Primary arrhythmogenic disorders of the heart are a major cause of sudden cardiac death in otherwise healthy individuals with structurally normal hearts. Mutations in several genes encoding ion channels and channel-interacting proteins have emerged over the last decade as the basis for a variety of inherited cardiac arrhythmias (1- 3). Different mutations in a single gene may account for various disorders depending on the position of the mutation, the nature of the amino acid substituted, the type of the mutation (non-sense, false-sense…), and its functional consequences (4- 5).

The SCN5A gene encodes the pore-forming subunit of the voltage-gated Na⁺ channel Nav1.5. In the heart, this channel plays a key role in rapid impulse propagation through the conduction system and in the excitability of atrial and ventricular cardiomyocytes. Mutations in SCN5A are responsible for a spectrum of hereditary arrhythmias, including type-3 long QT (LQT3) syndrome, the Brugada syndrome, cardiac conduction disease, sinus node dysfunction, atrial fibrillation, and dilated cardiomyopathy (see Tfelt-Hansen et al. [6] for a review) (5,7- 13). In addition to these syndromes, originally regarded as independent entities, recent evidence shows considerable clinical overlap, implying new disease entities known as overlap syndromes of the cardiac Na⁺ channel (14).

In this article, we report a novel autosomal dominant form of cardiac arrhythmia identified in 3 unrelated families and characterized by multifocal ectopic Purkinje-related premature contractions (MEPPC). This arrhythmia is due to the c.665G>A SCN5A mutation leading to a shift in p.Arg222Gln (R222Q) Nav1.5 voltage dependency.

An expanded Methods section is available in the Online Appendix.

This study was in agreement with the local guidelines for genetic research and has been approved by the local ethical committees. Informed, written consent was obtained from each family member who agreed to participate to the study. Standard 12-lead electrocardiography, Holter recording, and echocardiography were proposed to all participating family members. We also proposed electrophysiological studies to some patients.

Patients were considered affected when ventricular arrhythmia corresponding to MEPPC was detected.

Genomic DNA was extracted from peripheral blood lymphocytes using standard protocols. The proband (III.1) of Family 1 was screened for mutations in known genes responsible for dilated cardiomyopathy (DCM) and arrhythmia, including LMNA encoding lamin A/C, ABCC9 encoding SUR2A, and SCN5A (10,15- 16).

Ten microsatellite markers around the SCN5A gene (D3S1759, D3S2432, D3S3047, D3S3512, D3S1298, intragenic SCN5A marker, D3S3521, D3S3527, D3S3522, and D3S3559) were genotyped. We estimated the age of the mutation using the Genin et al.'s method (17).

Patch-clamp studies were performed on COS-7 cells transiently expressing human wild-type (WT) or R222Q Nav1.5 (NG_008934) and human β1 subunits, using the whole-cell configuration at room temperature (18).

Single-cell models of the human Purkinje cells (19) and left ventricular myocytes (20) were used. Both models were incorporated in a multicellular model and the propagation of the electrical waves in the cardiac tissues was described by a monodomain model (21). The numerical simulations were performed on a simplified 2D slice model described previously (22).

All data are presented as mean ± SEM. The statistical significance of the observed effects was assessed by the Student t test or 2-way analysis of variance (ANOVA) followed by a Tukey test for multiple comparisons when needed. A p value <0.05 was considered significant.

We identified 3 families affected by the same phenotype, segregating with an autosomal dominant pattern over 3 generations (Figure 1A). In Family 1, 11 of 18 family members were affected, 4 of 5 members were affected in Family 2, and 6 of 8 members were affected in Family 3. Patient characteristics are summarized in (Tables 1, 2).

