Background
Diabetes mellitus (DM) is considered to be a strong risk factor for sudden cardiac death [
1]. Diabetic heart exhibits pro-arrhythmogenic electrocardiographic abnormalities such as abnormal repolarization [
2] mostly related to potassium channels alterations [
3]. Therefore, the link between DM and related risk of arrhythmias should be straightforward. Intriguingly, there are some experimental studies [
4,
5] as well as clinical trials [
6,
7] suggesting less tendency of the diabetic heart to develop ventricular arrhythmias in certain conditions. Several explanations of this peculiar phenomenon were suggested [
4,
5], but its exact molecular mechanism is still not elucidated.
Cardiac HCN channels are of particular importance for heart rhythm. This has been documented by HCN2- or HCN4-deficient mice which exhibit sinoatrial dysfunction [
8‐
10] and also by development of bradycardic HCN channel blocking agent ivabradine [
11]. On the other hand, in ventricular myocardium, their overexpression contributes to cardiac pro-arrhythmogenic potential [
12]. All four isoforms of HCN channels are expressed in cardiac tissue but they manifest a species-dependent regional specific distribution [
10]. HCN2 and HCN4 isoforms are predominant ventricular HCN transcripts and together represent more than 90% of the ventricular HCN channels [
13]. However, in healthy atrial and ventricular cardiomyocytes which do not display spontaneous activity [
10], HCN channels are barely expressed when compared to their stable expression in pacemaker cells [
14]..An enhanced expression of myocardial HCN channels contributes to increased pacemaker current (
If) which relates to ventricular and atrial arrhythmias in failing hearts [
15]. Moreover, HCN channels, when overexpressed, were reported to prolong the repolarization of ventricular action potential and thereby increase the arrhythmogenic potential [
14]. HCN channels are supposed to functionally antagonize K
+ currents during late repolarization thanks to slow deactivation kinetics. Therefore, their upregulation in hypertrophic heart relates to prolonged QT interval and the increased arrhythmogenic potential [
13]. Furthermore,
If augmentation induces a diastolic influx of Na
+ cations leading to increased intracellular Ca
2+ due to the Na
+/Ca
2+ exchanger shift towards 'reverse mode' resulting in increased arrhythmogenicity [
16]. The overexpression of myocardial HCN occurs in various conditions, including cardiac hypertrophy [
17], acute myocardial infarction [
18] and heart failure [
15] while the reduction of myocardial HCN is rather related to atria and to impaired sinus rhythm [
19,
20].
On the other hand, blockade of cardiac HCN channels by ivabradine is clinically used to treat systolic heart failure and chronic angina indicating beneficial effect of attenuation of HCN channels activity in the heart disease [
11]. Furthermore, cardiac HCN channels blockade was shown to reduce lethal arrhythmias in dilated cardiomyopathy [
12]. Interestingly, although clinical studies reported efficacy and safety of ivabradine in diabetic patients [
21], cardiac effects of ivabradine have been reported to be weaker in rats with streptozotocin (STZ)-induced diabetic heart damage [
22]. Recently, reduced expression of HCN channels in the sinoatrial node of streptozotocin (STZ)-induced diabetic rats was found [
20]. Therefore, we hypothesize that altered cardiac electrogenesis in diabetes might be affected by changes in cardiac HCN channel expression.
Discussion
Experimental DM due to STZ administration in rats exhibits typical ECG changes which can be considered as common signs of developing diabetic cardiomyopathy [
2,
20]. Although implying pro-arrhythmogenic potential, studies showed rather reduced risk of triggered ventricular dysrhythmias in STZ model [
4,
5]. This is in line with certain evidence showing less tendency of diabetic hearts to ventricular arrhythmias regardless of pronounced risk of cardiovascular morbidity and mortality in diabetic patients [
4,
6,
34] but it is contradictory to the impaired cardiac repolarization in DM [
2,
35] suggesting effective compensatory mechanisms. The main finding of our study in this model is the downregulation of HCN2 channel exclusively in ventricular myocardium possibly contributing to regulation of electric stability of diabetic ventricles.
