utricle and saccule using a 63 × 1.4 NA oil objective lens (Carl Zeiss, z step = 0.36 μm, scaling per pixel: 0.068 μm × 0.068 μm; 2.5× zoom). Approximate striolar and extrastriolar regions were identified using calretinin-staining, which labels striolar calyxes and extrastriolar type II hair cells. For calretinin-positive calyx counts, confocal z-stacks of the entire striolar region were acquired from each utricle and saccule using the 63× oil objective (Carl Zeiss, z step = 0.36 μm, scaling per pixel: 0.099 μm × 0.099 μm; 0.5× zoom). CtBP2(+) puncta and calretinin(+) calyx counts were completed in Imaris Software using 3D projections to assess puncta via the “Spots” tool for each z-stack. Puncta counts were divided by the hair cell number to calculate the average number of CtBP2+ puncta/HC. For calretinin-positive calyx counts, all confocal z-stacks of a given utricle or saccule’s striolar region were aligned into a single image using the Zen Connect feature. All image quantifications were assessed in both males and females with no sex differences identified, and as a result, data were binned together accordingly.

Neonatal mice (P1) were anesthetized via a hypothermic ice bath for 5 min after applying topical lidocaine cream anesthetic Cat# 70512-030. Incision and puncture of the ear capsule were done under sterile conditions. Titers of 1.54 × 10¹³ gc/mL of AAV9-PHP.B-CMV-Tmc1ex1-WPRE, and 6.13 × 10¹³ gc/mL of the AAV9-PHP.B-CMV-eGFP-WPRE were injected into the left utricle at volumes of 1.2 μL per mouse. Injections were done in a small fenestration from a 30G needle (Lot#1253645) using a pulled glass capillary pipette with slow volumetric displacement from a 1 mL syringe pushing through a latex tube. Following injection, the incision site was sutured closed via surgeon’s knot and carefully trimmed. Mice recovered on a heating pad before being returned to the home cage as described previously (18).

All live experiments were conducted following IACUC approval protocol #00001240. Mice were anesthetized with ketamine/xylazine stock injected at a working concentration of 125 mg/kg mouse body weight. After 6–10 min, mice were tested for responses to foot press. A 23-gauge needle was used to pierce the neck skin behind the nuchal crest to thread 5 cm of stripped stainless-steel fiber (A-M Systems, Cat. 791500) and tie a secure loop. The other end was alligator clipped to a reference electrode. When fully anesthetized, mice were placed on their backs with ground electrode placed near left hip bone and head was secured with padded spring clip before placing the recording electrode inferior to the pinna at an angle to prevent movement during acquisition. All cables were taped down flat, and animal body temperatures were monitored using a rectal thermometer. Calibration was done for each mouse head size to ascertain a sensitivity reading ranging from −10 to −14 dB re: g/ms. The mouse’s head was subsequently translated with shaker and monitored by accelerometer increasing in 3 dB step sizes from −22.4 to 1.6 dB re: g/ms. Thresholds were called based on the presence of at least one of three distinct peaks occurring at ~1, 2, and 3 ms as previously described (19).

Tmc1 KO (n = 10), Tmc2 KO (n = 6), Tmc1/2 DKO (n = 10), and WT C57BL6 mice (WT, n = 23) were assessed. VOR responses to sinusoidal head rotation (0.2–4 Hz) (rVORs) and translation (0.2–2 Hz) (tVORs) were recorded utilizing infrared eye tracking (20, 21). Briefly, steady-state VOR responses were measured with horizontal rotations delivered at 0.2, 0.5, 1, 2, and 4 Hz (60°/s peak velocity). Horizontal translations were delivered along 45° right from the nasal-occipital direction (i.e., the translation direction is perpendicular to the visual axis of the left eye) at 0.2, 0.5, 1, and 2 Hz with 0.1 g peak acceleration. A minimum of 30 cycles/50 trials were recorded per condition. Horizontal and vertical signals for eye position, head rotation, and translation, were processed via an analytical pipeline has been described previously (7). All mice assessed were aged P60-P90 under similar conditions.

