Introduction

Top Summary Introduction Materials & Methods Results Discussion References

CCBs constitute an heterogeneous class of compounds with different chemical structures and varying potencies for blocking voltage-dependent Ca²⁺ channels (Hosey and Lazdunski, 1988). Their primary sites of action include cardiac muscle, systemic, and coronary arterial smooth muscle cells. Their largest domain of prescription is the management of hypertension, angina, and supraventricular arrhythmias. According to the type of voltage-dependent Ca²⁺ channel they block, one can distinguish T- or L-type calcium antagonists. Mibefradil is the only T-type calcium antagonist approved by the Food and Drug Administration (Clozel et al., 1997). L-type calcium antagonists include the 1,4-dihydropyridines such as nitrendipine and isradipine, the phenylalkylamines such as verapamil, and the benzothiazepines as typified by diltiazem. Many other drugs have effects on L-type Ca²⁺ channels, like the diarylaminopropylamine bepridil. Verapamil, diltiazem, bepridil, and mibefradil block Ca²⁺ channels in cardiac cells at clinically relevant concentrations, resulting in a decrease in heart rate and atrioventricular nodal conduction velocity. This forms the basis of their antiarrhythmic efficacy in reentrant arrhythmias or in slowing the ventricular rate in the case of atrial flutter or fibrillation (at least for phenylalkylamines and benzothiazepines) (Roden, 1996). With few exceptions, calcium antagonists do shorten the cell action potential duration; this is why they do not prolong the refractory period and all except bepridil (Campbell et al., 1990) are considered ineffective as ventricular antiarrhythmic agents. Hence, verapamil, diltiazem, and mibefradil can increase the rate and duration of experimentally induced ventricular tachycardia (Billman and Hamlin, 1996), and inappropriate use of verapamil has been shown to jeopardize the outcome of Wolff-Parkinson-White syndrome by further shortening the refractory period of the accessory pathway and increasing the ventricular rate in the case of atrial fibrillation (Gulamhusein et al., 1983). Conversely, verapamil also has been shown to be of use in certain forms of ventricular tachycardia known as "verapamil sensitive" that seem to be triggered by delayed afterdepolarizations (Lauer et al., 1992; Gill et al., 1993; Lee et al., 1996). Bepridil, because of its ability to prolong the refractory period of ventricular myocytes and the QT interval on the electrocardiogram, has been successfully used in ventricular arrhythmias (Roden, 1996). This classic feature of class III antiarrhythmic drugs results from a blockade of I_K, which is considered to be an important modulator of the cardiac action potential repolarization. In most mammalian species, including humans, I_K is recognized as being composed of at least two currents: I_Kr and I_Ks (Sanguinetti and Jurkiewicz, 1990; Li et al., 1996). I_Kr (rapid component) activates quickly and exhibits inward rectification. I_Ks (slow component) has much slower kinetics and shows no inactivation. Regarding class III antiarrhythmic drugs, with excessive prolongation of the QT interval, bepridil also can induce polymorphic ventricular tachycardias known as torsades de pointes, especially in the setting of hypokaliemia, bradycardia, or both (Manouvrier et al., 1986). The possibility for verapamil to share, with bepridil, some class III type of activity gives those calcium antagonists a role in modulating the action potential duration (i.e., shortening or lengthening the action potential duration according to the pathophysiological settings). This also raises concerns about their use in populations at risk, such as patients with long QT syndrome (Napolitano et al., 1994).

Because previous studies have suggested that CCBs can interact with cardiac voltage-dependent K⁺ channels (Hume, 1985), we aimed to extensively explore the potency to block I_K of the most prescribed CCBs available on the market. The two types of K⁺ channels involved in the generation of I_Kr (Sanguinetti et al., 1995) and I_Ks (Attali, 1996; Barhanin et al., 1996; Sanguinetti et al., 1996) have been cloned, and mutations in the corresponding genes have been associated with different manifestations of the long QT syndrome (Roden et al., 1996; Chouabe et al., 1997). The I_Kr and I_Ks currents are generated by, respectively, the human ether-a-go go-related gene HERG product (Sanguinetti et al., 1995) and a K⁺ channel resulting from the assembly of two different proteins, KvLQT1 and IsK (Attali, 1996; Barhanin et al., 1996; Sanguinetti et al., 1996). To evaluate separately the effects of calcium antagonists on the two components of I_K, we used an in vitro mammalian cell model (COS-7 cells) transfected with either HERG or KvLQT1 and IsK coding sequences.

