Effect of Salicylate on KCNQ4 of the Guinea Pig Outer Hair Cell

Abstract

Salicylate causes a moderate hearing loss and tinnitus in humans at high-dose levels. Salicylate-induced hearing loss has been attributed to impaired sound amplification by outer hair cells (OHCs) through its direct action on the OHC motility sensor and/or motor. However, there is a disparity of salicylate concentrations between the clinical and animal studies, i.e., extremely high extracellular concentrations of salicylate (from 1 to 10 mM) is required to produce a significant reduction of electromotility in animal studies. Such concentrations are above the clinical/physiological range for humans. Here, we showed that clinical/physiological concentration range of salicylate caused concentration-dependent and reversible reductions in I_K,n (KCNQ4) and subsequent depolarization of OHCs. Salicylate reduced the maximal tail current of the activation curve of I_K,n without altering the voltage-sensitivity (V_half). The salicylate-induced reduction of I_K,n was almost completely blocked by linopirdine (0.1 mM) and BaCl₂ (10 mM). Consistent with the finding in OHCs, salicylate significantly reduced KCNQ4-mediated current expressed in Chinese hamster ovarian (CHO) cells by comparable amplitude to OHCs without significantly shifting V_half. Nonstationary fluctuation analysis shows that salicylate significantly reduced the estimated single-channel current amplitude and numbers. Intracellular Ca²⁺ elevation resulting from cytoplasmic acidosis also contributes to the current reduction of I_K,n (KCNQ4) of OHCs. These results indicate a different model for the salicylate-induced hearing loss through the reduction of KCNQ4 and subsequent depolarization of OHCs, which reduces the driving force for transduction current and electromotility. The major mechanism underlying the reduction of I_K,n (KCNQ4) is the direct blocking action of salicylate on KCNQ4.

INTRODUCTION

Normal hearing in mammals depends on both sound detection by inner hair cells (IHCs) and sound amplification by outer hair cells (OHCs). The voltage-dependent somatic motility of OHCs, which is driven by membrane potential, is a major component of the cochlear amplification process (Ashmore 2008; He et al. 2006; Mellado Lagarde et al. 2008; Santos-Sacchi 2003; Santos-Sacchi et al. 2006; Zheng et al. 2003). The resting membrane potential (RP) of OHCs is dependent on I_K,n, a low-voltage activated K⁺ conductance, which is carried by the KCNQ4 channel and is specifically blocked by linopirdine (Chambard and Ashmore 2005; Holt et al. 2007; Kharkovets et al. 2000, 2006; Xu et al. 2007). KCNQ4 mutation that abolishes I_K,n results in depolarization of OHCs and a selective degeneration of OHCs, accounting for a human nonsyndromic deafness DFNA2, an autosomal dominant hearing loss (Kharkovets et al. 2006).

Salicylate, one of the most widely used nonsteroidal anti-inflammatory drugs in the world, has been effectively used to treat various diseases for centuries. However, salicylate has the unique side effect of inducing a temporary hearing loss and tinnitus (Cazals 2000; Myers and Bernstein 1965). Salicylate-induced tinnitus has recently been found to possibly be mediated by cochlear N-methyl-d-aspartate (NMDA) receptors on IHCs. Salicylate caused an increase of cochlear arachidonic acid level, which enhanced the NMDA response to glutamate at the basal pole of IHCs (Guitton et al. 2003; Ruel et al. 2008). Additionally, salicylate causes a moderate and reversible elevation of auditory thresholds (∼20–40 dB) and abolishes otoacoustic emissions (OAEs). These effects have been attributed to impaired sound amplification by OHCs through direct salicylate action on OHC electromotility (Kakehata and Santos-Sacchi 1996). Moreover, most animal studies have shown that the effective salicylate inhibitory concentrations on electromotility are extremely high, ranging from 1 to 10 mM (Hallworth 1997; Kakehata and Santos-Sacchi 1996; Lue and Brownell 1999; Tunstall et al. 1995). In chinchilla and guinea pigs, serum salicylate concentrations can reach around 1–3 mM after treatment of high doses of salicylate (Boettcher et al. 1990; Jastreboff et al. 1986). In contrast, the effective salicylate concentrations in human plasma that elicit hearing loss range from 0.1 to 1 mM. (Bonding 1979; Cazals 2000; Day et al. 1989; Jardini et al. 1978; McFadden and Plattsmier 1983; McFadden et al. 1984; Myers and Bernstein 1965; Ramsden et al. 1985). This paradox has previously been noted, but without satisfactory explanation (Cazals 2000).