Patient III.1 was identified at the age of 10 years after mild dyspnea while exercising. Her 12-lead surface electrocardiogram (ECG) showed a chaotic cardiac rhythm comprising narrow junctional and rare sinus beats competing with premature ventricular complexes (PVCs) showing various right bundle branch block (RBBB) patterns (Figure 2A). There was an overdrive suppression of the PVCs during treadmill exercising. Signal-averaged ECG did not show late potentials. A Holter monitoring device recorded more than 50,000 PVCs per 24 h. The transthoracic echocardiogram (TTE) revealed a mild dilation of the left ventricle (left ventricular diastolic diameter [LVDd] 54 mm or 40 mm/m², above 97th percentile) (23), but a normal left ventricular ejection fraction (LVEF) (62%). She remained asymptomatic until the age of 13 years, when she was referred for brief sudden loss of consciousness at rest. The ECG exhibited recurrent nonsustained ventricular tachyarrhythmias (NSVTs) with RBBB patterns, which were concomitant with several fainting episodes (Figure 2B). Several nonsustained supraventricular arrhythmias have been identified during the follow-up. New TTE showed a markedly enlarged LVDd (62 mm), and a decreased LVEF (32%). A dual-chamber implantable cardioverter-defibrillator (ICD) was then implanted associated with oral hydroquinidine treatment. Once the measured plasma level of the drug reached 1 mg/l (normal range 1 to 3 mg/l), this medication succeeded in markedly reducing the number of PVCs (Table 2). After 1 year under hydroquinidine treatment, the PVC burden was <1% and her heart was considered normal in TTE (LVDd 54 mm, LVEF 56%).

Clinical investigation of family members allowed the identification of 10 other affected individuals with similar ECG characteristics (5).

Patient III.1 was identified at age 6 years because of an irregular heart rate without any other symptoms. His twin brother was diagnosed at 12 years with the same arrhythmia after experiencing syncope while cycling. A complete clinical evaluation including a TTE and magnetic resonance imaging was normal, but his ECG showed frequent PVCs with a left bundle branch block (LBBB) pattern, and superior and inferior axial deviation (Figure 2C). In both patients, the Holter ECG recorded tens of thousands of PVCs per 24 h (Table 1). Exercise allowed a clear decrease of their number. Both patients were finally treated with hydroquinidine, leading to a dramatic decrease of the PVC number (Table 2).

Their father and grandmother had both been affected with DCM associated with frequent PVCs, and 1 uncle, known as affected by atrial flutter, had died suddenly at age 11 years.

The index patient II.3 was identified at the age of 18 years during a routine medical examination. Her 12-lead ECG showed premature atrial complexes (PACs), atrial fibrillation (AF), pre-excitation on the ECG, and frequent relatively narrow PVCs with both LBBB and RBBB patterns. Electrophysiological testing revealed ectopic foci from the atria, the atrioventricular junction, and the bundle branches. The accessory pathway was located at the right septal site. At the age of 48 years, Holter monitoring recorded a permanent chaotic supraventricular rhythm with >48,900 PACs and >7,000 isolated and bigeminal PVCs (35% and 5% of total QRS complexes, respectively). Flecainide therapy marginally reduced ventricular ectopies; however, it completely suppressed supraventricular arrhythmia (Holter monitoring recorded <150 PACs, representing <1% of total QRS complexes). Flecainide therapy was maintained because hydroquinidine is not available in the Netherlands, where Family 3 lives.

Within these 3 families, 21 individuals were affected by PVCs. Patient characteristics are summarized in (Tables 1, 2). Despite the various ages at diagnosis, from 24 weeks of gestation to 62 years (mean age 20 years; n = 20), the phenotype was remarkably constant: narrow sinus and junctional QRS complexes (n = 17) competing with various complexes showing an RBBB or LBBB patterns (n = 15), corresponding to PVCs with superior or inferior axes. We did not observe any QT prolongation (Table 1) or ST-segment elevation.

PVCs were isolated, or paired (bigeminism) with sinus or junctional QRS complexes. NSVTs were identified in 8 individuals. Five patients were affected by syncope or presyncope. Sudden death was reported in one 4-month-old boy (Patient III.12 in Family 1), an 11-year-old boy (Patient II.3 in Family 2), and 3 adult males (29, 50, and 71 years old; Patients II.7 and I.1 in Family 1 and Patient I.1 in Family 3). However, PVCs were clearly identified in only 2 of these cases (Patient II.7 in Family 1 and Patient I.1 in Family 3) before sudden death. Episodes of syncope or fainting during polymorphic ventricular tachycardia (VT) episodes (Holter or ICD monitorings) were reported in 3 females (Figure 2B). Two individuals in Family 1, 1 in Family 2, and 6 in Family 3 presented with atrial arrhythmia (PACs, nonsustained atrial tachyarrhythmia, or AF). One patient had AV conduction disturbances. Only 1 patient in Family 3 had an accessory pathway.

