ACh-induced hyperpolarization and decreased resistance in mammalian type II vestibular hair cells

Abstract

In the mammalian vestibular periphery, electrical activation of the efferent vestibular system (EVS) has two effects on afferent activity: 1) it increases background afferent discharge and 2) decreases afferent sensitivity to rotational stimuli. Although the cellular mechanisms underlying these two contrasting afferent responses remain obscure, we postulated that the reduction in afferent sensitivity was attributed, in part, to the activation of α9- containing nicotinic acetylcholine (ACh) receptors (α9*nAChRs) and small-conductance potassium channels (SK) in vestibular type II hair cells, as demonstrated in the peripheral vestibular system of other vertebrates. To test this hypothesis, we examined the effects of the predominant EVS neurotransmitter ACh on vestibular type II hair cells from wild-type (wt) and α9-subunit nAChR knockout (α9^−/−) mice. Immunostaining for choline acetyltransferase revealed there were no obvious gross morphological differences in the peripheral EVS innervation among any of these strains. ACh application onto wt type II hair cells, at resting potentials, produced a fast inward current followed by a slower outward current, resulting in membrane hyperpolarization and decreased membrane resistance. Hyperpolarization and decreased resistance were due to gating of SK channels. Consistent with activation of α9*nAChRs and SK channels, these ACh-sensitive currents were antagonized by the α9*nAChR blocker strychnine and SK blockers apamin and tamapin. Type II hair cells from α9^−/− mice, however, failed to respond to ACh at all. These results confirm the critical importance of α9nAChRs in efferent modulation of mammalian type II vestibular hair cells. Application of exogenous ACh reduces electrical impedance, thereby decreasing type II hair cell sensitivity.

NEW & NOTEWORTHY Expression of α9 nicotinic subunit was crucial for fast cholinergic modulation of mammalian vestibular type II hair cells. These findings show a multifaceted efferent mechanism for altering hair cell membrane potential and decreasing membrane resistance that should reduce sensitivity to hair bundle displacements.

INTRODUCTION

Vestibular efferent terminals appose type II hair cells, afferent bouton terminals, afferent fibers, and the outer surface of calyx terminals (Lysakowski and Goldberg 1997, 2004; Meredith and Roberts 1986; Purcell and Perachio 1997; Smith and Rasmussen 1968). There is compelling anatomical and physiological evidence in both mammalian and nonmammalian vertebrates that ACh is the primary neurotransmitter released by vestibular efferents (Guth et al. 1994; Holt et al. 2011; Housley and Ashmore 1991; Jordan et al. 2015). Vestibular efferent neurons express choline acetyltransferase (ChAT) (Jordan et al. 2015; Kong et al. 1998) and acetylcholinesterase (Carpenter et al. 1987; Gacek and Lyon 1974; Goldberg and Fernández 1980; Hilding and Wersall 1962) while ACh receptor antagonists block responses evoked by efferent stimulation (Holt et al. 2006, 2015a; Rossi et al. 1980; Sugai et al. 1992). Conversely, when cholinergic agonists are applied to vestibular organs, the responses mimic those observed in efferent stimulation experiments (Guth et al. 1986; Holt et al. 2001, 2003).

Vestibular efferent stimulation has a profound impact on afferent nerve activity. In mammals, efferent stimulation increases background discharge rates of vestibular afferents (Goldberg and Fernández 1980; Holt et al. 2015b; Marlinski et al. 2004; McCue and Guinan 1994) while reducing the sensitivity or gain of afferents to vestibular stimulation (Goldberg and Fernández 1980). Although the precise mechanism(s) underlying the increased background discharge rate has not been positively identified, the reduction in sensitivity is putatively due to the activation of ACh receptors on vestibular type II hair cells (Holt et al. 2011; Kong et al. 2005, 2007; Rabbitt et al. 2009). Similar observations have been made in the cochlea (Doi and Ohmori 1993; Geisler 1974; Glowatzki and Fuchs 2000; Nenov et al. 1996) and in type II hair cells of the toadfish (Boyle et al. 2009).

Several nicotinic ACh receptor subunits, including α9, α10, α4, and β2, are expressed by hair cells and afferents in the inner ear (Elgoyhen et al. 2001; Holt et al. 2003, 2015a; Luebke et al. 2005). α9-containing nicotinic ACh receptors (α9*nAChRs) are critical for efferent inhibition of auditory hair cell and auditory afferent activity (Vetter et al. 1999). Based on data from auditory hair cells, the “two-channel hypothesis” (Fuchs and Murrow 1992a, 1992b) proposes that ACh initiates a Ca²⁺ influx via α9*nAChRs that subsequently activates K⁺ efflux and resistance reduction via small-conductance, Ca²⁺-sensitive K⁺ (SK) channels (Elgoyhen and Katz 2012; Oliver et al. 2000), and in some cases by opening large-conductance (BK) channels (Rohmann et al. 2015). Opening of Ca²⁺-sensitive K⁺ channels lowers the membrane resistance and shunts the receptor current, thereby reducing sensitivity to hair bundle displacements (Geisler 1974; Rabbitt et al. 2009; Wiederhold and Peake 1966). In the mammalian vestibular system, it is unknown if α9*nAChRs and SK channels are also required for efferent control of hair cell sensitivity, but preliminary data suggest that mechanisms similar to auditory hair cells exist in mammalian type II vestibular hair cells (Kong et al. 2005; Poppi et al. 2014, 2015; Yu et al. 2015).

In this study we investigated cholinergic mechanisms at the efferent/vestibular type II hair cell synapse. We used α9-subunit knockout (α9^−/−) mice to confirm the critical importance of α9-subunit expression for efferent function in type II hair cells. ChAT immunohistochemistry was used to examine peripheral EVS morphology, whereas patch-clamp methodologies and pharmacology were utilized to characterize the response of type II hair cells to ACh. Similar to auditory hair cells, mouse type II vestibular hair cells respond to ACh using a two-channel mechanism. These results support the hypothesis that α9*nAChR-mediated reduction in type II hair cell membrane resistance contributes to the decreased sensitivity to motion stimuli observed in classical mammalian afferent recordings (Goldberg and Fernández 1980) following activation of the EVS.

MATERIALS AND METHODS

Ethics statement.

Prospective approval of all protocols and procedures on animal subjects was obtained from the University of Newcastle Animal Care and Ethics Committee (Animal Research Authority approval no. A-2013-325). Immunofluorescence work at the University of Rochester was done using fixed tissue samples originating from the University of Newcastle.

Wild-type and knockout mice.

