Analysis of the murine conduction system revealed very low Scn5a mRNA and Na_v1.5 protein expression in the sinoatrial node (SAN) and atrioventricular node (AVN). Previously, low Na_v1.5 expression levels in the central atrioventricular node versus high levels of Na_v1.5 in the peripherally located transitional cells have been observed in rat and rabbit, and thus may be considered species independent [33, 47]. The low abundance of Na_v1.5 in the SAN is in accordance with the relatively low expression of Scn5a mRNA observed in the adult mouse SAN region measured by quantitative RT-PCR [30], and the previously reported low degree of Na_v1.5 staining in the SAN of the embryonic and adult mouse heart [22, 28]. We have now confirmed the presence of low Scn5a mRNA expression in the SAN using in situ hybridization, further validating these observations. Functionally, the brain-type sodium channel, and not the cardiac-type, appears essential for heart rate regulation, at least in the mouse [28]. In addition, cardiac-type sodium channels have not been found in the center of the SAN of the human heart, but only in the periphery of the SAN, where they may be essential for proper conduction to the surrounding atrial tissue [16]. Based on these observations, the clinical phenotype of sick sinus syndrome or sinus dysfunction described in some SCN5A mutations may be explained by peripheral SAN dysfunction and impaired impulse propagation to the atria [5, 7, 42]. On the other hand, the persistent inward sodium current and altered channel inactivation kinetics resulting from the SCN5A mutation 1795insD have been proposed to underlie the observed bradycardia and sinoatrial pauses [42]. Furthermore, the SCN5A mutation E161K, associated with sick sinus syndrome, was predicted to cause a reduction in sinus rate by slowing of the diastolic depolarization rate and upstroke velocity of the sinus node action potential [39]. Thus, although the expression level of sodium channels in the SAN appears limited, they may still play a role in sinus node function and heart rate regulation, perhaps even more so in the setting of sodium channel dysfunction.

We observed abundant expression of Scn5a and Na_v1.5 in the atrioventricular bundle (AVB or His bundle), bundle branches and Purkinje fibers, predicting a substantial role for Na_v1.5 in fast propagation through the intraventricular conduction system. In previous studies by others, both high and low intensities of Na_v1.5 staining have been reported in these areas [17, 47]. This discrepancy in observations may be attributed to differences in tissue handling protocols used, which may have influenced tissue binding properties of the antibody. To address this controversy, we have used in situ hybridization of Scn5a mRNA as an independent tool for assessing the distribution pattern of the cardiac sodium channel in the intraventricular conduction system. The observed pattern of Scn5a mRNA distribution matches that of Na_v1.5 protein distribution, thereby ultimately proving the presence of abundant cardiac sodium channel expression in the AVB (or His bundle), bundle branches and Purkinje fibers. Since the cardiac sodium channel is hardly expressed in the atrioventricular node (AVN), the abundance of Na_v1.5 in the AVB (or His bundle), bundle branches and Purkinje fibers is likely to be responsible for the atrioventricular and intraventricular dysfunction observed in patients with sodium channel dysfunction due to SCN5A mutations [7, 43].

Another finding in the current study is the observation of a transmural gradient in Scn5a, Na_v1.5, and functional sodium channel availability. The presence and ionic mechanism of transmural differences in action potential repolarization have been investigated in detail, but studies of transmural differences in depolarization are limited. Transmural action potential heterogeneity of the left ventricle is well known, with subepicardial and midmyocardial action potentials displaying prominent phase-1 repolarization, while subendocardial action potentials do not [1]. In addition, midmyocardial cells have longer action potentials compared to subepicardial and subendocardial cells, while subepicardial action potentials are the shortest [1]. Various membrane channels contribute to these transmural differences in the time course of action potential repolarization, including the transient outward K⁺ current (I _to1) [27], the slow delayed rectifier K⁺ current (I _Ks) [26], and the sustained component of the sodium current (I _Na(sus)) [48]. Some investigators have reported reduced sodium channel density in isolated myocytes from the subepicardium as compared to the mid- and subendocardium [3, 40], whereas other studies showed either no difference [12] or rather an increased sodium current density in subepicardium versus subendocardium [21]. Similarly, inconsistent results on sodium gating properties of subendocardial versus subepicardial myocytes have been described. In the dog, subepicardial myocytes have a more negative half-inactivation voltage [12], which consequently will lead to a reduced sodium current availability at resting membrane potentials. In contrast, studies in rats revealed no differences in inactivation [3, 21], but one report showed a more negative half-activation potential in subendocardial compared to subepicardial myocytes [21]. These discrepancies may be attributed to differences in experimental protocol, myocyte isolation process and species studied. Our current immunohistochemistry results from whole mouse heart sections show the existence of a transmural gradient in Na_v1.5 distribution in both the left and right ventricle in situ. Importantly, Scn5a mRNA expression displayed a similar inhomogeneous distribution across the ventricular wall on in situ hybridization. Reduced Scn5a mRNA expression has also been demonstrated in epicardial compared to endocardial tissue in non-diseased human hearts, clearly suggesting that a transmural gradient in sodium channel availability is also present in human myocardium [18]. The functional relevance of the heterogeneous transmural Scn5a/Na_v1.5 distribution was subsequently confirmed by our demonstration of reduced functional sodium channel availability in the subepicardial compared to subendocardial cells from patch-clamp analysis in isolated myocytes. Although we did not measure sodium current characteristics in the current study, dV/dt _max is considered a suitable index for functional sodium current availability [6, 37], and these functional data are in complete accordance with our immunohistochemistry and in situ hybridization results.

