Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The congenital long QT syndrome (LQT) is an inherited cardiac disorder that causes ventricular arrhythmias, resulting in syncope and sudden death. One form of the disease, LQT3, is caused by mutations in SCN5A, the gene that encodes the voltage-dependent cardiac Na⁺ channel  subunit in humans (hH1) (Jiang, 1994; George, 1995; Wang et al., 1995a,b). A KPQ deletion (KPQ: lysine, proline, and glutamine at positions 1505-1507) in the linker between domains III and IV of hH1 is the most common mutation associated with LQT3. The KPQ mutant channel exhibits late channel openings caused by a defect in inactivation (Bennett et al., 1995; An et al., 1996; Dumaine et al., 1996; Wang et al., 1996c), and the resulting late Na⁺ current (I_Na) would prolong the action potential and cause QT prolongation on the surface ECG.

Gene-specific therapy for LQT3 (i.e., the use of drugs that target the Na⁺ channel, more specifically, late I_Na) is a logical approach. Most antiarrhythmic drugs block the Na⁺ channel in a use-dependent manner by preferential binding to either the inactivated state or the open state as described in the modulated receptor model (Hille, 1977; Hondeghem and Katzung, 1977). Inactivated state blockers of the Class Ib antiarrhythmic grouping (e.g., mexiletine) inhibit late I_Na at the cellular level (Wang et al., 1997), shorten action potential duration in a cellular model of LQT3 (Priori et al., 1996; Shimizu and Antzelevitch, 1997), and shorten the QT interval in LQT3 patients (Schwartz et al., 1995). We hypothesized that an open channel blocker would also be effective (perhaps more so) because of the prolonged dwell time of LQT3 channels in the open state. To test this hypothesis, we studied the Class Ic antiarrhythmic drug flecainide, a predominant open state blocker, comparing tonic block and use-dependent block (UDB) by flecainide for peak and late I_Na for the wild-type (WT) human cardiac Na⁺ channel and the KPQ mutant. Our findings may provide a molecular mechanism for the recently observed correction of the QTc interval in the electrocardiograms of KPQ LQT3 patients by flecainide (Windle et al., 1999).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Clones and Construction of KPQ Mutation. The human heart Na⁺ channel clone we used (hH1a) was kindly provided by Dr. H. Hartmann (Baylor College of Medicine, Houston, TX). The nucleotide and amino acid numbering follow Hartmann et al. (1994). The KPQ mutation was kindly provided by Drs. John W. Kyle and Gayle S. Tonkovich (University of Chicago, Chicago, IL). It was made by polymerase chain reaction techniques as described previously (Nagatomo et al., 1998). The entire polymerase chain reaction-generated region was completely sequenced to confirm the deletion and ensure that no unwanted changes were made in the channel.

Cell Preparation and Transfection. Cells from transformed human embryonic kidney cell line 293 were used. Approximately 5 × 10⁵ cells were seeded on a 60-mm diameter plate (Falcon 3001) with 3 ml of culture medium 1 day before the transfection. Culture medium was MEM complete medium containing: minimum essential medium (Eagle's salts and L-glutamine), 10% fetal bovine serum, 2 mM L-glutamine, 0.1 mM MEM nonessential amino acids solution, 1 mM MEM pyruvate solution, 10,000 U of penicillin and 10,000 µg of streptomycin. Transfection was carried out using a cationic liposome method. Details have been described previously (Nagatomo et al., 1998).

Electrophysiological Recordings. I_Na was recorded using the whole-cell patch clamp method. The bath solution contained 140 mM NaCl, 4 mM KCl, 1.8 mM CaCl₂, 0.75 mM MgCl₂, and 5 mM HEPES, pH 7.4, set with NaOH. The pipette solution contained 120 mM CsF, 20 mM CsCl₂, 5 mM EGTA, and 5 mM HEPES, pH 7.4, set with CsOH. Electrodes were made from borosilicate glass with a puller (P-87; Sutter Instrument, Novato, CA) and heat-polished with a microforge (MF-83; Narishige, Tokyo, Japan). They had a final resistance of less than 1.2 M in the whole-cell recording when filled with the pipette solution. Cells for study were placed in a Plexiglas chamber with continuously flowing bath solution mounted on an inverted microscope (Nikon, Tokyo, Japan) in a Faraday cage. Membrane currents were recorded with an Axopatch 200 amplifier (Axon Instruments Inc., Burlingame, CA) and data were acquired using pClamp v6.03. Data were digitized at 100 kHz and low pass filtered at 10 kHz.