The cardiac downregulation of HCN channels under pathologic conditions is not an isolated finding, though related mostly to atria. HCN downregulation in sinus node is reported in aged heart [
36] and heart failure and relates to sinus node dysfunction [
37] and even atrial tachyarrhythmia [
38]. Similarly, a decrease of HCN was found in the atria of metabolic murine model of mitochondrial dysfunction [
19] and in sinoatrial node of Goto-Kakizaki type 2 diabetic rats [
39]. Importantly, HCN downregulation was reported also in pacemaker cells [
20], in sinoatrial node [
20] and whole cardiac conduction system [
40] of rats with type 1 diabetes mellitus induced by STZ manifesting in lower intrinsic heart rate, a lengthened sinoatrial conduction time and rate-corrected maximal sinoatrial node recovery time in vivo as well as a longer cycle length in vitro [
20]. Also, knock out mouse models lacking Hcn2 or Hcn4 channels exhibit sinoatrial heart disturbances without reports of ventricular electric abnormalities [
8,
9] suggesting different roles of HCN channels in ventricles. Moreover, Hofmann et al. (2012) reported that HCN2/HCN4 deficiency results in a less pronounced prolongation of ventricular repolarization and a strong attenuation of pro-arrhythmogenic potential in settings of triggered ventricular hypertrophy. In light of this, the decrease of Hcn2 gene expression in ventricles is a novel and engaging finding that might contribute to electrical stability of diabetic ventricles in spite of presence of prolonged ventricular repolarization.
We observed altered expression of particular potassium channels-related genes particularly those significantly contributing to observed abnormal repolarization (
i.e. prolonged QT interval and increased T wave) such as Kcnh2 (gene encoding the Kv11.1 subunit of ERG channel; [
3]) and Kcnq1 (gene encoding the Kv7.1 subunit of KvLQT1 channel [
3]). However, repolarization abnormalities, presumably, mirror a complex orchestration of changes during development of diabetic heart damage [
41]. As mentioned above, HCN channels might be an integral part of cardiac electric remodelling and might play a role in cardiac repolarization as the enhanced activity of HCN channels is reported to disrupt ventricular repolarization and lengthen QT interval and double HCN2/HCN4 knockout in the ventricles of the hypertrophic hearts results in attenuated action potential and QTc interval prolongation [
13]. However, these ECG alterations were detected in animals with normal expression of K
+ channels what is different from our model where Hcn2 downregulation is a part of gene reprogramming. Since ventricular repolarization is mediated by K
+ channels, prolonged QTc interval in STZ rats is likely due to K
+ channels downregulation [
42] usually responsible for attenuation of ventricular repolarization reserve and consequent proarrhythmic risk [
41]. The Hcn2 downregulation in our study may be viewed as a compensatory phenomenon, supporting repolarization reserve.
Regulation of cardiac HCN channels expression may be affected by microRNAs [
43] regulating transcriptional and post-transcriptional gene expression. We focused on miR-1 and miR-133a as their dysregulation contributes to diabetic cardiomyopathy [
44] and moreover, also influences HCN2 and HCN4 expressions [
43]. However, alterations of HCN2 found in our study were not accompanied by altered levels of miR-1 nor miR-133a questioning their contribution to electric remodelling in ventricles of STZ rats and suggesting a different mechanism for HCN2 downregulation.
To investigate the particular role of hyperglycaemia for HCN channels expression, we used differentiated H9c2 cell line treated with high glucose concentration in the culture medium. H9c2 cells, as an in vitro model, have been used in research of HCN channels previously [
25] though they do not exhibit apparent endogenous pacemaker currents. Excess of glucose did not result in HCN channels expression changes in H9c2 cells indicating that hyperglycaemia per se is not sufficient to downregulate HCN2. We found a decrease in Glut4 glucose transporter mRNA expression after high glucose treatment which reflects disrupted glucose metabolism in H9c2 cells, similar to diabetic cardiomyopathy [
45]. However, H9c2 cells are not maturated cells and they do not spontaneously beat even after differentiation. As the mature cardiac myocytes are characterized by structural and functional entities involved in the generation and transmission of the action potential and the excitation–contraction coupling process, one would expect differences in genotype between H9c2 and mature cardiomyocytes. In fact, the specific organization of ion channels and transporters promoting action potential is key to the function of the cardiac myocytes [
46]. It was shown that mechanical stress [
47] and electrical stimulation [
48] which are physiologically relevant for cardiomyocytes increase expression of ion channels. Therefore, using H9c2 cell line to study electrophysiology is associated with certain limitations. On the other hand, the cell line expresses particular ion channels typically observed in cardiomyocytes and some electrophysiological properties were documented suggesting H9c2 cells are potentially valuable surrogates for the investigation of ion channel regulation [
49]. However, we did not detect any noticeable expression of transient outward current subunit Kcnd2 and ryanodine receptor which modulate the spontaneous beating rate of cardiomyocytes during development and therefore are crucial for cardiomyocyte function [
50]. Consequently, our results underline the fact, that observed HCN2 downregulation in diabetes is highly specific for ventricular myocytes and it is not observed in either atrial tissue or H9c2 cells (Additional file
1).
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.