L-Type Ca_v1.3 and HCN Channels Mediate Heart Rate Acceleration by Catecholamines

Abstract

BACKGROUND:

The ionic mechanism by which catecholamines increase the heart rate is incompletely understood. In this study, we have assessed the roles of sinoatrial node L-type Ca_v1.3 (α_1D) Ca²⁺ channels, phosphorylation of L-type channel regulatory partner protein Rad (Ras-related RGK GTP-binding protein), and cAMP-dependent regulation of hyperpolarization-activated HCN (hyperpolarization-activated cyclic nucleotide-gated) channels.

METHODS:

We studied β-adrenergic regulation of heart rate and sinoatrial pacemaker activity in mice lacking Ca_v1.3 channels and in mice expressing dihydropyridine-insensitive L-type Ca_v1.2 channels alone or concomitantly expressing cAMP-insensitive HCN4 subunits in a heart-specific and time-controlled manner. We also studied the chronotropic response to sympathomimetics of sinoatrial pacemaker myocytes under conditions of specific inhibition of cAMP-dependent regulation of HCN4 by the cyclic dinucleotide cyclic di-(3′,5′)-GMP and ablation of PKA (protein kinase A)–dependent phosphorylation of Rad.

RESULTS:

Mutant mice with knockout of Ca_v1.3 and cAMP-insensitive HCN4 subunits in the heart lacked diurnal variation in heart rate and failed to increase their heart rate after administration of catecholamines or during physical activity. Selective pharmacological inhibition of Ca_v1.3 prevented the enhancement of pacemaker activity by sympathomimetics or by direct activation of adenylate cyclase, as well as by phosphodiesterase inhibitors, when cAMP-dependent regulation of HCN was simultaneously silenced. Upregulation of Ca_v1.3 and HCN-mediated funny current (I_f) accounted for the total change in diastolic current on activation of β-adrenoceptors, explaining the loss of chronotropic effect of catecholamines. Preventing PKA phosphorylation of Rad abrogated the chronotropic response to sympathomimetics of intact hearts under HCN blockade, or in pacemaker myocytes on preventing cAMP-dependent regulation of HCN4, respectively.

CONCLUSIONS:

PKA phosphorylation of Rad, which disinhibits Ca_v1.3 channels and cAMP-dependent activation of HCN channels, are key effectors in β-adrenergic regulation of cardiac pacemaker activity and can sustain positive chronotropic effects independently. These findings on Rad-mediated regulation of Ca_v1.3 and HCN channels unravel the ionic mechanisms underlying the catecholaminergic acceleration of the heart rate.

Novelty and Significance.

What Is Known?

• Catecholamines accelerate heart rate during physical activity or under stress, a phenomenon referred to as the chronotropic response.

• The chronotropic response is due to activation of β-adrenoceptors and consequent quickening of spontaneous activity of pacemaker myocytes mediated by the increase in concentration of cAMP, CaMKII (Ca²⁺/calmodulin-dependent protein kinase II), and cAMP-dependent PKA (protein kinase A) activity.

• The role of ion channels and their regulations in mediating the chronotropic response of heart rate is still uncertain and debated.

What New Information Does This Article Contribute?

• During physical activity or under administration of sympathomimetics, the chronotropic response requires intact function of either L-type Ca_v1.3 (α_1D) calcium channels or cAMP-dependent activation of cardiac hyperpolarization-activated funny (HCN [hyperpolarization-activated cyclic nucleotide-gated] channel) channels. Interfering simultaneously with both Ca_v1.3 and HCN activity abrogates the chronotropic response in pacemaker myocytes and in mice.

• Ca_v1.3 channels or cAMP-mediated activation of HCN channels can autonomously generate a chronotropic response. They underlie a Ca_v1.3-reliant and HCN-reliant chronotropic mechanisms, respectively.

• The Ca_v1.3-reliant chronotropic mechanism is mediated by PKA-dependent phosphorylation of the small Ras-related protein Rad (Ras-related RGK GTP-binding protein), which disinhibits Ca_v1.3 channel gating on β-adrenoceptors activation.

The acceleration of the heart rate by catecholamines, referred to as chronotropic response, is a fundamental physiological mechanism to adapt to physical activity, or stressful situations. The chronotropic response is initiated when sympathetically released or circulating catecholamines activate β-adrenoceptors of pacemaker myocytes, stimulating the synthesis of cAMP and activation of cAMP-dependent protein kinase A (PKA). In this study, we show for the first time that the ionic mechanism of the chronotropic response is generated by 2 autonomous mechanisms, mediated by L-type Ca_v1.3 (α_1D) channels and hyperpolarization-activated cAMP-gated HCN channels. We report that loss of function of either channel renders the chronotropic response contingent on the functional channel, thereby identifying Ca_v1.3-reliant and HCN-reliant chronotropic mechanisms. Furthermore, we introduce PKA-dependent regulation of Ras-related Rad protein to Ca_v1.3 channels as a new key player in the chronotropic response. This study elucidates the ionic mechanism underlying the positive chronotropic response of heart rate to catecholamines. It identifies, Ca_v1.3 and HCN channels as key effectors of activated β-adrenoceptors in upregulation of cardiac pacemaker activity, opening the way to a deeper understanding of dysfunction in the chronotropic mechanism of the human heart.

Meet the First Author, see p e000743

Editorial, see Article by Xiao et al

The fight-or-flight response is an integrative physiological mechanism of survival that is triggered when an animal becomes aware of a harmful or stressful situation.¹ Numerous reflexes and organs are involved in the fight-or-flight response, yet the most perceptible reaction to circulating and sympathetically released catecholamines is a sudden acceleration of the heart rate. The role of ion channels in the positive chronotropic response of heart rate to catecholamines has been the subject of extensive research and debate for the last 40 years, as attempts to abrogate this mechanism by targeting ion channels have been unsuccessful.

The heart rate is determined by the frequency of sinoatrial node (SAN) action potentials, which are generated by spontaneously active, automatic pacemaker myocytes. Automaticity results from diastolic depolarization—the pacemaker potential—a spontaneous depolarization of the membrane voltage corresponding to the diastolic phase of the cardiac cycle, which encompasses the time from the end of repolarization to the threshold of the after action potential.^2,3 Catecholamines activate β-ARs (β-adrenoceptors) to accelerate heart rate by increasing the slope of diastolic depolarization.⁴ First reports in the late 1960s showed that adrenaline induced a change in the kinetics of membrane currents in a direction compatible with the acceleration of diastolic depolarization.⁵ In 1979, Brown et al⁴ showed that the acceleration of pacemaking induced by adrenaline in sinoatrial tissue was associated with an increase in the hyperpolarization-activated funny current (I_f) and in the slow inward current (I_si). These pioneering studies demonstrated that catecholamines positively regulated ionic currents involved in pacemaker activity but could not associate loss-of-function in I_f or I_si with disruption of the β-adrenergic chronotropic effect. I_si was then reinterpreted as a process of time-dependent overlap between activation of the cardiac L-type Ca²⁺current (I_CaL) and of the NCX (Na⁺/Ca²⁺ exchanger) current I_NCX,⁶ leaving the ionic mechanism of β-adrenergic–mediated chronotropic effect as one of the remaining outstanding research challenges. In the following years, the regulation of HCN (hyperpolarization-activated cyclic nucleotide-gated) channels encoding I_f by cAMP⁷ and the functional partnership between sarcoplasmic reticulum RyR2 (type 2 ryanodine receptor)-mediated Ca²⁺ release and the plasma membrane–bound cardiac NCX1 have been proposed to underlie the positive chronotropic response of heart rate to catecholamines.^8–13 However, adult mice carrying mutated HCN channels or silenced I_f conductance did not display a diminished chronotropic response.^14–18 Impairment of RyR2-mediated Ca²⁺ release by ryanodine reduced, but did not abolish, the positive chronotropic effect of isoprenaline in SAN myocytes.^10,19 Finally, mice carrying altered function CaMKII (Ca²⁺/calmodulin-dependent protein kinase II),²⁰ RyR2,²¹ or downregulation of NCX1^22,23 showed unaltered basal heart rate and only partial reduction of SAN chronotropic response. SAN myocytes constitute a unique cardiac cell type, as they coexpress 2 functionally distinct isoforms of L-type Ca²⁺ channels: the cardiac Ca_v1.2 (Cacna1C, α_1C) isoform, which couples excitation to contraction in the working myocardium,²⁴ and the Ca_v1.3 (Cacna1D,α_1D) isoform.

In this study, we have identified critical mechanisms mediating the positive chronotropic effect of catecholamines on heart rate. We show that regulation of Ca_v1.3 by its partner Rad (Ras-related RGK GTP-binding protein),²⁵ and cAMP-dependent activation of HCN channels are the effectors in the positive chronotropic response of SAN automaticity and heart rate to catecholamines.

Methods

Data Availability

Full methodological details are given in the Supplemental Methods. Data supporting the findings of this study are available from the corresponding authors on reasonable request. Please see the Major Resources Table in the Supplemental Material.

The study conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (Publication No. 85–23, revised 1996), the German law for the Protection of Animals and European directives (2010/63/EU). The experimental procedure was approved by the ethical committee of the University of Montpellier and the French Ministry of Agriculture (protocol no. 2017010310594939), by the Ministry of Science and Public Health of the City State of Hamburg, and by the institutional animal care and use committee of Columbia University. A protocol including the research question, key design features, and analysis plan was prepared before the study and approved by the Réseau d’Animaleries de Montpellier facility of Montpellier.

