Wild-type (WT, Charles River laboratory) and transgenic mice harboring Ca_v1.3^−/−, Ca_v1.3^+/−, or Girk4^−/− genotype on C57BL/6J genetic background were used in this study. Experiments were performed in accordance with the European Communities Council directive of November 1986 and conformed to the Guide for the Care and Use of Laboratory Animals” published by the US National Institutes of Health (NIH publication 8th Edition, 2011). All mice were maintained under environmentally controlled conditions (22 ± 2°C, 12 h light/12 h dark cycle) in the animal facility of the Institute. The surgical protocols were approved by the “Comité d’Ethique pour l’Expérimentation Animale Languedoc-Roussillon” (CEEA-LR) with the authorization number CE-LR-0814. For ex vivo experiments, the protocol “Study of physiological parameters and cardioprotection ex vivo” was approved by the SBEA (Structure Bien-être Animal) committee from the Institute according to the European directive 2010/63/EU and the French ministerial regulation of February 01, 2013. The “n” number of mice included in the study for each experiment is the “n” authorized by the ethics committee as being the minimal “n” number allowing a valid statistical analysis while respecting the 3R rule (Replacement, Reduction, Refinement) established in Europe.

Acute myocardial ischemia and reperfusion were induced in Ca_v1.3^−/−, Ca_v1.3^+/− and Girk4^−/−, as well as in WT controls as previously described (28). Adult male mice (22–32 g) were anesthetized with an intramuscular injection of ketamine (25 mg/kg; Imalgène 500, Merial, France), xylazine (10 mg/kg; Rompum 2%, Bayer, France), and chlorpromazine (1.25 mg/kg; Largactil 5 mg/ml, Sanofi-Aventis, France). Mice were ventilated via tracheal intubation with a Harvard rodent respirator (tidal volume, 7.2 µl/g body mass; respiratory rate, 200 breaths per min). Body temperature was maintained between 36.8°C and 37°C via a thermoregulated surgical table connected to a rectal probe. One-lead surface electrocardiograms (ECG) were recorded by Ag/AgCl gel-coated electrodes (Unomedical) attached to the superior right and both inferior limbs of mice and connected to a standard one-lead ECG amplifier module (EMKA Technologies, Paris, France). The recordings were carried out from 10 min before ischemia (baseline) until the end of the experience. Mice were randomly enrolled in the control group [NaCl 0.9%, intraperitoneal (IP) injection] or ivabradine-injected group (6 mg/kg, IP, 30 min before ischemia). Then, after a second injection of ketamine (25 mg/kg) and xylazine (10 mg/kg), the chest was opened by a left lateral thoracotomy, and a reversible coronary artery snare occluder was placed around the left coronary artery. All animals underwent 40 min of ischemia followed by 60 min of reperfusion by loosening of the knot. At the beginning of reperfusion, a third administration of ketamine/xylazine was performed (same dose). At the end of reperfusion, the artery was reoccluded, phthalocyanine blue was injected into the left ventricular (LV) cavity and allowed to perfuse the non-ischemic portions of the myocardium, and the heart was removed for infarct size (IS) analysis.

Adult male mice were anesthetized with an intraperitoneal dose of ketamine (100 mg/kg) and xylazine (10 mg/kg), followed by sodium pentobarbital (45 mg/kg; Ceva Santé Animale, France). An IP injection of heparin (7,500 IU/kg; Sanofi Aventis, France) was administered to avoid thrombus formation. A thoracotomy was performed and beating hearts were quickly removed and excised into ice-cold perfusion fluid. The aorta was cannulated under magnification and hearts were quickly mounted and perfused by the Langendorff method (29) on an isolated heart system apparatus (Isolated heart system, EMKA Technologies, France). The coronary circulation was perfused at 70 mmHg over a 1–4 ml/min flow range with modified Krebs-Henseleit buffer containing (mM): NaCl 116; KCl 5; MgSO₄ 1.1; NaH₂P0₄ 0.35; NaHCO₃ 27; CaCl₂ 1.8 and glucose 10. The buffer was gassed with 95% O_2% and 5% CO₂, giving a pH of 7.4 at 37°C. Before use, the buffer was initially passed through a 0.2 µm filter (Sarstedt, USA), and all perfusate delivered to hearts was filtered via an in-line 0.45 µm filter for removing microparticulates.

