Effects of the antifungal antibiotic clotrimazole on human cardiac repolarization potassium currents

Abstract

1. The antifungal antibiotic clotrimazole (CLT) shows therapeutic effects on cancer, sickle cell disease, malaria, etc. by inhibiting membrane intermediate-conductance Ca²⁺-activated K⁺ channels (IK_Ca). However, it is unclear whether this drug would affect human cardiac K⁺ currents. The present study was therefore designed to investigate the effects of CLT on transient outward K⁺ current (I_to1), and ultra-rapid delayed rectifier K⁺ current (I_Kur) in isolated human atrial myocytes, and cloned hERG channel current (I_hERG) and recombinant human cardiac KCNQ1/KCNE1 channel current (I_Ks) expressed in HEK 293 cells.

2. It was found that CLT inhibited I_to1 with an IC₅₀ of 29.5 μM, accelerated I_to1 inactivation, and decreased recovery of I_to1 from inactivation. In addition, CLT inhibited human atrial I_Kur in a concentration-dependent manner (IC₅₀=7.6 μM).

3. CLT substantially suppressed I_hERG (IC₅₀=3.6 μM), and negatively shifted the activation conductance of I_hERG. Moreover, CLT inhibited I_Ks (IC₅₀=15.1 μM), and positively shifted the activation conductance of the current.

4. These results indicate that the antifungal antibiotic CLT substantially inhibits human cardiac repolarization K⁺ currents including I_to1, I_Kur, I_hERG, and I_Ks. However, caution is recommended when correlating the observed in vitro effects on cardiac ion currents to the clinical relevance.

Introduction

Clotrimazole (CLT), a member of the imidazole family, is topically used for the treatment of mycoses (Weuta, 1974; Rodrigues et al., 1987; Yoshida & Aoyama, 1987). The antifungal effect is related to its inhibition of fungal sterol 14α-demethylase, a microsomal cytochrome P₄₅₀-dependent enzyme (Rodrigues et al., 1987; Yoshida & Aoyama, 1987). In addition, it was reported that CLT had therapeutic effects in sickle cell disease (Brugnara et al., 1996), secretory diarrhea, cancer, etc. These effects are related to the blockade of the intermediate-conductance Ca²⁺-activated potassium channel (IK_Ca) in human erythrocytes, colonic epithelium, and many tumor cells (Alvarez et al., 1992; Brugnara et al., 1996; Rufo et al., 1997; Vandorpe et al., 1998; Jensen et al., 1999; Khanna et al., 1999; Wulff et al., 2000). CLT was found also to block L-type Ca²⁺ channel current (I_Ca.L) in cardiac myocytes (Xiao et al., 1998; Thomas et al., 1999), voltage-gated K⁺ currents in mouse pancreatic β-cells (Welker & Drews, 1997), and several types of cloned Kv channels expressed in mammalian cell lines (Wulff et al., 2000).

Although systemic administration is suggested, limited information is available in the literature as to whether CLT affects human cardiac repolarization K⁺ currents. The present study was therefore designed to investigate the effects of CLT on transient outward K⁺ current (I_to1), ultra-rapid delayed rectifier K⁺ current (I_Kur) in human atrial myocytes, and cloned hERG channel current (I_hERG) and recombinant human cardiac KCNQ1/KCNE1 channel current (I_Ks) expressed in HEK 293 cells with whole-cell patch and/or perforated patch techniques.

Methods

Preparation of human atrial myocytes

Atrial myocytes were isolated from specimens of human right atrial appendage obtained from patients (52.4±2.7 years old; range from 24 to 75 years old) undergoing coronary artery bypass grafting. The procedure for obtaining the tissue was approved by the Ethics Committee of the University of Hong Kong, and a written consent was obtained from patients. All atrial specimens were grossly normal at the time of cardiac surgery, and all patients were free of supraventricular tachyarrhythmias and symptomatic congestive heart failure. After excision, the samples were quickly immersed in oxygenated, nominally Ca²⁺-free cardioplegic solution for transport to the laboratory. Atrial myocytes were enzymatically dissociated with a modified procedure as described previously (Du et al., 2004). Briefly, the myocardial tissue was minced with a sharp blade and then placed in a 15-ml tube containing 10 ml of the Ca²⁺-free Tyrode solution (36°C), gently agitated by continuous bubbling with 100% O₂. After 15 min (5 min at a time in fresh solutions), the chunks were incubated for 50 min in a similar solution containing 150–200 U ml⁻¹ collagenase (CLS II, Worthington Biochemical, Freehold, NJ, U.S.A.), 1.2 U ml⁻¹ protease (type XXIV, Sigma Chemical, St Louis, MO, U.S.A.), and 1 mg ml⁻¹ bovine serum albumin (Sigma). Subsequently, the supernatant was discarded and the chunks were re-incubated in a fresh enzyme solution with the same composition, but without protease. Microscope examination of the medium was performed every 5–10 min to determine the number and the quality of the isolated cells. When the yield appeared to be maximal, the chunks were suspended in a high K⁺ medium and gently blown with a pipette. The isolated myocytes were kept at room temperature in the medium for at least 1 h before study.

