Dilated cardiomyopathy (DCM) is associated with significant morbidity and mortality, and is caused by inherited gene variants in a substantial proportion of cases. Familial DCM is clinically variable and genetically heterogeneous, with at least 40 disease genes reported to date (1). Identification of the genetic basis of DCM provides an opportunity for early diagnosis and preventative intervention in genotype-positive family members. However, the current reality is that mutations in most of the known disease genes are uncommon, and the molecular defects underpinning DCM in the majority of families (>70%) are unknown ((1),2). Moreover, no effective disease gene-targeted therapies have been established for clinical use.

Mutations in the SCN5A gene that encodes the cardiac sodium (Na⁺) channel alpha-subunit, Nav1.5, cause a variety of arrhythmic disorders, including long QT syndrome, Brugada syndrome, ventricular tachycardia, sick sinus syndrome, atrial standstill, conduction system abnormalities (CD), and atrial fibrillation (AF). SCN5A mutations have also been associated with DCM that is typically preceded by a prodrome of CD or AF, a similar phenotype to that observed with mutations in the LMNA gene, which encodes nuclear lamin A/C ((3),(4),5).

A large Caucasian kindred with a clinical diagnosis of DCM and CD was referred to our laboratory for molecular genetics analysis. Detailed phenotype evaluation demonstrated that multiple electrocardiographic (ECG) abnormalities were present. In particular, frequent polymorphic ventricular ectopy was a prominent and early manifestation, raising the possibility of a causal relationship with the development of DCM. We re-sequenced the SCN5A and LMNA genes and identified a heterozygous missense R222Q SCN5A variant that segregated with disease status in the kindred. Functional characterization of the R222Q variant showed an activating effect on cardiac Na⁺ channel function, and in a ventricular cell model, these effects were predicted to induce premature ventricular complexes (PVCs) predominantly of Purkinje fiber cell origin. The availability of genotype results changed the medical management of affected family members, and administration of drugs with Na⁺ channel-blocking properties enabled both ventricular ectopy and DCM to be reversed.

Informed written consent was obtained from all participants, and the study protocol was approved by the institutional Human Research Ethics Committee. The proband and participating family members >16 years of age were evaluated by history and physical examination, 12-lead ECG, and transthoracic echocardiography. The results of additional cardiac investigations performed for clinical indications, including Holter monitor and electrophysiology studies (EPS), were obtained from medical records. One hundred unrelated healthy Caucasian volunteers comprised the control group.

Genomic DNA was isolated from peripheral blood samples. Protein-coding sequences of the SCN5A and LMNA genes were amplified by polymerase chain reaction and re-sequenced using an ABI PRISM 3730 DNA Analyzer (Applied Biosystems, Foster City, California). The R222Q SCN5A substitution results in loss of a Hinf1 site and was independently confirmed by restriction enzyme digestion.

Chinese hamster ovary (CHO) cells were transfected with cDNA clones encoding wild-type (WT) and R222Q Nav1.5 plus WT Navβ1. Cardiac Na⁺ channel currents (I_Na) were recorded using conventional patch-clamp techniques. To investigate the functional consequences of mutant channels, WT and R222Q Nav1.5 channels were modeled using Hodgkin-Huxley formalism as described (6), and WT and mutant models were incorporated into Purkinje cell (6) and ventricular cell (7) models (Online Appendix).

Forty-two individuals in 2 related large kindreds (Family HY) underwent clinical evaluation and genetic testing ((Figure 50_gr1),Table 1). The family phenotype was characterized by a high prevalence of atrial and ventricular arrhythmias, with DCM occurring mainly in males. A striking feature of the ECG tracings of family members studied in sinus rhythm was the relative paucity of normally conducted sinus beats, with the majority of beats being PVCs, including narrow PVCs of probably high septal origin that had varying morphology and axis, as well as wide PVCs of left and right bundle branch type ((Figure 50_gr1)B to Figure 50_gr1D). Premature atrial complexes (PACs) and accelerated junctional rhythms were also seen. Six individuals had documented AF, and 3 individuals received pacemakers for symptomatic bradycardia or complete heart block in later adult life. EPS results were available in 4 cases and uniformly showed multiple PVC foci in the left and right ventricles, with no inducible ventricular arrhythmias. Five individuals had prophylactic implantation of cardioverter-defibrillators (ICDs).

