Late sodium current contributes to the reverse rate-dependent effect of I_Kr inhibition on ventricular repolarization

Abstract

Background

The reverse rate dependence (RRD) of actions of I_Kr-blocking drugs to increase action potential duration (APD) and beat-to-beat variability (BVR) of APD is proarrhythmic. Therefore we determined if inhibition of endogenous, physiological late Na⁺ current (late I_Na) attenuates the RRD and proarrhythmic effect of I_Kr inhibition.

Methods and Results

Duration of the monophasic action potential (MAPD) was measured from female rabbit hearts paced at cycle lengths from 400 to 2000 ms and BVR was calculated. In the absence of drug, MAPD₉₀ and BVR increased as the cycle length was increased from 400 to 2000 ms (n=36 and 26, p<0.01). Both E-4031 (20 nmol/L) and d-sotalol (10 μmol/L) increased MAPD₉₀ and BVR at all stimulation rates and the increase was greater at slower than at faster pacing rates (n=19 and 11, 12 and 7, respectively, p<0.01). TTX (1 μmol/L) significantly attenuated the RRD of MAPD₉₀, reduced BVR, (p<0.01), and abolished torsade de pointes (TdP) in 5 of 6 hearts treated with either 20 nmol/L E-4031 or 10 μmol/L d-sotalol. Endogenous late I_Na in cardiomyocytes stimulated at cycle lengths from 500 to 4000 ms was greater at slower than at faster stimulation rates, and rapidly decreased during the first several beats at faster but not at slower rates (p<0.01, n=8). In a computational model, simulated RRD of APD caused by E-4031 and d-sotalol was attenuated when late I_Na was inhibited.

Conclusions

Endogenous late I_Na contributes to the RRD of I_Kr inhibitor-induced increases in APD and BVR and to bradycardia-related ventricular arrhythmias.

Introduction

Reverse rate (or frequency or use)-dependence (RRD) of action potential duration (APD) is an adaptive property of the normal heart wherein the APD shortens as heart rate increases.¹ Drug effects may also have a reverse rate-dependency, i.e., the effect of drug to prolong APD may be greater at slow vs. fast heart rates. Drug RRD is associated with ventricular proarrhythmic activity.^{1, 2} Inhibitions of rapidly-activating delayed rectifier K⁺ current (I_Kr) or inward rectifier K⁺ current (I_K1),³ and enhancement of L-type Ca²⁺-inward current (I_CaL)⁴ have been reported to prolong APD in a RRD manner,^{5, 6} whereas inhibition of I_Ks by chromanol 293B prolonged APD in a frequency-independent manner.^{3, 7}. RRD of APD caused by drugs that inhibit I_Kr is considered to be an undesired response because it results in a reduction of drug efficacy during tachycardia and may lead to proarrhythmic activity when the tachycardia is terminated. ^1,8

The mechanisms underlying the RRD of physiological and drug-induced enhanced RRD have not been clearly identified.⁵ In mammalian hearts in physiological conditions, I_Kr and I_Ks are major determinants of APD and I_Ks is increased at fast heart rates.⁹ An increase in I_Ks at faster heart rates due either to incomplete deactivation of current or to accumulation of extracellular K⁺ is considered a possible mechanism of physiological RRD.^10-12 Also, the time- and voltage-dependencies of both drug binding and disassociation from channels are presumed to be causes of drug-induced enhancement of RRD.^{13, 14} However, none of these mechanisms is fully supported by experimental and/or clinical evidence.⁵ Furthermore, an increased beat-to-beat variability of repolarization (BVR) is also associated with drug-induced ventricular tachycardia^15-19 and the mechanism responsible for the increased BVR and its rate dependency have not been identified.

The endogenous (or physiological) slowly-inactivating component of the inward tetrodotoxin (TTX)-sensitive Na⁺ current (late I_Na) in the heart is normally small.²⁰ This current is proarrhythmic when enhanced or when cardiac repolarization reserve is reduced by I_Kr blockers.^16-20 To our knowledge, the role of late I_Na in RRD of APD has not been previously investigated. We hypothesized that physiological late I_Na enhances the RRD and BVR of APD prolongation caused by I_Kr inhibitors in the heart, perhaps because the magnitude of late I_Na itself may be use-dependent.²⁰ In this study, the contribution of late I_Na to RRD of the I_Kr inhibitors sotalol and E-4031 was determined in rabbit isolated hearts and myocytes. TTX (1 μmol/L) and ranolazine (10 μmol/L) were used to provide selective inhibition of I_Na relative to other ion currents, and selective inhibition of late I_Na relative to peak I_Na, respectively.¹⁹ The contribution of late I_Na to RRD of APD was also simulated in a computational model of a rabbit ventricular myocyte AP.

