A detailed list of reagents is presented in the supplementary data.

CS was induced in 20 Yorkshire female pigs (37.8±5.2kg); the survival rate was 90%. At the end of the experimental procedure, 10 animals received 0.3mg/kg ivabradine and 10 received placebo (saline). Two animals (1 from each group) did not survive beyond 24hours after the CS.

Pigs were premedicated with intramuscular ketamine (10mg/kg) and midazolam (0.5mg/kg). Anesthesia was induced by isoflurane inhalation and maintained with continuous infusion of propofol (2mL/kg/h), fentanyl (50mg/kg/h), and diazepam (10mg/kg/h). Animals were intubated and ventilated with 100% oxygen saturation. The animals received 5000 IU of heparin and amiodarone (2mg/kg/h) to avoid blood clotting in catheters and malignant cardiac arrhythmias, respectively.

Ischemia/reperfusion was induced by left anterior descending artery occlusion for 45minutes using a JL 3 6-Fr catheter and balloon inflation at 5 atmospheres. Volume overload was induced by administration of 500mL hydroxyethyl starch and 1500mL of 0.9% saline serum. Hemodynamic measurements were recorded at 15, 30, and 45minutes of ischemia. When ventricular fibrillation/ventricular tachycardia occurred, we administered a biphasic DC shock (10-20J) combined with direct manual chest compressions. Noradrenaline (20-80μg/kg/h), dobutamine (240-660μg/kg/h), and physiological saline (1000-2000mL) were administered until the heart rate exceeded 90 bpm. After 45minutes of left anterior descending artery occlusion, the animals were then randomized to a control or ivabradine group.

Ivabradine powder was weighed on a high-precision laboratory scale and diluted to at least 12mg/mL in distilled water. The solution was then transferred to a 10-mL syringe for slow-bolus intravenous administration to the animal at a dose of 0.3mL/kg. The control group received an equivalent volume of physiological saline.

Myocardial infarction extension was evaluated by 5% Evans blue perfusion (in serum) and TTC staining as described previously12 (for details, see the supplementary data).

Plasma troponin I was determined with the commercial Human Cardiac Troponin 1 SimpleStep ELISA kit (Abcam, Spain) following the manufacturer's instructions.

Pig hearts were visualized by echocardiography using a Vivid Q ultrasound system from GE Healthcare (United States) equipped with a 1.9- to 4.0-MHz scan head, as described previously.13 Further information is provided in the supplementary data.

H9c2 and HL1B cells were grown as described previously14 and in the supplementary data. Cells were cultured under hypoxic conditions in hypoxia incubator chambers in a humidified atmosphere containing 1% oxygen, 5% carbon dioxide, and 94% nitrogen. Reoxygenation was performed with fresh medium.

Histological, immunohistochemical, and immunohistofluorescence procedures were performed as previously described.13,14

Isolation of protein lysates and immunoprecipitations and immunoblots were performed as described.13

Real-time polymerase chain reaction (RT-PCR) was performed as previously described.8 Additional information is provided in the supplementary data. The following primers were used: EMMPRIN-Forward: 5’-GGC ACC ATC GTA ACC TCT GT-3’; EMMPRIN-Reverse: 5’-CAC TGG CGT GTT CCG ATT TC-3’; β-actin-Forward: 5’-CTT AGT TGC GTT ACA CCC TTT CT-3’; β-actin-Reverse: 5’-CTG TCA CCT TCA CCG TTC CAG TT-3’.

All data were analyzed using SPSS 22.0 statistical software package (SPSS Inc, United States). All values are given as mean±standard deviation. Significance is reported at the 5% level. Whenever comparisons were made with a common control, the significance of differences was tested by analysis of variance followed by Dunnett's modification of the t test.

In pigs subjected to CS (figure 1A), intravenous administration of 0.3mg/kg ivabradine or saline was associated with a significant reduction in heart rate (figure 1B) but preserved stroke volume (figure 1C). The levels of cardiac troponin, indicative of ischemic injury, were reduced 24hours after CS (figure 1D). The effects of ivabradine on cardiac function were determined by cardiac ultrasound, which revealed a significant improvement in left ventricular ejection fraction by day 7 after CS in animals treated with ivabradine (figure 1E). This improvement was closely correlated with a decline in the left ventricular necrotic area, as detected by Evans blue/TTC staining (figure 1F).

Figure 1.

Ivabradine improves cardiac function in a porcine model of CS. A: schematic representation of the assay. Effects of 0.3mg/kg ivabradine or placebo on heart rate (B) and stroke volume (C) after CS. D: cardiac troponin before and 24hours after CS (N=9/group, mean±standard deviation; * P <.05 for ivabradine vs placebo). Left ventricular ejection fraction (N=9/group) (E) and necrotic area (N=5/group) (F) as a percentage of necrotic tissue (pale) vs the area at risk (red) 7 days after CS (mean±standard deviation; * P <.05 for ivabradine vs placebo). CS, cardiogenic shock.

