The potassium channel subunit KV1.8 (Kcna10) is essential for the distinctive outwardly rectifying conductances of type I and II vestibular hair cells

Abstract

In amniotes, head motions and tilt are detected by two types of vestibular hair cells (HCs) with strikingly different morphology and physiology. Mature type I HCs express a large and very unusual potassium conductance, g_K,L, which activates negative to resting potential, confers very negative resting potentials and low input resistances, and enhances an unusual non-quantal transmission from type I cells onto their calyceal afferent terminals. Following clues pointing to K_V1.8 (Kcna10) in the Shaker K channel family as a candidate g_K,L subunit, we compared whole-cell voltage-dependent currents from utricular HCs of K_V1.8-null mice and littermate controls. We found that K_V1.8 is necessary not just for g_K,L but also for fast-inactivating and delayed rectifier currents in type II HCs, which activate positive to resting potential. The distinct properties of the three K_V1.8-dependent conductances may reflect different mixing with other K_V subunits that are reported to be differentially expressed in type I and II HCs. In K_V1.8-null HCs of both types, residual outwardly rectifying conductances include K_V7 (Knq) channels. Current clamp records show that in both HC types, K_V1.8-dependent conductances increase the speed and damping of voltage responses. Features that speed up vestibular receptor potentials and non-quantal afferent transmission may have helped stabilize locomotion as tetrapods moved from water to land.

Keywords: hair cell; inner ear; mouse; neuroscience; potassium channel; utricle; vestibular; voltage gated.

Figure 1.. Kcna10^–/– type I hair cells (HCs) lacked g_K,L, the dominant conductance in mature Kcna10^{+/+,++/––} type I HCs.

Representative voltage-evoked currents in (A) a P22 Kcna10^+/– type I HC and (B) a P29 Kcna10^–/– type I HC. (A) Arrow, transient inward current that is a hallmark of g_K,L. Arrowheads, tail currents, magnified in insets. For steps positive to the midpoint voltage, tail currents are very large. As a result, K⁺ accumulation in the calyceal cleft reduces driving force on K⁺, causing currents to decay rapidly, as seen in A (Lim et al., 2011). Note that the voltage protocol (top) in B extends to more positive voltages. (C) Activation (G–V) curves from tail currents in A and B; symbols, data; curves, Boltzmann fits (Equation 1). (D) Fit parameters from mice >P12 show big effect of Kcna10^–/– and no difference between Kcna10^+/– and Kcna10^+/+. (D.1), Tukey’s test: +/+ vs –/–, p<1E-9; +/– vs –/–, p<1E-9. (D.2), Tukey’s test: +/+ vs –/–, p=9.4E-4. (D.3), Tukey’s test: +/+ vs –/–, p<1E-9; +/– vs –/–, p<1E-9. Asterisks: ***p < 0.001; and ****p < 0.0001. Line, median; Box, interquartile range; Whiskers, outliers. See Table 1 for statistics.

Figure 1—figure supplement 1.. Developmental changes in type I hair cell (HC) K_V conductances.

(A) Parameters from Boltzmann fits of tail G–V relations for type I HCs plotted against age. (B) Conductance density is similar in young (P5–P10) type I HCs that lack g_K,L. g_K,L is defined here as having a V_half negative to –55 mV. Kcna10^+/+,+/– with g_K,L, 17 ± 5 nS/pF (19); Kcna10^+/+,+/– without g_K,L, 3.7 ± 0.4 nS/pF (22); Kcna10^–/–, 1.8 ± 0.4 nS/pF (13). Kcna10^+/+,+/– with g_K,L vs Kcna10^–/–: p = 0.007, KWA, g 1.0. Asterisks: **p < 0.01. Line, median; Box, interquartile range; Whiskers, outliers.

Figure 2.. Kcna10^–/– type I hair cells (HCs) had much longer membrane charging times and higher input resistances (voltage gains) than Kcna10^+/+,+/– type I HCs.

(A) g_K,L strongly affects passive membrane properties: (A.1) V_rest, Tukey’s test p<1E-9, (A.2) R_in, input resistance, Tukey’s test p<1E-9, and (A.3) membrane time constant, ${\displaystyle {\tau }_{RC}=(R_{input}\ast C_m)}$, Tukey’s test p<1E-9. See Table 2 for all statistics. (B) Current clamp responses to the same scale from (B.1) Kcna10^+/– and (B.2) Kcna10^–/– type I cells, both P29. Filled arrowhead (B.2), sag indicating I_H activation. Open arrowhead, Depolarization rapidly decays as I_DR activates. (B.3) First 6 ms of voltage responses to 170 pA injection, normalized to steady-state value; curves, double-exponential fits (Kcna10^+/+, ${\displaystyle \tau }$ 40 μs and 2.4 ms) and single-exponential fits (Kcna10^–/–, ${\displaystyle \tau }$ 1.1 ms). Asterisks, ****p < 0.0001. Line, median; Box, interquartile range; Whiskers, outliers.

