Suppression of Transient Outward Potassium Currents in Mouse Ventricular Myocytes by Imidazole Antimycotics and by Glybenclamide

Materials and Methods

Preparation of Mouse Ventricular Myocytes.

The study has been carried out in accordance with the Declaration of Helsinki and with the institutional guides for the care and use of laboratory animals.

Single ventricular myocytes were obtained from adult mice. The animals were heparinized (250 IU, given intraperitoneally) and anesthetized with sodium pentobarbitone (Nembutal; 150–300 mg kg⁻¹, given intraperitoneally). The excised heart was cannulated via its aorta, mounted on a Langendorff system and perfused at 37°C and at constant flow (2.5 ml/min) for 2 to 5 min with an oxygenated normal Tyrode's solution. The heart was then perfused with Ca²⁺-free Tyrode's solution for 5 to 10 min, followed by a 15- to 20-min perfusion with a Ca²⁺-free Tyrode's solution containing 0.14 mg ml⁻¹ protease (type XIV; Sigma, St. Louis, MO) and 0.5 mg ml⁻¹ collagenase (type A; Roche Molecular Biochemicals, Mannheim, Germany), and a 10-min washing perfusion with Tyrode's solution in which the [Ca²⁺] was raised stepwise from 0.09 to 0.18 mM. The ventricle was cut into a few pieces in the 0.18 mM Ca²⁺ Tyrode's solution and cells were dispersed by gentle mechanical agitation. The cells were stored in the same solution at room temperature (21–22°C). Ca²⁺-tolerant rod-shaped ventricular myocytes with clear striations were selected randomly for the electrophysiological studies.

Electrophysiological Recordings and Data Analysis.

Membrane currents were measured as described before (Stengl et al., 1998) using the whole-cell patch-clamp technique (Hamill et al., 1981). Heat-polished borosilicate glass electrodes (horizontal puller; Zeitz Instrumente, Munich, Germany), with tip resistances of 0.5 to 1 MΩ when filled with the internal solution, were used. The electrodes were connected to an Axopatch 200B amplifier (Axon Instruments, Foster City, CA), and a DigiData 1200 (Axon Instruments) interface controlled by the pClamp 5.5.1 software (Axon Instruments) was used to generate command pulses and acquire data. All experiments were carried out at room temperature (21–22°C). The holding potential was set at −80 mV. Under our experimental conditions (0.18 mM Ca²⁺ in the extracellular solution) a small inward current component that could be attributed to Ca²⁺ current, I_Ca, was detectable in a small percentage of cells but was negligibly small, especially at positive potentials (<100 pA). Because of the small magnitude of I_Ca and the fact that typical blockers of L-type Ca²⁺ channel also block I_to (Gotoh et al., 1991) we decided not to use any Ca²⁺ channel blocker.

The time-dependent decay of the outward currents was fitted by a sum of exponentials:where I_fast and I_slow are the amplitudes of the fast decaying and slowly decaying components, respectively, τ_fastand τ_slow their respective time constants, and I_∞ the magnitude of the time-independent (noninactivating) component. To measure steady-state inactivation, prepulses lasting 5 s were given from the holding level to various potentials (between −120 and +70 mV, in 10-mV steps) before depolarizing to a test potential of +60 mV. Our preliminary experiments with prepulses lasting 0.4 to 2 s indicated that while such prepulses allowed steady-state inactivation of the fast component, longer prepulses were needed to allow steady-state inactivation of the slower component. Assuming full inactivation after a 5-s prepulse to the positive potentials, the lowest current at +60 mV following these prepulses was taken as baseline level. The total time-dependent current was measured as difference between peak current following a given prepulse and this baseline level, and was normalized relative to the current following the most negative prepulse (−120 or −100 mV). Similarly, the amplitudes of the fast (I_fast) and slow (I_slow) components at +60 mV following prepulses to various levels were obtained by exponential fitting and were normalized to the amplitude of the corresponding component following a prepulse to −120 or −100 mV. Normalized availability or inactivation curves were fitted using one single Boltzmann distribution function:(or a sum of two such functions) whereV_1/2 stands for the potential of half-maximum inactivation, and k for the slope factor.

Functions were fitted to data using Clampfit (Axon Instruments) or Origin (MicroCal, Northampton, MA). Average data are expressed as mean ± S.E.M. Statistical comparison was made using a two-tailed t test.