In the cochlea, there is a developmental switch from Tmc2 to Tmc1 expression during the first postnatal week, whereas, in the utricle, both Tmc1 and Tmc2 are expressed at least until P21 (5). Whether Tmc expression persists into adulthood in vestibular organs has not been clarified. To examine the relative abundance of Tmc1 and Tmc2 expression during postnatal development and into adulthood (P2, P14, P28, P60), utricles and saccules were dissected and analyzed separately via quantitative RT-PCR (qPCR). Tmc2 expression levels were highest in both organs at P2 and declined through P60. Tmc2 levels declined rapidly in the saccule and more gradually in the utricle (Figure 1). Interestingly, as Tmc2 declined, levels of Tmc1 expression rose, were ~1,000-fold higher than Tmc2, and increased as mice matured into adulthood and thus showed a similar but delayed trend compared to the cochlea (Figure 1). These data indicate that early neonatal vestibular development is characterized by a transitory wave of low levels of Tmc2 expression that dissipates during adulthood, while Tmc1 expression steadily rises into adulthood.

At neonatal stages, Tmc1 expression is predominant in the extrastriola, and the same pattern was evident at P28 (5). To further investigate the regional expression patterns of Tmc1 and Tmc2, we used fluorescent in situ hybridization hairpin chain reaction (FISH-HCR) to examine utricular Tmc1 and Tmc2 mRNA at P3. Tmc1 and Tmc2 mRNA signals were absent from Tmc1 KO and Tmc2 KO negative controls, respectively (Figures 2B,D,H,J). In WT utricles, Tmc1 transcripts co-localized with Gfi1 signals in vestibular hair cells (Figures 2A,C,G,I). To quantify regional expression of Tmc1, we estimated the number of HCR puncta within hair cells, defined by phalloidin and Gfi1 regions of interest (ROIs), as previously described (22, 23). Tmc1 expression was normalized to Gfi1 signal per hair cell ROI. Tmc1 mRNA was significantly reduced in the striola versus lateral extrastriola of the utricle (striolar mean ± SEM: 1.19 ± 0.33 vs. extrastriolar: 3.37 ± 0.33, n = 3, p = 0.00002) (Figures 2C′,E). Notably, Tmc2 utricle expression was substantially lower than Tmc1 and also showed slightly elevated extrastriolar expression (Figures 2D,F), which was quantifiably different (striolar mean ± SEM: 1.02 ± 0.08 vs. extrastriolar: 1.38 ± 0.08, n = 3, p = 0.00002).

Subsequently, we sought to address whether Tmc1 and Tmc2 mRNA levels varied in different regions of the saccule (Figures 2G–L). Mean striolar hair cell fluorescence of Tmc1 was higher than extrastriolar levels by FISH (striola: 2.87 ± 0.14; extrastriola:1.84 ± 0.14, n = 3, p = 0.0002) (Figure 2K). The striola of the saccule is bisected by the line of polarity reversal (LPR), where half the striolar hair cells express Emx2, a known enhancer of Tmc1 (23, 24). In contrast, Emx2 is notably absent from the striola of the utricle, which may contribute to the difference in Tmc1 expression between striolas of otolith organs. Mean Tmc2 expression levels were much lower in the saccule than Tmc1, and Tmc2 was enriched in the extrastriolar region (striola: 0.99 ± 0.15; extrastriola: 1.58 ± 0.15, n = 3, p = 0.0002) (Figure 2L). Our data show that Tmc1 expression was much higher than Tmc2 in saccular hair cells but demonstrated a less prominent regional difference in the saccule than in the utricle (Figures 2K,I).

The role of transient Tmc2 expression in early neonatal vestibular development is unclear; however, one possible consequence of the decline in Tmc2 expression in adulthood is an increased dependence on Tmc1, particularly in the saccule where Tmc2 expression declines more rapidly. To test this hypothesis functionally, a fixable styryl dye (FM1-43/FX) was injected intraperitoneally at P60. Since FM1-43/FX enters hair cells through mechanosensory transduction channels (17), it is a useful proxy for visualizing functional hair cells. To compare dye uptake between genotypes, we dissected and imaged inner ear tissue from FM1-43/FX-injected mice. WT mice showed robust FM1-43/FX uptake in hair cells of the utricle, saccule, and semicircular canals (Figures 3A–C), while there was no uptake in any of the vestibular organs excised from Tmc1/2 DKO mice (Figures 3A‴–C‴). At P60, Tmc1 KO saccular hair cells appeared to take up less dye than WT and Tmc2 KO counterparts (Figures 3A–A″). Additionally, at P60 the striola of Tmc2 KO saccules was stained with dye similar to WT, despite a lack of staining at P3 (Figures 3D–D″). Interestingly, P60 Tmc1 KO utricles resembled their P3 counterparts (5) but with even more diffuse dye labeling (Figures 3B–B″). These data suggest that the transitory wave of striolar Tmc2 expression that occurs in neonates is less critical for dye uptake in saccular hair cells than the persistent expression of Tmc1 in adulthood (P60).