Eight individuals had a diagnosis of DCM (Table 1). In 2 asymptomatic young males, DCM was detected only as a result of family screening. A history of palpitations that predated DCM diagnosis was elicited in all other cases. DCM was present in 7 of the 10 genotype-positive males but in only 1 of the 7 genotype-positive females. Myocardial fibrosis was excluded in 2 individuals with severe DCM by magnetic resonance imaging (IV-27) and cardiac biopsy (IV-34), respectively.

The coding regions of the SCN5A and LMNA genes were re-sequenced in the proband's DNA (III-10; Figure 50_gr1A) and a heterozygous 665G>A change in the SCN5A gene was identified that alters the amino acid at codon 222 from arginine (R) to glutamine (Q). The R222 residue is located in a voltage-sensing S4 region of Nav1.5 and has high homology across different members of the human Na⁺ channel gene family and different species (5). Fourteen individuals were genotype-positive, and 3 deceased individuals (II-5, III-5, III-7) were obligate carriers. Sixteen of these 17 gene variant carriers were clinically affected, with the exception being a 56-year-old male (III-12) who had a normal ECG and echocardiogram. R222Q is a recurrent rare variant identified in 4 DCM families ((8),(9),(10),11). This variant was not found in 200 control chromosomes (this study), in 506 control chromosomes in a previous report (8), or in the 1000 Genomes and dbSNP databases. None of the R222Q carriers had the common SCN5A variant, H558R, which has been shown to modify the functional effects of R222Q in vitro (12).

Given its location in a voltage-sensor domain, the R222Q variant is predicted to affect channel gating. CHO cells expressing WT Nav1.5 or R222Q Nav1.5 had similar mean current densities, but in R222Q channels, the midpoints of the voltage-dependence of activation and inactivation were left-shifted by 6.3 and 6.2 mV, respectively (Figure 50_gr2A). Other kinetic parameters were not altered (5). A consequence of the left-shifted activation and inactivation curves was that the area beneath the intersection of these curves (the “window” current) spanned a more negative membrane potential area (Figure 50_gr2A, inset). To determine the effects of increased window current, we examined the responses of WT and R222Q channels to a ramp voltage-protocol (13) analogous to the diastolic depolarization seen in Purkinje fiber cells. During a voltage ramp, in which the membrane potential was gradually changed between –120 and +40 mV at 0.16 mV/ms, the normalized current response of the R222Q channels opened at more negative potentials and had relatively greater peak current than that observed in WT channels (Figure 50_gr2B).

To investigate the functional consequences of having Na⁺ channels open at more negative potentials, we incorporated the R222Q Nav1.5 properties into a mathematical model of a Purkinje fiber cell (6). The original Stewart model displays automaticity with a frequency just below 1 Hz. When paced with 1-ms stimuli of 52 pA/pF at a frequency higher than the intrinsic rate of the model, each stimulus is captured and evokes an action potential. In contrast, the heterozygous WT/R222Q model showed a higher rate of spontaneous activity (79 beats/min), and pacing at 1 Hz resulted in grossly disorganized action potential patterns (Figure 50_gr2D). Similar to the WT model, pacing at higher rates in the R222Q model resulted in stimulus capture and a regular 1:1 relationship with each action potential (Figure 50_gr2E). These findings suggest that PVCs associated with the R222Q variant may result from increased automaticity of Purkinje fiber cells, and that these effects are rate dependent. To simulate the effects of Na⁺ channel-blocking drugs, decremental I_Na densities were evaluated in WT and WT/R222Q heterozygote model cells. A 33% reduction of I_Na decreased the spontaneous rate of the WT/R222Q heterozygote cell to 65 beats/min, and a 50% reduction decreased the rate to 50 beats/min, which enabled the cell to regain regular pacing with a 1-Hz stimulus (Figure 50_gr2D). To determine whether the effects of R222Q Nav1.5 were specific for Purkinje fiber cells, we also incorporated our modified Na⁺ channels into a ventricular cell model (7). There were no significant differences between WT and homozygous R222Q model cells when paced at rates between 2 and 0.5 Hz. We could, however, elicit a subtle difference by introducing an artificial current stimulus, analogous to a delayed afterdepolarization, during the diastolic interval. A smaller current stimulus could elicit a premature action potential in the R222Q but not in WT cells (5).

