Tetrodotoxin-Sensitive Neuronal-Type Na+ Channels: A Novel and Druggable Target for Prevention of Atrial Fibrillation

Abstract

Background Atrial fibrillation (AF) is a comorbidity associated with heart failure and catecholaminergic polymorphic ventricular tachycardia. Despite the Ca²⁺-dependent nature of both of these pathologies, AF often responds to Na⁺ channel blockers. We investigated how targeting interdependent Na⁺/Ca²⁺ dysregulation might prevent focal activity and control AF. Methods and Results We studied AF in 2 models of Ca²⁺-dependent disorders, a murine model of catecholaminergic polymorphic ventricular tachycardia and a canine model of chronic tachypacing-induced heart failure. Imaging studies revealed close association of neuronal-type Na⁺ channels (nNa_v) with ryanodine receptors and Na⁺/Ca²⁺ exchanger. Catecholamine stimulation induced cellular and in vivo atrial arrhythmias in wild-type mice only during pharmacological augmentation of nNa_v activity. In contrast, catecholamine stimulation alone was sufficient to elicit atrial arrhythmias in catecholaminergic polymorphic ventricular tachycardia mice and failing canine atria. Importantly, these were abolished by acute nNa_v inhibition (tetrodotoxin or riluzole) implicating Na⁺/Ca²⁺ dysregulation in AF. These findings were then tested in 2 nonrandomized retrospective cohorts: an amyotrophic lateral sclerosis clinic and an academic medical center. Riluzole-treated patients adjusted for baseline characteristics evidenced significantly lower incidence of arrhythmias including new-onset AF, supporting the preclinical results. Conclusions These data suggest that nNa_Vs mediate Na⁺-Ca²⁺ crosstalk within nanodomains containing Ca²⁺ release machinery and, thereby, contribute to AF triggers. Disruption of this mechanism by nNa_v inhibition can effectively prevent AF arising from diverse causes.

Keywords: atrial arrhythmias; atrial fibrillation; cardiac arrhythmias; neuronal‐type Na+ channel blockade.

Figure 1. Inhibition of tetrodotoxin (TTX)‐sensitive neuronal Na_Vs (nNa_vs) is sufficient to prevent induction of aberrant, repetitive Ca²⁺ oscillations in R33Q atrial myocytes.

A, Representative examples of the line scan images and corresponding Ca²⁺ transients recorded in field‐stimulated R33Q atrial cardiomyocytes paced at 0.5 Hz and loaded with the Ca²⁺ indicator, Fluo‐3 AM. Cells were treated with isoproterenol (ISO, 100 nmol/L) and subsequently TTX (100 nmol/L; green arrow indicates time when TTX was added) was rapidly applied. β‐Adrenergic stimulation with ISO promoted aberrant, repetitive Ca²⁺ oscillations (median pacing frequency was 2 Hz with the range of 0.5 to 4 Hz), while TTX (100 nmol/L) significantly decreased their incidence. Ca²⁺ oscillations were induced after drug washout (red arrow; number of cells tested depicted under the corresponding bars, N=8 animals for ISO and ISO‐TTX, N=6 animals for ISO‐washout, respectively; ***P<0.0001 McNemar test for ISO vs ISO‐TTX, P<0.0001 Fisher exact test for ISO‐TTX vs ISO‐washout). B, Treatment of ISO (100 nmol/L)‐exposed R33Q atrial myocytes with 10 μmol/L riluzole (Ril) significantly reduced the incidence of aberrant, repetitive Ca²⁺ oscillations, an effect that was washable (median pacing frequency was 2 Hz with the range of 0.5 to 7 Hz; number of cells tested depicted under the corresponding bars, N=10 animals for ISO and ISO‐Ril, N=8 animals for ISO‐washout, respectively; ***P<0.0001 McNemar test for ISO vs ISO‐Ril, P<0.0001 Fisher exact test for ISO‐Ril vs ISO‐washout).

Figure 2. Effect of neuronal Nav (nNav) blockade with riluzole (Ril) on isoproterenol (ISO)‐promoted inward Na⁺ currents (I_Na) in R33Q atrial myocytes.