Genetically Modified Mice

Mice expressing HCN4 channels with deleted C-terminally located cyclic nucleotide binding domain (CNBD) at 573X position of the human HCN4 gene (hHCN4-573X) in a heart-specific and time-controlled manner were generated, as described previously.¹⁷ To obtain mice carrying simultaneous ablation of Ca_v1.3 channels, or mice carrying dihydropyridine-resistant Ca_v1.2 channels (Ca_v1.2^DHP−/−) and expressing hHCN4-573X channels, we crossed mice deficient in Ca_v1.3 channels (Ca_v1.3^+/- and Ca_v1.3^−/−),²⁶ or Ca_v1.2^DHP−/− mice²⁷ with double transgenic HCN4-CNBD (cyclic nucleotide–binding domain) or HCN4-CNBDf mice (Figure S1A). The 4SA-Rad knock-in mouse line was generated as described previously.²⁸

In Vivo Telemetric Recordings of Heart Rate and ECG Recordings in Isolated Hearts

We performed telemetric recording of ECG in freely moving mice as previously described.²⁹ Heart rates were determined from ECG RR intervals and atrial rates from PP intervals, under control conditions or after a 4-hour baseline recording to evaluate the effects of drugs. We performed ECG recordings in isolated Langendorff-perfused mouse hearts as previously described.²⁹ To record heart rate and changes in membrane voltage in wild-type and 4SA-Rad hearts on the Langendorff system simultaneously, we loaded the hearts with the voltage-sensitive fluorescent dye 1-(3-Sulfonatopropyl)-4-(beta)(2-(di-n-butylamino)-6-naphthylvinyl)pyridinium betaine (Di-4-ANNEPS, 0.2 µmol/L; Sigma-Aldrich) for 20 minutes. All recordings were performed at 36.5 °C to 37 °C. Atrioventricular conduction velocity was determined between the leading site in the SAN and the ventricle.

Sharp Electrode Recordings of Automaticity in Intact SAN/Atria Preparations

We obtained SAN-atria preparations from excised hearts, as previously described.³⁰ We placed the entire SAN-atria preparation in warmed Tyrode’s solution (36 °C), under perfusion of either Tyrode’s solution, or Tyrode’s containing ivabradine or amlodipine. Intracellular electrodes had a resistance of 10 MΩ when filled with a 3 mol/L KCl solution.

Patch-Clamp Recordings of Pacemaker Activity and Ionic Currents

SAN pacemaker myocytes were isolated, and patch-clamp experiments were performed, as previously described.³¹ Myocytes were harvested in custom-made chambers and perfused with Tyrode’s solution warmed to 36 °C. Spontaneous action potentials in SAN myocytes, the I_f and diastolic currents were recorded using standard whole-cell variation of the patch-clamp technique.³² Whole-cell electrodes had a resistance of 3 to 5 MΩ when filled with K⁺-aspartate-based intracellular solution. In Figure S5, Ca_v1.3-mediated L-type current (I_Cav1.3) and Ca_v1.2-mediated L-type current (I_Cav1.2) were recorded using a standard whole-cell patch-clamp configuration, in Tyrode’s solution and intracellular K⁺-aspartate solution, to allow simultaneous recording of I_f. Otherwise, I_Cav1.3 and I_Cav1.2 were recorded, as described previously.³³

Confocal Imaging of Calcium

Confocal imaging of intracellular Ca²⁺ release in SAN myocytes was performed with the Ca²⁺ indicator Fluo-4 acetoxymethyl ester (Fluo 4-AM), as previously described.³⁴ SAN myocytes were seeded in glass fluorodishes, coated overnight with laminin, and allowed to settle for 1 hour before imaging. Fluorescence was excited at 488 nm, and emissions were collected at >505 nm. A 63× oil immersion objective was used to record [Ca²⁺]_i in isolated SAN myocytes.

Western Blotting of Wild-Type and 4SA-Rad SANs

To evaluate protein expression and phosphorylation of RyR2, PLN (phospholamban) and troponin, expression of NCX1, and tubulin, SANs from wild-type and 4SA-Rad were isolated and treated either with Tyrode’s solution or with Tyrode’s solution containing isoprenaline 0.1 µmol/L for 5 minutes. SANs treated with vehicle or isoprenaline were harvested and frozen in liquid nitrogen for western blot analysis. SANs from 2 wild-type or 4SA-Rad hearts, treated either with vehicle or isoprenaline, were combined and homogenized in 0.2 mL of lysis buffer (Tris-maleate [pH 7.4], 1 mmol/L EDTA, and complete protease inhibitors (Roche) on ice using a Dounce homogenizer. Immunoblots were developed using the following primary antibodies (see also Major Resources Table in the Supplemental Material): anti-RyR³⁵ (1:2500), anti–phospho-RyR2-pSer2808³⁶ (1:2500), anti–phospho-RyR2-pSer2815³⁷ (1:2500), anti-PLN (PA5-78410, 1:2500; Thermo Fisher), anti–phospho-PLN (S16, PA5-85740, 1:1000; Thermo Fisher), anti–phospho-PLN (Thr17, A010-13, 1:2500; Badrilla), anti-NCX1 (79350S, 1:2000; Cell Signaling Technology), and anti-tubulin (PA1-16947, 1:2500; Thermo Fisher).

Statistical Analysis

Throughout the study, N indicates the number of mice, or isolated hearts, whereas n indicates the number of SANs or individual myocytes. For general consistency, nonparametric tests were used to interpret and present data in figures and figure legends. Full details of statistical analysis can be found in the Supplemental Methods and statistical file in the Supplemental Material. Statistical tests used in each experiment are specified throughout the figure legends, and P values are indicated in panels. Statistical significance was defined as P<0.05.

Results

Mice Lacking L-Type Ca_v1.3 and cAMP-Dependent Regulation of HCN Channels Are Devoid of the Catecholaminergic Chronotropic Response of Heart Rate

Previous work has shown that conditional gene ablation,^14,15,38 silencing of cAMP-dependent regulation,^17,39,40 or transgenic block of HCN channel conductance¹⁶ do not prevent the chronotropic response of heart rate and of SAN pacemaker activity to β-ARs activation. Because we have shown that SAN L-type Ca_v1.3 channels contribute to diastolic depolarization,³³ we hypothesized that the chronotropic response of heart rate was reliant on Ca_v1.3 channels when HCN channels are disabled. To test this hypothesis, we studied heart rate regulation in mutant mice lacking Ca_v1.3 channels and expressing cAMP-insensitive HCN channels. We crossed Ca_v1.3 channel knockout mice (Ca_v1.3^−/−) with double-transgenic HCN4-CNBD^T/T (HCN4-CNBD) mice with Tet-off system–mediated heart-specific and time-controlled expression of dominant-negative hHCN4-573X channel subunits lacking the CNBD^17,41 (Figure S1A). Expression of hHCN4-573X subunits prevents cAMP-dependent regulation of endogenous HCN subunit isoforms.^17,41

Because a day-night variation exists in sympathetic input⁴² and circulating catecholamines,^43,44 we studied the day-night variation in the heart rate of control and double Ca_v1.3/HCN4-CNBD mutant mice. Control, Ca_v1.3^−/− and Ca_v1.3^±/HCN4-CNBD mutant mice displayed significant day-night variation in heart rate and in spontaneous locomotor activity. HCN4-CNBD mice exhibited a trend toward day-night variation in heart rate, but this trend did not reach statistical significance. In contrast, Ca_v1.3^−/−/HCN4-CNBD mice exhibited no significant day-night variation in heart rate, despite normal variation in locomotor activity (Figure 1A and 1B; Figure S1B and S1C). Preserved diurnal variation in locomotor activity suggested that loss of day-night regulation of heart rate in Ca_v1.3^−/−/HCN4-CNBD mice was independent of basal bradycardia (Figure S1D and S1E) but potentially due to decreased capability of sympathetic input and circulating catecholamines to control heart rate. To strengthen this hypothesis, we tested control and mutant mice for their ability to increase heart rate after 5 minutes of physical exercise on a treadmill. Exercise increased the heart rate of control, Ca_v1.3^−/−, HCN4-CNBD, and Ca_v1.3^±/HCN4-CNBD mice. In contrast, exercise failed to increase the heart rate in Ca_v1.3^−/−/HCN4-CNBD mice (Figure 1C). Lastly, we evaluated the chronotropic heart rate response in control and mutant mice after acute intraperitoneal administration of the nonspecific sympathomimetic β-AR agonist isoprenaline. Isoprenaline significantly increased the heart rate in control, Ca_v1.3^−/−, HCN4-CNBD and Ca_v1.3^±/HCN4-CNBD mice, but not in Ca_v1.3^−/−/HCN4-CNBD mice (Figure 1D and 1E).

Figure 1.

Chronotropic response in control and mutant mice lacking Ca_v1.3 and cAMP-dependent regulation of HCN (hyperpolarization-activated cyclic nucleotide-gated) channels. A, Diurnal variation in heart rate (HR) in control and mutant mice measured between Zeitgeber time (ZT) 0 (subjective dawn), ZT12 (subjective dusk), to the following subjective dawn (ZT24). Heart rates are averaged at ZT0 and then every hour. B, Single data points plot of heart rates in control and mutant mice recorded throughout light and dark periods in experiments shown in A for N=9 control, N=8 Ca_v1.3^−/−/HCN4-CNBD, N=12 Ca_v1.3^±/HCN4-CNBD, N=11 Ca_v1.3^−/− and N=6 HCN4-CNBD mice. Connected dots indicate an individual mouse. Statistics: Wilcoxon signed-rank test. C, Data points plot of heart rates of control and mutant mice as in A control (N=12, black dots) and mutant Ca_v1.3^−/−/HCN4-CNBD (N=8 dark red dots), Ca_v1.3^±/HCN4-CNBD (N=11, blue dots), Ca_v1.3^−/− (N=17, green dots), and HCN4-CNBD (N=15, violet dots) recorded at baseline (before) and during 5-minute physical exercise (after) on treadmill. Statistics: Wilcoxon signed-rank test. D, Sample traces of telemetric ECG recordings from control and mutant mice at baseline or after intraperitoneal injection of isoprenaline (ISO, 0.1 mg/kg). E, Corresponding single data points of heart rates recorded at baseline and after injection of isoprenaline (red dots) in control (N=20, black dots) and mutant Ca_v1.3^−/−/HCN4-CNBD (N=8, dark red dots), Ca_v1.3^±/HCN4-CNBD (N=12, blue dots), Ca_v1.3^−/− (N=16, green dots), and HCN4-CNBD (N=16, violet dots). Statistics: Wilcoxon signed-rank test.