A small aliquot of the solution containing the isolated cells was placed in an open perfusion chamber (1 ml) mounted on the stage of an inverted microscope. Myocytes were allowed to adhere to the bottom of the chamber for 5–10 min and were then superfused at 2–3 ml min⁻¹ with Tyrode solution. Only quiescent, rod-shaped cells showing clear cross-striations were selected for experiments.

Cell culture and gene transfection

A previous study had reported that human cardiac KCNQ1/KCNE1 channels transiently expressed in HEK 293 cells demonstrated the typical characteristics of I_Ks (Zhang et al., 2001). After co-transfecting hKCNQ1/pCEP4 and hKCNE1/pALTER-Max vectors (provided by Dr G.N. Tseng, Virginia Commonwealth University) into HEK 293 cells using Lipofectamine 2000™ (Invitrogen, Carlsbad, CA, U.S.A.) according to the instruction of the manufacturer, the cell line stably expressing the recombinant human cardiac KCNQ1/KCNE1 channels was selected in 200 μg ml⁻¹ hygromycin (Sigma-Aldrich, St Louis, MO, U.S.A.), and was maintained in Dulbecco's modified Eagle's medium (DMEM, Invitrogen) supplemented with 10% fetal bovine serum and 100 μg ml⁻¹ hygromycin. The vector of hERG/pcDNA3 provided by Dr Gail Robertson (University of Wisconsin-Madison, WI, U.S.A.) was transfected transiently into HEK 293 cells using Lipofectamine 2000™.

Solution and drugs

Ca²⁺-free cardioplegic solution for transport of specimens contained (in mM): KH₂PO₄ 50, MgSO₄ 8.0, adenosine 5.0, HEPES 10, glucose 140, mannitol 100, and taurine 10, with pH adjusted to 7.3 with KOH. The standard Tyrode's solution contained (in mM): NaCl 140, KCl 5.0, MgCl₂ 1.0, CaCl₂ 1.0, NaH₂PO₄ 0.33, HEPES 5.0, glucose 10, with pH adjusted to 7.4 with NaOH. For the atrial tissue wash, Ca²⁺ was omitted. High-K⁺ storage medium contained (in mM): KCl 10, K-glutamate 120, KH₂PO₄ 10, MgSO₄ 1.8, taurine 10, HEPES 10, EGTA 0.5, glucose 20, and mannitol 10, with pH adjusted to 7.3 with KOH. The pipette solution contained (in mM): KCl 20, K-aspartate 110, MgCl₂ 1.0, HEPES 10, EGTA 5.0, and GTP 0.1, Na₂-phosphocreatine 5.0, Mg₂-ATP 5.0, with pH adjusted to 7.2 with KOH. For I_to1 and I_Kur determination, BaCl₂ (200 μM) and CdCl₂ (200 μM) were added to the superfusion to block I_K1 and I_Ca. Atropine (1.0 μM) was used to minimize possible acetylcholine-activated K⁺ current (I_K,Ach) contamination during the current recording.

CLT (Sigma-Aldrich) was prepared as 50 mM stock solutions in dimethyl sulfoxide (DMSO, Sigma-Aldrich) and added to the bath solution at the indicated final concentrations. The DMSO concentration in the perfusion was <0.2% (v v⁻¹) and caused no effect on the membrane currents. Experiments were conducted at room temperature (21–22°C).

