In previous work (Kindt et al., 2021), we established that in sensory organs of the inner ear and lateral-line, Gpr156 expression is limited to HCs. In neonate otolith organs, we demonstrated that GPR156 is required for EMX2-positive HCs to adopt a reversed orientation compared to EMX2-negative HCs. We further showed that EMX2-positive HCs lacking GPR156 do not correct their orientation with time and remain non-reversed in young adults (P21-P28). Focusing on the utricular sensory epithelium, we showed normal macular surface area and number of HCs in all three zones (LES, striola and MES; Figure 1A) in adult Gpr156^del/del mutants. Here we begin by extending our morphological characterization of Gpr156^del/del adult macular organs to include the saccule as well as zonal and HC type profiles, before proceeding to functional and behavioral assessments.

We first verified that absence of GPR156 did not obviously interfere with striolar patterning in adult animals. We used oncomodulin (OCM) to immunolabel striolar HCs (Simmons et al., 2010) at 4 week of age (P28) and confirmed that striolar surface area and striolar HC density were unaffected in constitutive Gpr156^del/del mutant utricles and saccules compared to controls (Gpr156^del/+; Figure 1B–C). Each macular organ contains two main HC types, type I and type II. Therefore, we next quantified type I HCs in the LES, striola and MES zones of the adult utricle and in the anterior (ANT), striolar and posterior (POST) zones of the adult saccule. We used SPP1/Osteopontin, a marker that labels the neck region of approximately 90% of type I HCs (Figure 1D–E; McInturff et al., 2018). We observed no difference in SPP1-positive type I HC numbers between mutants and controls in any zone or organ (Figure 1F). We next used a SOX2 antibody to label type II HCs in the same zones (Figure 1G–H). Again, we observed no difference in type II HC numbers between mutants and controls (Figure 1I). In summary, loss of GPR156 prevents Emx2-positive HCs from reversing their orientation but does not affect macular patterning, including division into clear zones and appropriate numbers of type I and type II HCs throughout.

Gpr156 inactivation could be a powerful model to specifically ask how HC reversal contributes to vestibular function. However, GPR156 may have other confounding roles in HCs besides regulating their orientation, similar to EMX2, which impacts mechanotransduction in zebrafish HCs (Kindig et al., 2023) and afferent innervation in mouse and zebrafish HCs (Ji et al., 2022; Ji et al., 2018). After establishing normal numbers and types of mouse vestibular HCs, we thus first assessed whether HCs respond normally to hair bundle deflections in the absence of GPR156. For this assessment, we recorded whole-cell mechano-electrical transduction (MET) currents of HCs in the utricle of Gpr156^del/del mutants. To stimulate transduction, we deflected individual mechanosensory hair bundles of HCs from Gpr156^del/del mutants and heterozygous Gpr156^del/+ control utricles with a rigid probe coupled to the back (short edge) of the bundle (see Methods, Figure 2A). All bundles were positively deflected along their orientation axes. I_MET recordings were taken between P8 and P100 (median age for comparisons:~P20 for both null and heterozygous genotypes; see Supplementary files 1–5).

Figure 2

Download asset Open asset

We first compared properties of the evoked MET current (I_MET) across genotypes in the LES, the zone in which hair bundles fail to reverse in Gpr156 mutants.~70% (93/135) of LES HCs responded with detectable I_MET (within 20 pA) to hair bundle deflection, regardless of HC type and genotype (Chi-squared test, Supplementary file 1). We attribute the insensitivity of the remaining cells to damage during dissection of the tissue, given that damaged bundles can be seen in every preparation, zone, and bundle type, independent of genotype.

I_MET responses are shown for families of iterated 400 ms steps of displacement; subsets of these families are shown in Figure 2B–E. During bundle displacement, I_MET typically peaked at step onset. As noted before (Vollrath and Eatock, 2003), traces evoked by stimuli in the middle of the operating range adapted with a time course fit by 1–2 decaying exponential terms. Because type I and II HCs were voltage-clamped at different holding potentials (V_hold = –94 and –84 mV, respectively, see Methods), their MET currents experienced different driving forces. We normalized for driving force by converting current to conductance (G) and plotted G as a function of hair bundle displacement (X) for each HC type and genotype (Figure 2F–G). G(X) curves were fit with a simple Boltzmann function (Equation 1, Methods), yielding values for maximum conductance (G_max), the midpoint of the G(X) curve (X_1/2, displacement at half-maximal I_MET) and dx (slope factor), the displacement range corresponding to an e-fold increase in G for small displacements. Among the LES HCs tested, no significant differences were detected between genotypes for any Boltzmann parameter (G_max, X_1/2, or dx) (Figure 2F–G, Supplementary file 2). G_max values were in the range of 2.5–3.0 nS, X_1/2 was 200–300 nm and dx was 150–200 nm (Supp. File 2). From G(X) relations, we estimated the operating range (OR; displacement range yielding 10–90% growth of G_MET) to be ~800–900 nm for all LES HCs (Supp. File 2). In summary, the peak (non-adapted) displacement sensitivity of I_MET was not detectably affected in LES hair bundles of mutants.

Another salient feature of vestibular HC transduction is adaptation (response decay) during a steady hair bundle deflection. We fit the adaptation with one or more exponential decays, and also calculated its extent (% decay; Equation 4, Methods; Figure 2B–E). As reported previously for immature mouse utricular HCs (Vollrath and Eatock, 2003), these mature mouse utricular HCs had fast and slow adaptation components. The fast component was more prominent at small displacements and the slow component dominated at large displacements; often, both were evident in the middle of the operating range. We therefore compared adaptation time constants, τ_fast and τ_slow, at the midpoint of the operating range (X_1/2). For 400 ms steps, mean values are in the range of 4–10ms and 70–140ms, respectively, and the two components are responsible for comparable extents of the total adaptation. These results are consistent with previous reports with similar methods from immature mouse utricular HCs (Vollrath and Eatock, 2003) and immature rat saccular HCs (Songer and Eatock, 2013). In the current sample, however, adaptation at X_1/2 included a significant third, faster adaptation component (‘very fast’; τ_VF < 1 ms) in most type I HCs (12/13) and 2/10 type II HCs. In the LES, we detected no significant differences with genotype in adaptation rates or extent (Supplementary file 2).

MET current initiates the receptor potential, which in turn modulates substantial voltage-gated currents in the HC’s basolateral membrane, especially K⁺ currents, which further shape the receptor potential. For example, type I HCs make relatively fast, small receptor potentials compared to type II HCs because type I HCs have many more K channels that are open around resting potential. To test for effects of Gpr156 deletion on voltage-dependent currents, we applied voltage step protocols to whole-cell clamped type I and type II HCs in the LES of Gpr156^del/del and Gpr156^del/+ utricles (Figure 3A–D).

Figure 3

Download asset Open asset

The time course and voltage dependence of outwardly rectifying voltage-gated current from LES type I cells of Gpr156^del/del utricles resembled those of Gpr156^del/+ utricles (Figure 3A and C). Most of the current in type I HCs (Figure 3A and C) flows through a low-voltage-activated K⁺ conductance (g_K,L) that dominates the type I HC voltage response. As described in Methods, we used the current data to generate conductance density vs. voltage (G/Cm - V) curves (Figure 3E) and fitted them with the Boltzmann function (Equation 5, Methods) to show how the open probability of g_K,L varies with voltage. In both genotypes, approximately half of g_K,L was activated at resting membrane potential, V_rest (~–85 mV). As a consequence of their large K⁺ conductance at V_rest, type I cells had low input resistances (R_in ~50 MΩ). As is typical of type I HCs, the maximal K⁺ conductance density (G_max/C_m), which is proportional to the maximal number of open channels per unit membrane area, was much larger than for type II HCs (Figure 3E vs. 3 F, note difference in conductance scale; Supp. File 3).