Data Analysis. Leak subtraction of peak and late currents was calculated by extrapolating holding currents at subthreshold potentials (100 mV) to the potential of interest (Makielski et al., 1987). Data were fit to model equations using nonlinear regression (pClamp v6.03 or SigmaPlot 3.0). Parameter estimates are reported ± S.E. of the parameter estimate; statistical differences in parameters were determined by a t statistic. Goodness of fit was judged both visually and by the sum of squared errors. The number of exponential components that best fit the data were determined by an F-ratio test (P < .05) to account for the increased number of free parameters. Experiments at different concentrations of flecainide were done in separate cells at the same time after patch rupture. Summary data are expressed as means ± S.E. n represents the number of cells studied. In figures, the symbols represent mean data and the error bars are shown only when they exceed the size of the symbol.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Tonic Block for Peak and Late I_Na is Greater for KPQ. Tonic block is defined as the decrease in I_Na in the presence of drug and in the absence of previous depolarizations or with a frequency of depolarizations sufficiently slow that no UDB developed. I_Na in response to a 24-ms depolarization to 20 mV from a holding potential of 150 mV was measured in control solutions and 5 to 7 min after exposure to various concentrations of flecainide (see example records in Fig. 1, A and B). Tonic block of peak I_Na was assessed by dividing the peak I_Na for the first depolarization in flecainide by the control peak I_Na. The concentration dependence of block (Fig. 1, C and D) was then fit with a single site-binding equation. The half concentration (IC₅₀) for block of peak I_Na was 127 ± 5.6 µM for WT (n = 20) and 80 ± 9.0 µM for KPQ (n = 31), a significant difference (P < .001). To assess the tonic block of late current, the average value of I_Na between 20 ms and 23.5 ms after the depolarization was measured in control and flecainide (see Fig. 1, A and B, for examples) and analyzed in the same way as for peak I_Na (Fig. 1, C and D). The IC₅₀ value for block of late I_Na was 44 ± 1.6 µM for WT (n = 20) and 19 ± 1.6 µM for KPQ (n = 31), a significant difference (P < .001).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Tonic block for peak and late INa. A and B, representative current recordings in control and in the presence of 100 mM flecainide for WT and the KPQ mutant hH1a. INa was elicited by a 24-ms step depolarization from 150 mV to 20 mV. Insets show magnified current traces for comparison of the late component. C and D, concentration dependence of flecainide-induced tonic block of INa. Current was measured at peak () and "late" (; average value between 20 ms and 23.5 ms). The proportion of INa remaining in the presence of flecainide was normalized to INa in control solutions for different concentrations. Data points represent the mean for three to seven measurements; bars represent S.E. Solid lines represent fits to a single-site binding equation: Normalized INa = 1/(1 + ([flecainide]/IC50)) where [flecainide] is the millimolar concentration and IC50 is the half-blocking concentration. The IC50 values for WT were 127 mM and 44 mM at peak and late, respectively, and for KPQ, the IC50 values were 80 mM and 19 mM at peak and late, respectively.

Use-Dependent Block of Peak I_Na Is Greater for KPQ. Use-dependent block of peak I_Na was produced by imposing a pulse train of 200 depolarizations of 24-ms duration from 150 mV to 20 mV at 5Hz. Under control conditions, application of this pulse train produced negligible (<5%) change in peak I_Na (data not shown). Figure 2A shows an example of UDB by 100 µM flecainide for WT and KPQ. Figure 2B shows summary data for the relative amplitude of peak I_Na (compared with peak I_Na in the first pulse) in various concentrations of flecainide plotted versus the pulse number in the train. UDB of peak I_Na was deeper but developed more slowly for KPQ than for WT. The time course of UDB was evaluated by fitting the peak currents to a one- or two-component exponential decay as shown in Fig. 2B with summary data in Table 1.