Methods

Female rabbit isolated heart model

The use of animals in this investigation conformed to “Guide for the Care and Use of Laboratory Animals” (NIH publication No. 85-23). Animal use was approved by the Institutional Animal Care and Use Committee of Gilead Sciences (Palo Alto, CA, USA) and by the Administration of Regulation of Laboratory Animals (Hubei Province, China). Hearts from New Zealand White female rabbits weighing 2.5-3.5 kg were isolated, perfused by the method of Langendorff, and instrumented as previously described by Wu et al.¹⁷ Monophasic action potentials (MAPs) from the left ventricular epicardium and pseudo 12-lead electrocardiograms (ECGs) were recorded. The AV nodal area was thermally ablated to produce complete atrioventricular block, and hearts were paced at increasing cycle lengths (CLs) of 400, 500, 667, 1000 and 2000 ms for 3-4 minutes at each rate. After control (no drug) measurements were recorded, hearts were exposed to either 20 nmol/L E-4031 or 10 μmol/L d-sotalol in the absence and presence of 1 μmol/L TTX, allowing 20 min between increases in drug concentration to record a steady-state effect. The stimulation protocol at CL from 400 to 2000 ms was repeated before (control) and after each drug treatment.

BVR was determined by analysis of 30 consecutive stimulated beats using the following equation: Σ|MAPD_n+1-MAPD_n|/[30×√2], as described by Wu et al.^16-18 The slope of the restitution curves was calculated based on the fast slope portion (at stimulation CL of 400, 500 and 667 ms) of the relationship between the MAPD₉₀ and diastolic interval.

Whole cell patch clamp study

Rabbit left ventricular cardiomyocytes were isolated enzymatically as prescribed previously.²¹ Only cardiomyocytes with smooth and glossy edges and surfaces, clear transverse striations, and that could be stably voltage-clamped for 8 min without significant contraction were used. Seal resistance was above 1GΩ with less than 10 % variability. Currents were obtained with a patch-clamp amplifier (EPC-10, Heka Electronic, Lambrecht, Pfalz, Germany), filtered at 1 kHz and digitized at 10 kHz. During recording of late I_Na, the bath solution contained (in mmol/L):135 NaCl, 5.4 CsCl₂, 1.8 CaCl₂, 1 MgCl₂, 0.3 BaCl₂, 0.33 NaH₂PO₄, 10 glucose, 10 HEPES, 0.001 nicardipine, pH 7.3. The patch pipette solution contained (in mmol/L):120 CsCl₂, 1.0 CaCl₂, 5 MgCl₂, 5 Na₂ATP, 10 TEACl, 11 EGTA, 10 HEPES (pH 7.3, adjusted with CsOH). All experiments were carried out at a temperature of 23±1 °C.

To elicit a late I_Na in a single cell, twenty 300-ms depolarizing pulses to -20 mV from a holding potential of -90 mV were applied at CLs of 500, 667, 1000, 2000 and 4000 ms using a serial repeated-measurement protocol, with 3 min between each change of stimulation CL, during which membrane potential was held at -90 mV. In cardiomyocytes isolated from 8 rabbit hearts, the amplitude of late I_Na was measured at 200 ms after initiation of the depolarization step before (control, no drug) and 3 min after exposure of a myocyte to 1 and 4 μmol/L TTX, respectively.

Rabbit ventricular AP simulations

The rabbit ventricular AP was simulated using the Shannon-Bers model.^22,23 Simulation of the late I_Na was added to the original model using a Hodgkin-Huxley formalism as done by Hund et al.²⁴

$$I_{\mathit{NaL}}={\overline{G}}_{\mathit{NaL}}·m_L^3·h_L·(V-E_{\mathit{Na}})$$

Late I_Na activation (gating variable m) mimics the activation of the fast I_Na, whereas inactivation (gating variable h) was formulated as follow:

$$\begin{array}{lll}{h}_{L,\infty }& =& \frac{1}{1+\exp ((V_m+91)/6.1)}\\ & & {\tau }_{\mathit{hL}}=600\mathit{ms}\end{array}$$

Late I_Na maximal conductance (G_Na,L) was set to 0.012 mS/μF (i.e., 0.075% of the fast I_Na conductance) to simulate a small endogenous late I_Na in rabbit ventricular myocytes. Model differential equations were implemented in Matlab (Mathworks Inc., Natick, MA, U.S.A.) and solved numerically using a variable order solver (ode15s). The digital cell was stimulated with a current pulse (9.5 A/F, 5 ms) at various CLs from 400 to 2000 ms, and APD was measured as the interval between AP upstroke and 90% repolarization level (APD₉₀).

