Neuronal Na⁺ channel blockade suppresses arrhythmogenic diastolic Ca²⁺ release

Abstract

Aims

Sudden death resulting from cardiac arrhythmias is the most common consequence of cardiac disease. Certain arrhythmias caused by abnormal impulse formation including catecholaminergic polymorphic ventricular tachycardia (CPVT) are associated with delayed afterdepolarizations resulting from diastolic Ca²⁺ release (DCR) from the sarcoplasmic reticulum (SR). Despite high response of CPVT to agents directly affecting Ca²⁺ cycling, the incidence of refractory cases is still significant. Surprisingly, these patients often respond to treatment with Na⁺ channel blockers. However, the relationship between Na⁺ influx and disturbances in Ca²⁺ handling immediately preceding arrhythmias in CPVT remains poorly understood and is the object of this study.

Methods and results

We performed optical Ca²⁺ and membrane potential imaging in ventricular myocytes and intact cardiac muscles as well as surface ECGs on a CPVT mouse model with a mutation in cardiac calsequestrin. We demonstrate that a subpopulation of Na⁺ channels (neuronal Na⁺ channels; nNa_v) colocalize with ryanodine receptor Ca²⁺ release channels (RyR2). Disruption of the crosstalk between nNa_v and RyR2 by nNa_v blockade with riluzole reduced and also desynchronized DCR in isolated cardiomyocytes and in intact cardiac tissue. Such desynchronization of DCR on cellular and tissue level translated into decreased arrhythmias in CPVT mice.

Conclusions

Thus, our study offers the first evidence that nNa_v contribute to arrhythmogenic DCR, thereby providing a conceptual basis for mechanism-based antiarrhythmic therapy.

1. Introduction

Sudden arrhythmic death is a leading cause of mortality in the USA.¹ Arrhythmias caused by abnormal impulse generation occur both in patients with structural heart disease,² as well as in seemingly healthy individuals.³ In the latter, fatal arrhythmias are often associated with aberrant diastolic Ca²⁺ release (DCR) during exercise or stress-induced catecholamine surge.³ These are called Catecholaminergic Polymorphic Ventricular Tachycardia (CPVT).⁴ Recent evidence suggests that CPVT is related to mutations in proteins comprising the ryanodine receptor Ca²⁺ release channel (RyR2) complex, including the RyR2 itself and the sarcoplasmic reticulum (SR) Ca²⁺-binding protein calsequestrin (CASQ2).^5,6 These mutations appear to alter the ability of RyR2 to become refractory following systolic Ca²⁺ release, allowing it to remain open during diastole and leak SR Ca²⁺.^7–10 This can result in DCR, which in turn can be translated into membrane depolarization via electrogenic Na⁺/Ca²⁺ exchange (NCX) resulting in triggered ventricular arrhythmia.^8,10–12

Despite high response of CPVT to combination of Ca²⁺ channel- and β-blockers, the incidence of refractory cases is still significant.¹³ Surprisingly, based on the Ca²⁺-dependent mechanisms, such patients often respond to treatment with Na⁺ channel blockade.¹³ However, the precise contribution of Na⁺ influx to aberrant DCR and arrhythmogenesis is unclear. Data presented herein strongly suggest that a subpopulation of Na⁺ channels, neuronal Na⁺ channels (nNa_v), colocalize with RyR2. The electrogenic NCX couples Na⁺ fluxes in these compartments to Ca²⁺ influx and efflux as well as membrane potential. Disruption of coupling between nNa_v and RyR2 by nNa_v blockade reduced and desynchronized DCR in isolated cardiomyocytes as well as in tissue and decreased arrhythmias in CPVT mice. Thus, our study offers the first support for the concept that nNa_v contribute to the genesis and synchronization of arrhythmogenic DCR and identifies nNa_v as a target for mechanism-based antiarrhythmic therapy.

2. Methods

All animal procedures were approved by The Ohio State University Institutional Animal Care and Use Committee and conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 2011).