	Materials and Methods

Top Summary Introduction Materials & Methods Results Discussion References

Expression in COS cells. The pSI HERG expressing vector was kindly provided by D. J. Snyders (Snyders and Chaudhary, 1996). The human KvLQT1 and IsK coding sequences were amplified by polymerase chain reaction and respectively subcloned into the pCI and pCMV expression vector as described previously (Barhanin et al., 1996).

Electrophysiological methods. Electrophysiological recordings were carried out at 22 ± 2° (room temperature) in the whole-cell configuration of the patch-clamp technique (Hamill et al., 1981). A Petri dish containing cells was placed on the stage of an inverted microscope and superfused continuously with the standard extracellular solution. Patch pipettes with a resistance of 2-5 M were used routinely and connected electrically to the head stage of a RK300 patch-clamp amplifier (Biologic, Grenoble, France) with a 100-M feedback resistor. Junction potentials were zeroed with the pipette in the standard extracellular solution. After the whole-cell configuration was established, the capacitative transients elicited by symmetrical 10-mV voltage-clamp steps from 80 mV were recorded at 50 kHz (filtered at 10 kHz) for calculation of capacitative surface area and series resistance, which were 34.2 ± 0.8 pF and 4.8 ± 0.2 M, respectively (82 experiments). Membrane capacitance and series resistance were not compensated. Voltage commands and simultaneous signal recording were performed with pClamp software (Axon Instruments, Foster City, CA). A microperfusion system allowed local application and rapid change of the different experimental solutions. Superfusion flow rate was 50-100 µl/min.

Solutions and drugs. The intracellular pipette filling solution contained 150 mM KCl, 0.5 mM MgCl₂, 5 mM EGTA, and 10 mM HEPES/KOH, pH 7.2, and the standard extracellular solution contained 150 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂ and 10 mM HEPES/NaOH, pH 7.4. Nitrendipine (Bayer AG, Wuppertal, Germany), bepridil (Sigma), and mibefradil (F. Hoffmann-La Roche, Basel, Switzerland) were dissolved in dimethylsulfoxide (0.1 M). Diltiazem and verapamil (Sigma) were dissolved in distilled water (10 mM). The stock solutions were kept in the dark at 4°. The drug solutions were prepared fresh from these stock solutions and vortexed immediately before each use. The solvent concentration never exceeded 1:10,000.

Pulse protocols and analysis. The holding potential in all experiments was 80 mV. Current traces were uncorrected for the leak. In all experiments, recordings were started after a 5-min dialysis of the cell. The rundown of the current was negligible during the course of the experiments (generally <10% within 20 min for KvLQT1/IsK current, whereas HERG current was stable during this period). HERG currents were evoked every 7 sec by 2-sec depolarizing voltage steps ranging from 80 mV to +80 mV in 10-mV increments and then by repolarizing steps to 40 mV for 2 sec (sampling rate, 250 Hz; cutoff frequency, 80 Hz). KvLQT1/IsK currents were evoked every 10 sec by 4-sec depolarizing voltage steps ranging from 80 mV to +120 mV in 20-mV increments and then by repolarizing steps to 40 mV for 1 sec (sampling rate, 200 Hz; cutoff frequency, 60 Hz). For both HERG and KvLQT1/IsK currents, the current-voltage relationships were obtained by measuring the current amplitude at the end of the depolarizing voltage steps and at the peak of tail currents with respect to the zero of the A/D converter. The HERG and KvLQT1/IsK activation curves were determined by fitting peak values of tail currents (I_tail) versus test potential (V_t) to a Boltzmann function: I_tail = I_tail-max/(1 + exp[(V_0.5  V_t)/k], where I_tail-max is the maximum tail current. The voltage at which the current was half-activated (V_0.5) and the slope factor (k) were calculated from these data. The HERG inactivation curves were assessed by brief steps (20 msec) to various hyperpolarized potentials (ranging from 120 mV to +80 mV in 10-mV increments) from +50 mV. After allowing inactivation to relax to steady state at various test potentials, the membrane voltage was stepped to +50 mV to assess the relative number of channels available to activate. The HERG inactivation curves were determined by fitting the initial current measured on return to +50 mV versus the previous test potential to a Boltzmann function.