Solutions and Drugs.

The myocytes were superfused with a Tyrode's solution containing 135 mM NaCl, 5.4 mM KCl, 0.9 mM MgCl₂, 0.18 mM CaCl₂, 0.33 mM NaH₂PO₄, 10 mM HEPES, and 10 mM glucose; pH adjusted to 7.4 with NaOH. The internal solution contained 130 mM KCl or K-glutamate, 25 mM KCl, 1 mM MgCl₂, 5 mM Na₂ATP, 1 mM EGTA, 0.1 mM Na₂GTP, 5 mM HEPES; pH 7.25 (adjusted with KOH). 4-AP (Sigma) was made as a 1.5 M stock solution in distilled water (pH adjusted to 7.4 with HCl) and was added to the external solution to obtain the desired concentration. Solutions containing 4-AP were protected from light. Clotrimazole, miconazole, and glybenclamide were obtained from Sigma. Stock solutions were prepared by diluting an appropriate amount of compound in dimethyl sulfoxide. The dimethyl sulfoxide concentration in the perfusing solutions never exceeded 0.2% (v/v) and caused no effect of its own.

Results

Different Kinetic Components of I_to.

Figure1A shows currents elicited by 5-s voltage steps from the holding potential (−80 mV) to +60 mV, in the absence or in the presence of 50 μM 4-AP. The currents inactivated with time during the depolarizing step. Under control conditions, an initial rapid decrease was followed by a slow decay, and two exponentials were needed to satisfactorily fit the time course of the inactivation. At +60 mV τ_fast was 88.9 ± 6.80 ms and τ_slow was 1022 ± 45.27 ms (n = 47). These rates of decay did not change substantially with voltage >0 mV (data not shown). In the presence of 50 μM 4-AP the outward current was decreased in peak amplitude and decayed faster than in control, but the end-of-pulse current was practically unchanged. The 4-AP-sensitive current, i.e., the difference between the traces in control and in the presence of the drug, is presented in Fig. 1B, and the inset shows that its inactivation could be resolved by one exponential. Its amplitude (15.6 pA/pF) and time constant (716 ms) were of the same magnitude as those of the slow component of the total current (16.6 pA/pF, 931 ms), suggesting that there was a selective suppression of a slow component by 4-AP, with no or marginal effect on the fast component. These results are consistent with the findings of other studies (Fiset et al., 1997; London et al., 1998a,b; Zhou et al., 1998), which indicated a high sensitivity of the slow component of the transient current to 4-AP. For simplicity, we will assume below that the total outward current is made of a fast inactivating component (I_fast), of a slowly inactivating component (I_slow), and a noninactivating component (I_∞). The whole time-dependent current (sum of I_fast and I_slow) will be referred to as I_to.

I_to magnitude decreased when prepulses were given before the depolarization to +60 mV (Fig. 8A). With 5-s prepulses, the availability (or “inactivation”) curve could be fitted by a single Boltzmann distribution curve (see Figs. 5 and 8B). The potential for half-maximum inactivation (V_1/2) was −52 ± 0.7 mV and the slope factor of the inactivation curve was 11.5 ± 0.25 mV (n = 11). However, in experiments carried out at a preliminary stage of this study, in which 1 s or shorter prepulses were used, a sum of two Boltzmann equations (withV_1/2 of −53 and −22 mV) was needed to fit the inactivation curve, presumably as a result of an incomplete inactivation of I_slow. Under such conditions, after application of 25 to 50 μM 4-AP, one single Boltzmann distribution (V_1/2 of −52 mV) satisfactorily fitted the inactivation curve (data not shown), hence supporting the view that low 4-AP concentrations preferentially suppress a slow I_to component.

Preferential Block of I_slow by Miconazole and Clotrimazole.

