The experimental model of myocardial infarction-reperfusion that is described has already been used and has been extensively validated (Yoshimizu et al., 1986; Näslund et al., 1992; Kraitchman et al., 2000; Reffelmann et al., 2004; Krombach et al., 2005; Pérez de Prado et al., 2009). Compared to the occlusive infarction model, this infarction with reperfusion model offers the advantage of producing a more heterogeneous epicardial scar (Garcia-Dorado et al., 1987). Balloon occlusion of the mid left anterior descending artery, just distal to the first diagonal branch, was performed through percutaneous femoral access to achieve a Thrombolysis In Myocardial Infarction grade 0 flow for 150 min. Before occluding the artery, an IV bolus of lidocaine (1 mg/kg) was administered to prevent ischemia-induced ventricular arrhythmias. In the event animals presented ventricular tachyarrhythmias during the occlusion of the coronary artery, advanced cardiopulmonary resuscitation maneuvers were started following the resuscitation guidelines. After the infarction, pigs were started on aspirin 500 mg QD and clopidogrel 300 mg QD until their sacrifice and amiodarone 400 mg QD for 5 days postinfarction.

Two CMR studies were performed at 4 and 16 weeks after the induction of myocardial infarction with a 1.5 Tesla scanner (Intera, Philips, the Netherlands) using a 5-element dedicated cardiac coil. Following the same protocol that our working group has already used in previous works (Perez-David et al., 2011), the CMR study consisted of cine steady-state free-precession (SENSEx2, repetition time 2.4 ms, echo time 1.2 ms, average in-plane spatial resolution: 1.62 mm × 2 mm, 30 phases per cycle, 8-mm slice thickness without gap) and late gadolinium enhancement imaging (3D inversion-recovery turbo gradient echo sequence, pre-pulse delay optimized for maximal myocardial signal suppression; SENSE x2, flip angle 15°, repetition time 3.4 ms, echo time 1.3 ms, actual spatial resolution: 1.48 mm × 1.66 mm, interpolated spatial resolution 1.29 mm × 1.29 mm, 5-mm actual slice thickness, inversion time 200–300 ms; acquisition window was set to 150–170 ms). Images were obtained in short-axis views and 4-, 2-, and 3-chamber standard views. Late gadolinium enhancement images were obtained 10 min after injection of gadodiamide 0.2 mmol/kg. Cine and late gadolinium enhancement images were obtained to determine left ventricular end-diastolic volume (LVEDV), left ventricular end-systolic volume (LVESV), left ventricular ejection fraction (LVEF), total scar mass (TSM), and total myocardial mass (TMM) using specialized software (QMass MR version 7.0, MEDIS, the Netherlands). The scar was defined as those areas with voxel intensity values ≥2 standard deviations with respect to that of remote healthy myocardium (HM). (Yan et al., 2006; Perez-David et al., 2011).

Endocardial and epicardial scar growth was assessed by comparing the endocardial and epicardial signal intensity maps between the two CMR studies. Following the method previously published by our working group, signal intensity maps were obtained through the postprocessing of late gadolinium enhancement sequences, where the average signal intensity of the internal (endocardium) and external (epicardium) half of the myocardial thickness was projected into two separate 3D reconstructions (Perez-David et al., 2011; Arenal et al., 2014; Ávila et al., 2015). The left ventricular endo- and epicardial contours were defined on contiguous short-axis slices using the software QMass MR. Using custom software developed in the MATLAB environment (MathWorks, Natick, Massachusetts) 3D endo- and epicardial reconstructions were computed from an image volume integrating the short-axis, 4- and 2-chamber. The 3D visualization interface was implemented in Java (Sun Microsystems) using VTK (Kitware, Clifton Park, New York) visualization algorithms. Voxels with intensity values ≥2 but <3 were coded as heterogeneous tissue (HT), whereas voxels with values ≥3 were coded as dense scar (DS) (Yan et al., 2006; Perez-David et al., 2011). Dense scar and heterogeneous tissue were differentiated based on signal intensity, thus allowing the measurement of endocardial dense scar (EnDS), epicardial dense scar (EpDS), endocardial heterogeneous tissue (EnHT), and epicardial heterogeneous tissue (EpHT). The effect of CDCs was assessed by comparing the percentage change in EnDS, EpDS, EnHT, and EpHT. Each percentage change was calculated by dividing the difference of the variable between the first and second CMR by the value of the variable in the first CMR.

Four weeks after the creation of the myocardial infarction, animals were randomized to a treatment group (n = 10) that received intrapericardial allogenic CDCs (300,000 CDCs/kg in HypoThermosol solution) via an angiographically-guided subxiphoid puncture or to a control group (n = 9) where only the CDCs diluting solution was administered. CDCs were obtained from the hearts of other sacrificed healthy Large White pigs. Phenotypic and molecular characterization of CDCs were performed by flow cytometry and polymerase chain reaction, as detailed in previous works (Blázquez et al., 2016).

High-density electroanatomic mapping of the ventricle was performed with the CARTO3 system (Biosense Webster, CA, United States) using a PentaRay catheter (Biosense Webster, CA, United States) through retroaortic access. A filling density threshold of ≤5 mm was established for the regions with bipolar electrograms of amplitude ≤1.5 mV. Using the measurement tool included in CARTO3, the areas of total scar (bipolar electrograms ≤1.5 mV) and DS (≤0.5 mV) were calculated. Electrograms with amplitudes comprised between 0.5 and 1.5 mV were color-coded as HT.