Data acquisition and analysis

The whole-cell and/or perforated patch-clamp techniques were used as described previously. Briefly, borosilicate glass electrodes (1.2 mm OD) were pulled with a Brown-Flaming puller (model P-97, Sutter Instrument Co., Novato, CA, U.S.A.) and had tip resistances of 2–3 MΩ when filled with pipette solution. A 3-M KCl-agar salt bridge was used as reference electrode. Liquid junction potentials were compensated before the pipette touched the cell. After a gigaseal was obtained, the cell membrane was ruptured by gentle suction to establish the whole-cell configuration to record I_to1, I_Kur, and I_hERG. Perforated patch configuration was used to record I_Ks to minimize rundown of the current. The pipette solution included 200 μg ml⁻¹ amphotericin B (Sigma-Aldrich), and omitted GTP, ATP, and EGTA. The series resistance was electrically compensated to minimize voltage errors. The membrane currents were recorded with an EPC-10 amplifier and Pulse software (HEKA, Lambrecht, Germany). Command pulses were generated by a 12-bit digital-to-analog converter controlled by Pulse software. Current signals were low-pass filtered at 5 kHz and stored on the hard disk of an IBM computer.

Values are presented as mean±s.e. Nonlinear curve fitting was performed using Pulsefit (HEKA) and/or Sigmaplot (SPSS Science, Chicago, IL, U.S.A.). Paired and/or unpaired Student's t-tests were used to evaluate the statistical significance of differences between two group means. ANOVA was used for multiple groups. Values of P<0.05 were considered statistically significant.

Results

Effect of CLT on I_to1

Figure 1A shows voltage-dependent I_to1 elicited by 300-ms voltage steps from −50 to between −40 and +60 mV (as shown in the inset) in a representative human atrial myocyte during control, in the presence of CLT, and after the drug washout. I_to1 was substantially inhibited by the application of 10 μM CLT, and the effect was partially reversed by washout (10 min).

Figure 1.

Effects of CLT on I_to1 in human atrial myocytes. (A) Voltage-dependent I_to1 traces (capacitance compensated) recorded at 0.2 Hz in a representative myocyte with 300-ms voltage steps from −50 to between −40 and +60 mV (inset) during control (a), in the presence of 10 μM CLT (b), or after washout (c). (B) Time course of I_to1 recorded in a typical experiment in the absence and presence of 10 μM CLT.

Figure 1B illustrates the time-dependent effect of CLT on I_to1 activated by a 300-ms voltage step (the left inset). The current measured was from peak to the ‘quasi'-steady-state level. CLT at 10 μM gradually inhibited I_to1, and the effect reached a steady-state level within 10 min. The inhibition was partially reversed by washout. The original I_to1 traces at the corresponding time points are shown in the right inset of the panel.

Results from Figure 1 indicate that the sustained current (i.e. I_Kur) is also substantially reduced when I_to1 is inhibited by CLT. Although the I_to1 was measured from peak to the 'quasi'-steady-state level, it might be confounded by effects from other currents. We have recently found that verapamil inhibits I_Kur without reduction of I_to1 amplitude, while it induces an increase of measured I_to1 in human atrial myocytes (Gao et al., 2004). Therefore, verapamil at 10 μM was used to separate I_to1 from I_Kur as described below.

Figure 2A displays voltage-dependent I_to1 recorded in a representative myocyte with an identical protocol as shown in Figure 1A during control, in the presence of 10 μM verapamil, and co-presence of verapamil with 30 μM CLT. I_to1 amplitude was actually increased by the application of verapamil to inhibit I_Kur as described previously (Gao et al., 2004). CLT at 30 μM substantially suppressed I_to1. Figure 2B shows the I–V relationships of I_to1 in six cells in the presence of 10 μM verapamil to inhibit I_Kur (control), and after the application of 10 and 30 μM CLT. CLT significantly decreased I_to1 at test potentials from 0 to +50 mV (P<0.05 or 0.01 vs control). Figure 2C illustrates the concentration-response relationship for the inhibition of I_to1 by CLT. Data were fitted to the Hill equation:

where E is the effect at concentration C, E_max is the maximal effect, IC₅₀ is the concentration for half-maximal inhibitory effect, and b is Hill coefficient. The IC₅₀ (at +50 mV) for the inhibition of I_to1 by CLT was 29.5 μM, and the Hill coefficient was 1.7.

Figure 2.