In this study, we show that salicylate at a physiological concentration range (0.1–1 mM) caused a reversible reduction in I_K,n of OHCs and KCNQ4 current expressed in Chinese hamster ovarian (CHO) cells and subsequent depolarization of OHCs. This finding suggests that the reduction of I_K,n (KCNQ4) is the mechanism underlying physiological salicylate-induced hearing loss in humans.

METHODS

Ethical approval

All procedures in this study were approved by the Institutional Animal Care and Use Committee of Oregon Health and Science University.

OHC preparation

Adult guinea pigs (250–300 g) with positive Preyer's reflex were anesthetized by intramuscular injection of an anesthetic mixture (60 mg/ml ketamine, 2.4 mg/ml xylazine, and 1.2 mg/ml acepromazine in saline) at a dose of 1 ml/kg and were killed by decapitation. The cochleae were rapidly removed from the bulla and dissected in a petri dish filled with a standard artificial perilymph composed of (mM) 144 NaCl, 4 KCl, 1.3 CaCl₂, 0.9 MgCl₂, 0.7 Na₂HPO₄, 10 HEPES, and 5.6 glucose. The osmolarity of the solution was adjusted to 304 mosmol/l with glucose, and the pH was adjusted to 7.4 with NaOH. All procedures were performed at room temperature. The organ of Corti dissected was digested with dispase I (1.0 mg/ml) for 12 min. Dissociated OHCs, obtained by gentle titration, were placed in a petri dish and allowed to settle onto the glass bottom. They were continuously perfused with dissection solution. OHCs with lengths ranging from 65 to 68 μm, which present both typical I_K,n and I_K currents, were chosen for patch recordings.

Solutions for OHC recordings

All reagents were from Sigma-Aldrich (St. Louis, MO). Linopirdine and H-89 dihydrochloride (H-89) were predissolved in dimethylsulfoxide (DMSO), which was used at a final concentration of 0.1% DMSO. Salicylate, BaCl₂, and 4-AP were directly dissolved in solutions. The drugs were gravity delivered at a rate of 0.35 ml/min, through an array of parallel polyethylene tubes with an ID of ∼280 μm at the distal end. The polyethylene tube was positioned at a distance of ∼350 μm from the OHCs. Rapid switching of two different solutions was performed by shifting the loaded tubes. Bath solutions (in mM) containing 142 NaCl, 5 KCl, 1.5 CaCl₂, 2 MgCl₂, 10 HEPES, and 5.6 d-glucose were adjusted to a pH of 7.4 with NaOH and osmolarity to 304 mosmol/l with d-glucose. Regular pipette solutions (in mM) containing 148 KCl, 0.5 CaCl₂, 2 MgCl₂, 10 HEPES, and 1 EGTA (final Ca²⁺ concentration is 50 nM) were adjusted to pH of 7.4 with KOH and osmolarity to 298 mosmol/l with d-glucose. Ca²⁺-free pipette solutions (in mM) containing 140 KCl, 2 MgCl₂, 10 HEPES, and 5 EGTA were adjusted to a pH of 7.4 with KOH and osmolarity to 298 mosmol/l with d-glucose.

Whole cell recordings of OHCs

In current detection procedures, an Axopatch 1-D amplifier (Axon Instruments) was used with its low-pass filter bandwidth set to 1 kHz (4-pole Bessel). Membrane currents were recorded for episodic I-V commands with a Digidata 1322A (Axon Instruments) interface and pCLAMP 8 software (Axon Instruments) at a sampling rate of 10 kHz and for simultaneous gap-free recording with Minidigi digitizer and Axoscope 9.2 softeware (Axon Instruments) at a sampling rate of 50 Hz. Through a gap-free recording mode, we could continuously monitor the change of current at a holding potential (V_h) of −45 mV when salicylate was administered and decided on the time to apply step voltage commands for I-V plots. The whole cell configuration was achieved by rupturing the cell membrane with suction after forming a high-resistance seal (>1.5 GOhm). Stability of the patch was ascertained by monitoring the gap-free recording and the cell parameters [cell capacitance (C_cell), membrane resistance (R_m), and series resistance (R_s)] during the recordings. The patch pipettes were pulled with a puller (P80/PC, Sutter Instrument) in four steps from the borosilicate capillaries (WPI, 1B150F-4) with initial resistance of 4–5 Mohm in our regular Na⁺-rich bath and K⁺-rich pipette solutions. The uncompensated R_s, 17.1 ± 1.0 (Mohm; n = 20), was corrected off-line with the equation V_c = V_u − I × R_s (V_c, corrected clamping voltage; V_u, uncorrected clamping voltage; I, current) in Microsoft Office Excel spreadsheets and Origin 7.5 (OriginLab Technical, Northampton, MA) files. To stabilize the liquid junction potential (LJP), a salt bridge (3 M NaCl) with a ceramic tip was used as a reference electrode. The LJP (actual measurement) was 3 mV in regular Na⁺-rich bath and K⁺-rich pipette solutions. The LJP was corrected in Microsoft Office Excel spreadsheets and Origin 7.5 (OriginLab Technical) files. The data were analyzed by clampfit 9.0 (Axon Instruments, Union City, CA) and origin 7.5 (OriginLab Technical, Northampton, MA).