2.1. Myocyte isolation, confocal Ca²⁺ imaging, Na⁺ current, and membrane potential recordings

Ventricular myocytes were obtained by enzymatic isolation from 3- to 5-month-old R33Q (in C57BL/6 background) and age-matched C57BL/6 WT (Jackson Jax Lab) male mice. Mice were anaesthetized with isoflurane, and once a deep level of anaesthesia was reached, confirmed by lack of response to noxious stimuli, the heart was rapidly removed and perfused via a Langendorff as previously described.⁸ Single ventricular myocytes were isolated using Liberase Blendzymes (Roche, Applied Science, IN, USA). Whole-cell patch clamp recordings of action potentials were performed with an Axopatch 200B amplifier using external solution that contained (in mM): 140 NaCl, 5.4 KCl, 2.0 CaCl₂, 0.5 MgCl₂, 10 HEPES, and 5.6 glucose (pH 7.4). Patch pipettes were filled with a solution that contained (in mM): 90 K-aspartate, 50 KCl, 5 MgATP, 5 NaCl, 1 MgCl₂, 0.1 Tris GTP, 10 HEPES, and 0.1 Fluo-4 K-salt (Molecular Probes, Eugene, OR, USA) (pH 7.2). Sodium currents (I_Na) were recorded using internal solution containing (in mM): 10 NaCl, 20 TEACl, 123 CsCl, 1 MgCl₂, 0.1 Tris GTP, 5 MgATP, 10 HEPES, and 10 BAPTA (pH 7.2). The extracellular bathing solution contained (in mM): 10 NaCl, 130 TEACl, 4 CsCl, 0.4 CaCl₂, 2 MgCl₂, 0.05 CdCl₂, 10 HEPES, and 10 glucose; pH was maintained at 7.4 with CsOH. Whole-cell capacitance and series resistance compensation (≥60%) was applied along with leak subtraction. Electrical field stimulation experiments were performed using the following external solution (in mM): 140 NaCl, 5.4 KCl, 2.0 CaCl₂, 0.5 MgCl₂, 10 HEPES, and 5.6 glucose (pH 7.4). For Ca²⁺ spark recordings in permeabilized myocytes, the cardiac myocytes were permeabilized with saponin (0.01% for 20–30 s). The internal solution contained (in mM): 120 K-aspartate, 20 KCl, 0.81 MgCl₂, 1 KH₂PO₄, 0.5 EGTA (free [Ca²⁺] ∼50 nm), 3 MgATP, 10 phosphocreatine, 0.03 Fluo-4 K-salt, 20 HEPES (pH 7.2), and 5 U/mL creatine phosphokinase. To assess the SR Ca²⁺ load, 20 mM caffeine was applied at the end of the experiments. Intracellular Ca²⁺ cycling was monitored by a Nikon A1 laser scanning confocal microscope equipped with a ×60 1.4 NA oil objective. For intact myocytes, we used the cytosolic Ca²⁺-sensitive indicators Fluo-4 AM (Ca²⁺ sparks) and Fluo-4FF AM (Ca²⁺ waves). The fluorescent probes were excited with the 488 nm line of an argon laser, and emission was collected at 500–600 nm. Fluo-4/Fluo-4FF fluorescence was recorded in the line scan mode of the confocal microscope (0.207 µm per pixel, 1–5 ms per line). For Ca²⁺ wave recordings, myocytes were paced at 0.3 Hz using extracellular platinum electrodes to obtain a comprehensive distribution of the first DCR latencies. Any DCR event (i.e. wave, wavelet) that increased cell-wide fluorescence intensity above 10% of the signal generated by the preceding stimulated Ca²⁺ transient was included in the analysis. The fluorescence emitted was expressed as F/F₀, where F is the fluorescence at time t and F₀ represents the background signal. All experiments were performed at room temperature (26°C).

2.2. Cardiac muscles confocal Ca²⁺ and membrane potential imaging

Muscles for confocal Ca²⁺ and membrane potential imaging were prepared as previously described.⁸ Briefly, the hearts were removed and placed in a modified Krebs–Henseleit (KH) buffer, containing (in mmol/L) 120 NaCl, 5 KCl, 2MgSO₄, 1.2 NaH₂PO₄, 20 NaHCO₃, 0.25 CaCl₂, and 10 glucose (pH 7.4), equilibrated with 95% O₂–5% CO₂. Twenty millimolar 2,3-butanedione monoxime (BDM) was also added to the dissection buffer to minimize cutting injury. The recording chamber was placed on the stage of an Olympus FluoView 1000 laser scanning confocal microscope, equipped with a 5 mW He–Ne laser and a UPLSAPO 60X water objective (NA 1.20, working distance 0.28 mm). Muscles were loaded with the cytosolic Ca²⁺ indicator Rhod-2 AM for 45 min and di-4-ANBDQBS (Richard D. Berlin Center for Cell Analysis and Modeling, University of Connecticut Health Center, USA) for 10 min. The muscles were superfused with KH buffer containing 5 mM BDM and 10–20 µM blebbistatin to stop cell contraction. The Ca²⁺ concentration was also raised in the KH buffer to 1.5 mM. Line-scan images were acquired at the rate of 8 µs/pixel. The fluorescence emitted was expressed as F/F₀, where F is the fluorescence at time t and F₀ represents the background signal. All experiments were performed at room temperature.