Effects of verapamil and CLT on I_to1. (A) I_to1 traces recorded in a representative myocyte with the same voltage protocol as shown in the inset of Figure 1A during control (a), in the presence of 10 μM verapamil for 10 min (b), in the co-presence of verapamil and 30 μM clotrmazole (CLT) for 10 min (c). Verapamil induced an increase of measured I_to1 by selectively inhibiting I_Kur. (B) I–V relationships of I_to1 in the presence of 10 μM verapamil (control), co-presence of verapamil and 10 or 30 μM CLT (10 min for each concentration). CLT inhibited I_to1 in a concentration-dependent manner (n=6, P<0.05 or 0.01 at 0–+60 mV vs control). (C) Concentration–response relationship for the inhibition of I_to1 by CLT. Symbols are mean data at +50 mV, and solid line is the best-fit to Hill equation. IC₅₀ was 29.5 μM, Hill coefficient was 1.7, and E_max was 74.9%. The numbers in parentheses are numbers of experiments.

Figure 3a shows I_to1 traces recorded in a representative cell with a 300-ms voltage step from −50 to +50 mV in the presence of 10 μM verapamil (to inhibit I_Kur, control) and in the co-presence of verapamil and 10 μM CLT. I_to1 was well fitted to a mono-exponential function in the absence (control) and presence of CLT with the time constants shown. Figure 3b summarizes the averaged time constants, and the time constant of I_to1 inactivation was reduced from 40.3±4.2 ms in control to 24.5±1.9 and 21.2±3.2 ms (n=6, P<0.01 vs control), respectively, with 10 and 30 μM CLT. These results are consistent with the high-affinity open-channel block causing rapid current decay.

Figure 3.

Effects of CLT on the kinetics of I_to1. (a) I_to1 traces recorded from a representative cell upon a 300-ms voltage step from −50 to +50 mV in the presence of 10 μM verapamil (control) and co-presence of verapamil and 10 μM CLT. Raw data (points) of I_to1 were fitted to a mono-exponential function (solid lines, superimposed with raw data) with time constants shown. (b) Mean values of time constants at +50 mV during control, in the presence of 10 and 30 μM CLT. The time constant was reduced by the application of 10 and 30 μM CLT (n=6, ^**P<0.01 vs control). (c) Normalized voltage-dependent variables for I_to1 activation (Act.) and inactivation (Inact.) were fitted to Boltzmann distribution: y=1/{1+exp[(V_m−V_0.5)/S], where V_m is the membrane potential, V_0.5 is the midpoint, and S is the slope. V_0.5 for activation conductance of I_to1 was 10.9±1.0 mV for control and 12.1±1.3 mV for 10 μM CLT (n=6, P=NS), while S was 12.5±1.0 and 13.8±1.7 mV for control and CLT, respectively (P=NS). For inactivation, V_0.5 and S were −23.3±0.4 and −5.9±0.5 mV during control, −23.5±0.6 and −7.9±0.6 mV with 10 μM CLT (n=6, P=NS). (d) Recovery of I_to1 from inactivation was determined by 300-ms paired pulses from −80 to +50 mV after a 30-ms step of −40 mV (to inactivate I_Na) with varying P1 and P2 interval (inset). Data were fitted to bi-exponential functions. Recovery time constants (τ₁ and τ₂) of I_to1 were decreased by the application of 10 μM CLT (n=6, P<0.01).

Figure 3c displays voltage-dependent activation and inactivation of I_to1 evaluated in the absence and presence of 10 μM CLT with pretreatment with 10 μM verapamil. Voltage-dependent activation was determined from the I–V relationship for each cell in Figure 2B as described previously (Gao et al., 2005). Voltage-dependent inactivation was determined with the protocol as shown in the left inset (with 1-s conditioning pulses from voltages between −100 and +30 mV, followed by a 20-ms pulse to −40 mV to inactivate Na⁺ current and then a 300-ms test pulse to +50 mV). Data were fit to a Boltzmann distribution to obtain the half-activation or inactivation voltage (V_0.5) and the slope (S). The mean data for activation and inactivation in the absence and presence of 10 μM CLT are displayed in Figure 3c. V_0.5 for activation conductance of I_to1 was 10.9±1.0 mV in control, and 12.1±1.3 mV in CLT (n=6, P=NS), while S was 12.5±1.0 and 13.8±1.7 mV, respectively, for control and 10 μM CLT (P=NS). V_0.5 and S of I_to1 inactivation were −23.3±0.4 and −5.9±0.5 mV during control, and −23.5±0.6 and −7.9±0.6 mV with 10 μM CLT (n=6, P=NS).