Following the guidelines of the American Veterinary Medical Association, once the electrophysiological study was completed, pigs were euthanized. Animals under deep general anesthesia were placed in a supine position on the surgical table. A sternotomy was performed to expose the heart and the great vessels. Following heparinization (unfractionated heparin, 1 mg/kg) and cardiac arrest after a rapid IV injection of potassium chloride, heart extraction was performed expeditiously. After explanting the heart, a cannula was inserted into the ascending aorta to retrogradely perfuse the heart with a cardioplegic solution at 4°C (pH 7.4; solution composition in mM: 140 NaCl; 5.4 KCl; 1 MgCl²; 5 HEPES buffer; 11 glucose; 1.8 CaCl₂).

For the epicardial optical mapping, the heart was removed from the cold solution and suspended from the aortic cannula to initiate retrograde warm perfusion with 1 L of modified Krebs solution at 36.5°C (pH 7.4; solution composition in mM: 120 NaCl; 25 NaHCO₃; 1.8 CaCl₂; 5.4 KCl; 1 MgCl2; 5.5 glucose; 1.2 H2PO₂H₂O). When the Krebs solution was clear of blood, the circuit was closed to enable recirculation. The Krebs solution was oxygenated by constant bubbling of carbogen gas. If the heart presented ventricular tachyarrhythmias when it was transferred to the Langendorf system (Figure 2), repeated 5 J shocks were administered with external paddles to return to normal sinus rhythm. A dose of 10 mm 2,3-butanedione monoxime (Biotium, Inc. Hayward, California, United States) was administered to achieve electromechanical dissociation and avoid hypoxia and motion artifacts during optical mapping. Despite there are other available excitation-contraction uncoupling agents, like blebbistatin, which have their pros and cons compared with 2,3-butanedione monoxime, given the randomized nature of this study, the agent chosen is not expected to alter the results. A bolus of 100 μL of the voltage dye di-4-ANEPPS (Biotium) was administered and when its distribution was not homogeneous, repeated administrations were made.

To excite di-4-ANEPPS, hearts were illuminated with a filtered green LED light source (maximum power output 58 W; maximum wavelength 524 nm) (Luminus Devices, Billerica, United States) through a plano-convex lens (LA 1951; focal length = 25.4 mm) (Thorlabs, New Jersey, United States) and a green excitation filter (D540/25X; Chroma Technology, Bellows Falls, United States). Three light sources were used to achieve homogeneous illumination. Both the Langendorff system and the protocol for performing optical mapping have been used and validated by other working groups (Langendorff, 1895; Efimov et al., 2004; Herron et al., 2012; Boukens and Efimov, 2014; Pavey et al., 2018; Swift et al., 2019; O’Shea et al., 2020).

A custom software developed in the MATLAB (MathWorks, Natick, Massachusetts) environment was used to perform image recording (Electron Multiplygin Charge-Coupled Device Evolve 128 camera, 128 × 128 pixels, 24-μm square pixel, 16 bit) (Photometrics, AZ, United States) and processing during drive train pacing from the right ventricular epicardium at a 600 ms cycle length with rectangular pulses (amplitude 10 V and duration 2 ms) to ensure propagation was longitudinal to the epicardial fibers in the left ventricle (Figure 3). Conduction velocity (CV) was automatically estimated by calculating gradients of normal vector fields on isochronal propagation lines. Action potential duration (APD) of optical voltage signals was calculated at 80% repolarization. CV and APD were measured in the area covered by the camera where the HM and the HT were differentiated, the latter characterized by the presence of muscle fibers and fibrotic strands. Percentage changes in CV and APD were calculated by dividing the difference in the variable between the HT and HM by the value of the variable in the HM, i.e., ( H T − H M ) / H M , and compared between groups.

Explanted hearts were fixed in cuvettes with formalin at room temperature. The scar was identified macroscopically and cut into transverse slices from apex to base for inclusion in paraffin. These slices were further cut into 4 μm-thick slices and were used for Masson’s trichrome staining and immunohistochemistry with the different primary antibodies against the different antigens (von-Willebrand factor; endothelial antigen CD31; and connexin 43) (Bio-Rad, CA, United States) after deparaffinization.

With Masson’s trichrome stain three types of tissue were differentiated: 1) DS, defined by the absence of myocytes; 2) HT, characterized by the presence of myocytes and strands of fibrosis that separate the cardiomyocyte bundles; and 3) HM, defined by normal cardiomyocytes with no intervening fibrosis. Analyses of the HT in the endocardial and epicardial halves of the myocardial thickness were performed independently. Measurements were made with a 20 mm × 40 mm transparent grid divided into 1250 squares to allow the conversion of semi-quantitative variables into quantitative variables.

The following semi-quantitative variables were determined for each HT compartment: fibrosis extension, the quantity of connexin 43, connexin 43 correct polarity, and cardiomyocyte viability (i.e. myocytes with no signs of hypertrophy, cytolysis, vacuolar degeneration or edema). For each of the semi-quantitative variables, a numerical value was obtained in such a way that +, ++, +++, ++++, and +++++ became 20, 40, 60, 80, and 100%. Since the scar was examined in different histological sections for each of the animals, a measurement of the extension of the HT on each of them was made. Depending on the area, each of the variables was subsequently weighted to obtain an average value of each variable per animal.