Interaction of Riluzole with the Closed Inactivated State of Kv4.3 Channels

Voltage-gated K⁺ (Kv) channels are closed in the resting state, but they are opened upon membrane depolarization and then inactivated for a variety of time courses. These channels are the primary modulators of neuronal excitability by altering the repolarization of action potentials and neuronal firing frequencies. In this respect, the inactivation mechanism involves the fine-tuning of Kv channel activity and plays an important role in the modulation of the temporal pattern of the activity of excitable neurons. The two electrophysiologically different types of Kv channels can be distinguished based on inactivation kinetics: delayed rectifier and A-type K⁺ channels. Kv4.3, one of the major A-type currents, is rapidly activated and inactivated upon depolarization and is expressed at high levels in brain, heart, and smooth muscle (Ohya et al., 1997). This channel is activated at a subthreshold membrane potential, therefore modulating the threshold for the spike generation of neurons (Yuan et al., 2002). Thus, this channel may serve as an important target of certain drugs that are involved in the regulation of neuronal activity.

Riluzole (2-amino-6-trifluoromethoxy benzothiazole) is a neuroprotective agent and has been shown to be effective in the treatment of convulsion in humans (Bryson et al., 1996; He et al., 2002). The mechanisms of its therapeutic actions have been attributed to the interaction with the inactivated state of voltage-gated Na⁺ and Ca²⁺ channels (Benoit and Escande, 1991; Huang et al., 1997; Zona et al., 1998; Urbani and Belluzzi, 2000), and its blocking properties make it a potential candidate used as an anticonvulsant drugs (Romettino et al., 1991; Stutzmann et al., 1991). The actions of riluzole on these channels have been extensively investigated, but relatively little is known concerning its effects on Kv currents, especially on the A-type current. We and others have previously reported that riluzole interacts preferentially with the inactivated state of Kv1.5, cloned delayed rectifier Kv channels in CHO cells and outward K⁺ currents in cultured rat cortical neurons (Zona et al., 1998; Ahn et al., 2005). However, riluzole also modulates several types of K⁺ channels by alternate mechanisms (Duprat et al., 2000; Grunnet et al., 2001). For example, riluzole induces the slowing of the inactivation by oxidation of a cysteine in the N-terminal inactivation domain of Kv1.4, another A-type current (Xu et al., 2001). Because Kv channels display a variety of time courses and kinetic properties in the inactivation process, an understanding of the modulation of channel inactivation by drugs has been a focus of intense interest. The purpose of the present study was to characterize the action of riluzole on the inactivation kinetics of Kv4.3 and to compare its mechanisms of action with those for Kv1.4.

Materials and Methods

Cell Cultures. Rat Kv4.3 cDNA was kindly provided by Dr. Y. Imaizumi (Nagoya City University, Nagoya, Japan). The Kv4.3 cDNA was subcloned into a mammalian expression vector, pCR3.1 (Invitrogen, Grand Island, NY) as described previously (Ohya et al., 1997). CHO cells (American Type Culture Collection, Manassas, VA) were maintained in Iscove's modified Dulbecco's medium (IMDM; Invitrogen) supplemented with 10% fetal bovine serum, 0.1 mM hypoxanthine, and 0.01 mM thymidine in a humidified 5% CO₂ incubator at 37°C. Kv4.3 constructs were stably transfected into CHO cells using the Lipofectamine reagent (Invitrogen). The transfected CHO cells in IMDM containing 1 mg/ml geneticin (Invitrogen) were incubated for 48 h under a humidified 5% CO₂ incubator at 37°C. The selected cells were cultured in 96-well plates by serial dilution with IMDM containing 0.5 mg/ml geneticin, and stable transfectants were selected by a prolonged culture period. The transfected cells were exchanged with fresh IMDM containing 0.3 mg/ml geneticin and passed at 2- to 3-day intervals using a brief trypsin-EDTA treatment. The cells were seeded onto glass coverslips (12 mm in diameter; Fisher Scientific, Pittsburgh, PA) in a 35-mm dish 1 day before use. For the electrophysiological experiments, coverslips with attached cells were transferred to a continually perfused recording chamber (RC-13; Warner Instruments, Hamden, CT).