Recovery of I_to1 from inactivation was analyzed with a paired-pulse protocol, as shown in the inset of Figure 3d. The recovery curves were fitted to bi-exponential functions with the time constants τ₁ and τ₂ of 0.56±0.1 and 50.5±1.1 ms in control, and 4.7±0.4 and 89.2±6.9 ms after 10 μM CLT (n=6, P<0.01 vs control). The results indicate that recovery of I_to1 from inactivation was decreased by CLT.

CLT on I_Kur

I_Kur was recorded using a 100-ms prepulse to +40 mV to inactivate I_to1, followed by a 150-ms test pulse from −50 to between −40 and +50 mV, then to −30 mV as described previously (Gao et al., 2004; 2005). Figure 4A illustrates voltage-dependent I_Kur traces recorded in a representative cell with the voltage protocol as shown in the inset in the absence and presence of CLT. CLT at 10 μM substantially decreased both I_Kur and tail current, and the effect was reversed by washout. Figure 4B displays the time-dependent effect of CLT on I_Kur recorded in a typical experiment with the voltage protocol shown in the left inset. I_Kur was measured from zero level to the current at the end of the voltage step. I_Kur was gradually decreased by CLT, and the effect partially recovered upon washout. The original I_Kur traces at the corresponding time points are shown in the right inset of the panel.

Figure 4.

Effect of CLT on I_Kur. (A) Representative voltage-dependent I_Kur (capacitance compensated) recorded at 0.2 Hz in a typical experiment with a 100-ms prepulse to +40 mV to inactivate I_to1, followed by 150-ms test pulses from −50 to between −40 and +50 mV after a 10-ms interval, then to −30 mV (as shown in the inset) under control conditions (a), in the presence of 10 μM CLT (b). I_Kur was substantially suppressed by the application of CLT, and the effect was significantly recovered by washout of the drug for 20 min (c). (B) Time-dependent effects of 10 μM CLT on I_Kur elicited by a 150-ms voltage step from −50 to +50 mV (as shown in the left inset) delivered every 10 s. The original I_Kur traces at the corresponding time points are shown in the right inset.

Figure 5a shows the I–V relationships of I_Kur in the absence and presence of 5 and 10 μM CLT. I_Kur was substantially inhibited by CLT. At +50 mV, I_Kur was inhibited by 26.5±4.9 and 57.7±3.3% with 5 and 10 μM CLT (n=10, P<0.01 vs control), respectively, and the effect partially recovered on washout. No voltage-dependent effect was observed (data not shown). Figure 5b illustrates the concentration–response relationship for inhibition of I_Kur by CLT at +50 mV. The IC₅₀ was 7.4 μM with a Hill coefficient of 1.8.

Figure 5.

Concentration-dependent effect of CLT on I_Kur. (a) I–V relationships of I_Kur during control, in the presence of CLT at 5 and 10 μM, and after washout. CLT inhibited I_Kur in a concentration–dependent manner, and the effect was partially reversed by 77.1% upon drug washout. (b) Concentration–response relationship of I_Kur inhibition by CLT at +50 mV. Symbols are the mean values of inhibiting effect in cells exposed to different concentrations. Solid lines are the best-fit Hill equation. IC₅₀ was 7.4 μM, b was 1.8, and E_max was 87.1%. The numbers in parentheses are numbers of experiments.

CLT on I_hERG

Figure 6A shows voltage-dependent I_hERG traces recorded in a HEK 293 cell expressing hERG channels with 1-s voltage steps from −80 to between −60 and +60 mV, then back to −50 mV as shown in the inset at 0.1 Hz, in the absence and presence of CLT. The step I_hERG (I_hERG.step) and tail current (I_hERG.tail) were substantially suppressed by the application of 3 and 10 μM CLT, and the effect was partially reversed by washout. Figure 6B illustrates the time-dependent effect of CLT on I_hERG.step (measured from zero current to the level at the end of voltage step) recorded in a typical experiment by the voltage protocol shown in the left inset. The current was gradually suppressed by 3 μM CLT, and the effect significantly recovered upon washout. The original I_hERG tracings at the corresponding time points are shown in the right inset.

Figure 6.

Effects of CLT on I_hERG. (A) Voltage-dependent I_hERG traces recorded at 0.1 Hz in a HEK 293 cell expressing hERG channels with 1-s voltage steps from −80 to between −60 and +60 mV, then back to −50 mV as shown (inset) under control conditions (a), in the presence of 3 (b) or 10 (c) μM CLT, and washout (d). I_hERG was substantially inhibited by CLT (10 min exposure) at all voltages, and the effect partially recovered upon washout for 10 min. (B) Time-dependent effect of 3 μM CLT on the step I_hERG recorded in a typical experiment with the voltage protocol shown in the left inset. The original current traces at the corresponding time points are shown in the right inset.

