All studies and procedures involving animals were in strict accordance with the recommendations of the European Community Directive Article (2010/63/UE) for the protection of vertebrate animals used for experimental and other scientific purposes. The project was specifically approved by the regional ethics committee and by the French Ministry of Research (APAFIS N° 36487-2022040816108385 v.7). All experimental protocols for the patch clamp studies were approved by the Austrian Science Ministry (BMWFW-66.009/0175-WF/V/3b/2015).

Dystrophin-deficient mdx mice on the BL10 background (C57BL/10ScSn-Dmdmdx/J) (Ebner et al., 2020) were cross-bred with a transgenic mouse line (Cx40eGFP/+; BL10 background) expressing eGFP under the control of the connexin 40 (Cx40) gene (Miquerol et al., 2004). Only males aged between 1 and 12 months were used for the experiments. Genotyping of the mice was performed using standard PCR assays.

Mice were euthanized by cervical dislocation and hearts from 12-month-old animals were excised and immediately received a perfusion of PBS-KCl (50 mM) by the aorta.

For histological studies, adult hearts were fixed overnight in 4% paraformaldehyde (vol/vol) in PBS, washed in sucrose gradient, then embedded in OCT and cryosectioned. For immunofluorescence, sections were permeabilized in PBS 1X/0.2% Triton X100 for 20 min and incubated for 1 h in saturation buffer (PBS 1X/3% BSA/0.1% Triton X100). Primary antibodies were incubated in saturation buffer overnight at 4°C. Secondary antibodies coupled to fluorescent molecules were incubated in saturation buffer and after washes, hearts were observed under a Zeiss Apotome microscope.

For whole-mount immunofluorescence, the left ventricular wall of adult hearts was opened by scissors and pinned on a petri dish to expose the septal surface and fixed in 4% paraformaldehyde for 2 h at 4°C, washed in PBS, permeabilized in PBS 1X/0.5% Triton X100 for 1 h and incubated for 3 h in saturation buffer (PBS 1X/3% BSA/0.1% Triton X100). The primary antibodies were incubated in saturation buffer for 24 h at 4°C. Secondary antibodies coupled to fluorescent molecules were incubated in saturation buffer and after washes, hearts were observed under a Zeiss LSM780 confocal microscope.

Antibodies used in this study were specific to Contactin-2 (AF1714, R&D system), GFP (AbD Serotec), dystrophin (12715-1-AP, ProteinTech), WGA-Cy3 (29076-1, Clinisciences) and WGA-Cy5 (29024-1, Clinisciences). The antibody against Cx43 is homemade and previously described (Gros and Jongsma, 1996).

Wheat Germ Agglutinin (WGA) staining was quantified using Fiji (Fiji Is Just ImageJ). Images were converted to 8-bit grayscale, and a manual threshold was applied to isolate the WGA-positive signal. Threshold values were adjusted consistently across all samples, based on intensity histograms and visual observation, to ensure that the stained regions were accurately defined. A region of interest (ROI) was manually delineated to include only cardiac tissue, excluding background and non-specific areas. Within the ROI, the number of WGA-positive pixels was expressed as a percentage of the total number of tissue pixels: (number of WGA-positive pixels/total number of tissue pixels) × 100.

Surface ECGs were performed on anesthetized mice. An induction with 5% isoflurane was followed by maintenance at 1%–2% in a constant flow of oxygen at 700 mL/min. ECGs were recorded with a bipolar system in which the electrodes were placed subcutaneously at the right (negative) and left forelimb (reference) and the left hindlimb (positive) for lead II, at the right (reference) and left forelimb (negative) for lead III. Electrodes were connected to a Bioamp amplifier (AD Instruments) and were digitalized through a PowerLab 26T (AD Instruments). Digital recordings were analyzed with LabChart software version 8.1.13 (AD Instruments). Events were registered to 100 K/s and were filtered to 50 Hz. ECG recordings were obtained for 3 min after stabilization of the signal. Post-analysis was performed for heart rate, PR, QRS, QT intervals, T, R and S durations and T, R, S and QRS amplitudes. Body temperature was monitored using a temperature probe and maintained above 36°C and a warm pad (A-2101-00298, Intellibio). The two first electrocardiograms were recorded at 1 month and 3 months of age, then every 3 months until 12 months of age.