Consistent with our in vivo observations, Langendorff-perfused hearts from Ca_v1.3^−/−/HCN4-CNBD mice did not exhibit a significant positive chronotropic response to adrenaline, whereas the heart rates of the other mutant genotypes were responsive (Figure 2A through 2C). These data showed that the concomitant loss of Ca_v1.3 and cAMP-dependent regulation of HCN channels abolished heart rate acceleration by catecholamines in vivo and ex vivo. Notably, Ca_v1.3^±/HCN4-CNBD mice showed day-night variation in heart rate, an increase in heart rate on exercise, and a positive chronotropic effect to isoprenaline, which indicated that one functional Cacna1D allele, was sufficient for generating a significant positive chronotropic effect on activation of β-ARs.

Figure 2.

Chronotropic response to adrenaline in control and mutant hearts. A, Sample traces of ECG recordings of Langendorff-perfused hearts from control and mutant mice of different genotypes at baseline (upper line) and during 5-minute perfusion of adrenaline (ADR 0.1 µmol/L, bottom line). B. Comparison of dose-response curves of heart rates (HR) to adrenaline in hearts from control (N=7) and Ca_v1.3^−/−/HCN4-CNBD (N=5, left) or Ca_v1.3^−/− (N=8) HCN4-CNBD (N=7) and Ca_v1.3^±/HCN4-CNBD (N=11, right) mutant mice. Changes in heart rate are expressed as the increase in beats per minute with 100% being the resting heart rate. C, Corresponding bar graphs of rates at the maximal effect of adrenaline (0.1 µmol/L) on heart rates. Statistics: Wilcoxon signed-rank test. D through G, Chronotropic effect of isoprenaline (ISO) on sinoatrial node (SAN) beating rates in Langendorff-perfused wild-type (WT) and mutant hearts globally lacking Ca_v1.3 (Ca_v1.3^−/−) and T-type Ca_v3.1 (Ca_v3.1^−/−) channels, in the presence or absence of I_f inhibition by 7-minute perfusion of ivabradine (IVA 10 µmol/L). SAN rate is deduced from ECG PP intervals in N=13 WT, N=10 Ca_v3.1^−/−, N=10 Ca_v1.3^−/− and N=16 Ca_v1.3^−/−/Ca_v3.1^−/− at baseline (Ctrl), 5-minute perfusion of ISO (0.1 µmol/L), after IVA perfusion and after concomitant application of ISO and IVA (IVA+ISO). Statistics: Friedman test followed by Dunn multiple comparisons test. Insets represent changes in heart rate expressed as percentage of rate acceleration by ISO perfusion vs baseline and simultaneous IVA and ISO perfusion compared with IVA alone. Statistics: Wilcoxon signed-rank test. H. Comparison between percentages of rate acceleration by ISO in wild-type, Ca_v3.1^−/−, Ca_v1.3^−/− and Ca_v1.3^−/−/Ca_v3.1^−/− hearts. Statistics: Kruskal-Wallis test followed by Dunn multiple comparisons test.

To obtain evidence that the abolition of chronotropic response to isoprenaline in Ca_v1.3^−/−/HCN4-CNBD hearts was not due to atrial remodeling induced by the expression of hHCN4-573X channels, we used the HCN channel blocker ivabradine. In line with recordings obtained using Ca_v1.3^−/−/HCN4-CNBD hearts, the SAN rate measured as ECG PP intervals of hearts isolated from an independent colony of Ca_v1.3^−/− mice, was insensitive to isoprenaline under conditions of blockade of I_f by ivabradine, while wild-type hearts were normally responsive (Figure 2D and 2E). A previous study indicated that overexpression or genetic ablation of T-type Ca_v3.1 Ca²⁺ channels augmented or reduced the relative degree of chronotropic response, respectively, suggesting that these channels could contribute to the β-adrenergic effect on heart rate.⁴⁵ We previously demonstrated that genetic ablation of T-type Ca_v3.1 Ca²⁺ channels slows basal heart rate and SAN pacemaker activity^46,47 but does not affect the maximal heart rates reached during the active nighttime period.⁴⁶ Thus, we compared the effects of genetic ablation of Ca_v3.1 channels to those of Ca_v1.3 channels on the chronotropic response to isoprenaline in isolated hearts, under I_f blockade. Isoprenaline significantly increased the heart rate of wild-type and Ca_v3.1^−/− mice in vivo, despite I_f blockade by ivabradine (Figure S2A and S2B). Isoprenaline similarly increased the beating rate of Langendorff-perfused wild-type, Ca_v3.1^−/−, Ca_v1.3^−/− and Ca_v1.3^−/−/Ca_v3.1^−/− hearts (Figure 2D and 2F). In addition, I_f block by ivabradine did not significantly reduce the relative chronotropic response to isoprenaline of wild-type, Ca_v3.1^−/− hearts (Figure 2D through 2H, insets). In contrast, ivabradine abrogated the chronotropic response to isoprenaline of Ca_v1.3^−/− and of double-mutant Ca_v1.3^−/−/Ca_v3.1^−/− hearts (Figure 2E and 2G and insets), showing that deletion of Ca_v1.3 was necessary to abrogate the chronotropic response in Ca_v3.1^−/− hearts. Taken together, these data demonstrate the significant impact of deleting Ca_v1.3, compared with Ca_v3.1, on the loss of β-adrenergic chronotropic effect under conditions of I_f blockade. In addition, the loss of the chronotropic response to isoprenaline by ivabradine in Ca_v1.3^−/− and Ca_v1.3^−/−/Ca_v3.1^−/− hearts demonstrates that the chronotropic response on genetic ablation of Ca_v1.3 channels is reliant on I_f.

I_Cav1.3 Is Necessary to Generate the Chronotropic Response of Sinoatrial Pacemaker Activity to Catecholamines After Genetic Silencing of cAMP Sensitivity of HCN Channels

We investigated whether selective pharmacological inhibition of Ca_v1.3 channels could prevent the catecholaminergic chronotropic response of heart rate as observed using knockout Ca_v1.3^−/− mice. To achieve selective blockade of Ca_v1.3, we employed mice in which the sensitivity of Ca_v1.2 channels to dihydropyridines has been abolished by knocking in a point mutation in the channel’s dihydropyridine-binding site (Ca_v1.2^DHP−/−).^27,48 Current-to-voltage curves recorded in individual wild-type SAN myocytes showed blockade of I_CaL (I_Cav1.3 +I_Cav1.2) by the dihydropyridine nifedipine, while current-to-voltage curves recorded in Ca_v1.2^DHP−/− myocytes showed dihydropyridine-resistant I_Cav1.2 (Figure 3A). I_Cav1.2 was normally stimulated by isoprenaline in Ca_v1.2^DHP−/− SAN myocytes on blockade of I_Cav1.3, indicating that Ca_v1.3 inhibition did not affect the availability of I_Cav1.2 (Figure 3A). Heart rates recorded in Ca_v1.2^DHP−/− mice after injection of the L-type channel dihydropyridine antagonist isradipine were comparable to the basal heart rates of Ca_v1.3^−/− mice, indicating that the selective inhibition of I_Cav1.3 was quantitatively similar to the genetic deletion of Ca_v1.3 channels regarding heart rate reduction at baseline (Figure S3A). Ca_v1.2^DHP−/− mice underwent 5-minute physical exercise on a treadmill in control conditions, after individual administration of the dihydropyridine amlodipine or ivabradine, or after simultaneous injection of amlodipine and ivabradine (Figure 3B). Injection of amlodipine or ivabradine slowed basal heart rate, as expected by individual blockade of I_Cav1.3 or I_f, respectively. Physical exercise on a treadmill tended to increase the heart rate of mice injected with amlodipine or ivabradine. Yet no significant heart rate increase was observed after simultaneous injection of the 2 drugs, when both I_Cav1.3 and I_f were inhibited (Figure 3B and 3C). In line with this observation obtained in vivo, simultaneous perfusion of amlodipine and ivabradine prevented the positive chronotropic effect of adrenaline in Langendorff-perfused Ca_v1.2^DHP−/− hearts (Figure 3D). Sharp electrode intracellular recording of spontaneous action potentials in isolated SAN-atria preparations⁴⁷ from Ca_v1.2^DHP−/− mice showed that concomitant perfusion of amlodipine and ivabradine prevented the positive chronotropic response to isoprenaline of spontaneous SAN action potentials (Figure S3B). Taken together, these observations suggested that the positive chronotropic response to adrenaline recorded after blockade of I_Cav1.3 was reliant on HCN channels.

Figure 3.