Similarly, we detected no effect of genotype on the voltage-dependent outward currents of LES type II HCs (Figure 3B, D and F). The currents activated positive to resting potential (~–72 mV) and had activation midpoints (V_1/2 values) of –30 to –40 mV. Because of the small voltage-gated conductance at rest, input resistance was ~10-fold higher than in type I HCs (~500 MΩ vs. ~50 MΩ). The outward currents inactivated significantly within 400ms, indicating a substantial A-current component. The maximum conductance density (G_max/C_m) at 400ms (~steady state), which is proportional to the number of open channels per unit membrane area, was much smaller than in type I HCs. Again, no effect of genotype was detected on the differences between type I and II HCs of the LES (Supplementary file 3).

Together these results are consistent with previous reports on key physiological properties of type I and II HCs in amniote vestibular organs. The mature physiological status of each HC type appeared to be preserved in LES HCs that failed to reverse their orientation because GPR156 was inactivated.

Loss of GPR156 does not impact HC orientation in the striola and MES, but Gpr156 is transcribed in all HCs across the maculae (Kindt et al., 2021). We thus investigated transduction and adaptation in those zones as well (Figure 4). While there was heterogeneity across HCs, we saw no striking systematic differences with genotype in the voltage-gated or mechanotransduction currents from control vs. Gpr156 mutants (but note that the statistical power of the comparisons is low, Supplementary file 4).

Figure 4

Download asset Open asset

Similar to otolith organs of mammals, each neuromast within the posterior lateral-line system of zebrafish has two populations of HCs that enable the detection of anterior- or posterior-directed fluid flow (Figure 5A). To allow bi-directional detection, both Emx2 and Gpr156 are required to reverse the orientation of one HC population so that it can detect posterior flow (Figure 5A; Jiang et al., 2017; Kindt et al., 2021). Previous work has shown that in wild-type animals, the mechanosensitive properties of HCs that detect anterior flow (Emx2^-) are larger compared to those that detect posterior flow (Emx2⁺; Kindig et al., 2023). Further, this work demonstrated that loss or gain of Emx2 in all lateral-line HCs was able to not only alter HC orientation, but also increase and decrease mechanosensitive responses, respectively (Kindig et al., 2023). Our current results in the mouse utricle did not detect significant differences in the mechanotransduction properties of HCs lacking GPR156 (Figure 2). Whether loss of Gpr156 alters the mechanosensitive properties of lateral-line HCs is not known.

Figure 5

Download asset Open asset

To address this question, we used a membrane-localized GCaMP6s reporter in HCs (memGCaMP6s; Kindt et al., 2021) to perform in vivo calcium imaging (Lozano-Ortega et al., 2018) and by imaging GCaMP6s signals in an apical plane containing the hair bundles, we measured evoked, mechanosensation-dependent calcium signals during fluid-jet stimulation. We used this approach to stimulate in both the anterior and posterior direction and compared the magnitude of the mechanosensitive responses in lateral-line HCs in both control (gpr156^+/-) and gpr156 mutants. As shown previously, we could reliably identify the orientation of control HCs (whether they sense posterior to anterior flow, P to A) or the reverse, A to P using this approach (e.g. Figure 5B). We first verified that in control animals, the magnitude of the mechanosensitive responses of HCs that detect anterior flow was greater than those that detect posterior flow (e.g.: Figure 5B and Figure 5D and F). Consistent with previous work, in gpr156 mutants, we found that the majority of HCs only responded to P to A flow (e.g. Figure 5C; Kindt et al., 2021). We next examined the magnitude of the GCaMP6s response in these P to A HCs in gpr156 mutants. If Gpr156 does not function downstream of Emx2 to alter the mechanosensitive properties of HCs, we predicted that the response magnitude of these HCs would fall between that of P to A and A to P control HCs. However, if Gpr156 does act downstream of Emx2, we predicted that the response magnitude of the P to A HCs in gpr156 mutants would resemble that of P to A HCs in controls and be greater than that of A to P HCs in controls. Importantly, we found that the magnitude of the mechanosensitive responses in gpr156 mutants was comparable to that of control P to A HCs (e.g., Figure 5C and Figure 5E–F). Further, the mechanosensitive responses were significantly larger in gpr156 mutants than control A to P HCs (Figure 5F).

Our analysis thus indicates that in gpr156 mutants, lateral-line HCs show larger mechanosensitive responses. These responses are similar to HCs in control animals that respond to anterior flow (Emx2^-). Together, this data suggests that the EMX2-GPR156 mechanism driving HC orientation reversal also affects the mechanosensitive properties of HCs in zebrafish neuromasts.

After characterizing the functional properties of HCs in mouse and zebrafish Gpr156 mutants, we examined whether GPR156 plays a role in orientation-selective afferent innervation. Previous work has demonstrated that Emx2 expression in the LES of the mouse utricle and in HCs that sense posterior flow in the lateral line is required for afferents to innervate HCs of similar (rather than opposing) orientation (Ji et al., 2022; Ji et al., 2018). Whether GPR156 plays a similar role in direction-selective afferent innervation is unknown.

To investigate whether afferent organization in the LES is disrupted by loss of GPR156, we first compared the terminal fields of individual dye-filled calyx-bearing afferents in Gpr156^del/+ and Gpr156^del/del utricles. All LES calyces belong to dimorphic afferents, which form both calyces on type I HCs and boutons on type II HCs. To examine terminal fields, a fluorescent internal solution was injected into a calyx terminal within the LES during whole-cell recordings. The fluorescent dye was then allowed to diffuse into the calyx and from there throughout the terminal arbor.

Using this approach, we found that terminal arbors (receptive fields) in Gpr156^del/+ control utricles (n=12) comprised 1 or 2 calyces and as many as 65 boutons (median: 29; Figure 6A–C). The labeled afferent terminals were restricted to LES in all but one case (Figure 6A). In this single case, a terminal branch crossed the LPR, innervating 16 striolar HCs with opposite orientation to the 12 LES HCs also innervated. In some filled LES afferents, a thin branch originated below the epithelium (Figure 6C, Figure 6—video 1), as previously described for chinchilla extrastriola (Fernández et al., 1990). We also labeled a single striolar calyx from a control utricle near the striolar/MES boundary; its terminal field was excluded from the calbindin-negative MES (Figure 6D).

Figure 6 with 2 supplements see all

Download asset Open asset

When we performed dye labeling in Gpr156^del/del utricles, the innervation patterns and the size of terminal fields appeared similar to Gpr156^del/+ controls. In Gpr156^del/del utricles, the labeled LES afferents (n=8) were also all dimorphic, terminating in up to three calyces and 56 boutons (median: 23). The labeled receptive fields did not cross the striola/LES border (Figure 6E–G). We also did not observe fiber branches extending across the border from parent axons (Figure 6E–G). As observed in the lone Gpr156^del/+ striolar afferent (Figure 6D), Gpr156^del/del afferents with the main dendritic fiber terminating in the striola (n=4) contacted relatively few HCs, and receptive fields were restricted to the calbindin-positive region (Figure 6H).

To confirm and complement dye-filling results at the whole organ level, we used a new genetic approach to specifically label afferents innervating EMX2⁺ HCs in the LES. Interestingly, when breeding the Advillin-Cre strain (Zhou et al., 2010) with the Tigre Ai140 reporter (Daigle et al., 2018), we only observed Cre recombination and EGFP expression in vestibular ganglion neurons that contacted HCs in the lateral utricle region (Figure 7A). Close examination revealed clear labeling of HC calyces in the LES, and occasionally in OCM⁺ HCs in the lateral striola (Figure 7B, arrowheads). We confirmed that similar to OCM^- EGFP⁺ HCs in the LES, these OCM⁺ EGFP⁺ HCs were invariably oriented towards the medial edge (Figure 7C). This demonstrated that Advillin expression defines a population of neurons whose peripheral processes closely segregate with HC orientation, irrespective of striolar (OCM⁺) or extrastriolar (OCM^-) HC identity. Importantly, HC innervation by Advillin-positive neurons was unchanged in Gpr156^del/del mutants (Figure 7A–B). We observed equivalent numbers of OCM⁺ HCs innervated in the lateral striola in Gpr156^del/del mutants compared to controls (Mann-Whitney p=0.74; average OCM⁺/EGFP⁺ HCs per utricle: Gpr156^del/+: 20.25, Gpr156^del/del: 19.25; three animals and four utricles per genotype; no OCM⁺/EGFP⁺ HCs oriented towards the lateral edge were recorded in either genotype).