	Results

Top Summary Introduction Materials & Methods Results Discussion References

HERG and KvLQT1/IsK expression in COS cells. Cells were clamped at a holding potential of 80 mV. For HERG, depolarizing steps applied for 2 sec to voltages between 80 mV and +80 mV in 10-mV increments activated a time-dependent outward current that increased in amplitude from a threshold of 50 mV to peak at 0 mV (Fig. 1A). Depolarizing steps to more positive voltages resulted in an inward rectification because of a fast C-type inactivation (Sanguinetti et al., 1995). Deactivating outward tail currents elicited on repolarization to 40 mV were large as a result of the instantaneous removal of the inactivation process combined with a slow deactivation. HERG outward and tail currents were totally suppressed by the benzenesulfonamide antiarrhythmic agent E-4031 (5 µM, four cells; data not shown) as already described for I_Kr (Sanguinetti and Jurkiewicz, 1990). For KvLQT1/IsK, 4-sec depolarizing steps were applied to voltages ranging from 80 mV to +120 mV in 20-mV increments. A slowly time-dependent outward current developed from a threshold of 40 mV, and the amplitude increased linearly with more positive voltage steps (Barhanin et al., 1996; Chouabe et al., 1997) (Fig. 1B). Deactivating outward tail current recorded on repolarization to 40 mV increased in amplitude with depolarizing steps from 40 mV up to +100 mV, where they reached a maximum. The KvLQT1/IsK currents were E-4031 insensitive (5 µM, three cells; data not shown). The current-voltage relationships were evaluated for both currents at the end of the depolarizing steps and at the peak of tail currents. For HERG, the maximal amplitude of the outward current present at the end of the depolarizing step was of 1.1 ± 0.1 nA at 0 mV and the maximal amplitude of the tail current was of 2.0 ± 0.2 nA (31 cells; Fig. 1C). The amplitudes of KvLQT1/IsK at the end of the depolarizing steps to 0 mV and +60 mV were 1.0 ± 0.1 and 4.7 ± 0.3 nA, respectively, and the maximal amplitude of the tail current was of 1.9 ± 0.1 nA (14 cells; Fig. 1D). Because inactivation of the HERG channel was rapidly removed on repolarization, the plot of the normalized peak tail currents measured on repolarization at 40 mV versus the test potentials yielded the steady state activation curves (Fig. 1E). The mean control values of half-point activation and the corresponding slope factor were 13.0 ± 0.9 and 8.3 ± 0.2 mV, respectively (31 cells). The steady state HERG inactivation curves were assessed by brief steps to various hyperpolarized potentials from +50 mV. Normalized peak currents released from inactivation at +50 mV were plotted as a function of prepulse test potentials. The mean control values of half-point inactivation and the corresponding slope factor were 31.1 ± 3.1 and 16.9 ± 0.5 mV (29 cells). For KvLQT1/IsK, which does not inactivate, the relative activation curves were determined at the beginning of the deactivating tail currents after return to 40 mV (Fig. 1F). The mean control values of half-point activation and the corresponding slope factor were 26.9 ± 3.0 and 24.9 ± 1.3 mV (14 cells). Nontransfected COS cells or cells transfected with pCI plasmid alone exhibited no time-dependent current, as described (Barhanin et al., 1996; Chouabe et al., 1997) (Fig. 1B, inset).