In the cochlea, Tmc1 loss of function mutations lead to cochlear hair cell death as early as 1 month of age (5, 8, 25, 26). To investigate the possibility of hair cell loss in the utricle and saccule of Tmc mutant mice, tissues were labeled with Myosin 7a (Figure 4). The total number of hair cells was visually counted for all genotypes in the saccule at P60 and P180. No significant loss was apparent in Tmc1 KO, or Tmc2 KO at P60 or P180 (Figures 4A–A″,B–B″) relative to WT. Despite the apparent necessity of Tmc2 in the striola of the saccule during development, Tmc2 KO, Tmc1 KO, and Tmc1/2 DKO saccules retained normal numbers of type I striolar hair cells based on OCM staining at both P60 and P180 (Figures 5E, 6E). However, Tmc1/2 DKO saccules showed significant loss of hair cells, predominantly in the extrastriolar region internal to the line of polarity reversal (Figure 4B). In P60 Tmc1/2 DKO saccules there were 1,414 ± 124 Myo7a-positive hair cells (n = 3, p = 0.04), while at P180 there were significantly fewer Myo7a-positive cells (661 ± 205, n = 3, p = 0.02). Thus, the combined loss of both genes in the Tmc1/2 DKO mice led to significant loss of hair cells in the saccule but, surprisingly, not in the utricle. The mean number of Myo7a-positive cells in the utricle was unchanged across all genotypes except for Tmc1/2 DKO utricles, which had a slight increase at P60 (1,478 ± 91, n = 3, p = 0.02) and P180 (1,392 ± 78, n = 5, p = 0.04).

Interestingly, the lack of FM1-43 uptake in P3 Tmc2 KO maculas is predominantly in the striola where Tmc1 expression is absent (Figure 3D″). Accordingly, we were curious to assess whether any changes occurred in the cellular architecture in the striolas of Tmc2 KO mice. Oncomodulin (OCM) antibody labeling of type I striolar hair cells and calretinin antibody labeling of striolar calyx-only afferents were used to assess changes in the striolar region of WT, Tmc1 KO, Tmc2 KO, and Tmc1/2 DKO utricles (Supplementary Figure S1A) as previously described (27). The total number of calyx-bearing neuronal endings was evaluated using a βIII tubulin (Tuj1) antibody, which labels the striola domain (Supplementary Figures S1B,C). No difference in the total number of OCM-expressing type I striolar hair cells was apparent across genotypes (Supplementary Figure S1E). Calretinin expression in calyx-only afferents was quantified as a percentage of total calyx-containing βIII tubulin (+) cells within the striolar domain, the latter defined by OCM expression. No differences were identified in calretinin expression between genotypes (Supplementary Figures S1D,F). No differences in the total number of βIII tubulin (+) calyxes in the striola were identified across genotypes (Supplementary Figure S1G). In addition to the total number of neuronal endings and hair cells in the utricular striolas remaining unchanged, there was no change in the size of the striola region normalized to the total area of the sensory epithelia defined by phalloidin (Supplementary Figure S1H). The same analyses were completed at P180, showing no changes in the striola domain (Supplementary Figure S2). Taken together, these data indicate that loss of Tmc1 and/or Tmc2 does not alter the number of type I striolar hair cells or the number of calyceal afferent connections in the utricle.