Ca_v1.3-reliant and HCN (hyperpolarization-activated cyclic nucleotide-gated) channel–reliant chronotropic responses. A, Sample current traces (top line) and current-to-voltage (I-V) relationships (bottom line) of L-type Ca²⁺current (I_CaL) in wild-type (N=3, n=12) and Ca_v1.2^DHP−/− (N=4, n=6) sinoatrial node (SAN) myocytes in control solution (Ctrl, full line) and after perfusion of nifedipine (3 µmol/L, NIFE, green dotted line),and after concomitant perfusion of NIFE and 0.1 µmol/L isoprenaline (NIFE+ISO, red line). Statistics: Wilcoxon signed-rank test and Kruskal-Wallis test followed by Dunn multiple comparisons test. B, Time course of changes in averaged heart rates (HR) in N=5 Ca_v1.2^DHP−/− mice at baseline and after 5-minute physical exercise on a treadmill (TM). Mice were injected with amlodipine (10 mg/kg, AMLO, green line), ivabradine (IVA, 6 mg/kg, orange line), or AMLO and IVA (AMLO+IVA, dotted line) 20 minutes before the beginning of exercise (time=−20 minutes). C, Corresponding single data plot of exercise-induced changes in heart rates measured at the end of exercise. Statistics: Wilcoxon signed-rank test. D, Averaged change in heart rates (left) and sample ECG traces (right) recorded in isolated Ca_v1.2^DHP−/− hearts (N=19) on inhibition of I_Cav1.3 by 8-minute perfusion of AMLO (10 µmol/L), simultaneous inhibition of I_Cav1.3 by AMLO and I_f by IVA (10 µmol/L) and on perfusion of AMLO, IVA and adrenaline (0.1 µmol/L, ADR). Statistics: Friedman test followed by Dunn multiple comparisons test. Averaged change in heart rate and corresponding sample ECG traces in N=11 Ca_v1.2^DHP−/− (E) and N=15 Ca_v1.2^DHP−/−/HCN4-CNBDf (F) under control solution (baseline), adrenaline (0.3 µmol/L), return to control (wash), AMLO or simultaneous perfusion of AMLO and adrenaline. G, Averaged change in heart rate in N=10 HCN4-CNBDf hearts under baseline perfusion of adrenaline (0.3 µmol/L), wash, 8-minute perfusion of calciseptine (CAS, 0.3 µmol/L), simultaneous perfusion of CAS and adrenaline, and CAS, AMLO (3 µmol/L), and adrenaline. See color code for drug perfusions in E through G. Statistics (E through G): Friedman test followed by Dunn multiple comparisons test. Insets E through G, Corresponding percentages of heart rate acceleration on the different conditions. Statistics: Wilcoxon signed-rank test and Friedman test followed by Dunn multiple comparisons test. Tyr indicates Tyrode’s solution.

We then developed a mouse model in which the selective inhibition of I_Cav1.3 by dihydropyridine could be achieved in a genetic background where HCN channels were insensitive to cAMP. To accomplish this, we crossed Ca_v1.2^DHP−/− mice with a mouse line in which the expression of hHCN4-573X channels in the atrium after dox treatment is higher than in the HCN4-CNBD line,¹⁷ to obtain double-mutant (Ca_v1.2^DHP−/−/HCN4-CNBDf). Selective inhibition of I_Cav1.3 by amlodipine in Ca_v1.2^DHP−/− hearts significantly reduced but did not abolish the chronotropic response of heart rate to adrenaline (Figure 3E and inset; Figure S3C). In contrast, in Ca_v1.2^DHP−/−/HCN4-CNBDf hearts, inhibition of I_Cav1.3 abolished the positive chronotropic effect of adrenaline (Figure 3F and inset; Figure S3D), which indicated that the chronotropic response in HCN4-CNBDf hearts was reliant on Ca_v1.3 channels encoding I_Cav1.3. In line with these results, adrenaline accelerated pacemaker activity in HCN4-CNBDf, but not in Ca_v1.2^DHP−/−/HCN4-CNBDf SAN myocytes under perfusion of nifedipine (Figure S3E and S3F).

To compare the effects of blocking I_Cav1.3 with those of selective inhibition of I_Cav1.2, we employed the Ca_v1.2-selective blocker calciseptine.⁴⁹ Inhibition of I_Cav1.2 had no significant effect on either the basal heart rate or the positive chronotropic response to adrenaline of HCN4-CNBDf hearts. In contrast, concomitant block of I_Cav1.3 and I_Cav1.2 by amlodipine abrogated the positive chronotropic effect (Figure 3G and inset). Finally, we compared the effects of blocking I_Cav1.3 and I_Cav1.2 with those of selective inhibition of T-type channels by 3,5-dichloro-N-[1-(2,2-dimethyl-tetrahydro-pyran-4-ylmethyl)-4-fluoro-piperidin-4-ylmethyl]-benzamide (TTA-P2)⁵⁰ and TTX (tetrodotoxin)-sensitive channels underlying I_Na(TTX) by TTX.⁵¹ Inhibition of I_CaT, or of I_Na(TTX), significantly slowed the heart rate of isolated HCN4-CNBDf hearts (Figure S4A). However, neither TTA-P2 nor TTX significantly reduced the chronotropic response to adrenaline of HCN4-CNBDf hearts (Figure S4B). Taken together, these observations showed that only simultaneous pharmacological inhibition of I_Cav1.3 and of I_f, or concomitant I_Cav1.3 inhibition and genetic silencing of HCN channel regulation by cAMP, prevented the chronotropic response of heart rate to catecholamines.

Ca_v1.3 Channels Regulate the Magnitude of Diastolic Current on β-Adrenergic Activation of Pacemaker Activity

Because our experiments indicated that the positive chronotropic response of the SAN to β-adrenergic activation, after the abolition of cAMP-dependent regulation of HCN channels, was reliant on I_Cav1.3, we investigated how activation of this current generates the chronotropic response. An increase in the net inward current flowing during diastolic depolarization is required to induce a chronotropic response. We hypothesized that I_Cav1.3 and cAMP-dependent regulation of HCN4 generated the predominant fraction of the inward diastolic current induced by isoprenaline. We therefore expected the diastolic current to be insensitive to isoprenaline when I_Cav1.3 and cAMP-dependent regulation of HCN are simultaneously inhibited. Cyclic di-(3′,5′)-GMP (c-di-GMP) is a natural antagonist of cAMP-dependent regulation of HCN4 channels that can be used in the patch-clamp pipette.⁵² C-di-GMP binds to a pocket of the channel’s C-linker, preventing its rearrangement on cAMP binding to the CNBD and leading to inhibition of the cAMP effect⁵² (Figure 4A). Control experiments showed that c-di-GMP blocked the response of I_f to isoprenaline while leaving the isoprenaline-induced stimulation of total I_CaL and of the dihydropyridine-resistant I_Cav1.2 component unaffected (Figure S5A through S5E). We then employed c-di-GMP to selectively switch off the contribution of HCN4 to β-adrenergic regulation of pacemaker activity and of diastolic current in Ca_v1.2^DHP−/− SAN pacemaker myocytes under conditions of I_Cav1.3 inhibition. Pacemaker activity in Ca_v1.2^DHP−/− myocytes showed a normal response to isoprenaline, which persisted on perfusion of nifedipine (ie, blockade of I_Cav1.3; Figure 4B). Inclusion of c-di-GMP in the patch pipette, which causes blockade of cAMP-dependent regulation of HCN4, decreased basal pacemaker activity, which could still be positively regulated by isoprenaline. However, the positive chronotropic effect was abolished when c-di-GMP was added to the patch pipette with concomitant perfusion of nifedipine to Ca_v1.2^DHP−/− SAN myocytes (Figure 4C). In Ca_v1.2^DHP−/− SAN myocytes, nifedipine strongly reduced inward diastolic current, while isoprenaline reversed the effect of nifedipine, increasing diastolic current in the inward direction to control levels (Figure 4D through 4F). Net diastolic current dropped close to the zero-current line when nifedipine was applied in the presence of c-di-GMP (Figure 4G and 4H). In line with the observation that SAN myocytes exhibited a slow spontaneous beating rate under these conditions (Figure 4C), residual diastolic current was low in density, but still inward. Isoprenaline failed to increase significantly diastolic current, again in line with the lack of positive chronotropic effect of this agonist of β-ARs on spontaneous action potentials (Figure 4C). To further confirm that I_Cav1.3 was modulating the isoprenaline-induced diastolic current, we employed HCN4-CNBDf and Ca_v1.2^DHP−/−/HCN4-CNBDf SAN myocytes. Diastolic current in HCN4-CNBDf myocytes was increased by isoprenaline (Figure 4I and 4K). However, isoprenaline failed to increase the diastolic current in Ca_v1.2^DHP−/−/HCN4-CNBDf myocytes under the blockade of I_Cav1.3 (Figure 4J and 4L). In conclusion, these data show that the chronotropic effect of isoprenaline is reliant on I_Cav1.3 and I_f, and that I_Cav1.3 and cAMP-dependent regulation of HCN4 account for the net change of diastolic current under catecholaminergic activation of pacemaker activity in SAN myocytes under inhibition of I_Cav1.3.

Figure 4.

