Connexin-43 reduction prevents muscle defects in a mouse model of manifesting Duchenne muscular dystrophy female carriers

Duchenne muscular dystrophy (DMD) is a severe X-linked neuromuscular disorder that affects males. However, 8% of female carriers are symptomatic and underrepresented in research due to the lack of animal models. We generated a symptomatic mouse model of DMD carriers via injection of mdx (murine DMD) embryonic stem cells (ESCs) into wild-type (WT) blastocysts (mdx/WT chimera). mdx/WT chimeras developed cardiomyopathic features and dystrophic skeletal muscle phenotypes including elevated mononuclear invasion, central nucleation, fibrosis and declined forelimb grip strength. The disease was accompanied by connexin-43 (Cx43) aberrantly enhanced in both cardiac and skeletal muscles and remodeled in the heart. Genetic reduction of Cx43-copy number in mdx/WT-Cx43(+/−) chimeras protected them from both cardiac and skeletal muscle fiber damage. In dystrophic skeletal muscle, Cx43 expression was not seen in the fibers but in adjacent F4/80+ mononuclear cells. Ethidium Bromide uptake in purified F4/80+/CD11b+ mdx macrophages revealed functional activity of Cx43, which was inhibited by administration of Gap19 peptide mimetic, a Cx43 hemichannel-specific inhibitor. Thus, we suggest that Cx43 reduction in symptomatic DMD carrier mice leads to prevention of Cx43 remodeling in the heart and prevention of aberrant Cx43 hemichannel activity in the skeletal muscle macrophages neighboring Cx43 non-expressing fibers.


Results
Genetic reduction of Cx43 rescues the mdx/WT-Cx43(+/−) chimeric heart. We generated mdx/ WT and mdx/WT-Cx43(+/−) chimeras by injecting mdx (dystrophin−) murine embryonic stem cells (ESCs) into WT:Cx43(+/+) and WT:Cx43(+/−) (dystrophin+) blastocysts 14 . Seven day-old pups were genotyped for the presence of a heterozygous mutation in the Cx43 gene using DNA from tail tips. Half of the littermates contained the mutation and half did not (data not shown). Mdx ESCs contain a reporter DsRed transgene inserted into their genome whose ratio of band intensities compared to internal control, supplied by Jax labs, was used to determine the degree of ESC-derived cells (degree of chimerism) 14 . The presence of the DsRed transgene was assessed in pups tail tips (data not shown) 14 and then confirmed at time of sacrifice using heart (Fig. 1A), diaphragm ( Fig. 2A), and pectoralis ( Supplementary Fig. S3). Tissue from mice transgenic for DsRed and mdx mice were used as a positive (100%) control and a negative (0%) control, respectively. Mouse chimeras in the range of 10-30% (DsRed) contained 10-30% of ESC-derived cells and 90-70% of blastocyst-derived cells respectively. Results obtained from DsRed inversely correlated with those obtained by western blot analysis of the dystrophin protein in heart, diaphragm, and pectoralis extracts (Figs. 1B and 2B, Supplementary Fig. S3). Quantification of dystrophin in mdx/WT and mdx/WT-Cx43(+/−) chimeras compared to WT (positive control) and mdx (negative control) showed a range of dystrophin expression ( Supplementary Fig. S1). The presence of dystrophin+ and dystrophin− patches in the heart were observed in both mdx/WT and mdx/WT-Cx43(+/−) chimeras by immunofluorescence detection in cardiac cryosections (Fig. 1C). Variations in dystrophin levels became evident when comparative analysis was conducted across individual chimeras ( Supplementary Fig. S1). For example, chimera 1 showed similar levels of dystrophin expression in heart, diaphragm, and pectoralis, but chimera 4 showed higher levels of dystrophin expression in heart than those in diaphragm or pectoralis.
