Increased expression of HCN channels in the ventricular myocardium contributes to enhanced arrhythmicity in mouse failing hearts

Abstract

Background: The efficacy of pharmacological interventions to prevent sudden arrhythmic death in patients with chronic heart failure remains limited. Evidence now suggests increased ventricular expression of hyperpolarization-activated cation (HCN) channels in hypertrophied and failing hearts contributes to their arrythmicity. Still, the role of induced HCN channel expression in the enhanced arrhythmicity associated with heart failure and the capacity of HCN channel blockade to prevent lethal arrhythmias remains undetermined.

Methods and results: We examined the effects of ivabradine, a specific HCN channel blocker, on survival and arrhythmicity in transgenic mice (dnNRSF-Tg) expressing a cardiac-specific dominant-negative form of neuron-restrictive silencer factor, a useful mouse model of dilated cardiomyopathy leading to sudden death. Ivabradine (7 mg/kg per day orally) significantly reduced ventricular tachyarrhythmias and improved survival among dnNRSF-Tg mice while having no significant effect on heart rate or cardiac structure or function. Ivabradine most likely prevented the increase in automaticity otherwise seen in dnNRSF-Tg ventricular myocytes. Moreover, cardiac-specific overexpression of HCN2 in mice (HCN2-Tg) made hearts highly susceptible to arrhythmias induced by chronic β-adrenergic stimulation. Indeed, ventricular myocytes isolated from HCN2-Tg mice were highly susceptible to β-adrenergic stimulation-induced abnormal automaticity, which was inhibited by ivabradine.

Conclusions: HCN channel blockade by ivabradine reduces lethal arrhythmias associated with dilated cardiomyopathy in mice. Conversely, cardiac-specific overexpression of HCN2 channels increases arrhythmogenicity of β-adrenergic stimulation. Our findings demonstrate the contribution of HCN channels to the increased arrhythmicity seen in failing hearts and suggest HCN channel blockade is a potentially useful approach to preventing sudden death in patients with heart failure.

Keywords: HCN channel; arrhythmia; heart failure; ion channels.

Figure 1.

Ivabradine (iva) prolongs survival among dnNRSF‐Tg (Tg) mice. A, Representative I_f currents recorded in ventricular myocytes from a Tg mouse in the presence (red line; 2 minutes after the application of iva) and absence of iva (black line). Inward‐rectifier K⁺ current was suppressed by 0.5 mmol/L BaCl₂. Pulse protocol is shown in the top. B, Relative I_f amplitudes (%) measured at −140 mV in the absence (vehicle, indicated as veh) and presence of iva (+iva). A mean relative I_f amplitudes (%) in the absence of iva was assigned a value of 100 (n=6 each). *P<0.05 vs vehicle. The Wilcoxon signed‐rank test was used for the analysis. C, Heart rates in wild‐type (WT) and Tg mice at 20 weeks of age, with and without 12 weeks of iva treatment (n=9 for untreated control WT, n=5 for WT with iva, n=10 for untreated control Tg and n=6 for Tg with iva). Kruskal–Wallis nonparametric ANOVA followed by the Bonferroni correction was used for analysis among the 4 groups. NS, not significant. Graphs are shown in dot plots. D, Kaplan–Meyer survival curves for Tg mice, with or without ivabradine. Drug treatment began when the mice were 8 weeks of age and lasted 24 weeks: *P<0.05 (n=54 for Tg without drugs [control], 28 for Tg with ivabradine). E through G, Body weights (BW) (E), heart weight‐to‐body weight ratios (HW/BW) (F) and lung‐to‐body weight ratios (LW/BW) (G) in 20‐week‐old WT and Tg mice, with or without iva (for BW and HW/BW comparisons, n=9 for untreated WT, n=4 for WT treated with iva, n=11 for untreated Tg, and n=6 for Tg treated with iva; for LW/BW comparisons, n=3 in each group). Kruskal–Wallis nonparametric ANOVA followed by the Bonferroni correction was used for analysis among the 4 groups. *P<0.00833. NS, not significant. Data in E through G are shown as dot plots. dnNRSF‐Tg indicates dominant‐negative form of neuron‐restrictive silencer factor transgenic mice; ANOVA, analysis of variance; cont, control.