An Isoproterenol (ISO) stress test was performed at 12 months of age. After recording an ECG as described above, mice received a single intraperitoneal injection of ISO at a dose of 2 μg/g body weight. ISO was prepared as follows: DL-Isoproterenol hydrochloride (I5627, Sigma) was dissolved in ddH2O and vortexed to provide a 10 μg/μL ISO stock solution.

Vector cardiograms were obtained based on lead I (X-axis) and aVF (Y-axis) (lower) and represent the orientation of the main electrical axis of the heart.

Cardiac Purkinje fibers were isolated from wild-type (WT)- and mdx-Cx40 ^eGFP/+ mice as previously described (Ebner et al., 2020). The mice were anesthetized with isoflurane (2%, inhalation) and killed by cervical dislocation. After excision of the heart, a cannula was inserted into the aorta. The heart was then retrogradely perfused with calcium-free solution comprising 0.17 mg/mL Liberase TH (Roche) at 37°C for 18 min using a Langendorff setup. To further liberate Purkinje fibers, a second digestion step was applied, for which the ventricles were cut open along the aorta and placed in a culture dish containing 0.17 mg/mL Liberase TH (in calcium-free solution) for 8 min at room temperature. The tissue was then pulled into small pieces and incubated on a shaker at 37°C. Over 30 min, the Ca concentration was increased to 150 μM in four steps. The cells were then resuspended in Minimum Essential Medium (MEM)-α containing ITS media supplement (diluted 1:100), 2 mM L-glutamine, 100 U/mL penicillin, 0.1 mg/mL streptomycin and 17 μM blebbistatin (Sigma-Aldrich). Cells were then seeded on Matrigel (Corning)-coated 3.5 cm culture dishes for electrophysiological recordings. For experiments with ventricular cardiomyocytes of the working myocardium, hearts from mdx mice lacking Cx40-controlled eGFP expression were used for cell isolation. The procedure was as for Purkinje fiber isolation, except for the use of only a single digestion step, for which the hearts were perfused with 0.17 mg/mL Liberase TH (in calcium-free solution) at 37 °C for 10 min using a Langendorff setup (Hugo Sachs Elektronik).

I_Na in isolated cardiac Purkinje fibers and ventricular cardiomyocytes of the working myocardium was recorded up to 6 h after cell isolation utilizing the whole cell patch clamp technique. The measurements were performed at room temperature (22°C ± 1.5 °C) using an Axopatch 200B patch clamp amplifier, a Digidata1440 digitizer and Clampex 10.7 software (Axon Instruments, Union City, CA, United States). Patch pipettes were formed with a P-97 horizontal puller (Sutter Instruments, Novato, CA, United States) from aluminosilicate glass capillaries (A120-77-10; Science Products, Hofheim, Germany). They were filled with a solution consisting of (in mM) 5 NaCl, 110 CsF, 10 EGTA and 10 HEPES, adjusted to pH 7.3 with CsOH. Tip resistances lay between 1 and 1.5 MΩ. The bath solution contained (in mM) 5 NaCl, 135 N-methyl-D-glucamine, 2.5 KCl, 1 CaCl₂, 1 MgCl₂, 10 HEPES and 0.017 blebbistatin, adjusted to pH 7.4 with HCl. Fresh bath solution was continuously administered to the patched cells using a DAD-8-VC superfusion system (ALA Scientific Instruments, Westbury, NY, United States). Purkinje fibers were identified by their eGFP signal and morphological characteristics, as previously described (Ebner et al., 2020). I_Na was activated by 25 ms depolarizations ranging from −87 to −7 mV. Voltages were corrected for the liquid junction potential. Recordings were low-pass filtered with 10 kHz and digitized at 35 kHz. Analysis of the data was carried out with Clampfit 10.7 (Axon Instruments) and GraphPad Prism 8 (San Diego, CA, United States). I_Na peaks were measured and divided by the membrane capacitance to calculate current densities. To obtain current density-voltage relationships, these values were then plotted against the test pulse voltages. For curve fitting, the following function was used: I = G_max·(V-V_rev)/(1+exp ((V₅₀-V)/K)), where I is the current, G_max is the maximal conductance, V is the membrane potential, V_rev is the reversal potential, V₅₀ is the voltage at which the half-maximal activation occurred, and K is the slope factor.

Data are expressed as means ± standard deviation (SD). The I_Na data are expressed as means ± standard error (SE). Significant differences for electrocardiogram parameters were determined using two-way analysis of variance (ANOVA) followed by Sidak post hoc testing. Significant differences for heart size were determined using an unpaired t-test. All tests were made with Graphpad Prism software (Graphpad Prism 9.5.1, La Jolla, CA, United States). A nested analysis respecting the hierarchical data structure (Sikkel et al., 2017) was used for statistical comparisons of the patch clamp data. A p-value <0.05 was considered statistically significant.