During the stabilization phase, a 7.0 prolene ligature (Johnson & Johnson Ethicon, Belgium) was placed around the left coronary artery 1–2 mm distal from where it emerges beneath the left atrium. The ends of the ligature were passed through an occluder tube, initially without tension. ECG was continuously recorded by Ag/AgCl electrodes placed on the right atrium and near the apex. Heart temperature was continuously assessed by a probe positioned on the base of the heart and recorded by using a TH-5 monitoring thermometer (Phymep, USA). Perfused hearts were then immersed in a water-jacketed bath and continuously maintained at 37°C. After a stabilization phase of 15 min (baseline), regional normothermic ischemia was then induced by tightening the suture with the occluder tube for 40 min. After ischemia, reperfusion was allowed by loosening the ligature and perfusing the heart for 60 min. For atrial pacing, a bipolar stimulating electrode was placed on the right atrium and pacing was started during the stabilization phase.

At the end of reperfusion, the artery was reoccluded, the hearts were removed from the perfusion system and phthalocyanine blue was injected through the aortic cannula to perfuse the non-ischemic portions of the myocardium, allowing the area at risk to be delineated.

During the whole ex vivo IR protocol, the coronary flow from the isolated heart system apparatus was continuously digitized at 0.1 Hz and recorded using an IOX data acquisition system (EMKA Technologies, France). Excel software (Microsoft, USA) was used to perform offline analysis of the data recorded. For each mouse, a mean value of maximal coronary flow (MaxF) was calculated every 10 min. Values during ischemia and reperfusion were normalized with respect to the pre-ischemia phase (baseline) and were expressed in percentages.

For both in vivo and ex vivo models, signals were digitized continuously at 1 kHz and recorded using an IOX data acquisition system (EMKA Technologies, France). EcgAuto software (EMKA Technologies, France) was used to perform offline analysis of the data recorded. For each mouse the mean heart rate (HR) value during baseline, ischemia and reperfusion was calculated.

Infarct sizes were measured using the technique developed by Fishbein et al. (30). Hearts from in vivo and ex vivo experiments were dissected and the LVs were sliced transversally into 1 mm-thick sections and incubated in a 1% (w/v) solution of 2.3.5-triphenyltetrazolium chloride (TTC; Sigma-Aldrich, St Louis, Mo) for 15 min at 37°C. After fixation in a 4% paraformaldehyde-phosphate-buffered solution, the slices were weighted, and each side was photographed with an Olympus camera. The ischemic risk area (unstained in blue) and the infarcted area (unstained by TTC) were outlined on each photograph and measured by planimetry using Image J (Scion corp., Frederick, MD). The area of each region was averaged from the photographs of each side for each slice, and multiplied by the weight of that tissue section. Infarct size was expressed as a percentage of the ischemic risk area (AR). Mice with an AR of less than 40 (in % of LV mass) and those greater than 60 were excluded from the study to ensure that infarct size values were compared in a pool of animals with the same infarct severity.

All values are expressed as mean ± SD. For comparisons between two groups, the Mann–Whitney non-parametric test was used. Multiple comparisons between groups were assessed by the Kruskal–Wallis nonparametric test followed by the Dunn's post hoc test or ANOVA (when the Normality Shapiro–Wilk test passed) or by the three-way ANOVA when considering three different variables (time, mutation and pacing). Values of p < 0.05 were accepted as statistically significant. The p values, indicated in the text, were noted in the figures as: ns for p > 0.05, * for p < 0.05, ** for p < 0.001, *** p < 0.01 and **** p < 0.0001. All data were analyzed with GraphPad Prism (GraphPad Software, San Diego, California, USA). A linear relationship between HR and IS was fitted by linear regression model using the least squares method (GraphPad Prism software) based on the minimization of the sum of squares of the vertical distances of the points from the line.

