Late I_Na blocker GS967 Suppress Polymorphic VT in a Transgenic Rabbit Model of Long QT Type 2

Abstract

Background –

Long QT syndrome (LQTS) has been associated with sudden cardiac death likely caused by early afterdepolarizations (EADs) and polymorphic ventricular tachycardias (PVTs). Suppressing the late sodium current (I_NaL) may counterbalance the reduced repolarization reserve in LQTS and prevent EADs and PVTs.

Methods –

We tested the effects of the selective I_NaL blocker GS967 on PVT induction in a transgenic rabbit model of LQTS type 2 (LQT2) using intact heart optical mapping, cellular electrophysiology and confocal Ca²⁺ imaging, and computer modeling.

Results –

GS967 reduced (ventricular fibrillation) VF induction under a rapid pacing protocol (n=7/14 hearts in control vs. 1/14 hearts at 100 nM) without altering APD or restitution and dispersion. GS967 suppressed PVT incidences by reducing Ca²⁺-mediated EADs and focal activity during isoproterenol perfusion (at 30 nM, n=7/12 and 100 nM n=8/12 hearts without EADs and PVTs). Confocal Ca²⁺ imaging of LQT2 myocytes revealed that GS967 shortened Ca²⁺ transient duration via accelerating Na⁺/Ca²⁺ exchanger (I_NCX)-mediated Ca²⁺ efflux from cytosol, thereby reducing EADs. Computer modeling revealed that I_NaL potentiates EADs in the LQT2 setting through 1) providing additional depolarizing currents during AP plateau phase, 2) increasing intracellular Na⁺ (Na_i) that decreases the depolarizing I_NCX thereby suppressing the AP plateau and delaying the activation of slowly-activating delayed rectifier K⁺ channels (I_Ks), suggesting important roles of I_NaL in regulating Na_i.

Conclusions –

Selective I_NaL blockade by GS967 prevents EADs and abolishes PVT in LQT2 rabbits by counterbalancing the reduced repolarization reserve and normalizing Na_i.

Graphical Abstract

Introduction

Long QT syndrome type 2 (LQT2) is a congenital disease caused by the mutation of KCNH2 or the human ether-a-go-go related gene (hERG) α subunit, which encodes the rapid delayed rectifying potassium current (I_Kr)^{1, 2}. The loss of function of I_Kr in LQT2 hearts leads to prolongation of action potential duration (APD), and is associated with polymorphic ventricular tachycardia (PVT) and sudden cardiac death (SCD).¹

The hallmark of long QT syndrome is early afterdepolarizations (EADs).² EADs occur in the setting of delayed repolarization, which promotes regenerative depolarizing currents.^{3, 4} The main ion current responsible for the depolarization phase of the EADs is the window current of L-type Ca²⁺ channel (I_CaL).⁴ However, the late Na⁺ current (I_NaL) could also, along with the I_CaL window current, promote EADs by providing additional depolarizing force during the plateau phase of action potentials (APs).^5–8

I_NaL is carried by persistent openings (or re-openings) of small numbers of Na⁺ channels during the plateau phase of the AP. Although the amplitude of I_NaL is small (less than 0.1% of peak I_Na), its contribution to APD is not negligible⁵. Tetrodotoxin (TTX), known to block peak I_Na with a moderate preference for I_NaL (10 times higher selectivity to I_NaL)⁹, suppressed EADs in a drug-induced experimental model of LQTS.¹⁰ Ranolazine, an anti-anginal medication that preferentially blocks I_NaL, effectively reduced EADs and VTs in drug-induced animal models of LQT2 and LQT3.^{11, 12} GS-458967 (GS967), an I_NaL blocker with high specificity and selectivity for late vs peak I_Na, has been reported to abolish EADs and VTs in drug-induced animal models of LQT2 and LQT3 without deleterious effects.¹³ However, in a recent clinical trial (NCT02104583), the selective I_NaL blocker, GS6615, did not reduce the number of VTs in patients with implanted ICDs suggesting that I_NaL blockade should be tested in a genotype- and disease-specific context.

The purpose of the current study was to test the effect of the I_NaL blocker GS967 in suppressing EADs in a transgenic rabbit model of LQT2. LQT2 rabbits were previously shown to suffer SCD caused by frequent EADs and PVTs.^14–16 EADs in LQT2 rabbits are associated with remodeling in Ca²⁺ handling, particularly in the SR Ca²⁺ release channel, ryanodine receptor (RyR2), which causes elevation of intracellular Ca²⁺ (Ca_i) during the plateau phase of APs, thereby promoting EADs.¹⁷ Previous studies suggested that inhibiting I_NaL can reduce Na⁺ overload, which can enhance Ca²⁺ exclusion from cytosol via Na⁺/Ca²⁺ exchanger (NCX1), thereby reducing Ca²⁺ overload and EADs.^{13, 18} Here, we used GS967 to selectively block I_NaL in LQT2 rabbits and found that GS967 suppressed EADs and PVTs through modulating NCX1 and I_Ks. This study suggests that the selective inhibition of I_NaL could be explored as a novel effective therapy for preventing arrhythmias in LQT2 syndrome.