Ca_v1.3 and HCN (hyperpolarization-activated cyclic nucleotide-gated) channel–mediated diastolic currents generating the chronotropic response. A, Docking simulation of cyclic di-(3′,5′)-GMP (c-di-GMP)—yellow sticks with oxygen and nitrogen atoms colored in red and blue, respectively, in the X-ray structure of human HCN4 C-linker structure (green)—and CNBD (cyclic nucleotide–binding domain; gray). The fragment is solved in the presence of cGMP (Protopedia accession number: 4KL1; adapted from Lolicato et al⁵² with permission). The secondary structural elements are labeled. The key residues of the interaction with c-di-GMP are shown, labeled, and colored based on their chemical nature: blue and red for the positively and negatively charged residues, respectively. The cGMP interacting with CNBD is colored in black. B and C, Abrogation of the chronotropic effect of isoprenaline (ISO) on pacemaker activity in isolated sinoatrial node (SAN) myocytes by nifedipine (NIFE) and c-di-GMP. B, Sample traces of spontaneous action potentials and averaged chronotropic effect—in beats per minute—of 0.1 µmol/L isoprenaline in n=10 (N=8) SAN myocytes from Ca_v1.2^DHP−/− SANs in Tyrode’s solution (Tyr; control conditions, left) and after selective blockade of I_Cav1.3 by 3 µmol/L NIFE (n=15, N=6, right). Statistics: Wilcoxon signed-rank test and Friedman test followed by Dunn multiple comparisons test. C, Same experimental protocol as in B, but the chronotropic effect of ISO is evaluated on inhibition of cAMP-dependent regulation of HCN4 channels by intracellular 100 µmol/L c-di-GMP (N=7: Tyr n=12, c-di-GMP n=11, c-di-GMP + ISO n=7, left), or concomitant inhibition of I_Cav1.3 and cAMP-dependent regulation of HCN4 channels (N=2, n=6, right). Statistics: Kruskal-Wallis test and Friedman test followed by Dunn multiple comparisons test. D through L, Diastolic currents in Ca_v1.2^DHP−/−, HCN4-CNBDf and Ca_v1.2^DHP−/−/HCN4-CNBDf SAN myocytes as function of I_Cav1.3 and regulation of HCN4 channels by cAMP. D, Sample waveform of command voltage used to measure membrane currents. Sample traces of membrane currents in the absence (E), or presence (G) of c-di-GMP in the patch-clamp pipette. The dotted line indicates the zero current level; the solid black lines indicates currents recorded in Tyr. Currents recorded under other conditions follow the color code at bottom. Gray boxes indicate the range of voltages corresponding to diastolic depolarization and used to calculate time integrals of diastolic currents. Corresponding averaged time integrals of diastolic current densities are displayed in F for Ca_v1.2^DHP−/− SAN myocytes (N=7, n=16) recorded in the absence of c-di-GMP and in H, for n=7 (N=4) Ca_v1.2^DHP−/− myocytes with c-di-GMP in the whole-cell pipette. Statistics: Friedman test followed by Dunn multiple comparisons test. I through L, Diastolic currents recorded in HCN4-CNBDf (I) and Ca_v1.2^DHP−/−/HCN4-CNBDf (J) SAN myocytes, in Tyr, ISO, NIFE, and simultaneous perfusion of NIFE and ISO. K and L, Averaged time integrals of diastolic currents for recordings in n=27, N=10 HCN4-CNBDf (K) and n=7, N=2 Ca_v1.2^DHP−/−/HCN4-CNBDf (L) SAN myocytes. Statistics: Wilcoxon signed-rank test and Friedman test followed by Dunn multiple comparisons test.

Ca_v1.3 and HCN4 Channels Underpin the Chronotropic Response of Pacemaker Activity to Modulation of Adenylate Cyclase and Phosphodiesterases

Sinoatrial pacemaker activity is highly sensitive to basal and stimulated activity of AC (adenylate cyclase)^53,54 and of PDEs (phosphodiesterase).^55,56 We thus investigated if I_Cav1.3 and I_f also accounted for the positive chronotropic effects of direct AC activation or inhibition of PDE using Ca_v1.2^DHP−/− SAN myocytes. Stimulation of AC by forskolin increased the rate of spontaneous action potentials from baseline, when I_Cav1.3 or cAMP-dependent regulation of HCN4 were inhibited by nifedipine or c-di-GMP, respectively (Figure 5A through 5C). However, the effect of forskolin was abolished under simultaneous inhibition of I_Cav1.3 by nifedipine and of HCN4 channel regulation by c-di-GMP, showing that I_Cav1.3 was necessary for the positive chronotropic effect of forskolin under loss-of-function of HCN4 channels (Figure 5D). Similarly, c-di-GMP prevented the positive chronotropic effect of the general PDE inhibitor 3-isobutyl-1-methylxanthine alone or in combination with forskolin under conditions of I_Cav1.3 inhibition (Figure 5E through 5G). In sum, our data show that I_Cav1.3 and I_f underpin the positive chronotropic effects of AC activation and PDE inhibition.

Figure 5.

Ca_v1.3 and HCN (hyperpolarization-activated cyclic nucleotide-gated) channel in the chronotropic response to PDE (phosphodiesterase) inhibitors. A, Mean chronotropic effect of direct activation of adenylate cyclase by forskolin (FSK, 10 µmol/L) in n=6, N=2 Ca_v1.2^DHP−/− sinoatrial node (SAN) myocytes. Statistics: Wilcoxon signed-rank test. B through D, A similar experimental protocol as in A, but the effect of FSK is evaluated after selective blockade of I_Cav1.3 by nifedipine (NIFE: B, N=4, n=10), inhibition of cAMP-dependent regulation of HCN4 channels by intracellular cyclic di-(3′,5′)-GMP (c-di-GMP; C, N=3, n=6), or simultaneous inhibition of I_Cav1.3 and cAMP-dependent regulation of HCN4 channels (D, N=5, n=10) in Ca_v1.2^DHP−/− SAN myocytes. Statistics: Friedman test followed by Dunn multiple comparisons test. E. Chronotropic effect of phosphodiesterase inhibitor isobutyl-methyl-xanthine (IBMX, 100 µmol/L) on the frequency of spontaneous action potentials of n=18, N=8 Ca_v1.2^DHP−/− SAN myocytes. Statistics: Wilcoxon signed-rank test. F. Chronotropic effect of simultaneous perfusion of FSK and IBMX in n=9, N=3 SAN myocytes. Statistics: Wilcoxon signed-rank test. G. Chronotropic effect on spontaneous action potentials of Ca_v1.2^DHP−/− SAN myocytes at baseline in Tyrode’s solution (Tyr; n=16, N=5), in the presence of intracellular c-di-GMP (n=14, N=5) with perfusion of NIFE (n=14, N=5) and under simultaneous perfusion of NIFE, FSK, and IBMX (n=11, N=5). Statistics: Kruskal-Wallis test followed by Dunn multiple comparisons test.

β-Adrenergic Activation of Pacemaker Activity by L-type Ca_v1.3 Channels Does Not Require Intact RyR2-Dependent Ca²⁺ Release and Persists on Inhibition of NCX1 or Sarcoplasmic Reticulum Ca²⁺ ATPase

β-adrenergic–mediated activation and synchronization of diastolic local RyR2-dependent Ca²⁺ release events (LCRs) coupled with an elevation of NCX1 activity have been proposed to constitute a primary mechanism in the positive chronotropic effect of catecholamines on SAN pacemaker activity,^11,19 although disruption of this mechanism by interfering with PKA (protein kinase A)–dependent phosphorylation of RyR2,²¹ with RyR2-dependent Ca²⁺ release,¹⁹ and genetic downregulation of NCX1^22,23 do not abrogate the response of heart rate to catecholamines in mice. Our previous studies have shown that Ca_v1.3 channels play a crucial role in synchronizing diastolic RyR2-mediated LCRs during pacemaker activity, suggesting that RyR2-dependent Ca²⁺ release may be necessary for Ca_v1.3 to produce the chronotropic effect of β-ARs. Line-scan imaging of Ca²⁺ release showed that the frequency of spontaneous cell-wide Ca²⁺ transients was increased by isoprenaline in HCN4-CNBDf SAN myocytes, or in Ca_v1.2^DHP−/− myocytes in control conditions or on perfusion of nifedipine (Figure 6A). However, isoprenaline did not significantly increase the frequency of spontaneous Ca²⁺ transients in Ca_v1.2^DHP−/−/HCN4-CNBDf SAN myocytes under I_Cav1.3 inhibition by nifedipine (Figure 6B). Analysis of RyR2-mediated diastolic LCRs also showed that inhibition of I_Cav1.3 prevented isoprenaline-induced reduction of diastolic LCRs, which indicated loss of synchronization of LCRs on β-ARs activation (Figure S6A and S6B). These observations suggested that Ca_v1.3 channel blockade disabled the RyR2-dependent pacemaker mechanism. Consistent with this interpretation, the relative slowing effect of ryanodine on the frequency of spontaneous action potentials in Ca_v1.3^−/− SAN myocytes was strongly blunted compared with that of wild-type counterparts (Figure 6C and 6D). In addition, ryanodine did not prevent significant chronotropic effect in wild-type and Ca_v1.3^−/− myocytes, while addition of c-di-GMP in the patch-clamp pipette in the presence of ryanodine in Ca_v1.3^−/− myocytes abolished it (Figure 6E). Abolition of the chronotropic response to isoprenaline in Ca_v1.3^−/− SAN myocytes could not be attributed to ryanodine, because addition of c-di-GMP alone abrogated the chronotropic response without ryanodine (Figure S7), consistent with a role of I_f in generating this response. Taken together, these observations suggest that inhibition of I_Cav1.3 or genetic ablation of Ca_v1.3 channels prevents RyR2 from generating a positive chronotropic response and show that cAMP-dependent regulation of HCN4 channels is responsible for the residual response.

Figure 6.