Electrophysiological Recordings. The patch clamp was used in the whole-cell configuration at room temperature (22-24°C) using an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA). Patch pipettes were pulled from glass capillary tubing (PG10165-4; WPI, Sarasota, FL) using a puller (model P-97; Sutter Instrument Company, Novato, CA). The tip resistances of the recording pipettes in the bath solution were 2 to 3 MΩ. Series resistances were approximately 4 to 8 MΩ. Whole-cell capacitive currents were compensated with analog compensation. Series resistance compensation (80%) was used if the current exceeded 1 nA. The currents were low-pass filtered at 2 kHz (four-pole Bessel filter) and sampled at 5 kHz before being digitized. Data acquisition and analysis were performed with an IBM Pentium computer, using the pClamp 9.01 software (Axon Instruments).

Solutions and Drugs. The bath solution contained 140 mM NaCl, 5 mM KCl, 1 mM CaCl₂, 1 mM MgCl₂, and 10 mM HEPES and was adjusted to pH 7.3 with NaOH. The internal pipette solution contained 140 mM KCl, 1 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, and 10 mM EGTA and was adjusted to pH 7.3 with KOH. The measured osmolarity of the solutions was 300 to 340 mOsm. Riluzole (Tocris Cookson Inc., Bristol, UK) was dissolved in dimethyl sulfoxide to make a 100 mM stock solution and then appropriately diluted with bath solution.

Data Analysis. The Origin 7.0 software program (OriginLab Corp., Northampton, MA) was used for the analysis. The concentration-response data were fitted to the following logistic equation: y = 1/[1 + ([D]/IC₅₀)ⁿ], where IC₅₀ is the concentration of riluzole required to produce 50% inhibition, [D] is the concentration of riluzole, and n is the Hill coefficient. Kv4.3 currents were elicited by applying 500-ms depolarizing pulses from a holding potential of -80 mV to +40 mV at 10-s intervals. The voltage dependence for steady-state inactivation was investigated using a two-pulse voltage protocol; currents were measured by 500-ms depolarizing pulses to +40 mV, whereas 1-s preconditioning pulses were varied from -110 to +0 mV stepped by 5 mV at 10-s intervals in the absence and presence of the drug. The resulting steady-state inactivation data were fitted to the Boltzmann equation: (I - I_c)/(I_max - I_c) = 1/[1 + exp(V - V_1/2)/k] in which I_max represents the current measured at the most hyperpolarized preconditioning pulse, I_c is a nonzero current that was not inactivated at the most depolarized preconditioning pulse, k is the slope factor, V is the test potential, and V_1/2 is the potential at which the conductance was half-maximal. We eliminated the nonzero residual current by subtracting it from the actual value.

The data are expressed as the mean ± S.E. One-way analysis of variance, followed by Dunnett's test, was used to evaluate the statistical significance of the observed differences. Statistical significance was considered at p < 0.05.