(0.52MB).

Coronary ischemia/reperfusion induces the expression of several cardiac ECM-degrading enzymes, including MMPs. Administration of 0.3mg/kg ivabradine reduced the left ventricular levels of MMP-9 in the necrotic area by day 7 after CS, as detected by immunohistochemistry and immunoblotting with anti-MMP-9 antibody (figure 2A,B, respectively).

Figure 2.

Ivabradine reduces the levels of extracellular matrix-degrading enzymes in the heart. A: detection of MMP-9 in the necrotic areas of hearts subjected to CS in response to placebo or ivabradine (N=5/group, mean±standard deviation; * P <.05 for necrotic ivabradine vs necrotic placebo). B: immunoblotting of the 92-kDa pro-MMP-9 from healthy (H), at risk (R), and necrotic (N) sections of hearts subjected to CS (N=5/group, mean±standard deviation; * P <.05 for necrotic ivabradine vs necrotic placebo). CS, cardiogenic shock; MMP-9, matrix metalloproteinase-9.

(0.39MB).

The role of the MMP inducer EMMPRIN in the expression and activity of MMP-9 has previously been described.3 In our pig model, CS-induced expression of EMMPRIN mRNA was not inhibited by administration of 0.3mg/kg ivabradine (figure 3A). However, ivabradine played a significant posttranslational role, reducing the levels of EMMPRIN and EMMPRIN glycosylation in response to CS. Likewise, ivabradine did not affect the proteolysis of EMMPRIN, as seen by incubation of myocyte H9c2 cell lysates with the proteasome inhibitor MG-132 (figure 3B). Furthermore, incubation of resting H9c2 cells with ivabradine did not inhibit EMMPRIN glycosylation (a key step in EMMPRIN-mediated MMP expression and activity), as seen with treatment of cells with 5μg/mL of the N-linked glycosylation pharmacological inhibitor tunicamycin for 24hours (figure 3C). Similar results were obtained in myocyte HL1B cell cultures (figure 1 of the supplementary data). Conversely, in vivo, and in response to CS, ivabradine significantly reduced the high-glycosylated forms of EMMPRIN (HG-EMMPRIN), resulting in the accumulation of the low-glycosylated forms of EMMPRIN (LG-EMMPRIN), when compared with pigs subjected to CS and treated with placebo, through unknown mechanisms (figure 3D).

Figure 3.

Ivabradine reduces the levels of extracellular matrix-degrading enzymes in the heart. A: RT-PCR expression of EMMPRIN mRNA in the hearts of pigs treated with 0.3mg/kg ivabradine. GAPDH was used as a control (N=3/group, mean±standard deviation). B: immunoblotting of EMMPRIN in H9c2 cardiac cells incubated with MG-132. C: immunoblotting of EMMPRIN in H9c2 cardiac cells incubated with tunicamycin. The arrow indicates the low-molecular-weight 20-kDa native nonglycosylated EMMPRIN. D: expression of EMMPRIN in necrotic areas of hearts subjected to CS and incubated with ivabradine (N=3/group, mean±standard deviation; * P <.05 for HG- vs LG-EMMPRIN with CS + ivabradine). Beta-tubulin was used as a control (B-D).

(0.25MB).

Caveolae are sphingolipid- and cholesterol-enriched vesicles that harbor many signaling pathways playing key roles in cell survival, death, proliferation, and/or migration. In the heart, the principal component in caveolae is caveolin-3, which exerts cardioprotective functions by binding to several proteins. We previously demonstrated that nitric oxide induces cardiac protection by stabilizing the caveolin-3/LG-EMMPRIN complex and thereby preventing HG-EMMPRIN-dependent matrix metalloproteinase-induced cardiac necrosis.14 Here, we detected strong colocalization of caveolin-3 and EMMPRIN in response to 0.3mg/kg ivabradine (figure 4A, ivabradine merged panel, yellow). Protein coimmunoprecipitation assays from the same hearts used in the above experiments revealed an increase in LG-EMMPRIN bound to caveolin-3 (figure 4B) in response to ivabradine, suggesting that ivabradine may reduce CS-induced ECM degradation by preserving the caveolin-3/LG-EMMPRIN complex.

Figure 4.

Ivabradine prevents disruption of the EMMPRIN/caveolin-3 complex in pigs subjected to cardiogenic shock. A: confocal microscopy detection of caveolin-3 (Cy3, red) and EMMPRIN (FITC, green) in heart sections of pigs treated with ivabradine or placebo (colocalization in merged panels=yellow) (N=3/group, mean±standard deviation; * P <.002 for ivabradine vs placebo). B: immunoblotting of caveolin-3 and EMMPRIN in caveolin-3-immunoprecipitated extracts (N=3/group, mean±standard deviation; * P <.05 for CS vs CS IVA).

(0.52MB).