Figure 2, A and B, show the effect of miconazole (30 and 100 μM) and clotrimazole (30 μM), respectively, on I_to induced by steps to +60 mV. The peak amplitude of I_to was decreased and the current decayed faster in the presence of either drug. The time course of the drug effect is illustrated in Fig. 2C for a cell in which 10 μM miconazole was applied while giving voltage pulses consisting of a 1-s hyperpolarization to −120 mV followed by a 5-s depolarization to +60 mV. Peak current at +60 mV decreased progressively in the presence of miconazole, whereas there was no effect on the end-of-pulse current at +60 mV and only a marginal decrease at −120 mV (see Fig.3Aa). On washout of miconazole the effect could not be reversed. Pooled data from experiments (n= 4) such as those of Fig. 2C are summarized in Fig. 2D, which confirm that 10 μM miconazole decreased I_to (peak current at +60 mV: 64 ± 4.4 pA/pF in control and 43 ± 6.9 pA/pF with drug; P < 0.05) while having no effect on the noninactivating current or on I_K1 (current at −120 mV: −23 ± 2.4 pA/pF in control, and −21 ± 3.5 pA/pF with drug; P > 0.05). However, with higher concentrations and prolonged treatments, decreases in these latter currents could be observed but were not further investigated in the present study.

The miconazole- or clotrimazole-sensitive currents decayed monoexponentially (insets of Fig. 2, A and B, respectively), and their time constants (865 and 1044 ms, respectively) were of the same order of magnitude compared with the time constants of the slow component of the control current (in the same cells: 1125 and 1361 ms, respectively) and with that of the 4-AP-sensitive current (see above; Fig. 1B). This suggests that antimycotics caused an inhibition of I_slow and that the fast-decaying current remaining in the presence of the antimycotics is due to I_fast.

The effect on channels underlying I_slow could develop in the rested state, but similar results can also be obtained if, as a result of a channel block in the open or inactivated state (Carmeliet and Mubagwa, 1998), I_slow is modified so as to decay with a time constant close to and indistinguishable from that of I_fast. We therefore tested whether the block of I_slow is time-dependent during the voltage pulse, by isolating this component from the superimposed I_fast. I_slow can be isolated by depolarizations to potentials greater than −30 mV that are long enough to fully inactivate I_fast but short to only partly inactivate I_slow. Figure 3A compares the effect of miconazole on the current at +60 mV elicited after a prepulse to −90 mV (Fig. 3Aa) or after a 1-s prepulse to −30 mV (Fig. 3Ab). With the prepulse to −30 mV, the fast component was fully inactivated and the transient component induced at +60 mV was essentially due to I_slow. Under these circumstances, the current at +60 mV in the presence of miconazole did not show a fast decay (Fig. 3Ab; in contrast to the current generated following the prepulse to −90 mV, Fig. 3Aa), suggesting the absence of a slowly developing open or inactivated channel block on this component. These data indicate that the antimycotic blocked I_slow by interacting with the underlying channels in the rested state.

To further examine the preferential effect on I_slow, the effect of the antimycotics was tested in the presence of 50 μM 4-AP to suppress this current. Figure 3B illustrates that when clotrimazole was applied on top of 50 μM 4-AP, no effect was obtained on the fast-decaying current besides the further elimination of a small persisting slow component probably incompletely blocked by 4-AP.

To quantitatively assess the extent of drug-induced change in the I_to components, the currents in the absence or in the presence of the antimycotics were fitted by a sum of two exponentials. The magnitude of each component under steady state (after 10–20 min) in the presence of the drug was expressed relative to its magnitude in control conditions. The relative magnitude of the two I_to components in the presence of various miconazole concentrations is plotted in Fig.4. This analysis indicated that the slow component was selectively decreased by miconazole. The half-maximum inhibitory concentration (IC₅₀) for the slow component was 7.0 μM (n_Hill = 0.95). Time constants and amplitudes of the fast and slow components, as well as of the offset (noninactivating component), under control conditions and in the presence of 30 μM miconazole or 30 μM clotrimazole are presented in Table 1. Changes in the magnitude of the fast component were absent or relatively less marked compared with the effect on the slow component. Although a complete concentration-effect curve was not established for clotrimazole, it decreased the slow component by 32% at 30 μM (n = 4).

To further examine the possibility of a block in the inactivated state, we analyzed the effect of the antimycotics on the steady-state inactivation of I_to. Figure5 shows that the inactivation curve was unchanged in the presence of either 30 μM miconazole (Fig. 5A;n = 4) or 30 μM clotrimazole (Fig. 5B;n = 4). For the data with miconazole, the potential of half-maximum inactivation (V_1/2) was −51.9 ± 1.3 and −53.7 ± 1.3 mV (P > 0.05) before and during application of the drug. Similarly, there was no significant shift in V_1/2 in the presence of clotrimazole. (The differences inV_1/2 between miconazole and clotrimazole data are explained by the junction potential due to the different intracellular solution composition: KCl for miconazole experiments and K-glutamate for clotrimazole).