View larger version (28K): [in this window] [in a new window]   Fig. 1.   HERG and KvLQT1/IsK expression in transfected COS cells. A, Typical HERG currents elicited by depolarizing voltage steps. Currents are activated by 2-sec pulses applied in 20 mV from 60 mV to +60 mV. Current magnitude progressively increased during the pulse and then decreased with voltage according to a voltage-dependent inactivation. Tail currents elicited on repolarization to 40 mV (voltage protocol inset) increased progressively with voltage and saturated at +20 mV. B, Typical KvLQT1/IsK currents elicited by depolarizing voltage steps. Currents are activated by 4-sec pulses applied in 20 mV from 60 mV to +100 mV. Time-dependent outward currents slowly increased during the pulse according to the voltage. Tail currents recorded on repolarization to 40 mV (voltage protocol inset) increased progressively with voltage and saturated at +100 mV. Top traces, superimposed current traces recorded in a nontransfected COS cell showing the absence of K+ currents on depolarization. C and D, K+ currents measured at the end of the test pulses () and at the peak of the deactivating tail currents after return to 40 mV () plotted against test potentials. Data points are mean ± standard error for 31 cells for HERG (C) and 14 cells for KvLQT1/IsK (D). The maximal amplitude of the HERG current occurs at ~0 mV (~1 nA), whereas the tail currents saturate at ~2 nA from +20 mV on. The amplitude of the KvLQT1/IsK current linearly augments whereas the tail currents saturate at ~ 2 nA from +100 mV on. E and F, Gating properties of HERG (E) and KvLQT1/IsK (F) channels in control condition. Steady state activation curves were determined for both channels at the peak of the deactivating tail current on return to 40 mV. The voltage dependence of HERG channel inactivation were assessed by 20-msec steps to test potentials ranging from 120 mV to +80 mV in 10-mV increments from +50 mV. After allowing inactivation to relax to steady state at these various test potentials, the membrane voltage was stepped to +50 mV. Steady state inactivation curves were determined by fitting the initial current measured on return to +50 mV versus previous test potentials. Data points were fitted to a Boltzmann distribution. Averaged half-point activation values (corresponding slope factors) were 13.0 ± 0.9 mV (8.3 ± 0.2 mV) for HERG (31 cells) and 26.9 ± 3.0 mV (24.9 ± 1.3 mV) for KvLQT1/IsK (14 cells). The mean value of HERG half-point inactivation and the corresponding slope factor were 31.1 ± 3.1 and 16.9 ± 0.5 mV (29 cells).

Effects of nitrendipine, diltiazem, verapamil, bepridil, and mibefradil on HERG and KvLQT1/IsK currents. A 10 µM concentration of the different CCBs first was tested on HERG and KvLQT1/IsK currents to evaluate the range of efficacy of the different drugs. Verapamil, bepridil, and mibefradil turned out to be potent HERG blockers (Fig. 2A), inducing a sharp decrease in both the outward and tail currents. A representative effect of bepridil on HERG current is shown on Fig. 2B. Diltiazem had a much smaller effect (14.6 ± 2.7% block of tail current; four cells), whereas nitrendipine (8.7 ± 2.9%; three cells) and isradipine effects (3.1 ± 1.8%; two cells, data not shown) were merely negligible. KvLQT1/IsK outward and tail currents were effectively blocked only by bepridil (Fig. 2C) and mibefradil with little or no effect from nitrendipine (5.2 ± 2.4%; three cells), verapamil (9.6 ± 3.5%; four cells), and diltiazem (8.0 ± 1.8%; three cells) (Fig. 2D).

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Effects of nitrendipine, diltiazem, verapamil, bepridil, and mibefradil on HERG and KvLQT1/IsK currents. A and D, Relative blockade of HERG (A) and KvLQT1/IsK (D) tail currents by an equimolar 10 µM concentration of calcium antagonists. Bepridil and mibefradil share an equivalent potency to block both currents. Verapamil inhibits only HERG current significantly. Comparatively, diltiazem does not exert a relevant effect on HERG, whereas nitrendipine seldom blocks any of the current. B and C, Representative HERG (B) and KvLQT1/IsK (C) currents elicited in transfected COS cells according to the voltage protocol under control conditions () and after bepridil (10 µM) (). Bepridil decreased both the outward currents elicited on depolarization and the tails resulting from return to 40 mV.

