Introduction

Sodium channels play a central role in the initiation and propagation of the action potential (AP) in excitable cells, including cardiomyocytes (CMs) and neurons. Cardiac sodium channel (NaV1.5) loss of function is a critical mediator of cardiac conduction slowing, predisposing to ventricular arrhythmias and sudden cardiac death (SCD), both in acquired pathologies (ischemia, hypertrophy, heart failure) and inherited cardiac disorders caused by mutations in SCN5A, the gene encoding NaV1.5 [1, 2]. On the other hand, impaired NaV1.5 inactivation may induce a small inward sodium current (INa), the so-called late INa, that persists during the plateau and repolarization phase of the AP. Enhanced late INa can prolong AP duration (APD) and increase intracellular calcium (Ca2+) via altered Na+/Ca2+ exchanger activity, thus promoting arrhythmias [3]. While NaV1.5 is the main sodium channel isoform expressed in cardiac tissue, other sodium channel isoforms are also present in the heart, including NaV1.1/SCN1A, NaV1.2/SCN2A, NaV1.3/SCN3A and NaV1.6/SCN8A [4]. These isoforms are typically referred to as “neuronal” sodium channel isoforms due to their abundant expression and well-established function in neurons. While NaV1.5 is blocked only by micromolar concentrations of tetrodotoxin (TTX) [i.e. TTX-resistant], most neuronal isoforms are more TTX-sensitive and inhibited by nanomolar concentrations [4]. An exception is NaV1.8, encoded by the SCN10A gene. This isoform is mainly expressed in dorsal root ganglia, plays a role in pain perception [5] and is inhibited only by micromolar TTX concentrations, similar to NaV1.5 [4].

Several genome-wide association studies (GWAS) have suggested a role for SCN10A/NaV1.8 in modulating cardiac conduction parameters, such as PR and QRS interval [6,7,8,9,10]. SCN10A variants have also been associated with atrial fibrillation (AF) [11,12,13] and with Brugada syndrome [14,15,16], an inherited cardiac disease characterized by cardiac conduction slowing and increased risk for SCD. However, if and by which mechanism SCN10A/NaV1.8 impacts on cardiac electrical function is still a matter of debate. Inhibition of NaV1.8 by the blocker A-803467 has been reported to decrease late INa and shorten APD in mouse and rabbit cardiomyocytes [17], whereas we previously described the absence of functional NaV1.8 in murine cardiomyocytes [18]. Moreover, conflicting results have been reported in studies of mice deficient for Scn10a, with either similar or decreased APD observed in Scn10a−/− cardiomyocytes as compared to wild-type cardiomyocytes [17, 19]. We and others have shown that NaV1.8 is specifically expressed in murine, canine and human cardiac neurons [18, 20, 21], suggesting a function of the SCN10A gene product for cardiac conduction via modulation of AP firing in intracardiac neurons [18, 21, 22]. Additionally, the SCN10A variant rs6801957 has been shown to modulate SCN5A expression in cardiac tissue thereby potentially impacting on conduction [23]. Overall, the role of SCN10A/NaV1.8 in the heart and the mechanisms by which this gene and/or its gene product affects cardiac function remain only partially explained. In particular, electrophysiological studies in non-diseased human cardiomyocytes aimed at defining the physiological role of NaV1.8 in the human heart are limited. To address these issues, we here investigated the functional relevance of NaV1.8 in atrial and ventricular cardiomyocytes, focusing on the contribution of NaV1.8 to the peak and late INa under normal conditions in different species.

Methods

Isolation of Rabbit Left Ventricular Cardiomyocytes

Three–four-month-old male New Zealand White rabbits (Charles River Laboratories) were anaesthetized with 20 mg xylazine and 100 mg ketamine (intramuscularly) and heparinized with a bolus of 1000 IU heparin (intravenously). Subsequently, the animals were sacrificed, the thorax was opened and the heart was rapidly excised and connected to a Langendorff system. Left ventricular (LV) cardiomyocytes were isolated as previously described [24] (see Data Supplement).