Ca_v1.3-reliant chronotropic response and RyR2 (type 2 ryanodine receptor)/NCX (Na⁺/Ca²⁺ exchanger) function. A, Sample of confocal line-scan images of RyR2-dependent Ca²⁺ release (top line) in n=27, N=8 Ca_v1.2^DHP−/− and n=14, N=4 HCN4-CNBDf sinoatrial node (SAN) myocytes at baseline and during simultaneous perfusion of nifedipine (NIFE) and isoprenaline (ISO), with corresponding averaged data (bottom). Statistics: Wilcoxon signed-rank test. B, as in A, sample line-scan images (top line) and averaged frequencies of spontaneous cell-wide intracellular Ca²⁺ transients ([Ca²⁺]_i) in n=27, N=6 Ca_v1.2^DHP−/− and n=18, N=3 Ca_v1.2^DHP−/−/HCN4-CNBDf SAN myocytes at baseline in Tyrode’s solution (Ctrl) and after perfusion of NIFE or simultaneous NIFE and ISO. Statistics: Friedman test followed by Dunn multiple comparisons test. C and D, Averaged frequency of spontaneous action potentials recorded in n=14, N=6 wild-type (WT, C) and N=6 Ca_v1.3^−/− (Tyrode’s solution [Tyr] n=18, ryanodine [Ry] n=15, Ry+ISO n=12, D) SAN myocytes at baseline, after 3-minute perfusion of Ry (3 µmol/L), or after concomitant perfusion of Ry and ISO. Statistics: Friedman test and Kruskal-Wallis test followed by Dunn multiple comparisons test. E, Mean frequencies of spontaneous transients recorded in n=11, N=3 Ca_v1.3^−/− SAN myocytes in which cAMP-dependent regulation of HCN4 channels is inhibited by cyclic di-(3′,5′)-GMP (c-di-GMP) and then perfused with Ry and concomitant perfusion of Ry and ISO. Statistics: Friedman test followed by Dunn multiple comparisons test. F, Frequency of spontaneous action potentials recorded in n=10, N=4 HCN4-CNBDf SAN myocytes at baseline, after perfusion of Ry, and after simultaneous application of Ry and ISO. Statistics: Friedman test followed by Dunn multiple comparisons test. G and H. Same experiment as in F, but showing the effects on the rate of spontaneous action potentials of SEA (2-(4-((2,5-difluorobenzyl)oxy)phenoxy)-5-ethoxyaniline; 3 µmol/L) in n=8, N=3 (G), or of SERCA inhibitor cyclopiazonic acid (CPA 3 µmol/L, H) in n=11, N=5 HCN4-CNBDf SAN myocytes. Statistics: Friedman test followed by Dunn multiple comparison. I. Frequency of spontaneous action potentials in n=9, N=2 Ca_v1.3^−/−/HCN4-CNBDf at baseline, after perfusion of Ry and after concomitant perfusion of Ry and ISO. Statistics: Friedman test followed by Dunn multiple comparisons test.

We then investigated whether the regulation of diastolic RyR2-dependent Ca²⁺ release was required for I_Cav1.3 to generate a positive chronotropic effect by challenging HCN4-CNBDf SAN myocytes with isoprenaline after pretreatment with ryanodine. Ryanodine slowed pacemaker activity in HCN4-CNBDf SAN myocytes (Figure 6F). Nevertheless, isoprenaline could still induce a significant positive chronotropic effect even in the presence of ryanodine. Similarly, isoprenaline increased pacemaker activity in HCN4-CNBDf SAN myocytes in the presence of either NCX1 inhibitor SEA-0400 (2-[4-([2,5-difluorobenzyl]oxy)phenoxy]-5-ethoxyaniline)⁵⁷ (Figure 6G), or the SERCA (sarcoplasmic reticulum Ca²⁺ ATPase) inhibitor cyclopiazonic acid (Figure 6H). Consistent with results in Figure 6E and 6F, Ca_v1.3^−/−/HCN4-CNBDf SAN myocytes did not show a significant chronotropic response in the presence of ryanodine (Figure 6I). Together, these results suggest that even if RyR2 and NCX1 could constitute downstream effectors of I_Cav1.3, they are not obligatory mechanisms for Ca_v1.3 to generate a positive chronotropic response on activation of β-ARs.

β-Adrenergic Regulation of Sinoatrial Pacemaker Activity by Ca_v1.3 Requires PKA-Dependent Phosphorylation of Rad Protein

We showed that I_Cav1.3 was required for chronotropic responsiveness of pacemaker activity to β-ARs agonists when cAMP-dependent regulation of HCN4 channels is prevented (Figures 3 through 5). However, these data do not directly establish a role for PKA-dependent regulation of Ca_v1.3 channels in this process. Enhanced cardiac L-type Ca²⁺ channel activity by PKA depends on phosphorylation of the small Rad.^25,28 Unphosphorylated Rad acts as a tonic inhibitor of Ca_v1.2 and Ca_v1.3 channel gating. PKA-dependent phosphorylation of Rad causes the dissociation of Rad from the α1/β complex, increasing the Ca²⁺ channel activity.²⁵ We employed Rrad knock-in mice in which the 4 PKA phosphorylated serine (S) residues of murine Rad (Ser25, Ser38, Ser272, and Ser300) were replaced by alanine (A) residues (4SA-Rad). 4SA-Rad mice lack β-adrenergic regulation of Ca_v1.2-mediated atrial and ventricular I_CaL.²⁸ We thus investigated whether Rad phosphorylation underlies the positive chronotropic response to catecholamines and whether disabling PKA regulation of Rad mimics the effects of I_Cav1.3 blockade. Patch-clamp recordings of I_CaL showed that isoprenaline increased I_CaL in wild-type, but not in 4SA-Rad SAN myocytes, without affecting basal current density (Figure 7A and 7B), demonstrating that PKA-dependent phosphorylation of Rad was required for β-adrenergic upregulation of SAN I_CaL. 4SA-Rad hearts exhibited moderate slowing of basal heart rate (Figure 7C and inset) and a reduced degree of chronotropic response to adrenaline compared with wild-type counterparts (Figure 7D and inset). In addition, although I_f inhibition by ivabradine did not prevent the chronotropic effect of adrenaline in wild-type hearts, it completely abolished it in 4SA-Rad hearts (Figure 7C and 7D). Isoprenaline increased pacemaker activity in both wild-type and 4SA-Rad SAN myocytes (Figure 7E through 7G), but, similar to observations in isolated hearts, the chronotropic effect was significantly reduced in 4SA-Rad SAN myocytes (Figure 7H). Consistent with our observations using the Ca_v1.2^DHP−/− genotype, the chronotropic response to isoprenaline of 4SA-Rad SAN myocytes was abolished by inclusion in the patch-clamp pipette of c-di-GMP (Figure 7I). Taken together, these data indicate that PKA-dependent phosphorylation of Rad is required for β-adrenergic regulation of pacemaker activity generated by Ca_v1.3.

Figure 7.

The Ca_v1.3-reliant chronotropic effect of isoprenaline (ISO) requires phosphorylation of Rad (Ras-related RGK GTP-binding protein). A, Families of L-type Ca²⁺current (I_CaL) traces recorded in wild-type (WT) and homozygous 4SA-Rad sinoatrial node (SAN) myocytes recorded from holding potential of −60 mV to test potentials indicated, under control conditions (black line), or after perfusion of ISO (red line). B, Corresponding averaged current-to-voltage relationships of I_CaL in N=3, n=12 WT and N=3, n=7 homozygous 4SA-Rad SAN myocytes in control and perfusion of ISO. Statistical analysis on I_CaL density in WT myocytes at different voltages is shown in inset table. Statistics: Mann-Whitney U test. C. Mean heart rates (HR) of N=11 WT and N=10 4SA-Rad Langendorff-perfused hearts on baseline, perfusion of adrenaline (ADR), and simultaneous perfusion of ADR and ivabradine (IVA). Statistics: Friedman test followed by Dunn multiple comparisons test. The inset shows a comparison of the baseline heart rate of WT and 4SA-Rad hearts. Statistics: Wilcoxon signed-rank test. D, Comparison of percentages of heart rate acceleration by ADR in WT (black dots) and 4SA-Rad (blue dots) hearts in the absence or presence (WT, black hollows; 4SA-Rad, blue hollows) of HCN (hyperpolarization-activated cyclic nucleotide-gated) channels blockade by IVA. The inset in (D) shows a comparison of percentages of heart rate acceleration by ADR in WT and 4SA-Rad hearts. Statistics: Mann-Whitney U test. E, Sample traces of spontaneous action potentials at baseline or during perfusion of ISO in WT and homozygous 4SA-Rad SAN myocytes recorded with control whole-cell patch-clamp pipette solution, and 4SA-Rad SAN myocytes recorded with cyclic di-(3′,5′)-GMP (c-di-GMP) in the pipette solution. F, Averaged rate of spontaneous action potentials in n=12, N=3 WT SAN myocytes in basal conditions and ISO. G, Averaged rate of spontaneous action potentials in basal conditions and ISO in n=16, N=5 4SA-Rad SAN myocytes. H, Comparison of the chronotropic effect of ISO on spontaneous action potential firing frequency in n=9 WT and n=7 4SA-Rad SAN myocytes reported in (F) and (G). I, Mean chronotropic effect of ISO recorded in n=15, N=3 4SA-Rad SAN myocytes recorded in the presence or absence of c-di-GMP in the pipette solution. Statistics: Wilcoxon signed-rank test (F, G), Mann-Whitney U test (H) and Friedman test followed by Dunn multiple comparisons test (I). Tyr indicates Tyrode’s solution.