Methods

The data, analytic methods, and study materials will be/have been made available to other researchers for purposes of reproducing the results or replicating the procedure. The source code of computer modeling is available at https://github.com/MingwangZhong/Late-Sodium.

Heart preparation and optical mapping

Littermate control (LMC, n=7 females) and transgenic rabbits (LQT2, n=9 females and 5 males), averaging 9 months old / 3.33kg body weight / 8.46g heart weight, were euthanized with buprenorphene (0.03 mg/kg IM), acepromazine (0.5 mg kg⁻¹ IM), xylazene (15 mg kg⁻¹ IM), ketamine (60 mg kg⁻¹ IM), pentothal (35 mg kg⁻¹ IV), and heparin (200 U kg⁻¹). This investigation conformed to the current Guide for Care and Use of Laboratory Animals published by the National Institutes of Health (NIH Publication No. 85–23, revised 1996). Blebbistatin (5 μmol L⁻¹) was perfused to reduce movement artifact.¹⁹ The AV node was ablated to control heart rate, and hearts were paced at 350 ms basic cycle length (CL).

GS967, selective late Na⁺ channel blocker, was provided from Gilead Science. The optical apparatus and analysis of AP have been previously described¹⁶ (see detail in the Data Supplement). All hearts (LMC, LQT2) underwent three stimulation protocols in the following order: 1) S1S2 protocol to reduce CL by 10 ms for APD restitution and tissue refractory period, 2) 10 min rest followed by ramp pacing in which CL was decreased by 10 ms until loss of 1:1 capture or induction of VF, and 3) a bolus injection of isoproterenol (ISO) into the bubble trap (~140 nM for 20 seconds) to induce EADs and PVTs. The stock solution of GS967 (1 mM) was prepared in DMSO¹³ and added to Tyrode’s solution to have the final concentrations of 30 nM and 100 nM (30 and 100 μL of stock solution to 1 L of Tyrodes’ solution). The stimulation and isoproterenol protocols were repeated and EAD/PVT events were monitored by optical mapping and ECG recording. The vehicle group was treated with 100 μL DMSO only.

Confocal Ca²⁺ imaging and I_NaL measurements from single myocytes

Detailed description of single cell Ca²⁺ imaging and voltage clamp experiments are available in the Supplemental Material. Cytosolic Ca²⁺ changes from isolated cardiomyocytes were monitored using a Leica SP5 confocal system using the cell-permeable Ca²⁺ indicator Rhod-2 AM (Life Technologies, Grand Island, NY).¹⁷ Myocytes were perfused with 50 nM ISO (10 min.) and paced at 0.5 Hz. The time of the decay to 25% of Ca²⁺ transient amplitude after the peak (t₇₅) was used as a parameter describing the decay kinetics of Ca²⁺ transients in LQT2.¹⁷ The efflux of Ca²⁺ through NCX1 was assessed using the time of decay to 50% of peak amplitude of caffeine-induced Ca²⁺ transient (10 mM).

The I_NaL current was recorded at 34–36°C with Axopatch-200B amplifier, Digidata 1440A and pCalmp 10 (Axon Instruments). The sodium current was activated by 220 ms depolarizing pulses from −120 mV holding potential to +40 mV in 10 mV increments (see the Supplemental Material and Figure S1 for detail). For each cell, I_NaL was normalized to appropriate cell capacitance and expressed as the current density expressed in pA/pF.

Computational model of I_NaL

We used the rabbit ventricular myocyte detailed model of Zhong et al²⁰ with dynamically changed Na_i. We included the I_NaL equations described in²¹. Since the activation and inactivation time scales of the normal Na model are very small at voltage larger than −30 mV, we adopted the rapid kinetics approximation (m_∞ = m h_∞ = h and, Eq. 1). The expressions of m_∞ and h_∞ as a function of voltage (Figure S3 in the Supplemental Material) are quantitatively fitted with our patch-clamp measurements for LMC and LQT2 myocytes. The parameters of the late sodium model and the modified parameters of the detailed model are provided in the Supplemental Data.

$$I_{NaL}=g_{NaL}{m}_{\mathrm{\infty }}^3{h}_{\mathrm{\infty }}^2(V_m-E_{Na})$$

$$m_{\mathrm{\infty }}=\frac{1}{1+e^{-(V_m-V_1)/V_2}}$$

$$h_{\mathrm{\infty }}=\frac{1}{1+e^{(V_m-V_3)/V_4}}$$

Statistical Analysis

Continuous variables of optical mapping data are represented as mean ± standard deviation. The parameters of Ca²⁺ transients recorded from isolated myocytes are represented as mean ± SEM. The values of APD, APD dispersion, conduction velocity, alternans, and Ca²⁺ handling before and after GS967 were compared using ANOVA.