Figure 3.. Kcna10^–/– type II hair cells (HCs) in all zones of the sensory epithelium lacked the major rapidly inactivating conductance, g_A, and had less delayed rectifier conductance.

Activation and inactivation varied with epithelial zone and genotype. (A) g_A inactivation time course varied across zones. (A.1) Zones of the utricular epithelium: lateral extrastriola (LES), medial extrastriola (MES), and striola (S). (A.2) Normalized currents evoked by steps from –124 to +30 mV with overlaid fits of Equation 3. (A.3) ${\displaystyle {\tau }_{\mathrm{I}\mathrm{n}\mathrm{a}\mathrm{c}\mathrm{t},\mathrm{F}\mathrm{a}\mathrm{s}\mathrm{t}}}$ was faster in Kcna10^+/– (n=45) than Kcna10^+/+ (n=43) HCs (KWA, p=0.027), and faster in LES (n=56) than MES (n=23, KWA, p=0.002) or S (n=9, KWA, p=2E-4). Point label is number of cells. Brackets show post hoc pairwise comparisons between two zones (vertical brackets) and horizontal brackets compare two genotypes; see Table 3 for statistics on kinetics. (A.4) Fast inactivation was a greater fraction of total inactivation in LES (n=58) than striola (n=10, Tukey’s test p=0.0041). (B) Exemplars; ages, left to right, P312, P53, P287, P49, P40, P154. (C) % inactivation at 30 mV was much lower in Kcna10^–/– (n=37) than Kcna10^+/– (n=47, Tukey’s HSD, p<1E-9) and Kcna10^+/+ (n=44, Tukey’s HSD, p<1E-9). % inactivation was lower in striola (n=16) than LES (n=77, Tukey’s HSD, p=3E-5) and MES (n=36, Tukey’s HSD, p=0.0011). 2-way ANOVA detected interaction between zone and genotype, p=0.026 (Table 3). (D) Exemplar currents and G–V curves from LES type II HCs show a copy number effect. (D.1) Exemplar currents evoked by steps from –124 to +30 mV fit with Equation 3. (D.2) Averaged peak and steady-state conductance–voltage data points from LES cells (+/+, n=37; –/–, n=20) were fit with Boltzmann equations (Equation 1) and normalized by g_max in (D.3). Asterisks: *p < 0.05; **p < 0.01; ***p < 0.001; and ****p < 0.0001. Error bars, SEM. See Table 4 for statistics on voltage dependence.

Figure 3—figure supplement 1.. For type II hair cells (HCs) older than P12, K_V conductance activation and inactivation differed across zones and genotypes.

(A) In Kcna10^+/+ and Kcna10^+/– HCs, ${\displaystyle {\tau }_{Inact,Fast}}$ at 30 mV was faster in LES (n=56) than MES (n=23, KWA p=0.002) or S (n=9, KWA p=2E-4). ${\displaystyle {\tau }_{Inact,Fast}}$ was faster in Kcna10^+/– (n=45) than Kcna10^+/+ (n=43, KWA p=0.027, see Table 3). (B) Fast inactivation was a larger fraction of total inactivation in LES (n=56) than striola (n=9, Tukey’s p=0.0041). (C) ${\displaystyle {\tau }_{Act}}$ at 30 mV was slower in Kcna10^–/– (n=53) than Kcna10^+/+ (n=49, KWA p=0.0048) and Kcna10^+/– (n=59, KWA p=2E-7), and slower in S (n=16) than LES (n=75, KWA p=6E-4) and MES (n=33, KWA p=0.02). (D) Percent inactivation at 30 mV was lower in S (n=20) than LES (n=99, 2-way ANOVA Tukey’s p<0.0001) and MES (n=42, 2-way ANOVA Tukey’s p=0.001), and lower in Kcna10^–/– (n=43) than Kcna10^+/+ (n=58, 2-way ANOVA Tukey’s p=0.0048 and <0.0001) and Kcna10^+/– (n=60, 2-way ANOVA Tukey’s p<0.0001). Interaction between Zone and Genotype was significant (p=0.026). Asterisks: *p < 0.05; **p < 0.01; ***p < 0.001; and ****p < 0.0001. Line, median; Box, interquartile range; Whiskers, outliers.

Figure 3—figure supplement 2.. For type II hair cells (HCs) older than P12, K_V conductances were stable.

(A–C) Parameters from Boltzmann fits of peak G–V relations and (D) % inactivation at +30 mV plotted against age from all zones. Overlaid curves are smoothing cubic β-splines. Note the seven extrastriolar Kcna10^–/– type II HCs with % inactivation >30%.