A, Representative I_Na (top) were elicited by a protocol presented in the inset before (black) or after addition of ISO (100 nmol/L; red) or ISO (100 nmol/L)+Ril (10 μmol/L; purple). (Bottom) Corresponding current‐voltage relationship from control, ISO, and ISO+Ril (n=10, 9, and 7 cells from 5, 6, and 5 mice, respectively). Addition of ISO+Ril reduced I_Na density relative to ISO alone (peak I_Na at −40 mV of −37.1±3.2 vs −22.4±3.9 pA/pF for ISO and ISO+Ril, respectively; P=0.0102 ANOVA, P=0.0418 for ISO vs ISO+Ril). B, Representative traces of persistent I_Na elicited using the protocol shown in the inset were recorded before (black) or after addition of ISO (100 nmol/L; red) or ISO+Ril (10 μmol/L; purple). ISO enhanced persistent I_Na in R33Q cardiomyocytes, while Ril suppressed β‐adrenergic–mediated increase in persistent I_Na (n=9, 8, and 8 cells from 3, 3, and 4 mice for R33Q, ISO, and ISO‐Ril, respectively; P<0.0001 ANOVA, ***P=0.0004 for persistent I_Na integral and ***P<0.0001 for normalized late I_Na). Summary data are presented as persistent I_Na integral Amp‐ms/F (AmsF⁻¹; left) or normalized late I_Na (%; right), which were measured by either integrating I_Na between 50 and 450 ms (left) or normalizing mean persistent INa recorded between 250 and 450 ms by peak current generated by a step to ‐40 mV.

Figure 3. The extent of late inward Na⁺ currents (I_Na) inhibition corresponds to prevention of aberrant, repetitive Ca²⁺ oscillations in R33Q atrial myocytes.

A, (Left) I_Na obtained by step protocol illustrated in 1‐second intervals. (Right) In isoproterenol (ISO; 100 nmol/L)‐treated R33Q atrial myocytes, addition of ranolazine (Ran; 10 μmol/L, blue trace and bar; n=5 from N=4 animals) or R‐propafenone (R‐prop; 300 nmol/L, orange trace and bar; n=4 from N=3 animals) significantly reduced peak I_Na (P=0.0016 ANOVA, *P=0.0418 for ISO vs ISO+Ran and *P=0.0246 for ISO vs ISO+R‐prop) relative to ISO alone (n=7 from N=5 animals). Notably, there was no difference in peak I_Na reduction between the groups. B, (Left) Representative persistent I_Na elicited using the protocol shown in the inset. (Right) ISO (100 nmol/L) increased persistent I_Na, while Ran and R‐prop reduced it (n=13, 13, 7, and 7 cells from 6, 6, 6 and 4 mice for R33Q, ISO, ISO+Ran, and ISO+R‐prop, respectively; P<0.0001 ANOVA, ***P<0.0001, **P=0.0082, and *P=0.0480). Notably, R‐prop reduced ISO‐induced persistent I_Na to a greater extent than Ran (P=0.0480). C, Treatment of ISO (100 nmol/L)‐exposed R33Q atrial myocytes with Ran (10 μmol/L) significantly reduced the incidence of aberrant, repetitive Ca²⁺ oscillations, which was washable (number of cells tested depicted under the corresponding bars, N=5 animals; ***P=0.0009 McNemar test for ISO vs ISO‐Ran and for ISO‐Ran vs ISO‐washout). D, R‐prop abolished aberrant, repetitive Ca²⁺ oscillations, an effect that was only partially washable only after 10 minutes. (Number of cells tested depicted under the corresponding bars, N=4 animals; ***P=0.0005 McNemar test for ISO vs ISO–R‐prop, *P=0.0455 for ISO–R‐propafenone vs ISO‐washout).

Figure 4. Neuronal Na⁺ channel (nNa_V) and ryanodine receptor 2 (RyR2) colocalize to the same discrete subcellular regions.

Representative confocal micrographs of myocytes isolated from R33Q mice labeled for (A) RyR2 (red) and (B) Na⁺/Ca²⁺ exchange (NCX; red) with various Na⁺ channel (Na_v) isoforms (Na_v1.x, green). These often resulted in an overlap between the immunofluorescent signals (yellow) when overlaid. (Right) Close‐up views of regions highlighted by dashed white boxes. (Bottom) Representative fluorescent proximity ligation assay signal for RyR2 (A) and NCX (B) with different nNa_v isoforms (Na_V1.x).

Figure 5. Augmentation of tetrodotoxin (TTX)‐sensitive neuronal Nav (nNav) is sufficient to initiate aberrant, repetitive Ca²⁺ oscillations in wild‐type (WT) atrial myocytes.

A, Representative traces of persistent inward Na⁺ currents (I_Na) recorded in WT atrial cardiomyocytes before (black) and after (red) exposure to β‐pompilidotoxin (β‐PMTX; 40 μmol/L). β‐PMTX increased persistent I_Na relative to control (n=11 and 10 cells from 3 mice, respectively; ***P<0.0001 Wilcoxon rank sum test). B, (Left) Representative line scan images obtained from WT atrial cardiomyocytes exposed to isoproterenol (ISO; 100 nmol/L) and paced at 1 Hz. (Right) Rapid application of β‐PMTX (40 μmol/L) induced aberrant, repetitive Ca²⁺ oscillations. (Median pacing frequency was 0.5 Hz with the range of 0.5 to 2 Hz; number of cells tested depicted under the corresponding bars, from N=3 mice; ***P<0.0001 Fisher exact test for ISO vs ISO–β‐PMTX, ***P<0.0001 Fisher exact test for ISO–β‐PMTX vs ISO‐washout).