2.3. Confocal microscopy of immunolabelled myocytes and image analysis

Isolated ventricular myocytes from R33Q hearts were plated on laminin-coated glass coverslips, fixed with 4% paraformaldehyde (5 min), permeabilized with 0.1% Triton X-100, and washed with PBS. Endogenous immunoglobulin was blocked using a mouse-on-mouse blocking reagent (M.O.M. kit; Vector Laboratories, Burlingame, CA, USA) for 1 h at room temperature and subsequently incubated with primary antibodies overnight at 4°C. After washing, goat secondary antibodies (anti-mouse and anti-rabbit) conjugated to Alexa Fluor (488, 594; Life Technologies, Grand Island, NY, USA) were added for 1 h. Coverslips were mounted to using ProLong Gold Anti-Fade Mounting Kit (Molecular Probes).

Immunolabelled myocytes were imaged on a TCS SP8 laser scanning confocal microscope equipped with a ×63/1.4 numerical aperture oil objective (Leica, Buffalo Grove, IL, USA). Custom software written in Matlab (Mathworks, Natick, MA) was used to analyse images for colocalization and cross-correlation as well as to perform fast Fourier transform (FFT) morphological analysis. Briefly, colocalization was assessed from plots of red vs. green intensity as previously described.¹⁴ Additionally, the cross-correlation coefficient between the intensities of red and green immunofluorescent signals was calculated both row-wise as well as for the averaged intensity profile over a whole myocyte. Further morphological analysis was performed by calculating the two-dimensional (2D) FFT for the red and green channels of each image. The amplitude of the 2D FFT for the red and green channels, each itself in the form of a 2D image with the same dimensions as the original image, was then overlaid. Periodicity of immunofluorescent signal was assessed from the location and amplitude of peaks on the FFTs and the relative periodicities of the red and green immunofluorescent signals assessed from the overlaid FFTs of the respective channels. The antibodies used were rabbit anti-Na_V1.1, 1.3, 1.6 (Alomone, Jerusalem, Israel), 1.5 (generous gift from Dr Peter Mohler), and mouse monoclonal anti-RyR2 (Pierce Antibodies, Rockford, IL, USA).

2.4. ECG recordings

ECG recordings were performed before and after epinephrine and caffeine challenge. Continuous ECG recordings were obtained from mice anaesthetized with isoflurane, at minimum effective concentration (1–1.5%). Subcutaneous needle electrodes were applied to the left, right upper limb, and right lower limb for ECG recording (PL3504 PowerLab 4/35, ADInstruments). After baseline recording (5 min.), each mouse received intraperitoneally vehicle (DMSO), riluzole (10 mg/kg), or SN-6 (40 mg/kg). After 5–10 min from the initial intervention, animals were exposed to an intraperitoneally epinephrine (Epi, 1.5 mg/kg) and caffeine (Caff, 120 mg/kg) challenge, and ECG recording continued for 10 min. ECG recordings were analysed using the LabChart 7.3 program (ADInstruments). Ventricular tachycardia (VT) was defined as three or more premature ectopies, while arrhythmia was defined as frequent ectopies, bigeminies, and/or VT.

2.5. Reagents

Unless otherwise stated, all chemicals were purchased from Sigma (St Louis, MO, USA), Torcis (Bristol, UK), or Alomone. Fluorescent dyes were purchased from either Molecular Probes or Richard D. Berlin Center for Cell Analysis and Modeling (University of Connecticut Health Center, USA).

2.6. Data analysis

Membrane potential and I_Na analysis were performed using pCLAMP9 software (Molecular Devices, Sunnyvale, CA, USA). Line scanning images of Ca²⁺ and membrane potential were normalized for baseline fluorescence.⁸ Ca²⁺ and membrane potential imaging data were processed using ImageJ and Origin software. Statistical analysis of the data was performed using a two-tailed Student's t-test for paired and unpaired data or a single factor ANOVA. The Šidák correction was applied to adjust for multiple comparisons. A Fisher's exact and a Mantel–Haenszel test were used to test differences in nominal data. All values are reported as means ± SEM unless otherwise noted. A value of P < 0.05 was considered statistically significant.