Genetic reduction of Cx43 rescues the mdx/WT-Cx43(+/−) chimeric skeletal muscle. We also studied the dystrophic skeletal muscle in chimeric mice to test a potential rescue role of genetic Cx43 reduction. We focused primarily on the diaphragm as it is the most affected skeletal muscle tissue in DMD patients and mdx mice 35,36 and confirmed our observations, with similar results, in the pectoralis muscle (Supplementary Figs. S2 and S3). Histopathological analysis revealed augmented fibrosis (MT), central nucleation and mononuclear invasion (H&E) in the diaphragm muscle of mdx and mdx:Cx43(+/−) mice, relative to that of WT mice (Fig. 2D-F). This suggests that genetic reduction of Cx43 does not benefit the mdx skeletal muscle. However, fibrosis, mononuclear infiltration, and central nucleation were reduced in the diaphragm muscle of mdx/WT-Cx43(+/−) chimeras relative to that of mdx/WT chimeras ( Fig. 2D-F). This suggests that genetic reduction of Cx43 prevents pathology in the skeletal muscle of the mdx/WT-Cx43(+/−) chimeras. Accordingly, the skeletal muscle fiber cross-sectional area (CSA) distribution profile was assessed. While most of the fibers in WT diaphragm were in the range of 1,200-1,600 µm 2 , fibers in mdx diaphragm were in a broader range, spanning from 400 µm 2 to 2,000 µm 2 at the expense of a reduced number of fibers in the 1,200-1,600 µm 2 range. mdx/WT-Cx43(+/−) chimeric diaphragm muscle was alike to that of WT, in which most fibers were in the range of 1200-1600 µm 2 (Fig. 2G). mdx/WT had less medium caliber (1,200-1,600 µm 2 ), and more high caliber (2000+ µm 2 ) fibers, shifting towards the mdx phenotype (Fig. 2G). Lastly, grip strength was improved in mdx/WT-Cx43(+/−) chimeras but not in mdx/WT chimeras or mdx:Cx43(+/−) mice relative to WT mice (Fig. 2H). Taken all together, the results suggest that genetic reduction of Cx43 corrects muscular dystrophy in DMD carrier mice, but not in DMD mice.
Cx43 is enhanced and functional in mdx but not in WT skeletal muscle macrophages. Our results suggest that Cx43 is enhanced in the skeletal muscle of mdx mice and mdx/WT chimeras relative to WT mice ( Fig. 2C and Supplementary Fig. S1). Our results also suggest that Cx43 reduction prevents the development of skeletal muscle dystrophy in mdx/WT-Cx43(+/−) chimeras, but not in mdx:Cx43(+/−) mice ( Fig. 2D-H). This finding in the skeletal muscle contrasts with that in the heart, where Cx43 reduction leads to corrections in both mdx/WT-Cx43(+/−) chimeras and mdx:Cx43(+/−) mice hearts ( Fig. 1F-I). To begin to reconcile this dissimilar role of Cx43, we examined the Cx43 pattern of expression in skeletal muscle by immunofluorescence. We did not detect Cx43 on any part of the mdx or WT mature skeletal muscle fibers (Fig. 3A,B). This suggests that there may be other cell types distinct from the fibers that express Cx43 in the skeletal muscle, particularly in dystrophic mice, because Cx43 was seemingly in areas occupied by mononuclear cells (Fig. 3A,B). Due to previous reports of Cx43 in macrophages 37,38 we performed immunofluorescent analysis for Cx43 and F4/80 macrophage marker. This revealed that Cx43 was detected only in mdx F4/80+ mononuclear cells in the interstitial space between skeletal muscle fibers, areas where mononuclear invasion occurs (Fig. 3B).
Because cardiac macrophages were also observed in the dystrophic heart, we performed F4/80 and Cx43 immunofluorescence in WT and mdx hearts to study a potential overlap. However, unlike the presence of overlap in the dystrophic diaphragm ( Fig. 3B), there was no co-staining in the heart (Supplementary Fig. S4). Furthermore, macrophage-enriched FACS-sorted F4/80+/Cd11b+ cells purified from mdx hearts revealed negligible detection of Cx43 protein expression. Gravity-settled mdx heart myocytes were used as a Cx43 positive control ( Supplementary Fig. S4). Altogether, the results open the possibility that detrimental activity of enhanced Cx43 hemichannels from resident skeletal muscle macrophages may be mitigated by Cx43 reduction in a dystrophic, chimeric environment.

Discussion
In this study, we show that Cx43-copy number reduction in mdx/WT-Cx43(+/−) chimeras result in a rescue of DMD pathology in both the heart and the skeletal muscle. These results are aligned with previous studies in the heart, in which normalized Cx43 levels benefit mdx cardiac pathology 16,24 ; but are unlike the observations in mdx skeletal muscle, in which normalized Cx43 levels do not benefit skeletal muscle pathology. We suggest aberrant Cx43 hemichannel activity in mdx macrophages promotes the dystrophic phenotype in the skeletal muscle. We also suggest the lack of rescue in mdx:Cx43(+/−) skeletal muscle compared to mdx/WT-Cx43(+/−) chimeras can be explained by the presence of some dystrophin in the symptomatic DMD carrier and the lack of dystrophin in mdx:Cx43(+/−) mice. Complete absence of dystrophin in mdx:Cx43(+/−) skeletal muscle may lead to exacerbation of a more oxidative environment that may ultimately impinge on the state of the macrophages. Overall, we propose symptomatic mdx carrier skeletal muscle, in addition to the heart, benefit from Cx43 normalization in the presence of some dystrophin+ fibers.