Results

APD characteristics in LMC and LQT2 with GS967

We first determined whether GS967 changes the basic electrophysiological characteristics of LQT2 rabbit hearts. Hearts were paced at 350 ms CL, and GS967 was added to the perfusate at 30 and 100 nM concentrations. At 350ms CL, GS967 did not have a significant effect on the APD or APD dispersion in either LMC or LQT2 rabbits (p=0.366 and 0.313 for APD and APD dispersion respectively, Figure 1). Figure 2A and 2B shows APD and conduction velocity (CV) restitution curves from LMC and LQT2 rabbits under the S1S2 protocol. The left and right ventricular refractory periods were similar before and after GS967 administration in LMC rabbits (panel A, Δrefractoriness = 1±21 and 6±5 ms at 30 and 100 nM, respectively). APD restitution curves in the range from the basic cycle length of 350 ms to tissue refractoriness were not affected by GS967 in either LMC or LQT2 hearts. However, tissue refractoriness was slightly prolonged in LQT2 hearts despite no change in APD under 350 ms CL pacing (the increase in refractoriness = 13±11 and 29±15 ms at 30 and 100 nM respectively, paired t-test, p < 0.05). CV was not affected by GS967 in the broad range of diastolic interval (DI, 50 – 150 ms) but showed a tendency towards slowing at shorter DIs in both LMC and LQT2 hearts (Figure 2B). Table 1 summarizes the changes in APD and CV before and after GS967. APDs did not show statistically significant differences among 0, 30, and 100 nM at CL=350 and 250 ms. CV was modestly reduced at short CL by 100 nM GS967 (from 0.68±0.04 to 0.59±0.06 m/s, p < 0.05; see Table 1).

Figure 1.

Action potential traces and maps of activation and duration. A & B) Action potential traces before and after GS967 from LMC and LQT2 hearts. C & D) Maps of activation and APD from LMC and LQT2 hearts in the absence and the presence of 30 and 100 nM GS967.

Figure 2.

Lack of effect of GS967 on APD, CV restitution, or alternans in LMC (left column) and LQT2 hearts (right column). A) APD restitution in the absence and presence of 30 nM GS967. APD restitution curves did not show statistical differences in either LMC or LQT2 rabbits before or after GS967. Black; 0 nM, Red; 30 nM GS967. B) CV restitution from LMC and LQT2. The CV was not affected at CL=350 ms but showed slight reduction at short DIs. C) Sample traces of action potential alternans from LMC and LQT2 rabbits during ramp pacing. D) APD differences between odd and even beats (ΔAPD_alternans) before (black) and after 30 nM GS967 (red).

Table 1.

APD and CV in the absence and presence of GS967 at CL = 350 ms and 250 ms in LQT2 rabbits.

	0 nM		30 nM		100 nM
CL	350 ms	250 ms	350 ms	250 ms	350 ms	250 ms
APD (ms)	221±23	197 ± 12	237±27	203±15	227±31	201±8
CV (m/s)	0.75±0.06	0.68±0.04	0.75±0.03	0.64±0.03	0.74±0.03	0.59±0.06*

Values are mean ± STDEV obtained from 7 hearts. APDs were not affected by GS967, but CVs were modestly reduced (13%) at fast CL (250 ms) by 100 nM GS967 (n=7, * indicates paired t-test, p<0.05 between 0 nM and 100 nM at CL=250 ms).

APD alternans and VF induction under GS967

We investigated the effect of GS967 on APD and conduction under a ramp pacing protocol in which the stimulation CL was sequentially reduced in 10-msec steps. The pacing protocol often elicited 2:1 block at shorter CL in the presence of 100 nM GS967; hence we focused on the 30 nM concentration for the ramp pacing protocol. Figure 2C shows representative traces of alternans from LMC and LQT2 hearts under the ramp pacing protocol. The amplitude of APD alternans (ΔAPD_alternans) defined by |APD_odd – APD_even| was greater in LQT2 hearts and sharply increased at short CL below 200 ms. ΔAPD_alternans in the presence GS967 (red lines) shows a trend toward reduction under GS967 in both LMC and LQT2 hearts, particularly at short CL (panel D) but was not statistically significant. Of note, during the ramp pacing protocol, the episodes of VF were inducible in 7 of 13 LQT2 hearts (53.8%). At concentrations of 30 nM and 100 nM GS967, the incidence of VF induction during the ramp pacing was reduced to 5 (38.5%) and 1 (7.7%) out of 13 LQT2 hearts (fisher’s exact test between 0 and 100 nM GS967, p < 0.05, see Figure S2 in the Supplemental Material), respectively.