Isolation of Human Left Atrial Cardiomyocytes

Human left atrial appendages (LAAs) were obtained from patients in sinus rhythm (SR) without a history of AF undergoing cardiac surgery (coronary bypass grafting or valve surgery) and included in the multicenter PREDICT AF study [25]. Patient characteristics are reported in Supplemental Table 1. Part of the LAA tissue was immediately frozen in liquid nitrogen to be subsequently used for molecular analysis, while the other part was transported to the laboratory on ice and single cells were obtained by an enzymatic isolation modified from Dobrev et al. [26]. An expanded Methods section is available in the Data Supplement.

Differentiation of hiPSCs into Cardiomyocytes

A human-induced pluripotent stem cell (hiPSC) control line (iC113) previously generated and characterized [27] was used to generate cardiomyocytes (hiPSC-CMs) by adaptation of a previously described protocol [28]. hiPSC-CMs were used for electrophysiological analysis and RT-PCR. An expanded Methods section is available in the Data Supplement.

Electrophysiology

Data Acquisition and Analysis

Membrane currents [(INa, late INa and L-type calcium currents (ICaL)] and APs were measured with the ruptured and perforated patch-clamp technique, respectively, using an Axopatch 200B amplifier (Molecular Devices, San Jose, CA, USA). Voltage control, data acquisition and analysis of currents and APs were performed with pClamp10.6/Clampfit (Molecular Devices, San Jose, CA, USA) or a custom-made software. Borosilicate glass patch pipettes (Harvard Apparatus, Holliston, MA, USA) with a tip resistance of 2–2.5 MΩ were used. Series resistance (Rs) and cell membrane capacitance (Cm) were compensated for 80%. Peak INa, ICaL and APs were filtered at 5 kHz. INa and APs were digitized at 40 kHz, while ICaL was digitized at 20 kHz. Finally, late INa was filtered and digitized at 2 kHz and 1 kHz, respectively.

Sodium Current Measurements

Peak INa and late INa were measured in single cells using a pipette solution containing (in mM) 3.0 NaCl, 133 CsCl, 2.0 MgCl2, 2.0 Na2ATP, 2.0 TEACl, 10 EGTA, 5.0 HEPES; pH 7.2 (CsOH). For late INa measurements, hiPSC-CMs, rabbit and human CMs were superfused with a bath solution containing (in mM) 130 NaCl, 10 CsCl, 1.8 CaCl2, 1.2 MgCl2, 11.0 glucose, 5.0 HEPES, 0.005 nifedipine; pH 7.4 (CsOH). For peak INa recordings, a similar bath solution was used with the exception of a lower NaCl concentration for proper voltage control. Hence, NaCl was replaced by CsCl (for rabbit and human CMs, we used 7 mM NaCl and 133 mM CsCl; for hiPSC-CMs: 20 mM NaCl and 120 mM CsCl). Peak INa was measured at room temperature in response to depolarizing voltage steps from a holding potential of − 120 mV (cycle length of 5 s). INa was defined as the difference between peak and steady-state current (at 500 ms). Voltage dependence of activation and inactivation curves was fitted with Boltzmann function (y = [1 + exp.{(V-V1/2)/k}]−1), where V1/2 is the half-maximal voltage of (in) activation and k, the slope factor. Nav1.8-based late INa and total late INa were measured at 36 °C, as A-803467 (100 nM) and TTX (30 μM)-sensitive currents, respectively, using descending ramp protocols (cycle length of 5 s) as depicted in Fig. 2 and Supplemental Fig. 4. Current densities were calculated by dividing current amplitude by Cm. Cm was determined by dividing the decay time constant of the capacitive transient in response to 5 mV hyperpolarizing steps from − 40 mV, by the Rs. Potentials for peak INa and late INa recordings were not corrected for the estimated change in liquid junction potential. The NaV1.8 channel inhibitor A-803467 (Tocris Bioscience, Bristol, United Kingdom) was solubilized in DMSO at a stock solution of 10 mM and diluted to the final concentration of 100 nM before use. This dose was chosen based on previous IC50 data and to ensure maximal inhibition of Nav1.8-based current [29].