In this study, we used two wild-type (wt) strains, CBA/CaJ × 129SvEvTac (CBA) and C57BL/6, to compare the effect of genetic background (i.e., CBA/CaJ;129 vs. the readily available inbred strain C57BL/6). The CBA strain was chosen because it is genetically similar, although not the same, as the background of the noncongenic homozygous α9-subunit knockout (α9^−/−) strain provided by The Jackson Laboratory (CBACaJ;129S-Chrna9tm1Bedv; RRID:IMSR_JAX:005696). The α9^−/− genotype was confirmed using standard PCR and the primer sequences recommended for this strain by The Jackson Laboratory.

Immunofluorescent labeling of cholinergic terminals.

Mice (all strains, either sex, age 3–6 wk) were anesthetized with ketamine (100 mg/kg ip). The anterior canals were clipped to allow fluid entry, and the bony labyrinths were drop-fixed with fresh 4% paraformaldehyde for a minimum of 2 h at 4°C. Organs were placed in 0.1 M phosphate-buffered saline (PBS) and rinsed several times over the course of 24 h. For sections, cristae were dissected free, embedded in 4% agarose gel (Sigma), and sliced at 40 µm on a Leica CM1950 cryostat. Slices were then blocked with 10% normal goat serum (Sigma) and 0.5% Triton X-100 in 0.1 M PBS for 1 h. Tissue was incubated overnight in primary antibodies against ChAT (1:200 in 0.1 M PBS; Millipore catalog no. AB144P, lot no. 2280814; RRID:AB_11214092) and myosin VIIa (1:200 in 0.1 M PBS; Proteus catalog no. 25-6790). Samples were washed with 0.1 M PBS before the addition of anti-rabbit FITC (1:50 in PBS; Jackson IR catalog no. 711-095-152, lot no. 120791) and anti-goat Texas red (1:50 in PBS; Sigma catalog no. SAB3700290, lot no. RI32452) and reacted for 2 h at room temperature. After being washed with 0.1 M PBS, samples were washed with distilled water, mounted onto Superfrost Plus slides with SlowFade gold antifade mounting medium (ThermoFisher catalog no. P36930). Slides were subsequently imaged on a Nikon Eclipse C1 confocal laser scanning microscope. Low-power images were acquired using a ×40 objective, z-stack images were acquired with a ×100 Plan-Apo oil-immersion lens (z-slice thickness ~0.2 µm). Two-dimensional images were prepared using ImageJ software, and three-dimensional renders were prepared in FluoRender (Wan et al. 2017).

ChAT immunohistochemistry and terminal morphology were also examined in crista whole organs. Similarly to sections, whole crista were incubated overnight in antibodies against ChAT (1:100 in 0.1 M PBS; Millipore catalog no. AB144P, lot no. JC1618187; RRID:AB_11214092), followed by five 5-min washes with 0.1 M PBS before the addition of Alexa Fluor 488 donkey anti-goat secondary antibody (Invitrogen catalog no. A-11055; RRID:AB_2534102) for 2–4 h at room temperature. After another washing 5 × 5 min with 0.1M PB, DAPI (1 μg/ml) was added for 5 min. Samples were washed with 0.1M PB and mounted onto slides as detailed above. Whole crista slides were imaged on a FV1000 confocal laser scanning microscope, within the University of Rochester’s Confocal Shared Resource. Z-stack images were acquired with a ×100 Plan-Apo oil-immersion lens (z-slice thickness 1 μm). Most whole organ image stacks were 70–100 μm in total depth. To prevent counting and measuring the same varicosities or terminals (puncta) within a single z-plane, every 10th z-section was used for puncta quantification. On average, seven z-plane images per organ were exported as TIFF files and then opened in the Image Pro Plus analysis program (RRID:SCR_007369), where a manual threshold was set to select the varicosities and exclude the background. ChAT-positive puncta number and area were measured in Image Pro Plus using the count/measure feature within one to four non-overlapping, 400-μm² areas of interest manually placed on the section image. To reduce extraneous noise, the lower size of the puncta area was set at 0.25 μm, and following counting, the watershed algorithm was applied to help delineate between puncta that lay close to one another. Automated counts were comparable to blinded manual counts taken from same selected section. Data were exported into Microsoft Excel and GraphPad Prism (RRID:SCR_002798) for graphing and statistical analysis, respectively.

Patch-clamp recordings and ACh exposure.

Mice (all strains, either sex, age 3–6 wk) were anesthetized with ketamine (100 mg/kg ip). Ketamine is an NMDA receptor antagonist that we, and others, have found to promote neuronal viability in in vitro preparations (de Oliveira et al. 2010). Once deeply anesthetized, mice were decapitated and the bony labyrinth was isolated in ice-cold glycerol-modified Ringer’s solution containing (in mM) 26 NaHCO₃, 11 glucose, 250 glycerol, 2.5 KCl, 1.2 NaH₂PO₄, 1.2 MgCl₂, and 2.4 CaCl₂ and then bubbled with carbogen gas (pH 7.4; Ye et al. 2006). The vestibular triad, a semi-intact organ preparation comprising the anterior and horizontal crista and utricle (Fig. 1A), was excised, as described previously (Lim et al. 2011). The membranous roof overlying the neuroepithelium was removed, allowing access for whole cell patch-clamp recordings. The triad was transferred to the recording chamber and continually perfused with oxygenated Leibovitz’s L-15 cell culture medium (pH 7.55, osmolality 305 mmol/kg; Life Technologies Australia) at a rate of four to six bath exchanges per minute. Whole cell patch-clamp recordings were obtained from type II vestibular hair cells in the anterior and horizontal cristae (Fig. 1B). Type II hair cells receive direct efferent input. Recording pipettes were made from borosilicate glass (3–4 MΩ; King Precision Glass, Claremont, CA) filled with an internal solution containing (in mM) 42 KCl, 98 K-gluconate, 4 HEPES, 0.5 EGTA, 1 MgCl₂, and 5 Na-ATP. In a subset of experiments, 10 mM BAPTA tetrapotassium salt (ThermoFisher catalog no. B-1204) was substituted for EGTA in the internal solution. All experiments were done at room temperature (22°C). Type II hair cells were identified by their cylindrical profile, characteristic voltage-gated currents, and the absence of a low-voltage-activated K⁺ conductance (G_K,L).

Fig. 1.