DMD patients exhibit chronic inflammation of the heart and skeletal muscle in the form of fibrosis, necrosis, and functional deficiencies. In mdx mice there is an upregulation of Cx43 in the affected muscles, possibly to compensate for increased injury, satellite cell exhaustion and increased myotube regeneration of the muscle [39][40][41][42][43] . Similar to the heart, Cx43 is required for myoblast differentiation in vitro [44][45][46] . In contrast with the heart where gap junction intracellular communication (GJIC) is essential, mature skeletal muscle cells fuse to form a syncytium, therefore they do not rely on GJ formation 44,45 . Although absent from control, basal tissues Cx43 is expressed post-muscle injury and can be seen as punctate staining 30,31 . Under similar circumstances including ischemia/ reperfusion injury, denervation, and infection/sepsis, this study supports that macrophages upregulate Cx43 31-33 .
Current treatments only provide short-term relief for DMD patients and carriers, although promising findings have been shown to attenuate the inflammatory phenotype. One of the most commonly prescribed treatments, Dexamethasone (DEX), is a glucocorticoid used to build strength and muscle. However, DEX has led to an increase in Cx43 hemichannels that can exacerbate skeletal muscle atrophy 49 . In sepsis, Cx43 normalization via deletion or pharmacological inhibition is associated with a reduction of secreted necrotic factors, including cytokines and ATP, thereby increasing survival 38 . Peptide mimetics, like Gap19, have been successful because they are capable of inhibiting Cx43 hemichannel dysregulation, but maintain gap junction function in other cell types 37 .
In mdx mice, Cx43 is expressed in between mature skeletal muscle fibers, where lymphocytes, neutrophils, and macrophages reside. However, lymphocytes and neutrophils do not express Cx43 38 . Macrophages also have the capability to fuse with adjacent fibers, which may contribute to DMD pathology 37,46,50,51 . Aberrant hemichannel activity has also been observed in mdx mouse myofibers to cause apoptosis 22,33,48,52 . Based on localization of Cx43 to the mononuclear cells, we postulate that reduction of hemichannel activity in skeletal muscle macrophages is linked to decreased cell death, suppressed activation of pro-inflammatory factors, and therefore, slowed muscle deterioration 17,34,53 .
In our mouse model of symptomatic DMD carriers, we have shown that Cx43 has divergent roles in skeletal and cardiac tissue. Cardiomyocytes are post-mitotic, therefore post-translational effects of Cx43 better address the tissue environment 28,29 . One major contributor to stress signals in the DMD heart is oxidative stress 24,54,55 . In contrast, skeletal muscle is a syncytium in which fibers fuse and macrophages remain outside of the fibers. Insight for understanding the role of Cx43 may be further addressed by testing the response and modulation of reactive oxygen species in mdx versus WT macrophages. Data presented here suggest treatments limiting Cx43 and Cx43 hemichannel activity may also target Cx43+ skeletal muscle macrophages, which have roles in tissue repair and injury resolution. Macrophages have recently been shown to be essential for uptake and slow release of antisense oligonucleotides via Cx43 54,56,57 . The role as a reservoir is necessary for therapeutic administration of partially-functional dystrophin to skeletal fibers and muscle stem cell regeneration for DMD patients 39,43 . Cx43 is not enriched in mdx cardiac macrophages compared to the adjacent cardiomyocytes, reserving a larger role for Cx43 in the cardiomyocytes and skeletal-specific macrophages. We cannot rule out, however, a potential role for Cx43 in the cardiac macrophages, for example, at later stages of the disease.
Further investigation is necessary to determine if dystrophin from healthy fibers of a chimeric environment modulates Cx43 expression and function in neighboring macrophages, which in turn, modulate death of the neighboring fibers. In addition, we would like to investigate how this study translates to younger symptomatic carrier mice as we conducted studies in 6-month-old mdx/WT and mdx/WT-Cx43(+/−) chimeras with fibrosis in the former and no fibrosis in the latter (Supplementary Fig. S4).
As a limitation of the study, we cannot rule out an early developmental and postnatal contribution of Cx43 expression, which may still be highly expressed in the skeletal muscle, as the mixing of the WT blastocysts and the mdx ESCs takes place at pre-implantation stages of development. Thus, an early, developmental/postnatal effect of Cx43 may be a contributor to the long-term phenotype.
These results importantly contribute to research towards improved cardiomyopathic understanding and development towards target-specific (Cx43) skeletal muscle anti-inflammatory treatments for symptomatic DMD carriers.