L-Type Calcium Current Measurements

ICaL was measured in isolated rabbit left ventricular CMs at 36 °C. An expanded Methods section is available in the Data Supplement.

Action Potential Measurements

In single rabbit left ventricular CMs, human left atrial (LA) CMs and hiPSC-CMs, APs were measured at 36 °C using a modified Tyrode’s solution containing (in mM) 140 NaCl, 5.4 KCl, 1.8 CaCl2, 1.0 MgCl2, 5.5 glucose, 5 HEPES, pH 7.4 (NaOH). Pipettes were filled with (in mM) 125 K-gluconate, 20 KCl, 5 NaCl, 0.44 amphotericin-B, 10 HEPES, pH 7.2 (KOH). APs were elicited at 1 Hz by 3 ms, ≈ 1.2× threshold current pulses through the patch pipette. Typically, hiPSC-CMs have a small or even complete lack of the inward rectifying potassium current (IK1). Consequently, their resting membrane potential (RMP) is depolarized and they are frequently spontaneously active [30]. To overcome these conditions, which limit the functional availability of INa, transient outward potassium current and L-type Ca2+ current [31], we injected an in silico IK1 with kinetics of Kir2.1 channels through dynamic clamp [32]. Thus, cells became quiescent with a RMP of around − 80 mV. We analysed RMP, AP amplitude (APA), maximal AP upstroke velocity (Vmax) and AP duration (APD) at 50% and 90% repolarization (APD50 and APD90, respectively). Data from 10 consecutive APs were averaged and potentials were corrected for the calculated liquid junction potential of 15 mV [33].

Real-Time Polymerase Chain Reactions

Total RNA was isolated from left atrial appendages of five patients undergoing cardiac surgeries using TRIzol Reagent (Invitrogen, Waltham, MA, USA) and from hiPSC-CMs obtained from four independent differentiations using NucleoSpin RNA (MACHEREY-NAGEL ref.: 740955.50, Duren, Germany) following manufacture protocol. cDNA was synthesized from total RNA by SuperScript™ II Reverse Transcriptase (Invitrogen, Waltham, MA, USA). Real-time PCR was performed on the platform of Light Cycler 480 (Roche, Basel, Switzerland) using SYBR green I master mix (Roche, Basel, Switzerland) and the sets of primers reported in Supplemental Table 2. Gene expression was determined according to linear regression analysis using LinRegPCR software and normalized by the expression of hypoxanthine phosphoribosyltransferase (HPRT).

RNA Sequencing Data Analysis

SCN5A and SCN10A expression in human right and left atria [34], ventricular and atrial hiPSC-CMs [35] and rabbit left ventricular tissue were extrapolated from the RNA sequencing (RNA-Seq) datasets GSE31999, GSE111007 and GSE115605, respectively, which are publicly available online https://www.ncbi.nlm.nih.gov/geo/. For the analyses, read counts for SCN5A and SCN10A transcripts were normalized to millions of total reads generated per sample and to SCN5A (ENST00000413689.1) and SCN10A (ENST00000449082.2) transcript size (i.e. Fragment Per Kilobase Million, FPKM).

Statistical Analysis

Values are shown as mean ± SEM. Paired Student’s t test, unpaired Student’s t test, one-way repeated measures ANOVA followed by Holm-Sidak test for post hoc analyses and two-way repeated measures ANOVA were used when appropriate. Mann-Whitney U test and one-way repeated measures ANOVA on Ranks (Friedman test) followed by Tukey test for post hoc analyses were used for data not normally distributed. The level of statistical significance was set to p < 0.05.