In addition to EMMPRIN, caveolin-3 also binds to HCN4,15 the main target of ivabradine. Confocal microscopic visualization of caveolin-3 in ventricular sections of healthy hearts enabled the detection of a protein complex comprising caveolin-3, EMMPRIN, and HCN4 (figure 5A), although an even more intriguing discovery was the finding that, after CS, the association between caveolin-3/EMMPRIN and caveolin-3/HCN4 was increased in the presence of ivabradine vs placebo (figure 5B). This result suggests that caveolin-3 may be a target of ivabradine in cardioprotection by, on the one hand, preventing degradation of the ECM via increased binding to EMMPRIN and, on the other, stabilizing the complex with HCN4 and reducing the heart rate, as previously seen (figure 1). An interesting discovery is the formation of a HCN4 and EMMPRIN complex and the ability of ivabradine to uncouple the binding of the 2 proteins.

Figure 5.

Ivabradine regulates the binding of caveolin-3, HCN4, and EMMPRIN. A: confocal microscopy detection of caveolin-3 (FITC [green]), HCN4 (TRITC 561 [red]), and EMMPRIN (Alexa Fluor 350 [blue]) in heart sections of healthy pigs (A) or in pigs 7 days after CS (B) and treated with 0.3mg/kg ivabradine or placebo (N=3/group, mean±standard deviation; * P <.05 for HCN4/CAV3 placebo vs ivabradine; #P <.002 for HCN4/EMMPRIN placebo vs ivabradine; @P <.05 for HCN4/EMMPRIN/CAV3 placebo vs ivabradine). CS, cardiogenic shock; FITC, fluorescein isothiocyanate; TRITC, tetramethylrhodamine-isothiocyanate.

(1.3MB).

To test whether caveolin-3 can serve as a docking point between HCN4 and EMMPRIN, we silenced the expression of caveolin-3 with caveolin-3-specific siRNAs in H9c2 cardiac cells. With 10 nM siRNA, caveolin-3 expression was reduced by more than 75% vs nontransfected cells or cells transfected with scrambled siRNA (figure 6A). HCN4 bound to EMMPRIN in both caveolin-3-expressing and -silenced cells, suggesting the existence of a direct interaction (figure 6B). On the other hand, incubation with ivabradine yielded a significant uncoupling of HG-EMMPRIN bound to HCN4, independent of caveolin-3 expression (figure 6B). Likewise, the increased expression of MMP-9 in caveolin-3-silenced cells was downregulated by ivabradine (figure 6C), indicating that ivabradine may target caveolin-3 to prevent ECM degradation.

Figure 6.

Lack of caveolin-3 impairs formation of the EMMPRIN/HCN4 complex. A: immunoblotting of caveolin-3 from H9c2 cells transfected with Lipofectamine (no siRNA) or 1, 10, and 25nM caveolin-3 siRNA1. Beta-tubulin was used as a negative control. B, upper panel: immunoblotting of HCN4 from extracts immunoprecipitated with HCN4 from H9c2 cells silenced with 10nM caveolin-3 siRNA1; lower panel: immunoblotting of EMMPRIN from the same extracts (N=3). C: immunoblotting of MMP-9 in H9c2-silenced cells incubated with 0.3mg/kg ivabradine (N=3, mean±standard deviation; P <.001 for siRNA vs siRNA + ivabradine). MMP-9, matrix metalloproteinase-9; siRNA, small interfering RNA.

(0.19MB).

In the same way, we also used EMMPRIN-specific siRNAs to investigate the role of EMMPRIN in the complex between caveolin-3 and HCN4 (figure 7A). Inhibition of EMMPRIN reduced the levels of the HCN4/caveolin-3 complex, as determined by immunoblot detection of HCN4 in caveolin-3-immunoprecipitated cell lysates of H9c2 cells (figure 7B, line 3). This result indicates that EMMPRIN is required to maintain the HCN4/caveolin-3 complex and that their interaction is partly restored by incubation with ivabradine (figure 7B, lane 4), suggesting that EMMPRIN-independent binding of caveolin-3 and HCN4 occurs in response to ivabradine.

Figure 7.

Lack of EMMPRIN impairs formation of the caveolin-3/HCN4 complex. A: gene silencing of EMMPRIN with EMMPRIN-specific siRNA. Beta-tubulin was used as a loading control. B: immunoblotting of HCN4 from extracts immunoprecipitated with anti-caveolin-3 from H9c2 cells silenced with 10nM EMMPRIN siRNA1 (N=3, mean±standard deviation; P <.004 for HCN4 siRNA vs siRNA + ivabradine). siRNA, small interfering RNA.

(0.17MB).

One of the study limitations is the use of female pigs because females and males have different cardioprotective physiologies,32 which may compromise the level of the heart injury and heart failure related to the ischemic event. The use of rat H9c2 cell cultures is also a limitation. Isolation of pig cardiac myocytes would be the best experimental approach for testing the specific parameters addressed in this work. However, the extremely difficult isolation of living porcine cardiac myocytes led us to use H9c2 and HL1B cells to assess hypoxia and reoxygenation.