Based on the results of measurements of late I_Na in rabbit cardiomyocytes, applications of 1 and 4 μmol/L TTX were simulated by reducing G_NaL by 47 and 100 %, respectively. G_Kr was reduced by 55% to simulate the effect of 20 nmol/L E-4031 and by 45% to simulate the effect of 10 μmol/L d-sotalol application.

Statistical analyses

Data were plotted and analyzed using Prism version 5 (Graph Pad Software, San Diego, CA) and expressed as mean ± SEM. The significance of differences of measures before and after interventions in the same hearts was determined by repeated measure one-way analysis of variance (ANOVA) followed by Student-Newman-Keuls test. When treatment values were obtained at different rates from different groups of hearts, two-way ANOVA of repeated measures was used. A paired or un-paired student t test was used to determine the statistical difference between values of two means obtained from the same or different experiments, respectively.

Results

Relationships between pacing heart rate and MAPD or BVR in control hearts

In the absence of drug (control), MAPD₉₀ and BVR were significantly increased in a reverse rate-dependent manner from 148.7±1.67 and 0.20±0.01 to 192.9±3.1 and 0.31±0.01 ms (n=36 and 26, p<0.01, Figs. 1-3, open symbols) as the pacing CL was prolonged from 400 ms to 2000 ms.

Figure 1.

Rate-dependent changes of MAPD and BVR in absence (control, ○) and presence of 1 μmol/L TTX (●). MAPD₉₀ (A), BVR (B) and QRS interval (A, inset) were measured from rabbit isolated hearts paced at CLs of 400, 500, 667, 1000 and 2000 ms, respectively. *, p<0.05 compared with values at pacing CL of 400 ms, p<0.05. †, compared with control at same pacing CL, p<0.05.

Figure 3.

Effects of TTX and ranolazine (Ran) on the reverse rate-dependent changes of BVR of left ventricular MAPD induced by I_Kr inhibitors. BVR was calculated in absence (control, ○) and presence of either E-4031 (20 nmol/L, ●), E-4031 plus TTX (1 μmol/L, ■, A), E-4031 plus Ran (10 μmol/L, B), or d-sotalol (10 μmol/L) plus TTX (C). Each point indicates a mean of 4, 7 and 11 hearts, respectively. The TTX (or Ran) sensitive components of BVR are plotted in panel D. *, compared to CL of 400 ms, p<0.05-0.001, p<0.05-0.001; †, Compared to control at the same CL; p<0.05-0.001; ‡, compared to either E-4031 (A and B) or d-sotalol (C) at same CL, p<0.05-0.001.

TTX (1 μmol/L) alone did not change MAPD₉₀ or BVR at any tested pacing rate (n=7 and 6, compared to control, p>0.05, Fig. 1). The QRS interval was prolonged by 1 μmol/L TTX by 2.4±0.3 and 1.7±0.5 ms at CL of 400 and 2,000 ms (n=7, compared to control at same rate, p<0.05, inset in Fig. 1A).

RRD effects of I_Kr inhibitors to increase MAPD and BVR

E-4031 (20 nmol/L) and d-sotalol (10 μmol/L) significantly prolonged MAPD₉₀ (n=19 and 11, Fig. 2) and increased BVR (n=12 and 7, Fig. 3) at all stimulation rates (p<0.05-0.01) and the increase was greater at longer (2000 ms, 26.2±2.2 and 0.04±0.02 ms for E-4031, 17.3±66.9 and 0.07±0.02 for d-sotalol) than at shorter CLs (400 ms, 82.6±9.2 and 0.56±0.08 ms for E-4031, 66.9±9.2 and 0.19±0.02 for d-sotalol, p<0.01); i.e., the actions of E-4031 and d-sotalol on MAPD and BVR were reverse rate-dependent.

Figure 2.