In order to evaluate conduction defects in mdx mice we crossed mice carrying the mdx mutation with a Cx40-GFP allele to generate double transgenic mice mdx::Cx40-GFP (C57BL/10ScSn-Dmdmdx/J). Cx40-GFP mice express the GFP reporter gene under the control of the gap junction alpha 5 (Gja5) promoter (Miquerol et al., 2004). Gja5 encodes for Connexin 40 specifically expressed in atrial cardiomyocytes and in the ventricular conduction system (VCS). Firstly we evaluated the survival rate of these mice to study the age-related progression of DMD associated cardiac defects. We observed that 27% of mdx::Cx40-GFP mice died after 1 year while 100% of control animals survived (Figure 1a). A follow-up of the cardiac activity was made through six-lead surface electrocardiogram (ECG) recordings on mice anesthetized with isoflurane. The two first electrocardiograms were recorded at 1 month and 3 months of age, then every 3 months until 12 months of age (Figure 1b). The mdx::Cx40-GFP mice presented a statistically significant increased PR interval compared to controls at 12-month-old, indicative of first degree atrioventricular block (36.6 ± 1.4 ms for WT vs. 41.5 ± 4.6 ms for mdx). QRS-II duration in mdx::Cx40-GFP mice was significantly increased compared to the controls at 3-month (16.1 ± 1.2 ms for WT vs. 17.1 ± 0.9 ms for mdx) and at 12-month (16.3 ± 0.8 ms for WT vs. 17.9 ± 1.8 ms for mdx), indicating a slower ventricular activation in the mdx::Cx40-GFP mice. The other ECG parameters did not present any differences between mdx and WT mice (Table 1). As expected, mdx::Cx40-GFP mice presented a dystrophic myocardial phenotype. Normalization of the heart size by the body weight of each mouse revealed a greater heart length (0.17 ± 0.01 for WT vs. 0.20 ± 0.02 for mdx) and a greater heart width (0.15 ± 0.01 for WT vs. 0.18 ± 0.02 for mdx) in mdx::Cx40-GFP mice (Figure 1c).

During ECG follow-up, we scored spontaneous ventricular arrhythmic events and found that mdx::Cx40-GFP but not WT mice presented premature ventricular complexes (PVC) under basal conditions (Figures 2a,b). To mimic physiological stress, the mice were subjected to β-adrenergic stimulation by injection of Isoproterenol (2 μg/g) at the age of 12 months. The increase in heart rate after Isoproterenol injection was not significantly different between WT and mdx::Cx40-GFP mice (28.8% ± 9.0% of increase for WT vs. 26.66% ± 10.3% of increase for mdx). β-adrenergic stimulation increased the number of mice with PVC in both groups, with a greater prevalence in mdx::Cx40-GFP mice (33% for WT vs. 65% for mdx) (Figure 2b). One mdx::Cx40-GFP mouse displayed ventricular tachycardia (VT) (Figure 2a). In summary, the progressive onset of cardiomyopathy in mdx::Cx40-GFP mice is associated with a larger heart, ventricular conduction defects and arrhythmias at 12 months of age.

Since Purkinje fiber degeneration has been observed in human and canine DMD (Urasawa et al., 2008; Echigoya et al., 2017; Nomura and Hizawa, 1982), we used the Cx40-GFP mouse line to study the morphology and histology of Purkinje fibers in mdx::Cx40-GFP mice. In these mice GFP is expressed in the entire VCS including the His bundle, bundle branches and the Purkinje fiber network as seen in a luminal view of the endocardial surface of an opened WT left ventricle (Figure 3a). The PF network in mdx::Cx40-GFP mice is similar to WT mice in terms of number, distribution and structure of ellipsoids. Thus, there was no difference in the morphology of the Purkinje fiber network between control and mdx::Cx40-GFP mice (Figure 3a). As recent data have suggested that Cx43 lateralization contributes to DMD arrhythmogenesis in mdx mice (Gonzalez et al., 2015), we performed whole-mount immunostaining with a Cx43 antibody to examine the distribution of Cx43 gap junctions in working cardiomyocytes and in Purkinje fibers. Under high magnification, Cx43 gap junctions are present mainly at the intercalated discs (ID) of working cardiomyocytes, whereas they are distributed all along the plasma membrane of Purkinje fibers in WT (Figure 3a). We found a similar distribution of Cx43 gap junction localisation in mdx::Cx40-GFP hearts in both PF or working myocardium (Figure 3a). These data show that the Purkinje fiber network structure and the overall distribution of Cx43 are not affected in mdx::Cx40-GFP mice at 12 months of age when the arrhythmia-associated cardiomyopathy is in place.