Figure 7

Download asset Open asset

In summary, we detected no anomaly in afferent receptive fields in utricles lacking GPR156 and thus lacking a LPR, showing that the tendency of afferents to innervate HCs on one side of the LPR is not controlled by GPR156 or HC orientation.

Our innervation data suggested that HCs in zones demarcated by the LPR are still innervated by distinct populations of afferent neurons in Gpr156^del/del mutants. We next investigated whether the LES afferent population retains normal physiology. Striolar afferents have highly irregular inter-spike intervals, while extrastriolar afferents in both LES and MES have much more regular inter-spike intervals (Goldberg, 1991; Goldberg, 2000; Goldberg et al., 1990). Associated with the difference in spike timing regularity are zonal differences in response dynamics and information encoding capacity (e.g. Jamali et al., 2019). The spike timing difference is considered to reflect differences in afferents’ excitability as conferred by their intrinsic ion channels, with regular afferents being more excitable (Kalluri et al., 2010; Ventura and Kalluri, 2019). To test for effects of Gpr156 inactivation on afferent excitability, we injected small depolarizing current steps into LES calyces through the whole-cell recording pipette and measured current threshold for spiking, I_thresh, and the numbers of spikes evoked by each current step.

As reported (González-Garrido et al., 2021; Ono et al., 2020), I_thresh evoked 1-to-several spikes but increments beyond I_thresh evoked more sustained firing (Figure 8A–B). No difference was observed across genotypes (Supplementary file 5): in both Gpr156^del/+ afferents (n=18) and Gpr156^del/del afferents (n=20), mean I_thresh was ~100 pA and ~95% of afferents produced sustained spiking at 3×I_thresh. These results suggest that afferent firing patterns, and likely the underlying ion channel expression, are unaltered in the LES of Gpr156^del/del utricles, despite abnormal HC orientation.

Figure 8

Download asset Open asset

We next expanded our analysis to determine whether loss of Gpr156 impacts afferent selectivity in the lateral-line. For this analysis, we used mosaic expression of neuroD:tdTomato to label single lateral-line afferent neurons. After identifying gpr156 mutant and control (gpr156^+/-) animals with labeled fibers, we immunolabeled to visualize orientation (phalloidin), HC bodies (MYO7), and cell type (presence or absence of Emx2). We then used confocal microscopy to image the innervation pattern of single afferents.

Consistent with previous results, some afferent fibers selectively contacted Emx2^- HCs with hair bundles oriented to respond to anterior flow in control animals (Figure 9A–H). Other afferent fibers selectively contacted Emx2⁺ HCs with hair bundles oriented to respond to posterior flow (Figure 9—figure supplement 1A–H). We then examined the innervation patterns of afferent fibers in gpr156 mutants. Similar to controls we were able to identify fibers with a clear preference for Emx2^- or Emx2⁺ HCs (Figure 9I–P, Figure 9—figure supplement 1I–P). Importantly, in gpr156 mutants, afferent fibers innervated HCs based on the presence or absence of Emx2, despite all HCs responding to anterior flow.

Figure 9 with 1 supplement see all

Download asset Open asset

Following this initial assessment, we quantified the percentage of HCs per neuromast that were innervated by each afferent fiber. We found that gpr156 mutant fibers innervated a similar percentage of HCs per neuromast compared to controls (Figure 9Q, control: 38.90% ± 2.40, n=18, gpr156 ko: 41.31% ± 2.33, n=23, p=0.48). We also quantified how selective each afferent fiber was for Emx2⁺ or Emx2^- HC. For each afferent fiber, we calculated the percentage of either Emx2⁺ or Emx2^- HCs innervated out of the total number of innervated HCs. In both control and gpr156 mutants, afferent fibers showed a high selectivity (Figure 9R, control: 85.06% ± 4.32, n=18, gpr156 ko: 81.29% ± 3.17, n=23, p=0.48). We split this dataset and examined the selectivity of afferent fibers that preferentially contacted either Emx2^- or Emx2⁺ HCs (Figure 9—figure supplement1S–T). This analysis revealed that in control and gpr156 mutants, both afferent fiber types were highly selective (afferents contacting Emx2^- cells, control: 81.41% ± 7.90, n=8, gpr156 ko: 79.37% ± 4.73, n=10, p=0.82; afferents contacting Emx2⁺ cells, control: 87.99%±4.75, n=10, gpr156 ko: 82.80% ± 4.38, n=13, p=0.43). Overall, our single fiber labeling revealed that Gpr156 is not required for afferent fibers to selectively innervate HCs based on the presence or absence of Emx2.

To extend this analysis, we examined synaptic pairing in gpr156 mutants. Previous research found that along with inappropriate innervation, there is an increase in the amount of unpaired pre- and post-synapses in emx2 mutants (Ji et al., 2018). We immunostained control along with gpr156 and emx2 mutant HCs with Ribeye b to label presynapses and pan-MAGUK to label postsynapses (Figure 10A–C). We then quantified the number of complete immunolabeled synapses per HC. Among all genotypes, the number of complete synapses was unaltered (Figure 10D, control 3.56 ± 0.06, n=9 neuromasts, gpr156 ko: 3.61 ± 0.10, n=16 neuromasts, P=0.74; control 3.51±0.27, n=9 neuromasts, emx2 ko: 3.67±0.08 n=9 neuromasts, p=0.50). We also quantified unpaired pre- and post-synapses. Similar to previous results, there were more unpaired pre- and post-synapses per neuromast in emx2 mutants compared to controls (Figure 9E–F, presynapses, emx2^+/+ or emx2^+/- control 2.00±0.53, n=9 neuromasts, emx2 ko: 4.22±0.70, n=9 neuromasts, p=0.02; postsynapses, control 1.44 ± 0.56, n=9 neuromasts, emx2 ko: 6.56±0.78, n=9 neuromasts, p<0.0001). In contrast, in gpr156 mutants, there was no increase in unpaired pre- or post-synapses per neuromast compared to controls (Figure 10E–F, presynapses, gpr156^+/- control 2.11±0.31, n=9 neuromasts, gpr156 ko: 1.68±0.23, n=16 neuromasts, p=0.29; postsynapses, control 0.33±0.17, n=9 neuromasts, gpr156 ko: 0.75±0.17, n=16 neuromasts, p=0.12). A lack of unpaired pre- and post-synapses in gpr156 mutants further confirms that Emx2 but not Gpr156 impacts afferent innervation in lateral-line HCs.

Figure 10

Download asset Open asset

In summary, Gpr156 inactivation does not appear to impact afferent selectivity in either mouse or zebrafish. This is in sharp contrast with results in Emx2 mutants, where orientation-specific HC afferent innervation is profoundly altered in mouse and zebrafish (Ji et al., 2022; Ji et al., 2018). It follows that HC orientation per se does not dictate afferent selectivity, with EMX2 likely using a different effector(s) than GPR156 in this context. Thus, afferent studies confirm that the Gpr156^del model is well-suited to investigate specifically how the reversal of HC orientation in otolith and lateral line organs serves organ function.

In mammals, only otolith organs (utricle, saccule) have a LPR. Therefore, the absence of a LPR in Gpr156 mutants should only affect otolith-driven reflexes and behaviors. In order to test this prediction we compared Gpr156 control and mutant mice on a number of tests that can discriminate between otolith and canal inputs.

To assess how loss of the LPR affects vestibular functions, we first determined whether Gpr156 mutant mice showed postural/balance impairments. Visual and quantitative assessments did not reveal circling or head tremor behaviors in Gpr156^del/del mice at rest (Figure 11—figure supplement 1). Additionally, we conducted basic standard behavioral tests that involve subjective scoring: (1) tail hanging, (2) air righting, and (3) contact inhibition of righting. Both control and Gpr156^del/del mice scored 0 (where 0=normal, see Methods) on each of these behavioral tests (Figure 11—figure supplement 2A–C).