GS967 Prevented EAD and PVT Events in LQT2

We sought to determine whether GS967 could prevent EADs and PVT events at concentrations of 30 nM and 100 nM. A bolus injection of ISO to LQT2 hearts causes numerous EADs and PVTs (Figure 3). As shown in panel E, a total of 62 EAD/PVT events occurred among 12 LQT2 hearts. After adding GS967 to the perfusate, EAD/PVT events were markedly reduced (30 and 7 EAD/PVT events at 30 nM and 100 nM GS967, respectively). Note that the reduction in EAD/PVT events was associated with noticeable APD shortening by GS967 in the presence of ISO (panel D). This is in contrast to no noticeable changes in APD in the absence of ISO (see Figure 1 and Figure 2). GS967 at 30 nM prevented EADs in most LQT2 hearts (n=7 of 12 hearts), and 100 nM of GS967 further reduced EADs and PVTs in 1 out of 5 hearts compared with 30 nM (panel E). In 4 out of 12 LQT2 hearts the EADs led to the development of monomorphic VT in the presence of 100 nM GS967 (panel F). However, 300 nM GS967 suppressed these EADs and the VT in all hearts (averaged EAD/VT events per heart = 8.2±9.6, 4.7±7.2, 1.4±2.1, and 0.3±0.94 for 0, 30, 100, and 300 nM GS967, ANOVA test p<0.05 between 0 and 30, 100 nM, 300 nM). None of the LMC hearts (n=7) showed EAD/PVT events during ISO in the presence of 0, 30, or 100 nM GS967.

Figure 3.

Prevention of EADs and pVTs in LQT2 rabbits by GS967. A-C) Representative traces of LQT2 AV-node-ablated heart with continuous pacing at 0.5 Hz in the absence and presence of 0 nM, 30 nM, and 100 nM GS967. D) APDs at 0.5 Hz after isoproterenol (n=12, 0 nM; 490.0±28.4 ms, 30 nM; 427.0±43.1 ms, 100 nM; 321.2±37.2 ms, p<0.05). E) Total number of EAD/PVT events in LQT2 rabbits (n=12) administered 0 nM, 30 nM, and 100 nM of GS967. F) Percentage of LQT2 rabbit hearts that showed EAD/PVT events at 0 nM, 30 nM, and 100 nM GS967 (8/12 at 0 nM, 5/12 at 30 nM, and 4/12 at 100 nM). G) Representative traces and activation maps in the absence and presence of 30 nM GS967 in LQT2 hearts. PVT in the absence of GS967 demonstrated complex activation patterns of mixed focal activity (star in the middle panel) and reentry. After 30 nM GS967 administration, PVT showed relatively repetitive activation patterns (panel G bottom activation maps and also see online supplementary movies S1 and S2). H) Reduced focal activity in 30 nM GS967 (t-test, p<0.05, n=10 hearts).

GS967 Suppresses EADs and Multi-focal Activity

We investigated the activation pattern and dynamics of PVTs before and after administration of GS967 to better understand its protective mechanisms. Due to the limited number of PVT events in the presence of 100 nM GS967, only PVT events that occurred in the presence of 30 nM GS967 were analyzed. Under the control condition (0 nM), PVTs showed complex wave dynamics with multiple focal activity and beat-to-beat changes in activation pattern (Figure 3G upper panels). The traces of PVTs show that V_m oscillates at a relatively high plateau phase under control conditions. However, GS967 at 30 nM (Figure 3B) markedly changed action potential shapes and activation patterns of PVTs, shortened APD of the first beat (from 596 ms to 296 ms, see Figure 3D and G), reduced small-amplitude V_m oscillations (panel G), and reduced the complexity in activation patterns of PVTs (panel G, activation maps). The activation maps during PVTs frequently showed focal activity (star in the 3^rd activation map under 0 nM). However, focal activity was markedly reduced (panel H) under 30 nM GS967. This result indicates that the blockade of I_NaL by GS967 is beneficial by suppressing EADs and PVTs via reducing focal activity.