Semi-intact preparation of the mouse vestibular organs and cholinergic efferent varicosities. A: semi-intact preparation of the mouse vestibular triad: anterior (AC) and horizontal cristae (HC), utricle (U), and vestibular nerve (VIII). Recording (R) and perfusion (P) pipettes are shown. Scale, 300 µm. B: infrared differential interference contrast optics image showing an expanded view of the recording pipette (R) in contact with a type II hair cell and a perfusion pipette (P) filled with 300 µM ACh dissolved in L-15. Asterisk indicates eminentia or torus of the anterior crista. Scale, 25 µm. C–E: low-power (×40) micrographs of canal cristae from all 3 strains immunostained with antibodies to choline acetyltransferase (ChAT; red) and myosin VIIa (green). Scale bars, 50 µm. F–H: higher power (×100) z-projections of canal cristae from all 3 strains immunostained with antibodies to ChAT (red) and myosin VIIa (green). Scale bars, 10 µm. I–K: higher power (×100) 3-dimensional FluoRender images of hair cells and cholinergic terminals from all 3 strains. Scale bars, 5 µm in planes x, y, z.

ACh was applied within 10–20 µm of the recorded cell via a picospritzer-driven pipette (Fig. 1B) at concentrations of 100 µM, 300 µM, and 1 mM (Sigma catalog no. A6625). The picospritzer was programmed to deliver a bolus of ACh over 100 ms, to elicit reproducible ACh responses and minimize postsynaptic receptor desensitization. Intrinsic membrane properties, voltage-activated currents, and baseline holding current (or membrane potential) of type II hair cells were recorded before, during, and after ACh exposure. Recordings where series resistance (R_s) exceeded 20 MΩ or changed by >20% during the course of the recording session were excluded from analysis. Holding potentials varied from −96 to +4 mV to determine the voltage dependence of ACh-evoked currents. Antagonists (strychnine, apamin, Sigma-Aldrich, Castle Hill, Australia; iberiotoxin, Abcam, Melbourne, Australia; and tamapin, Alomone, Jerusalem, Israel) were bath applied. Data were collected with Multiclamp 700B or Axopatch 1D amplifiers. Signals were sampled at 20 kHz, filtered at 10 kHz, and digitized using Instrutech ITC16/USB16 analog-to-digital boards. R_s was compensated by 60% in all recordings, and voltages were corrected for liquid junction potentials (~6 mV for KCl-gluconate and ~8 mV for BAPTA-containing internals). Data were acquired and analyzed using AxoGraphX software (AxoGraphX, Sydney, Australia) and Igor Pro 6.3 (WaveMetrics, Lake Oswego, OR).

Membrane properties: resistance and capacitance measurements.

We also measured R_s, membrane resistance (R_m), and membrane capacitance (C_m). This was achieved using both time-domain and frequency-domain methods. In the time domain, we applied a standard 5-mV hyperpolarizing step pulse at the beginning, during, and at the completion of each cell recording to ensure recording conditions remained constant. In a subset of experiments, we continuously recorded R_m before, during, and after ACh exposure by superimposing one of two protocols onto the cell’s holding potential. The time-domain protocol consisted of alternating ±5-mV step pulses (3-ms duration, 2-ms separation). This protocol was used to monitor R_m every 10 ms. The frequency-domain protocol consisted of a summation of three interrogation voltage sine waves (325, 525, 725 Hz) with a combined maximum peak-to-peak amplitude of 15 mV (±7.5 mV) superimposed on top of the voltage command as described previously (Rabbitt et al. 2016). Similar to other methods (Farrell et al. 2006; Santos-Sacchi 2004), Fourier analysis of the current and voltage perturbations at the three interrogation frequencies was used to determine total access resistance R_s and whole cell membrane impedance (effective capacitance C_m and resistance R_m) during the voltage command protocol. Briefly, the real and imaginary Fourier components of the current and voltage were extracted from the time-domain data using a sliding Hanning window (over 15 periods, 40 ms) for each of the three interrogation frequencies. Fourier components at the three interrogation frequencies were subtracted from the voltage command and whole cell current when whole cell current was displayed. Data collected by using a model cell with known electrical properties were used to determine the frequency-dependent transfer function necessary to calibrate the instrumentation for the specific voltage-clamp protocols used. Separate transfer functions were calculated for the Axopatch and Multiclamp amplifiers and associated hardware. The method was verified by recording from alternative model cells and comparing the frequency-domain Fourier analysis results to transient resistance and capacitance responses and known model cell parameters. The use of three perturbation frequencies provides an overdetermined set of equations (6 degrees of freedom) to estimate the three unknown parameters. Parameter estimation was done offline using custom nonlinear least squares software (Igor Pro 6.3).

RESULTS

Presence of cholinergic efferent terminals in wt and α9^−/− mice.

In this study, we used a semi-intact preparation of the anterior and horizontal cristae (Fig. 1, A and B) that preserves much of the microstructure of the vestibular neuroepithelium (Lim et al. 2011). Cristae were collected from two wt strains (C57BL/6 and CBA) and α9^−/− animals and were processed for ChAT immunohistochemistry. In both wt (C57BL/6 and CBA) and α9^−/− cristae, ChAT immunohistochemistry showed an extensive network of fibers and small en passant and terminal spherical varicosities (Fig. 1, C–H), qualitatively similar to observations made in other mouse strains (Jordan et al. 2015; Luebke et al. 2014; Morley et al. 2017). On gross inspection, there were no obvious differences in varicosity morphology among these three strains of mice. Consistent with this assertion, the average number of ChAT-positive varicosities per 400 μm² in CBA cristae was not significantly different from counts performed in α9^−/− mice cristae (16.1 ± 0.98 vs. 14.5 ± 0.96, unpaired t-test, P = 0.35; n = 6). Our varicosity measurements, however, did suggest that mean ChAT-positive varicosity area in α9^−/− cristae was significantly larger than mean varicosity area in CBA cristae (0.77 ± 0.05 vs. 1.013 ± 0.06 μm²; unpaired t-test, P = 0.03; n = 6).

Characterizing cholinergic responses in type II vestibular hair cells.

We used whole cell patch-clamp recordings from both wt strains (CBA, n = 62; C57BL/6, n = 41) and α9^−/− mice (n = 88) to characterize ACh responses in type II vestibular hair cells. Type II vestibular hair cells were identified by their characteristic voltage-activated currents (Fig. 2A), and the absence of both the low-voltage-activated K⁺ current, I_K,L, and a voltage-activated Na⁺ current, I_Nav, that are prominent features of type I hair cells and calyx terminals, respectively (see review, Eatock and Lysakowski 2006). R_m, C_m, and voltage-activated current properties in type II hair cells were similar in all three mouse strains (Table 1 and Fig. 2B). Recordings from type II hair cells in both wt strains (CBA and C57BL/6) were equivalent, and therefore these data were pooled.

Fig. 2.