Figure 6. Modulation of tetrodotoxin (TTX)‐sensitive neuronal Nav (nNav) channel correspondingly modulates atrial arrhythmias in mice.

A, Simultaneous surface ECG (lead II) and intracardiac atrial electrograms with frequent, rapid P waves and irregular RR intervals suggestive of atrial arrhythmia such as atrial flutter and atrial fibrillation in R33Q mice after burst pacing. Pretreatment with riluzole (Ril; 15 mg/kg IP), targeting plasma concentrations of ~10 μmol/L,33 reduced the atrial arrhythmia inducibility (n=7 mice; **P=0.0160 Wilcoxon signed rank test). B, Representative surface ECG recordings of wild‐type (WT) mice treated (top, red ECG) or untreated (bottom, black ECG) with β‐pompilidotoxin (β‐PMTX; 40 mg/kg IP) and exposed to catecholamine challenge with epinephrine (Epi, 1.5 mg/kg) and caffeine (Caff, 120 mg/kg). Since increased heart rate has been linked to reduced arrhythmia inducibility in calsequestrin null mice, and WT mice show higher heart rate relative to calsequestrin null mice,32 all WT animals were pretreated with ivabradine (3 mg/kg) for 10 minutes before any intervention. Epi+Caff challenge during β‐PMTX exposure precipitated repetitive P waves and irregular RR intervals suggestive of atrial arrhythmia in over 50% of WT mice, which is a 3‐fold increase relative to β‐PMTX–untreated mice (number of mice tested and those positive for atrial arrhythmias depicted under the corresponding bars; *P=0.0410 Fisher exact test).

Figure 7. Riluzole (Ril) reduces enhanced, persistent inward Na⁺ currents (I_Na) and prevents induction of aberrant, repetitive Ca²⁺ oscillations in canine heart failure (HF) atrial myocytes.

A, Representative traces of persistent I_Na integral elicited using the protocol shown in the inset. Recordings were made in control (top) and failing (bottom) atrial myocytes before (black) and after exposure to isoproterenol (ISO; 100 nmol/L, red) and after treatment with Ril (10 μmol/L, purple). At baseline, atrial cardiomyocytes from failing hearts showed a larger persistent I_Na integral relative to control. ISO (100 nmol/L) enhanced persistent I_Na only in control cardiomyocytes. Ril reduced persistent I_Na integral in both control and failing atrial myocytes. B, Summary data are presented as persistent I_Na integral Amp‐ms/F (AmsF⁻¹), which was measured by integrating I_Na between 50 and 450 ms (n=7 and 9 cells from 5 control and 3 failing dogs; P<0.0001 Kruskal–Wallis test; *P=0.0126 for control vs ISO, *P=0.0209 for control vs ISO+Ril, *P=0.0268 for control‐ISO vs HF‐ISO, ***P<0.0001). C, Representative examples of the line‐scan images and corresponding Ca²⁺ transients recorded in canine HF atrial cardiomyocytes loaded with Ca²⁺ indicator, Fluo‐3 AM, and paced at 0.5 Hz with field stimulation. (Top) Cells were treated with ISO (100 nmol/L) and subsequently Ril (10 μmol/L; purple bar indicates time when Ril was added) was rapidly applied. (Bottom) Resumption of field stimulation failed to induce Ca²⁺ oscillations during concomitant exposure to ISO and Ril; however, washout of Ril resulted in their reinitiation. D, Ril significantly reduced the incidence of aberrant, repetitive Ca²⁺ oscillations, an effect that was washable (median pacing frequency was 0.5 Hz with the range of 0.5 to 1 Hz; number of cells tested depicted under the corresponding bars, N=3 animals for ISO and ISO‐Ril, N=2 animals for ISO‐washout, respectively; **P=0.0009 McNemar test for ISO vs ISO‐Ril, P<0.0001 Fisher exact test for ISO‐Ril vs ISO‐washout incidence; ***P=0.0005 Friedman rank sum test for Ca²⁺ oscillations frequency).

Figure 8. Riluzole prevents cardiac arrhythmias in patients with amyotrophic lateral sclerosis (ALS).

Two retrospective cohorts of ALS one exposed to riluzole vs no riluzole (controls), were compared by Cox proportional hazard models. The time‐to‐first composite arrhythmic events was analyzed using Kaplan–Meier production limit estimator. A, Overall cohort, (B) US cohort. HR indicates hazard ratio.