Figure 2.

Effects of ivabradine (iva) on the water consumption and histology in dnNRSF‐Tg (Tg) mice. A, Water consumption (g/day) in 12‐week‐old WT and Tg mice, with or without iva (n=8 for untreated [cont] WT mice, n=6 for WT mice treated with iva, n=6 for untreated (cont) Tg mice, n=6 for Tg mice treated with iva). Kruskal–Wallis nonparametric ANOVA followed by the Bonferroni correction was used for analysis among the 4 groups. B, Histology of WT and Tg hearts from 20‐week‐old mice treated with or without iva: H‐E, Hematoxylin‐Eosin staining; scale bars, 100 μm. C, Graphs show the percentage of collagen area in untreated (cont) WT mice, Tg mice treated without (cont) or with iva (n=5 for untreated WT mice, n=9 for untreated Tg mice, n=7 for Tg mice treated with iva). Kruskal–Wallis nonparametric ANOVA followed by the Bonferroni correction was used for the analysis among the 3 groups. *P<0.0166. All data are shown as box plots. NS indicates not significant; WT, wild type; cont, control; dnNRSF‐Tg, dominant‐negative form of neuron‐restrictive silencer factor transgenic mice; ANOVA, analysis of variance.

Figure 3.

Effects of ivabradine (iva) on the gene expression in dnNRSF‐Tg (Tg) mice. A through F, Relative levels of ANP (A), BNP (B), SERCA2 (C), HCN2 (D), HCN4 (E), and CACNA1H (F) mRNA in hearts from 20‐week‐old WT and Tg mice treated with or without iva. n=3 in each group. Mice treated without iva are indicated as cont. Kruskal–Wallis nonparametric ANOVA followed by the Bonferroni correction was used for analysis among the 4 groups. All data are shown as dot plots. NS indicates not significant; ANP, atrial natriuretic peptide; cont, control; WT, wild type; dnNRSF‐Tg, dominant‐negative form of neuron‐restrictive silencer factor in transgenic mice; BNP, brain natriuretic peptide; SERCA2, sarcoplasmic/endoplasmic reticulum calcium ATPase 2; HCN2, hyperpolarization‐activated cyclic nucleotide‐gated channel 2; CACNA1H, calcium channel, voltage‐dependent, T type, alpha 1H subunit; ANOVA, analysis of variance.

Figure 4.

Effects of ivabradine (iva) on the fibrosis‐related genes expression in dnNRSF‐Tg (Tg) mice. A through F, Relative levels of Col1a1 (A), Col3a1 (B), FN1 (C), MMP2 (D), MMP9 (E), and Tgfb1 (F) mRNA in hearts from 20‐week‐old WT and Tg mice treated with or without iva. n=4 for WT without iva, n=4 for WT with iva, n=5 for Tg without iva, and n=3 for Tg with iva. Mice treated without iva are indicated as cont. Kruskal–Wallis nonparametric ANOVA followed by the Bonferroni correction was used for the analysis. All data are shown as dot plots. NS indicates not significant; dnNRSF‐Tg, dominant‐negative form of neuron‐restrictive silencer factor transgenic mice; ANOVA, analysis of variance; WT, wild type; cont, control; Col1a1, collagen type1 α1; Col3a1, collagen type3 α 1; FN1, fibronectin 1; MMP2, matrix metallopeptidase 2, MMP9, matrix metallopeptidase 9; Tgfb1, transforming growth factor‐ β 1.

Figure 5.