Type II hair cell recordings and responses to exogenous ACh application. Activation profiles (A) and current-voltage plot (B) of wt and α9^−/− type II hair cells show no differences in voltage-activated currents. C: example records of wt type II hair cell responses to 100 µM, 300 µM, and 1 mM ACh application at membrane holding potential (V_m) = −66 mV using KCl-gluconate internal solution. D: current responses to ACh (300 µM) at different holding potentials from −96 to −46 mV using KCl-gluconate internal solution. E: current-clamp recording in type II hair cell at resting membrane potential of −49 mV shows 300 µM ACh-evoked hyperpolarization. F, left: current responses to ACh (300 µM) application at V_m = −46, −66, and −96 mV using KCl-gluconate internal solution. Right, attenuated responses at V_m = −48, −68, and −98 mV with addition of 10 mM BAPTA in the internal solution. Orange bar in all panels indicates onset and duration of ACh application (100 ms).

Table 1.

Comparison of intrinsic membrane properties in type II hair cells of C57BL/6, CBA, and α9^−/− mice

	n	Rm, MΩ	Cm, pF	Peak IK, nA
C57BL/6	41	518.6 ± 26.8	5.7 ± 0.4	3.0 ± 0.1
CBA/CaJ;129SvEVTac	62	593.4 ± 34.0	5.4 ± 0.3	3.1 ± 0.1
CBA129;A9−/−	88	599.5 ± 20.6	5.1 ± 0.3	3.0 ± 0.1

Membrane resistance (R_m) and capacitance (C_m) were measured in all type II hair cell recordings with a ±5-mV test pulse at a holding potential of −66 mV using KCl-gluconate internal solution. Peak K⁺ current (I_K) values were measured in all type II hair cells with a 80-mV depolarizing step from a holding potential of −66 mV. Values are means ± SE. R_m, C_m, and peak I_K values were not significantly different across the 3 mouse strains (P = 0.0583, P = 0.2224, and P = 0.3635, respectively; Kruskal-Wallis test). Because intrinsic membrane properties from wild-type strains (CBA and C57BL/6) were equivalent, these data were pooled and referred to as wt.

To determine the most appropriate concentration of ACh that would repeatedly evoke consistent responses over extended periods in type II hair cells, three initial concentrations of ACh (100 µM, 300 µM, and 1 mM) were applied by picospritzer onto wt type II hair cells. The lowest ACh concentration used, 100 µM, elicited responses that were smaller in amplitude relative to those evoked using higher ACh concentrations (n = 13; see Fig. 2C). The highest ACh concentration (1 mM) elicited currents that were similar in amplitude but longer in duration than responses evoked by 300 µM ACh (n = 7; Fig. 2C). Because there was no significant increase in amplitude during 1 mM exposures, we concluded that 300 µM ACh was sufficient to stimulate complete and reproducible ACh responses while minimizing the potential for receptor desensitization. The application of 300 μM ACh on a vestibular type II hair cell, at different holding potentials, is shown in Fig. 2D. At −66 mV (Fig. 2D, black trace) the whole cell current was biphasic, with a faster inward component followed by a slower outward component. The whole cell current was fully inward at potentials more hyperpolarized than −70 mV and predominantly outward at holding potentials of −50 mV and above. In current-clamp recordings, ACh hyperpolarizes type II hair cells (Fig. 2E). These observations are consistent with cholinergic signaling in other hair cell recordings whereby the early inward current represents α9nAChR activation whose Ca²⁺ influx subsequently activates SK channels, giving rise to the slow outward current.

The dependence of the outward current on Ca²⁺ influx was revealed when 10 mM BAPTA was used in the internal solution instead of 0.5 mM EGTA. The resulting ACh-induced inward currents were reduced by 90% and 41% at −68 and −48 mV, respectively. In addition, the outward currents were completely abolished by internal BAPTA at both potentials (n = 5 BAPTA, n = 5 EGTA; Fig. 2F, right). The absolute area under the curve (charge) evoked by ACh was also reduced by 90% and 98% at −68 and −48 mV, respectively. These altered responses are due to fast intracellular Ca²⁺ chelation by BAPTA. We conclude that Ca²⁺ entry through α9*nAChRs is critical to the normal ACh-induced currents shown in Fig. 2, D and F, left, as well as the concomitant change in membrane impedance discussed below.

Cholinergic responses in type II hair cells require α9nAChRs.

To confirm a role for α9*nAChRs, we characterized the response of vestibular type II hair cells to ACh under two conditions: 1) in wt mice before and during the application of the α9nAChR antagonist strychnine (Rothlin et al. 1999) and 2) in α9^−/− mice lacking the α9nAChR subunit. In wt mice, brief 100-ms pulses of 300 µM ACh on type II hair cells triggered extended current responses with at least two components that were dependent on membrane potential. At −46 mV, the response to ACh was composed of a small and brief inward current (Fig. 3A, inset), followed by a much larger outward current (Fig. 3A). At more negative potentials (−66 and −96 mV), the inward current dominated the response. In the presence of 1 µM strychnine, the ACh-evoked response was abolished at all holding potentials (Fig. 3A; blue traces), suggesting ACh responses are dependent on α9*nAChRs. This was confirmed in α9^−/− mice, where such ACh responses in type II hair cells were absent (Fig. 3B).

Fig. 3.

Whole cell ACh responses in wt and α9^−/− type II hair cells. A: example records showing a wt type II hair cell response to 300 µM ACh application (100 ms; orange bar) at different membrane holding potentials (V_m = −46, −66, and −96 mV; black traces). At −46 mV, ACh triggered a fast, small, inward current followed by a larger, slower, outward current (inset represents expansion of dashed rectangle). At −66 mV, a large inward current is followed by a relatively small outward current. At −96 mV, only inward current is observed. All ACh-induced currents were blocked with 1 µM strychnine (STR; blue traces). B: example record showing an α9^−/− type II hair cell response to 300 µM ACh (100 ms; orange bar) at V_m = −46, −66, and −96 mV (red traces). Little or no responses were observed at the different holding potentials. All traces are the average of 3 consecutive repetitions.

Ca²⁺ influx through α9*nAChR activates SK channels in type II hair cells.

In the cochlea, activation of α9*nAChRs is coupled to two different types of Ca²⁺-activated K⁺ channels, termed BK and SK (Rohmann et al. 2015). To determine whether BK and/or SK channels might be involved in efferent signaling in vestibular organs, we examined the effects of blocking these two channels. The large outward current was insensitive to the BK channel antagonist iberiotoxin (100 nM; n = 8; Fig. 4A, left trace) but was very sensitive to the SK channel antagonist apamin (0.5–100 nM; n = 17; Fig. 4A, middle trace, and Fig. 4B) and to tamapin (50–100 nM; n = 3; Fig. 4A, right), a preferentially SK2-selective antagonist. To reveal the underlying SK component, we subtracted the average traces recorded in the presence of apamin from control traces (Fig. 4C). These data suggest the reversal potential of the SK-mediated current was between −70 and −75 mV and confirm K⁺ is the major charge carrier for the apamin-sensitive current.