The rate-dependence predicted by the modeling studies was supported by clinical observations in affected family members with PVC frequency increased during periods of low heart rate at rest and at night, and reduced by high heart rates during exercise. During EPS, acute reductions in PVC numbers were seen following rapid pacing (cycle length <450 ms), or intravenous administration of isoproterenol, metoprolol (doses >5 mg), or flecainide. All family members diagnosed with DCM were treated with standard heart failure therapy, beta-blockers, and/or angiotensin-converting enzyme inhibitors, but these drugs had modest or no benefits ((Figure 50_gr3)E and (Figure 50_gr3)F, 5). In contrast, addition of amiodarone (IV-10, IV-27) or flecainide (IV-21, IV-34), drugs that have Na⁺ channel-blocking properties ((14),15), resulted in a dramatic reduction of PVC numbers and recovery of normal left ventricular function in those individuals with DCM ((Figure 50_gr3)A to (Figure 50_gr3)F, 5). The effects on PVCs occurred early, after treatment was started, whereas left ventricular remodeling changes occurred more slowly and took approximately 6 months.

R222Q is a recurrent SCN5A variant identified to date in 4 large kindreds with DCM. In this study, we found that the clinical phenotype associated with R222Q SCN5A included multiple arrhythmias, suggestive of enhanced automaticity in the atrium and ventricles. Our data provided new perspectives on the biophysical properties of R222Q Nav1.5 and showed that it has activating effects on I_Na, which in the ventricle predispose to PVCs mainly of Purkinje fiber origin. These findings raise interesting questions about the cause of DCM and highlight the value of genotype information as a guide to family management.

The high prevalence of PVCs was a prominent feature of this family's phenotype. PVCs are ectopic impulses that arise from ventricular myocardium and can result from focal sites with increased automaticity, re-entrant circuits that occur in regions of heterogeneous conduction (e.g., border zones between infarcted and normal tissue) or from triggered activity due to early or late afterdepolarizations. Although PVCs are considered benign in healthy adults ((16),17), PVCs that are frequent (>60/h) or complex (multiple morphologies or repetitive patterns) have been associated with an increased long-term risk of malignant ventricular arrhythmias and sudden death ((17),18). PVCs have a higher prevalence in patients with reduced ejection fraction, and although often regarded as a complication of DCM, there has been increasing recognition that a high burden of PVCs may independently impair cardiac contraction and promote DCM (“ventricular ectopy-induced DCM”) ((19),20). The risk of developing DCM increases with PVC burden, with a cutoff value of >24% estimated in 1 study (20). This threshold level can be variable, however, and lower numbers of PVCs may be sufficient to promote DCM in individual cases. The precise mechanisms underpinning ectopy-induced DCM are unclear, and may be related to factors such as dyssynchronous myocardial contraction or increased sympathetic nervous system activation (21).

PVCs of left and right ventricular origin have previously been associated with SCN5A mutations (22). Patch-clamp and modeling data in Family HY point to hyperexcitability of Purkinje fiber cells due to increased I_Na window current as a mechanism for PVC genesis. This differs from gain-of-function SCN5A mutations associated with long QT syndrome 3 that disrupt fast inactivation of I_Na, prolonging the action potential duration and increasing the propensity for early afterdepolarizations (15). In Family HY, there were multiple PVC morphologies, with a high prevalence of narrow complexes of probable high septal fascicular origin (23), as well as wider complexes suggesting distal Purkinje fiber or ventricular origin. These multiple morphologies made estimation of the daily PVC numbers difficult to assess on Holter monitor recordings because ventricular and supraventricular complexes were not always readily distinguished. PVCs may also be confused with aberrantly conducted PACs, which were demonstrated during EPS in 1 family member. Nevertheless, ECG and Holter tracings clearly showed a massive number of PVCs, raising the possibility that ventricular ectopy could have an important role in causing DCM. Factors supporting ventricular ectopy-induced DCM include the history of palpitations that preceded DCM in many cases, and the dramatic reductions in PVC numbers before improvements in contraction as a result of amiodarone or flecainide therapy. PVCs are commonly seen in patients with myopathic hearts, and it is equally plausible that at least some of the PVCs, especially those of ventricular origin, might have arisen as a consequence of DCM. Similar arguments can be made for the contribution of AF, which may be a cause or effect of DCM. Only 2 of the 6 family members with AF developed DCM, and there was no clear history of AF predating DCM in either case.