Electrophysiological testing was performed in 7 cases ((Figure 3), 5). Pre-systolic potentials recorded at ectopic sites during PVCs either corresponded to Purkinje potentials during normal sinus beats or to the proximal extension of the left anterior and posterior fascicles (Figure 3B and 3C). In all tested patients, numerous ectopic foci were identified along the left conduction system and its proximal extension. We did not record any PVCs from the myocardium itself as the local ventricular electrograms were always preceded by a Purkinje potential and we did not record any retrograde Purkinje potential activation or slow conduction area in the Purkinje network. Reentrant mechanisms were excluded because no PVCs or VTs could be pacing induced and no mid diastolic potentials were recorded during VT (24- 26).

When tried on Patient II.4 of Family 1, radiofrequency applications did not eradicate PVCs as the whole Purkinje tissue was involved in triggering PVCs from varying sites. Whereas we proposed electrophysiological studies to most of the patients, not all of them accepted. Among affected members in the 3 families, PVC features were very similar with a sharp slope of the first ventricular deflection and close characteristics to authentic LBBB or RBBB patterns with right or left axis. In patients who underwent an electrophysiological study, we identified those PVC morphologies as being triggered from Purkinje fibers. We therefore inferred that similar PVC patterns may originate from similar Purkinje regions in the other patients.

Transthoracic echocardiograms were performed in all affected patients and magnetic resonance imaging was performed in 3 cases. Low LVEF with dilation of the left ventricle was identified in 7 patients. Several antiarrhythmic treatments were attempted within the families (Table 2). Amiodarone treatment was partially successful in 2 cases (Patient II.6 in Family 1 and Patient II.1 in Family 2). For patient II.1 in Family 2, amiodarone and cardiac resynchronization therapy defibrillator treatment were introduced simultaneously, partially normalizing the LVEF (from 20% to 45%). Hydroquinidine was successful in 5 cases with a dramatic decrease in the number of PVCs per 24 h (Table 2), and dramatically decreased the PVC number and improved the ventricular function in all cases. Patient III.1 was the first person to be treated by hydroquinidine in Family 1 and no adverse effects were noticed after 4 years of follow-up. Flecainide therapy was used in 5 patients and greatly reduced supraventricular and ventricular arrhythmias, with a clear decline in ventricular tachycardia episodes, in 3 of them. Other antiarrhythmic drugs were used in 4 patients but were successful in 3 cases only (Table 2).

An ICD was implanted in 4 patients, 2 of which had a cardiac resynchronization therapy defibrillator. Only 1 patient (Family 3) had atrioventricular conduction disturbances and was implanted with a double chamber pacemaker (Table 2).

Because the first proband identified (Family 1, Patient III.1) had DCM, mutations in LMNA encoding lamin A/C and ABCC9 encoding SUR2A, genes reported to be associated with DCM, were searched in this patient although unsuccessfully. The SCN5A gene, reported to be associated with DCM and various arrhythmogenic diseases, was also sequenced. Genetic testing of Family 1's proband revealed that she was heterozygous for a c.665G>A transition in SCN5A exon 6 (Figure 1B), resulting in the substitution of an arginine for a glutamine at position 222 (R222Q). This substitution is located in the voltage-sensing S4 segment of domain I of the cardiac Na⁺ channel Nav1.5 (Figure 1D). Subsequently, the mutation was detected in Family 2's (Patient III.1) and Family 3's (Patient II.3) probands who did not present DCM. The mutation was 100% penetrant and strictly segregated with the cardiac arrhythmia in all the families, consistent with a mutation-related disease (Figure 1A). It was absent in 600 control chromosomes. The Arg222 (R222) amino acid is highly conserved across species (Figure 1C). This mutation is reported on dbSNP (rs45546039) but at 0% in all populations and is not reported in the National Heart, Lung, and Blood Institute Exome variant server.

To investigate whether a common ancestral chromosome accounted for the recurrence of the R222Q mutation in unrelated Families 1, 2, and 3, we constructed mutation-associated haplotypes by genotyping family members for 10 microsatellite markers (Online Fig. 3, Online Appendix: Results).