Fig. 4.

Ca²⁺-activated K⁺ conductance in type II hair cells following ACh exposure. A: example records showing 3 different wt type II hair cell responses to 300 µM ACh (100 ms; orange bars) at V_m = −46 mV. Outward current was insensitive to iberiotoxin (IBTX; 0.1 µM; left) but blocked by both apamin (APA; 0.1 µM; middle) and tamapin (TAM; 0.5 nM; right). B: example record showing a wt type II hair cell response to 300 µM ACh (100 ms; orange bars) at V_m = −46, −66, and −96 mV (black traces). APA (0.1 µM) blocked Ca²⁺-activated small-conductance K⁺ current (SK), revealing the primary α9nAChR (Ca²⁺) current (green traces). C: at V_m = −46, −66, and −96 mV, the APA record was subtracted from control traces to show the contribution of the SK current (gray traces). Each record is the average of 3 consecutive repetitions.

Membrane resistance and capacitance measurements in wt and α9^−/− type II hair cells.

To further explore the signaling capacity of ACh-mediated responses, we examined changes in whole cell resistance (∆R_m) and capacitance (∆C_m) using a multi-sine method. We first confirmed that the multi-sine capacitance interrogation signal did not alter responses by comparing ACh-evoked whole cell currents to results obtained using standard voltage clamp. Currents recorded using multi-sine protocol for wt (Fig. 5A, thick black trace) and α9^−/− (Fig. 5A, red trace) at −66 mV compare favorably with standard voltage-clamp recordings (e.g., Fig. 4, A and B, at −66 mV). When held at −66 mV, wt type II hair cells exhibited a significant ACh-induced decrease in whole cell resistance R_m (Fig. 5B). This reduction in R_m was similar to that observed in toadfish type II hair cells during electrical stimulation of efferent fibers in vivo (Boyle et al. 2009), suggesting similar responses might be present under physiological conditions in mice. The average reduction in R_m for wt type II hair cells is illustrated in Fig. 5B (thick black trace) with maximum reduction after the stimulus of 511 ± 202 MΩ (mean ± SD, gray band; n = 8). The total duration of ΔR_m matched the total duration of average whole cell change in membrane currents (ΔI_m). (See vertical dotted gray time line in Fig. 5 that indicates similar durations of ΔI_m, ΔR_m, ΔC_m.) ΔI_m was biphasic, initially dominated by α9nAChRs and subsequently by SK channels, whereas ΔR_m was strictly negative because both α9nAChRs and SK channels reduce the membrane resistance when activated. There were negligible changes in I_m or R_m in α9^−/− type II hair cells during the same ACh exposure (Fig. 5, A and B, respectively, thin red trace). Changes were also near zero in wt type II hair cells when vehicle only (L-15) was applied via the same picospritzer perfusion system (Fig. 5, A–C, dashed blue trace). In summary, ACh-evoked a marked decrease in R_m in wt mammalian type II hair cells that was absent in α9^−/− type II hair cells.

Fig. 5.

ACh evokes whole cell changes in membrane current (ΔI_m), membrane resistance (ΔR_m), and membrane capacitance (ΔC_m). A: ACh (300 µM, 100 ms; orange bar) applied to type II hair cells, held at −66 mV. The average whole cell ΔI_m for wt (wt ACh Avg; n = 8; thick dark gray trace), extracted from the multi-sine wave protocol, was the same as those collected with standard voltage protocol (see Fig. 2D at −66 mV). The dark gray trace shows the familiar ACh-evoked combination of inward and outward ionic currents. This response is in stark contrast to the average of α9^−/− responses (α9^−/− ACh Avg; n = 5; red trace), where no detectable change in I_m occurred. The shaded light gray (wt) and pink (α9^−/−) areas reflect variability of responses across preparations, and this shading also applies to all subsequent panels. Thin black traces refer to a single example of wt response (wt ACh). Intracellular BAPTA (10 mM), a Ca²⁺ chelator, significantly affected ΔI_m in response to ACh (wt ACh BAPTA; dashed green trace). The inward current was diminished, whereas the outward current was abolished. This suggests that intracellular Ca²⁺ is essential for normal ACh response in wt type II hair cells but affects SK activation more than α9nAChRs. No response is elicited when vehicle only (L-15; dashed light blue trace) is used. B: whole cell R_m significantly decreased in wt ACh Avg during ACh application. The transient time course and duration (dotted gray vertical time line) of the change in wt R_m paralleled the time course of wt ΔI_m (see A). Note the wt ACh Avg current trace briefly passes through zero net current (∆I_m, arrowhead in A) but still maintained a non-zero net change in membrane resistance (∆R_m). Intracellular BAPTA reduced ∆R_m (dashed green line) and had a similar time course to ∆I_m in the presence of BAPTA. Also, there was no ∆R_m response in α9^−/− cells or wt type II hair cells exposed to L15. C: ACh-evoked transient increases in ΔC_m in wt type II cells (black and dark gray traces) had the same time course as ∆I_m and ∆R_m. There was no change in α9^−/− type II hair cells (red trace) or wt cells exposed to vehicle only (L-15, light blue trace). D: expanded time course of steady-state capacitance (60 s after ACh application) shows a maintained average ΔC_m increase in wt cells (dark gray traces) but no change from baseline in α9^−/− cells (red trace), vehicle only, or intracellular BAPTA. The average difference in ACh-evoked ΔC_m between wt and α9^−/− cells was 30.9 ± 13 fF (mean ± SD). ΔC_m was significantly greater in wt than in α9^−/− cells (wt = 22.1 ± 5.3 fF, n = 8 vs. α9^−/− = −19.4 ± 18.2 fF, n = 4; means ± SD; Wilcoxon rank test, 2-tailed P < 0.05). E: transient ΔC_m increase in wt cells at V_m = −46 (gray trace), −66 (red trace), and −91 mV (green trace) in response to ACh (100 ms; orange bar), showing the transient capacitance changes are independent of holding potential and may even be larger under hyperpolarized conditions (−91 mV).

In wt type II hair cells, the ∆C_m response consisted of two components: 1) an initial ∆C_m transient increase (∆C_m = 282 ± 157 fF, mean ± SD, n = 8; Fig. 5C) with peak time and duration that paralleled the time course of ΔR_m and 2) a smaller long-lasting increase (∆C_m = 30.9 ± 13 fF, n = 8; Fig. 5D) that extended more than 60 s after ACh exposure. Alteration of capacitance due to mechanical membrane deformation associated with the application technique can be ruled out because there were no detectable C_m changes in response to application of vehicle alone (n = 3; Fig 5D, dashed blue trace) or in α9^−/− hair cells. Both the transient and long-lasting increases in C_m were eliminated with BAPTA in the recording pipette, demonstrating that intracellular Ca²⁺ signaling was required for both components.

