**2. Clinical Evidence**

Early studies reported pathogenic/likely pathogenic *SCN5A* variants in nearly 2% of all DCM cases [14]. Both sporadic cases and familial forms with autosomal dominant inheritance have been reported [12,14]. Variants associated with DCM have been localized to cytoplasmic, extracellular and transmembrane domains (DI-DIV) of NaV1.5 (Figure 1). The main findings that support a potential role in DCM are the familial aggregation of the trait and the segregation of *SCN5A-*variants with clinical phenotype and/or histological characteristics of DCM, with or without associated electrical abnormalities (although with reduced penetrance). A recent study, however, reported no excess variation in *SCN5A* in DCM cases versus an Exome Aggregation Consortium (ExAC) control population, suggesting that most variants in this gene are unlikely to cause DCM [17]. Instead, it has been noted that *SCN5A* is in fact one of the genes with highest background variation. Thus, establishing a causal role of an *SCN5A* variant in DCM requires very strong functional evidence of pathogenicity and/or segregation with phenotype in large pedigrees.

**Figure 1.** Rare variants in cardiac NaV1.5 voltage-gated sodium channel (*SCN5A*) reported in association with dilated cardiomyopathy. Two additional variants, c.2550-2551insTG and c.3318dupC, causing truncation of the encoded protein in patients with dilated cardiomyopathy, are not shown. Pathogenic/likely pathogenic variants are shown in red, variants with insufficient evidence of pathogenicity in dilated cardiomyopathy are shown in red/white.

#### **3. Experimental Evidence**

Most of the reported DCM-associated *SCN5A* variants are missense variants, with a predilection for location in the S3 and S4 transmembrane domains, implicating a disruption of voltage-sensing mechanisms [15]. In vitro studies have shown that these variants commonly have loss-of-function, or infrequently gain-of-function, or rarely combined loss-of-function and gain-of-function effects on NaV1.5 activity [16,18–21]. Additionally, conserved variants R225P and R814W localized at the S4 of DI and DII, respectively, which were associated with an atypical phenotype combining cardiac conduction disturbances, Brugada syndrome or long QT phenotype and DCM, were shown to result in the creation of an alternative permeation pathway through the normally non-conductive voltage sensor domain (gating pore current) [22]. Yet, there is practically no in vivo or in vitro evidence directly linking *SCN5A* defects to DCM. In other words, the existing evidence on biophysical phenotypes of rare *SCN5A* variants demonstrates their potential as a substrate for arrhythmias and conduction disturbances, but does not clearly point to the pathogenesis of DCM. Digging into the thought-provoking hypotheses

regarding how *SCN5A* defects might lead to ventricular dysfunction and heart failure might shed some light on this uncertainty.