Concentration dependence of HERG blockade by verapamil. Verapamil reversibly blocked HERG currents in a dose-dependent manner (Fig. 3A). Both the outward and tail currents were suppressed in the presence of verapamil. Each given concentration (0.1-10 µM) was maintained until steady state block was attained (Fig. 3B). Blockade occurred rapidly and was completely reversible within 2 min of washout. The concentration dependence of the blockade of HERG tail currents was best fitted to the Hill equation that yielded an EC₅₀ value of 0.83 µM (Fig. 3C).

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Concentration-dependent blockade of HERG current by verapamil. A, HERG current was evoked according to the voltage protocol (inset). Top trace, recording of tail current on repolarization to 40 mV in control conditions. Increasing concentrations of verapamil (0.1-10 µM) dose-dependently decreased the current and its corresponding tail. B, Plot of the maximal amplitude of the tail current against time from the same cell as above during control condition and superfusion with increasing concentrations of verapamil. The blockade produced by verapamil was totally reversible within 2 min of washout (WO) with standard solution. C, Verapamil induced a concentration-dependent blockade of HERG tail currents (mean ± standard error, six cells). Data were best fitted with the Hill equation (inset, chemical structure of verapamil). The EC50 value is 0.83 µM (Hill coefficient, 1.19).

Verapamil influences HERG voltage-dependent gating. To evaluate the voltage-dependence characteristics of verapamil blockade, a concentration close to verapamil EC₅₀ (1 µM) was applied on HERG current. Verapamil significantly blocked both the outward and tail currents (Fig. 4, A and B). Verapamil shifted HERG steady state activation curves by ~8 mV to more negative values (Fig. 4C and Table 1). Half-point activation value was 13.9 ± 1.1 mV (control) and 21.9 ± 2.1 mV with verapamil (p < 0.05; 11 cells), with no significant change in slope factors [8.4 ± 0.3 and 7.9 ± 0.3 mV for control and verapamil, respectively; p > 0.1 (NS)]. Steady-state inactivation curves also were shifted to more negative values by ~22 mV. Half-point inactivation value was 26.1 ± 4.6 mV (control) and 48.4 ± 4.5 mV with verapamil (p < 0.05, 11 cells), again with no significant change in slope factors [17.8 ± 0.6 and 20.2 ± 0.7 mV for control and verapamil, respectively; p > 0.1 (NS)].

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Effects of verapamil on HERG currents. A, Current traces of HERG were recorded by using depolarizing pulses to seven different potentials with 20-mV steps between 60 mV and +60 mV from a holding potential of 80 mV (voltage protocol inset). Left traces, control conditions. Right traces, recorded in the presence of verapamil (1 µM), which decreased significantly the outward and tail currents. B, Current-voltage relationships for the currents measured at the end of the 2-sec depolarizing pulse () and at the peak of the deactivating tail currents () on return to 40 mV in control conditions (mean ± standard error). Closed symbols, currents measured after verapamil (11 cells). Verapamil significantly decreased the tail currents according to the voltage and shifted the peak of the outward current by 10 mV to hyperpolarized values. Inset, normalized HERG tail currents () and relative tail current blockade magnitude () according to the test pulse potential. The intensity of verapamil blockade parallels the steady state activation of the current. C, Steady state activation and inactivation curves determined by Boltzmann equation in control conditions () and after verapamil (, 11 cells) characterized a shift toward more hyperpolarized values. Data are mean ± standard error (vertical bars). D, HERG current profile during a cardiac action potential. The cardiac action potential was simulated as a brief pulse (10 msec) from 90 mV to +50 mV followed by a ramp (400 msec) from +50 mV to 90 mV. This waveform (gray trace) was applied as voltage command. HERG currents (black traces) were recorded in control condition, in steady state condition during application of 1 µM verapamil, and after washing out with control solution.

Bepridil and mibefradil blockade of HERG current. Regarding verapamil, bepridil and mibefradil also decreased dose-dependently and reversibly both the HERG outward and tail currents. Similar analysis of the data obtained from each drug (six cells in both cases) resulted EC₅₀ values of 0.55 µM for bepridil and 1.43 µM for mibefradil (Fig. 5, A-D). Again, bepridil and mibefradil blockade of HERG tail current seemed to be voltage dependent (Fig. 6, A-D). Bepridil and mibefradil decreased the tail current amplitude by 4.4 ± 1.2% and 20.6 ± 6.0% at 20 mV and by 46.6 ± 8.0% and 52.8 ± 2.0% at +20 mV, respectively (10 cells, p < 0.05 in each case). Similar modifications of the HERG gating properties were observed with both drugs (Fig. 6, E and F, and Table 1).