While the saccules across genotypes showed no change in the total number of OCM(+) hair cells (Figures 5B,E), calretinin expression was diminished in afferent striolar neurons of Tmc2 KO and Tmc1/2 DKO saccules. No differences were observed in Tmc1 KO (Figure 5F). By P60, 47.00 ± 0.08% of striolar calyx-only afferents expressed calretinin, compared to 98.00 ± 0.01% of WT (n = 4, p = 0.0007). Comparable loss of calretinin was observed in Tmc1/2 DKO saccules where 60.00 ± 0.70% of striolar calyxes expressed calretinin (n = 4, p = 0.0012). Although statistically significant, Tmc1 KO calretinin expression was reduced by just 2.70 ± 0.01%, suggesting that it plays less of an effect on striolar afferents than Tmc2 (Figure 5F) (n = 4, p = 0.047). Despite a significant reduction in calretinin expression in Tmc2 KO saccules, the total number of βIII tubulin(+) calyxes in the striola was unchanged, indicating otherwise normal innervation (Figure 5G). Likewise, the total area of the striola was also unchanged across genotypes in the saccule (Figure 5H). Tmc2 appears to be uniquely important for expression of calretinin in striolar calyx-only afferents in the saccule, while Tmc1 is less critical. Examination of P180 saccular striolas revealed that calretinin loss appears to be progressive, with only 28.00 ± 0.04% of Tmc2 KO and 25.00 ± 0.03% of Tmc1/2 DKO striolar calyxes having detectable calretinin by P180 (n = 4, p < 0.0005) (Figures 6C,D,F). Together with the lack of change in calretinin expression in the utricle, these results suggest that transient neonatal expression of Tmc2 in the striola is more critical for the saccule.

Since Tmc1 and Tmc2 expression affects synaptic development and maturation in the cochlea (28), we wondered whether regional expression differences of Tmc1 and Tmc2 may affect presynaptic ribbon distribution within striolar versus extrastriolar hair cells in vestibular macules. Ribbon counts from the approximate striolar and extrastriolar regions were conducted using CtBP2 and calretinin antibody labeling. The regional demarcations are approximate due to the loss of calretinin from Tmc2 KOs and Tmc1/2 DKOs in the saccule. Reduced numbers of CtBP2(+) puncta were observed predominantly in the extrastriola of Tmc1/2 DKO saccules (Supplementary Figures S3C,D). To normalize ribbon counts to regional differences in hair cell densities and variable hair cell survival, the average number of ribbons per hair cell was calculated. Consistent with Myo7a staining, hair cell loss was evident in Tmc1/2 DKO saccules by P180 (Supplementary Figures S3C,D). However, there were no significant differences in average ribbon counts across groups and time points (Supplementary Figures S3E,F). Loss of Tmc1 or Tmc2 does not appear to alter ribbon density in vestibular hair cells.

Tmc1 has been characterized as a necessary gene and protein for hearing and cochlear hair cell survival (5, 25), but less is known regarding its necessity for vestibular function. TMC1 and TMC2 proteins are localized to the tips of stereovilli in vestibular hair cells (29), and recordings from type II extrastriolar hair cells of Tmc1 KO mouse utricles show differences in conductance and Ca²⁺ permeability relative to Tmc2 KO and WT hair cells (6). To evaluate otolith function in Tmc mutant mice, we measured VsEPs (Figure 7A). Notably, Tmc1 KO mice failed to produce VsEP waveforms during head translation events at even the highest stimulus intensity, and thus the mice lacked VsEP thresholds similar to Tmc1/2 DKO mice (Figure 7B). Tmc2 KO mice, on the other hand, produced characteristic VsEP waveforms with mean thresholds that were not significantly different from WT (WT: −14.50 ± 0.65 dB; Tmc2 KO: −13.50 ± 1.94 dB, n = 4, p = 0.55) suggesting Tmc2 is not required for linear VsEPs (Figure 7C).

Interestingly, both Tmc1 KO and Tmc2 KO vestibulo-ocular reflexes exhibited distinct phenotypic differences compared to WT mice. Tmc1 KO mice exhibited decreases in both rotational VOR (rVOR) and translational (tVOR) gains (Figures 7D–F), which assay semicircular canal and otolith function, respectively. In contrast, Tmc2 KO mice exhibited nearly the same tVOR gains as WT mice (Figure 7F). Similar to their failure to elicit VsEP signals, Tmc1 KO and Tmc1/2 DKO mice had significantly reduced tVOR gains relative to WT and Tmc2 KO mice (Figure 7F). Taken together, the VsEP and VOR data suggest that Tmc1 and Tmc2 have distinct contributions to vestibular function; in particular, Tmc1 contributes strongly to otolith organ function in adult mice.