3. Results

3.1. Neuronal Na⁺ channel modulation regulates DCR

To determine whether nNa_v possess a unique action on intracellular Ca²⁺ handling that may serve as a substrate for triggered arrhythmias, we assessed the impact of nNa_v blockade on Ca²⁺ sparks and DCR in the form of Ca²⁺ waves in cardiomyocytes isolated from hearts expressing a CPVT-associated CASQ2 mutation (R33Q) and exposed to isoproterenol (Iso).¹⁵ In accord with previous observations,¹⁶ R33Q cardiomyocytes evidenced a high rate of SR Ca²⁺ leak that corresponded to high Ca²⁺ spark frequency and prolonged duration of release. Tetrodotoxin (TTX) at nanomolar concentration is known to selectively inhibit nNa_v without significantly affecting cardiac-type Na⁺ channels (Na_v1.5) or directly inhibiting RyR2.^17–20 Line-scan confocal Ca²⁺ imaging showed that 100 nM TTX significantly decreased the frequency of Iso-promoted Ca²⁺ sparks and DCR in resting and paced cardiomyocytes, respectively (Figure 1A–E). We also assessed DCR synchrony, a marker of Ca²⁺ release refractoriness.^8,10 TTX significantly reduced DCR synchrony, as demonstrated by a broadened distribution of DCR first latencies and an increase in the S.D. of latency (Figure 1F). Notably, TTX's antiarrhythmic effects were not accompanied by alterations in the SR Ca²⁺ content or Ca²⁺ transient amplitude (see Supplementary material online, Figure S1A–C), which is consistent with a previous report that utilized TTX concentrations of up to 5 µM.²¹ The sensitivity of DCR in R33Q ventricular cardiomyocytes to 100 nM TTX might have resulted from an adaptive up-regulation of nNa_v. To address such possibility, we assessed peak Na⁺ current (I_Na) in R33Q and WT myocytes. There was no significant difference in I_Na density between the two study cohorts (Figure 2A and B).

Figure 1.

Neuronal Na⁺ channel blockade disrupts Na⁺/Ca²⁺ signalling and suppresses DCR. (A) Representative examples of the line-scan images of Ca²⁺ sparks recorded in R33Q ventricular cardiomyocytes. Cells were treated with TTX (100 nM), riluzole (Ril, 10 µM), and SN6 (5 µM) (B) TTX (n = 1054 events from 47 cells isolated from four mice), Ril (n = 1675 events from 59 cells isolated from three mice), and SN6 (n = 2811 events from 84 cells isolated from three mice) significantly decreased Ca²⁺ spark frequency in Iso-treated R33Q myocytes (n = 2442 events from 90 cells isolated from seven mice) (C) resulting in a reduction of SR Ca²⁺ load normalized spark-mediated SR Ca²⁺ leak (=spark mass × spark frequency/SR Ca²⁺ content) (n = 38, 19, 22, and cells isolated from 3–4 mice for TTX, Ril, and SN6, respectively) compared with Iso alone (n = 41 cells isolated from seven mice, *P < 0.05). (D) Representative examples of the line-scan images of Iso-stimulated R33Q and WT ventricular cardiomyocytes paced at 0.3 Hz. (E) TTX (n = 66 cells from four mice), Ril (n = 67 cells from two mice), and SN6 (n = 50 cells from three mice) significantly decreased DCR events in the form of Ca²⁺ waves in Iso-treated R33Q myocytes (n = 93 cells isolated from eight mice), (F) delayed the latency to the first DCR, as well as reduced DCR synchronicity as demonstrated by a broadened distribution of DCR first latencies and (inset) an increase in the median latency and S.D. of the latency compared with Iso alone (TTX, n = 208 events; Ril, n = 118 events; SN6, n = 129 events vs. Iso alone, n = 237 events, *P < 0.05).

Figure 2.

Effect of neuronal Na⁺ channel blockade with riluzole on peak Na⁺ current. (A) Inward Na⁺ currents obtained by 250 ms depolarization steps from holding potential −120 to 0 mV in 5 mV increments at 3 s intervals. (B) Corresponding peak I/V relationship (n = 9 cells from five mice for WT and n = 6 cells from two mice for R33Q). (C) Reduction of Na⁺ current by 100 nM TTX and 10 µM Ril (n = 5 cells from two mice for TTX, n = 7 cells from two mice for Ril in WT, and n = 4 cells from two mice for Ril in R33Q, *P < 0.05). (D) The decaying phase of Na⁺ current traces was fitted to 2-exponential function yielding two time constants, one slow (τ_is) and the other to fast component (τ_if). The effect of 10 µM Ril in R33Q cardiomyocytes (n = 4 cells from two mice, *P < 0.05) on the decaying phase of Na⁺ current with two components, slow (τ_is) and a fast (τ_if). (E) Reduction of the slow (τ_is) component of decaying phases of Na⁺ current by 100 nM TTX and 10 µM Ril in WT cardiomyocytes (n = 4 cells from two mice and n = 6 cells from two mice, respectively, *P < 0.05).

Next, we investigated the antiarrhythmic effects of nNa_v inhibition using riluzole,²² an agent with an established clinical efficacy in management of amyotrophic lateral sclerosis (ALS).²³ First, we assessed whether riluzole inhibits the TTX-sensitive component of I_Na. To this end, we compared I_Na reduction between 100 nM TTX and 10 µM riluzole. In WT cardiomyocytes, we observed no significant difference in the reduction of peak I_Na (32.5 ± 7.7 vs. 35.1 ± 4.2%) between the two pharmacological interventions (Figure 2C). Furthermore, both riluzole and TTX facilitated I_Na inactivation as evidenced by a reduction in the slow component of the decaying phase of I_Na (Figure 2D and E). Importantly, effects of 10 µM riluzole on excitability (Supplementary material online, Figure S2A and B), Ca²⁺ release (Figure 1A–E; Supplementary material online, Figure S1A–C), and Ca²⁺ release synchrony (Figure 1E) were similar to those observed with TTX.^20,24 Lastly, riluzole did not affect Ca²⁺ spark frequency in permeabilized cardiomyocytes (8.73 ± 0.49 vs. 7.69 ± 0.36 Sparks/100 µm/s, P = ns, n = 634–700 events), suggesting that riluzole lacks a direct effect on RyR2.