Due to the lack of reversibility of the miconazole effect (Fig.2C), we tested whether the drug was acting on an internal site by adding 30 μM miconazole to the patch pipette solution. The slow component of I_to was not suppressed in three cells dialyzed with the drug in this way, and the cells still responded to extracellularly applied miconazole (data not shown). This result excludes the possibility that the miconazole effect was due to an interaction with some intracellular site from which it was difficult to remove the drug by extracellular washout.

Block of Both I_fast and I_slow by Glybenclamide.

Given the reported acceleration of I_to inactivation by glybenclamide (Schaffer et al., 1999), we examined whether the drug acts in the same way as the antimycotics by preferentially suppressing I_slow. Figure 6A illustrates the effect of glybenclamide (100 μM) on currents induced by a 1-s step to −120 mV followed by a 5-s depolarization to +60 mV. Glybenclamide caused no or marginal changes of the current during the hyperpolarizing pulse, suggesting that it had no or little effect on I_K1(Fig. 6, C and D). In contrast, the drug caused a decrease of the peak amplitude of I_to induced at +60 mV and accelerated its inactivation, but had no effect on the noninactivating current at the end of the pulse. Figure 6B shows the glybenclamide-sensitive current, i.e., the difference current obtained by subtracting the trace in the presence of glybenclamide from the control trace. The inset illustrates that the difference current could not be resolved by one single exponential (τ_fast = 120 ms, τ_slow= 965 ms; compared with τ_fast = 116, τ_slow = 1047, for the control tracing), hence indicating that the glybenclamide effect could not be readily attributed to the selective suppression of a current component that decays monoexponentially. The time course of the glybenclamide effect is illustrated in Fig. 6C. The effect developed rapidly, reached a steady-state within 2 to 4 min with 50 μM glybenclamide, and could be readily reversed on drug washout from the extracellular solution. Figure 6C and pooled data from nine experiments in Fig. 6D, further illustrate that the effect on I_to peak occurred in the absence of significant effect on the current (largely I_K1) at negative potentials or on the noninactivating current at depolarized levels (current at −120 mV: −30 ± 3.0 pA/pF in control, −28 ± 3.1 pA/pF during exposure to the drug; P > 0.05, n = 9).

To assess the extent of drug-induced change in the I_to components, the currents at +60 mV before drug application or in the presence of glybenclamide were fitted by a sum of two exponentials and one constant. The magnitude of the two I_to components in the presence of various glybenclamide concentrations is plotted in Fig.7. The data indicate that glybenclamide concentration dependently decreased both the fast and the slow components of I_to, and that the IC₅₀ values for the effect on both components were of the same magnitude (38–49 μM glybenclamide). Time constants and amplitudes of the fast and slow components, as well as of the offset (noninactivating component), under control conditions and in the presence of 50 μM glybenclamide are also included in Table 1.

We also examined whether glybenclamide interacts with the inactivated channel. Figure 8A shows current traces recorded at +60 mV following 5-s prepulses to various levels, in control conditions, in the presence of 50 μM glybenclamide, and after washout of the drug. In the presence of glybenclamide prepusles to −60 mV and to −50 mV caused a more marked depression of I_to, and this effect was reversible on drug washout. Figure 8B shows that the inactivation curve was shifted in the negative direction in the presence of glybenclamide:V_1/2 of inactivation was −52 ± 1.3 mV in baseline conditions and was shifted to −67 ± 5.1 mV during application of glybenclamide (n = 6). These data indicate that the effect of glybenclamide were due, at least in part, to an enhancement of I_to inactivation.

Effects on Action Potential.

Given the prominent effects of either the antimycotics or glybenclamide on I_to, one expects a slowing of the ventricular repolarization in the presence of these drugs. Action potentials were recorded in nine cells with the patch electrode under the same experimental conditions used for voltage clamp, except that the external solution contained 1.8 mM Ca²⁺. In three cells, the action potential duration at 50% repolarization (15.2 ± 3.71 ms under control conditions at 1-Hz pacing) was prolonged by 55 ± 12.0% in the presence of glybenclamide (100 μM) without change in the resting potential (−81.4 ± 1.60 mV). This effect was completely reversed with 5 to 10 min of drug washout. Miconazole (n = 5) or clotrimazole (n = 1), each at 10 μM, caused a progressive increase in the delay between stimulus and action potential upstroke, a decrease of action potential amplitude (hence precluding a quantitative assessment of its effect on action potential duration), and induced a complete loss of excitability despite maintained resting potential.