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Concentration-dependent blockade of HERG current by bepridil and mibefradil. A and B, HERG current was evoked with 2-sec depolarizing pulses to +50 mV from a holding potential of 80 mV (voltage protocol inset). Tail current was recorded on repolarization to 40 mV. Increasing concentrations (0.1-10 µM) of bepridil (A) and mibefradil (B) dose-dependently decreased the current and its corresponding tail. In each case, the top trace corresponds to control conditions. C and D, Concentration-dependent blockade by bepridil (C, six cells) and mibefradil (D, six cells) of HERG tail currents (insets, chemical structures of bepridil and mibefradil). The Hill equation was fitted to the data with 100% blockade taken as the fixed maximal effect. EC50 values are 0.55 and 1.43 µM for bepridil and mibefradil, respectively. Hill coefficients are 1.35 (bepridil) and 1.57 (mibefradil). Data are expressed as mean ± standard error (vertical bars).

View larger version (35K): [in this window] [in a new window]   Fig. 6.   Effects of bepridil and mibefradil on HERG currents. A, Current traces of HERG were recorded by using depolarizing pulses to five different potentials with 20-mV steps between 40 mV and +40 mV from a holding potential of 80 mV (voltage protocol inset). Top traces, control conditions. Bottom traces, recorded in the presence of bepridil (0.5 µM). B, HERG traces in control conditions (top) and after mibefradil (1 µM) (bottom). The same voltage protocol is used as in A. C and D, Current-voltage relationships for the currents measured at the end of the 2-sec depolarizing pulse () and at the peak of the deactivating tail currents () on return to 40 mV in control conditions (mean ± standard error). Closed symbols, the currents measured after bepridil (C, 10 cells) and mibefradil (D, 10 cells). E and F, Steady state activation and inactivation curves determined by Boltzmann equation in control conditions () and after drug () for bepridil (E, 10 cells) and mibefradil (F, 8-10 cells) show a shift toward more negative value of both curves. Data are expressed as mean ± standard error (vertical bars).

Bepridil and mibefradil blockade of KvLQT1/IsK current. Increasing concentrations of bepridil and mibefradil (1-100 µM) were tested on KvLQT1/IsK currents. Bepridil and mibefradil dose-dependently blocked both the outward and tail currents (Fig. 7A). At each given concentration, the drug was maintained until steady state block was attained (Fig. 7B). Blockade occurred rapidly and was fully reversible when the drug was washed out. The concentration-effect relationships were fitted to the Hill equation and yielded EC₅₀ values of 10.0 and 11.8 µM for bepridil and mibefradil, respectively (Fig. 7, C and D).

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Concentration-dependent blockade of KvLQT1/IsK current by bepridil and mibefradil. A, KvLQT1/IsK current was evoked with 4-sec depolarizing pulses to +30 mV from a holding potential of 80 mV (voltage protocol inset). Tail current was recorded on repolarization to 40 mV. Top trace, control conditions. Increasing concentrations of bepridil (1-30 µM) dose-dependently decreased the current and its corresponding tail. B, Plot of the tail current amplitude against time from cell (A) before, during, and after superfusion with increasing concentrations of bepridil. WO, washout. C and D, Concentration-dependent blockade by bepridil (C, seven cells) and mibefradil (D, eight cells) of KvLQT1/IsK tail currents. The Hill equation was fitted to the data with 100% blockade taken as the fixed maximal effect. EC50 values are 10.0 and 11.8 µM for bepridil and mibefradil, respectively. Hill coefficients are 1.42 (bepridil) and 0.97 (mibefradil). Data are expressed as mean ± standard error (vertical bars).