Results

Concentration Dependence of Kv4.3 Inhibition.Figure 1A shows representative recordings of Kv4.3 currents elicited by 500-ms depolarizing pulses to +40 mV under control conditions and in the presence of different concentrations of riluzole. Under control conditions, Kv4.3 began to become activated at a potential of -30 mV and was fully activated at +40 mV and then became completely inactivated during the application of depolarizing pulses, as reported previously (Ohya et al., 1997). Thus, a depolarization pulse of +40 mV was used to test the effect of riluzole. Riluzole significantly decreased the peak amplitude of Kv4.3 in a concentration-dependent manner. The concentration-response curve for riluzole on Kv4.3 at +40 mV is shown in Fig. 1B. The normalized inhibition was plotted as a function of drug concentration and was fit to a Hill equation, giving an IC₅₀ of 115.6 ± 7.6 μM and a Hill coefficient of 2.4 ± 0.3 (n = 6). In the present study, riluzole decreased the peak amplitude of Kv4.3 in a concentration-dependent manner, but its most obvious effect seemed to be a decrease in the rate for current decay, resulting in the crossover of the current traces obtained under control conditions and after the addition of riluzole (Fig. 2A). To study the effect of riluzole on the inactivation kinetics of Kv4.3, the time constants for the inactivation were measured. The time course of inactivation of Kv4.3 at +40 mV under control conditions was fitted to a double exponential function, with a fast time constant (τ_f) of 14.8 ± 0.3 ms and a slow time constant (τ_s) of 107.6 ± 5.4 ms (n = 6). Riluzole had no effect on τ_f of Kv4.3, but it caused a significant decrease in the rate of slow inactivation, with a τ_s value of 113.0 ± 6.5, 130.9 ± 10.5, and 145.5 ± 9.3 ms for 10, 30, and 100 μM (n = 6; p < 0.05), respectively (Fig. 2B). Therefore, riluzole increased the slow component of inactivation from the control value of 10.3 ± 0.7% to 12.0 ± 0.8, 14.6 ± 1.1, and 20.6 ± 0.7% (n = 6) for 10, 30, and 100 μM riluzole, respectively. The activation time constants were calculated by fitting a single exponential to the final 50% of activation (Snyders et al., 1993). Under control conditions, the time constant of activation was 0.51 ± 0.06 ms (n = 6). In the presence of riluzole, the time constants of activation were 0.49 ± 0.05 and 0.46 ± 0.06 ms (n = 6) for 10 and 30 μM, respectively, indicating that the kinetics of activation were not significantly affected by riluzole.

Voltage Dependence of the Inhibition by Riluzole.Figure 3A displays current-voltage relationships for Kv4.3 under control conditions and after the application of 100 μM riluzole. Riluzole reduced Kv4.3 at all test potentials where Kv4.3 was activated (Fig. 3B). However, the inhibition of Kv4.3 was larger at more depolarized potentials. To quantify the voltage dependence of Kv4.3 inhibition, the relative currents before and after the application of riluzole were plotted as a function of test potential (Fig. 3C). In the presence of riluzole, the inhibition increased steeply between -20 and +20 mV. At a test potential of -20 mV, riluzole inhibited the Kv4.3 current by 5.7 ± 9.5% (n = 5). This inhibition increased to 39.5 ± 4.7% at +20 mV (n = 5; p < 0.05). Between +30 and +60 mV, where the maximal conductance is reached, however, the degree of inhibition was uniform.

Shift of the Steady-State Inactivation Curves. To assess the effect of riluzole on the voltage dependence of channel availability, steady-state inactivation curves were generated using standard double-pulse protocols (Fig. 4A). The data points were fitted to a Boltzmann function to estimate V_1/2 and k. The steady-state inactivation curves for Kv4.3 in the control had V_1/2 values of -50.7 ± 1.6 mV and k of 4.6 ± 0.3 mV (n = 8). Riluzole significantly shifted the inactivation curve (V_1/2) to a hyperpolarized potential in a concentration-dependent manner (-63.3 ± 2.6 mV at 30 μM and -72.3 ± 2.4 mV at 100 μM; n = 8; p < 0.05). However, no significant change in k for the curve in the presence of riluzole was found (4.8 ± 0.3 mV at 30 μM and 4.9 ± 0.2 mV at 100 μM; n = 8). Although the apparent dissociation constant, K_R, for the riluzole-induced inhibition of Kv4.3 in the closed state can be estimated from the reduction in the peak amplitude of the Kv4.3 current (Fig. 1), the apparent dissociation constant, K_i, for the riluzole-induced inhibition in the inactivated state can be estimated from the concentration-dependent shift in the steady-state inactivation curve (Bean et al., 1983). The theoretical value of K_i was calculated to be 1.2 ± 0.5 μM (n = 8) (Fig. 4B). Thus, the affinity of riluzole for the inactivated state of Kv4.3 is more than 90 times that of the closed state.