To assess whether modulation of nNa_v inactivation²⁵ (Figure 2D and E) can in part contribute to the generation of DCR in quiescent (non-stimulated) myocytes, we facilitated as well as slowed nNa_v inactivation with riluzole (10 µM) and β-pompilidotoxin (β-PMTX, 50 µM), respectively.²⁶ In catecholamine-free wild-type (WT) myocytes, we observed riluzole-mediated decrease in Ca²⁺ spark frequency (Figure 3A–C). On the other hand, β-PMTX increased frequency of Ca²⁺ sparks in WT myocytes despite the omission of catecholamines (Figure 3A–C). Furthermore, the enhanced Na⁺ flux at rest translated into increased rate of DCR frequency in stimulated, catecholamine-exposed WT cardiomyocytes (Figure 3D and E). Additionally, such rise in DCR was coupled to abbreviated latency to the first DCR and reduced SR Ca²⁺ content (Figure 3F; Supplementary material online, Figure S1E and F). This reduction in SR Ca²⁺ content can perhaps be ascribed to a large reverse mode NCX (Na⁺-out and Ca²⁺-in) in the junctional cleft prompted by enhanced Na⁺ influx via the incompletely inactivated nNa_v. Of note, the effects of β-PMTX on Ca²⁺ cycling were reversed with riluzole (Figure 3A–E; Supplementary material online, Figure S1D–F). Importantly, the effects of nNa_v modulation were independent of global phosphorylation state of the cardiomyocyte. Neither Ca²⁺/calmodulin-dependent protein kinase II inhibition with KN-93 (5 µM) in the presence of Iso nor the complete omission of Iso from the superfusate prevented nNa_v modulation of aberrant Ca²⁺ release (Figure 3A–C; Supplementary material online, Figure S2C and D). Taken together these results strongly suggest that a specific subpopulation of Na⁺ channels (i.e. nNa_v) exert a modulatory effect on DCR, which likely involves NCX.

Figure 3.

Neuronal Na⁺ channel contribute to Na⁺/Ca²⁺ signalling and DCR. (A) Representative examples of the line-scan images of Ca²⁺ sparks recorded in WT ventricular cardiomyocytes. nNa_v blockade with Ril in WT myocytes reduced Ca²⁺ spark frequency, while augmentation of the current passed by nNa_v with β-PMTX increased Ca²⁺ spark frequency. (B) Ril decreased (n = 199 events from 49 cells isolated from two mice) while β-PMTX increased Ca²⁺ spark frequency (n = 2413 events from 159 cells isolated from six mice) relative to catecholamine-free control (n = 1363 events from 115 cells isolated from eight mice). (C) Ril (n = 27 cells isolated from two mice) decreased while β-PMTX (n = 36 cells isolated from six mice) increased SR Ca²⁺ load normalized spark-mediated SR Ca²⁺ leak relative to catecholamine-free control (n = 29 cells isolated from eight mice, *P < 0.05). Ril and SN6 significantly reversed the effects of β-PMTX on Ca²⁺ spark properties (n = 293 and 753 events from 51 and 65 cells isolated from three mice for Ca²⁺ spark, respectively, and n = 27 and 24 cells isolated from three mice for SR Ca²⁺ load normalized spark-mediated SR Ca²⁺ leak, respectively, ^#P < 0.05). (D) Representative examples of the line-scan images of Iso-stimulated WT ventricular cardiomyocytes paced at 0.3 Hz. (E) Augmentation of the current passed by nNa_v with β-PMTX in WT myocytes (n = 95 cells isolated from seven mice for β-PMTX and n = 38 cells isolated from five mice for Iso alone, *P < 0.05) increased DCR frequency and (F) shortened the latency to the first DCR relative to Iso alone (n = 309 events for PMTX and n = 121 for Iso alone, *P < 0.05). Ril and SN6 significantly reversed the effects of β-PMTX on DCR frequency (n = 30 and 86 cells isolated from three mice, respectively, ^# P < 0.05, * P < 0.05 vs. Iso alone) and latency (n = 97 and 343 events, respectively).