View larger version (26K): [in this window] [in a new window]   Fig. 8.   Effects of bepridil and mibefradil on KvLQT1/IsK currents. A, Current traces of KvLQT1/IsK were recorded by using depolarizing pulses to five different potentials with 40-mV steps between 60 mV and +100 mV from a holding potential of 80 mV (voltage protocol inset). Top traces, control conditions. Bottom traces, recorded in the presence of bepridil (10 µM). B, KvLQT1/IsK traces in control conditions (top) and after mibefradil (10 µM) (bottom). The same voltage protocol is used as in A. C and D, Current-voltage relationships for the currents measured at the end of the 4-sec depolarizing pulse () and at the beginning of the deactivating tail currents () on return to 40 mV in control conditions (mean ± standard error). Closed symbols, currents measured after bepridil (C, eight cells) and mibefradil (D, six cells). E and F, Steady state activation curves determined in control conditions () and after drug () at the beginning of the deactivating tail currents on return to 40 mV (mean ± standard error) for bepridil (E, eight cells) and mibefradil (F, six cells).

	Discussion

Top Summary Introduction Materials & Methods Results Discussion References

The cardinal feature of this study is that calcium antagonists have various potentialities to block I_K. Bepridil and mibefradil block both HERG and KvLQT1/IsK currents, verapamil only blocks HERG current, and nitrendipine and diltiazem seldom block any of these currents.

Among the five CCBs that were tested, only verapamil, bepridil, and mibefradil blocked HERG current with an EC₅₀ value relevant to therapeutic concentrations. Open channel block by verapamil of native K⁺ currents different from I_Kr has been reported previously in humans (Pancrazio et al., 1991) and animal cell models (DeCoursey, 1995). Similar results also have been observed with an expression model of cloned K⁺ channels (Kv1.5 and Kv1.3) other than HERG (Rampe et al., 1993; Rauer and Grissmer, 1996). In this study, we have shown the effect of verapamil on I_Kr in a transfected COS cell model and found similar results with bepridil and mibefradil. These drugs, at their respective EC₅₀ values, produced similar changes in steady state activation and inactivation curves. The midpoint potential (V_0.5) of the activation curves was shifted ~8 mV in the hyperpolarizing direction with no significant change in the slope (k). For the inactivation curves, there was a greater hyperpolarization shift of V_0.5 (~22 mV) without any change in k. The greater shift of inactivation compared with activation only partially (~30%) accounts for the reduction in HERG currents during depolarizing voltage steps. Hence, in tail currents reflecting mainly the activation process, their blockade cannot be explained solely by the shift of inactivation. In fact, the degree of channel block directly correlates with channel activation (Fig. 4B, inset). The blockade of activation and the voltage shifts can be separate mechanisms of action as more clearly shown by using the simulated action potential protocol (Fig. 4D). Besides the blocking effect exerted by verapamil, the delay in the peak current during the ramp repolarization reflects the shift of the inactivation curve to more hyperpolarized voltages. In addition, verapamil seems to be an open-channel blocker of HERG channel (Fig. 4A).

Only bepridil and mibefradil block KvLQT1/IsK current with an EC₅₀ value close to 10 µM, which is comparable to that observed with the I_Ks blocker chromanol 293B (Loussouarn et al., 1997) or the antiarrhythmic drug amiodarone (Balser et al., 1991) and greater than that with azimilide, a new I_Ks blocker currently under development (Salata and Brooks, 1997). The blockade exerted by bepridil or mibefradil did not seem to be either voltage dependent or accompanied by a significant change in the channel gating. Nitrendipine, diltiazem, and verapamil did not have any effect on this current.

Ca²⁺ and K⁺ currents, flowing through different time- and voltage-dependent channels, are major determinants of the plateau and repolarization phases of the cardiac action potential. In most mammalian species, the I_K is composed of two main currents: a rapidly activating component, I_Kr, and a slowly activating component, I_Ks (Sanguinetti and Jurkiewicz, 1990). Similarly, I_Kr and I_Ks have been found in human cardiac cells (Li et al., 1996). Numerous attempts have been made to discriminate between these two components of the cardiac delayed rectifier, including the use of specific I_Kr blockers, Ca²⁺ channel blockers, -adrenoreceptor agonists, and low external Ca²⁺ and K⁺ (Sanguinetti and Jurkiewicz, 1990; Salata et al., 1996). Most of these conditions interfere with interpretation of the results. In our model, mammalian COS cells are transfected with either HERG or KvLQT1/IsK human coding sequences, therefore expressing separately each component of the I_K and avoiding any unneeded interfering condition.