The functional implications of transmural heterogeneity in Na_v1.5 expression may be substantial. In the subepicardium, decreased inward sodium current will on the one hand slow conduction, and on the other hand allow for a greater impact of the repolarizing potassium currents, resulting in action potential shortening and subsequent transmural dispersion in repolarization. In the setting of myocardial ischemia, a gradient of Na_v1.5 expression predisposes to inhomogeneous conduction slowing, greatly increasing the risk of life-threatening arrhythmias [9]. In patients with Brugada syndrome, the typical ECG pattern of ST-segment elevation in the right precordial leads has been attributed to a transmural heterogeneity of repolarization in the right ventricle [2], where the transient outward K⁺ current (I _to1) responsible for phase-1 action potential repolarization is more prominent compared to the left ventricle [15]. In loss-of-function SCN5A mutations, present in approximately 15–30% of Brugada syndrome patients [36, 38], sodium current reduction combined with the large transient outward current (I _to) in the right ventricular subepicardium is thought to give rise to a prominent early repolarization phase and an even larger transmural repolarization gradient [2]. Brugada syndrome is also thought to arise from right ventricular conduction delay due to localized conduction slowing [4, 13]. The inhomogeneous Na_v1.5 distribution across the ventricular wall observed in our immunohistochemistry experiments may contribute to both mechanisms. Since the subepicardial layer comprises a relatively large fraction of the total right ventricular wall (as compared to the left ventricle), the effects of further sodium channel reduction is expected to be most notable in the right ventricle. Thus, in the current study, we have identified transmural heterogeneity in cardiac sodium channel expression as a novel additional pathophysiological player in the Brugada syndrome.

Cardiac expression and function of certain ion channels may be differentially regulated during murine heart development [17, 20, 32]. In our study, the observed distribution patterns of Scn5a mRNA and Na_v1.5 protein expression were similar in adult and embryonic murine hearts (ED14.5). Interestingly, the spatio-temporal expression pattern of Scn5a is similar to that of Cx40, and together they are likely responsible for fast conduction through the AVB, bundle branches and Purkinje fibers [11, 19]. Cx40, at least in embryonic hearts, displays a transmural gradient across the ventricular wall in a pattern similar to Scn5a, and its expression is known to be regulated by the T-box transcriptional activator Tbx5 [8, 19]. Sodium channel expression in the SAN is controlled by the T-box transcriptional repressor Tbx3, since Na_v1.5 is ectopically expressed in the SAN in Tbx3-deficient mice [22]. Therefore, Tbx3 and Tbx5 may act in concert to determine the unique distribution pattern of Scn5a and Na_v1.5 within the heart, with Tbx3 repressing expression in the SAN and AVN, and Tbx5 promoting expression in the atria, ventricle, AVB and bundle branches. The homeodomain transcription factor Irx5 also displays a transmural gradient similar to Scn5a and has been shown to determine the gradient of potassium channel expression and repolarization across the ventricular wall [10, 14]. Thus, it is tempting to speculate that Tbx5 and/or Irx5 are also responsible for the heterogeneous transmural distribution of the cardiac sodium channel.