After this initial assessment, we next subjected mice to more challenging balance tests including walking along a narrow balance beam and swimming. On the balance beam, both Gpr156^del/del and Gpr156^del/+ mice showed normal coordination (Figure 11A). However, when Gpr156^del/del mice were placed in water, they demonstrated severely impaired swimming compared to Gpr156^del/+ controls. This impairment was characterized by underwater tumbling and the inability to maintain normal upright posture in the water. All mice were rescued within 60 s and the average time of rescue was significantly lower in Gpr156^del/del compared to control mice (Figure 11B; Gpr156^del/+ controls = 49.9±3.0s; Gpr156^del/del = 17.4±2.7s; p<0.0001). Overall, these results indicated that GPR156 is required for vestibular function important for swimming behavior.

Figure 11 with 2 supplements see all

Download asset Open asset

We hypothesized that markedly impaired swimming performance could be due to deficits in the otolith rather than semicircular canal dependent pathways. To test this proposal, we quantified eye movements generated by two different vestibulo-ocular reflex behaviors, one driven by activation of the semicircular canals (angular vestibular-ocular reflex, or angular VOR) and the other driven by activation of the otoliths (off-axis vertical rotation, or OVAR).

We first quantified the angular vestibulo-ocular reflex in Gpr156^del/del and Gpr156^del/+ control mice in darkness (VORd). The angular VOR was evoked by rotating the animal sinusoidally around earth vertical axis. To quantify the angular VORd we computed the gain and phase for each testing frequency (0.2, 0.4, 0.8, 1, and 2 Hz,±16°/s). During VORd, both control and Gpr156^del/del mice displayed robust compensatory eye movements, which increased as a function of frequency (Figure 12A–B). Further, to confirm that there was no visual deficit in Gpr156^del/del mice, we also tested their optokinetic reflex (OKR) by rotating a patterned visual surround sinusoidally about earth vertical axis at the same testing frequencies, while the animal remained stationary (see Methods). The OKR responses of Gpr156^del/del and control mice were comparable, confirming that there was no visual deficit in Gpr156 mutants (Figure 12C–D). For the sake of completeness, we quantified the angular VOR evoked in the light (VORl) by rotating the animal sinusoidally around earth vertical axis, while the patterned visual surround remained stationary. Again, we found no difference in the VORl response of Gpr156^del/del versus control mice in this lit environment (Figure 12E–F). Finally, we tested whether Gpr156^del/del mice demonstrated normal ability to adapt their angular VOR. We used a standard angular VOR motor learning protocol (see Methods) in which we carried out 30-minute-long VOR gain-down training for both control and Gpr156^del/del mice. Head restrained mice were rotated with the visual surround in phase. The training stimulus was 2 Hz with a peak velocity of 16°/s. The learning efficacy was assessed by quantifying change in the VOR gain after training. In both control and Gpr156^del/del mice, this VOR motor learning protocol produced a significant reduction in VOR gain after training (Figure 12—figure supplement 1A–B). Overall, Gpr156^del/del and control mice demonstrated a comparable percent change of gain decrease at each testing frequencies (Figure 12—figure supplement 1C). Together, our VOR and OKR results indicated normal semicircular canals-driven responses in Gpr156^del/del mice.

Figure 12 with 1 supplement see all

Download asset Open asset

As no defects were apparent in semicircular canal-mediated behavior, we next tested whether otolith-mediated vestibulo-ocular reflex behaviors were impaired in Gpr156^del/del mice. For this assessment, we recorded eye movements while the mouse was tilted 17° off-the vertical axis (i.e. off-vertical axis rotation (OVAR)), and then rotated at a constant velocity (50°/s for 72 s). The eye velocity evoked by this paradigm comprised two different responses: (1) a transient canal-mediated response that decayed over the first 10–15 s and (2) an otolith-mediated steady-state response in which eye velocity oscillated around a constant bias with a sinusoidal waveform (Figure 12G). Overall, our analysis revealed that transient OVAR responses did not differ between control and Gpr156^del/del mice, supporting the proposal that absence of GPR156 did not affect canal-related ocular responses. In contrast, steady-state OVAR responses demonstrated a significantly impaired bias in Gpr156^del/del mice as compared to controls (Figure 12G–H). The OVAR of Gpr156^del/del mice was characterized by a significantly reduced bias relative to controls (Gpr156^del/del: bias = 0.039 ± 0.26 versus control: bias = 3.85 ± 0.54, p=4.03E-5). Thus taken together, these results show that the absence of HC reversal in Gpr156^del/del alters otolith-mediated OVAR responses. This in turn may explain the swimming deficits observed (Figure 11B).

Taken together, our observations in utricles lacking GPR156 agree with other evidence that in mouse vestibular epithelia, HC orientation is controlled separately from other zonally differentiated properties of either HCs or afferents. Prior evidence includes data from Emx2 knockouts also exhibiting loss of the LPR without altering zoning (Jiang et al., 2017) and data showing that inactivating cytochrome P450 family 26 deletes striolar identity without affecting HC orientation (Ono et al., 2020). This conclusion stands in contrast to zebrafish lateral line results, where transduction properties vary with HC orientation depending on Emx2 (Kindig et al., 2023) and Gpr156 (this study).

Zebrafish lateral-line HCs use a combination of 2 mechanosensitive channels, Tmc2a and Tmc2b (Chou et al., 2017). In the posterior lateral line, Emx2⁺ HCs that sense posterior flow rely on primarily on Tmc2b, while Emx2^- HCs that sense anterior flow rely on both Tmc2a and Tmc2b. Recently, a functional asymmetry was discovered between anterior- and posterior-sensitive HCs: Emx2⁺ HCs that rely primarily on Tmc2b have smaller mechanosensitive responses compared to Emx2^- HCs that rely on both Tmc2a and Tmc2b (Kindig et al., 2023). This work raises the possibility that the transcription factor Emx2 inhibits tmc2a transcription to reduce mechanosensitive responses in HCs sensing posterior flow. In our present study we find that functional asymmetry was also lost in gpr156 mutants (Figure 5). In the context of mouse HC orientation, EMX2 is necessary and sufficient to enrich and polarize GPR156 at HC apical junctions and reverse HC orientation (Kindt et al., 2021). As the Gpr156 receptor is unlikely to act transcriptionally, zebrafish Emx2 may act through Gpr156 to impact Tmc2a/2b function. For example, polarization of Gpr156 at apical junctions may limit scaffolding or transport of Tmc proteins to homomeric channels composed of Tmc2b. Overall, our work in zebrafish indicates that both Emx2 and Gpr156 are important to dictate HC orientation and the mechanosensitive properties of HCs with opposing orientations.

We did not see a clear impact of GPR156 loss on the transduction properties of Emx2⁺ HCs in the mouse utricle (Figures 2—4). Subtle effects remain possible, however, given the variance in single-cell electrophysiological data from both control and mutant mice. Nevertheless, current results are consistent with normal HC function in the Gpr156 mouse mutant, a prerequisite to interrogate how non-reversed HCs affects vestibular behavior.

The brain combines information from the maculae and cristae with inputs from other modalities, including the somatosensory and visual systems, to compute a representation of our self-motion and spatial orientation (reviewed in Cullen, 2019). Our present findings demonstrate that GPR156 is not required for macular HCs to selectively contact different afferents across the LPR. Accordingly, we predicted that HCs that fail to reverse in the lateral utricle of Gpr156 mutants would lead to deficits in paradigms that test macular contributions to behavioral responses. Indeed, Gpr156 mutant mice demonstrated impaired performance on two tests that engage the otolith organs: swimming and off-vertical-axis rotation (OVAR). First, during swimming, impaired spatial orientation results from the combined altered otolith input and reduced somatosensory feedback due to the aqueous environment. Second, Gpr156 mutant mice displayed impaired sustained responses to off-vertical axis rotation (OVAR). While initial eye movement response to OVAR is mediated by stimulation of semicircular canal HCs, the canal response decays such that sustained eye movement is solely driven by the otolith stimulation (Beraneck et al., 2012; Ono et al., 2020). Quantification of the sustained responses revealed a marked reduction in this otolith-mediated reflex response. Confirming that these deficits are specific to otolith function, the Gpr156 mutant mice did not show overt vestibular dysfunction such as spinning or head tilt that can arise with global dysfunction of vestibular epithelia. Further, they displayed normal performance during testing that selectively engaged superior canal organs, namely angular VOR in the light and dark. This latter finding was predicted; because HC orientation is unchanged in Gpr156 mutant cristae, the coding properties of semicircular canal afferents should be unchanged and would therefore produce a normal angular VOR upon specific stimulation of canal ampullae.