Since loss of Rad-dependent upregulation of I_CaL may influence sarcoplasmic reticulum Ca²⁺ load and Ca^2+-dependent intracellular cAMP synthesis, we investigated the effects of 4SA-Rad mutation on the function of proteins involved in RyR2-dependent Ca²⁺ release, sarcoplasmic reticulum Ca²⁺ load, and uptake (Figure 8A through 8D). Line-scan imaging of Ca²⁺ release showed similar averaged frequencies of spontaneous Ca²⁺ transients in wild-type and 4SA-Rad SAN myocytes (Figure 8A). Pacing wild-type and 4SA-Rad myocytes at 3 Hz, showed that isoprenaline did not significantly affect the amplitude of spontaneous Ca²⁺ transients (Figure 8B) and that the Rad mutation did not affect the absolute ratio between the amplitude of Ca²⁺ transients in control conditions and on perfusion of isoprenaline (Figure 8B and inset). Experiments of caffeine-induced fast Ca²⁺ unload showed a moderate, but significant reduction of sarcoplasmic reticulum Ca²⁺ load on perfusion of isoprenaline in wild-type, but not in 4SA-Rad SAN myocytes in comparison to the control condition (Figure 8C). This phenomenon may be due to the reduction of isoprenaline-induced upregulation of diastolic Ca²⁺ release on loss of Rad-dependent regulation of I_Cav1.3, as we previously observed in Ca_v1.3^−/− SAN myocytes.³⁴ We did not observe a significant difference between the decay time of caffeine-induced Ca²⁺ transient, suggesting similar SERCA pump function in the 2 genotypes (Figure 8D).

Figure 8.

Rad (Ras-related RGK GTP-binding protein)–dependent regulation of L-type Ca²⁺ current (I_CaL) and phosphorylation of sarcoplasmic reticulum (SR) proteins. A, Sample of confocal line-scan images of baseline RyR2 (type 2 ryanodine receptor)-dependent Ca²⁺ release (top) and corresponding comparison between averaged frequencies of spontaneous Ca²⁺ transients (bottom) in n=17, N=5 wild-type (WT) and n=10, N=4 4SA-Rad sinoatrial node (SAN) myocytes. Statistics: Mann-Whitney U test. B, Comparison between averaged Ca²⁺ transient amplitude in n=14, N=5 WT and n=13, N=5 4SA-Rad in 3 Hz field-paced SAN myocytes. Statistics: Wilcoxon signed-rank test. Inset, Comparison between relative changes in Ca²⁺ transient amplitudes induced by isoprenaline (ISO) in WT and 4SA-Rad SAN myocytes shown in B. Statistics: Mann-Whitney test. C and D, Comparison between averaged SR Ca²⁺ load (C) and recovery time of cell-wide Ca²⁺ transient elicited by a 5-second puff of 10 mmol/L caffeine (D) in WT and 4SA-Rad SAN myocytes after field pacing at 3 Hz. Statistics: (C), n=17, N=3 WT in Tyrode solution (Tyr) and n=13, N=5 WT in ISO; n=14, N=5 4SA-Rad myocytes in Tyr and n=13, N=5 in ISO 4SA-Rad myocytes, Mann-Whitney U test; (D) n=16, N=3 WT in Tyr and n=8, N=3 WT in ISO; n=13, N=5 4SA-Rad myocytes in Tyr and n=11, N=5 4SA-Rad myocytes in ISO, Mann-Whitney U test. E, Sample western blots of expression and phosphorylation of RyR2 at PKA (protein kinase A)–dependent Ser2808 and CaMKII (Ca²⁺/calmodulin-dependent protein kinase II)–dependent Ser2815 phosphosites in WT and 4SA-Rad SANs, in control conditions and after 5-minute perfusion of ISO (0.1 µmol/L). F, Expression and phosphorylation of PLN (phospholamban) at PKA-dependent Ser16 (PS16) and CaMKII-dependent Thr17 (PT-17) phosphorylation sites in WT and 4SA-Rad SANs, in control conditions and after perfusion of ISO. G, Sample western blots of expression of NCX (Na⁺/Ca²⁺ exchanger) in WT and 4SA-Rad SANs. Expression of tubulin have been used as control. In E through G, 2 SANs have been pooled in each blot line. H, Averaged RyR2, PLN ratios expressed in arbirary units (AU), between total protein and protein phosphorylated at PKA-dependent and CaMKII-dependent phosphorylation sites are indicated and corresponding to E and F. NCX1 protein expression in SAN samples treated or not with ISO corresponding to G is also shown. I. Averaged fold changes in phosphorylation of RyR2 and PLN at PKA- and CaMKII-dependent phosphorylation sites induced by ISO in WT (black dots) and 4SA-Rad (blue dots) SANs, corresponding to E and F. Statistics: Multiple Mann-Whitney U test. J, Schematic figure summarizing the transition between basal pacemaker activity (left) and stimulated state (right) after activation of β-adrenoceptors by catecholamines. Ion channels, membrane receptors and the SR are depicted. Red lines indicate targets of phosphorylation by activated PKA, CaMKII, or increased ionic currents and exchange during the heart chronotropic response. P indicates phosphorylation site; and SERCA, sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase. Created with BioRender.com.

We then tested if the 4SA-Rad mutation affected the expression and phosphorylation of SAN RyR2 and PLN. Because 4SA-Rad SAN myocytes lacked PKA-dependent upregulation of I_CaL-mediated Ca²⁺ entry, we tested phosphorylation of RyR2 and PLN at both PKA-dependent and CaMKII-dependent sites. We tested the degree of phosphorylation at Ser2808, as a reporter of PKA-dependent phosphorylation of RyR2, recognizing that another PKA phosphorylation site, Ser2030, has also been reported.^58,59 We did not observe significant differences in RyR2 protein expression and phosphorylation at PKA-dependent Ser2808 and CaMKII-dependent Ser2815 on perfusion of intact SAN tissue with isoprenaline (Figure 8E, 8H, and 8I). Similarly, there was no significant difference between expression and phosphorylation of PLN at the PKA phosphorylated Ser16 and the CaMKII phosphorylated Thr 17 sites (Figure 8F, 8H, and 8I). Finally, the 4SA-Rad mutation did not significantly affect expression of the NCX1 protein (Figure 8G and 8H).

Taken together, our results indicate that reduction in chronotropic response in 4SA-Rad hearts (Figure 7C and 7D) and SAN myocytes (Figure 7H) was not due to reduced function or expression RyR2, PLN and NCX1, to reduction in capability of sarcoplasmic reticulum to regulate Ca²⁺ load, or to reduced degree of phosphorylation of RyR2 or of PLN, but to loss of upregulation of I_Cav1.3 on activation of β-ARs.

Discussion

In this study, we have identified L-type Ca_v1.3 (α_1D) and HCN channels as key effectors in the heart rate–accelerating effects of β-adrenergic stimulation. We causally associate loss-of-function in PKA/cAMP-dependent regulation of these channels to abrogation of the catecholaminergic chronotropic effect on pacemaker activity and demonstrate that the role of PKA in SAN chronotropism requires Rad-dependent regulation of Ca_v1.3. We show that Ca_v1.3 and HCN channels can operate independently to mediate the effect of catecholamines. Consequently, heart rate can be upregulated by either a Ca_v1.3-reliant or HCN-reliant chronotropic mechanism when the regulation of either of these channels is prevented. Our study also indicates that although several ion channels, including Ca_v1.3, HCN, T-type Ca_v3.1 Ca²⁺ channels, and TTX-sensitive Na⁺ channels, contribute to basal heart rate (Figures 1 and 3; Figures S3 and S4), the positive chronotropic response to catecholamines relies on Ca_v1.3 and HCN channels.

We show that the activities of Ca_v1.3 and HCN4 channels are required for a significant fraction of β-adrenergic stimulated diastolic current, which is responsible for the chronotropic effect. We found that when the cAMP-dependent regulation of HCN channels is prevented by the expression of HCN4-573X subunits, the basal diastolic current falls by ≈50% of its control value (Figure 4F and 4K). This observation is explained by assuming that basal cAMP contributes to pacemaker activity at rest by regulating the tonic activation of HCN channels.⁶⁰ On β-adrenergic activation, the diastolic current attributable to HCN4 channels, recorded with I_Cav1.3 inhibition, doubles in density. This restores the spontaneous action potential rate to basal values recorded before I_Cav1.3 inhibition (Figure 4F). This conclusion is in line with a previous study, which showed that dynamic clamp injection of predicted β-adrenergic activated I_f increases pacemaker activity of SAN myocytes by 40%.⁶¹ Similarly, when cAMP-dependent regulation of HCN4 channels is prevented, the diastolic current doubles on β-adrenergic activation (Figure 4I and 4K). Since this current is blocked by nifedipine, it is attributable to phosphorylation of Rad and disinhibition of Rad-bound Ca_v1.3 channels (Figure 4J and 4L). Taken together, our data indicate that Ca_v1.3 and HCN4 channels contribute to the β-adrenergic chronotropic response independently by supplying quantitatively comparable diastolic currents. Nevertheless, more directed study is required to establish whether early activation of HCN channels during diastolic depolarization contributes to accelerated recruitment of Ca_v1.3 channels or if Ca_v1.3-activated Ca²⁺-dependent pathways influence the regulation of HCN4 channels. Furthermore, in addition to the increase in I_Cav1.3-mediated Ca²⁺ entry induced by release of Rad inhibition of Ca_v1.3, the recorded diastolic I_Cav1.3 could also include the dihydropyridine-sensitive sustained Na⁺ current component (I_st), of which Ca_v1.3 is a required molecular determinant.⁶² In addition, a contribution from Ca_v1.3-stimulated NCX1 activity or I_st to dihydropyridine-sensitive diastolic current could explain our observation that nifedipine-sensitive I_Cav1.3 flows across the full range of diastolic depolarization (Figure 4E), despite the apparent threshold of baseline I_Cav1.3 evoked in the absence of extracellular Na⁺ being around −50 and −55 mV on β-adrenergic activation with isoprenaline.³³