Our data demonstrate that a founder effect for the R222Q mutation in these 3 families is very unlikely.

To investigate the functional consequences of the R222Q mutation on the Na⁺ channel activity, we used the whole-cell configuration of the patch-clamp technique (Online Appendix: Results). The presence of the mutation did not modify the Na⁺ current density ((Figure 4)A and 4B, 5). The activation curve was shifted toward more negative potentials in the presence of the mutation ((Figure 4)A, 5). Activation kinetics were accelerated in the mutant (5) (2-way ANOVA, p < 0.001 vs. WT). Inactivation voltage sensitivity was also changed ((Figure 4)B, 5), which, combined with the activation curve shift, predicted an increase of the window current availability (5A). The maximum conductance of the tetrodotoxin-sensitive window current elicited by depolarizing-voltage ramps was not significantly different in R222Q Nav1.5 expressing cells, but the voltage at which the maximum conductance was measured was significantly shifted toward more negative values and the window current availability was increased when expressed as the area under the current-voltage curve (t test p < 0.001 and p < 0.01, respectively) ((5), Figure 4C). Each of these changes caused a gain of function of the mutant channel. However, inactivation onset kinetics were also accelerated, causing a loss of function (5B) (2-way ANOVA p < 0.001 vs. WT). Finally, kinetics of recovery from inactivation were not significantly modified by the R222Q mutation (5).

In order to understand the effects of hydroquinidine treatment on PVCs occurrence, we evaluated the WT and R222Q Na⁺ current sensitivity to quinidine, its active form. At a concentration at which the WT peak I_Na was about half-reduced (30 μM in our conditions) ((5), Figure 4D), WT and R222Q peak currents were similarly inhibited (2-way ANOVA p < 0.001 vs. control). When quinidine was tested on the window current, the inhibition was also similar in both groups when measured at the maximum conductance potential of each group ((Figure 4)D, 5) (2-way ANOVA p < 0.001 vs. control). Therefore, WT and R222Q Na⁺ currents were equally sensitive to quinidine.

Because of the multiple and opposite effects of the mutation on channel activation and inactivation, its net effect on the action potential (AP) is not straightforward. To get insights on this effect, we carried out computer simulations and evaluated the impact of the mutation on AP in single-cell models of human Purkinje fibers and ventricles. Single-cell models of Purkinje and ventricular cell AP were run in WT and heterozygous conditions. In heterozygous conditions, incomplete repolarization occurred in Purkinje cell AP, correlated with an increased late Na⁺ current (5). The ventricular AP was much less affected (5).

To evaluate the possibility of a causal role of the mutation on the generation of PVCs, we built a multicellular model incorporating both human cell models. In the WT condition, 1-Hz stimulation of the Purkinje fibers generated AP that propagated to the ventricle ((Figure 5)Ba,Online Video 1A). Interestingly, in the heterozygous condition, incomplete repolarization in the Purkinje fibers triggered premature APs, propagating into the ventricles ((Figure 5)Bb, Online Video 1B). Consistent with the disappearance of clinically observed PVCs during exercise, the triggered ventricular APs disappeared at higher pacing frequencies (2 Hz) ((Figure 2)B, Online Video 2B). Altogether, these results: 1) strongly suggest that the gain of function of the Na⁺ current predominantly affects the Purkinje cells; 2) explain the fact that the entire Purkinje system was affected resulting in a wide variety of ectopic foci; and 3) explain the frequency dependency of the PVCs.