Ivabradine (iva) reduces arrhythmicity in dnNRSF‐Tg (Tg) hearts. A and B, Numbers of PVCs (A) and VTs (B) recorded using a telemetry system in Tg mice treated with or without iva. Data are shown as box plots. Mann–Whitney test was used for analysis. *P<0.05 (n=7 for Tg without iva, 6 for Tg with iva). C, Frequency of mice with inducible VTs during intracardiac electrophysiology studies among Tg mice treated for 12 weeks with or without iva. VT, numbers of mice with inducible VT; total, total numbers of mice tested. D, Average power of the low frequency (LF) and high frequency (HF) components of HRVs recorded over a 24‐hour period in untreated WT and Tg mice treated with or without iva. Mice treated without iva are indicated as cont. n=4 for WT without iva, n=6 for Tg without iva, n=4 for Tg with iva. Kruskal–Wallis nonparametric ANOVA followed by Bonferroni correction was used for the analysis. Data are shown as dot plots. E, Average resting membrane potentials recorded from ventricular myocytes isolated from 20‐week‐old untreated WT and Tg mice treated with or without iva: Kruskal–Wallis nonparametric ANOVA followed by the Bonferroni correction was used for the analysis. NS, not significant. *P<0.0166 vs WT (n=12 for untreated WT, 12 for untreated Tg, 14 for Tg with iva). Mice treated without iva are indicated as cont. Data are shown as box plots. F, Representative traces showing that ivabradine (3 μmol/L) reduces the frequency of spontaneous action potentials in the presence of isoproterenol (0.3 μmol/L) are shown. Arrows show larger pictures of action potentials. G, Graphs show numbers of spontaneous action potentials in the presence of isoproterenol (0.3 μmol/L) in ventricular myocytes from dnNRSF‐Tg in the absence (veh) or presence of iva. Shown are the numbers of spontaneous action potentials (AP/min) occurring in the presence of isoproterenol (0.3 μmol/L) during the second minute after addition of iva or vehicle (veh). Data are shown as box plots. Mann–Whitney test was used for the analysis (n=6 for control and 5 for iva). *P<0.05. NS indicates not significant; dnNRSF‐Tg indicates dominant‐negative form of neuron‐restrictive silencer factor transgenic mice; PVC, premature ventricular contraction; VT, ventricular tachycardia; ANOVA, analysis of variance; WT, wild type; cont, control; HRV, heart rate variability.

Figure 6.

Generation of cardiac‐specific HCN2 transgenic mice. A, Scheme of the construct for HCN2 transgenic mouse. hGH, human growth hormone. B and C, Relative levels of HCN2 (B) and HCN4 (C) mRNA in hearts from 12‐week‐old WT and HCN2‐Tg mice. The Mann–Whitney test was used for the analysis. *P<0.05 vs WT. n=3 in each group. Data are shown as dot plots. D, Representative Western blots for HCN2 and β‐actin in ventricular myocytes from WT and HCN2‐Tg mice. E, Representative I_f currents recorded in ventricular myocytes from HCN2‐Tg mice. Inward‐rectifier K⁺ current was suppressed by 0.5 mmol/L BaCl₂. F, Current‐voltage relationship for I_f in ventricular myocytes from HCN2‐Tg mice. The amplitudes of time‐dependent components activated by hyperpolarizing pulses were normalized by cellular capacitance (n=5). G, Effect of 3 μmol/L ivabradine (iva) (red line; 2 minutes after the application of iva) or vehicle (cont) on I_f in ventricular myocytes from HCN2‐Tg mice. H, Graphs show the suppressive effect of iva on I_f amplitude at −80 mV (left panel, n=4 each) and −140 mV (right panel, n=5 each) in ventricular myocytes from HCN2‐Tg mice. The mean relative I_f amplitudes (%) in the absence of iva were assigned a value of 100. The Wilcoxon signed‐rank test was used for the analysis. *P<0.05 vs cont. HCN2‐Tg indicates hyperpolarization‐activated cyclic nucleotide‐gated channel 2 transgenic mice; WT, wild type; cont, control; αMHC, α‐myosin heavy chain.

Figure 7.