Previous studies and our observations using pharmacological blockers (Figure S4) support the notion that TTX-sensitive Na⁺ channels^51,63 and T-type Ca_v3.1 channels⁴⁶ contribute to baseline heart rate. In line with this view, residual diastolic current recorded on concomitant blockade of I_Cav1.3 and cAMP-dependent regulation of HCN4, or expression of HCN4-CNBDf subunits, is inward, and likely to be generated by I_Cav3.1, I_Na(TTX) and basal cAMP-independent activity of HCN channels. In addition, residual Ca_v1.3-independent RyR2-dependent Ca²⁺ release coupled to NCX1^9,11 can also contribute to diastolic current. Finally, other channels previously reported to play a role in baseline SAN pacemaker activity, such as TRPM4^64,65 and store-operated TRPC⁶⁶ channels, may also contribute to the residual diastolic current. We also show that SAN L-type Ca_v1.2 channels are not essential for setting basal heart rate or generating the chronotropic response to catecholamines. Consequently, β-adrenergic modulation of native sinoatrial Ca_v1.2 channels cannot compensate for loss-of-function or inhibition of the Ca_v1.3 isoform. In this context, our data also highlight differential roles for cardiac L-type channel isoforms in generating the positive inotropic effect by stimulation of I_Cav1.2 in the ventricles and the positive chronotropic effect of catecholamines on the heart rate in the SAN, as induced by stimulation of I_Cav1.3. Consistent with our model, which posits that Rad serves as a brake on Ca_v1.3/Ca_v1.2 activity in the SAN, Rad-null mice demonstrated elevated heart rate during the sleep phase.⁶⁷ A previous study indicated that overexpression or knockout of T-type Ca_v3.1 channels heightened or reduced the β-adrenergic chronotropic effect on heart rate, respectively.⁴⁵ In our study, genetic ablation (Figure 2), or selective pharmacological inhibition (Figure S4) of Ca_v3.1 channels did not affect the chronotropic response to catecholamines. Hearts perfused with the T-channel blocker TTA-P2 show an unaffected chronotropic response. This observation suggests that preserved chronotropic response in our Ca_v3.1^−/− mice and isolated hearts cannot be explained by compensatory remodeling in knockout animals. However, a potential role for Ca_v3.1 in the chronotropic effect may not be incompatible with our data. It is possible that I_Cav1.3 and Rad-dependent regulation of Ca_v1.3 are required for upregulation of T-type Ca_v3.1 channels on activation of β-ARs. This hypothesis deserves further study.

Previous studies employing molecular targeting of HCN channels alone, RyR2, or NCX1 did not show abrogation of the positive chronotropic effect on heart rate. This suggested a missing molecular determinant of the catecholaminergic effect able to explain the roles of I_f and of RyR2 in pacemaker activity and of the chronotropic effect of activated β-ARs on heart rate. We identify Ca_v1.3 channels and their regulation by phosphorylation of Rad as these formerly missing effectors. We propose a model in which activation of β-ARs switches basal activity to stimulated pacemaker activity by PKA-mediated unbinding of Rad from Ca_v1.3 channels and facilitation of HCN4 channel openings by cAMP. In this model, unbinding of Rad stimulates I_Cav1.3, increasing the diastolic current via Ca²⁺ entry, accelerating the slope of diastolic depolarization (Figure 8J). Consistent with this model, we find that the chronotropic effect of catecholamines on heart rate and SAN pacemaker activity is reliant on functional Ca_v1.3 channels or Rad phosphorylation, when cAMP-dependent activation of HCN channels is prevented. Our data indicate that it is the stimulation of I_Cav1.3/I_Cav1.2 by Rad unbinding that is required for the chronotropic effect, because 4SA-Rad SAN myocytes have preserved basal I_Cav1.3/I_Cav1.2 (Figure 7A and 7B). In this regard, despite Ca_v1.3 being an important regulator of diastolic RyR2-mediated LCRs, I_Cav1.3 could generate a positive chronotropic response to isoprenaline even on impairment of RyR2-dependent Ca²⁺ release, as well as inhibition of SERCA or NCX1. In addition, our observation that the loss of Rad-dependent regulation of I_Cav1.3/I_Cav1.2 reduces the chronotropic response, whereas not preventing phosphorylation of RyR2 and PLN at both PKA- and CaMKII-dependent phosphorylation sites, suggests that RyR2 cannot compensate for the impaired regulation of L-type channels. In sum, our data suggest that phosphorylation of RyR2 and activation of NCX1 are not sufficient for generating positive chronotropism.

Although the pacemaker mechanism in the human SAN is not yet fully understood, molecular and clinical studies are strongly suggestive of an important role of Ca_v1.3 channels. A multiomics study of human heart chambers revealed that the CACNA1D gene is one of the most upregulated genes in the cellular cluster of SAN pacemaker myocytes.⁶⁸ In this regard, the new role of Ca_v1.3 channels in the chronotropic effect of catecholamines that we describe in this study may help explain clinical observations on inherited and acquired forms of bradycardia in humans. Adverse interaction between the widely used antiarrhythmic drug amiodarone and the antiviral sofosbuvir in the Ca_v1.3 channel pore has been proposed as a mechanism to explain severe bradycardia in patients treated with these drugs.⁶⁹ Furthermore, several clinical observations on inherited forms of SAN dysfunction show that both Ca_v1.3 and HCN4 can mediate a positive chronotropic response under ablation or inhibition of the other channel, despite bradycardia at baseline. For example, a mutation that induces a complete loss-of-function in Ca_v1.3 channel conductance in humans underlies the sinus node dysfunction and deafness syndrome, resulting in resting bradycardia and heart block, underscoring the importance of these channels for human SAN pacemaking.⁷⁰ The compensatory action of cAMP-dependent regulation of human sinoatrial HCN channels could now explain preserved β-adrenergic chronotropic competence in patients with sinus node dysfunction and deafness syndrome. Similarly, carriers of a mutation in the CNBD domain of human HCN4, inducing loss-of-function in cAMP-dependent channel regulation, display resting bradycardia, but a normal capability to accelerate heart rate on a bicycle ergometer,⁴¹ or on pharmacological challenge with β-adrenergic agonist,⁷¹ possibly because of compensatory action of Ca_v1.3 channels.

In conclusion, our study contributes to the development of a new multiscale model of the ionic mechanism regulating heart pacemaking by β-ARs. This model can serve as a basis for improving understanding of clinical conditions caused by SAN chronotropic incompetence, for example, in heart failure and aging. Cardiac aging is characterized by diminished maximal heart rate and an age-related decrease in the chronotropic response.⁷² Age-related loss in chronotropic response may be explained by simultaneous disruption and remodeling of Ca_v1.3 and HCN expression or dysfunction in the gating regulation of these channels.

Article Information

Acknowledgments

The authors thank Jörg Striessnig (University of Innsbruck, Austria) for sharing Ca_v1.3^−/− and Ca_v1.2^DHP−/− mouse lines; Frank Schwede (Biolog Life Science, Institute GmbH & Co.KG) for the generous gift of cyclic di-(3′,5′)-GMP (c-di-GMP); and the personnel of the Réseau d’Animaleries de Montpellier and the IExplore mouse facilities of the Institut de Génomique Fonctionnelle, Montpellier, France. Requests for materials should be sent to M.E Mangoni or P. Mesirca.

Author Contributions

M.E. Mangoni, D. Isbrandt, S.O. Marx, A. D’Souza contributed to conceptualization and design. P. Mesirca, E. Torre, M. Faure, I. Bidaud, B. Engeland, B.-x. Chen, S. Reiken, M. Gaillardon, D. Isbrandt, S.O. Marx, A. D’Souza, A. Moroni, and A. Saponaro designed the methodology. E. Torre, M. Faure, I. Bidaud, P. Mesirca, M. Baudot, L. Talssi, S. Laarioui, A. Saponaro, W. Pereira de Vasconcelos, M. Gaillardon, and S. Reiken contributed to investigation and acquisition of data. M. Faure, P. Mesirca, E. Torre, and I. Bidaud contributed to visualization. M.E. Mangoni, S.O. Marx, A. Saponaro, and A.R. Marks contributed to funding acquisition. M.E. Mangoni, P. Mesirca, S.O. Marx, D. Isbrandt contributed to supervision. M.E. Mangoni contributed to writing the original draft. M.E. Mangoni, P. Mesirca, D. Isbrandt, S.O. Marx, A. Moroni, D. Isbrandt, A. D’Souza, A. R. Marks contributed to writing—reviewing and editing for important intellectual content.

Sources of Funding

This study is supported by Fondation Leducq TNE Fighting Against Sinus Node Dysfunction and Associated Arrhythmias (FANTASY,19CV03 to M.E. Mangoni, A. Moroni, and A. D’Souza); Fondation pour la Recherche Médicale Physiopathologie Cardiovasculaire (DPC20171138970 to M.E. Mangoni), Fondation pour la Recherche Médicale « Equipe FRM » (EQU202403018013 to M.E. Mangoni); Agence Nationale de la Recherche (ANR-15-CE14-0004-01 and ANR-23-CE14-0009-01, to M.E. Mangoni); and National Institutes of Health grants R01 HL146149, R01 HL140934, R01 HL155377, P01 HL164319, and R01 HL121253 to S.O. Marx. The Institut de Génomique Fonctionnelle group is a member of the Laboratory of Excellence (Labex) Ion Channel Science and Therapeutics, supported by a grant from Agence Nationale de la Recherche (ANR-11-LABX-0015). A. D’Souza was supported by a British Heart Foundation Intermediate Fellowship (FS/19/1/34035). E. Torre was supported by a Fondation Lefoulon Delalande postdoctoral fellowship.

Disclosures

None.

Supplemental Material

Expanded Materials and Methods

Figures S1–S7

Major Resources Table

ARRIVE Guidelines Table

Supplemental Statistical file

Unedited/Uncut Original Western Blots

Peer Review Report

Supplementary Material

res-138-e327497-s001.pdf ^{(5.6MB, pdf)}