GS967 Normalizes Defective Ca²⁺ Homeostasis

A previous study by our group showed that abnormal Ca²⁺ handling in LQT2 myocytes potentiates EADs.¹⁷ Therefore, using confocal Ca²⁺ imaging of single myocytes, we investigated whether blocking I_NaL has a major impact on Ca²⁺ handling in LQT2 rabbits, thereby suppressing EADs in LQT2 hearts. Figure 4A shows representative traces of Ca²⁺ transients from LQT2 myocytes treated with ISO (50 nM, 10 min) in the presence (red) and absence (black) of GS967 (100 nM). GS967 accelerated Ca²⁺ transient decay (panel B), resulting in a significant shortening of Ca²⁺ transients without affecting Ca²⁺ transient amplitude.

Figure 4.

GS967 accelerates Ca²⁺ transient decay in LQT2 in the presence of 50 nM ISO. A) Representative Ca²⁺ transients recorded from isolated cardiac myocytes before (black) and after (red) 100 nM GS967 at CL = 2000 ms. B) Ca²⁺ transient decay at 0.5 Hz. (τ₇₅=523.8±35.0 vs. 300.7±19.3 for 0 and 100 nM GS967 respectively. N=20 cells from 3 LQT2 hearts in each group, p < 0.05). C-D) Caffeine-induced Ca²⁺ release and its decay. GS967 (red) accelerated Ca²⁺ decay (τ₅₀ =1.051±0.179 vs. 0.540±0.070 s for 0 and 100 nM GS967 respectively. n=9 and 14 cells from 3 LQT2 hearts. p<0.05).

Previous studies by Kornyeyev et al.¹⁸ showed that blocking I_NaL with GS967 reduces intracellular [Na⁺] (Na_i), which increases the driving force of Na⁺ ions for NCX1, in turn accelerating Ca²⁺ extrusion. To verify that GS967 increases Ca²⁺ efflux through NCX1, Ca²⁺ release was elicited by 10 mM caffeine, and the decay rate of the caffeine-induced Ca²⁺ transient was measured as described in the Methods section. Panel C in Figure 4 shows representative traces of caffeine-elicited Ca²⁺ release. The amplitudes of caffeine-induced Ca²⁺ transients, i.e., the measure of SR Ca²⁺ content, were not statistically significantly (normalized caffeine transients = 97.9±7.3 vs. 121.0±11% for control and 100 nM GS967, p=0.07). However, the decay rate of caffeine transients was about twice as fast in the presence of GS967 (see Figure 4D), supporting the contention that accelerating NCX1 by GS967 can play an important role in the suppression of EADs in LQT2 rabbits.

INaL modulates APDs through altering Na_i

Our experimental data shows that I_NaL blockade slightly affects APDs at physiological heart rate, but largely affects APDs at slow heart rate during which EADs occur, associated with suppressing PVTs robustly through altering Ca²⁺ handling in LQT2 rabbits. We next examined I_NaL in LQT2 rabbits using patch clamp technique (see supplemental material).⁶⁴ Figure 5A and B shows representative traces of the GS967-sensitive current from isolated LQT2 myocytes during a depolarizing −20 mV pulse (panel A) under 50 nM isoproterenol and IV-curve of I_NaL.

Figure 5.

The effect of I_NaL on APDs in computer modeling. A) Representative traces of I_NaL by measuring GS967-sensitive current during a depolarizing 200 ms pulse to −20 mV from the holding potential −120 mV measured from LQT2 myocytes. B) The IV curve of I_NaL, which was fitted to equation (1). I_NaL has no significant difference between LMC and LQT2 when ISO was absent, and I_NaL was 32% greater in LQT2 than LMC when ISO was present (−0.32±0.11 vs −0.24±0.05 pA/pF, two way ANOVA, factor genotype, p=0.003). C) The effect of I_NaL on APDs at CL=350 ms in LQT2. Left panel: AP traces with g_NaL=0 (red), 0.01 mS/μF (blue). Right panel: APDs in the parameter space of g_NaL vs Na_i at CL=350 ms. The red circle and blue square correspond to the results on the left panel. D) AP traces and APDs of computer simulation at CL=2 sec in LQT2 myocytes without ISO.

We then used computer modeling to investigate the influence of I_NaL on action potential dynamics. We developed an I_NaL equation²⁵ similar to the Na channel model, with parameters chosen to fit the patch clamp data in panel B. The simulation results show that the effect of I_NaL on APDs is greatly dependent on Na_i modulation. When the pacing cycle length is at 350 ms, the steady state Na_i is around 14.5 mM and APDs are similar for g_NaL = 0.01 mS/μF vs. 0, masking the effect of additional depolarizing currents by I_NaL (panel C) as seen experimentally in Table 1 and Figure 1. When the pacing cycle length is set to 2 sec, blockade of I_NaL (by reducing g_NaL to 0) shortens APDs (panel D) as seen experimentally Figure 3D. These computational modeling results demonstrate the important roles of I_NaL regulating APD through modulating Na_i. Similar behaviors were also observed in LMC myocytes (Figure S4 in the Supplemental Material).