Results

Effect of A-803467 on AP Properties in Atrial and Ventricular Cardiomyocytes

We first assessed the effects of the NaV1.8 blocker A-803467 on AP properties. Figure 1a, c, e shows typical AP recordings obtained from rabbit left ventricular CMs, hiPSC-CMs and human left atrial CMs under physiological conditions (baseline), in the presence of 100 nM A-803467 and after wash-out of the drug. On average, maximal upstroke velocity, AP amplitude and resting membrane potential were not affected by A-803467 exposure (Fig. 1b, d, f, Supplemental Table 3). In rabbit ventricular CMs, we observed a small, yet significant, APD shortening induced by A-803467 treatment (APD reduction of 4.8% for APD50 and of 3.5% for APD90, Fig. 1b, Supplemental Table 3). However, as shown in detail in Supplemental Fig. 1a and b, the observed APD reduction was (partly) reversible upon wash-out only in a minority of cells. In the majority of cases, either A-803467 did not affect APD or the effect was non-reversible. Moreover, time-matched control experiments in rabbit ventricular CMs showed APD50 and APD90 shortening occurring over time similar to that observed with A-803467 (Supplemental Figs. 2 and 3). In human left atrial CMs, exposure to A-803467 did not change APD (Fig. 1e, f, Supplemental Fig. 1e, f and Supplemental Table 3), and in hiPSC-CMs, only APD50 was significantly reduced but not APD90 (Fig. 1c, d, Supplemental Table 3). The effect of A-803467 was reversible upon wash-out in the majority of hiPSC-CMs, but the blocker did not affect APD in all cells (Supplemental Fig. 1c and d).

Fig. 1
figure 1

Effect of A-803467 treatment on action potential (AP) properties in ventricular and atrial cardiomyocytes. a, c, e Examples of APs recorded at the stimulation frequency of 1 Hz in rabbit left ventricular (LV) cardiomyocytes (CMs) (a), human-induced pluripotent stem cell-derived CMs (hiPSC-CMs) (c) and human left atrial (LA) CMs (e) under physiological conditions (baseline), after 5-min wash-in of 100 nM A-803467 and after 5-min wash-out of the drug. b, d, f Average data at 1 Hz for maximal upstroke velocity (Vmax), AP amplitude (APA), resting membrane potential (RMP), AP duration at 50% and 90% repolarization (APD50 and APD90), before (baseline) and after wash-in and wash-out of 100 nM A-803467 in rabbit LV CMs, hiPSC-CMs and human LA CMs. *p < 0.05 baseline vs A-803467, #p < 0.05 baseline vs wash-out; one-way repeated measures ANOVA followed by Holm-Sidak test for post hoc analyses or one-way repeated measures ANOVA on Ranks (Friedman test)  followed by Tukey test for post hoc analyses when data were not normally distributed

Full size image

Absence of NaV1.8-Based Late INa in Ventricular and Atrial Cardiomyocytes

We next investigated the effects of A-803467 on late INa in rabbit left ventricular CMs, hiPSC-CMs and human left atrial CMs using descending ramps after a 200 ms pre-pulse to 40 mV (see inset of Fig. 2a, c, e). The advantage of using a ramp protocol instead of a single step protocol is that the ramp protocol allows measurements of late INa across a dynamic voltage range simulating a plateau and repolarization phase of an AP [36]. Figure 2 a, c, and e show typical examples of Nav1.8-based late INa recordings under basal conditions (baseline, black line) and after 5 min wash-in of 100 nM A-803467 (red line). NaV1.8-dependent late INa, measured as A-803467-sensitive current, was obtained by subtraction of the current recorded in the presence of A-803467 from the current recorded in the absence of the compound. NaV1.8-dependent late INa was not detected in any of the three cell types (Fig. 2b, d, f). In a subset of rabbit left ventricular CMs, A-803467 perfusion was followed by 30 μM TTX application, and total late INa was measured as TTX-sensitive current obtained by subtraction of the current recorded in the presence of TTX from the current recorded earlier in the absence of TTX (Supplemental Fig. 4). Average total late INa was around − 0.2 pA/pF, while the A-803467 sensitive current was undetectable. Hence, these experiments demonstrate that functional NaV1.8-based late INa is absent under basal conditions in atrial and ventricular cardiomyocytes.