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Time course of use-dependent block of peak INa in response to pulse trains for WT and KPQ hH1a channels. Pulse trains consisted of depolarizations of 24-ms duration from 150 mV to 20 mV applied at 5 Hz. A, superimposed current recordings in response to a pulse train in the presence of flecainide 100 mM. INa from the 1st, 2nd, 5th, 10th, 20th, 50th, 100th, and 200th pulse in the train are shown. B, time course of UDB for WT and KPQ channels. Drug concentrations of 1 () (for KPQ only), 3 (), 10 (), 30 (), 100 (), and 300 µM () flecainide. Relative amplitude of INa for each pulse number was calculated as the peak INa for the pulse number divided by the peak INa for the first pulse in that train. Symbols represent the mean of from three to five measurements; S.E. is shown as a bar. Solid lines represent fitting with a one or two exponential function with the parameters of the fit shown in Table 1.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Dose-response relationship for the UDB of peak and late INa by flecainide. Data points represent the mean relative decrease in amplitude of the peak current at the end of the pulse train for the results in Fig. 2 for WT () and KPQ () peak INa. The solid lines were obtained by fitting data to a single site binding equation: Fractional UDB = [flecainide]/([flecainide] + IC50) where [flecainide] is the micromolar concentration and IC50 is the half-blocking concentration. The IC50 values for WT () and KPQ () peak INa were 29 µM and 11 µM, respectively. For KPQ UDB, the late INa was measured as the relative decrease in amplitude for an average of the last 10 pulses in a 5 Hz train (see Fig. 5 at 3 mM) and mean data for 1, 3, 10, and 30 mM (D, n = 18) are shown. UDB of late INa at 100 and 300 µM for KPQ, and for all concentrations in WT, could not be measured because the current was too small.

Recovery from Use-Dependent Block of Peak I_Na Is Slower for KPQ. We measured recovery rates after UDB induced by a train of 100 24-ms depolarizing steps to 20 mV at 25 Hz in 10 µM flecainide. The pulse train was followed by a variable recovery interval (T) and a test depolarization (Fig. 4, inset). Summary results for normalized peak I_Na in response to the test depolarization are plotted versus a logarithmic scale of the recovery interval (Fig. 4). For control conditions, two components of recovery are apparent, with >95% of the recovery occurring within 10 ms for both WT and KPQ. This is consistent with rapid recovery from inactivation of unblocked channels (An et al., 1996; Wang et al., 1996c). The small amplitude of the slow recovery component in control (<5%) indicates that the relatively short (24 ms) conditioning depolarizations are not sufficiently long to induce slow recovery (Shander et al., 1995). In 10 µM flecainide, three components of recovery are apparent (Fig. 4) for both WT and KPQ. The fastest component (<10 ms) is consistent with recovery of unblocked channels from inactivation, an intermediate component (10 ms-1000 ms) is consistent with a rapid recovery from flecainide block, and the slowest component (>1000 ms) is consistent with a slow recovery from flecainide block. For recovery intervals <10 ms, recovery was faster for KPQ compared with WT consistent with recovery from inactivation of unblocked channels. For recovery intervals >10 ms, the recovery rate from flecainide block was decreased for KPQ compared with WT, especially for the first component of recovery. Results from recovery in the presence of 3 µM flecainide were similar to those of 10 µM (data not shown). This slower recovery for KPQ is consistent with the deeper UDB found for the mutant (Fig. 2) and it is also consistent with the slower development of block. Quasi-equilibrium for the amplitude of UDB is reached when the number of channels blocked during a pulse is equal to the number of channels being unblocked during the recovery interval. Slower rates of interaction during either interval will slow the process of achieving this quasi-equilibrium (Starmer, 1987).

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Recovery from UDB of peak INa. UDB was induced by trains of 100 pulses at 25 Hz followed by a single test pulse after a variable recovery interval DT (see protocol inset). To compare the time courses of recovery, current amplitudes were normalized to the peak INa for the test pulse at 3 s in control and 300 s in the presence of 10 µM flecainide. Symbols represent means; bars represent S.E. The averaged data were fitted with a two- (control) or a three- component (flecainide) exponential function: normalized INa = 1  [A1 × exp-t/1 + A2 × exp-t/2] or INa = 1  [A1 × exp-t/1 + A2 × exp-t/2 + A3 × exp-t/3], where t is the recovery time, 1-3 are recovery time constants, and A1-3 are fractional amplitudes of each component. The fitted parameters were as follows: control (A1, 1, A2, 2): WT (n = 8, ), 2.6 ms, 0.96, 909 ms, 0.04; KPQ (n = 10, ), 1.85 ms, 0.97, 833 ms, 0.03; and in flecainide (A1, 1, A2, 2, A3, 3): WT (n = 5, ), 4 ms, 0.57, 55.6 ms, 0.1, 38.5 s, 0.33; KPQ (n = 10, ), 2.2 ms, 0.52, 111 ms, 0.04, 52.6 s, 0.44.