To assess the role of NCX in the interaction between nNa_v and RyR2, we inhibited reverse mode NCX (Na⁺-out and Ca²⁺-in) with SN6 (5 µM).^27,28 Effects on DCR similar to those observed with nNa_v blockade were obtained with SN6 (Figures 1 and 3; Supplementary material online, Figure S1). Furthermore, no significant effect was observed on forward mode NCX (Ca²⁺-out and Na⁺-in) as assessed by the decay of the caffeine-induced Ca²⁺ transient (Supplementary material online, Figure S1C, inset). These data suggest that NCX can regulate Ca²⁺ release²⁹ and is an integral part of the pro-arrhythmic interaction between nNa_v and RyR2 within R33Q cardiomyocytes.

3.2. Neuronal Na⁺ channels and RyR2 colocalize

Experiments in isolated myocytes suggest that localization of nNa_v in close proximity to SR Ca²⁺-release machinery is consistent with their role in precipitating DCR. Therefore, we performed confocal immunolocalization experiments in ventricular myocytes isolated from R33Q hearts. Over 80% of the nNa_v (Na_v1.1, 1.3, 1.6) immunofluorescent signal coincided with RyR2 labelling (Figure 4A–C and E–G) and exhibited a periodic pattern consistent with T-tubular localization (Supplementary material online, Figure S3A–C, E–G, S4, and S5A).^30,31 On the other hand, very little spatial cross-correlation was observed between regions labelled with cardiac-type Na⁺ channels (Na_v1.5) and RyR2 (Figure 4D and H; Supplementary material online, Figure S3D, H, and S5B). These results strongly suggest that nNa_v colocalize with RyR2 in the same subcellular regions in the R33Q mouse ventricle and in part may explain nNa_v contribution to the regulation of DCR in a setting of leaky RyR2.

Figure 4.

Neuronal Na⁺ channels and RyR2 colocalize to the same discrete subcellular regions. Confocal micrographs of myocytes from R33Q mutant mice with red labelling for RyR2s (top left) and green labelling (bottom left) for (A) Na_v1.1, (B) Na_v1.3, (C) Na_v1.6, and (D) Na_v1.5 and the overlay of the two signals (right). Scale bars represent 30 µm. The dashed white box marks the region displayed in high magnification in the insets. Results of colocalization analysis for RyR2 (red) and (E) Na_v1.1, (F) Na_v1.3, (G) Na_v1.6, and (H) Na_v1.5. Images show colocalizing pixels in white, pixels with only RyR2 signal in red, and only Na_v1.x signal in green. Inset plots show green (Na_v1.x) signal intensity plotted against red (RyR2) signal intensity. Pixels with colocalizing signals are marked in yellow, only red signal in red, only green signal in green, and weak signal intensity (background fluorescence) in black.

3.3. Neuronal Na⁺ channel blockade desynchronizes DCR in tissue

We examined the role of nNa_v in DCR and focal excitation in tissue. We hypothesized that in a setting of leaky RyR2,^32,33 nNa_v could facilitate temporal alignment or synchronization of DCRs, and the resultant membrane depolarization in neighbouring cells accounting for the generation of ventricular extrasystole across the myocardium. Therefore, we performed simultaneous confocal imaging of membrane potential and cytosolic Ca²⁺ in cardiac muscle preparations from R33Q mice during nNa_v blockade. Both riluzole (10 µM) and TTX (100 nM) significantly reduced frequency and synchronicity of Iso-promoted DCR (Figure 5A–E). Correspondingly, both drugs reduced the frequency of extrasystolic action potentials (Figure 5E). Thus, nNa_v blockade with riluzole can reduce DCR, its synchronicity, and the resultant triggered arrhythmias in cardiac muscle.

Figure 5.

Neuronal Na⁺ channel blockade desynchronizes DCR in tissue. Representative Ca²⁺ (top) and voltage (bottom) line-scan images of Iso-treated R33Q muscle preparations paced at 1 Hz before (A) and after (B) exposure to riluzole (Ril, 10 µM), along with the corresponding Rhod-2 fluorescence profiles from each myocyte and average di-4-ANBDQBS signal. Dashed white lines outline myocyte borders. Red asterisks indicate tissue-wide extrasystolic Ca²⁺ release (DCR) that resulted in triggered activity (TA). (C) Temporal distribution of DCR in cardiac muscles (n = 264 events for Iso alone, n = 355 for Ril and n = 372 for 100 nM TTX). (D) Median latencies to first DCR (solid bars) and their S.D. (hashed bars) were increased by Ril (n = 355 events) and TTX (n = 372 events) relative to Iso alone (n = 264 events). (E) DCRs and TA frequency per second (Hz) before and after treatment with Ril (10 µM; n = 26 cells from four preparations isolated from three mice, *P < 0.05) or TTX (100 nM; n = 29 cells from three preparations isolated from three mice, *P < 0.05).