Features of cardiac‐specific HCN2‐Tg mice. A through C, Body weights (BW) (A), heart weight‐to‐body weight ratios (HW/BW) (B) and lung‐to‐body weight ratios (LW/BW) (C) in 12‐week‐old WT and HCN2‐Tg mice are shown as dot plots (for BWs, n=4 for each group; for HW/BW ratios, n=3 for each group; and for LW/BW ratios, n=3 for each group). The Mann–Whitney test was used for the comparison between WT and HCN2‐Tg. NS, not significant. D, Histology of WT and HCN2‐Tg hearts from 12‐week‐old mice: Sirius‐red staining. Magnification, ×400; scale bars, 100 μm. E, Graph showing the the percentage of collagen area in WT and HCN2‐Tg mice (n=4 for each group). The Mann–Whitney test was used for the analysis. NS, not significant. Data are shown as dot plots. F through H, Relative levels of ANP (F), BNP (G), and SERCA2 (H) mRNA in hearts from 12‐week‐old WT and HCN2‐Tg mice (n=3 in each group). The Mann–Whitney test was used for the analysis. NS, not significant. Data are shown as dot plots in F through H. HCN2‐Tg indicates hyperpolarization‐activated cyclic nucleotide‐gated channel 2 transgenic mice; WT, wild type.

Figure 8.

Effects of β‐adrenergic stimulation in HCN2‐Tg mice. A and B, HW/BW ratios (A) and blood pressures (B) in 24‐week‐old WT and HCN2‐Tg mice, with or without 1 week of isoproterenol (iso) administration (15 mg/kg per day; subcutaneous infusion): *P<0.05. (for HW/BW ratios, n=3 for WT without iso, n=5 for WT with iso, n=3 for Tg without iso, n=5 for Tg with iso; for blood pressure, n=3 in each group). Mice treated without iso are indicated as cont. Kruskal–Wallis nonparametric ANOVA followed by the Bonferroni correction was used for analysis among the 4 groups. NS, not significant. All data are shown as dot plots. C, Water consumption (g/day) in 24‐week‐old WT and Tg mice treated for 1 week with iso (15 mg/kg per day; subcutaneous infusion) (n=4 for WT mice, n=3 for Tg mice). The Mann–Whitney test was used for the analysis. NS, not significant. Data are shown as dot plots. D, Heart rate assessed by ambulatory electrocardiography in 24‐week‐old WT and HCN2‐Tg mice, with or without 1 week of isoproterenol (iso) administration (15 mg/kg per day; subcutaneous infusion): Kruskal–Wallis nonparametric ANOVA followed by the Bonferroni correction was used for analysis among the 4 groups. n=3 for each group. Data are shown as dot plots. HCN2‐Tg indicates hyperpolarization‐activated cyclic nucleotide‐gated channel 2 transgenic mice; WT, wild type; cont, control; HW/BW, heart weight‐to‐body weight ratios; ANOVA, analysis of variance.

Figure 9.

β‐adrenergic stimulation‐induced ventricular arrhythmias in HCN2‐Tg hearts. A, Representative ECG traces showing sinus rhythm in WT mice treated with isoproterenol (upper panel) and ventricular tachycardia (VT) with atrioventricular dissociation (middle panel) and isolated PVC (lower panel) in HCN2‐Tg mice treated with isoproterenol. B and C, Numbers of isolated PVCs (B) and VTs (C) recorded using a telemetry system in WT and HCN2‐Tg mice treated with isoproterenol (iso). Mann–Whitney test was used for the analysis. *P<0.05 (n=3 for each group). D, Representative traces of action potentials recorded in ventricular myocytes isolated from WT and HCN2‐Tg hearts. E, Representative traces of spontaneous action potentials in ventricular myocytes from HCN2‐Tg hearts recorded in the presence of 0.3 μmol/L isoproterenol, to which 3 μmol/L ivabradine was subsequently added, as indicated. Arrows show larger pictures of action potentials. F, Graph showing the numbers of spontaneous action potentials (AP) in the presence of 0.3 μmol/L isoproterenol (iso) in ventricular myocytes from HCN2‐Tg mice with or without 3 μmol/L ivabradine (iva). The Wilcoxon signed‐rank test was used for the analysis. Shown are the numbers of action potentials occurring during the minute just before (pre‐iva) and the second minute after (post‐iva) addition of ivabradine. *P<0.05 (n=6). HCN2‐Tg indicates hyperpolarization‐activated cyclic nucleotide‐gated channel 2 transgenic mice; WT, wild type; PVC, premature ventricular contraction.