Use-Dependent Block of Late I_Na Is Greater for KPQ. Fig. 5 shows the time course of UDB of peak and late (averaged between 20 ms and 23.5 ms) I_Na for 3 mM flecainide. At this concentration, little tonic block was exhibited and it is near the levels achieved in patients treated for arrhythmia. For KPQ, the late component of I_Na was more sensitive to UDB by flecainide than was peak I_Na as also shown in Fig. 3.

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Time course of UDB of late INa in response to pulse trains. UDB at peak INa () in the presence of 3 µM flecainide was compared with the late INa (), which was determined as the average current value between 20 ms and 23.5 ms. The data were normalized to the current amplitude of the first pulse in the train. The pulse train conditions are given in the legend to Fig. 2. Symbols represent means; bars represent S.E. (n = 5). The averaged data of the decrease in late INa was best fit with a double exponential function: Normalized INa = Af × exp-t/f + As × exp-t/s + base, where t is the pulse number, f and s are time constants (in pulses) of fast and slow components, and Af and As are fractional amplitudes of fast and slow components. The parameters of the fit were 0.30, 8.4, 0.39, 352, 0.31, for Af, f, As, s, baseline, respectively.

Flecainide Blocks the Open State of hH1. To confirm the affinity of flecainide for open state channels, we compared the amount of UDB for peak I_Na with various pulse durations in 100 µM flecainide. If flecainide has important additional affinity for the inactivated state, then prolonging the depolarization would be expected to increase the level of block. Pulse trains of depolarizations to 20 mV with various pulse durations (1, 2, 5, 10, 20, 100, 200 ms) were applied from a holding potential of 150 mV and the recovery interval was kept constant at 200 ms. The fractional UDB in both WT and KPQ saturated within the first 10 ms of the depolarization and less than 5% of block occurred after further pulse prolongation (Fig. 6). Although data such as these have traditionally been interpreted as indicating open state block (Anno and Hondeghem, 1990; Nitta et al., 1992), an alternative explanation that cannot be excluded includes preferential binding to another transiently available state, such as a preopen state.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Effect of pulse durations on UDB by 100 µM flecainide. After 200 1-ms depolarizing pulses, trains of 50 depolarizing pulses with various durations (2, 5, 10, 20,100, 200 ms) were applied with a constant recovery interval of 200 ms. Fractional UDB of peak INa for WT (n = 5, ) and KPQ (n = 7, ) are shown plotted for each of the pulse durations. The amount of fractional UDB was determined from the fitting of block development as described in Fig. 2. In a few cases (10-ms and 20-ms durations), the curve fits did not reflect the final level of block well and the average value of the steady-state normalized current amplitude in the last 10 pulses was used.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This study is the first description of the effects of a Vaughan-Williams Class Ic antiarrhythmic drug, the predominantly open channel blocker flecainide, on I_Na through human cardiac Na⁺ channels and on I_Na through the KPQ mutant of the LQT3 syndrome. Consistent with previous studies of flecainide block in nonhuman Na⁺ channels (Anno and Hondeghem, 1990), our results show that flecainide blocked I_Na in both a tonic and use-dependent manner and that recovery from block was slow and had multiple components. For the first time, we have characterized block of the late component of I_Na, the current flowing after peak I_Na that influences action potential duration especially for the KPQ channel where this late component is increased. For KPQ compared with WT, the open channel blocker flecainide 1) preferentially inhibited the late component of I_Na in both a tonic and use-dependent manner, 2) blocked peak and late I_Na with higher affinity, and 3) caused deeper UDB consistent with a demonstrated decrease in the recovery rate for flecainide for KPQ current.