3.4. Effect of neuronal Na⁺ channel blockade on CPVT

We next examined the therapeutic efficacy of disrupting the crosstalk between nNa_v and the leaky RyR2 on the cellular level that desynchronized DCR in intact tissue. To this end, we assessed arrhythmia inducibility in R33Q mice. A caffeine and epinephrine challenge induced frequent ventricular bigeminy (Figure 6A), which degenerated into polymorphic VT. After riluzole treatment (10 mg/kg), half of the R33Q animals remained in sinus rhythm (Figure 6B) despite the caffeine and epinephrine challenge. Overall, riluzole reduced by half ventricular arrhythmias, including frequent ventricular extrasystoles and bigeminies (Figure 6D), while completely suppressing VTs in all R33Q mice tested (Figure 6E). Consistent with observations in isolated cardiomyocytes, riluzole's antiarrhythmic effect in R33Q mice was modulated by perturbations of NCX-dependent Na⁺/Ca²⁺ signalling. NCX inhibition with SN-6 (40 mg/kg, Figure 6C–E) resulted in a reduced overall arrhythmia burden. Taken together, these data suggest that perturbing local crosstalk between nNa_v and RyR2 via nNa_v blockade exerts an antiarrhythmic effect in vivo by reducing and desynchronizing DCR.

Figure 6.

Effect of neuronal Na⁺ channel blockade on CPVT. (A) Representative ECG recordings of R33Q mutant mice before (top) and after (bottom) injection (i.p.) of epinephrine (Epi, 1.5 mg/kg) and caffeine (Caff, 120 mg/kg). As illustrated in the ECG trace, Epi + Caff challenge resulted in numerous repetitive ventricular extrasystoles (arrows), bigeminy, and the induction of ventricular tachycardia (VT). (B) ECG of R33Q mice after administration of Ril (10 mg/kg), targeting plasma concentrations comparable to those achieved in experiments conducted in isolated cardiomyocytes,³⁴ before (top) and after (bottom) Epi + Caff challenge. (C) ECG of SN-6 (40 mg/kg) treated mice 10 min before (top) and after (bottom) Epi + Caff challenge. (D) Arrhythmia and (E) ventricular tachycardia (VT) incidence (%) in R33Q mice exposed to Epi + Caff during various interventions (n = 6–11 mice, *P < 0.05). (F) Na⁺/Ca²⁺ signalling: colocalization to the same discrete subcellular region allows crosstalk between neuronal Na⁺ channels (nNa_v) and leaky Ca²⁺ release channels (RyR2) through the NCX resulting in aberrant DCR that underlie ventricular extrasystoles, which in turn trigger VT.

4. Discussion

Here we studied CPVT-associated arrhythmogenesis resulting from aberrant DCR.^15,35 Previous studies showed that CPVT-linked mutations disrupt RyR2 gating, thereby precipitating synchronous DCR and triggered arrhythmias.^5,8,36 In recent years, Na⁺ channel blockade has emerged as a potential antiarrhythmic strategy in genetic and acquired models of Ca²⁺-mediated arrhythmias.^13,19,21 Previous studies have suggested two mechanisms for the antiarrhythmic effect of Na⁺ channel blockers such as flecainide: (i) direct inhibition of the RyR2³⁷ or (ii) decreased membrane excitability due to cardiac-type Na⁺ channel (Na_v1.5) blockade uncoupling DCR from membrane depolarization.³⁸ Our results demonstrate a novel mechanism for Ca²⁺-mediated arrhythmogenesis which involves Na⁺/Ca²⁺ signalling independent of direct RyR2 block and reduced excitability. Importantly, disruption of the Na⁺/Ca²⁺ signalling via inhibition of neuronal Na⁺ channels (nNa_v) effectively suppressed arrhythmias in CPVT mice, thus suggesting a new therapeutic approach in management of CPVT.

Initially, the antiarrhythmic effect of Na⁺ channel blockers, such as flecainide, was attributed to the direct inhibition of the RyR2.³⁷ Yet, riluzole did not reduce spark frequency in permeabilized myocytes in the present study arguing against its interacting directly with RyR2. Subsequent studies have suggested that decreased membrane excitability as a result of Na⁺ channel blockade uncouples DCR from membrane depolarization.³⁸ However, neither 10 µM riluzole in our study nor 100 nM TTX in previous ones altered membrane excitability.^19,20,39 Importantly, our findings point to a distinct microdomain where nNa_v are localized near RyR2s (Figure 4), forming the functional unit underlying arrhythmogenic DCR in CPVT (Figure 6F). This is in accord with previous studies that have demonstrated the presence of nNa_v in the T-tubule.^31,40–43 We further demonstrate using pharmacological modulation of nNa_v (i.e. inhibition with TTX and riluzole and augmentation with β-PMTX) as well as NCX activity (SN6) that the unique interplay between Na⁺ and Ca²⁺ handling within this domain is key to the pro-arrhythmic process. More precisely, in CPVT where cardiomyocytes exhibit increased baseline pro-arrhythmic RyR2 Ca²⁺ leak, nNa_v, and NCX blockade reduced DCR and suppressed arrhythmic activity on cellular level (Figures 1 and 3) and in tissue isolated from CPVT hearts (Figure 5). The results of these functional experiments taken together with colocalization of nNa_v (but not cardiac isoform of Na⁺ channel) with RyR2 make a very compelling case for the involvement of nNa_v in the modulation of the arrhythmic DCR. However, this does not preclude the cardiac isoform of the Na⁺ channel from participating in the arrhythmogenic process. Future studies will need to address the involvement of this particular subpopulation of Na⁺ channels in aberrant Ca²⁺ handling.