I_NaL blockade prevents EADs through modulating Na_i and I_NCX

To dissect the two potential contributions of I_NaL to EAD formation through Na_i and additional depolarizing current, we implemented computer simulations with fixed Na_i values, which are chosen from steady-states of simulations where Na_i is unclamped (Figure S5 in the Supplemental Material). Figure 6A shows EAD formation in the parameter space of g_NaL vs. Na_i. Multiple EADs were observed when g_NaL is greater than 0.007 mS/μF in the range of 7 ~ 12 mM Na_i. The voltage traces corresponding to the squares and circles in Figure 6A are shown in Figure 6B, where the squares and thick lines represent normal I_NaL in LQT2 at different Na_i levels, and the circles and thin lines represent fully blocked I_NaL (100 nM GS967). When Na_i was fixed at 10.75 mM, the blockade of I_NaL eliminates EADs (blue thin line). The complete abolishment of EADs can also be obtained by reducing Na_i to 5.34 mM (red thick line) or the combination of blocking I_NaL and reducing Na_i. Note that when Na_i is reduced to 5.34 mM, the plateau V_m is elevated (blue vs. red lines in panel B) due to changes in I_NCX. Panel D shows the outward NCX current (reverse mode, blue lines) at 10.75 mM Na_i. I_NCX becomes inward current (forward mode, red lines) throughout plateau when Na_i is reduced to 5.34 mM, which elevates V_m during the initial AP plateau phase, causing rapid activation of I_Ks (panel E, red lines) to repolarize APs quickly and prevent EADs. This computer modeling study emphasizes the important roles of I_NaL in Na_i homeostasis: the blockade of I_NaL alleviates Na_i overload in addition to depolarizing currents during AP plateau phase. The same mechanism was observed when I_NaL was reduced to the LMC level by reducing α_NaL from 1 to 0 (Fig. S6 in the Supplemental Material).

Figure 6.

Computer modeling study of EADs reproduces I_NaL-dependent EADs in LQT2 myocytes with 50 nM ISO. A) EAD formation in the parameter space of g_NaL vs. Na_i. Representative traces corresponding to the squares and circles are shown in panels B-F. The Na_i values are chosen from steady states for different α_NaL values. The pink region represents infinite plateaus. B) V_m traces for different g_NaL and Na_i. Both lowering g_NaL from 0.01 mS/μF (thick blue line) and reducing Na_i from 10.75 mM to 5.34 mM eliminate EADs. C-F) Similar to panel B, I_Na, I_NCX, I_Ks, and I_Ca during pacing. Lowering Na_i to 5.34 mM (blue lines) increases forward mode I_NCX, which elevates V_m during plateau and largely activates I_Ks, resulting in rapid repolarization without EADs.

Discussion

Targeting I_NaL to accelerate repolarization and suppress EADs has been tested in many different animal models of cardiac diseases.^{5–8, 13, 22–28} In the current study, we tested GS967, a selective I_NaL blocker, to explore its anti-arrhythmic potential in the transgenic rabbit model of LQT2. GS967 markedly reduced the incidence of EADs and PVTs in LQT2 rabbit hearts by increasing Ca²⁺ efflux. These anti-arrhythmic effects were accompanied by minimal changes in APD, CV, or their restitution kinetics, revealing the important role of I_NaL in LQT-related arrhythmias and inhibition of I_NaL as a potential therapeutic target for LQT2.

Proarrhythmic role of I_NaL in LQT2 rabbits

The proarrhythmic role of I_NaL has been implicated in many pathological conditions in which I_NaL is enhanced such as in the ischemic and/or failing heart and LQT3.^{7, 8, 11, 28} I_NaL is also thought to influence V_m and EAD genesis when repolarization reserve such as I_Kr is reduced.^{24, 29} Ranolazine, an antianginal drug, is known to block I_NaL preferentially relative to the peak of I_Na and effectively reduce arrhythmias.⁵ However, the efficacy of ranolazine has been interpreted with caution due to a) non-specific blockade such as I_Kr^{24, 30}, b) use-dependent blockade of peak I_Na³⁰ and c) anti-adrenergic effect.³¹. Studies with the more-selective I_NaL blocker GS967 showed promising results in preventing arrhythmia with negligible effects on peak I_Na and I_Kr.^{13, 26} In the present study testing GS967 on a transgenic rabbit model of LQT2, the results clearly demonstrated that I_NaL is an important contributor to triggering EADs and PVTs which can be effectively abolished by selective inhibition of I_NaL with GS967.