Interaction with the Closed Inactivated State. Because previous studies have shown that Kv4.3 channels are predominantly inactivated from the closed state (Beck and Covarrubias, 2001; Wang et al., 2005), we examined the possible effect of riluzole on the kinetics of closed state inactivation (Fig. 5A). We used subthreshold depolarizing pulses (-60 mV), because the inactivation occurs predominantly from the close state at this holding potential. Under control conditions, a 12-s conditioning pulse to -60 mV inactivated 53.2 ± 4.3% (n = 8) of the Kv4.3 channels (Fig. 5B). The time course for closed state inactivation was fitted to a single exponential function with a time constant of 4.2 ± 0.5 s (n = 8). In the presence of 100 μM riluzole, the rate of closed stated inactivation was accelerated and inactivation was almost complete after a 1-s conditioning pulse. Closed state inactivation was also fitted to a single exponential function with a time constant of 0.2 ± 0.03 s (n = 8; p < 0.05).

Use-Dependent Inhibition. The use dependence of riluzole action was evaluated with 10 consecutive pulses to +40 mV at 1 and 2 Hz. Under control conditions (Fig. 6A), the Kv4.3 currents displayed a slight decline at 1 Hz (6.2 ± 2.4%; n = 6). This percentage increased to 30.2 ± 4.5% (n = 6) at a frequency of 2 Hz. In the presence of 100 μM riluzole, the amplitude of the peak current declined more rapidly with successive pulses, reaching a steady-state level equivalent to 39.4 ± 4.4% of the control after five pulses at 1 Hz. The degree of frequency-dependent inhibition increased more markedly at 2 Hz. Thus, the potency of riluzole seems to be enhanced by increasing the frequency of the depolarizing stimulation.

Recovery from Inactivation of Kv4.3. The effects of riluzole on the kinetics of Kv4.3 recovery from steady-state inactivation are shown in Fig. 7A. Under control conditions, the recovery from the inactivation of Kv4.3 was complete after an interpulse interval of 1 s and was best fit to a single exponential function with a time constant of 258.4 ± 25.6 ms (n = 6) (Fig. 7B). In the presence of 100 μM riluzole, however, the recovery process was best fit to a biexponential function with a fast time constant of 379.4 ± 51.7 ms and a slow time constant of 2001.6 ± 164.9 ms (n = 6). Both time constants differed significantly from the time constant under control conditions (n = 6; p < 0.05), indicating the slower recovery kinetics of Kv4.3 currents.

Reversibility of the Effects of Riluzole.Figure 8A shows the results of a typical experiment in which a single pulse to +40 mV was repeated while 100 μM riluzole, separated by washout periods, was applied. Riluzole caused a decrease in the amplitude of Kv4.3 by approximately 50% within 1 min. The repeated application of riluzole caused the rapid inhibition of Kv4.3, but the peak amplitudes of the current were completely restored to the control values upon washout (Fig. 8B). Riluzole slowed the inactivation kinetics of Kv4.3 slightly (τ_f of 16.2 ± 1.7 ms and τ_s of 116.2 ± 8.6 ms; n = 4) after a washout, but the values were not statistically different from the control values (Fig. 2B), indicating that the effects of riluzole were completely reversible upon washout.

Discussion

The findings herein show that riluzole is a potent inhibitor of the Kv4.3 current, which is stably expressed in CHO cells. The inhibition is characterized by a reduction in peak amplitude of Kv4.3 and a concentration-dependent shift in steady-state inactivation curves in the hyperpolarizing direction. These results suggest that the inhibition of Kv4.3 by riluzole is mainly caused by the binding of riluzole to the inactivated state of the channel.