The observation that transient (<5 s) and long-lasting (>60 s) ∆C_m components were absent in α9^−/− hair cells (Fig. 5, red traces), is consistent with results obtained using BAPTA in the pipette and further support the contention that Ca²⁺ influx via efferent ACh receptors is required for ACh-evoked C_m changes. The observations that transient C_m time-dependent changes 1) closely followed those of the SK conductance, 2) were eliminated by BAPTA, and 3) were absent when α9 receptors had been knocked out suggest the transient correlates with SK channel opening, but the specific origin of the displacement current was not investigated. Because the transient ∆C_m required intracellular Ca²⁺ concentration modulation and exhibited no significant voltage dependence (Fig. 5E), the displacement current was not a gating charge coupled to pore opening (Bezanilla 2000). As noted above, the long-lasting ∆C_m increases were present 60 s after the stimulus, whereas the transient had a duration of <5 s. Interestingly, we observed similar, if not larger, ACh-evoked increases in ∆C_m at hyperpolarized holding potential of −91 mV (n = 3; Fig. 5E).

It should be noted that the lack of long-lasting ∆C_m in α9^−/− type II hair cells was not due to compromised capacitance changes associated with depolarization steps. These changes are thought to be the result of exocytosis, because wt and α9^−/− type II hair cells both exhibited C_m increases in response to simple depolarizing voltage steps (Fig. 6). In an example from a wt type II hair cell, capacitance increases in response to 60-mV depolarizing pulses of 200- to 500-ms duration are shown in Fig. 6A. Cumulative increases in C_m for multiple depolarizations were the same for both wt and α9^−/− type II hair cells (Fig. 6B), suggesting exocytosis machinery was functional in all strains.

Fig. 6.

Depolarization evoked neurotransmitter release in type II hair cells of wt and α9^−/− mice. A: example record showing an increase in capacitance in a wt type II hair cell evoked by 60-mV, 200- to 500-ms depolarizing voltage steps from a holding potential of −66 mV. Asterisks indicate transient increases in capacitance, and square brackets indicate “steady-state” capacitance values between depolarizing steps. B: the cumulative increase in “steady-state” capacitance (fF) with depolarization step duration (s) was indistinguishable between α9^−/− (red circles; n = 3) and wt strains (gray triangles; n = 5), suggesting transmitter release evoked by depolarization steps is normal in α9^−/− mice.

Effects of intracellular Ca²⁺ chelation.

As described above, intracellular BAPTA (10 mM) markedly reduced the ACh-evoked initial α9*nAChR inward current in type II hair cells by 77% and completely abolished the secondary, SK channel outward current when measured in the time domain (n = 5; Fig. 2F). This pattern of altered ACh-evoked current response was replicated when the frequency-domain protocol was used (∆I_m; Fig. 5A). In addition, the multi-sine protocol also revealed that BAPTA markedly reduced ∆R_m and blocked all ∆C_m responses to ACh application (Fig. 5, B and C).

DISCUSSION

Recent behavioral work has demonstrated that VOR adaptation and compensation are compromised in α9^−/− mice, suggesting the EVS and α9*nAChRs might be required for these important vestibular functions in mammals (Hübner et al. 2015, 2017). In primitive vertebrates, the EVS is known to play an important role in volitional movements and attention (Tricas and Highstein 1990), and similar functions involving α9*nAChRs might be relevant to mammals. Our study has examined the role of these receptors at the cellular level in mammalian vestibular type II hair cells, using α9^−/− animals and two strains of control mice.

Cholinergic varicosities and ChAT expression.

The loss of ACh responsiveness in type II hair cells from the α9^−/− mice do not appear to be associated with any gross alterations in the peripheral EVS innervation patterns, at least with respect to canal cristae. We routinely observed vestibular efferent varicosities in the cristae of all three strains (Fig. 1). At face value, it was difficult to reconcile any obvious differences in ChAT-positive EVS neurons and varicosities in α9^−/− mouse crista compared with the two control mouse strains examined in this study or with ChAT staining of EVS neurons in other mouse models, including a second, independent α9^−/− strain where the morphological and ultrastructural organization of peripheral EVS varicosities appeared normal (Jordan et al. 2015; Luebke et al. 2014; Morley et al. 2017). In the mouse cochlea, loss of the α9nAChR subunit has been associated with changes in the number and size of efferent varicosities (Simmons and Morley 2011; Vetter et al. 1999, 2007). Although our quantitative analysis failed to identify any difference in the density of efferent varicosities between α9^−/− and CBA mice, it did reveal that efferent varicosities in α9^−/− animals may be larger. Background genetics inherent to maintaining the α9^−/− mutation on a CBA/CaJ, a crossed CBA/CaJ × 129/SvEv, or a 129/SvEv line do not appear to contribute to the cochlear efferent phenotype observed in α9^−/−mice (Vetter et al. 1999), suggesting that differences in the size of cochlear efferent varicosities are a function of the loss of the α9nAChR subunit. However, given we have only characterized the EVS morphology and innervation patterns in CBA/CaJ × 129SvEvTac animals, we cannot eliminate the possibility that larger EVS varicosities in α9^−/− mice may be attributed to differences in background instead of, or in addition to, the loss of the α9nAChR subunit.

Hair cell recordings and ACh exposure.

Our results show α9-subunit expression is necessary for normal cholinergic signaling at efferent/vestibular type II hair cell synapses in mice. We used 100-ms-duration ACh applications, which triggered current and voltage responses that lasted several seconds (Fig. 2, C and D). This is consistent with work in turtles and toadfish, where responses in afferents and hair cells substantially outlasted the duration of efferent stimulation (Boyle et al. 2009; Brichta and Goldberg 2000).

It is known that α9*nAChRs are highly permeable and preferentially selective for Ca²⁺ (Doi and Ohmori 1993; Katz et al. 2000; Weisstaub et al. 2002), and intracellular Ca²⁺ concentration is elevated in vestibular hair cells following ACh exposure (Housley et al. 1990; Ohtani et al. 1994; Yamashita et al. 1993). The significant dampening effect of 10 mM internal BAPTA on whole cell ACh responses (Fig. 2F) demonstrates the critical role of intracellular Ca²⁺ signaling in cholinergic modulation of vestibular hair cells.

The essential α9 nicotinic receptor subunit.