Discussion

The presence of distinct components of the inactivating outward current (here simply called I_to) in mouse ventricular myocytes was first reported by Benndorf et al. (1987) andBenndorf and Nilius (1988), who observed two to three channel populations with different conductances and kinetics in cell-attached patches. Recently, more evidence has been presented for the existence of distinct channels at the molecular level. In myocytes of transgenic mice overexpressing a truncated Kv1.1 channel, the expression of Kv1.5 was markedly depressed and the slowly inactivating component of I_to was absent (London et al., 1998a,b). This prompted the authors to suggest that the slow component (which they called I_slow) is encoded by Kv1.5. Barry et al. (1998) found that in cardiomyocytes from Kv4.2W362F mice the fast inactivating I_to (I_fast) was eliminated, suggesting that members from the Kv4 family underlie the fast component. More recently, Xu et al. (1999a,b) reported four different components, and Guo et al. (1999) found that in ventricular myocytes from mice with a targeted deletion of the Kv1.4 gene, I_slow was absent, hence demonstrating that Kv1.4 or a related protein is the molecular correlate of I_slow.

In the present study we have corroborated the presence of two I_to components. The decay of I_to could be resolved into two exponentials with time constants differing by one order of magnitude as also found by previous studies (Benndorf et al., 1987; Wang and Duff, 1997; London et al., 1998a,b; Zhou et al., 1998). The two components were present in all our cells, although Xu et al. (1999a,b) found no fast component in 10% of myocytes randomly dispersed from left and right ventricles.

In our study, V_1/2 of the inactivation curve was about −50 mV. Differences in experimental conditions (concentration of divalent cations, presence of junction potential) among the various studies make a comparison ofV_1/2 values obtained by different groups difficult. In the study of Zhou et al. (1998)V_1/2 for the slow component was −35 mV but they included 1 mM Ca²⁺, 1 mM Mg²⁺, and 2 mM Co²⁺ in the extracellular solution. V_1/2 for a fast component was −24 mV in the studies of Xu et al. (1999a,b), while for the slow component (I_Kslow) they found a biphasic inactivation curve with V_1/2values of −73 and −19 mV in the presence of 1 mM Ca²⁺, 2 mM Mg²⁺, and 5 mM Co²⁺. In contrast, aV_1/2 of −66 mV has been reported in day 1 neonatal mouse, in which only the fast component is present (Wang and Duff, 1997). Thus, even after correction for divalent cation effects (Stengl et al., 1998) and for junction potentials major differences remain between different studies.

Channel-selective drugs are a useful tool for separating various channel populations. As also found by others (Fiset et al., 1997;London et al., 1998a,b; Zhou et al., 1998; Xu et al., 1999b), low 4-AP concentrations in our study blocked the slow component (IC₅₀ of ≈ 3 μM) while leaving intact a fast component. Millimolar concentrations were needed to block this fast component (data not shown). This differential sensitivity to 4-AP between fast and slowly inactivating components is an additional argument to support the view that I_to is composed of at least two distinct channels.

Based on the finding that ketoconazole blocks Kv1.5 expressed inX. laevis oocytes (Dumaine et al., 1998), we expected an effect of two ketoconazole-related fungicides, miconazole or clotrimazole, on native I_to of mouse myocardial cells. The magnitude of I_to in the presence of either drug was decreased and the inactivation was apparently accelerated. It is shown that the effect of the antimycotics can be attributed to a selective suppression of a slow component of I_to (I_slow) as is seen in the presence of micromolar 4-AP concentrations. The IC₅₀ (7 μM) for I_slowinhibition by miconazole is nearly one order of magnitude lower than for the ketoconazole inhibition of Kv1.5 expressed in oocytes (105 μM). In this study we did not use ketoconazole. If similar potencies are assumed for the antimycotics, the observed disparity in IC₅₀ values may be accounted for by a genuine difference between native I_slow and Kv1.5, or by an influence on drug sensitivity of the expression system (oocyte for Kv1.5) used. This effect of the antimycotics is important not only from a theoretical point of view but also from a clinical standpoint.