Figure 7 illustrates the concentration and voltage dependence of CLT effects on I_hERG. Figure 7a and b displays the I–V relationships of relative I_hERG.step and I_hERG.tail in the absence and presence of 1, 3, and 10 μM CLT. The I_hERG.step and I_hERG.tail in the presence of CLT were relative to the maximum I_hERG.step and I_hERG.tail during control. Both I_hERG.step and I_hERG.tail were reduced with 1 and 3 μM CLT at potentials of +10 to +60 mV (n=9, P<0.01 vs control), and 10 μM CLT at −10 to +60 mV (P<0.01 vs control). Figure 7c shows the concentration–response relationship for inhibition of I_hERG.step by CLT at +30 mV. The IC₅₀ was 3.6 μM, and the Hill coefficient was 0.99. In addition, the IC₅₀ for I_hERG.tail was 6.4 μM with a Hill coefficient of 1.03.

Figure 7.

Concentration- and voltage-dependent effects of CLT on I_hERG. (a) I–V relationships of step I_hERG (I_hERG.step) with 1, 3 and 10 μM CLT relative to the maximum current during control. (b) I–V relationships of I_hERG tail (I_hERG.tail) with 1, 3, and 10 μM CLT relative to the maximum current during control. (c) Concentration–response relationship of I_hERG.step block by CLT at +30 mV. Symbols are mean values of inhibiting action exposed to different concentrations. Solid lines are best-fit Hill equation. For blockade of I_hERG.step, IC₅₀ was 3.6 μM, and b was 0.99, E_max was 93%. The numbers in parentheses are numbers of experiments. (d) Activation conductance of I_hERG was fitted to Boltzmann distribution. V_0.5 was negatively shifted with 3 μM CLT by 4.8 mV (from 2.8±1.8 to −2.0±1.7 mV, P<0.05), while S was not significantly altered (10.6±0.6 for control, and 11.2±0.5 for CLT, P=NS).

Voltage dependence of I_hERG activation was determined by normalizing tail current. The activation curves were fitted to a Boltzmann distribution. V_0.5 of the activation was negatively shifted by 4.8 mV (from 2.8±1.8 to −2.0±1.7 mV, P<0.05) with 3 μM CLT, while S was not significantly altered (10.6±0.6 for control and 11.2±0.5 for CLT, P=NS).

CLT on I_Ks

Figure 8A shows the time-dependent effect of CLT on I_Ks recorded with perforated configuration by a 3-s voltage step from −80 to +40 mV, then to −40 mV (left inset) in a representative HEK 293 cell stably expressing human recombinant I_Ks (KCNQ1/KCNE1). The voltage step-activated I_Ks (I_Ks.step) was measured from the significant activation of time-dependent current to the level at the end of depolarization step. The current was gradually decreased by 10 μM CLT, and reached a steady-state level within 8 min, and the effect recovered upon washout. The original I_Ks traces at the corresponding time points are shown in the right inset. Figure 8B displays voltage-dependent I_Ks recorded by the voltage protocol shown in the inset at 0.1 Hz (20-mV increment) in the absence and presence of CLT. I_Ks was substantially decreased by the application of 10 μM CLT, and the inhibitory effect was partially reversed by washout.

Figure 8.

Effects of CLT on I_Ks stably expressed in HEK 293 cells. (A) Time-dependent effect of 10 μM CLT on I_Ks-step recorded in a typical experiment with 3-s voltage step from −80 to +40 mV, then back to −40 mV as shown in the left inset. The original I_Ks traces at corresponding time points are shown in the right inset. (B) Voltage-dependent I_Ks traces recorded in a representative cell with the voltage protocol shown in the inset during control (a), in the presence of 10 μM CLT (b) and after washout (c). I_Ks was substantially inhibited by CLT (10 min exposure) at all voltages, and the effect recovered upon drug washout for 10 min.