Effect of TTX and ranolazine (Ran) on the reverse rate-dependent changes of left ventricular epicardial MAPD₉₀ caused by I_Kr inhibitors. Values of MAPD₉₀ were measured in the absence (control, ○) and presence of either E-4031 (20 nmol/L, ●), E-4031 plus TTX (1 μmol/L, ■, A), E-4031 plus Ran (10 μmol/L, B), or d-sotalol (10 μmol/L) plus TTX (C). Each point indicates a mean of 7, 9 and 11 hearts, respectively. The TTX (or Ran) sensitive component of MAPD₉₀ prolongation by each intervention is plotted in panel D. Insets represent the incidence of torsade de pointes in control, E-4031 or d-sotalol in absence and presence of TTX or Ran. *, compared to CL of 400 ms, p<0.05-0.001; †, Compared to control at the same CL; p<0.05-0.001; ‡, compared to either E-4031 (A and B) or d-sotalol (C) alone at the same CL, p<0.05-0.001.

Reversal by TTX and ranolazine of the RRD effects of I_K inhibitors

TTX (1 μmol/L) significantly (p<0.05) attenuated the rate-dependent increases in MAPD₉₀ and BVR in the presence of either E-4031(n=9 and 7, Fig. 2A and 3A) or d-sotalol (n=11 and 7, Fig. 2C and 3C) by 36±10 and 68±17%, and 24±8 and 46±17 % at CL of 667 ms (150 bpm), and 63±5 and 75±7 %, and 49±5 and 51±12 %, at CL of 2000 ms (30 bpm), respectively.

The TTX-sensitive component of MAPD and BVR was minimal in control hearts (Figs. 2D and 3D) but was significantly increased in the presence of either E-4031 or d-sotalol, especially at the slower pacing rate of 30 bpm (Figs. 2D and 3D).

In the presence of 20 nmol/L E-4031, the late I_Na inhibitor ranolazine (Ran, 10 μmol/L) also attenuated the RRD of MAPD and BVR, especially at the longer pacing CLs (n=10 and 6, p<0.05, Figs. 1B and 3B), in spite of the fact that ranolazine is also reported to block I_Kr as well as late I_Na.²⁵

Values of QT interval, Tpeak-Tend, and the index of Tpeak-Tend/QT interval had similar RRD as did MAPD and BVR (supplemental figure), and the effects of TTX and ranolazine on these parameters were also similar to the effects of TTX and ranolazine on the rate dependence of MAPD (supplemental figure).

When CL was increased from 500 to 1000 ms (i.e., when stimulation rate was decreased for 120 to 60 beats per minute), MAPD and BVR were increased by 20.6±1.2 (n=36) and 0.04±0.01 ms (n=26, p<0.01) in normal hearts (no drug). In the presence of E-4031 and d-sotalol, MAPD and BVR were further increased to 45.0±4.7 and 0.22±0.04 ms (n=19 and 12, respectively, p<0.01), and 38.1±2.4 and 0.11±0.02 ms (n=11 and 7, p<0.05), respectively. TTX (1 μmol/L) reduced the increases in MAPD and BVR in the continued presence of either E-4031 or d-sotalol to 35.7±4.7 and 0.22±0.08 ms (p<0.05), and 28.0±2.2 and 0.08±0.02 ms (p<0.05), respectively.

The plot of the dependence of MAPD on the duration of the diastolic interval is known as the restitution curve,²⁶ and is similar to the dependence of MAPD on CL that is shown in Fig. 2. Inhibition of I_Kr by either E-4031 or d-sotalol increased the steepness of the slope of the restitution curve from 0.103±0.021 (control) to 0.158±0.017 and from 0.112±0.024 to 0.157±0.023, respectively (p<0.05, Fig. 4). TTX (1 μmol/L) flattened the restitution curves in the continuous presence of either E-4031 (Fig. 4A) or d-sotalol (Fig. 4B) and decreased the slope of the curves to 0.109±0.029 and 0.131±0.028, respectively (p<0.05).

Figure 4.

MAPD restitution curve in control (no drug) and in presence of either drug alone (E-4031 in A or d-sotalol in B), or drug plus 1 μmol/L TTX. *, Compared with CL of 400 ms, p<0.05; †, Compared to control at the same CL; p<0.05; ‡, compared to either E-4031 (A and B) or d-sotalol (C) at same CL, p<0.05.