Fluorescent imaging of intracellular Ca²⁺ of OHCs

The organ of Corti was incubated with fluorescent dye, Fluo-4 (10 μM), for 30 min. After Fluo-4 was thoroughly washed by perilymph solution, the organ of Corti was digested with dispase and gently titrated. Dissociated OHCs were placed in a petri dish and allowed to settle onto the glass bottom. The perfusion pipette is usually from ∼5 to 6 mm away from OHCs to avoid moving cells when solution is switched. A confocal laser microscope (Olympus Fluoview FV1000) was used to acquire time-lapse imaging of fluorescence signals (with 488-nm excitation and 520-nm emission filters). The fluorescence intensity was normalized by the intensity in control bath before capsaicin application.

Tissue culture and transfection

CHO cells were obtained from American Type Culture Collection (ATCC, Manassas, VA), and maintained in DMEM containing 10% FBS (Gibco BRL, Bethesda, MD) and 1% penicillin/streptomycin. Cell cultures were kept at 37°C in a 5% CO₂ incubator. The cells were trypsinized, plated at a concentration of 1.5 × 10⁵ cells/ml in 2 ml of culture medium in 35-mm dishes and transfected with 1 μg of total KCNQ4 DNA per dish (Xu et al. 2007). Transfections were performed using Lipofectamine and following the manufacturer's protocol (Invitrogen, Carlsbad, CA). The cells were rinsed in fresh culture medium and incubated for 24 h before patch-clamp recordings. Transfected cells were identified for recording by visualization of the enhanced green fluorescent protein (EGFP) co-transfection (BD Bioscience, Clontech, Mountain View, CA).

Electrophysiology of KCNQ4 expressed in CHO cells

Whole cell voltage-clamp recordings were performed from CHO cells at room temperature (20–22°C) using an Axopatch 200A amplifier (Axon Instruments, Union City, CA). Fire-polished electrodes (3–5 MΩ) were pulled from borosilicate glass. The electrodes containing (in mM) 140 KCl, 1 MgCl₂, 10 HEPES, 10 EGTA, 1 CaCl₂, and 4 K₂ATP was adjusted to pH of 7.2 with KOH. The external bathing solution was constantly perfused (∼2–3 ml/min) and contained (in mM) 145 NaCl, 4 KCl, 1.8 CaCl₂, 0.5 MgCl₂, 10 HEPES, and 5 d-glucose, pH 7.40, with NaOH. Outward KCNQ4 current traces were generated with depolarizing voltage steps from a holding potential of −70 mV and stepped to varying positive potentials (ΔV = 10 mV). Currents were measured with capacitance compensation and series resistance compensation (>90%), filtered at 2 kHz using an eight-pole Bessel filter, and sampled at 5 kHz. Whole cell K⁺ current amplitude at varying test potentials was measured at the peak and steady-state levels using a peak and steady-state detection routine; the current was divided by the cell capacitance (pF) to generate the current density-voltage relationship.

Nonstationary fluctuation analysis was used to estimate the number (N) of functional channels in the membrane. For a homogeneous population of channels gating independently, the mean macroscopic current (I) is defined as follows: I = N × i × p_o. The macroscopic variance σ² is defined as follows: σ² = N × i² × p_o × [1 − p_o], where i is the single-channel current amplitude and p_o represents the open probability of the channel. The variance was averaged over all the records collected (Tsien et al. 1986). Provided that I and σ² are determined for a range of open probabilities, I and N can be estimated by a plot of variance versus mean current fit by the following parabolic function: σ² = i × I − I²/N (Sigworth 1980). Analyses of data were performed using custom-made software and Microcal Origin (Northampton, MA) programs. Pooled data are presented as means ± SD. Unless otherwise indicated, reagents were obtained from Sigma-Aldrich (St. Louis, MO).

Statistics

Data are presented as means ± SD from n observations. Student's t-test (paired or nonpaired as appropriate) was performed using Microsoft Office Excel software; P ≤ 0.05 represents a significant difference.