The “two-channel hypothesis” describes the biphasic action of ACh on auditory hair cells (Fuchs and Murrow 1992a; Martin and Fuchs 1992), a highly conserved mechanism across all vertebrates (Lustig 2006). ACh triggers Ca²⁺ influx via α9*nAChRs and a secondary Ca²⁺-activated K⁺ current that hyperpolarizes the hair cell (Glowatzki and Fuchs 2000; Housley and Ashmore 1991; Oliver et al. 2000) and decreases R_m (Boyle et al. 2009). We demonstrate a similar mechanism also applies to mammalian vestibular type II hair cells. Between −66 and −46 mV, the ACh response comprises interplay between at least two components: an initial, fast inward current and a secondary, slower outward current (Fig. 3A and inset). We interpret the initial inward current as primarily Ca²⁺ through α9*nAChRs and the outward current as K⁺ since at holding potentials more hyperpolarized than K⁺ ion reversal (E_K = −85 mV), all ACh-induced current is inward (Fig. 3A). In wt mice, strychnine abolished ACh responses, suggesting that they are dependent on α9*nAChR activation (Fig. 3A), as shown in frog and pigeon vestibular hair cells (Holt et al. 2001, 2003; Li and Correia 2011). That α9^−/− mice lack these responses (Fig. 3B) confirms the critical importance of α9*nAChRs for initiating the ACh response in vestibular type II hair cells.

A role for Ca²⁺-sensitive K⁺ channels.

In mouse vestibular cristae, the outward K⁺ component of the ACh response was highly sensitive to the SK blockers apamin and tamapin, whose picomolar concentrations favor SK2 (KCNN2, KCa2.2). In contrast, there were negligible effects of the BK blocker iberiotoxin (Fig. 4A). Our observations are in contrast to those made in isolated guinea pig type II vestibular hair cells, which showed sensitivity to iberiotoxin, but not to apamin (Kong et al. 2005). These contradictory results may be due to their enzymatic isolation of hair cells and/or prolonged (seconds) ACh exposure. If BK was activated via Ca²⁺-induced Ca²⁺ release (CICR), the prolonged ACh exposure and enzyme treatment might account for BK recruitment over and above that we observed in our experiments. We used a semi-intact preparation, without enzyme treatment (Lim et al. 2011), and only brief (100 ms), locally applied ACh to avoid receptor desensitization (Lee et al. 2015).

Although our ACh exposure was brief, the ensuing responses (∆I_m, ∆R_m, and transient ∆C_m) could last for several seconds or several tens of seconds (long-lasting ∆C_m). These extended effects may be the result of additional mechanisms at play. In addition to SK channels, initial Ca²⁺ influx, via α9*nAChRs, may also trigger intracellular CICR (Castellano-Muñoz et al. 2016; Sridhar et al. 1997). Efferent-evoked CICR is thought to operate primarily through the closely associated “synaptoplasmic cisterns” that store Ca²⁺ and amplify cholinergic responses, as described in cochlear hair cells (Evans et al. 2000; Kennedy and Meech 2002; Lioudyno et al. 2004). The close proximity of vestibular efferent synapses to an analogous structure may provide an additional pool of intracellular Ca²⁺ (Lysakowski and Goldberg 1997) to drive the SK response. Although our current data support ACh-induced changes in intracellular Ca²⁺ concentration in vestibular hair cells, the potential involvement of Ca²⁺ stores was not addressed in this study.

In α9^−/− type II hair cells, the lack of functional α9*nAChRs means SK channel conductances could not be triggered by ACh exposure, eliminating the intracellular Ca²⁺ signal required for activation of Ca²⁺-activated K⁺ channels. Moreover, there was no compensatory upregulation of other nAChR subtypes that triggered SK or BK channels in α9^−/− hair cells. Similarly, there were no detectable changes in voltage-activated currents (Fig. 4, A and B) or passive membrane properties (Table 1) in α9^−/− hair cells compared with wt.

Membrane resistance and ACh.

In previous toadfish, frog, and burbot studies, efferent stimulation decreased R_m of vestibular or lateral line hair cells by activation of basolateral conductances. Consistent with previous findings in toadfish (Boyle et al. 2009), present results in mice demonstrate that the drop in R_m arises primarily from opening of Ca²⁺-activated K⁺ channels and with voltage-activated channels playing a much smaller role. EVS activation generates an electrical shunt in the basolateral membrane of the type II hair cell that reduces sensitivity by reducing voltage modulation in response to physiological mechanoelectrical transduction (MET) currents (Boyle et al. 2009; Flock and Russell 1976; Sugai et al. 1992). This reduced sensitivity highlights a critical feature controlling the sensitivity of type II hair cells to ACh. Except for studies in toadfish semicircular canals (Boyle et al. 2009) and outer hair cells (Geisler 1974; Rabbitt et al. 2009), this decreased gain response has perhaps been overshadowed by the excitatory action of ACh on vestibular afferent background discharge (Goldberg and Fernández 1980; Holt et al. 2015a) and the emphasis on the hyperpolarizing effects in auditory hair cells (Glowatzki and Fuchs 2000; Goutman et al. 2005; Marcotti et al. 2004; Roux et al. 2011). The present report demonstrates that one source of the reduced afferent sensitivity after EVS activation in mammals in type II hair cells is likely due to α9*nAChR/SK-dependent reduction in hair cell resistance.

Membrane capacitance, ACh, and Ca²⁺ signaling.

The importance of Ca²⁺ signaling to ACh responses of type II hair cells was demonstrated by buffering intracellular Ca²⁺ with BAPTA. Calcium buffering eliminated the ACh-evoked SK current and, like apamin (Fig. 4B), revealed the α9*nAChR current (green dashed trace, Fig. 5A). BAPTA also eliminated the transient capacitance increase, demonstrating a correlation between SK opening and ∆C_m (Fig. 5C). The present study did not examine causality or specific charges responsible for the transient change in capacitance.

Intracellular BAPTA also eliminated the long-lasting (>60 s) capacitance increase (Fig. 5, C and D). The long-lasting ACh-evoked capacitance increase implies an increase in membrane surface area, similar to the increase evoked by depolarizing voltage pulses (Fig. 6A). This raises the possibility of a link between efferent activation and hair cell neurotransmitter exocytosis. In immature cochlear inner hair cells, α9*nAChR expression was needed for normal maturation of the ribbon synapse (Johnson et al. 2013). However, it is not known whether Ca²⁺ influx through α9*nAChR activation influences neurotransmitter exocytosis at the ribbon synapse. It has been shown previously in auditory hair cells that neurotransmitter vesicle release from ribbon synapses is related to available intracellular Ca²⁺ concentrations and CICR (Schnee et al. 2011). In the present experiments, long-lasting ACh-induced capacitance increases were present under whole cell voltage-clamp conditions even at hyperpolarized holding potentials (e.g., −91 mV; Fig. 5E), minimizing the possibility of any Ca²⁺ influx near the ribbon synapse through voltage-activated Ca²⁺ channels. A consistent hypothesis is that ACh-evoked Ca²⁺ entry through α9á*nAChRs might have triggered neurotransmitter exocytosis, leading to long-lasting capacitance increases.