Fig. 2
figure 2

Absence of NaV1.8-based late sodium current (INa) in ventricular and atrial cardiomyocytes. a, c, e Representative Nav1.8-based late INa traces recorded during a ramp protocol (see insets) in rabbit left ventricular (LV) cardiomyocytes (CMs) (a), human-induced pluripotent stem cell-derived CMs (hiPSC-CMs) (c) and human left atrial (LA) CMs (e) at baseline and after 5-min application of 100 nM A-803467. A-803467 sensitive current was obtained by subtraction of the current recorded in the presence of A-803467 from the current recorded earlier in the absence of the compound. b, d, f Average current-voltage (I-V) relationships for NaV1.8-based late INa measured as A-803467 sensitive current in rabbit LV CMs (b), hiPSC-CMs (d) and human LA CMs (f)

Full size image

Absence of NaV1.8-Based Peak INa in Ventricular and Atrial Cardiomyocytes

In addition to late INa, we also investigated the effects of A-803467 on peak INa density and voltage dependency of activation and inactivation. Figure 3a, c, e shows typical peak INa recordings obtained from rabbit left ventricular CMs, hiPSC-CMs and human left atrial CMs under basal conditions (baseline) and after 5 min wash-in of 100 nM A-803467. Average peak INa densities were unchanged after exposure to A-803467 in all cell types analysed (Fig. 3b, d, f, Supplemental Table 4).

Fig. 3
figure 3

Nav1.8 does not contribute to peak sodium current (INa) in ventricular and atrial cardiomyocytes. a, c, e Representative peak INa traces recorded from rabbit left ventricular (LV) cardiomyocytes (CMs) (a), human-induced pluripotent stem cell-derived CMs (hiPSC-CMs) (c) and human left atrial (LA) CMs (e) under physiological conditions (baseline) and after 5-min wash-in of 100 nM A-803467. b, d, f Average current-voltage (I-V) relationships at baseline and in the presence of 100 nM A-803467 in rabbit LV CMs (b), hiPSC-CMs (d) and human LA CMs (f). Insets: voltage protocols. Statistical test applied: two-way repeated measures ANOVA

Full size image

INa voltage dependence of activation and inactivation, assessed as the half voltage of (in)activation (V1/2) and the slope factor k, was not affected by A-803467 in rabbit CMs and hiPSC-CMs (Fig. 4a–d, Supplemental Table 4). A-803467 caused a small negative shift in V1/2 of activation and inactivation (− 2.8 mV for both activation and inactivation curve) in human LA cardiomyocytes (V1/2 activation − 40.5 ± 1.1 mV vs − 43.3 ± 1.2 mV, p < 0.05, paired Student’s t test; V1/2 inactivation − 90.6 ± 0.9 mV vs − 93.4 ± 1.0 mV, p < 0.05, paired Student’s t test; Supplemental Table 4). Although significant, the biological meaning of such a small change is questionable. Moreover, in a subset of cells where wash-out experiments were also performed, we were unable to reverse these effects of A-803467, and a further negative shift of V1/2 of (in)activation was observed upon wash-out of the compound (Supplemental Fig. 5). These results suggest a time-dependent shift of (in)activation, rather than a A-803467-dependent effect on INa kinetics [37]. Taken together, these findings demonstrate the absence of functional Nav1.8-based peak INa in atrial and ventricular cardiomyocytes.