Riluzole is a neuroprotective agent and its mechanism of action may involve, in part, the blocking of voltage-gated Na⁺ channels (Taylor and Meldrum, 1995; Doble, 1996). Riluzole also modulates several types of voltage-gated K⁺ and Ca²⁺ channels (Huang et al., 1997; Wu and Li, 1999; Duprat et al., 2000; Grunnet et al., 2001; Cao et al., 2002). The most important aspect of the mechanism of blocking voltage-gated ion channels by riluzole is its high affinity for the inactivated state, compared with the closed or open state of the channels (Benoit and Escande, 1991; Huang et al., 1997; Zona et al., 1998; Ahn et al., 2005). In the present study, the following observations suggest that riluzole inhibits Kv4.3 through a preferential interaction with the inactivated state of the channel. First, the voltage dependence of the steady-state inactivation curves was shifted in the hyperpolarizing direction in a concentration-dependent manner. Second, the inhibitory action of riluzole was use-dependent. The reshaping of the recovery kinetics with a slower component may result from the slower dissociation of this drug from its binding site of Kv4.3, which may explain the use-dependent inhibition by riluzole. Third, riluzole produced a voltage-dependent inhibition of Kv4.3 over the voltage range corresponding to channel opening. This voltage dependence of riluzole inhibition seems to be due to the voltage dependence of inactivation of Kv4.3 currents. Therefore, our observations confirm that the mechanism by which riluzole inhibits Kv4.3 is similar to those observed for several other delayed rectifier K⁺ channels (Zona et al., 1998; Ahn et al., 2005). However, other molecular mechanisms of action have been reported for riluzole. For example, riluzole slows the inactivation of Kv1.4 channels via the oxidation of a cysteine residue in the N-terminal inactivation ball (Xu et al., 2001). Channel inactivation by membrane depolarization prevents the riluzole-induced oxidation of a cysteine in the N-terminal domain of Kv1.4 channels. In the present study, however, the inhibitory effects of riluzole were more pronounced when Kv4.3 channels were inactivated. It is known that Kv1.4 and Kv4.3 are typical A-type currents, which are rapidly activated and inactivated during depolarization. These types of channels are expressed primarily in pre- and postsynaptic membranes in central and peripheral neurons (Sheng et al., 1992). Whereas Kv1.4 channels undergo an N-type inactivation (Lee et al., 1996), Kv4.3 channels lack the ball moiety that causes an N-type inactivation, do not follow typical N-type inactivation, but include the putative concerted action of the cytoplasmic N- and C-terminal regions (Jerng and Covarrubias, 1997). In addition, although the oxidation of cysteine residues in N-terminal domain affects the inactivation kinetics of Kv1.4 (Ruppersberg et al., 1991b; Stephens et al., 1996), the modulation of inactivation by an oxidation/reduction mechanism has not been reported for Kv4.3 channels. More importantly, whereas the inactivation of Kv1.4 channels occurs only from the open state (Demo and Yellen, 1991), Kv4.3 channels are inactivated from both the open and closed states and can enter the inactivated state directly from closed states without activation (Beck and Covarrubias, 2001; Wang et al., 2005). Thus, the previous observations are extended and clarified by demonstrating that riluzole binds preferentially to the closed inactivated states.

Another characteristic of the riluzole-induced inhibition of Kv4.3 was that the peak amplitude of the Kv4.3 currents was reduced and the subsequent inactivation of the current was slowed, resulting in the crossover of the current traces obtained under control conditions and after the addition of riluzole. The instantaneous inhibition at the onset of a depolarizing pulse could be attributed to an interaction of riluzole with the closed state of the Kv4.3 channel. Although the fast inactivation time course for Kv4.3 was not altered by riluzole, the slow time course was slowed substantially. One possible explanation is that the slowing of inactivation could also due to delayed activation and inactivation. These effects have been described in A-type current I_to with a similar mode (Campbell et al., 1993). In ferret ventricular myocytes, 4-AP not only reduced the peak amplitude of I_to but also slowed the apparent rate of both the activation and inactivation processes, resulting in the crossover of the current. These results suggest that 4-AP inhibits I_to through a closed state blocking mechanism. Because the activation kinetics was not affected by riluzole in the present study, the rate at which channels are delivered to the open state from the closed state was slowed due to delayed drug dissociation. Thus, riluzole must unbind from the closed channel during the late phase of the depolarizing pulse, and the channel enters into the open state after the dissociation of the drug and inactivates at a normal rate, which led to a crossover phenomenon. Similar results have been reported for 4-AP on Kv4.2 (Kirsch et al., 1986) and (-)-epigallocatechin-3-gallate on Kv1.5 (Choi et al., 2001): the closed state inhibition can be relieved during membrane depolarization due to the unbinding of the drug from the channel at depolarized potentials. Because Kv4.3 channels can enter the inactivated state directly from the closed state, bypassing the open state during subthreshold depolarization (Beck and Covarrubias, 2001; Wang et al., 2005), riluzole interacts with both the closed and closed inactivated states of the channel. The binding of riluzole to the closed and closed inactivated states, which contribute to a reduction in the total channel availability for the activation of Kv4.3, caused a decrease in the peak amplitude of the Kv4.3 currents upon depolarization. Thus, a reduction in the peak Kv4.3 current reflects the sum of the inhibition of the closed and closed inactivated states. Although reopening from inactivation during recovery has been reported for Kv1.4 (Ruppersberg et al., 1991a), the inactivated state of Kv4.3 is generally thought to be an absorbing state and recovers by directly bypassing the open state (Bahring et al., 2001; Beck and Covarrubias, 2001). After riluzole dissociates from the closed or closed inactivated states during depolarization, Kv4.3 channels can open mainly from the closed state, but not from the closed inactivated state without deactivation at the hyperpolarized membrane potential, as is the case for Na⁺ channels (Kuo and Bean, 1994). Thus, these results suggest that the mechanism of slowing of the inactivation kinetics by riluzole involves a reversal of closed channel inhibition during membrane depolarization.