Figure 3—figure supplement 3.. A minority of extrastriolar Kcna10^–/– type II hair cells (HCs) had a very small fast-inactivating outward rectifier current.

(A) All extrastriolar Kcna10^+/+,+/– type II HCs inactivated by >30%. Most mature (>P12) extrastriolar Kcna10^–/– type II HCs inactivated by <30% but some inactivated by >30% (7/30, 23%) because they had fast inactivation (B). (B) Exemplar residual fast inactivation (${\displaystyle {\tau }_{\mathrm{I}\mathrm{n}\mathrm{a}\mathrm{c}\mathrm{t},\mathrm{F}\mathrm{a}\mathrm{s}\mathrm{t}}}$ = 10 ms at +30 mV). For the seven cells in this group, ${\displaystyle {\tau }_{\mathrm{I}\mathrm{n}\mathrm{a}\mathrm{c}\mathrm{t},\mathrm{F}\mathrm{a}\mathrm{s}\mathrm{t}}}$ = 30 ± 6 ms, amplitude of fast inactivation = 310 ± 70 pA; activation peak V_half = –15 ± 2 mV and slope factor = 12.4 ± 0.9 mV. These parameters are similar to g_A but for the much smaller conductance (one-way ANOVAs).

Figure 4.. Kcna10^–/– type II hair cells (HCs) had larger, slower voltage responses and more electrical resonance.

(A) Passive membrane properties near resting membrane potential: (A.1) Resting potential. R_input (A.2) and ${\displaystyle {\tau }_{RC}}$ (A.3) were obtained from single-exponential fits to voltage responses <15 mV. R_input and ${\displaystyle {\tau }_{RC}}$ were higher in Kcna10^–/– (n=13) than Kcna10^+/+ (n=22, KWA p=0.015; p=0.016) and Kcna10^+/– (n=33, KWA p=0.002; p=0.008; see Table 5). (B) Exemplar voltage responses to iterated current steps (bottom) illustrate key changes in gain and resonance with K_V1.8 knockout. (B.1) Kcna10^+/– type II HC (P24, LES) and (B.2) Kcna10^–/– type II HC (P53, LES). Arrowheads, depolarizing transients. (C) Range of resonance illustrated for Kcna10^–/– type II HCs (left, pink curves fit to Equation 5) and controls (right, blue fits). (C.1) Resonant frequencies, left to right: 19.6, 18.4, 34.4, and 0.3 Hz. Leftmost cell resonated spontaneously (before step). (C.2) Tuning quality (Q_e; Equation 6) was higher for Kcna10^–/– (n=26) type II HCs (KWA: p = 0.0064 vs Kcna10^+/+, n=23; p = 7E-8 vs Kcna10^+/–, n=45). (D) Kcna10^–/– type II HCs had higher, slower peaks and much slower rebound potentials in response to large (170 pA) current steps. (D.1) Normalized to show initial depolarizing transient (filled circles, times of peaks; horizontal arrows, peak width at half-maximum). (D.2) Longer time scale to highlight how null mutation reduced post-transient rebound. (E) In Kcna10^–/– HCs (n=19), depolarizing transients evoked by a +90 pA step were slower to peak (E.1) than in Kcna10^+/+ (n=19, 2-way ANOVA Tukey’s p<1E-9) and Kcna10^+/– (n=34, 2-way ANOVA Tukey’s p<1E-9) and (E.2) larger than in Kcna10^+/+ (n=19, KWA p=0.006) and Kcna10^+/– (n=34, KWA p=2E-4). Asterisks: *p < 0.05; **p < 0.01; ***p < 0.001; and ****p < 0.0001. Line, median; Box, interquartile range; Whiskers, outliers.

Figure 5.. Type I and II hair cell (HC) basolateral membranes show specific immunoreactivity to Kv1.8 antibody (magenta).

Antibodies for K_V7.4 (A, green) and calretinin (B, cyan) were used as counterstains for calyx membrane (Kv7.4), type II HC cytoplasm (calretinin) and cytoplasm of striolar calyx-only afferents (calretinin). (A) Left, Cartoon showing K_V7.4 on the calyx inner face membrane (CIF) and K_V1.8 on the type I HC membrane. SC, supporting cell nuclei. A.1–3, Adult mouse utricle sections. K_V7.4 antibody labeled calyces on two K_V1.8-positive type I HCs (A.1), four K_V1.8-positive type I HCs (A.2), and two K_V1.8-negative type I HCs from a Kcna10^–/– mouse (A.3). (B) Left, Cartoon showing cytoplasmic calretinin stain in calyx-only striolar afferents and most type II HCs, and K_V1.8 on membranes of both HC types. In wildtype utricles, K_V1.8 immunolocalized to basolateral membranes of type I and II HCs (extrastriola, B.1). K_V1.8 immunolocalized to type I HCs (striola, B.2). Staining of supporting cell (SC) membranes by Kv1.8 antibody was non-specific, as it was present in Kcna10^–/– tissue (striola, B.3 and B.4). All scale bars 5 µm.