To further investigate this, we performed histological analysis of Purkinje fibers on cryosections of WT and mdx hearts. Purkinje fibers were identified using a contactin-2 (Cntn2) antibody and cardiomyocytes were identified using DMD and WGA (Wheat germ agglutinin) staining. In WT hearts, DMD is present on the membrane of all cardiomyocytes with higher levels found in Purkinje fibers (Figure 3b). As expected, no staining with DMD antibody was detected in mdx hearts (Figure 3b). WGA staining showed that the membrane and the size of cardiomyocytes are overall larger in mdx hearts compared to WT. Cntn2 is a well-known marker of mature Purkinje fibers and play an important role in the electrical propagation in the heart (Pallante et al., 2010). Cntn2 staining was similar in WT and mdx hearts (Figure 3b), demonstrating the preservation of mature Purkinje fibers in the murine DMD model.

Slowed ventricular conduction, as represented by QRS interval prolongation in the ECGs of mdx mice, may result from reduced I_Na in mdx compared to WT in ventricular cardiomyocytes and Purkinje fibers. Here, we addressed whether the age of mice (range: neonatal – 1-year-old) has an impact on the severity of I_Na loss in the dystrophic mdx mouse heart. In Koenig et al. (2011), we reported that ventricular cardiomyocytes isolated from neonatal mdx mice show less severe I_Na loss than ventricular myocytes from adult (4-6-month-old) mdx mice. This suggested that, although already present at neonatal age, I_Na loss in dystrophic ventricular cardiomyocytes becomes more pronounced at adulthood. This result was in line with a normal QRS interval in neonatal, but prolonged QRS interval in adult mdx compared to WT mice (Koenig et al., 2011).

Here, in Figures 4a–d, the I_Na properties of Purkinje fibers derived from 11–13-week-old and 19–21-week-old WT and mdx mice were compared. We observed that the current densities of Purkinje fibers in WT mice were independent of the age of the animal. At 11–13 weeks, current densities in mdx Purkinje fibers were only slightly reduced compared to those in WT Purkinje fibers (Figure 4c). In contrast, at 19–21 weeks, mdx Purkinje fibers showed a significantly reduced current density (Figure 4d). This suggests an increase in I_Na loss in Purkinje fibers with age in dystrophic mdx mice.

Finally, we tested whether the loss of I_Na was even more severe in old mdx mice, an age known to be associated with the onset of arrhythmogenic cardiomyopathy (Quinlan et al., 2004). In Figure 4e the I_Na densities of ventricular cardiomyocytes derived from 16–17 and 50–51-week-old mdx mice were similar.

Taken together, our I_Na recordings suggest that current loss in dystrophic ventricular cardiomyocytes and Purkinje fibers is already present in mdx mice at a very young age, worsens from juvenile to full adulthood, and finally persists with similar severity until 1 year of age, when arrhythmogenic cardiomyopathy is present.

To detect any other cardiac activation defects, we calculated the angle of the main electrical axis of the heart from ECG recordings for each mouse. At 1 month of age, all WT and mdx mice presented a normal axis comprised between 0° and 120° (Figure 5a). At 6 months of age, three out of the 26 mdx mice presented a left deviation of the electrical axis whereas none of the WT mice had a deviated axis (Figure 5a). At 12-month-old, 4 out of the 26 mdx mice presented a left deviation of their electrical axis (Figure 5a). Moreover, the angle of the main activation axis was highly variable (Figure 5a) though remaining in a normal range (0°–120°), showing that the stereotypical depolarization pattern is slightly affected in mdx hearts. To better understand the origin of the axis deviation, we analyzed the presence of fibrosis in these hearts using WGA staining which can serve as a readout of fibrosis (Emde et al., 2014). At 12 months of age, 5/7 of the mdx hearts showed ventricular fibrosis (Figures 5b,c). However, there was no significant correlation between fibrosis density and cardiac axis deviation (Figure 5d; r = 0.58; p = 0.06). Thus, the deviation of the main electrical axis may not result only from the presence of fibrosis in mdx hearts.