Finally, the effects of quinidine were tested on the multicellular model. Clinically relevant effects of this drug are thought to be due to alteration of the Na⁺ current, the transient outward current I_to (27- 28), and the repolarizing delayed rectifier current I_Kr (29- 30). In addition, quinidine also inhibits Ca²⁺ and other K⁺ currents, although in a lower extent (31- 32). Therefore, inhibition of these latter currents was not considered. In the model, the K⁺ currents were reduced in parallel with the heterozygous Na⁺ current. For the first test, we chose 10 μM quinidine, the maximal therapeutic dose (i.e., preserving 50% I_Na, 30% I_Kr [(33)], and 30% I_to [28]). In this case, we observed a normalization of the Purkinje and ventricular AP course at 1 Hz (not shown). To evaluate the effects of a more “therapeutic” dose of quinidine, we ran the model with lower current inhibition levels. As illustrated for the ventricular AP in (Figure 5)B, the normalization was still observed when the computed quinidine dose was reduced to preserve 75% I_Na and 45% I_Kr and I_to. However, when the computed quinidine dose was further reduced to preserve 85% I_Na and 50% I_Kr and I_to, triggered APs reappeared. Qualitatively, if not quantitatively, the multicellular model mimics the cardiac response to hydroquinidine treatment.

We report a novel autosomal dominant form of cardiac arrhythmia showing MEPPC. MEPPC syndrome is characterized by frequent PVCs originating from various ectopic foci along the fascicular-Purkinje system occasionally associated with NSVTs and sudden death. Some patients (n = 9) also presented with PACs, nonsustained atrial tachyarrhythmias, or paroxysmal AF. In some patients, arrhythmias were associated with mild DCM. Hydroquinidine markedly reduced the number of PVCs and normalized the left ventricular function in patients with DCM. We demonstrate that a fully penetrant SCN5A gain-of-function mutation is responsible for this new syndrome.

Ventricular arrhythmias originating from left ventricular fascicles have been described previously in patients with or without structural heart disease (ischemic or idiopathic cardiomyopathies) (24- 25,34- 35). Accounting for up to 30% to 50% of arrhythmias in nonischemic DCM, bundle branch re-entrant VTs usually have an LBBB pattern and enlarged QRS complexes with a prolonged H-to-V interval during sinus rhythm (34). Such conduction abnormalities were not identified in any of our patients.

Fascicular VT is the most common form of idiopathic left ventricular tachycardia occurring more frequently in young males without structural heart disease and with a particular sensitivity to verapamil (24). The ECG characteristics of our patients were different. MEPPC syndrome was distinguished from PVCs originating from the posterior papillary muscle in the left ventricle (36) mainly by the presence of high-frequency potentials preceding local myocardial activation suggesting that Purkinje system was involved. Idiopathic VTs and bundle branch VTs typically present a single ECG morphology originating either near the posterior or the anterior left fascicle. In our patients, ECG characteristics demonstrated the varying origins of the PVCs, in either the left or right ventricle. Electrophysiological testing using a 3D endocardial navigation system clearly demonstrated that the PVCs originated from more than 10 different foci along the extension of the left anterior and posterior fascicles. This phenomenon was responsible for important changes in PVCs morphology (shape and axis) from one beat to another. Similar to triggered activities, the firing was fairly increased by isoproterenol infusion despite the increase in sinus rhythm at a lesser extent. As opposed to idiopathic fascicular VTs, which are well known to be highly sensitive to verapamil, this treatment had no effect on the tested affected family members. Atrial arrhythmias were less frequently observed than PVCs.

Several findings support the causative role of the R222Q SCN5A mutation in the pathogenesis of the disease in the 3 reported families: 1) the complete cosegregation of the mutation with the disease phenotype; 2) the identification of the same mutation in 3 unrelated families with a remarkably uniform phenotype; 3) the absence of the mutation in 300 ethnically matched control individuals; and 4) the high conservation of the R222 residue across species (Figure 1C).

R222Q has been reported in a single patient with a Brugada syndrome (37). With the help of Dr. Xavier Waintraub (Hôpital Pitié-Salpêtrière, Paris, France) and the kind permission of Dr. Pascale Guicheney (INSERM UMR S956, Paris, France), we had the opportunity to re-evaluate the phenotype of this patient. He was clearly not affected by the Brugada syndrome, his ECG presenting no ST-segment elevation in the precordial leads. His phenotype appeared to be very closed to that of MEPPC syndrome with permanent junctional activity, very frequent multifocal ventricular premature beats (57,000 thin ventricular premature beats per day coming from left anterior and posterior fascicles), atrial arrhythmias, and dilated cardiomyopathy with moderated alteration of the ventricular contraction (LVEF 56%) (X. Waintraub, personal communication, September 2011). His familial history found that his mother and grandmother were also affected by DCM associated with frequent PVCs. None of the ECGs performed in patients harboring the SCN5A R222Q mutation met the ST-segment criteria for a Brugada syndrome.