Fig. 4
figure 4

Effect of A-803467 treatment on sodium current (INa) voltage dependence of activation and inactivation in ventricular and atrial cardiomyocytes. Average INa voltage dependence of activation (a, c, e) and inactivation (b, d, f) in rabbit left ventricular (LV) cardiomyocytes (CMs), human-induced pluripotent stem cell-derived CMs (hiPSC-CMs) and human left atrial (LA) CMs under basal conditions (baseline) and after 5-min exposure to 100 nM A-803467. Insets: voltage protocols. Statistical test applied: paired Student’s t test (see Supplemental Table 4)

Full size image

To explore a potential off-target effect of A-803467, we also measured its effects on the L-type calcium current (ICaL) in rabbit left ventricular CMs. Typical example of ICaL traces recorded under basal conditions (baseline) and in the presence of A-803467 are shown in Supplemental Fig. 6a. Exposure to A-803467 affected neither ICaL density (Supplemental Fig. 6b, Supplemental Table 5) nor ICaL voltage dependence of activation and inactivation (Supplemental Fig. 6c, Supplemental Table 5).

Low SCN10A mRNA Transcript Levels in hiPSC-CMs and in Human Left Atrial Appendages

We finally assessed the mRNA expression levels of SCN10A in hiPSC-CMs and human LAAs using quantitative real-time PCR (RT-PCR). In LAAs tissue, RT-PCR was performed on the same samples used for AP and late INa measurements. As expected, both hiPSC-CMs and LAAs tissue showed robust expression of the cardiac sodium channel isoform SCN5A (relative to the reference gene HPRT). In contrast, SCN10A transcript levels were very low in both hiPSC-CMs and human LAAs (Fig. 5). Similarly, low to almost absent expression of SCN10A as compared to SCN5A was observed in online RNA-Seq datasets of rabbit left ventricular tissue (GSE115605) (Fig. 6a), atrial and ventricular hiPSC-CMs (GSE111007) (Fig. 6b, c) [35] and human left and right atria (GSE31999) (Fig. 6d, e) [34]. These observations are in line with our patch-clamp data showing the absence of functional NaV1.8-based sodium channels under basal conditions in atrial and ventricular CMs.

Fig. 5
figure 5

SCN10A and SCN5A expression levels in human-induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) and human left atrial tissue. SCN5A and SCN10A mRNA levels in hiPSC-CMs (a) and in human left atrial appendages (LAAs) (b). Gene expression was normalized to the reference gene HPRT

Full size image
Fig. 6
figure 6

RNA sequencing (RNA-Seq) data analysis for SCN5A and SCN10A expression levels in rabbit left ventricle, induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) and human atria. SCN5A and SCN10A expression levels in rabbit left ventricle (LV) (a), atrial and ventricular hiPSC-CMs (b, c) and human left (LA) and right atria (RA) (d, e) were extrapolated from analysis of online RNA sequencing (RNA-Seq) raw datasets previously published [34, 35]. Read counts for SCN5A and SCN10A transcripts were normalized to millions of total reads generated per sample (six samples for rabbit left ventricular tissue and four samples for hiPSC-CMs and human atria) and to SCN5A and SCN10A transcript size (i.e. fragment per kilobase million, FPKM). *p < 0.05; unpaired Student’s t test or Mann-Whitney U test when data were not normally distributed

Full size image

Discussion

While several studies have previously implicated NaV1.8 in modulating cardiac electrophysiology and arrhythmia susceptibility, the underlying mechanism(s) are still a matter of debate. To address this, we here investigated the functional relevance of NaV1.8 in atrial and ventricular cardiomyocytes, focusing on the contribution of NaV1.8 to the peak and late INa under physiological conditions. Using detailed patch-clamp analyses of atrial and ventricular myocytes from different species, we observed a lack of effect of the NaV1.8 blocker A-803467 on peak and late INa in cardiomyocytes. In line with these observations, molecular investigation showed a virtual absence of SCN10A mRNA in human atrial tissue and hiPSC-CMs. Similarly, analysis of online RNA-Seq datasets of rabbit ventricular tissue, ventricular and atrial hiPSC-CMs and human right and left atria revealed low to almost absent expression of SCN10A as compared to SCN5A. Hence, our results demonstrate the absence of functional NaV1.8 channels in non-diseased atrial and ventricular cardiomyocytes, which is of particular relevance when extrapolating findings on SCN10A mutations and (common) variants.