Previously, our lab has shown in vitro adenoviral-mediated delivery of Tmc1 transcripts restored mechanosensory transduction in type II utricle hair cells (5, 6). Additionally, we demonstrated that Tmc2 gene replacement can restore VOR function in Tmc1/2 DKO mice (7). However, no in vivo functional restoration experiments have been evaluated for Tmc1 gene replacement in vestibular hair cells. Thus, we sought to introduce the Tmc1 coding sequence into the vestibular system through left utricular injection of 1 μL of AAV9-PhP.B-CMV-Tmc1ex1-WPRE in P0-P1 mice. Tmc1 KO mice in the negative control group were injected with 1 μL of AAV9-PhP.B-CMV-eGFP-WPRE. Previous work demonstrated that the injection technique and viral capsid efficiently and safely drive expression of GFP in utricle and saccule hair cells (18) and that AAV9-PhP.B-CMV-Tmc1ex1-WPRE restores auditory function in Tmc1 KO mice (8, 26).

After neonatal injection and maturation of the vestibular system, we found robust GFP expression in both saccules and utricles of injected mice at P60 (Figure 8A). AAV9-PhP.B-CMV-Tmc1ex1-WPRE injection restored functional mechanosensory transduction in vestibular hair cells, as evidenced by FM1-43/FX uptake (Figure 8B), which was comparable to WT (Figures 3A–D). To test recovery of vestibular function we recorded VsEPs from injected WT and Tmc1 KO mice (Figure 8C). Results showed that a single inner ear injection of AAV9-PhP.B-CMV-Tmc1ex1-WPRE was sufficient to restore VsEP peaks and thresholds in Tmc1 KO mice, whereas uninjected mice or Tmc1 KO mice injected with AAV9-PhP.B-CMV-eGFP-WPRE had no measurable VsEP responses (Figures 8C,D). WT mice injected with either AAV9-PhP.B-CMV-eGFP-WPRE or AAV9-PhP.B-CMV-Tmc1ex1-WPRE had no change in VsEP thresholds, indicating the injection technique and viral vectors had no deleterious consequences on vestibular function. To track the durability of the response, VsEPs were tested at P60, P120, and P180. VsEPs were not recorded later than P180 due to previous reports that VsEP thresholds are significantly elevated in C57B6 mice, which may not be representative of normal vestibular function (19). Tracking individual mouse thresholds over time showed that gene replacement of Tmc1 through viral delivery effectively restored and maintained thresholds near WT levels for up to 6 months without need for a second treatment (Figure 8E).

In principle, functional restoration of vestibular signaling is expected to correlate with the efficiency of gene delivery to target tissues. Accordingly, we hypothesized that the degree of VsEP threshold recovery in Tmc1 KO mice would vary with Tmc1 expression level following gene replacement therapy. To assess the efficacy of AAV delivery to vestibular hair cells, neonatal (P0-P1) Tmc1 KO mice injected with AAV9-PhP.B-CMV-Tmc1ex1-WPRE were assessed for VsEP recovery at P60, and ears were collected for wholemount FISH HCR analysis (Figures 9A–C). Tmc1 expression level was then plotted as a function of VsEP thresholds level to evaluate correlation between Tmc1 mRNA and the level of functional vestibular recovery (n = 4 mice, 109 striolar and 111 extrastriolar cells). Results showed a linear relationship between VsEP thresholds in Tmc1 KO mice and the amount of normalized Tmc1/Gfi1 in striolar hair cells (R² = 0.97) (Figure 9D). Likewise, extrastriolar levels of normalized Tmc1 correlated with improved recovery of VsEP thresholds (R² = 0.97). No difference was observed between Tmc1 expression in striolar versus extrastriolar regions within each mouse.

These data are consistent with a model in which transient striolar expression of Tmc2 early in neonatal vestibular development at a time point when low-level Tmc1 expression resides primarily in extrastriolar hair cells of the vestibular maculas (Figure 10A). Subsequently, qPCR analysis and FISH HCR imaging show that Tmc2 declines as mouse maculas mature while Tmc1 expression continues to increase and expand into the striolar domain. As a result, the saccule likely becomes more functionally dependent on Tmc1 expression with age. This is supported by FM1-43 dye uptake efficiency diminishing in extrastriolar hair cells over time and the loss of responsiveness to lateral translational movement (tVOR). As a result, Tmc1 KO mice lack sufficient levels of Tmc2 to maintain VsEP responsiveness in adulthood (Figure 10B). In contrast, Tmc2 KO vestibular organs fail to take up FM1-43 initially at P3, but by P60, Tmc1 expression has both increased and expanded into the striolar domain, presumably compensating for the loss of Tmc2 and enabling near WT VsEP and tVOR responses (Figure 10C).