Figure 6.. Inactivation curve of g_A in extrastriolar type II hair cells (HCs).

(A) Modified voltage protocol measured accumulated steady-state inactivation at the tail potential. 100 μM ZD7288 in bath prevented contamination by HCN current. (B) Voltage dependence of g_A’s steady-state inactivation (h_∞ curve) and peak activation are consistent with K_V1.4 heteromers. Curves, Boltzmann fits (Equation 1). Average fit parameters from Kcna10^+/+,+/– type II HCs, P40–P210, median P94. Inactivation: V_half, –42 ± 2 mV (n = 11); S, 11 ± 1 mV. Activation: V_half, –23 ± 1 mV (n = 11); S, 11.2 ± 0.4 mV.

Figure 7.. A K_V7-selective blocker, XE991, reduced residual delayed rectifier currents in Kcna10^–/– type I and II hair cells (HCs).

(A) XE991 (10 μM) partly blocked similar delayed rectifier currents in type I and II Kcna10^–/– HCs and a type II Kcna10^+/+ HC. (B) Properties of XE991-sensitive conductance, _DR(K_V7). (B.1) % Block of steady-state current. (B.2) Mean tail G–V curves for Kcna10^–/– type I HCs (n = 8), Kcna10^–/– type II HCs (9), and Kcna10^+/+ type II HCs (5); shading is ± SEM. (B.3) V_half was less negative in Kcna10^+/+ type II than Kcna10^–/– type I HC (p = 0.01, KWA). (B.4) Conductance density was similar in all groups (ANOVA), non-significant at 0.4 power (left), 0.2 power (right). Asterisks: *p < 0.05 and ***p < 0.001. Line, median; Box, interquartile range; Whiskers, outliers.

Figure 7—figure supplement 1.. A minority of striolar Kcna10^–/– type I hair cells (HCs) had a small low-voltage-activated outward rectifier current in addition to a more positively activating outward rectifier.

(A) Low-voltage-activated current from one cell was isolated by subtraction with 10 μM XE991 (P39), indicating that it was carried by K_V7 channels. Deactivation of XE991-sensitive current evoked by step from –64 to –124 mV (arrow) was fit with exponential decay (${\displaystyle \tau }$ = 21 ms). (B) XE991-sensitive tail G–V curve of the XE991-blocked conductance (A) was fit with a sum of two Boltzmann equations: G(V) = A₁/(1 + exp((V_half,1 – V)/S₁)) + A₂/(1 + exp((V_half,2 – V)/S₂)). (C) The low-voltage-activated V_half,1 component was only seen in striolar Kcna10^–/– type I HCs, and even there in the minority: 5/23; 22%; P6–P370. It was always seen together with a more positively activating outward rectifier. Average Boltzmann parameters (n=5, including B): A₁/(A₁ + A₂) = 0.15 ± 0.04, V_half,1 = –106 ± 5 mV, S₁ = 3.8 ± 0.8 mV, V_half,2 = –41±1 mV, S₂ = 7 ± 1 mV. Ages: P11, 39, 202, 202, 202. No extrastriolar type I HCs (0/45; P6–277) had a double activation tail G–V curve.

Figure 7—figure supplement 2.. No difference was detected in H (HCN) and Kir (fast inward rectifier) currents between Kcna10^+/+ and Kcna10^–/– hair cells (HCs), consistent with a specific involvement of K_V1.8 in Kcna10 expression.

(A) Hyperpolarizing voltage steps evoked I_Kir and I_HCN in Kcna10^{+/+,+/–,–/–} type I and II HCs. (A.1) Arrows, deactivation of g_K,L. (A.2–A.4) Bracket, the sum of I_H and I_Kir were measured as total inward current after 250 ms at –124 mV. (B) Summed I_KIR and I_H density was smaller in striola than extrastriola (type I HC KWA, p=4E-9; type II HC 2-way ANOVA Tukey’s p=0.006, see Supplementary file 1c). Data labels are number of cells. (C) Inward currents seen at the onset of hyperpolarization, including I_Kir, were larger in type II HCs than type I. Magnification of boxed in inset from the extrastriolar cells in A.2–A.4. Open arrowhead, activation of fast inward rectifier, I_Kir; filled arrowhead, slower activation of I_HCN. Asterisks: **p < 0.01 and ****p < 0.0001. Line, median; Box, interquartile range; Whiskers, outliers.