Gpr156 does not appear to be significantly expressed in the central nervous system according to the Allen Brain Atlas (Lein et al., 2007). In addition, Advillin-Cre tracing and dye-filling experiments (Figures 6, 7) showed that vestibular ganglion neurons still selectively innervated LES HCs even when lacking GPR156. We thus speculate that HC signals in the LES zone of Gpr156 mutants are relayed to the cerebellum (Ji et al., 2022; Maklad et al., 2010) via the central projection of Advillin⁺ neurons as usual, but with abnormal (non-reversed) response polarity. Overall, it appears unlikely that vestibular deficits observed in our constitutive Gpr156 mouse mutants reflect defective neuronal activity in addition to a missing LPR. That said, we cannot entirely rule out another role of GPR156 beside its regulation of HC orientation. This could be addressed in a follow-up study by limiting Gpr156 inactivation to HCs.

It is interesting to compare vestibular-related behaviors between the Gpr156 and Emx2 mouse models in light of established similarities and differences in protein distribution and function (Ji et al., 2022; Jiang et al., 2017; Kindt et al., 2021). Regarding behavior, this study and work by Ji et al., 2022 agree that vestibular deficits in each model are relatively mild (normal balance beam, VOR and OKR). Emx2 inactivation in HCs (Gfi1-Cre) showed impaired swimming ability, but defects were less severe and different in nature compared to constitutive Gpr156 mutants. Gpr156 mutants displayed substantial difficulty in maintaining their orientation in water, and frequently rolled over such that they needed to be rescued. In contrast, Emx2 mutants did not need rescue and were described as demonstrating ‘frantic’ swimming behavior, spending more time trying to climb out of the water as compared to their littermate controls. OVAR highlighted otolith-specific defects in Gpr156 mutants but was not performed in Emx2 mutants.

GPR156 acts downstream of EMX2 to regulate hair bundle orientation, and perhaps mechanosensitivity in zebrafish. Another significant new result in the current study is to show that afferent projections segregate with HC orientation and not with zonal identity (Figure 7; Advillin neuron labeling), yet segregation does not depend on HC orientation since it is unchanged in mouse and zebrafish Gpr156 mutants (Figures 6—10). It follows that EMX2 must influence afferent segregation (Ji et al., 2022; Ji et al., 2018) via other effectors unrelated to HC orientation. Accordingly, Ji and colleagues concluded that EMX2 regulates HC orientation and afferent patterning independently, with Emx2 expression in support cells critical for proper afferent contacts with HCs (Ji et al., 2022). This aligns well with previous evidence that Gpr156 is only expressed in HCs (Kindt et al., 2021).

Ji and colleagues also used specific inactivation of the transduction protein TMIE in the LES (Emx2-Cre) to silence the Emx2⁺ HC population. Interestingly, as with Emx2 mutants, they did not observe severe behavioral defects, and swimming ability in particular was not obviously compromised in contrast to Gpr156 mutants. It should be noted however that while the Gpr156 KO and Emx2-Cre; Tmie cKO models both lack a normal complement of bidirectional responses, they provide different forms of aberrant linear acceleration signals. In the Tmie cKO model, LES HC inputs to the cerebellum are missing but in the Gpr156 model they are retained and of opposite nature since LES HCs fail to reverse their orientation.

Several considerations could potentially explain why the Gpr156 model appears to display more severe behavioral defects compared to the Emx2 models. First, HC or afferent properties were not tested in Emx2 mutants (Ji et al., 2022) and may be defective. Second, compounded HC orientation and afferent defects in the Emx2 mutants could dampen the deleterious consequence of having LES HCs drive a neuronal response of reversed polarity. Afferent defects in Emx2 mutants were reported during embryogenesis (E16.5) and innervation could further degenerate in adults, further dampening the consequence of non-reversed HC activity in the LES. It is worth noting that aberrant vestibular afferent signals are clinically more deleterious than absent signals, which perhaps explains why properly relayed yet aberrant signals from LES HCs in Gpr156 mutants result in behavioral defects worse than observed in Emx2 or Emx2-Cre; Tmie cKO models.

In conclusion, the Gpr156 mutant mouse model is well-suited to specifically interrogate how HC orientation reversal and the resulting LPR influence vestibular function because it appears to lack confounding afferent patterning defects. In future studies, this model could be improved by limiting Gpr156 inactivation to HCs.

The Gpr156^del strain (C56BL/6N-Gpr156tm1.1(KOMP)Vlcg/J; MGI:5608696) was produced by the Knockout Mouse Project (KOMP) and studied previously in Kindt et al., 2021. The Advillin-Cre strain is B6.129P2-Avil^{tm2(cre)Fawa/J} (MGI: 6196038) (Zhou et al., 2010). The Tigre Ai140 strain is B6.Cg-Igs7^{tm140.1(tetO-EGFP,CAG-tTA2)Hze} (MGI: 5904001; Daigle et al., 2018). Some whole-cell patch clamp recordings were obtained from CD1 mice (Charles River Laboratories USA). All animals were maintained in standard housing and all animal work was reviewed for compliance and approved by the Animal Care and Use Committees of The Jackson Laboratory (anatomical experiments), the University of Chicago (cellular electrophysiology and afferent labeling experiments), and the Johns Hopkins University School of Medicine (behavioral experiments).

Temporal bones of adult animals were isolated, the oval and round windows were cleared and the cochlear apex was punctured to allow better access to the fixative. Samples were then fixed in paraformaldehyde (PFA 4%) overnight at 4 °C, rinsed in PBS and incubated overnight in 4% Ethylenediaminetetraacetic acid (EDTA) at room temperature for decalcification. Both the “trio” (utricle, anterior crista, horizontal crista) and the saccule were dissected from the bone and their epithelium exposed. After dissection, samples were permeabilized and blocked in PBS with 0.5% Triton-X100 and bovine serum albumin (1%) for at least at 1 hr at room temperature. For SPP1/Osteonpontin staining, temporal bones were acutely dissected to isolate the trio and the saccule and expose their epithelium, and an eyelash was used to scrape off the otoconia. Samples were then fixed in PFA 4% for 1 hr at room temperature before being permeabilized and blocked as described previously. Primary and secondary antibodies were incubated overnight at 4 °C in PBS, and fluorescent dye-conjugated phalloidin was added to secondary antibodies. Samples were washed three times in PBS +0.05% Triton-X100 after each antibody incubation before being finally post fixed in PFA 4% for at least 1 hr at room temperature. Primary antibodies used were:

• Mouse anti-MYO7A (Developmental Studies Hybridoma Bank/DSHB 138–1, 1:1000)

• Goat anti-OCM (ThermoFisher Scientific, PA547832, 1:200)

• Goat anti-SPP1 (R&D Systems, AF808, 1:200)

• Rabbit anti-PCNT (Biolegend, PRB432C, 1:400)

• Goat anti-SOX2 (R&D Systems, AF2018, 1:200)

Secondary antibodies were raised in donkey and coupled with AlexaFluor (AF) 488, 555, or 647 (donkey anti-rabbit 488 (A-21206), donkey anti-mouse 647 (A-31571), donkey anti-goat 555 (A-21432); ThermoFisher Scientific). Fluorescent conjugated phalloidin was used to reveal F-actin (CF405 (00034); Biotum).

Confocal images were captured on a line scanning confocal microscope (LSM800) using Zen 2.6 software, the Airyscan detector in confocal mode, and either a 20× or a 63× 1.4 NA oil objective lens (Carl Zeiss AG). Images show a single optical plane unless stated otherwise in the figure legend.