RESULTS

Salicylate blocked I_K,n in OHCs

In our regular Na⁺-rich bath and K⁺-rich pipette solutions, OHCs presented a steady-state outward current in gap-free recording mode, when held at –45 mV (Fig. 1A). Salicylate (0.1 mM) caused a reversible reduction of the outward current by shifting it negatively. With the step voltage protocols, OHCs presented two previously described K⁺ currents in the control bath (Fig. 1B) (Housley and Ashmore 1992); I_K (in green), a slowly developing outward current showing slight inactivation (activated above −40 mV), and I_K,n (in red), a current that is activated and shows relaxation during large hyperpolarizing steps (−120 mV). Salicylate (0.1 mM) reduced the steady-state currents (the average bound by arrows) of the control bath over the testing voltages from −45 to +20 mV (Fig. 1, B and C), but reduced the currents with the large I_K,n before relaxation (9 ms after each step started) over a larger voltage range (Fig. 6). The reversal potential (V_r) of the net salicylate-sensitive current (I_sal) was −74.6 ± 3.4 (n = 8), suggesting a high K⁺ selectivity (Fig. 1, B and C [the K⁺ equilibrium potential (E_K) was −88 mV]). Salicylate-sensitive current (I_sal) at each voltage step was defined as a difference in the current between before and after salicylate treatment; i.e., the current in salicylate treatment (b in Fig. 1B) is subtracted from the current in the control condition (a in Fig. 1B) at each step potential.

Fig. 1.

Salicylate blocks a K conductance. A: representative current elicited by salicylate (0.1 mM) at V_h of −45 mV. Current spikes, a, b, and c are currents elicited by step voltage command protocols in B. B: currents induced by step voltage commands at control bath (a and c) and salicylate application (b). In a, outer hair cell (OHC) presented a typical I_K (in green), a slowly developing outward current showing slight inactivation, and I_K,n (in red), which is activated and shows relaxation during large hyperpolarizing steps (−120 mV). The salicylate sensitive net current (I_sal) is a − b. The 0 current is shown as a solid horizontal bar preceding the step currents. C: representative I-V curves from current spike a, b, c, and c − b. The current for each corrected step voltage (V) is the average bounded by arrows as shown in C.

Fig. 6.

Effect of intracellular Ca²⁺ ([Ca²⁺]_i) on salicylate (sal)-induced current reduction. I-V curves were produced from the currents (at 9 ms after each step voltage started) in control and after salicylate (sal) application under different [Ca²⁺]_i, Ca²⁺ free (left), and 50 nM (right). The data were plotted as mean ± SD. The step voltage command protocol in Fig. 1 was used. *Significant current reduction (P < 0.05, paired t-test).

Figure 2 shows the effect of salicylate on the steady-state activation curve of I_K,n, which was determined from tail currents with the voltage protocol in Fig. 1. Salicylate (0.1 mM) reduced I_max (the fully activated current amplitude at the tail-current potential) of I_K,n by 35.5% (from 1.00 ± 0.02 to 0.648 ± 0.02, n = 8), with little change of V_half (half-maximal activation potential). The V_half (mV) and slope factor (S) (mV) are −84.2 ± 8.7 (n = 8) and 29.8 ± 5.4 (n = 8), respectively, for salicylate treatment, versus −85.2 ± 4.4 (V_half) and 24.2 ± 2.8 (S) for control. The V_half of I_K,n was consistent with previous findings (Housley and Ashmore 1992; Jagger and Ashmore 1999).

Fig. 2.

Effect of salicylate (0.1 mM) on steady-state activation curve of I_K,n. Tail-current voltage protocol in Fig. 1 was used to determine the activation curve. Normalized tail-current amplitudes vs. prepulse potentials were fitted with a 1st-order Boltzmann function (n = 8): I = I_max/{1 + exp[(V − V_half)/S]}, where I_max is the fully activated current amplitude at the tail-current potential, V_half is the potential at half-maximal activation (black dotted line for ctrl and gray dotted line for salicylate treatment), V is prepulse potentials, and S is the slope factors. Currents were normalized to I_max for each paired experiment.

At the holding voltage (V_h) of −45 mV where I_K,n is predominantly activated, salicylate, with IC₅₀ = 10 μM, produced a dose-dependent reduction of mean I_K,n (Fig. 3) by 3.8 (10⁻⁷ M), 8.0 (10⁻⁶ M), 14.3 (10⁻⁵ M), 28.3 (10⁻⁴ M), 32.0 (10⁻³ M), and 33.6% (10⁻² M). The concentration-response curve was fitted with the Hill equation, V = V_max [1 − C^h/(IC₅₀^h + C^h)], with IC₅₀ = 10⁻⁵ M, h = 0.53, and V_max = 0.35 (the goodness of fit is 0.992), where V_max is the maximal response, C is concentration, and h is the Hill coefficient. Here, salicylate sensitive reduction of I_K,n at V_h of −45 mV was determined by the difference in current between the steady-state current in the control bath and that with salicylate treatment. The steady-state current at the V_h of −45 mV after salicylate treatment was normalized by the steady-state current in the control bath, which was set at 100% for each experiment. The currents after salicylate treatment relative to control level were 0.962 ± 0.008 (10⁻⁷ M, n = 3), 0.920 ± 0.020 (10⁻⁶ M, n = 5), 0.857 ± 0.027 (10⁻⁵ M, n = 6), 0.7169 ± 0.0254 (10⁻⁴ M, n = 8), 0.6796 ± 0.007 (10⁻³ M, n = 5), and 0.6636 ± 0.035 (10⁻² M, n = 6). Because repeated application and washout reduced the response of some OHCs to salicylate, only responses from the first one or two applications of salicylate (different concentrations) were included in the statistics for each OHC.