The hyperpolarization-activated current (I_f) plays an important role in the generation of heart rate. Therefore, we first investigated the effects of the selective I_f inhibitor ivabradine on IR injury (Figure 1A). Anesthetized WT mice were subjected to an in vivo surgery protocol. ECG was continuously recorded along the surgery protocol (Figure 1A; Supplementary Figure S1). In comparison with control WT mice (WT; 0.9% NaCl, IP, n = 6), ivabradine treated mice (WT Iva; 6 mg/kg; IP injection, 30 min before coronary ligation; n = 7) showed a significant decrease in both basal heart rate (179 ± 19 bpm vs. 240 ± 40 bpm, **p = 0.0047; 25%-reduction; Figure 1B) and infarct size (24.73% ± 5.25% vs. 36.37% ± 3.56%, **p = 0.0047; 32%-reduction; Figure 1C). The area at risk of ischemia (AR/LV mass) was similar among groups (p^ns= 0.9452; Supplementary Figures S2A).

Because ivabradine was described to provide also heart rate-independent benefit in cardioprotection (31), we evaluated the cardioprotective effects of heart rate reduction obtained through genetic deletion of Ca_v1.3 channels, a major contributor in the generation of diastolic depolarization. To this aim, anesthetized Ca_v1.3^+/− and Ca_v1.3^−/− mice were both subjected to the in vivo surgical protocol presented in Figure 2A with a continuous ECG recording along the protocol (Supplementary Figure S1). Figure 2B shows that heterozygous Ca_v1.3^+/− mice exhibit moderate heart rate reduction (22%) (187 ± 22 bpm, n = 6 for Ca_v1.3^+/− vs. 240 ± 40 bpm, n = 6 for Ctrl; p^ns= 0.3714) that was associated with a 26.7%-decrease in infarct size (26.67% ± 6.16%, n = 6 for Ca_v1.3^+/− vs. 36.37% ± 3.56%, n = 6 for WT; p* = 0.0267). In homozygous Ca_v1.3^−/− mice (n = 7), characterized by a marked reduction (54%) of the basal heart rate in comparison to WT counterparts (n = 6) (109 ± 14 bpm, n = 7 for Ca_v1.3^−/− vs. 240 ± 40 bpm, n = 6 for Ctrl; p*** = 0.0004), a 28.8%-decrease in infarct size vs. WT was observed (25.90% ± 3.28%, n = 7 for Ca_v1.3^−/− vs. 36.37% ± 3.56%, n = 6 for WT; p* = 0.0190; Figure 2C). Area at risk of ischemia (AR/LV mass) was similar among groups (p^ns= 0.0511; Supplementary Figure S2B), indicating that all hearts were exposed in vivo to quantitatively similar ischemic challenge. These results suggest that moderate heart reduction induced by genetic ablation of one allele of Ca_v1.3 channels exerted potent cardioprotective effect against IR injury.

In order to elucidate the correlation between heart rate and infarct size, we used genetically modified mice harboring different basal heart rates. We thus measured infarct size in Ca_v1.3^+/−, Ca_v1.3^−/− and Girk4^−/− mice subjected to the same IR protocol (Figure 3A) in control conditions (WT) and after ivabradine (WT Iva; 6 mg/kg) treatment. Figure 3B shows variation in basal heart rate according to the different genotypes, with heart rate of Ca_v1.3^−/− being the lowest and that of Girk4^−/− the highest among groups (Supplementary Figure S1). Mean heart rates of Ca_v1.3^−/−, Ca_v1.3^+/− and WT Iva were lower than that of WT counterparts (109 ± 14 bpm, n = 6; p**** < 0.0001 for Ca_v1.3^−/−; 187 ± 22 bpm, n = 6; p = 0.0282 for Ca_v1.3^+/− and 179 ± 19 bpm, n = 7; p** = 0.0069 vs. 240 ± 40 bpm, n = 6, for WT). By contrast, mean heart rates from Girk4^−/− and Girk4^−/− Iva were higher compared to WT [360 ± 54 bpm, n = 5; p****< 0.0001 for Girk4^−/− and 271 ± 32 bpm, n = 5; p^ns= 0.356 for Girk4^−/− Iva vs. WT (240 ± 40 bpm, n = 6). Figure 3B]. Figure 3C shows infarct size analysis among the various genotypes and the clear correlation between infarct size and heart rate. Girk4^−/− mutant mice exhibit the largest (47.26% ± 4.67%, n = 5) and Ca_v1.3^−/− the lowest (25.9% ± 3.28%, n = 7) infarct sizes. A moderate heart rate reduction in Girk4^−/− mice induced by a bolus injection of ivabradine (40.5% ± 7.09%, n = 7) yielded infarct sizes that were in between those recorded in WT and Girk4 ^−/− mice.