All images were processed using Adobe Photoshop (CC2020), and the same image treatment was applied across genotypes for each experiment. All quantifications included at least three animals of each genotype, and all values plotted in the study, as well as animal cohort size (N), HC number and stereocilia number (n) are detailed in legends. To quantify striolar surface, striolar density, as well as HC Type I and II density by region, images were captured using a DM5500B fluorescence microscope using the Leica Application Suite (LASX) and a 20 x objective (Leica). Striolar surface was measured using the polygon selection and area tools in Fiji/ImageJ based on the Oncomodulin staining. Striolar density was measured by using the cell counter tool in Fiji/ImageJ to obtain the total number of Oconmodulin positive HCs. That number was then normalized by the total striolar surface previously measured to obtain a final number of HCs per 1000 µm². To measure Type I and Type II HC density by region, three regions of interest (ROI) were defined centrally of either 130x50 µm (utricle) or 150x40 µm (saccule). The lateral extrastriolar (LES) domain was placed 20 µm inside the macula starting from the lateral edge of the utricle, encompassing most of the region lateral to the line of polarity reversal (LPR). The striolar domain was defined immediately adjacent to the LES domain based on the LPR and oncomodulin signal in controls. The medial extrastriolar (MES) domain started 50 µm medial to the striolar domain. The anterior extrastriolar domain (ANT) was placed 10 µm from the anterior edge of the saccule. The striolar domain was defined 20 µm posterior to the ANT domain based on oncomodulin staining in controls, and the posterior extrastriolar domain (POST) was defined 50 µm posterior to the striolar domain. The location of these domains is illustrated in Figure 1A. The same fields and spacing were next used for the Gpr156^del/del mutants. The cell counter plugin in Fiji/ImageJ was used to count HCs in each ROI.

All data were plotted in Prism 9 (GraphPad). Striolar surface, HC density and number of Type I and type II HCs in each region of the maculae were plotted individually. Their distribution was framed by 25–75% whisker boxes where exterior lines show the minimum and the maximum, the middle line represents the median, and + represents the mean. Statistical significance was tested using Mann-Whitney U (non-parametric, unpaired t-test) for striolar surface and HC density, and 2-way ANOVA with Sidak’s multiple comparison for type I and type II HCs.

We followed procedures described in González-Garrido et al., 2021 to prepare and record from HCs and calyceal afferent terminals in semi-intact utricles comprising the sensory epithelium and attached distal vestibular nerve including the vestibular ganglion. Recordings were analyzed for 83 cells from 57 Gpr156 heterozygotes and 93 cells from 73 Gpr156-null animals.

Analysis of I_MET properties is described elsewhere (Songer and Eatock, 2013; Vollrath and Eatock, 2003). Briefly, conductance-displacement [G(X)] activation curves were calculated by dividing I_MET (averaged across 2–4 presentations) by the driving force (HP – I_MET reversal potential, 0.2 mV; Corns et al., 2014) and plotting G_MET against probe displacement (X). Probe displacement as a function of driving voltage was calibrated off-line with a photodiode (Songer and Eatock, 2013). G(X) curves were averaged across HCs and fitted with a Boltzmann function (Equation 1):

where G_max and G_min are maximum and minimum MET conductance, respectively, X_1/2 is displacement that evokes half-maximal ${\displaystyle G_{MET}}$ and S is displacement corresponding to an e-fold rise in G_MET. We used the Boltzmann fits to calculate operating range (OR): displacement range corresponding to 10–90% of the response range, ${\displaystyle G_{max}-G_{min}}$.

We characterized transducer adaptation for responses at X_1/2, shown previously to reveal both fast and slow components of adaptation in mouse utricular hair cells (Vollrath and Eatock, 2003). To reduce noise, current responses were averaged (n usually 3) and low-pass filtered at 1.5 kHz. Fast and slow adaptation components were estimated from double-exponential fits of the first 300ms of the step response (Equation 2). Deviation of those fits at the onset response signaled a very fast transducer adaptation component, which was separately estimated with a single-exponential fit of the first 1.5 ms of the response (Equation 3).

Here, τ_vf, τ_f and τ_s are the very short (1ms or less), short (from 1 to 10ms) and long (>10ms) time constants, respectively; A_vf, A_f and A_s are the amplitudes of the corresponding exponential terms; and I_SS is the steady-state current. All HCs had a very fast or fast component or both; some lacked a slow component that was detectable in the 400 ms step.

We calculated extent of adaptation (% decay) at X_1/2 as follows (Equation 4):

where I₀, I_peak, and I_SS are pre-step, steady-state, and peak currents, respectively.

In mouse utricular HCs, the dominant voltage-gated currents are carried by outwardly rectifying K_V channels (Martin et al., 2024). By the end of the first postnatal week, these currents provide distinctive electrophysiological signatures in type I and type II mouse utricular HCs, which include more negatively activating K⁺ currents and more negative resting potentials for type I HCs (Rüsch and Eatock, 1996; Rüsch et al., 1998). To characterize the voltage dependence of the currents, we applied families of iterated test voltage steps (400ms duration) from holding potentials of –74 mV (type I HCs) and –64 mV (type II HCs). The different holding potentials reflect the difference in resting potentials of the two HC types. Steady-state voltage dependence was measured by constructing conductance-voltage [G(V)] relations from tail currents, measured 1ms after the 400 ms test voltage steps, at a common membrane voltage of –39 mV (chosen to reduce contamination of tail currents by HCN currents). Tail current was converted to ‘tail conductance’ by dividing by driving force (V_m – V_rev), where V_rev is approximated by the K⁺ equilibrium potential (–84 mV in our solutions). Tail conductance was plotted against the iterated test step voltage and fit with the Boltzmann function:

where V_1/2 is the voltage of half-maximal activation and S is the voltage change over which G changes e-fold at voltages negative to V_1/2. Voltage was corrected for voltage errors caused by residual R_S and for the liquid junction potential by subtracting 4 mV. Input resistance (R_in) was estimated from current clamp data as the slope of a linear regression of the V_m(I) relation for small currents around resting potential.

To investigate whether afferent receptive fields (terminal arbors) were disrupted by the lack of bundle reversal in the lateral extrastriola (LES), we recorded in whole-cell mode from calyces near the LES-striolar border. To label the terminal arbors for imaging, we included Alexa Fluor 594 (#A10438 Thermo Fisher Scientific) instead of sulforhodamine in the pipette solution (50 micromolar). After whole-cell recording was achieved, we waited 20 min as the fluorescent pipette solution diffused into the calyx and throughout the terminal arbor, then removed the pipette and fixed the utricular epithelium with 4% paraformaldehyde for 10 min at room temperature (~22°). Fixed utricles were immersed in blocking buffer (PBS with 4% normal donkey serum and 0.2% triton) then treated with primary antibodies diluted with blocking buffer overnight at 4 °C. Primary antibodies were:mouse monoclonal anti-βII spectrin (1:500; BD Transduction, #612562), to label hair bundles and show the location of the kinocilium by absence of label; rabbit polyclonal anti-calbindin (1:250; Invitrogen, #711443) to label the striola and juxtastriola (Leonard and Kevetter, 2002). Specimens were then washed extensively and incubated with secondary antibodies conjugated with fluorescent proteins (donkey anti-mouse or rabbit IgG (H + L) antibody, ThermoFisher Scientific) for 1  hr at 4 °C. Stained specimens were mounted whole in mounting medium (Vector Laboratories, #H-1500), imaged with spinning-disc confocal microscopy (Marianas SDC 3i), and analyzed with Adobe Photoshop 2021 and FluoRender (ver. 2.26.3).

To document excitability of afferents in the LES, we recorded from calyces in whole-cell current clamp mode, injected current steps and noted the thresholds for spiking and the firing patterns. Most LES afferents are dimorphic, forming both calyces and boutons on type I and II HCs, respectively. We rejected calyces with resting potentials less negative than −50 mV. We evoked spikes with current steps to determine if afferents had step-evoked firing patterns consistent with expected zonal patterns of spiking regularity in mammalian vestibular epithelia (Goldberg et al., 1990; Iwasaki et al., 2008; Kalluri et al., 2010). Families of 500 ms current steps were applied from zero-current potential, beginning at −200 pA and incremented by 50 pA. Threshold current (I_thresh) was taken as the minimum current (±50 pA) required to evoke any spikes.