Simulated effect of late I_Na inhibition on RRD of APD₉₀

We used the Shannon-Bers rabbit ventricular AP model to recapitulate our MAPD₉₀ results from rabbit hearts. Representative AP traces are shown in Figure 5D. The control APD was 252 ms at a CL of 2000 ms, and 185 ms at a CL of 400 ms (Fig 5A, open circles). The rate-dependent properties of late l_Na contributed to APD adaptation, as the simulated current amplitude was larger at slow vs. fast pacing rates (not shown). When the effect of TTX was simulated by G_Na,L reductions of 47 % (1 μmol/L TTX) and 100 % (4 μmol/L TTX), APD₉₀ was decreased (Fig 5). APD adaptation to change of stimulation rate in the absence and presence of 1 μmol/L TTX was not markedly different (APD₉₀ shortened by ~23 and ~27% with and without TTX respectively, when decreasing the CL from 2000 to 400 ms). TTX shortened APD₉₀ more prominently at longer CLs (1000 and 2000 ms), especially when simulating complete late I_Na block by 4 μmol/L TTX (Fig. 5A).

Figure 5.

Simulated effect of TTX on the reverse rate-dependent changes of rabbit action potential duration (APD₉₀). A, APD₉₀ prolongation as pacing CL was increased from 400 to 2000 ms in absence (open circles) and presence of TTX (1 and 4 μmol/L). B. RRD of APD₉₀, predicted when reducing I_Kr conductance to mimic E-4031 administration (solid circles), was greatly attenuated when late I_Na was inhibited by TTX (1 and 4 μmol/L, solid symbols). C. RRD of APD₉₀ predicted when reducing I_Kr conductance to mimic d-sotalol administration (solid circles) was reduced or nearly abolished when late I_Na was inhibited by 1 and 4 μmol/L TTX (solid symbols). D. Representative AP traces obtained when a digital cell was paced at CLs of 500 ms (left) and 2000 ms (right) in control, in presence of I_Kr block E-4031 alone, and in presence of simultaneous inhibitions of l_Kr and late I_Na.

Simulation of I_Kr block by E-4031 predicted a marked prolongation of the rabbit ventricular AP showing clear RRD (Fig. 5B, squares), being larger at 2000 than at 400 ms CL (100 vs. 32 ms). When late I_Na inhibition was simultaneously simulated, the model predicted a reduction of RRD in the presence of E-4031 (Fig. 5B, closed circles). Similar results were found when I_Kr block by d-sotalol was simulated (Fig. 5C), i.e., APD prolongation induced by G_Kr reduction was more pronounced at slower than at faster heart rates, and this effect was markedly reduced when 47 and 100 % block of late I_Na was simulated.

Inhibition of endogenous late I_Na abolished slow rate-related TdP in the presence of E-4031 and d-sotalol

Neither early afterdepolarizations (EAD), extra-ventricular beats (EVB) nor TdP was observed in 28 control hearts at any tested pacing rate. In contrast, in 15 of 17 (88 %) hearts treated with 20 nmol/L E-4031, EAD, frequent EVBs and episodes of TdP were observed when CL was increased from 400 to 2000 ms. Similar arrhythmic activities were observed 6 of 11 (55 %) hearts treated with d-sotalol. These results are consistent with a previous report of the proarrhythmic effects of QT prolonging drugs. EADs, EVBs and episodes of TdP caused by 20 nmol/L E-4031 were abolished by 1 μmol/L TTX and 10 μmol/L ranolazine in 5 of 6 (83 % Fig. 2A, inset) and 8 of 8 (100 %, Fig. 2B, inset) hearts, respectively. Similarly, 1 μmol/L TTX abolished TdP caused by 10 μmol/L d-sotalol in 5 of 6 hearts (83 %, Fig. 2C, inset).

Effect of stimulation rate on magnitude of endogenous late I_Na in rabbit ventricular myocytes

The magnitude of late I_Na in rabbit isolated cardiomyocytes decreased with increasing stimulation rate (CL from 4000 to 500 ms) and with use (i.e., from first pulse to 20^th pulse) (see representative traces in Fig. 6 and summary data in Fig. 7). The amplitude of late I_Na recorded during the first pulse of each pulse train was similar: 48.1±3.8, 48.2±3.8, 48.1±4.3, 48.2±4.3 and 47.9±4.1 pA, (n=8, p=NS, Figs.7, 8). TTX (1 and 4 μmol/L) caused a significant decrease in late I_Na at all stimulation rates, as the rate-late I_Na response relationship was shifted down in parallel (p<0.01 for each rate, Figs. 6-8). The steady-state amplitudes of late I_Na recorded from cells paced at 2.0, 1.5, 1.0, 0.5 and 0.25 Hz were -18.8±3.9, -24.3±3.8, -32.0±4.2 and -41.4±4.4 pA, respectively (n=8, p<0.05-0.01). TTX at concentrations of 1 and 4 μmol/L attenuated late I_Na at all stimulating rates (Figs. 6-8).

