Figure 1. mRNA expression in transcripts per kilobase million (TPM) assessed by RNA-sequencing. A) KCNA5 (Kv1.5); B) KCNJ2 (Kir2.1); C-E) KCNN1-3 (SK1-3). Sample provenance: left ventricle (LV; black), right atrial appendage (RAA; red), left atrium (LA; blue). Patient’s health status: healthy (donor; triangles), dilated cardiomyopathy (DCM), ischemic cardiomyopathy (ICM), coronary artery disease (CAD), heart valve disease (HVD), atrial fibrillation (AF). Statistical significance was assessed by the Kruskal-Wallis test and multiple comparisons were conducted with Dunn's test.



In search for more efficacious and safe pharmacological treatment of atrial fibrillation (AF), atria-selective antiarrhythmic drugs have been developed, that target ion channels primarily expressed in atria. This concept allows utilising antiarrhythmic effects in atria, while at the same time protecting the ventricles against unwanted proarrhythmic side effects. Recently published evidence suggests that cardiac small conductance Ca²⁺-activated K⁺ (SK) channels represent such an atria-selective target in mammals including humans. However, there are conflicting results concerning the expression of SK channels in sustained AF. Moreover, latest findings suggest that SK channels would lose their atria-selectivity when patients develop heart failure.

We performed RNA-sequencing to compare expression levels of the three isoforms of SK channels (KCNN1, KCNN2, KCNN3) in human ventricular and atrial tissue from donor hearts (no cardiac disease), patients in sinus rhythm (with cardiac disease) and patients in AF. Kv1.5 (KCNA5) and Kir2.1 (KCNJ2) served as controls for atria-selective ion channel expression.

In atrial samples, expression levels of KCNA5 (p=0.001) and KCNJ2 (p=0.019) were lower and higher, respectively, than in ventricular tissue, confirming the known differential expression of these channels (Fig. 1 A and B). The SK channel subtype KCNN1 was significantly higher expressed in atrial than in ventricular donor samples (p=0.020), although overall expression was low (Fig. 1 C). For the subtypes KCNN2 (p=0.663) and KCNN3 (p=0.986) expression was not different in atrial and ventricular samples (Fig. 1 D and E). Moreover, we detected downregulation of KCNN2 (p=0.030) in atrial tissue from AF patients, as previously reported but not for KCNN3 (p=0.920). When comparing heart failure to donor samples, we found significantly increased ventricular expression of KCNN3 (p=0.019) as previously published, but not for KCNN2 (p=0.689).