We classified neuronal firing patterns according to the number of spikes evoked by a 400 ms step at I_m = 3 × I_thresh (Ono et al., 2020): a pattern of 1–2 spikes at step onset was considered ‘transient’; if more spikes were evoked, the pattern was ‘sustained’. These patterns have been associated with the mechanisms that give rise to well-known differences in the regularity of firing across zones. The transient step-evoked pattern is associated with striolar, irregular afferents and the sustained step-evoked pattern with extrastriolar, regular afferents. The difference in step-evoked firing patterns is attributed to zonal differences in expression of K_V channels that influence spike regularity (Iwasaki et al., 2008; Kalluri et al., 2010) and, more importantly, encoding strategy of afferent fibers (Jamali et al., 2019).

Data analysis, statistical tests, curve fitting and graphing were done with OriginPro (2022; OriginLab, Northampton, MA), Clampfit (Molecular Probes) and Matlab (Mathworks, Natick, MA). For samples that had homogeneous variance, significance was tested with either a two-tailed t-test or a one-way ANOVA with post-hoc Tukey test. When samples had unequal variance, we used either Welch’s t-test or Kruskal-Wallis ANOVA with post-hoc Mann-Whitney U test. For significant differences (p<0.05), we calculated effect size with Hedge’s g equation. For non-significant (NS) differences, we calculated statistical power post hoc (OriginPro 2022) and these values are entered in tables (Supplementary files 1–5) along with p-values.

We tested the angular vestibulo-ocular reflex (aVOR), the optokinetic reflex (OKR), and the otolith vestibulo-ocular reflex (off-vertical axis rotation, OVAR). Specifically, aVOR tests semicircular canal function, whereas OVAR tests otolith function. For completeness we also tested each animal’s OKR as a control of the integrity of visuo-motor function.

Surgical techniques and experimental setup have been previously described (Beraneck and Cullen, 2007). Eye movement data were collected using an infrared video system (ETL-200, ISCAN system). The rotational velocity of the turntable (head velocity) was measured using a MEMS sensor (MPU-9250, SparkFun Electronics). Eye movements during the OKR were evoked by sinusoidal rotations of a visual surround (vertical black and white stripes, visual angle width of 5°) placed around the turntable at frequencies 0.2, 0.4, 0.8, 1, 2, and 3 Hz with peak velocities of ±16°/s. To record VOR responses, the turntable was rotated at sinusoidal frequencies 0.2, 0.4, 0.8, 1, 2, and 3 Hz with peak velocities of ±16°/s in both light and dark. In both conditions, the visual surround remained stationary, but in the dark condition all lighting was extinguished so that surroundings were made dark. Head and eye movement signals were low-passed filtered at 125 Hz and sampled at 1 kHz. Eye position data were differentiated to obtain velocity traces. Cycles of data with quick phases were excluded from the analysis. Least-square optimization determined the VOR and OKR gains, and phases plotted as mean ± standard error of the mean (SEM) against all frequencies for all mice.

Techniques employed for measurement of eye movements during OVAR in alert mice were described elsewhere (Beraneck et al., 2012). Briefly, recordings made after fixating mice on a rotating platform, which was tilted 17° with respect to the ground. The platform speed was increased from 0 to 50°/s in 500ms and maintained its constant velocity for 72 s (10 complete rounds) before being stopped. Eye movements were measured using the same Infrared video system (ETL-200, ISCAN system) used for angular VOR and OKR recordings. Quick phases were identified as previously described and excluded from subsequent analysis. We then estimated the time constant of the OVAR slow-phase eye velocity response decay, as well as the amplitude and phase of its sinusoidal modulation using a linear regression approach.

A large container (26.25×16.25 × 14.38′) was filled with water (24–26°C) at the height of at least 15 cm. The mouse was placed into the water and observed for up to 1 min. Its performance was rated using the following scale (Hardisty-Hughes et al., 2010):

• 0=swims, body elongated, and tail propels in flagella-like motion.

• 1=immobile floating.

• 2=underwater tumbling.

In a time-to-rescue trial, if an animal exhibited distressed swimming (underwater tumbling) before 60 s, the animal was rescued and time of rescue was marked. If not, the trial was marked as lasting 60 s.

A 6-mm-wide and 40-cm-long beam was used for balance beam testing. The mice traversed 40 cm to a dark box. Walking speed was measured by recording the time the mouse took to reach the goal box from the opposite end of the beam. Mice were scored ‘time out’ if they failed to reach the endpoint in 2 min.

The animal was picked up by the tail and lowered to an even surface, and its posture rated using the following scale (Rabbath et al., 2001):

• 0=straight body posture with extension of forelimbs toward the earth (normal).

• 1=slightly bending the body ventrally (intermediate response).

• 2=persistently bending the body (severe response).

The mouse was picked up by the tail and lowered into a small container so that all four feet were in contact with the bottom, then the top of the container closed in contact with the mouse’s back. The container was then quickly inverted so that the mouse became supine while the surface remains in contact with the soles of the mouse’s feet. The mouse’s reflex was rated using the following scale (Rabbath et al., 2001):

• 0=animal rights successfully (normal).

• 1=partial righting (intermediate response).

• 2=complete loss of righting (severe response).

Mice were placed in a cylinder (9 cm diameter and 21.5 cm height) that limited their motion, so mice maintained their steady posture. Head movements in six dimensions were recorded for 2 min using a miniature head motion sensor (MPU-9250, SparkFun Electronics, Niwot, CO, United States) affixed on the top of the skull, which comprises a three-dimensional (3D) accelerometer (measures linear acceleration; fore/aft, lateral, and vertical) and 3D gyroscope (measures angular velocity: roll, pitch, and yaw). Data was acquired at 200 Hz using windows-based CoolTerm software. We then computed the power spectral densities (pwelch function, MATLAB, MathWorks) using Welch’s averaged periodogram with nfft = 4096 and a Bartlett window (4096ms duration) for all six dimensions of movement.

Data are reported as the mean ± SEM. Nonparametric Mann-Whitney U-test was performed to test significance for time to traverse (balance beam) and time to rescue (swimming) between two groups. Two-way repeated-measures ANOVA followed by Bonferroni post hoc comparison tests was used for the VOR data across frequencies. For power spectra analysis we used an independent sample permutation test to test significant differences between the two groups. Prism 9 (GraphPad) or MATLAB was used for statistical analyses. The same independent sample permutation test was also used to test significance for bias and gain of OVAR responses.

Zebrafish (Danio rerio) were grown at 30  °C using a 14 hr light, 10 hr dark cycle. Larvae were raised in E3 embryo medium (5  mM NaCl, 0.17  mM KCl, 0.33  mM CaCl₂, and 0.33  mM MgSO₄, pH 7.2). Larvae were examined at 5–6 days post fertilization (dpf). The following previously established mutant and transgenic zebrafish strains were used in this study: gpr156^idc15, emx2^idc5 (Jiang et al., 2017; Kindt et al., 2021), Tgmyo6b:memGCaMP6s^idc1 (Kindt et al., 2021). For comparisons, emx2 mutants were compared to wild-type or emx2^+/- animals. Gpr156 mutants were compared to gpr156^+/- animals. Zebrafish work was performed at the National Institute of Health (NIH) and approved by the Animal Use Committee at the NIH under animal study protocol #1362–13.