Despite effective suppression of EADs and PVTs by GS967, it is worth noting that in a few cases, the PVTs morphed into transient monomorphic VTs caused by transient single rotors in the field of view (Figure 3G and see supplemental movie S2). This phenomenon is most likely caused by GS967 suppression of EADs (and thereby multi-focal activity) allowing single reentry formation. In the structurally normal heart such as in LQT2 rabbits, this single rotor moved out of boundary to terminate the monomorphic VTs. Further studies are required to evaluate the advantages and risks for inhibiting I_NaL in other cardiac diseases.

Modulation of APDs by I_NaL through Na_i

The minimal effect on APD under basal conditions and CV restitution is a major advantage of GS967. Earlier studies in rabbits reported that GS967 did not alter QRS duration or activation time in the wide range of physiological heart rates.^{26, 27} Our further examinations using S1S2 and the ramp pacing protocol revealed that APD and APD dispersion were not affected by GS967. Previous experiments in drug-induced LQT2 models using E4031 demonstrated that inhibition of I_NaL by ranolazine or GS967 shortens APDs, and this phenomenon was associated with suppression of EADs.²⁹ GS967 also shortened APDs in transgenic LQT2 rabbit hearts but only during slow heart rate and ISO perfusion (Figure 3D) while APDs were not affected at normal heart rate. This heart-rate dependent APD shortening by GS967 in LQT2 rabbit hearts is most likely due to complex APD modulation by I_NaL through modulating Na_i. Previous single cell and computer modeling studies^{32, 33} reported that the elevated Na_i drives I_NCX and I_NaK in the outward direction during AP plateau phase to lower the AP plateau. Therefore, the overall effect of I_NaL blockade on APD will be the result of two competing effects, direct reduction of I_NaL and also indirect increase of I_NCX and I_NaK during the AP plateau. Since Na_i increases with heart rate, the effect of Na_i on APDs becomes greater at fast heart rate, compensating I_NaL’s direct impact on APDs. Therefore, I_NaL blockade by GS967 may have resulted in a negligible effect on APDs in LQT2 rabbit hearts unless the heart rate is slow as shown experimentally in Figure 1 and by computer simulation in Figure 5.

GS967 suppresses EADs via normalizing defective Na⁺ and Ca²⁺ homeostasis

Abnormal Ca²⁺ cycling has been suspected as a major contributing factor for triggering EADs in LQTS.^{34, 35} The proposed mechanism is that the prolonged APD in LQTS increases Ca²⁺ influx through L-type Ca²⁺ channels, resulting in increased intracellular Ca²⁺. Animal models of LQT2 with I_Kr blockers showed that Ca²⁺ elevation often preceded the upstrokes of EADs³⁵ and depletion of SR Ca²⁺ suppressed EADs, supporting the role of Ca²⁺ in triggering EADs. Importantly, our previous study showed that unlike in drug-induced LQT2 animal models, cardiomyocytes from transgenic LQT2 rabbits exhibit no Ca²⁺ overload in SR.¹⁷ In this transgenic LQT2 model, Ca²⁺ transient amplitude is decreased in parallel with a significant decrease in SR Ca²⁺ content due to the increased SR Ca²⁺ loss mediated by hyperactive RyR2s. The increased propensity to generate Ca²⁺-dependent EADs in hereditary LQT2 has been linked to accelerated reactivation of RyR2 clusters after Ca²⁺ transient peak that via NCX1 contribute to prolonged maintenance of AP plateau at the voltage range suitable for reactivation of L-type Ca²⁺ channels.¹⁷

The importance of NCX1 in EAD genesis has been confirmed by numerous studies.^{36, 37} We previously reported that blocking NCX1 with SEA0400 abolished EADs in LQT2 myocytes,¹⁷ supporting its important role in EAD genesis. The activity of the NCX1 depends on parameters such as the transmembrane concentration gradients of Na⁺ and Ca²⁺, as well as the membrane potential. NCX1-mediated removal of [Ca²⁺] from the dyadic space has been shown to effectively suppress localized spontaneous Ca²⁺ releases from the RyR2 clusters.³⁸ The present study indicates that the effective suppression of EADs by GS967 is associated with the accelerated NCX1-mediated extrusion of intracellular Ca²⁺ from dyadic space during the peak and recovery phase of Ca²⁺ transients.