The same SCN5A mutation was also detected in a cohort of 2,500 unrelated cases referred for LQTS genetic testing (38). However, none of the patients from the 3 families here presented a prolonged QTc. This may be explained by the relatively less impaired biophysical properties of the R222Q channel, mostly a negative shift of the window current, when compared with that of already reported LQT3-related mutants (5,39).

R222Q effects that we report on channel parameters were similar to those measured by Cheng et al. (40). In addition, we show that these effects are intermediate in the heterozygous state and that they also impair the window current, crucial during the AP plateau phase.

Looking for a mechanistic link between the genotype and the phenotype, we simulated the effect of the mutated channel activity on the Purkinje and ventricular electrical activity. Two points suggest that the shift in Nav1.5 voltage dependency is responsible for the premature ventricular contractions originating from the fascicular-Purkinje system. First of all, 1 of the particularities of this arrhythmia is that ectopic foci were observed in erratic locations of the Purkinje system, which made radiofrequency applications ineffective. The multicellular model shows that the impaired function of the Na⁺ channels, anywhere in the Purkinje system, is sufficient enough to trigger premature APs in the surrounding ventricular tissue. Second of all, consistent with the disappearance of MEPPC clinically observed during exercise, the triggered ventricular APs disappeared at higher pacing frequencies.

Therefore, the model confirms the mechanism by which the Na⁺ channel alteration is responsible for the patients' phenotype.

SCN5A R222Q mutation was already detected in a few patients referred for familial or idiopathic DCM (13,40- 42). However, in all previously reported R222Q carriers, DCM was always associated with MEPPC-related ECG patterns. McNair et al. (13) identified 5 SCN5A mutations among 338 patients affected by DCM (13). Among them, 2 were carriers of the R222Q mutation and in both cases patients were affected by frequent PVCs (>1,000/h) and NSVT. Morales et al. (42) investigated patients affected by peripartum cardiomyopathy and identified a family carrying the R222Q mutation, affected by DCM (42). In this family, 4 of 6 carriers were also affected by ventricular or supraventricular tachycardia (for 1 patient no information was available). In the families reported here, DCM was diagnosed in 6 individuals, and is very likely a consequence of the arrhythmia, and not directly linked to the mutation. Indeed, 13 patients affected by MEPPC, carriers of the R222Q SCN5A mutation, were not affected by DCM. Most importantly, the cardiomyopathy recovered at least partially under antiarrhythmic treatment after having reduced the number of PVCs. The low penetrance of the associated DCM phenotype in these families suggests that the genetic background may play a role in the clinical evolution of the disease as now admitted even for monogenic cardiac arrhythmias (43).

Several other SCN5A mutations associated with familial DCM-arrhythmia syndrome have also been described (R814W, D1275N, T220I, and D1595H) (10- 12). In these mutation carriers, predominant clinical findings were mainly sinus bradycardia, atrioventricular block, atrial fibrillation, and/or atrial flutter, phenotypes quite different from the MEPPC syndrome. Therefore, our work does not rule out that DCM may be a primary consequence of other SCN5A mutations.

Because the treatment of the affected patients with hydroquinidine has been proven to be effective on PVCs, we investigated the effects of quinidine on the Na⁺ currents and observed equivalent inhibition of WT- and R222Q Nav1.5–mediated currents. Despite the fact that we did not consider the use dependence of quinidine activity (44- 45), the multicellular model recapitulates the effects of quinidine on PVCs suppression. Altogether, these results reinforce our hypothesis of MEPPC being the primary consequence of SCN5A R222Q mutation. Furthermore, it offers a potentially efficient therapy to limit the symptoms for SCN5A R222Q carriers. However, it remains to be shown that the treatment may limit the risk of sudden death.

The authors thank André Terzic, University of Minnesota, Minneapolis, Minnesota, for ABCC9 sequencing and Béatrice Leray, l'institut du thorax, Nantes, France, for expert technical assistance with cell culture.

For expanded Methods and Results sections, as well as supplemental tables, figures, and videos, please see the online version of this article.