Altered I_Na activity may directly impair ventricular function independently of effects of PVCs or AF. Many of the SCN5A variants associated with DCM reduce I_Na density or exhibit rate-dependent reductions of Na⁺ current ((12),(24),(25),26). Reductions in myocardial I_Na have been observed in human heart failure (27) and in transgenic mice with DCM due to overexpression of the transcriptional repressor, Snail (28). Previous cellular electrophysiology studies have shown that the R222Q variant does not reduce I_Na when expressed alone in transfected cells (12), and our data are concordant with this. R222Q Nav1.5 was reported to reduce I_Na when co-expressed with the common H558R SCN5A variant (12). However, because none of the genotype-positive members of Family HY were H558R carriers, the modifying effects of this variant were not clinically relevant. Gain-of-function effects on I_Na have been seen with some DCM-causing SCN5A variants, including the R814W variant, which increases window current (26), analogous to our findings with R222Q. It has been proposed that the increased window current results in augmented diastolic Na⁺ fluxes, and that this could promote DCM by causing regional heterogeneity in myocardial impulse propagation or by altering intracellular Na⁺ and Ca²⁺ homeostasis. Although myocardial structural defects can contribute to DCM associated with SCN5A variants (29), ventricular fibrosis was excluded in 2 members of Family HY with severe DCM who underwent magnetic resonance imaging and cardiac biopsy, respectively.

An intriguing question in Family HY is why only half of the genotype-positive individuals developed DCM. There was a clear sex difference, with DCM occurring in 6 of the 7 living clinically affected males, but only in 1 of 6 females. Sex differences in ion channel gene expression (30) and in ECG parameters (31) have been observed, and may account for differential susceptibility to ventricular arrhythmias. From adolescence onwards, women have faster heart rates and longer QTc intervals, with less intraventricular conduction disturbances, QRS prolongation, and QT dispersion than men. Although these changes may increase the risk of arrhythmias associated with long QT-causing SCN5A variants (32), they may be relatively protective in the setting of the R222Q variant, which was predicted to have less Purkinje fiber excitability at high heart rates.

The cause of the CD in a subset of family members is unclear. Conduction system defects are often seen with loss-of-function SCN5A mutations, but may occur with activating SCN5A mutations ((24),(25),33). Age-related degenerative effects and/or changes in Nav1.5 expression, as well as additional genetic variants may be contributing factors.

Clinical guidelines for familial DCM recommend that affected individuals receive standard pharmacological management as indicated by the severity of heart failure and its complications. No specific gene-based treatment strategies have been devised (34). Conventional heart failure therapies were relatively ineffective, however, in members of Family HY, whereas drugs with Na⁺ channel-blocking properties enabled the phenotype to be reversed. Amiodarone has multiple actions, including effects on K⁺, Ca²⁺, and Na⁺ channels (14), and is often used to treat ventricular ectopy. Its long-term use may not be tolerated, however, due to the high risk of adverse effects (35). Flecainide is an alternative option, because it is a more selective Na⁺ channel blocker with a lower side-effect profile overall, but this drug is generally contraindicated in patients with reduced left ventricular function due to potential proarrhythmic effects (35). Flecainide was well-tolerated and markedly reduced PVC numbers in 2 family members, 1 of whom had DCM and an ICD. ICDs were implanted prophylactically in 5 family members with frequent complex PVCs, but the availability of phenotype-reversing drug treatment may obviate this need for early ICD implantation. One prudent approach to management of family members without ICDs would be to start with amiodarone until DCM resolves, then switch to flecainide for ongoing maintenance therapy. The responses to amiodarone or flecainide therapy were quite remarkable, and the kindred described here provided an exemplary case in which elucidation of the genetic basis for familial forms of DCM can enable effective disease-modifying therapy to be implemented.