One must consider the relevance of the concentrations that were used in this study. The therapeutic plasma levels of calcium antagonists range from 1 to 10 nM for dihydropyridines and up to several micromolar concentrations for verapamil (Rampe et al., 1993) and mibefradil (Clozel et al., 1997), which is well within the range of concentrations inducing a blockade of HERG and KvLQT1/IsK currents in our study.

It is recognized that agents prolonging cardiac repolarization, such as d-sotalol or amiodarone, might be antiarrhythmic because they inhibit the delayed rectifier I_K (Balser et al., 1991) and do prolong the refractory period of the ventricular cardiomyocytes. This mechanism can represent a relevant alternative to the suppression of the delayed afterdepolarization in the verapamil sensitivity of certain ventricular tachycardias (Lauer et al., 1992).

Conversely, the prolongation of the action potential has the ability to induce early afterdepolarizations that may trigger a complex form of ventricular arrhythmias known as torsades de pointes (Napolitano et al., 1994). This is the case for bepridil, which prolongs the action potential duration (Campbell et al., 1990) and was shown to be, apart from a potent antiarrhythmic agent, promoting polymorphic ventricular arrhythmias (Manouvrier et al., 1986). Verapamil does not seem per se to trigger such cardiac adverse events, and the slight differences observed in the EC₅₀ values of mibefradil and bepridil for HERG and KvLQT1/IsK currents hardly explain the widely different clinical profiles of these compounds (Massie, 1997), especially concerning the bepridil arrhythmogenicity. Several reasons might explain such a discrepancy. (1) Verapamil targets only the rapid component of the cardiac delayed rectifier, which renders bepridil and mibefradil likely to be more potent blockers of I_K by blocking both of its components. (2) The ratio of potassium and calcium channel blockade potency may be of importance in favoring prolongation over shortening of the action potential duration; in this study, bepridil at therapeutic concentrations has the same propensity to block Ca²⁺ channels (Yatani et al., 1986) than HERG channel. (3) Other ancillary properties, such as Na⁺ channel blockade, can modify the myocardial cell refractory period. Unlike many of the other CCBs, bepridil also inhibits fast Na⁺ channels (Yatani et al., 1986), thus imparting with a class I antiarrhythmic activity the same class proarrhythmogenicity (CAST, 1989). This may explain why verapamil decreases the occurrence of early afterdepolarizations (Singh, 1989; Cosio et al., 1991), whereas bepridil does not (Campbell et al., 1990).

Drug ancillary properties also are of importance in the therapeutic goal to achieve in patients. Amiodarone, for example, is a multifaceted drug that bears class I-IV antiarrhythmic efficacy (Nattel et al., 1992). It also has been shown to better preserve patients with heart failure from death (Pinto et al., 1997) than its more specific "pure" class I (CAST, 1989) or class III (Waldo et al., 1996) counterparts. This could be the case also for calcium antagonists bearing K⁺ channel blockade properties (DAVIT, 1990).

Study limitations should be considered. Obviously, species differences concerning the respective prominence of the delayed rectifier and the inward calcium currents in the process of repolarization should be considered. Nevertheless, mutations of HERG, KvLQT1, or IsK in the setting of the congenital long QT syndromes have emphasized the importance of the delayed rectifier in human cardiac repolarization (Roden et al., 1996; Chouabe et al., 1997). Regarding bepridil, it is possible that some individuals may be more susceptible to torsades de pointes than others. They may possess a mutation in a channel or regulatory element responsible for repolarization that renders them more susceptible to K⁺ channel blockade (Roden et al., 1996).

In conclusion, the dual approach of the I_K provided by our model permitted us to differentially quantify the potassium channel blockade exerted by calcium antagonists. The fact that drugs have the ability to block both K⁺ and Ca²⁺ currents renders them potentially useful "action potential modulators," for example, with arrhythmias complicating essential hypertension or stable angina.