Figure 9a shows the I–V relationships of I_Ks step current in the absence and presence of CLT. I_Ks was substantially inhibited at potentials of 0 to +60 mV by the application of 10 and 30 μM CLT (n=11, P<0.01 at −20 to +60 mV vs control). The effect was reversed by washout. Figure 9b displays the voltage dependence of I_Ks activation determined by normalizing I_Ks tail current in the absence and presence of CLT. V_0.5 of I_Ks activation was positively shifted by 10.4 mV (from 19.8±1.2 mV in control to 30.2±2.3 mV in CLT, n=10, P<0.01 vs control) with 10 μM CLT, while S was not affected (14.3±0.8 mV in control and 12.1±0.9 mV in CLT, P=NS). Figure 9c illustrates the concentration–response relationship of CLT for inhibiting I_Ks. IC₅₀ of CLT for I_Ks.step was 15.1 μM, Hill coefficient was 0.92. In addition, the measured IC₅₀ for I_Ks.tail was 11.7 μM, with a coefficient of 0.9.

Figure 9.

Concentration-dependent effect of CLT on I_Ks. (a) I–V relationships of I_Ks.step under control conditions, in the presence of 10, and 30 μM CLT, and washout. CLT inhibited I_Ks with increasing concentration, and the effect was reversed by 77% upon washout. (b) Activation conductance of I_Ks in the absence and presence of 10 μM CLT was fitted to Boltzmann distribution. V_0.5 of I_Ks activation was positively shifted by 10.4 mV with 10 μM CLT (from 19.8±1.2 mV in control to 30.2±2.3 mV in CLT, n=10, P<0.01 vs control), while S was not affected (14.3±0.8 in control and 12.1±0.9 in CLT, P=NS). (c) Concentration–response relationship of I_Ks-step inhibition by CLT at +40 mV. Symbols are the mean values of inhibiting effect. Solid lines are the best-fit Hill equation. IC₅₀ was 15.1 μM, b was 0.92, and E_max was 89%. The numbers in parentheses are numbers of experiments.

Discussion

CLT is a widely used imidazole antimycotic in the treatment of patients with systemic and/or topical mycosis. The antimycotic effect is attributed to its inhibition of cytochrome P₄₅₀ enzyme. CLT exerts its fungicidal effect by inhibiting fungal sterol 14α-demethylase, a microsomal cytochrome P₄₅₀-dependent enzyme (Rodrigues et al., 1987; Yoshida & Aoyama, 1987). Recently, CLT was found to have therapeutic effects in sickle cell disease, and cancers (Brugnara et al., 1996; Brugnara, 2003), and has been used as a potential immunosuppressant for the treatment of autoimmune disorders, such as rheumatoid arthritis, inflammatory bowel disease, and multiple sclerosis (Jensen et al., 2002). In addition, CLT was reported to be a promising antimalarial agent suitable for clinical study (Tiffert et al., 2000). Moreover, CLT was found to prevent experimental restenosis after angioplasty (Kohler et al., 2003). These later effects are generally believed to be based on its selective blockade of the intermediate-conductance Ca²⁺-activated K⁺ channels (IK_Ca) in different types of cells.

CLT was also reported to inhibit other ionic channels, including I_to1 in mouse ventricular myocytes (Hernandez-Benito et al., 2001), voltage-gated K⁺ currents in smooth muscle cells from rabbit portal vein, and cloned Kv1.5 channel expressed in HEK 293 cells (Iftinca et al., 2001), I_Ca.L in guinea pig (Thomas et al., 1999), and rat (Xiao et al., 1998) ventricular myocytes. The IC₅₀ of CLT for reducing I_Ca.L was 3.5 μM in rat ventricular myocytes. The present study provides additional information that CLT inhibits human cardiac repolarization K⁺ currents, including I_to1 and I_Kur in human atrial myocytes, and cloned I_hERG and I_Ks expressed in HEK 293 cells. Taken together, the order of IC₅₀ for different cardiac ion channel currents was I_Ca.L (3.5 μM)<I_hERG (3.6 μM)<I_Kur (7.4 μM)<I_Ks (15.1 μM)<I_to1 (29.5 μM).

I_to1 plays an important role in human atrial repolarization. CLT significantly inhibited I_to1 in myocytes isolated from human atrium (Figures 1 and 2), and increased the inactivation of I_to1, suggesting an open-channel blocking action. CLT at 10 μM had no effect on voltage-dependent activation and inactivation of I_to1 (Figure 3). These properties are consistent with those observed in mouse ventricular myocytes (Hernandez-Benito et al., 2001). In addition, CLT significantly slowed the recovery of I_to1 from inactivation in human atrial myocytes.