Of note, higher r² values of the linear regression were obtained for heart rate values recorded at baseline and during ischemia (r²= 0.6170 and r²= 0.6111, respectively; Figures 3D,E) than those recorded during reperfusion (r²= 0.3800; Figure 3F). Irrespective of heart rates recorded at baseline and the genotype of mice tested, the heart rate was significantly decreased during ischemia, while increased during reperfusion (Figure 3C; Supplementary Figure S1). The relationship between heart rate recorded at baseline and infarct size was preserved for heart rates recorded during ischemia (Figures 3D,E) and, to a lesser extent, during reperfusion (Figure 3F). AR/LV mass measured for each heart was similar among groups (p^ns= 0.0506) (Supplementary Figure S3).

Since the infarct size correlated with heart rate irrespective of the genotype in vivo, we investigated whether increase in heart rate prevented cardioprotection. To this aim, atrial pacing ex vivo was performed to increase heart rate in the absence of the vagus nerve influence (Figure 4). Hearts underwent the same protocol as in vivo through the implementation of a reversible ligation on the left coronary during 40 min ischemia followed by 60 min reperfusion (Figure 4A). In the atrial-paced group, hearts were placed at 475 bpm. Since Ca_v1.3^−/− hearts ex vivo exhibit complex atrioventricular conduction block (32), which could interfere in experiments of atrial pacing, we used hearts from Ca_v1.3^+/− associated with strong cardioprotection. Ca_v1.3^+/− hearts with a mean heart rate of 341 ± 20 bpm (n = 6) were paced at 475 bpm (n = 5), representing a 39%-increase. Infarct size in paced hearts was 44%-increased (74.22% ± 2.99%, n = 5 in paced vs. 51.52% ± 8.32%, n = 6 in un-paced Ca_v1.3^+/−; p** = 0.0043; Figure 4B). Same experiments were performed in C57Bl6 WT mice in order to overcome the influence genetic modifications. Heart rate at baseline was 398 ± 50 bpm in the WT group (n = 9) and upon atrial pacing (20% increment in heart rate), infarct size was increased by 32% in those hearts (56.33% ± 7.45%, n = 7 vs. 42.78% ± 7.45%, n = 9; p** = 0.0052; Figure 4C). Consistently, the relationship between infarct size and heart rate was maintained ex vivo by pacing WT hearts to lower heart rate values including 200, 300 and 400 bpm. Figure 4D shows a linear correlation established between infarct size and heart rate in the range of 200–475 bpm (Supplementary Figure S5).

Since heart rate reduction prolongs the ventricular diastole allowing to augment the coronary perfusion, we recorded the coronary flow during ex vivo experiments (Figure 4E; Table 1). Our results show that coronary flow dropped during the ischemic phase (≈50%) in WT and Ca_v1.3^+/− paced or un-paced hearts. However, during reperfusion, the coronary flow of un-paced WT and Ca_v1.3^+/− hearts recovered and stabilized above 100% of the control baseline level, whereas the flows in paced hearts of both genotypes did not fully recover after stabilization and was reduced by 25% compared to the un-paced condition (p**** < 0.0001).

Taken together, these results demonstrate positive correlation between heart rate and infarct size ex vivo and suggest that the cardioprotective effect associated to a decreased heart rate might be related to a better coronary circulation during reperfusion.