Afterward, muscle sections were washed in PBS-T and mounted using ProLong TM Diamond antifade mountant with DAPI (P36962, Thermo Fisher Scientific). At least 3-5 sections (40x or 100x)were examined. Images were taken using a Nikon Eclipse T1 (Melville, NY, http://www.nikonusa.com) for immunoflourescence images, and an Olympus BX51 microscope for histological samples and processed using NIS-Elements BR software in a blinded fashion. Cross-sectional area (CSA) was measured using at least 200 fibers across 4-5 sections (100x and 200x). Colocalization was analysed using an Olympus Fluoview 1000 Confocal Laser Scanning Microscope and Z-stacks 0.5 µm thick at 600x magnification using Fluorview software and then processed in Fiji for maximum intensity. Cell counting and CSA were determined in Fiji in a blinded fashion. H&E and Masson trichrome. Skeletal and heart muscles were fixed in 4% paraformaldehyde overnight and embedded in paraffin, following euthanasia in 10-14-month-old mice. Sections were cut at 6 μm from each tissue and stained for histopathology utilizing an H&E kit (Vector, Torrance, CA) or Masson Trichrome kit (Richard-Allen Scientific, Thermo Fisher Scientific, Waltham, MA). Images were taken using a Nikon Eclipse T1 microscope and processed using NIS-Elements BR software. Fibrosis was quantified, blinded, by subtracting blue fibrotic regions from the total muscle area in Fiji and repeated in 5 images. Grip strength. An assessment of muscular function was recorded using a grip strength meter in a blinded fashion (Columbus Instruments). The grip strength meter was positioned horizontally, and mice, held by their tails, were lowered toward the pull bar. Mice (10-12 months) were allowed to get a good grip on the triangular bar with only their forelimbs. Once they had a good grasp, the mouse was pulled backward parallel to the device, until the mouse let go. This was completed in 2 runs: 5 pulls each. Run #1 trained and conditioned the mice to the apparatus. The top 3 scores from run #2 were averaged for the final value. This process was repeated 3-times over the 10-12 month period for each mouse, and the average used in statistical analysis. The force that was applied to the bar at the time of release was recorded as Gram-force. www.nature.com/scientificreports www.nature.com/scientificreports/ Fluorescent activated cell sorting (FACS). Following euthanasia, diaphragms were dissected from 8-10-month-old WT and mdx mice, washed in 1% Pen/Strep in PBS, and minced. Digestion was completed using Liberase TL (Sigma) for 2 hours at 37 °C on an orbital shaker (20 rpm). Heart: the heart was dissected from 13-16-month-old mdx mice using a Lagendorff-free method 60 describe below. Cells were filtered through a 40 μm cell strainer and incubated with antibodies F4/80, Cd11b and DAPI (Thermo Fisher Scientific) for 20 minutes on the benchtop covered from light. Samples were washed and sorted by the Flow Cytometry/Cell Sorting Rutgers Core Facility.

Ethidium bromide uptake in vitro.
Cells post-FACS-sort, from 8-10-month-old WT and mdx mice, were concentrated, mounted, and allowed to attach for 20-minutes on a chamber for confocal imaging. Ethidium Bromide (EthBr) was added during the first few frames of imaging (20 μM). A scanning confocal microscope (Olympus, FluoView1000) with a 600x objective lens captured an image every 10 seconds for 20 minutes (120 total images).
Gap19 Cx43 hemichannel-specific blocker (Sigma peptide mimetic) was added to chamber cells at the time of mounting for 20-minute incubation. While blinded to the treatment group, Fiji software was used to measure the fluorescence of a circle drawn around each cell in each frame (F). Values were normalized to the 3 rd frame (F 0 ) after the addition of EthBr to account for variability induced by different rates of diffusion. Results are displayed as F/F 0 .
Laggendorff-free isolation method for cardiac myocytes and nonmyocytes. The heart was dissected from 13-16-month-old mdx mice (N = 3). Single ventricular cardiomyocytes and mononuclear cells were isolated with Collagenase Type II (Worthington LS004176) and Protease XIV (Sigma P5147) by Laggendorf-free digestion of heart ventricle 60 . Myocyte and non-myocyte populations were separated by 20 minute of gravity filtration. Langendorff-free supernatant cell suspension was filtered through a 40 μm cell strainer and underwent the FACS preparation and sorting. The settled pellet of cardiomyocytes underwent two additional 15 minute gravity settlings in 4 mL of Perfusion buffer in a 15 mL tube 60 . The final pellet of myocytes was gently homogenized to obtain a protein sample used in western blot ( Supplementary Fig. S4B,C).
Statistics. Data was analyzed for statistical significance using parametric analysis in GraphPad Prism 8.