The KPQ Mutant Has an Intrinsically Increased Affinity for Flecainide. The modulated receptor (Hille, 1977; Hondeghem and Katzung, 1977) and the guarded receptor (Starmer, 1987) models provide two frameworks for understanding antiarrhythmic drug block in terms of intrinsic association and dissociation rates for ion channels. Our results show different affinities of flecainide for WT versus KPQ channels and for peak versus late I_Na in KPQ. Do these results represent intrinsic affinity differences for drug binding to the channel protein, or are they instead secondary to the altered kinetics of KPQ? Flecainide has been considered to be an open channel blocker (Anno and Hondeghem, 1990; Nitta et al., 1992), having higher affinity for the open state than the resting or inactivated states. Our data confirm for hH1 that flecainide has preferential affinity for a state transiently available at the beginning of the depolarization, consistent with open state block (Fig. 6). We originally hypothesized that flecainide might have a greater blocking effect on KPQ because of an increased dwell time in the open state, rather than an intrinsic change in drug binding affinity. The results shown in Figure 6, however, suggest that use-dependent flecainide block occurred predominantly within the first 10 ms of depolarization for both WT and KPQ channels. Therefore, additional binding to a persistent open state is unlikely to be an important mechanism for the higher affinity block in KPQ channels. This suggests, instead, that the KPQ channel has an intrinsically higher open state affinity for flecainide than WT.

Implications for the Antiarrhythmic Drug Binding Site. These data also provide additional evidence that the III-IV linker of the Na⁺ channel participates as part of the binding site for antiarrhythmic drugs in addition to the previously implicated Domain IV S6 (Ragsdale et al., 1996). The binding site for flecainide is thought to be the same, or to at least overlap, the binding site for other antiarrhythmics and local anesthetic drugs such as lidocaine (Ragsdale et al., 1996). Modification of three key amino acids (IMF to QQQ) on the III-IV linker of the channel produced a channel that inactivated slowly and had a reduced affinity for lidocaine (Bennett et al., 1995). A naturally occurring mutation (T1313M) in the III-IV linker of the skeletal muscle Na⁺ channel also slowed inactivation and decreased lidocaine affinity and UDB (Fan et al., 1996). For T1313M, the decreased block was shown to be caused by an intrinsic change in channel affinity for the drug rather than secondary to the changes in inactivation properties, implicating the III-IV linker as a part of the binding site. Our results with flecainide show an increased intrinsic (that is, not explained by the altered channel kinetics) affinity difference for KPQ over WT, supporting the view that the III-IV linker forms part of the binding site for the drug.

Clinical Implications and Limitations of the Study. Mexiletine, generally considered to be an inactivated state blocker, has attracted the most attention as a therapeutic agent for LQT3 patients (Schwartz et al., 1995; Shimizu and Antzelevitch, 1997; Wang et al., 1997). The present study provides direct evidence that supports open channel block as a pharmacological model for the treatment of LQT3 patients. Compared with antiarrhythmic drugs such as flecainide, Class Ib drugs, such as mexiletine, provide less UDB at slow heart rates because of more rapid recovery kinetics. These kinetic differences offer a theoretical advantage for flecainide as a gene-specific therapy in patients with LQT3. Flecainide has been beneficial in a skeletal muscle disease in which Na⁺ channel mutants have increased late I_Na (Rosenfeld et al., 1997). Recently, it has been shown to correct the QTc interval in the electrocardiograms of KPQ LQT3 patients (Windle et al., 1999). Caution must be used in extrapolating the IC₅₀ values reported here to the clinical setting because of the experimental conditions required by the voltage clamp technique. These include the hyperpolarized holding potential (150 mV) used to minimize the effects of time-dependent kinetic shifts (Makielski et al., 1987; Wang et al., 1996a) and the lower temperature (Johns et al., 1989; Makielski and Falleroni, 1991) used to facilitate voltage control. Two additional considerations may also affect the clinical application of these results. First, in the CAST study, flecainide was associated with increased mortality in patients after myocardial infarction (CAST investigators, 1989). Patients with LQT, however, generally do not have structural heart diseases and thus flecainide might be used safely. Second, flecainide is not entirely specific for I_Na. At relatively high concentrations, flecainide suppressed other currents, such as ATP-sensitive potassium current (Wang et al., 1995b), rapidly activating delayed-rectifier potassium current (Follmer et al., 1992; Wang et al., 1996b), the transient outward current (Wang et al., 1995b), and L-type calcium current (Scamps et al., 1989). Nonetheless, the efficacy of flecainide to shorten the QT interval in patients (Windle et al., 1999) suggests that late I_Na block is a predominant mechanism contributing to the shortening.