Our findings are in contrast to the study of Yang et al. [17], which suggested that NaV1.8 is a component of late INa in non-diseased cardiomyocytes, and as such may modulate arrhythmia susceptibility [17]. Yang et al. showed that in mouse and rabbit ventricular cardiomyocytes, application of A-803467 reduced late INa and shortened APD, without affecting peak INa density [17]. In contrast, we did not detect any NaV1.8-based late INa in our cardiomyocytes. This discrepancy could be due to species differences (mouse vs human) and/or to different experimental conditions such as different temperature (room temperature [17] versus physiological temperature used by us). Indeed, in a recent study, Poulet et al. [38] reported a significant increase in late INa in human right atrial cardiomyocytes from AF patients as compared to patients in SR when experiments were conducted at room temperature. However, at physiological temperature, the difference in late INa amplitudes between SR and AF cells was less pronounced and did not reach the level of statistical significance [38]. In our study, A-803467 reduced APD90 by only 3.5% in rabbit ventricular CMs, while in the study of Yang et al. [17], the reduction was ~ 30%. Again, different experimental conditions may underlie these discrepancies, for instance the use of the perforated patch vs rupture patch, differences in recording solutions and temperature. Moreover, in the study of Yang et al., wash-out of A-803467 was not investigated, leaving the possibility that part of the observed effect was a time-dependent effect. Indeed, in time-matched AP recordings in rabbit cardiomyocytes, we observed an APD reduction over time similar to that observed with A-803467 suggesting that the apparent APD shortening induced by A-803467 in this species is a non-specific effect independent of the blocker. A similar small APD reduction was also observed in mouse ventricular CMs (but not in mouse atrial CMs) in our previous study [18]. Finally, despite the absence of NaV1.8-based peak and late INa, we still observed a small yet significant and mostly reversible decrease in APD in hiPSC-CMs induced by A-803467. This AP shortening could be due to a potential off-target effect of A-803467. Although we observed no effects of A-803467 on ICaL in rabbit, we cannot completely rule out effects of A-803467 on other ion channels, cautioning its use in electrophysiological studies aimed at establishing the functional relevance of NaV1.8 in, e.g. arrhythmogenesis.

A number of previous studies have suggested a role for SCN10A/NaV1.8 in modulating cardiac conduction and arrhythmogenesis. Perhaps the most compelling evidence on a potential role for NaV1.8 came from studies in mice deficient for Scn10a (Scn10a−/−). In ventricular CMs isolated from Scn10a−/− mice, APs were shorter than those in wild-type mice and A-803467 had no effect on peak or late INa, nor on APD, thus supporting the idea of a contribution of NaV1.8 to late INa [17]. However, in a follow-up study, the same authors reported similar APD in wild-type and Scn10a−/− ventricular cardiomyocytes at baseline conditions, and only under extreme experimental conditions, e.g. after pre-treatment with the late INa enhancer ATX II, a reduced late INa was observed in knockout ventricular CMs [19]. Nevertheless, ATX-II administration in anesthetized mice and Langendorff-perfused hearts prolonged QTc and induced arrhythmias to the same extent in wild-type and Scn10a−/− mice. Finally, no Scn10a transcript was detected in either wild-type or Scn10a−/− ventricular CMs and ECG parameters were similar in both wild-type Scn10a−/− mice [19], further underlining the limited relevance of NaV1.8 in cardiomyocytes under physiological conditions.