I_Kur is reported to be present in the atrium, but not in the ventricle of human heart (Li et al., 1996), and is believed to be encoded by Kv1.5 (Feng et al., 1997). In the present study, I_Kur was inhibited by CLT in a concentration-dependent manner, with an IC₅₀ of 7.4 μM in human atrial myocytes (Figure 5). The IC₅₀ observed in the present study was higher than that observed in cloned rabbit portal vein Kv1.5 expressed in HEK cells (K_d=2 μM) (Iftinca et al., 2001), but close to that in cloned hKv1.5 expressed in HEK 293 cells (K_d=8.0 μM) (Wulff et al., 2000). However, sustained K⁺ current was not affected by CLT in mouse ventricular myocytes (Hernandez-Benito et al., 2001), suggesting a genuine difference between I_Kur/hKv1.5 and the sustained K⁺ current in mouse ventricular cells. The inhibition of I_to1 and I_Kur in human atrial myocytes may be beneficial for inhibiting supraventricular arrhythmias.

However, the present study demonstrated that CLT inhibited cloned I_hERG expressed in HEK 293 cells with an IC₅₀ of 3.6 μM, showing a close similarity to I_hERG inhibition by another imadazole antimycotic (i.e. miconazole, IC₅₀=2.1 μM) in HEK 293 cells (Kikuchi et al., 2005), and a higher sensitivity, compared with that of I_hERG expressed in Xenopus oocytes by ketoconazole (IC₅₀=49 μM) (Dumaine et al., 1998). It is well known that blockade of I_Kr (i.e. hERG channel) and/or I_Ks channels may cause long QT syndrome and Torsade de Pointe, which trigger life-threatening ventricular arrhythmias (Roden, 2004). The IC₅₀ of CLT for blocking I_hERG was 3.6 μM, which is close to clinical blood plasma concentrations required to act as antisickling agent (1–2 μM) (Brugnara et al., 1995; Rifai et al., 1995). CLT also inhibited I_Ks with an IC₅₀ of 15.1 μM, close to that observed for I_Ks channels expressed in CHO cells (IC₅₀=11.2 μM). Moreover, CLT shifted the activation conductance of the current to more positive potentials. These effects of CLT, at concentrations which are in the same range used to induce antiproliferative action (Benzaquen et al., 1995; Raicu et al., 2000; Smith et al., 2000) on cardiac I_hERG and I_Ks would suggest a potential cardiac side effect; however, there are no case reports suggesting that CLT would induce QT prolongation of the ECG, although other imidazole antifungals (i.e. ketoconazole and miconazole) were reported to induce QT prolongation by blocking hERG channels (Dumaine et al., 1998; Wassmann et al., 1999; Kikuchi et al, 2005). However, since the most common use of this drug is as a topical cream for the treatment of fungal infections, systemic exposure to the agent would not be expected. Thus, for this clinical application, the effects on cardiac ion channels described herein may be irrelevant.

On the other hand, when CLT was used systematically as an antisickling agent, the blood plasma concentration of the compound would reach micromolar levels (Brugnara et al., 1995); however, it was shown to be 99% bound by plasma (Rifai et al., 1995). The 1% free CLT may not induce torsadogenic effect. In addition, CLT had an IC₅₀ of 3.5 μM for I_Ca.L suppression (Xiao et al., 1998), similar to that (3.6 μM) of hERG inhibition. Such a profile is similar to the I_Ca.L channel blocker verapamil. Verapamil significantly blocks I_hERG (Zhang et al., 1999), but has no torsadogenic effect (Redfern et al., 2003), which may be the case for CLT; however, this requires further investigation. Several potential uses of systemic administration for CLT are suggested (Tiffert et al., 2000; Jensen et al., 2002; Brugnara, 2003); further clinical studies are required to determine the efficacious dose or plasma drug level for each individual disease that needs a higher dose of CLT.

In summary, the present study provides the novel information that CLT significantly inhibits human cardiac repolarization potassium currents, including I_to1, I_Kur, I_hERG, and I_Ks in a concentration-dependent manner in isolated cells. However, caution is recommended when extrapolating the observed in vitro effects on cardiac ion currents to the clinical relevance.

Abbreviations

hERG

human ether-á-go-go-related K⁺ channel

I_hERG

hERG channel current

I_Ks

slowly activating delayed rectifier K⁺ current

I_Kur

ultra-rapid delayed rectifier K⁺ current

I_to1

transient outward K⁺ current

IC₅₀

the concentration for half-maximal inhibitory effect