The concept of Na⁺/Ca²⁺ signalling in the vicinity of RyR2 that our data suggest is supported by previous observations that, reverse mode NCX (Na⁺-out/Ca²⁺-in) activated by the TTX-sensitive inward Na⁺ current early during the action potential, can contribute to stimulated Ca²⁺ release.^17,44–47 These studies demonstrated that, during a normal heart beat, Na⁺ channels modulate the efficiency of Ca²⁺ release by supplying the necessary Na⁺ for reverse NCX, which in turn can facilitate RyR2 function (i.e. priming RyR2). The present study extends this concept to elucidate the mechanism of Ca²⁺-dependent arrhythmias by providing the first evidence that the crosstalk between nNa_v along with NCX and RyR2 is critical to aberrant DCR and triggered activity. Specifically, it has been previously demonstrated that nNa_v inactivate at relatively high potentials⁴⁸ and carry a substantial residual Na⁺ current during depolarization.²⁴ nNa_v have also been shown to exhibit substantial activity at negative membrane potentials,²⁵ where they are prone to activation and do not completely inactivate.⁴⁸ Thus, nNa_v activity in both stimulated and quiescent cardiomyocytes could account for the nNa_v-dependent effects observed herein. Our results further suggest that Na⁺ influx through the not fully inactivated nNa_v (Figure 2D)²⁴ in the wake of electrical stimuli (Figures 1D–F and 3D–F) or those that open at rest (Figures 1A–C and 3A–C) may facilitate DCR by enhancing NCX-dependent Ca²⁺ accumulation in the space between the sarcolemma and the RyR2. This microdomain Ca²⁺ accumulation in turn promotes Ca²⁺-induced Ca²⁺ release via RyR2 that are sensitized due to the presence of a pathogenic mutation in CASQ2 (Figures 1, 3, and 6F).^7–10 While a direct measurement of microdomaine Na⁺ and Ca²⁺ flux in a physiologically relevant setting might be presently not feasible, our functional and structural assays targeting the key components of the Na⁺/Ca²⁺ signalling provide strong evidence for its role in CPVT-associated arrhythmogenesis. Importantly, they reveal Na⁺/Ca²⁺ signalling and in particular nNa_v as a therapeutic target. Future studies will have to, however, address the specific subtypes of nNa_v involved in this pro-arrhythmic process.

An additional unanswered question that our study attempts to tackle is how DCR can occur synchronously in sufficient amount of myocytes to elicit a ventricular extrasystole within the entire myocardium. An extrasystolic trigger must provide a depolarizing current source large enough to overcome the electrical load of connected neighbouring myocytes serving as current sinks.⁴⁹ Our data suggest that in addition to facilitating arrhythmogenic DCR on the cellular level, Na⁺/Ca²⁺ signalling synchronizes aberrant DCR across the myocardium which results in triggered activity in intact cardiac tissue (Figure 5). The latter effect enables DCR-induced depolarization to overcome the source-sink mismatch, thereby generating ventricular extrasystole.

Interestingly, over 25 years ago Na⁺ channel blockade has emerged as a promising strategy in management of Ca²⁺-mediated arrhythmias due to heart failure.⁵⁰ However, reductions of electrical excitability have proved to be pro-arrhythmic in these patients, and class Ic antiarrhythmics are contraindicated in the population with structural heart disease due to an increased risk of arrhythmic death.^51,52 Importantly, the reduction in electrical excitability in the present as well as our previous work was not observed during selective blockade of nNa_v,¹⁹ which nonetheless proved to be antiarrhythmic. Taken together, these data suggest a novel arrhythmogenic mechanism and provide clues to its structural underpinnings. Importantly, they also form the basis for a novel mechanism-based antiarrhythmic therapy utilizing a clinically available nNa_v blocker without compromising electrical excitability, thus causing pro-arrhythmic responses.^19,51

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Funding

This work was supported by National Institutes of Health (NIH) Grants (HL074045, HL063043) to S.G., (HL098039-03) to P.B.R. and (HL084583, HL083422, HL114383) to P.J.M.; and ACCP Cardiology PRN Investigator Development Research Award to P.B.R.