It should also be noted that both the transient and long-lasting ∆C_m components are dependent on the presence of α9-subunit expression. Similarly to the intracellular BAPTA results in wt mice, there was no net ∆C_m in α9^−/− type II hair cells under the same conditions (Fig. 5C). This lack of ACh-evoked ∆C_m in α9^−/− type II hair cells was not due to a transgenic alteration in the vesicular release mechanisms, because depolarizing steps evoked ∆C_m increases in type II hair cells of all strains used, including α9^−/− (Fig. 6B). This supports the possibility of a Ca²⁺-dependent link between α9*nAChRs and exocytosis in wt vestibular hair cells. If true, Ca²⁺-dependent neurotransmitter release from type II hair cells could contribute to transient discharge rate increases in vestibular afferent neurons, particularly in calyx-bearing, functionally dimorphic afferents (Fig. 7A) during efferent activation (Goldberg and Fernández 1980; Holt et al. 2015a; Rabbitt et al. 2010).

Fig. 7.

A summary of anatomical contacts and ACh-evoked events associated with mammalian type II hair cells. A: type II hair cells contact bouton afferents and functional dimorphic afferents. In mammals, functional dimorphic afferents include conventional dimorphic (1) arrangement, comprising calyx and bouton terminals (arrowheads), whereas unconventional dimorphic (2) arrangement consists of calyx terminals with outer face (en face) contacts (arrow) with type II hair cell (Lysakowski and Goldberg 1997). Although the prevalence of en face contacts has yet to be determined in mice, conventional dimorphic units comprise nearly 80% of some rodent species and therefore could be directly affected by proposed neurotransmitter exocytosis from type II hair cells. B: ACh released from efferent terminals onto type II hair cells during efferent activation (orange terminal) triggers influx of Ca²⁺ through α9 receptors (red), which results in the efflux of K⁺ through SK channels (blue). This ACh-evoked response has 3 effects: 1) a hyperpolarization of membrane potential (↓V_m), 2) a reduction in membrane resistance (↓R_m), and 3) increased membrane capacitance (↑C_m). Hyperpolarization of V_m and reduced R_m would decrease type II hair cell sensitivity to MET channel currents. Neurotransmitter exocytosis from type II hair cells, indicated by long-lasting ∆C_m, could contribute to elevation of calyx-bearing afferent background discharge by exciting their associated afferent bouton contacts and/or via direct en face synaptic contact with calyx terminals.

Intracellular Ca²⁺and buffering.

When intracellular Ca²⁺ is considered, there are two points to that need to be made. First, although the responses presented in this study are similar to those shown in other hair cell preparations, the exogenous application of 300 µM ACh may well evoke a response in the hair cell that differs from that which would occur following endogenous release of ACh from efferent terminals. It is plausible that 300 µM ACh is driving a larger Ca²⁺ entry that overwhelms the cell’s intrinsic intracellular Ca²⁺-buffering capabilities and therefore triggers neurotransmitter release where under normal physiological conditions, it may not. The actual concentrations of ACh release at efferent terminals may be lower. Second, even though the methods and internal solutions used are similar to other hair cell preparations, the intracellular Ca²⁺-buffering capacity is unknown and therefore may be altered by our whole cell recording technique. Therefore, it remains to be determined if the mechanisms described will occur under physiological conditions.

Decreased gain, increased activity.

Our data begin to reconcile three broadly accepted observations in response to vestibular efferent activation: 1) putative type II hair cell hyperpolarization in all mammalian species (Ashmore and Russell 1983; Holt et al. 2003; Housley et al. 1990; Kong et al. 2007), 2) decrease in irregular afferent sensitivity to physiological stimulation in mammals (Goldberg and Fernández 1980), and 3) excitation in irregular afferent background discharge (Goldberg and Fernández 1980).

Type II hair cells contact the majority of vestibular afferents, either as the exclusive hair cell inputs to bouton afferents or as adjunct inputs to functional dimorphic afferent terminals (Fig. 7A). Therefore, any changes in type II hair cell activity will significantly influence peripheral output. Our results demonstrate that ACh released during efferent stimulation likely “shunts” type II hair cells by opening α9*nAChRs and SK conductances, thereby decreasing R_m. Together with outward K⁺ current, these responses would result not only in hair cell hyperpolarization but also, importantly, reduced voltage sensitivity to hair bundle MET currents driven by the reduction in R_m. In type II hair cells, Ca²⁺ influx through α9*nAChRs evoked long-lasting increases in C_m under whole cell voltage clamp, possibly due to an increase in membrane surface area triggered by Ca²⁺ entry via ACh receptors (Fig. 7B) and subsequent activation of CICR (Castellano-Muñoz et al. 2016; Sridhar et al. 1997). Hence, efferent contact on type II vestibular hair cells could cause 1) a decrease in sensitivity to physiological stimulation and 2) a transient increase in neurotransmitter exocytosis, resulting in a transient increase in discharge rate of afferent neurons that receive type II hair cell inputs.

GRANTS

This study was funded by National Health and Medical Research Council of Australia Grant 1011159, by the Garnett Passe & Rodney Williams Memorial Foundation (Australia), and by National Institutes of Health Grant R01 DC006685.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

L.A.P., P.J., R.J.C., P.M.J., J.C.H., R.D.R., R.L., and A.M.B. conceived and designed research; L.A.P., H.T., H.R.D., P.M.J., and J.C.H. performed experiments; L.A.P., H.T., P.J., P.M.J., J.C.H., R.D.R., R.L., and A.M.B. analyzed data; L.A.P., H.T., P.J., R.J.C., A.A.M., P.M.J., J.C.H., R.D.R., R.L., and A.M.B. interpreted results of experiments; L.A.P., P.M.J., J.C.H., R.D.R., and A.M.B. prepared figures; L.A.P., R.L., and A.M.B. drafted manuscript; L.A.P., H.T., R.J.C., A.A.M., P.M.J., J.C.H., R.D.R., R.L., and A.M.B. edited and revised manuscript; L.A.P., H.T., P.J., R.J.C., A.A.M., P.M.J., J.C.H., R.D.R., R.L., and A.M.B. approved final version of manuscript.