Based on the shift in steady-state inactivation curves, we estimated the K_i for binding to inactivated channels to be 1.2 μM. Thus, the affinity of riluzole for the inactivated state seems to be more than 90 times higher than that for the closed state. At the recommended daily dose, the average plasma levels of riluzole are in the range between 1 and 2 μM (Le Liboux et al., 1997). The K_i of 1.2 μM for the inactivated state of Kv4.3 is comparable to the previously reported value for blocking the inactivated Na⁺ channels in rat cortical neurons (Urbani and Belluzzi, 2000), and this concentration can easily be reached in brain tissue. A-type K⁺ channels are expressed in a variety of neurons throughout the central nervous system. In general, this type of current seems to play a prominent role in the regulation of the duration and frequency of the action potential and neuronal firing pattern. Kv4.3 is primarily expressed in the soma and dendrites of neurons (Serodio and Rudy, 1998) and is activated at subthreshold membrane potentials, thus limiting the back-propagation of the action potential into dendrites and regulating membrane excitability in the hippocampus (Yuan et al., 2002). For example, the inhibition of dendritic A-type channels can modulate the action potential threshold and increase dendritic excitability and the back-propagation of the action potential in the dendrites of hippocampal pyramidal neurons (Hoffman and Johnston, 1998). Therefore, effect of riluzole on Kv4.3 channels observed in the present study could affect the neuronal excitability, which would be adverse to the neuroprotective action of the drug. However, correlating experimental data with clinical implications must proceed with great caution, because riluzole is a nonselective ion channel inhibitor and affects several other ionic currents that are responsible for the configuration of the action potentials. The overall effects of riluzole on neuronal excitability are determined by the sum of the complex interactions of the drug with multiple ion channels through different neuronal membranes. Indeed, riluzole decreases current-evoked firing discharge in striatal spiny (Centonze et al., 1998) and cortical neurons (Urbani and Belluzzi, 2000), consistent with the use of this drug as a neuroprotective agent. It should also be noted that the pore-forming α-subunit of the Kv4.3 channel protein by itself can form a voltage-gated K⁺ channel in heterologous expression systems, but Kv4.3 coassembles with a large variety of auxiliary subunits, including K⁺ channel interacting proteins in native cells. The interaction of these proteins alters the biophysical properties and drug sensitivity of Kv4.3 channels (Beck et al., 2002; Bett et al., 2006). Thus, the Kv4.3 currents expressed in CHO cells were kinetically and pharmacologically different from the native A-type current. Further studies are clearly needed to clarify the mechanisms underlying the modulation of riluzole's sensitivity of Kv4.3 by K⁺ channel-interacting proteins.

In conclusion, the results reported herein are consistent with general mechanisms in which riluzole binds to the inactivated state of Kv4.3 channels and thus provides further information on the mechanism of action of this drug, in that it interacts preferentially with the closed inactivated state of this type channel.