A common genetic variant within SCN10A modulates cardiac SCN5A expression

Abstract

Variants in SCN10A, which encodes a voltage-gated sodium channel, are associated with alterations of cardiac conduction parameters and the cardiac rhythm disorder Brugada syndrome; however, it is unclear how SCN10A variants promote dysfunctional cardiac conduction. Here we showed by high-resolution 4C-seq analysis of the Scn10a-Scn5a locus in murine heart tissue that a cardiac enhancer located in Scn10a, encompassing SCN10A functional variant rs6801957, interacts with the promoter of Scn5a, a sodium channel-encoding gene that is critical for cardiac conduction. We observed that SCN5A transcript levels were several orders of magnitude higher than SCN10A transcript levels in both adult human and mouse heart tissue. Analysis of BAC transgenic mouse strains harboring an engineered deletion of the enhancer within Scn10a revealed that the enhancer was essential for Scn5a expression in cardiac tissue. Furthermore, the common SCN10A variant rs6801957 modulated Scn5a expression in the heart. In humans, the SCN10A variant rs6801957, which correlated with slowed conduction, was associated with reduced SCN5A expression. These observations establish a genomic mechanism for how a common genetic variation at SCN10A influences cardiac physiology and predisposes to arrhythmia.

Figure 1. Contact profiles of enhancers and promoters of the Scn10a-Scn5a locus in combination with ChIP-seq data.

(A) UCSC genome browser view of chr3:38,465,426–38,861,154 (hg18), demonstrating distinct LD blocks within the SCN10A-SCN5A locus with SNPs tagged in GWAS studies (green) (–9). (B) Mouse region (mm9; chr9:119,303,698–119,662,489) depicted with 4C analysis showing contact profiles of EnhA in Scn10a (blue), Scn5a promoter C (green), and EnhB downstream of Scn5a (red). Blue, green, and red arrows correspond to the position of the different viewpoints. Black arrows represent the position of the promoter of Scn10a and alternate promoters of Scn5a (promoters A and B). From the EnhA viewpoint, interactions can be observed between EnhA, EnhB, and the Scn5a promoter regions. From the Scn5a promoter C viewpoint, contacts can be observed with EnhA and EnhB and weakly with the Scn10a promoter region. From EnhB, contacts with the Scn5a promoter region and, weakly, with EnhA can be seen. See Supplemental Figure 1 for quantitative assessment of these interactions. (C) UCSC genome browser views of ChIP-seq data of the TBX3 (21) and of Pol2 and p300 (–29) binding profiles aligned with the 4C traces, showing the correspondence between contacts and factor occupancy. (D) Model for the interactions of EnhA (blue) and EnhB (red) with the Scn5a and Scn10a promoters (green and gray, respectively). The promoter of flanking gene Exog does not interact.

Figure 2. Scn5a and Scn10a expression analysis.

(A) Expression of Scn5a in the 12-week-old adult mouse heart was 145-fold higher than that of Scn10a (53.81 ± 5.39 versus 0.37 ± 0.08 rpm) by mRNA-seq. (B) CCS expression of Scn5a in the 12-week-old adult mouse heart was significantly higher than that of Scn10a, whose expression was undetectable by qRT-PCR. All samples were analyzed in triplicate. (C) mRNA-seq analysis of adult human heart tissue. SCN5A was expressed 1,357-fold higher than SCN10A in the whole heart (326 ± 149 versus 0.24 ± 0.22 rpm), 1,753-fold higher in the atria (205 ± 138 versus 0.12 ± 0.18 rpm), and 1,299-fold higher in the ventricles (376 ± 126 versus 0.29 ± 0.21 rpm). Data are mean ± SD. ***P < 0.001, Student’s t test.

Figure 3. EnhA and EnhB are necessary for Scn5a cardiac expression.

(A) Modified murine BAC RP23-198L19 with LacZ inserted into the endogenous Scn5a translational start site. (B–F) Stable BAC transgenic lines, shown in whole-mount and cross-section histology; arrows indicate AV bundle (blue) and distal bundle branches (red). Pie charts show the distribution of LacZ expression in each genotype class. All studies were performed at 12 weeks of age; the number of independent transgenic lines examined is indicated (each analyzed in triplicate). (B) The WT enhancer demonstrated robust CCS and myocardial expression. (C) EnhA deletion eliminated proximal CCS and myocardial expression. (D) EnhB deletion eliminated distal CCS and myocardial expression. (E) Deletion of both enhancers eliminated CCS and myocardial expression entirely. (F) The minor allele at rs6801957 markedly altered Scn5a-LacZ expression, which was either entirely absent or absent from the atrioventricular bundle and confined to the distal ventricular septum in the majority of cases.

Figure 4. SNP rs6801957 modulates EnhA activity.

Transient transgenic embryos at E14.5 harboring human 2.2-kb EnhA with major (A) and minor (B) allele of SNP rs6801957. Human EnhA with the major allele (n = 9) was sufficient to drive CCS expression (A), but human EnhA with the minor allele (n = 4) was not (B). (C) Correlation of SCN5A expression in human hearts with SNP rs6801957 genotype. SCN5A expression was assessed by RNA-seq in 42 human cardiac tissue samples, of which 19 were homozygous GG, 18 were heterozygous GA, and 6 were homozygous AA at locus rs6801957. Homozygous GG individuals had significantly higher SCN5A expression (428 ± 128 rpm) than GA (331 ± 104 rpm) or AA (322 ± 66 rpm) individuals. Each A allele reduced expression by 12% ± 6% (P = 0.01). A linear additive regression model was used to determine statistical significance. Lines within boxes represent median value; boxes represent interquartile range; whiskers represent 5th and 95th percentiles.