Immunohistochemistry was performed on whole larvae at 5 dpf. Whole larvae were fixed with 4% paraformaldehyde in PBS at 4 °C for 3.5 hr. For pre- and post-synaptic labeling all wash, block and antibody solutions were prepared with 0.1% Tween in PBS (PBST). For Emx2 labeling performed on sparse afferent labeling (see below) all wash, block and antibody solutions were prepared with PBS  + 1% DMSO, 0.5% Triton-X100, 0.1% Tween-20 (PBDTT). After fixation, larvae were washed 4×5 min in PBST or PBDTT. For synaptic labeling, larvae were permeabilized with Acetone. For this permeabilization larvae were washed for 5 min with H₂O. The H₂O was removed and replaced with ice-cold acetone and placed at −20 °C for 5 min, followed by a 5 min H₂O wash. The larvae were then washed for 4×5 min in PBST. For all immunostains larvae were blocked overnight at 4 °C in blocking solution (2% goat serum, 1% bovine serum albumin, 2% fish skin gelatin in PBST or PBDTT). Larvae were then incubated in primary antibodies in antibody solution (1% bovine serum albumin in PBST or PBDTT) overnight, nutating at 4 °C. The next day, the larvae were washed for 4×5 min in PBST or PBDTT to remove the primary antibodies. Secondary antibodies in antibody solution were added and larvae were incubated for 2 hr at room temperature, with minimal exposure to light. Secondary antibodies were washed out with PBST or PBDTT for 4×5 min. Larvae were mounted on glass slides with Prolong Gold (ThermoFisher Scientific) using No. 1.5 coverslips. Primary antibodies used were:

• Rabbit anti-Myo7a (Proteus 25–6790; 1:1000)

• Mouse anti-Ribeye b (IgG2a) (Sheets et al., 2011)

• Mouse anti-pan-MAGUK (IgG1) (Millipore MABN7; 1:500)

• Mouse anti-Myo7a (DSHB 138–1; 1:500)

• Rabbit anti-Emx2 (Trans Genic KO609; 1:250)

The following secondaries were used at 1:1000 for synaptic labeling: goat anti-rabbit Alexa 488, goat anti-mouse IgG2a Alexa 546, goat anti-mouse IgG1 Alexa 647, along with Alexa 488 Phalloidin (Thermofischer; #A-11008, #A-21133, #A-21240, #A12379). For Emx2 co-labeling the following secondaries were used at 1:1000: goat anti-mouse Alexa 488, and goat ant-rabbit Alexa 647, along with Alexa 488 Phalloidin (Thermofischer; #A12379, #A28175, #A27040).

Fixed samples were imaged on an inverted Zeiss LSM 780 laser-scanning confocal microscope with an Airyscan attachment (Carl Zeiss AG, Oberkochen, Germany) using an 63×1.4 NA oil objective lens. Airyscan z-stacks were acquired every 0.18  µm. The Airyscan Z-stacks were processed with Zen Black software v2.1 using 2D filter setting of 6.0. Experiments were imaged with the same acquisition settings to maintain consistency between comparisons. Processed imaged were further processed using Fiji.

Images were processed in ImageJ. Researcher was not blinded to genotype. Hair bundle orientation was scored relative to the midline of the muscle somites. HC number per neuromast were quantified based on Myo7a labeling and presence of a paired/complete synapse. For quantification of Emx2 labeling, HCs were scored as Emx2 positive if they labeled with both Emx2 and Myo7a. To quality as a ribbon or presynapse, the following minimum size filters were applied to images: Ribeye b: 0.025 μm², MAGUK: 0.04 μm². A complete synapse was comprised of both a Ribeye b and MAGUK puntca. An unpaired presynapse consisted of only a Ribeye b puntca, while an unpaired postsynapse consisted of only a MAGUK puncta. In each neuromast all HCs (~15 HC per neuromast) were examined for our quantifications.

GCaMP6s-based calcium imaging in zebrafish hair bundles has been previously described (Lozano-Ortega et al., 2018). Briefly, individual 5–6 dpf larvae were first anesthetized with tricaine (0.03% Ethyl 3-aminobenzoate methanesulfonate salt, SigmaAldrich). To restrain larvae, they were then pinned to a Sylgard-filled recording chamber. To suppress the movement, alpha-bungarotoxin (125 μM, Tocris) was injected into the heart. Larvae were then rinsed and immersed in extracellular imaging solution (in mM: 140 NaCl, 2 KCl, 2 CaCl₂, 1 MgCl₂ and 10 HEPES, pH 7.3, OSM 310+/-10) without tricaine. A fluid jet was used to mechanically stimulate the apical bundles of HCs of the A-P neuromasts of the primary posterior lateral-line. To stimulate the two orientations of HCs (A to P and P to A) a 500ms ‘push’ was delivered. Larvae were rotated 180° to deliver a comparable ‘push’ stimulus to both the A to P and P to A HCs.

To image calcium-dependent mechanosensitive responses in apical hair bundles, a Bruker Swept-field confocal system was used. The Bruker Swept-field confocal system was equipped with a Rolera EM-C2 CCD camera (QImaging) and a Nikon CFI Fluor 60×1.0 NA water immersion objective. Images were acquired in 5 planes along the Z-axis at 0.5 μm intervals (hair bundles) at a 50 Hz frame rate (10 Hz volume rate). The 5 plane Z-stacks were projected into one plane for image processing and quantification. The method to create spatial ∆F heatmaps has been described (Lozano-Ortega et al., 2018). For GCaMP6s measurements, a circular ROI with a~1.5 μm (hair bundles) diameter was placed on the center of each individual bundle. The mean intensity (∆F/F₀) within each ROI was quantified. F₀ represents the GCaMP6s intensity prior to stimulation. We examined the GCaMP6s signal in each hair bundle to determine its orientation. The GCaMP6s responses for each neuromast were averaged to quantify the magnitude of the A to P and P to A responses.

To visualize the innervation pattern of single afferent neurons, a neuroD1:tdTomato plasmid was injected into zebrafish embryos at the one-cell stage. This plasmid consists of a 5 kb minimal promoter, neurod1, that drives tdTomato expression in lateral-line afferents (Ji et al., 2018). This plasmid was pressure at a concentration of 30 ng/µl. At 3 dpf larvae were anesthetized with 0.03% ethyl 3-aminobenzoate methanesulfonate (Tricaine), to screen for tdTomato expression. Larvae were screened for mosaic expression of tdTomato expression in the lateral-line afferents using a Zeiss SteREO Discovery V20 microscope (Carl Zeiss) with an X-Cite 120 external fluorescent light source (EXFO Photonic Solutions Inc). After selecting larvae with tdTomato expression, larvae were prepared for immunostaining at 5 dpf, and imaged as outlined above.

Statistical analyses and data plots were performed with Prism 9 (Graphpad, San Diego, CA). Values of data with error bars on graphs and in text are expressed as mean ± SEM unless indicated otherwise. All experiments were performed on a minimum of three animals, six neuromasts (primary posterior lateral-line neuromasts with A-P orientations: L1-L5). For 5–6 dpf larvae, each neuromast represents analysis from 12 to 20 HCs and 41–68 synapses. All replicates are biological. No animals or samples were excluded from our analyses unless control experiments failed–in these cases all samples were excluded. No randomization or blinding was used for our animal studies. Where appropriate, data was confirmed for normality using a D’Agostino-Pearson normality test. Statistical significance between conditions was determined by an unpaired t -test.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

This work was supported by the National Institute on Deafness and Other Communication

Disorders (NIDCD) grant R01DC018304 (to RAE, KEC and BT), R01DC015242 (to BT) and NIDCD Intramural Research Program Grant 1ZIADC000085-01 (to KK).

All mouse work was reviewed for compliance and approved by the Animal Care and Use Committees of The Jackson Laboratory (anatomical experiments), University of Chicago (cellular electrophysiology and afferent labeling experiments) and JohnsHopkins University School of Medicine (behavioral experiments). Zebrafish work was performed at the National Institute of Health (NIH) and approved by the Animal Use Committee at the NIH under animal study protocol #1362-13.

1. Sent for peer review: March 26, 2024
2. Preprint posted: March 29, 2024
3. Reviewed Preprint version 1: June 18, 2024
4. Reviewed Preprint version 2: October 9, 2024
5. Version of Record published: November 12, 2024

You can cite all versions using the DOI https://doi.org/10.7554/eLife.97674. This DOI represents all versions, and will always resolve to the latest one.

This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.