Fig. 3.

Concentration-response curve of salicylate. Currents elicited by salicylate at −45 mV were normalized by the steady-state outward current at −45 mV in control bath before salicylate application.

Figure 4 shows that I_K,n-specific blockers, linopirdine (0.1 mM) and BaCl₂ (10 mM), blocked most of I_sal (0.1 mM salicylate) over the whole range of testing V_h (Fig. 4, A–D). At a V_h of −45 mV, 0.1 mM linopirdine (a specific KCNQ blocker) blocked I_sal (0.1 mM salicylate) by 94.7%; from 91.3 ± 4.1 to 4.8 ± 1.5 pA (n = 5, P < 0.05; Fig. 4E); 10 mM BaCl₂ (a I_K,n blocker) blocked I_sal (0.1 mM salicylate) at a V_h of −45 mV by 88.6%, from 99.9 ± 6.7 to 11.4 ± 1.5 pA (n = 3, P < 0.05; Fig. 4E).

Fig. 4.

Linopirdine and BaCl₂ block the I_sal. A and B: representative current recordings at V_h of −45 mV in gap-free mode with salicylate (sal) (0.1 mM) in the gray bar, and linopirdine (0.1 mM) (A) and BaCl₂ (10 mM) (B) in white bars. C: salicylate (100 μM)-sensitive net currents elicited by step voltage command protocols as shown in Fig. 1 in control bath, bath with linopirdine (100 μM), and BaCl₂ (10 mM). Zero current is shown as a solid horizontal bar preceding the step currents. D: representative I-V curves of net currents in C. The current for each V_h is the average bounded by the arrows. E: bar graph summary: the I_sal (0.1 mM salicylate) at V_h of −45 mV in the control, linopirdine (100 μM), and BaCl₂ (10 mM) bath. *P < 0.05, n = 5 and 3.

Figure 5 shows the effect of linopirdine and 4-AP on I_K,n and I_K. Linopirdine (100 μM), the specific KCNQ4 blocker, blocks most of I_K,n without changing I_K, the slowly developing outward current with slight inactivation (Fig. 5, A and B). Consistent with previous reports, 4-AP (100 μM), the known specific I_K blocker, blocks all the slowly developing outward kinetics of I_K without changing I_K,n (Fig. 5, A and B) (Mammano and Ashmore 1996).

Fig. 5.

Effect of linopirdine and 4-AP on I_K,n and I_K. A: representative currents induced by step-voltage commands following application of linopirdine (100 μM) and 4-AP (300 μM). The control current is for linopirdine (d from Fig. 4). A: the control for 4-AP are not shown here. The 0 current is shown as a solid horizontal bar preceding the step currents. B: net currents are after subtraction from control currents (c from Fig. 4). A: in bath. C: the representative I-V curves were produced from linopirdine- (c − d) and 4-AP–sensitive currents.

Effect of OHC [Ca²⁺]_i on I_K,n reduction elicited by salicylate

At the regular basal level of [Ca²⁺]_i, 50 nM, salicylate significantly reduced I_K,n at a membrane potential (V_h) of −40 mV by 28.8 (10⁻⁴ M), 31.7 (10⁻³ M), and 33.1% (10⁻² M) (P < 0.05; Fig. 6). In comparison, at Ca²⁺-free [Ca²⁺]_i, salicylate significantly reduced I_K,n at a V_h of −40 mV by 11.2 (10⁻³ M) and 39.3% (10⁻² M) (P < 0.05); 10⁻⁴ M salicylate did not significantly reduce I_K,n (Fig. 6). A time-lapse [Ca²⁺]_i imaging study (Fig. 7) shows that 1 mM salicylate significantly increased OHC [Ca²⁺]_i by 6% (P < 0.05, n = 4) at 70 s after salicylate application at both supranuclear and basal cytoplasmic regions. A relatively long distance between the perfusion pipette and the cell resulted in a relatively slow application rate for salicylate, which caused [Ca²⁺]_i to increase at a slower rate compared with current changes in whole cell recording.

Fig. 7.