Figure 6.

Representative recordings of the rate (2.0, 1.5, 1.0, 0.5 and 0.25 Hz) dependent changes of late Na⁺ current (late I_Na) in a single rabbit ventricular myocyte. Traces of I_Na at different stimulation cycle length (CL in ms) in absence (control, A) and presence of TTX (1 μmol/L in B; 4 μmol/L in C) from impulses # 1, 2, 5, 10 and 20. The depolarization pulse protocol to elicit late I_Na is shown in the inset.

Figure 7.

TTX (1 and 4 μmol/L) reduced the beat-to-beat (from pulse 1 to pulse 20) changes of late sodium current (I_Na) at shorter (from 500 to 2000 ms in CL, A-D) but not at longer (4000 ms in CL, E) stimulation cycle length. The TTX-sensitive component of late I_Na decreased with a decrease in stimulation CL, as shown in panel F. Each point indicates a mean of measurements from 10 cells. Absolute values of amplitudes of late I_Na are shown. *, Compared to pulse #1, p<0.05.

Figure 8.

The rate-dependent decrease of late I_Na (measured as TTX-sensitive current). TTX (1 and 4 μmol/L) was used to inhibit late I_Na. Late I_Na at pulse #1 (A) and #20 (B), the TTX sensitive component (control-TTX) at pulse #20 (C), and the reduction of late I_Na between pulses #1 and 20 (D) are plotted. *, Compared with CL of 500 ms; †, compared with control at same rate; ‡, compared to 1 μmol/L TTX at same rate; p<0.05. Absolute values of amplitudes of late I_Na are shown.

The amplitude of late I_Na during a train of 20 pulses decreased from the first to the 20^th pulse, and the decrease was greater when the stimulation rate was increased (see Fig. 8D for summary data). Compared to pulse #1, the amplitude of late I_Na at pulse #20 was significantly decreased to 19.8±4.0 (Fig. 7A), 22.6±4.0 (Fig. 7B), 25.2±4.0 (Fig. 7C) and 32.4±4.4 (Fig. 7D) pA at CLs of 500, 667, 1000, and 2000 Hz, respectively, (n=8, p<0.05) but did not change at CL of 4000 ms (42.5±4.6 pA, Fig. 7E, n=8, p=NS). A higher rate of stimulation was associated with a faster decrease in late I_Na. The greatest decrease of late I_Na occurred during the first few pulses at each rate (Fig. 7).

Discussion

The findings in this study suggest that physiological late I_Na may play an important role in the enhanced RRD of APD and BVR caused by QT-prolonging agents that inhibit I_Kr. RRD of MAPD and BVR were demonstrated using female rabbit isolated hearts, and RRD of APD was simulated in a computer model. In the continuous presence of I_Kr inhibitors, TTX (1 μmol/L) and ranolazine (10 μmol/L) reduced the RRD of MAPD and BVR induced by I_Kr inhibitors, and abolished slow rate-related episodes of TdP in the presence of I_Kr inhibitors. Indeed, the amplitude of endogenous late I_Na was larger at slower than faster heart rates. Because TTX and ranolazine are known to block cardiac late I_Na (NaV1.5) at concentrations of 1 and 10 μmol/L, respectively, ^{19, 20} these results suggest that inhibition of late I_Na may diminish the RRD of the APD/QT interval prolongation and BVR, and therefore may antagonize the slow rate- and pause-triggered ventricular arrhythmias that may be caused by disease or drug-induced inhibition of I_Kr.