The slowly inactivating current accounts for about 35 to 40% of the peak total current, and for more than 50% of I_to(Table 1). Its kinetic properties and high sensitivity to 4-AP, miconazole, and clotrimazole support recent views that this current is due to Kv1.5 or a related channel. Since intracellularly applied miconazole was without effect, the drug does not act via an intracellular mechanism (e.g., an inhibition of cytochrome P450-dependent processes) but probably interacts directly with the channel at a site accessible from the extracellular side of the membrane. The mechanism of block was not clarified in the present study. However, it is unlikely that the drugs cause a fast open/inactivated channel block (Carmeliet and Mubagwa, 1998) since no time-dependent change was obtained when the slow current component was isolated after full inactivation of the fast component. Similarly, the absence of an effect on steady-state inactivation suggests that there is no preferential action on the inactivated channels. Thus, our data suggest an effect on the rested state. The noted decrease of the fast time constant in the presence of the drugs (Table 1) could indicate some effect of the antimycotics on I_fast. Such an effect cannot be completely excluded, especially with prolonged drug application and at high drug concentrations. However, a substantial overestimation of τ_fast can be caused by limitations in accurately determining this parameter in the presence of a substantial slow component by the least-square algorithm used.

Imidazole antimycotics such as ketoconazole, clotrimazole, and miconazole have been associated with QT interval prolongation and torsades de pointes (Monahan et al., 1990). A few reports have attributed miconazole-induced cardiac arrhythmias to a rapid intravenous administration with insufficient dilution (Huijgens et al., 1975; Fainstein and Bodey, 1980; Coley and Crain, 1997). The mechanism underlying the antimycotic-induced rhythm changes is unknown. Clotrimazole has been shown to inhibit L-type calcium current in guinea pig ventricular myocytes (Thomas et al., 1999), but this effect is unable to account for the QT prolongation. A short report suggested that ketoconazole also has a direct blocking effect on the delayed rectifier and transient outward currents in feline ventricular myocytes (Chen and Woosley, 1993), but these effects on K⁺currents have not been examined in detail. Our study indicates that I_slow is a potential target to explain the toxic effects of the drugs. Although under our experimental conditions the antimycotics readily suppressed action potentials, indicating that the effects on other (Na⁺, Ca²⁺) channels may play a more critical role in the arrhythmogenesis, in conditions where excitability is not fully suppressed, the I_to suppression by antimycotics is likely to contribute to the arrhythmogenesis by causing action potential prolongation and increasing action potential dispersion in the ventricle. An effect on human ether-a-go-go-related gene-related native currents (Dumaine et al., 1998) could constitute an additional factor, but it was not examined in the present study.

As for glybenclamide, its effects on outward currents have been reported in human cardiomyocytes (Schaffer et al., 1999). In atrial cells, peak amplitude was decreased and inactivation was accelerated, as also found in our study on mouse ventricular cells. These results were interpreted as caused by a suppression of I_to,1 and I_Kur (Schaffer et al., 1999). In contrast to the decrease of the end-of-pulse current (interpreted as an effect on I_ss) in atrial cells, we did not find such an effect in mouse ventricular cells. However, since short (0.3-s) pulses were used in the atrial study, it is possible that the end-of-pulse current contained noninactivated I_slow or I_Kur. The noted acceleration of I_to inactivation could result from a selective suppression of a slow component with no effect on the fast component. This seems not to be the case for glybenclamide, for which the exponential analysis indicates an effect on both slow and fast components. In the present study, we show that glybenclamide shifts the inactivation curve of I_to in the negative direction, hence suggesting that it interacts with channels in the inactivated state.

It is now established that glybenclamide acts on many channels, and that its selectivity for I_K-ATP channels is only obtained at submicromolar concentrations (Schaffer et al., 1999). Since the IC₅₀ for the effect on I_to is about 50 μM, plasma concentrations usually reached under well controlled therapy are unlikely to cause cardiac complications due to an effect on I_to. Many ion channels (swelling-activated, cAMP-activated, Ca²⁺-activated, etc.; Schaffer et al., 1999) on which glybenclamide acts at high concentrations do not play a role in the basal cardiac electrical activity. In contrast, I_to participates to the normal repolarization of the cardiac action potential, hence the potential effect of glybenclamide on this current requires that higher concentrations be avoided both under conditions where a selective effect on I_K-TP is desired.