Mutations or variants in SCN10A (associated with both gain and loss of Nav1.8-based sodium channel function) have been associated with inherited arrhythmia syndromes such as Brugada syndrome [14,15,16], as well as increased AF susceptibility [11,12,13]. However, functional relevance of these identified mutations and variants has only been assessed in heterologous expression systems such as HEK-293 cells which differ significantly from the cardiomyocyte environment [10, 12, 13, 15, 16]. Nonetheless, even if a mutation in SCN10A is found to alter function of NaV1.8-based channels in expression systems, this does not automatically imply a functional impact on the cardiomyocyte level. In fact, our current and previous findings demonstrate very low expression levels of SCN10A and the consequent absence of functional NaV1.8-based current in atrial and ventricular CMs. We therefore propose that the potential electrophysiological and pro-arrhythmic effects of SCN10A variants/mutations do not occur on the cardiomyocyte level, but instead are consequent to the actions of SCN10A/NaV1.8 in other cell types such as intracardiac neurons. Indeed, we and others have previously demonstrated that NaV1.8 is specifically expressed and functionally relevant in murine, canine and human cardiac neurons [18, 20, 21], suggesting a function of the SCN10A gene product for cardiac conduction via regulation of action potential firing in intracardiac neurons [18, 21, 22].

In recent years, various GWAS studies have suggested a potential modulatory effect of SCN10A common genetic variants on ECG parameters such as PR and QRS interval [6,7,8,9,10] in addition to susceptibility to AF [11,12,13] and Brugada syndrome [14,15,16]. However, it was subsequently demonstrated that the SCN10A variant rs6801957 (associated with QRS duration [9]) is located within a cardiac enhancer region which interacts with the promotor of SCN5A. As such, rs6801957 was shown to decrease SCN5A expression in the heart, explaining the observed associations of this SCN10A variant with cardiac conduction [23]. Based on these observations and our current findings, it is therefore highly likely that the SCN10A locus identified in various GWAS studies exerts its modulatory effects indirectly through their impact on SCN5A expression and/or neuronal activity, rather than through a direct effect on cardiomyocyte electrophysiology.

While our findings indicate a lack of functional relevance for NaV1.8 in CMs under physiological conditions, they do not rule out a potential function during pathophysiological situations. Recently, increased SCN10A/NaV1.8 expression in human ventricular tissue isolated from heart failure and hypertrophic patients as compared to non-failing and healthy myocardium, has been demonstrated [39, 40]. NaV1.8 inhibition with the specific blockers A-803467 and PF-01247324 decreased late INa magnitude, abbreviated APD and reduced cellular-spontaneous Ca2+−release and proarrhythmic events in human failing and hypertrophic CMs [39, 40]. Of note, in both these studies, no electrophysiological experiments were performed in non-failing and non-hypertrophic CMs, thus precluding comparison of the effects of Nav1.8 inhibition in human non diseased CMs under comparable experimental settings [39, 40]. A modulatory role for SCN10A/NaV1.8 has furthermore been suggested in AF, with A-803467 administration preventing AF recurrence in a fast-pacing canine model [21]. Increased late INa has been observed in right atrial appendage cardiomyocytes from AF patients as compared to individuals in SR [38, 41]. Whether alterations in SCN10A/NaV1.8-based contribute to this increased late INa in the setting of AF will require further investigation. Interestingly, injection of A-803467 into canine cardiac ganglionated plexi (GP) [22] and canine left stellate ganglion (LSG) [42] suppressed vagal-mediated AF and ischemia-induced ventricular arrhythmia, respectively, most likely by inhibiting the neuronal activity of GP and LSG. This further underlines the potential functional involvement of SCN10A/NaV1.8 in intracardiac neurons.

In conclusion, our study demonstrates the (functional) absence of SCN10A/NaV1.8-based channels in human and rabbit atrial and ventricular CMs under basal, non-remodeled conditions. We therefore propose that the association of SCN10A variants with cardiac electrophysiology is likely the result of indirect effects on SCN5A expression and/or NaV1.8 activity in cell types other than CMs, including (intracardiac) neurons.