Effect of salicylate on [Ca²⁺]_i in time-lapse photography. The [Ca²⁺]_i fluorescence intensities at basal and supranuclear cytoplasmic regions of 4 OHCs were acquired at 1.7 interval before and after application of salicylate (1 mM) and control bath (both in gray bars). Each [Ca²⁺]_i fluorescence intensity was normalized by the intensity in control bath (set as 1) before salicylate or control bath application. The data were plotted as mean ± SD (vertical bar) from 4 OHCs. Representative imaging of OHC before (A) and 75 after salicylate (1 mM) (B) application.

Salicylate depolarized OHCs

In current-clamp mode, salicylate (1 mM) significantly depolarized OHCs (Fig. 8, A and B) by 14.3 ± 1.9 mV, shifting the resting potential (RP) from −64.6 ± 2.1 to −50.3 ± 3.0 mV (n = 8, P < 0.05), consistent with the reduction in I_K,n, the main driving current of RP. The depolarization was completely reversible after a 8–10 min wash. Pretreatment with linopirdine (0.1 mM) depolarized the OHCs by 28.9 ± 5.6 mV (n = 3). Application of salicylate (1 mM) did not produce significant additional change of RP (mV), from −41.90 ± 4.90 (linopirdine) to −41.93 ± 4.70 (salicylate) (n = 3, P < 0.05).

Fig. 8.

Salicylate depolarizes OHC. A: representative recordings of membrane potential with salicylate (1 mM) in the gray bar and linopirdine (0.1 mM) in the white bar. B: bar graph summary of membrane potentials in control bath (Ctrl) and salicylate (1 mM); *P < 0.05, n = 8.

Salicylate blocked KCNQ4 current expressed in CHO cells

To confirm that the effects of salicylate are mediated through direct drug effect on the KCNQ4 channels, we expressed the cloned channel from the inner ear and determined the pharmacology of the channel in CHO cells. Salicylate produced a concentration-dependent reduction of KCNQ4 current at membrane potential of −40 mV: by 6.0 (10⁻⁶ M), 16.0 (10⁻⁵ M), 22.0 (10⁻⁴ M), 30 (10⁻³ M), and 44% (10⁻² M) (Fig. 9, A, B, C, and E). The estimated IC₅₀ was 0.9 ± 0.1 mM (n = 6), and the Hill coefficient was 0.68. The activation curves did not present a significant change in V_half and slope factor. V_half for control was –24.4 ± 1.0 mV and the slope factor, S, was 14.1 ± 0.9 mV. After application of salicylate, the V_half was –22.3 ± 1.0 mV and the slope factor, S, was 14.2 ± 0.8 mV (n = 9). We further evaluated the effects of PKA inhibitor (H89) on the KCNQ4. In accord with previous reports (Jagger and Ashmore 1999), H89 (1 μM) reduced the magnitude of the KCNQ4 current as shown in Fig. 10. A reduction in whole cell current amplitude may ensue from altered unitary current amplitude, probability of channel openings, and changes in the number of functional channels. Using nonstationary fluctuation analysis as shown in Fig. 11, under control conditions, the estimated single-channel current amplitude was 1.1 ± 0.3 pA (n = 5), and the number of channels was estimated to be 1,500 + 200 (n = 5). After application of salicylate, the estimated single-channel amplitude dropped to 0.4 ± 0.3 pA (n = 5), and the estimated number of functional channels also reduced to 500 ± 80 (n = 5). Because, salicylate did not alter the voltage-dependent activation curves substantially, it can be inferred that the effects of the drug may be conferred through reduction in the unitary current amplitude and/or changes in the number of functional channels.

Fig. 9.

A: examples of outward KCNQ4 current traces recorded from a holding potential of −70 mV to step potentials ranging from −100 to 30 mV using voltage increment of 10 mV steps. The tail current was measured at −40 mV. B: similar recording conditions in the same cell after application of 100 μM salicylate. C: current density (in pA/pF)-voltage relations of summary data from 9 cells are shown. *Voltage at which there were significant differences (P < 0.05, paired t-test) in the current magnitude. D: the activation curves were generated from the tail currents and fitted with a Boltzmann function; the V_half for control was –24.4 ± 1.0 mV and the slope factor, S, was 14.1 ± 0.9 mV. After application of salicylate, the V_half was –22.3 ± 1.0 mV, and the slope factor, S, was 14.2 ± 0.8 mV. The data represent mean ± SD from 9 cells. E: the dose-response curve of the effects of salicylate. The estimated IC₅₀ was 0.9 ± 0.1 mM (n = 6).

Fig. 10.