The potential mechanism underlying the protective effect of GS967 may be related to the role of I_NaL in modulating Na_i homeostasis,^{13, 18} thereby accelerating NCX1. The enhanced I_NaL in cardiomyopathy has been linked to Na_i overload, leading cytosolic Ca²⁺overload and afterdepolarizations that can be abolished with ranolazine.³⁹ A previous study by Kornyeyev et al.¹⁸ showed that prolongation of APD by E4031 slightly increases Na_i (~ 1 mM change from 4 to 5 mM in rabbit myocytes). Since small changes in Na_i can greatly alter NCX1 kinetics, it is possible that I_NaL blockade by GS967 is sufficient to lower Na_i, which accelerates forward mode of NCX1 to increase AP plateau as well as activate I_Ks, thereby suppressing EADs. Further studies are needed to verify the role of I_NaL in modulating Na_i homeostasis and NCX1 kinetics. Importantly, our transgenic LQT2 rabbit model shows major remodeling of cardiac excitation: 1) prolongation of APD in LQT2 rabbits is much smaller compared to drug-induced LQT2 model (18% vs. 40%), which is associated with downregulation of I_Ca and I_Ks,¹⁴ 2) E4031 perfusion in isolated rabbit hearts triggers EADs and PVTs without sympathetic stimulation while transgenic LQT2 rabbits (similar to humans) did not show EADs and PVTs at baseline and required ISO or low K⁺ to trigger EADs or PVTs,^{15, 40} 3) LQT2 rabbits have remodeling of Ca²⁺ handling proteins.¹⁷ The reduction in protein phosphatase 2 tethering to RyR complex in LQT2 rabbit myocytes result in hyperactive RyRs (leaky RyRs), leading to unique EAD genesis through sustained Ca²⁺ release during AP plateau phase and may potentially underlie diastolic dysfunction. It is possible that similar loss of PP2a¹⁷ due to posttranslational modifications which is known to increase I_NaL⁴¹ may have altered I_NaL to have higher efficacy of suppressing PVTs by GS967 through Na_i regulation. Since multiple pathways can impact I_NaL such as glycosylation, phosphorylation on Tyr residues, Arg methylation, and mechanosensitivity (see the review by⁴²), it is difficult to identify a single pathway that underlies I_NaL remodeling in LQT2 rabbits. Overall, these unique features of transgenic LQT2 rabbits may have caused EAD formation and PVT induction more sensitive to Na_i and Ca²⁺ homeostasis shown in our computational modeling.

Conclusion

We found that the selective I_NaL blocker GS967 abolished EAD/PVT events in LQT2 rabbits. The inhibition of I_NaL by GS967 normalized Ca²⁺ homeostasis by accelerating cytosolic Ca²⁺ extrusion via NCX1 at the single-cell level. Our result suggests that I_NaL is an important current modulating Ca²⁺ cycling and bi-excitability, and a potential therapeutic target to prevent triggered activity and PVT events in LQTS. However, transient reentry formation during suppression of focal activity by GS967 could be proarrhythmic under certain circumstances

What is known?

• Long-QT syndrome type 2 (LQT2), an inherited disease caused by loss-of-function mutations in KCNH2 (hERG) underlying rapidly activating delayed rectifier K+ current (IKr), is associated with sudden cardiac arrest triggered by early afterdepolarizations (EADs).

• The late Na+ current (INaL) could also, along with the L-type Ca2+ current (ICaL), promote EADs by providing an additional depolarizing force during the plateau phase of action potentials (APs).

What the study adds?

• It tests whether the selective INaL blocker GS967 can prevent EADs and polymorphic ventricular tachycardia (PVT) in a transgenic rabbit model of LQT2.

• It shows that the inhibition of INaL by GS967 lessens [Na+]i overload. Decreased [Na+]i accelerates the forward mode of the Na+/Ca2+ exchanger, which elevates action potential plateau Vm and then activates the slow activating delayed rectifier K+ current (IKs), thereby suppressing EADs and PVT in LQT2 rabbits.

• It identifies important proarrhythmic roles of INaL via regulating [Na+]i and [Ca2+]i homeostasis in the LQT2 setting.

Nonstandard Abbreviations and Acronyms:

AP

Action potential

APD

action potential duration

Ca_i

intracellular Ca²⁺ concentration

CL

Cycle length

CV

Conduction Velocity

EAD

Early afterdepolarizations

hERG

human ether-a-go-go gene

I_CaL

L-type Ca²⁺ current

ICD

Implantable Cardioverter Defibrillator

I_Kr

the rapidly activating delayed rectifying K⁺ current

I_Na

Na⁺ current

I_NaK

Na⁺/K⁺ ATPase current

I_NaL

Late Na⁺ current

I_NCX

Na⁺/Ca²⁺ exchanger

ISO

Isoproterenol

LMC

Littermate control

LQT2

Long QT Syndrome Type 2

LQTS

Long QT Syndrome

Na_i

Intracellular Na⁺ concentration

PVT

Polymorphic Ventricular Tachycardia

RyR

Ryanodine Receptor

SCD

Sudden cardiac death

VF

Ventricular Fibrillation