RRD is an implicit property of the heart that reflects the underlying rate dependence of individual cardiac ion channel currents ^{5, 27}, and thus their contribution to membrane resistance during the AP plateau. In this study, the magnitude of late I_Na was also found to be reverse rate-dependent. Thus, reduction of the reverse rate-dependent inward late I_Na by TTX or ranolazine “offset” reduction of outward K⁺ current by reverse rate-dependent I_Kr-blocking drugs. Inhibition of late I_Na by either 1 μmol/L TTX or 10 μmol/L ranolazine reduced the RRD of MAPD in hearts treated with the I_Kr blockers E-4031 and d-sotalol (Figs. 2, 3). Simulated data from a computer model of rabbit heart electrical activity (including changes in intracellular Na⁺, Ca²⁺ and late I_Na) recapitulated the RRD of APD in the presence of I_Kr blockers, and predicted that a reduction of I_Kr accentuates the RRD of APD. Thus, both experimental data and model predictions provide evidence that the RRD of effects of E-4031 and d-sotalol were reduced when late I_Na was inhibited, and that the contribution of late I_Na to prolongation of ventricular repolarization was greater at slower heart rates. The latter result is consistent with a previous study of LQT3 ΔKPQ mutant Na⁺ channels expressed in HEK293 cells, wherein it was reported that the amplitude of ΔKPQ late I_Na was strongly rate-dependent.²⁸

Increased BVR is proarrhythmic and has been associated with interventions that either block I_K or augment late I_Na, especially at slower heart rates.¹⁵ Late I_Na blockade with 1 μmol/L TTX did not significantly alter baseline RRD of BVR in rabbit hearts, but attenuated the increased RRD of BVR during administration of E-4031 and d-sotalol. Although inhibition of physiological late I_Na by TTX and lidocaine is sufficient to shorten APD in normal cardiac tissue.^{19, 20, 29}, our findings suggest that late I_Na contributes more to BVR when repolarization reserve is reduced and the net transmembrane current is small (i.e., during the long AP plateau in the presence of an I_Kr inhibitor) and support previous reports that endogenous late I_Na is proarrhythmic in hearts with reduced repolarization reserve.^{16, 18, 20} We speculate that reduction of a depolarizing current during the AP plateau shortens the ‘vulnerable phase’ in which stochastic fluctuations in ion channel or transporter currents, or local Ca²⁺ concentrations, can have significant impact on the duration of ventricular repolarization.¹⁵ Indeed, Ca²⁺-dependent mechanisms are thought to underlie BVR, as exogenous intracellular Ca²⁺ buffering suppresses it. ³⁰

The increased density of late I_Na recorded from rabbit cardiomyocytes paced at slow vs. fast rates is consistent with the RRD of MAPD and BVR in isolated hearts and in the computational model.¹⁷ The increase in late I_Na at slower heart rates may be attributed to decreased inactivation of Na⁺ channels at these slow rates,³⁰ as the greatest decrease in late I_Na was found during the first several beats when the pacing rate was increased (Fig. 7). In the computer model, APD adaptation (i.e., shortening) to an increase of pacing rate was linked to increased intracellular Na⁺ accumulation at fast rates, and its consequences to increase outward Na⁺ pump and NCX currents.^{27, 31} A potential positive feedback between increased late I_Na and reduced I_Kr can explain the effect of an enhanced late I_Na to cause arrhythmias when repolarization reserve is reduced.¹⁹ An increase in late I_Na may also cause changes in Ca²⁺ handling that alter APD adaptation by multiple mechanisms.^{20, 27, 32}

The observation that late I_Na plays an important role in the enhanced RRD of I_Kr inhibitors may explain previous experimental and clinical findings that RRD is greater in Purkinje fibers and mid-myocardial cells where the late I_Na is greater than in epicardial cells,³³ and that I_Kr-blocking drugs such as amiodarone and ranolazine that also block late I_Na have no RRD and no or minimal proarrhythmic risk.^{16, 17, 34} On the other hand, due to the synergistic effect of increased late I_Na and inhibition of I_Kr to increase APD and BVR,^{17, 35} an increase of late I_Na, as occurs in the ischemic heart and other structure heart diseases,³⁶ can potentiate the proarrhythmic effect of an I_Kr blocker.

Conclusion

Endogenous physiological late I_Na plays an important role in the RRD of APD and BVR caused by I_K inhibitors. Inhibition of endogenous late I_Na may diminish the rate-dependent prolongation of the APD/QT interval by either drugs or pathological conditions that decrease I_Kr, and may decrease the occurrence of slow rate- or pause-triggered cardiac arrhythmias. If late I_Na were to be exacerbated by pathological conditions, it is possible that it’s impact to increase the RRD of I_Kr blocking drugs would also be augmented.