Current traces were generated from KCNQ4-tranfected CHO cells. The current traces were elicited using voltage protocols similar to the one described in Fig. 9. I-V relationships of KCNQ4 currents recorded in the absence (■) or presence (•) of the protein kinase A (PKA) inhibitor H89. In the presence of H89, the magnitude of the KCNQ4 current plummeted, as shown in the I-V relationships (C).

Fig. 11.

Nonstationary fluctuation analysis of tail currents of KCNQ4 currents. The mean current at 60 mV is plotted vs. variance. The cell capacitance for the example shown is ∼10.1 pF. Data from 200 consecutive current traces collected at 5 s intervals are plotted. The lines represent best fits to the function σ² = i × I – (I²/N), where σ², i, I, and N represent variance, single-channel current amplitude, macroscopic current, and total number of channels, respectively. The goodness of fit is 0.98 for control and 0.99 for salicylate treatment. Under control conditions, the estimated single-channel current amplitude was 1.1 ± 0.3 pA (n = 5); the number of channels was estimated to be 1,500 ± 200 (n = 5). After application of salicylate, the estimated single-channel amplitude dropped to 0.4 ± 0.3 pA (n = 5), and the estimated number of functional channels also reduced to 500 ± 80 (n = 5).

DISCUSSION

Salicylate inhibited I_K,n (KCNQ4)

The effect of salicylate on the whole cell current of OHCs has been previously observed (Shetata et al. 1991; Tunstall et al. 1995). However, the effect was not completely characterized. The salicylate-sensitive current (I_sal) was not isolated, and its concentration-response relationship was not established. Consequently, the potential role of membrane current change in salicylate-induced hearing loss has not been recognized. In this study, we showed that salicylate, analogous to linopirdine and Ba²⁺, partially blocked I_K,n (KCNQ4-mediated current). The supporting evidence is 1) salicylate caused a substantial reduction of a K⁺ conductance of OHCs at Vh of −45 mV where I_K,n is fully activated and is the dominant K⁺ current while I_K is not activated; 2) I_sal was almost completely blocked by the specific I_K,n (KCNQ4) blocker, linopirdine, and Ba²⁺; and 3) salicylate caused a subsequent depolarization of OHCs. The RP of OHCs has been found to be dependent on I_K,n, a current carried by the KCNQ4 channel (Chambard and Ashmore 2005; Holt et al. 2007; Kharkovets et al. 2000, 2006). OHC depolarization will reduce the net driving force for transduction current and subsequently reduce OHC electromotility. Salicylate, unlike 4-AP (the specific I_K blocker), did not block the characteristic slowly developing outward I_K (Figs. 1B and 5).

In previous studies, a V_h of 0 mV was generally used to evaluate changes of outward current of guinea pig OHCs (Shetata et al. 1991; Tunstall et al. 1995), because it creates a large driving force for K⁺ currents. However, a V_h of 0 mV activates both I_K and I_K,n (Housley and Ashmore 1992). Moreover, both I_K and I_K,n are prone to decay from unhealthy OHCs, which may account for the variability of current recovery after washout of salicylate in previous studies (Shetata et al. 1991; Tunstall et al. 1995). In this study, full recovery of I_K,n after washout (monitored at a V_h of −45 mV through gap-free recording mode) was accomplished for most of healthy OHCs (Fig. 1), even in high salicylate concentration ranges (1 and 10 mM).

Salicylate inhibited I_K,n in the physiological concentration range

There is a disparity between functional salicylate concentrations used in animal OHC electromotility studies and the plasma levels found in clinical observations of human hearing loss. Most animal studies used extremely high concentrations of salicylate (from 1 to 10 mM) (Hallworth 1997; Kakehata and Santos-Sacchi 1996; Lue and Brownell 1999; Tunstall et al. 1995); whereas, human plasma concentrations ranged from 60 (0.38 mM) to 500 mg/l (3.13 mM), equivalent to the perilymph concentrations from 0.1 to 0.9 mM, which were significantly lower compared with those used in animal studies (Bonding 1979; Cazals 2000; Day et al. 1989; Jardini et al. 1978; McFadden and Plattsmier 1983; McFadden et al. 1984; Ramsden et al. 1985). A serum concentration of 4 mM (∼1.3 mM in perilymph, which is near 30% of plasma concentration) can be lethal (Cazals 2000; Myers and Bernstein 1965).

Therefore, it is reasonable that another mechanism may contribute to the salicylate-induced hearing loss in addition to direct impairment of OHC electromotility. In this study, we proposed that the mechanism is the reversible and partial suppression of I_K,n. The reduction of I_K,n by 14.3, 28.3, and 32% for 0.01, 0.1, and 1 mM salicylate concentrations, respectively, overlaps the concentration range producing reversible and mild or moderate hearing loss in clinical observations. It should be noted that we observed repeated application and washout of salicylate (>3 times) reduced the response of OHCs to salicylate (data not shown). We were thus unable to apply all the concentrations on one OHC to generate a concentration-response curve. In this study, we used the first one or two responses (different concentrations) to salicylate for each healthy OHC and normalized the current with I_K,n in a pretreatment control bath. This approach showed that I_K,n reduction induced by salicylate was dose dependent. In addition, consistent with OHC findings, salicylate at a physiological range (0.01, 0.1, and 1 mM) produced a significant reduction of the current of KCNQ4 expressed in CHO cells. The reduction in the current amplitudes is comparable to that observed in OHCs. However, the IC₅₀ (0.9 mM) is relatively higher for CHO cells, which is reasonable because I_K,n is generated by KCNQ4 and its auxiliary subunits. It is conceivable that the in vivo accessory subunits of the channel may alter the pharmacology (Heitzmann et al. 2007; Radicke et al. 2008).

Mechanism underlying salicylate suppression of I_K,n (KCNQ4)

It has been found that I_K,n (KCNQ4) is regulated by background phosphorylation through cAMP and protein kinase A (PKA) and dephosphorylation by protein phosphatase (Chambard and Ashmore 2005; Jagger and Ashmore 1999), e.g., suppression of cAMP or protein PKA pathway changes the voltage sensitivity (shifting V_half ) and accordingly downregulates KCNQ4. Because there was no significant shift of the V_half of the activation curve of I_K,n by salicylate, we surmise the effects of the drug precludes the cAMP/PKA-dependent mechanism.

Moreover, it is likely the salicylate that elicited whole cell current reduction results from altered unitary current amplitude and changes in the number of functional channels through salicylate's direct action on KCNQ4 channel alone without interfering with the other second messenger mechanism present in OHCs. In this regard, we tested the effect of salicylate on KCNQ4 expressed in CHO cells. Indeed, we found that salicylate at a physiological range significantly reduced the current of KCNQ4 expressed in CHO cells by a comparable amplitude to OHCs without significantly shifting V_half. In addition, nonstationary fluctuation analysis shows that salicylate significantly reduced the estimated single-channel current amplitude and numbers. These data suggest that salicylate has a direct blocking action on KCNQ4 channel, which is the major mechanism of the drug.

Salicylate-elicited intracellular acidosis is another possible mechanism to be considered. Salicylate is a weak acid and can acidify OHC cytoplasm (Tunstall et al. 1995). It is known that acidification can release free Ca²⁺ from its bound form and accordingly increase free [Ca²⁺]_i. Indeed, our time lapse Ca²⁺ imaging study showed that salicylate increased [Ca²⁺]_i at both supranuclear and basal cytoplasmic regions of OHC where KCNQ4 is mainly located. There are several studies showing that increase of [Ca²⁺]_i downregulates I_K,n of OHC, KCNQ4, and other KCNQ family members (KCNQ2/3) through calmodulin (CaM) and/or other unidentified pathways (Chambard and Ashmore 2005; Gamper and Shapiro 2003; Xu et al. 2007). Therefore it is likely that salicylate-elicited [Ca²⁺]_i elevation contributes to the reduction of I_K,n (KCNQ4) of OHCs. To test this hypothesis, we used two different sets of [Ca²⁺]_i: 50 nM (the regular basal level) and Ca²⁺ free. Indeed, under free intracellular Ca²⁺_, salicylate can inhibit I_K,n (KCNQ4) at only higher concentration levels (1 and 10 mM).

It has been reported that salicylate could be absorbed in the artificial phospholipid bilayer membranes and increased potassium permeability through producing a negative electrostatic surface potential (McLaughlin 1973). However, it does not seem to account for our finding, because salicylate decreased KCNQ4 permeability in this study.

It should be mentioned that I_K,n was also present in IHCs at a small size (∼45% that of OHCs) (Marcotti and Kros 1999; Marcotti et al. 2003; Oliver et al. 2003). It is quite likely that salicylate also produced a reduction of I_K,n in IHCs, which may also contribute to the moderate clinical hearing loss induced by salicylate.

In summary, salicylate produced a reversible reduction of I_K,n (KCNQ4) of OHCs at the concentration range known to elicit hearing loss in the clinic. The subsequent depolarization of OHCs would consequently reduce the driving force for transduction current and electromotility, resulting in the moderate clinical hearing loss induced by salicylate. The direct inhibitory action of salicylate on KCNQ4 is the major mechanism. Intracellular Ca²⁺ elevation resulting from cytoplasmic acidosis also contributes to the current reduction of I_K,n (KCNQ4) of OHCs.

GRANTS

This study was supported by National Institute on Deafness and Other Communication Disorders Grants DC-005983, DC-000141 to A. L. Nuttall, DC-007592, and DC-010386 to E. N. Yamoah.